Package Details: linux-ck 4.14.5-1

Git Clone URL: https://aur.archlinux.org/linux-ck.git (read-only)
Package Base: linux-ck
Description: The Linux-ck kernel and modules with the ck1 patchset featuring MuQSS CPU scheduler v0.162
Upstream URL: https://wiki.archlinux.org/index.php/Linux-ck
Licenses: GPL2
Provides: linux-ck=4.14.5
Submitter: graysky
Maintainer: graysky
Last Packager: graysky
Votes: 371
Popularity: 4.731636
First Submitted: 2011-07-22 14:51
Last Updated: 2017-12-10 13:36

Sources (15)

Latest Comments

pedrogabriel commented on 2017-11-26 22:29

You plan to include the new runqueue patch?

johnnybegood commented on 2017-11-25 17:44

Thanks a lot for the ck patches for bfq-sq!!! :)

Osleg commented on 2017-11-25 16:21

Kernel panic solved, was unrelated to kernel itself

Osleg commented on 2017-11-25 13:44

Also this since upgrade to >4.14
Kernel panic with "No working init found"
http://storage9.static.itmages.com/i/17/1125/h_1511617326_7762244_e6000b236e.jpg
Any idea what could be the cause?

Osleg commented on 2017-11-25 13:38

Yeah It downloads the sources twice and it builds it twice.
Just finished compilation using plain old and reliable makepkg, worked like a charm w/o duplication. :)
Also, @graysky, was looking for you at IRC but since can't find you I'll ask you here.
I have a server that standing idling for no reasons for quite long time and it will keep standing there. Would you like to host the repo-ck.org/$arch there?
Cuz afaik lots of people have problems downloading the binary -ck images from your servers. But it's quite anoying to re-build it every time on my machine :)
I can give you access to the machine if you'd like?

graysky commented on 2017-11-25 13:05

No idea how yaourt works but if it is running the build function twice, the author of it should be notified as that is wasteful. Are you sure that it isn't just building once, and packaging the kernel then headers? That is how makepkg does it.

Osleg commented on 2017-11-25 12:36

@sir_lucjan
Thanks for responding, already found my question in documentation, this PKGBUILD is multi package for both headers and the kernel itself.
The question tho is that possible to build both headers and kernel w/o re-downloading and re-building the entire kernel?

Right now I'm installing using yaourt. So it download sources twice and building twice. How could I prevent that?

sir_lucjan commented on 2017-11-23 19:40

@Osleg

Could you tell us what do you mean?

Osleg commented on 2017-11-23 19:37

I wonder what is the difference between this pkg and linux-ck as their PKGBUILDs are identical

Terence commented on 2017-11-23 15:05

@artafinde @graysky After some more investigations, I figured out those were just warnings with the purpose to warn you they will not be included by default in the kernel, and you can safely ignore them as dkms take care of building them.
Thank you very much for your help, I hope this could eventually help someone confused like me.

QuartzDragon commented on 2017-11-23 10:03

Attention @graysky and @everyone:

Urgently needed update!

FS#56404 - [linux] Using bcache will destroy filesystems (4.14.X) ~ https://bugs.archlinux.org/task/56404

and:

4.14.1-2 patch ~ https://git.archlinux.org/svntogit/packages.git/commit/trunk?h=packages/linux&id=a45f53296994bfbc2581656fa91f81f74fa21b83

I've already updated manually for this.

Terence commented on 2017-11-22 23:11

@artafinde: Thanks for your complete answer.
I rigorously followed your advices, booting back to the ARCH kernel and making sure once more dkms modules are blacklisted and not stored, but whatever I do, I still get the same error with localmodconfig when either compiling on ARCH or -ck. I don't understand what's wrong...
Note that I also use dkms modules on the regular kernel.

artafinde commented on 2017-11-22 18:05

@Terence: modprobed-db is snapshotting your loaded modules and adds them into a database. Provided you run the system long enough and use it you should have all the modules you are using.
localmodconfig is a tool to configure the kernel to use the modules which are currently used from the kernel
dkms is using the kernel-headers to create new modules for nvidia/virtualbox etc.

During linux-ck compilation and if you have set the localmodconfig parameter in PKGBUILD we load all modules which are referenced in modprobed-db prior to calling localmodconfig. Then kernel is configured according to the loaded modules and compiles etc.

If you want to make sure you have a clean modprobed-db follow these steps:
1. clean up modprobed-db from modules which don't belong on kernel (nvidia, vbox etc)
2. add those to IGNORE list (optionally test by calling modprobed-db store nd check they are not added)
3. reboot and load -ARCH kernel (to make sure all the modules are available)
4. edit PKGBUILD and set localmodconfig - compile and install
5. dkms should build new modules (nvidia, vbox, exfat etc)
6. reboot to your new kernel

The above process is what I follow and works fine with nvidia, vbox and exfat -dkms packages. As a fallback I have -ARCH kernel always installed.

Terence commented on 2017-11-22 17:42

@artafinde sorry I didn't make it clear enough but that's what I did :)
To me it's like modprobe-db just loads all the modules it detected has been loaded at some point and then localmodconfig just lsmod, which includes dkms modules...
There has to be a way to exclude them.

artafinde commented on 2017-11-22 17:38

Open the modprobed-db database (it's a file) and remove them from the list.

Terence commented on 2017-11-22 17:30

@artafinde @graysky Thanks but I knew that, that's why they are built using dkms and I took care of blacklisting all of them, making sure they are not appearing in the modprobe-db file, but problem persists.

artafinde commented on 2017-11-22 17:25

@Terence: The nvidia is external same as virtualbox modules. If you have dkms packages it should trigger a build after you install (provided you have headers). modprobed-db should probably exclude them with IGNORE list - see wiki.

graysky commented on 2017-11-22 17:19

@Terence - These are either provided by some other package or not included in the kernel anymore. See man modprobed-db.

Terence commented on 2017-11-22 17:10

I've never been able to find anywhere a solution to this problem with localmodconfig, modprobed-db and dkms modules:

==> If you have modprobed-db installed, running it in recall mode now
Modprobed-db v2.37

164 modules currently loaded per /proc/modules
168 modules are in /home/terence/.config/modprobed.db

Attempting to modprobe the contents of /home/terence/.config/modprobed.db

164 modules are now loaded per /proc/modules

==> Running Steven Rostedt's make localmodconfig now
using config: '.config'
vboxpci config not found!!
nvidia config not found!!
nvidia_uvm config not found!!
vboxnetflt config not found!!
vboxnetadp config not found!!
vboxdrv config not found!!
8812au config not found!!
nvidia_drm config not found!!
nvidia_modeset config not found!!

Any insight?

sir_lucjan commented on 2017-11-21 21:04

@johnnybegood

I also recommended blk-mq patches.

johnnybegood commented on 2017-11-21 20:55

Works perfect for me as well. However, I still miss the old single queue bfq, the now called "bfq-sq" that was never merged into the official kernel, so I compiled ck with sirlucjan latest patch for 4.14 (https://raw.githubusercontent.com/sirlucjan/kernel-patches/master/4.14/4.14-bfq-sq-mq-git-20171114.patch)

I find the combination of -ck and bfq-sq unbeatable. Maybe it's just me but it could be an addition to this excellent kernel.

Graysky, thanks for your work and CK's!

zebulon commented on 2017-11-21 12:07

Works for me. Need to use scsi_mod.use_blk_mq=1 init option for blq as expected.
@silvik: I do not need to use elevator=bfq, scsi_mod.use_blk_mq=1 is sufficient. I need however to set udev rules to set bfq.

graysky commented on 2017-11-20 21:37

CK pushed MuQSS v0.162 into ck1 for 4.14 so it's official. I merged with ARCH changes and here we go. Please test/provide feedback. I will hold off on pushing to [repo-ck] for at least 24 h to get some feedback.

silvik commented on 2017-11-20 20:12

BFQ works for me with scsi_mod.use_blk_mq=1 elevator=bfq kernel options.

francoism90 commented on 2017-11-04 21:28

@kwe: Do you have a source?

We should try to keep the Wiki up-to-date, contributes are welcome. :)

Terence commented on 2017-11-04 20:22

@kwe Thanks and sorry for the off-topic.

kwe commented on 2017-11-04 20:20

@Terence
1. this is not really related to this AUR package anymore
2. don't take the wiki for the golden words from god
3. what the note means is that all sq schedulers are unavailable and (probably from pre-4.12 Kernels) that there were no mq-schedulers. 4.12 added Kyber and BFQ

Terence commented on 2017-11-04 20:14

@kwe Ok my bad, I forgot it's not deadline anymore but mk-deadline.
Why is the wiki saying that I can't do that though?
https://wiki.archlinux.org/index.php/improving_performance#Tuning_IO_schedulers

kwe commented on 2017-11-04 20:10

@Terence Deadline should be an option:
$ cat /sys/block/sda/queue/scheduler
mq-deadline kyber [bfq] none

Terence commented on 2017-11-04 20:09

@kwe Ok, I'm indeed using grub and forgot about that, thanks.
One more question: I have my root partition on my SSD and my home/var partition on my HDD. I was before using deadline on the SSD and BFQ on the HDD. It seems I can't do that anymore now and it's confirmed by the arch wiki. Will it's still perform ok with the SSD compared to deadline or is there another alternative for my system configuration?

kwe commented on 2017-11-04 20:02

@Terence: From what I can tell, the "elevator" parameter is being slowly phased out and might only set the sq scheduler. Using elevator=[mq-scheduler] seems to crash. There is a patch for the Linux kernel that fixes this crash by introducing a relevant check, but it won't be introduced until 4.15 IIRC.
You don't need to boot a live system if you're using GRUB. You can edit the start parameters via the E key and then fix your GRUB config from within your normal system.

Terence commented on 2017-11-04 19:58

@kwe Thanks for your relevant comment, I tried to use the 'elevator' kernel parameter but it crashed the kernel at boot and I needed to boot to the live usb to be able to fix it.
I then have 2 questions:
1) Why are there still patches for BFQ included?
2) Can the package be updated with "CONFIG_SCSI_MQ_DEFAULT = y"?

kwe commented on 2017-11-04 18:59

@everybody who is not involved, @Terence
I've had some troubles figuring out what's up, so here is my tl;dr of the issue:
We used to have BFQ as a part of the ck kernel patches. With a recent kernel version, you have BFQ merged into the vanilla kernel.org kernel tree. Hence, BFQ was removed from the ck patchset.
However, the BFQ in the kernel is for blk-mq, whereas previously we had bfq for non-mq / sq.
Because this kernel defaults to sq (CONFIG_SCSI_MQ_DEFAULT = n), the mq BFQ from the Linux kernel is not available. The solution seems to be to either build with that option enabled or alternatively boot the kernel with scsi_mod.use_blk_mq=1

Terence commented on 2017-11-04 18:52

I can't get BFQ to be enabled with the latest version.

keepitsimpleengr commented on 2017-10-31 20:37

@graysky got it

graysky commented on 2017-10-31 19:29

@keepitsimp - It's due to the fact that nvidia-utils-387.22-1 is out but I know and thus didn't update nvidia-ck. It has been done now. In the future please flag the nvidia-ck package out-of-date.

mrkline commented on 2017-10-31 18:04

Apparently I'm blind and/or need more coffee. My apologies.

sir_lucjan commented on 2017-10-31 17:49

@mrkline

# Patch source with ck patchset
patch -Np1 -i "../${_ckpatchname}"
patch -Np1 -R -i "../0015-Enable-and-make-BFQ-default-IO-scheduler-overriding-.patch"
patch -Np1 -i "../unfuck_bfq_v2.patch"

Do you see "patch -R" or not?

mrkline commented on 2017-10-31 17:36

@sir_lucjan - What does `patch -R` have to do with this? I'm posting because 4.13.10-1 allowed me to select BFQ as a scheduler, while 4.13.10-2 and 4.13.10-3 do not seem to.

@artafinde - I've been following along for the most part. Previously BFQ wasn't the default, but I was able to switch to it on boot using a udev rule. I'm confused why this changed.

keepitsimpleengr commented on 2017-10-31 16:51

:: nvidia-ck-k10: installing nvidia-utils (387.22-1) breaks dependency 'nvidia-utils=387.12'
:: nvidia-ck-sandybridge: installing nvidia-utils (387.22-1) breaks dependency 'nvidia-utils=387.12'

artafinde commented on 2017-10-31 11:04

@mrkline: This has been in discussion in forums official post: https://bbs.archlinux.org/viewtopic.php?id=111715&p=124 (read last two pages at least)

sir_lucjan commented on 2017-10-31 10:59

@mrkline:

You should read 'patch --help.'

-R --reverse Assume patches were created with old and new files swapped.

mrkline commented on 2017-10-31 00:45

@graysky: The latest two builds (4.13.10-2 and 4.13.10-3) don't seem to be building in BFQ. I see that you added CK's Enable-and-make-BFQ-default-IO-scheduler-overriding-.patch, but after building and installing, I see:

$ cat /sys/block/sda/queue/scheduler
noop deadline [cfq]

Perhaps the patches are being applied in the wrong order?

timofonic commented on 2017-10-30 19:09

Hello. I would love to build ck patches, but need to use drm-tip. I might get one interesting patch you have and maybe scrap for some more, but I doubt patches could apply to this too new version. If it does, I'll happily do a variant of it.

Anyway, could you help me to improve my PKGBUILD? I'll do an -anbox variant with the binder and the other subsystem enabled (I don't remember the name right now). I got busy IRL and a little lost about what to improve or continue changing.

https://paste.pound-python.org/show/ScHsBnIJVxdZP6D2MGYY/


Thanks in advance.

j1simon commented on 2017-10-24 06:59

I have a problem with 4.13.x. I have opened a bug because this occurs with all kernels 4.13.x that I've tested: stock, zen, ck.

https://bugs.archlinux.org/task/56094

artafinde commented on 2017-10-23 09:09

@kyak: as mentioned [here], build yourself from aur
[here]: https://bbs.archlinux.org/viewtopic.php?pid=1744646#p1744646

kyak commented on 2017-10-23 08:44

Hi graysky,

The virtualbox-ck-host-modules-ivybridge doesn't seem to exist anymore in repo-ck. Is it as intended, and if so, what is the suggested replacement?

graysky commented on 2017-10-22 17:30

All - again, the old behavior causes boot freezes thus requiring: 1) users to tweak with udev or the like, 2) something I do with the help of loqs, 3) the best option is for CK to do it and release ck2 with the fix.

I know 4.13.x has been delayed and I am no happier about that than you guys are. I emailed ck earlier today asking about his decision re: this fix in ck2 (recall the discussion on his blog last week), but haven't heard back yet. As well, if you're following the repo-ck thread on the bbs, I have asked loqs for his best solution as I neither have the hardware nor to the time to dig into this further right now. Stay tuned and thanks for your time to test and provide feedback.

nTia89 commented on 2017-10-22 17:19

@graysky, since changelog says:
-CONFIG_DEFAULT_CFQ=y
+CONFIG_DEFAULT_BFQ=y
-CONFIG_DEFAULT_IOSCHED="cfq"
+CONFIG_DEFAULT_IOSCHED="bfq"

I expect "bfq" enabled without ANY trick (both kernel param., udev rule or any other...).
Correct me if I am wrong.

Tjuh commented on 2017-10-22 17:17

Nice, thanks for that, most appreciated.

FarStar commented on 2017-10-22 17:09

Hi,
Personally, I use an udev rule:
/etc/udev/rules.d/60-scheduler.rules
ACTION=="add|change", KERNEL=="sd[a-z]", ATTR{queue/scheduler}="bfq"

And I get:
$ cat /sys/block/sda/queue/scheduler
mq-deadline kyber [bfq] none

Tjuh commented on 2017-10-22 17:05

A bit confused, so how does one enable BFQ nowadays? Wiki says setting elevator=bfq as a kernel parameter does not work.

graysky commented on 2017-10-22 16:28

@nTia89 and Tjuh - I believe that is the expected output, no?

Tjuh commented on 2017-10-22 15:29

Same as nTia89.

nTia89 commented on 2017-10-22 15:25

I get this, no kernel parameters or anything else (e.g udev rules):
$ cat /sys/block/sda/queue/scheduler
[mq-deadline] kyber bfq none

graysky commented on 2017-10-22 13:46

Please try 4.13.9-1 which has a workaround incorporated. No special kernel parameters should be needed. Works for me without them. Feedback please!

QuartzDragon commented on 2017-10-20 08:56

Hey @graysky and everyone,

The reason for all of the BFQ woes came from SCSI_MQ_DEFAULT defaulting to disabled upstream.

I have BFQ automatically set via udev rules, and enabling SCSI_MQ_DEFAULT allowed BFQ to work as usual.

This was a problem previously, and I'm surprised no-one seems to have remembered this... unless it was presumed that SCSI_MQ_DEFAULT was still enabled by default.

sir_lucjan commented on 2017-10-17 13:20

@zoidberg

Could you read previous comments?

zoidberg commented on 2017-10-17 13:17

As of version 4.13.7-2, I am not able to select bfq as a scheduler. It is not available as a scheduler option. Is this intentional?

$ cat /sys/block/sda/queue/scheduler
noop deadline [cfq]

Edit: Sorry, did not read previous comments re. the problem with bfq.

artafinde commented on 2017-10-16 21:14

Works fine, good job both. (also nvidia-dkms works fine I blame my debug stripped package)

sir_lucjan commented on 2017-10-16 21:02

No problem! I'm always at your service, John :D

graysky commented on 2017-10-16 21:00

Thanks to sir_lucjan's suggestion, 4.13.7-2 boots and works as expected. Please build and test for yourself.

artafinde commented on 2017-10-16 20:57

Hey lazy people, share the patch :)

sir_lucjan commented on 2017-10-16 20:56

@graysky - I'm lazier than you.... :D

graysky commented on 2017-10-16 20:51

@sir_lucjan - Yes, but I'm too lazy to use that lame captha on the blog :D

sir_lucjan commented on 2017-10-16 20:47

@graysky: As I wrote - of course you can quote me.

graysky commented on 2017-10-16 20:39

@sir_lucjan - Building what you proposed in the email now and will push 4.13.7-2 if it works. Would you mind pasting your email to ck's blog? Let him also consider it.

artafinde commented on 2017-10-16 20:21

@graysky: confirmed bfq to blame. (side note: nvidia-dkms fails also).

sir_lucjan commented on 2017-10-16 20:17

@graysky

Please check your mail.

graysky commented on 2017-10-16 20:03

It seems as though bfq is to blame. For me, booting elevator=cfq in the boot loader works as expected. Can others confirm?

artafinde commented on 2017-10-16 18:33

How does the debug symbols impact performance? It would be handy to have backtraces information when this happens but it means we need to compile with debug symbols which probably defeats the purpose of optimizations (MNATIVE etc). It would be handy for CK though.. My kernel panic on i7-6700 here: https://ptpb.pw/kmqk

nTia89 commented on 2017-10-16 17:52

Same here (not too old hardware, i5-5200U).
Compiled with options: MNATIVE and X86_P6_NOP
I attach a photo, maybe it can be useful: https://imgur.com/a/xxPdn
I have to say that respond to command: CTRL+ALT+CANC reboot the pc

Saren commented on 2017-10-16 15:56

Can confirm it's not just you. I got kernel panic on boot

graysky commented on 2017-10-16 13:40

I just pushed 4.13.7-1 but I am unable to boot after compiling on an admittedly old system. Requesting feedback from other users.

zerophase commented on 2017-10-15 04:35

@ozmartian in /usr/lib/modules-load.d just comment out crypto_user in bluez.conf.

I wasn't having issues with booting; the module just wouldn't load for me.

ozmartian commented on 2017-10-14 22:14

@zerophase could you provide more detail re crypto_user and what you edited and for which bluez package, pretty please? my linux-ck partitions are stalling on boot and crypto_user was the only error I could find in journal/logs so would save me having to figure it out myself until 4.13 is out... am temporarily using standard arch 4.13 kernel as a result and it feels like i am cheating on CK :p

thanks!

zebulon commented on 2017-10-11 21:07

@zerophase: thanks, always interesting to hear experience from others. Hopefully the wait will not be too long.

zerophase commented on 2017-10-09 18:57

For supporting journaled quotas is another switch needed to be turned on other than QFMT_V2? Switched it on, but still cannot mount my disk with usrjquota and grpjquota turned on.

zerophase commented on 2017-10-06 18:56

@zebulon I just needed to edit one of the bluez packages to not load crypto_user, since that doesn't exist until 4.13.

I understand ck has a busy life, and in the past has spent too much time hacking the kernel. I'm just trying to see what's up.

graysky commented on 2017-10-06 18:53

I emailed CK asking about the timelines for MuQSS for 4.13.x and he said 1-2 weeks. I also asked him to post a blog entry since others probably want to know too.

zebulon commented on 2017-10-06 11:29

@zerophase: Con Klivas' blog and repository mention the latest patch for 4.12. We need to wait. What kind of issues do you have?

zerophase commented on 2017-10-06 09:37

Has anyone heard from ck on 4.13 being released? Being on 4.12 still is starting to cause bugs with other arch packages.

keepitsimpleengr commented on 2017-10-04 14:46

:: nvidia-ck-k10: installing nvidia-utils (387.12-1) breaks dependency 'nvidia-utils=384.90'

keepitsimpleengr commented on 2017-09-28 15:12

nvidia-ck-sandybridge: installing nvidia-utils (384.90-1) breaks dependency 'nvidia-utils=384.69'

willianholtz commented on 2017-09-23 22:17


Hello, I have the same SD card problem as the following link.

https://bbs.archlinux.org/viewtopic.php?pid=1734477#p1734477

If the SD is plugged in when starting the kernel, it simply hangs, if it is not plugged in, when entering the DE does not mount!

graysky commented on 2017-09-23 17:49

That is to be expected with nvme.

zerophase commented on 2017-09-23 17:20

Should bfq be the default scheduler without setting anything, now? Just noticed my nvme drive is running with none by default.

Eschwartz commented on 2017-09-15 13:11

Apparently around 2012 linux-ck-headers started providing linux-headers as a truly ugly workaround for some community package that had a hard dependency on linux-headers. AFAICT no package still requires linux-headers when they work for -ck as well, since we have generally moved to dkms. Can you please remove this provides?

graysky commented on 2017-09-06 19:44

@archzz - please do not flag out of date until ck releases a patchset suitable for the 4.13.0 series.

graysky commented on 2017-09-05 16:21

Due to x.69 being pushed to [extra]. Updated now.

keepitsimpleengr commented on 2017-09-05 16:09

nvidia-ck-k10: installing nvidia-utils (384.69-1) breaks dependency 'nvidia-utils=384.59'

hotty commented on 2017-08-26 16:05

Since it is weekend and I finally had some time I found the culprit in the form of my internal sd-card reader. If I have a sd-card inserted udev times out resulting in a blinking cursor. Without a card linux-ck boots successful. Also inserting a card after booting does not detect the sd-card but instead crashes systemd-udev with the kernel blaming the error on block/elevator.c
Using an external card-reader works.

To report the bug and send in the full error message, is it an bug in udev or in ck?

Tharbad commented on 2017-08-22 19:00

It's appears that there's a conflict between scsi_mod.use_blk_mq=0 and any of the multi queue I/O schedulers. So either set scsi_mod.use_blk_mq to 1 or use the "none" scheduler.
Choosing option n2 makes this kernel worthless.
I haven't tried it yet.

Found it here: https://bbs.archlinux.org/viewtopic.php?pid=1730573#p1730573

vp1981 commented on 2017-08-22 12:21

@hotty, @Tharbad, guys would you mind to provide more details? For example, your system configuration (HDD/SSD, FS), kernel cmdline, what queue scheduler do you use and etc. And don't forget that graysky only build the kernel with CK patches, so report any problems to CK, https://ck-hack.blogspot.com/.

Tharbad commented on 2017-08-22 12:08

System freezes sometime during the graphical login. CPU is at 100%.
ctrl+alt+F# and magic keys don't work.

hotty commented on 2017-08-20 21:23

Still got a non-bootable kernel, both from here or from repo-ck freezes at boot since version 4.12.7. Normal arch kernel boots without any issues. Just a blinking cursor and after 2 minutes it shows "Triggering uevents" but thats it.

graysky commented on 2017-08-17 19:13

I bumped to 4.12.8-1 but will hold on building for [repo-ck] until CK verifies the recent merges (github) will not be released to ck3 any time soon.

mrkline commented on 2017-08-15 21:38

@graysky: See https://groups.google.com/forum/#!msg/bfq-iosched/j4QUK4IPOhw/CY1LmmhcAwAJ, as previously mentioned - apparently other kernel devs have removed compile-time scheduler defaults.

I've been able to "fix" it for now with a udev rule, slightly modified from what @vp1981 posted:

ACTION=="add|change", KERNEL=="sd[a-z]", ATTR{queue/scheduler}="bfq"

(The wildcard was causing several false positives and errors to show up in the system log while booting.)

graysky commented on 2017-08-15 21:20

@mrkline - Interesting... I get the same.
EDIT: https://ck-hack.blogspot.com/2017/08/linux-412-ck2-muqss-version-0160-for.html?showComment=1502832329876#c3067299809122898386

mrkline commented on 2017-08-15 20:20

Scratch that, 4.12.7-2 seems to have the same issue.

mrkline commented on 2017-08-15 18:40

Apparently ck-2 contains changes to make BFQ the default scheduler: http://ck-hack.blogspot.com/2017/08/linux-412-ck2-muqss-version-0160-for.html

So we should be good to go as soon as @graysky bumps this.

vp1981 commented on 2017-08-15 12:32

@mrkline, you may ask post-factum to translate new on linux.org.ru, see https://www.linux.org.ru/news/kernel/13581989, but this is receipt from the news:

[russian]
Планировщики ввода-вывода при использовании blk-mq нельзя указывать через командную строку ядра в загрузчике. Вместо этого используйте правило udev, например:

ACTION=="add|change", KERNEL=="sd[a-z]*", ATTR{queue/scheduler}="bfq"

Чтобы включить подсистему blk-mq, ядро нужно собрать с опцией CONFIG_SCSI_MQ_DEFAULT=y или передать параметр scsi_mod.use_blk_mq=1 из загрузчика.
[/russian]

(my translation)
... use udev rule, for example:

ACTION=="add|change", KERNEL=="sd[a-z]*", ATTR{queue/scheduler}="bfq"

In order to enable blk-mq subsystem the kernel has to be configured with option CONFIG_SCSI_MQ_DEFAULT=y or pass parameter scsi_mod.use_blk_mq=1 in boot loader.

P.S. Gee, I found the news on post-factum site:
https://pf.natalenko.name/news/, see "v4.12-pf2 is out" (https://pf.natalenko.name/news/?p=291)

mrkline commented on 2017-08-15 07:27

@graysky 4.12.7-1 boots and runs fine on several of my boxen.

The only curious tidbit is that even though BFQ is set as the default:
$ zcat /proc/config.gz | grep -i BFQ
CONFIG_DEFAULT_BFQ=y
CONFIG_DEFAULT_IOSCHED="bfq"
CONFIG_IOSCHED_BFQ=y
CONFIG_BFQ_GROUP_IOSCHED=y

The kernel is defaulting to mq-deadline:
$ cat /sys/block/sda/queue/scheduler
[mq-deadline] kyber bfq none

Is this what @kwe was discussing below?

EDIT: See https://groups.google.com/d/msg/bfq-iosched/j4QUK4IPOhw/CY1LmmhcAwAJ
What's the simplest way to switch all disks to BFQ on boot in the meantime?

artafinde commented on 2017-08-15 06:49

Building 4.12.7-1. Apart for the changes in config so you have to configure lots more than before we are missing the CPU optimizations. Update: Got a non-bootable kernel. Trying a second build, config seems better now as expected (I'll blame ccache which I disabled now). #2 Update: Non bootable again, stuck at "Triggering udevents"

pedrogabriel commented on 2017-08-15 02:03

4.12.7-1 works like a charm.

kwe commented on 2017-08-14 19:58

@graysky That config option will disappear in the future because the behavior will be enabled by default:

https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/drivers/scsi?id=5c279bd9e40624f4ab6e688671026d6005b066fa&h=master

Apparently this is planned for 4.13:

linux/drivers/scsi$ git tag --contains 5c279bd9e40624f4ab6e688671026d6005b066fa
v4.13-rc1
v4.13-rc2
v4.13-rc3
v4.13-rc4
v4.13-rc5

So, when it comes around, you might want to remove it again. :)

beest commented on 2017-08-14 19:25

4.12.7-1 boots, everything seems kosher so far.

graysky commented on 2017-08-14 18:31

@zebulon - Please try 4.12.7-1 which contains a suggestion by CK to fix. Does it boot and run? I cannot build and test for some time.

graysky commented on 2017-08-13 19:07

@amoka - There is not yet a stable ck1 patchset for 4.12.

pedrogabriel commented on 2017-08-13 05:11

Got a non-bootable system too.

zebulon commented on 2017-08-12 19:25

For your info, when you wait long enough, you get error messages: "systemd-udevd blocked for more than 120 seconds".

post-factum commented on 2017-08-12 19:16

-ck is still buggy, unfortunately. I'd recommend downgrade for the time being.

graysky commented on 2017-08-12 16:27

@zerophase - Yes, I too get a non-bootable. I emailed CK.
@kwe - Agreed, thanks for pointing that out. Will be edited in the -3 release.

kwe commented on 2017-08-12 16:09

@graysky How much sense does the _BFQ_enable_ option have with 4.12?

$ grep -i bfq config.x86_64
CONFIG_DEFAULT_BFQ=y
CONFIG_DEFAULT_IOSCHED="bfq"
CONFIG_IOSCHED_BFQ=y
CONFIG_BFQ_GROUP_IOSCHED=y

zerophase commented on 2017-08-12 16:01

I just get a blinking cursor.

zebulon commented on 2017-08-12 15:43

@graysky: unfortunately I cannot boot. It compiled fine (with Skylake optimisation) and seemed to install. But for the first time with the ck kernel, it hangs at the very beginning of the boot sequence. I can only see systemd version, then nothing happens.

zebulon commented on 2017-08-12 15:27

@graysky: thanks. I was wondering why MuQSS had disappeared from config.i686.

graysky commented on 2017-08-12 15:05

@zebulon - Fuck, you're correct. Fixed in 4.12.6-2 just now.

zebulon commented on 2017-08-12 14:53

@graysky: thanks for maintaining this package. One thing I noticed: up to 4.11 it asked via a numbered menu for the kind of CPU optimisation the user wanted. This time it did not, is this normal?

puithove commented on 2017-07-29 11:37

Looks like CK was busy for a while playing with his new Tesla :) Good for him.

zerophase commented on 2017-07-28 19:31

4.12 is up on core. Are we waiting on the ck patch?

Junio_Calu commented on 2017-07-16 11:37

About P6_NOPS, should I enable it?

graysky commented on 2017-07-13 22:34

@gnu - done, will be in the next update, thks

beest commented on 2017-07-13 15:02

Just for the sake of pedantry, next release you may want to change the URL preceding _BFQ_enable_ to https://wiki.archlinux.org/index.php/Improving_performance#Tuning_IO_schedulers as the fragment in the original comment is dead and the linux-ck page links there now anyway.

r08 commented on 2017-07-11 20:35

@QuartzDragon
@artafinde
Thanks for the feedback. I guess I'm not the only one... Will test this using CONFIG_HZ_PERIODIC=y to see if I get the same latency issues..

artafinde commented on 2017-07-09 07:17

@QuartzDragon: I have similar behaviour with my laptop here (Intel(R) Core(TM) i7-4510U CPU @ 2.00GHz). I was setting the tick to 1000Hz before MuQSS but on the first iterations of MuQSS I think it was better. Do you get same freezes with linux from core? Is it fixed if you compile your own -ARCH and set tick to 1000Hz?

QuartzDragon commented on 2017-07-08 22:38

When I change CONFIG_HZ to 1000, CONFIG_HZ_100_MUQSS seems to be automatically unset.

artafinde commented on 2017-07-08 08:57

MuQSS 0.150 reverted the Hz back to 100Hz and I think that's the recommended value from ck. @QuartzDragon, r08 do you change both the below:

CONFIG_HZ_100_MUQSS=y
CONFIG_HZ=100

graysky commented on 2017-07-08 07:37

@Quartz and r08 - Post to ck's blog?

QuartzDragon commented on 2017-07-08 06:54

In the past, I too had noticed weird stuttering using 100Hz, so now I always set it to 1000Hz. System is smooth. :)

r08 commented on 2017-07-07 10:51

Interesting how I noticed slow downs using the default MuQSS@100Hz while using Firefox and running Steam games like Team Fortress 2.
Setting it to 1000Hz fixed the latency issues I was having. Should 1000Hz tick be the recommended setting for a "normal desktop user?"
Running old Intel Core 2 T9500@2.60Ghz

zerophase commented on 2017-07-05 22:23

Anyone else getting mce errors at boot, occasionally, on Haswell? I only get them when I hard reboot if the computer stalls after displaying time till next fsck.

nTia89 commented on 2017-07-05 17:58

Have you read it?
https://wiki.archlinux.org/index.php/Makepkg#Creating_optimized_packages

vp1981 commented on 2017-07-05 02:43

@j1simon: I think SuperIce97 meant that if you compare output for 'native' for both Skylake and Kaby Lake then you don't see any difference. I have two hosts one with Skylake and other with Kaby Lake. I compared they 'native' output and the difference:

$ diff -y --suppress-common-lines native-sl.out native-kl.out
-march= skylake | -march= broadwell
-mtune= skylake | -mtune= generic

For example, this is difference between Sandy Bridge and Skylake:

$ diff -y --suppress-common-lines native-sb.out native-sl.out
-mabm [disabled] | -mabm [enabled]
-madx [disabled] | -madx [enabled]
-march= sandybridge | -march= skylake
-mavx2 [disabled] | -mavx2 [enabled]
-mbmi [disabled] | -mbmi [enabled]
-mbmi2 [disabled] | -mbmi2 [enabled]
-mclflushopt [disabled] | -mclflushopt [enabled]
-mf16c [disabled] | -mf16c [enabled]
-mfma [disabled] | -mfma [enabled]
-mfsgsbase [disabled] | -mfsgsbase [enabled]
-mlzcnt [disabled] | -mlzcnt [enabled]
-mmovbe [disabled] | -mmovbe [enabled]
-mprfchw [disabled] | -mprfchw [enabled]
-mrdrnd [disabled] | -mrdrnd [enabled]
-mrdseed [disabled] | -mrdseed [enabled]
-msgx [disabled] | -msgx [enabled]
-mtune= sandybridge | -mtune= skylake
-mxsavec [disabled] | -mxsavec [enabled]
-mxsaves [disabled] | -mxsaves [enabled]

Sandy Bridge:
Intel(R) Core(TM) i5-2410M CPU @ 2.30GHz

Skylake:
Intel(R) Core(TM) i5-6200U CPU @ 2.30GHz

Kaby Lake:
Intel(R) Core(TM) i5-7400 CPU @ 3.00GHz

@SuperIce97: at first I was confused why 'native' reports that arch is 'broadwell' and even thought that GCC don't support this new processor. But if you are right about behavior of GCC when one passes 'native' vs 'NAME' then it is related only to 'generic' build for Kaby Lake and that processor is fully supported by GCC.

SuperIce97 commented on 2017-07-04 21:57

@j1simon: I believe this is because gcc optimizes in a way to make sure that all CPUs from that architecture will run correctly, meaning that if 1 CPU is missing an instruction, it will not be included in the optimization. This leads to differences as some CPUs in the same architecture supports newer instructions. Here is the diff between the broadwell arch optimizations and native for an Intel 3205u Celeron Broadwell CPU:
diff -y --suppress-common-lines broadwell.out native.out
-mabm [disabled] | -mabm [enabled]
-mcx16 [disabled] | -mcx16 [enabled]
-mfsgsbase [disabled] | -mfsgsbase [enabled]
-mfxsr [disabled] | -mfxsr [enabled]
-mlzcnt [disabled] | -mlzcnt [enabled]
-mmmx [disabled] | -mmmx [enabled]
-mmovbe [disabled] | -mmovbe [enabled]
-mno-sse4 [enabled] | -mno-sse4 [disabled]
-mpclmul [disabled] | -mpclmul [enabled]
-mpopcnt [disabled] | -mpopcnt [enabled]
-mprfchw [disabled] | -mprfchw [enabled]
-mrdrnd [disabled] | -mrdrnd [enabled]
-mrdseed [disabled] | -mrdseed [enabled]
-msahf [disabled] | -msahf [enabled]
-msse [disabled] | -msse [enabled]
-msse2 [disabled] | -msse2 [enabled]
-msse3 [disabled] | -msse3 [enabled]
-msse4 [disabled] | -msse4 [enabled]
-msse4.1 [disabled] | -msse4.1 [enabled]
-msse4.2 [disabled] | -msse4.2 [enabled]
-mssse3 [disabled] | -mssse3 [enabled]

As you can see, on my Broadwell CPU, broadwell optimization != native optimizations. Also notice how "native" only enables optimizations instead of disabling. This is why I usually recommend -march=native, as you will be able to optimize your CPU as much as possible. For Kaby Lake vs Skylake, here is the Wikipedia page listing the differences (which are all clock speed/process, socket/chipset, and GPU changes, not microarchitecture changes):
https://en.wikipedia.org/wiki/Kaby_Lake#Architecture_changes_compared_to_Skylake

j1simon commented on 2017-07-04 12:23

@SuperIce97 I don't sure of that.
My CPU: Intel(R) Core(TM) i7-7700 CPU @ 3.60GHz
I execute:
$ gcc -march=skylake -mtune=skylake -Q --help=target > skylake.out
and:
$ gcc -march=native -mtune=native -Q --help=target > native.out

and these are the differences:
$ diff -y --suppress-common-lines skylake.out native.out
-mabm [disabled] | -mabm [enabled]
-madx [disabled] | -madx [enabled]
-maes [disabled] | -maes [enabled]
-march= skylake | -march= broadwell
-mavx [disabled] | -mavx [enabled]
-mavx2 [disabled] | -mavx2 [enabled]
-mbmi [disabled] | -mbmi [enabled]
-mbmi2 [disabled] | -mbmi2 [enabled]
-mclflushopt [disabled] | -mclflushopt [enabled]
-mcx16 [disabled] | -mcx16 [enabled]
-mf16c [disabled] | -mf16c [enabled]
-mfma [disabled] | -mfma [enabled]
-mfsgsbase [disabled] | -mfsgsbase [enabled]
-mfxsr [disabled] | -mfxsr [enabled]
-mhle [disabled] | -mhle [enabled]
-mlzcnt [disabled] | -mlzcnt [enabled]
-mmmx [disabled] | -mmmx [enabled]
-mmovbe [disabled] | -mmovbe [enabled]
-mno-sse4 [enabled] | -mno-sse4 [disabled]
-mpclmul [disabled] | -mpclmul [enabled]
-mpopcnt [disabled] | -mpopcnt [enabled]
-mprfchw [disabled] | -mprfchw [enabled]
-mrdrnd [disabled] | -mrdrnd [enabled]
-mrdseed [disabled] | -mrdseed [enabled]
-mrtm [disabled] | -mrtm [enabled]
-msahf [disabled] | -msahf [enabled]
-msgx [disabled] | -msgx [enabled]
-msse [disabled] | -msse [enabled]
-msse2 [disabled] | -msse2 [enabled]
-msse3 [disabled] | -msse3 [enabled]
-msse4 [disabled] | -msse4 [enabled]
-msse4.1 [disabled] | -msse4.1 [enabled]
-msse4.2 [disabled] | -msse4.2 [enabled]
-mssse3 [disabled] | -mssse3 [enabled]
-mtune= skylake | -mtune= generic
-mxsave [disabled] | -mxsave [enabled]
-mxsavec [disabled] | -mxsavec [enabled]
-mxsaveopt [disabled] | -mxsaveopt [enabled]
-mxsaves [disabled] | -mxsaves [enabled]

I think it isn't the same.

sir_lucjan commented on 2017-06-27 10:14

@zhenyu:

1. gpg --receive-keys 79BE3E4300411886 38DBBDC86092693E

2. repo-ck.com, not ck-repo.com.

zhenyu commented on 2017-06-27 03:30

two errors:
first,two unknown keys 79BE3E4300411886 38DBBDC86092693E,fix it plz.
second,the ck-repo.com can not update,exec 'pacman -Syy' return me 'can not find ck-repo.db'.
plz tell me how to fix it.thx!

graysky commented on 2017-06-23 21:30

https://wiki.archlinux.org/index.php/Arch_User_Repository wrote:
"Please do not leave a comment containing the version number every time you update the package. This keeps the comment section usable for valuable content mentioned above."

I will therefore stop commenting as I have :)

francoism90 commented on 2017-06-16 17:23

@graysky: sorry for delay, output of Kaby Lake system:
$ gcc -c -Q -march=native --help=target | grep 'march\|mtune'
-march= broadwell
-mtune-ctrl=
-mtune= generic

$ cat /proc/cpuinfo | grep 'model name' | uniq
model name : Intel(R) Core(TM) i3-7100 CPU @ 3.90GHz

@graysky @vp1981: weird it shows broadwell :S

vp1981 commented on 2017-06-12 11:26

I have host with

$ cat /proc/cpuinfo | grep 'model name' | uniq
model name : Intel(R) Core(TM) i5-7400 CPU @ 3.00GHz

and

$ gcc -c -Q -march=native --help=target | grep 'march\|mtune'

gives me

-march= broadwell
-mtune-ctrl=
-mtune= generic

While on host with

$ cat /proc/cpuinfo | grep 'model name' | uniq
model name : Intel(R) Core(TM) i5-6200U CPU @ 2.30GHz

the same 'gcc...' gives me

-march= skylake
-mtune-ctrl=
-mtune= skylake

I faced with this mismatch for Kaby Lake when was installing Archlinux on that new host. After some investigation I decided that -march|-mtune=skylake is appropriate for Kaby Lake processor.

P.S. It is curious enough that both processor on Intel site are categorized as 'Products formely...', not sure what that mean. See http://ark.intel.com/products/88193/Intel-Core-i5-6200U-Processor-3M-Cache-up-to-2_80-GHz and http://ark.intel.com/products/97147/Intel-Core-i5-7400-Processor-6M-Cache-up-to-3_50-GHz

graysky commented on 2017-06-11 19:30

Are you sure? On your kaby lake what is the output of:

gcc -c -Q -march=native --help=target | grep 'march\|mtune'

francoism90 commented on 2017-06-11 18:53

@SuperIce97: Ah thanks, I'll give it a try. :)

SuperIce97 commented on 2017-06-11 18:38

@francoism90: Kaby Lake has an identical architecture to Skylake (they just improved the manufacturing process which simply allows for higher clock speeds), so you could just use use the Skylake optimized version. You could compile it with native optimizations.

francoism90 commented on 2017-06-11 16:27

@graysky: thanks, will install generic instead. :)

graysky commented on 2017-06-11 16:25

There is no KL support in gcc v7.1.1. https://gcc.gnu.org/onlinedocs/

francoism90 commented on 2017-06-11 16:21

Hi graysky,

Thanks for providing the kernel and packages. :)
Is it possible to add support for Kaby Lake?

graysky commented on 2017-05-26 21:52

@mareexx and Davikar - Best to post to ck's blog as well since that is more or less the bug tracker for the patchset.

mareex commented on 2017-05-26 21:26

I get really high cpu load with 4.11. These processes have huge cpu spikes:

rcu_preempt
kworker/u8:<some_number>
irq/279-s-iwlwi
ksoftirqd/1

My fans are screaming. Stock arch kernel is calm. Can anybody else confirm?

Davikar commented on 2017-05-26 17:04

Ever since I started using 4.11 I keep getting these kernel panics:

------------[ cut here ]------------
WARNING: CPU: 1 PID: 449 at net/ipv4/tcp_input.c:2819 tcp_fastretrans_alert+0x8e7/0xad0
Modules linked in: ip6table_nat nf_nat_ipv6 ip6t_REJECT nf_reject_ipv6 ip6table_mangle ip6table_raw nf_conntrack_ipv6 nf_defrag_ipv6 xt_recent ipt_REJECT nf_reject_ipv4 xt_comment xt_multiport xt_conntrack xt_hashlimit xt_addrtype xt_mark xt_nat xt_tcpudp xt_CT iptable_raw nf_log_ipv6 xt_NFLOG nfnetlink_log xt_LOG nf_log_ipv4 nf_log_common nf_nat_tftp nf_nat_snmp_basic nf_conntrack_snmp nf_nat_sip nf_nat_pptp nf_nat_proto_gre nf_nat_irc nf_nat_h323 nf_nat_ftp nf_nat_amanda ts_kmp nf_conntrack_amanda nf_conntrack_sane nf_conntrack_tftp nf_conntrack_sip nf_conntrack_pptp nf_conntrack_proto_gre nf_conntrack_netlink nfnetlink nf_conntrack_netbios_ns nf_conntrack_broadcast nf_conntrack_irc nf_conntrack_h323 nf_conntrack_ftp ip6table_filter ip6_tables iptable_filter iptable_mangle ipt_MASQUERADE
nf_nat_masquerade_ipv4 iptable_nat nf_conntrack_ipv4 nf_defrag_ipv4 nf_nat_ipv4 nf_nat nf_conntrack libcrc32c crc32c_generic btrfs xor adt7475 hwmon_vid iTCO_wdt gpio_ich iTCO_vendor_support evdev mac_hid raid6_pq nouveau led_class mxm_wmi wmi video psmouse ttm i2c_i801 drm_kms_helper lpc_ich skge sky2 drm syscopyarea sysfillrect asus_atk0110 sysimgblt fb_sys_fops i2c_algo_bit button shpchp intel_agp intel_gtt acpi_cpufreq tpm_tis tpm_tis_core tpm sch_fq_codel coretemp msr nfsd auth_rpcgss oid_registry nfs_acl lockd grace sunrpc ip_tables x_tables ext4 crc16 jbd2 fscrypto mbcache sd_mod ata_generic pata_acpi serio_raw atkbd libps2 uhci_hcd ehci_pci ehci_hcd ahci libahci usbcore usb_common pata_jmicron mpt3sas raid_class libata scsi_transport_sas scsi_mod i8042 serio
CPU: 1 PID: 449 Comm: irq/19-enp5s4 Tainted: G W 4.11.3-1-ck-core2 #1
Hardware name: System manufacturer System Product Name/P5B-Deluxe, BIOS 1238 09/30/2008
Call Trace:
<IRQ>
dump_stack+0x63/0x83
__warn+0xcb/0xf0
warn_slowpath_null+0x1d/0x20
tcp_fastretrans_alert+0x8e7/0xad0
tcp_ack+0xe57/0x14f0
tcp_rcv_established+0x11f/0x6f0
? sk_filter_trim_cap+0xb7/0x270
tcp_v4_do_rcv+0x130/0x210
tcp_v4_rcv+0xb39/0xcc0
ip_local_deliver_finish+0xa1/0x200
ip_local_deliver+0x5d/0x100
? inet_del_offload+0x40/0x40
ip_rcv_finish+0x1eb/0x3f0
ip_rcv+0x2b3/0x3c0
? ip_local_deliver_finish+0x200/0x200
__netif_receive_skb_core+0x507/0xa70
? tcp4_gro_receive+0x11a/0x1c0
? try_preempt+0x160/0x190
__netif_receive_skb+0x18/0x60
netif_receive_skb_internal+0x81/0xd0
napi_gro_receive+0xdb/0x150
skge_poll+0x397/0x880 [skge]
net_rx_action+0x242/0x3d0
__do_softirq+0x104/0x2e1
? irq_finalize_oneshot.part.2+0xe0/0xe0
do_softirq_own_stack+0x1c/0x30
</IRQ>
do_softirq.part.4+0x41/0x50
__local_bh_enable_ip+0x88/0xa0
irq_forced_thread_fn+0x59/0x70
irq_thread+0x12f/0x1c0
? wake_threads_waitq+0x30/0x30
kthread+0x108/0x140
? irq_thread_dtor+0xc0/0xc0
? kthread_create_on_node+0x70/0x70
ret_from_fork+0x2c/0x40
---[ end trace a181bdf0ee69c250 ]---

graysky commented on 2017-05-25 19:03

Bump to v4.11.3-1
Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.11.3

mrturcot commented on 2017-05-21 03:03

Thanks guys, I was able to upgrade with @grayskys instructions. Also duly noted
@mrkline, cheers.

mrkline commented on 2017-05-20 21:41

@mrturcot nvidia-dkms is also a viable option if you don't mind building it building the driver locally with each kernel update.

graysky commented on 2017-05-20 13:54

Bump to v4.11.2-1
Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.11.2

graysky commented on 2017-05-20 10:24

I can't read any content from that hatebin provider. Try using a gist on github or another provider. In any case, you will need to build both packages first on a machine that isn't running linux-ck with nvidia-ck installed due to the versioning dependencies. Or, uninstall all of your ck packages first and boot into the standard arch kernel, then build them stepwise. I hope you understand.

mrturcot commented on 2017-05-20 10:16

@graysky not quite srry if I was not clear. My pending updates are Linux ck and
Nvidia ck. It does not matter now I initiate the upgrade, all ways lead to this error
https://hastebin.com/ukekaxojom.sql. Tried to upgrade just Linux-ck first by itself
and it fails, tried nvidia-ck & it installs linux-ck as a dependency first and fails
with that error as well. Thanks

graysky commented on 2017-05-20 09:51

@mrt - You built and installed linux-ck 4.11.1-2 and then were unable to build nvidia-ck?

mrturcot commented on 2017-05-20 05:30

Hey getting error when updating with nvidia-ck installed...
:: nvidia-ck: installing linux-ck (4.11.1-2) breaks dependency 'linux-ck<4.11'
More info - https://hastebin.com/utetikakav.coffeescript
Please help... i'm stuck thanks.

graysky commented on 2017-05-20 01:37

Bump to v4.11.1-2
Changelog: fix pre-made config for cpu type
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2017-05-20 01:26

@mrkline - It's not a dumb question at all. Mistake on my part. Fixed in 4.11.1-2. Thank you for the report.

mrkline commented on 2017-05-20 00:18

@graysky: Perhaps this is a dumb question, but is there a reason the build process no longer prompts for your CPU type? I now have to go set CONFIG_MNATIVE and CONFIG_X86_P6_NOP with makenconfig again - the menu you introduced in December has been much more convenient.

graysky commented on 2017-05-19 22:14

Bump to v4.11.1-1
Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.11.1
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog
Linux changes: https://kernelnewbies.org/Linux_4.11

Notes: I need some feedback on the 4.11 release, specifically the extra packages such as nvidia-304xx-ck, nvidia-340xx-ck, broadcom-wl-ck, and virtualbox-ck modules. I have successfully build all of these but can only test nvidia-ck and virtualbox-ck since I do not have older nvidia hardware nor broadcom-wl hardware. Once I get some positive feedback, I will build up the 4.11.x series in [repo-ck] but for now, will hold. Also 4.11.2 is due out tomorrow.

QuartzDragon commented on 2017-05-15 05:38

Do you have the PKGBUILD for 4.11 sitting around? Would love to test it. :)

graysky commented on 2017-05-14 15:46

The major show stopper are the nvidia drivers at the moment. The kernel builds fine but I don't want mess up users of nvidia-xxx.

hekel commented on 2017-05-14 14:49

@graysky I see, thanks. I've only recently started using your repo/pkgbuild again. I guess I spoiled myself by building my own asap for so long.

graysky commented on 2017-05-14 14:19

@hekel - https://wiki.archlinux.org/index.php/Linux-ck#Release_cycle and see https://lists.archlinux.org/pipermail/arch-dev-public/2017-May/thread.html (threads titled Kernel 4.11 status).

hekel commented on 2017-05-14 13:25

BFQ has been out for 4.11 since 05/01
4.11-ck1 has been out since 05/11
4.11.1 just dropped today and 4.11.0 has been out since 4/30

…Is there a particular reason this was updated for 4.10.16 instead of 4.11.1?

graysky commented on 2017-05-14 12:20

Bump to v4.10.16-1
Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.10.16
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

monotykamary commented on 2017-05-13 10:49

https://devtalk.nvidia.com/default/topic/1005209/fully-working-patch-for-nvidia-driver-340-102-compiler-installer-file-and-linux-kernel-4-11/

There is one for 340xx in the nvidia forums, but haven't found one for 304xx.

graysky commented on 2017-05-12 21:59

Link to a patch for 304xx and 340xx?

monotykamary commented on 2017-05-12 21:31

@graysky Gotcha. For dkms, virtualbox-host-dkms builds correctly here. nvidia-dkms requires a MODULE_LICENSE edit for nvidia-drm/nvidia-drm-linux.c. I have not tested broadcom.

graysky commented on 2017-05-12 19:09

@monotykamary - Need to verify that all broadcom, nvidia-xxxx, and vbox modules build against 4.11.x before I bump it.

keepitsimpleengr commented on 2017-05-11 14:32

nvidia-ck-sandybridge: installing nvidia-utils (381.22-1) breaks dependency 'nvidia-utils=378.13'

https://www.archlinux.org/packages/multilib/x86_64/lib32-nvidia-utils/

graysky commented on 2017-05-08 18:31

Bump to v4.10.15-1
Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.10.15
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2017-05-03 18:03

Bump to v4.10.14-1
Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.10.14
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2017-04-27 19:12

Bump to v4.10.13-1
Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.10.13
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2017-04-21 22:17

Bump to v4.10.12-1
Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.10.12
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

artafinde commented on 2017-04-21 14:11

BFQ scheduled for mainline - yoohoo! https://groups.google.com/forum/#!topic/bfq-iosched/D-7wDP47iBI

graysky commented on 2017-04-18 07:26

Bump to v4.10.11-1
Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.10.11
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

kyak commented on 2017-04-14 05:29

Having some problems with Docker and Linux-ck, posted here: https://bbs.archlinux.org/viewtopic.php?pid=1704244

graysky commented on 2017-04-12 19:02

Bump to v4.10.10-1
Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.10.10
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

bacondropped commented on 2017-04-08 18:10

Hm, I'm getting an occasional "kernel bug" crash with BFQ enabled on a per-device basis (4.10.8 and 4.10.9 with NUMA enabled). I've disabled BFQ, let's see if it persists.

graysky commented on 2017-04-08 11:21

Bump to v4.10.9-1
Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.10.9
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

sidro commented on 2017-04-07 18:06

Other errors:

drivers/media/dvb-frontends/tda10071.c:1267:1: fatal error: opening dependency file drivers/media/dvb-frontends/.tda10071.o.d: No such file or directory
MODULE_FIRMWARE(TDA10071_FIRMWARE);
^~~~~~~~~~~~~~~
compilation terminated.
make[3]: *** [scripts/Makefile.build:301: drivers/media/dvb-frontends/tda10071.o] Error 1
make[2]: *** [scripts/Makefile.build:553: drivers/media/dvb-frontends] Error 2
make[1]: *** [scripts/Makefile.build:553: drivers/media] Error 2
make: *** [Makefile:992: drivers] Error 2
==> ERROR: A failure occurred in build().
Aborting...
==> ERROR: Makepkg was unable to build linux-ck.

SuperIce97 commented on 2017-04-07 18:03

@sidro: This is just how AUR/Arch packages work. There is no way to "fix" the package. It is expected that you import the gpg keys before you build the package.

sir_lucjan commented on 2017-04-07 18:03

gpg --recv-keys 79BE3E4300411886

gpg --recv-keys 38DBBDC86092693E

Code works fine. PEBKAC.

sidro commented on 2017-04-07 18:00

Because original Build Script return errors.

==> Verifying source file signatures with gpg...
linux-4.10.tar ... FAILED (unknown public key 79BE3E4300411886)
patch-4.10.8 ... FAILED (unknown public key 38DBBDC86092693E)
==> ERROR: One or more PGP signatures could not be verified!
==> ERROR: Makepkg was unable to build linux-ck.

Fix your code.

Other errors:

drivers/media/dvb-frontends/tda10071.c:1267:1: fatal error: opening dependency file drivers/media/dvb-frontends/.tda10071.o.d: No such file or directory
MODULE_FIRMWARE(TDA10071_FIRMWARE);
^~~~~~~~~~~~~~~
compilation terminated.
make[3]: *** [scripts/Makefile.build:301: drivers/media/dvb-frontends/tda10071.o] Error 1
make[2]: *** [scripts/Makefile.build:553: drivers/media/dvb-frontends] Error 2
make[1]: *** [scripts/Makefile.build:553: drivers/media] Error 2
make: *** [Makefile:992: drivers] Error 2
==> ERROR: A failure occurred in build().
Aborting...
==> ERROR: Makepkg was unable to build linux-ck.

sir_lucjan commented on 2017-04-07 17:04

@sidro: You're wrong.

SuperIce97 commented on 2017-04-07 17:02

@sidro: How is it wrong?

sidro commented on 2017-04-07 16:59

Greg Kroah-Hartman sign is wrong.
Fix it.

sir_lucjan commented on 2017-04-07 16:25

gpg --recv-keys 79BE3E4300411886 38DBBDC86092693

sidro commented on 2017-04-07 16:13

==> Verifying source file signatures with gpg...
linux-4.10.tar ... FAILED (unknown public key 79BE3E4300411886)
patch-4.10.8 ... FAILED (unknown public key 38DBBDC86092693E)
==> ERROR: One or more PGP signatures could not be verified!
==> ERROR: Makepkg was unable to build linux-ck.

graysky commented on 2017-03-31 17:58

Bump to v4.10.8-1
Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.10.8
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2017-03-30 11:14

Bump to v4.10.7-1
Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.10.7
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2017-03-26 14:59

Bump to v4.10.6-1
Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.10.6
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2017-03-22 18:33

Bump to v4.10.5-1
Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.10.5
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2017-03-18 12:48

Bump to v4.10.4-1
Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.10.4
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2017-03-15 19:20

Bump to v4.10.3-1
Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.10.3
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2017-03-13 18:15

Bump to v4.10.2-2
Changelog: apply fix for CVE-2017-2636
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2017-03-12 18:33

@bugs - Manjaro != Arch. It's not surprising that Arch packages might not work on it.

bugscze commented on 2017-03-12 17:05

Manjaro won't boot with this kernel. In default, Kernel panic shows up.

Adding this line into /boot/grub/grub.cfg: initrd /boot/initramfs-linux-ck-ivybridge.img

kernel boots but cannot startx (with and without nvidia-ck-ivybridge)

Anyone knows how to solve this issue?

graysky commented on 2017-03-12 11:18

Bump to v4.10.2-1
Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.10.2
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2017-03-11 10:16

@ArchNemo: Arch != Antergos.

QuartzDragon commented on 2017-03-11 00:28

Have you tried this command?

cat /sys/block/<drive>/queue/scheduler

ArchNemo commented on 2017-03-11 00:08

How to persistently enable bfq scheduler? I tried adding "elevator=bfq" in kernel parameter as suggested by Arch's wiki, but it didn't do anything. Thanks!

Updates
I just found out Antergos by default has a script trying to set scheduler in:
/etc/udev/rules.d/60-schedulers.rules

So I commented everything in that file and it's good now, nothing is wrong with linux-ck.
I guess Antergos isn't real Arch, eh? :D

mradermaxlol commented on 2017-03-03 21:05

@graysky nope, it's *probably* not gonna be there because of AMD devs' habits, I think, but that's a good idea to try to push a backport to 4.10.X. Will do that ASAP.

graysky commented on 2017-03-03 20:22

@mrader - I tend to not put out-of-tree patches in the PKGBUILD and [repo-ck]... Do you know if this is planned to be back-ported in the 4.10 stable queue?[1] I currently see nothing there...

1. https://git.kernel.org/cgit/linux/kernel/git/stable/stable-queue.git/tree/queue-4.10

mradermaxlol commented on 2017-03-03 20:13

@graysky is it possible to include this - https://gist.github.com/mradermaxlol/bb82874aaddec6794d62ff98187ad639 - patch in the next PKGREL? It works for me and people on freedesktop.org bugtracker, so it should be just fine to include it. It's a cherry-pick from the in-development 4.11 kernel, and thus would be useless and safe to remove after you update the package to that version.

graysky commented on 2017-03-01 19:42

@cooljay - again, recommend that you open a FS against the ARCH kernel

cooljay032 commented on 2017-03-01 13:28

commit tells us otherwise
-CONFIG_LIRC_SERIAL=m
-CONFIG_LIRC_SERIAL_TRANSMITTER=y

not sure if there are other (new) options to run serial devices!?!
+CONFIG_IR_SERIAL=m
+CONFIG_IR_SERIAL_TRANSMITTER=y

think i have to dig in here...

graysky commented on 2017-02-28 21:16

@cooljay - It's not in the official ARCH kernel either. I simply mirror ARCH's config (with the ck changes). The module has either been dropped upstream or we aren't configured to build it... you might wanna open up a flyspray against the kernel.

When I look at the 4.9 -> 4.10 commit, I don't see that we dropped the config option containing 'serial' though: https://git.archlinux.org/svntogit/packages.git/commit/trunk?h=packages/linux&id=192b948e730c25abf0269068755ef934486aec54

cooljay032 commented on 2017-02-28 20:19

seems the lirc_serial driver missing!?!

-- Unit systemd-modules-load.service has begun starting up.
Feb 28 21:07:41 vdr systemd-modules-load[801]: Inserted module 'acpi_cpufreq'
Feb 28 21:07:41 vdr systemd-modules-load[801]: modprobe: ERROR: could not find module by name='lirc_serial'
Feb 28 21:07:41 vdr systemd-modules-load[801]: modprobe: ERROR: could not insert 'lirc_serial': Unknown symbol in module, or unknown parameter (see dmesg)
Feb 28 21:07:41 vdr systemd[1]: systemd-modules-load.service: Main process exited, code=exited, status=1/FAILURE
Feb 28 21:07:41 vdr systemd[1]: Failed to start Load Kernel Modules.

SuperIce97 commented on 2017-02-28 00:29

@mrkline That's a bit odd, yeah. nvidia-libgl is now on version 378.13-3 but nvidia-dkms is 378.13-2. You need 378.13-3 for Linux 4.10 support, so I wonder why they haven't pushed that dkms update yet.

mrkline commented on 2017-02-28 00:26

Ah - turns out the patch needed for 4.10 is in nvidia-dkms 378.13-3, which is currently only available in the testing repo. Installing it by hand from https://www.archlinux.org/packages/testing/x86_64/nvidia-dkms/ did the trick.

mrkline commented on 2017-02-28 00:11

nvidia-dkms fails to build against 4.10.1-1. Is there a known workaround? Or do we now have to use nvidia-ck? (I've been using nvidia-dkms with linux-ck for quite a while without any trouble.)

From the build log:

CC [M] /var/lib/dkms/nvidia/378.13/build/nvidia/nv-vtophys.o
/var/lib/dkms/nvidia/378.13/build/nvidia/nv-pat.c: In function ‘nvidia_cpu_callback’:
/var/lib/dkms/nvidia/378.13/build/nvidia/nv-pat.c:213:14: error: ‘CPU_DOWN_FAILED’ undeclared (first use in this function)
case CPU_DOWN_FAILED:
^~~~~~~~~~~~~~~
/var/lib/dkms/nvidia/378.13/build/nvidia/nv-pat.c:213:14: note: each undeclared identifier is reported only once for each function it appears in
/var/lib/dkms/nvidia/378.13/build/nvidia/nv-pat.c:220:14: error: ‘CPU_DOWN_PREPARE’ undeclared (first use in this function)
case CPU_DOWN_PREPARE:
^~~~~~~~~~~~~~~~
In file included from /var/lib/dkms/nvidia/378.13/build/nvidia/nv-pat.c:15:0:
/var/lib/dkms/nvidia/378.13/build/nvidia/nv-pat.c: In function ‘nv_init_pat_support’:
/var/lib/dkms/nvidia/378.13/build/common/inc/nv-linux.h:391:34: error: implicit declaration of function ‘register_cpu_notifier’ [-Werror=implicit-function-declaration]
#define register_hotcpu_notifier register_cpu_notifier
^
/var/lib/dkms/nvidia/378.13/build/nvidia/nv-pat.c:258:17: note: in expansion of macro ‘register_hotcpu_notifier’
if (register_hotcpu_notifier(&nv_hotcpu_nfb) != 0)
^~~~~~~~~~~~~~~~~~~~~~~~
/var/lib/dkms/nvidia/378.13/build/nvidia/nv-pat.c: In function ‘nv_teardown_pat_support’:
/var/lib/dkms/nvidia/378.13/build/common/inc/nv-linux.h:388:36: error: implicit declaration of function ‘unregister_cpu_notifier’ [-Werror=implicit-function-declaration]
#define unregister_hotcpu_notifier unregister_cpu_notifier
^
/var/lib/dkms/nvidia/378.13/build/nvidia/nv-pat.c:283:9: note: in expansion of macro ‘unregister_hotcpu_notifier’
unregister_hotcpu_notifier(&nv_hotcpu_nfb);
^~~~~~~~~~~~~~~~~~~~~~~~~~
CC [M] /var/lib/dkms/nvidia/378.13/build/nvidia/os-interface.o
cc1: some warnings being treated as errors
make[2]: *** [scripts/Makefile.build:294: /var/lib/dkms/nvidia/378.13/build/nvidia/nv-pat.o] Error 1
make[2]: *** Waiting for unfinished jobs....
make[1]: *** [Makefile:1494: _module_/var/lib/dkms/nvidia/378.13/build] Error 2
make[1]: Leaving directory '/usr/lib/modules/4.10.1-1-ck/build'
make: *** [Makefile:81: modules] Error 2

graysky commented on 2017-02-27 22:43

Bump to v4.10.1-1
Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.10.1
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2017-02-26 11:31

Bump to v4.9.13-1
Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.9.13
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

QuartzDragon commented on 2017-02-26 00:24

Cannot reproduce here, even after many hours of system activity. I also have an Ivy Bridge CPU.

simona commented on 2017-02-25 19:42

Anyone has freeze problem with all CK (CK standard, CK-ivybridge) ?

graysky commented on 2017-02-23 20:32

Bump to v4.9.12-1
Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.9.12
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2017-02-23 20:32

@kyak - Rebuilt a few days ago. Thanks for reporting.

kyak commented on 2017-02-21 18:14

It seems that virtualbox-ck-host-modules-ivybridge requires rebuild:

Failed to insert 'vboxnetadp': Invalid argument
Failed to insert 'vboxnetflt': Invalid argument
Failed to insert 'vboxnetadp': Invalid argument
Failed to insert 'vboxnetflt': Invalid argument


Kernel 4.9.11-1-ck-ivybridge from repo-ck.

Gelmo commented on 2017-02-21 04:58

@dkaylor Thanks for the heads up. I didn't realize this was following the Core release exclusively, and I've removed the flag. Regardless, I'm leaving the PKGBUILD; it builds 4.10 just fine :D

dkaylor commented on 2017-02-20 09:16

@gelmo Please read graysky's wiki for linux-ck before flagging out-of-date.

He's been doing a very good job with this package for years now, your updated PKGBUILD is not needed...

Gelmo commented on 2017-02-20 07:11

Here's a working PKGBUILD for 4.10:

http://pastebin.com/raw/iz4E0wH7

Updated versioning, commented out the upstream patches, commented out the console loglevel patch (implemented now; I got this option presented to me pre-nconfig).

Edit: The loglevel option is now available in nconfig under Kernel Hacking

Edit 2: I also had to comment out the fpu patch on my setup. If you do this, don't forget to comment it out in the sources and prepare sections. Currently writing this from 4.10.0-1-ck, no issues.

graysky commented on 2017-02-18 17:20

Bump to v4.9.11-1
Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.9.11
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2017-02-15 22:45

Bump to v4.9.10-1
Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.9.10
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2017-02-09 22:34

Bump to v4.9.9-1
Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.9.9
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

cdown commented on 2017-02-08 16:31

@VerruckteFuchs @fekle This is a bug in pacaur. I've just put up a patch to fix it at https://github.com/rmarquis/pacaur/pull/646.

VerruckteFuchs commented on 2017-02-06 20:58

@graysky @artafinde A simple makepkg -sri after clearing out the src directory sorted it out. It looks like something broke after pacaur got updated.

artafinde commented on 2017-02-06 20:07

Just follow the wiki: https://wiki.archlinux.org/index.php/Arch_User_Repository#Installing_packages

fekle commented on 2017-02-06 20:00

@graysky to wait and see if other people experience this problem ;) you're right though, just tried installing manually and it worked just fine - oops.

graysky commented on 2017-02-06 18:54

Why would trying in a few days make a difference? Don't build with an AUR helper. Those things are way more trouble than they're worth. Just build with makepkg and profit.

VerruckteFuchs commented on 2017-02-06 18:40

@artafinde I've run

cd ~/.cache/pacaur/linux-ck/src/linux-4.9
make clean #I also tried make mrproper (used in the PKGBUILD) and make distclean

I also tried
pacaur -Sc
to see if that did anything to help. As far as I can tell skimming over the PKGBUILD, in the prepare() step make mrproper comes after the initial patching and should clean the build environment. Oddly enough I had no issues upgrading linux-ck on my laptop, but on my desktop it's having trouble with the prepare() step for some reason. It might be because I upgraded pacaur before upgrading linux-ck and there is a change in PKGBUILD syntax/formatting or something else. On my laptop I'm pretty sure I upgraded linux-ck and pacaur at the same time, which means the previous version of pacaur handled the installation. That's all I can think of that's different.

fekle commented on 2017-02-06 16:57

@VerruckteFuch I've got the same problem: http://pastebin.com/raw/4kT3quL6
Tried deleting .cache/pacaur, choosing both yes and no when asked (see paste), to no avail.
Anyway, I'll try again in a few days, and as always, thanks for the great work @graysky!

artafinde commented on 2017-02-06 16:01

@VerruckteFuchs clean your build environment.

VerruckteFuchs commented on 2017-02-06 15:14

I'm getting a couple of errors in the prepare() step:

The next patch would create the file arch/x86/include/asm/asm-prototypes.h,
which already exists! Assume -R? [n]

The next patch would create the file include/asm-generic/asm-prototypes.h,
which already exists! Assume -R? [n]
I keep getting a failure occurring in prepare() regardless if I answer 'y' or 'n' to the question and to the "Apply anyway? [n]" question.

I've started with a fresh download of everything by deleting the ~/.cache/pacaur/linux-ck/ folder. I've had pacaur redownload everything using this method a few times and the failure persists.

graysky commented on 2017-02-04 23:39

Bump to v4.9.8-1
Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.9.8
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

nicegamer7 commented on 2017-02-04 21:23

@graysky Sorry to bother then, only wanted to make sure :)

walkingrobot commented on 2017-02-03 03:32

@zerophase no I do not. I've never tried it.

zerophase commented on 2017-02-03 02:16

@walkingrobot thanks. I turned the SCSI one on, and I noticed I was able to boot successfully once in a bluemoon without udev compiled into my initramfs. Do you know if there is currently a way to boot without udev, and an nvme drive in the pcie slot?

Smasher816 commented on 2017-02-03 00:44

I am getting this error. If I manually extract the .xz file, then I end up receiving conflicts during the patching.

patch: **** Can't open patch file XXXXXXX/linux-ck/src/patch-4.9.7 : No such file or directory

graysky commented on 2017-02-02 21:18

@niceg - It's correct as-is.

nicegamer7 commented on 2017-02-02 14:45

On the PKGBUILD...

nicegamer7 commented on 2017-02-02 14:45

I believe there is a typo on line 168.
It is "s'/..." instead of "'s/...".

xsmile commented on 2017-02-02 14:04

@graysky:
I would like to request a patch addition to fix low CPU freuqencies after standby, most probably limited to Dell notebooks. Please see https://bugzilla.kernel.org/show_bug.cgi?id=90041#c32 for the patch and for more details on the bug tracker. Thanks for the consideration.

walkingrobot commented on 2017-02-02 10:05

@zerophase Here are the NVME options I have.
CONFIG_NVME_CORE=m << this one is missing from your list
CONFIG_BLK_DEV_NVME=m
# CONFIG_BLK_DEV_NVME_SCSI is not set
CONFIG_NVME_FABRICS=m
CONFIG_NVME_RDMA=m
CONFIG_NVME_TARGET=m
CONFIG_NVME_TARGET_LOOP=m
CONFIG_NVME_TARGET_RDMA=m
CONFIG_NVMEM=m

graysky commented on 2017-02-02 00:12

Sorry about that... fixed but didn't bump pkgver.

SuperIce97 commented on 2017-02-01 22:07

The config files are failing to pass the sha256sums

zerophase commented on 2017-02-01 21:41

@graysky I'm just running into issues with booting the drive outside of fallback initramfs, but this isn't unique to CK.

graysky commented on 2017-02-01 21:39

Bump to v4.9.7-1
Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.9.7
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2017-02-01 21:38

@zerophase - you shouldn't need anything special

zerophase commented on 2017-01-31 23:40

Does anyone know which NVME switches I should turn on in the kernel?

I turned on CONFIG_BLK_DEV_NVME, and CONFIG_NVME_TARGET. I just left NVME_TARGET_LOOP, and CONFIG_BLK_DEV_NVME_SCSI off.

Are there any other NVME settings I should turn on?

QuartzDragon commented on 2017-01-27 04:01

Finally updated to 4.9 after watching the bugs get fixed. Nice and stable so far!

Thanks graysky! :)

graysky commented on 2017-01-27 00:39

Bump to v4.9.6-1
Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.9.6
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2017-01-20 20:59

Bump to v4.9.5-1
Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.9.5
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2017-01-15 13:28

Bump to v4.9.4-1
Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.9.4
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

enihcam commented on 2017-01-15 02:18

Since the issue was worked around by enabling suspend-to-RAM, I did not want to spend time in repro it with makepkg. That was for your information only, not a request.

graysky commented on 2017-01-15 01:58

Does that mean it compiles without modification? Please use makepkg and report back.

enihcam commented on 2017-01-15 01:24

@graysky the variable is declared if and only if suspend-to-RAM is enabled, otherwise compiler cannot find the variable therefore complains. I manually compiled the kernel sources with patches you provided.

snack commented on 2017-01-14 18:50

With linux-ck-core2 4.9.3-1 and nvidia-340xx-ck-core2 340.101-3 I cannot start the X server. I get an error "Error: driver Nvidia is already registered. Aborting". No problem for me to stay with linux 4.8, this is just a warning for those wishing to upgrade.

graysky commented on 2017-01-14 15:14

Disabling suspend to ram and affect compilation? Did you try building with makepkg?

enihcam commented on 2017-01-14 14:06

@graysky Disabling suspend-to-RAM can repro the issue.

graysky commented on 2017-01-14 13:24

How does that pm freezing lik you referenced apply to your situation? Did you try building as-is with makepkg?

enihcam commented on 2017-01-14 13:17

@graysky The issue is discussed here:
http://ck-hack.blogspot.com/2016/12/linux-49-ck1-muqss-version-0150.html?showComment=1481533077370#c7648396274643602658

Seems like one patch is missing.

graysky commented on 2017-01-14 12:42

Stop using an AUR helper which I believe you're using and just build with makepkg.

enihcam commented on 2017-01-14 06:41

Both of my two machines failed in compiling Linux-ck 4.9.3 with GCC 6.3.1. Anyone got the same issue?

<-Log Start-------------->

LD fs/built-in.o
==> ERROR: A failure occurred in build().
Aborting...
==> ERROR: Makepkg was unable to build linux-ck.
==> Restart building linux-ck ? [y/N]
==> ---------------------------------
==>

<-Log End------------->

SanskritFritz commented on 2017-01-13 21:55

Users of nvidia-304xx have held back the last version anyway since current version doesn't work at all :D

graysky commented on 2017-01-13 21:47

Bump to v4.9.3-1
Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.9.3
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog
Linux changes: https://kernelnewbies.org/Linux_4.9
Notes: Users of nvidia-304xx will have to pull the corresponding utils package from [testing] until 4.9 is pushed into [core]. Sorry about the inconvenience but way more users of linux-ck want the 4.9 series kernel and I didn't want to hold it up any longer.

graysky commented on 2017-01-12 21:19

@artafinde - READ 2 POSTS DOWN

artafinde commented on 2017-01-12 21:05

Is there a reason the linux 4.9.x is not pushed to repos yet? Someone mentioned in ML that's it's unstable - and seen reports of crashes on BTRFS ML.

graysky commented on 2017-01-12 19:35

For those wanting the 4.9.3 release - http://repo-ck.com/PKG_source/next/linux-ck-493.tar.xz

graysky commented on 2017-01-12 07:24

@enihcam - Not out-of-date, plz read comments.

graysky commented on 2017-01-07 19:25
graysky commented on 2016-12-22 19:02

zerophase commented on 2017-01-11 05:22

@graysky sounds like an efi issue.

graysky commented on 2017-01-10 09:10

@pedrogabriel - http://repo-ck.com/PKG_source/next/linux-ck-492.tar.xz
@zero - Haswell here and no problems.

zerophase commented on 2017-01-09 23:29

@pedrogabriel I believe 4.9 currently has issues with some old and new intel chips. That's why it hasn't had a release for arch yet.

pedrogabriel commented on 2017-01-09 23:25

Could you release a pastebin of 4.9?

graysky commented on 2017-01-09 09:14

Bump to v4.8.17-1
Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.8.17
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2017-01-07 19:25

@nTia89 - Fixed thanks.
@Saren - Whenever the Arch maintainer of nvidia-304xx updates it which is needed to build against the 4.9 code base. See my comments from 2016-12-22 19:02

Saren commented on 2017-01-07 18:45

Hi there, when will be 4.9 arriving? Thanks.

nTia89 commented on 2017-01-07 18:03

missing a word in the enable* patch:

...
It is not recommended to compile on Atom-CPUs with the 'native' option.[4] The
recommendation is to *** the 'atom' option instead.
...

graysky commented on 2017-01-06 19:57

Bump to v4.8.16-1
Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.8.16
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

kogone commented on 2016-12-28 22:13

this is a free for all

graysky commented on 2016-12-25 15:04

The PKGBUILDs in the AUR are free for all to modify.

petterk commented on 2016-12-25 14:44

I have created my own package from this (forked basically) adding support for Reiser4 and with a config optimized solely for the Dell XPS 9550 laptops (in its various configurations).

If its OK for you I'd like to name this Linux-CK-Reiser4 but with clear text that this is not made by you (in case shit hits the fan) but you will ofcourse gain the credit for the inital package the fork is based on.

So Im basically using your PKGBUILD but with other options, the same patches but with a couple ones added, and grouped as ck-skylake to support your pre-compiled modules at repo-ck. Is this OK?

graysky commented on 2016-12-22 19:02

I updated linux-ck and all its daughter packages (broadcom-ck nvidia-x-ck, and virtualbox-ck) but currently nvidia 304xx doens't build without an update to 304.134 and currently nvidia-304xx-utils in [extra] is stuck on 304.132. This means I can't push the 4.9 update until our devs put 304.134 in the repos.

graysky commented on 2016-12-22 08:38

@BS - Yes, I saw it. Just need to verify that all modules build ok. Stay tuned.

BS86 commented on 2016-12-22 07:29

finally, 4.9 is out:
https://git.archlinux.org/svntogit/packages.git/commit/trunk?h=packages/linux&id=26d5f8c9110f0fca96a7ef05c235dd005b1049a7

graysky commented on 2016-12-21 17:00

@Saren - I see... sorry, you'll need to edit the PKGBUILD. You can do it with a call to sed if you want.

Saren commented on 2016-12-21 16:07

@graysky - lol no, I want to enable NUMA.

graysky commented on 2016-12-21 13:58

@Saren - You shouldn't need to modify the PKGBUILD to disable NUMA... that has been defaulted for some time now.

Saren commented on 2016-12-21 13:30

Hi, since it's now possible to configure cpu march now, is it possible to add whether to disable numa as well so that multi cpu system users don't have to edit PKGBUILD every time linux-ck upgrades? Thanks.

mrkline commented on 2016-12-19 18:23

@Rainmaker: I've seen that if pacman is upgrading bash in another process. I would assume it's just a temporary problem - you'd have some really serious issues if /bin/sh went missing. :)

QuartzDragon commented on 2016-12-19 14:59

@Rainmaker

Cannot reproduce here. That's a weird error I've never seen before...

What else is happening on your machine while trying to compile?

Rainmaker commented on 2016-12-19 10:50

4.8.15-2 fails because it's trying to call /bin/sh during the build:

Cyclomatic Complexity 2 sound/core/hwdep.c:snd_hwdep_proc_read
Cyclomatic Complexity 1 sound/core/hwdep.c:alsa_hwdep_init
Cyclomatic Complexity 1 sound/core/hwdep.c:alsa_hwdep_exit
make[2]: /bin/sh: Command not found
make[2]: *** [scripts/Makefile.build:427: sound/core/snd-hwdep.o] Error 127
make[1]: *** [scripts/Makefile.build:440: sound/core] Error 2
make: *** [Makefile:974: sound] Error 2
==> ERROR: A failure occurred in build().
Aborting...

blitz commented on 2016-12-18 23:07

@graysky - Kernel config allows for menu driven cpu arch selection with minimal interaction and pre-selected MNATIVE.
(taken from own PKGBUILD)

_CPU_NATIVE=y
...
### Optionally set processor type for kernel
if [ -n "${_CPU_NATIVE}" ]; then
msg "Setting cpu arch to gcc native..."
if [ "${CARCH}" = "x86_64" ]; then
sed -i -e 's/^CONFIG_GENERIC_CPU=y/# CONFIG_GENERIC_CPU is not set/' \
-i -e 's/^# CONFIG_MNATIVE is not set/CONFIG_MNATIVE=y/' ./.config
else
sed -i -e 's/^CONFIG_M686=y/# CONFIG_M686 is not set/' \
-i -e 's/^# CONFIG_MNATIVE is not set/CONFIG_MNATIVE=y/' ./.config
fi
fi

graysky commented on 2016-12-18 19:38

@mrkline - Update the checksum (again without bumping the pkgver). No real easy way to automate the selection of native.

mrkline commented on 2016-12-18 19:29

@graysky - The menu is quite nice! Hopefully I can stop messing around in nconfig soon. On that note, is it possible to automate the selection? I tried adding a `CONFIG_MNATIVE=y` line to config.x86_64, but I was still prompted when I ran makepkg.

Also, enable_additional_cpu_optimizations_for_gcc_v4.9+_kernel_v3.15+.patch.gz needs its SHA updated.

graysky commented on 2016-12-18 17:36

OK, just updated without bumping the pkgver with this change. Feedback is welcomed.

graysky commented on 2016-12-18 13:42

@artafinde - see my comment from 2016-12-18 10:44 for the example.

artafinde commented on 2016-12-18 12:37

@graysky Can you provide a PKGBUILD (maybe through gist) to see how much interaction will be needed? I personally have to edit PKGBUILD anyway since I'm building with modprobed and BFS enabled by default (and using "-march=native")

nTia89 commented on 2016-12-18 11:43

yes, missing of optimizations without -march=native is my worry! (I know, this seems false but it is true... https://bbs.archlinux.org/viewtopic.php?id=210740)
hence, any modification is welcome, even if this mean to press some key

my doubt is another: who builds this package from AUR, I imagine, wants the best optimizations overall, so, why "-march=native" is not the default choice?

zerophase commented on 2016-12-18 10:59

@graysky depends on how it's implemented. But, in general if a config file has new additions in a package bump it interrupts kernel compilation till I resolve the issue. I imagine whatever you implement will work with yaourt as well.

graysky commented on 2016-12-18 10:44

A modification I am considering is to have the AUR package provide unmodified config files for x86_64 and for i686 with respect to the gcc patch. What this means is that whenever you build linux-ck yourself, makepkg (or am I assuming your AUR helper) will pause and wait for you to select from the list of CPU optimizations (with default being a simple enter key).

This will circumvent the need to modify the PKGBUILD enabled the nconfig option, and subsequent poking around in the nconfig itself. Potentially way faster and easier for the user.

1) How do people feel about this?
2) Will the added interactivity break AUR helpers (I am a pure makepkg person myself)?

Example output:
% makepkg -src
==> Making package: linux-ck 4.8.15-2 (Sun Dec 18 05:42:02 EST 2016)
...
Processor family
1. AMD Opteron/Athlon64/Hammer/K8 (MK8)
2. AMD Opteron/Athlon64/Hammer/K8 with SSE3 (MK8SSE3) (NEW)
3. AMD 61xx/7x50/PhenomX3/X4/II/K10 (MK10) (NEW)
4. AMD Barcelona (MBARCELONA) (NEW)
5. AMD Bobcat (MBOBCAT) (NEW)
6. AMD Bulldozer (MBULLDOZER) (NEW)
7. AMD Piledriver (MPILEDRIVER) (NEW)
8. AMD Steamroller (MSTEAMROLLER) (NEW)
9. AMD Jaguar (MJAGUAR) (NEW)
10. Intel P4 / older Netburst based Xeon (MPSC)
11. Intel Atom (MATOM)
12. Intel Core 2 (MCORE2)
13. Intel Nehalem (MNEHALEM) (NEW)
14. Intel Westmere (MWESTMERE) (NEW)
15. Intel Silvermont (MSILVERMONT) (NEW)
16. Intel Sandy Bridge (MSANDYBRIDGE) (NEW)
17. Intel Ivy Bridge (MIVYBRIDGE) (NEW)
18. Intel Haswell (MHASWELL) (NEW)
19. Intel Broadwell (MBROADWELL) (NEW)
20. Intel Skylake (MSKYLAKE) (NEW)
> 21. Generic-x86-64 (GENERIC_CPU)
22. Native optimizations autodetected by GCC (MNATIVE) (NEW)
choice[1-22?]:

graysky commented on 2016-12-18 10:07

@SuperIce97 @mrkline - Good points about usability and about niche population with aged hardware. I reverted the commit removing native updating 4.8.15-2 without bumping the pkgver since I don't want to rebuild for [repo-ck] since this change has 0 impact to the repo packages.

QuartzDragon commented on 2016-12-18 06:20

@mrkline @SuperIce97

Also agreed! :)

mrkline commented on 2016-12-18 05:29

+1 for bringing back `-march=native`. If you look at what flags are actually passed to the actual compiler (cc1), it provides much more machine-specific information (e.g., cache size, etc.) than `-march=<CPU family>`.

As other users have suggested, it seems like there are workarounds for CPUs where it caused trouble. Even if workarounds don't exist, why remove the option from *all* users for the sake of a minority? A simple warning to avoid the flag for the few setups where it causes trouble should suffice.

SuperIce97 commented on 2016-12-17 21:02

Is there anything wrong with the version of your patch before the last commit that removed the native option? If you choose native, it automatically gives you the option for P6 NOPS but it's disabled by default. I think that's a pretty safe way to do it. The help message also has the pretty straightforward message of "Say Y if you have Intel CPU newer than Pentium Pro, N otherwise." at the end, so I believe that anyone who enables Native optimizations would know what to do there.

For the issue with Atom CPUs having issues when set to native, you could just add "[causes issues on some Atom CPUs]" to the name of the native option. I don't think this needs to be too complex.

graysky commented on 2016-12-17 19:57

@SuperIce97 - I am not skilled enough to implement the logic for the X86_P6_NOP needed. While true that the native option may pass more tokens to make, if you look at the unpatched Makefile for a selection of core2, they are merely passing march=core2 option[1].

1. https://github.com/torvalds/linux/blob/master/arch/x86/Makefile#L110

SuperIce97 commented on 2016-12-17 19:53

Also, I noticed that on my SandyBridge laptop, march=native enables a ton more instructions than march=sandybrdige (my CPU is a 2630QM which is a SandyBridge). That means that we are not getting the full potential by specifying an arch vs native arch

SuperIce97 commented on 2016-12-17 19:40

Looking at issue #17 on the gcc patch's github page though, I don't think it would be that difficult to enable native optimizations and just have a X86_P6_NOP as an additional option like tpruzina suggested. The reason I prefer using native instead of the specific arch is because some lower end CPUs (at least with Intel) don't seem to actually support every instruction of their arch for some reason. I used to have a Chromebook C720 with the Haswell Celeron 2955u (I now have a C740 with a Broadwell Celeron 3205u) and I was using the Haswell kernel until it decided that it wouldn't boot (not sure if a change in the kernel or gcc caused it; it was a long time ago and I was too lazy to hunt down the issue) if I compiled it with Haswell. With native on the other hand, the kernel worked just fine. Would it be possible to reimplement march=native but have the extra configurable as suggested by tpruzina?

graysky commented on 2016-12-17 18:57

@nTia89 - I removed it as handling dependencies for newer/older hardware would be too difficult[1]... just select the hardware option that corresponds to your CPU[2] and functionally, there is no difference.

1. https://github.com/graysky2/kernel_gcc_patch/commit/84aad1d8bf5e0c2505b8fb9882394a262a7307e8
2. https://wiki.archlinux.org/index.php/Repo-ck#Selecting_the_correct_CPU_optimized_package

nTia89 commented on 2016-12-17 17:24

@graysky, how does new GCC patch affect "-march=native" users ?

graysky commented on 2016-12-17 16:20

Bump to v4.8.15-2
Changelog: Updated version of gcc patch
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2016-12-15 23:51

@Rainmaker - I like to use the official PKGBUILD as a template for this one since it minimizes both the chance of errors and my time to keep it in sync. Plus, I think many might not want to have the extra bulk of the headers package since it's only really needed to build modules. Not sure what you mean by split out the config.

@BSB6 - It's the holidays people travel... seems to get this this pretty consistently this time of year.

QuartzDragon commented on 2016-12-15 22:51

@BS86

So what? Perhaps there's a good reason for it?

QuartzDragon commented on 2016-12-15 22:50

@Rainmaker

I believe that the linux-ck PKGBUILD already does this?

BS86 commented on 2016-12-15 22:40

lol - arch still has only 4.8.13 in their official repositories, and no 4.9 - even Manjaro already has 4.8.14 and 4.9.0 in the stable repos -.-

Rainmaker commented on 2016-12-15 22:07

First of all, thank you for being very on top of all updates.

If I may suggest 2 improvements:
- You can build multiple packages in a single PKGBUILD. Because linux-ck and linux-ck-headers are likely to be installed together, may I suggest merging both packages to a single package.
Building multiple packages in the same PKGBUILD is something that, for example, clion does. https://aur.archlinux.org/cgit/aur.git/tree/PKGBUILD?h=clion.
Effectively, it is setting pkgname= to a space separated list.
- Could you split out the config? Each time I do a git pull, I need to open up PKGBUILD, and edit the parameters. By having a separate configuration file, I can setup git to ignore that file (with assume-unchanged), which saves me from having to set the parameters over and over.

graysky commented on 2016-12-15 18:57

Bump to v4.8.15-1
Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.8.15
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2016-12-12 11:42

@metaphorex0 - Not out-of-date until 4.9.0 is written.[1]

1. https://wiki.archlinux.org/index.php/Linux-ck#Release_cycle

PotHix commented on 2016-12-11 17:49

@graysky thanks for your reply. Just saw the solution on @nlopez's post:
gpg --recv-keys 79BE3E4300411886 38DBBDC86092693E

graysky commented on 2016-12-11 11:59

Bump to v4.8.14-2
Changelog: Bug patch fix following Arch upstream.
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2016-12-11 11:55

@PotFix - This is by default, see the official PKGBUILD.
@mareex - That was due to a diff in the md5sums of the local/remote packages. I didn't catch the error. Fixed now, thanks.

mareex commented on 2016-12-11 10:04

@graysky
What has happened with linux-ck-skylake from your repos? Maybe I am blind, but I cannot find it anymore (http://repo-ck.com/x86_64/).

PotHix commented on 2016-12-11 02:28

There is a problem with GPG signature verification

==> Validating source files with sha256sums...
linux-4.8.tar.xz ... Passed
linux-4.8.tar.sign ... Skipped
patch-4.8.14.xz ... Passed
patch-4.8.14.sign ... Skipped
patch-4.8-ck8.xz ... Passed
enable_additional_cpu_optimizations_for_gcc_v4.9+_kernel_v3.15+.patch.gz ... Passed
config.x86_64 ... Passed
config ... Passed
99-linux.hook ... Passed
linux.preset ... Passed
change-default-console-loglevel.patch ... Passed
net_handle_no_dst_on_skb_in_icmp6_send.patch ... Passed
==> Verifying source file signatures with gpg...
linux-4.8.tar ... FAILED (error during signature verification)
patch-4.8.14 ... FAILED (error during signature verification)
==> ERROR: One or more PGP signatures could not be verified!

graysky commented on 2016-12-10 18:51

Bump to v4.8.14-1
Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.8.14
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2016-12-08 20:43

Bump to v4.8.13-2
Changelog: Add net_handle_no_dst_on_skb_in_icmp6_send.patch from Arch upstream
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2016-12-08 20:29

Bump to v4.8.13-1
Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.8.13
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

SuperIce97 commented on 2016-12-08 20:06

@niceman The mainline kernel does include the msr module, as does basically every kernel in the main and AUR repos (including this one). If you enable nconfig for this kernel, you can see under "Processor Type and Features -> /dev/cpu/*/msr - Model-specific register support" that msr is indeed enabled as kernel module. Perhaps your device does not support msr? (I have an old Intel Atom machine that I use as a micro server that does not support msr and thus does not support i7z, which I would have liked to have been able to use)

niceman commented on 2016-12-08 19:17

hello, where can I find the msr module ? powertop seems to need it.

The mainline kernel doesn't seem to include it too, is the module built in the kernel ?

graysky commented on 2016-12-06 23:02

Bump to v4.8.12-2
Changelog: Security patch fix following Arch upstream.
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

zerophase commented on 2016-12-02 23:53

@mlc thanks. I just ended up rm -fing the misc file, and reinstalling.

graysky commented on 2016-12-02 19:41

Bump to v4.8.12-1
Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.8.12
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

mlc commented on 2016-11-30 12:40

@zerophase: you can firstly remove linux-ck and linux-ck-headers

$ sudo pacman -Rdd linux-ck linux-ck-headers

Then delete /lib/modules/4.8.11-1-ck/kernel/drivers/usb/misc if this file still exists and reinstall theses two packages:

$ sudo pacman -S linux-ck linux-ck-headers

It should be alright.

zerophase commented on 2016-11-30 10:10

@mlc think it's safe to delete that file on a running kernel?

mlc commented on 2016-11-30 05:46

@zerophase: no, it's a directory here.

zerophase commented on 2016-11-30 05:43

Is /lib/modules/4.8.11-1-ck/kernel/drivers/usb/misc installing as a file, rather than a directory for anyone else?

I tried reinstalling ck, and it still looks corrupted.

graysky commented on 2016-11-26 09:44

Bump to v4.8.11-1
Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.8.11
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2016-11-22 08:36

Bump to v4.8.10-2
Changelog: Bump to ck8 with MuQSS 0.144
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2016-11-21 20:17

Bump to v4.8.10-1
Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.8.10
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog
Note: CK will be releasing ck8 in several hours so I am holding off on building 4.8.10-1 for [repo-ck]. AUR users are of course free to do it.

graysky commented on 2016-11-20 14:34

Bump to v4.8.9-3

Changelog: Added a warning about initramfs rename for [repo-ck] users.
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2016-11-20 12:01

@snack - Yes, this naming scheme is the default now with upstream changes (hook and PKGBUILD). Since linux-ck-core2 from [repo-ck] has a pkgbase=linux-ck-core2, all users of [repo-ck] or of the AUR building a non-standard pkgbase will need to update their boot loader.

See: https://bbs.archlinux.org/viewtopic.php?pid=1670684#p1670684

If I was more savvy about this, I would have added a pacman warning to the post install, but even then, you only reach a subset of users.

snack commented on 2016-11-20 09:42

I fixed by removing all the old linux-ck files from /boot (keeping the linux-ck-core2 ones) and substituting "linux-ck" with "linux-ck-core2" everywhere in /boot/grub/grub.cfg

snack commented on 2016-11-20 09:31

Update to linux-ck-core2 4.8.9-1/2 from repo broke my system. Boot fails because my grub configuration (which used to work up to 4.8.6) searches for /boot/vmlinuz-linux-ck, while with 4.8.9 I got /boot/vmlinuz-linux-ck-core2.
Fixing it in the grub command line the boot fails again complaining that /lib/modules/4.8.9-2-ck-core2/modules.devname is missing, while the file itself is present. The boot media is an ext4 ssd.
Should I update something in my system configuration or is the linux-ck-core2 package broken?

QuartzDragon commented on 2016-11-20 04:18

@graysky

Now I understand better; I was just confused by the changes and how they worked.

Thanks, graysky! :)

graysky commented on 2016-11-19 19:57

Bump to v4.8.9-2

Changelog: Fix checksums and bump for repo rebuild.
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

sidneycrestani commented on 2016-11-19 19:21

@marceliq yes, patch-4.8.9.xz sha256sum fails

marceliq commented on 2016-11-19 14:53

Hi, is it just me or sha256sums of patch-4.8.9.xz doesn't match for you either?

sir_lucjan commented on 2016-11-19 14:13

@graysky - you have absolutely right.

graysky commented on 2016-11-19 12:32

@fafryd - I don't think it's a problem since the PKGBULD renames the hook on the filesystem. See my comment in FS#51818 for more.

graysky commented on 2016-11-19 12:27

@fafryd - Yes, linux-ck is shipping 99-linux.hook since 4.8.8-2. I posted in the bug report you referenced asking if the hook needs a unique name. Right now linux-ck shares the name with linux which might be a problem. We can make the change if needed. Thank you for bring this to my attention.

fafryd commented on 2016-11-19 12:03

>Every linux package (linux, linux-zen, linux-lts, linux-grsec, ... AUR packages >like linux-git, ...) has to ship its own hook.
https://bugs.archlinux.org/task/51818

edit: http://pastebin.com/DLTbkhZS

graysky commented on 2016-11-19 10:55

Bump to v4.8.9-1

Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.8.9
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2016-11-19 10:55

@QD - No sure what you mean. You mention errors but provide no output and you mention references but provide no line numbers. Finally, you mention larger scope that PKGBUILD but provide no examples... I can't really act on anything you're reporting.

If the references to the vanilla kernel are some variables in the PKGBUILD or linux.{preset,install} make sure you inspect them in a shell as many are substitutions that simply don't get substituted. For example, PKGBUILD line 262.

QuartzDragon commented on 2016-11-19 05:48

Hey graysky,

I was looking at the PKGBUILD, due to build errors, and there are many references that only apply to the vanilla Arch Linux kernel.

And it's not just the PKGBUILD that needs fixing... justing letting you know.

Cheers,
QD

graysky commented on 2016-11-17 23:46

Bump to v4.8.8-2

Changelog: Updates to PKGBUILD per Arch upstream.
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2016-11-17 23:36

Yep, already updated but was waiting for 4.8.9 to bump[1].
EDIT: here you go... better to hear about any bugs/errors before I do the repo rebuild so maybe this 24-36 h period will unearth some.

1. https://lkml.org/lkml/2016/11/17/287

SuperIce97 commented on 2016-11-17 22:15

The mainline kernel has some interesting changes with 4.8.8-2 and above. I'm not sure if these should be integrated into the ck kernel, but I would say it's worth a look (mainly some sed cleanups and using a new trigger to generate the initramfs as a post-transaction hook): https://git.archlinux.org/svntogit/packages.git/log/trunk?h=packages/linux

graysky commented on 2016-11-15 20:07

Bump to v4.8.8-1

Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.8.8
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

rob77 commented on 2016-11-15 09:53

@QuartzDragon didn't think of that as I have git cloned this packages URL, I just git pull and re build the package. I did build the package so my comment was an FYI for anyone else facing same issue. But when next update comes will delete the file I have locally and re download.

QuartzDragon commented on 2016-11-15 07:59

@rob77

Try deleting the file so it can be redownloaded? I did that just in case.

rob77 commented on 2016-11-15 00:25

sha256 check failed on enable_additional_cpu_optimizations_for_gcc_v4.9+_kernel_v3.15+.patch.gz had to amend line 74 of PKGBUILD to pass the error to cf0f984ebfbb8ca8ffee1a12fd791437064b9ebe0712d6f813fd5681d4840791. Alternatively double check the sha256sum by running sha256sum enable_additional_cpu_optimizations_for_gcc_v4.9+_kernel_v3.15+.patch.gz

graysky commented on 2016-11-12 10:36

Bump to v4.8.7-2

Changelog: Bump to ck7 with MuQSS 0.140
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

QuartzDragon commented on 2016-11-12 01:28

https://ck-hack.blogspot.com.au/2016/11/linux-48-ck7-muqss-version-0140.html

:)

graysky commented on 2016-11-11 20:52

Bump to v4.8.7-1

Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.8.7
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

mlc commented on 2016-11-10 10:44

@QuartzDragon & artafinde:
https://ck-hack.blogspot.fr/2016/10/linux-48-ck2-muqss-version-0114.html
=> "In addition to the most up to date version of MuQSS replacing BFS, this is the first release with BFQ included. It is configurable and is set by default in -ck though it is entirely optional".

QuartzDragon commented on 2016-11-10 09:27

@artafinde

I suspect so:
http://ck.kolivas.org/patches/4.0/4.8/4.8-ck6/patches/0012-BFQ-version-4.8.0-v8r4-by-Paolo-Valente.patch

I applaud Con for adding BFQ to the CK patchset! :D

Though, I wonder why he did it silently? Or was there an announcement somewhere that we missed?

artafinde commented on 2016-11-09 23:50

@graysky: after https://aur.archlinux.org/cgit/aur.git/commit/?h=linux-ck&id=8a91519b68ce6683a379ea8d37fed029d6de6fcf the BFQ patches are excluded. But the scheduler is still available. Is this now included with -ck patch or merged on main kernel?

graysky commented on 2016-11-07 20:08

@Tahu - Please report to CK's blog if you haven't already.

Tahu363 commented on 2016-11-07 14:43

Latest changes break wireless from working after suspend. Full reboot required to gain wireless functionality. ifconfig up/down has no effect.

graysky commented on 2016-11-05 19:10

Thanks updated without bumping the pkgver since purely cosmetic.

glitsj16 commented on 2016-11-05 19:05

Suggestions for your most recent PKGBUILD:

--- a/PKGBUILD
+++ b/PKGBUILD
@@ -118,7 +118,7 @@
# this as the default choice for it leads to more throughput and power
# savings as well.
#
- # http://ck-hack.blogspot.com/2016/10/linux-48-ck5-muqss-version-0120.html
+ # http://ck-hack.blogspot.com/2016/11/linux-48-ck6-muqss-version-0135.html
sed -i -e 's/^CONFIG_HZ_300=y/# CONFIG_HZ_300 is not set/' \
-i -e 's/^# CONFIG_HZ_100 is not set/CONFIG_HZ_100=y/' \
-i -e 's/^CONFIG_HZ=300/CONFIG_HZ=100/' .config
@@ -149,7 +149,7 @@
# modprobe configs
zcat /proc/config.gz > ./.config
else
- warning "You kernel was not compiled with IKCONFIG_PROC!"
+ warning "Your kernel was not compiled with IKCONFIG_PROC!"
warning "You cannot read the current config!"
warning "Aborting!"
exit

WFV commented on 2016-11-05 00:55

4.8.5-3 and 4.8.6-1 ck-piledriver, Virtualbox guests sluggish to the point unable to use them (and WinXP guest blue-screens @ launch). Problem isn't present in stock kernel 4.8.6-1, all guests function normal. (Asus M5A88-M, FX8350 AM3+). No other host problems noticed in either ck rev.
EDIT: problem persists in 4.8.6-2 ck-piledriver. Does it have to do with MuQSS and CGROUPS shared real time (for virtualbox)?

graysky commented on 2016-11-04 23:04

Bump to v4.8.6-2

Changelog: Bump to ck6 with MuQSS 0.135
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

zerophase commented on 2016-11-04 10:34

Definitely noticed MuQSS does a much better job of utilizing cpu resources than BFS. The only side effect I've noticed is if compiling large projects (Unreal Engine) The desktop has a tendency to become unresponsive.

graysky commented on 2016-11-01 17:12

@nTia - This sort of report is not very useful without a comparison back to the ARCH kernel of the same version at a minimum. As always, make sure that whatever you're experiencing is reproducible and not present on the non-ck kernel of the same version and report to CK's blog.

nTia89 commented on 2016-11-01 16:54

just a warning, on my system (dell xps 13, broadwell) latest version (4.8.6-1) is unstable
boot and shutdown slowed down, a bunch of errors about audio ASoC, etc...

graysky commented on 2016-10-31 18:44

Bump to v4.8.6-1

Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.8.6
Notes: This release also includes two of CK's pending patches per his recommendation as a ck6 is not on the nearterm horizon.
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

BS86 commented on 2016-10-30 14:28

no issues with disabled NUMA here on my Piledriver FX 8350 AM3+ CPU

SuperIce97 commented on 2016-10-30 04:11

@zerophase No, full dynticks is not recommended. I miscommented earlier. Also, I was referring to CPU accounting, not CPU ticks.

zerophase commented on 2016-10-30 04:04

@SuperIce97 Is Full dynticks still recommended?

graysky commented on 2016-10-29 15:53

@zosodk - I can mirror if needed but it's up for me at this point in time...

zosodk69 commented on 2016-10-29 14:05

Is there an alternative download source for patch-4.8-ck5.xz? There appear to be issues with ck.kolivas.org right now.

graysky commented on 2016-10-29 11:58

@Tjuh - Please post to CK's blog

Tjuh commented on 2016-10-29 11:51

I'm getting this msg in journallog with the latest version:

BFQ WARNING:last 4611686022722355494 budget 13835058059577131310 jiffies 4294967598 diff 4611686018427387904

Saren commented on 2016-10-29 11:15

ck.kolivas.org is down for everyone ATM

graysky commented on 2016-10-29 09:02

Bump to v4.8.5-3

Changelog: 100Hz is default by CK now so hardset this in PKGBUILD
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2016-10-29 09:02

@Saren - Correct: dual socket users will need to build from the AUR since NUMA is disabled by default (recommended by CK).

@QuartzD - Yes, it would be... I currently build 20 different package sets for [repo-ck] and toggling this on/off would double that to 40 and only serve to confuse users. After all, what percentage of [repo-ck] users have a dual socket motherboard?

@SuperIce and @QuartzD - Yes, corrected in 4.8.5-3.

QuartzDragon commented on 2016-10-29 08:13

Would it be a hassle to maintain NUMA and non-NUMA versions in Repo-CK?

Saren commented on 2016-10-29 06:08

With NUMA disabled by default, as a dual socket workstation user, does it mean I have to recompile kernel every time when upgrading it?

QuartzDragon commented on 2016-10-29 05:04

It's being set to 100 Hz, it's just that the "_1K_HZ_ticks" variable is confusing. Should be renamed to "_100_HZ_ticks", now, to reflect the changes.

Just guessing you missed it somehow, graysky?

SuperIce97 commented on 2016-10-29 04:46

Can you set the ticks to 100Hz as recommended by ck now?

graysky commented on 2016-10-29 01:19

@vishwin - http://ck-hack.blogspot.com/2016/10/linux-48-ck5-muqss-version-0120.html

vishwin commented on 2016-10-29 00:28

ck in his -ck4 announcement post said that he's cutting -ck5 and MuQSS 120 today. They have just been tagged on his github and I bet he's prepping the tarballs as I type…

mrkline commented on 2016-10-29 00:18

I'm late to the party, but I've run with NUMA disabled since I started using this package (4+ months?). Zero issues on all my machines (Skylake, Haswell, and Ivy Bridge).

graysky commented on 2016-10-28 18:08

Bump to v4.8.5-1

Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.8.5
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2016-10-27 19:15

Thanks for the comments, all.

Saren commented on 2016-10-27 13:58

Dual haswell xeon cpu. 4.8.4-4, NUMA and MuQSS 116 enabled. 0 issue so far and great performance.

artafinde commented on 2016-10-27 07:04

I was one of the persons who had issues with NUMA on an AMD FX 8100. About 6 months ago I changed to i7 6700 and disabled NUMA. No issues so far. I did had btrfs corruption with linux-ck early MuQSS versions (v110 I think) so I was a late adopter but now running pretty stable for days. I think the NUMA issue were related to AM3+ socket.

mlc commented on 2016-10-27 04:28

I'm running linux-ck with NUMA disabled since about two weeks. I haven't yet experienced any problems with my Skylake CPU.

graysky commented on 2016-10-26 21:16

@QD - It's a small increase using a make endpoint in my experience (see the flyspray I note in the PKGBUILD comments). It could be that other endpoints reveal more substantial gains as well. In any case, NUMA is really for servers with multiple sockets; it has no point on a single socket motherboard is my understanding. Linux-ck had it disabled for a long time until some combination of upstream/BFS + NUMA disabled was believed to be responsible for problems which is when I disabled the code to disable it. Now that MuQSS has replaced BFS, this may not be the case any more. Several users have posted to the AUR reporting stability with it disabled as is my experience as well.

QuartzDragon commented on 2016-10-26 20:50

How much of a speed increase does disabling NUMA actually give, anyway? I've never really felt the difference.

graysky commented on 2016-10-26 19:17

Greg just tagged 4.8.5-rc1 and has it scheduled to release on Friday/noon UTC[1]. We can give disabling NUMA a whirl in 4.8.5-1 if there are no reports linking this setting to bad behavior with MuQSS (I am not experiencing anything bad on my Haswell).

1. https://lkml.org/lkml/2016/10/26/646

metaphorex0 commented on 2016-10-26 10:45

I've been running linux-ck with NUMA disabled since you added that option back to PKGBUILD. I've had 0 issues so far.

graysky commented on 2016-10-25 15:49

I can but I would like to get some additional feedback from others who build with NUMA disabled since it caused problems in the past. I too have it disabled on my workstation without ill effects. Can other AUR users please comment. If you have NUMA disabled, are you running a stable 4.8.4-4?

mareex commented on 2016-10-25 15:45

hi graysky, would you mind to disable NUMA in your repo builds? Never had any problems on four machines with NUMA disabled.

graysky commented on 2016-10-25 10:28

@mono - Please report upstream: http://ck-hack.blogspot.com/2016/10/linux-48-ck4-muqss-cpu-scheduler-v0116.html#comment-form

monotykamary commented on 2016-10-25 06:42

No more complete freezes, but it now slows to a crawl on the same application.

graysky commented on 2016-10-24 19:33

Bump to v4.8.4-4

Changelog: Update with ck4/MuQSS v0.116. Updates to [repo-ck] are building now.
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

eduardoeae commented on 2016-10-24 10:11

linux-4.8-ck4, MuQSS CPU scheduler v0.116

https://ck-hack.blogspot.com.ar/2016/10/linux-48-ck4-muqss-cpu-scheduler-v0116.html

graysky commented on 2016-10-23 16:53

@mono - the difference between AUR and [repo-ck] right now is that AUR uses 4.8.4-3 which has ck3 (MuQSS) whereas [repo-ck] uses 4.8.4-1 with ck1 (bfs). I have a dual core system that black screened on booting with ck3 last I checked but there are two patches in 'pending' that I need to try first. I think if they fix that problem, I will go ahead and build 4.8.4-3 for [repo-ck] but I won't be able to test until tomorrow.

monotykamary commented on 2016-10-23 16:12

There's a bug that completely freezes my computer when opening or playing OpenGL games for a period of time (with NVIDIA), from the linux-ck-ivybridge package group.
The AUR linux-ck package doesn't have that bug, but now specifically crashes osu! on wine everytime upon load (or a few seconds after load). I haven't seen it crash on any other native or wine application so far.

---
WINEARCH=win32 WINEPREFIX=~/.wineprefixes/.osu winetricks -q dotnet46 cjkfonts gdiplus wininet winhttp
osu! folder size: 3.9G

Installed Packages: linux-ck, linux-ck-headers, nvidia-ck, broadcom-wl-ck
BFQ and MuQSS enabled; bug also occurs with I/O scheduler CFQ.

CPU: Intel Core i5-3317U CPU @ 1.7GHz
GPU: GeForce GT 640M LE

kogone commented on 2016-10-23 00:34

everything is been stable on ivybridge. have built every update so far! hard to keep up =P

graysky commented on 2016-10-22 23:59

Bump to v4.8.4-3

Changelog: Added two key pending patches from CK.
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog
Notes: I suspect CK will ready a ck4 patchset and continue to hold on pushing MuQSS to [repo-ck] a little longer. AUR users are encouraged to test and communicate via CK's blog results as MuQSS matures.

graysky commented on 2016-10-22 19:51

Bump to v4.8.4-2

Changelog: Fix EXTRAVERSION variable.
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog
Notes: I would like to get some feedback from AUR users around the stability with the MuQSS before I go building [repo-ck] packages with 4.8-ck3. Please test and report back good or bad.

graysky commented on 2016-10-22 19:48

@blitz - Thanks for the reminder... accidentally deleted that code.

blitz commented on 2016-10-22 19:11

linux-4.8 ck patchset introduced new versioning schema:
patch-4.8-ck3
diff --git a/Makefile b/Makefile
+CKVERSION = -ck3
+CKNAME = MuQSS Powered
+EXTRAVERSION := $(EXTRAVERSION)$(CKVERSION)

From kernel version '4.8' patch '4' extraversion '1' this schema translates into 4.8.4-1-ck3-ck

To keep both ck3 and ck is extraneous and not nesessary, imho.

# set extraversion to pkgrel
# remove ckX version
msg "Setting kernel extra version"
sed -ri -e "s|^(EXTRAVERSION =).*|\1 -${pkgrel}|" \
-e "s|^(EXTRAVERSION :=).*|\1 -${pkgrel}|" Makefile

Tjuh commented on 2016-10-22 18:02

I dunno yet, keep having mini system freezes when playing certain games on steam with the 4.8 series. No issues whatsoever with 4.7.8.

FarStar commented on 2016-10-22 13:39

Everything is behaving fine with linux-ck 4.8.4 and MuQSS v0.115 (like v0.114). My CPU is an i7-6700K.

nTia89 commented on 2016-10-22 12:40

if with the first release of the 4.8 series, now (4.8.3-3) my system (Dell XPS 13, Broadwell i5-5200U) boots fine and works as in the past days

BS86 commented on 2016-10-22 12:39

No issues here with MuQSS on my piledriver. on 4.8.4-1

I am trying out some advanced amdgpu package, which checks for a specific config. Would it be much effort to include it, so that this check does not fail?
Kernel 4.8.4-1-ck3-ck not supported
CONFIG_DRM_AMDGPU_CIK is missing

graysky commented on 2016-10-22 11:24

Bump to v4.8.4-1

Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.8.4
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog
Notes: I would like to get some feedback from AUR users around the stability with the MuQSS before I go building [repo-ck] packages with 4.8-ck3. Please test and report back good or bad.

mlc commented on 2016-10-22 08:56

Hello graysky,
I use linux-ck with BFQ and MuQSS enabled since v106. It works very fine with my Skylake CPU. Performances and responsiveness are great with v114/115 :)
Thank you for all!

graysky commented on 2016-10-22 07:01

Bump to v4.8.3-3

Changelog: Included 4.8-ck3 which uses MuQSS v0.115.
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog
Notes: I would like to get some feedback from AUR users around the stability with the MuQSS before I go building [repo-ck] packages with 4.8-ck3. Please test and report back good or bad.

vishwin commented on 2016-10-22 03:21

-ck3 just released, includes MuQSS v0.115

http://ck-hack.blogspot.com/2016/10/linux-48-ck3-muqss-version-0115.html

graysky commented on 2016-10-21 19:22

Bump to v4.8.3-2

Changelog: Included 4.8-ck2 which uses MuQSS v0.114.
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog
Notes: I would like to get some feedback from AUR users around the stability with the MuQSS before I go building [repo-ck] packages with 4.8-ck2. Please test and report back good or bad.

graysky commented on 2016-10-21 13:21

Please post these details to ck's blog so he can assist. http://ck-hack.blogspot.com/2016/10/linux-48-ck2-muqss-version-0114.html#comment-form

xevrem commented on 2016-10-21 12:17

I'm experiencing a kernel panic shortly after boot. Currently I am using the linux-ck-piledriver kernel.

from journalctl -k -b -1:

Oct 21 07:03:49 vanir kernel: AMD-Vi: Completion-Wait loop timed out
Oct 21 07:03:49 vanir kernel: AMD-Vi: Event logged [IOTLB_INV_TIMEOUT device=07:00.0 address=0x000000040cfa2380]
Oct 21 07:03:49 vanir kernel: AMD-Vi: Completion-Wait loop timed out

in this log its complaining about device 07:00.0, but its complained about other devices as well that work perfectly fine in other kernels... I would do a dump of those dmesg logs if i could, but it appears journalctl didnt capture anything from those boot cycles... only that one and it never catchs the kernel stack trace dump >_<

here is info about my CPU from /proc/cpuinfo if that helps:

processor : 0
vendor_id : AuthenticAMD
cpu family : 21
model : 101
model name : AMD FX-9800P RADEON R7, 12 COMPUTE CORES 4C+8G
stepping : 1
microcode : 0x6006113
cpu MHz : 1400.000
cache size : 1024 KB
physical id : 0
siblings : 4
core id : 0
cpu cores : 2
apicid : 16
initial apicid : 0
fpu : yes
fpu_exception : yes
cpuid level : 13
wp : yes
flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc rep_good acc_power nopl nonstop_tsc extd_apicid aperfmperf eagerfpu pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs xop skinit wdt lwp fma4 tce nodeid_msr tbm topoext perfctr_core perfctr_nb bpext ptsc mwaitx cpb hw_pstate vmmcall fsgsbase bmi1 avx2 smep bmi2 xsaveopt arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic overflow_recov
bugs : fxsave_leak sysret_ss_attrs null_seg
bogomips : 5392.44
TLB size : 1536 4K pages
clflush size : 64
cache_alignment : 64
address sizes : 48 bits physical, 48 bits virtual
power management: ts ttp tm 100mhzsteps hwpstate cpb eff_freq_ro acc_power [13]

It should be said, i do not experience these issues with stock kernel nor the zen kernel. I've also tried the generic linux-ck kernel and experience similar issues...

Any suggestions or additional info i can grab for you that would help, just let me know :)

Taijian commented on 2016-10-21 11:46

I have a question regarding some modules I am missing when I try to compile the kernel locally. These modules also fail to load with the repo-ck 4.8.3 kernel. The 4.7.6 kernel from repo-ck still had them.

Specifically, I have a Dell Inspiron laptop, so I am using i8kutils to get my temp-sensors and my fan to work. According to modprobed-db, this is pulling in the following modules:
- dell_laptop
- dell_smbios
- dell_smm_hwmon
- dell_wmi

However, when trying to build the kernel, these modules are missing/nowhere to be found. Can anybody help point me in the right direction, where I can pull these in from?

ooo commented on 2016-10-21 10:56

@graysky, Fair enough. Altough The pending/ patches for bfs512 aren't really 'development' patches, but fixes to bugs that were discovered after bfs512/ck1 release. My understanding is that ck simply didn't want to make a new release, but still recommends adding the fixes. Since some people seem to be having issues with bfs512, it would make sense to me to apply those.

I have no problems myself, but I guess anyone having issues with linux-ck-4.8* should try if adding the pending bfs512 patches helps.

nTia89 commented on 2016-10-21 10:09

@SuperIce97 I posted here just to advice other users...

graysky commented on 2016-10-21 09:41

@ooo - I don't want to include pending patches without CK specifically asking since this is not a development package. You are free to modify the PKGBUILD on your own of course.

@eduardoeae - I saw the release but am a bit reluctant to bump and push to [repo-ck] since a few reports to the blog you linked indicate issues still. Perhaps the AUR can have it but I will hold off on publishing to the repo initially.

eduardoeae commented on 2016-10-21 01:33

-ck2 is out. I'm not flagging as out of date because MuQSS is now the default. Probably would want to wait a bit before jumping into it.

https://ck-hack.blogspot.com.ar/2016/10/linux-48-ck2-muqss-version-0114.html

ooo commented on 2016-10-20 22:55

How about including the pending patches for bfs512? They contain bugfixes, and are recommended by ck.
http://ck.kolivas.org/patches/bfs/4.0/4.8/Pending/

rob77 commented on 2016-10-20 21:48

linux-ck 4.8.3-1 working fine here, with MuQSS enabled, BFQ enabled and compiled with intel haswell optimizations.

graysky commented on 2016-10-20 20:26

Please report issues with the patchset to CK via his blog. I am just the packager :p

initnull commented on 2016-10-20 19:34

I get a Black screen on boot with repo-ck/linux-ck-skylake 4.8.2-1.
Setup is Kaby Lake processor with /boot and /home crypted

stock 4.8.2-1-ARCH works fine.


graysky commented on 2016-10-20 18:50

Bump to v4.8.3-1

Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.8.3
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

SuperIce97 commented on 2016-10-20 17:10

Everybody who is having problems: Have you tried running the stock 4.8.2 kernel from the main Arch repos? If that has problems as well, there's nothing that we can do.

Also, if you are running a 4.7 kernel, please update immediately to 4.7.9. A large security vulnerability was found and fixed. The exploit using this vulnerability has been found in the wild (and the bug has existed for 11 years). For people running 4.8.2, I expect @graysky to bump the version soon. Please update immediately when he does.

http://www.phoronix.com/scan.php?page=news_item&px=Linux-Updates-CVE-2016-5195

Saren commented on 2016-10-20 15:33

Getting boot freeze on my haswell-ep system as well since 4.8.2-1 with muqss on, no problem if no muqss.

nTia89 commented on 2016-10-20 14:01

I've just updated to ck-4.8.2 and my system (Dell XPS 13, broadwell 9343) stucks on boot (during the systemd boot process...)

the error is related to "i2c designware..."

going back to latest 4.7

NB: no extra options enabled from the AUR package, just this boot parameters: elevator=bfq and nmi_watchdog=0

bacondropped commented on 2016-10-20 13:02

Jesus, @graysky, you're rolling this package out faster than I update my system. Immense props for that.

rob77 commented on 2016-10-20 09:59

@Tjuh http://forum.slitaz.org/topic/error-unable-to-locate-ioapci-for-gsi

It's an error message introduced in Linux 2.6.35 and up. Basically it happens when your system doesn't support APIC. It's a harmless error, but doesn't look all that nice.

You can make it go away by adding noapic to the end of your grub boot line.
Adding it from Tazpanel:
1. Boot--> Boot Loader
2. Click on View or edit menu.lst
3. Click Edit
4. Add noapic to the end of the kernel line so it looks something like this:
kernel /boot/vmlinuz-2.6.37-slitaz root=/dev/hda1 quiet noapic
5. Click Save and you're done.

The error should be gone the next time you boot

Not upgraded kernel yet but will check to see if that fixes it this evening

Tjuh commented on 2016-10-20 09:27

Getting this weird error -ERROR: Unable to locate IOAPIC for GSI 37- in my journallogs with MuqSS, 1k_HZ_ticks, NUMAdisable, localmodcfg and BFQ_enable enabled. No such error or any other issues with the same settings enabled on the previous version.

zerophase commented on 2016-10-20 08:56

@VerruckteFuchs I'm on Haswell as well. I've been up for close to 8 hours without issue, compiling projects and running services that stress the cpu.

I don't run mpd, though. Have you noticed any issues in terms of cpu load and the crash?

snack commented on 2016-10-20 07:38

It seems that virtualbox-ck-host-modules-core2 (and possibly other virtualbox-ck-host-modules packages) in repo have been built with a wrong dependency from linux-ck-core2 version since pacman says:

$ sudo pacman -Su
:: Starting full system upgrade...
resolving dependencies...
looking for conflicting packages...
error: failed to prepare transaction (could not satisfy dependencies)
:: virtualbox-ck-host-modules-core2: installing linux-ck-core2 (4.8.2-1) breaks dependency 'linux-ck-core2<4.8'

VerruckteFuchs commented on 2016-10-20 01:37

Trying to run mpd results in an instant freeze/lockup of my Haswell machine. I do have a previous playlist queue from when it was last open, if that info helps at all.

graysky commented on 2016-10-19 20:30

Bump to v4.8.2-1

Changelog: bfs512/ck1, upstream release 4.8.2, and optional (AUR only) MuQSS 0.112
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog
Linux changes: https://kernelnewbies.org/Linux_4.8

graysky commented on 2016-10-16 16:53

Bump to v4.7.8-1

Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.7.8
Commit: https://wiki.archlinux.org/index.php?title=Linux-ck/Changelog

graysky commented on 2016-10-15 14:32

@Saren - Once CK incorporates MuQSS into the ck1 patchset, yes. Right now it is considered in development, so AUR users can build it if they wish, but I don't want to server-up experimental code in [repo-ck].

Saren commented on 2016-10-15 14:23

@graysky Will packages on repo-ck adopt MuQSS updates as well?

graysky commented on 2016-10-13 22:07

4.7.7-1 has been updated with MuQSS 111.

graysky commented on 2016-10-11 18:56

4.7.7-1 has been updated with MuQSS 110.

graysky commented on 2016-10-09 16:09

@agm28011997 - CK has stuff in his pending queue beyond MuQSS108 but I don't want to add continually add these since the rate of development of MuQSS is as rapid as it has been. Once MuQSS finalized and replaces BFS, this package will be here applying it.

Saren commented on 2016-10-08 07:23

Noticeable performance loss... compared to 4.7.1-ck since BFS 8...
https://browser.geekbench.com/v4/cpu/compare/677991?baseline=649259

pedrogabriel commented on 2016-10-08 02:14

It's working nicely with or withouth the MuQSS, I'm actually not using MuQSS because it someway brokes the fan driver of my notebook (with or withouth NUMA), but exceptly that, it's working nicely.

graysky commented on 2016-10-07 17:35

Bump to v4.7.7-1

Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.7.7
Commit: https://wiki.archlinux.org/index.php?title=Linux-ck/Changelog

graysky commented on 2016-10-07 17:34

@Roberth - Not out of date until BFQ is ready for 4.8

graysky commented on 2016-10-07 15:31

Good to hear, pedro... this is with MuQSS disabled, correct? I would like to hear from more users before I enable it globally again (and in the repo).

pedrogabriel commented on 2016-10-07 15:27

I'm using it with NUMA disabled, no problems so far. (Intel Haswell chip)

zerophase commented on 2016-10-07 05:07

@graysky those nvme patches are listed on the archwiki for fixing an issue with powersaving with nvme disks.

I'm just trying to figure out if I should add those patches to the pkgbuild when I pick up an nvme drive in a couple weeks. I'll at least test to see if there are any incompatibility issues. If they work without issue, should they get added into the ck repo till reaching mainline?

https://wiki.archlinux.org/index.php/Solid_State_Drives/NVMe

graysky commented on 2016-10-07 00:07

@jwhickman - Dunno about NUMA. Have to ask some users to test and verify that it is or is not a problem. Anyone willing to test? Also, I don't think MuQSS is stable at this point. We'll probably have BFS until the 4.9 release if I read CK's blog correctly. In either case, the linux-ck brand is here to stay; MuQSS will still have the ck1 brand as I understand it.

jwhickman commented on 2016-10-06 12:52

Thanks @graysky! I modified this to build MQS for k4.8; no problems besides having to comment out the BFQ patches due to a compile failure, not unsurprisingly. I was hoping the k4.7 BFQ patches would work on k4.8, since the patches applied without issue. :)
Question for you; is the 'NUMA disable issue' still relevant, as it's been around a long time, plus now this is MQS. Also, I noticed in my k4.8 build, the config showed a new tick rate of 250 Hz, so was also curious about that.
/=======================================================
# Running with a 1000 HZ tick rate (vs the ARCH 300) is reported to solve the issues with the bfs/linux 3.1x and suspend. [...]
_1k_HZ_ticks=y
### Do not disable NUMA until CK figures out why doing so causes panics for some users! [...]
#_NUMAdisable=
=======================================================/
In commented original discussion on the tick rate, ck has said, "(16 September 2013 at 18:55) If it still needs 1000Hz to work correctly, then there is still a bug." Anyway, thanks again!

graysky commented on 2016-10-06 09:05

OK... unofficial patches. I was wondering if something was in the merge queue on the topic that I didn't see.

zerophase commented on 2016-10-06 09:03

@graysky These patches https://github.com/damige/linux-nvme

graysky commented on 2016-10-06 08:58

What patches are you referring to?

zerophase commented on 2016-10-06 03:15

Anyone know if the Linux-nvme kernel patches have any issues with ck? Going to be picking up an nvme drive in a couple week and just want to check if anyone knows of any issues.

graysky commented on 2016-10-05 18:09

@Pedrogabriel - Yes, if you enable the build option. I just updated 4.7.6-1 to pull in MuQSS v106 (disabled by default).

pedrogabriel commented on 2016-10-05 15:45

Does MuQSS 106 apply to 4.7?

graysky commented on 2016-10-04 22:12

105 is out but it's jacked up and does not cleanly apply to either 4.7 or 4.8. See CK's latest blog. I have unflagged; please flag again when the 105 is updated.

pedrogabriel commented on 2016-10-04 22:00

MuQSS 105 is out now.

pedrogabriel commented on 2016-10-04 13:27

The kernel don't recognize my fans after compile it with MuQSS.

graysky commented on 2016-10-03 18:36

@vishwin - Sorry, but not out of date until BFQ for 4.8 is released. I do not see MuQSS104 that patches against the 4.7 code-base (although CK is working on 105 as we speak)... plus, 4.8-ARCH is probably still in the works as it hasn't been build nor hit [testing] just yet.

artafinde commented on 2016-10-03 13:45

I actually won't be enabling this unless it turns out to be stable enough and give something to the <= 8 CPUs. I've enabled it on my laptop and screwed up the BTRFS ctree. Seems like I'm reformatting today. I can't really blame MuQSS since there's a big WARNING on top, but it's a warning to users here.
Could be something else but all started after I compiled and booted the MuQSS enabled kernel.

The combination was: CPU Haswell, BTRFS single metadata SSD 1 device, BFQ enabled by default, localmodules enabled and MuQSS.

kogone commented on 2016-10-03 13:40

maybe we should create a rubric of tests to run to see the difference of muqss running on cpus with <= 8 cores?

artafinde commented on 2016-10-02 09:26

According to the blog post CK mentions only CPU not cores. I'm not entirely sure how the runqueues are defined (per CPU or per Core) but my understanding from the post is to target servers and make the BFS more comparable to mainline CFS.

zerophase commented on 2016-10-02 09:05

@artafinde Does MuQSS benefit more than 1 cpu die, or multiple cores? Say 8+?

artafinde commented on 2016-10-02 08:58

MuQSS is really targeting multi CPUs. As per the comment from CK "it will make a difference in 16+ CPUs and with high loads". I suppose we can still use it for testing purposes but unlikely we will produce any lockups to help surface any bugs with the desktop/laptop PCs. 103 is out btw.

graysky commented on 2016-10-01 20:05

@SuperIce - Thanks updated to 102 still in 4.7.6-1.

SuperIce97 commented on 2016-10-01 17:14

@graysky Version 102 is out now as well. According to ck: "I've uploaded version 102 which has a a lot of bugfixes and makes it boot properly again for me."

graysky commented on 2016-10-01 15:39

@SuperIce97 - Thanks for the heads-up about version 101 of muqss. Updated without bumping the pkgver since so few people are actually building and since the repo packages do not use this and I don't want to trigger a rebuild.

graysky commented on 2016-10-01 09:39

Bump to v4.7.6-1

Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.7.6
Commit: https://wiki.archlinux.org/index.php?title=Linux-ck/Changelog
Note: This release allows AUR users to optionally build CK's MuQSS (Multiple Queue Skiplist Scheduler)[1]. See the commented lines in the PKGBUILD.

1. http://ck-hack.blogspot.com/2016/10/muqss-multiple-queue-skiplist-scheduler.html

Saren commented on 2016-10-01 04:38

No freeze observed so far since v4.7.5-1.

graysky commented on 2016-09-24 09:45

Bump to v4.7.5-1

Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.7.5
Commit: https://wiki.archlinux.org/index.php?title=Linux-ck/Changelog

Rainmaker commented on 2016-09-23 22:40

@graysky. You're right.

makepkg only builds once.
makepkg -C DOES build twice.

But that behavior has nothing to do with this package specifically. Thanks.

graysky commented on 2016-09-23 08:28

Bump to v4.7.4-2

Changelog: Update to new BFS[1]
Commit: https://wiki.archlinux.org/index.php?title=Linux-ck/Changelog

1. http://ck-hack.blogspot.com/2016/09/bfs-502-linux-47-ck5.html

zerophase commented on 2016-09-19 23:01

@graysky Is the extra patch in the pending directory getting applied? I guess that's supposed to fix the problem with Haswell.

Saren commented on 2016-09-19 16:27

@zerophase That's why I am staying on 4.7.1-1.

zerophase commented on 2016-09-18 23:56

Definitely, getting those crashes on Haswell E as well. The only thing I noticed in the logs was:

ENERGY_PERF_BIAS: Set to 'normal', was 'performance'

Could cpupower not be setting the cpu governor correctly? I also get that same message while running the stock kernel.

graysky commented on 2016-09-18 13:12

Makepkg does not build twice, it packages linux-ck then linux-ck-headers. Are you sure you are not mischaracterizing?

Rainmaker commented on 2016-09-18 12:51

Well, I'm using makepkg.

I have a local directory on which I do a git pull every time there is an update.

But, I am doing makepkg -C. Maybe I should try cleaning up the previous source manually before building.

Currently, it DOES do two builds. It finishes building linux-ck, and then proceeds with unpacking linux-ck-headers and compiling that. It does warn about the PKGBUILD being a split package.

When using yaourt, it does built twice, but I guess that is to be expected.

graysky commented on 2016-09-18 12:45

@Saren - Thanks for clarify.
@Rainmaker - As suggested by others, don't use an aur helper. Just build with makepkg and you'll be fine (ie one build, multiple packages). Which one are you using by the way? The fact that is build twice is just wrong.

francoism90 commented on 2016-09-18 10:29

@Saran: How do you build this package? By an AUR wrapper (e.g. pacaur/yaourt)?
If so, please just download/unpack the tar and run makepkg instead.

I don't have any issues with multiple PKG builds.

Saren commented on 2016-09-18 07:33

@graysky I think you got it wrong. What @Rainmaker said is when we build linux-ck and linux-ck-headers packages, the kernel will be compiled twice, which is unnecessary and waste of time and power.

When we build linux-ck-headers, linux-ck will be also compiled and provided. We can just accept installing both packages instead of linux-ck-headers only when pacman prompt us.

graysky commented on 2016-09-16 18:44

The headers package has nothing to do with it; edit /etc/mkinitcpio.d/linux-ck.preset and disable the fallback image.

Saren commented on 2016-09-16 13:42

@Rainmaker Install linux-ck-headers only.

Rainmaker commented on 2016-09-16 12:01

When you install linux-ck and linux-ck-headers, the kernel gets built twice.

Is there a way to only build it once?

graysky commented on 2016-09-15 22:16

Disable ccache....

artafinde commented on 2016-09-15 22:07

@rko same thing to me with ccache had to clean. Seems like it's a pattern now (build failed due to object fixdep-in.o again) :(

jschorr commented on 2016-09-15 20:12

4.7.3-5 (linux-ck-headers-4.7.3-5 w/ linux-ck-piledriver-4.7.3-5) fail to boot for me- it just freezes partway through so I had to go back to 4.7.3-4. I have an AMD 9590 CPU.

graysky commented on 2016-09-15 19:05

Bump to v4.7.4-1

Changelog:https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.7.4
Commit: https://wiki.archlinux.org/index.php?title=Linux-ck/Changelog

boteium commented on 2016-09-15 17:48

Both my baytrail z3775 (silvermont) and E3 1265l v3 (haswell) freeze on 4.7.3.

artafinde commented on 2016-09-15 12:54

I just want to point out that on my Skylake i7 6700 I had no panics for some time now and none with 4.7.x Could be only related to older hardware?

agm28011997 commented on 2016-09-15 12:21

please post problems here http://ck-hack.blogspot.com/2016/09/bfs-497-linux-47-ck4.html

kogone commented on 2016-09-15 11:36

freeze happening with my Silvermont on 4.7.3-5 but no problems with my Ivybridge.

Saren commented on 2016-09-15 11:26

I compiled linux-ck 4.7.3-5 from AUR last night with native optimizations as set in nconfig. I still got kernel panic when I was playing games inside VM.

I am using 4.7.1-1 for now and see if crashes still occur.

EDIT: Crashes seem to no longer occur at least for now after 3 hours consecutive VM gaming.

zerophase commented on 2016-09-15 08:49

@agm28011997 Odd, I'm not getting any freezes with Haswell-e that I believe are caused by the kernel. I just freeze when I have CLion, and a Windows vm up. (I give that up to changing the CLion vm options for parsing a large CMake file)

I'm pretty sure I didn't use up all of my ram or swap. Maybe, it is the kernel.

evilpot commented on 2016-09-15 08:06

its crashing all over the place...

agm28011997 commented on 2016-09-14 17:04

incredible man, I don't know what is it the difference or what have the new code but in my i5 4690 haswell with intel hd graphics from git all is perfect with one hour of gaming now and any freeze, it is something weird because it is supposed to be here yet the freezes

rko commented on 2016-09-14 16:55

Random Comment:
I was getting build failures with this, and it turns out that I needed to clear my ccache to get the kernel to build.

graysky commented on 2016-09-14 13:14

Even though 4.7.4 is due out today, I have built and published 4.7.3-5 and all related packages to [repo-ck]. CONFIG_CGROUP_SCHED is enabled per CK's suggestion.

agm28011997 commented on 2016-09-13 22:55

@gravysky thanks for your work, both of you, I wish I could help more than testing packages, when you put the new version in the repo I will test for report the issues, but I think that the freezes will continue due to the other comment in the page..
@saren what is BTW? I have seen the journalctl with the hour of freezes and nothing.. the freeze make my computer to do nothing and repeat 2 last seconds of music more or less and the mosue to not move and the keyboard the same

graysky commented on 2016-09-13 19:40

Bump to v4.7.3-5

Changelog: update to ck4/bfs497 and reenable CONFIG_CGROUP_SCHED.
Commit: https://wiki.archlinux.org/index.php?title=Linux-ck/Changelog
Note: I am waiting to hear back from CK before I build the repo packages since he may suggest disabling CONFIG_CGROUP_SCHED. I like to think that the Arch community plays a large role in helping CK to refine bfs so AUR users please continue provide feedback to ck's blog as usual: http://ck-hack.blogspot.com/2016/09/bfs-497-linux-47-ck4.html

Saren commented on 2016-09-13 19:38

@agm28011997 Haswell-EP is the xeon version of Haswell cpus, linux-ck-haswell should be 100% compatible with it. I am downgrading once again from 4.7.2-2 to 4.7.2-1 because of your comment. :/

BTW, can confirm the lockup is kernel panic.

agm28011997 commented on 2016-09-13 19:30

@saren what is haswell-EP? my freezes starts at 4.7.2-2 with the bfs 472 patch, and other thing, post your problem in the ck blogspot of con kolivas
https://ck-hack.blogspot.com.es/
@th3voic3 other like you it is me, if you want you can downgrade the version to 4.7.2-1 with patch bfs472 that is very stable for me
@gravysky try to use cgroups enable if you want because for haswell desktop users there are too loockups for use this kernel, we'll wait until a solution

Saren commented on 2016-09-13 19:24

Haswell-EP user here, using linux-ck-haswell. Getting 3 random lockups already since 4.7.3. 2 of the lockups in the GPU Passthrough gaming, and the remaining one happened while watching youtube videos.

I am downgrading to 4.7.2 as suggested by the comments and see if the issue will be going away.

th3voic3 commented on 2016-09-13 19:09

I'm using a Haswell cpu and a qemu vm with PCI passthrough.
With the recent kernels my entire host would lockup as soon as I shutdown my VM.
I just compiled the kernel with the new patch version and enabled
CONFIG_CGROUP_SCHED.

Just had a lockup again. Although I had lockups with or without
CONFIG_CGROUP_SCHED. Haven't tested this patch yet with the option disabled.

For now I'm using the mainline kernel. No issues there.

graysky commented on 2016-09-13 19:04

Just waiting to hear back from CK regarding his recommendation to or not to disable CONFIG_CGROUP_SCHED with ck4/bfs497. Will update once this happens.

puithove commented on 2016-09-13 18:50

I'm on a SandyBridge machine and using the SandyBridge package from the repo.

I ended up having the same complete system freezes as well as KP on shutdown/reboot even after updating to 4.7.3-4

As of yesterday evening I rolled back to 4.7.2-1 and I seem to be stable again.

agm28011997 commented on 2016-09-12 12:44

CONFIG_CGROUP_PIDS=y
CONFIG_CGROUP_FREEZER=y what means this lines with cgroups disabled??

agm28011997 commented on 2016-09-12 11:52

@gravysky after a few tests I have yet freeze problems while gaming browsing etc..
something wird because is a problem that affects haswell desktops only .. based on read forums and ck blog, I tried to use generic package of linux ck in the repotoo, but it has freezes too

pnehem commented on 2016-09-11 22:52

Thanks @graysky, running great so far!

agm28011997 commented on 2016-09-11 19:43

thanks @gravysky

graysky commented on 2016-09-11 14:47

Bump to v4.7.3-4

Changelog: Disable CONFIG_CGROUP_SCHED per CK's suggestion.
Commit: https://wiki.archlinux.org/index.php?title=Linux-ck/Changelog

bacondropped commented on 2016-09-10 08:02

@QuartzDragon @pnehem @graysky I've recompiled the kernel without CONFIG_CGROUP_SCHED as well (GCC 6.2.1), and for me, it DOES seem to solve the persistent kernel oops. I'll try to test it more than just poking it with a stick, meanwhile, you can use this config patch (I know it's like sixty lines with context, shut up): https://gist.github.com/bacondropped/b96cbce0382babf6e9046e143fecfd43

QuartzDragon commented on 2016-09-10 00:43

I've disabled CONFIG_CGROUP_SCHED and am rebuilding to see if the problem persists.

bacondropped commented on 2016-09-09 23:28

@pnehem It seems to happen around the same time to me as well, specifically during Xorg launch sequence (I'm running KDE on Manjaro, so to me, it happens during/after the Plymouth splash).

Another person reported an oops with a stack trace that looks almost like mine (sans maybe two entries in the middle): https://ck-hack.blogspot.ru/2016/09/bfs-490-linux-47-ck3.html?showComment=1473328786493#c6823661886449353973

Does your stack trace look like the one in that gist? I'm currently rebuilding the kernel with GCC 5.4.0 instead of 6.2.1 to eliminate the toolchain, I'll let you know how that goes.

UPDATE: @graysky Do you think disabling cgroups might be worth considering? CK writes they're unstable and are known to cause problems/panics: https://ck-hack.blogspot.ru/2016/09/bfs-490-linux-47-ck3.html?showComment=1473438769973#c6466606596661012936

UPDATE: rebuilding with GCC 5.4.0 does NOT fix the problem (reminder, here's the problem I'm trying to fix, bug report not mine, but very similar stack trace: https://gist.github.com/pfactum/8c65b2c01f7182eb578dbd9b2ef8ffd3)

pnehem commented on 2016-09-09 20:34

Hello, I'm getting a kernel panic as well, but I was trying to figure out if it where it was happening at because I can get to my login screen, just fine. But when I go to log in then it kernel panics and locks up the computer. I was just getting ready to start reading logs.
--update: I updated my system but still getting a seg fault/kernel panic. I get to the login/sddm screen just fine, But the I try to go farther the screen goes black and I can see the cursor and nothing responds locks up the whole system sort of -I can do ctl-alt-F2 and login from the "command line" but I can't make heads or tails of the logs.

agm28011997 commented on 2016-09-09 17:51

@gravisky con kollivas has commented that you use the patch of cgroups but he told too that it is unstable yet, I don't think the problem is this but can someone that have my problem try to recompile the kernel without this patch for seeing if the problem is fixed?
other thing the people that have my problem has c states in bios enabled? it is an idea..

agm28011997 commented on 2016-09-09 14:19

@gravysky what changes have you put in the kernel ck betwin the 4.7.2-1 and 4.7.2-2 with bfs 480? I revert to older kernel and I have no problems with freezes with cpufreq and pstate, but I am the only person that appear to have this problem, cgroups patch is enabled in the repo?

agm28011997 commented on 2016-09-09 14:11

@QuartzDragon kernel opps? What do you mean?

agm28011997 commented on 2016-09-09 12:57

I don't know why but I changed the p_state to cpufreq and the freezes continues here.., I used the pstate powersave and the powersave cpufreq and nothint both of them and I think the rest too give me random freezes, for ram? for memory? for cache? I don't know..

QuartzDragon commented on 2016-09-09 10:38

I consistently get kernel oops with BFS 490 ck3 whenever I start my X server. I know it's not BFQ, because it also happens with CFQ.

I'll post on his blog, anyways.

bacondropped commented on 2016-09-09 03:21

@agm28011997 @graysky Can confirm a kernel oops on v4.7.3-3.

Update 2016-09-09 12:14+000: @QuartzDragon, yep, happens on the X server start to me as well.

agm28011997 commented on 2016-09-09 02:16

I continues having problems with that kernel, I dont know if it is a problem with p_state driver, someone have this problems too???? it is supposed to be fixed.. but for me .. not

graysky commented on 2016-09-08 20:09

Bump to v4.7.3-3

Changelog: Fix config.x86_64 (my mistake), otherwise unchanged from 4.7.3-2.
Commit: https://wiki.archlinux.org/index.php?title=Linux-ck/Changelog

graysky commented on 2016-09-08 20:08

@rko - Fuck! You are correct. Fixed in 4.7.3-3. Thank you for the observation.

rko commented on 2016-09-08 20:01

I think the config.x86_64 got copied over with config (which is i386)?

Looking at the diffs, this does not seem correct and is specifying 32 bit version?

-CONFIG_64BIT=y
-CONFIG_X86_64=y
+# CONFIG_64BIT is not set
+CONFIG_X86_32=y

graysky commented on 2016-09-08 19:59

Is anyone else experiencing a kernel panic upon booting into 4.7.3-2 or is it just me?

graysky commented on 2016-09-08 19:02

Bump to v4.7.3-2

Changelog: update bfq v8r2-->v8r3 and incorporate Arch upstream changes to kernel config
Commit: https://wiki.archlinux.org/index.php?title=Linux-ck/Changelog

sflor commented on 2016-09-08 08:09

My laptop's keyboard doesn't work with 4.7.3, but did with 4.7.2. I can't even enter my password.

https://bugs.archlinux.org/task/50700

agm28011997 commented on 2016-09-08 00:53

@zerophase I test the latest kernel and for the moment I only had one freeze since I update.. I am not sure if the problem is the memory, but I don't think so with 8 GB and swap of 6GB the only thing is that a I have the tmp files mount in ram but.. not sure.. before, the pc not freezes..

zerophase commented on 2016-09-08 00:04

@agm28011997 I just run out of memory if I'm doing things with Unreal. It's because I don't have enough swap space. (going to fix that once I get an NVME ssd, and just use my old ssd for swap) I set my cpu to performance mode, so I don't encounter issues with frequencies changing.

kogone commented on 2016-09-07 19:46

with this bump from 4.7.2 to 4.7.3, make oldconfig asked whether to enable CGROUP_SCHED. should I do it?

agm28011997 commented on 2016-09-07 19:44

thabks very much graysky, I'll wait until the repo update:)

graysky commented on 2016-09-07 19:31

Bump to v4.7.3-1

Changelog: bfs490/ck3[1] and upstream release 4.7.3[2]
Commit: https://wiki.archlinux.org/index.php?title=Linux-ck/Changelog

1. http://ck-hack.blogspot.com/2016/09/bfs-490-linux-47-ck3.html
2. https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.7.3

agm28011997 commented on 2016-09-07 18:00

i dont know zerophase, if you use intel new cpu you use p-state driver and for me causes freezes.. and for out of memory the kernel not freezes too..

eduardoeae commented on 2016-09-07 12:22

BFS 490 (-ck3) is also out.
https://ck-hack.blogspot.com.ar/2016/09/bfs-490-linux-47-ck3.html

graysky commented on 2016-09-04 18:54

You guys really want to be reporting this on ck's blog if you aren't doing it already: http://ck-hack.blogspot.com/2016/09/bfs-480-linux-47-ck2.html

zerophase commented on 2016-09-04 18:27

Which driver is p state? I'm getting freezes, but thought it was from running out of memory when running virtual machines.

agm28011997 commented on 2016-09-04 18:07

someone with haswell desktop p-state driver and the lastest kernel of ck with freeze problems? i am getting freezes all time with that kernel, the older for me don't make me this and in other laptop with amd cpu it didn't do this too

graysky commented on 2016-09-02 17:18

Bump to v4.7.2-2

Changelog: bfs480 and ck2, see: http://ck-hack.blogspot.com/2016/09/bfs-480-linux-47-ck2.html
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2016-08-29 18:55

@blitz - Thanks for that info, see: https://bbs.archlinux.org/viewtopic.php?pid=1650577

According the wikipedia[1], these aren't really out there yet for testing purposes. As well, the AMD stuff isn't out until 2017 but still good to keep watch.

1. https://en.wikipedia.org/wiki/AVX-512#CPUs_with_AVX-512

blitz commented on 2016-08-29 08:18

Regarding enable_additional_cpu_optimizations_for_gcc_v4.9+_kernel_v3.15+.patch.gz.

GCC 6 Series target specific improvements for Intel Skylake -march=skylake-avx512 and AMD Zen (family 17h) processors -march=znver1 and -mtune=znver1 is now available.
https://gcc.gnu.org/gcc-6/changes.html :: New Targets and Target Specific Improvements

kyak commented on 2016-08-21 06:24

@Distorted That would be great, unless i'm the only one using bbswitch-ck and it is not worth effort.

Distorted commented on 2016-08-20 19:46

@kyak I removed bbswitch-ck from the AUR because of dkms, do you want me to put it back?

kyak commented on 2016-08-20 17:42

It looks like bbswitch-ck is gone from AUR. I don't want to use bbswitch-dkms, are there any other options?

graysky commented on 2016-08-20 16:57

Bump to v4.7.2-1

LinuxChanges: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.7.2
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2016-08-19 07:51

@asuh - CK gave the go-ahead so I will include it in 4.7.2-1 once it is available.

asuh commented on 2016-08-17 21:27

@graysky
http://ck.kolivas.org/patches/bfs/4.0/4.7/Pending/bfs472-fix_set_task_cpu.patch

graysky commented on 2016-08-17 19:10

If that's true (the lack of support for SMT_NICE) you will need to patch the source prior to building or email CK to ask him to include it in the ck1 patchset as I think your hardware is in the minority of users. I don't want to just apply a patch that without CK's blessing particularly when the code doesn't help 99% of the users :)

enihcam commented on 2016-08-17 01:25

@graysky thanks for pointing it out. Are you suggesting the panic can only be workarounded by setting CONFIG_SMT_NICE=y given that AMD processors (not zen) do not support SMT actually?

graysky commented on 2016-08-16 23:48

Well, they are certainly in the sources:
% grep CONFIG_SMT config.x86_64
CONFIG_SMT_NICE=y

% grep CONFIG_SMT config
CONFIG_SMT_NICE=y

And in the [repo-ck] packages but I can't control what you do to the PKGBUILD on your own machine :p

% uname -r
4.7.1-1-ck
% zgrep CONFIG_SMT /proc/config.gz
CONFIG_SMT_NICE=y

enihcam commented on 2016-08-16 23:48

By the way, why would I need to set SMT on an AMD processor?

enihcam commented on 2016-08-16 23:45

@graysky, I didn't find CONFIG_SMT_NICE in /proc/config.gz. Are there any dependencies I need to turn on first?

graysky commented on 2016-08-16 23:31

@enihcam - Regarding configs without CONFIG_SMT_NICE set? Linux-ck and the Arch kernel both have this set. I don't want to add a conditional in the PKGBUILD that inspects $srcdir/linux-$pkgver/.config and patches if so for example.

enihcam commented on 2016-08-16 23:08

@graysky

Please refer to http://ck-hack.blogspot.com/2016/07/bfs-0472-linux-47-ck1.html?showComment=1471307713101#c1188572232112009672

graysky commented on 2016-08-16 19:36

Bump to v4.7.0-1

LinuxChanges: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.7.1
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2016-08-16 13:31

Is this in CK's pending set? Can you get him to articulate a timeline to implement into the next bfs? If the code is not tested enough for example, and if he feels more time and testing is needed, I do not want to push unstable code to the aur and by analogy to the 4000+ users of [repo-ck].

enihcam commented on 2016-08-16 06:19

@SuperIce97 @graysky

The issue exists in bfs code. The following patch fixes it.

--- linux-4.7.orig/kernel/sched/bfs.c
+++ linux-4.7/kernel/sched/bfs.c
@@ -1418,6 +1418,9 @@ static void try_preempt(struct task_stru
* a different CPU set. This means waking tasks are
* treated differently to rescheduling tasks.
*/
+#ifndef CONFIG_SMT_NICE
+ cpu = cpu_of(highest_prio_rq);
+#endif
set_task_cpu(p, cpu);
resched_curr(highest_prio_rq);
}

SuperIce97 commented on 2016-08-15 23:50

@enihcam I haven't had any issues with it, and 4.7 is what the mainline Arch kernel is at right now. Do you have issues with the default arch kernel now? If you do, report it to the arch bug tracker so they can try to get it fixed

Distorted commented on 2016-08-15 11:03

@enihcam the aur is git based, just checkout an older version

enihcam commented on 2016-08-15 11:00

Hi Graysky,
could you please fallback this package to v4.6.6?
I think 4.7 is not stable for now, because two of my boxes failed to boot.
One freezes at logon screen, another panics.

Thanks a lot!

enihcam commented on 2016-08-15 09:37

"In addition to the resync, a few minor changes have made their way into this release with respect to the way tasks preempt other tasks. See bfs470-updates.patch for details." --<<BFS 472, linux-4.7-ck1>>

http://ck.kolivas.org/patches/bfs/4.0/4.6/Pending/bfs470-updates.patch

ryant0000 commented on 2016-08-12 16:50

Made it here as a warning. Anyways, recent update resolved issue.

graysky commented on 2016-08-11 21:26

Bump to v4.7.0-1

LinuxChanges: https://kernelnewbies.org/Linux_4.7
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2016-08-11 09:58

The AUR is probably not the best place for your report... Try communicating with Palolo, the lead developer.

ryant0000 commented on 2016-08-10 22:14

My machine kernel panicked after updating to 4.6.6-1. I do use BFQ for my ext4 root drive. I did not kernel panic on previous versions.

graysky commented on 2016-08-10 17:39

Bump to v4.6.6-1

Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.6.6
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

Distorted commented on 2016-08-10 14:48

Does anyone have issues with the new BFQ scheduler? The first v8 did some damage so i want to be careful

graysky commented on 2016-08-08 22:07

Bump to v4.6.5-4

Changelog: v8r2 of BFQ
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2016-08-05 22:02

Bump to v4.6.5-3

Changelog: v8r1 of BFQ
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

Distorted commented on 2016-08-05 20:39

Paolo updated the patch to bfq 4.6.0-v8r1, which causes the current url to be incorrect as wel.

graysky commented on 2016-08-02 17:36

@fafik1234 - Not out of date yet, see[1].

1. https://wiki.archlinux.org/index.php/Linux-ck#Release_cycle

graysky commented on 2016-07-31 08:10

OK... I recommend using an alternative IO scheduler until Paolo resolves the issue.

glavin commented on 2016-07-31 08:06

yes getting kernel panic with bfq enabled on ext4 filesystem, in my case.

graysky commented on 2016-07-31 07:59

For those getting panics, please verify that it is because you are using BFQ on a device.

Saren commented on 2016-07-31 07:43

Getting kernel panic on Haswell-E as well, I thought it was btrfs's problem, done btrfs check and fixed nothing, then I found the kernel panic message is about bfq and got here.

glavin commented on 2016-07-31 06:59

Getting kernel panic on sandybridge as well, freeze on moderate I/O and sometimes get panic messages on shutdown.

zosodk69 commented on 2016-07-30 23:01

The panics are almost certainly BFQv8 related. Reverting to BFQv7r11 makes them go away. I was able to get good panic output by logging to the serial port. I'm not a linux-ck user, rather a standalone BFQ user. I've started a thread on paolo's newsgroup to work through this issue:
https://groups.google.com/forum/#!topic/bfq-iosched/80V4U6ak57w

Because of these stability issues, I'd recommend linux-ck revert to BFQv7r11.

zerophase commented on 2016-07-30 22:49

@graysky I'm definitely getting kernel panics after the system has been up for a few hours. the numlock key doesn't even respond, once the system locks up. Running the standard kernel for now to see if it's still an issue.

graysky commented on 2016-07-29 19:06

@rucci - That is the default for the PKGBUILD except for _BFQ_enable_=y (and NUMA code has been commented for ages now). The only diff you have is enabling BFQ by default but that shouldn't trigger a bug as you activated it through other means before.

rucci commented on 2016-07-29 16:28

I recompiled again with "_1k_HZ_ticks=y", "_BFQ_enable_=y", "_NUMAdisable=y", gcc patch and now it works without kernel panic...

zerophase commented on 2016-07-29 15:17

@graysky I haven't had any panics yet. Running the kernel compiled for Haswell, with '_1k_hz_ticks' and bfq enabled.

Cake commented on 2016-07-29 14:12

@graysky Yes, I only installed one disk (/dev/sda) and and enabled BFQ on it. Several hours ago, I checked it with
cake-pc# cat /sys/block/sda/queue/scheduler
noop deadline cfq [bfq]

I am sorry that I did not try recompiling `linux-ck=4.6.5` with `_1k_HZ_ticks=`. I just changed `grub.cfg` and rebooted with the stock kernel ... Everything seems to be working fine ...

I am wondering what is going on when the system freezes. So if you require any other information, please let me know.

graysky commented on 2016-07-29 13:19

So two reports of the new bfq causing panics? Can you both verify that you have a drive set to use bfq? I don't see how tick rate plays into the equation.

rucci commented on 2016-07-29 09:20

It works if I disable "_1k_HZ_ticks" option

Cake commented on 2016-07-29 08:39

@rucci I have the same problem (Gnome BFQ Ivybridge) after updating to 4.6.5-2. I have to restart by pressing the reset button on my machine when the system freezes...

rucci commented on 2016-07-29 08:28

Kernel panic after update 4.6.5-1 -> 4.6.5-2 (KDE BFQ IvyBridge)

mradermaxlol commented on 2016-07-29 05:35

@graysky if you wanna know, everything's just fine after the latest update - stuff is buildable, at least for me. I can't understand why -ck patch was ruining the build, though; maybe I'm just dumb.

graysky commented on 2016-07-28 18:58

3.6.5-2 has been updated the corrected name of the 4th patch; the checksums remain the same since upstream did not change the content, only the name.

mradermaxlol commented on 2016-07-28 15:27

I'm getting patching errors (likely) during prepare() execution which result in makepkg build failure. It's happening at -ck patchset applying, and I have completely no idea what's causing it - maybe these patches are outdated, maybe it's just 4.6.5-incompatible (although I don't think so as errors happen in bfq.c/bfq.h files, for example). Right now my best guess is that the fourth BFQ patch is causing that, because that makes sense - it might be okay a while ago, but now its name in PKGBUILD is useless as it's wrong; I've changed it as @herraiz suggested and, well, everything's okay except patching X)

Does anybody else experience this issue as well?

EDIT: After a little research I've found out (I hope so) that the thing responsible for that problem is... Yup, the -ck patchset itself. Ta-daa!

herraiz commented on 2016-07-28 14:23

@nTia89: the third patch has changed, you have probably an old copy of the third patch in your sources dir.

nTia89 commented on 2016-07-28 13:16

other than corrected forth name's patch,
in order to have a working pkgbuild,
we need to correct the checksum for the third patch

graysky commented on 2016-07-28 09:25

I will update the source array for the renamed patch but this is technically not out of date since the content of that patch did not change. See my post a few down (we knew this change was coming).

herraiz commented on 2016-07-28 08:25

I think there is a typo in one of the URLs. I am getting the following error:

curl: (22) The requested URL returned error: 404 Not Found
==> ERROR: Failure while downloading http://algo.ing.unimo.it/people/paolo/disk_sched/patches/4.6.0-v8/0004-blkck-bfq-turn-BFQ-v7r11-for-4.7.0-into-BFQ-v8-for-4.patch

The correct URL is http://algo.ing.unimo.it/people/paolo/disk_sched/patches/4.6.0-v8/0004-block-bfq-turn-BFQ-v7r11-for-4.7.0-into-BFQ-v8-for-4.patch

(blkck -> block)

FadeMind commented on 2016-07-28 02:54

Hi @graysky

https://groups.google.com/forum/#!topic/bfq-iosched/wAKntB66OHs

BFQ changelog.

Regards

FadeMind

cryzed commented on 2016-07-28 01:03

@grasky, yes: pacaur. Should I run makepkg manually? I want to note that I have been using pacaur previously to install and update linux-ck without any problems.

EDIT: Running it manually right now using "makepkg -fsic".
EDIT 2: Well that worked. That's the first time an AUR helper bit me, especially pacaur which is supposed to be very reliable (https://wiki.archlinux.org/index.php/AUR_helpers#Comparison_table). Sorry about this.

graysky commented on 2016-07-28 00:56

Are you using an aur helper?

cryzed commented on 2016-07-28 00:36

When updating to the latest version, I get the following error after the seemingly successful compilation process:

> ==> Finished making: linux-ck 4.6.5-2 (Do 28. Jul 02:33:40 CEST 2016)
> ==> Cleaning up...
> :: Installing linux-ck,linux-ck-headers package(s)...
> :: linux-ck,linux-ck-headers package(s) failed to install. Check .SRCINFO for mismatching data with PKGBUILD.

The .SRCINFO looks like this: https://gist.github.com/anonymous/596f391daa1fca584798684e5b99cd70.

I did the following modifications to the PKGBUILD:
> _1k_HZ_ticks=
> _BFQ_enable_=y
I don't use the sleep/suspend feature, hence the first change.

My /etc/makepkg.conf has the following modifications:
> CFLAGS="-march=native -O2 -pipe -fstack-protector-strong"
> CXXFLAGS="-march=native -O2 -pipe -fstack-protector-strong"
> MAKEFLAGS="-j$(nproc)"

graysky commented on 2016-07-27 21:16

@blitz - Yes, I assumed the filename will change, and I'll have to update it :p If you catch it before someone else posts here, please email me privately so I can push the change and minimize impact of users building from the AUR. Thanks.

blitz commented on 2016-07-27 21:04

@graysky 2016-07-27 21:01
That was fast, while /me wrote mail message to Paolo Valente with regard to broken filenaming sequence, -blkck- vs -block-

graysky commented on 2016-07-27 21:01

Bump to v4.6.5-2

Changelog: 4.6.5-1 neglected to apply one of the BFQ patches; 4.6.5-2 fixes it (thanks to Lucjan Lucjanov for pointing this out).
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2016-07-27 18:21

Bump to v4.6.5-1

Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.6.5
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2016-07-11 21:57

Bump to v4.6.4-1

Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.6.4
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2016-06-25 15:33

Bump to v4.6.3-1

Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.6.3
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog
LinuxChanges: http://kernelnewbies.org/Linux_4.6

Paolo has assured me that BFQ v7 is fine for the 4.6.x series of kernels; the bug that is affecting it applies equally to 4.5.x kernel users so there is no harm in including it against the 4.6.x series. He is testing prev8 of BFQ now but would not predict a release date. Enjoy the 4.6 series.

artafinde commented on 2016-06-20 19:52

That's fine and I agree - just mentioned my findings (though if I knew about the g-groups I'd already know that :D ). Probably the wait will be worth it with the new BFQ v8 [https://groups.google.com/forum/#!topic/bfq-iosched/ReUuq4jtZ84]

graysky commented on 2016-06-20 19:32

Paolo has a testing version of BFQ for 4.6 but it and the 4.5 series have problems with hierarchical support[1]. We have 4,000+ [repo-ck] users so I would rather wait fo the ink to dry on the BFQ 4.6-final version before including old or testing patches.

1. https://groups.google.com/forum/#!topic/bfq-iosched/7gJMK1DttYQ

artafinde commented on 2016-06-20 07:55

Interestingly the patches of BFQ 4.5 apply successfully on 4.6.2. Although I haven't test those as I don't use BFQ.

graysky commented on 2016-06-08 21:03

@dlh - Not of of date... see the comments from 2016-06-08

graysky commented on 2016-06-08 14:50

@PieGuy128 - Not out of date... see the comments from 2016-06-08

graysky commented on 2016-06-08 11:17

Bump to v4.5.7-1

Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.5.7
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2016-06-08 11:05

@Fusion - Yes, but we still need to wait for Paolo to release a BFQ for 4.6: http://algo.ing.unimo.it/people/paolo/disk_sched/sources.php

fusion809 commented on 2016-06-08 03:13

4.6 is out http://ck.kolivas.org/patches/4.0/4.6/

graysky commented on 2016-06-01 21:18

Bump to v4.5.6-1

Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.5.6
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2016-05-26 19:06

@Tjuh - Yes, that is likely your problem. I highly recommend using modprobed-db if you want to make localmodconfig for this purpose. Just know that it's only as good as your database and if you want to have a new module that you haven't used before, you will have to boot into a kernel that actually has this module available for modprobed-db to learn that you have an interest. Hope that makes sense. See the wiki for more.

Distorted commented on 2016-05-26 12:46

@Tjuh
Disable localmodcfg, it causes the kernel to only compile the modules you are currently using.

Tjuh commented on 2016-05-26 11:07

Ye I also find it weird, cause I checked the config.x86_64 and saw that it is enabled as a module. Hmm maybe I should disable the localmodprobe option in the PKGBUILD. And yes rebooted and recompiled many times.

graysky commented on 2016-05-25 20:56

@Tjuh - Dunno why, it's in the config and it loads on my system running 4.5.5-ck-1
% zgrep -i zram /proc/config.gz
CONFIG_ZRAM=m
CONFIG_ZRAM_LZ4_COMPRESS=y

Have you perhaps not rebooted since updating to 4.5.5?

Tjuh commented on 2016-05-25 20:21

I'm trying to get zram working follwing the wiki, but the module seems to be missing;

tjuh ~ $ modprobe zram
modprobe: FATAL: Module zram not found in directory /lib/modules/4.5.5-1-ck

Any suggestions?

graysky commented on 2016-05-19 06:52

Bump to v4.5.5-1

Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.5.5
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

francoism90 commented on 2016-05-17 14:47

@sir_lucjan: OK, thanks.. hope it will be available soon. :)

sir_lucjan commented on 2016-05-17 13:08

@francoism90:

Nope.

http://ck.kolivas.org/patches/bfs/4.0/

francoism90 commented on 2016-05-17 13:07

Hi graysky, any news about Linux 4.6? Is a patch available? Thanks :)

graysky commented on 2016-05-11 22:54

Bump to v4.5.4-1

Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.5.4
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2016-05-04 22:46

Bump to v4.5.3-1

Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.5.3
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

herraiz commented on 2016-04-29 17:16

@graysky you are right, I had an old copy of that patch in my sources dir. Thanks!

ScarecrowDM commented on 2016-04-29 14:56

@Tjuh, same problem here.
Disabling BUILDDIR=/tmp/makepkg fixed it for me. :)

graysky commented on 2016-04-28 18:44

@herraiz - It's on your end :) Delete the offending file to allow makepkg to redownload it.

herraiz commented on 2016-04-28 17:19

One of the patches fails to pass the checksum:

0003-block-bfq-add-Early-Queue-Merge-EQM-to-BFQ-v7r11-for.patch ... FAILED

graysky commented on 2016-04-27 22:12

@Tjuh - You have something wrong with your setup... compiles for me just fine. I recommend that you try it in a clean chroot. Try clean-chroot-manager here in the AUR.

Tjuh commented on 2016-04-27 21:53

I did lols, everytime it fails, I unpack the linux-ck.tar.gz and start a fresh makepkg. And furthermore those dirs src and pkg aren't even present when that error occurs.

graysky commented on 2016-04-27 21:33

@francoism - https://wiki.archlinux.org/index.php/Linux-ck#Package_defaults
@Tjuh - Clear your src dir before you build (ie rm -rf pkg src) assuming you're using makepkg and not an AUR helper.

Tjuh commented on 2016-04-27 21:14

So I've tried to compile this atleast a dozen times, but it fails everytime at applying a patch;

...
==> Patching source with ck1 including BFS v0.469
patching file arch/powerpc/platforms/cell/spufs/sched.c
The next patch would create the file Documentation/scheduler/sched-BFS.txt,
which already exists! Skipping patch.
1 out of 1 hunk ignored
patching file Documentation/sysctl/kernel.txt
patching file fs/proc/base.c
patching file include/linux/init_task.h
patching file include/linux/ioprio.h
patching file include/linux/sched.h
patching file init/Kconfig
patching file init/main.c
patching file kernel/delayacct.c
patching file kernel/exit.c
patching file kernel/sysctl.c
patching file lib/Kconfig.debug
patching file include/linux/jiffies.h
patching file drivers/cpufreq/cpufreq.c
patching file drivers/cpufreq/cpufreq_ondemand.c
The next patch would create the file kernel/sched/bfs.c,
which already exists! Skipping patch.
1 out of 1 hunk ignored
patching file include/uapi/linux/sched.h
patching file drivers/cpufreq/cpufreq_conservative.c
patching file kernel/time/Kconfig
patching file kernel/sched/Makefile
The next patch would create the file kernel/sched/bfs_sched.h,
which already exists! Skipping patch.
1 out of 1 hunk ignored
patching file kernel/sched/stats.c
patching file arch/x86/Kconfig
Hunk #2 succeeded at 1294 (offset 1 line).
Hunk #3 succeeded at 1314 (offset 1 line).
Hunk #4 succeeded at 2012 (offset 1 line).
Hunk #5 succeeded at 2041 (offset 1 line).
patching file include/linux/sched/prio.h
patching file drivers/cpufreq/intel_pstate.c
patching file kernel/sched/idle.c
patching file kernel/time/posix-cpu-timers.c
patching file kernel/trace/trace_selftest.c
patching file kernel/Kconfig.preempt
patching file kernel/Kconfig.hz
patching file Makefile
==> ERROR: A failure occurred in prepare().
Aborting...
...

francoism90 commented on 2016-04-27 19:59

@graysky: I mean patches that aren't in the added to the Linux kernel yet, like the BFS patch. :)
But that's maybe a bit off-topic, think I'll better ask on the forums.

graysky commented on 2016-04-27 19:40

@francoism - Not sure what you're asking?

francoism90 commented on 2016-04-27 19:37

@graysky: Thanks going to upgrade tomorrow. :D

Off-topic: is there a repo somewhere with recommended patches and working with this kernel? :)

graysky commented on 2016-04-27 19:17

Bump to v4.5.2-1

Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.5.2
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog
LinuxChanges: http://kernelnewbies.org/Linux_4.5

sir_lucjan commented on 2016-04-27 18:44

@FadeMind - a little bit to late, BFQ for 4.5 is out :D

FadeMind commented on 2016-04-27 15:20

Who is interesting Linux 4.5 with CK patches can download ready PKGBUILD and patches from my repo. BFQ for 4.5 kernel included.

git clone https://github.com/FadeMind/archpkgbuilds.git

https://github.com/FadeMind/archpkgbuilds/tree/master/AUR/linux-ck

Kernel working for me like a charm. See README.md file for details.

Regards.

FadeMind

francoism90 commented on 2016-04-22 20:43

@graysky: thanks for the skylake update! :)

graysky commented on 2016-04-20 21:21

Bump to v4.4.8-1

Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.4.8
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2016-04-19 19:55

@QuartzDragon - Yes, Paolo told me as much in a private email but I don't want to force all the repo users to be guinea pigs since the 4.5 series of BFQ should be released very shortly per Paolo.

@kyak - No problem, thanks for pointing that out.

@francoism - No testing is needed, it's just that adding support required tweaks to the entire set of packages (ie adding conflicts arrays) the skylake packages will be included in the next build (either 4.4.8 or 4.5.2 or whenever Paolo releases BFQ for 4.5.x - whichever comes first).

francoism90 commented on 2016-04-19 18:01

Hi graysky,

Thanks for providing Linux-ck, really like the BFQ scheduler and also like the easy tweaking/customizing. :)

Sorry if this is already ask before, but if testing is needed for the v4.5.x series, please let me know. I'm running on Skylake platform, maybe this platform needs more testing? :P

Thanks again!

kyak commented on 2016-04-19 14:54

@graysky - thank you, all works fine now!
As for rename, i'm not pushing at all; this is just something to keep in mind.
Thanks again!

QuartzDragon commented on 2016-04-19 12:59

Greetings graysky,

Some users on the BFQ mailing list have tested the v7r11 4.4.0 patches on 4.5.0 kernel and have found they work without any issues. I've also tested for myself with a modified PKGBUILD and found they work stably and without any issues or quirks compared to CFQ, no matter my system load or uptime.

I guess your plans are to wait for the official patches?

graysky commented on 2016-04-18 20:48

@kyak - Thanks, try refreshing and downloading now. As to the renaming of the package, that would require a rewrite of my build script which isn't in the cards for the short term.

kyak commented on 2016-04-18 17:47

@graysky, thanks! But virtualbox-ck-host-modules-ivybridge and virtualbox-host-modules-arch are conflicting now. How should we handle this?

By the way, does it make sense to change the naming convention in accordance to upstream - i.e. virtualbox-host-modules-ck-ivybridge ?

graysky commented on 2016-04-18 16:51

@kyak - Done.

kyak commented on 2016-04-18 15:57

hi graysky,

can we get the virtualbox-host-modules-ck-* back now that the upstream has returned the binary modules?

zipeldiablo commented on 2016-04-17 18:40

Hi guys, question might be weird but i run windows 10 in virtual machine with vga passthrough (qemu-kvm).
Should i expect any performance gain on my benchmark or will i just have more perf on my host system?

graysky commented on 2016-04-12 21:08

Bump to v4.4.7-1

Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.4.7
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2016-04-06 19:43

NOTE -- Even though CK has released BFS for linux v4.5.x, there is not a corresponding BFQ I/O scheduler for this series yet so please don't flag out-of-date until Paolo has done so.

graysky commented on 2016-04-06 11:37

@tech4david - See the pinned comment :p

zerophase commented on 2016-04-02 15:34

@pedrogabriel, I tried the steamos-xpad driver.

Error! There are no instances of module: steamos-xpad
20160103 located in the DKMS tree.
error: command failed to execute correctly

I tried deleting the xpad.ko.gz file from 4.4.6-1-ck kernel with no luck.

pedrogabriel commented on 2016-04-02 02:52

@zerophase, use the steamos driver.
https://aur.archlinux.org/packages/steamos-xpad-dkms/

zerophase commented on 2016-04-02 01:28

I managed to get my 360 controller working by enabling xpad, but can't figure out how to get the led ring to stop blinking with the ck kernel? Anyone able to help.

graysky commented on 2016-03-26 11:01

@kyak - https://wiki.archlinux.org/index.php/Repo-ck#Installation

kyak commented on 2016-03-26 05:49

Hi graysky,

Where has virtualbox-ck-host-modules-ivybridge gone?
I only see generic virtualbox-ck-host-modules now. Should i use that?

Nordlicht commented on 2016-03-25 20:33

Runs without problems here with nvidia-ck (AUR).

graysky commented on 2016-03-25 19:04

OK all, please build and test 4.4.6-1 which is running fine on my machine. I will hold off a bit on building/pushing to the repo until I get some feedback from you guys. Thanks.

graysky commented on 2016-03-25 18:27

Bump to v4.4.6-1

Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.4.6
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog
LinuxChanges: http://kernelnewbies.org/Linux_4.4

NOTE -- Even though CK has released BFS for linux v4.5.x, there is not a corresponding BFQ I/O scheduler for this series yet so please don't flag out-of-date until Paolo has done so.

Distorted commented on 2016-03-25 12:07

Running it now without issues

graysky commented on 2016-03-25 09:02

Dammit. You will have to rename the file to simply linux-ck-4.4.6-1.src.tar after you download it as it wasn't run through gzip my mistake. Also, you will need to update the checksums as config will fail as-is.

graysky commented on 2016-03-25 08:07

Guys - I don't have time to verify this builds or test so I am holding off pushing up to the AUR until later today or over the weekend depending on my schedule . Please test 4.4.6 and report back. Source tarball: http://repo-ck.com/PKG_source/linux-ck/linux-ck-4.4.6-1.src.tar.gz

inti commented on 2016-03-19 17:56

@Peter1992 I'm not sure if this is what you meant, but graysky doesn't create the ck patchset; Con Kolivas does. See: http://users.tpg.com.au/ckolivas/kernel/

graysky puts together the AUR linux-ck PKGBUILD once Con Kolivas releases updated versions of the patchset.

Peter1992 commented on 2016-03-14 21:28

@graysky Right,graysky didn't know about that.Well just for curiosity how long would you think it would take for you to put the ck1 patchsets for the 4.4.x kernels?

graysky commented on 2016-03-14 19:31

@Peter1992 - Not out of date... there is no ck1 for 4.4.x.

FadeMind commented on 2016-03-09 08:59

@ thevictoryiswon

gpg --recv-keys ABAF11C65A2970B130ABE3C479BE3E4300411886 647F28654894E3BD457199BE38DBBDC86092693E

Distorted commented on 2016-03-09 08:59

@thevictoryiswon
Look at some of the older comments, your issue has come up numerous times.

thevictoryiswon commented on 2016-03-09 04:00

It will not build for me. Has it been corrupted?

==> Verifying source file signatures with gpg...
linux-4.3.tar ... FAILED (unknown public key 79BE3E4300411886)
patch-4.3.6 ... FAILED (unknown public key 38DBBDC86092693E)
==> ERROR: One or more PGP signatures could not be verified!

WFV commented on 2016-03-06 23:13

Thanks graysky and FadeMind, rebuild worked.

FadeMind commented on 2016-03-06 21:50

@WFV kernel images have stored systemd version from time creating of image kernel.
For bump this run just

sudo mkinitcpio -P and reboot.

Alternative:
ignore this or add to grub conf file

/etc/default/grub

in line GRUB_CMDLINE_LINUX

rd.udev.log-priority=3

for prevent print systemd version during boot.

Update grub after made changes in a conf file OFC.

graysky commented on 2016-03-06 21:49

@WFV - Rebuild your initramfs image... nothing to do with the packages.

WFV commented on 2016-03-06 21:27

Running 4.3.6-1-ck (piledriver) and current -Syu as of 3/3/16. I notice during boot the message still shows "starting version 228" instead of 229. Doesn't appear to affect anything, but was thinking the version refers to systemd?

Distorted commented on 2016-02-21 21:45

@farwaver i have the same issue, but im still on 4.3.5, so i think its unrelated.

graysky commented on 2016-02-21 21:44

@fawyayer - You need to compare 4.3.6-1-ck to 4.3.6-1-ARCH (which you need to build yourself) to provide a comparison. Comparing to 4.4.1-1-ARCH is the wrong experiment.

farwayer commented on 2016-02-21 21:43

Look like there is some problem with working intel i965 driver with 4.3.6-1-ck kernel. Xorg cursor does not render. However there are no errors in xorg log.
Changing acceleration method to old uxa fixs the problem. All ok with arch vanila kernel (4.4.1).

graysky commented on 2016-02-20 00:12

Bump to v4.3.6-1

Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.3.6
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

nlopez commented on 2016-02-15 17:40

gpg --recv-keys 79BE3E4300411886 38DBBDC86092693E # to properly validate the signatures on this package. As the page @Distored mentions, makepkg uses your user keyring: http://allanmcrae.com/2015/01/two-pgp-keyrings-for-package-management-in-arch-linux/

sinatosk commented on 2016-02-07 15:33

nevermind... I see that in order for CGROUP_CPUACCT it to show... BFS needs to be disabled... just noticed

dogabone commented on 2016-02-04 02:28

Sent you an e-mail with a small change to the PKGBUILD. Hope it helps.

graysky commented on 2016-02-02 22:30

@hydraz - Not out-of-date until CK releases the BFS for 4.4 :p

Anonymous comment on 2016-02-01 17:26

https://gist.githubusercontent.com/AnAverageHuman/6ff2cb85240a34a182f6/raw/aaee0224ce09bdc449b9eda9735dc8df61fdd1a6/gistfile1.txt

It seems like linux-ck-headers was installed twice and linux-ck wasn't installed?

graysky commented on 2016-02-01 00:05

Bump to v4.3.5-1

Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.3.5
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

Distorted commented on 2016-01-29 01:35

@sa1t
https://wiki.archlinux.org/index.php/Makepkg#Signature_checking
TL;DR version: use 'makepkg --skippgpcheck'

sa1t commented on 2016-01-29 01:19

==> Verifying source file signatures with gpg...
linux-4.3.tar ... FAILED (unknown public key 79BE3E4300411886)
patch-4.3.4 ... FAILED (unknown public key 38DBBDC86092693E)
==> ERROR: One or more PGP signatures could not be verified!
==> ERROR: Makepkg was unable to build linux-ck.
==> Restart building linux-ck ? [y/N]

FadeMind commented on 2016-01-24 20:11

http://algo.ing.unimo.it/people/paolo/disk_sched/patches/4.4.0-v7r11/

;)

but for 4.4.0, ck1 is still pending...

eduardoeae commented on 2016-01-24 20:02

BFQ has been updated to v7r11 fixing some bugs.

graysky commented on 2016-01-23 12:42

Bump to v4.3.4-1

Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.3.4
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

Distorted commented on 2016-01-23 10:24

Would it be possible to update to 4.3.4?

graysky commented on 2016-01-20 15:36

Bump to v4.3.3-3

Changelog: fix #47820 CVE-2016-0728
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2016-01-18 20:53

I saw it, FM, thanks. Just need ck1...

FadeMind commented on 2016-01-18 18:26

@graysky
http://algo.ing.unimo.it/people/paolo/disk_sched/patches/4.4.0-v7r10/

graysky commented on 2016-01-13 08:43

@gutigen - Not out of date until:
-Arch updates to 4.4
-CK releases a BFS for 4.4
-Paolo releases a BFQ for 4.4

graysky commented on 2015-12-24 15:35

Bump to v4.3.3-2

Changelog: fix #46968
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2015-12-24 15:35

Sorry about that. Fixed now (didn't bump pkgver since I just rebuilt the repo).

whatot commented on 2015-12-24 14:53

missing the sha256sum of 0001-disabling-primary-plane-in-the-noatomic-case.patch after the eighth.

pedrogabriel commented on 2015-12-24 14:50

==> ERROR: Integrity checks (sha256) differ in size from the source array.
==> ERROR: Makepkg was unable to build linux-ck.

It's not just you.

Dijuna commented on 2015-12-24 14:19

==> ERROR: Integrity checks (sha256) differ in size from the source array.
==> ERROR: Makepkg was unable to build linux-ck.

Is it just me?

graysky commented on 2015-12-15 08:37

Bump to v4.3.3-1

Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.3.3
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog
Note from CK: http://ck-hack.blogspot.com/2015/12/bfs-467-linux-43-ck3.html

graysky commented on 2015-12-14 12:02

@XenGi - I try to be as true to the Arch config as I can be for linux-ck. You are of course free to customize/build your own. My suggestion is for you to open a flyspray again the official Arch linux package asking for this. Might want to include your reasoning why it would be important.

XenGi commented on 2015-12-14 10:51

hi,

Could you set CONFIG_CHECKPOINT_RESTORE=y for all ck kernels? I'm using the broadwell build from repo-ck.
This would enable fancy process names for LXC. The setting is not set atm.

```
$ zgrep 'CONFIG_CHECKPOINT_RESTORE' /proc/config.gz
# CONFIG_CHECKPOINT_RESTORE is not set
```

Here's a thread explaining the issue: https://lists.linuxcontainers.org/pipermail/lxc-users/2015-November/010411.html

graysky commented on 2015-12-14 07:29

I did find this suggesting it is not related to linux-ck: https://bugzilla.kernel.org/show_bug.cgi?id=107421

graysky commented on 2015-12-14 07:27

Recommend that you post to CK's blog.

enihcam commented on 2015-12-14 07:14

@renodesper I got the exactly same issue.

renodesper commented on 2015-12-14 01:58

I got this weird message when entering boot process:

[ 47.772679] r8169 0000:03:00.0 enp3s0: rtl_counters_cond == 1 (loop: 1000, delay: 10).
[ 52.807884] r8169 0000:03:00.0 enp3s0: rtl_counters_cond == 1 (loop: 1000, delay: 10).
[ 57.842769] r8169 0000:03:00.0 enp3s0: rtl_counters_cond == 1 (loop: 1000, delay: 10).
[ 62.877924] r8169 0000:03:00.0 enp3s0: rtl_counters_cond == 1 (loop: 1000, delay: 10).
[ 67.912998] r8169 0000:03:00.0 enp3s0: rtl_counters_cond == 1 (loop: 1000, delay: 10).
[ 72.947899] r8169 0000:03:00.0 enp3s0: rtl_counters_cond == 1 (loop: 1000, delay: 10).
[ 77.982966] r8169 0000:03:00.0 enp3s0: rtl_counters_cond == 1 (loop: 1000, delay: 10).
[ 83.017963] r8169 0000:03:00.0 enp3s0: rtl_counters_cond == 1 (loop: 1000, delay: 10).
[ 87.661070] r8169 0000:03:00.0 enp3s0: rtl_counters_cond == 1 (loop: 1000, delay: 10).
[ 88.053001] r8169 0000:03:00.0 enp3s0: rtl_counters_cond == 1 (loop: 1000, delay: 10).

It still showing up when I check from dmesg. The message only appear when I use ck kernel. Is there something wrong with ck kernel?

graysky commented on 2015-12-11 13:21

Bump to v4.3.2-1

Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.3.2
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2015-12-11 13:01

Bump to v4.3.1-3

Changelog: Incorporate BFS 466 in ck2[1]
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

1. http://ck-hack.blogspot.com/2015/12/bfs-466-linux-43-ck2.html

graysky commented on 2015-12-10 18:27

Bump to v4.3.1-2

Changelog: Mirroring changes to official package's config.
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2015-12-09 21:08

Bump to v4.3.1-1

Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.3.1
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

crobe2 commented on 2015-12-07 21:41

The unimo.it page seems down :(

petterk commented on 2015-11-24 22:04

For all those with this problem in dmesg/bootup:

[drm:intel_pipe_config_compare [i915]] *ERROR* mismatch in has_drrs (expected 1, found 0)

Theres also a warning, I dont know if it is related:

WARNING: CPU: 1 PID: 6 at drivers/gpu/drm/i915/intel_display.c:12700 intel_atomic_commit+0xdea/0x1390 [i915]()

Bug submittet here:
https://bugzilla.kernel.org/show_bug.cgi?id=108431

kyak commented on 2015-11-13 18:18

@Distorted - of course, and thanks again!

Distorted commented on 2015-11-13 18:15

@kyak I dont use the repo-ck, so i'm just guessing here, if linux-ck-ivybridge is not version 4.3 you're absolutely right, but until then you could use an older commit of bbswitch-ck.

FadeMind commented on 2015-11-13 18:13

kyak

linux-ck is main package for GENERIC (ALL) CPU's and OFC can be adjust BEFORE compile.
Customized by CPU names are designed for powerfull CPU's AMD and Intel
See: https://github.com/graysky2/kernel_gcc_patch
are available on repo-ck.com repo site.

IF You have installed linux-ck-ivybridge You have two options:
1. Download sources of linux-ck and custom config file for enable valid CPU family for GCC and compile it.
2. Purge whole -ck packages (linux-ck-ivybridge, bbswitch-ck, and nvidia-ck-ivybridge)
3. Purge and install again bbswitch-ck package (major linux version bumped from 4.1 to 4.3 and package NEED to be recompilled against 4.3)
4. Install new version of nvidia-ck version ready for 4.3

OR

JUST WAIT for upload proper packages in to repo-ck repository.

SUBFORUM of AUR packages IS designed for report ISSUES with PKGBUILD/package build issues. FOR help with Your misunderstanding of the issues please write directly on BBS Arch Forums: https://bbs.archlinux.org/viewtopic.php?id=111715&p=104

Reference https://wiki.archlinux.org/index.php/Repo-ck#Kernels_and_related_packages

And one more: stop spamming and confuse users of other packages (bbswitch-ck).

kyak commented on 2015-11-13 18:11

@Distorted i think i get it now. I should wait for the latest version of linux-ck-ivybridge to hit the repo-ck. After that i'd have something that provides linux-ck=4.3 and be able to install bbswitch-ck. I didn't realize i was still using linux-ck-ivybridge 4.1.3

Distorted commented on 2015-11-13 18:01

@kyak also, did you rebuild the bbswitch-ck package? i updated it for the new version.

Distorted commented on 2015-11-13 18:00

@kyak You could probably edit the bbswitch-ck PKGBUILD file and change depends=linux-ck to depends=linux-ck-ivybridge
Unless they changed the actual name of the kernel, not just the package.
If this is not the case, the linux-ck-ivybridge pkgbuild should say provide='linux-ck=4.3' to solve these kinds of problems.

kyak commented on 2015-11-13 17:54

Well, wait a second.. bbswitch-ck requires linux-ck, which conflicts with linux-ck-ivybridge that i have installed. What to do? It used to work before

kyak commented on 2015-11-13 17:51

ok, finally.. had to uninstall bbswitch-ck. It doesn't work woth 4.3 kernel. Thanks for the hint, Distorted.

kyak commented on 2015-11-13 17:47

but.. i didn't build anything from aur. I'm using repo-ck

graysky commented on 2015-11-13 17:25

Just remove all repo-ck packages before you try installing one you built yourself from the AUR: pacman -Rs ck-ivybridge

kyak commented on 2015-11-13 17:01

$ LANG=C pacaur -R linux-ck
error: target not found: linux-ck

Yes, duh. I don't have linux-ck installed, only linux-ck-ivybridge. I suspect a pacaur bug.

sir_lucjan commented on 2015-11-13 16:42

Read carefully:

conflicts=('kernel26-ck' 'linux-ck-corex' 'linux-ck-p4' 'linux-ck-pentm' 'linux-ck-atom' 'linux-ck-core2' 'linux-ck-nehalem' 'linux-ck-sandybridge' 'linux-ck-ivybridge' 'linux-ck-broadwell' 'linux-ck-haswell' 'linux-ck-kx' 'linux-ck-k10' 'linux-ck-barcelona' 'linux-ck-bulldozer' 'linux-ck-piledriver' 'linux-ck-silvermont')

Duh? :)

Distorted commented on 2015-11-13 15:28

@kyak There are probably some packages that depend on linux-ck-ivybridge that are the cause this error.

kyak commented on 2015-11-13 15:16

Running pacaur -Syu:

:: linux-ck and linux-ck-ivybridge are in conflict (linux-ck). Remove linux-ck-ivybridge? [y/N]

:: unresolvable package conflicts detected
:: failed to prepare transaction (conflicting dependencies)
:: linux-ck and linux-ck-ivybridge are in conflict


Duh?

FadeMind commented on 2015-11-12 21:12

Self compilled with custom modules list (modprobed_db).
Linux version 4.3.0-1-ck (tomasz@arch) (gcc version 5.2.0 (GCC) ) #1 SMP PREEMPT Thu Nov 12 21:46:41 CET 2015

custom modules list for my device: https://pastebin.com/raw.php?i=u0hhFnmA
kernel tree: https://pastebin.com/raw.php?i=bzXse8Rk
enabled blk-mq via grub flag (scsi_mod.use_blk_mq=1)
(working like a charm - better than noop)
dmesg: https://pastebin.com/raw.php?i=03cnEfz5
config: https://pastebin.com/raw.php?i=Y6VcKyKX

Good work CK and Graysky!

graysky commented on 2015-11-12 19:41

Bump to v4.3-1

Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.3
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog
LinuxChanges: http://kernelnewbies.org/Linux_4.3
Note: I am working on the related packages now (broadcom, nvidia, vbox) and I will upload to the repo once I get some feedback from AUR users as to stability so please use the AUR page to report in if something is bad (or good)!

yannick commented on 2015-11-12 08:25

It's building now, thx

vishwin commented on 2015-11-11 07:53

I just 'SKIP'ped that file for now, will wait for the next release haha.

graysky commented on 2015-11-10 08:45

@yannick and vishwin - Try now.

yannick commented on 2015-11-10 08:41

This part of the patch is not good:

- sha256sums = 819961379909c028e321f37e27a8b1b08f1f1e3dd58680e07b541921282da532
+ sha256sums = cf0f984ebfbb8ca8ffee1a12fd791437064b9ebe0712d6f813fd5681d4840791

vishwin commented on 2015-11-10 00:24

"enable_additional_cpu_optimizations_for_gcc_v4.9+_kernel_v3.15+.patch.gz ... FAILED"
…yet the file itself appears to never have been modified and has passed before 4.1.13-1?

graysky commented on 2015-11-09 23:03

Bump to v4.1.13-1

Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.1.13
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

artafinde commented on 2015-10-28 20:48

Hmm seems like the BFQ patches host [1] is down. Mirror from Oleksandr Natalenko at https://pf.natalenko.name/mirrors/bfq/4.1.0-v7r8/ same SHA256 sum
[1]: http://algo.ing.unimo.it

graysky commented on 2015-10-27 21:12

Bump to v4.1.12-1

Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.1.12
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2015-10-27 11:58

Thanks for the comments, FM, and SL. I have been following Con's blog and am waiting to hear back from him that the proposed solution is kosher before I apply it to the AUR and repo. Stay tuned.

sir_lucjan commented on 2015-10-27 07:22

@FadeMind:

http://ck-hack.blogspot.com/2015/08/bfs-464-linux-41-ck2.html?showComment=1445687952915#c1684323780488525819

"Just a heads up that 4.1.12 will break compilation of BFS because of a change in the common scheduler API. I already sent a patch to Con to fix it, so please don't panic. :)"

FadeMind commented on 2015-10-27 06:58

https://cdn.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.1.12

Main URL to sources was changed to https://cdn.kernel.org

Sample: https://cdn.kernel.org/pub/linux/kernel/v4.x/patch-4.1.12.xz

graysky commented on 2015-10-22 23:07

Bump to v4.1.11-1

Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.1.11
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

glitsj16 commented on 2015-10-14 00:13

Hi graysky, would you consider modifying the links to kernel.org inside the PKGBUILD cfr. https://www.kernel.org/introducing-fastly-cdn.html to improve download speeds for folks located outside North America?

graysky commented on 2015-10-07 18:43

Bump to v4.1.10-3

Changelog: Fix potential deadlock in reqsk_queue_unlink[1]
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

1. https://bbs.archlinux.org/viewtopic.php?pid=1568197#p1568197

graysky commented on 2015-10-05 19:22

Bump to v4.1.10-1

Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.1.10
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

kyak commented on 2015-10-02 11:38

Hi graysky,
The repo download speed is doing pretty badly. Even with several Server lines in pacman.conf. Even worse, the download doesn't resume from the same place.

At maximum i was able to download around 1000 Kb, and the download starts over and over again from the beginning of the file.

It's been like that for a couple of days now (can't update to v4.1.9-1).

It would be great if you could have a look. Thank you!

graysky commented on 2015-09-29 19:33

Bump to v4.1.9-1

Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.1.9
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2015-09-22 00:49

Bump to v4.1.8-1

Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.1.8
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2015-08-31 05:29

@ali - Why out of date? No BFQ for 4.2 and no ck1 for 4.2.

NullDivision commented on 2015-08-30 18:30

Yeah. Sorry. My mistake. Forgot to add the key ID to the chain. ^.^

sir_lucjan commented on 2015-08-30 17:19

@NullDivision
No.

NullDivision commented on 2015-08-30 17:02

It would appear that the GPG signatures for linux-4.1.tar and patch-4.1.6 are out of date.

Tjuh commented on 2015-08-24 14:35

Yes, I do. That's interesting, I'm gonna recompile it with localmodconf disabled. Thanks.

marceliq commented on 2015-08-24 08:01

@Tjuh: Do you use localmodconf with modprobed-db? Because I found out that if I don't use it then it works without panics.

Tjuh commented on 2015-08-23 15:26

I am regularly getting a kernel panic, with and without NUMA disabled :S

graysky commented on 2015-08-17 21:58

Bump to v4.1.6-1

Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.1.6
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2015-08-11 18:38

Bump to v4.1.5-1

Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.1.5
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

artafinde commented on 2015-08-11 08:20

This should work https://gist.github.com/spyropoulosl/2e1abf1c5cdb6f898864 (it builds but didn't boot it yet)

FadeMind commented on 2015-08-11 06:00

4.1.5 released
https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.1.5

graysky commented on 2015-08-09 07:23

Bump to v4.1.4-2

Changelog: Bump to ck2 patchset
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

nemesys commented on 2015-08-08 03:16

I consistently disable NUMA on my Lenovo T420 and have not had any recognizable issues. I also customize my kernel pretty heavily to remove support for things I don't use or need.

artafinde commented on 2015-08-06 22:53

I'm running with NUMA disabled and with ck patch for reverting unplugged [1]. I'm fairly stable with BFQ enabled and on BTRFS. I'll post if I got issues. Test: Normal work day, scrubs on BTRFS, kernel compilation.
[1] http://ck-hack.blogspot.co.uk/2015/08/bfs-463-41-ck1.html?showComment=1438818814282#c3678691556058830907

dkaylor commented on 2015-08-05 10:53

I may have spoken too soon. System had been up and running for about 30 hours with no trouble. Then did a pacman -Syu (with linux-4.1.4 and systemd-224) and got a panic right in the middle of the kernel image generation. Reboot, cold boot, nothing helped. Panic then hang. And generic kernel image generation didn't finish, so system was toast. Will try a repair with ISO before, but wiki instructions seem dodgy. Probably should have saved pacman work for the generic kernel, lesson learned.

marceliq commented on 2015-08-05 08:50

I have NUMA disabled and today kernel panic appears during boot process. After hard reset it booted without problems.

artafinde commented on 2015-08-05 08:28

I've recompiled with disabled NUMA. After half a dozen reboots no panic so far.

pedrogabriel commented on 2015-08-05 04:24

@graysky - The kernel is working fine in my system with NUMA disabled.

dkaylor commented on 2015-08-04 23:52

Additional to my previous post: SMT is turned off in my config, since processor is 4 cores == 4 threads. Doubt it's relevant, but just in case.

graysky commented on 2015-08-04 23:03

@dk - That's good to hear. I guess we should solicit others to try disabling NUMA and checking for stability. Any takers?

dkaylor commented on 2015-08-04 22:13

Regarding NUMA - I've been running a custom linux-ck 4.1.4-1 with NUMA disabled and haven't had any hangs. Everyday workloads, several suspend/resume cycles. On the same hardware, I had to turn NUMA off on 4.0.x to get it to run. Machine is Core2 Quad.

artafinde commented on 2015-08-04 14:00

Maybe it's better to follow up at forums [1]
[1]:https://bbs.archlinux.org/viewtopic.php?pid=1550487#p1550487

marceliq commented on 2015-08-04 12:29

Thanks FadeMind.
Sorry, I didnt notice that. Somehow it worked in previous versions.

Now I have same problem as artafinde, but without nouveau :)

FadeMind commented on 2015-08-04 09:57

marceliq

modprobed_db use _ instead of - for naming modules.

You mixed up this.

Corrected modules list:

dm_bio_prison
dm_bufio
dm_cache
dm_cache_mq
dm_crypt
dm_log
dm_mirror
dm_mod
dm_persistent_data
dm_region_hash
dm_snapshot

marceliq commented on 2015-08-04 09:31

Hi,

I run into some trouble. Im using localmodconf. In my modprobed.db file I can see modules:

dm_bio_prison
dm_bufio
dm-cache
dm-cache-mq
dm_crypt
dm_log
dm_mirror
dm_mod
dm_persistent_data
dm_region_hash
dm-snapshot

But after compilation, while installing new package when building new init image I get this:
>>> Updating module dependencies. Please wait ...
>>> Generating initial ramdisk, using mkinitcpio. Please wait...
==> Building image from preset: /etc/mkinitcpio.d/linux-ck.preset: 'default'
-> -k /boot/vmlinuz-linux-ck -c /etc/mkinitcpio.conf -g /boot/initramfs-linux-ck.img
==> Starting build: 4.1.4-1-ck
-> Running build hook: [base]
-> Running build hook: [udev]
-> Running build hook: [autodetect]
-> Running build hook: [modconf]
-> Running build hook: [block]
-> Running build hook: [encrypt]
-> Running build hook: [lvm2]
==> ERROR: module not found: `dm-snapshot'
==> ERROR: module not found: `dm-mirror'
==> ERROR: module not found: `dm-cache'
==> ERROR: module not found: `dm-cache-mq'
-> Running build hook: [filesystems]
-> Running build hook: [keyboard]
-> Running build hook: [fsck]
==> Generating module dependencies
==> Creating lzop-compressed initcpio image: /boot/initramfs-linux-ck.img
==> WARNING: errors were encountered during the build. The image may not be complete.

Any suggestions?

Thanks.

graysky commented on 2015-08-03 19:51

Bump to v4.1.4-1

Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.1.4
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2015-08-03 08:42

Post to ck's blog.

artafinde commented on 2015-08-03 07:41

I had a kernel panic while shutdown with latest build [1]. This is build from AUR with BFQ, nouveau and localmodconfig.
[1]: https://ptpb.pw/Mn2Z

graysky commented on 2015-08-02 22:49

Just enable the _makenconfig in the PKGBUILD and build it. You'll need to select the CPU from the nconfig therein.

artafinde commented on 2015-08-02 21:48

@ummakybes: If you have configured makepkg accordingly yes, see [1].
[1]: https://wiki.archlinux.org/index.php/Makepkg#Creating_optimized_packages

ummakynes commented on 2015-08-02 21:27

Does this PKGBUILD automatically apply a CPU architecture optimization?

graysky commented on 2015-08-02 17:20

Bump to v4.1.3-1

LinuxChanges: http://kernelnewbies.org/Linux_4.1
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2015-08-02 17:19

No idea...

th3voic3 commented on 2015-08-02 14:42

Recently I've been having problems using steamos-xpad-dkms with linux-ck. I get an error: "unknown symbol xpad-send-led-command"

graysky commented on 2015-07-24 22:50

Bump to v4.0.9-2

Changelog: Rebuild against gcc 5.2.0-1.
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2015-07-24 22:50

Bump to v4.0.9-1

Changelog: Rebuild against gcc 5.2.0-1.
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2015-07-22 08:02

Bump to v4.0.9-1

Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.0.9
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2015-07-17 19:12

@spotfire - how ood? No ck patch set means not ood.

Gacha commented on 2015-07-07 19:55

When you see "ERROR: One or more PGP signatures could not be verified!" then import the required keys with: gpg --keyserver keys.gnupg.net --recv-keys ABAF11C65A2970B130ABE3C479BE3E4300411886 647F28654894E3BD457199BE38DBBDC86092693e

graysky commented on 2015-06-30 19:30

Bump to v4.0.7-1

Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.0.7
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

zoopp commented on 2015-06-25 17:28

I might have found something regarding the Kernel Panic.

I usually compile (with the NUMA patch enabled) using native optimizations detected by gcc and I disable kernel debugging. Today I updated and forgot to disable kernel debugging and for the first time since this issue arose I got the kernel panic. I then recompiled with kernel debugging disabled and the kernel panic was gone.

Not sure if this is going to be of any use but I'll leave it here just in case.

graysky commented on 2015-06-23 02:05

Bump to v4.0.6-1

Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.0.6
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2015-06-08 22:04

Bump to v4.0.5-2

Changelog: update BFQ to v7r8.
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

Det commented on 2015-06-08 05:39

http://algo.ing.unimo.it/people/paolo/disk_sched/patches/4.0.0-v7r8/

graysky commented on 2015-06-06 16:23

Bump to v4.0.5-1

Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.0.5
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

FadeMind commented on 2015-06-06 15:53

4.0.5 https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.0.5

graysky commented on 2015-05-22 17:37

Bump to v4.0.4-3

Changelog: sync up with ARCH (ext4 data corruption patch)
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2015-05-20 20:11

Bump to v4.0.4-2

Changelog: sync up config with ARCH trunk.
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2015-05-17 19:51

Bump to v4.0.4-1

Changelog: https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.0.4
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

FadeMind commented on 2015-05-17 19:34

Bump 4.0.4

https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.0.4

sir_lucjan commented on 2015-05-14 07:25

Changelog:

https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.0.3

ishitatsuyuki commented on 2015-05-14 07:22

Your version bot f*ck out on changelog link

graysky commented on 2015-05-13 19:32

Bump to v4.0.3-1

Changelog: https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-4.0.3
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2015-05-12 20:05

Bump to v4.0.2-2

Comment: 4.0.2-2 has been reported by 6 users now to be stable (running for 3-4 days under various conditions). Please post to the repo-ck discussion thread if you experience a problem.
Changelog: https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-4.0.2
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2015-05-12 20:04

Bump to v4.0.2-2

Changelog: https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-4.0.2
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2015-05-07 00:31

Bump to v3.19.7-1

Changelog: https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.19.7
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2015-05-06 19:02

Yes, it's in the description as you know so it's a cosmetic mistake. I have corrected it and will update when 3.19.7 goes live. Thank you.

qqqqqqqqq9 commented on 2015-05-06 13:00

actually it should read modprobed-db

qqqqqqqqq9 commented on 2015-05-06 12:58

Hi
there's a d missing in line 201:
modprobe_db should read modprobed_db

graysky commented on 2015-05-05 19:36

See the official repo-ck thread for a discussion and summary of debugging with CK's help.

XenGi commented on 2015-05-05 11:55

@gravsky Sorry, clicked before thinking. Waiting for 4.0-ck to arrive. Take your time. Maybe look at linux-pf. Seems they have fixed some bfq problems in 4.0.

graysky commented on 2015-05-05 09:08

@XenGi - why out of date?

graysky commented on 2015-04-29 19:34

Bump to v3.19.6-1

Changelog: https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.19.6
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2015-04-29 19:33

OK. Given all the reports of problems with ck1 and 4.0.x, I have decided to push version 3.19.6 to the AUR (and all associated packages now depend on 3.19.x). If you experienced a kernel panic using linux-ck v4.0.x, please see this post and consider if you can help Con debug: https://bbs.archlinux.org/viewtopic.php?pid=1524428#p1524428

For those interested in the 4.0.x version of PKGBUILDs, I have made them available on repo-ck at this address: http://repo-ck.com/PKG_source/next

PLEASE DO NOT FLAG OUT-OF-DATE UNTIL CK FIXES THE PANICS - I WILL NOT PUSH THE AUR OR REPO UNTIL IT IS STABLE!

graysky commented on 2015-04-26 09:15

Previous versions are available from: http://repo-ck.com/old

May I ask that those of you experiencing problems like this please follow and post in the repo-ck thread in the bbs? Here is where the recent discussion about 4.0 starts: https://bbs.archlinux.org/viewtopic.php?pid=1523304#p1523304

graysky commented on 2015-04-26 09:06

Previous versions are available from: http://repo-ck.com/old

Please report these issues to Con on his blog or via email: http://ck-hack.blogspot.com

rat commented on 2015-04-26 08:31

I have the same situation, panic of kernel during the boot,

Where I can find the previous version of linux-ck 3.19.5 ??

rat commented on 2015-04-26 08:30

I have the same situation, panic of kernel during the boot,

Where I can find the precarious version of linux-ck 3.19.5 ??

tomaszc commented on 2015-04-26 07:25

After update linux-ck-nehalem x86_64 from 3.19.5-1 to 4.0-1 there is a kernel panic at boot or later.

linux 4.0.0-2-ARCH from testing repo working properly,

Jak zdobędę więcej informacji to umieszczę je tutaj,

pedrogabriel commented on 2015-04-25 15:34

Why doesn't it provide the linux-ck-headers?

graysky commented on 2015-04-24 19:48

Bump to v4.0-1

Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog
LinuxChanges: http://kernelnewbies.org/Linux_4.0

graysky commented on 2015-04-23 00:02

BFQ for 4.0 is now released. For those wishing to play with version 4.0 now, see the source here: http://repo-ck.com/PKG_source/next/linux-ck-4.0-1.src.tar.gz

I plan to upload to the AUR shortly (24 h or so) due to updates to other packages.

brikler commented on 2015-04-22 14:33

i'am not able to build the ck kernel. there was never a problem since 3.12 up to 3.19.3...

=> kernel/Makefile:132: *** No X.509 certificates found ***
my problem in short: how can i get the needed keys?

[tom@frija linux-ck]$ makepkg -rsie --skippgpcheck
==> Making package: linux-ck 3.19.5-1 (Tue Apr 21 20:24:34 CEST 2015)
==> Checking runtime dependencies...
==> Checking buildtime dependencies...
==> WARNING: Using existing src/ tree
==> Removing existing pkg/ directory...
==> Starting build()...
CHK include/config/kernel.release
CHK include/generated/uapi/linux/version.h
CHK include/generated/utsrelease.h
CALL scripts/checksyscalls.sh
CHK include/generated/compile.h
SKIPPED include/generated/compile.h
kernel/Makefile:132: *** No X.509 certificates found ***
LD arch/x86/realmode/rm/realmode.elf
arch/x86/realmode/rm/header.o: In function `real_mode_header':
(.header+0x8): undefined reference to `pa_trampoline_start'
arch/x86/realmode/rm/header.o: In function `real_mode_header':
(.header+0xc): undefined reference to `pa_trampoline_status'
arch/x86/realmode/rm/header.o: In function `real_mode_header':
(.header+0x10): undefined reference to `pa_trampoline_header'
arch/x86/realmode/rm/header.o: In function `real_mode_header':
(.header+0x14): undefined reference to `pa_trampoline_pgd'
arch/x86/realmode/rm/header.o: In function `real_mode_header':
(.header+0x18): undefined reference to `pa_wakeup_start'
arch/x86/realmode/rm/header.o: In function `real_mode_header':
(.header+0x1c): undefined reference to `pa_wakeup_header'
arch/x86/realmode/rm/header.o: In function `real_mode_header':
(.header+0x20): undefined reference to `pa_machine_real_restart_asm'
arch/x86/realmode/rm/trampoline_64.o: In function `trampoline_start':
/tmp/makepkg/linux-ck/src/linux-3.19/arch/x86/realmode/rm/trampoline_64.S:77: undefined reference to `pa_startup_32'
arch/x86/realmode/rm/trampoline_64.o: In function `startup_32':
/tmp/makepkg/linux-ck/src/linux-3.19/arch/x86/realmode/rm/trampoline_64.S:95: undefined reference to `pa_tr_cr4'
/tmp/makepkg/linux-ck/src/linux-3.19/arch/x86/realmode/rm/trampoline_64.S:99: undefined reference to `pa_trampoline_pgd'
/tmp/makepkg/linux-ck/src/linux-3.19/arch/x86/realmode/rm/trampoline_64.S:103: undefined reference to `pa_tr_efer'
/tmp/makepkg/linux-ck/src/linux-3.19/arch/x86/realmode/rm/trampoline_64.S:104: undefined reference to `pa_tr_efer'
/tmp/makepkg/linux-ck/src/linux-3.19/arch/x86/realmode/rm/trampoline_64.S:118: undefined reference to `pa_startup_64'
arch/x86/realmode/rm/trampoline_64.o: In function `tr_gdt':
(.rodata+0x2): undefined reference to `pa_tr_gdt'
arch/x86/realmode/rm/reboot.o: In function `machine_real_restart_asm':
/tmp/makepkg/linux-ck/src/linux-3.19/arch/x86/realmode/rm/reboot.S:27: undefined reference to `pa_tr_gdt'
/tmp/makepkg/linux-ck/src/linux-3.19/arch/x86/realmode/rm/reboot.S:33: undefined reference to `pa_machine_real_restart_paging_off'
arch/x86/realmode/rm/reboot.o: In function `machine_real_restart_paging_off':
/tmp/makepkg/linux-ck/src/linux-3.19/arch/x86/realmode/rm/reboot.S:46: undefined reference to `pa_machine_real_restart_idt'
/tmp/makepkg/linux-ck/src/linux-3.19/arch/x86/realmode/rm/reboot.S:53: undefined reference to `pa_machine_real_restart_gdt'
arch/x86/realmode/rm/reboot.o: In function `machine_real_restart_gdt':
(.rodata+0x12): undefined reference to `pa_machine_real_restart_gdt'
arch/x86/realmode/rm/wakeup_asm.o: In function `wakeup_gdt':
(.rodata+0x2): undefined reference to `pa_wakeup_gdt'
arch/x86/realmode/rm/Makefile:50: die Regel für Ziel „arch/x86/realmode/rm/realmode.elf“ scheiterte
make[3]: *** [arch/x86/realmode/rm/realmode.elf] Fehler 1
arch/x86/realmode/Makefile:18: die Regel für Ziel „arch/x86/realmode/rm/realmode.bin“ scheiterte
make[2]: *** [arch/x86/realmode/rm/realmode.bin] Fehler 2
scripts/Makefile.build:402: die Regel für Ziel „arch/x86/realmode“ scheiterte
make[1]: *** [arch/x86/realmode] Fehler 2
make[1]: *** Es wird auf noch nicht beendete Prozesse gewartet...
CC security/integrity/iint.o
Makefile:943: recipe for target 'arch/x86' failed
make: *** [arch/x86] Error 2
make: *** Waiting for unfinished jobs....
CHK kernel/config_data.h
CC security/keys/gc.o
LD security/integrity/integrity.o
LD security/integrity/built-in.o
CC security/yama/yama_lsm.o
CC security/keys/key.o
CC security/keys/keyring.o
CC security/keys/keyctl.o
LD security/yama/yama.o
LD security/yama/built-in.o
CC security/commoncap.o
CC security/keys/permission.o
CC security/keys/process_keys.o
CC security/keys/request_key.o
CC security/keys/request_key_auth.o
CC security/min_addr.o
CC security/keys/user_defined.o
CC security/security.o
CC security/keys/compat.o
CC security/keys/proc.o
CC security/keys/sysctl.o
CC security/capability.o
CC security/keys/persistent.o
CC security/keys/big_key.o
CC security/inode.o
CC security/device_cgroup.o
LD security/keys/built-in.o
LD security/built-in.o
==> ERROR: A failure occurred in build().
Aborting...

graysky commented on 2015-04-19 14:36

Bump to v3.19.5-1

Changelog: https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.19.5
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

FadeMind commented on 2015-04-19 09:08

3.19.5
https://git.kernel.org/cgit/linux/kernel/git/stable/stable-queue.git/commit/?id=85a3a82533565abdf0f1550e28cafd1d5732286a

graysky commented on 2015-04-18 19:27

@Buddlespit - Why out-of-date? See my comment from 2015-04-16 21:11.

graysky commented on 2015-04-17 19:14

@fademind - dunno what to tell you; running fine on my systems (haswell and ivy with onboard video).

@all - heard back from Paolo about BFQ for 4.0. Arianna and he are testing a new patch that is specifically for 4.0 so I don't plan to bump linux-ck until that happens. You can continue using the version I have linked below if you wish until then.

FadeMind commented on 2015-04-17 19:11

@graysky
After disabing zramswap.

Boot, sddm splash, loading desktop, no wallpaper - black background, freeze mouse and keyboard, capslock key led flash in pulse.

I can't do anything except hard poweroff. Keyboard and mouse not respond.

Kernel panic is before SDDM startup...

IMO it is Intel DRM regression... weird.

graysky commented on 2015-04-17 18:53

@fademind - I dunno about your setup but I see somr zram lines in your dmesg. Disable zram and try again? Also, if you boot to the shell and not to sddm, does it panic? If not, can you start your xsession without sddm by running `xinit` and see if it happens. Trying to isolate which part is causing the hang.

FadeMind commented on 2015-04-17 09:59

@graysky
Here it is: http://ck-hack.blogspot.com/2015/04/bfs-462-linux-40-ck1.html?showComment=1429264555246#c9059219542602497285

Linux 3.19.4-ck boot fine. Here is dmesg from it: https://pastebin.com/pKitMCcK
Lsmod: https://pastebin.com/zwm3gfHS
inxi -Fxz https://pastebin.com/GQiRMaER

I can't boot fine on 4.0-ck anyway

Regards

graysky commented on 2015-04-17 09:16

@fademind - what video are you using and please report your findings on CK's blog

FadeMind commented on 2015-04-17 05:25

@graysky

I'm used Your SRC. After compilling, reboot and boot from linux-ck I have freeze under X11. I saw mouse and SDDM login splash begin animation but hangs. Only hard poweroff (push 5 sec. power button) was usefull for poweroff.

After power on and boot from stock kernel I had Recovery Jornal again.

It is something wrong with CK for 4.0.

I patched -mainline kernel with GCC patch and working fine - without any issues.

For SSD I used NOOP. For HDD I used CFQ (in stock) or BFQ (in -ck, -mainline) and I don't had Issues. Dynamic changing scheduler udev rules script.

Note: I booted from 3.19.4-ck kernel and PASS fine.

graysky commented on 2015-04-16 21:11

I generated a PKGBUILD+related files for linux-ck-4.0 using ck1-4.0 and using the 3.19.x series of BFQ since there is one report on the BFQ forum that this version works fine in the 4.0 code[1].

You may download it and build/test from this file: http://repo-ck.com/PKG_source/next/linux-ck-4.0-0.src.tar.gz

I also emailed Paolo to see if this is the official word from his group. I don't want to put linux-ck-4.0 live in the AUR/repo until we hear back from him. Please test this out if you wish and report back good or bad.

1. https://groups.google.com/forum/#!topic/bfq-iosched/lUAAQQG6P4A

graysky commented on 2015-04-16 20:57

@FadeMind - So the only difference is ck1 patchset? Are you using BFQ at all?

FadeMind commented on 2015-04-16 14:53

I built own kernel 4.0 with lastest ck patch and I had random kernel panic during boot with hangs. After reboot I had recovery journal of disk. Sometimes is fine, sometimes just hangs and blow up.

Safe is waiting for 4.0.1 release update 4.0 kernel series.

ON Stock 4.0 sersies ARCH kernel this never happened.

PerfectGentleman commented on 2015-04-16 13:33

yep, BFQ for 3.19 works fine on 4.0

AnAkkk commented on 2015-04-16 10:03

Apparently BFQ for 3.19 works just fine on 4.0 according to their mailing list.

AnAkkk commented on 2015-04-16 10:03

Apparently BFQ 3.19 works just fine on 4.0 according to their mailing list.

graysky commented on 2015-04-16 10:02

I saw it, butBFQ has not yet been released for v4.0. Please flag again once that happens if I don't beat you to it.

ishitatsuyuki commented on 2015-04-16 07:45

http://ck.kolivas.org/patches/4.0/4.0/4.0-ck1/
It's coming this time.

graysky commented on 2015-04-14 19:54

Bump to v3.19.4-1

Changelog: https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.19.4
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

brikler commented on 2015-04-14 15:04

the 3.19.4 pkgbuild doesn't work for me

[code]LC_ALL=C makepkg -rsie --skippgpcheck
==> Making package: linux-ck 3.19.4-1 (Tue Apr 14 17:01:00 CEST 2015)
==> Checking runtime dependencies...
==> Checking buildtime dependencies...
==> WARNING: Using existing src/ tree
==> Removing existing pkg/ directory...
==> Starting build()...
CHK include/config/kernel.release
CHK include/generated/uapi/linux/version.h
CHK include/generated/utsrelease.h
CALL scripts/checksyscalls.sh
CHK include/generated/compile.h
kernel/Makefile:132: *** No X.509 certificates found ***
CHK kernel/config_data.h
LD arch/x86/realmode/rm/realmode.elf
arch/x86/realmode/rm/header.o: In function `real_mode_header':
(.header+0x8): undefined reference to `pa_trampoline_start'
arch/x86/realmode/rm/header.o: In function `real_mode_header':
(.header+0xc): undefined reference to `pa_trampoline_status'
arch/x86/realmode/rm/header.o: In function `real_mode_header':
(.header+0x10): undefined reference to `pa_trampoline_header'
arch/x86/realmode/rm/header.o: In function `real_mode_header':
(.header+0x14): undefined reference to `pa_trampoline_pgd'
arch/x86/realmode/rm/header.o: In function `real_mode_header':
(.header+0x18): undefined reference to `pa_wakeup_start'
arch/x86/realmode/rm/header.o: In function `real_mode_header':
(.header+0x1c): undefined reference to `pa_wakeup_header'
arch/x86/realmode/rm/header.o: In function `real_mode_header':
(.header+0x20): undefined reference to `pa_machine_real_restart_asm'
arch/x86/realmode/rm/trampoline_64.o: In function `trampoline_start':
/tmp/makepkg/linux-ck/src/linux-3.19/arch/x86/realmode/rm/trampoline_64.S:77: undefined reference to `pa_startup_32'
arch/x86/realmode/rm/trampoline_64.o: In function `startup_32':
/tmp/makepkg/linux-ck/src/linux-3.19/arch/x86/realmode/rm/trampoline_64.S:95: undefined reference to `pa_tr_cr4'
/tmp/makepkg/linux-ck/src/linux-3.19/arch/x86/realmode/rm/trampoline_64.S:99: undefined reference to `pa_trampoline_pgd'
/tmp/makepkg/linux-ck/src/linux-3.19/arch/x86/realmode/rm/trampoline_64.S:103: undefined reference to `pa_tr_efer'
/tmp/makepkg/linux-ck/src/linux-3.19/arch/x86/realmode/rm/trampoline_64.S:104: undefined reference to `pa_tr_efer'
/tmp/makepkg/linux-ck/src/linux-3.19/arch/x86/realmode/rm/trampoline_64.S:118: undefined reference to `pa_startup_64'
arch/x86/realmode/rm/trampoline_64.o: In function `tr_gdt':
(.rodata+0x2): undefined reference to `pa_tr_gdt'
arch/x86/realmode/rm/reboot.o: In function `machine_real_restart_asm':
/tmp/makepkg/linux-ck/src/linux-3.19/arch/x86/realmode/rm/reboot.S:27: undefined reference to `pa_tr_gdt'
/tmp/makepkg/linux-ck/src/linux-3.19/arch/x86/realmode/rm/reboot.S:33: undefined reference to `pa_machine_real_restart_paging_off'
arch/x86/realmode/rm/reboot.o: In function `machine_real_restart_paging_off':
/tmp/makepkg/linux-ck/src/linux-3.19/arch/x86/realmode/rm/reboot.S:46: undefined reference to `pa_machine_real_restart_idt'
/tmp/makepkg/linux-ck/src/linux-3.19/arch/x86/realmode/rm/reboot.S:53: undefined reference to `pa_machine_real_restart_gdt'
arch/x86/realmode/rm/reboot.o: In function `machine_real_restart_gdt':
(.rodata+0x12): undefined reference to `pa_machine_real_restart_gdt'
arch/x86/realmode/rm/wakeup_asm.o: In function `wakeup_gdt':
(.rodata+0x2): undefined reference to `pa_wakeup_gdt'
arch/x86/realmode/rm/Makefile:50: die Regel für Ziel „arch/x86/realmode/rm/realmode.elf“ scheiterte
make[3]: *** [arch/x86/realmode/rm/realmode.elf] Fehler 1
arch/x86/realmode/Makefile:18: die Regel für Ziel „arch/x86/realmode/rm/realmode.bin“ scheiterte
make[2]: *** [arch/x86/realmode/rm/realmode.bin] Fehler 2
scripts/Makefile.build:402: die Regel für Ziel „arch/x86/realmode“ scheiterte
make[1]: *** [arch/x86/realmode] Fehler 2
Makefile:943: recipe for target 'arch/x86' failed
make: *** [arch/x86] Error 2
make: *** Waiting for unfinished jobs....
==> ERROR: A failure occurred in build().
Aborting...
[/code]

graysky commented on 2015-04-13 08:48

the point is that this package is not outdated even with the release of linux 4.0 or 4.0.1 etc. since there has not yet been a release of ck1 for the 4.0 codebase.

ishitatsuyuki commented on 2015-04-13 08:30

Sorry, flagged with my mistouch.

FadeMind commented on 2015-04-13 07:46

4.0 will be really STABLE in 4.0.1 release after fixing missed regressions. It has always been. Don't rush.

graysky commented on 2015-04-13 07:00

@cryptoluks - Why out of date? 3.19.4 is not released yet, and there is not ck patchset yet for 4.0.

Scimmia commented on 2015-04-09 23:53

There are other ways, too. First off, you don't have to trust the keys. Second, you can automate downloading them in gpg.conf

--skippgpcheck is a joke. If you're too lazy to figure out how to use gpg, you shouldn't be building your kernel, to say the least.

Twilight_Genesis commented on 2015-04-09 23:47

2 Ways To fix PGP key errors:

(Recommended):
Import the PGP keys in the validpgpkeys array in the PKGBUILD and trust them.
gpg --recv-keys ABAF11C65A2970B130ABE3C479BE3E4300411886 647F28654894E3BD457199BE38DBBDC86092693E

For each key run gpg --edit-key <KEY>, where <KEY> is the PGP key then type trust and hit enter, then 5 and hit enter to trust it ultimately, and finally type quit to quit

(Ignore PGP signatures)(Not recommended)
Have makepkg skip PGP signature checking just pass the --skippgpcheck argument to makepkg

makepkg --skippgpcheck -r -s

Twilight_Genesis commented on 2015-04-09 23:45

2 Ways To fix PGP key errors:

(Recommended):
Import the PGP keys in the validpgpkeys array in the PKGBUILD and trust them.
gpg --recv-keys ABAF11C65A2970B130ABE3C479BE3E4300411886 647F28654894E3BD457199BE38DBBDC86092693E

For each key run gpg --edit-key <KEY>, where <KEY> is the PGP key then type trust and hit enter, then 5 and hit enter to trust it ultimately, and finally type quit to quit


(Ignore PGP signatures)
Have makepkg skip PGP signature checking just pass the --skippgpcheck argument to makepkg

makepkg --skippgpcheck -r -s

Twilight_Genesis commented on 2015-04-09 23:41

2 Ways To fix PGP key errors:

(Recommend):
Import the PGP keys in the validpgpkeys array in the PKGBUILD and trust them.
gpg --recv-keys ABAF11C65A2970B130ABE3C479BE3E4300411886 647F28654894E3BD457199BE38DBBDC86092693E

For each key run gpg --edit-key then type trust and hit enter, then 5 and hit enter to trust it ultimately, and finally type quit to quit


(Ignore PGP signatures)
Have makepkg skip PGP signature checking just pass the --skippgpcheck argument to makepkg

makepkg --skippgpcheck

Twilight_Genesis commented on 2015-04-09 23:19

@radialhat: You have to either import the PGP keys in the validpgpkeys array in the PKGBUILD and trust them or if you don't care about the PGP signatures then you can pass the --skippgpcheck argument to makepkg and have it ignore the PGP signatures.

graysky commented on 2015-04-06 08:49

@alonhar - https://bbs.archlinux.org/viewtopic.php?id=195705

alonhar commented on 2015-04-06 00:43

How can I uninstall linux-ck.. return to the normal kernel and not break arch?

graysky commented on 2015-04-05 12:26

Please take makepkg/gnupg discussions to the forums; I doubt others in the AUR like receiving notifications on these topics which have nothing to do with the package. Thank you.

radialhat commented on 2015-04-05 11:50

What kind of fix do you refer to? I've done nothing to my gnupg setup... It works with other packages

Scimmia commented on 2015-04-05 03:41

@radialhat, if you're having problems, don't work around it. Fix your GnuPG setup.

Works fine here.

radialhat commented on 2015-04-05 01:35

Seriously, what is wrong with the PGP keys? Do we have to workaround everytime now?

graysky commented on 2015-03-30 19:25

@sp1d3rmxn - No idea what you're talking about; PKGBUILD is solid as-is.

dkaylor commented on 2015-03-30 12:25

@sp1d3rmxn Uh.... What?

sp1d3rmxn commented on 2015-03-30 11:45

Funny you all are having problems with the pgp key. I brought this up some time ago and was given a bunch of BS about why it "should" work as if I already didn't know it "should".

Anyways... I've been seeing this wonderful additional problem with the "enable patch" that for some reason is also being disregarded so here's how to fix it.

When you compile the package it's nice to see someone has left or did something stupid (yet again) where the file "enable_additional_cpu_optimizations_for_gcc_v4.9+_kernel_v3.15+.patch.gz" is actually not ending in ".gz". Awesome huh?

So what you have to do is this. Rename the file removing the ".gz" part and copy it to the "src" directory where you're building (you can't do this through the yaourt build because you need to intervene beyond editing the PKGBUILD).

Recap:

- run makepkg -rs (if this is how you do it)

- ignore errors since this is becoming commonplace with this package anyway

- edit PKGBUILD and remove .sig and .sign PGP crap since no one wants to implement this right anyway

- mv enable_additional_cpu_optimizations_for_gcc_v4.9+_kernel_v3.15+.patch.gz enable_additional_cpu_optimizations_for_gcc_v4.9+_kernel_v3.15+.patch

- update sha256sums => (4th line down) "e5b0f1882861cb1d0070e49c41a3e7390cf379f183be0bddb54144ecc68fb8f9"

- cp enable_additional_cpu_optimizations_for_gcc_v4.9+_kernel_v3.15+.patch to ./src (you should be in the directory you're compiling)

- Sit back and wait for fresh kernel and enjoy

Tjuh commented on 2015-03-28 14:01

Unfortunately none of the suggestions below worked, I had to run makepkg with the --skippgpcheck flag to get it to compile.

sir_lucjan commented on 2015-03-27 14:07

1. nano /home/tjuh/.gnupg/gpg.conf

2. Add 'keyserver hkp://keys.gnupg.net'

3. Try again

Tjuh commented on 2015-03-27 13:57

tjuh ~ $ gpg --keyserver hkp://keys.gnupg.net --recv-keys 00411886 6092693E
gpg: directory '/home/tjuh/.gnupg' created
gpg: new configuration file '/home/tjuh/.gnupg/gpg.conf' created
gpg: WARNING: options in '/home/tjuh/.gnupg/gpg.conf' are not yet active during this run
gpg: keybox '/home/tjuh/.gnupg/pubring.kbx' created
gpg: keyserver receive failed: No keyserver available

Same warning :S

FadeMind commented on 2015-03-27 13:12

Execute this command:
gpg --keyserver hkp://keys.gnupg.net --recv-keys 00411886 6092693E

Scimmia thanks for tip.

FadeMind commented on 2015-03-27 13:10

Good to know this. Thanks Scimmia.

Scimmia commented on 2015-03-27 12:36

You shouldn't need to sign them.

FadeMind commented on 2015-03-27 12:29

Execute this command:
gpg --keyserver hkp://keys.gnupg.net --recv-keys 00411886 6092693E
gpg --lsign ABAF11C65A2970B130ABE3C479BE3E4300411886
gpg --lsign 647F28654894E3BD457199BE38DBBDC86092693E

Tjuh commented on 2015-03-27 11:45

==> Verifying source file signatures with gpg...
linux-3.19.tar ... FAILED (unknown public key 79BE3E4300411886)
patch-3.19.3 ... FAILED (unknown public key 38DBBDC86092693E)
==> ERROR: One or more PGP signatures could not be verified!
==> Removing installed dependencies...
checking dependencies...

Packages (1) bc-1.06.95-1

Total Removed Size: 0.18 MiB

:: Do you want to remove these packages? [Y/n] y
(1/1) removing bc [######################] 100%
tjuh ~/Downloads/linux-ck $ gpg --recv-keys 79BE3E4300411886
gpg: keyserver receive failed: No keyserver available
tjuh ~/Downloads/linux-ck $ gpg --recv-keys 38DBBDC86092693E
gpg: keyserver receive failed: No keyserver available

Anyway to fix this? I tried running dirmngr </dev/null as root and then those 2 gpg commands again, but the keyserver is still not available.

Anonymous comment on 2015-03-26 21:21

Thanks Sir_Lucjan, Graysky and Sir.Tiddlesworth -- removed my comments for errors, it was my mistakes. Fixed download interruption using tip on repo-ck wiki and had to add keys as root (not sure why using sudo wasn't working).

Finally I am on ck-kx :)

sir_lucjan commented on 2015-03-26 21:19

@jackpot

as normal user

sir_lucjan commented on 2015-03-26 21:19

jackpot

Anonymous comment on 2015-03-26 21:18

#dirmngr </dev/null

#gpg --recv-keys ABAF11C65A2970B130ABE3C479BE3E4300411886
#gpg --recv-keys 647F28654894E3BD457199BE38DBBDC86092693E

gpg: keyserver receive failed: No keyserver available

graysky commented on 2015-03-26 21:17

Bump to v3.19.3-1

Changelog: https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.19.3
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

sir_lucjan commented on 2015-03-26 20:50

@jackpot:

gpg --recv-keys ABAF11C65A2970B130ABE3C479BE3E4300411886
gpg --recv-keys 647F28654894E3BD457199BE38DBBDC86092693E

Anonymous comment on 2015-03-26 20:47

Thanks SirTiddle. It worked but repo-ck download speed is pain in the eyes. When doing it via AUR, its again:

==> Verifying source file signatures with gpg...
linux-3.19.tar ... FAILED (unknown public key 79BE3E4300411886)
patch-3.19.3 ... FAILED (unknown public key 38DBBDC86092693E)

:S

Sir.Tiddlesworth commented on 2015-03-26 11:39

@jackpot
Run the following as root, then try again.
# dirmngr </dev/null

Anonymous comment on 2015-03-20 10:59

@graysky: Thanks for reply. Actually I posted after every option I could find failed. When I try to add keys as mentioned in repo-ck.com I get following:

Quote:

gpg: connecting dirmngr at '/root/.gnupg/S.dirmngr' failed: IPC connect call failed
gpg: keyserver receive failed: No dirmngr
==> ERROR: Remote key not fetched correctly from keyserver.

Unquote

This is happening since 2 weeks (i.e. the time I have been trying different things)

graysky commented on 2015-03-19 22:58

@jackpot - See the instructions on http://repo-ck.com which should fix you up.

Anonymous comment on 2015-03-19 22:17

I am unable to add key for repo-ck (server error) and also having similar problem using yaourt. Followed comments of Fademind but again having remote server error.

Please suggest solution. Thanks!

graysky commented on 2015-03-18 22:17

Bump to v3.19.2-1

Changelog: https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.19.2
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2015-03-12 08:42

If I remember correctly, I did this back when I bought a RPi. No real use for it now... In a drawer somewhere :p

ryanvade commented on 2015-03-12 01:34

Can the CK patchset be used on Arm devices?

FadeMind commented on 2015-03-08 12:21

milan385 add these PGP keys:
gpg --recv-keys ABAF11C65A2970B130ABE3C479BE3E4300411886
gpg --recv-keys 647F28654894E3BD457199BE38DBBDC86092693E
gpg --lsign ABAF11C65A2970B130ABE3C479BE3E4300411886
gpg --lsign 647F28654894E3BD457199BE38DBBDC86092693E



sir_lucjan commented on 2015-03-08 12:08

==> Verifying source file signatures with gpg...
linux-3.19.tar ... Passed
patch-3.19.1 ... Passed
==> Entering fakeroot environment...
==> Creating source package...
-> Adding PKGBUILD...
-> Generating .SRCINFO file...
-> Adding config.x86_64...
-> Adding config...
-> Adding linux-ck.preset...
-> Adding change-default-console-loglevel.patch...
-> Adding install file (linux-ck.install)...
-> Compressing source package...
==> Leaving fakeroot environment.
==> Source package created: linux-ck (Sun Mar 8 13:07:38 CET 2015)

milan385 commented on 2015-03-08 11:59

Have those keys in my keyring...

sir_lucjan commented on 2015-03-08 11:52

@milan385

You're wrong.

http://allanmcrae.com/2015/01/two-pgp-keyrings-for-package-management-in-arch-linux/

and

https://wiki.archlinux.org/index.php/Makepkg#Signature_checking

milan385 commented on 2015-03-08 11:51

Now it's stuck in this moment:
"Verifying source file signatures with gpg...
linux-3.19.tar ... "

graysky commented on 2015-03-08 11:29

Grrr... sorry about that. I refreshed the PKGBUILD for the correct checksum. The issue was that it was in fact correct, but at that time I didn't push the local copy to the repo. I then manually did so when pedrogabriel posted but the sum in the PKGBUILD was reindex against the old version. Anyway, it's right now.

clfarron4 commented on 2015-03-08 11:01

@DaMoo @milan385: Have you downloaded the GCC patch fresh? It was updated to reflect the addition of the Silvermont family of processors.

milan385 commented on 2015-03-08 07:04

Same think that DaMoo have. Failed validity check.

Lompik commented on 2015-03-08 02:47

same issue as DaMoo. Changed sha256 as per below:

--- PKGBUILD 2015-03-07 20:44:57.330121134 -0600
+++ PKGBUILD.orig 2015-03-07 20:45:26.940364828 -0600
@@ -78,7 +78,7 @@
'3dbf80df9a81a285baa5188ea8d768110f24a3e4fe8bd37e1c9d7410d60a680b'
'SKIP'
'6d3043360485bbf3b8b6b780d62ff529074489e6a4d0086607de873d1278c031'
- 'deacee3a3d9b06bc2aae74d908cef183dd39c4f3049567c488950f019ec95d79'
+ '5c3067b83d526f02be9173b92451a4b259e6245a1ee22f854cbf75b4001037d1'
'aad4a85e81b26bb7e4f44f7f1d307e812b2d02672673363a8c7acdd1174b99be'
'99993989d38c452388458648ee354679d1e7763e216a06573e0b579cbf787e69'
'2b3ebf5446aa3cac279842ca00bc1f2d6b7ff1766915282c201d763dbf6ca07e'

DaMoo commented on 2015-03-08 01:29

I'm having an issue where a patch is failing to pass a sha256 integrity check. Here's the output from my AUR helper:

==> Validating source files with sha256sums...
linux-3.19.tar.xz ... Passed
linux-3.19.tar.sign ... Skipped
patch-3.19.1.xz ... Passed
patch-3.19.1.sign ... Skipped
patch-3.19-ck1.xz ... Passed
enable_additional_cpu_optimizations_for_gcc_v4.9+_kernel_v3.15+.patch.gz ... FAILED
config.x86_64 ... Passed
config ... Passed
linux-ck.preset ... Passed
change-default-console-loglevel.patch ... Passed
0001-block-cgroups-kconfig-build-bits-for-BFQ-v7r7-3.19.patch ... Passed
0002-block-introduce-the-BFQ-v7r7-I-O-sched-for-3.19.patch ... Passed
0003-block-bfq-add-Early-Queue-Merge-EQM-to-BFQ-v7r7-for-3.19.0.patch ... Passed
==> ERROR: One or more files did not pass the validity check!
:: failed to verify linux-ck integrity

graysky commented on 2015-03-07 16:25

Bump to v3.19.1-1

Comments: This release includes an updated gcc patch which allows for the optimization Silvermont processors.
Changelog: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

pedrogabriel commented on 2015-03-07 16:20

Thank you.

graysky commented on 2015-03-07 16:16

Fixed the checksum. Sorry about that.

pedrogabriel commented on 2015-03-07 15:55

linux-3.19.tar.xz ... Passed
linux-3.19.tar.sign ... Skipped
patch-3.19.1.xz ... Passed
patch-3.19.1.sign ... Skipped
patch-3.19-ck1.xz ... Passed
enable_additional_cpu_optimizations_for_gcc_v4.9+_kernel_v3.15+.patch.gz ... FAILED
config.x86_64 ... Passed
config ... Passed
linux-ck.preset ... Passed
change-default-console-loglevel.patch ... Passed
0001-block-cgroups-kconfig-build-bits-for-BFQ-v7r7-3.19.patch ... Passed
0002-block-introduce-the-BFQ-v7r7-I-O-sched-for-3.19.patch ... Passed
0003-block-bfq-add-Early-Queue-Merge-EQM-to-BFQ-v7r7-for-3.19.0.patch ... Passed
ERROR: One or more files did not pass the validity check!

graysky commented on 2015-03-07 15:52

Bump to v3.19.1-1

Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog
LinuxChanges: http://kernelnewbies.org/Linux_3.19.1

sir_lucjan commented on 2015-03-01 14:25

@dkaylor: http://ck.kolivas.org/patches/3.0/3.19/3.19-ck1/patch-3.19-ck1.xz

dkaylor commented on 2015-03-01 00:01

Downloading from ck.kolivas.org has gone flaky again, can't grab patch-3.18-ck1.bz2

Lompik commented on 2015-02-28 21:34

gpg --recv-keys ABAF11C65A2970B130ABE3C479BE3E4300411886 647F28654894E3BD457199BE38DBBDC86092693E

Solved.

sir_lucjan commented on 2015-02-28 21:27

@Lompik

http://allanmcrae.com/2015/01/two-pgp-keyrings-for-package-management-in-arch-linux/

and

https://wiki.archlinux.org/index.php/Makepkg#Signature_checking

Lompik commented on 2015-02-28 21:24

Hello,

got this error today:

==> Verifying source file signatures with gpg...
linux-3.19.tar ... FAILED (unknown public key 79BE3E4300411886)
==> ERROR: One or more PGP signatures could not be verified!
==> ERROR: Makepkg was unable to build linux-ck.

Rgds

graysky commented on 2015-02-28 00:10

Bump to v3.19-1

Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog
LinuxChanges: http://kernelnewbies.org/Linux_3.19

graysky commented on 2015-02-27 20:54

Bump to v3.18.8-1

Changelog: https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.18.8
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2015-02-27 20:45

@protake - I see it. Just need to get 3.18.8 squared away and update all package for the 3.19 kernel. I will bump shortly.

sflor commented on 2015-02-27 17:02

BFS 461 is now available: http://ck-hack.blogspot.de/2015/02/bfs-461-linux-319-ck1.html

FadeMind commented on 2015-02-27 15:22

hepha. In PKGBUILD of linux-ck all is fine.

[tomasz@arch ~]$ export LC_ALL=C;pacman -Qo modprobed-db
/usr/bin/modprobed-db is owned by modprobed-db 2.26-1
[tomasz@arch ~]$ export LC_ALL=C;pacman -Ql modprobed-db
modprobed-db /usr/
modprobed-db /usr/bin/
modprobed-db /usr/bin/modprobed-db
modprobed-db /usr/bin/modprobed_db
modprobed-db /usr/share/
modprobed-db /usr/share/licenses/
modprobed-db /usr/share/licenses/modprobed-db/
modprobed-db /usr/share/licenses/modprobed-db/LICENSE
modprobed-db /usr/share/man/
modprobed-db /usr/share/man/man8/
modprobed-db /usr/share/man/man8/modprobed-db.8.gz
modprobed-db /usr/share/modprobed-db/
modprobed-db /usr/share/modprobed-db/modprobed-db.skel
modprobed-db /usr/share/zsh/
modprobed-db /usr/share/zsh/site-functions/
modprobed-db /usr/share/zsh/site-functions/_modprobed-db


hepha commented on 2015-02-27 15:16

hello
/usr/bin/modprobed_db not /usr/bin/modprobed-db

if [ -e /usr/bin/modprobed-db ]; then
[[ ! -x /usr/bin/sudo ]] && echo "Cannot call modprobe with sudo. Install via pacman -S sudo and configure to work with this user." && exit 1
sudo /usr/bin/modprobed-db recall
fi

graysky commented on 2015-02-17 20:58

Been following the discussion on the ck blog but am unwilling to update the package with non-ck patches.

AnAkkk commented on 2015-02-17 20:57

Apparently there is a patch to make BFS work on 3.19:
https://bitbucket.org/alfredchen/linux-gc/commits/37f1af26a73c71d70d98104d5f58c250f57a56ee?at=v3.19-gc

Unofficial but CK already used patches from this guy (Alfred Chen) to make it work on new kernels before.

kyak commented on 2015-02-15 12:02

Thanks, graysky!

graysky commented on 2015-02-15 11:27

@kyak - See the repo-ck wiki for advice on dealing with godaddy's shitty internet services. I haven't considered a donate button; this is my way of giving back to the community.

kyak commented on 2015-02-15 06:57

Hi graysky,
When using your repo, download if often aborted with "transfer closed with XXX bytes remaining to read" - due to slow speed.
Previsously it was possible to just "pacman -Syu" again and the download would resume where it had been stopped. But now the download is always started from the beginning. I'm struggling now to download 10 Mb of linux-ck-ivybridge-3.18.7-2 - even for such small size, i had to "pacman -Syu" several times, untill it succeded (the download is always started from the beginning).

Anyway. Did you think about "Donate" button that would probably help you improve the infrastructure a little bit?

graysky commented on 2015-02-14 22:06

Bump to v3.18.7-2

Changelog: Included ck's patch to correct locked plugged IO[1].
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

1. http://ck-hack.blogspot.fi/2014/12/bfs-460-linux-318-ck1.html?showComment=1419271635895#c5781461364809108545

graysky commented on 2015-02-11 20:17

Bump to v3.18.7-1

Changelog: https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.18.7
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

sir_lucjan commented on 2015-02-11 09:11

3.18.7 is out

clfarron4 commented on 2015-02-09 14:30

@khron: Of course, it goes without saying that if you've done this that you should make note that you've done this.

Good thing, as @sir_lucjan points out, is that https://pf.natalenko.name/mirrors/bfq/ hosts all the (current) versions of BFQ, so there's that as well.

khron commented on 2015-02-08 23:07

Easy way, just host your own mirror and use your favourite AUR tool:

$ echo "127.0.0.1 algo.ing.unimo.it" >> /etc/hosts
$ mkdir ~/patches
$ cd ~/patches
$ mkdir -p people/paolo/disk_sched/patches/3.18.0-v7r7/
(download patches from http://repo-ck.com/source/mirror there)
$ python2 -m SimpleHTTPServer 80
$ pacaur -S linux-ck

:)

graysky commented on 2015-02-07 12:30

I am hosting the BFQ patches until their site is back up. Access them like so:

1) Download the linux-ck source tarball, untar it, enter the linux-ck directory
2) sed -i 's|^_bfqpath=.*$|_bfqpath=\"http://repo-ck.com/source/mirror\"|' PKGBUILD

Tjuh commented on 2015-02-07 11:34

The 0001-block-cgroups-kconfig-build-bits-for-BFQ-v7r7-3.18.patch fails to download during makepkg-ing, I've tried it numerous times.

graysky commented on 2015-02-06 20:12

Bump to v3.18.6-1

Changelog: https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.18.6
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

sir_lucjan commented on 2015-02-06 15:33

3.18.6 is out

sir_lucjan commented on 2015-02-02 19:36

@clfarron4:

https://pf.natalenko.name/mirrors/bfq/3.18.0-v7r7/

clfarron4 commented on 2015-02-02 19:34

I think http://algo.ing.unimo.it/ is down again...

graysky commented on 2015-01-30 20:29

Bump to v3.18.5-1

Changelog: https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.18.5
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

clfarron4 commented on 2015-01-28 17:33

@scabdates: Does that happen with the [core] kernel in the repositories as well, or just ck?

scabdates commented on 2015-01-28 15:52

The 3.18.4-1 update caused my init to stop working, just FYI.

graysky commented on 2015-01-27 19:15

Bump to v3.18.4-1

Changelog: https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.18.4
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

Xaero252 commented on 2015-01-18 02:53

@graysky:
Yup, seems to be resolved here as well. Must have been some momentary downtime somewhere in the pipes preventing it. ¯\(°_o)/¯

graysky commented on 2015-01-17 19:49

@Xaero - No problems downloading sources here.

Xaero252 commented on 2015-01-17 18:23

Correction, all files appear to be failing from the mirror in the PKGBUILD, falling back on repo-ck until things are resolved.

Xaero252 commented on 2015-01-17 18:21

One of the patches fails to download from the algo.ing.unimo.it mirror:
==> ERROR: Failure while downloading 0001-block-cgroups-kconfig-build-bits-for-BFQ-v7r7-3.18.patch

Suggest using the repo-ck mirror for this until things are fixed.

graysky commented on 2015-01-17 14:06

https://wiki.archlinux.org/index.php/Makepkg#Signature_checking

prakharsingh95 commented on 2015-01-17 13:38

@Tjuh, http://allanmcrae.com/2015/01/two-pgp-keyrings-for-package-management-in-arch-linux/

Tjuh commented on 2015-01-17 13:15

Trying to build this, but it fails at verifying some files:

==> Verifying source file signatures with gpg...
linux-3.18.tar ... FAILED (unknown public key 79BE3E4300411886)
patch-3.18.3 ... FAILED (unknown public key 38DBBDC86092693E)
==> ERROR: One or more PGP signatures could not be verified!
==> Removing installed dependencies...

graysky commented on 2015-01-17 12:13

Bump to v3.18.3-1

Changelog: https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.18.3
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2015-01-15 21:18

There are known issues with users of 3.18.x and the broadcom-wl-* driver.

petterk commented on 2015-01-15 19:54

Stock kernel (3.18.2) causes major kernel lockups on my Asus laptop.

tried with this kernel, same issues.

Switched to linux-lts.

melanopsis commented on 2015-01-09 22:08

Just to let you know that my desktop no longer locks up with 3.18.2-1-ck

graysky commented on 2015-01-09 21:32

https://wiki.archlinux.org/index.php/Makepkg#Signature_checking

prakharsingh95 commented on 2015-01-09 21:32

@sir_lucjan

Thanks! I came across this article but I was trying the keys present in PKGBUILD. Had to go through comments as well. It worked out.

sir_lucjan commented on 2015-01-09 21:09

@prakharsingh95

http://allanmcrae.com/2015/01/two-pgp-keyrings-for-package-management-in-arch-linux/

prakharsingh95 commented on 2015-01-09 21:05

Hi,

The installation broke with this new tarball.

==> Verifying source file signatures with gpg...
linux-3.18.tar ... FAILED (unknown public key 79BE3E4300411886)
patch-3.18.2 ... FAILED (unknown public key 38DBBDC86092693E)
==> ERROR: One or more PGP signatures could not be verified!

I built the previous tarball for this version, and I have installed that, but not rebooted. I deleted all my previous builds. That should boot right? I mean sig files are just for verification.

graysky commented on 2015-01-09 19:56

Accidentally left the sig files for upstream source and patch commented in the PKGBUILD. I have just remedied this now without bumping the pkgrel.

graysky commented on 2015-01-08 19:50

Bump to v3.18.2-1

Changelog: https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.18.2
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog
Note: This PKGBUILD now conforms to the pacman-4.2 standards of being able to verify pgp signatures of upstream files. See the following for info on how to add the needed keys or how to disable this feature entirely,[1]

1. https://wiki.archlinux.org/index.php/Makepkg#Signature_checking

graysky commented on 2015-01-04 14:21

I am now hosting the BFQ patches until their site is back up. Access them like so:

1) Download the linux-ck source tarball, untar it, enter the linux-ck directory
2) sed -i 's|^_bfqpath=.*$|_bfqpath=\"http://repo-ck.com/source/mirror\"|' PKGBUILD

sir_lucjan commented on 2015-01-04 13:59

https://pf.natalenko.name/mirrors/bfq/3.18.0-v7r7/

sp1d3rmxn commented on 2015-01-04 13:54

I'm lovin it:
---------------------
-> Downloading 0001-block-cgroups-kconfig-build-bits-for-BFQ-v7r7-3.18.patch...
% Total % Received % Xferd Average Speed Time Time Time Current
Dload Upload Total Spent Left Speed
0 0 0 0 0 0 0 0 --:--:-- --:--:-- --:--:-- 0Warning: Transient problem: HTTP error Will retry in 3 seconds. 3 retries
Warning: left.
0 0 0 0 0 0 0 0 --:--:-- --:--:-- --:--:-- 0Warning: Transient problem: HTTP error Will retry in 3 seconds. 2 retries
Warning: left.
0 0 0 0 0 0 0 0 --:--:-- --:--:-- --:--:-- 0Warning: Transient problem: HTTP error Will retry in 3 seconds. 1 retries
Warning: left.
0 0 0 0 0 0 0 0 --:--:-- 0:00:59 --:--:-- 0curl: (22) The requested URL returned error: 503 Service Unavailable
==> ERROR: Failure while downloading 0001-block-cgroups-kconfig-build-bits-for-BFQ-v7r7-3.18.patch
Aborting...

sp1d3rmxn commented on 2015-01-04 13:53

I'm lovin' it:

-> Downloading 0001-block-cgroups-kconfig-build-bits-for-BFQ-v7r7-3.18.patch...
% Total % Received % Xferd Average Speed Time Time Time Current
Dload Upload Total Spent Left Speed
0 0 0 0 0 0 0 0 --:--:-- --:--:-- --:--:-- 0Warning: Transient problem: HTTP error Will retry in 3 seconds. 3 retries
Warning: left.
0 0 0 0 0 0 0 0 --:--:-- --:--:-- --:--:-- 0Warning: Transient problem: HTTP error Will retry in 3 seconds. 2 retries
Warning: left.
0 0 0 0 0 0 0 0 --:--:-- --:--:-- --:--:-- 0Warning: Transient problem: HTTP error Will retry in 3 seconds. 1 retries
Warning: left.

prakharsingh95 commented on 2015-01-03 14:32

Hi, I have been using 3.18.1-3-ck since Dec 26, and I have (thank God) not experienced any lockups, freezes, or wierd stuff. Can anyone elaborate on this issue, or steps to replicate as I would degrade my kernel version. I am using broadcom-wl-ck. I have an i7-3610QM.

prakharsingh95 commented on 2015-01-03 14:29

fackamato commented on 2015-01-03 14:05

Black screen on sandy bridge even with nomodeset on i915. Single mode works. Any ideas?

graysky commented on 2014-12-23 15:59

As with every report of instability: does the corresponding ARCH kernel of the same version cause issues? In this case you will need to install 3.18.1-1-ARCH from [testing] to properly answer this question.

petterk commented on 2014-12-23 14:20

Running linux-ck-haswell with the update to 3.18.1 locks up the cpu bad.

First thought it was the update to overlayfs in PSD that was causing this, but it seems its a kernel issue.

Changed to stock linux solves the issue.

graysky commented on 2014-12-22 18:49

Bump to v3.18.1-3

Changelog: Build against new BFQ patchset (v7r7).
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

Note: I am not yet implimenting the linux-ck equivalent of 'verify source signatures[1]' due to a bug affecting mkaurball[2].
1. https://projects.archlinux.org/svntogit/packages.git/commit/trunk?h=packages/linux&id=2c463587ade65ef28a62f33d54531519f330f351
2. https://bugs.archlinux.org/task/43199

graysky commented on 2014-12-22 18:48

Bump to v3.18.1-3

Changelog: Build against new BFQ patchset (v7r7).
Note: I am not yet implimenting the linux-ck equivalent of 'verify source signatures[1]' due to a bug affecting mkaurball[2].
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

1. https://projects.archlinux.org/svntogit/packages.git/commit/trunk?h=packages/linux&id=2c463587ade65ef28a62f33d54531519f330f351

2. https://bugs.archlinux.org/task/43199

sir_lucjan commented on 2014-12-22 15:09

http://algo.ing.unimo.it/people/paolo/disk_sched/patches/3.18.0-v7r7/

melanopsis commented on 2014-12-19 05:43

@graysky - No, I am not using the broadcom-wl. It is most likely this issue http://www.phoronix.com/scan.php?page=news_item&px=MTg2NzA
http://www.phoronix.com/scan.php?page=news_item&px=MTg1MDc

eduardoeae commented on 2014-12-19 01:50

Mine is an older Intel and also panics. I am using broadcom-wl

graysky commented on 2014-12-19 01:12

@melanopsis - Are you using the broadcom-wl driver by chance? 3.18.1-2-ck is rock solid for me...

melanopsis commented on 2014-12-19 01:11

Just a heads up, my ivy bridge desktop locks up pretty badly with this kernel. I had to downgrade to 3.17.7-1-ck...

graysky commented on 2014-12-18 09:57

Bump to v3.18.1-2

Changelog: Fix FS#43143
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2014-12-17 18:41

Bump to v3.18.1-1

Changelog: https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.18.1
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog
LinuxChanges: http://kernelnewbies.org/Linux_3.18

graysky commented on 2014-12-17 17:28

@mareex - I saw it on my mobile. Thanks. Will make the switch shortly.

mareex commented on 2014-12-17 16:51

linux 3.18.1-1 is in testing now.

graysky commented on 2014-12-16 20:00

Bump to v3.17.7-1

Changelog: https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.17.7
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2014-12-16 19:50

Update3: 3.18.1 has just been released and I have updated the source tarball for those of you wishing to compile it up on your own. A reminder that all of the other *-ck packages I maintain have now been updated and are also available on repo-ck BOTH as source tarballs and (for generic x86_64 only) as compiled packages - note that the compiled package of 3.18.1-ck-1 will be updated shortly.

*Source tarballs: http://repo-ck.com/PKG_source/next/
*Pre-build 3.18-ck packages (x86_64 only): http://repo-ck.com/PKG_source/next/x86_64

graysky commented on 2014-12-15 19:12

Update2: I now have all packages updated to build against 3.18-ck-1 including broadcom-wl-ck. Find the source package for it in the same place as I indicated below. Note - I do not have a broadcom wireless chipset so I cannot test the module. I can modprobe it and it does insert. Can someone with this hardware try it and report back?

graysky commented on 2014-12-14 17:51

Update: All packages except for broadcom-wl-ck have been updated to work with 3.18.0-1-ck. Again, I will wait to diff the linux-ck PKGBUILD and associated files against the official ARCH version once it publishes, but those wishing to upgrade to 3.18.0-1-ck now can either build your own from the src tarballs or use pre-build the packages you can download directly with `pacman -U http://repo-ck.com.../pkg` if you wish. Otherwise, just keep waiting for the official release.

*Source tarballs: http://repo-ck.com/PKG_source/next/
*Pre-build 3.18-ck packages (x86_64 only): http://repo-ck.com/PKG_source/next/x86_64

graysky commented on 2014-12-14 03:05

While we wait the ARCH devs to create a proper PKGBUILD for 3.18.0, I have pieced together one for those wishing to test out the package in the following directory of repo-ck: http://repo-ck.com/PKG_source/next

Note that I have simply done a `make oldconfig` against 3.17.6-1's config files and selected the modules for new features unless that wasn't an option in which case I went with the upstream defaults for new features. Also know that I am noting planning to do an early update to the accompanying packages (broadcom, nvidia*, and vbox*) so if you have a system that depends on them, you will have to modify them yourself until 3.18.x-1-ARCH goes live. Enjoy!

graysky commented on 2014-12-13 15:15

While we wait the ARCH devs to create a proper PKGBUILD for 3.18.0, I have pieced together one for those wishing to test out the package in the following directory of repo-ck: http://repo-ck.com/PKG_source/next

Note that I have simply done a `make oldconfig` against 3.17.6-1's config files and selected the modules for new features unless that wasn't an option in which case I went with the upstream defaults for new features.

Also know that I am noting planning to do an early update to the accompanying packages (broadcom, nvidia*, and vbox*) so if you have a system that depends on them, you will have to modify them yourself until 3.18.x-1-ARCH goes live.

Enjoy!

graysky commented on 2014-12-11 18:19

Got em. Just need to wait for the ARCH dev team to push their config and PKGBUILD etc. into svn. I unflagged it; please flag it out of date again once 3.18-1 is disclosed[1].

1. https://projects.archlinux.org/svntogit/packages.git/log/trunk?h=packages/linux

TheAifam5 commented on 2014-12-11 17:16

http://ck.kolivas.org/patches/3.0/3.18/3.18-ck1/
http://algo.ing.unimo.it/people/paolo/disk_sched/patches/3.18.0-v7r6/

sir_lucjan commented on 2014-12-11 09:52

http://ck-hack.blogspot.de/2014/12/bfs-460-linux-318-ck1.html

graysky commented on 2014-12-11 00:20

Samsagax - why out of date?

clfarron4 commented on 2014-12-10 12:28

@AnAkkk: There's also the issue about which extra kernels should be provided in the main ArchLinux repositories (which is always a fun one).

graysky commented on 2014-12-09 21:08

Did that a while back and was not voted in; limited time for it now... what's wrong with [repo-ck]? I have been maintaining it since 2011. See the wiki page for more if you haven't already.

AnAkkk commented on 2014-12-09 21:03

Can't you ask to become a TU and do that yourself ? :)

graysky commented on 2014-12-09 20:54

Oh... that's not up to me. I believe a TU would need to adopt the package and bring it in toe [community]. No way would he or she build all the CPU versions that I supply in [repo-ck] though...

AnAkkk commented on 2014-12-09 20:53

I think what he is saying is that you could have your linux-ck package in community maybe, like linux-grsec.

graysky commented on 2014-12-09 20:51

The dev team will not incorporate an out-of-tree scheduler into the official distro kernel package. That's why we have the AUR.

dradec commented on 2014-12-09 19:46

Why is this not in main?

graysky commented on 2014-12-08 10:34

Bump to v3.17.6-1

Changelog: https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.17.6
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2014-12-07 12:46

Bump to v3.17.5-1

Changelog: https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.17.5
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2014-11-28 11:59

@Det - You're welcome. Enjoy.

Det commented on 2014-11-28 11:54

Just came to express my appreciation for your beautiful prebuilt packages.

I was already tired of constantly rebuilding my own local kernel, and why should I, when we have such a wonderful man like you.

That's all. Remember you're a force for good.

graysky commented on 2014-11-21 21:40

Bump to v3.17.4-1

Changelog: https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.17.4
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2014-11-18 20:20

Bump to v3.17.3-3

Changelog: Bump to latest ck patchset: ck2
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2014-11-18 20:20

Bump to v3.17.3-3

Changelog: Bump to latest ck patchset: ck2.
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2014-11-18 10:34

All - I know ck2 has been released. I will update this package later this afternoon.

graysky commented on 2014-11-14 23:54

Bump to v3.17.3-2

Changelog: Added fix FS#42689
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2014-11-14 21:36

Bump to v3.17.3-1

Changelog: https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.17.3
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2014-11-11 22:48

Bump to v3.17.2-1


Changelog: https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.17/2
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog
LinuxChanges: http://kernelnewbies.org/Linux_3.17

jbanks commented on 2014-11-11 12:45

linux-3.17-ck1 is out.

graysky commented on 2014-10-30 18:53

Bump to v3.16.7-1

Changelog: https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.16.7
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2014-10-23 19:35

Dunno how long as CK does this in his spare time. For reference:
BFS for v3.16 = 13 days
BFS for v3.15 = 26 days
BFS for v3.14 = 36 days
BFS for v3.13 = 18 days
BFS for v3.12 = 15 days
BFS for v3.11 = 7 days
BFS for v3.10 = 10 days
BFS for v3.9 = 8 days
BFS for v3.8 = 13 days
BFS for v3.7 = 4 days
BFS for v3.6 = 12 days
BFS for v3.5 = 10 days
BFS for v3.4 = 12 days

AnAkkk commented on 2014-10-23 19:28

I guess we'll have to wait a long time then :/

graysky commented on 2014-10-23 19:27

I'll update to the 3.17 series once ck officially releases the patch. I don't think many users would want a user-created bfs.

AnAkkk commented on 2014-10-23 09:35

There's a patch to make BFS work on 3.17:
http://ck-hack.blogspot.fr/2014/08/bfs-453454455456-and-316-ck2.html?showComment=1413266570176#c813606578912178649

Any chance you can update to 3.17? :)

graysky commented on 2014-10-23 00:13

Bump to v3.16.6-3

Changelog: enable early microcode updates
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2014-10-19 23:45

Bump to v3.16.6-2

Changelog: new version of BFQ
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

sir_lucjan commented on 2014-10-19 22:56

graysky: http://algo.ing.unimo.it/people/paolo/disk_sched/patches/3.16.0-v7r6/

artafinde commented on 2014-10-17 16:38

Agree, thanks.

graysky commented on 2014-10-17 16:03

Not a linux-ck issue. Please take discussion to the bbs so as not to clutter up the AUR comments.

artafinde commented on 2014-10-17 15:37

I believe it's an issue with ccache and btrfs on SSD. When I disable the ccache it works. The problem is I'm getting a kernel panic so compilation stops half-way. After hard-reset I clean out the src and pkg directory and restart the compilation. Then I am assuming the objects are half created and I hit again a compilation error (not kernel-panic). It's an obscure case so comments in AUR might not be the proper place to solve this.

clfarron4 commented on 2014-10-17 08:32

@artafinde: What is the exact error? Might help us work out what the problem actually is.

prakharsingh95 commented on 2014-10-17 07:01

I compile with 8 threads, but without ccache. Works fine.

artafinde commented on 2014-10-17 06:30

I get a compilation error while using 8 threads and ccache. Anyone else can verify?

graysky commented on 2014-10-16 19:39

@mareex - Of course, but not until the corresponding intel-ucode package (currently 20140913-1) makes it into [extra] as there are key changes therein.

mareex commented on 2014-10-16 19:30

Will linux-ck enable early microcode updates like the Arch stock kernel?

https://projects.archlinux.org/svntogit/packages.git/commit/trunk?h=packages/linux&id=09b497ba4fd2674bd6488aa4115b4a38cdcbb64e

graysky commented on 2014-10-15 22:07

Bump to v3.16.6-1

Changelog: https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.16.6
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

Tjuh commented on 2014-10-10 18:28

Fixed now, thanks. I ran a new modprobed_db with the external hdd on and recompiled.

prakharsingh95 commented on 2014-10-10 18:00

@Tjuh. I faced a similar issue. Just boot into arch default, plug in all your peripherals and get them working, and then build linux-ck with localmodconfig=y. Everything will work out of the box them.

@graysky. Is it possible for you to enable SCSI, USB, etc by default (ie add an option to PKGBUILD like minimal_support=y)?

graysky commented on 2014-10-10 16:25

Odds are that using that option, you did not compile in USB support. I suggest using modprobe-db if you haven't already done so. Boot into a kernel with all modules, load those you will need, ie insert the external USB drive, populate the db and try rebuilding.

Tjuh commented on 2014-10-10 13:34

Yes, localmodcfg=y in the PKGBUILD. The latest linux-ck-haswell kernel works just fine from the repo-ck. The previous kernel v3.16.4-1 works just fine with the same PKGBUILD settings.

graysky commented on 2014-10-10 13:24

Did you modify the PKGBUILD or build with localmodconfig or the like? What happens if you use one of the kernel packages from repo-ck?

Tjuh commented on 2014-10-10 10:20

After compiling the latest linux-ck my external usb hdd is not being recognized anymore. I had no problem with the previous kernel.

graysky commented on 2014-10-09 20:30

Bump to v3.16.5-1

Changelog: https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.16.5
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

synergy commented on 2014-10-08 06:26

@AnAkkk
What you're most likely witnessing isn't a bug. By design, BFQ sacrifices throughput for interactivity.

prakharsingh95 commented on 2014-10-08 05:18

@nemesys and @clfarron4 I had to manually enable USB SCSI option. I was just surprised that localmodcfg skipped this.

prakharsingh95 commented on 2014-10-08 05:17

@nemesys and @clfarron4 I had to manually enable USB SCSI option. I was just surprised that localmodcfg skipped this (Deleted my previous comment and leaving this one in case anyone else encounters this issue).

nemesys commented on 2014-10-07 18:46

@prakharsingh95: Please read the kernel documentation for what different build options do before attempting to install a custom kernel. The fact that you are compiling the main working software component on your system should be warning enough to RTFM.

clfarron4 commented on 2014-10-07 18:39

@prakharsingh95: A warning for what?

I know there is a clear warning for what localmodcfg does and how it will only compile probed modules in the PKGBUILD.

prakharsingh95 commented on 2014-10-07 14:59

Please please, add a warning in big bold letters for people like me who just want to use localmodcfg and minimal nconfig. I compiled 7 times before, got kernel failed to load errors, googled, fixed them myself, and still no USB.
Finally I realized that the USB announce and support options were unchecked.
Phew.

graysky commented on 2014-10-05 22:18

Bump to v3.16.4-1

Changelog: https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.16.4
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2014-09-28 13:57

If you feel strongly enough about it, report your findings upstream: http://algo.ing.unimo.it/people/paolo/disk_sched/sources.php

AnAkkk commented on 2014-09-28 12:40

I haven't timed, but it's quite noticeable. I've tried with and without elevator=bfq, my KDE session opens pretty much instantly without, and there's a delay with BFQ.

graysky commented on 2014-09-28 12:02

No slower here. Have you actually timed between boots?

coderkun commented on 2014-09-28 09:57

@AnAkk: I don’t know which scale you are talking about but I think the BFQ is a litte bit slower by design.

AnAkkk commented on 2014-09-28 09:35

Has anyone else noticed slower boot times when BFQ is enabled (on a SSD)? My KDE session takes a longer time to load with BFQ.

coderkun commented on 2014-09-19 15:04

@graysky: Thanks, works fine now.

graysky commented on 2014-09-18 18:54

Sorry gang, my bad. Fixed now, but did not bump the pkgver to avoid rebuilding the repo (there is no need to do so).

graysky commented on 2014-09-18 18:53

Sorry gang, my bad. Fixed now.

coderkun commented on 2014-09-18 12:30

@clfarron4: I get the same sha256sum as you but the PKGBUILD contains a different one (4c44ad820fae8afaf3ad55fa7b191cc15e61829293a074126be780a35275b7c6).

clfarron4 commented on 2014-09-18 11:13

@coderkun: What do you get as the sum for that file? I get this:

% sha256sum enable_additional_cpu_optimizations_for_gcc_v4.9+_kernel_v3.15+.patch.gz
a9ca0e6fc01a3d34058b6cc9cdb560f62233f2eca5917db65dc0b29772bb236e enable_additional_cpu_optimizations_for_gcc_v4.9+_kernel_v3.15+.patch.gz

coderkun commented on 2014-09-18 07:35

Checksum of “enable_additional_cpu_optimizations_for_gcc_v4.9+_kernel_v3.15+.patch.gz” does not match.

graysky commented on 2014-09-17 19:37

Bump to v3.16.3-1

Changelog: https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.16.3
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2014-09-06 11:14

Bump to v3.16.2-1

Changelog: https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.16.2
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2014-08-25 00:24

Bump to v3.16.1-2

Changelog: Update to ck2 with bfs v456 and fix FS#41683 (intel ringbuffer).
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

AnAkkk commented on 2014-08-20 15:51

According to the kernel config, 1000 Hz is best for desktop systems anyway, so it might be better to just leave it like that.
https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/tree/kernel/Kconfig.hz?id=refs/tags/v3.17-rc1

graysky commented on 2014-08-18 12:18

I haven't checked running at 300 Hz. If you would like to try and report back that would be great. AFAIK ck recommends 1000 Hz.

AnAkkk commented on 2014-08-18 10:48

Haven't the suspend issues been solved since a long time now? 1000 Hz should no longer be required?
Aren't there some downsides from using 1000Hz instead of 300Hz?

dkaylor commented on 2014-08-16 18:22

OK. As always, thanks for the work you do on linux-ck, been using it for years now.

graysky commented on 2014-08-16 18:14

No idea about the answer to your SMT question. I just use modprobed-db on my own system and disable a few things that I don't use, NUMA, virtualization, and hibernation.

dkaylor commented on 2014-08-16 18:12

Thanks graysky. Which leads me to a question - SMT nice aside, if I have a non-SMT processor (4 cores = 4 threads, nice and simple) does disabling SMT in the kernel config accomplish anything, other than saving a little overhead? What's your opinion on heavily customized kernel configs in general?

graysky commented on 2014-08-16 17:55

No, I thought about this... it's currently on (ie hardcoded in the .config files). You will need to manually disable it. I could add this to the PKGBUILD as an option, but have no idea how many users would even use it.

dkaylor commented on 2014-08-16 17:53

so as bfs 450 now includes the smt nice work, will I be able to turn it off in the kernel config? I have a non-hyperthreaded core2 quad, so I always turn off SMT scheduling, although I'm not sure it matters much. ck mentions a bfs450-nosmt-buildfix.patch to make it impossible to enable SMT nice on a non-SMT kernel, but not sure it's included in the PKGBUILD. I'm fine with just disabling it in the config.

graysky commented on 2014-08-16 10:09

Bump to v3.16.1-1

Changelog: https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.16.1
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

sir_lucjan commented on 2014-08-16 10:03

patch-3.16-ck1 is out!

graysky commented on 2014-08-15 19:16

@sir_lucjan - I see it. Updated but did not bump pkgver since rebuilding the repo packages is silly with this defaulted to off.

sir_lucjan commented on 2014-08-15 17:50

@graysky:

smtnice 6 is out

sir_lucjan commented on 2014-08-15 08:58

Probably:

http://ck.kolivas.org/patches/bfs/3.0/3.15/test/3.15-ck1-smtnice5.patch

graysky commented on 2014-08-15 06:46

@ishitatsuyuki - Why out-of-date?

graysky commented on 2014-08-15 06:46

Bump to v3.15.10-1

Changelog: https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.15.10
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2014-08-08 06:58

Bump to v3.15.9-2

Changelog: Fixed _kernver mistake in 3.15.9-1
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2014-08-08 00:50

Bump to v3.15.9-1

Changelog: https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.15.9
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

clfarron4 commented on 2014-08-05 21:46

Looks like it's back up now.

ilkyest commented on 2014-08-04 03:09

yeah it's down... have you seen something about? another link to put on "pkgbuild"

ilkyest commented on 2014-08-04 01:51

Graysky... patch from con kolivas, resulting in timeout
-> patch-3.15-ck1.bz2...

ilkyest commented on 2014-08-04 01:40

Thanks so much by the kernel Graysky. I'm a folding at home user (http://folding.stanford.edu/), that try improves F@H performance, every time trying new kernels.

Sometime ago, about 3.10.1 your kernel was not good to folding. At this time, only in one test, it were the worst,

I'm trying NUMA=ON and NUMA=OFF differences, and, withing your changes, and "kernel's studying" in a while, only ck will fold on my pc.


But... when you will bump the pkgver? Any date?

graysky commented on 2014-08-01 13:42

Note that CK just posted an experimental patch that in his words, "Works for me flawlessly on all my stuff, but it's still pretty new." You can read about it at the link below. I don't want to enable it by default in the repo but do want to allow users a chance to try it out in the PKGBUILD. See the first set of comments to enable it. As always, post your experiences both good and bad to CK's blog so he can make ck1 and BFS more and more powerful and robust for everyone to enjoy.

Note-I am not going to bump the pkgver to 2 since I just rebuild the repo overnight and since this is really no chance unless you enable it.

http://ck-hack.blogspot.com/2014/08/smthyperthreading-nice-and-scheduling.html

graysky commented on 2014-08-01 02:36

Bump to v3.15.8-1

Changelog: https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.15.8
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2014-07-30 19:22

Yeah, I saw them and in the past CK has considered the patch experimental and as noted in his blog post, "this patch by itself does nothing unless other code uses the locks." I don't want to add an experimental patch to the package must less to repo packages people depend on for stable systems.

clfarron4 commented on 2014-07-30 19:10

I assume the flagger saw a new post on the CK blog and flagged out-of-date without reading.

That said, the new post on the CK blog refers to these patches:
http://ck.kolivas.org/patches/bfs/3.0/3.15/urwlocks/urw-locks.patch
http://ck.kolivas.org/patches/bfs/3.0/3.15/urwlocks/bfs449-grq_urwlocks.patch

Just after I build 3.15.7-1-ck-pax as well (with no time left today or tomorrow to muse over these new patches).

clfarron4 commented on 2014-07-30 19:09

I assume the flagger saw a new post on the CK blog and flagged out-of-date without reading.

That said, the new post on the CK blog refers to these patches:
http://ck.kolivas.org/patches/bfs/3.0/3.15/urwlocks/urw-locks.patch
http://ck.kolivas.org/patches/bfs/3.0/3.15/urwlocks/bfs449-grq_urwlocks.patch

graysky commented on 2014-07-30 16:17

Why out of date?

graysky commented on 2014-07-28 20:07

Bump to v3.15.7-1

Changelog: https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.15.7
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

clfarron4 commented on 2014-07-25 11:45

And algo.ing.unimo.it is back up again.

graysky commented on 2014-07-24 18:59

I am now hosting the BFQ patches until their site is back up. Access them like so:

1) Download the linux-ck source tarball, untar it, enter the linux-ck directory
2) sed -i 's|^_bfqpath=.*$|_bfqpath=\"http://repo-ck.com/source/mirror\"|' PKGBUILD

fx101 commented on 2014-07-24 15:23

For whatever reason, algo.ing.unimo.it is offline. Any mirrors?

graysky commented on 2014-07-19 02:24

Bump to v3.15.6-1

Changelog: https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.15.6
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

mareex commented on 2014-07-18 09:41

Out of date

Linux 3.15.6 is out on kernel.org

https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.15.6

graysky commented on 2014-07-11 20:46

Bump to v3.15.5-2

Changelog: add patch for #33745 efistub breakage
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2014-07-11 20:46

Bump to v3.15.5-1

Changelog: add patch for #33745 efistub breakage
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2014-07-10 20:04

Bump to v3.15.5-1

Changelog: https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.15.5
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

mareex commented on 2014-07-09 19:30

Out of date: 3.15.5 is out on kernel.org

graysky commented on 2014-07-08 10:43

Bump to v3.15.4-1

Changelog: https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.15.4
Changelog: bfs v0.449 update
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

clfarron4 commented on 2014-07-07 14:57

I see it now: http://ck.kolivas.org/patches/bfs/3.0/3.15/bfs448-449.patch

It would be nice for people to post here to say what's out of date.

Also, 3.15.4 is out now.

clfarron4 commented on 2014-07-07 14:42

I see it now: http://ck.kolivas.org/patches/bfs/3.0/3.15/bfs448-449.patch

dkaylor commented on 2014-07-07 03:15

I'm not the one flagging it out of date, but it's probably because ck released BFS 449. graysky stated earlier that he was busy with family stuff this weekend, so he just hasn't had a chance to do 3.15.3-2 yet.

fx101 commented on 2014-07-06 20:48

Who keeps flagging this out of date? Kernel.org lists 3.15.3 as the latest release.

clfarron4 commented on 2014-07-05 22:45

How is this out-of-date? Seriously, we need functionality like the main package website so that we understand how it is out-of-date.

graysky commented on 2014-07-02 23:53

Bump to v3.15.3-1

Changelog: https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.15.3
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

Note - I just pushed the 3.15.3-1 release series but will not have time for any updates until after the holiday due to family being in town.

graysky commented on 2014-07-01 19:08

Bump to v3.14.10-1

Changelog: https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.14.10
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

Note - CK should be releasing 3.15-ck1 shortly. We are having family come out to visit for the 4th of July holiday which may prevent me from updating all the linux-ck packages (nvidia, broadcom, vbox, etc.) to the 3.15 release until after they leave.

clfarron4 commented on 2014-07-01 18:46

Looks like 3.14.10 now includes 0012-fix-saa7134.patch.

sir_lucjan commented on 2014-07-01 09:30

3.14.10 is out!

graysky commented on 2014-06-26 21:45

Bump to v3.14.9-1

Changelog: https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.14.9
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

clfarron4 commented on 2014-06-26 19:59

3.14.9 is out!

graysky commented on 2014-06-20 19:24

Bump to v3.14.8-2

Changelog: BFQ updated to v7r5 (thanks for the note clfarron4)
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

clfarron4 commented on 2014-06-20 18:25

Oop, it looks like BFQ v7r5 is out!

http://algo.ing.unimo.it/people/paolo/disk_sched/patches/3.14.0-v7r5/

illis commented on 2014-06-19 02:19

@clfarron4: Sorry this could actually be something on my end.

Tried downloading the file another few times and was still getting the mismatched hash. Looked at it a bit further and it turns out that the file was in plain text (not gz) - somewhere along the line it was being uncompressed automatically.

Have a feeling it may be something to do with my ISP as I teathered up my mobile issues the same curl -O ... command, and got the correct response.

clfarron4 commented on 2014-06-18 12:12

@illis: I just downloaded it fresh, and I get this:

sha256sum enable_additional_cpu_optimizations_for_gcc_v4.9+.patch.gz
c6c4a9f77683b95c37636b20c4bc8a1f8214c87feef7fc469e58534fcc32fb4a

The sha256sum in the PKGBUILD is correct.

clfarron4 commented on 2014-06-18 12:06

@illis: I just downloaded it fresh, and I get this:

sha256sum enable_additional_cpu_optimizations_for_gcc_v4.9+.patch.gz
c6c4a9f77683b95c37636b20c4bc8a1f8214c87feef7fc469e58534fcc32fb4a

illis commented on 2014-06-17 23:03

sha256 match seems to be failing for the gcc patch.

sha256sum enable_additional_cpu_optimizations_for_gcc_v4.9+.patch.gz
4883a4d45fdcb02600af976be8b9b89f459775ac8916712dfdefd29419c3eacf enable_additional_cpu_optimizations_for_gcc_v4.9+.patch.gz

turtil commented on 2014-06-17 21:31

@graysky, @clfarron4, thanks, i removed the src/, and it worked without a charm, must have been during my couple of goes trying to get it up and running, and it patching itself a couple of times or something.

Thanks, all working now!

clfarron4 commented on 2014-06-17 17:35

@turtil: Either the below comment or you're building in a directory that needs to the /src and /pkg folders to be removed.

graysky commented on 2014-06-17 08:05

@turtil - Perhaps you have a cached version of the source or something that needs to be cleared? Check your /etc/makepkg.conf

Schmeidenbacher commented on 2014-06-17 07:33

Just compiled this kernel, went through without an issue.

turtil commented on 2014-06-16 23:50

patching file include/uapi/linux/sched.h
patching file include/linux/sched/rt.h
patching file kernel/stop_machine.c
patching file drivers/cpufreq/cpufreq_conservative.c
patching file kernel/time/Kconfig
patching file kernel/sched/Makefile
The next patch would create the file kernel/sched/bfs_sched.h,
which already exists! Skipping patch.
1 out of 1 hunk ignored
patching file kernel/sched/stats.c
patching file arch/ia64/kvm/kvm-ia64.c
patching file arch/powerpc/kvm/book3s_hv.c
Hunk #1 succeeded at 1509 (offset -2 lines).
patching file arch/x86/Kconfig
patching file arch/x86/kvm/x86.c
Hunk #1 succeeded at 6114 (offset -1 lines).
patching file kernel/Kconfig.preempt
patching file kernel/Kconfig.hz
patching file Makefile
==> ERROR: A failure occurred in prepare().
Aborting...

Generally speaking, sure this is an issue with the Makefile being patched, has anyone else had this issue? If its just me, I dont mind trying to patch the patch, and provide back.

graysky commented on 2014-06-16 22:03

Bump to v3.14.8-1

Changelog: https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.14.8
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

clfarron4 commented on 2014-06-16 21:56

https://www.kernel.org/pub/linux/kernel/v3.x/patch-3.14.8.xz

mariojuniorjp commented on 2014-06-14 11:45

Seriously, I just wanted to know why it is not working out the compilation of the kernel. Done again, and it was until the end. Once finished installing the kernel, simply aur started again with the same procedure. Canceled, ran the grub, it detected the new kernel, I sent restart and when I went to look, simply appears only the default system kernel.

I've compiled again (manual mode), and it worked. I would just download the files, and compile without using the damn yaourt?

mariojuniorjp commented on 2014-06-14 00:15

@Schemeidenbacher

I changed the folder for the compilation. Now I have to wait to do it again, and see if will work this time. Many thanks for all the help! =D

Schmeidenbacher commented on 2014-06-14 00:02

Or you uncomment the BUILDDIR stanza in /etc/makepkg.conf and set it to a place with more space.

mariojuniorjp commented on 2014-06-14 00:00

@Schemeidenbacher

13.9 MB of 908.7 MB (98% used)

That was it. Now I have to figure out how to increase the size of / tmp folder, because what is taking up all the space is precisely the kernel compilation. xD

Schmeidenbacher commented on 2014-06-13 23:54

It's building in /tmp/makepkg then (the default), so yeah, depends on what's mounted there.

mariojuniorjp commented on 2014-06-13 23:52

@graysky @Scimmia

I'm building using Yaourt. :P

I did not know it would cause an error in compiling the kernel that way. I've only set up the PKGBUILD, other things too, and then sent compile the kernel.

@Schemeidenbacher

#-- Specify a directory for package building.
#BUILDDIR=/tmp/makepkg

Is this?

Schmeidenbacher commented on 2014-06-13 23:36

@mariojuniorjp:

Also check the file /etc/makepkg.conf for the BUILDDIR stanza. That will tell you where makepkg is building it. Check for space there.

Scimmia commented on 2014-06-13 23:35

The fact that it's asking you if you want to restart means you're using a helper of some sort. From the looks of it, I would say yaourt. By default, yaourt builds in /tmp, which is on tmpfs. That's your problem.

graysky commented on 2014-06-13 23:32

Seems like something thinks otherwise... where are you building (type pwd in the build dir)... also, are you using an aur helper? Don't do that. Just try makepkg -srci and see if that gives problems.

mariojuniorjp commented on 2014-06-13 23:31

I'm getting an error on the part of the kernel build, and do not know how to solve. The error is this one:

patch: **** write error : No space left on device
==> ERROR: A failure occurred in prepare().
Aborting...
==> ERROR: Makepkg was unable to build linux-ck.
==> Restart building linux-ck ? [y/N]


Have plenty of space on the partition, then do not know what the logic of this error. Oo

graysky commented on 2014-06-11 20:40

Bump to v3.14.7-1

Changelog: https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.14.7
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2014-06-11 20:40

I think it's because repo-ck moves lots of data and they [godaddy] want me to get sick of them and drop my unlimited service they have me in at an old rate. They have offered to upgrade me to a dedicated server for higher reliability several times now for a large increase in my monthly fee. Each time I remind them that my plan is unlimited and that they can upgrade me without a charge. Bastards. I am working slowly with barrikad to setup a mirror on his service. I have been very busy with work/family shit and traveling though. I am booked this weekend as well but hope to do some testing and setup with him shortly.

dlh commented on 2014-06-09 10:41

http://godaddy.com/

snack commented on 2014-06-09 10:12

@clfarron4: I don't understand your comment about "godaddy" (english is not my native speaking, sorry). I am indeed able to resume the download, but it's a bit annoying when you have to launch pacman 4-5 times to update the system...

clfarron4 commented on 2014-06-09 10:11

@snack: That's godaddy being godaddy. You should be able to recommence the download with pacman and just repeat until it finishes.

snack commented on 2014-06-09 07:21

Since some days I get this kind of error while retrieveing packages from repo-ck:

error: failed retrieving file 'linux-ck-core2-3.14.6-2-x86_64.pkg.tar.xz' from repo-ck.com : transfer closed with 55761293 bytes remaining to read

Am I the only one?

graysky commented on 2014-06-08 14:25

Bump to v3.14.6-2

Changelog: https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.14.6
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

sir_lucjan commented on 2014-06-08 11:21

https://projects.archlinux.org/svntogit/packages.git/commit/trunk?h=packages/linux&id=ff50f2d3bb61aa7a00a6a83ca70fc6e0d26077d9

dlh commented on 2014-06-08 10:53

Thanks for your work, I'll check that patch in the meantime.

graysky commented on 2014-06-08 10:42

OK. I see that 3.14.6-1 is now in svn. I have further adjusted the gcc patch and am building 3.14.6-2 now. I will upload the source tarball to the AUR in several hours once I see that the various new arches using the patch build just fine.

For those wanting a tarball early, find it here[1] and please be sure to post to the AUR your experiences building (good or bad) and for which arch!

1. http://repo-ck.com/PKG_source/linux-ck/linux-ck-3.14.6-2.src.tar.gz

graysky commented on 2014-06-08 00:10

OK... I am building with the new patch but will not index the repo until I get some feedback that everything is ok from you guys. So:

1) If you want to build 3.14.6-1-ck yourself, download the source tarball here.1
2) If you want to try a repo-ck package, just browse manually to http://repo-ck.com and navigate to either x86_64 or i686 and manually download the package for your processor. Remember that I am building now and will not have a complete set for another 6 hours or so.

1. http://repo-ck.com/PKG_source/linux-ck/linux-ck-3.14.6-1.src.tar.gz

graysky commented on 2014-06-07 23:25

Ivy works for me... no access to older stuff that I can mess with at this point in time.

graysky commented on 2014-06-07 23:06

I need some savvy users of newer intel cpus (any core ix like nehalem, sandy, ivy, or haswell) to try out the new patch I have placed in my 'unstable' branch. Just substitute the line in this PKGBUILD for it. Please build the package selecting the nconfig option for your hardware, boot into it and let me know that my stuff works.

https://github.com/graysky2/kernel_gcc_patch/blob/unstable/enable_additional_cpu_optimizations_for_gcc_v4.9%2B.patch

graysky commented on 2014-06-07 21:27

I know that 3.14.6 was just released but I want to wait to make the linux-ck release to see what tpowa does with the official ARCH package as it seems a few new config options are introduced. Please do not flag linux-ck out-of-date until the ARCH 3.14.6-1 is pushed to svn or repo. Thanks!

graysky commented on 2014-06-07 19:37

I see now that this is confirmed: https://gcc.gnu.org/gcc-4.9/changes.html

I will need to adjust the patch accordingly.

graysky commented on 2014-06-07 19:29

Interesting that with the new version of gcc, these "older" options for march and mtune seem to have been made more generic. For example, on my ivy:

% gcc -c -Q -march=native --help=target | grep march
-march= ivybridge

Older gcc version returned
-march= core-avx-i

graysky commented on 2014-06-07 19:23

@dlh - For haswell: CONFIG_MCOREAVX2 gets set which uses -march=core-avx2 -mtune=core-avx2.

See: https://github.com/graysky2/kernel_gcc_patch/blob/master/enable_additional_cpu_optimizations_for_gcc.patch

elektorronikci commented on 2014-06-07 19:20

@dlh, sure linux-ck-haswell is not compiled with that options and I think that is another story if you are compiling from source, right? Did I understand wrong?

If GCC is wrong with some options when you issue -native, you may have missed some other options too, other than -mno-avx -mno-avx2.

elektorronikci commented on 2014-06-07 19:15

@dlh, I don't know how it is compiled. I said it works with i7-4770 just fine.

dlh commented on 2014-06-07 12:35

@elektorronikci: I have G3220 and with native I have -mno-avx -mno-avx2
But you are re saying that linux-ck-haswell is compiled with those options, right?
That is why I am unable to run it, correct me if I am wrong?

elektorronikci commented on 2014-06-06 17:07

@sash-arch, your CPU doesn't have AVX2 set: http://ark.intel.com/products/80800/intel-celeron-processor-g1840-2m-cache-2_80-ghz

elektorronikci commented on 2014-06-06 17:01

I installed your ck-haswell compilation, now it is running. Not much time running, so not tested enough I think, but it really seems fine. If any problem happens, I will acknowledge.

graysky commented on 2014-06-06 09:59

As a test can either of you guys [tweeserhead or elektorronikci] try the ck-haswell packages on your haswell systems to confirm the problem is not with them + the hardware?

elektorronikci commented on 2014-06-06 08:16

I have i7-4770. I don't use ck-haswell repo but compile with -march=native. I use bfs&bfq. (bfqv7r3 had some issues before) There is no problem.

walkingrobot commented on 2014-06-06 00:04

Sorry, I forgot to mention it's a haswell processor

walkingrobot commented on 2014-06-06 00:02

Well this kinda misses the point but I compile with -march=native and CONFIG_MNATIVE

graysky commented on 2014-06-03 19:03

Not sure what to tell you guys... are other haswell users out there using the ck-haswell repo packages without issue?

dlh commented on 2014-06-03 14:03

@sash-arch I have the same problem as you, therefore I switched to generic linux-ck package.

sash-arch commented on 2014-06-02 22:36

Haswell...




sascha@server ~ % gcc -c -Q -march=native --help=target | grep march
-march= haswell

graysky commented on 2014-06-02 22:35

Not sure what to make of that, sash. Ivy gets CONFIG_MCOREAVXI which haswell gets CONFIG_MCOREAVX2 which are the correct flags. What does the following return for you?

gcc -c -Q -march=native --help=target | grep march

sash-arch commented on 2014-06-02 22:30

I have an Celeron G1840 CPU. It's an "Haswell" CPU but the Linux-ck-haswell Package doesn't work with CPU. It stops working while loading the initial ramdisk. If i switch to linux-ck-ivybridge, everything works fine.

graysky commented on 2014-05-31 22:42

Bump to v3.14.5-1

Changelog: https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.14.5
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

janck commented on 2014-05-30 20:21

Thank you very much:)

graysky commented on 2014-05-30 18:32

Just refreshed 3.14.4-2 which is a proper split package now that the AUR supports them.

@janck - Your suggestion is live now.

koshak commented on 2014-05-30 07:11

@graysky
Sorry, seems like local proxy server gives me unzipped version of this patch for some reason, so sha256sum is different and its contents is plain-text... On-the-fly PKGBUILD editing fixed this particular issue.

graysky commented on 2014-05-29 19:45

Let me take them in order:
clfarron4 - Yes, that patch is needed for building with SMP but this is the first time I can remember it coming up. In the interest of simplicity when I merge this PKGBUILD with the official ARCH PKGBUILD on major version bumps, I want to minimize the potential for errors on my part and thus am rejecting the request to add the SMP patch. CK may roll it into his patchset which is fine, but there aren't many people building this package on a single core system.

@janck - I see, you're saying that the modification is needed to keep the build from FAILING in the case of you stripping out the extraneous config options. OK. I will implement your fix and for my sanity comment the PKGBUILD as such so I don't nuke it on an major version merge.

@koshak - The sums match for me, not sure why your sums are different. I also don't understand your comment about the gz being txt. Are you editing with something smart like vim? For example, vim will decompress the gz when you open the gz transparently.

I hope that addresses everyone's concerns.

docwild commented on 2014-05-29 13:20

as clfarron4 mentioned below I suggested a CONFIG_SMP patch on the forum repo topic by mistake. That post is now gone but I'll put it here:

Adding
[code]
if grep --quiet "# CONFIG_SMP is not set" "${startdir}/config.last" 2&>0 /dev/null || grep --quiet "# CONFIG_SMP is not set" "${startdir}/config.x86_64.last" 2&>0 /dev/null
then
msg "Patch NOSMP patch"
patch -Np1 -i "${srcdir}/ck-nosmp.patch"
fi
[/code]
at the very end of the prepare() function, after populating the config.old file allows users with uniprocessors to disable symmetric multiprocessing support in kernel config (it has some overhead).

The "ck-nosmp.patch" file here: https://gist.githubusercontent.com/pfactum/9332896/raw/0001-ck-3.12-fix-BFS-compiling-with-CONFIG_SMP-n.patch

Mentioned on cks blog here: http://ck-hack.blogspot.co.uk/2014/05/bfs-0447-314-ck1.html

koshak commented on 2014-05-29 09:40

wrong sum for http://repo-ck.com/source/gcc_patch/enable_additional_cpu_optimizations_for_gcc.patch.gz

prepare() is looking for enable_additional_cpu_optimizations_for_gcc.patch, which is absent.

instead enable_additional_cpu_optimizations_for_gcc.patch.gz is present but it is not a gzip file. It is pure text.

WTF?

janck commented on 2014-05-29 09:39

After removing all dvb shit from kernel config, include/config/dvb does not exist in the source tree anymore, and consequently the build fails at line 432.

This fixes my particular problem (have been testing that it works this time):

if [ -d include/config/dvb/ ]; then
mkdir -p "${pkgdir}/usr/lib/modules/${_kernver}/build/include/config/dvb/"
cp include/config/dvb/*.h "${pkgdir}/usr/lib/modules/${_kernver}/build/include/config/dvb/"
fi

clfarron4 commented on 2014-05-29 09:29

Some has suggested this in the BBS, to fix something to do with SMP: https://bbs.archlinux.org/viewtopic.php?pid=1420124#p1420124

clfarron4 commented on 2014-05-29 09:29

Some has suggested this in the BBS: https://bbs.archlinux.org/viewtopic.php?pid=1420124#p1420124

graysky commented on 2014-05-28 22:22

@janck - But again, the PKGBUILD as-is should accomplish this:

mkdir -p "${pkgdir}/usr/lib/modules/${_kernver}/build/include/config/dvb/"
cp include/config/dvb/*.h "${pkgdir}/usr/lib/modules/${_kernver}/build/include/config/dvb/"

janck commented on 2014-05-28 22:13

Found a fix from a previous build.

# this line will allow the package to build if a user disables the dvb shit
find include/config/dvb -name '*.h' -exec cp {} "${pkgdir}/usr/src/linux-${_kernver}/include/config/dvb/" \;

graysky commented on 2014-05-28 18:55

@janck - Guess I'm confused... your statement says if include/config/dvb exists copy include/config/dvb/*.h to "${pkgdir}/usr/lib/modules/${_kernver}/build/include/config/dvb/"

The current PKGBUILD just does that, first making the destination on line 432 and doing the same copy on line 433. What am I missing?

janck commented on 2014-05-28 04:58

sorry to confuse. What needs to be done is to repalce line 432 with this one:
[[ -e include/config/dvb/ ]] && cp include/config/dvb/*.h "${pkgdir}/usr/lib/modules/${_kernver}/build/include/config/dvb/"

graysky commented on 2014-05-27 21:40

Yes, those lines are in the PKGBUILD. I am confused.

janck commented on 2014-05-27 20:52

yes... did you check?

graysky commented on 2014-05-27 20:03

@janck - That is contained in this PKGBUILD, no?

janck commented on 2014-05-27 19:25

line 432:

# and...
# http://bugs.archlinux.org/task/11194
mkdir -p "${pkgdir}/usr/lib/modules/${_kernver}/build/include/config/dvb/"
cp include/config/dvb/*.h "${pkgdir}/usr/lib/modules/${_kernver}/build/include/config/dvb/"

graysky commented on 2014-05-26 18:23

Shit... did I merge out that line again? I need to review some past PKGBUILDs. I have a nasty habit of comparing linux-ck to linux the ARCH package to make sure I capture all the new changes. I think I may have merged out our change. I will look into it; faster if you can and point me to it given how busy I am these days.

http://pkgbuild.com/git/aur-mirror.git/log/linux-ck

janck commented on 2014-05-26 17:36

All dvb config has been diabled, and I get this error when compiling:

==> Starting package_linux-ck-headers()...
cp: cannot stat ‘include/config/dvb/*.h’: No such file or directory
==> ERROR: A failure occurred in package_linux-ck-headers().

kyak commented on 2014-05-20 16:37

Hi graysky,

It seems that nvidia-ck needs to be updated:

error: failed to prepare transaction (could not satisfy dependencies)
:: nvidia-ck-ivybridge: requires nvidia-utils=337.12

Thank you!

graysky commented on 2014-05-18 13:05

Bump to v3.14.4-2

Changelog: Upstream bump of BFQ to v7r4.
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

clfarron4 commented on 2014-05-18 12:43

BFQ v7r4 is out!

http://algo.ing.unimo.it/people/paolo/disk_sched/patches/3.14.0-v7r4/

sash-arch commented on 2014-05-14 06:49

Since i've updatet to 3.14.3-1 i have the same issue, like it's discribed in: https://bbs.archlinux.org/viewtopic.php?id=180139

Is it possible to compile nls_iso8859-1 into the Kernel like it's been since 3.13?

I don't know how to fix this issue without an Kernel with nls_iso8859- builtin support.

test0 commented on 2014-05-13 20:49

Thanks a lot, graysky! :)

graysky commented on 2014-05-13 18:55

Bump to v3.14.4-1

Changelog: https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.14.4
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

sir_lucjan commented on 2014-05-13 17:32

3.14.4 is out.

test0 commented on 2014-05-13 09:25

Please apply patch to fix CVE-2014-0196 (https://projects.archlinux.org/svntogit/packages.git/commit/trunk?h=packages/linux&id=6be263f1dfd1deb20b81a3d0f6b53a0508e66523).

dlh commented on 2014-05-07 18:42

@graysky: yes, the generic package works fine, but linux-ck-haswell not.

graysky commented on 2014-05-07 18:33

@dlh - If you install the generic packages from repo-ck does that render you with a bootable system? Post #1805 in this thread reports no problem with haswell.

https://bbs.archlinux.org/viewtopic.php?id=111715&p=73

dlh commented on 2014-05-07 15:48

[repo-ck.com] linux-ck-haswell doesn't boot at all, the issue not apply for the linux-ck

graysky commented on 2014-05-06 20:15

Bump to v3.14.3-1

Changelog: Bump to new version.
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

sir_lucjan commented on 2014-05-06 12:32

http://ck-hack.blogspot.de/2014/05/bfs-0447-314-ck1.html

graysky commented on 2014-04-26 00:34

Bump to v3.13.11-2

Changelog: Bump for new BFQ.
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2014-04-24 09:04

So it is... I will update later today. Wish I knew this last night before I built 3.13.11-1 but what can you do...

felixonmars commented on 2014-04-24 08:15

BFQ v7r3 is out!

clfarron4 commented on 2014-04-23 22:53

@brando56894: Nice catch indeed. I've updated those parts in my packages as well!

graysky commented on 2014-04-23 19:53

Bump to v3.13.11-1

Changelog: https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.13.11
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2014-04-23 19:52

Good catch, updated.

graysky commented on 2014-04-23 19:52

Good catch, updated.

brando56894 commented on 2014-04-23 19:48

This section of the package build's comments needs to be updated:

"To keep track of which modules are needed for your specific system/hardware,
# give module_db script a try: http://aur.archlinux.org/packages.php?ID=41689
# This PKGBUILD will call it directly to probe all the modules you have logged!"

needs to be changed to

To keep track of which modules are needed for your specific system/hardware,
# give modprobed_db script a try: https://aur.archlinux.org/packages/modprobed-db/
# This PKGBUILD will call it directly to probe all the modules you have logged!

graysky commented on 2014-04-12 22:19

I am now hosting the BFQ patches until their site is back up. Access them like so:

1) Download the linux-ck source tarball, untar it, enter the linux-ck directory
2) sed -i 's|^_bfqpath=.*$|_bfqpath=\"http://repo-ck.com/source/mirror\"|' PKGBUILD

Now you can build it.

chosig commented on 2014-04-12 19:24

http://unimo.it redirects to http://unimore.it -- which seems to have wrecked havoc with the subdomains, I guess the move isn't completed yet.

Anonymous comment on 2014-04-12 16:07

algo.ing.unimo.it seems to be offline; is there an alternative source of the patches hosted by it?

dlh commented on 2014-04-09 19:05

Here I described my issue, it's actually not related with kvm, I was confused cause I'm using this feature too https://bbs.archlinux.org/viewtopic.php?pid=1402406

coderkun commented on 2014-04-08 15:23

@adam777, @graysky: On my laptop KVM (Qemu) works flawless with linux-ck 3.13.9-1.

@dlh: Maybe it is also related to something else?

adam777 commented on 2014-04-06 18:51

@coderkun, @dlh - In here, 3.13.9-1 from the repo works fine with KVM.
I'm using the following:

qemu-system-x86_64 -machine type=pc,accel=kvm -cpu host -m 1G -net nic,model=virtio -net user -vga qxl -drive file=*******/XP.qcow2,if=virtio -rtc base=localtime -balloon virtio -spice port=5900,disable-ticketing,image-compression=off,jpeg-wan-compression=never,playback-compression=off -device virtio-serial-pci -device virtserialport,chardev=spicechannel0,name=com.redhat.spice.0 -chardev spicevmc,id=spicechannel0,name=vdagent &

dlh commented on 2014-04-06 15:52

@coderkun: I will try tomorrow, since I am currently on machine with different processor.

coderkun commented on 2014-04-06 15:47

@graysky: I disabled the KVM-path in the PKGBUILD and recompiled linux-ck but KVM is still not working for me. So I’m not sure if the path makes it better or worse. What about you @dlh?

graysky commented on 2014-04-06 15:33

coderkun and dlh - adam777 requested that patch in the bbs and replied that all is well with kvm.[1] I asked him to comment here. If the patch is causing breakage for users, I will disable it. I just need to know since I do not use KVM.

1. https://bbs.archlinux.org/viewtopic.php?pid=1401303#p1401303

dlh commented on 2014-04-06 15:31

Maybe it was fixed upstream

coderkun commented on 2014-04-06 15:12

@graysky: I have no evidence that the patch really causes this but KVM stopped working for me at around the same time.

graysky commented on 2014-04-06 14:22

@dlh - Yes, since v3.13.6-1 that patch has been included per a user request. Are you saying that your KVM functionality has been broken since that release, in other words, that patch has caused the regression?

dlh commented on 2014-04-06 12:40

graysky do you build your kernel in repo-ck with http://ck.kolivas.org/crap/kvm-fix.patch ? From version 3.13.6 it's stopped working.

graysky commented on 2014-04-04 18:58

Bump to v3.13.9-1

Changelog: https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.13.9
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

sir_lucjan commented on 2014-04-04 14:47

graysky - sorry, my mistake.

graysky commented on 2014-04-04 00:28

Can update until after work tomorrow but thanks for the heads up. Odd that Greg pushed it on a Thurs night, must be something critical... Will check the readme once lkml is back up.

clfarron4 commented on 2014-04-03 22:39

Dear lord, 3.13.9 is out all ready...

dlh commented on 2014-04-02 15:01

Does anyone experience https://bbs.archlinux.org/viewtopic.php?id=179151 ?

graysky commented on 2014-03-31 23:58

Bump to v3.13.8-1

Changelog: https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.13.8
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

clfarron4 commented on 2014-03-31 17:34

3.13.8 is out: www.kernel.org

clfarron4 commented on 2014-03-31 17:34

3.13.8 is out: kernel.org

graysky commented on 2014-03-24 19:46

Bump to v3.13.7-1

Changelog: https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.13.7
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

snack commented on 2014-03-24 09:20

I'm experiencing random hangs of touchpad, mouse and keyboards at boot after kdm start. I'm unable to do anything but a hard reboot. This is what I find in the boot log:

mar 24 09:58:53 elric kernel: BUG: unable to handle kernel NULL pointer dereference at (null)
mar 24 09:58:53 elric kernel: IP: [<ffffffffa0c9b317>] mousedev_open_device+0x77/0x100 [mousedev]
mar 24 09:58:53 elric kernel: PGD b646b067 PUD b6476067 PMD 0
mar 24 09:58:53 elric kernel: Oops: 0000 [#1] PREEMPT SMP
mar 24 09:58:53 elric kernel: Modules linked in: mousedev(+) psmouse serio_raw atkbd libps2 uvcvideo videobuf2_vmalloc videobuf2_memops videobuf2_core videodev media coretemp arc4 microcode(+) iwldvm mac80211 pcspkr i2c_i801 lpc_ich snd_hda_codec_hdmi r592 memstick iwlwi
mar 24 09:58:53 elric kernel: CPU: 0 PID: 367 Comm: acpid Tainted: P O 3.13.6-1-ck #1
mar 24 09:58:53 elric kernel: Hardware name: Sony Corporation VGN-SR21M_S/VAIO, BIOS R1110Y1 08/14/2008
mar 24 09:58:53 elric kernel: task: ffff8800b77dda30 ti: ffff8800b64c0000 task.ti: ffff8800b64c0000
mar 24 09:58:53 elric kernel: RIP: 0010:[<ffffffffa0c9b317>] [<ffffffffa0c9b317>] mousedev_open_device+0x77/0x100 [mousedev]
mar 24 09:58:53 elric kernel: RSP: 0018:ffff8800b64c1c10 EFLAGS: 00010202
mar 24 09:58:53 elric kernel: RAX: 0000000000000000 RBX: ffff8800b655a000 RCX: ffff8800b655a068
mar 24 09:58:53 elric kernel: RDX: 0000000000000000 RSI: 0000000000000000 RDI: 0000000000000246
mar 24 09:58:53 elric kernel: RBP: ffff8800b64c1c28 R08: 0000000000000000 R09: ffff88013b001600
mar 24 09:58:53 elric kernel: R10: 0000000000000000 R11: 0000000000000004 R12: 0000000000000000
mar 24 09:58:53 elric kernel: R13: ffff8800b655a080 R14: ffff8800b65761a8 R15: ffff8801395cf300
mar 24 09:58:53 elric kernel: FS: 00007fae54adb700(0000) GS:ffff88013fc00000(0000) knlGS:0000000000000000
mar 24 09:58:53 elric kernel: CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
mar 24 09:58:53 elric kernel: CR2: 0000000000000000 CR3: 00000000b7726000 CR4: 00000000000007f0
mar 24 09:58:53 elric kernel: Stack:
mar 24 09:58:53 elric kernel: ffff880138dda000 ffff8800b655a000 ffff8800b655a078 ffff8800b64c1c60
mar 24 09:58:53 elric kernel: ffffffffa0c9c0cc ffff8800b655a340 ffff8800b65761a8 ffff8801395cf300
mar 24 09:58:53 elric kernel: ffffffffa0c9ce80 ffff8801395cf310 ffff8800b64c1c98 ffffffff8118555f
mar 24 09:58:53 elric kernel: Call Trace:
mar 24 09:58:53 elric kernel: [<ffffffffa0c9c0cc>] mousedev_open+0xcc/0x150 [mousedev]
mar 24 09:58:53 elric kernel: [<ffffffff8118555f>] chrdev_open+0x9f/0x1d0
mar 24 09:58:53 elric kernel: [<ffffffff8117ec97>] do_dentry_open+0x1b7/0x2c0
mar 24 09:58:53 elric kernel: [<ffffffff8118bf41>] ? __inode_permission+0x41/0xb0
mar 24 09:58:53 elric kernel: [<ffffffff811854c0>] ? cdev_put+0x30/0x30
mar 24 09:58:53 elric kernel: [<ffffffff8117f0b1>] finish_open+0x31/0x40
mar 24 09:58:53 elric kernel: [<ffffffff8118ed72>] do_last+0x572/0xe90
mar 24 09:58:53 elric kernel: [<ffffffff8118c236>] ? link_path_walk+0x236/0x8d0
mar 24 09:58:53 elric kernel: [<ffffffff8118f74b>] path_openat+0xbb/0x6b0
mar 24 09:58:53 elric kernel: [<ffffffff81190e5a>] do_filp_open+0x3a/0x90
mar 24 09:58:53 elric kernel: [<ffffffff8119d527>] ? __alloc_fd+0xa7/0x130
mar 24 09:58:53 elric kernel: [<ffffffff81180284>] do_sys_open+0x124/0x220
mar 24 09:58:53 elric kernel: [<ffffffff8118039e>] SyS_open+0x1e/0x20
mar 24 09:58:53 elric kernel: [<ffffffff814fd16d>] system_call_fastpath+0x1a/0x1f
mar 24 09:58:53 elric kernel: Code: f0 8b 85 e0 5b 44 89 e0 41 5c 41 5d 5d c3 66 0f 1f 44 00 00 4c 89 ef 41 bc ed ff ff ff e8 d2 8b 85 e0 eb e0 48 8b 15 c9 21 00 00 <8b> 02 8d 48 01 85 c0 89 0a 75 c6 48 8b 05 37 1f 00 00 48 8d 98
mar 24 09:58:53 elric kernel: RIP [<ffffffffa0c9b317>] mousedev_open_device+0x77/0x100 [mousedev]
mar 24 09:58:53 elric kernel: RSP <ffff8800b64c1c10>
mar 24 09:58:53 elric kernel: CR2: 0000000000000000
mar 24 09:58:53 elric kernel: ---[ end trace fffdd7fde71f92ba ]---

I'm not really sure the culprit is the kernel, but I don't have time now to test with a stock kernel. So I'd like to know if any other user of linux-ck experiences this (I'm using linux-ck-core2)

graysky commented on 2014-03-22 13:31

Bump to v3.13.6-2

Changelog: Fix mass storage problems.
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

Since 3.13.7 is due out within the next 24 h.[1] According to the 3.13.7-rc1 patchset, both of these patches are included therein. I will bump 3.13.6 to -2 and include them both, but I don't want to rebuild all the repo packages since I will likely be doing it again when upstteam releases shortly.

1. https://lkml.org/lkml/2014/3/20/910

graysky commented on 2014-03-22 12:57

Bump to v3.13.6-2

Changelog: Fix mass storage problems.
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2014-03-22 12:55

@sir_lucjan - 3.13.7 is due out within the next 24 h.[1] According to the 3.16.7-rc1 patchset, both of these patches are included therein. I will bump 3.13.6 to -2 and include them both, but I don't want to rebuild all the repo packages since I will likely be doing so again tomorrow morning when 3.13.7 goes gold.

1. https://lkml.org/lkml/2014/3/20/910

sir_lucjan commented on 2014-03-22 11:08

Please apply two patches:

https://projects.archlinux.org/svntogit/packages.git/commit/trunk?h=packages/linux&id=297450e858cc2a2463857b2cefa3dd0a5a9419f6

corro commented on 2014-03-16 11:38

Please nevermind my earlier comment. It has nothing to do with the package and was probably in the wrong place anyway. Thanks for the hint about CONFIG_FHANDLE though.

graysky commented on 2014-03-16 10:25

Ah, didn't see his/her comment. I understand now. I can tell you that unless corro modded, it, the provided config has this set:

% uname -r
3.13.6-1-ck

% zgrep CONFIG_FHA /proc/config.gz
CONFIG_FHANDLE=y

TurboLama commented on 2014-03-16 08:37

graysky: no, it's not. I was just reacting to comments posted by "corro", since he uses his custom config, and is having trouble with systemd, most probably because it requires CONFIG_FHANDLE from version 209 onwards (see: https://bbs.archlinux.org/viewtopic.php?id=177773).

graysky commented on 2014-03-15 17:15

@Turbo - What is corro's config? I do not believe this is something my package provides or that I have any control over, no?

TurboLama commented on 2014-03-15 17:12

graysky: yes, i know, but it's probably missing in corro's config, hence the problem with systemd 210.

graysky commented on 2014-03-15 13:31

@turbo - Is it...
% grep CONFIG_FHA config*
config:CONFIG_FHANDLE=y
config.x86_64:CONFIG_FHANDLE=y

TurboLama commented on 2014-03-15 13:12

corro: enable CONFIG_FHANDLE

graysky commented on 2014-03-07 21:09

Bump to v3.13.6-1

Changelog: https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.13.6
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

corro commented on 2014-03-06 22:45

I tried the default kernel config instead of my tailored one and it worked with the current ck kernel and systemd. I still don't know where the problem is (I did not modify the config in months), but I've come to the conclusion that it must be my (or better my config's) fault.

Schmeidenbacher commented on 2014-03-06 09:31

Now that you mention it, i run my system from an encrypted lvm configuration. My mkinitcpio.conf has the following MODULES and HOOKS configurations:

MODULES="dm_mod ext2 ext4"

HOOKS="base udev v86d autodetect modconf block keyboard keymap lvm2 encrypt filesystems shutdown fsck"

Note, that my system is not configured in english, so i had to add keymap and keyboard early on to get my encryption password entered correctly.

xcomcmdr commented on 2014-03-06 06:31

I've tried every way to build linux-ck (without any modification but the ck, with the stock configuration, only with probed modules, ...). Still the same problem :
Kernel panic - not syncing : VFS: Unable to mount root fs on unknown-block(0,0)

Google suggests to include the ext4 module in the initrd image... Interesting.
Since last night, while using the stock kernel and copying 35 GB of MP3s from /dev/sda2 (ext4) to /dev/sda1 (ntfs3g) firefox and audacious hanged a few times, which means that simple multitasking doesn't work (!), I wanted to try linux-ck's BFQ I/O scheduler once again...

Schmeidenbacher commented on 2014-03-05 22:26

Mine works fine. The only issue i encountered with systemd 210-2 was the sudden urge of my quad-core desktop computer to start systemd-backlight@eeepc-wmi.service for some insane reason – scaring me with a nice red "FAILED" during boot – and systemd-logind.service to complain that it can't apply ACLs. Neither managed to hinder boot otherwise.

My kernel is custom build using mostly the modules and drivers i need on my system. The kernel config can be found here: http://pastebin.com/9LE1VeGR

xcomcmdr commented on 2014-03-05 22:10

same issue here, since ages ago. Linux-ck can't mount / and panics.

corro commented on 2014-03-04 22:52

It looks like I've encountered some troubles during boot with the 3.13.5 kernel and systemd 210-2. It only happens with the ck kernel, neither default ARCH nor AUR mainline kernel seem to be affected. The boot process always times out when mounting the local file systems and drops to the rescue shell. After a downgrade to systemd 208-11, the current ck kernel works fine again. You can find the boot log here: http://pastebin.com/zxsPpEmz. Am I overlooking something or is there anyone experiencing the same problem?

graysky commented on 2014-03-04 20:39

Updated nvidia-ck and fixed the postinstall scriptlet in virtualbox-ck-foo-modules. Thanks for reporting.

kyak commented on 2014-03-04 14:12

hi graysky,

it seems that virtualbox-ck-host-modules-ivybridge contains something after EOF in post_upgrade script...

(18/21) обновление virtualbox-ck-host-modules-ivybridge [#####################################] 100%
/tmp/alpm_gxCJK2/.INSTALL: line 23: warning: here-document at line 7 delimited by end-of-file (wanted `EOF')
/tmp/alpm_gxCJK2/.INSTALL: line 24: ошибка синтаксиса: неожиданный конец файла
/usr/bin/bash: post_upgrade: команда не найдена

sir_lucjan commented on 2014-03-04 10:31

graysky - sorry, my mistake.

dlh commented on 2014-03-04 06:55

:: nvidia-ck-core2: requires nvidia-utils=331.49

local/nvidia-utils 331.49-1
NVIDIA drivers utilities

dlh commented on 2014-03-04 06:54

:: nvidia-ck-core2: requires nvidia-utils=331.49

eduardoeae commented on 2014-03-03 23:27

We still love you <3

graysky commented on 2014-03-03 22:45

NP. Very angry about that stupid mistake (my bad with vimdiff updating the PKGBUILD against the ARCH PKGBUILD and an older copy of linux-ck's) having build 3/4 of the damn repo packages and now having to start over... >:(

sir_lucjan commented on 2014-03-03 22:42

Thank you, graysky.

graysky commented on 2014-03-03 22:39

Bump to v3.13.5-2

Changelog: BFQ version v7r2.

graysky commented on 2014-03-03 22:39

Grrrr.... thanks for pointing that out, fixed in -2.

sir_lucjan commented on 2014-03-03 21:22

Why did you use BFQ v7?

http://algo.ing.unimo.it/people/paolo/disk_sched/patches/3.13.0-v7r2/

The newest version is v7r2.

graysky commented on 2014-03-03 21:15

Bump to v3.13.5-1

Changelog: https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.13.5
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

eduardoeae commented on 2014-03-03 17:55

And BFQ v7r3 "may be around the corner"
https://groups.google.com/forum/#!topic/bfq-iosched/UsIMqP3zxFo

eduardoeae commented on 2014-03-03 17:52

BFS 446 for 3.13 available
http://ck-hack.blogspot.com.ar/2014/03/bfs-0446-313-ck1.html

graysky commented on 2014-03-01 10:30

@shillshocked, Fixed, thanks. modprobed-db still works if you call it via modprobed_db due to a symlink I put in its Makefile:

# symlink for compatibility due to name change
ln -s $(PN) "$(DESTDIR)$(BINDIR)/modprobed_db"

shillshocked commented on 2014-03-01 07:22

@siamer With this kernel I didn't need it. Perhaps you do, though. If you use Nvidia or Intel the parameter is slightly different.

shillshocked commented on 2014-03-01 05:14

Optional dependency "modprobed-db" should read "modprobed-db", not "modprobed_db"

shillshocked commented on 2014-03-01 05:14

Optional dependency "modprobed-db" should read "modprobed_db"

shillshocked commented on 2014-03-01 04:32

@siamer You may have to turn off power management with the latest kernel under some configurations. I did (not using this kernel, but a plain vanilla one) under 3.13.5-1-ARCH. In my case all I had to do was add radeon.runpm=0 to the kernel boot options.

siamer commented on 2014-02-24 14:34

I can't boot with last version (3.12.13-1). getting black boot screen without any error or message... using laptopa lenovo y580 with uefi. any ideas what's the problem ?

graysky commented on 2014-02-22 22:02

Bump to v3.12.13-1

Changelog: https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.12.13
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2014-02-22 21:57

Will bump as soon as upstream code is available.

clfarron4 commented on 2014-02-22 21:50

https://www.kernel.org/ have released 3.12.13.

graysky commented on 2014-02-22 15:24

No problem. I really wish the 'flag out-of-date' button required a reason to be entered like the one on the official packages to help maintainers understand why. For example, this package could be dated for several reasons:
*kernel code
*ck1 code
*BFQ code

sash-arch commented on 2014-02-22 14:29

Sorry graysky, you're right - i missunderstood that.

graysky commented on 2014-02-22 00:22

@sash-arch - Why out-of-date? No ck1 patchset for the 3.13 tree yet.

graysky commented on 2014-02-20 23:16

Bump to v3.12.12-1

Changelog: https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.12.12
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2014-02-15 02:23

BFS is v0.444. Do you think it should be otherwise?

RubenKelevra commented on 2014-02-14 22:41

You should update the description, its telling Brain Fuck Scheduler v0.444...

graysky commented on 2014-02-14 21:39

Bump to v3.12.11-2

Changelog: Added BFQ v7r2.
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2014-02-14 21:38

Bump to v3.12.11-1

Changelog: https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.12.11
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2014-02-14 21:27

eduardoeae - got it in 3.12.11-2-ck1

eduardoeae commented on 2014-02-14 21:18

BFQ v7r2 is available

graysky commented on 2014-02-08 12:50

I am now hosting the BFQ patches until their site is back up. Access them like so:

1) Download the linux-ck source tarball, untar it, enter the linux-ck directory
2) sed -i 's|^_bfqpath=.*$|_bfqpath=\"http://repo-ck.com/source/mirror\"|' PKGBUILD

Now you can build it.

dlh commented on 2014-02-08 12:00

algo.ing.unimo.it down some time ago.

Schmeidenbacher commented on 2014-02-08 11:57

Can't download anything form algo.ing.unimo.it. First i was getting timeouts, then the ip adress changed several times in a few minutes (i did "follow" it with several nslookups), and after it stabilized i got a "network unreachable" error.

Several "is it down" sites on the net also report it as down.

graysky commented on 2014-02-07 08:56

Bump to v3.12.10-2

Changelog: Added BFQ v7r1.
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

felixonmars commented on 2014-02-07 02:17

FYI, a new release of BFQ just out: v7r1

graysky commented on 2014-02-06 22:55

Bump to v3.12.10-1

Changelog: https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.12.10
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2014-02-06 21:29

No ck patch set = not out-of-date. Please flag again once ck releases the new version for 3.13.x. Thanks.

graysky commented on 2014-02-06 21:27

Det - Please flag-out-of-date when a new version of nvidia-utils hits [extra]; I checked svn and nothing is changed for the official package. Thanks.

graysky commented on 2014-02-02 18:50

Bump to v3.12.9-3

Changelog: Added patch fixing FS#38737 (security).
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2014-01-29 14:31

Bump to v3.12.9-2

Changelog: Added BFQ v7.
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

Raziel23 commented on 2014-01-29 14:23

BFQ-v7 is out:
https://groups.google.com/forum/?fromgroups#!topic/bfq-iosched/n4A5602SETA

graysky commented on 2014-01-25 20:05

Bump to v3.12.9-1

Changelog: https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.12.9
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2014-01-25 15:05

@AnAkkk - For many, yes, but a few still reported issues as I recall. Perhaps we can test 300 Hz when bfs for 3.13 is released.

AnAkkk commented on 2014-01-25 14:58

Is there any reason to keep _1k_HZ_ticks=y now? Weren't the suspend and resume problems fixed? :)

graysky commented on 2014-01-17 19:19

Bump to v3.12.8-1

Changelog: https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.12.8
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2014-01-12 22:57

Bump to v3.12.7-2

Changelog: Mirror official version bump - fix 15 seconds nfs hang, #37886 kernel oops[1]
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

1. https://projects.archlinux.org/svntogit/packages.git/commit/trunk?h=packages/linux&id=e4f568d6e0d36df218cd461f2cb4dc5e01edb695

graysky commented on 2014-01-12 22:57

Bump to v3.12.7-1

Changelog: Mirror official version bump - fix 15 seconds nfs hang, #37886 kernel oops[1]
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

1. https://projects.archlinux.org/svntogit/packages.git/commit/trunk?h=packages/linux&id=e4f568d6e0d36df218cd461f2cb4dc5e01edb695

graysky commented on 2014-01-11 12:53

Bump to v3.12.7-1

Changelog: https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.12.7
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

paulfurtado commented on 2014-01-06 01:39

Just rebuilt it and can confirm that I have BFQ again. Thanks for the quick fix!

graysky commented on 2014-01-05 16:46

...thanks for letting me know by the way!

graysky commented on 2014-01-05 16:42

I see the problem... a mistake on my part when I fixed a previous mistake regarding the tick rate a few versions ago. I have fixed it now but have not bumped the pkgver since this is the only change (ie no BFQ in manually built packages of 3.12.6-1-ck). Repo-ck users have the scheduler due to the build script being smarter than its author :p

paulfurtado commented on 2014-01-05 14:39

Thanks for the responses, graysky and coderkun.

$ uname -a
Linux paul-desktop 3.12.6-1-ck #1 SMP PREEMPT Sat Jan 4 17:40:03 EST 2014 x86_64 GNU/Linux

$ cat /sys/block/sda/queue/scheduler
noop deadline [cfq]

graysky commented on 2014-01-05 13:24

@paul - BFQ is available as a module unless you modified that behavior. As coderkun posted, you too post the output of:

cat /sys/block/sda/queue/scheduler

coderkun commented on 2014-01-05 13:00

@paulfurtado: BFQ is available for me.

$ uname -r
3.12.6-1-ck

$ cat /sys/block/sda/queue/scheduler
noop deadline cfq [bfq]

paulfurtado commented on 2014-01-05 12:44

After upgrading from 3.12.5 to 3.12.6, the bfq IO scheduler is no longer available on both my desktop and laptop. Looking at the diff (https://wiki.archlinux.org/index.php/Linux-ck/Changelog), it doesn't look like this is intentional. Is anyone else experiencing this?

coderkun commented on 2013-12-26 11:22

I have a lot of zombie-processes from different parent processes lately. May it be the kernels fault?

graysky commented on 2013-12-22 11:42

@Raziel23 - Fixed, but didn't bump the pkgver to avoid rebuilding since 1k is defaulted. Thanks for reporting.

Raziel23 commented on 2013-12-22 09:32

The _1k_HZ_ticks option is not working properly anymore, because in the config files the tick rate was setup to 1000 HZ (previously it was 300 HZ). I presume that it's the minor omission :)

Best regards.

graysky commented on 2013-12-20 22:41

Sorry about that... had a copy in the builddir when I generated the sums. Fixed now without bumping the pkgver. Thanks for reporting.

Jonhoo commented on 2013-12-20 22:38

Checksum for gcc_patch/enable_additional_cpu_optimizations_for_gcc.patch.gz fails.

graysky commented on 2013-12-20 21:34

Bump to v3.12.6-1

Changelog: https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.12.6
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2013-12-12 14:28

Bump to v3.12.5-1

Changelog: https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.12.5
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2013-12-11 17:13

@qpalz - http://repo-ck.com/source/ck-patches/patch-3.12-ck2.bz2

qpalz commented on 2013-12-11 12:22

ck.kolivas.org is down again. Could you please put it to repo-ck.com?

graysky commented on 2013-12-10 19:41

Bump to v3.12.4-1

Changelog: https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.12.4
Commit: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2013-12-06 00:23

Bump to v3.12.3-1

Changelog: https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.12.3
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=cffd5589ff10b99bca0e992dc1b3eb2904275f34

graysky commented on 2013-12-05 20:14

Bump to v3.12.3-1

Changelog: https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.12.3
Commit: Looks like pkgbuild.com isn't updating.

graysky commented on 2013-12-04 02:02

Bump to v3.12.2-2
Changelog: Version 0.444 of BFS in ck2.
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=677c8bb6088cd5673deb95dd4ba57b0def340b5a

graysky commented on 2013-12-04 02:02

Bump to v3.12.2-2
*Changelog: Version 0.444 of BFS in ck2.
*Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=677c8bb6088cd5673deb95dd4ba57b0def340b5a

graysky commented on 2013-11-27 00:36

Bump to v3.12.1-3

Changelog: Updated x86_64.config per Arch devs; no change to i686 config.
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=133009feae1652c1501c84bc503de7cd27ca6238

graysky commented on 2013-11-22 01:31

Bump to v3.12.1-2

Changelog: Added btrfs patch (suggestion by wonderwoofy), see https://bugs.archlinux.org/task/37867
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=777be749193553e3632e6941857572e5b73c90d3

graysky commented on 2013-11-21 01:55

Bump to v3.12.1-1

Changelog: https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.12.1
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=8cc6734615ef4f6362e507636b207dc7e811bcec

graysky commented on 2013-11-19 16:36

Bump to v3.12.0-1

LinuxChanges: http://kernelnewbies.org/Linux_3.12
Changelog: https://lkml.org/lkml/2013/11/3/160
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=27bbb0cca481e840790335f44ec3678348f37d13

Schmeidenbacher commented on 2013-11-18 22:48

Thanks. Yep, it's all there now. Excellent.

graysky commented on 2013-11-18 22:45

@Schmeidenbacher - Try now, thanks for reporting.

Schmeidenbacher commented on 2013-11-18 22:41

I'm getting a 404 for
http://repo-ck.com/source/gcc_patch/kernel-312-gcc48-1.patch.gz

PerfectGentleman commented on 2013-11-18 15:28

@graysky, yep, patсhes for 3.12 are ready

PerfectGentleman commented on 2013-11-18 13:22

@graysky, yep, pathes for 3.12 are ready

graysky commented on 2013-11-14 15:53

@AnAkkk - why out of date.

graysky commented on 2013-11-13 20:09

Bump to v3.11.8-1

Changelog: https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.11.8

graysky commented on 2013-11-05 20:38

Bump to v3.11.7-1

Changelog: https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.11.7
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=91d96e2ff023637119e67b49658b3d176b07d5e7

graysky commented on 2013-11-05 00:03

Bump to v3.11.7-1

Changelog: https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.11.7
Commit: --- pkgbuild.com/git/aur-mirror.git is down ---

graysky commented on 2013-10-18 23:08

Bump to v3.11.6-1

Changelog: https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.11.6
Commit: --- pkgbuild.com/git/aur-mirror.git is down ---

graysky commented on 2013-10-14 16:26

Odd, on my test boxes, the old versions worked just fine. In any case, PKGBUILDs have been updated and repo package are syncing now. Refresh in 10 min or so to pull down the new package, then:

# rmmod nvidia
# modprobe nvidia

Then restart your X server and all should be well... or if you're lazy, just update and reboot.

csmk commented on 2013-10-14 16:11

@graysky: after upgradig to 3.11.5-1 my nvidia module has become unusable.

nvidia: disagrees about version of symbol module_layout
modprobe: ERROR: could not insert 'nvidia': Exec format error

I'm using nvidia-ck-k10 from repo-ck.
Please, can you help me?

graysky commented on 2013-10-14 10:45

Bump to v3.11.5-1

Changelog: https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.11.5 and miscell. config changes by tpowa that mirror the official kernel package.
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=1ba9d64268d0f8b82e6d120eefb1041b9612801a

graysky commented on 2013-10-14 10:45

Bump to v3.11.5-1

Changelog: https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.11.5 and miscell. config changes by tpowa that mirror the official kernel pacakage.
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=1ba9d64268d0f8b82e6d120eefb1041b9612801a

craigcabrey commented on 2013-10-09 16:26

In case anyone had any problems with a disappearing mouse cursor in GNOME 3.10 after installing this package, try running

`gsettings set org.gnome.settings-daemon.plugins.cursor active false`

See bug https://bugzilla.gnome.org/show_bug.cgi?id=694758.

graysky commented on 2013-10-08 18:34

Seems as though the server is back up.

fishburn commented on 2013-10-08 08:07

Link still down...

Is there some reason these patches aren't kept in the PKGBUILD tarball?

hadrons123 commented on 2013-10-07 15:25

==> ERROR: Failure while downloading 0001-block-cgroups-kconfig-build-bits-for-BFQ-v6r2-3.11.patch

Link is down. Is there any other alternative?

graysky commented on 2013-10-05 16:24

Bump to v3.11.4-1

Changelog: https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.11.4
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=cb0eeef388e2344efaefdb445eef2a371df151ea

graysky commented on 2013-10-01 20:04

Bump to v3.11.3-1

Changelog: https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.11.3
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=400c9cfcc3965334fc814a4247684ac464554d00

graysky commented on 2013-09-29 08:54

@jubjub - You'll need to post the output; works for me.

Anonymous comment on 2013-09-29 03:52

==> ERROR: A failure occurred in prepare().
Aborting...

...happens every time even with vanilla PKGBUILD

graysky commented on 2013-09-27 23:23

Bump to v3.11.2-3

Changelog: https://lkml.org/lkml/2013/9/26/723

hepha commented on 2013-09-22 12:34

@graysky

my bad
i open KERNEL_LZ4 but i no install community/lz4
so A failure occurred in build ...

graysky commented on 2013-09-21 14:32

@hepha - Not sure why that error is present... have you modified the PKGBUILD? Which options have you enabled? Post the output of:

head -n48 PKGBUILD| sed '/#/d'

hepha commented on 2013-09-21 14:29

can you help me
thanks

pkgver=3.11.1
pkgrel=2
arch=('x86_64')
LD [M] sound/pci/hda/snd-hda-intel.ko
LD [M] sound/soundcore.ko
==> ERROR: A failure occurred in build().
Aborting...

kyak commented on 2013-09-19 18:33

I was too fast with my comment, sorry for that. This fix only makes difference for normal kernel, where it makes hibernation work. Still i face sporadic handing with -ck when hibernating. -\

kyak commented on 2013-09-19 18:26

This might have something to do with alx driver which recently has went upstream.. After applying the rmmod/modprobe fix as described here: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1173952 (i actually used alx.sh fix from https://aur.archlinux.org/packages/dkms-alx/), hibernation seems to work. "Seems" because i only tried couple of times, let's see if it fails during the net few days.

Anonymous comment on 2013-09-18 13:38

@graysky

Yes suspend works fine. I have the same problem in 3.10, and it still goes with 3.11.

graysky commented on 2013-09-17 00:35

CK's blog is the best place. Does suspend work? Hibernate is different from suspend.

Anonymous comment on 2013-09-17 00:22

@kyak

hibernate to RAM is buggy for -ck AUR since 3.10, I have not using -ck for some critical text editing. However if someone could tell me how to record the bug ? The system freeze when hibernate and only a hard reset can do.

My box is X230, i7 with HD3000 GPU.

kyak commented on 2013-09-16 19:37

hi graysky, thanks for pushing this change (it doesn't make any bad anyway).
Unfortunately, it doesn't fix the problem for me. First time, it hibernates fine; the second time my laptop freezes hard -(
Won't try any more so as not to rape my laptop

graysky commented on 2013-09-16 17:48

@lano- Until CK or upstream is able to fix the underlying cause of the suspend lockup, this serves as a functional fix, no?

lano1106 commented on 2013-09-16 17:34

I always recompile linux-ck with CONFIG_HZ_1000 but you do realize that this does not really fix the problem.

All it does, is change the timing of events that hides the original race condition that creates the lock up.

dlh commented on 2013-09-16 14:55

@graysky: 1000Hz works for me

graysky commented on 2013-09-16 13:26

@misc - I agree with you in general but in this case, it was a necessary evil so that this specific change could be pushed out the ~2,500 repo users. If indeed the tick rate fixes the suspend problems, many will rejoice.[1]

1. https://bbs.archlinux.org/viewtopic.php?pid=1325218#p1325218

misc commented on 2013-09-16 13:00

I'd prefer if that sort of update wasn't accompanied by a pkgrel bump. Hardly anyone is affected by this, yet still everyone has to recompile resp. manually postpone it on every yaourt/etc. run til the next "real" bump.

graysky commented on 2013-09-16 10:15

----
Bump to v3.11.1-2

Changelog: Added an option in the PKGBUILD to set the default tick rate from 300 to 1000 per reports (2 people now) on CK's blog that this fixes the suspend/resume issues people have been reporting.[1] Also cleaned-up the comments in the PKGBUILD.
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=425cec6886d91723fa3d67b972c5419e31a524be

1. http://www.blogger.com/profile/12086606957383598779

graysky commented on 2013-09-16 10:15

@dlh - Yes and no. See: http://ck.kolivas.org/patches/bfs/bfs-faq.txt

"THESE ARE OPTIONAL FOR LOWEST LATENCY. YOU DO NOT NEED THESE!
Configure your kernel with 1000Hz, preempt ON and disable dynamic ticks."

dlh commented on 2013-09-16 10:11

@graysky: I thought that 1000 Hz is recommended for BFS.

graysky commented on 2013-09-16 09:14

@dlh - I am building packages now for the repo that will have 1k ticks enabled just like the default PKGBUILD recommends. If indeed this solves the suspend issues people are having, I think upstream needs to do this in the ck1 patchset itself... either that or figure out why some machines hang when suspending/waking up with the patchset.

graysky commented on 2013-09-16 09:13

@dlh - I am building packages now for the repo that will have 1k ticks enabled just like the default PKGBUILD recommends. If indeed this solves the suspend issues people are having, I think upstream needed to do this in the ck1 patchset itself... either that or figure out why some machines hang when suspending/waking up with the patchset.

dlh commented on 2013-09-16 09:06

@graysky: do you compile your packages in repo-ck with "CONFIG_HZ_1000"?

graysky commented on 2013-09-14 16:05

Bump to v3.11.1-1

Changelog: https://lkml.org/lkml/2013/9/14/73
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=364a26f076da66e54e66071c7ab7f1157d5ba211

graysky commented on 2013-09-14 16:05

Bump to v3.11.1-1'''

Changelog: https://lkml.org/lkml/2013/9/14/73
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=364a26f076da66e54e66071c7ab7f1157d5ba211

graysky commented on 2013-09-09 13:26

Bump to v3.11-1

LinuxChanges: http://kernelnewbies.org/Linux_3.11
Changelog: https://lkml.org/lkml/2013/9/2/402
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=c7529b6849ae49a74c270eef9499711e48ba4d27

maxrd2 commented on 2013-09-09 12:54

CK patches for 3.11 are out: http://ck.kolivas.org/patches/3.0/3.11/3.11-ck1/ :D

graysky commented on 2013-09-08 10:58

Bump to v3.10.11-1

Changelog: https://lkml.org/lkml/2013/9/8/7
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=5430bd538516a7333f6fa81f30a970e933862900

graysky commented on 2013-09-07 00:44

BFQ is ready for 3.11 series. For those of you [misc] who like to see or build/run the pre-release of the next linux-ck as we wait for CK to patch into the new tree, I am hosting the incomplete (no ck1) stuff here:

http://repo-ck.com/PKG_source/ck-next/ck-next-1.0.tar.xz

Note that all the PKGBUILDs/files for the xxx-ck stuff are in the tar file:

% tar tf ck-next-1.0.tar.xz| grep PKGBUILD
./lirc-ck/PKGBUILD
./virtualbox-ck-modules/PKGBUILD
./broadcom-wl-ck/PKGBUILD
./nvidia-ck/PKGBUILD
./nvidia-304xx-ck/PKGBUILD
./linux-ck/PKGBUILD

SanskritFritz commented on 2013-09-05 07:13

@dlh Yes, my mouse doesn't stutter when CPU load is high or copying large files in the background.

dlh commented on 2013-09-05 07:09

Do you guys noticed any difference using ck kernel in context of workflow?

skydrome commented on 2013-09-04 04:14

its also a good idea not to jump on new kernels like this the instant they are released. Waiting for a .1 or .2 is more pleasant. Which is likely when BFS will be adopted to it.

runnytu commented on 2013-09-03 10:05

Ok i understand, thanks for the info, had not given the opportunity to see a change in kernel version in the early days until now, had always been after, when the packages are released.

graysky commented on 2013-09-03 09:52

Example, virtualbox: https://mailman.archlinux.org/pipermail/arch-general/2013-September/034088.html

graysky commented on 2013-09-03 09:50

You can help by keeping an eye out for the ck1 patchset. It usually takes CK 1-2 weeks to get a patchset together for a major version bump. 3.11-1-ARCH has been published, so configs and patches, etc. aren't an issue. The only other must have is a BFQ patchset for the 3.11 tree. Paolo et al usually do a pretty good job timewise. Another wildcard is are the speed at which devs patch needed modules to build with 3.11. Nvidia and virtualbox for example.

runnytu commented on 2013-09-03 09:41

@graysky I flagged the package before read your post, sorry for the inconvenients, i don't know what are the differents procedures when a major release of kernel are taken, sorry for that, i wait until upstream code and all patches for ck and bfs are released.

graysky commented on 2013-09-03 09:00

@runnytu -did you read my last post??? Why are you flagging this package when upstream code doesn't yet exist?

graysky commented on 2013-09-02 21:50

Guys - I realize that 3.11.0 was just mainlined, but we don't yet have several of the prereqs to make a proper linux-ck release[1]:
*ck1
*bfq

Please flag it again once we have them in case I miss it.

1. https://wiki.archlinux.org/index.php/Linux-ck#Release_Cycle

graysky commented on 2013-09-01 15:00

Use this sed oneliner IN THE SAME DIRECTORY AS THE PKGBUILD to redirect to repo-ck.com's mirror for the patch:

sed -i 's|^\"http://ck.kolivas.*$|\"http://repo-ck.com/source/mirror/patch-3.10-ck1.bz2\"|' PKGBUILD

vic.pozd commented on 2013-09-01 14:09

http://ck.kolivas.org - now down.
I can't get patch-3.10-ck1.bz2 :(

Anonymous comment on 2013-08-30 08:28

@graysky

nop, my problem is not like [1]. It does not report any thing and the screen goes blank after I select the arch-ck boot entry.

I am using UEFI boot mode with CSM disabled. My box is a lenovo x230 with intel HD4000 gpu cards. 3.10.9-1 can boot ok but I suffers from random freeze/lockup so I think upgrade to 3.10.10-1 might solve the problem.

Anonymous comment on 2013-08-30 08:26

@graysky

ok ! You mean disable BFQ by

(1) modifying PKGBUILD _BFQ_enable_=n or
(2) modify the kernel boot line : elevator=bfq => elevator=noop ?

BTW : I donnot know whether this patch [1] in [2](#19) related to the problem

[1] : https://bugzilla.kernel.org/attachment.cgi?id=107199&action=diff
[2] : https://bbs.archlinux.org/viewtopic.php?id=167518

Anonymous comment on 2013-08-30 08:25

@graysky

ok ! You mean disable BFQ by

(1) modifying PKGBUILD _BFQ_enable_=n or
(2) modify the kernel boot line : elevator=bfq => elevator=noop ?

BTW : I donnot know whether this patch [1] in [2](#19) related to the problem

[1] : https://bugzilla.kernel.org/attachment.cgi?id=107199&action=diff
[2] : https://bbs.archlinux.org/viewtopic.php?id=167518

graysky commented on 2013-08-30 07:42

@liubenyuan - Is it the same problem described here[1]? Are you booting with BFQ enabled? What if you disable it? Paolo will need to know if his new patch does not fix the problem.

1. https://bbs.archlinux.org/viewtopic.php?id=168583

Anonymous comment on 2013-08-30 07:38

@graysky

anyway, i failed to boot 3.10.10-1 and I am waiting for your magic update like 3.10.8 -> 3.10.9 :)

And one problem I need report is that once I upgrade to 3.10.9, the system suffers with random lockup and a hard reset must be done.

graysky commented on 2013-08-30 07:14

@liubenyuan - Not according to Paolo:

On 08/28/2013 03:33 AM, Paolo Valente wrote:
> I'm not sure I fully understood your question, anyway:
> to add BFQ-v6r2 to a 3.10.8+ kernel from scratch, the patchset [1] is all you need
>
> I hope this addresses your issue,
> Paolo
>
> Il giorno 27/ago/2013, alle ore 21:04, member graysky ha scritto:
>
>> > Hi Paolo. I see you posted some new BFQ code[1]. Tell me, are these considered "stable?" Also, I noticed that I can still patch the included one from that google group link on top of these 3 you posted. Doesn't that mean that your patches are different from the one there? Should I still patch your 3 plus the one by Arianna Avanzini[2]?
>> >
>> > 1. http://algo.ing.unimo.it/people/paolo/disk_sched/patches/3.10.8+-v6r2/
>> > 2. https://groups.google.com/group/bfq-iosched/attach/46d6e501b377bdef/0001-block-Switch-from-BFQ-v6r2-for-3.10.0-to-BFQ-v6r2-fo.patch?part=4&authuser=1&view=1
>> >

Anonymous comment on 2013-08-30 05:51

@graysky

do we still need the following patch ?
1. https://groups.google.com/forum/#!msg/bfq-iosched/f4Lg5INzQ-k/EFbHNzfjFTsJ

'cause my box failed to boot after upgrading to 3.10.10-1

graysky commented on 2013-08-29 20:25

Bump to v3.10.10-1

Changelog: https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.10.10 and updated BFQ patchset
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=dacf417b9f9fafee7493fcc1bb03e691633c9dd6

Gringo commented on 2013-08-22 13:04

Thanks for the quick update. I had the same hard lockup at boot, which forced me to move back to the vanilla kernel. linux-ck 3.10.9 works like charm however!

OrdinaryMagician commented on 2013-08-21 22:48

Oh, I'll be taking that patch too, then...

graysky commented on 2013-08-21 22:21

Bump to v3.10.9-1

Changelog: https://lkml.org/lkml/2013/8/20/677 and added patch for BFQ[1]
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=ff178cde2a2adc84e515dae379dc45322b4b5941

1. https://groups.google.com/forum/#!msg/bfq-iosched/f4Lg5INzQ-k/EFbHNzfjFTsJ

runnytu commented on 2013-08-21 17:00

With the patch works ok, i'm running 3.10.9-ck ok.

graysky commented on 2013-08-21 16:49

I will look into that BFQ patch when I are lease 3.10.9-1 later this afternoon, my understanding through reading anecdotal reports is that indeed, BFQ and 3.10.8 do not play well together.

PS sorry about the multiple posts. Posting from a phone sucks,

graysky commented on 2013-08-21 16:47

I will look into that BFQ patch when I release 3.10.9-1 later this afternoon, my understanding through reading ancidotial reports is that indeed, BFQ and 3.10.8 do not play well together.

graysky commented on 2013-08-21 16:47

I will look into that BFQ patch when I are lease 3.10.9-1 later this afternoon, my understanding through reading ancidotial reports is that indeed, BFQ and 3.10.8 do not play well together.

Anonymous comment on 2013-08-21 15:07

Somebody got the patch from:
https://groups.google.com/forum/#!msg/bfq-iosched/f4Lg5INzQ-k/EFbHNzfjFTsJ
to work?
When should it be applied?

xcomcmdr commented on 2013-08-21 14:22

@clayman
I've had hard lockups at boot-time with linux-ck and the vanilla kernel for a long time. It's rare and random, however. :/

3.10-8-1-ck gives me the same error as BorgHunter. I reverted to 3.10-7-1-ck and it works fine. I do not have "elevator=something" in my kernel boot line.

clayman commented on 2013-08-21 12:14

@Cosmik Thanks, that's an interesting observation. I have just recompiled the kernel and managed to boot up but it seems I forgot to set BFQ as the default queue scheduler, which explains it.

csmk commented on 2013-08-21 11:29

@clayman, me too. The hard lockup is gone after I removed the "elevator=bfq" from the kernel boot line.

csmk commented on 2013-08-21 11:24

@clayman, me too. The hard lockup is gone after I removed the "elevator=bfq" from kernel boot line.

clayman commented on 2013-08-21 11:18

@BorgHunter Thanks, I have the same issue. I have already tried 3.10.9 by editing PKGBUILD but the problem persists. :(

clayman commented on 2013-08-21 11:07

@BorgHunter Thanks, I have the same issue. I have already tried 3.10.9 by editing PKGBUILD but the problem persists. :(

BorgHunter commented on 2013-08-21 11:02

3.10.8 definitely causes problems for me. Hard lockup at boot with the message "Fixing recursive fault but reboot is needed" or something like that (can't find any logs for it). Looks like, according to the changelog, 3.10.9 might solve this. I'm on the vanilla kernel (3.10.7) until ck is updated.

clayman commented on 2013-08-21 10:47

Since 3.10.8 I'm experiencing hard lockups (according to the watchdog) at boot. Does this happen to anybody else?

ajs124 commented on 2013-08-21 00:05

https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.10.9

graysky commented on 2013-08-20 21:16

Bump to v3.10.8-1

Changelog: https://lkml.org/lkml/2013/8/20/418
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=6b9ca6543b53836ff5d5becda24ea975332afba2

graysky commented on 2013-08-20 19:01

Compiles fine for me using makepkg... it looks like you're using a helper. Try using `makepkg -src` and post back.

LeetShiva commented on 2013-08-20 12:11

I forgot to mention that I'm running AMD APU A6 3670k with built in IGP Radeon HD 6530D

LeetShiva commented on 2013-08-20 12:07

I'm getting seg. fault when compiling linux-ck 3.10.7-2 all other 3.10 versions and I compiled each of them have worked properly and compiled properly.

How and where to file bug report about this?
This is paste of this compile error http://pastebin.com/LUT2c3cC
Anyone have any hints about what may cause a problem or maybe had similar issue ?

graysky commented on 2013-08-16 18:43

Guys - Please use CK's blog[1] for discussions about suspend; I am not a kernel developer, just a packager of this kernel.

1. http://ck-hack.blogspot.com/2013/07/bfs-0440-ck-for-linux-310.html#comment-form

notthesun commented on 2013-08-16 18:40

suspend and hibernation not works intel nehalem

graysky commented on 2013-08-15 20:23

Bump to v3.10.7-2

Changelog: Added patch for logitech receivers per official Arch package.
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=73ed08838b578de92231b76a317ff8e671e6bbd7

graysky commented on 2013-08-15 20:23

Bump to v3.10.7-2
Changelog: Added patch for logitech receivers per official Arch package.
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=73ed08838b578de92231b76a317ff8e671e6bbd7

Anonymous comment on 2013-08-13 21:44

suspend & hibernation still broken for me (intel sandy bridge, ck-sandybridge, 3.10.6-2-ck) :(

graysky commented on 2013-08-13 14:34

Bump to v3.10.6-2

Changelog: fixed flayspray #35991 and #36296
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=3dcb0c9920cec68fe946d7b390e1fcb5ed4a7f0c

Jonhoo commented on 2013-08-13 13:52

I guess recompiling from scatch is the way to go then. At least ccache is helping a bit :)

graysky commented on 2013-08-13 13:07

Bump to v3.10.6-2

Changelog: fixed flayspray #35991 and #36296
Commit:

graysky commented on 2013-08-13 12:32

@jonhoo - I don't believe this is a limitation of the linux-ck package but of how makepkg works. I am not an expert; I recommend that you open a thread asking how to compile a kernel as you want to do for someone with real knowledge about it to comment.

dlh commented on 2013-08-13 12:23

Well in gentoo linux, first of all you downloading the whole source of proper series, for instance 3.10 and then set of patches. Then scripts for managing updates will unpack 3.10 and then apply all patches starting from 3.10.1 to one which is demand in this case 3.10.6 so 6 patches will be applied during this process.

qpalz commented on 2013-08-13 12:18

@Jonhoo

Kernel patches is not totally incremental. Patch for 3.10.6 is for 3.10.0 -> 3.10.6, but NOT for 3.10.5 -> 3.10.6. You cannot update the src from 3.10.5 to 3.10.6 by simply applying the patch of 3.10.6.

Also, as far as I know, you should always recompile everything even if you have only changed some configuration. There is some bugs in the Makefile. (Not sure whether this is true now)

Jonhoo commented on 2013-08-13 10:43

@graysky: As I wrote, removing src/ is what I'm currently doing, but this means make has to recompile every single file, which is a bit of a waste. The kernel patches are incremental, so applying them on top of an existing source tree should upgrade that source tree without needing to run make from scratch. I don't know why this isn't working with linux-ck though...

graysky commented on 2013-08-13 10:16

Bump to v3.10.6-1

Changelog: https://lkml.org/lkml/2013/8/11/143
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=2f5a2c4a30883b51da063e753473decf6e173aba

graysky commented on 2013-08-13 10:15

@Jonhoo - You need to remove your src dir prior to building:
% rm -rf src pkg && makepkg -srci

Jonhoo commented on 2013-08-12 19:26

Quite often when a new patch level is released and I run makepkg, I see:

The next patch would create the file ...,
which already exists! Assume -R? [n]

This is rather annoying, as the only solution I've found thus far is to rm -r src/. Not entirely sure why this is happening? This time the file in question was "drivers/acpi/acpi_cmos_rtc.c", but I'm fairly certain other files have been the problem before.

graysky commented on 2013-08-11 11:01

@Det - Ah, thanks. It technically is but the difference there is purely cosmetic[1] and will be included in 3.10.6-1 whenever Greg releases that. It seems too that tpowa has been tweaking the Arch configs and added a patch[2] which I have incorporated into linux-ck, again these will be included in 3.10.6-1. When you guys hit the flag button, I always have to review the various patchsets (ck, BFQ, upstream, etc.) to figure out why it's out-of-date :p

1. https://github.com/graysky2/kernel_gcc_patch/commit/8a8b2af1c9891d13665088eb09d93c08f9272657

2. https://projects.archlinux.org/svntogit/packages.git/log/trunk?h=packages/linux

Det commented on 2013-08-11 10:54

I just thought your own patch was out of date.

graysky commented on 2013-08-10 21:12

@Det - Why out-of-date? 3.10.6 isn't due out for at least 4 more hours[1]. When I checked kernel.org, 3.10.6-rc1 is still current. Did I miss something?

1. https://lkml.org/lkml/2013/8/8/775

Herbstmensch commented on 2013-08-07 11:44

Same here, suspend and hibernate doesn't work on archlinux 3.10.5-1-ck-sandybridge.

Anonymous comment on 2013-08-07 08:31

Same as sPHERE, suspend and hibernate doesn't work on archlinux 3.10.5-1-ck-sandybridge.

petterk commented on 2013-08-06 20:42

Suspend and hibernate still doesnt work most of the time.

archlinux 3.10.5-1-ck-ivybridge

graysky commented on 2013-08-06 19:34

@kyak - This can happen when the Arch devs push a major version update of nvidia-utils to the official servers and I am away from my work station or unable to update the nvidia-ck package to match. It is rare and the version conflict will prevent any breakage. I just updated nvidia-ck package to 325.15 and populated the repo with them as well. If you refresh and update you should be good now.

In the future, simply flagging the nvidia-ck package out-of-date will alert me to the Arch update. This is true for nvidia-304xx-ck, lirc-ck, and virtualbox-ck-modules too.

graysky commented on 2013-08-06 19:32

@kyak - This can happen when the Arch devs push a major version update of nvidia-utils to the official servers and I am away from my work station or unable to update the nvidia-ck package to match. It is rare and often only for a few hours. I just updated nvidia-ck package to 325.15 and populated the repo with them as well. If you refresh and update you should be good now.

kyak commented on 2013-08-06 19:14

Hi graysky,

nvidia-ck-ivybridge requires nvidia-utils=319.32, but the arch repo already contains nvidia-utils 325.15-1.

Therefore, i get conflict and can't satisfy dependencies during upgrade.

graysky commented on 2013-08-04 15:09

Bump to v3.10.5-1

Changelog: https://lkml.org/lkml/2013/8/4/20
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=fb03793514ad2d2d5be3dd52c8876b8d83e938d5

graysky commented on 2013-07-29 20:20

Bump to v3.10.4-1

Changelog: https://lkml.org/lkml/2013/7/29/6
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=02ec0dadb8c7357833d514b7563a0e53ef392008
Bullshit to deal with: Servers go down everyday; the world is an imperfect place. Since http://algo.ing.unimo.it is down, I am temp hosting the patched on repo-ck. Just run this sed oneliner in the same dir as the PKGBUILD to adjust the source line to hit my mirror:

sed -i 's|_bfqpath=.*$|_bfqpath="http://repo-ck.com/source/mirror"|' PKGBUILD

graysky commented on 2013-07-29 20:19

Bump to v3.10.4-1

Changelog: https://lkml.org/lkml/2013/7/29/6
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=02ec0dadb8c7357833d514b7563a0e53ef392008
Bullshit to deal with: Servers go down everyday; the world is an imperfect place. Since http://algo.ing.unimo.it is down, I am temp hosting the patched on repo-ck. Just run this sed oneliner in the same dir as the PKGBUILD to adjust the source line to hit my mirror:

*Bullshit to deal with: Servers go down everyday; the world is an imperfect place. Since http://algo.ing.unimo.it is down, I am temp hosting the patched on repo-ck. Just run this sed oneliner in the same dir as the PKGBUILD to adjust the source line to hit my mirror:

sed -i 's|_bfqpath=.*$|_bfqpath="http://repo-ck.com/source/mirror"|' PKGBUILD

test0 commented on 2013-07-29 14:39

It's not just you! http://algo.ing.unimo.it looks down from here.

surfatwork commented on 2013-07-29 08:04

update: shutdown works with gdm.

surfatwork commented on 2013-07-28 20:22

suspend works with 3.10.3-1-ck. didnt work with 3.10-2, even with applying the cpufreq patch.
however not able to shutdown - issues with SLIM login manager.

kyak commented on 2013-07-26 21:08

Same here, my laptop (Asus N76VZ) only hibernates maybe 25% of total attempts. The problem is not reproducible, so i can't really report it. The fact is, starting some time from 3.10 series and up to the latest 3.10.3 i can't rely on hibernation to work..

petterk commented on 2013-07-26 11:02

Cant suspend/hibernate with the newest kernel.

Im using linux-ck-ivybridge on my Asus K55VJ.

graysky commented on 2013-07-26 10:35

Glad to hear it. Note that v3.10.3-1-ck does not contain any out-of-tree patches for this at all. It is pure upstream in this regard.

Anonymous comment on 2013-07-26 08:45

Version 3.10.3-1 seems to have fixed my suspend issues. I had previously tried uncommenting the patch in 3.10.2, but that didn't fix it.

dlh commented on 2013-07-26 06:54

From 3.10.3: cpufreq: Revert commit a66b2e to fix suspend/resume regression

Let's try it.

graysky commented on 2013-07-26 00:49

Bump to v3.10.3-1'''

Changelog: https://lkml.org/lkml/2013/7/25/655
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=9f627b06d90ee4288055397989dd7e6709f17c76

Anonymous comment on 2013-07-25 22:11

Reported. I am using linux-ck-3.10.1 and linux-ck-3.10.2 for two weeks, how ever the -ck patch set suffers from suspend issues for my thinkpad X230 box. About 25% chance when you pressed Fn + F4 (suspend-to-ram), the system locked (may be kernel panic) and the power/sleep led keep on flashing fast. you have to hold the power button to shutdown it.

#1. commented out the line 110 of the PKGBUILD of linux-ck 3.10.2 does not resolve this problem. it still could not function suspend-to-ram properly

#2. I switched to the linux-3.10.2 ARCH kernel yesterday, the suspend-to-ram functioned well so far, 11/11 success.

I do not know whether it is a 3.10 mainline bug or related to the BFS/ck patchset. I installed 3.11-rc2 today and keep on reporting the results, if it is usefull. and how could I save the kernel log when the system hangs ? I found no evidence of such failure in journalctl or dmesg log.

graysky commented on 2013-07-25 19:12

Guys - I have seen 1/2 dozen patches claiming to fix suspend issues. As stated before, I am not adding experimental patches to the PKGBUILD that are enabled. You guys should be savvy enough to just add a patch line to the PKGBUILD to include your favorite ones beyond what I am including. Please do so. I am happy to add a patch to the kernel once it has been accepted by upstream, particularly if it fixes a problem people are having. For the most part, I do not what to be the enabling force add commented lines to apply patches. Please post if you have applied the patch yourself and tested in a robust fashion. In other words, run it for a few days, if it fixes a particular issue, pound on your system trying to reproduce the bug a number of times etc. Above all else, please open a flyspray against the Arch kernel package if you know the issue is contained in mainline since we're all in the same boat, ck patchset or not.

/off my soapbox

jim1960 commented on 2013-07-25 13:01

I'm using linux-mainline v3.11rc2 and nvidia-beta v325.08, just work fine.
It fixed suspend/resume issue with linux v3.10.x (linux-ARCH & linux-ck)

the patch for linux-mainline v3.11rc2, follow the #4 of posts (kernel_v3.10.patch.txt):
https://devtalk.nvidia.com/default/topic/557110/linux/linux-3-11_rcx-compatibility/

CPU: AMD Athlon(tm) II X4 651 Quad-Core Processor (651K)
GPU: nVidia GTS 450

dlh commented on 2013-07-25 09:17

I can say that in 3.10 kernel suspend often work and so in 3.9.x, but sometimes I left with blinking cursor, but 3.10.x is completely broken.

I roll back to 3.9.9 from official repo for now and I'm monitoring the current state.

graysky commented on 2013-07-24 22:45

@indie - I just refreshed 3.10.2-1-ck without bumping the pkgver. What changed? I added the patch you recommended but left it comments out. There is therefore no difference between this version and the original unless the user uncomments line 110. Why? I am won't add an experimental patch to the PKGBUILD that is enabled by default. Please uncomment line 110 and test. Hopefully, if this patch indeed fixes this issue, it will be merged for inclusion in 3.10.3 proper. Also, if I works, I would encourage you to open a bug report against the official Arch kernel package linking the patch and describing your experiences with it.

indie commented on 2013-07-24 21:42

Hi graysky,

I have suspend/resume bug in latest 3.10.x kernel. Looks like the issue is resolved in mainline. Could you please add patches in the package mentioned in http://comments.gmane.org/gmane.linux.kernel/1529561 ?

Thanks

indie commented on 2013-07-24 21:40

Hi graysky,

I have suspend/resume bug in latest 3.10.x kernel. Looks like the issue is resolved in mainline. Could you please add patches in the package mentioned in http://comments.gmane.org/gmane.linux.kernel/1529561 ?

Thanks

graysky commented on 2013-07-22 21:16

Bump to v3.10.2-1

Changelog: https://lkml.org/lkml/2013/7/21/181
Changelog: Added patch to correct Intel graphic freezes. See commented lines in PKGBUILD.
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=8155496e3c7c84451b6f1fe0e50e0a33ad34eb0d

graysky commented on 2013-07-20 19:37

@dlh - Very odd. As you know the ck1 patchset has not changed since its release. 3.10.2 is due out shortly. Hope that'll fix your woes.

dlh commented on 2013-07-20 18:31

>So 3.10.0-1-ck works but 3.10.1-2-ck does not work?
That's right, any of 3.10.1* not working.

graysky commented on 2013-07-20 18:13

So 3.10.0-1-ck works but 3.10.1-2-ck does not work?

dlh commented on 2013-07-20 14:24

>Can you suspend the ARCH kernel of the same version?
Yes, I can

Optimized version of 3.10.0-1-ck for core2 works fine for me (suspend)

dlh commented on 2013-07-18 21:07

>Can you suspend the ARCH kernel of the same version?
Yes, I can

graysky commented on 2013-07-18 11:31

I've read mixed reports about this... Can you suspend the ARCH kernel of the same version? I would recommend that you st on the ck blog.

http://ck-hack.blogspot.com/2013/07/bfs-0440-ck-for-linux-310.html?m=1

dlh commented on 2013-07-18 11:00

I didn't know where to post, I couldn't find bugzilla or something similar. So I decided to do it right here.

I'm using precompiled kernel linux-ck-core2 from your repository and I no more able to suspend my machine.

lip 18 10:12:48 HALF-BLOOD systemd[1]: Stopped A simple open wireless connection.
lip 18 10:12:48 HALF-BLOOD systemd[1]: Started netctl sleep hook.
lip 18 10:12:48 HALF-BLOOD systemd[1]: Starting Sleep.
lip 18 10:12:48 HALF-BLOOD systemd[1]: Reached target Sleep.
lip 18 10:12:48 HALF-BLOOD systemd[1]: Starting Suspend...
lip 18 10:12:48 HALF-BLOOD systemd-sleep[4388]: Suspending system...

Looks like everything is fine, but I end up with flashing dash and all I can do is manually switch power off.

xzy3186 commented on 2013-07-16 03:24

Hi graysky,

I see... Because I enabled CONFIG_HIGHMEM64G in the config of 32-bit linux-ck, there maybe some conflicts between CONFIG_HIGHMEM64G and CONFIG_OF_I2C... Thanks very much.

graysky commented on 2013-07-15 19:38

@xyz - Well, it seems that this is i686 only and does exist in the my ck1 builds.

% ls -l /usr/lib/modules/3.10.1-2-ck/kernel/drivers/of
total 8
-rw-r--r-- 1 graysky users 2233 Jul 14 08:52 of_i2c.ko.gz
-rw-r--r-- 1 graysky users 2776 Jul 14 08:52 of_mdio.ko.gz

Dunno what to say about suspend. There is actually an Arch flyspray[1] about this so others are having problems without the ck1 patchset I think.

1. https://bugs.archlinux.org/task/36107

xzy3186 commented on 2013-07-15 15:17

Hi graysky,

Thanks for your reply. Yes, the full name of the module is called of_i2c. I am using a 32-bit kernel with which I have the following output:
yaourt -Ql linux | grep of_i2c
linux /usr/lib/modules/3.10.1-1-ARCH/kernel/drivers/of/of_i2c.ko.gz

Another problem I am suffering with linux-ck is that I could not suspend my system. If I do so, the system will get frozen (no way to wake it up without turning off the power). However, everything is working properly under official kernel...

graysky commented on 2013-07-14 12:16

Bump to v3.10.1-2

Changelog: Set the CONFIG_INTEL_MEI* from hardcoded to modules following flyspray #36144[1] and the ubuntu bug report/lkml discussion therein. This may have positive effects for users experiencing graphical corruptions/black screens from suspend.
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=e6d9e50667e2439217e15c3a8ca5b53985147c6c

1. https://bugs.archlinux.org/task/36144

graysky commented on 2013-07-14 11:26

@xzy - I just copy the config/config.x86_64 from the ARCH set, and apply the ck1+bfq to it. What is the full name of the module? Is it 'of_i2c'? If I extract the 3.9.9-1-ARCH source and find for it, I do not find it.

% find /scratch/old/usr/lib/modules -name 'of_i2c*'

If I compare 3.10.1-2-ck to 3.10.1-1-ARCH, there are 107 modules that contain the term 'i2c':

% find /scratch/arch/usr/lib/modules -name '*.ko.gz' | grep i2c | wc -l
107

% find /scratch/ck/usr/lib/modules -name '*.ko.gz' | grep i2c | wc -l
107

xzy3186 commented on 2013-07-14 02:08

Hi graysky,

Do you have any idea why the kernel module of of_i2c is never compiled for linux-ck, including 3.10 and previous versions? I checked the official kernel and it is there. Furthermore, CONFIG_OF_I2C is set to 'm' in the config of linux-ck. In my sense, it should be compiled.

graysky commented on 2013-07-13 22:51

Bump to v3.10.1-1

Changelog: https://lkml.org/lkml/2013/7/13/115
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=edc146d99f984adec3b10b65f1a0858b79d0feef

graysky commented on 2013-07-13 09:32

Bump to v3.10-2

Changelog: Disabled the 'enable_haswell_pstate_driver.patch' by default. Users may enable it in the PKGBUILD if they like. Please see discussion thread wherein one Haswell user observed some funky behavior[1].
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=bd706801e7fbbb9aa6081c83c93791d6dd1f5032

1. https://bbs.archlinux.org/viewtopic.php?id=166623

graysky commented on 2013-07-10 18:24

Bump to v3.10-1

LinuxChanges: http://kernelnewbies.org/Linux_3.10
Changelog: LKML is down
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=c7b3308a14a380fdc7f328d6a57407193403f027

Please report any problems, particularly if the system hangs on a reboot or shutdown (this is previously identified in the 3.9 version).

graysky commented on 2013-07-10 11:49

Cool. I didn't know about it; nothing othe blog and I haven't been stalking the FTP site in a while. I will update and post shortly.

Raziel23 commented on 2013-07-10 09:12

There already is ck patchset:
http://ck.kolivas.org/patches/3.0/3.10/3.10-ck1/

P.S. I didn't flag package out of date :)

graysky commented on 2013-07-10 08:51

Why out-of-date? There is no ck patchset yet.

graysky commented on 2013-07-05 07:56

BFQ is ready for 3.10 series. For those of you [misc] who like to see or build/run the pre-release of the next linux-ck as we wait for CK to patch into the new tree, I am hosting the incomplete (no ck1) stuff here:

http://repo-ck.com/PKG_source/linux-ck-next/linux-ck-3.10-0.src.tar.gz

graysky commented on 2013-07-04 09:50

Bump to v3.9.9-1

Changelog: https://lkml.org/lkml/2013/7/3/344
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=b99ecbb1d9ea91219c178f2012031b5c1fde8741

graysky commented on 2013-07-02 00:40

@misc - Did you open a bug with upstream by chance? I saw you logged one against the ARCH package[1].

Little worry about this affecting 3.10-1-ck since it takes CK ~1-1/2 weeks to release a BFS once a new major kernel is released (I came up with this number based on the last six bfs/ck-1 releases). Also, see the his blog wherein he talks about his travel plans extending the usual time out further[2].

3.9 8 days
29-Apr-2013
07-May-2013

3.8 13 days
19-Feb-2013
04-Mar-2013

3.7 4 days
11-Dec-2012
15-Dec-2012

3.6 12 days
01-Oct-2012
13-Oct-2012

3.5 10 days
21-Jul-2012
31-Jul-2012

3.4 12 days
21-May-2012
02-Jun-2012

1. https://bugs.archlinux.org/task/35991
2. http://ck-hack.blogspot.com/2013/07/extra-delays-for-linux-310.html

misc commented on 2013-07-01 23:55

3.10 has a regression that renders Logitech unified receivers useless, ie. my mouse. I've reapplied this workaround/patch

https://patchwork.kernel.org/patch/1562431/raw/

which was removed for 3.10 here

https://git.kernel.org/cgit/linux/kernel/git/jikos/hid.git/commit/?id=8af6c08830b1ae114d1a8b548b1f8b056e068887

to get it working again.

graysky commented on 2013-07-01 22:55

Bump to v3.9.8-2

Changelog: Added Haswell support for intel_pstates. This is the only difference between the -1 release so I will not rebuild the repo packages as only the ck-haswell set is affected.
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=6b0a35a54df31f40f2181e233253abe3b7873a6d

graysky commented on 2013-07-01 19:08

@misc - Cool. I will at the next version bump.

misc commented on 2013-07-01 18:11

My Haswell has been running with your patch since the 21th now — no crashes, at least none that I'd trace to it (one or two with btrfs, two unrelated ones at shutdown). Should be fine to include it.

graysky commented on 2013-06-27 20:28

Bump to v3.9.8-1

Changelog: https://lkml.org/lkml/2013/6/27/535
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=d8d1929119b64ee3a64062fa789a91adb79a2e9f

misc commented on 2013-06-21 21:05

dkms-alx works indeed with that autoinstall patch removed. Thanks!

kyak commented on 2013-06-21 18:53

I'm using linux-ck with dkms-alx just fine. Just had to remove that stupid patch that had been introduced to dkms-alx package recently (see my comments on dkms-alx page).

graysky commented on 2013-06-21 18:50

Why can't dkms-alx work with linux-ck? So long as you have the headers package you should be able to build, no?

misc commented on 2013-06-21 12:15

Patch appears to work fine. Didn't test it extensively (just played a video) since I can't get [dkms-]alx to work, which leaves me without ether/internet in -ck. Looks like I'm stuck with the vanilla kernel until at least 3.10 :(

graysky commented on 2013-06-20 20:13

Bump to v3.9.7-1

Changelog: https://lkml.org/lkml/2013/6/20/579
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=77d7f46d84b9853d4aa93d08332d761bacd76ffb

graysky commented on 2013-06-19 22:36

@misc - Try this patch and lemme know[1]. It is based on the CPUIDs in turbostat.c from Linus' github[2].

1. https://gist.github.com/graysky2/5818740
2. https://github.com/torvalds/linux/blob/master/tools/power/x86/turbostat/turbostat.c#L1459

graysky commented on 2013-06-19 20:30

@misc - Seems like upstream would be eager to add this support to linux-next. If it's the same for Haswell as it is for ivy, it is a one-line patch as you pointed out. As of 19-Jun-2013, it doesn't look like there is support for Haswell:

https://github.com/torvalds/linux/blob/master/drivers/cpufreq/intel_pstate.c#L524

misc commented on 2013-06-19 15:29

So no love/pstates for my Haswell yet. :(

Guess I might just try and find its CPU ID to add it myself, though.

graysky commented on 2013-06-16 14:33

Bump to v3.9.6-2

Changelog: Updates to BFQ. See, https://groups.google.com/forum/?fromgroups=#!topic/bfq-iosched/BcT3HBmQO5M
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=4cf8c1fa57906ce2b308e3f8a50bb603d645cc4a

graysky commented on 2013-06-16 14:33

@misc - Yes, because in the 3.9 tree, only Sandy is defaulted to use pstates. The patch adds Ivy so both Sandy and Ivy are using pstates. Without the patch, only Sandy gets pstates.

misc commented on 2013-06-16 13:50

That pstate_driver patch states that it'll become standard also for Sandy and Haswell, but the PKGBUILD only names Ivy — ?

graysky commented on 2013-06-16 13:05

No idea andre. I don't see this on my box. I suggest you post to CK's blog1 or email him directly.

1. http://ck-hack.blogspot.com/

andre.vmatos commented on 2013-06-14 04:30

Hi. I noted a issue with CK. Switching to another VT or logout in current VT (spawning login process in VT) is too slow, sometimes takes up to 3s. Arch Kernel doesn't has this issue. Any hint? Maybe something with cgroups?

graysky commented on 2013-06-13 21:51

Bump to v3.9.6-1

Changelog: https://lkml.org/lkml/2013/6/13/388
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=b852da3ce14e805897f1947bff147acb96f8c1e1

graysky commented on 2013-06-09 11:32

Bump to v3.9.5-2

Changelog: Sync'ed PKGBUILD with that of official ARCH. Functionally, there is no difference between 3.9.5-1 and 3.9.5-2; there is no need for users of 3.9.5-1 to update to 3.9.5-2.
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=883660d4b2487bfe4c67cd9c924bd0ba8c6b4ab6

graysky commented on 2013-06-08 00:26

Bump to v3.9.5-1

Changelog: https://lkml.org/lkml/2013/6/7/707
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=087c852bbfb9a56679e1ebd6bc509a586ad13551

Anonymous comment on 2013-06-01 13:41

@mantiz thanks! :)

mantiz commented on 2013-05-30 15:07

@cinch: You can tell makepkg to cache the sources by setting "SRCDEST" in /etc/makepkg.conf

Anonymous comment on 2013-05-30 14:34

when there's an update: is there a way to tell the build script to not download the linux source again? it feels like a hog on bandwidth.

graysky commented on 2013-05-26 16:32

Bump to v3.9.4-2

Changelog: Arch upstream enabled several VFIO config options; this release syncs the config files.
Arch changelog: https://projects.archlinux.org/svntogit/packages.git/commit/trunk?h=packages/linux&id=610466be4620cf634766b706e8ac6864fbfbd235
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=a761fc8bc087727bf1dd9455ca1a99eb26935554

dkaylor commented on 2013-05-26 10:26

OK, thanks guys!

graysky commented on 2013-05-26 09:58

OK... I am now mirroring the patchset; you can directly modify the PKGBUILD like this or simply download it and manually move it into the build directory.

1) Download AUR source and extract
2) cd linux-ck
3) sed -i 's/^\"http:\/\/ck.kolivas.org\/patches\/3.0\/3.9/"http:\/\/repo-ck.com\/source\/ck1/' PKGBUILD

http://repo-ck.com/source/ck1/patch-3.9-ck1.bz2

SirWuffleton commented on 2013-05-26 09:36

@dkaylor Just finished a compile on one of my boxes and still had it around. I've got a temporary mirror here: http://wuffleton.com/tmp/patch-3.9-ck1.bz2

I'll leave it up until either ck.kolivas.org comes back up or graysky hosts a mirror on repo-ck.

dkaylor commented on 2013-05-26 07:56

http://ck.kolivas.org/ has done a disappearing act, again. Anyone have patch-3.9-ck1.bz2 laying around?

clayman commented on 2013-05-25 21:46

@graysky Thanks, I'll check out both options.

graysky commented on 2013-05-24 23:19

Bump to v3.9.4-1

Changelog: https://lkml.org/lkml/2013/5/24/376
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=8b3571c28481820006370e5d0d5d3e71c36dddc5

graysky commented on 2013-05-24 23:13

@Nierro - I have no experience with that error but will look at your thread.

Nierro commented on 2013-05-24 10:20

@gravsky: please take a look here https://bbs.archlinux.org/viewtopic.php?id=163796 . System sometimes hangs during shutdown with "disabling #irq16" message (after normal systemd shutdown messages)

graysky commented on 2013-05-23 21:56

@clayman - You should post on the bbs. I believe your interest is identifying a method to have the pstate driver honor niced processes? You might have better luck posting on the google+ thread: https://plus.google.com/117091380454742934025/posts/2vEekAsG2QT?cfem=1

@hepha - Are you reporting that your machine is freezing at shutdown like in the flyspray you posted?

hepha commented on 2013-05-23 08:26

All DM devices detached.
Storage is finalized.

FS#35237 : [linux] linux-3.9.2-1-ARCH Failed to poweroff/halt - https://bugs.archlinux.org/task/35237

hepha commented on 2013-05-23 07:53

All DM devices detached.
Storage is finalized.

FS#35237 : [linux] linux-3.9.2-1-ARCH Failed to poweroff/halt - https://bugs.archlinux.org/task/35237

clayman commented on 2013-05-22 17:58

@Scimmia Yes, they are over 70 when without the ignore_nice_load flag set. (I was running comfortably in 50 - 60 with it.) I run BOINC on my notebook at work and the fan noise is annoying.

Scimmia commented on 2013-05-22 10:36

fry your CPU? Have you even checked the temps yet?

clayman commented on 2013-05-22 06:35

@graysky -- in the meantime, is there *any* way to mimic the "old" behavior where the governor would ignore niced processes? I'd like my CPU not to fry if at all possible. Thanks very much for the info!

clayman commented on 2013-05-22 04:48

@graysky -- Right, thanks a lot. I'll look into it. :)

graysky commented on 2013-05-21 20:08

@sphere - How are you reading the operating frequencies? My understanding is that a util like i7z in [community] accurately reports frequencies for sandy/ivybridge processors but older utils do not. What is the output of:

`sudo i7z`

It could be related to the pstate driver. This is the future of Linux power management and is included by upstream for sandybridge CPUs in 3.9 and has been slated (I think) for inclusion for ivybridge CPUs in 3.10 but this patch has been applied to linux-ck by default in 3.9.3. If you are experiencing problems with it using the ivybridge processor you can disable the patch in the PKGBUILD (just delete the 'y' in the line starting with _pstates_ivy=y).

You can test to see if you are using pstates with this command:

`cat /sys/devices/system/cpu/cpu0/cpufreq/scaling_driver`

If it returns a value of 'intel_pstate' then you are using them.

petterk commented on 2013-05-21 18:21

After this update, my cpu is reportedly running at 3.1ghz even with powersave governor (same with performance, up to 3,4ghz), even tho the stock freq is 2.4ghz (even says so on a sticker on the front of the laptop).

Previously it was running at around 1200mhz with ondemand and 2.4ghz with performance.

Regards

petterk commented on 2013-05-21 17:29

btw using ck-ivybridge on my i7-3630qm

petterk commented on 2013-05-21 17:29

After the update to 3.9.3-1, my cpu is running at 3.1ghz instead of the stock 2.4ghz.

Also, previously I had it running on on-demand when on battery, and Performance when on AC-adapter. Isnt the cpu using governors anymore?

If thats the case, how can I set up a similar setup with LMT?

Regards

graysky commented on 2013-05-21 07:50

Not a stupid question. Both Sandy and Ivybridge CPUs get a more sophisticated method of managing frequency called pstates. This is true for linux-ck and linux (Arch standard). You aren't using the ondemand any more. Search for pstates in google and on the forums for more.

clayman commented on 2013-05-21 06:58

Hi, probably a stupid question but hey -- with 3.9.3 I've been missing the /sys/devices/system/cpu/cpufreq/ondemand/ignore_nice_load setting. I'm running BOINC and this thing keeps my CPU temps low. Where did it go?

I use standard config in this regard.

mantiz commented on 2013-05-20 19:41

@graysky: I know that. ;-) And I already compiled the kernel with the PKGBUILD and patches provided here. But as you see in my last comment, the ck-patch takes care that the options that I think I need can not be enabled at the same time with BFS. As soon as I have some more time that I can spend on this I will contact Con Kolivas why these options are disabled and what the problems are in detail. I don't wanna harm my system or even worse my data. ;-)
After that I will give this another try. But to be honest, I don't want a kernel that I have to compile on my own with each update.
And since the only reason for me using the ck-kernel are large file copies because with the normal cfq-scheduler they cause many freezes on my machine I will use the "normal" kernel from the core-repo until then.
Right now my priority is lxc and large file copies are not that often. ;-)

But I am really thankful for your comments, helped me a lot otherwise I would still be trying to get the ck-kernel working and neglect my work that I really have to do now before anything else. :)

graysky commented on 2013-05-20 19:17

@mantiz - Nothing is stopping you from compiling the arch kernel from ABS and enabling whatever patches you like in this PKGBUILD. Or by that same logic, replacing the config/config.x86_64 in this pkg with the stock arch ones, and simply commenting out the line that patches the code with the CK patchset.

mantiz commented on 2013-05-20 16:11

@graysky: I unsuccessfully built this package on my own. Well, the build was successful and I was able to boot this kernel but the required features weren't there. After looking at the ck-patch it is obvious that this couldn't work because these settings are modified by "depends on !SCHED_BFS". So, I think you were right that these settings are conflicting with BFS.

I will stick with the core-repo-kernel for now and see if I still need to use the bfq-scheduler or if I can live without it.

And I will contact "him" in order to ask if there is any possibility to support these options or any workaround for this.

I will post the answer here as soon as I have the time to write this mail and get an answer.

Thanks for your help so far.

graysky commented on 2013-05-20 13:55

@mantiz - Thanks for the link. I still recommend that you ask on CK's blog. He knows the code better than I :p

mantiz commented on 2013-05-20 13:10

Ok, I don't think that user sessions are involved ("User namespace" is missing in the core-repo-kernel, too) but I really don't know systemd and cgroups very well, so I don't know which features are related to each other. I will try to add these options manually and see if this works.
Thanks for the info, I didn't know that BFS has any problems with cgroups although it is mentioned in the wiki (my fault).

You can get informations about lxc at https://wiki.archlinux.org/index.php/Linux_Containers if you are interested.
It is also referred to as 'chroot on steriods'. ;-)

graysky commented on 2013-05-20 12:49

Not sure what an archlinux-lxc-container is but if it is related to user sessions which depend on the cgroups you mentioned, they and BFS are mutually exclusive. In other words, BFS can't have them to the best of my knowledge. You might try posting this question to CK's blog to get an official answer from him.

http://ck-hack.blogspot.com/2013/05/bfs-0430-ck1-for-linux-39x.html

mantiz commented on 2013-05-20 12:45

I am having problems with starting an archlinux-lxc-container (with systemd) by using the ck-kernel(s). I am using linux-ck-nehalem and linux-ck-ivybridge on my systems.
It turned out that the kernel from the core-repo works fine and I think the problem is that the ck-kernels are missing "CONFIG_CGROUP_CPUACCT" and "CONFIG_CGROUP_SCHED" which are enabled by the core-repo-kernel.

Would it be possible to enable these options with the ck-kernels?

I would really appriciate this. ;-)

graysky commented on 2013-05-20 00:41

Bump to v3.9.3-1

Changelog: https://lkml.org/lkml/2013/5/19/96
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=c260ccda764b74d1a5e00c147a998ad8c720263b

gmenguez commented on 2013-05-19 14:52

Ok, thank you, I will do that.

graysky commented on 2013-05-19 13:02

Suggest that you open a thread in the forum. No idea about your specific problem. If you can log in via ssh you should have full access to logs which you can post in your thread.

gmenguez commented on 2013-05-19 11:41

Hi,

I'm having an strange issue in 3.9, I can't login in a tty (I don't start in graphical mode by default), the login cursor doesn't appear or blinks and keyboard seems not to be working (I can't switch between ttys). I can login through ssh and shutdown the system. I have compiled with 300 Hz instead of 1 kHz tick rate, but the same happens. Is something like that happening to anybody? I've read about hangs in laptops here...

graysky commented on 2013-05-18 07:35

...think I am changing my mind about enabling the ivybridge pstates patch by default; I have read many positive things about folks using this for >4 weeks now with no ill-effects. Sunday or Monday when 3.9.3 gets released, I will have it enabled by default for repo users too.

graysky commented on 2013-05-17 20:23

I misunderstood your question. There is no need to rename the package. Just enable the corresponding options (including the nconfig) and select your CPU type for the nconfig menu. You may select either the 3rd gen i7 option, or if building on the same machine, the native option.

andre.vmatos commented on 2013-05-17 20:17

Yes, I know. I just want to know how to make a package (with this patch enabled) exactly like ivy-bridge packages in your repo. I need to edit pkgname and set ivy-bridge in nconfig, or there's something ready to make it?

graysky commented on 2013-05-17 20:13

This is just a patch. Edit the PKGBUILD and set the variable in the first few commented lines to enable it. Instructions and info are contained in the same section as well. Again, note that I am not enabling this be default, so repo users will not have this enabled.

andre.vmatos commented on 2013-05-17 20:07

Graysky, how to build this package with ivy-bridge optimization, maybe with compatible pkgname with your repos? May I change CPU in nconfig and pkgname in PKGBUILD, or there's a more standard way?

graysky commented on 2013-05-17 20:02

Rebump of v3.9.2-2

Notes: I just updated v3.9.2-2 with an optional feature to enable the new Intel Pstate driver for Ivybridge CPUs. See the g+ thread[1] which is containing in this bbs thread[2] for the source of this patch and a good discussion of what it is and why you may want it. Note that I am not bumping the pkgver since I have this disabled by default and I do not want to rebuild all the repo packages to keep the AUR in sync with the repo. Again, this is 100 % optional.

1. https://plus.google.com/117091380454742934025/posts/2vEekAsG2QT
2. https://bbs.archlinux.org/viewtopic.php?id=163524

Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=5a8d0a9fd91a8c8fd348f09fb5a9590f609129fc

graysky commented on 2013-05-17 19:50

I just updated v3.9.2-2 with an optional feature to enable the new Intel Pstate driver for Ivybridge CPUs. See the g+ thread which is containing in this bbs thread[1] for the source of this patch and a good discussion of what it is and why you may want it.

Note that I am not bumping the pkgver since I have this disabled by default and I do not want to rebuild all the repo packages to keep the AUR in sync with the repo.

1. https://bbs.archlinux.org/viewtopic.php?id=163524

graysky commented on 2013-05-12 20:05

Bump to v3.9.2-2

LinuxChanges: http://kernelnewbies.org/LinuxChanges_3.9
Changelog: https://lkml.org/lkml/2013/5/11/136
Commit:http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=ad38592612a42503818f5c5cfa8c267a5b8872df
Notes: I tracked down the cause of the freezes to the tick rate in the config; it has been changed for to 1k Hz for the 3.9 series. This is been independently confirmed now by some users of linux-pf (also 1k Hz) and by CK as well.

I will be placing 3.8.13-1-ck along with all related package into the Archive Section on repo-ck should users want access to the now deprecated 3.8 series: http://repo-ck.com/archive.htm

The 3.9.2-2-ck build should be finished around midnight GMT and the repo populated with fresh 3.9.2-2-ck and related packages.

Enjoy!

graysky commented on 2013-05-12 18:23

Bump to v3.9.2-2

LinuxChanges: http://kernelnewbies.org/LinuxChanges_3.9
Changelog: https://lkml.org/lkml/2013/5/11/136
Commit:
Notes: I tracked down the cause of the freezes to the tick rate in the config; it has been changed for to 1k Hz for the 3.9 series. This is been independently confirmed now by some users of linux-pf (also 1k Hz) and by CK as well.

graysky commented on 2013-05-12 12:04

Bump to v3.8.13-1

Changelog: https://lkml.org/lkml/2013/5/11/135
BFQ gets updated to 3.8.0-v6r1
Commit:http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=03b4d81fab0110e9b73e7d86c314f5133a7f2f5c

Notes: I am still trying to figure out why the freeze-on-reboot on my laptop (and now on my workstation) using the 3.9.2-1-ck code I posted about here on 2013-05-11 21:05.

Det commented on 2013-05-12 11:09

v6r1 and .13 are up.

graysky commented on 2013-05-11 21:05

EDIT: Upstream has just pushed 3.9.2 so I wanted to update everything accordingly. I also added some links to pre-compiled generic packages for 3.9.2-1-ck for those wishing to test but not wishing to compile. Note that these are kernel packages and headers only; no nvidia, vbox, broadcom util this issues is sorted out.

Precompiled packages: http://repo-ck.com/bench.htm#Linux-ck_Next_Package
Source to compile it yourself: http://repo-ck.com/PKG_source/linux-ck-next/linux-ck-3.9.2-1.src.tar.gz

Original post: Paolo et al have released BFQ for the 3.9 tree today. I have update the PKGBUILD which is now 100 % from a code perspective. I am however reluctant to push it to the AUR and to update all other -ck packages until an issue that affects my laptop gets resolved. What happens? When I reboot or shutdown, the system hangs after everything goes down. CK has edited his blog announcement acknowledging the behavior[1]. To date, I do not believe this issue is resolved.

How to reproduce the error: on my system I only need to reboot or shutdown. This can happen either from xfce4's menu or from the shell, i.e. `sudo reboot`. Here is a pic of what is displayed on the hang:

http://s19.postimg.org/riaulcbcj/freeze.jpg

Note - 3.9.2-1-ARCH throws the same text errors[2], but the system power cycles. That is in contrast to 3.9.2-1-ck which also complains about the connection being refused, (likely unrelated to the actual freeze), but then hangs until I power if off.

1. http://ck-hack.blogspot.com/2013/05/bfs-0430-ck1-for-linux-39x.html
2. https://bbs.archlinux.org/viewtopic.php?id=162905

graysky commented on 2013-05-11 20:41

EDIT: Upstream has just pushed 3.9.2 so I wanted to update everything accordingly. I also added some links to pre-compiled generic packages for 3.9.2-1-ck for those wishing to test but not wishing to compile. Note that these are kernel packages and headers only; no nvidia, vbox, broadcom util this issues is sorted out. Also note that I am building the packages as I type this, so expect them there in about 1/2 h if all goes well with my upstream bandwidth to godaddy (21:15 GMT).

Precompiled packages: http://repo-ck.com/bench.htm#Linux-ck_Next_Package
Source to compile it yourself: http://repo-ck.com/PKG_source/linux-ck-next/linux-ck-3.9.2-1.src.tar.gz

Original post:
OK, Paolo et al have released BFQ for the 3.9 tree today. I have update the PKGBUILD which is now 100 % from a code perspective. I am however reluctant to push it to the AUR and to update all other -ck packages until an issue that seems to be affect my laptop is resolved. What happens? When I reboot or shutdown, the system hangs after everything goes down. CK has edited his blog announcement acknowledging the behavior[1]. To date, I do not believe this issue is resolved.

If you're willing to test, find the 'official' 3.9.2-1-ck source on repo-ck here: http://repo-ck.com/PKG_source/linux-ck-next/linux-ck-3.9.2-1.src.tar.gz

How to reproduce the error: on my system I only need to reboot or shutdown. This can happen either from xfce4's menu or from the shell, i.e. `sudo reboot`. Here is a pic of what is displayed on the hang:

http://s19.postimg.org/b60svlx0z/freeze.jpg

Note - 3.9.2-1-ARCH throws the same text errors[2], but the system power cycles. That is in contrast to 3.9.2-1-ck which also complains about the connection being refused, (likely unrelated to the actual freeze), but then hangs until I power if off.

1. http://ck-hack.blogspot.com/2013/05/bfs-0430-ck1-for-linux-39x.html
2. https://bbs.archlinux.org/viewtopic.php?id=162905

graysky commented on 2013-05-11 19:52

OK, Paolo et al have released BFQ for the 3.9 tree today. I have update the PKGBUILD which is now 100 % from a code perspective. I am however reluctant to push it to the AUR and to update all other -ck packages until an issue that seems to be affect my laptop is resolved. What happens? When I reboot or shutdown, the system hangs after everything goes down. CK has edited his blog announcement acknowledging the behavior.[1] To date, I do not believe this issue is resolved.

If you're willing to test, find the 'official' 3.9.1-1-ck source on repo-ck here: http://repo-ck.com/PKG_source/linux-ck-next/linux-ck-3.9.1-1.src.tar.gz

How to reproduce the error: on my system I only need to reboot or shutdown. This can happen either from xfce4's menu or from the shell, i.e. `sudo reboot`. Here is a pic of what is displayed on the hang:

http://s19.postimg.org/b60svlx0z/freeze.jpg

Note - 3.9.1-1-ARCH throws the same text errors[2], but the system power cycles. That is in contrast to 3.9.1-1-ck which also complains about the connection being refused, (likely unrelated to the actual freeze), but then hangs until I power if off.

1. http://ck-hack.blogspot.com/2013/05/bfs-0430-ck1-for-linux-39x.html
2. https://bbs.archlinux.org/viewtopic.php?id=162905

andre.vmatos commented on 2013-05-10 19:45

3.9.1-0 compiled fine and running well in my ivy bridge i5.

graysky commented on 2013-05-09 20:54

Yep.

Anonymous comment on 2013-05-09 19:03

Yeah it compiles every time there is an update using AUR right now. How do I switch to the repo? just ininstall this package and reinstall through the repo?

graysky commented on 2013-05-09 17:57

@Sharopo - If "updating" means you are compiling it then, yes. Depending on the speed of your CPU, 20-30 min would be about right. If you do not wish to compile yourself, use my unofficial repo: http://repo-ck.com

Anonymous comment on 2013-05-09 16:11

Is updating always expected to take 20-30 minutes? It's a pain to have to sit there and wait for it to finish and be running my cpu maxed out that entire time.

graysky commented on 2013-05-08 20:33

@andre - My bad... thought I pulled the offending outdated patch from this (the alsa fix). I just reuploaded 3.9.1-0 overwriting the old one. Please download it again and it should build.

andre.vmatos commented on 2013-05-08 20:11

I'm trying to compile 3.9.1, but it's failing with:
...
patching file sound/usb/usx2y/usx2yhwdeppcm.c
patching file include/sound/emu10k1.h
Reversed (or previously applied) patch detected! Skipping patch.
1 out of 1 hunk ignored -- saving rejects to file include/sound/emu10k1.h.rej
patching file sound/pci/emu10k1/emu10k1_main.c
Reversed (or previously applied) patch detected! Skipping patch.
4 out of 4 hunks ignored -- saving rejects to file sound/pci/emu10k1/emu10k1_main.c.rej

Any ideas?

graysky commented on 2013-05-08 20:07

@Andre - I see. I can tell you that the nvidia-304xx and virtualbox modules do not show anything bad in my dmesg; both of these have not been compiled against 3.8.12. I did have to rebuild the broadcom-wl-ck module however.

andre.vmatos commented on 2013-05-08 20:01

Hmm, IMHO, I think he means that 3rd-part modules (e.g. virtualbox) compiled for 3.8 series are showing errors, often symbol errors, upon insertion, avoiding it to load properly in 3.8.12, even when a compilation for 3.8 series are supposed to be binary-compatible with all 3.8 series (symbols and binary compatibility should change only in major (3.x) upgrades, requiring a recompilation or even patching).

graysky commented on 2013-05-08 20:00

Meant to post this link to the pre-release of 3.9.1-ck which is nearly 100 %. It is just missing the official yet-to-be-released BFQ which Paolo tells me should not be different from the current release which patches into 3.9 just fine.

http://repo-ck.com/PKG_source/linux-ck-next/linux-ck-3.9.1-0.src.tar.gz

graysky commented on 2013-05-08 19:55

Thanks for that link, andre. I guess I do not know what tpowa meant by his comment, "module breaker."

https://mailman.archlinux.org/pipermail/arch-dev-public/2013-May/024896.html

andre.vmatos commented on 2013-05-08 19:46

This warnings isn't related with binary modules in 3.8.12, but with mkinitcpio 0.14: https://mailman.archlinux.org/pipermail/arch-dev-public/2013-May/024864.html

graysky commented on 2013-05-08 19:27

Bump to v3.8.12-1

Changelog: https://lkml.org/lkml/2013/5/8/12
Commit:

Notes: Tpowa noted that the 3.8.12 release is a "binary module breaker" which I assume explains the mkinitcpio output when it built my fallback image:
==> WARNING: Possibly missing firmware for module: aic94xx
==> WARNING: Possibly missing firmware for module: bfa

Dunno the scope of this and how it will affect users.

graysky commented on 2013-05-07 19:14

OK. The ck1 patchset for 3.9 has gone gold. We are just waiting for Paolo to do the same with the BFQ set and are on track for a release any time now.

For those who can't wait: http://repo-ck.com/PKG_source/linux-ck-next/linux-ck-3.9-0.src.tar.gz

Det commented on 2013-05-04 20:59

@Kalrish, was it fair for kernel.org to be attacked :)?

graysky commented on 2013-05-03 19:30

CK has released an rc quality patch featuring v0.430 of BFS[1]. I am running it here and wanted to give folks the chance to tinker with it as well:

http://repo-ck.com/PKG_source/linux-ck-next/linux-ck-3.9-0.src.tar.gz

Notes:
1) This is a pre-release so only download and compile the source if you accept that fact.
2) There is no BFQ yet, since Paolo has yet to release an updated patch. The old version of BFQ actually patches into the 3.9 tree without errors, but I do not want to run it until I hear back from him as to whether it is safe.
3) If you run this and find problems or you run this and find that it works just fine, please consider giving CK some feedback on the his blog I linked.

1. http://ck-hack.blogspot.com/2013/03/bfs-0428-for-linux-38x.html

graysky commented on 2013-05-03 19:28

CK has released an rc quality patch featuring v0.430 of BFS. Running it here and wanted to give folks the chance to tinker with it as well:

http://repo-ck.com/PKG_source/linux-ck-next/linux-ck-3.9-0.src.tar.gz

1) There is no BFQ since Paolo has yet to release an updated patch.
2) The old version of BFQ actually patches into the 3.9 tree without errors, but I do not want to run it until I hear back from him as to whether it is safe.

Kalrish commented on 2013-05-03 19:07

@graysky Thanks a lot, it worked like a charm -as always!-.
@Det In my opinion, including a temp-host URL in a PKGBUILD would not be very... fair? I think it's up to the user to download the patch from an unofficial (yet useful) server. (In any case, http://ck.kolivas.org is up again.)

Det commented on 2013-05-03 00:30

Hmm, but why not just include that in the PKGBUILD?

You wouldn't even have to remove the previous one, if you'd just comment it.

graysky commented on 2013-05-02 21:31

Bump to v3.8.11-1

Changelog: https://lkml.org/lkml/2013/5/1/193
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=694e770f5bb922296a3a09028b715855e54f00e7

graysky commented on 2013-05-02 18:22

@misc - http://repo-ck.com/PKG_source/linux-ck-next/linux-ck-3.9-1.src.tar.gz
@Kalrish - Yeah, ck has been experiencing some addition problems with his VPS. I am hosting. Run this sed line on the PKGBUILD and you are good to go:

sed 's/^\"http:\/\/ck.kolivas.org\/patches\/3.0\/3.8/"http:\/\/repo-ck.com\/source\/mirror/' PKGBUILD

Kalrish commented on 2013-05-02 16:28

http://ck.kolivas.org/patches/3.0/3.8/3.8-ck1/patch-3.8-ck1.bz2 is down at this moment

misc commented on 2013-05-02 15:09

3.9-0 would be much appreciated.

graysky commented on 2013-05-02 14:20

Bump to v3.8.11-1

Changelog:
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=694e770f5bb922296a3a09028b715855e54f00e7

graysky commented on 2013-05-02 13:29

I know that 3.8.11 was released today but there has been some activity in the ARCH upstream; I want to make sure that the config files are right before pushing something.

graysky commented on 2013-04-30 10:27

Plz do not flag out of date until all needed components are updated.

From the wiki: https://wiki.archlinux.org/index.php/Linux-ck#Release_Cycle

"Linux-ck roughly follows the release cycle of the official ARCH kernel. The following are requirements for its release:
Upstream code
CK's Patchset
BFQ Patchset
ARCH config/config.x86_64 sets for major version jumps only"

Anonymous comment on 2013-04-30 09:42

Don't think it's out-of-date... CK didn't release ck-patchset for Linux 3.9 yet...

graysky commented on 2013-04-27 11:30

Bump to v3.8.10-1

Changelog: https://lkml.org/lkml/2013/4/26/474
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=2ee46d8d473141b30b5bf8918cfbb7ad46c9914e

graysky commented on 2013-04-26 20:12

Bump to v3.8.9-1

Changelog: https://lkml.org/lkml/2013/4/26/2
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=7264ab8cd521b90d9f68fb72c06de76ce36928a8

graysky commented on 2013-04-26 07:59

Ok. I know 3.8.9 is out but I will not have time to test it properly before releasing right now. Expect it later today though.

graysky commented on 2013-04-25 08:09

FYI - I am aware of the 3.8.8-2-ARCH bump yesterday, but the 3.8.9 patch is due today so I'd rather not bump at this late hour.

https://projects.archlinux.org/svntogit/packages.git/commit/trunk?h=packages/linux&id=04fca80d591b2d40afb3cfeba701214307978dde

graysky commented on 2013-04-18 22:21

Bump to v3.8.8-1

Changelog: https://lkml.org/lkml/2013/4/17/53
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=f179be73161dde40eb195a90e1722aadce600d70

graysky commented on 2013-04-18 08:52

Someone flagged it by mistake or out of ignorance. It is not out of date.

lano1106 commented on 2013-04-18 03:40

Is it really already out of date on the same day it is released or just a mistake?

graysky commented on 2013-04-17 17:11

Bump to v3.8.8-1

Changelog: https://lkml.org/lkml/2013/4/17/53
Commit: aur-mirror.git is down

Det commented on 2013-04-16 02:03

You can actually do that whole global pkgname with an .AURINFO: https://mailman.archlinux.org/pipermail/aur-dev/2013-March/002428.html

Never knew about it either back when they released 2.1.

flamusdiu commented on 2013-04-14 16:43

@graysky - of course. I was already using customizepkg for something else so used it for this as well. Just having and issue between customizepkg and aura but that's for another day. :P

flamusdiu commented on 2013-04-14 16:41

@graysky - works now. Using customizepkg (in AUR) with a one liner:

replace#global#http:\/\/ck.kolivas.org\/patches\/3.0\/3.8/#http:\/\/repo-ck.com\/source\/ck1_temp/

=)

graysky commented on 2013-04-14 16:34

Thanks for the report. Guess I'd just recommend the sed oneliners I posted but to each his own...

flamusdiu commented on 2013-04-14 16:32

@graysky - works now. Using customizepkg (in AUR) with a one liner:

replace#global#http:\/\/ck.kolivas.org\/patches\/3.0\/3.8/#http:\/\/repo-ck.com\/source\/ck1_temp/

=)

graysky commented on 2013-04-14 12:55

@flamusdiu - My bad... works now.

1) Download AUR source and extract
2) sed -i 's/^\"http:\/\/ck.kolivas.org\/patches\/3.0\/3.8/"http:\/\/repo-ck.com\/source\/ck1_temp/' PKGBUILD
3) makepkg -src

FYI-CK knows his shit is down; his VPS is doing a poor job fixing it.

graysky commented on 2013-04-14 12:55

@flamusdiu - My bad... works now.

1) Download AUR source and extract
2) sed -i 's/^\"http:\/\/ck.kolivas.org\/patches\/3.0\/3.8/"http:\/\/repo-ck.com\/source\/ck1_temp/' PKGBUILD
3) makepkg -src

flamusdiu commented on 2013-04-14 11:35

@graysky - seems that the patch is missing on the repo-ck site now. :'(

graysky commented on 2013-04-13 11:04

@fincan - Yeah, it's down alright.

I am temp hosting the files on repo-ck.com. Run this on the dir with the PKGBUILD to build this until their servers are back up:

sed -i 's/^\"http:\/\/ck.kolivas.org\/patches\/3.0\/3.8/"http:\/\/repo-ck.com\/source\/ck1_temp/' PKGBUILD

fincan commented on 2013-04-13 05:55

http://ck.kolivas.org/ is down? I could not download the patch

graysky commented on 2013-04-12 22:24

Bump to v3.8.7-1

Changelog: https://lkml.org/lkml/2013/4/12/283
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=a6624337b2e39d651574b82e9671740a144d97e0

graysky commented on 2013-04-08 20:07

Bump to v3.8.6-3

Changelog: Several to follow.
*Add required build-time dependencies [bc will be needed in 3.9] (FS#34600).
*Updated config files so as not to require user input as a function of the gcc patch.
*Re-wrote PKGBUILD to use prepare function and opened [https://bugs.archlinux.org/task/34688 FS#34688] to match my changes.
*Added missing keys file in headers package.
*Added "save configuration for later reuse" to match ARCH PKGBUILD.

Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=c1433862d6c7c6f8dc5f0ef05914179a07f95869

graysky commented on 2013-04-05 20:03

Bump to v3.8.6-1

Changelog: https://lkml.org/lkml/2013/4/5/362
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=165f9066879dd0684a1f6377e516f55542f3838d

graysky commented on 2013-04-02 18:44

Yeah, looks like upstream's website is down. I am temp hosting the files on repo-ck.com. Run this on the dir with the PKGBUILD to build this until their servers are back up:

sed -i 's/_bfqpath=.*/_bfqpath=\"http:\/\/repo-ck.com\/source\/bfq_temp\"/' PKGBUILD

MikeDacre commented on 2013-04-02 17:27

I am having the same problem as getsiu with algogroup.unimo.it (Failed connect to www.algogroup.unimo.it:80; Connection timed out), the host apparently does not exist anymore.

getsiu commented on 2013-04-02 17:00

I'm getting an error while building this package:

-> Загрузка 0001-block-cgroups-kconfig-build-bits-for-BFQ-v6-3.8.patch...
% Total % Received % Xferd Average Speed Time Time Time Current
Dload Upload Total Spent Left Speed
0 0 0 0 0 0 0 0 --:--:-- --:--:-- --:--:-- 0curl: (7) Failed connect to www.algogroup.unimo.it:80;

Seems like host "www.algogroup.unimo.it" is down:

% ping www.algogroup.unimo.it -c 4
PING smf-pc.mat.unimo.it (155.185.5.26) 56(84) bytes of data.
From mate-p2-hp.unimo.it (155.185.255.156) icmp_seq=1 Destination Host Unreachable

getsiu commented on 2013-04-02 17:00

I'm getting an error while building this package:

-> Загрузка 0001-block-cgroups-kconfig-build-bits-for-BFQ-v6-3.8.patch...
% Total % Received % Xferd Average Speed Time Time Time Current
Dload Upload Total Spent Left Speed
0 0 0 0 0 0 0 0 --:--:-- --:--:-- --:--:-- 0curl: (7) Failed connect to www.algogroup.unimo.it:80;

Seems like host "www.algogroup.unimo.it" is down:

% ping www.algogroup.unimo.it -c 4
PING smf-pc.mat.unimo.it (155.185.5.26) 56(84) bytes of data.
From mate-p2-hp.unimo.it (155.185.255.156) icmp_seq=1 Destination Host Unreachable

graysky commented on 2013-03-31 10:42

@gaugalos - See my comment from 2012-05-14 21:09.

qpalz commented on 2013-03-31 03:06

@gaougalos
See:
https://aur.archlinux.org/packages/linux-uksm-ck/

gaougalos commented on 2013-03-31 02:39

Can you use uksm patch?

graysky commented on 2013-03-28 22:23

Bump to v3.8.5-1

Changelog: https://lkml.org/lkml/2013/3/28/471
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=b558b6da20c5c4b154fc817a701228101be42eaa

graysky commented on 2013-03-20 22:18

Bump to v3.8.4-1

Changelog: https://lkml.org/lkml/2013/3/20/545
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=f0a06808a05920913cf57e49fef13078b0c4006a

graysky commented on 2013-03-20 22:18

Bump to v3.8.4-1

*Changelog: https://lkml.org/lkml/2013/3/20/545
*Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=f0a06808a05920913cf57e49fef13078b0c4006a

graysky commented on 2013-03-18 20:00

Bump to v3.8.3-2

Changelog: Several i915 patches applied to fix graphic issues.
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=58d7de36e45e8f7411ff8a534699e4012dd90efb

graysky commented on 2013-03-17 19:19

Bump to v3.8.3-2

Changelog: Several i915 patches applied to fix graphic issues.
Commit:

graysky commented on 2013-03-17 18:27

No unless someone flags this out-of-date. I'm on it now. Should be a few (wanna test build and reboot into it myself before I push the to AUR).

Anonymous comment on 2013-03-17 17:41

ah cool ;) not in a hurry. Thanks!

Scimmia commented on 2013-03-17 17:33

Upstream problem, already fixed. Arch just added the patches for 3.8.3-2, give greysky some time.

Anonymous comment on 2013-03-17 17:22

I'm having trouble with my resolution (intel G45) after compiling this. Where should bug reports be searched/posted?

graysky commented on 2013-03-15 20:33

Bump to v3.8.3-1

Changelog: https://lkml.org/lkml/2013/3/14/493
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=0258a41065a93892e8f7d2cd7337a67f63a65a24

graysky commented on 2013-03-14 20:05

3.8.3 has been released but I plan to wait for the ARCH package to see if tpowa et al incorporate any patches/fixes of their own before pushing. Plz flag this when 3.8.1-1-ARCH goes live if you see it before I do.

AdmiralAkber commented on 2013-03-13 04:55

Yes Scimmia, it appears to be a 3.8 problem. I'll take the issue to the appropriate place. My Apologies.

Scimmia commented on 2013-03-13 03:40

You're comparing a 3.7 kernel to a 3.8 kernel. Does it happen with the official Arch 3.8 kernel too? If it does, it's not a -ck problem and doesn't really belong here.

AdmiralAkber commented on 2013-03-13 03:35

I notice over a 50% drop in wireless bit rate using this kernel compiled for 2nd gen i7 i.e. i7-avx. Signal level and link quality are the same.

~~~~~~~~~~
Wifi Card: Intel Centrino 1000 N
Driver: iwlwifi
Laptop: Lenovo Ideapad y560p
Diagnostic tool: iwconfig
~~~~~~~~~~
Symptoms:
> 3.7.10-1-ARCH Kernel:
--> Wireless connects quickly and sits at 150Mbps

> 3.8.2-1 CK Kernel from AUR compiled for 2nd gen i7:
--> Wireless sits around the 15-75Mbps rate, it starts low and builds its way up. Upon reaching ~75Mbps, the wireless bit-rate remains constant. It also takes longer to associate. Signal level and Link Quality are the same.
~~~~~~~~~
Steps to reproduce:
1. Install 3.8.2-1 CK Kernel for i7-avx with BFQ on default.
2. Use iwlwifi with Intel Centrino 1000 N wireless card.
3. Connect to wireless N router.

graysky commented on 2013-03-10 19:52

@Sajmon - See: https://wiki.archlinux.org/index.php/Linux-ck#Package_Defaults

Sajmon commented on 2013-03-10 19:38

How the config file is set up, as opposed to arch kernel?

Sajmon commented on 2013-03-10 19:24

How the config file is set up, as opposed to arch kernel?

graysky commented on 2013-03-04 22:20

Bump to v3.8.2-1

Changelog: https://lkml.org/lkml/2013/3/3/127
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=ed5d6edd2e6d61d97128baf7f45b71bcf8cf4d12

graysky commented on 2013-03-04 21:55

@PG - Sure. Like SW said, any of the menu based methods are fine.

SirWuffleton commented on 2013-03-04 15:52

@PerfectGentleman: It would be perfectly valid to do such a thing on your end. Although I personally find nconfig simpler and more intuitive, you could modify your local PKGBUILD to include any of the *config options if you find one of the X-based tools to be more comfortable.

PerfectGentleman commented on 2013-03-04 12:48

i have KDE. so, is it okay to change "make nconfig" to "make xconfig" in PKGBUILD ?

PerfectGentleman commented on 2013-03-04 12:33

i have KDE. so, is it okay to change "make nconfig" to "make xconfig" in PKGBUILD ?

crondog commented on 2013-03-04 10:02

3.8.2 is now available :)

graysky commented on 2013-03-04 08:49

OK.. I understand now: the path changed. Thank you crondog. Updated but not bumping the pkgver.

For others - Plz don't flag out-of-date if a source path changes; it makes me think I missed something upstream or with ARCH.

graysky commented on 2013-03-04 08:44

@crondog - Look at line 64 in the PKGBUILD I released yesterday :D
@blingblong123 - Why out-of-date? The 77 patches in 3.8.2 are still under review https://lkml.org/lkml/2013/3/1/377

crondog commented on 2013-03-04 01:49

Just to let you know the ck patch is located at http://ck.kolivas.org/patches/3.0/3.8/3.8-ck1/patch-3.8-ck1.bz2

graysky commented on 2013-03-03 19:19

@Misc - BFS is still superior using a non-latency endpoint, my make-based benchmark tested now on two machines:

i7-3770K @ 4.5 GHz (9 threads)
i7-2620M @ 2.7 GHz (5 threads)

Conclusion: BFS v0.428 _clearly_ differentiates itself from CFS in mainline 3.8.1.

Results on machine #1: http://s19.postimage.org/jkil3zx83/3770_K.png
Results on machine #2: http://s19.postimage.org/uy54f77qr/2620_M.png

These analyses are an ANOVA showing you the time it took to compile (y-axis in sec) each of the 9 runs under each kernel (x-axis).

Black dots = time points.
Green diamonds = 95 % confidence intervals.
The circles to the right are comparison circles. If these are separated, the differences in average times are statistically significant. In other words, BFS is still faster than CFS. This is true for both machines tested. Note that I will expand this to a few other machines in the next few days.

Code to benchmark script is 'make_bench' and can be found on my
github: https://github.com/graysky2/bin

The benchmark simply runs `make -jx bzImage` 9 times and logs the total time to build to a log file for analysis. This the same analysis I did in my study of them a few months ago which is available here: http://repo-ck.com/bench/cpu_schedulers_compared.pdf

P.S. Sorry for multiple posts. Cannot edit as you know.

graysky commented on 2013-03-03 19:13

@Misc - BFS is still superior using a non-latency endpoint, my make-based benchmark tested now on two machines:

i7-3770K @ 4.5 GHz (9 threads)
i7-2620M @ 2.7 GHz (5 threads)

Conclusion: BFS v0.428 _clearly_ differentiates itself from CFS in
mainline 3.8.1.
Results on machine #1: http://s19.postimage.org/jkil3zx83/3770_K.png
Results on machine #2: http://s19.postimage.org/uy54f77qr/2620_M.png

Code to benchmark script is 'make_bench' and can be found on my
github: https://github.com/graysky2/bin

The benchmark simply runs `make -jx bzImage` 9 times and logs the total time to build to a log file for analysis. This the same analysis I did in my study of them a few months ago which is available here: http://repo-ck.com/bench/cpu_schedulers_compared.pdf

Note that I will expand this to a few other machines in the next few days.

graysky commented on 2013-03-03 19:12

@Misc - BFS is still superior using a non-latency endpoint, my make-based benchmark tested now on two machines:

i7-3770K @ 4.5 GHz (9 threads)
i7-2620M @ 2.7 GHz (5 threads)

Conclusion: BFS v0.428 _clearly_ differentiates itself from CFS in
mainline 3.8.1.

Code to benchmark script is 'make_bench' and can be found on my
github: https://github.com/graysky2/bin

The benchmark simply runs `make -jx bzImage` 9 times and logs the total time to build to a log file for analysis. This the same analysis I did in my study of them a few months ago which is available here: http://repo-ck.com/bench/cpu_schedulers_compared.pdf

Note that I will expand this to a few other machines in the next few days.

misc commented on 2013-03-03 18:54

Given the improvements to CFS in 3.8, it would be interesting to see how BFS compares to it now.

graysky commented on 2013-03-03 16:40

Bump to v3.8.1-1

LinuxChanges: http://kernelnewbies.org/LinuxChanges_3.8
Changelog: https://lkml.org/lkml/2013/2/28/219
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=96b741a6eb56b2a5c2d9dc89cf28181424c0e1a8

graysky commented on 2013-03-01 20:42

@misc - Nothing was actually wrong with the patch, I mistakenly changed the source in the PKGBUILD to point to the patch for the 3.7 tree, not the 3.8 tree. This has been fixed and works with both arches.

@radioact - No need, I have done this in the -0 release of linux-ck-3.8. See my comment below (2013-02-25 20:45) for the announcement and link. Still no update from CK by the way.

Radioactiveman commented on 2013-03-01 19:28

If you want 3.8.1 you need to disable both BFS patches and the patch for CVE-2013-1763.

misc commented on 2013-03-01 01:34

kernel-37-gcc47-2's hunk #6 fails in arch/x86/Kconfig.cpu for 3.8.1.

MatejLach commented on 2013-03-01 01:04

@graysky - Builds fine here now, many thanks for fixing this.

graysky commented on 2013-02-28 21:02

@PerfectGentleman - Try now.

graysky commented on 2013-02-28 10:26

Ok. Just need to add the following. I will update later today:

[[ -d include/config/dvb ]] && find ...

PerfectGentleman commented on 2013-02-28 03:39

==> Starting package_linux-ck-headers()...
find: `include/config/dvb': No such file or directory
==> ERROR: A failure occurred in package_linux-ck-headers().

graysky commented on 2013-02-27 23:19

Bump to v3.7.10-1

Changelog: lkml is down
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=63b322e023549522cad284171882b919543c0b8c

graysky commented on 2013-02-27 21:24

@Raziel23 - Very helpful, thank you. Updated the PKGBUILD to reflect this new location.

@MatejLach - I updated the PKGBUILD per your feedback using a find statement that shouldn't trip makepkg if nulls are found. I tested it forcing a delete of $srcdir/linux-3.7/include/config/dvb/*.h and it worked fine. Please build with it disabling the dvb shit as you do and report back.

Note that I did not bump the pkgver version.

MatejLach commented on 2013-02-27 19:59

I'm building 3.7.9-2 BTW.

MatejLach commented on 2013-02-27 19:56

Indeed, when I delete this:

mkdir -p "${pkgdir}/usr/src/linux-${_kernver}/include/config/dvb/"
[[ -e include/config/dvb/ ]] && cp include/config/dvb/*.h "${pkgdir}/usr/src/linux-${_kernver}/include/config/dvb/"

It is enough, to compile the kernel without problems.
Were there any changes there compared to previous PKGBUILD?

MatejLach commented on 2013-02-27 19:55

@graysky - I think that there's a bug in the PKGBUILD where it is expected that I have dvb support enabled in the kernel (and therefore there is a file 'include/config/dvb/*.h'), but if I have it disabled, the build (of the *headers package) simply fails, because of that expectation.


I think it's this block in the PKGBUILD:

# add dvb headers for external modules
# in reference to:
# http://bugs.archlinux.org/task/9912
mkdir -p "${pkgdir}/usr/src/linux-${_kernver}/drivers/media/dvb-core"
cp drivers/media/dvb-core/*.h "${pkgdir}/usr/src/linux-${_kernver}/drivers/media/dvb-core/"
# and...
# http://bugs.archlinux.org/task/11194
mkdir -p "${pkgdir}/usr/src/linux-${_kernver}/include/config/dvb/"
[[ -e include/config/dvb/ ]] && cp include/config/dvb/*.h "${pkgdir}/usr/src/linux-${_kernver}/include/config/dvb/"


More precisely this bit:

mkdir -p "${pkgdir}/usr/src/linux-${_kernver}/include/config/dvb/"
[[ -e include/config/dvb/ ]] && cp include/config/dvb/*.h "${pkgdir}/usr/src/linux-${_kernver}/include/config/dvb/"

Can you have a second glance at it?
Thanks,

MatejLach commented on 2013-02-27 19:06

@graysky - I think that there's a bug in the PKGBUILD where it is expected that I have dvb support enabled in the kernel (and therefore there is a file 'include/config/dvb/*.h'), but if I have it disabled, the build (of the *headers package) simply fails, because of that expectation.


I think it's this block in the PKGBUILD:

# add dvb headers for external modules
# in reference to:
# http://bugs.archlinux.org/task/9912
mkdir -p "${pkgdir}/usr/src/linux-${_kernver}/drivers/media/dvb-core"
cp drivers/media/dvb-core/*.h "${pkgdir}/usr/src/linux-${_kernver}/drivers/media/dvb-core/"
# and...
# http://bugs.archlinux.org/task/11194
mkdir -p "${pkgdir}/usr/src/linux-${_kernver}/include/config/dvb/"
[[ -e include/config/dvb/ ]] && cp include/config/dvb/*.h "${pkgdir}/usr/src/linux-${_kernver}/include/config/dvb/"


More precisely this bit:

mkdir -p "${pkgdir}/usr/src/linux-${_kernver}/include/config/dvb/"
[[ -e include/config/dvb/ ]] && cp include/config/dvb/*.h "${pkgdir}/usr/src/linux-${_kernver}/include/config/dvb/"

Can you have a second glance at it?
Thanks,

Raziel23 commented on 2013-02-27 08:14

BFQ move to a new website. The old one will not be updated and in the near future will probably be shut down.
Here is the relevant message on mailing list from Paolo:
https://groups.google.com/forum/?fromgroups=#!topic/bfq-iosched/avSJ0HLbmHg

graysky commented on 2013-02-26 23:32

@MatejLach - You building 3.7.9-2 or 3.8-0? I cannot reproduce your error; builds fine here (both). Are you using any modifications to the config?

MatejLach commented on 2013-02-26 23:00

==> Starting package_linux-ck-headers()...
cp: cannot stat ‘include/config/dvb/*.h’: No such file or directory
==> ERROR: A failure occurred in package_linux-ck-headers().
Aborting...
The build failed.

graysky commented on 2013-02-26 00:06

Bump to v3.7.9-2

Changelog: Security fix and updated config files.
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=61720e7b3541a8b801b333f8d7a40ce44223cdd1

graysky commented on 2013-02-25 20:45

Done. Since this is still a 'pre-pre-release' I did not bump the pkgver; it will stay at 0. You'll find the patch and updated config files therein:

http://repo-ck.com/PKG_source/linux-ck-next/linux-ck-3.8-0.src.tar.gz

graysky commented on 2013-02-25 20:44

Done. Since this is still a 'pre-pre-release' I did not bump the pkgver; it will stay at 0.

http://repo-ck.com/PKG_source/linux-ck-next/linux-ck-3.8-0.src.tar.gz

Scimmia commented on 2013-02-25 20:38

How about the security patch that Arch just applied to it's stock kernels?

https://projects.archlinux.org/svntogit/packages.git/tree/trunk/CVE-2013-1763.patch?h=packages/linux

misc commented on 2013-02-20 23:02

Yes, that's me :)

As always, thanks.

graysky commented on 2013-02-20 17:06

Isn't it misc who is always interested in the pre-release of package? The ground work has been laid for the 3.8 release. We just need CK to publish.

For those who MUST have the code, it is below. REMEBER that there is no bfs in this yet!

http://repo-ck.com/PKG_source/linux-ck-next/linux-ck-3.8-0.src.tar.gz

graysky commented on 2013-02-20 11:43

Plz don't flag out of date until CK releases a patchset that will apply to the 3.8 tree.

graysky commented on 2013-02-18 02:38

Bump to v3.7.9-1

Changelog: https://lkml.org/lkml/2013/2/17/115
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=0e242898b5be25addce190b9b5c225a02947673c

graysky commented on 2013-02-14 23:15

Bump to v3.7.8-1

Changelog: https://lkml.org/lkml/2013/2/14/356
Commit:

SirWuffleton commented on 2013-02-12 22:28

@Senth: The i686 architecture is still fully supported, I'm running the latest version of linux-ck on my i686 box with no issues whatsoever.

If you are referring to the comment referencing PAE and i686, the only thing that comment was asking was to enable HIGHMEM64/PAE to allow more memory to be used by the i686 kernel without additional configuration on the user's end. In no way was it inferring that the i686 architecture is no longer being supported.

Senth commented on 2013-02-12 21:54

then.. i686 architecture is not supported ?

graysky commented on 2013-02-11 22:13

Bump to v3.7.7-1

Changelog: http://lkml.org/lkml/2013/2/11/362
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=f13186c9a66a013b2514810635d769d71af430d6

graysky commented on 2013-02-11 22:12

Bump to v3.7.7-1

Changelog: http://lkml.org/lkml/2013/2/11/362
http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=f13186c9a66a013b2514810635d769d71af430d6

graysky commented on 2013-02-07 23:56

Bump to v3.7.6-2

Changelog: Update BFQ to v6.
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=3ea74f82742de78b72ed2237b6b60c769f66573f

graysky commented on 2013-02-07 21:18

Bump to v3.7.6-2

Changelog: Update BFQ to v6.
Commit:

Raziel23 commented on 2013-02-07 07:55

BFQ version 6 is out.

graysky commented on 2013-02-05 10:49

Sorry dude, I do not want to modify the ARCH config files. This package should stay as close to the ARCH package is possible. That said, you are free to edit the config by enabling the nconfig option in the PKGBUILD.

V3n3RiX commented on 2013-02-05 10:42

Can you please enable PAE on i686 builds? I have 4G memory, and i don't wanna switch to x86_64...yet!

graysky commented on 2013-02-05 09:58

Because the maintainer needs to add the patch to allow it. See the nvidia-ck package and https://bugs.archlinux.org/task/33710

clayman commented on 2013-02-05 06:13

Hmm, I seem to have problems compiling nvidia-beta drivers against this kernel, the driver says the usual "*** Unable to determine the target kernel version. ***". Anyone can confirm?

graysky commented on 2013-02-04 15:42

Bump to v3.7.6-1

Changelog: https://lkml.org/lkml/2013/2/3/421
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=a45f4a5142881cbd2c4a3fd0c39ae49fd9572121

lano1106 commented on 2013-01-29 16:13

The bug fixed by the latest patch has been first reported on ArchLinux forum:

https://bbs.archlinux.org/viewtopic.php?id=154594

graysky commented on 2013-01-29 09:10

Bump to v3.7.5-2

Changelog: New version of bfs - http://ck-hack.blogspot.com/2013/01/bfs-0427-for-linux-37x.html
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=d744ec526f00a4984cb00fa4355a401e7e7ed342

graysky commented on 2013-01-29 03:03

Bump to v3.7.5-2

Changelog: New version of bfs - http://ck-hack.blogspot.com/2013/01/bfs-0427-for-linux-37x.html
Commit:

graysky commented on 2013-01-28 22:12

Bump to v3.7.5-1

Changelog: https://lkml.org/lkml/2013/1/27/302
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=352811607633d0fb193c63ab0ea30545b07659ee

roentgen commented on 2013-01-27 08:41

MikeDacre, this will probably be solved by installing linux-headers or in your case linux-ck-[something]-headers.

MikeDacre commented on 2013-01-27 04:18

Hi,

How can I get all of the sources? I am trying to build alsa-driver.hda-intel.hda-codec-realtek-git and getting the following error:

The file /lib/modules/3.7.4-1-ck/build/include/INCLUDE_VERSION_H does not exist.
Please install the package with full kernel sources for your distribution
or use --with-kernel=dir option to specify another directory with kernel
sources (default is /lib/modules/3.7.4-1-ck/build).

Thanks,

Mike

graysky commented on 2013-01-22 02:45

Bump to v3.7.4-1

Changelog: http://lkml.org/lkml/2013/1/21/485
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=5f83b741ae7244aa07ff66bddb366a0415c9fd13

graysky commented on 2013-01-21 21:30

Bump to v3.7.4-1

Changelog: http://lkml.org/lkml/2013/1/21/485
Commit: pkgbuild.com/git/aur-mirror.git is down

graysky commented on 2013-01-20 21:28

@hadrons - An easy way to tell is to run the corresponding kernel from [testing] and see if the same errors occur for your hardware since both are v3.7.3.

hadrons123 commented on 2013-01-20 18:06

nomodset took me to ttys in fedora kernel 3.7.2.204 64 bit. But Arch Testing 3.7.3 kernel or your kernel doesn't go beyond fsck. It just goes black. There has been a bunch of updates for intel KMS in this kernel. Not sure if that's the culprit. Not sure if anyone else is having the same issue too.

graysky commented on 2013-01-19 15:45

Not sure about your specific hardware, sorry. Perhaps someone else can help.

hadrons123 commented on 2013-01-19 15:06

I cannot boot into your kernel. The screen goes black after the fsck. I have lenovo y580.

I have dkms-bbswitch, hack-lenovo dkms module, nvidia-bumblebee, nvidia-utils-bumbleebee and bumblebee installed.

I thought it was something to do with nvidia modules but couldn't finish boot even if i switched off nvidia in BIOS.

hadrons123 commented on 2013-01-19 12:56

I cannot boot into your kernel. The screen goes black after the fsck. I have lenovo y580.

I have dkms-bbswitch, hack-lenovo dkms module, nvidia-bumblebee, nvidia-utils-bumbleebee and bumblebee installed.

I thought it was something to do with nvidia modules but couldn't finish boot even if i switched off nvidia in BIOS.

graysky commented on 2013-01-17 22:10

Bump to v3.7.3-1

Changelog: http://lkml.org/lkml/2013/1/17/329
Commit: --- pkgbuild.com/git/aur-mirror.git is down ---

kyak commented on 2013-01-12 12:54

@graysky thanks a lot, very much appreciated!

graysky commented on 2013-01-12 12:41

@kyak - Correct. I am building up 3.7.2-2 right now. Even on my cluster, it take about 7 min per package. As you know, there are 25 packages across i686/x86_64 and all the cpu flavors to build. Should be online in 2-3 h from now.

kyak commented on 2013-01-12 12:33

@graysky it seems that repo-ck hasn't got latest updates. I'm still on 3.7.1-3-ck and pacman reports "repo-ck" repo as up to date.

graysky commented on 2013-01-12 11:53

Bump to v3.7.2-2

Changelog: Sync'ed up with ARCH pacakge, add CONFIG_ARPD, add CONFIG_DYNAMIC_DEBUG and fix GPU hang.
Commit: --- pkgbuild.com/git/aur-mirror.git is down ---

graysky commented on 2013-01-12 11:52

Bump to v3.7.2-2

Changelog: Sync'ed up with ARCH pacakge, add CONFIG_ARPD, add CONFIG_DYNAMIC_DEBUG, fix GPU hang.
Commit: (http://pkgbuild.com/git/aur-mirror.git still not updating)

graysky commented on 2013-01-11 20:31

Bump to v3.7.2-1

Changelog: https://lkml.org/lkml/2013/1/11/361
Commit:

qpalz commented on 2013-01-04 14:59

@cbowman57

OK, I will make them separate.

cbowman57 commented on 2013-01-04 14:58

I say keep it clean qpalz.

qpalz commented on 2013-01-04 06:31

@graysky

I have tried making linux-uksm-ck be able to disable uksm and become linux-ck, but in a very ugly way since if you disable ck or uksm, the kernel suffix remains -uksm-ck. I can only fix this problem in a very ugly way, so I would rather not to use a extremely ugly method to solve a ugly problem.

The extremely ugly method is to modify linux-uksm-ck.install linux-uksm-ck.preset dynamically when makepkg is running. But it results in the wrong checksum when you want to rebuild the package. Then a more more more ugly method can be used, that is to change the checksum dynamically...

So, is it better not to merge them? What do you think?

Anonymous comment on 2013-01-03 14:13

@graysky,thanks.i see.

qpalz commented on 2013-01-03 10:04

@jr0501
I am now maintaining it.

graysky commented on 2013-01-03 10:02

I checked out that patch set a while ago and made the choice not to add it to the package. I am still uninterested. Sorry.

Anonymous comment on 2013-01-03 05:08

linux-uksm-ck is orphan,could you merge with it?

thanks.

graysky commented on 2012-12-21 12:47

@Thaodan - Yes, that it there to remove extra info from versioning just like in the ARCH kernel's PKGBUILD. If you are curious about it, simply copy the file that it modifies to a new name, run the sed line, then diff the two files. You will see.

Thaodan commented on 2012-12-21 02:28

Im' updating my modifed linux-pf PKGBUILD (added propper package-naming description) and addded some changes that you done to your PKGBUILD (for example docs, turn of NUMA). And I found 'sed -ri "s|^(EXTRAVERSION =).*|\1 -${pkgrel}|" Makefile' in your PKGBUILD, for what is it?

graysky commented on 2012-12-21 02:09

Bump to v3.7.1-3

Changelog: Added ARCH fix or watchdog regression, see: https://bugs.archlinux.org/task/33095
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=bc06b904a8d071e0549217dcdafe47f9e0eba5b7

graysky commented on 2012-12-20 22:19

Bump to v3.7.1-3

Changelog: Added ARCH fix or watchdog regression, see: https://bugs.archlinux.org/task/33095
Commit:

graysky commented on 2012-12-20 22:19

Bump to v3.7.1-3

Changelog: Added ARCH fix or watchdog regression, see: https://bugs.archlinux.org/task/33095
Commit:

graysky commented on 2012-12-20 17:01

@vic - You need to provide details but I suspect you didn't compile the usb drivers you need. Suggest you open a forum thread to keep the AUR clean.

vic.pozd commented on 2012-12-20 16:55

2 graysky
I have problem with usb mouse (kernel compiled with HIGHMEM64 option)
Did you know how to solve this problem?

vic.pozd commented on 2012-12-20 16:50

2 graysky
I have problem with usb mouse (kernel compiled with HIGHMEM64 option)
Did you know how to solve this problem?

graysky commented on 2012-12-19 01:19

...what?

I you do not want to run the 3.7 tree of this package, you can find the source for the 3.6 series (including 3.6.11-1) here: http://repo-ck.com/PKG_source/

M4rQu1Nh0S commented on 2012-12-19 01:06

the kernel 3.7 is UNSTABLE and are in testing session on package of Arch Linux.
I set make on source of kernel 3.7, but in makepkg on BUILDPKG of this tarball I received erro from build().

how to solve this?

graysky commented on 2012-12-18 20:46

Bump to v3.7.1-2

Changelog: Added nfs41 support mirroring the changes in the official ARCH kernel.
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=df906aaadf5bf92d8b94b3500a23af3a1668ad35

graysky commented on 2012-12-18 18:34

# DRTAILS FOR _NUMA_off=yes
# Since 99.9% of users do not have multiple CPUs but do have multiple cores in one CPU
# see, https://bugs.archlinux.org/task/31187

Anonymous comment on 2012-12-18 15:34

Why NUMA is disabled by default? (_NUMA_off=yes)

graysky commented on 2012-12-17 21:17

Bump to v3.7.1-1

Changelog: https://lkml.org/lkml/2012/12/17/361
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=4ca16d6cef9d9f08289e545882a8fc75a3738c49

fincan commented on 2012-12-17 19:36

3.7.1

graysky commented on 2012-12-15 15:02

Refreshed v3.7.0-1

Changelog: Updated URL to ck1. There is no functional change.
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=e711b74e859f19ee6da58fc64ab00fdd17a98e48

graysky commented on 2012-12-14 01:21

Bump to v3.7.0-1

Changelog: https://lkml.org/lkml/2012/12/10/688 and http://kernelnewbies.org/LinuxChanges
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=2b602c4160d4aa622d08e3485059deb39e1e9dca

graysky commented on 2012-12-14 01:02

Bump to v3.7.0-1

*Changelog: https://lkml.org/lkml/2012/12/10/688 and http://kernelnewbies.org/LinuxChanges
*Commit:

snack commented on 2012-12-12 14:00

@graysky: key downloaded & signed, everything works now. Thanks.

graysky commented on 2012-12-12 13:52

Looks like Paolo already released a BFQ for 3.7 and I have been in email contact with CK. The ck1 patchset is currently being tested and should be released soon.

graysky commented on 2012-12-12 12:57

@kyak - Yes, it should. Thank you for reporting. It was a typo on my part in my repo build script which has been corrected a few min ago: https://github.com/graysky2/repo-ck/commit/fb41c335aeff95c5873b53ec5e5e1bb065fbe946

@mrbit - This is an unfortunate limitation of the ncurses interface. You have two options:
1) Make your terminal larger before your run makepkg
2) Change line 183 to: make menuconfig

This should do it for you: sed -i 's/make nconfig/make menuconfig/' PKGBUILD

mrbit commented on 2012-12-12 12:40

_makenconfig="y"


==> Running make nconfig
HOSTCC scripts/kconfig/nconf.gui.o
HOSTCC scripts/kconfig/nconf.o
HOSTLD scripts/kconfig/nconf
scripts/kconfig/nconf Kconfig
Your terminal should have at least 20 lines and 75 columns
make[1]: *** [nconfig] Errore 1
make: *** [nconfig] Errore 2
==> ERRORE:

kyak commented on 2012-12-12 12:36

Hi graysky,

I think that virtualbox-ck-host-modules-core2 should replace virtualbox-ck-host-modules-corex. Otherwise i have to resolve conflict between virtualbox-ck-host-modules-corex and linux-ck-core2 manually (i.e. remove virtualbox-ck-host-modules-corex, then replace linux-ck-corex with repo-ck/linux-ck-core2, then install virtualbox-ck-host-modules-core2).

graysky commented on 2012-12-12 12:25

@mrbit - You will need to edit the PKGBUILD and change the nconfig option to "y" before you compile. Once you start makepkg to compile, you will be presented with the usual nconf. The cpu options can be enabled under:
Processor type and features>Processor family

Note that I am pretty sure the i5-2400 is a sandy bridge (gen2 core processor). You should google around to be sure.

mrbit commented on 2012-12-12 11:13

how to change the config.x86_64 for
Intel(R) Core(TM) i5-2400 CPU @ 3.10GHz ??

graysky commented on 2012-12-12 10:38

@snack - Yes, I did change my key. Sorry, I posted this to the support thread in the bbs but not here:

"Already all, hosting issues are fixed and repo is populated with new packages. Note that I generated a new gpg key and revoked the old one. Pacman should grab it automatically. YOU WILL NEED TO SIGN THE KEY. Do so after pacman imports it or after you manually import it.

% sudo pacman-key --lsign-key 5EE46C4C
==> Updating trust database...

Enjoy!"

snack commented on 2012-12-12 03:30

While trying to install 3.6.10 I get:

errore: linux-ck-core2: key "88A032865EE46C4C" is unknown
errore: key "88A032865EE46C4C" could not be looked up remotely
errore: linux-ck-core2-headers: key "88A032865EE46C4C" is unknown
errore: key "88A032865EE46C4C" could not be looked up remotelyù

I didn't get this error with previous updates. Graysky, did you change your key? Or is there something wrong in my pacman keyring?

pub 2048R/6176ED4B 2011-11-07
uid graysky <graysky@archlinux.us>
sub 2048R/9B453E32 2011-11-07

graysky commented on 2012-12-12 01:23

@misc - Ha, sure. I haven't tested it, but did diff it against the official one. Plz test and lemme know if bad.

Note to other users - This is PRERELEASE package and does not contain ck1 or bfq! http://repo-ck.com/PKG_source/linux-ck-next/linux-ck-3.7-0.src.tar.gz

To answer your question, no, there are changes to the PKGBUILD as well. See: https://projects.archlinux.org/svntogit/packages.git/log/trunk?h=packages/linux

graysky commented on 2012-12-12 01:22

@misc - Ha, sure. I haven't tested it, but did diff it against the official one. Plz test and lemme know if bad.

Note to other users - This is PRERELEASE package and does not contain ck1 or bfq! http://repo-ck.com/PKG_source/linux-ck-next/linux-ck-3.7-0.src.tar.gz

misc commented on 2012-12-12 01:06

As always, a 3.7.0-0 would be appreciated. Or is it enough if one just applies the changes from the stock kernel? In particular, can one just copy its two config files?

graysky commented on 2012-12-11 21:35

Plz don't flag out-of-date until:

1) CK releases a ck1 patchset for the 3.7 tree.
2) Paolo releases a BFQ for the 3.7 tree.

graysky commented on 2012-12-10 22:40

Bump to v3.6.10-1

Changelog: Package gets more options for CPUs; see https://bbs.archlinux.org/viewtopic.php?id=154333
Upstream Changelog: https://lkml.org/lkml/2012/12/10/506
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=1ac516c54772351addc823e56bf64d539f869cfe

graysky commented on 2012-12-09 02:09

Not sure... you'll give to dig into it.

http://en.gentoo-wiki.com/wiki/Safe_Cflags/AMD#Athlon_X2_7x50.2C_Phenom_X3.2FX4.2C_Phenom_II.2C_Athlon_II_X2.2FX3.2FX4.2C_Turion_II

http://gcc.gnu.org/onlinedocs/gcc-4.7.2/gcc/i386-and-x86_002d64-Options.html#i386-and-x86_002d64-Options

whoops commented on 2012-12-08 20:27

Sorry, I don't fully understand the categories.
Is Phenom II X6 = K10 = Opteron = Option 1?
Or is it a case for "Generic-x86-64"?

Thanks!

graysky commented on 2012-12-08 16:23

I tested the 'kernel-36-gcc47-1.patch' written by André Ramnitz using three different machines running a generic x86-64 kernel and an otherwise identical kernel running with the optimized gcc options.

Conclusion:
There are small but real speed increases using a make endpoint to running with this patch.

Details:
1) Three test machines: Intel Xeon X3360, Intel i7-2620M, Intel Core i7-3660K.
2) All ran the make benchmark (linked below) 35 times while booted into a 'generic' kernel. Then all ran the same make benchmark 35 times after booting into an optimized kernel. Below are the optimizations chosen for each machine.
2a) X3360 = core2
2b) i7-2620M = corei7-avx
2c) i7-3660K = core-avx-i
3) Analyzed resulting distributions for statistical significance via ANOVA plots that clearly show statistically significant albeit small differences.

Links to ANOVA plots:
http://s19.postimage.org/68urcofzn/corei7_avx.png
http://s19.postimage.org/ozwomuak3/core_avx_i.png
http://s19.postimage.org/d0l6fj4z7/core2.png

References:
Bash script that controls the benchmark: https://github.com/graysky2/bin/blob/master/bench
Log file generated by script: http://repo-ck.com/bench/compile_time_optimization.txt.gz

graysky commented on 2012-12-08 16:22

Bump to v3.6.9-3

Changelog: Added the kernel-36-gcc47-1.patch by André Ramnitz that adds additional CPU family support.
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=5e6d83e6cc2c6027f876509611368d74b469fedd

graysky commented on 2012-12-08 16:20

Thanks for the info, misc.

xzy3186 commented on 2012-12-04 16:39

You are right, linux-3.6.9 from testing repo have exactly the same problem. Thank you very much, I will delete my reply later.

misc commented on 2012-12-04 14:01

Well, I think that has resolved itself. localmodconfig *is* streamline_config.pl — as is localyesconfig. Both are simply an "alias" then also set as variable (see scripts/kconfig/Makefile) that have the script behave differently.

graysky commented on 2012-12-04 12:17

'''Bump to v3.6.9-2'''

*Changelog: Incorporated fat-3.6.x.patch as does the ARCH package.
*Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=2d661e58d8f8983755e86ed66e7e1630dbdd4d41

graysky commented on 2012-12-04 10:25

@misc - If memory serves me, Steven told me that the perl script you mentioned differs from the call to `make localmodconfig` somehow. Plz follow-up with him and let me know.

@xzy3186 - Not sure... plz try with the linux package from [testing] and verify that it isn't an upsteam bug.

xzy3186 commented on 2012-12-04 06:10

After upgrade to linux-ck 3.6.9-1, gdm can not load at boot on my laptop. I have to run "systemctl stop gdm.service, systemctl start gdm.service" by hand. Any idea on this problem?

misc commented on 2012-12-04 00:53

Seeing that streamline_config.pl is still being actively maintained ( http://bit.ly/YJ1T4X ), wouldn't it be worth to try out how it performs? Only as precursor to localmodcfg, not on its own, of course.

I mean, if the Steven Rostedt still truly considered it such a bad idea, wouldn't they have removed it by now?

graysky commented on 2012-12-03 23:06

Bump to v3.6.9-1

Changelog: https://lkml.org/lkml/2012/12/3/489
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=c6e7ed7dc8c486fd8ec66a681c5b215461873034

graysky commented on 2012-11-26 21:02

Bump to v3.6.8-1

Changelog: https://lkml.org/lkml/2012/11/26/669
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=b8ca3028322bf05c25cd94c6f875578ea4ce0fd3

graysky commented on 2012-11-22 05:25

Thanks for the report. Fixed.

kalpik commented on 2012-11-22 01:37

Problem with the repo

warning: cannot resolve "linux-ck-corex-corex", a dependency of "linux-ck-corex-headers"
:: The following package cannot be upgraded due to unresolvable dependencies:
linux-ck-corex-headers

graysky commented on 2012-11-21 23:28

Bump to v3.6.7-3

Changelog: Updated conflicts array for better AUR/repo support.

graysky commented on 2012-11-20 15:07

Bump to v3.6.7-2

Changelog: https://wiki.archlinux.org/index.php/Linux-ck/Changelog

graysky commented on 2012-11-20 15:06

Bump to v3.6.7-2

Changelog: Included fix for FS#32615 - Check for valid irq_cfg pointer in smp_irq_move_cleanup_interrupt
Commit:

kyak commented on 2012-11-17 13:03

thanks graysky

graysky commented on 2012-11-17 12:25

@kyak - https://wiki.archlinux.org/index.php/Repo-ck#Package_Trivia.2FRepo_Statistics

kyak commented on 2012-11-17 11:28

@graysky,

Thanks, i understand that these CPU-specific options are "under Processor type and features>Processor family or by setting-up the .config file accordingly."

However, i'm wondering how do you build linux-ck-corex from this PKGBUILD? Are you using some script to modify .config or you are choosing the processor manually every time? Also, the package name is "linux-ck-corex", not just "linux-ck", so you definitely change the PKGBUILD. Could you tell more about how you do it, so i could rebuild linux-ck-corex here?

For now all i got is this PKGBUILD, and it is not straigtforward how to get binary linux-ck-corex out of it.

graysky commented on 2012-11-15 19:54

@Kyak - https://wiki.archlinux.org/index.php/Repo-ck#Details

kyak commented on 2012-11-15 18:23

Hi graysky,

Just wondering, how do you build linux-ck-corex from this PKGBUILD?
I'm currently using linux-ck-corex from repo, but need to build from source to enable several options.

Thanks!

skydrome commented on 2012-11-08 21:34

biltong, you would want to add a line: KCFLAGS="options here", but you can specify the cpu to build for in the kernel config so -march is redundant in this case

bassu commented on 2012-11-08 14:40

Brain Fuck Scheduler fucking RULES!
Thank you for all your efforts and time, Graysky!

Anonymous comment on 2012-11-07 19:25

Is there any way to include a custom DSDT? nconfig does not enable the necessary options to be edited. Also I'm assuming if we want to compile with optimizations we would put them in $MAKEFLAGS, correct? For example on my 2600k I would use -j9 and --mtune=native, where would those go?

graysky commented on 2012-11-05 19:43

Bump to v3.6.6-1

Changelog: https://lkml.org/lkml/2012/11/5/124
Commit:

graysky commented on 2012-11-01 19:07

Bump to v3.6.5-1

Changelog: https://lkml.org/lkml/2012/10/31/504
Commit:

sudokode commented on 2012-10-31 19:05

Ah, nvm. Looks like it's a general problem according to this bug: https://bugs.archlinux.org/task/32307. Only applies to AMD with >=8G of memory.

sudokode commented on 2012-10-31 19:00

@graysky Hm. Looks like 3.6.4-1-ARCH (now in [core]) also fails.

graysky commented on 2012-10-31 10:14

@sudokode - Is this true for both linux-ck and linux from [testing]?

sudokode commented on 2012-10-31 03:21

Not sure what's going on here, but 3.6.4 doesn't boot for me. Nothing in the kernel log, but it's definitely the kernel.

graysky commented on 2012-10-30 01:14

@jqx - OK. Glad to hear that it is again not a problem with the ck patches, but with the upstream. You should spend some time googling for the specific crash. On one of my boxes, I found that the 3.6 series crashes if I power down with my lirc modules probed. If I rmmod them and then shutdown, no issues at all.

jqww2002 commented on 2012-10-30 00:39

@graysky Just now update with testing. I'm very sad that the [testing] linux 3.6.4 has the same problem as linux-ck 3.6.4, gets a panic at booting.
Maybe this is a kernel issue, waiting for v3.6.5 ?

Anonymous comment on 2012-10-29 20:52

RULEZ!! Thanks for your work graysky!

graysky commented on 2012-10-29 16:38

@jqx - Are you 'jim1960' by chance? https://aur.archlinux.org/packages.php?ID=50911
If not, this guy has the same problem as you which seems to be independent of the ck patchset.

graysky commented on 2012-10-29 15:09

@jqx - Looks like 3.6.4-1 hit [testing] about 37 min ago. Go forth and test :)

ryley commented on 2012-10-29 11:44

The issue I had earlier with corrupted file transfers to the iPhone is no longer present in 3.6.4. Thanks graysky for keeping linux-ck so up to date!

jqww2002 commented on 2012-10-29 11:37

@graysky The official repos is still v3.6.3. I will have a try when it being v3.6.4.

graysky commented on 2012-10-29 11:18

@jqx - First question I always ask is: have you also tested version x from the official repos... in this case however, the devs have not bumped 3.6.3-->3.6.4 yet. As soon as they do, I would ask you to test it and see if you get the same problem. I actually have yet to see a panic that was caused by ck's stuff.

jqww2002 commented on 2012-10-29 08:23

I got a panic after update to v3.6.4 (x86) today from repo-ck.com.
The panic message is "PANIC:early exception 08 rip246:10 error 81041d86 cr2 0".
I try google and find that adding mem=4000M to grub, all it's ok. But I have 6G mem, v3.6.3 is all ok.
Can any help?

graysky commented on 2012-10-28 21:06

Bump to v3.6.4-1

Changelog: https://lkml.org/lkml/2012/10/28/99
Commit:

graysky commented on 2012-10-22 21:58

Bump to v3.6.3-2

Changelog: Incorporated ARCH patches and readded memgroups per ARCH commit https://projects.archlinux.org/svntogit/packages.git/commit/trunk?h=packages/linux&id=af56e226b41a0bb4714a3861bbfaafe2346c66f5
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=41e2d71562a9f73c32afee2e6a02a9a9c1e63578 and http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=6f2b745c8cad16488a61eb6e0e212b6146caa89d

graysky commented on 2012-10-22 20:04

Bump to v3.6.3-2

Changelog: Incorporated ARCH patches and readded memgroups per ARCH commit https://projects.archlinux.org/svntogit/packages.git/commit/trunk?h=packages/linux&id=af56e226b41a0bb4714a3861bbfaafe2346c66f5
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=41e2d71562a9f73c32afee2e6a02a9a9c1e63578 and

ryley commented on 2012-10-22 05:22

Glad to see its not specific to me or iPhones. I just tried out linux-mainline from AUR (3.7-rc2) and it does not have any issues with copying files though nautilus/gvfs-afc, so I guess the issue is only in the 3.6 series atm.

coderkun commented on 2012-10-22 05:01

@ryley: same problem with ck-kernel 3.6.2 and an android (4.0) phone.

ryley commented on 2012-10-22 01:03

Just upgraded from 3.5.6 and it seems copying files to an iPhone thrugh nautilus/gvfs-afc now produces garbage, unusable files. Reverting to 3.5.6 and everything is fine. Is anyone else having this problem?

graysky commented on 2012-10-21 22:04

Bump to 3.6.3-1

Changelog: https://lkml.org/lkml/2012/10/21/107
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=45dd6d2b16363b963ec99bf32a6b9ea70053a19f

Anonymous comment on 2012-10-15 08:30

@graysky: that's correct. I have a specific BUILDDIR setup in makepkg.conf and I think this is a valid thing to do. It helps keep the older builds around (for debugging, speeding up the builds and so on). So, it's good if the PKGBUILD explicitly references the patches/files to be used.

graysky commented on 2012-10-15 07:56

@naveen - If you look in the source array you'll see that the correct versions are downloaded. I suspect that you are reusing the same directory each time you build and therefore keep older versions of patches.

Anonymous comment on 2012-10-15 07:26

Can you modify the PKGBUILD to specifically look for the 3.6 BFQ patches? My source directory happens to have both the older 3.5 patches and the newer 3.6 patches for BFQ.
I guess it is probably better to hardcode patch names in PKGBUILd.

unknwn commented on 2012-10-14 12:10

that's true it is xorg related.

hermes14 commented on 2012-10-14 09:43

@unknwn
There's a topic about that: https://bbs.archlinux.org/viewtopic.php?pid=1175368
Make sure whether the issue is related to the xorg or to the linux-ck upgrade.

unknwn commented on 2012-10-14 07:58

With 3.6.2-1 brightness and volume controls through fn+key doesn't work anymore on my laptop.

felixonmars commented on 2012-10-14 00:56

@graysky
Oops, thanks a lot :)

graysky commented on 2012-10-13 17:32

Bump to v3.6.2-1

Changelog: https://lkml.org/lkml/2012/10/12/415
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=0f8252e6a57a80642a39df5590ddcb7d6eecf453
What's new in the 3.6 series: http://kernelnewbies.org/LinuxChanges

graysky commented on 2012-10-13 16:42

@felixonmars - I went a head and created a 3.5.7-1 package for you which you can grab here: http://repo-ck.com/PKG_source/linux-ck/linux-ck-3.5.7-1.src.tar.gz

felixonmars commented on 2012-10-13 14:49

Just a FYI: I'm stick at 3.5.x because there is a serious problem [1] of zram in the 3.6 tree. A patch was submitted two days ago but not seen in the latest 3.6.2 build.
And @graysky, there is a 3.5.7 release, maybe it's appropriate to update to it?

[1] http://code.google.com/p/compcache/issues/detail?id=102

graysky commented on 2012-10-13 14:43

@OK100 - Not out-of-date. I am aware of the release; see my previous comment. I am in email contact with Con now attempting to diagnose and fix.

graysky commented on 2012-10-13 13:34

OK! CK just released[1] an OFFICIAL CK1 with bfs v0.425 that patches into the 3.6 tree, however it has a slight problem and will not build with guest virtualization enabled which as you know is part of the stock ARCH config. I emailed him and hope to have a fix soon. I will keep you updated.

[1] http://ck-hack.blogspot.com/2012/10/the-vicissitudes-of-life.html

graysky commented on 2012-10-13 13:33

OK! CK just released an OFFICIAL CK1 with bfs v0.425 however it has a slight problem and will not build with guest virtualization enabled which as you know is part of the stock ARCH config. I emailed him and hope to have a fix soon. I will keep you updated.

graysky commented on 2012-10-11 01:06

Just wanted to update the group - After much testing, it seems that the build errors ONLY occurs when I build having guest virtualization enabled. If I disable it, I can build the _full set_ of modules just fine using the ck1 patch Martin provided. I am reluctant put to push this to the AUR and to the repo as-is. For those interested, I have the source package 3.6.1-4 which you can download and build yourself. This version has the aforementioned option disabled and builds just fine for me (both x86_64 and i686).

Download it from here: http://repo-ck.com/PKG_source/testing/linux-ck-3.6.1-4.src.tar.gz

graysky commented on 2012-10-11 00:50

Well guys - After much testing, it seems that the build errors ONLY occurs when I build having guest virtualization enabled. If I disable it, I can build the _full set_ of modules just fine using the ck1 patch Martin provided. I am reluctant put to push this to the AUR and to the repo as-is.

[ ] Processor type and features --->Paravirtualized guest support --->

Just wanted to update the group.

graysky commented on 2012-10-10 16:49

Critical problems forced me to abandon this. Downgrading to 3.5.6 at this time. Discussion: https://bbs.archlinux.org/viewtopic.php?id=111715&p=40

graysky commented on 2012-10-10 16:17

Ack... looks like something is fucked-up with the package as it built fine on my work station but not on my server. Hold off on downloading until I figure it out and update to -3.

graysky commented on 2012-10-07 21:01

Bump to v3.5.6-1
Changelog: https://lkml.org/lkml/2012/10/7/110
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=14469ae78302348d5985c2244a2b2f6c9b5a04d3

For those of you [misc] wanting to use linux-ck 3.6.1 WITHOUT the ck1 patchset (an inert PKGBUILD basically): http://repo-ck.com/PKG_source/testing/linux-ck-3.6.1-0.src.tar.gz

misc commented on 2012-10-04 20:56

Runs fine. Thanks.

graysky commented on 2012-10-04 19:37

@misc - Right... looks like we're just waiting on ck1. If you will be kind enough to test out 3.6.0-0 find it here (builds fine for me): http://repo-ck.com/PKG_source/testing/linux-ck-3.6-0.src.tar.gz

misc commented on 2012-10-04 19:05

With those patches that are available. Right now that's only BFQ, I think.

graysky commented on 2012-10-04 18:52

@Misc - You want an inert one (i.e. one without CK1 or BFQ)?

misc commented on 2012-10-04 12:57

I can has 3.6.0-0 PKGBUILD plz?

graysky commented on 2012-10-03 21:28

@mikronimo - Please don't post repo questions to the AUR. Instead use this thread if this does not work for you:

# pacman-key -r 6176ED4B && pacman-key --lsign-key 6176ED4B

Repo discussion thread: https://bbs.archlinux.org/viewtopic.php?id=111715

mikronimo commented on 2012-10-03 21:02

It's some time i use linux-ck-p4 without any problem, but today, during attempting the upgrade to the 3.5.5-1 version i had this problem:

acchetti (2): linux-ck-p4-3.5.5-1 linux-ck-p4-headers-3.5.5-1

Dimensione totale dei pacchetti da installare: 91,71 MiB
Dimensione netta dell'aggiornamento: 0,00 MiB

Vuoi procedere con l'installazione? [S/n] s
(2/2) verifica dell'integrità dei pacchetti [########################################] 100%
errore: linux-ck-p4: signature from "graysky <graysky@archlinux.us>" is unknown trust
errore: linux-ck-p4-headers: signature from "graysky <graysky@archlinux.us>" is unknown trust
errore: impossibile eseguire l'operazione richiesta (il pacchetto non è valido oppure è corrotto (firma PGP))
Si sono verificati degli errori, nessun pacchetto è stato aggiornato.
[root@bridge-live mikronimo]#

graysky commented on 2012-10-03 09:05

Bump to v3.5.5-1

Changelog: https://lkml.org/lkml/2012/10/2/356
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=95e060e1f2dd135df3f26365a71e90bd591b3475

Acknowledgments to stlifey for pointing out the fix in linux-git to the mistake in include/linux/sched.h that kept ck1 from applying to 3.5.5!

graysky commented on 2012-10-03 09:04

Bump to v3.5.5-1

Changelog: https://lkml.org/lkml/2012/10/2/356
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=95e060e1f2dd135df3f26365a71e90bd591b3475

graysky commented on 2012-10-03 01:29

@stlifey - Nice. Although the web interface seemed to give me a backwards patch, applying with the -R switch is just fine. Will test a bit and then upload. Until then, check my math:

Patch: http://pastebin.com/MJGy9c0b
Proposed 3.5.5-1 for this package: http://repo-ck.com/PKG_source/testing/linux-ck-3.5.5-1.src.tar.gz

graysky commented on 2012-10-03 01:13

@stlifey - Nice. Although the web interface seemed to give me a backwards patch, applying with the -R switch is just fine. Will test a bit and then upload. Until then, check my math:

http://pastebin.com/MJGy9c0b

You need to apply with: patch -Rp1 -i $srcdir/unfuck.patch

Anonymous comment on 2012-10-03 01:12

if we use bfs sche,we can revert this commit.

graysky commented on 2012-10-03 00:57

@stlifey - I am confused... these patches appear to be applied as part of the 3.5.5 release.

graysky commented on 2012-10-03 00:50

@stlifey - wait for 3.5.6 ?

Anonymous comment on 2012-10-03 00:08

@graysky
https://git.kernel.org/?p=linux/kernel/git/stable/linux-stable.git;a=commit;h=4f83989550ace0aa91464051cbaddc10e1b85778

graysky commented on 2012-10-02 19:36

@fincan - I saw it but ck1 [bfs] does not patch into it due to some upstream changes:

==> Patching source with ck1 including bfs v0.424
patching file arch/powerpc/platforms/cell/spufs/sched.c
patching file Documentation/scheduler/sched-BFS.txt
patching file Documentation/sysctl/kernel.txt
patching file fs/proc/base.c
patching file include/linux/init_task.h
Hunk #1 succeeded at 141 (offset 9 lines).
Hunk #2 succeeded at 269 (offset 10 lines).
patching file include/linux/ioprio.h
patching file include/linux/sched.h
Hunk #3 FAILED at 1240.
Hunk #4 succeeded at 1360 (offset 3 lines).
Hunk #5 succeeded at 1596 (offset 3 lines).
Hunk #6 succeeded at 1671 (offset 3 lines).
Hunk #7 succeeded at 2055 (offset 3 lines).
Hunk #8 succeeded at 2771 (offset 3 lines).
1 out of 8 hunks FAILED -- saving rejects to file include/linux/sched.h.rej
patching file init/Kconfig
patching file init/main.c
patching file kernel/delayacct.c
patching file kernel/exit.c
patching file kernel/posix-cpu-timers.c
patching file kernel/sysctl.c
patching file lib/Kconfig.debug
patching file include/linux/jiffies.h
patching file drivers/cpufreq/cpufreq.c
patching file drivers/cpufreq/cpufreq_ondemand.c
patching file drivers/cpufreq/cpufreq_conservative.c
patching file arch/x86/Kconfig
patching file kernel/sched/bfs.c
patching file kernel/sched/Makefile
patching file mm/vmscan.c
Hunk #2 succeeded at 928 (offset 32 lines).
Hunk #3 succeeded at 1921 (offset 32 lines).
Hunk #4 succeeded at 2873 (offset 32 lines).
Hunk #5 succeeded at 2885 (offset 32 lines).
patching file include/linux/swap.h
patching file mm/memory.c
patching file mm/swapfile.c
patching file mm/page-writeback.c
patching file kernel/Kconfig.hz
patching file arch/x86/kernel/cpu/proc.c
patching file arch/x86/kernel/smpboot.c
patching file include/linux/nfsd/stats.h
patching file include/net/inet_timewait_sock.h
patching file init/calibrate.c
patching file kernel/Kconfig.preempt
patching file Makefile
==> ERROR: A failure occurred in build().

fincan commented on 2012-10-02 19:16

3.5.5 out

agapito commented on 2012-10-01 10:53

Sorry!! You are right.

graysky commented on 2012-10-01 08:20

@agapito - Not out of date until:
*Arch package hits [testing].
*CK patchset is released.
*BFQ patchset is released.

See: https://wiki.archlinux.org/index.php/Linux-ck#Release_Cycle

graysky commented on 2012-09-20 00:59

@Jestern - no bother; glad we got to the bottom of it.

Anonymous comment on 2012-09-16 17:30

Yeah, my system was a little bit out-of-date. Sorry for bothering you ;)

graysky commented on 2012-09-16 14:44

@Jestern - This is by design. Is your system up-to-date? I suspect you have an old kmod on your system. Update with a sync'ed mirror and try again.

References:
http://mailman.archlinux.org/pipermail/arch-dev-public/2012-September/023530.html
https://projects.archlinux.org/svntogit/packages.git/commit/trunk?h=packages/linux&id=906c4499f7c48efe2a95bfd5c632d484d6b7ef2f

graysky commented on 2012-09-16 14:44

@Jestern - This is by design. Is your system up-to-date? I suspect you have on old kmod.

http://mailman.archlinux.org/pipermail/arch-dev-public/2012-September/023530.html
https://projects.archlinux.org/svntogit/packages.git/commit/trunk?h=packages/linux&id=906c4499f7c48efe2a95bfd5c632d484d6b7ef2f

Anonymous comment on 2012-09-16 14:03

Hello, with the latest PKGBUILD the order for depmod and "move module tree /lib -> /usr/lib" was exchanged and it gives this error:

DEPMOD 3.5.4-1-ck
ERROR: could not open directory /home/user/linux-ck/pkg/linux-ck/usr/lib/modules/3.5.4-1-ck: No such file or directory
FATAL: could not search modules: No such file or directory

Undoing the change, it builds correctly.

graysky commented on 2012-09-15 10:09

Bump to v3.5.4-1

Changelog: https://lkml.org/lkml/2012/9/15/10
Arch-specific Changelog: add FHANDLE, X86_32 and VIA_FB #31426 #31515 #31270 and reverse module move and depmod calls
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=64baad69a39ba0a6f9c591f0f4206008f2419632

graysky commented on 2012-09-15 10:09

Bump to v3.5.4-1
Changelog: https://lkml.org/lkml/2012/9/15/10
Arch-specific Changelog: add FHANDLE, X86_32 and VIA_FB #31426 #31515 #31270 and reverse module move and depmod calls
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=64baad69a39ba0a6f9c591f0f4206008f2419632

fincan commented on 2012-09-15 05:36

3.5.4

graysky commented on 2012-08-31 09:05

@alium - For users with just linux-ck kernel, I believe that provides statement needs to be there for packages that require 'linux-headers' or else the ARCH linux-headers package is downloaded. For example, community/virtualbox-source. Can you think of another option?

alium commented on 2012-08-28 07:29

linux-ck-headers provides also "linux-headers=${pkgver}", but I think it should not, especially if you have more kernels (I have installed linux and linux-ck)

For example, compiling r8168-all ended with error message "make: *** / usr/lib/modules/3.5.3-1-ARCH/build: Directory or file does not exist." (i have not installed linux-headers), because the kernel-ARCH required modules and the build -ARCH.

To make "universal" 'linux-headers' could not be linked linux-ck-headers to a specific version, as it is now.

alium commented on 2012-08-28 07:28

linux-ck-headers provides also "linux-headers=${pkgver}", but I think it should not, especially if you have more kernels (I have installed linux and linux-ck)

For example, compiling r8168 ended with error message "make: *** / usr/lib/modules/3.5.3-1-ARCH/build: Directory or file does not exist." (i have not installed linux-headers), because the kernel-ARCH required modules and the build -ARCH.

To make "universal" 'linux-headers' could not be linked linux-ck-headers to a specific version, as it is now.

graysky commented on 2012-08-27 21:13

Bump to v3.5.3-2

Changelog: Added patches from ARCH kernel: i915-i2c-crash-3.5.x.patch, watchdog-3.5.x.patch, and minor tweaks to PKGBUILD/install file per ARCH devs. Thanks misc.
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=b824bdcfdeaf487e9b80a1218b8993a85ce60296

graysky commented on 2012-08-27 20:26

Bump to v3.5.3-2

Changelog: Added patches from ARCH kernel: i915-i2c-crash-3.5.x.patch, watchdog-3.5.x.patch, and minor tweaks to PKGBUILD/install file per ARCH devs. Thanks misc.
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=f223bd79ee9b470f1853c1ecfefa01bc858a792c

misc commented on 2012-08-27 14:37

Arch kernel released with additional patches.

graysky commented on 2012-08-26 09:03

Bump to v3.5.3-1

Changelog: https://lkml.org/lkml/2012/8/25/203
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=f223bd79ee9b470f1853c1ecfefa01bc858a792c

graysky commented on 2012-08-26 08:47

@bluephoenix - plz open a discussion thread about modprobed_db... this has no relation to linux-ck

fincan commented on 2012-08-26 05:15

3.5.3 out

wilbowma commented on 2012-08-25 17:51

I decided to give the modprobed_db + localmodcfg a try.

I'm using lvm2 with dm-crypt, and run into errors when mkinicpio. dm-snapshot and dm-mirror, which are appearently required, are not found.

When I manually load these via modeprobe, they don't appear after I run modprobed_db store.

Any one have thoughts on this? My current solution it to also run nconfig and manually enable it.

graysky commented on 2012-08-25 10:26

Bump to v3.5.2-4

Changelog: Added five patches to fix bugs in ext4 filesystems which have been included in the 3.5 stable tree after having been affected by one of these myself.
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=f14264725afb7185f56a2710234d1773c1be7f0e

graysky commented on 2012-08-25 02:12

Bump to v3.5.2-4

Changelog: Added five patches to fix bugs in ext4 filesystems which have been included in the 3.5 stable tree after having been affected by one of these myself.
Commit:

corro commented on 2012-08-25 00:49

graysky/VladikSS: The patch made it into 3.6-rc3, see http://git.kernel.org/?p=linux/kernel/git/torvalds/linux.git;a=commit;h=535b6c51fe8293c88ce919cdfc4390c67a1cb6d1

ValdikSS commented on 2012-08-23 03:34

graysky, ok, I will. The developer guy in alsa-devel who made this patch said, that it will be included in upstream. I think it will be included in 3.6. I just found that audio codec is used 100% of time as powertop says and this cannot be fixed without reboot or suspend, but everything was OK for me in 3.4.x.

corro commented on 2012-08-23 00:50

I can confirm a noticable improvement in battery life with this patch applied. When measuring power consumption with powertop, it says it's about 2-3 Watts less than before. The latest kernel 3.6-rc3 has the same problem. ValdikSS, you should definitely report this issue.

graysky commented on 2012-08-22 22:58

Bump to v3.5.2-3

Changelog: Added minor patch to fix powersave issue with ALSA.
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=4d760ac08ea4190e352b91e0968ac8d4714bc5a9

graysky commented on 2012-08-22 22:57

Bump to v3.5.2-3

PKG Changelog: Added minor patch to fix powersave issue with ALSA.
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=4d760ac08ea4190e352b91e0968ac8d4714bc5a9

graysky commented on 2012-08-22 20:41

@ValdikSS - I changed my mind about your suggestion and have added it to v3.5.2-3. Can I ask that you open a bug report against the linux package in [testing] reporting the known issue and directing them to the patch you referenced? Will upload new PKGBUILD in a few.

graysky commented on 2012-08-21 19:21

@ValdikSS - I know nothing of this and am reluctant to add it. Feel free to add a few lines to the PKGBUILD on your own system to add support.

ValdikSS commented on 2012-08-21 03:00

Hello, graysky! Can you add ALSA powersave patch, as it's broken under some circumstances in 3.5.x for me.
http://mailman.alsa-project.org/pipermail/alsa-devel/2012-August/054532.html (the latest one only)
Thanks!

graysky commented on 2012-08-19 12:26

Bump to v3.5.2-2

PKG Changelog: Disabled NUMA since 99.9999 % of users do not have machines with multiple PHYSICAL processors. Performance boosts are tangible as measured by reduction in gcc compile times. For supporting data, see http://ck-hack.blogspot.com/2012/08/35-ck1-bfs-424-for-linux-35.html?showComment=1345343902486#c4511102253913582401
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=7311828d1358de95fd77ea5ec1255f6f5629fa51

graysky commented on 2012-08-18 13:19

Bump to v3.5.2-1

Sorry for the double bump, but I wanted to add more verbosity to the original announcement from 16 Aug 2012 21:18:39

PKG Changelog: Adapted PKGBUILD to conform to new ARCH standard.
Upstream changelog: https://lkml.org/lkml/2012/8/15/337
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=8ff60963c7fcbe5f32e4e7f85e619c4d0c6fcb77

Summary of BFS changes: none beyond CK fixing the tickless option and porting bfs v0.424 to the 3.5 tree. I therefore reverted the temp fix and have RE-ENABLED the ARCH default of tickless enabled. Just thought that I would also share some some tangible evidence of efficiency gains due to BFS:

http://ck-hack.blogspot.com/2012/08/35-ck1-bfs-424-for-linux-35.html?showComment=1345151880595#c5789006603939353411
and
http://ck-hack.blogspot.com/2012/08/35-ck1-bfs-424-for-linux-35.html?showComment=1345246929324#c5703441061197524272

Enjoy!

graysky commented on 2012-08-16 21:18

Bump to v3.5.2-1

PKG Changelog: Adapted PKGBUILD to conform to new ARCH standard.
Upstream changelog: https://lkml.org/lkml/2012/8/15/337
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=8ff60963c7fcbe5f32e4e7f85e619c4d0c6fcb77

Here is some tangible evidence of efficiency gains due to BFS: http://ck-hack.blogspot.com/2012/08/35-ck1-bfs-424-for-linux-35.html?showComment=1345151880595#c5789006603939353411

Enjoy!

graysky commented on 2012-08-16 20:08

Bump to v3.5.2-1

PKG Changelog: Adapted PKGBUILD to conform to new ARCH standard.
Upstream changelog: https://lkml.org/lkml/2012/8/15/337
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=8ff60963c7fcbe5f32e4e7f85e619c4d0c6fcb77

Here is some tangible evidence of efficiency gains due to BFS: http://ck-hack.blogspot.com/2012/08/35-ck1-bfs-424-for-linux-35.html?showComment=1345147322305#c3994016153175174657

Enjoy!

graysky commented on 2012-08-16 20:04

Bump to v3.5.2-1

PKG Changelog: Adapted PKGBUILD to conform to new ARCH standard.
Upstream changelog: https://lkml.org/lkml/2012/8/15/337
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=8ff60963c7fcbe5f32e4e7f85e619c4d0c6fcb77

xzy3186 commented on 2012-08-16 07:50

ck patch for linux-3.5 has been released
http://ck-hack.blogspot.jp/2012/08/35-ck1-bfs-424-for-linux-35.html

graysky commented on 2012-08-15 21:15

Bump to v3.4.9-1

Changelog: https://lkml.org/lkml/2012/8/15/413
Commit: pkgbuild.com/git/aur-mirror.git is down

fincan commented on 2012-08-15 16:34

3.4.9 and 3.5.2 out

qpalz commented on 2012-08-13 05:01

Here is patch-3.5-ck0.3, with the new bfs ported to 3.5: http://ck-hack.blogspot.com/2012/07/bfs-and-ck-delays-for-linux-350.html?showComment=1344805740367#c514497747808461131
and some fix by me.
The stability is higher now. (but still no guarantees. This patch is just for playing...)

qpalz commented on 2012-08-13 03:52

Here is patch-3.5-ck0.3, with the new bfs ported to 3.5: http://ck-hack.blogspot.com/2012/07/bfs-and-ck-delays-for-linux-350.html?showComment=1344805740367#c514497747808461131
and some fix by me.
Still unstable.

qpalz commented on 2012-08-12 12:27

@graysky
I agree with you because the stability of this patch is really bad.
But the stability of the patch for 3.4.6+(enable CONFIG_NO_HZ) is great. I am using it for several days and everything works great.
"http://dl.dropbox.com/u/20066450/Linux/kernel/ck-patch/patch-${_basekernel}.6+-ck${_ckpatchversion}"
If you do not enable CONFIG_NO_HZ, there will not be any changes.

graysky commented on 2012-08-12 10:16

@qpalz - Since so many users depend on the stability of this PKG, it would be irresponsible of me to use a non-CK-authored patch.

qpalz commented on 2012-08-12 05:26

bfq cannot be used in 3.5(.1). Compilation error.

qpalz commented on 2012-08-12 03:34

You may change ##"http://ck.kolivas.org/patches/3.0/3.4/${_basekernel}-ck${_ckpatchversion}/${_ckpatchname}.bz2" ->
"http://dl.dropbox.com/u/20066450/Linux/kernel/ck-patch/patch-${_basekernel}-ck${_ckpatchversion}" (but no guarantees)
and the _ckpatchversion is... 0.2 (version 1 should be made by ck himself)

graysky commented on 2012-08-11 15:00

I have updated the non-ck patched version for early adopters to reflect the changes in the PKGBUILD from the ARCH devs: http://repo-ck.com/PKG_source/linux-ck-pre/linux-ck-3.5.1-2.src.tar.gz

graysky commented on 2012-08-09 23:26

Bump to v3.4.8-1

Changelog: http://lkml.org/lkml/2012/8/9/544
Commit: pkgbuild.com/git/aur-mirror.git is down

Should also mention that 3.5.1 has been released; I have updated the non-ck patched version for early adopters: http://repo-ck.com/PKG_source/linux-ck-pre/linux-ck-3.5.1-1.src.tar.gz

Anonymous comment on 2012-08-09 23:14

Changelog: http://lkml.org/lkml/2012/8/9/544

graysky commented on 2012-08-09 20:51

Bump to v3.4.8-1

Changelog: Can't find one on lkml.org
Commit: pkgbuild.com/git/aur-mirror.git is down

fincan commented on 2012-08-09 17:21

3.4.8

qpalz commented on 2012-08-07 12:34

After reboot, everything works great.

qpalz commented on 2012-08-07 12:09

I am using the kernel which has been modified by me now. Although it can be run, it seems to be unstable. At first, everything makes responds slowly. Something even hang but I cannnot stop it.

When I was typing this word, the whole kde hangs. Luckily, Chrome saves the text automatically. I change to console, then restart kdm, until now, everything works well!

I am not sure what causes this problem. It may be the problem of the nvidia driver because a game cannot run well(After restarting the game, it can run well now). I am going to do some more checks and tests.

qpalz commented on 2012-08-07 12:07

I am using the kernel which has been modified by me now. Although it can be run, it seems to be unstable. Everything makes responds slowly. Something even hang but I cannnot stop it.

When I was typing this word, the whole kde hangs. Luckily, Chrome saves the text automatically. I change to console, then restart kdm, until now, everything works well!

I am not sure what causes this problem. It may be the problem of the nvidia driver because a game cannot run well(After restarting the game, it can run well now). I am going to do some more checks and tests.

qpalz commented on 2012-08-07 10:57

Well, after simply commenting the call of calc_load_enter_idle() and calc_load_exit_idle(), compilation has been completed with no errors. I do not know whether there will be bugs or not after doing so.
I guess that these two function is used for matching with the new core.c, but bfs.c does not have any changes. So, tick-sched.c need not be changed. Deleting the call of the functions MAY not cause problems.

I am afraid of trying it...

graysky commented on 2012-08-07 00:20

Bump to v3.4.7-2

Changelog: Provides array for headers package updated to provide 'linux-headers' which fixes latest virtualbox pains with dkms.
Commit:

graysky commented on 2012-08-06 06:44

@qpalz - Dunno, still waiting for CK to fix this.
@all - here is a prerelease of 3.5-1 which does NOT contain a ck1 patchset nor does it contain a BFS. This is an incomplete package but if some of you want to use v3.5 under the linux-ck name, go for it. Note that you need to enable [testing]. On my test system, the USB mouse didn't work without doing so.

http://repo-ck.com/PKG_source/linux-ck-pre/linux-ck-3.5-1.src.tar.gz

qpalz commented on 2012-08-06 04:42

For this error:
LD init/built-in.o
LD .tmp_vmlinux1
kernel/built-in.o: In function `tick_nohz_stop_sched_tick.isra.9':
tick-sched.c:(.text+0x49c29): undefined reference to `calc_load_enter_idle'
kernel/built-in.o: In function `tick_nohz_idle_exit':
(.text+0x49fa7): undefined reference to `calc_load_exit_idle'
make: *** [.tmp_vmlinux1] Error 1

In the ck patch, there are the following changes:
--- linux-3.4-ck3.orig/kernel/sched/Makefile 2012-03-20 17:39:43.000000000 +1100
+++ linux-3.4-ck3/kernel/sched/Makefile 2012-07-03 14:00:08.131680971 +1000
@@ -11,10 +11,14 @@ ifneq ($(CONFIG_SCHED_OMIT_FRAME_POINTER
CFLAGS_core.o := $(PROFILING) -fno-omit-frame-pointer
endif

+ifdef CONFIG_SCHED_BFS
+obj-y += bfs.o clock.o
+else
obj-y += core.o clock.o idle_task.o fair.o rt.o stop_task.o
-obj-$(CONFIG_SMP) += cpupri.o
obj-$(CONFIG_SCHED_AUTOGROUP) += auto_group.o
-obj-$(CONFIG_SCHEDSTATS) += stats.o
obj-$(CONFIG_SCHED_DEBUG) += debug.o
+endif
+obj-$(CONFIG_SMP) += cpupri.o
+obj-$(CONFIG_SCHEDSTATS) += stats.o


Index: linux-3.4-ck3/mm/vmscan.c

core.o will not be compiled when BFS is used, but calc_load_enter_idle is implemented in core.c. bfs.c does not contain calc_load_enter_idle. This may be the problem.
Is there any ideas to solve this problem?

graysky commented on 2012-07-31 18:57

Bump to v3.4.7-1

Changelog: https://lkml.org/lkml/2012/7/30/246
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=2410215f4d34258b9a5eb382d5eafd4df3743b68

skydrome commented on 2012-07-31 16:36

well this is ok since i just read there seems to be a huge power regression in 3.5
so perhaps sticking with 3.4.7 is better anyways

graysky commented on 2012-07-31 15:41

Update on ck1 for 3.5 from CK himself: http://ck-hack.blogspot.com/2012/07/bfs-and-ck-delays-for-linux-350.html

Anonymous comment on 2012-07-27 16:14

Well, I got 3.5 working, I didn't use localmod this time and went through manually, I still don't want to admit I missed something the first time though :)
Not worthy of a comment but oh well. Thanks for the update.

graysky commented on 2012-07-27 08:51

Bump to v3.4.6-2

Changelog: Updated version of BFQ. See upstream's changelog for details. http://algo.ing.unimo.it/people/paolo/disk_sched/patches/3.4.0-v4/README.BFQ
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=51d559ace7afbb06659f59ca7822162bf9cf8231

graysky commented on 2012-07-27 05:08

@sky - Saw it, thanks. Paolo also ported it into the 3.5 tree:
http://algo.ing.unimo.it/people/paolo/disk_sched/patches/3.5.0-v4/README.BFQ

Just need ck to port bfs/ck1 into 3.5 and we're golden. I'll update 3.4.6-1 --> 3.4.6-2 shortly.

skydrome commented on 2012-07-27 03:05

bfq bumped to 3.4.0-v4
minor changes http://algo.ing.unimo.it/people/paolo/disk_sched/patches/3.4.0-v4/README.BFQ

Anonymous comment on 2012-07-25 11:27

@Eithrial
This has already been covered.
Disable dynamic ticks, aka: # CONFIG_NO_HZ. See comments from 22 Jul.

@Graysky
Well I'm busy this week, but my plan was to try again later in the week, double and triple check through and try 3.5 again, failing that I was going to pull arch 3.5 from git, have a look through and build it. I have a feeling because I also used localmodconf I may have missed something.

Also, using systemd 186-2 had no booting issues for me.

graysky commented on 2012-07-25 11:24

@eithrial -

Comment by: graysky on Sun, 22 Jul 2012 12:22:02 +0000
Bump to v3.4.6-1

Changelog: We'll need to wait for ck to release a bfs that compiles against upstream with NO_HZ as an option. For now, I have removed it and set the clock to 300 MHz. If you are unhappy with this, do not update to 3.4.6 and stay with 3.4.5-1.
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=f48012d059d9c273e56b0cc66782e40042c98fdb

Anonymous comment on 2012-07-25 11:02

I can't compile 3.4.6-1. I got this error.

LD init/built-in.o
LD .tmp_vmlinux1
kernel/built-in.o: In function `tick_nohz_stop_sched_tick.isra.9':
tick-sched.c:(.text+0x49c29): undefined reference to `calc_load_enter_idle'
kernel/built-in.o: In function `tick_nohz_idle_exit':
(.text+0x49fa7): undefined reference to `calc_load_exit_idle'
make: *** [.tmp_vmlinux1] Error 1
==> ERROR:(B A failure occurred in build().(B
 Aborting...(B

felixonmars commented on 2012-07-25 06:42

I'm using systemd 187-2 and have no trouble at all on 3.5.0-0.

SirWuffleton commented on 2012-07-25 06:19

@graysky: Just enabled [testing] and -Syyu'd. After a reboot, both linux 3.5-1 and systemd 187-2 worked fine with no issues whatsoever. A full build of linux-ck 3.5-0 without localmodconfig is also working fine on my setup now that I've removed the btrfs_advanced hook.

graysky commented on 2012-07-25 05:42

@DarkSniper - can you boot the linux package in [testing]? Also, 187-2 is in [testing] as well. Might wanna enable it and -Syyu
@Psykorgasm - same question but does your stuff work there?

SirWuffleton commented on 2012-07-25 04:58

@misc: Yikes, thanks for that information. I definitely won't be looking forward to that moving into core if those issues aren't resolved by then.

It looks like the nonfunctional usb keyboard was due to a missing module (I used localmodconfig on my first build of it), and the early-userspace error was somehow the fault of btrfs_advanced, since specifying my root partition on the kernel's command line and disabling that hook resolved that problem. For now, systemd 186 seems to be working fine, seeing as I was able to successfully boot after the init issue was taken care of.

misc commented on 2012-07-24 19:30

Haven't been able to boot with systemd since 187 regardless of the used kernel. It will just stop to boot after apparently mounting my partitions once more (it repeats the same massage mentioning my HDD as often as there are Linux-relevant partitions — all ext4 — on it), without any further explanation. All that's possible then is to try to reboot, upon which readahead will complain about a read-only partition and some fsck are run, then it halts again.

IOW, right now I'm back to classic initscripts.

SirWuffleton commented on 2012-07-24 18:53

I was unable to boot 3.5.0-0, it claimed that /sbin/init did not exist, and kicked me to a recovery shell, and my USB keyboard did not work (but my laptop's built one did). Not sure if it could contribute to this, but I'm using systemd and have a btrfs root (via the btrfs_advanced hook).

misc commented on 2012-07-24 13:48

My USB mouse works, but now I have to enable a "Generic HID driver" in addition to the "Special HID" Logitech one. It's a laptop, so no idea how that affects USB keyboards.

BTW I couldn't successfully boot at the first time, now it's ok again.

Anonymous comment on 2012-07-23 23:40

Hi Graysky, I gave it a test myself - Does my USB mouse work? No idea because I couldn't get in to X because my USB keyboard didn't function at all, lol.
Do you have USB keyboard? I'm sure I didn't overlook anything when configuring, never experienced this before. Back to 3.4.5 and another fresh config and all is well.

graysky commented on 2012-07-23 21:59

@misc - http://repo-ck.com/PKG_source/linux-ck-pre/linux-ck-3.5-0.src.tar.gz

Works fine on my machine expect that my usb mouse does not work. If you build, please report back. Does your usb mouse work? Remember that this -0 release has no BFS nor does it have a BFQ patchset. It is just the 3.5 kernel in the linux-ck formatted PKGBUILD.

graysky commented on 2012-07-23 21:58

@misc - http://repo-ck.com/PKG_source/linux-ck-pre/linux-ck-3.5-0.src.tar.gz

Works fine on my machine expect that my usb mouse does not work. If you build, please report back. Does your usb mouse work?

misc commented on 2012-07-23 15:06

Thanks!

graysky commented on 2012-07-23 13:37

Yeah, I will spin one up later today and host it on repo-ck.com for you.

misc commented on 2012-07-23 13:20

graysky, do you provide preliminary 3.5-0 build scripts?

Because I'd very much prefer to run a "ck-less" 3.5 linux-ck for a few days which still was nonetheless compiled by myself (esp. with makenconfig & localmodcfg) than the default Arch kernel or 3.4.6-ck, and such -0 should only differ by a few lines from -1 anyway.

hermes14 commented on 2012-07-23 06:16

Ok, I'm going to disable dynamic ticks for now, and wait for 3.5-ck to be released.
Thanks

graysky commented on 2012-07-22 20:12

@hermes14- you cannot enable dynamic ticks with 3.4.6 due to an incompatibility in bfs v0.424 and linux v3.4.6. CK will very likely have this fixed in the bfs version that will patch against 3.5 which is just been mainlined today. Did you modify this option in your nconfig? Test by:

$ grep src/linux-3.4/.config
# CONFIG_NO_HZ is not set

hermes14 commented on 2012-07-22 19:04

I get this error:

LD vmlinux.o
MODPOST vmlinux.o
GEN .version
CHK include/generated/compile.h
UPD include/generated/compile.h
CC init/version.o
LD init/built-in.o
LD .tmp_vmlinux1
kernel/built-in.o: In function `tick_nohz_stop_sched_tick.isra.9':
tick-sched.c:(.text+0x49c49): undefined reference to `calc_load_enter_idle'
kernel/built-in.o: In function `tick_nohz_idle_exit':
(.text+0x49fc7): undefined reference to `calc_load_exit_idle'
make: *** [.tmp_vmlinux1] Errore 1

Any idea?

fincan commented on 2012-07-22 14:25

btw 3.5 is out and waiting for patches :)

graysky commented on 2012-07-22 12:22

Bump to v3.4.6-1

Changelog: We'll need to wait for ck to release a bfs that compiles against upstream with NO_HZ as an option. For now, I have removed it and set the clock to 300 MHz. If you are unhappy with this, do not update to 3.4.6 and stay with 3.4.5-1.
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=f48012d059d9c273e56b0cc66782e40042c98fdb

graysky commented on 2012-07-21 02:48

Patches welcome :)

fackamato commented on 2012-07-21 00:59

How about adding [[ -e ...dvb/stuff ]] && cp .. to the DVB header section? I don't use DVB, so disable all of it in the config. I have to edit the PKGBUILD to skip the copy of these otherwise the build will fail, because the files are not to be found.

Keep up the good work!

graysky commented on 2012-07-19 21:58

Sorry gang - looks like there is a problem compiling ck patched kernel against 3.4.6 so I am keeping 3.4.5-1 until this is resolved.

http://ck-hack.blogspot.com/2012/07/bfs-424-linux-34-ck3.html?showComment=1342735027935#c472134017088459259

fincan commented on 2012-07-19 19:44

3.4.6

graysky commented on 2012-07-16 22:27

Bump to v3.4.5-1

Changelog: https://lkml.org/lkml/2012/7/16/332
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=0de4792fba28d3c5110268f649e08d332acb9303

fincan commented on 2012-07-16 19:58

3.4.5

graysky commented on 2012-07-14 14:12

Bump to v3.4.4-5

Notes: Read http://www.archlinux.org/news/the-lib-directory-becomes-a-symlink and never force an update.
Special note: You CANNOT build 3.4.4-5 unless you first update your system. To update your system you need to first remove linux-ck 3.4.4-4 and ANY other ck packages, then update, then rebuild, and finally reboot. If this sounds too risky for you, simply install the ARCH kernel package and boot into it before removing the ck stuff.
Changelog: Package compliant with new Arch standards of lib --> usr/lib
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=3972cf7a45e931e9e03ac84665d4a536c455004d

graysky commented on 2012-07-04 22:38

@ghen - Yep... dunno why the Arch devs didn't add that to the ARCH release.

ghen commented on 2012-07-04 20:42

Just make it depend explicitly on kmod>=9-2 to avoid update mistakes - the update will be refused for people not on [testing].

graysky commented on 2012-07-04 14:50

Thanks, misc. Actually made one on my own and verified that I can build it in a clean chroot with [testing] enabled. Here is the src package for those interested:

http://repo-ck.com/PKG_source/linux-ck/linux-ck-3.4.4-5.src.tar.gz

YOU MUST HAVE [testing] ENABLED TO USE THIS!

misc commented on 2012-07-04 14:31

Meanwhile, here's the diff; running its result ATM:

http://pastie.org/4199135

graysky commented on 2012-07-04 11:16

Thnx for the link. This is bigger than an upstream patch... I will need to closely follow the timing of the official ARCH packages from [testing] to [core] to know when to upload the modified PKGBUILD here and when to rebuild the repo packages.

skydrome commented on 2012-07-04 09:17

https://mailman.archlinux.org/pipermail/arch-dev-public/2012-July/023178.html

graysky commented on 2012-07-04 09:12

Thanks for the heads up, guys. I made the modifications to the PKGBUILD per the ARCH 3.4.4-3 changes that skydrome posted, but find that my build blows up. I need some sleep now because my mind is crap. While I'm snoozing, does anyone understand the source of the error after making the changes in the PKGBUILD? $pkgdir/lib is moved to $pkgdir/usr just fine but the depmod fails with this error:

+depmod -b /scratch/linux-ck/pkg -F System.map 3.4.4-5-ck
ERROR: could not open directory /scratch/linux-ck/pkg/lib/modules/3.4.4-5-ck: No such file or directory
FATAL: could not search modules: No such file or directory

Why doesn't depmod know to look in /scratch/linux-ck/pkg/usr/lib/modules/3.4.4-5-ck rather than the old path?

EDIT: I cannot build the ARCH package via ABS (from [testing]) either; ir errors out in the same exact spot

INSTALL /scratch/linux/pkg/linux/lib/firmware/yamaha/ds1e_ctrl.fw
DEPMOD 3.4.4-3-ARCH
ERROR: could not open directory /scratch/linux/pkg/linux/lib/modules/3.4.4-3-ARCH: No such file or directory
FATAL: could not search modules: No such file or directory
==> ERROR: A failure occurred in package_linux().

... Is this because I am NOT using [testing]?

graysky commented on 2012-07-04 09:10

Thanks for the heads up, guys. I made the modifications to the PKGBUILD per the ARCH 3.4.4-3 changes that skydrome posted, but find that my build blows up. I need some sleep now because my mind is crap. While I'm snoozing, does anyone understand the source of the error after making the changes in the PKGBUILD? $pkgdir/lib is moved to $pkgdir/usr just fine but the depmod fails with this error:

+depmod -b /scratch/linux-ck/pkg -F System.map 3.4.4-5-ck
ERROR: could not open directory /scratch/linux-ck/pkg/lib/modules/3.4.4-5-ck: No such file or directory
FATAL: could not search modules: No such file or directory

Why doesn't depmod know to look in /scratch/linux-ck/pkg/usr/lib/modules/3.4.4-5-ck rather than the old path? Is this because I am NOT using [testing]?

graysky commented on 2012-07-04 09:09

Thanks for the heads up, guys. I made the modifications to the PKGBUILD per the ARCH 3.4.4-3 changes that skydrome posted, but find that my build blows up. I need some sleep now because my mind is crap. While I'm snoozing, does anyone understand the source of the error after making the changes in the PKGBUILD? $pkgdir/lib is moved to $pkgdir/usr just fine but the depmod fails with this error:

+depmod -b /scratch/linux-ck/pkg -F System.map 3.4.4-5-ck
ERROR: could not open directory /scratch/linux-ck/pkg/lib/modules/3.4.4-5-ck: No such file or directory
FATAL: could not search modules: No such file or directory

Why doesn't depmod know to look in /scratch/linux-ck/pkg/usr/lib/modules/3.4.4-5-ck rather than the old path?

skydrome commented on 2012-07-04 08:34

yes, the patch to the -ARCH kernel is what needs to be applied if you are using testing right now
https://projects.archlinux.org/svntogit/packages.git/commit/trunk?h=packages/linux&id=392ebae775235e2a8c0e6a738f1163dc93f6959d

apply that, update system, build kernel, reboot, happy

felixonmars commented on 2012-07-04 08:31

Maybe we need some changes to support arch's removal of /lib?
Refer to [arch-dev-public], a snapshot can be found here: http://permalink.gmane.org/gmane.linux.arch.tur.user/25500

xzy3186 commented on 2012-07-04 00:01

failed to build on my laptop (for months it was working fine except today)
errors are:

arch/x86/include/generated/asm/syscalls_32.h:327:1: error: expected ‘]’ before ‘:’ token
make[1]: *** [arch/x86/kernel/asm-offsets.s] Error 1
make: *** [prepare0] Error 2
==> ERROR: A failure occurred in build().
Aborting...

look like there is a typo in the header file. Any idea?

graysky commented on 2012-07-03 19:10

Bump to v3.4.4-4

Changelog: Addition of ck3 patchset including bfs v0.424. See: http://ck-hack.blogspot.com/2012/07/bfs-424-linux-34-ck3.html
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=59d348e45d3fbbe9f718e4d7c3799d2302ef89a0

graysky commented on 2012-07-03 16:35

Yeah, I will update soon.

skydrome commented on 2012-07-03 15:54

maybe he was referring to http://ck.kolivas.org/patches/3.0/3.4/3.4-ck3 ?

Anonymous comment on 2012-07-03 12:00

Not sure why this was flagged out of date, BFS 424 is listed right up there in the Sources list.

fincan commented on 2012-07-03 08:58

BFS 424 officially out

fincan commented on 2012-07-03 08:13

BFS 424 officially out

graysky commented on 2012-07-02 01:17

Bump to v3.4.4-3

Changelog: OPTIONAL setting to enable experimental BFS v0.424. Edit the PKGBUILD and enable the corresponding option in the header as you would for other options. Note that since this is experimental at this point, I will not be building it for the repo packages. Once CK has deemed it stable, I will. For more on the new BFS, see CK's blog: http://ck-hack.blogspot.com/2012/07/bfs-424-test.html

Summary: 3.4.4-3 and 3.4.4-2 are IDENTICAL unless you enable the experimental BFS prior to building!

Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=6851d92e49601ddb7b10dbbd25526e2468a10da2

graysky commented on 2012-07-01 20:20

Bump to v3.4.4-3

Changelog: OPTIONAL setting to enable experimental BFS v0.424. Edit the PKGBUILD and enable the corresponding option in the header as you would for other options. Note that since this is experimental at this point, I will not be building it for the repo packages. Once CK has deemed it stable, I will. For more on the new BFS, see CK's blog: http://ck-hack.blogspot.com/2012/07/bfs-424-test.html

Summary: 3.4.4-3 and 3.4.4-2 are IDENTICAL unless you enable the experimental BFS prior to building!

graysky commented on 2012-07-01 20:19

Bump to v3.4.4-3

Changelog: OPTIONAL setting to enable experimental BFS v0.424. Edit the PKGBUILD and enable the corresponding option in the header as you would for other options. Note that since this is experimental at this point, I will not be building it for the repo packages. Once CK has deemed it stable, I will. For more on the new BFS, see CK's blog: http://ck-hack.blogspot.com/2012/07/bfs-424-test.html

blackout24 commented on 2012-07-01 13:05

Worked like a charm.

graysky commented on 2012-06-28 20:24

@oblique - Nope. See ck's comment about this on his blog: http://ck-hack.blogspot.com/2012/06/upgradeable-rwlocks-and-bfs.html

oblique commented on 2012-06-28 18:09

can you also add RIFS-ES scheduler? https://lkml.org/lkml/2012/6/16/123

graysky commented on 2012-06-28 09:03

Sorry dude, I haven't messed with that in a long time. Consult google or the author?

Anonymous comment on 2012-06-28 02:25

Actually, I just edited your PKGBUILD to use `make localyesconfig` before `make nconfig`, and the options are still marked as modules. Any ideas? Or is this a question for Steven Rostedt?

Anonymous comment on 2012-06-27 14:19

It looks like you ran into similar issues about two years ago! https://bbs.archlinux.org/viewtopic.php?pid=845113

I'm not sure if the script has changed at all, but it looks like it's behaving as intended for the most part. The logic could be improved, but not without breaking some user's configs.

I don't think it says it's supposed to mark all current modules to be compiled in, so that would still have to be done manually.

graysky commented on 2012-06-26 20:36

@Klink - You are not misunderstanding. You'll need to email the script's author Steven Rostedt to ask him why it is behaving this way.

Anonymous comment on 2012-06-26 14:13

graysky, I'm a little confused by some of the options in the PKGBUILD.
If I use _localmodcfg="y" , it loads all of the modules I'd ever need from modprobed_db (love that tool, by the way. Thanks for writing it!)

But when combined with _makenconfig="y", I notice that things that ought to be compiled in are marked as modules, and certain areas that wouldn't need anything enabled (modprobed_db has figured out all of the sound modules I need-- no need for any others) still have everything enabled.

Are these options incompatible, or am I just misunderstanding the point of localmodcfg? Does it just make sure to mark all of the modules I need, and nothing else?

graysky commented on 2012-06-25 22:04

Bump to v3.4.4-2

Changelog: fix backlight and nouveau
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=e8323533a643f40d50116d27046e91051fa22b59

graysky commented on 2012-06-23 11:30

Bump to v3.4.4-1 (12 h ago)

Changelog: https://lkml.org/lkml/2012/6/22/359
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=b6959328be9eff864b964545ceb52fe3f974533f

graysky commented on 2012-06-23 11:19

@jim - http://repo-ck.com/PKG_source/

jim1960 commented on 2012-06-23 10:21

I need linux-ck 3.3.8, where can I found it? Thanks!

graysky commented on 2012-06-21 06:00

Bump to v3.4.3-1

Changelog: Followed ARCH fixes for 3.4.2-rpc_pipefs.patch and linux-ck.install PATH fix as well as updated to latest upstream stable patchset (3.4.3) whose change log is here, https://lkml.org/lkml/2012/6/17/96
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=e2fcf21d19b1d24fb6b8bfec619d18d78cc5e124

graysky commented on 2012-06-18 22:42

Thanks for taking care of the community while my 'net connection is down, misc.

misc commented on 2012-06-18 19:46

Until the proper update and for those interested, here's the diff from the current tarball to build "3.4.3-0", incorporating the only change from testing's linux:

https://pastebin.com/raw.php?i=dTNLxcnr

Just "patch -i [what you saved it as]" in the folder of the extracted tarball. patch might give a harmless warning since pastebin ate the trailing empty line.

misc commented on 2012-06-18 19:45

Until the proper update and for those interested, here's the diff from the current tarball to build "3.4.3-0", incorporating the only read change from testing's linux:

https://pastebin.com/raw.php?i=dTNLxcnr

Just "patch -i [what you saved it as]" in the folder of the extracted tarball. patch might give a harmless warning since pastebin ate the trailing empty line.

alucryd commented on 2012-06-18 16:21

@graysky: It seems inverting the two names will also invert the order in which the package functions are processed. However, the headers function depends on variables created in the linux-ck function, so for this to work, you need to copy the few lines that define those variables from header to linux-ck functions, this is the modified PKGBUILD I used: http://pastebin.com/2zTbr0kh. It should be possible to remove the lines from the linux-ck function rather than having them in both, but I haven't tested.

graysky commented on 2012-06-17 21:31

Sorry guys, ISP needs to rerun a line from the street to my cable box, so no updates until this happens. Doubtful until later in the week.

alucryd commented on 2012-06-15 14:29

@graysky: Thx for considering it, and yes right now I'm doing it manually, have to change some "n" to "y" anyway^^

graysky commented on 2012-06-14 11:14

@Alucryd - should be no problem but let me verify that it doesn't jack up my repo build script before I push it. For now, can manually make that switch and have a happy system?

alucryd commented on 2012-06-14 07:46

@skydrome: yeah you would have to make a single package() with everything that's under the two split package functions, it's easier to just switch the two package names since it would behave exactly the same, except the headers would be installed first.

skydrome commented on 2012-06-14 04:15

scratch that, that wont work without some modifications to the pkgbuild

skydrome commented on 2012-06-14 02:24

@alucryd, you could also just comment that line out and the package wont be split, the headers will be included into just the single package.

alucryd commented on 2012-06-13 09:31

@graysky: Since I'm installing the official kernel from the repo (just to have a functional kernel in case I screw up), I'm not having this problem, linux-headers is in my SyncFirst array and will install before linux, allowing the hooks to build their modules properly.
I did some tests today, switching the two package functions will not have the desired effect. BUT turns out changing "true && pkgname=(linux-ck linux-ck-headers)" to "true && pkgname=(linux-ck-headers linux-ck)" will, plus it doesn't involve modifying the order the packages are built. Could you make this small change?

graysky commented on 2012-06-13 02:05

Bump to v3.4.2-3

Changelog: Added ARCH pactch to fix potential issue with nfs4.

graysky commented on 2012-06-12 19:51

@Alucryd - I modeled the PKGBUILD after the official ARCH PKGBUILD and am reluctant to change it. Do the hooks work with the ARCH kernel?

alucryd commented on 2012-06-12 08:33

Hi graysky, would it be possible to put package_linux-ck-headers before package_linux-ck in the PKGBUILD so the headers are installed first? Catalyst and vbox hooks are useless the way linux-ck is. Thx in advance!

graysky commented on 2012-06-11 22:16

Bump to v3.4.2-2

Changelog: Updated with ck2 including bfs v0.423 and a new option to enabled an experimental upgradeable rwlocks feature that CK has released as beta. For more info on this see his blog: http://ck-hack.blogspot.com/2012/06/upgradeable-rwlocks-and-bfs.html

Note: Repo users will not get this feature enabled since it is considered experimental.

Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=44af1fdb54e890f75902c99552e29b9fcda68d55

graysky commented on 2012-06-11 21:36

Bump to v3.4.2-2

Changelog: Updated with ck2 including bfs v0.423 and a new option to enabled an experimental upgradeable rwlocks feature that CK has released as beta. For more info on this see his blog: http://ck-hack.blogspot.com/2012/06/upgradeable-rwlocks-and-bfs.html

Note: Repo users will not get this feature enabled since it is considered experimental.

graysky commented on 2012-06-11 21:29

Bump to v3.2.4-2

Changelog: Updated with ck2 including bfs v0.423 and a new option to enabled an experimental upgradeable rwlocks feature that CK has released as beta. For more info on this see his blog: http://ck-hack.blogspot.com/2012/06/upgradeable-rwlocks-and-bfs.html

Note: Repo users will not get this feature enabled since it is considered experimental.

graysky commented on 2012-06-11 13:56

Thanks... I get it now. Will update in a few hours.

dkaylor commented on 2012-06-11 13:29

Guardiano might have been reffering to this:

http://ck-hack.blogspot.com/2012/06/bfs-0423-34-ck2.html

Would've been nice for him to specify!

graysky commented on 2012-06-11 10:01

Your package linux-ck has been flagged out of date by Guardiano [1].

Guys, please post some text indicating to me why this is out of date?
1) 3.4.2 is current per lkml
2) I see no modifications from the Arch debs beyond what is already in 3.4.2-1

graysky commented on 2012-06-11 09:11

@berz-that script is based on whatever you have probed at this time you run it. I believe the code in my PKGBUILD is fine.
@AsAck-Why did you flag out of date? 3.4.2 is current per lkml.

berz commented on 2012-06-10 19:03

some options seem to be missing in .config when using localmodcfg

==> If you have modprobe_db installed, running it in recall mode now
Attempting to modprobe entire database...
Done!
==> Running Steven Rostedt's make localmodconfig now
using config: '.config'
[...]
Yama support (SECURITY_YAMA) [N/y/?] n
Integrity Measurement Architecture(IMA) (IMA) [N/y/?] n
EVM support (EVM) [N/y/?] (NEW) aborted!

Console input/output is redirected. Run 'make oldconfig' to update configuration.

==> Running make bzImage and modules
scripts/kconfig/conf --silentoldconfig Kconfig
[...]
Yama support (SECURITY_YAMA) [N/y/?] n
Integrity Measurement Architecture(IMA) (IMA) [N/y/?] n
EVM support (EVM) [N/y/?] (NEW) aborted!

Console input/output is redirected. Run 'make oldconfig' to update configuration.

make[2]: *** [silentoldconfig] Error 1
make[1]: *** [silentoldconfig] Error 2
make: *** No rule to make target `include/config/auto.conf', needed by `include/config/kernel.release'. Stop.
make: *** Waiting for unfinished jobs....
make[1]: Nothing to be done for `all'.
make[1]: Nothing to be done for `relocs'.
==> ERROR: A failure occurred in build().
Aborting...

graysky commented on 2012-06-09 20:02

Bump to v3.4.2-1

Changelog: https://lkml.org/lkml/2012/6/9/105

kyak commented on 2012-06-08 15:33

@DarkSniper thanks, i was looking at linux-ck wiki page, didn't guess to look at repo-ck page.

SirWuffleton commented on 2012-06-08 07:43

@kyak: If you haven't found out by now, it's on the wiki page for repo-ck (pretty much your standard pacman-key import and locally sign/trust procedure): https://wiki.archlinux.org/index.php/Repo-ck#Setup

kyak commented on 2012-06-06 16:36

So far i have to use it like this:

[repo-ck]
Server = http://repo-ck.com/$arch
SigLevel = Never

kyak commented on 2012-06-06 16:33

@graysky: could you guide on how to import your repo-ck.com PGP key?

i get:

error: linux-ck: signature from "graysky <graysky@archlinux.us>" is unknown trust

Thank you!

graysky commented on 2012-06-05 16:02

Bump to v3.4.1-2

Changelog: The ARCH devs made a minor modification to the kernel configs enabling an option for iSCSI Boot Firmware Table Attributes. This is the only change between 3.4.1-1 and 3.4.1-2.
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=964dd058cb99330f21f14b1686700861a5dbbba5

graysky commented on 2012-06-04 23:10

Bump to v3.4.1-1

Changelog: https://lkml.org/lkml/2012/6/4/400
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=bef74cc8abb1b38cf65e2ace20ab3a1cce652336

graysky commented on 2012-06-04 01:25

Read it but the PKG defaults does not require this patch.

fackamato commented on 2012-06-04 01:00

Hey, you might want bfs422-nohotplug_fix.patch from http://ck.kolivas.org/patches/bfs/3.4.0/bfs422-nohotplug_fix.patch (see the blog post).

graysky commented on 2012-06-02 12:28

Bump to v3.4.0-1

Changelog: http://kernelnewbies.org/Linux_3.4
Changelog: http://ck-hack.blogspot.com/2012/06/bfs-0422-34-ck1.html
Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=519b31441471005ad539c89f0b76e098511c0ff0

fincan commented on 2012-06-02 05:48

ck patches are ready...

graysky commented on 2012-05-30 18:31

Thx... Now we just need bfs.

skydrome commented on 2012-05-30 17:36

http://algo.ing.unimo.it/people/paolo/disk_sched/patches/3.4.0-v3r4

Anonymous comment on 2012-05-27 15:58

Yup, this patch works fine for me.
Thanks for taking this forward - that is very nice of you.

graysky commented on 2012-05-26 15:50

Also - https://bugs.archlinux.org/task/30028

graysky commented on 2012-05-26 15:34

I actually do not like DKMS... can you verify that the patch you supplied produces the desired results? I am glad to add it if so (code is dependent on that pkg being installed).


Anonymous comment on 2012-05-26 13:00

Can you kindly include the below patch to linux-ck.install so that DKMS autoinstall service runs after each kernel install? I prefer this method rather than invoking DKMS service on each system boot.

--- linux-ck.install.old 2012-05-26 18:23:07.240130815 +0530
+++ linux-ck.install.new 2012-05-26 18:26:36.502330478 +0530
@@ -8,6 +8,13 @@
echo ">>> Generating initial ramdisk, using mkinitcpio. Please wait..."
mkinitcpio -p linux${KERNEL_NAME}

+ pacman -Q dkms &>/dev/null
+ hasdkms=$?
+ if [ $hasdkms -eq 0 ]; then
+ echo ">>> Running the DKMS autoinstall service..."
+ dkms autoinstall -k ${KERNEL_VERSION}
+ fi
+
# compat symlinks for the official kernels only
if [ -z "${KERNEL_NAME}" -o "${KERNEL_NAME}" = "-lts" ]; then
loaders="$(find /boot -name syslinux.cfg -or -name extlinux.conf -or -name grub.cfg -or -name menu.lst)"
@@ -53,6 +60,13 @@
depmod ${KERNEL_VERSION}
echo ">>> Generating initial ramdisk, using mkinitcpio. Please wait..."
mkinitcpio -p linux${KERNEL_NAME}
+
+ pacman -Q dkms &>/dev/null
+ hasdkms=$?
+ if [ $hasdkms -eq 0 ]; then
+ echo ">>> Running the DKMS autoinstall service..."
+ dkms autoinstall -k ${KERNEL_VERSION}
+ fi
}

post_remove() {

Anonymous comment on 2012-05-23 02:36

@felixonmars
@ValdikSS

I think uksm is now official supported here: https://aur.archlinux.org/packages.php?ID=58802

And I've seen this interface on the user manual the first day I installed uksm, I am actually thinking this idea(remote debugging) is cool. Every other APPs have this feature nowadays, I just hate to dump my all messy stuff(i.e. /proc info, dmesg, core dumps....etc) if I see a bug and told to do so ---- that sounds more like they are debugging me! :P

felixonmars commented on 2012-05-23 02:19

@ValdikSS
Hi, I e-mailed author of uksm and got his reply. This is documented but not that easy to find :P

“usr_spt_enabled: Testing interface for future feature of “remote bug reporting”. Currently this interface is enabled by default for testing —- this will make uksmd send a meaningless udp message from your machine every 2 mintues. You can disable it by echo 0 > usr_spt_enabled.
(via:http://kerneldedup.org/en/projects/uksm/uksmdoc/usage/).”

And he stated they will split these codes so it won't be too confused.

ValdikSS commented on 2012-05-22 17:29

Oh-oh-oh, what's this in uksm?

+ uksm_usr_spt_addr.sin_family = AF_INET;
+ uksm_usr_spt_addr.sin_port = htons(UKSM_USR_SPT_PORT);
+ uksm_usr_spt_addr.sin_addr.s_addr = in_aton("114.212.190.16");
+
+ uksm_usr_spt_iov.iov_base = &uksm_usr_spt;
+ uksm_usr_spt_iov.iov_len = sizeof(uksm_usr_spt);
+ uksm_usr_spt_msg.msg_name = (struct sockaddr *)&uksm_usr_spt_addr;
+ uksm_usr_spt_msg.msg_iov = &uksm_usr_spt_iov;
+ uksm_usr_spt_msg.msg_iovlen = 1;
+ uksm_usr_spt_msg.msg_control = NULL;
+ uksm_usr_spt_msg.msg_controllen = 0;
+ uksm_usr_spt_msg.msg_namelen = sizeof(uksm_usr_spt_addr);
+ uksm_usr_spt_msg.msg_flags = MSG_DONTWAIT;
+
+ sock_sendmsg(uksm_usr_spt_sock, &uksm_usr_spt_msg, sizeof(uksm_usr_spt));
+ uksm_usr_spt_last = jiffies;

ValdikSS commented on 2012-05-22 17:20

Oh-oh-oh, what's this in uksm?

+ uksm_usr_spt_addr.sin_family = AF_INET;
+ uksm_usr_spt_addr.sin_port = htons(UKSM_USR_SPT_PORT);
+ uksm_usr_spt_addr.sin_addr.s_addr = in_aton("114.212.190.16");
+
+ uksm_usr_spt_iov.iov_base = &uksm_usr_spt;
+ uksm_usr_spt_iov.iov_len = sizeof(uksm_usr_spt);
+ uksm_usr_spt_msg.msg_name = (struct sockaddr *)&uksm_usr_spt_addr;
+ uksm_usr_spt_msg.msg_iov = &uksm_usr_spt_iov;
+ uksm_usr_spt_msg.msg_iovlen = 1;
+ uksm_usr_spt_msg.msg_control = NULL;
+ uksm_usr_spt_msg.msg_controllen = 0;
+ uksm_usr_spt_msg.msg_namelen = sizeof(uksm_usr_spt_addr);
+ uksm_usr_spt_msg.msg_flags = MSG_DONTWAIT;
+
+ sock_sendmsg(uksm_usr_spt_sock, &uksm_usr_spt_msg, sizeof(uksm_usr_spt));
+ uksm_usr_spt_last = jiffies;

graysky commented on 2012-05-22 11:51

ghen - why out of date?

graysky commented on 2012-05-22 01:30

Sorry guys, a single character typo caused mkinitcpio NOT to run. I have corrected it and re-uploaded 3.3.7-1 without bumping the pkgver. Sorry for the inconvenience.

graysky commented on 2012-05-21 23:05

Bump to v3.3.7-1
Changelog: https://lkml.org/lkml/2012/5/21/395
PKG Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=eb049a0cbcb676039312ec4ffbba61840711c341

As to the release of v3.4.0-1 ... CK emailed me to say that he is swamped with his personal life at the moment so it is unclear when we can expect a ck1 patchset for the 3.4 tree. As usual, please use the out-of-date button when CK pushes something.

graysky commented on 2012-05-21 22:44

Of course not, Latjay. Updating and testing. Should post here soon.

lukpod commented on 2012-05-21 22:26

@graysky

3.3.7 released.

I hope you won't mind if this package is flagged out of date when stable patches are released.

lukpod commented on 2012-05-21 22:23

@graysky

3.3.7 released.

I hope you won't mind if this package will be flagged out of date when stable patches are released.

graysky commented on 2012-05-21 22:06

Thanks for the enthusiasm, but package is not out-of-date just yet...
1) No BFS
2) No ck1
3) No BFQ

:p

As usual, I will provide an initial release of source once we have something from CK. In case I miss it, please use the out-of-date button when CK pushes something.

graysky commented on 2012-05-21 22:05

Thanks for the enthusiasm, but package is not out-of-date just yet...
1) No BFS
2) No ck1
3) No BFQ

:p

misc commented on 2012-05-21 16:20

@graysky: As with 3.3, please provide a preliminary tarball to build 3.4 (ie, "pkgrel=0").

fabertawe commented on 2012-05-16 10:29

[Edit: I've deleted the build failure post.] Please accept my apologies for wasting your time. From sorting out another failed build the problem was an empty file named 'touch' which had been created in $HOME/bin. Very odd.

fabertawe commented on 2012-05-16 09:51

Hi... trying to build v3.3.6-1 (from 3.2.14-2-ck, x86_64) but the build is failing without any explanation. The last few lines...
CC [M] drivers/staging/zram/zram_sysfs.o
LD drivers/staging/zram/built-in.o
CC [M] drivers/staging/vt6656/int.o
LD [M] drivers/staging/xgifb/xgifb.o
LD [M] drivers/staging/zram/zram.o
LD [M] drivers/staging/vt6656/vt6656_stage.o
LD drivers/staging/built-in.o
LD drivers/built-in.o
==> ERROR: A failure occurred in build().
Aborting...

Never had a fail before, any ideas? Thanks.

graysky commented on 2012-05-14 21:09

Updated v3.3.6-1

Changelog: Removed optional uksm since there is already a linux-ck-uksm package which is maintained by a member of the uksm dev team. No point in me duplicating this in linux-ck.
PKG Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=c2412cde953b772b20fee22ea84a2a62e24d2916

Lawliet commented on 2012-05-14 02:18

@graysk:

not different from the linux-ck, but the linux-uksm-ck package main major is UKSM patch, and defalut enable.
we are UKSM dev and support team, so the package unlike linux-ck, it is BFS main major, not UKSM.



and UKSM have new patch "0002"
the fix libsandbox crash with libc at Gentoo(maybe some case will happen with Arch..)
http://kerneldedup.org/download/uksm/0.1.1.1/patches/0002-UKSM-add-VM_MERGEABLE-to-vm_flags-before-vma_merge.patch

dkaylor commented on 2012-05-13 18:57

Actually, I think you've earned our trust in terms of the repo packages and the PKGBUILD on AUR, just to clarify. There's something about UKSM that makes me nervous....

dkaylor commented on 2012-05-13 18:54

@graksky - I get your point, and I trust your judgement in regards to the repo package. But I would urge real caution when it comes to UKSM, atleast for now.

Anonymous comment on 2012-05-13 17:53

FWIW I did try this for a week and noticed _NO_ improvement in mem usage but the 'uksm' process was always in the top 5 cpu usage, almost constantly running at 2% - 5%. Just my unscientific 2 pence.

graysky commented on 2012-05-13 17:29

No worries, the default is NOT compiled in as you can see in the PKGBUILD. As I have previously stated, the repo packages will NOT get this due to the former. Arch is about choice, so go ahead and enable and compile is you wish.

dkaylor commented on 2012-05-13 16:34

"I don't know enough about that patch to know if it is 100 % dormant without `echo 1 >/sys/kernel/mm/uksm/run` since the page and forums are in Chinese."

And therein lies a big problem. That, combined with the responses on LKML almost 15 months ago, combined with the lack of follow up by xianai here or on LKML since then makes me think it's not worth the maintenance hassle or the risk.
There are those, me formerly being one of them, who will assume "hey BFS and BFQ are performance related (in the sense of interactive responsiveness on a contemporary desktop) so UKSM must be also, therefore I will compile it in".
And if graysky himself can't really understand the interactions, due to no documentation in the dominant kernel development language, then I wouldn't touch it, not even as an option.

graysky commented on 2012-05-13 10:49

@kyak - I don't know enough about that patch to know if it is 100 % dormant without `echo 1 >/sys/kernel/mm/uksm/run` since the page and forums are in Chinese. I therefore do not want to force it on all the repo users. BFQ by comparison is 100 % dormant unless the user chooses to active it.

Can you point me to documentation?

graysky commented on 2012-05-13 10:48

@kyak - I don't know enough about that patch to know if it is 100 % dormant without `echo 1 >/sys/kernel/mm/uksm/run` since the page and forums are in Chinese. Can you point me to documentation?

kyak commented on 2012-05-13 06:50

@skydrome Thanks, but i don't have /sys/kernel/mm/uksm/run on the repo-ck kernel!

There is /sys/kernel/mm/ksm/run however, and it's disabled by default.

@graysky Can we do the same for uksm? (I.e. provide it, but disable by default for repo-ck users?)

graysky commented on 2012-05-12 20:01

Bump to v3.3.6-1

Changelog: https://lkml.org/lkml/2012/5/12/5
PKG Commit: http://pkgbuild.com/git/aur-mirror.git/commit/linux-ck?id=f3ecfad6ad86ded2888e09e0e152fa8b9d5da102

graysky commented on 2012-05-12 16:23

Updated v3.3.5-1

Changelog: Updated with an upstream patch for users of the optional uksm option.

graysky commented on 2012-05-11 07:37

Because the packager is sloppy... thanks for pointing that out/fixed.

karabaja4 commented on 2012-05-11 07:14

Um... why is UKSM patch enabled by default?

graysky commented on 2012-05-10 14:38

@qcts33 - No need. That flag will make the needed changes to the config for you:

sed -i -e '/CONFIG_DEFAULT_IOSCHED/ s,cfq,bfq,' -i -e s'/CONFIG_DEFAULT_CFQ=y/# CONFIG_DEFAULT_CFQ is not set\nCONFIG_DEFAULT_BFQ=y/' ./.config

You can verify which is your default with this command:

# cat /sys/block/sda/queue/scheduler
noop deadline cfq [bfq]

Anonymous comment on 2012-05-10 14:34

I want to know if I set _BFQ_enable_="y" and compile the package from source, do I need to add "elevator=bfq" in /etc/default/grub like they said in wiki?

graysky commented on 2012-05-09 19:08

@dkay - no idea on performance benefit. I don't think performance is the goal; more free memory is the goal as I understand it. I don't think the CPU scheduler matters.

skydrome commented on 2012-05-09 15:45

@ryak - Yes, echo [0|1] > /sys/kernel/mm/uksm/run
to turn off/on respectively

dkaylor commented on 2012-05-09 12:56

@graysky - regarding the current state of uksm patches, what's your take on any performance benefit at all, especially when compilied in with BFS?

@xianai - any thoughts on the above? Do you still have plans for mainlining, or are you resigned to the BFS/BFQ route?

graysky commented on 2012-05-09 09:20

No possible. Needs to be compiled into the kernel at this point as far as I can tell. If you want that code, you'll have to compile from the AUR.

kyak commented on 2012-05-09 08:04

Do you guys know if it is possible to activate uksm from userland (or at least from kernel command line)?

I'm asking because i'm using repo-ck kernel (so i can't enable uksm at build time) and uksm is definitely lacking documentation (in English, at least) - so i can't any information.

graysky commented on 2012-05-09 01:26

Thanks for the info. I updated 3.3.5-1 with the new uksm patch and again did not bump the PKGVER. Enjoy.
@Lawliet - what does the linux-ck-uksm package offer that the linux-ck package doesn't offer? They seem to be redundant, no? Would you be supportive to have the AUR devs remove it?

Lawliet commented on 2012-05-08 08:49

graysky:

Sorry, the URL is wrong.. Sorry, my fail...
right URL is:
http://kerneldedup.org/download/uksm/0.1.1.1/patches/uksm-0.1.1.1-for-v3.3.ge.5.patch

Lawliet commented on 2012-05-08 08:48

graysky:

Sorry, the URL is wrong.. Sorry, my fail...
right URL is:
kerneldedup.org/download/uksm/0.1.1.1/patches/uksm-0.1.1.1-for-v3.3.ge.5.patch

xianai commented on 2012-05-08 08:38

BTW,

The name uksm-0.1.1.1-for-v3.3.ge.5.patch means this uksm-0.1.1.1 patch is for upstream kernel v3.3.y where y is currently >= 5 ("ge" means "greater than or equal to").
e.g. We will file a new patch named uksm-0.1.1.1-for-v3.3.ge.10.patch in case there is a patch confilct from upstream v3.3.10.

xianai commented on 2012-05-08 08:38

BTW,