Package Base Details: linux-ck

Git Clone URL: https://aur.archlinux.org/linux-ck.git (read-only)
Submitter: graysky
Maintainer: graysky
Last Packager: graysky
Votes: 394
Popularity: 3.474389
First Submitted: 2011-07-22 14:51
Last Updated: 2018-12-08 17:22

Packages (2)

Det commented on 2011-08-28 10:21

@t3ddy, doesn't happen with the packages from his repository? Are we talking about a binary repo for this package? If that's the case, and you haven't modified the kernel config (or PKGBUILD or anything else), then there's only 1 possibility I can think of and that's the gcc flags of the kernel image.

So instead of:
make ${MAKEFLAGS} KCFLAGS="-march=native -O2 -pipe -fstack-protector --param=ssp-buffer-size=4 -D_FORTIFY_SOURCE=2" KCPPFLAGS="-march=native -O2 -pipe -fstack-protector --param=ssp-buffer-size=4 -D_FORTIFY_SOURCE=2" bzImage modules

You should try with:
make ${MAKEFLAGS} bzImage modules

And see what happens.

graysky commented on 2011-08-28 08:08

Post your grub.cfg to pastebin

t3ddy commented on 2011-08-28 07:46

I use grub2, it says "loading initial ramdisk" and then the pc reboot itself, although this doesn't happen with the packages from your repository (without making any change to grub.cfg)

graysky commented on 2011-08-27 18:13

@t3ddy - hmm... how far into the boot process does it get before rebooting? You using grub or grub2? Post the menu.lst or grub.cfg for your system to pastebin, then link it here. Or you can start a discussion thread on the bbs, just be sure you use the work "linux-ck" in the title so I see it :p

t3ddy commented on 2011-08-27 16:39

I've got a problem with this pkgbuild.
I can't understand why, but if I compile it, then on boot the pc reboot itself continuously (I think due to initramfs problems).
Although the packages from repository works well and with other kernels I've got no problem to initramfs.

Det commented on 2011-08-26 11:11

Well that explains it. But what about my changes? :)

You convinced yet?

graysky commented on 2011-08-26 07:35

I emailed Paolo about BFQ. Here is the exchange:

> See here: http://pf.natalenko.name/sources/3.0/
> The pf patchset includes BFQ for 3.0? Is it official?

Did not know it

> Did you release it yet?

Unfortunately no, I did not yet find the time to complete the new version. The bright side is that it will not be just a port to 3.0, but it will contain new improvements.

Det commented on 2011-08-25 11:50

Well, I don't exactly know _where_ they are defined like that. I just know that they are.

1) The "-O2" flag is told to be used by default in the "CONFIG_CC_OPTIMIZE_FOR_SIZE=y" config option's description:
"Enabling this option will pass "-Os" instead of "-O2" to gcc resulting in a smaller kernel."

2) "-fstack-protector" on the other hand is mentioned in the said bug report (https://bugs.archlinux.org/task/18864).

[1] could at least be checked by finding out where all those CONFIG_ options are defined and see what _this_ one does, while [2] I have no idea about. It's likely in the kernel sources since I don't see it being mentioned anywhere in the PKGBUILD or the configs.

You can test build without either flag and see for yourself that the size of the kernel image doesn't indeed change (it would be quicker to just test both at the same time). I can already confirm that this is the case.

graysky commented on 2011-08-25 01:36

@Det - missed some of that, sorry. I took your suggestion about the make modules_install, thanks for pointing it out. As to the the your other two points regarding switches, it's not immediately clear to me where the -O2 and -fstack-protector are defaulted in the linux package. Can you point me to a location?

Det commented on 2011-08-24 21:40

@graysky, these: http://pastebin.com/6jr7yPzS

>To echo what I posted earlier, if 2.6 --> 3.0 was that simple, I would think that Paolo would have done it already.
I wouldn't know about that but it's possible that Paolo simply hasn't got the time yet to even look at it (even if it was that simple a port). For what it's worth I'm using BFQ here (through the -pf patchset) and have had no issues so far.