Arch Linux

Good synchro is the key :)
I will update this package when 3.5.5 or 3.6.1 will be out.

XD that was fast!

patch-3.0.4.bz2 here:
http://ftp.ua.postgresql.org/pub/mirrors/kernel.org/linux/kernel/v3.0/patch-3.0.4.bz2
edit PKGBUILD.

patch-3.0.4.bz2 not found!

Merging into linux-one.

Find kernel-one 3.0.1 here: https://aur.archlinux.org/packages.php?ID=51552

I cannot guarantee it to work. It's neither tested against that hardware nor can I speak for hardware compatibility. According to the hardware details given on the Acer Aspire One page of the wiki the D225 features different network devices, both wired and wireless. This kernel however features only the r8169 and ath5k drivers. You may try if they work for you.

just wondering if this linux-one kernel is valid for a AAO d255.

There's no kernel3 package for arch yet. This package won't get an upgrade until 3.0 has been pushed to core to stay in line with other arch packages. So, don't expect an update. An btw, there's not even such a package in testing yet...

New kernel version is out.

Sorry for the delay in my reply. Very odd, I remember a similar error when first trying to build a 2.6.39 package but fixed it prior to release. Apparentely I forgot to fix it on the release version. I'm very sorry. Fixed on the -2 release.

However the -dev version shouldn't bear from that error. I'll double check, yet my version of it is compiling flawlessly.

I have problem install kernel26-one 2.6.39.1-1 currenly using kernel26-one 2.6.37.2-1
I got this error msg "make[1]: *** No rule to make target `/tmp/yaourt-tmp-root/aur-kernel26-one/pkg/lib/firmware/./', needed by `/tmp/yaourt-tmp-root/aur-kernel26-one/pkg/lib/firmware/ti_3410.fw' . Stop.
make[1]: *** Waiting for unfinished jobs....
MKDIR /tmp/yaourt-tmp-root/aur-kernel26-one/pkg/lib/firmware/ttusb-budget/
make: *** [_modinst_post] Error 2
same error with kernel26-one-dev

The 2.6.39 package might take a while. My build environment has become completely unstable and I have not managed a single build successfully without the system softresetting. Until I've been able to test a 2.6.39 version, I cannot release simply it (and maybe leave your systems unbootable). I apologize for the inconvenience.

Sorry, I have been rather busy with work. I will update the -dev package somewhen over the Easter holidays and this under the week, if the -dev package is working. Thanks for your patience guys.

Tbh, I had never gotten it working, so I can't say there's a change for me. I always had to have an SD card inserted in the right slot upon booting to be able use the slot.

At what kernel version did it work without it for you?

It works well, but I've one question: has there been any changes to the pciehp module? Later kernels doesn't seem to be able to hotplug SD cards.

Sorry for the long wait for an update. I don't have much time due to work atm. Therefore I couldn't figure why the power consumption from 2.6.36 to 2.6.37 again rose by 1 W. Thus I will stick with the former kernel version myself until I'll figure that out. I'll bump this and as I didn't get any negative feedback on the one-dev kernel I guess it should work for those who want the upgrade.

Sadly the wiki page is both outdated and has become a huge bloat. The page is often contradictory, so even when it is correct you might not find the right info. I however recommend asking specific questions on the forum, not here, since this "board" is meant for specific issues with (building) the package or fixing errors. You can use the AAO thread found here: https://bbs.archlinux.org/viewtopic.php?id=68982

Hi, I'm just getting to grips with my new Arch install and was wondering how up to date the wiki entry is? Are the entries on how to configure things for the custom kernels still accurate?

only want to thanks for the great work with this package ;)

A largely revamped config file up for testing in the kernel26-one-dev package. I included the recommened options by powertop (e.g. PM_RUNTIME, USB_SUSPEND et. al.) but personally experienced an increase in power consumption instead of a decline. But crosschecking with different hw setups would be nice. So feel free to give it a try and provide your feedback there. Thank you!

Updated to latest kernel patch. Also march=prescott has been replaced by march=atom which has become available with gcc 4.5.

Ran for me out of the box. Maybe it fixes your issues too, Athaba?

Hmpffff. That's so odd. Exactly what happened to me on earlier versions and why I claimed I didn't get it running. Now it seemed fixed for me. What AAO version are you running? SSD or HDD?

Causes a panic here.

panic occurred, switching back to to text console
BUG; scheduling while atomic: swapper/1/0x00000003
Modules linked in:
Pid: 1, comm: swapper Not tainted 2.6.35-on #1

2.6.35 is finally up. This version does not yet contain optimizations, changes to powersaving settings or wlan tweaks/patches. I will provide them once I can recharge my AAO - I lost my AC adapter during holiday. ^^

Back from my holiday I didn't manage to get 2.6.35-one to work (while kernel26 from core is working) but I've provided my testing bed open to test for everyone: I've adopted the orphaned kernel26-one-dev package and will from now on use it as a development package for this. Both packages should be installable next to each other flawlessly. Check it out here if you want to experiment: http://aur.archlinux.org/packages.php?ID=22794

Please note that kernel26-one-dev is not meant as a stable replacement! So DO NOT INSTALL it if you want a working laptop.


On a second note: after checking Miblo's link, sorry, no way to include that fix since it's no kernel patch. As stated there it requires manual tweaking of the system after boot. So I have to take my comment of 10 days ago back sadly.

@Miblo Yes, I would include such a fix. Had been experimenting with a previous unsupported jumbo fix before but the discussion threads on the issue said the issue was fixed in 2.6.34 (which also according to my dmesg is not). If I ever manage to get 2.6.35 working I'll include the patch.

@Miblo lspci says the same for my wireless adapter. My connection used to go down frequently, I assumed my crappy router was to blame, but it seems ok since updating to 2.6.35-ARCH and dmesg no longer says unsupported jumbo.
You seem right about the model numbers, the box mine came in says both AOA 150-Ab and ZG5, so this should work for me, but I'll wait for 2.6.35 to test.

On an unrelated note, does everybody's graphics work well? OpenArena plays fine, but glxgears runs at just over 50 fps and torcs at about 3 fps.

Folks, if my Aspire One is anything to go by, the model numbers AOA150-Ab and ZG5 are interchangeable, since both appear on my machine.

Regarding wifi, `lspci` lists my wireless adaptor as the "Atheros Communications Inc. AR5001 Wireless Network Adapter (rev 01)". Does this correspond with everybody else's, and do you get a stable wifi connection with this kernel? I don't, and wonder if it would be possible to somehow stabilise the connection. Just a little research has taken me to this page (http://wireless.kernel.org/en/users/Drivers/ath5k#Eliminating_.60ath5k:_unsupported_jumbo.60_bug) which seems to provide a solution to the dropped-connections bug. I'll try it and see if it works. If it does, would it be possible to include such a fix in kernel26-one, or is it something that must be done manually?

Heads up on kernel 2.6.35: my first tests with different configs have been fairly unsuccessful leaving the system unbootable. :-( I have yet to figure out what setting is causing that. Since I will be on holiday for two weeks there won't any updates anytime soon. Please be patient. And also if any else is experimenting with 2.6.35 and manages to get it running, feel free to drop me a message.

Regarding the ZG5 if the hardware used (wifi & ethernet chipset, card readers etc.) is the same as the A110/150 hardware, there won't be any problems. So as Athaba said, feel free to give it a try. I personally just can't support anything despite the A110L.

@test0: Think, it should work. Why not simply try it?

It is an Acer Aspire One, just not an A110L, but I'll check out kernel-netbook instead. Thanks for the link.

I wouldn't recommend using this kernel with anything else but an Acer Aspire One since the kernel is limited to the hardware and functionality of this specific netbook. It might not work with other models. And I also won't support that. For a more general kernel for any netbook maybe you want to look at the kernel-netbook package: http://aur.archlinux.org/packages.php?ID=34625

Do I need to edit anything for this to work on a ZG5?

IIRC those options had been disabled on purpose by gothicknight (my config is still based on his original work in many respects). I had been experimenting especially with USB suspend myself some time ago and decided to leave them deactivated, don't remember the reasons though. Might have been that they actually didn't have any measurable effect. I'll give them a try again and maybe include them on the next update. Thanks for the hint.

Hi, thanks for your work, it's been really useful for me. However, there are some config options that should be enabled for powersaving purposes. They are (as suggested by powertop): CONFIG_CPU_FREQ_STAT and CONFIG_USB_SUSPEND. Thanks again.

CONFIG_LBDAF enabled, PKGBUILD adapted to new package() structure

The kernel doesn't have CONFIG_LBDAF enabled. Without it the EXT4 file system doesn't work correctly. Please enable it!

Thanks for reporting this error. Wouldn't have noticed that since oddly it did compile flawlessly for me a week ago. Fixed (together with a couple more errors) and updated.

I have a problem when trying to update from 2.6.33-one to 2.6.34-one. I get the error:

"cp: cannot stat <<include/asm-x86>>: No such file or directory"

Anyone else who gets this message and knows how to fix it?

Once again: DO NOT FLAG THE PACKAGE OUT OF DATE UNLESS A NEW VER HITS THE CORE REPO. This package will stay in line with core. .34 has only just hit testing... I won't dare breaking compatibility with this package. kernel26-one-dev is for that - or if you really want to try its quite easy to adapt my PKGBUILD to a testing kernel...

Yes, so? Guess how I test. It's fairly easy to adapt the PKGBUILD yourself if you wish to use them.

Still the released -one kernel should not be dependant on testing packages (like kernel26-firmware then would be). There is a kernel26-dev for that to be ahead of releases. But generally a new kernel version does usually stay in testing for a reason. Ir'd be even worse if an AUR package broke compatability with other arch packages. So, no, I won't publish anything until it hits core.

You do realize that the official arch patches are also offered for the kernel26 in testing, right?

Once again, do not flag this package out of date unless kernel26 has been updated in core since it depends on the offical arch patch.