unofficial linux-ck repo with cpu optimized packages is ONLINE

MrCode · 2012-12-21 22:33:24

graysky wrote:

Plz try v3.7.1-3 which should fix this issue for you:
[…]
EDIT: Repo has been populated.

I had actually updated my kernel to the then-latest ck-nehalem, but turned off NMI watchdog control in laptop-mode-tools to prevent the kernel panic/hard freeze. Will try turning it on now that I have the newest package.

EDIT: Awesome, no freeze!

Last edited by MrCode (2012-12-21 22:36:56)

akspecs · 2012-12-22 02:12:53

yess thank you for staying on top of things with 3.7.1-3 - the lockup/panic bug no longer affects me.

bassu · 2012-12-24 07:04:44

There's an apparent sound driver problem (intel) in 3.7.1-3 creating repeated stuttering after system wakes up from resume.
Just to let you guys know!

bassu · 2012-12-24 08:04:11

bassu wrote:

There's an apparent sound driver problem (intel) in 3.7.1-3 creating repeated stuttering after system wakes up from resume.
Just to let you guys know!

P.S. Added a little info on downgrading here at https://wiki.archlinux.org/index.php/Li … owngrading

Army · 2012-12-28 10:01:46

Hey graysky, didn't you have a virtualbox-ck package in the repo a little while ago?

edit: Nevermind, like always, I should have read the wiki first: https://wiki.archlinux.org/index.php/li … h_Linux-ck

Last edited by Army (2012-12-28 10:10:18)

lano1106 · 2012-12-29 00:03:11

Hi grayski,

the gcc arch test returns k8-sse3.

I assume that the best linux-ck would be the kx one. Do you agree?

graysky · 2012-12-29 00:33:35

@lano - Yep, go for it.

dodo3773 · 2013-01-02 19:11:27

Love the package graysky. Using sandybridge kernel. Wanted to ask is there a place for polls / new features etc.. Curious if you guys have talked about compression at all. I've been reading about lzo compression and it seems nice but this would pull in one or two extra dependencies. I grepped the .config and it looks like we are using gzip. Everyone has gzip for sure so that seems safe. Just wondering if different options etc.. have been talked about. Just curious that's all.

graysky · 2013-01-02 19:42:18

Have been using lzo myself for years. Faster. I don't wanna customize it too much beyond the ARCH kernel. As you pointed out, gzip is the de facto standard it seems.

dodo3773 · 2013-01-02 19:53:49

graysky wrote:

Have been using lzo myself for years. Faster. I don't wanna customize it too much beyond the ARCH kernel. As you pointed out, gzip is the de facto standard it seems.

Figured it was something like this. If the whole community agreed though (maybe if we had a poll or something) maybe someday it could be used as default? I vote for the switch. Not sure why anyone wouldn't want this unless they were that tight on space for the extra few MBs dependencies (in which case my guess is they would be rolling their own). Unless it conflicts with something else I am unaware of I am for it. Just throwing it out there. If you decide to throw something like that in the repo I would be happy to test it as well.

graysky · 2013-01-02 22:46:48

By popular request, I updated the gcc patch to allow for -march=amdfam10 builds and built up a corresponding package set for you AMD guys. New group is called 'ck-k10' so if you have supported hardware, plz try it out. Packages are online now.

More on wiki: https://wiki.archlinux.org/index.php/Repo-ck

swanson · 2013-01-03 04:23:41

Just installed and rebooted with your new ck-k10, kernel, headers and nvidia-ck-k10. Works really great! Thanks for the effort Graysky! Been using your kernels and nvidia for years.

Nepherte · 2013-01-03 19:14:36

Same here. Just installed ck-k10, headers and nvidia. Works just fine.

sxe · 2013-01-04 09:49:58

Hi,

i'm a log time linux-ck user and "suddenly" i get this error shortly after my computer booted and my desktop is fully visible ( i can even click some windows).
It does not happen all the time but often enough to get really annoying because the whole X server is dead after that error and i have to ssh to my computer and reboot it.

When using the stock kernel this does not happen as far as i know.

Here is the error message:

Jan 04 10:09:26 main kernel: NVRM: GPU at 0000:01:00.0 has fallen off the bus.
Jan 04 10:09:26 main kernel: NVRM: os_pci_init_handle: invalid context!
Jan 04 10:09:26 main kernel: NVRM: os_pci_init_handle: invalid context!
Jan 04 10:09:26 main kernel: NVRM: GPU at 0000:01:00.0 has fallen off the bus.
Jan 04 10:09:26 main kernel: NVRM: os_pci_init_handle: invalid context!
Jan 04 10:09:26 main kernel: NVRM: os_pci_init_handle: invalid context!
Jan 04 10:09:27 main kernel: irq 16: nobody cared (try booting with the "irqpoll" option)
Jan 04 10:09:27 main kernel: Pid: 0, comm: BFS/0 Tainted: P           O 3.7.1-3-ck #1
Jan 04 10:09:27 main kernel: Call Trace:
Jan 04 10:09:27 main kernel:  <IRQ>  [<ffffffff810ce536>] __report_bad_irq+0x36/0xe0
Jan 04 10:09:27 main kernel:  [<ffffffff810ce833>] note_interrupt+0x1a3/0x1f0
Jan 04 10:09:27 main kernel:  [<ffffffff810cc13f>] handle_irq_event_percpu+0xbf/0x260
Jan 04 10:09:27 main kernel:  [<ffffffff810cc328>] handle_irq_event+0x48/0x70
Jan 04 10:09:27 main kernel:  [<ffffffff810cf38a>] handle_fasteoi_irq+0x5a/0x100
Jan 04 10:09:27 main kernel:  [<ffffffff81017742>] handle_irq+0x22/0x40
Jan 04 10:09:27 main kernel:  [<ffffffff81497ada>] do_IRQ+0x5a/0xe0
Jan 04 10:09:27 main kernel:  [<ffffffff8148e9ad>] common_interrupt+0x6d/0x6d
Jan 04 10:09:27 main kernel:  <EOI>  [<ffffffff812ae94d>] ? intel_idle+0xed/0x150
Jan 04 10:09:27 main kernel:  [<ffffffff812ae92e>] ? intel_idle+0xce/0x150
Jan 04 10:09:27 main kernel:  [<ffffffff8135c659>] cpuidle_enter+0x19/0x20
Jan 04 10:09:27 main kernel:  [<ffffffff8135cd46>] cpuidle_idle_call+0xa6/0x340
Jan 04 10:09:27 main kernel:  [<ffffffff8101eeff>] cpu_idle+0xbf/0x130
Jan 04 10:09:27 main kernel:  [<ffffffff81468b9b>] rest_init+0x7b/0x80
Jan 04 10:09:27 main kernel:  [<ffffffff818b0c40>] start_kernel+0x3cc/0x3d9
Jan 04 10:09:27 main kernel:  [<ffffffff818b0673>] ? repair_env_string+0x5e/0x5e
Jan 04 10:09:27 main kernel:  [<ffffffff818b0356>] x86_64_start_reservations+0x131/0x135
Jan 04 10:09:27 main kernel:  [<ffffffff818b045a>] x86_64_start_kernel+0x100/0x10f
Jan 04 10:09:27 main kernel: handlers:
Jan 04 10:09:27 main kernel: [<ffffffffa000b560>] usb_hcd_irq [usbcore]
Jan 04 10:09:27 main kernel: [<ffffffffa0e42030>] nv_kern_isr [nvidia]
Jan 04 10:09:27 main kernel: Disabling IRQ #16

I use the linux-ck repository and the latest nvidia driver from the same repository as well.

I had this problem before, some month ago and i was able to fix it by executing this at boot "/usr/bin/nvidia-smi -pm 1" to set the GPU to persistence mode but this does not help anymore as it seems.

Would be nice to get some help.

Greetings,
Andy

Last edited by sxe (2013-01-04 09:53:27)

graysky · 2013-01-04 09:58:14

Sxe - I am going to guess that your problem is not related to the ck patch set but to something else. Can you please try the ARCH kernel from [testing] and see if you can reproduce the panic condition?

sxe · 2013-01-04 10:44:42

Hi graysky,

i'm not sure updating to the arch kernel from testing will help because i got this error with all of the latest -ck kernels.

Anyway, to rule that out i will install the kernel from testing and see how that goes.

Thanks for your help!

Greetings

sxe · 2013-01-04 11:28:08

ok, i did some more testing.

I booted each kernel 5 times and on both it did not froze even once.
I have no idea how to debug this further. If you have an idea that would be awesome.

graysky · 2013-01-04 18:49:43

I doubt this problem is related to linux-ck at all. I am not an expert in debugging. You can atart by searching in google for key words in your panic and see what you find.

sxe · 2013-01-04 19:10:18

Thats what i did the whole day without being successful.
btw, it happened again with the linux-ck kernel while playing a game in WINE. I played 2 hours after that with the normal kernel without a crash. MAybe its the nvidia-ck package? I have no idea.

DarkSkyGhost · 2013-01-05 06:45:03

I have this weird thing happening.

Whenever I use molecular dynamics simulation and cpu's both cores are constantly loaded (but hdd gets written into briefly only at checkpoints every couple of minutes), almost all programs are responsive and working properly, dynamics program backs off when another program wants some cpu.

BUT, two programs have weird behaviour. One is ktorrent, another is clementine.

KTorrent doesn't want to ul/dl, everything is stalled. As soon as simulation stops, KTorrent proceeds to work.
And clementine from time to time hangs for ten or so minutes.

Can it be that ck kernel and/or bfq are somehow responsible for this? (Eh, probably it would be faster just to test it with normal kernel.)
Or these two programs just think that they are doing a good job for not working when cpu is loaded? Or is it kde's problem?

Could anyone point me in any direction where to dig deeper? I'd appreciate it.

WonderWoofy · 2013-01-05 07:12:06

DarkSkyGhost wrote:

Eh, probably it would be faster just to test it with normal kernel.

You should probably take your own advice and start here. It sounds like you have done no investigation so far... you need to at least gather some information.

DarkSkyGhost · 2013-01-05 09:57:14

And so I did.

Booted default kernel, launched simulation, started ktorrent, and it was dl/ul'ing properly. Although system was less responsive and stuttery, especially music playback.

Booted ck-k10, started ktorrent, launched simulation, dl/ul'ing stopped. As soon as simulation was stopped, ktorrent started working normally. And as soon as simulation would be restarted, ktorrent would immediately stop. But system was more responsive and music playback wasn't stuttering.

Where do I go from here?

Or is it just simply a trade - I get responsive system, but some things don't work, or I get less responsive system, but with everything working?
Or do I have to manually compile kernel to lessen the drawbacks and make it more to my liking?

I tried playing with nice levels, setting simulation for lowest priority while setting ktorrent for highest, but to no avail.

Last edited by DarkSkyGhost (2013-01-05 10:00:01)

graysky · 2013-01-05 12:44:22

@dark - Please continue this discussion in a new thread; this thread is for repo issues. I will gladly post to your new thread with my thoughts (for whatever they are worth).

latjay · 2013-01-09 01:23:49

I'm running a current kernel from the repo and here's what I find in it's config:

$ cat /proc/config.gz | zgrep CONFIG_HZ_
# CONFIG_HZ_100 is not set
# CONFIG_HZ_250_NODEFAULT is not set
CONFIG_HZ_300=y
# CONFIG_HZ_1000 is not set
# CONFIG_HZ_1500 is not set
# CONFIG_HZ_2000 is not set
# CONFIG_HZ_3000 is not set
# CONFIG_HZ_4000 is not set
# CONFIG_HZ_5000 is not set
# CONFIG_HZ_7500 is not set
# CONFIG_HZ_10000 is not set

One of the patches in the ck patchset defaults this to 1000 and it's not changed in the config in graysky's sources. Does anyone else has the same?

graysky · 2013-01-09 01:28:15

latjay wrote:

One of the patches in the ck patchset defaults this to 1000 and it's not changed in the config in graysky's sources. Does anyone else has the same?

Which patch are your referencing in that statement?

% zgrep CONFIG_HZ_ /proc/config.gz
# CONFIG_HZ_100 is not set
# CONFIG_HZ_250 is not set
CONFIG_HZ_300=y
# CONFIG_HZ_1000 is not set

Last edited by graysky (2013-01-09 01:31:24)

Arch Linux

#1151 2012-12-21 22:33:24

Re: unofficial linux-ck repo with cpu optimized packages is ONLINE

#1152 2012-12-22 02:12:53

Re: unofficial linux-ck repo with cpu optimized packages is ONLINE

#1153 2012-12-24 07:04:44

Re: unofficial linux-ck repo with cpu optimized packages is ONLINE

#1154 2012-12-24 08:04:11

Re: unofficial linux-ck repo with cpu optimized packages is ONLINE

#1155 2012-12-28 10:01:46

Re: unofficial linux-ck repo with cpu optimized packages is ONLINE

#1156 2012-12-29 00:03:11

Re: unofficial linux-ck repo with cpu optimized packages is ONLINE

#1157 2012-12-29 00:33:35

Re: unofficial linux-ck repo with cpu optimized packages is ONLINE

#1158 2013-01-02 19:11:27

Re: unofficial linux-ck repo with cpu optimized packages is ONLINE

#1159 2013-01-02 19:42:18

Re: unofficial linux-ck repo with cpu optimized packages is ONLINE

#1160 2013-01-02 19:53:49

Re: unofficial linux-ck repo with cpu optimized packages is ONLINE

#1161 2013-01-02 22:46:48

Re: unofficial linux-ck repo with cpu optimized packages is ONLINE

#1162 2013-01-03 04:23:41

Re: unofficial linux-ck repo with cpu optimized packages is ONLINE

#1163 2013-01-03 19:14:36

Re: unofficial linux-ck repo with cpu optimized packages is ONLINE

#1164 2013-01-04 09:49:58

Re: unofficial linux-ck repo with cpu optimized packages is ONLINE

#1165 2013-01-04 09:58:14

Re: unofficial linux-ck repo with cpu optimized packages is ONLINE

#1166 2013-01-04 10:44:42

Re: unofficial linux-ck repo with cpu optimized packages is ONLINE

#1167 2013-01-04 11:28:08

Re: unofficial linux-ck repo with cpu optimized packages is ONLINE

#1168 2013-01-04 18:49:43

Re: unofficial linux-ck repo with cpu optimized packages is ONLINE

#1169 2013-01-04 19:10:18

Re: unofficial linux-ck repo with cpu optimized packages is ONLINE

#1170 2013-01-05 06:45:03

Re: unofficial linux-ck repo with cpu optimized packages is ONLINE

#1171 2013-01-05 07:12:06

Re: unofficial linux-ck repo with cpu optimized packages is ONLINE

#1172 2013-01-05 09:57:14

Re: unofficial linux-ck repo with cpu optimized packages is ONLINE

#1173 2013-01-05 12:44:22

Re: unofficial linux-ck repo with cpu optimized packages is ONLINE

#1174 2013-01-09 01:23:49

Re: unofficial linux-ck repo with cpu optimized packages is ONLINE

#1175 2013-01-09 01:28:15

Re: unofficial linux-ck repo with cpu optimized packages is ONLINE

Board footer