View previous topic :: View next topic |
Author |
Message |
rahulthewall Veteran
Joined: 01 Nov 2007 Posts: 1264 Location: Zürich
|
Posted: Sun Dec 12, 2010 8:40 pm Post subject: |
|
|
I ditched git for zen-sources ebuild from portage. Everything works now. _________________ Who shall guard the guards? |
|
Back to top |
|
|
mbar Veteran
Joined: 19 Jan 2005 Posts: 1990 Location: Poland
|
Posted: Mon Dec 13, 2010 9:28 am Post subject: |
|
|
mbar wrote: | And I think the ath9k update in zen sources is broken (no one can connect to my hostapd on my in-kernel ath9k).
EDIT: yessss, confirmed by installing 2.6.36-gentoo-r4 (ath9k works) and r5 (ath9k doesn't work). This is a problem with 2.6.36.2 update -- where should I report this (I have not sent any kernel bug report yet)? |
...and under 2.6.37-rc5-git3 ath9k works OK for me again
Could you guys backport ath9k from .37 to zen-stable? |
|
Back to top |
|
|
ponciarello Apprentice
Joined: 22 Jul 2008 Posts: 223 Location: beach of slack
|
Posted: Tue Dec 14, 2010 8:49 am Post subject: |
|
|
FYI, after the re-merge
Code: | Kernel: arch/x86/boot/bzImage is ready (#1)
Building modules, stage 2.
MODPOST 520 modules
[...]
ERROR: "unlzma" [fs/squashfs/squashfs.ko] undefined!
make[1]: *** [__modpost] Errore 1
make: *** [modules] Errore 2 |
EDIT: I tried applying bfs-357-360 to master and seems to work (just a -I think not blocking- .rej) |
|
Back to top |
|
|
jw5801 Apprentice
Joined: 12 Jun 2008 Posts: 251 Location: Melbourne, Australia
|
Posted: Fri Dec 17, 2010 11:28 pm Post subject: |
|
|
Is SLQB around any more? It seems to have disappeared since 2.6.35. Currently using zen-stable and getting absolutely horrendous performance under any semblance of disk load. I remember this used to happen and went away completely when I started using BFS, BFQ and SLQB. At the moment I'm using BFS, BFQ and SLUB. |
|
Back to top |
|
|
kernelOfTruth Watchman
Joined: 20 Dec 2005 Posts: 6111 Location: Vienna, Austria; Germany; hello world :)
|
|
Back to top |
|
|
jw5801 Apprentice
Joined: 12 Jun 2008 Posts: 251 Location: Melbourne, Australia
|
Posted: Fri Dec 17, 2010 11:35 pm Post subject: |
|
|
kernelOfTruth wrote: | so switch to SLAB and you'll be way more happier |
Done. Will report back in a little while.
Report: So far so good. MUCH improved on SLUB.
Report 2: Although still locks up occasionally under disk load. Seems to happen most when copying to a slow USB connected drive.
Report 3: Scratch report 2, I'd accidentally booted back into the old kernel. Definitely performs significantly better! |
|
Back to top |
|
|
Ant P. Watchman
Joined: 18 Apr 2009 Posts: 6920
|
Posted: Sun Dec 19, 2010 5:14 pm Post subject: |
|
|
Er, this is weird. I just noticed the "priority" column in htop is reporting insane values (> 3 million) for a few programs. It's not causing (visible) problems, but I've never seen it before. |
|
Back to top |
|
|
one_and_only Apprentice
Joined: 13 May 2007 Posts: 250 Location: PL/Krakow
|
Posted: Fri Dec 24, 2010 9:44 am Post subject: |
|
|
Does anybody have problems with reiser4 with latest kernel? Everytime i try to do emerge -NuDvta --keep-going @world computer hangs and the only think I can do is hard reset. I got:
Code: |
Dec 23 20:16:09 localhost kernel: kernel BUG at fs/reiser4/block_alloc.c:151!
Dec 23 20:16:09 localhost kernel: invalid opcode: 0000 [#1] PREEMPT SMP
Dec 23 20:16:09 localhost kernel: last sysfs file: /sys/devices/pci0000:00/0000:00:19.0/net/eth0/carrier
Dec 23 20:16:09 localhost kernel: CPU 0
Dec 23 20:16:09 localhost kernel: Modules linked in: sco bnep rfcomm l2cap crc16 fuse vboxnetflt vboxnetadp vboxdrv tp_smapi hdaps thinkpad_ec cpufre
q_userspace cpufreq_powersave cpufreq_ondemand cpufreq_conservative snd_hda_codec_analog nvidia(P) ath5k mac80211 snd_hda_intel ath snd_hda_codec snd
_hwdep snd_pcm cfg80211 snd_timer snd_page_alloc firewire_ohci firewire_core thinkpad_acpi nvram iTCO_wdt i2c_i801 btusb bluetooth rfkill sdhci_pci s
dhci mmc_core yenta_socket pcmcia_core pcmcia_rsrc led_class
Dec 23 20:16:09 localhost kernel:
Dec 23 20:16:09 localhost kernel: Pid: 2883, comm: emerge Tainted: P 2.6.36-zen2 #9 6459CTO/6459CTO
Dec 23 20:16:09 localhost kernel: RIP: 0010:[<ffffffff811689af>] [<ffffffff811689af>] sub_from_ctx_grabbed.clone.8+0xf/0x20
Dec 23 20:16:09 localhost kernel: RSP: 0018:ffff88010b9a1a80 EFLAGS: 00010297
[...]
Dec 23 20:16:09 localhost kernel: ---[ end trace 3a97c5d5b2565a46 ]---
Dec 23 20:16:09 localhost kernel: note: emerge[2883] exited with preempt_count 2
Dec 23 20:16:09 localhost kernel: BUG: scheduling while atomic: emerge/2883/0x10000003
|
Moreover, my computer seems much more unresponsive while disk access than with 2.6.35. Maybye I got something messed up in my kernel? Could somebody tell me what was the conclusion from thread about unresponsiveness on amd64? What should I set/patch if I'm using lastest git? |
|
Back to top |
|
|
kernelOfTruth Watchman
Joined: 20 Dec 2005 Posts: 6111 Location: Vienna, Austria; Germany; hello world :)
|
Posted: Fri Dec 24, 2010 10:16 am Post subject: |
|
|
one_and_only wrote: | Does anybody have problems with reiser4 with latest kernel? Everytime i try to do emerge -NuDvta --keep-going @world computer hangs and the only think I can do is hard reset. I got:
Code: |
Dec 23 20:16:09 localhost kernel: kernel BUG at fs/reiser4/block_alloc.c:151!
Dec 23 20:16:09 localhost kernel: invalid opcode: 0000 [#1] PREEMPT SMP
Dec 23 20:16:09 localhost kernel: last sysfs file: /sys/devices/pci0000:00/0000:00:19.0/net/eth0/carrier
Dec 23 20:16:09 localhost kernel: CPU 0
Dec 23 20:16:09 localhost kernel: Modules linked in: sco bnep rfcomm l2cap crc16 fuse vboxnetflt vboxnetadp vboxdrv tp_smapi hdaps thinkpad_ec cpufre
q_userspace cpufreq_powersave cpufreq_ondemand cpufreq_conservative snd_hda_codec_analog nvidia(P) ath5k mac80211 snd_hda_intel ath snd_hda_codec snd
_hwdep snd_pcm cfg80211 snd_timer snd_page_alloc firewire_ohci firewire_core thinkpad_acpi nvram iTCO_wdt i2c_i801 btusb bluetooth rfkill sdhci_pci s
dhci mmc_core yenta_socket pcmcia_core pcmcia_rsrc led_class
Dec 23 20:16:09 localhost kernel:
Dec 23 20:16:09 localhost kernel: Pid: 2883, comm: emerge Tainted: P 2.6.36-zen2 #9 6459CTO/6459CTO
Dec 23 20:16:09 localhost kernel: RIP: 0010:[<ffffffff811689af>] [<ffffffff811689af>] sub_from_ctx_grabbed.clone.8+0xf/0x20
Dec 23 20:16:09 localhost kernel: RSP: 0018:ffff88010b9a1a80 EFLAGS: 00010297
[...]
Dec 23 20:16:09 localhost kernel: ---[ end trace 3a97c5d5b2565a46 ]---
Dec 23 20:16:09 localhost kernel: note: emerge[2883] exited with preempt_count 2
Dec 23 20:16:09 localhost kernel: BUG: scheduling while atomic: emerge/2883/0x10000003
|
Moreover, my computer seems much more unresponsive while disk access than with 2.6.35. Maybye I got something messed up in my kernel? Could somebody tell me what was the conclusion from thread about unresponsiveness on amd64? What should I set/patch if I'm using lastest git? |
please revert reiser4 and download the latest patch from kernel.org - Edward meanwhile updated it as he explained on the reiserfs mailing list
Edward wrote: | Could you please try the attached patch?
This patch is against reiser4-for-2.6.36.
Also please make sure you have the latest one
(I have updated the stuff 11-Dec-2010). |
reiser-for-2.6.36 + reiser4-fix-entd_flush.patch _________________ https://github.com/kernelOfTruth/ZFS-for-SystemRescueCD/tree/ZFS-for-SysRescCD-4.9.0
https://github.com/kernelOfTruth/pulseaudio-equalizer-ladspa
Hardcore Gentoo Linux user since 2004 |
|
Back to top |
|
|
one_and_only Apprentice
Joined: 13 May 2007 Posts: 250 Location: PL/Krakow
|
Posted: Sat Dec 25, 2010 11:44 am Post subject: |
|
|
kernelOfTruth wrote: |
please revert reiser4 and download the latest patch from kernel.org - Edward meanwhile updated it as he explained on the reiserfs mailing list
Edward wrote: | Could you please try the attached patch?
This patch is against reiser4-for-2.6.36.
Also please make sure you have the latest one
(I have updated the stuff 11-Dec-2010). |
kernelOfTruth wrote: |
reiser-for-2.6.36 + reiser4-fix-entd_flush.patch |
|
OK, I reverted zen resier4 patch, apliend Edward's - works fine - thanks! To be sure, do I need sth more than this http://www.kernel.org/pub/linux/kernel/people/edward/reiser4/reiser4-for-2.6/ path? Is reiser4-fix-entd_flush.patch included in Edward's?
Last edited by one_and_only on Sat Dec 25, 2010 1:35 pm; edited 1 time in total |
|
Back to top |
|
|
kernelOfTruth Watchman
Joined: 20 Dec 2005 Posts: 6111 Location: Vienna, Austria; Germany; hello world :)
|
|
Back to top |
|
|
Ant P. Watchman
Joined: 18 Apr 2009 Posts: 6920
|
Posted: Sun Dec 26, 2010 11:07 pm Post subject: |
|
|
Can we please have snd_aloop (alsa-driver module)? I've tried using alsa-driver-1.0.23, alsa-driver-9999 and compiling it manually, but the best I can get is hard application lockups from it. Using latest zen-stable kernel. |
|
Back to top |
|
|
Ormaaj Guru
Joined: 28 Jan 2008 Posts: 319
|
Posted: Wed Dec 29, 2010 7:58 am Post subject: |
|
|
kernelOfTruth wrote: | jw5801 wrote: | Is SLQB around any more? It seems to have disappeared since 2.6.35. Currently using zen-stable and getting absolutely horrendous performance under any semblance of disk load. I remember this used to happen and went away completely when I started using BFS, BFQ and SLQB. At the moment I'm using BFS, BFQ and SLUB. |
SLUB sux and this is fact
so switch to SLAB and you'll be way more happier | What makes you say that? I suppose this is outdated? http://kernel-seeds.org/settings-01.html |
|
Back to top |
|
|
kernelOfTruth Watchman
Joined: 20 Dec 2005 Posts: 6111 Location: Vienna, Austria; Germany; hello world :)
|
|
Back to top |
|
|
wrc1944 Advocate
Joined: 15 Aug 2002 Posts: 3435 Location: Gainesville, Florida
|
Posted: Wed Dec 29, 2010 4:46 pm Post subject: |
|
|
That kernel-seeds.org page Ormaaj mentions also doesn't like cgroups (to say the least). Is this also now considered outdated? Quote: | [ ] Group CPU scheduler
[ ] Control Group support --->
Important!Consider both of these settings as poison to your computer. While the idea has merit, at the time of this writing, there remain no userspace programs to interface with the setup. As such, these settings and their analogues tend to slow systems down. For this reason, these settings are turned off by default, and their use is highly discouraged. Kernel seeds are about speed and efficiency, and the above settings are neither fast or efficient. |
I was thinking of adopting cgroups (either the full kernel patch or the 4 line ~/.bashrc file userspace method by Red Hat dev Lennart Poettering). http://www.webupd8.org/2010/11/alternative-to-200-lines-kernel-patch.html . This page goes on with users discussing several variations of this alternative method, pros/cons etc. However, after reading all this (and many other info sources) I'm as confused as ever.
I read OpenSuse has gone with the supposedly great 200 line cgroup kernel patch as a default.
http://news.opensuse.org/2010/12/23/opensuse-announces-fifth-development-milestone-with-kernel-interactivity-patch
The suse guys must know what they're doing, so I'm perplexed by the apparent contradictions being expressed by various "kernel experts" on cgroups as well as the above mentioned slub/slab disagreement. What's the consensus about either of these kernel options (if any) on this thread? I'd assume anyone using zen kernels knows at least as much as I do at the moment.
I've always wondered if any of these type options really make any real-world difference if you have something like 4-8GB ram, enable 'Sysctl support' in the kernel, and set vm.swappiness=5 in /etc/sysctl.conf. Quote: | Linux moves memory pages that have not been accessed for some time to the swap space even if there is enough free memory available. By changing the percentage in /proc/sys/vm/swappiness you can control the swapping behavior, depending on the system configuration. A high swappiness value means that the kernel will be more apt to unmap mapped pages. A low swappiness value means the opposite, the kernel will be less apt to unmap mapped pages. In other words, the higher the vm.swappiness value, the more the system will swap.
vm.swappiness takes a value between 0 and 100 to change the balance between swapping applications and freeing cache. At 100, the kernel will always prefer to find inactive pages and swap them out; in other cases, whether a swapout occurs depends on how much application memory is in use and how poorly the cache is doing at finding and releasing inactive items. | Is this also "outdated"? _________________ Main box- AsRock x370 Gaming K4
Ryzen 7 3700x, 3.6GHz, 16GB GSkill Flare DDR4 3200mhz
Samsung SATA 1000GB, Radeon HD R7 350 2GB DDR5
OpenRC Gentoo ~amd64 plasma, glibc-2.36-r7, gcc-13.2.1_p20230304
kernel-6.8.4 USE=experimental python3_11 |
|
Back to top |
|
|
kernelOfTruth Watchman
Joined: 20 Dec 2005 Posts: 6111 Location: Vienna, Austria; Germany; hello world :)
|
|
Back to top |
|
|
wrc1944 Advocate
Joined: 15 Aug 2002 Posts: 3435 Location: Gainesville, Florida
|
Posted: Wed Dec 29, 2010 8:35 pm Post subject: |
|
|
I've been a long-time user of Zen kernels, as well as trying most of the "performance" type kernel patch sets for many years. Years ago, the old ck "staircase" patches were noticable on my athlon systems, and recently the bfs scheduler has been similar in that you could really notice a difference between vanilla and bfs.
I do recall emailing Con Kolivas about whether or not the "staircase" or similar kernel patches really made much difference if you had ample ram, and IIRC he tended to agree.
Lately, I've been running either vanilla and vanilla-rc's with bfs (when I could get it to apply to the rc's), or vanilla git-sources (rcx-gitx), and with my usage patterns even plain vanilla's seem to get pretty decent desktop interactiveness. Of course that might be influenced by a pretty decent cpu/motherboard upgrade with 8GB ram and vm.swappiness=5 in /etc/sysctl.conf, not to mention an improving default kernel scheduler.
However, I'm still a big zen kernel fan, and will always give the releases a good try. _________________ Main box- AsRock x370 Gaming K4
Ryzen 7 3700x, 3.6GHz, 16GB GSkill Flare DDR4 3200mhz
Samsung SATA 1000GB, Radeon HD R7 350 2GB DDR5
OpenRC Gentoo ~amd64 plasma, glibc-2.36-r7, gcc-13.2.1_p20230304
kernel-6.8.4 USE=experimental python3_11 |
|
Back to top |
|
|
kernelOfTruth Watchman
Joined: 20 Dec 2005 Posts: 6111 Location: Vienna, Austria; Germany; hello world :)
|
Posted: Wed Dec 29, 2010 9:37 pm Post subject: |
|
|
wrc1944 wrote: | I've been a long-time user of Zen kernels, as well as trying most of the "performance" type kernel patch sets for many years. Years ago, the old ck "staircase" patches were noticable on my athlon systems, and recently the bfs scheduler has been similar in that you could really notice a difference between vanilla and bfs.
I do recall emailing Con Kolivas about whether or not the "staircase" or similar kernel patches really made much difference if you had ample ram, and IIRC he tended to agree.
Lately, I've been running either vanilla and vanilla-rc's with bfs (when I could get it to apply to the rc's), or vanilla git-sources (rcx-gitx), and with my usage patterns even plain vanilla's seem to get pretty decent desktop interactiveness. Of course that might be influenced by a pretty decent cpu/motherboard upgrade with 8GB ram and vm.swappiness=5 in /etc/sysctl.conf, not to mention an improving default kernel scheduler.
However, I'm still a big zen kernel fan, and will always give the releases a good try. |
even vm.swappiness=60 works for me _________________ https://github.com/kernelOfTruth/ZFS-for-SystemRescueCD/tree/ZFS-for-SysRescCD-4.9.0
https://github.com/kernelOfTruth/pulseaudio-equalizer-ladspa
Hardcore Gentoo Linux user since 2004 |
|
Back to top |
|
|
rx3 n00b
Joined: 02 Jan 2011 Posts: 2
|
Posted: Sun Jan 02, 2011 3:04 pm Post subject: |
|
|
kernelOfTruth wrote: | Ormaaj wrote: | kernelOfTruth wrote: | jw5801 wrote: | Is SLQB around any more? It seems to have disappeared since 2.6.35. Currently using zen-stable and getting absolutely horrendous performance under any semblance of disk load. I remember this used to happen and went away completely when I started using BFS, BFQ and SLQB. At the moment I'm using BFS, BFQ and SLUB. |
SLUB sux and this is fact
so switch to SLAB and you'll be way more happier | What makes you say that? I suppose this is outdated? http://kernel-seeds.org/settings-01.html |
it's my own day-to-day experience and reading the comments & reports from the devs and users on lkml |
And yet...
https://bbs.archlinux.org/viewtopic.php?id=36019
http://www.webhostingtalk.com/showthread.php?t=913888
+1 for SLQB, if only someone ported the patch to 2.6.37... |
|
Back to top |
|
|
wrc1944 Advocate
Joined: 15 Aug 2002 Posts: 3435 Location: Gainesville, Florida
|
Posted: Sun Jan 02, 2011 8:35 pm Post subject: |
|
|
IMHO it's worth considering that the Arch Linux page is more than 3 years old. In computer software code terms, it's kind of like comparing 21st century technologies to to their 18th century counterparts.
The other page is one year old, and mainly seems to be concerned about web hosting servers running apache.
I'm not saying slub is no good, just that the performance of any version of any allocator may be good or bad for a given task, or subject to who knows what future kernel patches/changes may bring. In any case, I'm always open to changing any kernel option/setting according to whatever currently produces the best result.
I guess the only thing we can say for sure is that everything will eventually become outdated. With that, guess I'm off to try the new 2.6.37-rc8 bfs patch from Con Kolivas. _________________ Main box- AsRock x370 Gaming K4
Ryzen 7 3700x, 3.6GHz, 16GB GSkill Flare DDR4 3200mhz
Samsung SATA 1000GB, Radeon HD R7 350 2GB DDR5
OpenRC Gentoo ~amd64 plasma, glibc-2.36-r7, gcc-13.2.1_p20230304
kernel-6.8.4 USE=experimental python3_11 |
|
Back to top |
|
|
rx3 n00b
Joined: 02 Jan 2011 Posts: 2
|
Posted: Wed Jan 05, 2011 10:24 am Post subject: |
|
|
wrc1944 wrote: | IMHO it's worth considering that the Arch Linux page is more than 3 years old. In computer software code terms, it's kind of like comparing 21st century technologies to to their 18th century counterparts. The other page is one year old, and mainly seems to be concerned about web hosting servers running apache. |
These are more recent (2010):
http://lwn.net/Articles/394048/
http://lwn.net/Articles/394211/
wrc1944 wrote: | With that, guess I'm off to try the new 2.6.37-rc8 bfs patch from Con Kolivas. |
Thanks to CK's work we can choose BFS instead of CFS, thanks to Paolo & Francesco we have BFQ instead of CFQ, now if only Nick Piggin had continued his work on SLQB we at least had an alternative to SLAB & SLUB... |
|
Back to top |
|
|
wrc1944 Advocate
Joined: 15 Aug 2002 Posts: 3435 Location: Gainesville, Florida
|
Posted: Wed Jan 05, 2011 4:24 pm Post subject: |
|
|
rx3,
Thanks for the links. This 2010 info looks very encouraging regarding SLUB+Q. _________________ Main box- AsRock x370 Gaming K4
Ryzen 7 3700x, 3.6GHz, 16GB GSkill Flare DDR4 3200mhz
Samsung SATA 1000GB, Radeon HD R7 350 2GB DDR5
OpenRC Gentoo ~amd64 plasma, glibc-2.36-r7, gcc-13.2.1_p20230304
kernel-6.8.4 USE=experimental python3_11 |
|
Back to top |
|
|
tranquilcool Veteran
Joined: 25 Mar 2005 Posts: 1179
|
Posted: Mon Jan 10, 2011 7:21 am Post subject: |
|
|
LD arch/x86/built-in.o
CC kernel/sched.o
In file included from kernel/sched.c:2:0:
kernel/sched_bfs.c:2011:6: error: conflicting types for ‘calc_global_load’
include/linux/sched.h:152:13: note: previous declaration of ‘calc_global_load’ was here
make[1]: *** [kernel/sched.o] Error 1
make: *** [kernel] Error 2
errors with latest pull. _________________ this is a strange strange world. |
|
Back to top |
|
|
wrc1944 Advocate
Joined: 15 Aug 2002 Posts: 3435 Location: Gainesville, Florida
|
Posted: Mon Jan 10, 2011 10:02 am Post subject: |
|
|
Running 2.6.37-ck1 on three Gentoo installs. Runs fine and had no problems patching and compiling.
2.6.37-ck1 is my current kernel of choice, but thinking of trying a 2.6.37 with latest cgroups version.
Anyone tried comparing ck1 with a cgroups kernel? _________________ Main box- AsRock x370 Gaming K4
Ryzen 7 3700x, 3.6GHz, 16GB GSkill Flare DDR4 3200mhz
Samsung SATA 1000GB, Radeon HD R7 350 2GB DDR5
OpenRC Gentoo ~amd64 plasma, glibc-2.36-r7, gcc-13.2.1_p20230304
kernel-6.8.4 USE=experimental python3_11 |
|
Back to top |
|
|
broch Tux's lil' helper
Joined: 09 Jul 2005 Posts: 94
|
Posted: Mon Jan 10, 2011 6:18 pm Post subject: |
|
|
tranquilcool wrote: | LD arch/x86/built-in.o
CC kernel/sched.o
In file included from kernel/sched.c:2:0:
kernel/sched_bfs.c:2011:6: error: conflicting types for ‘calc_global_load’
include/linux/sched.h:152:13: note: previous declaration of ‘calc_global_load’ was here
make[1]: *** [kernel/sched.o] Error 1
make: *** [kernel] Error 2
errors with latest pull. |
Hello,
I am getting the same error after today pull
thank you |
|
Back to top |
|
|
|