2.6.20-viper9: Photon Torpedoes

[vipernicus]

Viper Sources is pretty much my playground for new patches.
The main goal of this patchset is pretty much to offer the latest features, decrease latencies, decrease overhead, and improve interactivity. The only difference is experimental versions of everything. I boot test each release and run for an extended amount of time.

/* Notes */
04/24/2007
No splitsched, just pure SD 0.46, plus an extra patch for boosting priorities of priviledged tasks. Default settings is '0', but setting it to '-10' could give you much better interactivity on the desktop.
04/22/2007
Back to splitsched, thanks to predatorfreak. Staircase is updated to v17.1, SD is updated to 0.45, and you can expect Nicksched v33 soon. One very important feature to note is the new ext4 git release is a rerollup of ext4 for 2.6.20 by the developers to fix bugs with the old one. (Hope it fixes the timestamp issues for you x86_64 guys.) I also rolled up delayed allocation support for ext4 in the patch, so you can 'mount -o delalloc' now. New suspend2. Several patches from -dark like execshield and cpusupport. (We trade patches alot, obviously.)
04/07/2007
Updated to 2.6.20.6, SD 0.39 (included in the ck patchset), newer genpatches, suspend2 fix for 2.2.9.10, updates for ext4 including a fix for the timestamp issue on x86_64, and I've also included a new reiser4 patch used in the benchmarks from http://linuxhelp.150m.com/resources/fs-benchmarks.htm

According to Con, for best desktop interactivity:
renice X to -10 and 'echo 6 > /proc/sys/kernel/rr_interval'

Read more here:
http://article.gmane.org/gmane.linux.kernel.ck/7044

Extended changelog, downloads, and support requests have been moved to http://forums.nesl247.org/viewforum.php?f=18

2.6.20-viper9: Photon Torpedoes

Code: Select all

/* Patch List */ 
patch-2.6.20.7
archlinux-patches-rollup-2.6.20.patch
genpatches-rollup-2.6.20-7.patch
2.6.20.7-sd-0.46-1.patch
2.6.20.7-sd-0.46-1-boost.patch
2.6.20.7-sd-0.46-1-boost-tunable.patch
ck-desktop-rollup-2.6.20.7.patch
cpu-support-rollup.patch
rcu-preempt-rollup-2.6.20.patch
adaptive-readahead-2.6.20.patch
hrt-dyntick-2.6.20.patch
realtime-lsm.patch
suspend2-2.2.9.12-for-2.6.20.patch
linux-phc-0.2.9-for-2.6.20.patch
thinkpad-for-2.6.20-2.patch
fs-partitions-cleanup.patch
fs-ext4-2.6.20-git3.patch
fs-reiser4-for-2.6.20.patch
fs-unionfs-2.0-for-2.6.20.patch
driver-rtc-framework-for-cmos-rtcs-2.6.20.patch
driver-sdhci-handle-resumed-devices-2.6.20.patch
decrease-isn-wraparound-time.patch
tulip-fixes.patch
warning-fixes-2.patch
cleanups-from-mm.patch
daconfig-2.3.1-2.6.20.patch
viper-version.patch

Enjoy!

Patch:
http://vipernicus.org/files/patches/2.6 ... viper9.bz2

Ebuild:
http://vipernicus.org/files/patches/2.6 ... -r9.ebuild

Broken-Out:
http://vipernicus.org/files/patches/2.6 ... ut.tar.bz2

Broken-Out (Directory):
http://vipernicus.org/files/patches/2.6 ... roken-out/


Need Help?
Kernel How-to


If you want to chat with us, we can now be found in #nesl247 on irc.freenode.org.

[_sys/sid]

I've just rebooted... and all seems working without problem

Code: Select all

Linux hashley 2.6.20-viper9 #1 SMP PREEMPT Tue Apr 24 19:19:27 CEST 2007 i686 Intel(R) Pentium(R) 4 CPU 2.60GHz GenuineIntel GNU/Linux

Thanks vipernicus!

[vipernicus]

I've just rebooted... and all seems working without problem

Code: Select all

Linux hashley 2.6.20-viper9 #1 SMP PREEMPT Tue Apr 24 19:19:27 CEST 2007 i686 Intel(R) Pentium(R) 4 CPU 2.60GHz GenuineIntel GNU/Linux

Thanks vipernicus!

Did you try out the SD_BOOST? I recommend -10. Renices more than X.

[buddabrod]

I'm using it right now and it everything is working, including SD_BOOST^^

[immudium]

Yeah SD_BOOST is great as I have to admit that I was a little unsure how to properly initialize the nice value for X on restart so I wasn't really using that feature until now. But it's really working great so far. And I'm just getting everything started on my desktop again after the reboot, but so far it feels much more snappy and responsive. Thanks vipernicus!

[santaclaws]

Yes. Another release.


But how do I reice X permanently e.g. on boot?

[vipernicus]

Yes. Another release.


But how do I reice X permanently e.g. on boot?

You can either do it in KDM/GDM or you can do it with the SD_BOOST Kconfig setting.

[santaclaws]

Yes. Another release.


But how do I reice X permanently e.g. on boot?

You can either do it in KDM/GDM or you can do it with the SD_BOOST Kconfig setting.

WOW. Thanks for the quick response, vipernicus.

I am just emergeing viper9 - lets see what comes out.

Greez

[AaronPPC]

Works great here! SD_BOOST=-10.

[santaclaws]

AAARGHHHH! BUG??

Code: Select all

localhost linux # genkernel --kernel-cc=/usr/lib/ccache/bin/gcc --xconfig all
* Gentoo Linux Genkernel; Version 3.4.8
* Running with options: --kernel-cc=/usr/lib/ccache/bin/gcc --xconfig all

* Linux Kernel 2.6.20-viper9 for x86...
* kernel: >> Running mrproper...
* config: Using config from /usr/share/genkernel/x86/kernel-config-2.6
*         Previous config backed up to .config.bak
*         >> Running oldconfig...
* kernel: >> Cleaning...
* config: >> Invoking xconfig...
*         >> Compiling 2.6.20-viper9 bzImage...
* ERROR: Failed to compile the "bzImage" target...

* -- Grepping log... --

  SHIPPED scripts/kconfig/lex.zconf.c
  SHIPPED scripts/kconfig/zconf.hash.c
  HOSTCC  scripts/kconfig/zconf.tab.o
  HOSTLD  scripts/kconfig/conf
scripts/kconfig/conf -o arch/i386/Kconfig
.config:210:warning: trying to assign nonexistent symbol REGPARM
.config:213:warning: trying to assign nonexistent symbol HZ_250
.config:303:warning: trying to assign nonexistent symbol X86_SPEEDSTEP_CENTRINO_TABLE
.config:403:warning: trying to assign nonexistent symbol IP_ROUTE_FWMARK
.config:526:warning: trying to assign nonexistent symbol IP_NF_MATCH_HASHLIMIT
.config:2175:warning: trying to assign nonexistent symbol SND_AC97_BUS
.config:2367:warning: trying to assign nonexistent symbol USB_HIDINPUT
.config:2841:warning: trying to assign nonexistent symbol UNWIND_INFO
--
    2. MMConfig (PCI_GOMMCONFIG)
    3. Direct (PCI_GODIRECT)
  > 4. Any (PCI_GOANY)
  choice[1-4?]: 4
  PCI Express support (PCIEPORTBUS) [Y/n/?] y
    Root Port Advanced Error Reporting support (PCIEAER) [Y/n/?] y
--
SCSI media changer support (CHR_DEV_SCH) [M/n/y/?] m
*
* Some SCSI devices (e.g. CD jukebox) support multiple LUNs
*
Probe all LUNs on each SCSI device (SCSI_MULTI_LUN) [N/y/?] n
Verbose SCSI error reporting (kernel size +=12K) (SCSI_CONSTANTS) [N/y/?] n
--
  IP-over-InfiniBand (INFINIBAND_IPOIB) [M/n/?] m
    IP-over-InfiniBand debugging (INFINIBAND_IPOIB_DEBUG) [N/y/?] n
  InfiniBand SCSI RDMA Protocol (INFINIBAND_SRP) [M/n/?] m
  iSCSI Extensions for RDMA (iSER) (INFINIBAND_ISER) [M/n/?] m
*
* EDAC - error detection and reporting (RAS) (EXPERIMENTAL)
*
EDAC core system error reporting (EXPERIMENTAL) (EDAC) [N/m/y/?] n
--
/usr/qt/3/bin/moc -i scripts/kconfig/qconf.h -o scripts/kconfig/qconf.moc
  HOSTCC  scripts/kconfig/kconfig_load.o
  HOSTCXX scripts/kconfig/qconf.o
  HOSTLD  scripts/kconfig/qconf
scripts/kconfig/qconf arch/i386/Kconfig
Session management error: Authentication Rejected, reason : None of the authentication protocols specified are supported and host-based authentication failed
--
  LD      fs/xfs/built-in.o
  LD      fs/built-in.o
* Gentoo Linux Genkernel; Version 3.4.8
* Running with options: --kernel-cc=/usr/lib/ccache/bin/gcc --xconfig all

* ERROR: Failed to compile the "bzImage" target...

* -- End log... --

* Please consult /var/log/genkernel.log for more information and any
* errors that were reported above.

* Report any genkernel bugs to bugs.gentoo.org and
* assign your bug to genkernel@gentoo.org. Please include
* as much information as you can in your bug report; attaching
* /var/log/genkernel.log so that your issue can be dealt with effectively.
* 
* Please do *not* report compilation failures as genkernel bugs!

Any hint?

[vipernicus]

AAARGHHHH! BUG??

Code: Select all

scripts/kconfig/qconf arch/i386/Kconfig
Session management error: Authentication Rejected, reason : None of the authentication protocols specified are supported and host-based authentication failed

Any hint?

Same as before, authentication issue, try using menuconfig and it will work. If you are using sudo, switch to su, or sux.

[genestyler]

hum another bug ? ;o)

Code: Select all

* ERROR: Failed to compile the "modules" target...

* -- Grepping log... --

  SHIPPED scripts/kconfig/lex.zconf.c
  SHIPPED scripts/kconfig/zconf.hash.c
  HOSTCC  scripts/kconfig/zconf.tab.o
  HOSTLD  scripts/kconfig/conf
scripts/kconfig/conf -o arch/i386/Kconfig
.config:210:warning: trying to assign nonexistent symbol REGPARM
.config:213:warning: trying to assign nonexistent symbol HZ_250
.config:303:warning: trying to assign nonexistent symbol X86_SPEEDSTEP_CENTRINO_TABLE
.config:403:warning: trying to assign nonexistent symbol IP_ROUTE_FWMARK
.config:526:warning: trying to assign nonexistent symbol IP_NF_MATCH_HASHLIMIT
.config:2175:warning: trying to assign nonexistent symbol SND_AC97_BUS
.config:2367:warning: trying to assign nonexistent symbol USB_HIDINPUT
.config:2841:warning: trying to assign nonexistent symbol UNWIND_INFO
--
    2. MMConfig (PCI_GOMMCONFIG)
    3. Direct (PCI_GODIRECT)
  > 4. Any (PCI_GOANY)
  choice[1-4?]: 4
  PCI Express support (PCIEPORTBUS) [Y/n/?] y
    Root Port Advanced Error Reporting support (PCIEAER) [Y/n/?] y
--
SCSI media changer support (CHR_DEV_SCH) [M/n/y/?] m
*
* Some SCSI devices (e.g. CD jukebox) support multiple LUNs
*
Probe all LUNs on each SCSI device (SCSI_MULTI_LUN) [N/y/?] n
Verbose SCSI error reporting (kernel size +=12K) (SCSI_CONSTANTS) [N/y/?] n
--
  IP-over-InfiniBand (INFINIBAND_IPOIB) [M/n/?] m
    IP-over-InfiniBand debugging (INFINIBAND_IPOIB_DEBUG) [N/y/?] n
  InfiniBand SCSI RDMA Protocol (INFINIBAND_SRP) [M/n/?] m
  iSCSI Extensions for RDMA (iSER) (INFINIBAND_ISER) [M/n/?] m
*
* EDAC - error detection and reporting (RAS) (EXPERIMENTAL)
*
EDAC core system error reporting (EXPERIMENTAL) (EDAC) [N/m/y/?] n
--
COMMAND: make j1  menuconfig
* kernel: >> Making dependencies...
COMMAND: make -j3 CC="gcc" LD="ld" AS="as"
  HOSTLD  scripts/kconfig/conf
scripts/kconfig/conf -s arch/i386/Kconfig
*** Warning: make dep is unnecessary now.
--
* Running with options: all
* Clearing cache dir contents from /var/cache/genkernel/x86
*        >> removing busybox-1.1.3+gentoo-x86.bz2
*        >> removing insmod-x86-static-2.6.bz2

* ERROR: Failed to compile the "modules" target...

* -- End log... --

[vipernicus]

Try building WITHOUT genkernel, since the errors that are being posted tell us nothing.

[Donman]

Boots fine here, but the timestamp issues persist. When I boot, it does not complain about file time modifications in the future, but some of the files that I check say they were last accessed or modified in the future. For example the output of "ls -l --time=atime":

Code: Select all

 ls --time=atime -l
total 164
drwxr-xr-x 4 sathaten sathaten  4096 Aug  9  2415 Development
-rw-r--r-- 1 sathaten sathaten 39526 Mar  1 23:19 Gonna_get_raped_cat.jpg
drwxr-xr-x 2 sathaten sathaten 16384 Apr 19 21:54 Music
drwxr-xr-x 8 sathaten sathaten  4096 Apr 19 21:57 Pictures
-rw-r--r-- 1 root     root         0 Apr 24 17:03 mat
drwxr-xr-x 2 sathaten sathaten  4096 Aug  9  2415 mat habbo layout
drwxr-xr-x 2 sathaten sathaten  4096 Apr 19 21:53 misc
drwxr-xr-x 2 sathaten sathaten  4096 Apr 24 13:37 movie
-rwxr-xr-x 1 sathaten sathaten 77069 Feb  7 11:18 poster.jpg
drwxr-xr-x 2 sathaten sathaten  4096 May 27  2143 quickplay
drwxr-xr-x 6 sathaten sathaten  4096 Aug  9  2415 schoolwork
drwxr-xr-x 2 sathaten sathaten  4096 Apr 19 22:05 wallpaper

This issue seems to also affect Python as well, as emerge spits this error when I run it:

Code: Select all

sathaten@Zarakwi ~ $ emerge
'import site' failed; use -v for traceback
Traceback (most recent call last):
  File "/usr/bin/emerge", line 23, in <module>
    import os, stat
  File "/usr/lib64/python2.5/os.py", line 49, in <module>
    import posixpath as path
  File "/usr/lib64/python2.5/posixpath.py", line 14, in <module>
    import stat
OverflowError: modification time overflows a 4 byte field

So I suppose I am back to gentoo-sources, but if you have anything you would like me to try out to help solve the issue I will be more than willing to do so.

[Rion]

compiled fine but i have the same error on boot as here

PS viper7 boots fine

[vipernicus]

compiled fine but i have the same error on boot as here

PS viper7 boots fine

http://forums.gentoo.org/viewtopic-p-40 ... ml#4018567

[Rion]

http://forums.gentoo.org/viewtopic-p-40 ... ml#4018567

i recompiled kernel without modules and initrd with this patch. doesn't help
in some hours i'll recompile it with modules and post results here

[jespera]

Enjoy!

I compiled the kernel and modules and am now running a nice new kernel on my Core Due laptop.

Code: Select all

Linux jmachine 2.6.20-viper9 #1 SMP PREEMPT Wed Apr 25 09:29:35 CEST 2007 i686 Genuine Intel(R) CPU T2400 @ 1.83GHz GenuineIntel GNU/Linux

However, when I run the gnome system monitor (the one with the nice colored graphs) I see that the second core (CPU2) is always at 0% usage. This was not the case for viper7! (in fact I now see that I have the same 0% on CPU2 using viper8).

Any hints as to what I can do to make use of both cpu cores? (here's my kernel config.gz: http://www.jel-lej.dk/download/config.gz.

Anyone have the same problem?

Oh! Right and thanks for making the viper kernels in the first place!

[deno]

I have had the same situation.

http://forums.gentoo.org/viewtopic-t-55 ... ight-.html

Maybe you could try different scheduler. In my case it is not that second cpu is not recognized, and actually its activity goes up to 1-2 % sometimes but it is seldom. And yes I almos forgot, frequency policy worked only on first cpu. Second keeps policy which it gets when cpufrequtils starts first time.

EDIT:

Should said when cpufrequtils starts . It starts only once.

[jespera]

I have had the same situation.

http://forums.gentoo.org/viewtopic-t-55 ... ight-.html

Maybe you could try different scheduler. In my case it is not that second cpu is not recognized, and actually its activity goes up to 1-2 % sometimes but it is seldom. And yes I almos forgot, frequency policy worked only on first cpu. Second keeps policy which it get when cpufrequtils starts first time.

I tried that with viper8. I had the choice between three: ingosched, staircase, and sd. Neither would end up with the second core being utilized according to "top" and "gnome-system-monitor".

I copied the .config from my viper7 directly and did "make oldconfig". Also according to dmesg both cpus are activated.

If all fails I can go back to viper7, but it seems like it would be worth it to find the error.

[charlieg]

I'm impressed at the longevity of this patchset... how long have you been doing this now?

[vipernicus]

I'm impressed at the longevity of this patchset... how long have you been doing this now?

2.6.12-dark2: Learn To Swim - Jun 25, 2005
2.6.14-rc5-viper1 "Mango Madness" - Oct 24, 2005
2.6.16-emission1: Mission to the Horsehead Nebulae - May 10, 2006

I guess I started -viper in late 2005.

[vipernicus]

Any hints as to what I can do to make use of both cpu cores? (here's my kernel config.gz: http://www.jel-lej.dk/download/config.gz.

This may be a suspend2 bug. Try building your kernel WITHOUT suspend2. If this fixes your problem, I think an upgrade to suspend2-2.2.9.13 will fix your problem. This may be one of the SMP bugs of suspend2-2.2.9.10-2.2.9.12.

[jespera]

Any hints as to what I can do to make use of both cpu cores? (here's my kernel config.gz: http://www.jel-lej.dk/download/config.gz.

This may be a suspend2 bug. Try building your kernel WITHOUT suspend2. If this fixes your problem, I think an upgrade to suspend2-2.2.9.13 will fix your problem. This may be one of the SMP bugs of suspend2-2.2.9.10-2.2.9.12.

Indeed! Disabling suspend2 gave me back my second core.

I will report back whether upgrading to suspend2-2.2.9.13 also fixes the problem asap.

Actually, as I am rather new to this custom patching of a kernel: what is the most easy way for me to do this?

Thanks!

[vipernicus]

Any hints as to what I can do to make use of both cpu cores? (here's my kernel config.gz: http://www.jel-lej.dk/download/config.gz.

This may be a suspend2 bug. Try building your kernel WITHOUT suspend2. If this fixes your problem, I think an upgrade to suspend2-2.2.9.13 will fix your problem. This may be one of the SMP bugs of suspend2-2.2.9.10-2.2.9.12.

Indeed! Disabling suspend2 gave me back my second core.

I will report back whether upgrading to suspend2-2.2.9.13 also fixes the problem asap.

Actually, as I am rather new to this custom patching of a kernel: what is the most easy way for me to do this?

Thanks!

Don't worry, I'll provide the hotfixes!