Gentoo Forums
Gentoo Forums
Gentoo Forums
Quick Search: in
2.6.36_TOI-plus kernel patch & patchset
View unanswered posts
View posts from last 24 hours

 
Reply to topic    Gentoo Forums Forum Index Unsupported Software
View previous topic :: View next topic  
Author Message
kernelOfTruth
Watchman
Watchman


Joined: 20 Dec 2005
Posts: 6111
Location: Vienna, Austria; Germany; hello world :)

PostPosted: Wed Oct 27, 2010 11:21 pm    Post subject: 2.6.36_TOI-plus kernel patch & patchset Reply with quote

Hi guys,

here's another revision of my patchset:

this time it's based on 2.6.36 with partial ck-patchset (the patch without BFS doesn't work with all ck-patches)

in a nutshell: tuxonice, ck-patches (partial / full-blown ck-patchset in combination with BFS), BFQ, several bugfixes and improvements (improved latency and reclaim efficiency; fsync performance for small files; load balancing for tasks with large weight differential; dm-crypt scaling to multiple CPUs [WIP so might worsen latency issues and interactivity during heavy writing] cutting transfer times half or reducing to 1/3 or even less (!); vfs, jbd, ext4 fixes)

hosting right now limited to multiupload.com (thanks guys !) - other mirrors will come up when I have more time ...

current/up-to-date files:



http://www.multiupload.com/4F6TGBM01Y <-- linux-2.6.36_TOI_plus-r2.patch (2.15 MB) (this one's for the folks out there where the reduced hangcheck timing doesn't work)

md5sum: 805ab792d76711d6bbb885e856d6f6c4
sha256sum: fe91076a6c72bac9ebae483abf599447a6f18461a53555f9036819edc3624b16

http://www.multiupload.com/LOG55UJILQ <-- linux-2.6.36_TOI_plus-r2_BFS.diff (615.6 KB) (this one's for the folks out there where the reduced hangcheck timing doesn't work; corresponding patch to upgrade to BFS)

md5sum: 659aeb3cc46884fe8fa03586145309a4
sha256sum: 8e284b0a868dc194d4c9e875401c64218c693571d14bc78b2b75559a57f116e4




http://www.multiupload.com/VD6J3YFBJM <-- linux-2.6.36_TOI_plus-r2_BFS_full.patch (2.75 MB) ] (full patchset with full ck-patchset, including 2.6.36-sched-bfs-357-1.patch); (without reduced hangcheck since it seems to be broken for some folks)

md5sum: 21a52f8d7f9b2d9db0fd6131054f3288
sha256sum: 0d8a5a71d9ee1b1bb08e406bd7bc32d0e5774577df369820245a449bcec0ec3a


http://www.multiupload.com/MJFMF8J2KY <-- 2.6.36_TOI_plus_broken-out.tbz2 (382.27 KB) broken out patches of this patchset, some patches may need to be corrected due to rejects (they are vanilla/unmodified)

md5sum: 4e1cbc216b5790bd067878cac807144a
sha256sum: 84581c72881e7bf135e5061850d1e390f84093bb6a6889ed204cc67596841b1f


as options you can add (apply those patches):

* (folder 20): glibc's memcpy - that might speed up operation of btrfs and other filesystems that make heavy use of it

* (folder 21): enhanced NX protection for kernel data [caveat: this might break tuxonice and suspend-to-ram]


more descriptions coming up ...



broken files: (need additional patches/fixes)

http://www.multiupload.com/9RLQSU637X <-- linux-2.6.36_TOI_plus.patch (2.15 MB)

md5sum: 8df020383b704062ee44d1ce64ac519b
sha256sum: b509dd294616a5d331cd0c27b1f1342cffccf75b0541139728ebdd46582f517f


needs:
Code:
diff -Nrupad linux-2.6.36_TOI_plus_bak//include/linux/sched.h linux-2.6.36_TOI_plus_bak_fix1//include/linux/sched.h
--- linux-2.6.36_TOI_plus_bak//include/linux/sched.h   2010-10-28 17:14:55.943182964 +0200
+++ linux-2.6.36_TOI_plus_bak_fix1//include/linux/sched.h   2010-10-28 17:17:38.369182963 +0200
@@ -38,6 +38,7 @@
 #define SCHED_BATCH      3
 /* SCHED_ISO: reserved but not implemented yet */
 #define SCHED_IDLE      5
+#define SCHED_IDLEPRIO      SCHED_IDLE
 /* Can be ORed in to make sure the process is reverted back to SCHED_NORMAL on fork */
 #define SCHED_RESET_ON_FORK     0x40000000
 


http://www.multiupload.com/XMWR2VNNR3 <-- linux-2.6.36_TOI_plus_BFS.diff (614.83 KB) [this adds the remaining ck-patches and BFS to the patchset]

md5sum: a90738fed43692af1380a785ce407505
sha256sum: 5d16e253560874477ccb0448510318d667837aa854dedf3f14c83343437715eb



older stuff:



http://www.multiupload.com/61LJ5UEN4I <-- linux-2.6.36_TOI_plus-r1.patch (2.15 MB) (with reduced hangcheck)

md5sum: 54d4b07f3dd7606c7e83a9080e6d7e50
sha256sum: e142d275f870d239ddcee05ac90c4247ae2f78bab54c5364cb38ee11c745a8bd

http://www.multiupload.com/HXBTLCOIS0 <-- linux-2.6.36_TOI_plus-r1_BFS.diff (615.6 KB) (with reduced hangcheck)

md5sum: 63351f392838ba97d5b386170a1b5f31
sha256sum: 1b7f9248c225974f69bd56c19a24a088fa3b3d6ce871e5b50e135ef3b3e93a7c


http://www.multiupload.com/9P9WVH6BP8 <-- linux-2.6.36_TOI_plus_BFS_full.patch (2.74 MB) (full patchset with full ck-patchset, including 2.6.36-sched-bfs-357-1.patch) [with reduced hangcheck]

md5sum: c72d0782514bc2296e3a58f30cc16fd5
sha256sum: 848d99c9160a375d56d741447e4231cc73a560f131de39cd10a6365625374ef2


_________________
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 :D


Last edited by kernelOfTruth on Fri Oct 29, 2010 9:40 pm; edited 13 times in total
Back to top
View user's profile Send private message
kernelOfTruth
Watchman
Watchman


Joined: 20 Dec 2005
Posts: 6111
Location: Vienna, Austria; Germany; hello world :)

PostPosted: Thu Oct 28, 2010 1:21 am    Post subject: Reply with quote

2.6.36 fix for reiserfs users that need quotas: [PATCH] reiserfs: Fix locking in reiserfs_quota_on()
_________________
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 :D
Back to top
View user's profile Send private message
pross
Apprentice
Apprentice


Joined: 30 Jul 2003
Posts: 156
Location: Ireland

PostPosted: Thu Oct 28, 2010 2:36 pm    Post subject: Reply with quote

build error:
Code:
   CHK     include/linux/version.h
  CHK     include/generated/utsrelease.h
  CALL    scripts/checksyscalls.sh
  CHK     include/generated/compile.h
  CC      mm/vmscan.o
mm/vmscan.c: In function ‘effective_sc_prio’:
mm/vmscan.c:1875: error: ‘SCHED_IDLEPRIO’ undeclared (first use in this function)
mm/vmscan.c:1875: error: (Each undeclared identifier is reported only once
mm/vmscan.c:1875: error: for each function it appears in.)
make[1]: *** [mm/vmscan.o] Error 1
make: *** [mm] Error 2


using linux-2.6.36_TOI_plus.patch

and with the second patch added:
Code:
  CC      block/blk-exec.o
block/blk-exec.c: In function ‘blk_execute_rq’:
block/blk-exec.c:101: error: ‘sysctl_hung_task_timeout_secs’ undeclared (first use in this function)
block/blk-exec.c:101: error: (Each undeclared identifier is reported only once
block/blk-exec.c:101: error: for each function it appears in.)
make[1]: *** [block/blk-exec.o] Error 1
make: *** [block] Error 2


my config (its a basic pappy seed) http://paste.pocoo.org/show/282686/
_________________
My Overlay [pross] | bfs-sources
Back to top
View user's profile Send private message
kernelOfTruth
Watchman
Watchman


Joined: 20 Dec 2005
Posts: 6111
Location: Vienna, Austria; Germany; hello world :)

PostPosted: Thu Oct 28, 2010 3:22 pm    Post subject: Reply with quote

sorry !

the following fix goes on top of linux-2.6.36_TOI_plus.patch:

Code:
diff -Nrupad linux-2.6.36_TOI_plus_bak//include/linux/sched.h linux-2.6.36_TOI_plus_bak_fix1//include/linux/sched.h
--- linux-2.6.36_TOI_plus_bak//include/linux/sched.h   2010-10-28 17:14:55.943182964 +0200
+++ linux-2.6.36_TOI_plus_bak_fix1//include/linux/sched.h   2010-10-28 17:17:38.369182963 +0200
@@ -38,6 +38,7 @@
 #define SCHED_BATCH      3
 /* SCHED_ISO: reserved but not implemented yet */
 #define SCHED_IDLE      5
+#define SCHED_IDLEPRIO      SCHED_IDLE
 /* Can be ORed in to make sure the process is reverted back to SCHED_NORMAL on fork */
 #define SCHED_RESET_ON_FORK     0x40000000
 



I'll re-create the BFS-diff patch

meanwhile you could try out the full BFS patch (that should work)

thanks for trying out/using !
_________________
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 :D
Back to top
View user's profile Send private message
kernelOfTruth
Watchman
Watchman


Joined: 20 Dec 2005
Posts: 6111
Location: Vienna, Austria; Germany; hello world :)

PostPosted: Thu Oct 28, 2010 6:03 pm    Post subject: Reply with quote

a word of warning for those interactivity addicts out there:

if you want a responsive system even under heavy I/O - you better revert the patchset:

dm-crypt scaling to multiple CPUs (2 patches)


so you either have the option of throughput (2-4 times better while using several processors or even more) or interactivity/latency


this issue should be fixed in the upcoming 2.6.37 or at latest 2.6.38 kernel
_________________
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 :D
Back to top
View user's profile Send private message
pross
Apprentice
Apprentice


Joined: 30 Jul 2003
Posts: 156
Location: Ireland

PostPosted: Thu Oct 28, 2010 6:07 pm    Post subject: Reply with quote

still failing with r1 patches:

steps to reproduce:

untar 2.6.36
patch -p1 < ~/linux-2.6.36_TOI_plus-r1.patch
make alldefconfig
make

fails too if i add the second bfs patch.

Code:
block/blk-exec.c: In function ‘blk_execute_rq’:
block/blk-exec.c:101: error: ‘sysctl_hung_task_timeout_secs’ undeclared (first use in this function)
block/blk-exec.c:101: error: (Each undeclared identifier is reported only once
block/blk-exec.c:101: error: for each function it appears in.)
make[1]: *** [block/blk-exec.o] Error 1
make: *** [block] Error 2
make: *** Waiting for unfinished jobs....


edit: just tried full patch, fails same error
_________________
My Overlay [pross] | bfs-sources
Back to top
View user's profile Send private message
kernelOfTruth
Watchman
Watchman


Joined: 20 Dec 2005
Posts: 6111
Location: Vienna, Austria; Germany; hello world :)

PostPosted: Thu Oct 28, 2010 6:48 pm    Post subject: Reply with quote

pross wrote:
still failing with r1 patches:

steps to reproduce:

untar 2.6.36
patch -p1 < ~/linux-2.6.36_TOI_plus-r1.patch
make alldefconfig
make

fails too if i add the second bfs patch.

Code:
block/blk-exec.c: In function ‘blk_execute_rq’:
block/blk-exec.c:101: error: ‘sysctl_hung_task_timeout_secs’ undeclared (first use in this function)
block/blk-exec.c:101: error: (Each undeclared identifier is reported only once
block/blk-exec.c:101: error: for each function it appears in.)
make[1]: *** [block/blk-exec.o] Error 1
make: *** [block] Error 2
make: *** Waiting for unfinished jobs....


edit: just tried full patch, fails same error


thanks for reporting - hm, why is it failing for everyone except me ? 8O

you guys must be enabling something that I don't use/triggers this error:

please revert the following patch:

[PATCH] block: Prevent hang_check firing during long I_O.txt

Code:
During long I/O operations, the hang_check timer may fire,
trigger stack dumps that unnecessarily alarm the user.

Eg.  hdparm --security-erase NULL /dev/sdb  ## can take *hours* to complete

So, if hang_check is armed, we should wake up periodically
to prevent it from triggering.  This patch uses a wake-up interval
equal to half the hang_check timer period, which keeps overhead low enough.

Signed-off-by: Mark Lord <mlord@pobox.com>

--- old/block/blk-exec.c   2010-09-20 19:56:53.000000000 -0400
+++ linux/block/blk-exec.c   2010-09-24 09:43:32.342604574 -0400
@@ -80,6 +80,7 @@
    DECLARE_COMPLETION_ONSTACK(wait);
    char sense[SCSI_SENSE_BUFFERSIZE];
    int err = 0;
+   unsigned long hang_check;
 
    /*
     * we need an extra reference to the request, so we can look at
@@ -95,7 +96,13 @@
 
    rq->end_io_data = &wait;
    blk_execute_rq_nowait(q, bd_disk, rq, at_head, blk_end_sync_rq);
-   wait_for_completion(&wait);
+
+   /* Prevent hang_check timer from firing at us during very long I/O */
+   hang_check = sysctl_hung_task_timeout_secs;
+   if (hang_check)
+      while (!wait_for_completion_timeout(&wait, hang_check * (HZ/2)));
+   else
+      wait_for_completion(&wait);
 
    if (rq->errors)
       err = -EIO;


--
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

_________________
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 :D
Back to top
View user's profile Send private message
pross
Apprentice
Apprentice


Joined: 30 Jul 2003
Posts: 156
Location: Ireland

PostPosted: Thu Oct 28, 2010 7:17 pm    Post subject: Reply with quote

Code:
pross@dhcppc0 ~/kernel-testing/linux-2.6.36 $ patch -R -p1 < ../revert.patch
patching file block/blk-exec.c
Hunk #1 FAILED at 80.
Hunk #2 FAILED at 95.
2 out of 2 hunks FAILED -- saving rejects to file block/blk-exec.c.rej

_________________
My Overlay [pross] | bfs-sources
Back to top
View user's profile Send private message
kernelOfTruth
Watchman
Watchman


Joined: 20 Dec 2005
Posts: 6111
Location: Vienna, Austria; Germany; hello world :)

PostPosted: Thu Oct 28, 2010 7:41 pm    Post subject: Reply with quote

pross wrote:
Code:
pross@dhcppc0 ~/kernel-testing/linux-2.6.36 $ patch -R -p1 < ../revert.patch
patching file block/blk-exec.c
Hunk #1 FAILED at 80.
Hunk #2 FAILED at 95.
2 out of 2 hunks FAILED -- saving rejects to file block/blk-exec.c.rej


huh ?

works here:
Quote:
patch -p1 -R --verbose --dry-run < /home/kernelOfTruth/bak/gentoo/kernel-development/2.6.36_TOI_plus_broken-out/9_hang_check\ long\ i_o/\[PATCH\]\ block\:\ Prevent\ hang_check\ firing\ during\ long\ I_O.txt
Hmm... Looks like a unified diff to me...
The text leading up to this was:
--------------------------
|During long I/O operations, the hang_check timer may fire,
|trigger stack dumps that unnecessarily alarm the user.
|
|Eg. hdparm --security-erase NULL /dev/sdb ## can take *hours* to complete
|
|So, if hang_check is armed, we should wake up periodically
|to prevent it from triggering. This patch uses a wake-up interval
|equal to half the hang_check timer period, which keeps overhead low enough.
|
|Signed-off-by: Mark Lord <mlord@pobox.com>
|
|--- old/block/blk-exec.c 2010-09-20 19:56:53.000000000 -0400
|+++ linux/block/blk-exec.c 2010-09-24 09:43:32.342604574 -0400
--------------------------
Patching file block/blk-exec.c using Plan A...
Hunk #1 succeeded at 80.
Hunk #2 succeeded at 95.
Hmm...patch unexpectedly ends in middle of line
Ignoring the trailing garbage.
done

_________________
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 :D
Back to top
View user's profile Send private message
pross
Apprentice
Apprentice


Joined: 30 Jul 2003
Posts: 156
Location: Ireland

PostPosted: Thu Oct 28, 2010 7:59 pm    Post subject: Reply with quote

i applied the 2 smaller patches, revert fails, i tried the big patch , revert fails...
_________________
My Overlay [pross] | bfs-sources
Back to top
View user's profile Send private message
kernelOfTruth
Watchman
Watchman


Joined: 20 Dec 2005
Posts: 6111
Location: Vienna, Austria; Germany; hello world :)

PostPosted: Thu Oct 28, 2010 8:01 pm    Post subject: Reply with quote

pross wrote:
i applied the 2 smaller patches, revert fails, i tried the big patch , revert fails...


I just did the same - both times it worked for me :o

*creates another patch with hang-check removed*

edit:

hope it works for ya now:

http://www.multiupload.com/4F6TGBM01Y <-- linux-2.6.36_TOI_plus-r2.patch (2.15 MB) (this one's for the folks out there where the reduced hangcheck timing doesn't work)

md5sum: 805ab792d76711d6bbb885e856d6f6c4
sha256sum: fe91076a6c72bac9ebae483abf599447a6f18461a53555f9036819edc3624b16

http://www.multiupload.com/LOG55UJILQ <-- linux-2.6.36_TOI_plus-r2_BFS.diff (615.6 KB) (this one's for the folks out there where the reduced hangcheck timing doesn't work; corresponding patch to upgrade to BFS)

md5sum: 659aeb3cc46884fe8fa03586145309a4
sha256sum: 8e284b0a868dc194d4c9e875401c64218c693571d14bc78b2b75559a57f116e4
_________________
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 :D
Back to top
View user's profile Send private message
pross
Apprentice
Apprentice


Joined: 30 Jul 2003
Posts: 156
Location: Ireland

PostPosted: Thu Oct 28, 2010 8:35 pm    Post subject: Reply with quote

its building properly now :) shall i add it the ebuild to my overlay?
_________________
My Overlay [pross] | bfs-sources
Back to top
View user's profile Send private message
kernelOfTruth
Watchman
Watchman


Joined: 20 Dec 2005
Posts: 6111
Location: Vienna, Austria; Germany; hello world :)

PostPosted: Thu Oct 28, 2010 8:50 pm    Post subject: Reply with quote

pross wrote:
its building properly now :) shall i add it the ebuild to my overlay?


yes please :)
_________________
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 :D
Back to top
View user's profile Send private message
pross
Apprentice
Apprentice


Joined: 30 Jul 2003
Posts: 156
Location: Ireland

PostPosted: Thu Oct 28, 2010 11:41 pm    Post subject: Reply with quote

Added, feel free to test
_________________
My Overlay [pross] | bfs-sources
Back to top
View user's profile Send private message
wrc1944
Advocate
Advocate


Joined: 15 Aug 2002
Posts: 3260
Location: Gainesville, Florida

PostPosted: Fri Oct 29, 2010 9:42 am    Post subject: Reply with quote

Hmm. I get this when patching. Will the failures actually matter, or should I forget the xconfig qt4 patch and try again? Qt4 patch works fine with zen and ck on recent versions. (Just can't live without xconfig, and having qt3 installed presents other problems)

I've been having weird problems with 2.6.36-ck1, both portage zen-sources and a manually built version. They both compile, but when trying to boot from grub they both will randomly get a kernel panic not syncing, unable to mout root fs on unknown-block(8,3), sr0 detection, and swapper not tainted problems. Memtest+ says all my ram is fine.

Curiously, both will sometimes boot normally after a few tries (or switching to the other one), and function fine. Anyway, here's my patching attempt:
Code:
wrc@gentoo /mnt/data/kern/linux-2.6.36 $ patch -p1 <2900_xconfig-with-qt4.patch
patching file scripts/kconfig/Makefile
patching file scripts/kconfig/qconf.cc
patching file scripts/kconfig/qconf.h
patching file scripts/kconfig/Makefile
wrc@gentoo /mnt/data/kern/linux-2.6.36 $ patch -p1 <linux-2.6.36_TOI_plus-r2_BFS.diff
patching file arch/powerpc/platforms/cell/spufs/sched.c
patching file arch/x86/Kconfig
patching file arch/x86/Kconfig.orig
patching file arch/x86/kernel/cpu/proc.c
patching file arch/x86/kernel/smpboot.c
patching file arch/x86/kernel/smpboot.c.orig
patching file Documentation/scheduler/sched-BFS.txt
patching file Documentation/sysctl/kernel.txt
patching file drivers/cpufreq/cpufreq_ondemand.c
Hunk #1 FAILED at 28.
Hunk #2 succeeded at 482 (offset -31 lines).
1 out of 2 hunks FAILED -- saving rejects to file drivers/cpufreq/cpufreq_ondemand.c.rej
patching file drivers/cpufreq/cpufreq_ondemand.c~
patching file fs/proc/base.c
patching file include/linux/init_task.h
patching file include/linux/ioprio.h
patching file include/linux/mmzone.h
Hunk #4 succeeded at 638 (offset -8 lines).
The next patch would create the file include/linux/mmzone.h.orig,
which already exists!  Assume -R? [n] n
Apply anyway? [n] y
patching file include/linux/mmzone.h.orig
Hunk #1 FAILED at 1.
1 out of 1 hunk FAILED -- saving rejects to file include/linux/mmzone.h.orig.rej
patching file include/linux/nfsd/stats.h
patching file include/linux/sched.h
Hunk #1 FAILED at 36.
Hunk #2 succeeded at 142 (offset -1 lines).
Hunk #3 succeeded at 267 (offset -1 lines).
Hunk #4 succeeded at 1166 (offset -1 lines).
Hunk #5 succeeded at 1285 (offset -1 lines).
Hunk #6 succeeded at 1510 (offset -1 lines).
Hunk #7 succeeded at 1582 (offset -1 lines).
Hunk #8 succeeded at 1909 (offset -1 lines).
1 out of 8 hunks FAILED -- saving rejects to file include/linux/sched.h.rej
patching file include/linux/sched.h.orig
patching file include/net/inet_timewait_sock.h
patching file init/calibrate.c
patching file init/Kconfig
patching file init/main.c
Hunk #1 succeeded at 828 (offset -1 lines).
The next patch would create the file init/main.c.orig,
which already exists!  Assume -R? [n] n
Apply anyway? [n] y
patching file init/main.c.orig
Hunk #1 FAILED at 1.
1 out of 1 hunk FAILED -- saving rejects to file init/main.c.orig.rej
patching file kernel/delayacct.c
patching file kernel/exit.c
patching file kernel/Kconfig.hz
patching file kernel/Kconfig.preempt
patching file kernel/kthread.c
patching file kernel/posix-cpu-timers.c
patching file kernel/sched_bfs.c
patching file kernel/sched_bfs.c.orig
patching file kernel/sched.c
Hunk #2 succeeded at 9191 (offset -2 lines).
patching file kernel/sched.c.orig
Hunk #1 succeeded at 2003 (offset 1 line).
patching file kernel/sysctl.c
patching file lib/Kconfig.debug
patching file Makefile
Hunk #1 FAILED at 1.
1 out of 1 hunk FAILED -- saving rejects to file Makefile.rej
patching file mm/vmscan.c
Hunk #2 succeeded at 128 (offset -6 lines).
Hunk #3 succeeded at 2302 (offset -126 lines).
Hunk #4 succeeded at 2354 (offset -126 lines).
Hunk #5 succeeded at 2536 (offset -140 lines).
patching file mm/vmscan.c.orig
Hunk #2 succeeded at 2302 (offset -126 lines).
Hunk #3 succeeded at 2354 (offset -126 lines).
Hunk #4 succeeded at 2422 (offset -128 lines).
Hunk #5 succeeded at 2482 (offset -130 lines).
Hunk #6 succeeded at 2512 (offset -130 lines).
Hunk #7 succeeded at 2546 (offset -130 lines).
wrc@gentoo /mnt/data/kern/linux-2.6.36 $

EDIT: Nope- leaving out the xconfig qt4 patch and i still get exactly the same failures. :(
_________________
Main box- AsRock x370 Gaming K4
Ryzen 1700, 3.0GHz, 16GB GSkill Flare DDR4 3200mhz
Samsung SATA 1000GB, Radeon HD R7 350 2GB DDR5
Gentoo ~amd64 plasma, glibc-2.30-r2, gcc-9.2.0 kernel-5.4.1-gentoo USE=experimental
Back to top
View user's profile Send private message
wrc1944
Advocate
Advocate


Joined: 15 Aug 2002
Posts: 3260
Location: Gainesville, Florida

PostPosted: Fri Oct 29, 2010 11:43 am    Post subject: Reply with quote

OK- I just compiled a vanilla 2.6.36 patched with bfs-357.1 from Con's repo, and had no problems. Works nice. :D

The random booting problems with ck1 described above have disappeared (so far), so it must be something else in ck1 besides bfs.
_________________
Main box- AsRock x370 Gaming K4
Ryzen 1700, 3.0GHz, 16GB GSkill Flare DDR4 3200mhz
Samsung SATA 1000GB, Radeon HD R7 350 2GB DDR5
Gentoo ~amd64 plasma, glibc-2.30-r2, gcc-9.2.0 kernel-5.4.1-gentoo USE=experimental
Back to top
View user's profile Send private message
kernelOfTruth
Watchman
Watchman


Joined: 20 Dec 2005
Posts: 6111
Location: Vienna, Austria; Germany; hello world :)

PostPosted: Fri Oct 29, 2010 1:40 pm    Post subject: Reply with quote

of course you're getting the same errors :roll:

you need to apply on top of 2.6.36 - either:

linux-2.6.36_TOI_plus-r2.patch and then linux-2.6.36_TOI_plus-r2_BFS.diff

or

linux-2.6.36_TOI_plus_BFS_full.patch only (has the reduced hangcheck timing)


after that you'd get only the following rejects:

Quote:
patch -p1 --dry-run < /home/kernelOfTruth/Downloads/2900_xconfig-with-qt4.patch.txt
patching file scripts/kconfig/Makefile
Hunk #1 succeeded at 238 (offset 15 lines).
patching file scripts/kconfig/expr.h
Hunk #1 succeeded at 172 (offset 1 line).
patching file scripts/kconfig/qconf.h
Hunk #3 succeeded at 57 with fuzz 1 (offset 3 lines).
Hunk #4 succeeded at 67 (offset 3 lines).
Hunk #5 succeeded at 79 (offset 3 lines).
Hunk #6 succeeded at 87 (offset 3 lines).
Hunk #7 succeeded at 130 with fuzz 1 (offset 5 lines).
Hunk #8 FAILED at 135.
Hunk #9 succeeded at 207 (offset 6 lines).
Hunk #10 succeeded at 228 (offset 6 lines).
Hunk #11 FAILED at 253.
Hunk #12 succeeded at 283 (offset 7 lines).
Hunk #13 succeeded at 311 (offset 7 lines).
Hunk #14 succeeded at 319 (offset 7 lines).
Hunk #15 succeeded at 343 (offset 7 lines).
2 out of 15 hunks FAILED -- saving rejects to file scripts/kconfig/qconf.h.rej
patching file scripts/kconfig/qconf.cc
Hunk #2 succeeded at 65 with fuzz 2.
Hunk #3 succeeded at 80 (offset -1 lines).
Hunk #4 succeeded at 406 (offset 10 lines).
Hunk #5 succeeded at 420 (offset 10 lines).
Hunk #6 succeeded at 443 (offset 10 lines).
Hunk #7 succeeded at 507 (offset 10 lines).
Hunk #8 succeeded at 548 (offset 10 lines).
Hunk #9 succeeded at 567 (offset 10 lines).
Hunk #10 succeeded at 592 (offset 10 lines).
Hunk #11 succeeded at 656 (offset 10 lines).
Hunk #12 succeeded at 719 (offset 10 lines).
Hunk #13 succeeded at 781 (offset 10 lines).
Hunk #14 succeeded at 806 (offset 10 lines).
Hunk #15 succeeded at 824 (offset 10 lines).
Hunk #16 succeeded at 833 (offset 10 lines).
Hunk #17 succeeded at 841 (offset 10 lines).
Hunk #18 succeeded at 925 (offset 16 lines).
Hunk #19 succeeded at 979 (offset 16 lines).
Hunk #20 FAILED at 984.
Hunk #21 succeeded at 1161 (offset -12 lines).
Hunk #22 succeeded at 1184 (offset -12 lines).
Hunk #23 succeeded at 1201 (offset -12 lines).
Hunk #24 succeeded at 1221 (offset -12 lines).
Hunk #25 FAILED at 1320.
Hunk #26 succeeded at 1388 (offset 1 line).
Hunk #27 succeeded at 1397 (offset 1 line).
Hunk #28 succeeded at 1413 (offset 2 lines).
Hunk #29 succeeded at 1448 (offset 2 lines).
Hunk #30 succeeded at 1459 (offset 2 lines).
Hunk #31 succeeded at 1475 (offset 2 lines).
Hunk #32 succeeded at 1489 (offset 2 lines).
Hunk #33 succeeded at 1498 (offset 2 lines).
Hunk #34 succeeded at 1703 (offset 2 lines).
2 out of 34 hunks FAILED -- saving rejects to file scripts/kconfig/qconf.cc.rej


which should be easy to fix

edit:

in fact it doesn't seem it would work: import stuff of scripts/kconfig/qconf.cc is missing to make sense and put those changes in *scratches head*
_________________
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 :D
Back to top
View user's profile Send private message
Veldrin
Veteran
Veteran


Joined: 27 Jul 2004
Posts: 1945
Location: Zurich, Switzerland

PostPosted: Fri Oct 29, 2010 7:37 pm    Post subject: Reply with quote

built... currently testing the kernel. stable so far.

is there a particular reason, that the broken complete patch is still available? (at least the link in the initial post returns a broken patchset as reported by pross)

cheers
V.
Back to top
View user's profile Send private message
kernelOfTruth
Watchman
Watchman


Joined: 20 Dec 2005
Posts: 6111
Location: Vienna, Austria; Germany; hello world :)

PostPosted: Fri Oct 29, 2010 9:27 pm    Post subject: Reply with quote

Veldrin wrote:
built... currently testing the kernel. stable so far.

is there a particular reason, that the broken complete patch is still available? (at least the link in the initial post returns a broken patchset as reported by pross)

cheers
V.


yeah, I hadn't beaten myself yet to build yet another full revision ;)

in the past I had problems using the full-blown patchset with CFS instead of BFS so I wanted to separate both: so you really should go with the 1st patch to be on the safe side - if you're adventurous try it out ! if it works for you then: more power to you :wink:

(it could also be that stuff changed from 2.6.35 to 2.6.36 so that it now works without any problems)

also: it (the one with the reduced hangcheck) doesn't seem to be broken for everyone

thanks for testing/using :)

here ya go:

http://www.multiupload.com/VD6J3YFBJM <-- linux-2.6.36_TOI_plus-r2_BFS_full.patch (2.75 MB) ] (full patchset with full ck-patchset, including 2.6.36-sched-bfs-357-1.patch); (without reduced hangcheck since it seems to be broken for some folks)

md5sum: 21a52f8d7f9b2d9db0fd6131054f3288
sha256sum: 0d8a5a71d9ee1b1bb08e406bd7bc32d0e5774577df369820245a449bcec0ec3a


edit:

changed the 1st post to reflect the updates/changes

moved initial versions and r1-releases to the "broken" section
_________________
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 :D
Back to top
View user's profile Send private message
wrc1944
Advocate
Advocate


Joined: 15 Aug 2002
Posts: 3260
Location: Gainesville, Florida

PostPosted: Sat Oct 30, 2010 4:18 am    Post subject: Reply with quote

Just patched and compiled your latest linux-2.6.36_TOI_plus-r2_BFS_full.patch (2.75 MB), with no problems- smooth operation.

However, it wouldn't boot, and had the same "kernel panic not syncing, unable to mout root fs on unknown-block(8,3), sr0 detection, and swapper not tainted" problem mentioned above. Tried booting my other 2.6.36 gentoo kernels (see my Gentoo grub.conf section, here)
Code:
 default 0
timeout 15 # amount of time (seconds) before grub boots the selected option
splashimage (hd0,0)/grub/splash.xpm.gz

title=[2.6.35.4-zen2 KMS-enabled Gentoo kde-4.5.2]
root (hd0,0)
kernel /boot/2.6.35-zen2 root=/dev/sda3 video=1280x1024-24@75 rootflags=commit=5

title=[2.6.36-ck1 KMS-enabled Gentoo kde-4.5.2]
root (hd0,0)
kernel /boot/2.6.36-ck1 root=/dev/sda3 video=1280x1024-24@75 rootflags=commit=5

title=[2.6.36-ck KMS-enabled Gentoo kde-4.5.2]
root (hd0,0)
kernel /boot/2.6.36-ck root=/dev/sda3 video=1280x1024-24@75 rootflags=commit=5

title=[2.6.36-bfs-357-1 KMS-enabled Gentoo kde-4.5.2]
root (hd0,0)
kernel /boot/2.6.36-bfs-357-1 root=/dev/sda3 video=1280x1024-24@75 rootflags=commit=5

title=[2.6.36-TOI_plus-r2_BFS KMS-enabled Gentoo kde-4.5.2]
root (hd0,0)
kernel /boot/2.6.36-TOI_plus-r2_BFS root=/dev/sda3 video=1280x1024-24@75 rootflags=commit=5


The first 2.6.35.4-zen2 always boots, and the 3rd 2.6.36-ck (portage ck-sources) booted last. I compiled the TOI kernel while booted to the 4th 2.6.36-bfs-357-1 kernel.

After the TOI failed 4 straight times, I tried the 2.6.36-bfs-357-1 again several times, and it wouldn't boot. :roll: Neither would any other 2.6.36 kernel boot, but all have booted and run fine several times before. It seems a random thing if they boot or not.

Finally, I remembered that this time with TOI, I decided to enable BFQ as the default scheduler. I edited the grub kernel line at the boot screen with elevator=CFQ, and TOI now boots, and runs fine. Since my other kernels have default CFQ and at least boot sometimes, it seems TOI is having a problem with BFQ as default.

What I haven't understood is why these 2.6.36 kernels are randomly booting, or NOT booting. There are no heat or RAM memory problems on this box (that was my first thought), and all kernels use the same config file. No 2.6.25 kernel has ever exhibited this weird booting problem. It remains to be seen if I can boot TOI kernel with BFQ, or if it too will have the random booting problem with CFQ- will try again tomorrow, but I'm pleased it does boot with CFQ, at least this first time. :roll:

Anyone else having the TOI BFQ boot problem, or any idea of what is causing it? A known 2.6.36 bug of some kind? Haven't seen any mention of it anywhere.

UPDATE: OK- Maybe this will offer a clue. A reboot to the TOI kernel with CFQ failed the first time, (same panic), but an immediate 2nd boot attempt succeeded. So it seems the random boot failure problem with CFQ is also present on TOI kernel.

This can't be a hardware problem, because this has never happened on this box with Arch linux, Mint, Ubuntu, PCLOS, or Mandriva. Only on Gentoo 2.6.36 kernels, from portage or manually patched/configured. One thing I haven't done is try a plain vanilla 2.6.36- that needs to be done. If that doesn't have the problem it must be something in the ck/bfs/TOI patches.
_________________
Main box- AsRock x370 Gaming K4
Ryzen 1700, 3.0GHz, 16GB GSkill Flare DDR4 3200mhz
Samsung SATA 1000GB, Radeon HD R7 350 2GB DDR5
Gentoo ~amd64 plasma, glibc-2.30-r2, gcc-9.2.0 kernel-5.4.1-gentoo USE=experimental
Back to top
View user's profile Send private message
pross
Apprentice
Apprentice


Joined: 30 Jul 2003
Posts: 156
Location: Ireland

PostPosted: Sat Oct 30, 2010 11:16 am    Post subject: Reply with quote

con released ck2
_________________
My Overlay [pross] | bfs-sources
Back to top
View user's profile Send private message
kernelOfTruth
Watchman
Watchman


Joined: 20 Dec 2005
Posts: 6111
Location: Vienna, Austria; Germany; hello world :)

PostPosted: Sat Oct 30, 2010 12:10 pm    Post subject: Reply with quote

pross wrote:
con released ck2


thanks !

from what I understood my patchset should already include the latest ck with the one-line change

@wrc1944:

currently I'm out of ideas why it might fail :?

I tried out BFQ and wouldn't use it since I get worse interactivity/responsiveness with it under load than with CFQ :idea:

it surely however depends on your workload ...

thanks for testing/using !
_________________
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 :D
Back to top
View user's profile Send private message
wrc1944
Advocate
Advocate


Joined: 15 Aug 2002
Posts: 3260
Location: Gainesville, Florida

PostPosted: Sat Oct 30, 2010 1:34 pm    Post subject: Reply with quote

Yeah- it's a mystery to me. Just tried booting all my 2.6.36 kernels (all have ck/bfs patches, which also includes your TOI). All failed to boot. I could keep trying to boot them, and am pretty sure one or more would eventually boot, but first I'm going to compile a vanilla 2.6.36 with CFQ to try and confirm my problem is actually ck/bfs related. If CFQ fails, I'll also try elevator=deadline. If vanilla has no booting problem, I'll try the new ck2.

However from what the ck-hacking blog says, and since this box isn't wireless or a laptop, it seems unlikely that one-line change in the ck code would affect me. However, while not wireless itself (and no wireless adapter installed), this box is hard-wired through an ethernet port out of a older linksys wireless router I'm using for other boxes in the house.

UPDATE: OK, FWIW a plain vanilla 2.6.36 CFQ boots fine, and a 2nd boot to confirm also worked. So, my problem must be something in the ck1 and/or bfs patches, and thus also TOI patches. Will try ck2 with another vanilla 2.6.36 later when I have time later today.

UPDATE 2: OK, couldn't wait- compiled a 2.6.36-ck2, and it's booted 2 consecutive times, so I'm happy. Won't be totally convinced the problem is fixed until it boots a few more times. :wink: AFAIK, the only change from ck1/bfs has been that one-liner mentioned on the blog.
_________________
Main box- AsRock x370 Gaming K4
Ryzen 1700, 3.0GHz, 16GB GSkill Flare DDR4 3200mhz
Samsung SATA 1000GB, Radeon HD R7 350 2GB DDR5
Gentoo ~amd64 plasma, glibc-2.30-r2, gcc-9.2.0 kernel-5.4.1-gentoo USE=experimental
Back to top
View user's profile Send private message
kernelOfTruth
Watchman
Watchman


Joined: 20 Dec 2005
Posts: 6111
Location: Vienna, Austria; Germany; hello world :)

PostPosted: Sat Oct 30, 2010 4:04 pm    Post subject: Reply with quote

http://www.multiupload.com/6JXJTNO5L5 <-- 2.6.36-btrfs-unstable.patch (152.4 KB)

md5sum: 4f09dfa8dc55cf047250ed22b9e97df1
sha256sum: d9250508d395411eca56cabd4f52be47aea2e455325206840de0b9db9d8b549a

based on git pull from Chris Mason

another pull request might be becoming so it's not yet the final for 2.6.37
_________________
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 :D
Back to top
View user's profile Send private message
wrc1944
Advocate
Advocate


Joined: 15 Aug 2002
Posts: 3260
Location: Gainesville, Florida

PostPosted: Fri Nov 05, 2010 6:01 pm    Post subject: Reply with quote

wrc1944 wrote
Quote:
UPDATE 2: OK, couldn't wait- compiled a 2.6.36-ck2, and it's booted 2 consecutive times, so I'm happy. Won't be totally convinced the problem is fixed until it boots a few more times. AFAIK, the only change from ck1/bfs has been that one-liner mentioned on the blog.

Well, same problem mentioned above ocurrs on the next reboot, 2 days later (been updating a few other boxes). It just makes no sense that it's randomly booting/not booting on any 2.6.36 kernel, except for a plain vanilla CFQ.

It's always the same message at failure:
Code:
0816 7815591  sdb6

(Then similar lines down to sdb18, then a line like this, apparently referring to the dvd burner:

0b00 1048575 sr0  Driver: sr
Kernel panic -not syncing: VFS unable to mount root FS on unknown-block-(8,3)     (block numbers depend on which Gentoo install partition I'm booting to)
Pid : 1, comm:  swapper not tainting   (then lists the specific kernel)


sdb is my old unmounted winXP drive. My fstab is all correct. I've tried disconnecting both the sdb disk AND dvd burner, but still get the same failure. I've checked and re-checked my kernel configs, and recompiled. I've run memtest86 plus, and checked my Seagate hard drive health- all seems OK. I build all my boxes from scratch, so I know the hardware well, thus my configs are finely tuned and always work well.

It seems like it has to be only a 2.6.36 and ck, bfs, and/or TOI patches problem, but nobody else seems to be having this happen. :roll: All other kernels for years have worked perfectly on this hardware (at least booting-wise), and grub is working perfectly. I've compiled many hundreds of testing kernels over the years, and just don't see how I've made any mistakes here. I'm out of ideas, except to simply wait for 2.6.36.1, and ck3/bfs updates.

If anyone spots something I'm missing, please advise.

EDIT: This is literally insane. :? After a few hours messing around booted to 2.6.35-zen2, I again couldn't resist re-booting to 2.6.36-ck2, and of course it booted perfectly. I did nothing but reboot the machine. I know this sounds suspiciously like a hardware problem, but I'm virtually certain I've eliminated and tested everything for that. All checks out perfect, and all other Gentoo kernels on three installs on this box, as well as 5 other distros on different partitions also are booting OK with non-ck/bfs/TOI kernels, and some with their default 2.6.36 kernels (non ck/bfs).
_________________
Main box- AsRock x370 Gaming K4
Ryzen 1700, 3.0GHz, 16GB GSkill Flare DDR4 3200mhz
Samsung SATA 1000GB, Radeon HD R7 350 2GB DDR5
Gentoo ~amd64 plasma, glibc-2.30-r2, gcc-9.2.0 kernel-5.4.1-gentoo USE=experimental
Back to top
View user's profile Send private message
Display posts from previous:   
Reply to topic    Gentoo Forums Forum Index Unsupported Software All times are GMT
Page 1 of 1

 
Jump to:  
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum