2.5.69-mm8

2.5.69-mm8

[Andrew Morton]

ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.5/2.5.69/2.5.69-mm8/

. One anticipatory scheduler patch, but it's a big one.  I have not stress
  tested it a lot.  If it explodes please report it and then boot with
  elevator=deadline.

. The slab magazine layer code is in its hopefully-final state.

. Some VFS locking scalability work - stress testing of this would be
  useful.



Changes since 2.5.69-mm7:


-ipv6-compile-fix.patch
-sysfs_create_link-fix.patch
-subarch-circular-dependency-fix.patch
-genarch_clustered_io_apic.patch
-voyager_callout.patch
-sound-core-memalloc-build-fix.patch
-ppp-warning-fix.patch
-misc.patch
-large-dma_addr_t-PAE-only.patch
-3c59x-irq-fix.patch
-reiserfs-multiple-block-insertion.patch
-reiserfs_file_write-6.patch
-apm-module-fix.patch
-vmtruncate-PAGE_SIZE-fix.patch
-v4l-1.patch
-v4l-2.patch
-v4l-3.patch
-v4l-4.patch
-v4l-5.patch
-v4l-6.patch
-v4l-7.patch
-bttv-compile-fix.patch
-radeon-fb-64-bit-fix.patch
-cs4281-cleanup.patch
-stanford-memcy-size-fixes.patch
-BUG-to_BUG_ON.patch
-3c59x-id.patch
-suspend-asm-fix.patch
-handle-sparse-physical-apid-ids.patch
-put_page_testzero-fix.patch
-via-agp-fix.patch
-DAC960-oops-fix.patch
-it87-memset-fix.patch

 Merged

+kmalloc_percpu-interface-change-warning-fix.patch

 kmalloc_percpu() interface change

+ppc64-addnote-warning-fix.patch
+ppc64-fp-warning-fix.patch
+ppc64-do_signal32-fix.patch
+ppc64-xics-warning-fix.patch
+ppc64-prom-warning-fix.patch
+ppc64-compat-build-fix.patch
+ppc64-ioctl32-warning-fix-2.patch
+ppc64-setup-warning-fix.patch
+ppc64-traps-warning-fixes.patch
+ppc64-lpar-warning-fixes.patch
+tty_io-warning-fix.patch
+siocdevprivate_ioctl-warning-fix.patch

 Compile warning massacre

+apm-set_cpus_allowed-fix.patch

 Race fix.

+as-sync-async.patch

 Anticipatory scheduler change

-sched-2.5.68-B2.patch
-sched_idle-typo-fix.patch
-kgdb-ga-idle-fix.patch
-sched-2.5.64-D3.patch

 The HT scheduler (and its fallout) broke.  It needs updating.

+slab-reclaimable-accounting.patch
+slab-reclaimable-accounting-fs.patch

 Account for reclaimable slab pages in slab itself (for vm_enough_memory)

+aio-06-bread_wq-fix.patch

 AIO fixlet

+misc2.patch

 Misc fixes (mainly more warnings)

+io_stats-documentation.patch

 Document the disk accounting stats

+voyager-do_fork-fix.patch

 Fix voyager crash

+cpia-fp-removal.patch

 Don't use float in kernel

+vfsmount_lock.patch
+lockfree-lookup_mnt.patch

 Take some pressure off dcache_lock.




All 129 patches:


linus.patch

mm.patch
  add -mmN to EXTRAVERSION

kgdb-ga.patch
  kgdb stub for ia32 (George Anzinger's one)

kmalloc_percpu-interface-change.patch
  kmalloc_percpu: interface change.

kmalloc_percpu-interface-change-warning-fix.patch
  nail a warning

irqreturn-drivers-net.patch

slab-magazine-layer.patch
  magazine layer for slab

slab-magazine-tuning.patch
  tuning for slab-magazine-layer.patch

slabinfo-rework.patch
  new statistics for slab

aio-random-cleanups.patch

config_spinline.patch
  uninline spinlocks for profiling accuracy.

ppc64-ioctl-pci-update.patch
  From: Anton Blanchard <anton@samba.org>
  Subject: ppc64 stuff

ppc64-reloc_hide.patch

ppc64-aio-32bit-emulation.patch
  32/64bit emulation for aio

ppc64-scruffiness.patch
  Fix some PPC64 compile warnings

ppc64-xics-irq-fix.patch
  PPC64 irq return fix

ppc64-addnote-warning-fix.patch
  Squash warning in ppc64 addnote tool

ppc64-fp-warning-fix.patch
  Squash implicit declaration warning in ppc64 align.c

ppc64-do_signal32-fix.patch
  ppc64 do_signal32 warning fix

ppc64-xics-warning-fix.patch
  Squash warning in ppc64 xics.c

ppc64-prom-warning-fix.patch
  Unused variables in ppc64 prom.c

ppc64-compat-build-fix.patch
  ppc64 build fix

ppc64-ioctl32-warning-fix-2.patch
  ppc64 ioctl32 warning fix

ppc64-setup-warning-fix.patch
  nail warnings in arch/ppc64/kernel/setup.c

ppc64-traps-warning-fixes.patch
  arch/ppc64/kernel/traps.c warning fixes

ppc64-lpar-warning-fixes.patch
  ppc64: more warning fixes

tty_io-warning-fix.patch
  tty_io warning fix

siocdevprivate_ioctl-warning-fix.patch
  siocdevprivate_ioctl warning fix

aic-errno-removal.patch
  aic7xxx build fix

aic-non-i386-build-fix.patch
  aic7xxx non-i386 build fix

aic7xxx-fixes.patch

sym-do-160.patch
  make the SYM driver do 160 MB/sec

irqreturn-snd-via-fix.patch
  via sound irqreturn fix

irq_cpustat-cleanup.patch
  irq_cpustat cleanup

config-PAGE_OFFSET.patch
  Configurable kenrel/user memory split

irq-check-rate-limit.patch
  IRQs: handle bad return values from handlers

irq_desc-others.patch
  Fix up irq_desc initialisation for non-ia32

dcache_lock-vs-tasklist_lock-take-3.patch
  Fix dcache_lock/tasklist_lock ranking bug

apm-set_cpus_allowed-fix.patch
  APM does unsafe conditional set_cpus_allowed

buffer-debug.patch
  buffer.c debugging

ext3-truncate-ordered-pages.patch
  ext3: explicitly free truncated pages

VM_RESERVED-check.patch
  VM_RESERVED check

reiserfs-inode-attribute-support.patch
  reiserfs: inode attributes support.

rcu-stats.patch
  RCU statistics reporting

ext3-journalled-data-assertion-fix.patch
  Remove incorrect assertion from ext3

make-KOBJ_NAME-match-BUS_ID_SIZE.patch
  Make KOBJ_NAME_LEN match BUS_ID_SIZE

xirc2ps_cs-irqreturn-fix.patch
  xirc2ps_cs irq return fix

ide_setting_sem-fix.patch

readdir-return-value-fix.patch
  Fix readdir error return value

reslabify-pgds-and-pmds.patch
  re-slabify i386 pgd's and pmd's

linux-isp.patch

isp-update-1.patch

list_del-debug.patch
  list_del debug check

airo-schedule-fix.patch
  airo.c: don't sleep in atomic regions

synaptics-mouse-support.patch
  Add Synaptics touchpad tweaking to psmouse driver

inode-unhashing-fix-2.patch
  Don't remove inode from hash until filesystem has deleted it

resurrect-batch_requests.patch
  bring back the batch_requests function

kblockd.patch
  Create `kblockd' workqueue

cfq-infrastructure.patch

elevator-completion-api.patch
  elevator completion API

as-iosched.patch
  anticipatory I/O scheduler

as-proc-read-write.patch
  AS: pgbench improvement

as-discrete-read-fifo-batches.patch
  AS: discrete read fifo batches

as-sync-async.patch
  AS sync/async batches

unplug-use-kblockd.patch
  Use kblockd for running request queues

cfq-2.patch
  CFQ scheduler, #2

cfq-iosched-dyn.patch
  CFQ: update to rq-dyn API

unmap-page-debugging.patch
  unmap unused pages for debugging

fremap-all-mappings.patch
  Make all executable mappings be nonlinear

sched-numa-warning-fix.patch
  scheduler warning fix for NUMA

acpi-irq-ret-fix.patch
  acpi irq return value fix

sound-irq-hack.patch

show_task-free-stack-fix.patch
  show_task() fix and cleanup

put_task_struct-debug.patch

ia32-mknod64.patch
  mknod64 for ia32

ext2-64-bit-special-inodes.patch
  ext2: support for 64-bit device nodes

ext3-64-bit-special-inodes.patch
  ext3: support for 64-bit device nodes

64-bit-dev_t-kdev_t.patch
  64-bit dev_t and kdev_t

oops-dump-preceding-code.patch
  i386 oops output: dump preceding code

lockmeter.patch

slab-reclaimable-accounting.patch
  slab: account for reclaimable caches

slab-reclaimable-accounting-fs.patch
  mark shrinkable slabs as being reclaimable

security-process-attribute-api.patch
  Process Attribute API for Security Modules

proc-pid-attr-fix.patch
  Process Attribute API for Security Modules (fixlet)

proc-pid-security-labels.patch
  /proc/pid inode security labels

time-interpolation-infrastructure.patch
  improved core support for time-interpolation

time-interpolation-infrastructure-fix.patch
  make timer interpolation patch compile

thread-info-in-task_struct.patch
  allow thread_info to be allocated as part of task_struct

reinstate-task-freeing-hack-for-ia64.patch
  reinstate lame task_struct (non)-refcounting hack/fix

ext3-no-bkl.patch

journal_dirty_metadata-speedup.patch

journal_get_write_access-speedup.patch

ext3-concurrent-block-inode-allocation.patch
  Subject: [PATCH] concurrent block/inode allocation for EXT3

ext3-orlov-approx-counter-fix.patch
  Fix orlov allocator boundary case

ext3-concurrent-block-allocation-fix-1.patch

ext3-concurrent-block-allocation-hashed.patch
  Subject: Re: [PATCH] concurrent block/inode allocation for EXT3

reboot_on_bsp.patch

kexec-revert-NORET_TYPE.patch

apic_shutdown.patch

i8259-shutdown.patch

hwfixes-x86kexec.patch

kexec-warning-fixes-2.patch

CONFIG_FUTEX.patch
  FUTEX support should be optional

CONFIG_EPOLL.patch
  eventpollfs configuration option

invalidate_mmap_range.patch
  Interface to invalidate regions of mmaps

devpts-xattr-handler.patch
  devpts xattr handler for security labels 2.5.69-bk

unregister_netdev-cleanup.patch

vt8237.patch
  vt8237 IDE support

aio-01-retry.patch
  AIO: Core retry infrastructure

aio-01-retry-cleanup.patch
  aio trivia

aio-02-lockpage_wq.patch
  AIO: Async page wait

aio-03-fs_read.patch
  AIO: Filesystem aio read

aio-04-buffer_wq.patch
  AIO: Async buffer wait

aio-05-fs_write.patch
  AIO: Filesystem aio write

aio-06-bread_wq.patch
  AIO: Async block read

aio-06-bread_wq-fix.patch

aio-07-ext2getblk_wq.patch
  AIO: Async get block for ext2

aio-poll.patch
  aio_poll

aio-poll-cleanup.patch
  aio-poll: don't put extern decls in .c!

overcommit-root-margin.patch
  overcommit root margin

rpc-ifdef-fix.patch
  net/sunrpc/sunrpc_syms.c typo fix

notify_count-for-de_thread.patch
  add notify_count for de_thread

extend-check_valid_hugepage_range.patch
  rename check_valid_hugepage_range()

misc2.patch
  misc fixes

io_stats-documentation.patch
  Documentation for disk iostats

voyager-do_fork-fix.patch
  do_fork fixes for voyager x86 subarch

cpia-fp-removal.patch
  Remove floating point use in cpia.c

vfsmount_lock.patch
  From: Maneesh Soni <maneesh@in.ibm.com>
  Subject: [patch 1/2] vfsmount_lock

lockfree-lookup_mnt.patch
  lockfree lookup_mnt



--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"aart@kvack.org"> aart@kvack.org </a>

Re: 2.5.69-mm8

[Philippe Gramoullé]

Hello Andrew,

It works fine here on a Dell Inspiron 8000 using elevator=as

I have a minor warning for the i8k module though:

WARNING: /lib/modules/2.5.69-mm8/kernel/drivers/char/i8k.ko needs unknown symbol SET_MODULE_OWNER

Thanks,

Philippe

--

Philippe Gramoulle
philippe.gramoulle@mmania.com
Lycos Europe - NOC France



On Thu, 22 May 2003 02:16:52 -0700
Andrew Morton <akpm@digeo.com> wrote:

  | 
  | ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.5/2.5.69/2.5.69-mm8/
  | 
  | . One anticipatory scheduler patch, but it's a big one.  I have not stress
  |   tested it a lot.  If it explodes please report it and then boot with
  |   elevator=deadline.
  | 
  | . The slab magazine layer code is in its hopefully-final state.
  | 
  | . Some VFS locking scalability work - stress testing of this would be
  |   useful.
  | 

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"aart@kvack.org"> aart@kvack.org </a>

Re: 2.5.69-mm8

[Helge Hafting]

Andrew Morton wrote:
> ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.5/2.5.69/2.5.69-mm8/
> 
> . One anticipatory scheduler patch, but it's a big one.  I have not stress
>   tested it a lot.  If it explodes please report it and then boot with
>   elevator=deadline.
> 
> . The slab magazine layer code is in its hopefully-final state.
> 
> . Some VFS locking scalability work - stress testing of this would be
>   useful.


It seems to work fine for UP and survives a kernel compile.

Helge Hafting


--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"aart@kvack.org"> aart@kvack.org </a>

Re: 2.5.69-mm8

[William Lee Irwin III]

On Thu, May 22, 2003 at 02:16:52AM -0700, Andrew Morton wrote:
> ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.5/2.5.69/2.5.69-mm8/
> . One anticipatory scheduler patch, but it's a big one.  I have not stress
>   tested it a lot.  If it explodes please report it and then boot with
>   elevator=deadline.
> . The slab magazine layer code is in its hopefully-final state.
> . Some VFS locking scalability work - stress testing of this would be
>   useful.


Looks like this bit fell out from mainline; required for CONFIG_NUMA
to compile and identical to mainline.

-- wli

diff -prauN mm8-2.5.69-1/kernel/sched.c mm8-2.5.69-2/kernel/sched.c
--- mm8-2.5.69-1/kernel/sched.c	2003-05-22 04:54:59.000000000 -0700
+++ mm8-2.5.69-2/kernel/sched.c	2003-05-22 07:35:01.000000000 -0700
@@ -1084,6 +1084,9 @@ static void balance_node(runqueue_t *thi
 
 static void rebalance_tick(runqueue_t *this_rq, int idle)
 {
+#ifdef CONFIG_NUMA
+	int this_cpu = smp_processor_id();
+#endif
 	unsigned long j = jiffies;
 
 	/*
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"aart@kvack.org"> aart@kvack.org </a>

Re: 2.5.69-mm8

[Alexander Hoogerhuis]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Andrew Morton <akpm@digeo.com> writes:
>
> [SNIP]
>

Still some guff in the log on shutdown regarding the USB hub, but now
it doesn't do instant improvisations of a doorstop :) Other than that,
it's still alive after a few hours of banging on it, even toying with
vmware 4.0 :)

mvh,
A
- -- 
Alexander Hoogerhuis                               | alexh@ihatent.com
CCNP - CCDP - MCNE - CCSE                          | +47 908 21 485
"You have zero privacy anyway. Get over it."  --Scott McNealy
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.2 (GNU/Linux)
Comment: Processed by Mailcrypt 3.5.8 <http://mailcrypt.sourceforge.net/>

iD8DBQE+zQUrCQ1pa+gRoggRAjePAJ96nazhfgP2ZZXOH4rZFYfrI4IitACdHsWG
xPJbUVuXJQ/r2iJA//q8/UI=
=5hye
-----END PGP SIGNATURE-----
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"aart@kvack.org"> aart@kvack.org </a>

Re: 2.5.69-mm8

[Felipe Alfaro Solana]

On Thu, 2003-05-22 at 11:16, Andrew Morton wrote:
> ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.5/2.5.69/2.5.69-mm8/
> 
> . One anticipatory scheduler patch, but it's a big one.  I have not stress
>   tested it a lot.  If it explodes please report it and then boot with
>   elevator=deadline.
> 
> . The slab magazine layer code is in its hopefully-final state.
> 
> . Some VFS locking scalability work - stress testing of this would be
>   useful.

Running on it right now... Compiles and boots. I'm sure it won't explode
on my face :-)

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"aart@kvack.org"> aart@kvack.org </a>

Re: 2.5.69-mm8

[Paul Larson]

[-- Attachment #1: Type: text/plain, Size: 967 bytes --]

On Thu, 2003-05-22 at 13:53, Felipe Alfaro Solana wrote:
> On Thu, 2003-05-22 at 11:16, Andrew Morton wrote:
> > ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.5/2.5.69/2.5.69-mm8/
> > 
> > . One anticipatory scheduler patch, but it's a big one.  I have not stress
> >   tested it a lot.  If it explodes please report it and then boot with
> >   elevator=deadline.
> > 
> > . The slab magazine layer code is in its hopefully-final state.
> > 
> > . Some VFS locking scalability work - stress testing of this would be
> >   useful.
> 
> Running on it right now... Compiles and boots. I'm sure it won't explode
> on my face :-)
2.5.69-mm8 is bleeding for me. :)  See bugs #738 and #739.  I don't
*think* they are the same but apologies in advance if they are.  #738
appears to have been produced mostly by running LTP and #739 I got with
a combination of ftest07 and aio01 from LTP and previously just by
compiling LTP.

-Paul Larson

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 240 bytes --]

Re: 2.5.69-mm8

[Martin J. Bligh]

>> > ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.5/2.5.69/2.5.69-mm8/
>> > 
>> > . One anticipatory scheduler patch, but it's a big one.  I have not stress
>> >   tested it a lot.  If it explodes please report it and then boot with
>> >   elevator=deadline.
>> > 
>> > . The slab magazine layer code is in its hopefully-final state.
>> > 
>> > . Some VFS locking scalability work - stress testing of this would be
>> >   useful.
>> 
>> Running on it right now... Compiles and boots. I'm sure it won't explode
>> on my face :-)
> 2.5.69-mm8 is bleeding for me. :)  See bugs #738 and #739.  I don't
> *think* they are the same but apologies in advance if they are.  #738
> appears to have been produced mostly by running LTP and #739 I got with
> a combination of ftest07 and aio01 from LTP and previously just by
> compiling LTP.

Also seems to hang rather easily. When it gets into that state, it's difficult
to tell what works and what doesn't ... I can login over serial, but not 
start new ssh's and "ps -ef" hangs for ever. I'll try to get some more
information, and assemble a less-totally-crap bug report.

M.
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"aart@kvack.org"> aart@kvack.org </a>

Re: 2.5.69-mm8

[Paul Larson]

[-- Attachment #1: Type: text/plain, Size: 411 bytes --]

On Thu, 2003-05-22 at 14:39, Martin J. Bligh wrote:
> Also seems to hang rather easily. When it gets into that state, it's difficult
> to tell what works and what doesn't ... I can login over serial, but not 
> start new ssh's and "ps -ef" hangs for ever. I'll try to get some more
> information, and assemble a less-totally-crap bug report.
ssh and ps -ef seem to work fine on my machine

-Paul Larson

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 240 bytes --]

Re: 2.5.69-mm8

[Andrew Morton]

Paul Larson <plars@linuxtestproject.org> wrote:
>
> 2.5.69-mm8 is bleeding for me. :)  See bugs #738 and #739.

#739 seems to be the b_committed_data race.  Alex is cooking up a fix for
that.  Sorry, I didn't realise it was that easy to trigger.

I'm fairly amazed about #738.  The asertion at fs/jbd/transaction.c:2023
(J_ASSERT_JH(jh, kernel_locked())) is bogus and should be removed.
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"aart@kvack.org"> aart@kvack.org </a>

Re: 2.5.69-mm8

[Dave Hansen]

On Thu, 2003-05-22 at 12:39, Martin J. Bligh wrote:
> Also seems to hang rather easily. When it gets into that state, it's difficult
> to tell what works and what doesn't ... I can login over serial, but not 
> start new ssh's and "ps -ef" hangs for ever. I'll try to get some more
> information, and assemble a less-totally-crap bug report.

Give sysrq 't' a shot

echo t > /proc/sysrq-trigger

-- 
Dave Hansen
haveblue@us.ibm.com

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"aart@kvack.org"> aart@kvack.org </a>

Re: 2.5.69-mm8

[Paul Larson]

[-- Attachment #1: Type: text/plain, Size: 1190 bytes --]

On Thu, 2003-05-22 at 15:14, Andrew Morton wrote:
> Paul Larson <plars@linuxtestproject.org> wrote:
> >
> > 2.5.69-mm8 is bleeding for me. :)  See bugs #738 and #739.
> 
> #739 seems to be the b_committed_data race.  Alex is cooking up a fix for
> that.  Sorry, I didn't realise it was that easy to trigger.
> 
> I'm fairly amazed about #738.  The asertion at fs/jbd/transaction.c:2023
> (J_ASSERT_JH(jh, kernel_locked())) is bogus and should be removed.
Yep, a few quick tests suggest that #739 looks to be fixed by Alex's
patch, and it removes the assert from 2023, so if that's bogus then his
patch fixes that too.  I did see a hang with his patch and I wasn't able
to get any output or sysrq, so I'm going to go back and try with
nmi_watchdog to see if it'll tell me anything new.  I'm not sure if I
just didn't get far enough to see this without his patch, or if it comes
in the door with it though.  Also a few extra "sleeping function called
from illegal context" goodies on boot with it.  I'm going to get this
test kicked of right now but I'll be out until Tuesday so if it doesn't
show up again quickly I won't be seeing it until then.

Thanks,
Paul Larson

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 240 bytes --]

Re: 2.5.69-mm8

[Rudmer van Dijk]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Thursday 22 May 2003 14:07, Helge Hafting wrote:
> Andrew Morton wrote:
> > ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.5/2.5.69/2.5.
> >69-mm8/
> >
> > . One anticipatory scheduler patch, but it's a big one.  I have not
> > stress tested it a lot.  If it explodes please report it and then boot
> > with elevator=deadline.
> >
> > . The slab magazine layer code is in its hopefully-final state.
> >
> > . Some VFS locking scalability work - stress testing of this would be
> >   useful.
>
> It seems to work fine for UP and survives a kernel compile.

also for me, UP no preempt and it is running for 11 hours now without 
problems. It survived a kernel compile, compilation of the kde-network 
package and every other normal desktop-system usage. 

	Rudmer
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.2 (GNU/Linux)

iD8DBQE+zT9ShvANkaSdp/IRAh/IAJ4wuUoONk96noYpbLJOBbhvDsmNwwCeKsNa
S9VGQ6HCiwrlQJv2rEjOBMA=
=386g
-----END PGP SIGNATURE-----

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"aart@kvack.org"> aart@kvack.org </a>

Re: 2.5.69-mm8

[Martin J. Bligh]

> ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.5/2.5.69/2.5.69-mm8/
> 
> . One anticipatory scheduler patch, but it's a big one.  I have not stress
>   tested it a lot.  If it explodes please report it and then boot with
>   elevator=deadline.
> 
> . The slab magazine layer code is in its hopefully-final state.
> 
> . Some VFS locking scalability work - stress testing of this would be
>   useful.

Well, unsure about the problems I reported earlier - seems to be related
to modem disconnects during SDET runs ... the hung session seems to lock
up the system somehow. But that could have been around for ages - I'll
try to be more scientific about reproducing it at some point.

SDET results are about the same, kernel compile is down a bit on systime
(16-way NUMA-Q)

Kernbench: (make -j vmlinux, maximal tasks)
                              Elapsed      System        User         CPU
               2.5.69-mm7       46.58      117.00      578.47     1492.00
               2.5.69-mm8       46.09      115.11      570.74     1487.25

      1004     2.0% default_idle
       272     8.3% __copy_from_user_ll
       129     1.7% __d_lookup
        79     7.5% link_path_walk
...
       -50    -1.3% find_get_page
       -55    -1.5% zap_pte_range
       -66    -6.5% file_move
       -74    -1.2% page_add_rmap
       -80    -0.6% do_anonymous_page
      -110    -6.9% schedule
      -139    -7.0% atomic_dec_and_lock
      -698    -0.4% total
     -1139    -4.6% page_remove_rmap

Not sure quite what that's all about, but there it is ;-)

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"aart@kvack.org"> aart@kvack.org </a>

Re: 2.5.69-mm8

[Dave Hansen]

On Thu, 2003-05-22 at 23:18, Martin J. Bligh wrote:
>       1004     2.0% default_idle
>        272     8.3% __copy_from_user_ll
>        129     1.7% __d_lookup
>         79     7.5% link_path_walk

I have to wonder if these are cache effects, or just noise.  Can you
give oprofile a try with one of the cache performance counters?

-- 
Dave Hansen
haveblue@us.ibm.com

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"aart@kvack.org"> aart@kvack.org </a>

Re: 2.5.69-mm8

[Martin J. Bligh]

>>       1004     2.0% default_idle
>>        272     8.3% __copy_from_user_ll
>>        129     1.7% __d_lookup
>>         79     7.5% link_path_walk
> 
> I have to wonder if these are cache effects, or just noise.  Can you
> give oprofile a try with one of the cache performance counters?

No, but you can ;-)

M.

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"aart@kvack.org"> aart@kvack.org </a>

Re: 2.5.69-mm8

[William Lee Irwin III]

At some point in the past, Dave Hansen removed Martin Bligh's attribution from:
>>>       1004     2.0% default_idle
>>>        272     8.3% __copy_from_user_ll
>>>        129     1.7% __d_lookup
>>>         79     7.5% link_path_walk

At some point in the past, Martin Bligh removed Dave Hansen's attribution from:
>> I have to wonder if these are cache effects, or just noise.  Can you
>> give oprofile a try with one of the cache performance counters?

On Fri, May 23, 2003 at 07:32:31AM -0700, Martin J. Bligh wrote:
> No, but you can ;-)

Cache things are usually link order and .config dependent, in 2.4.x at
least they were highly dependent on cache color conflicts between task
structures and hot codepaths and similar bizarre phenomena. i.e. exact
binutils, compiler, kernel source, and .config matches are required to
reproduce. And sometimes even that isn't enough and it's not reproducible
across runs. You also want instruction-level multiplicative differential
profiling to find cacheline bounces, not function-level additive
differential profiling, with some method of correlating assembly to source.

-- wli
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"aart@kvack.org"> aart@kvack.org </a>

Re: 2.5.69-mm8

[Martin J. Bligh]

>       1004     2.0% default_idle
>        272     8.3% __copy_from_user_ll
>        129     1.7% __d_lookup
>         79     7.5% link_path_walk
> ...
>        -50    -1.3% find_get_page
>        -55    -1.5% zap_pte_range
>        -66    -6.5% file_move
>        -74    -1.2% page_add_rmap
>        -80    -0.6% do_anonymous_page
>       -110    -6.9% schedule
>       -139    -7.0% atomic_dec_and_lock
>       -698    -0.4% total
>      -1139    -4.6% page_remove_rmap
> 
> Not sure quite what that's all about, but there it is ;-)

WRT consistency, a second set of runs indicates it's very consistent.
And the user time is down significantly too. These are all averages
of 5 runs to start with (well, averages of the median 3 runs of 5).
Same for the profiles.

Kernbench: (make -j vmlinux, maximal tasks)
                              Elapsed      System        User         CPU
               2.5.69-mm7       46.58      117.00      578.47     1492.00
             2.5.69-mm7-2       47.39      117.24      578.58     1468.75
               2.5.69-mm8       46.09      115.11      570.74     1487.25
             2.5.69-mm8-2       45.91      115.00      571.22     1494.25

The copy to/from user stuff may be noise, but page_remove_rmap and
atomic_dec_and_lock are certainly happier. Second set of runs gives:

      3708     8.0% default_idle
      1285     0.8% total
       161     4.9% __copy_from_user_ll
        95     2.5% find_get_page
        77     5.8% kmem_cache_free
        58     4.7% release_pages
        57     5.4% link_path_walk
...
       -52   -14.0% .text.lock.filemap
       -62   -14.3% .text.lock.file_table
       -63    -3.8% do_page_fault
       -67    -3.6% path_lookup
       -70    -0.5% do_anonymous_page
       -84   -10.0% pte_alloc_one
       -91    -5.7% schedule
      -111   -11.2% clear_page_tables
      -118    -6.0% atomic_dec_and_lock
      -200    -3.7% __copy_to_user_ll
      -338    -4.1% __d_lookup
      -931    -3.8% page_remove_rmap

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"aart@kvack.org"> aart@kvack.org </a>