http://www.zip.com.au/~akpm/linux/patches/2.5/2.5.59/2.5.59-mm9/
. Adam's cleanup and cutdown of devfs has been put back in again. We
really need devfs users to test this and to report, please. (And not just
to me! I'll only bounce it to Adam J. Richter <[email protected]>
anyway)
Adam has a userspace helper app `devfs-helper' at
ftp://ftp.yggdrasil.com/pub/dist/device_control/devfs/
. The I/O scheduler changes here are stable now. All-round peformance is
OK.
I've been using it on the desktop for a week. Certain operations are
noticeably different; for example a big `cvs co' would previously have had
several-second stalls every 20 seconds or so. Now, things slow down but it
keeps going. I'm sure the overall runtime will be slightly increased, but
overall the decreased read latency is good.
. dcache_rcu has been put back.
Changes since 2.5.59-mm8:
-reiserfs-readpages.patch
-fadvise.patch
-auto-unplug.patch
-less-unplugging.patch
-kirq.patch
-kirq-up-fix.patch
-agp-warning-fix.patch
-prune-icache-stats.patch
-vma-file-merge.patch
-mmap-whitespace.patch
-read_cache_pages-cleanup.patch
-remove-GFP_HIGHIO.patch
-wli-11_pgd_ctor.patch
-wli-11_pgd_ctor-update.patch
-smaller-slab-batches.patch
-printk-locking.patch
-hangcheck-timer.patch
-jbd-documentation.patch
-sendfile-security-hooks.patch
-mmzone-parens.patch
-no_space_in_slabnames.patch
-remove-will_become_orphaned_pgrp.patch
-MAX_IO_APICS-ifdef.patch
-dac960-error-retry.patch
-topology-remove-underbars.patch
-put_user-warning-fix.patch
-hash-warnings.patch
-mark_inode_dirty-race.patch
-lost-tick.patch
-seq_file-page-defn.patch
-scsi-iothread.patch
-writeback-sync-cleanup.patch
-dont-wait-on-inode.patch
-unlink-latency-fix.patch
-pin_page-fix.patch
-pin_page-pmd.patch
-frlock-xtime.patch
-frlock-xtime-i386.patch
-frlock-xtime-ia64.patch
-frlock-xtime-other.patch
-seqlock.patch
-do_gettimeofday-speedup.patch
-default_idle-speedup.patch
-pte_chain_alloc-fixes.patch
-hugetlbfs-set_page_dirty.patch
-compound-pages.patch
-compound-pages-hugetlb.patch
-hugetlbfs-get_unmapped_area.patch
-hugetlbfs-truncate-fix.patch
-hugetlbfs-i_size-fix.patch
-hugetlbfs-cleanup.patch
-hugetlbfs-nopage-cleanup.patch
-hugetlbfs-fault-fix.patch
-hugetlbpage-cleanup.patch
-hugetlb_vmtruncate-fixes.patch
-hugetlb-mremap-fix.patch
-mremap-cleanup.patch
-up-spinlock-debugging.patch
Merged
+seqlock-fixes.patch
seqlock tweak
+profiler-per-cpu.patch
Use per-cpu data for the kernel profiler
+disassociate_tty-fix.patch
TTY layer race fix
+crc32-speedup.patch
Faster crc32 implementation
+report-lost-ticks.patch
Add some controls to the ia32 lost tick detector.
-scheduler-update.patch
+sched-2.5.59-F3.patch
Ingo's latest HT-aware CPU scheduler changes
+generic_file_aio_write_nolock-overflow-fix.patch
Fix huge writes on 64-bit machines.
+lockd-lockup-fix.patch
Maybe fix a knfsd deadlock
+ll_rw_block-fix.patch
Fix obscure incorrectness when ll_rw_block() is used for IS_SYNC
operations.
+rcu-stats.patch
Resusrrect RCU statistics in /proc
+dcache_rcu-fast_walk-revert.patch
+dcache_rcu-main.patch
Use RCU technology for relieving dcache_lock stress
+cyclone-fixes.patch
+enable-timer_cyclone.patch
Add support for the x440's timer
+vxfs-memleak-fix.patch
Fix a memory leak
+smalldevfs.patch
Simplified devfs
+remove-journal_try_start.patch
Remove the ext3 code which attempts to avoid blocking kswapd on
journal_start().
+dac960-range-fix.patch
DAC960 fix.
+DAC960-maintainer.patch
Add David
+sys_exit_group-warning-fix.patch
Fix a compile warning
+nforce2-support.patch
Support nforce2 IDE controllers
All 43 patches:
linus.patch
seqlock-fixes.patch
kgdb.patch
devfs-fix.patch
ptrace-flush.patch
Subject: [PATCH] ptrace on 2.5.44
buffer-debug.patch
buffer.c debugging
warn-null-wakeup.patch
htlb-2.patch
hugetlb: fix MAP_FIXED handling
ext3-truncate-ordered-pages.patch
ext3: explicitly free truncated pages
profiler-per-cpu.patch
Subject: Re: [patch] Make prof_counter use per-cpu areas patch 1/4 -- x86 arch
oprofile-p4.patch
oprofile_cpu-as-string.patch
oprofile cpu-as-string
disassociate_tty-fix.patch
Subject: [PATCH][RESEND 3] disassociate_ctty SMP fix
epoll-update.patch
epoll timeout and syscall return types ...
mandlock-oops-fix.patch
ftruncate/truncate oopses with mandatory locking
reiserfs_file_write.patch
Subject: reiserfs file_write patch
user-process-count-leak.patch
fix current->user->processes leak
misc.patch
misc fixes
numaq-ioapic-fix2.patch
NUMAQ io_apic programming fix
deadline-np-42.patch
(undescribed patch)
deadline-np-43.patch
(undescribed patch)
batch-tuning.patch
I/O scheduler tuning
starvation-by-read-fix.patch
fix starvation-by-readers in the IO scheduler
crc32-speedup.patch
crc32 improvements for 2.5
scheduler-tunables.patch
scheduler tunables
sched-2.5.59-F3.patch
HT scheduler, sched-2.5.59-F3
rml-scheduler-update2.patch
rml scheduler tree
report-lost-ticks.patch
make lost-tick detection more informative
generic_file_aio_write_nolock-overflow-fix.patch
generic_file_write() overflow fix
lockd-lockup-fix.patch
Subject: Re: Fw: Re: 2.4.20 NFS server lock-up (SMP)
ll_rw_block-fix.patch
Fix ll_rw_block() when used for data integrity
rcu-stats.patch
RCU statistics reporting
dcache_rcu-fast_walk-revert.patch
dcache_rcu: revert fast_walk code
dcache_rcu-main.patch
dcache_rcu
cyclone-fixes.patch
Cyclonetimer fixes
enable-timer_cyclone.patch
Enable timer_cyclone code
vxfs-memleak-fix.patch
fix leaks in vxfs_read_fshead()
smalldevfs.patch
smalldevfs
remove-journal_try_start.patch
ext3: Remove journal_try_start()
dac960-range-fix.patch
Subject: [PATCH] DAC960 Stanford Checker fix
DAC960-maintainer.patch
Add David Olien MAINTAINERs for DAC960
sys_exit_group-warning-fix.patch
nforce2-support.patch
nforce2 IDE support for the amd74xx driver
Andrew Morton <[email protected]> wrote:
>
> http://www.zip.com.au/~akpm/linux/patches/2.5/2.5.59/2.5.59-mm9/
I've taken this down.
Ingo, there's something bad in the signal changes in Linus's current tree.
mozilla won't display, and is unkillable:
mnm:/home/akpm> ps aux|grep moz
akpm 1462 0.0 2.5 44568 23244 ? S 02:26 0:00 /usr/lib/mozilla-1.3a/mozilla-bin
akpm 1463 0.0 2.5 44568 23244 ? S 02:26 0:00 /usr/lib/mozilla-1.3a/mozilla-bin
akpm 1469 0.0 2.5 44568 23244 ? S 02:26 0:00 /usr/lib/mozilla-1.3a/mozilla-bin
akpm 1470 0.0 2.5 44568 23244 ? S 02:26 0:00 /usr/lib/mozilla-1.3a/mozilla-bin
akpm 1471 0.0 2.5 44568 23244 ? S 02:26 0:00 /usr/lib/mozilla-1.3a/mozilla-bin
akpm 9024 0.0 0.0 3260 556 pts/19 S 02:32 0:00 grep moz
mnm:/home/akpm> kill -9 1462 1463 1469 1470 1471
mnm:/home/akpm> ps aux|grep moz
akpm 1462 0.0 2.5 44568 23244 ? S 02:26 0:00 /usr/lib/mozilla-1.3a/mozilla-bin
akpm 1463 0.0 2.5 44568 23244 ? S 02:26 0:00 /usr/lib/mozilla-1.3a/mozilla-bin
akpm 1469 0.0 2.5 44568 23244 ? S 02:26 0:00 /usr/lib/mozilla-1.3a/mozilla-bin
akpm 1470 0.0 2.5 44568 23244 ? S 02:26 0:00 /usr/lib/mozilla-1.3a/mozilla-bin
akpm 1471 0.0 2.5 44568 23244 ? S 02:26 0:00 /usr/lib/mozilla-1.3a/mozilla-bin
akpm 9028 0.0 0.0 3260 556 pts/19 S 02:33 0:00 grep moz
mnm:/home/akpm> ps axo pid,comm,wchan|grep moz
1462 mozilla-bin schedule_timeout
1463 mozilla-bin rt_sigsuspend
1469 mozilla-bin rt_sigsuspend
1470 mozilla-bin rt_sigsuspend
1471 mozilla-bin rt_sigsuspend
That's just from bringing up X and starting mozilla 1.13a. Happens every time.
On Fri, Feb 07, 2003 at 03:03:50AM -0800, Andrew Morton wrote:
> Andrew Morton <[email protected]> wrote:
> >
> > http://www.zip.com.au/~akpm/linux/patches/2.5/2.5.59/2.5.59-mm9/
>
> I've taken this down.
>
> Ingo, there's something bad in the signal changes in Linus's current tree.
>
> mozilla won't display, and is unkillable:
Yeah, I'm seeing hangs in rt_sigsuspend under GDB also. Thanks for
saying that they show up without ptrace; I hadn't been able to
reproduce them without it.
Something is causing realtime signals to drop.
--
Daniel Jacobowitz
MontaVista Software Debian GNU/Linux Developer
Andrew Morton <[email protected]> writes:
> . Adam's cleanup and cutdown of devfs has been put back in again. We
> really need devfs users to test this and to report, please. (And not just
> to me! I'll only bounce it to Adam J. Richter <[email protected]>
> anyway)
Ok, I patched 2.5.59 with Adam's patch and it boots fine. I miss the
file /dev/.devfsd - Debian uses that file to detect devfs and act
accordingly. Still in the first minutes, I'll get back when Linus has
released 2.5.60 (I'll probably switch back to a stock kernel then).
Jochen
--
#include <~/.signature>: permission denied
Daniel Jacobowitz <[email protected]> wrote:
>
> On Fri, Feb 07, 2003 at 03:03:50AM -0800, Andrew Morton wrote:
> > Andrew Morton <[email protected]> wrote:
> > >
> > > http://www.zip.com.au/~akpm/linux/patches/2.5/2.5.59/2.5.59-mm9/
> >
> > I've taken this down.
> >
> > Ingo, there's something bad in the signal changes in Linus's current tree.
> >
> > mozilla won't display, and is unkillable:
>
> Yeah, I'm seeing hangs in rt_sigsuspend under GDB also. Thanks for
> saying that they show up without ptrace; I hadn't been able to
> reproduce them without it.
>
> Something is causing realtime signals to drop.
OK. Looks like Linus is hot on the trail.
BTW, some nice people have been sending in smalldevfs testing results
(successful). I've put that patch back up at
http://www.zip.com.au/~akpm/linux/patches/2.5/2.5.59/smalldevfs.patch
for other testers. It applies to 2.5.59 base.
And it is not clear why I copied Ingo on the signal thing, when it is not he
who is working that code. Sorry about that.
Jochen Hein wrote:
> Andrew Morton <[email protected]> writes:
>
>
>>. Adam's cleanup and cutdown of devfs has been put back in again. We
>> really need devfs users to test this and to report, please. (And not just
>> to me! I'll only bounce it to Adam J. Richter <[email protected]>
>> anyway)
>
>
> Ok, I patched 2.5.59 with Adam's patch and it boots fine. I miss the
> file /dev/.devfsd - Debian uses that file to detect devfs and act
> accordingly. Still in the first minutes, I'll get back when Linus has
Slackware does the same.
> released 2.5.60 (I'll probably switch back to a stock kernel then).
--
Georgi Chorbadzhiyski
http://georgi.unixsol.org/
On 08 Feb 2003, Georgi Chorbadzhiyski wrote:
>Jochen Hein wrote:
>> Andrew Morton <[email protected]> writes:
>>>. Adam's cleanup and cutdown of devfs has been put back in again. We
>>> really need devfs users to test this and to report, please. (And not just
>>> to me! I'll only bounce it to Adam J. Richter <[email protected]>
>>> anyway)
>>
>> Ok, I patched 2.5.59 with Adam's patch and it boots fine. I miss the
>> file /dev/.devfsd - Debian uses that file to detect devfs and act
>> accordingly. Still in the first minutes, I'll get back when Linus has
>Slackware does the same.
Distributions that want to do something different for devfs
can parse /proc/mounts, or, less reliably, do statfs("/dev", &statfs_buf)
and look at statfs_buf.f_type. So, you still have this ability without
the need for additional kernel code.
It's worth noting that some devfs users may want the non-devfs
behavior (which I assume means creating /dev files during some
installation process) because they may have a script to save /dev
before shutdown and restore their additional /dev nodes at boot, so
you probably want to centralize this decision in some little script
anyhow. The devfsd (for the stock devfs) has a couple of commands
designed for this, although this can just as easily be done in scripts
for boot and shutdowns.
Also, I suppose that checking for /dev/.devfsd is an easy way
to detect _which_ devfs you are using, although I don't know if such a
check is useful, since you could start devfsd unconditionally and it
should just exit if the old devfs is not present.
Adam J. Richter __ ______________ 575 Oroville Road
[email protected] \ / Milpitas, California 95035
+1 408 309-6081 | g g d r a s i l United States of America
"Free Software For The Rest Of Us."