On 12/01/2006 12:56 a.m., Jens Axboe wrote:
On Thu, Jan 12 2006, Reuben Farrelly wrote:
On 12/01/2006 12:13 a.m., Jens Axboe wrote:
On Wed, Jan 11 2006, Andrew Morton wrote:
Neil thinks that an IO got lost. In the git2->git3 diff we have:
b/drivers/scsi/Kconfig | 10
b/drivers/scsi/ahci.c | 1
b/drivers/scsi/ata_piix.c | 5
b/drivers/scsi/libata-core.c | 145 +
b/drivers/scsi/libata-scsi.c | 48
b/drivers/scsi/libata.h | 4
b/drivers/scsi/sata_mv.c | 1
b/drivers/scsi/sata_promise.c | 1
b/drivers/scsi/sata_sil.c | 1
b/drivers/scsi/sata_sil24.c | 1
b/drivers/scsi/sata_sx4.c | 1
b/drivers/scsi/scsi_lib.c | 50
b/drivers/scsi/scsi_sysfs.c | 31
b/drivers/scsi/sd.c | 85 -
b/fs/bio.c | 26
Jens, Jeff: were any of those changes added in the final day or two, not
included in the trees which I pull?
Reuben, do you have any barrier= options in your fstab for any reiser
file system?
None whatsoever:
/dev/md0 / reiserfs defaults 0 0
none /dev/pts devpts gid=5,mode=620 0 0
none /dev/shm tmpfs defaults 0 0
none /proc proc defaults 0 0
sysfs /sys sysfs defaults 0 0
/dev/sda1 /boot ext3 defaults 1 2
#/dev/sdb1 /boot-2 ext3 defaults 1 2
/dev/md1 /home reiserfs defaults 0 0
/dev/md2 /var reiserfs defaults 0 0
/dev/md3 /var/www/cgi-bin reiserfs defaults 0 0
/dev/md4 /tmp reiserfs defaults 0 0
/dev/md5 /backup reiserfs defaults 0 0
/dev/sda8 /var/spool/squid-1 reiserfs noatime,notail 0 0
/dev/sdb8 /var/spool/squid-2 reiserfs noatime,notail 0 0
/dev/sda9 swap swap defaults 0 0
/dev/sdb9 swap swap defaults 0 0
/dev/sdc1 /store reiserfs defaults 0 0
/dev/shm /var/spool/amavisd/tmp tmpfs
defaults,size=25m,mode=700,uid=508,gid=509, 0 0
/dev/fd0 /media/floppy auto
pamconsole,exec,noauto,managed 0 0
Then the barrier changes from git2 -> git3 should not have anything to
do with it. Strange... I guess you should try the git bisect method to
narrow it down.
Ok push came to shove, so I spent the evening (early hours of the morning
really) learning how to git my way around a little and use git bisect. Not bad,
people who come up with clever stuff like that would probably be clever enough
to be able to do kernel development or something ;-)
Anyway, humour aside, I've bisected down to six revisions:
[root@tornado linux-2.6]# git bisect good
Bisecting: 6 revisions left to test after this
[93c9338713d4e11102cd09b4670ad42a336b06a3] [BLOCK] update libata to use new
blk_ordered for barriers
[root@tornado linux-2.6]#
however I'm not sure I can go a lot further now as the tree is failing to
compile at that point:
include/asm/mpspec_def.h:78: warning: 'packed' attribute ignored for field of
type 'unsigned char[5u]'
block/ll_rw_blk.c:2421: error: conflicting types for 'blk_execute_rq_nowait'
include/linux/blkdev.h:617: error: previous declaration of
'blk_execute_rq_nowait' was here
make[1]: *** [block/ll_rw_blk.o] Error 1
make: *** [block] Error 2
[root@tornado linux-2.6]#
I'm guessing there are a block of changes that all go together around this point.
Here's my BISECT_LOG:
git-bisect start
# bad: [0aec63e67c69545ca757a73a66f5dcf05fa484bf] Fix posix-cpu-timers
sched_time accumulation
git-bisect bad 0aec63e67c69545ca757a73a66f5dcf05fa484bf
# good: [2e3e13f8e9d9b2111404cdccaa4e1b988b70acce] i2c: i2c-i801 explicitly
enables/disables PEC
git-bisect good 2e3e13f8e9d9b2111404cdccaa4e1b988b70acce
# good: [9bbc8346fb21fad3f678220b067450e436e45dbf] s390: fix invalid return code
in sclp_cpi
git-bisect good 9bbc8346fb21fad3f678220b067450e436e45dbf
# bad: [221fc10ec89834329e5613e3cab4569ba22da410] fs/ufs: debug mode compilation
failure
git-bisect bad 221fc10ec89834329e5613e3cab4569ba22da410
# good: [ddaf22abaa831763e75775e6d4c7693504237997] md: attempt to auto-correct
read errors in raid1
git-bisect good ddaf22abaa831763e75775e6d4c7693504237997
# good: [d9d166c2a9d5d01af34396793950aa695883eed4] md: allow array level to be
set textually via sysfs
git-bisect good d9d166c2a9d5d01af34396793950aa695883eed4
# bad: [e650c305ec3178818b317dad37a6d9c7fa8ba28d] [SCSI] scsi_end_async() needs
to take an uptodate parameter
git-bisect bad e650c305ec3178818b317dad37a6d9c7fa8ba28d
# good: [64100099ed22f71cce656c5c2caecf5c9cf255dc] [BLOCK] mark some block/
variables cons
git-bisect good 64100099ed22f71cce656c5c2caecf5c9cf255dc
I'll leave the setup as it is right now so if there's an easy way to narrow it
down even further I can continue tomorrow.
Incidentally, I also tested - the problem is still in -mm3 also.
reuben
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
[Index of Archives]
[Kernel Newbies]
[Netfilter]
[Bugtraq]
[Photo]
[Stuff]
[Gimp]
[Yosemite News]
[MIPS Linux]
[ARM Linux]
[Linux Security]
[Linux RAID]
[Video 4 Linux]
[Linux for the blind]
[Linux Resources]
