Skip to contentSkip to conversation

Linux 2.6.13-rc1

DISCUSSION Ronny V. Vindenes 3 participants 4 messages Jun 29, 2005 1 message not in the archive

lore
Share & tools
lore

A message that is not in the archive (sent off-list or not yet received): [email protected]

#1 lore raw
Quoted 8 lines · Linus Torvalds

ARM, x86[-64], ppc, sparc updates, networking, sound, infiniband, input layer, ISDN, MD, DVB, V4L, network drivers, pcmcia, isofs, jfs, nfs, xfs, knfsd.. You name it.

Git trees and traditional patches/tar-balls are out there, or at least slowly mirroring out. Go wild,

Linus

On x86_64 with reiserfs and ext3 on dm (using cfq scheduler) the log is full of this:

Badness in blk_remove_plug at drivers/block/ll_rw_blk.c:1424

Call Trace:<ffffffff8022e9a2>{blk_remove_plug+49} <ffffffff8022f093>{__generic_unplug_device+21}
       <ffffffff80230da3>{get_request_wait+155} <ffffffff8013e6d0>{autoremove_wake_function+0}
       <ffffffff80231252>{__make_request+789} <ffffffff8022f3e9>{generic_make_request+502}
       <ffffffff8013e6d0>{autoremove_wake_function+0} <ffffffff8016a378>{bio_clone+130}
       <ffffffff8803cadc>{:dm_mod:clone_bio+37} <ffffffff8803cdf5>{:dm_mod:__split_bio+327}
       <ffffffff8803d11e>{:dm_mod:dm_request+171} <ffffffff8022f3e9>{generic_make_request+502}
       <ffffffff8013e6d0>{autoremove_wake_function+0} <ffffffff8014d81c>{mempool_alloc+49}
       <ffffffff8022f4ba>{submit_bio+188} <ffffffff80169dd0>{bio_alloc_bioset+232}
       <ffffffff801674fd>{submit_bh+254} <ffffffff80169158>{__block_write_full_page+419}
       <ffffffff8016c0e1>{blkdev_get_block+0} <ffffffff8018308f>{mpage_writepages+418}
       <ffffffff8016b39b>{blkdev_writepage+0} <ffffffff8013afe6>{worker_thread+0}
       <ffffffff8014ac31>{__filemap_fdatawrite_range+81} <ffffffff881c6752>{:reiserfs:flush_async_commits+103}
       <ffffffff881c66eb>{:reiserfs:flush_async_commits+0}
       <ffffffff8013b13f>{worker_thread+345} <ffffffff8012aef0>{default_wake_function+0}
       <ffffffff8013afe6>{worker_thread+0} <ffffffff8013e533>{kthread+191}
       <ffffffff8013afe6>{worker_thread+0} <ffffffff8010eee7>{child_rip+8}
       <ffffffff8013afe6>{worker_thread+0} <ffffffff8013e358>{keventd_create_kthread+0}
       <ffffffff8013e474>{kthread+0} <ffffffff8010eedf>{child_rip+0}
Signature · 1 line
Ronny V. Vindenes <[email protected]>
#2 lore raw
Quoted 9 lines · Ronny V. Vindenes
Quoted 7 lines · Linus Torvalds

ARM, x86[-64], ppc, sparc updates, networking, sound, infiniband, input layer, ISDN, MD, DVB, V4L, network drivers, pcmcia, isofs, jfs, nfs, xfs, knfsd.. You name it.

Git trees and traditional patches/tar-balls are out there, or at least slowly mirroring out. Go wild,

Quoted 6 lines
Quoted 1 line

Linus

On x86_64 with reiserfs and ext3 on dm (using cfq scheduler) the log is full of this:

Badness in blk_remove_plug at drivers/block/ll_rw_blk.c:1424

Also on x86, Reiserfs on LVM2, cfq, on sata_sil; Preemption set to
Voluntary.

Badness in blk_remove_plug at drivers/block/ll_rw_blk.c:1424

 [<c027adf1>] blk_remove_plug+0x61/0x70
 [<c027ae14>] __generic_unplug_device+0x14/0x30
 [<c027b6d5>] get_request_wait+0xd5/0x100
 [<c01322f0>] autoremove_wake_function+0x0/0x50
 [<c01322f0>] autoremove_wake_function+0x0/0x50
 [<c02837f0>] cfq_merge+0x0/0xc0
 [<c027bf03>] __make_request+0x93/0x470
 [<c027c617>] generic_make_request+0x107/0x1f0
 [<c027b61b>] get_request_wait+0x1b/0x100
 [<c01322f0>] autoremove_wake_function+0x0/0x50
 [<c01322f0>] autoremove_wake_function+0x0/0x50
 [<c01615bc>] bio_clone+0xcc/0xe0
 [<c02ca72c>] __map_bio+0x3c/0x100
 [<c02cab0b>] __clone_and_map+0x22b/0x240
 [<c02cabb7>] __split_bio+0x97/0x100
 [<c02cac7f>] dm_request+0x5f/0x90
 [<c027c617>] generic_make_request+0x107/0x1f0
 [<c01322f0>] autoremove_wake_function+0x0/0x50
 [<c01322f0>] autoremove_wake_function+0x0/0x50
 [<c027c74b>] submit_bio+0x4b/0xe0
 [<c01612c1>] bio_alloc_bioset+0x181/0x200
 [<c0160bba>] submit_bh+0xda/0x130
 [<c01b0369>] write_ordered_chunk+0x29/0x50
 [<c01b03cb>] add_to_chunk+0x3b/0x40
 [<c01b0673>] write_ordered_buffers+0xb3/0x1a0
 [<c01b0340>] write_ordered_chunk+0x0/0x50
 [<c01b0c2c>] flush_commit_list+0x37c/0x460
 [<c01b5469>] do_journal_end+0x8d9/0x930
 [<c01477c0>] pdflush+0x0/0x20
 [<c01b425c>] journal_end_sync+0x4c/0xb0
 [<c01a2e55>] reiserfs_sync_fs+0x45/0x60
 [<c01629cb>] sync_supers+0xbb/0xd0
 [<c0146e85>] wb_kupdate+0x25/0x120
 [<c0387eea>] schedule+0x2fa/0x5d0
 [<c01477c0>] pdflush+0x0/0x20
 [<c01476e6>] __pdflush+0x96/0x170
 [<c01477da>] pdflush+0x1a/0x20
 [<c0146e60>] wb_kupdate+0x0/0x120
 [<c01477c0>] pdflush+0x0/0x20
 [<c0131f05>] kthread+0x95/0xd0
 [<c0131e70>] kthread+0x0/0xd0
 [<c010134d>] kernel_thread_helper+0x5/0x18
Signature · 2 lines
Tomasz Torcz                "Funeral in the morning, IDE hacking
[email protected]    in the afternoon and evening." - Alan Cox
#3 lore raw
Quoted 9 lines · Tomasz Torcz
Quoted 5 lines · Ronny V. Vindenes

On x86_64 with reiserfs and ext3 on dm (using cfq scheduler) the log is full of this:

Badness in blk_remove_plug at drivers/block/ll_rw_blk.c:1424

Also on x86, Reiserfs on LVM2, cfq, on sata_sil; Preemption set to
Voluntary.

You should find the patch I posted just a little earlier fixes all that...

get_request is now expected to be holding on to queue_lock, with interrupts disabled, when it returns NULL; but one path forgot that, causing all kinds of nastiness under swap load - badness backtraces, strange failures, BUGs.

drivers/block/ll_rw_blk.c +2 -3
@@ -1917,10 +1917,9 @@ get_rq:
 	 * limit of requests, otherwise we could have thousands of requests
 	 * allocated with any setting of ->nr_requests
 	 */
-	if (rl->count[rw] >= (3 * q->nr_requests / 2)) {
-		spin_unlock_irq(q->queue_lock);
+	if (rl->count[rw] >= (3 * q->nr_requests / 2))
 		goto out;
-	}
+
 	rl->count[rw]++;
 	rl->starved[rw] = 0;
 	if (rl->count[rw] >= queue_congestion_on_threshold(q))
#4 lore raw

ons, 29,.06.2005 kl. 16.07 +0100, skrev Hugh Dickins:

Quoted 6 lines · Tomasz Torcz
Quoted 5 lines · Ronny V. Vindenes
Quoted 4 lines

On x86_64 with reiserfs and ext3 on dm (using cfq scheduler) the log is full of this:

Quoted 6 lines
Quoted 4 lines
Quoted 1 line

Badness in blk_remove_plug at drivers/block/ll_rw_blk.c:1424

Also on x86, Reiserfs on LVM2, cfq, on sata_sil; Preemption set to
Voluntary.

You should find the patch I posted just a little earlier fixes all that...

Confirmed.

Signature · 1 line
Ronny V. Vindenes <[email protected]>