Re: Soft update panic while running perl 5.12 tests

From: Alexander Best <arundel_at_freebsd.org>
Date: Wed, 29 Sep 2010 17:42:20 +0000
On Wed Sep 29 10, Dimitry Andric wrote:
> Hi,
> 
> I just encountered the following soft update panic while running perl
> 5.12's tests:
> 
> panic: indir_trunc: Index out of range -148 parent -2061 lbn -305164
> cpuid = 3
> KDB: enter: panic
> [ thread pid 19 tid 100047 ]
> Stopped at      kdb_enter+0x3a: movl    $0,kdb_why
> db> bt
> Tracing pid 19 tid 100047 td 0xc72e9b40
> kdb_enter(c0cd6a90,c0cd6a90,c0cfe8ab,e6be2b58,3,...) at kdb_enter+0x3a
> panic(c0cfe8ab,ffffff6c,fffff7f3,ffffffff,fffb57f4,...) at panic+0x136
> indir_trunc(fffff7f3,ffffffff,c85507c0,c834d200,c733de00,...) at 
> indir_trunc+0x4be
> handle_workitem_indirblk(4,c0cd520e,df,c834d200,c834d200,...) at 
> handle_workitem_indirblk+0x64
> handle_workitem_freeblocks(0,e6be2c74,2,5dc,1e0,...) at 
> handle_workitem_freeblocks+0x95
> process_worklist_item(c0fb2f98,0,c0cfdf8d,54a,c72e9b40,...) at 
> process_worklist_item+0x21c
> softdep_process_worklist(c732aca8,0,c0cfdf8d,4cd,64,...) at 
> softdep_process_worklist+0x8c
> softdep_flush(0,e6be2d28,c0cd1c8a,349,c72ed550,...) at softdep_flush+0x2a0
> fork_exit(c0afb670,0,e6be2d28) at fork_exit+0xb8
> fork_trampoline() at fork_trampoline+0x8
> --- trap 0, eip = 0, esp = 0xe6be2d60, ebp = 0 ---

i encountered a similar panic while updating /usr/src over subversion. the
amount of data coming in was quite a lot, since i'm tracking HEAD, but haven't
updated /usr/src for almost 2 months.

i decided to disable SUJ, because in addition to the panic i was experiencing
random reboots in connection with chromium (+10 tabs open).

i've been running plain UFS2+SU for almost 3 day now and haven't had a single
panic or spontanious reboot.

unfortunately i wasn't able to capture any information from the panic i
mentioned beforehand, but my guess is that it was something similar that
dimitry experienced.

cheers.
alex

> 
> It is consistently reproducible.  This is on a -current system, at
> r213139, on i386.  Settings for the affected filesystem:
> 
> tunefs: POSIX.1e ACLs: (-a)                                disabled
> tunefs: NFSv4 ACLs: (-N)                                   disabled
> tunefs: MAC multilabel: (-l)                               disabled
> tunefs: soft updates: (-n)                                 enabled
> tunefs: soft update journaling: (-j)                       enabled
> tunefs: gjournal: (-J)                                     disabled
> tunefs: maximum blocks per file in a cylinder group: (-e)  2048
> tunefs: average file size: (-f)                            16384
> tunefs: average number of files in a directory: (-s)       64
> tunefs: minimum percentage of free space: (-m)             8%
> tunefs: optimization preference: (-o)                      time
> tunefs: volume label: (-L)                                 dimtest1
> 
> The panic is apparently caused by perl's "op/lfs.t" test, which tests
> perlio with big files.  You can test it by building the perl 5.12 port,
> and then running:
> 
> cd /usr/ports/lang/perl5.12/work/perl-5.12.2/t
> ../miniperl op/lfs.t
> 
> Alternatively, just run "make test" in /usr/ports/lang/perl5.12.

-- 
a13x
Received on Wed Sep 29 2010 - 15:42:20 UTC

This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:40:07 UTC