Date: Sun, 19 May 2002 15:21:07 -0700 (PDT) From: Paul Herman <pherman@frenchfries.net> To: freebsd-stable@FreeBSD.ORG Subject: Re: panic: softdep_disk_write_complete: lock is held Message-ID: <20020519150606.G443-100000@mammoth.eat.frenchfries.net> In-Reply-To: <20020517123019.V1458-100000@mammoth.eat.frenchfries.net>
next in thread | previous in thread | raw e-mail | index | archive | help
On Fri, 17 May 2002, I wrote:
> panic: softdep_disk_write_complete: lock is held
>
> syncing disks... panic: softdep_lock: locking against myself
> Uptime: 1m59s
> [...]
OK, I'm seeing something strange here. As you can see, the panic
happens here at frame 10:
#7 0xc022066f in sync (p=0xc048f380, uap=0x0)
at /usr/src/sys/kern/vfs_syscalls.c:576
#8 0xc01f0c82 in boot (howto=256) at /usr/src/sys/kern/kern_shutdown.c:235
#9 0xc01f130c in poweroff_wait (junk=0xc03dad00, howto=-1015284636)
at /usr/src/sys/kern/kern_shutdown.c:595
#10 0xc02f7a64 in softdep_disk_write_complete (bp=0xc37995ac)
at /usr/src/sys/ufs/ffs/ffs_softdep.c:3228
#11 0xc0218571 in biodone (bp=0xc37995ac) at /usr/src/sys/kern/vfs_bio.c:2703
#12 0xc021a670 in cluster_callback (bp=0xc376fef8)
at /usr/src/sys/kern/vfs_cluster.c:549
#13 0xc0218544 in biodone (bp=0xc376fef8) at /usr/src/sys/kern/vfs_bio.c:2698
#14 0xc015d5af in ad_interrupt (request=0xc105d8c0)
where the offending code in softdep_disk_write_complete() is:
3226 #ifdef DEBUG
3227 if (lk.lkt_held != -1)
3228 panic("softdep_disk_write_complete: lock is held");
3229 lk.lkt_held = -2;
3230 #endif
but in this very frame...
(kgdb) print lk
$2 = {lkt_spl = 6867008, lkt_held = -1}
So, according to lk, no lock is held, but the if condition is
satisfied anyway. Is there a race condition somewhere? biodone()
is called twice (don't know if that's normal) which might cause the
lock to be freed twice? Or is something else happening? I have no
clue here.
I've narrowed this panic down, and I'm able to 100% reliably
reproduce it in single user mode with only a shell and ftp running.
It happens while transfering a large file over a wi0 wireless link.
During the transfer, wi0 produces about 460 interrupts per second.
Again, I'm running 4.6-PRERELEASE.
-Paul.
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-stable" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20020519150606.G443-100000>