Re: Today's special: nessusd panics -CURRENT
On Thu, May 10, 2001 at 11:29:57AM -0700, David Wolfskill wrote: > >Ok, I see what's broken. I don't know how you are out of witness's though. > >We don't have enough types of mutexes for that to happen. Try this patch: <...> > Since I didn't use the "-l" flag to patch, it didn't apply, so I did it > by hand (and subsequently thought to check for whitespace issues). Yes same with me:-) > But the patch appears to work; I'm up & running: Well unfortunately no such luck. It solves the first part of the problem (the "system call open retunrning with nutex(s) held" panic) but even after applying the patch and toning down MAXUSERS to 32 (I suspected it might be somewhat related) I still get the triple panic, which now looks like: witness_get: witness exhausted panic: witness_restore: lock (sleep mutex) Giant not locked After 'c', it says (attempting to sync the disks) witness_save: panic: lock (sleep mutex) Giant not locked And finally after attempting to dump: ata0: resetting devices: panic: lock (sleep mutex) Giant not locked Here come the traces. The third trace also contains the first and the second a bit further down. The third trace: Debugger panic witness_save mawait ata_command ad_reinit ata_reinit addump dumpsys boot panic=> From here is part of the second trace too. witness_save boot panic=> From here this is the first trace too. witness_restore msleep vm_pageout fork_exit fork_trampoline So basically the first trace is: Debugger panic witness_restore msleep vm_pageout fork_exit fork_trampoline And the second: Debugger panic witness_save boot panic witness_restore msleep vm_pageout fork_exit fork_trampoline This problem *only* appears when running nessusd. Any other app I tried (including X) runs fine. If you need more info let me know. -- Regards: Szilveszter ADAM Szeged University Szeged Hungary To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
RE: Today's special: nessusd panics -CURRENT
>Date: Thu, 10 May 2001 09:52:03 -0700 (PDT)
>From: John Baldwin <[EMAIL PROTECTED]>
>> witness_get: witness exhausted
>> exclusive(sleep mutex) Giant(0xc044a760) locked @ ../../i386/i386/trap.c:1169
>> panic: system call open returning with mutex(s) held
>Ok, I see what's broken. I don't know how you are out of witness's though.
>We don't have enough types of mutexes for that to happen. Try this patch:
>Index: subr_witness.c
>===
>RCS file: /usr/cvs/src/sys/kern/subr_witness.c,v
>retrieving revision 1.70
>diff -u -r1.70 subr_witness.c
>--- subr_witness.c 2001/05/04 17:15:16 1.70
>+++ subr_witness.c 2001/05/10 16:39:27
>@@ -1222,6 +1222,9 @@
>("%s: p != curproc and we aren't in the debugger", __func__));
>KASSERT(!witness_cold, ("%s: witness_cold", __func__));
>+ if (!db_active && witness_dead)
>+ return (0);
>+
>nheld = witness_list_locks(&p->p_sleeplocks);
>/*
>--
Since I didn't use the "-l" flag to patch, it didn't apply, so I did it
by hand (and subsequently thought to check for whitespace issues).
But the patch appears to work; I'm up & running:
FreeBSD m147.whistle.com 5.0-CURRENT FreeBSD 5.0-CURRENT #59: Thu May 10 04:06:04 PDT
2001 root@:/common/C/obj/usr/src/sys/LAPTOP_30W i386
(I had encountered the witness panic before applying the patch.)
Only annoyance I'm encountering with it so far is involving my use of ssh;
perhaps I'm doing something Really Weird, but in any case, that probably
deserves a different Subject:.
Cheers,
david
--
David H. Wolfskill [EMAIL PROTECTED]
As a computing professional, I believe it would be unethical for me to
advise, recommend, or support the use (save possibly for personal
amusement) of any product that is or depends on any Microsoft product.
To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-current" in the body of the message
RE: Today's special: nessusd panics -CURRENT
On 10-May-01 Szilveszter Adam wrote:
> Hello everybody,
>
> I have stumbled across some nice giraffe today.
>
> Look, this is cute (transcribed by hand for your enhanced viewing pleasure):
> (sorry, no serial console handy... should be easy to reproduce though)
>
># nessusd -D
>
> witness_get: witness exhausted
> exclusive(sleep mutex) Giant(0xc044a760) locked @ ../../i386/i386/trap.c:1169
> panic: system call open returning with mutex(s) held
Ok, I see what's broken. I don't know how you are out of witness's though.
We don't have enough types of mutexes for that to happen. Try this patch:
Index: subr_witness.c
===
RCS file: /usr/cvs/src/sys/kern/subr_witness.c,v
retrieving revision 1.70
diff -u -r1.70 subr_witness.c
--- subr_witness.c 2001/05/04 17:15:16 1.70
+++ subr_witness.c 2001/05/10 16:39:27
@@ -1222,6 +1222,9 @@
("%s: p != curproc and we aren't in the debugger", __func__));
KASSERT(!witness_cold, ("%s: witness_cold", __func__));
+ if (!db_active && witness_dead)
+ return (0);
+
nheld = witness_list_locks(&p->p_sleeplocks);
/*
--
John Baldwin <[EMAIL PROTECTED]> -- http://www.FreeBSD.org/~jhb/
PGP Key: http://www.baldwin.cx/~john/pgpkey.asc
"Power Users Use the Power to Serve!" - http://www.FreeBSD.org/
To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-current" in the body of the message
