Date: Thu, 14 Jan 1999 15:37:50 -0600 (CST) From: Greg Rowe <greg@uswest.net> To: freebsd-smp@FreeBSD.ORG, freebsd-scsi@FreeBSD.ORG Subject: Fatal Trap - Update Message-ID: <XFMail.990114153750.greg@uswest.net>
next in thread | raw e-mail | index | archive | help
I've done some more testing on this problem and have at least narrowed it down
somewhat. The failure happens repeatedly by just un-tarring a tar file. No
matter what the tar file contains, it seems to always drop into the debbugger
at approximately the same point in the file ??? I've done this around 30 times
with different tar files and it always crashes within a file or two of the
previous crash. It seems to only be able to handle X amount of IO from tar, but
multiple copies of bonnie run fine.
The problem doesn't seem to be hardware related (at least failing hardware) as
it happens on another system with the same mainboard and CPU's. The problem
does not happen in non-SMP mode. The problem also does not exist in SMP mode on
3.0-RELEASE (straight CD install with SMP-GENERIC kernel), so the problem was
introduced in -CURRENT. (I'm running yesterdays -CURRENT now) Adaptec SCSI
controller revision, number of drives, type of drives, and number of
controllers do not effect the problem, I've tried it all ways. Also, the page
fault will sometimes report cpuid=0 and sometimes cpuid=1. The problem
doesn't exist on a quad Xeon system running the same -CURRENT.
So, I guess the question of the day is "what does a Tyan Tiger 100 mainboard,
3.0 CURRENT, SMP, and tar have in common" ? Any ideas what I should try next ?
Thanks,
Greg
-----FW: Fatal Trap-----
Date: Mon, 11 Jan 1999 16:48:53 -0600 (CST)
From: Greg Rowe <greg@uswest.net>
To: freebsd-scsi@FreeBSD.ORG, freebsd-smp@FreeBSD.ORG
Subject: Fatal Trap
I have a Tyan mb dual 400mhz system I'm building that's getting repeatable
crashes under disk load. This didn't occur until I compiled an SMP kernel. I'm
running -current as of last Friday and the system has 512mb of memory and 4
drives on Adaptec controllers (1 - 4gig on one controller and 3 - 9gig on the
other). Untarring a file will interestingly enough crash the system at around
the same appoximate place in the tar image every time, although it's crashed
while building ports also. The system drops to the debugger with the following:
Fatal trap 12: page fault while in kernel mode
mp_lock = 01000002; cpuid=1; lapic.id = 01000000(I've also seen it for cpuid0)
fault virtual address = 0x0
fault code = supervisor write, page not present
instruction pointer = 0x8:0xf0216357
stack pointer = 0x10:0xff618ce8
frame pointer = 0x10:0xff618d0c
code segment = base 0x0, limit 0xfffff, type 0x1b
= DPL 0, pres 1, def32 1, gran1
processor eflags = interrupt enabled, resume, IOPL = 0
current process = 243 (tar)
interrupt mask = net tty bio cam <-SMP:XXX
kernel: type 12 trap, code 0
Stopped at generic_bzero+0xf: repe stosl %es:(%edi)
Any ideas as to what to look for ? Thanks.
Greg
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-scsi" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?XFMail.990114153750.greg>