Date: Wed, 10 Mar 1999 12:18:51 -0600 (CST) From: Greg Rowe <greg@uswest.net> To: freebsd-hackers@FreeBSD.ORG Subject: SMP Woes Message-ID: <XFMail.990310121851.greg@uswest.net>
next in thread | raw e-mail | index | archive | help
I'm still struggling to get this last batch of SMP systems working correctly.
Originally I was having Fatal Trap errors under high IO load on my Tyan dual
CPU systems but somewhere close to 3.1 Release level, the problems cleared up
and they seem to be running fine. Unfortunately, around the same build level my
Quad Xeon box started up with the Fatal Traps ??? It's starting to drive me
crazy !!!
As far as I can tell, 3.0 Release worked fine on all the SMP systems here.
Somewhere between 3.0 Release and 3.1 Release something changed that's causing
these failures. I went up to 4.0 current yesterday to see if Matt's recent
changes had any effect, but they didn't.
I only get the failures when running in SMP mode on the system. Single CPU
works fine. The failure is easily reproducable by running a couple cpio's or
even a make world (I have to go to single CPU to upgrade). The Xeon is a SC450NX
with a SCSI backplane. I've tried both the on-board NCR and Adaptec
controllers. The DDB output is always the same, except the CPU ID changes.
Again, it works at 3.0 Release, so I don't think it's a hardware failure.
Attached is the DDB output and trace from a terminal server window. I'm
more than willing to try any suggestions and can provide access to crash dumps
or the system. Thanks.
Fatal trap 12: page fault while in kernel mode
mp_lock = 03000002; cpuid = 3; lapic.id = 02000000
fault virtual address = 0x0
fault code = supervisor write, page not present
instruction pointer = 0x8:0xf020ec9f
stack pointer = 0x10:0xfe5a3c34
frame pointer = 0x10:0xfe5a3c58
code segment = base 0x0, limit 0xfffff, type 0x1b
= DPL 0, pres 1, def32 1, gran 1
processor eflags = interrupt enabled, resume, IOPL = 0
current process = 245 (cpio)
interrupt mask = net tty bio cam <- SMP: XXX
kernel: type 12 trap, code=0
Stopped at generic_bzero+0xf: repe stosl %es:(%edi)
db> trace
generic_bzero(f3283f80,0,f4801900,fe5a3c90,fe5a3c98) at generic_bzero+0xf
zalloci(f3283f80,f488bb00,f4801900,6f802,fe541180) at zalloci+0x29
getnewvnode(1,f33f0200,f3266200,fe5a3cfc,100) at getnewvnode+0x2f8
ffs_vget(f33f0200,6f802,fe5a3d7c,ff779d00,fe5a3edc) at ffs_vget+0xa5
ufs_lookup(fe5a3dd4,fe5a3de8,f016f6d4,fe5a3dd4,fe553c1d) at ufs_lookup+0x936
ufs_vnoperate(fe5a3dd4,fe553c1d,ff779d00,fe5a3edc,0) at ufs_vnoperate+0x15
vfs_cache_lookup(fe5a3e30,fe5a3e40,f0171ae9,fe5a3e30,fe53b640) at
vfs_cache_lookup+0x248
ufs_vnoperate(fe5a3e30,fe53b640,fe5a3edc,fe5a3eb8,0) at ufs_vnoperate+0x15
lookup(fe5a3eb8,fe541180,f0250848,fe541180,1) at lookup+0x2c1
namei(fe5a3eb8,fe541180,f0250848,0,8057000) at namei+0x133
lstat(fe541180,fe5a3f94,8057000,ffffffff,3) at lstat+0x44
syscall(2f,efbf002f,3,ffffffff,efbfdc70) at syscall+0x187
Xint0x80_syscall() at Xint0x80_syscall+0x4c
db>
Greg Rowe <greg@uswest.net> US WEST - Internet Service Operations
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-hackers" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?XFMail.990310121851.greg>