linuxolator - ltp testing: panic with fsync02 test

Scot Hetzel swhetzel at gmail.com
Sat Apr 21 02:16:41 UTC 2007


I was trying to run the ltp tests on a recent -CURRENT kernel that was
checked out using:

cvs -q update -P -d -D "04/20/2007 00:01 UTC"

But when the tests run the fsync02 test, it causes the following:

Unread portion of the kernel message buffer:
panic: NOT MPSAFE and Giant not held
cpuid = 0
KDB: enter: panic
Uptime: 3h47m36s
Physical memory: 1008 MB
Dumping 84 MB: 69 53 37 21 5

(kgdb) bt
#0  doadump () at pcpu.h:171
During symbol reading, Incomplete CFI data; unspecified registers at
0xffffffff804508fc.
#1  0xffffffff804511fb in boot (howto=0x104) at
../../../kern/kern_shutdown.c:409
#2  0xffffffff80450c87 in panic (fmt=0xffffffff80728e38 "NOT MPSAFE
and Giant not held")
    at ../../../kern/kern_shutdown.c:563
#3  0xffffffff804c1373 in namei (ndp=0xffffffffa434f820) at
../../../kern/vfs_lookup.c:182
#4  0xffffffffa3f303e1 in linprocfs_domtab (td=0xffffff0024ac9a80,
p=0x0, pn=0xffffff002b60c500,
    sb=0xffffff0028470800, uio=0x0) at
/usr/src/7x/sys-test/modules/linprocfs/../../compat/linprocfs/linprocfs.c:317
#5  0xffffffff80407ea8 in pfs_read (va=0x0) at pseudofs_internal.h:137
#6  0xffffffff806a2794 in VOP_READ_APV (vop=0x0, a=0xffffffffa434f9b0)
at vnode_if.c:637
#7  0xffffffff804d5fd4 in vn_read (fp=0xffffff0027cbb708,
uio=0xffffffffa434faf0, active_cred=0x0, flags=0x0,
    td=0xffffff0024ac9a80) at vnode_if.h:343
#8  0xffffffff80487b44 in dofileread (td=0xffffff0024ac9a80, fd=0x4,
fp=0xffffff0027cbb708, auio=0xffffffffa434faf0,
    offset=0x0, flags=0x0) at file.h:242
#9  0xffffffff80487db0 in kern_readv (td=0xffffff0024ac9a80, fd=0x4,
auio=0xffffffffa434faf0)
    at ../../../kern/sys_generic.c:180
#10 0xffffffff80487eaa in read (td=0x0, uap=0x0) at
../../../kern/sys_generic.c:110
#11 0xffffffff8069c109 in ia32_syscall (frame=0xffffffffa434fc70) at
../../../amd64/ia32/ia32_syscall.c:187
#12 0xffffffff8064b100 in Xint0x80_syscall () at ia32_exception.S:65
#13 0x000000000b0209de in ?? ()
Previous frame inner to this frame (corrupt stack?)

By #ifdef .. endif out sections of code in the
ltp-20061226/testcases/kernel/syscalls/fsync.fsync02.c test program, I
was able to narrow down the cause of the panic to the fstatvfs
function.

I then looked at the linuxolator sources, and didn't find that function.

Any ideas as to what could be the cause of this panic?

Scot
-- 
DISCLAIMER:
No electrons were mamed while sending this message. Only slightly bruised.


More information about the freebsd-emulation mailing list