Kernel Panic on 9.0 and 9.1 with carp on BCE network interface

[Jean-Luc Dupont]

Hello,

  We are using Freebsd 9.0 on dell poweredge servers with broadcom (bce) integrated interfaces and intel pci-x cards.

This server is used as gateway/firewall (IPFW) with several VLANs and each interface and carp for redundancy with another identical server.

We encounter a kernel panic and reboot when some users connect to an external FTP (just after 
accepting the password).

We upgraded to 9.1-prerelease and 9.1-stable, without any change, still same systematic panic.

Here is the backtrace :

cpuid = 3
KDB: stack backtrace:
#0 0xffffffff808342b6 at kdb_backtrace+0x66
#1 0xffffffff807fe40d at panic+0x1cd
#2 0xffffffff80a21c6f at ufs_dirbad+0x4f
#3 0xffffffff80a2334a at ufs_lookup_ino+0x68a
#4 0xffffffff8088077f at vfs_cache_lookup+0xff
#5 0xffffffff80b3b380 at VOP_LOOKUP_APV+0x40
#6 0xffffffff80887e94 at lookup+0x464
#7 0xffffffff80888fa9 at namei+0x4e9
#8 0xffffffff808a307b at vn_open_cred+0x3bb
#9 0xffffffff808a2117 at kern_openat+0x1e7
#10 0xffffffff80ad1fa6 at amd64_syscall+0x546
#11 0xffffffff80abd967 at Xfast_syscall+0xf7
Uptime: 6m4s

(kgdb) #0  doadump (textdump=Variable "textdump" is not available.
) at /usr/src/sys/kern/kern_shutdown.c:271
#1  0xffffffff807fdf02 in kern_reboot (howto=260)
    at /usr/src/sys/kern/kern_shutdown.c:448
#2  0xffffffff807fe3e3 in panic (fmt=0x104 <Address 0x104 out of bounds>)
    at /usr/src/sys/kern/kern_shutdown.c:636
#3  0xffffffff80a21c6f in ufs_dirbad (ip=Variable "ip" is not available.
)
    at /usr/src/sys/ufs/ufs/ufs_lookup.c:773
#4  0xffffffff80a2334a in ufs_lookup_ino (vdp=0xfffffe00b0aee5a0, 
    vpp=0xffffff834bb30a10, cnp=0xffffff834bb30a38, dd_ino=0x0)
    at /usr/src/sys/ufs/ufs/ufs_lookup.c:393
#5  0xffffffff8088077f in vfs_cache_lookup (ap=Variable "ap" is not available.
) at vnode_if.h:80
#6  0xffffffff80b3b380 in VOP_LOOKUP_APV (vop=0xffffffff8104e960, 
    a=0xffffff834bb306f0) at vnode_if.c:123
#7  0xffffffff80887e94 in lookup (ndp=0xffffff834bb309d0) at vnode_if.h:54
#8  0xffffffff80888fa9 in namei (ndp=0xffffff834bb309d0)
    at /usr/src/sys/kern/vfs_lookup.c:297
#9  0xffffffff808a307b in vn_open_cred (ndp=0xffffff834bb309d0, 
    flagp=0xffffff834bb309cc, cmode=0, vn_open_flags=Variable "vn_open_flags" is not available.
)
    at /usr/src/sys/kern/vfs_vnops.c:195
#10 0xffffffff808a2117 in kern_openat (td=0xfffffe000ff75470, fd=-100, 
    path=0x809ff80c0 <Address 0x809ff80c0 out of bounds>, 
    pathseg=UIO_USERSPACE, flags=1, mode=Variable "mode" is not available.
)
    at /usr/src/sys/kern/vfs_syscalls.c:1132
#11 0xffffffff80ad1fa6 in amd64_syscall (td=0xfffffe000ff75470, traced=0)
    at subr_syscall.c:135
#12 0xffffffff80abd967 in Xfast_syscall ()
    at /usr/src/sys/amd64/amd64/exception.S:387
#13 0x0000000801e5366c in ?? ()
Previous frame inner to this frame (corrupt stack?)
(kgdb) 


We have the coredump and the kernel.debug, if there is anything we can do to help you identifying the origin of this issue.

Thank you very much