Current-Users archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

EVDO card issue: more data



I have some more data on the crashes; I hope someone can suggest where
I can look for the problem.

First -- the crash at termination are due to I/O completion when
there's a buffer active that isn't thought to be active.  This suggests
a race condition or lack of locking somewhere.  Most significantly, the
problem can be associated with device close, not removal.  Furthermore,
the system *always* crashes if pppd terminates when card I/O has
blocked.  Again, this sounds like a locking issue.  

However -- there's another problem that's rather more mysterious.  If I
successfully exit pppd *twice*, at some point shortly thereafter the
system will crash.  It isn't a immediate crash -- two days ago, it took
20 minutes -- but it *will* happen, and it doesn't happen under other
circumstances.  Here's a traceback from one crash:

#1  0xc04a41eb in panic ()
#2  0xc054bf83 in trap ()
#3  0xc010cd98 in calltrap ()
#4  0xc0543271 in db_read_bytes ()
#5  0xc01f0857 in db_get_value ()
#6  0xc0543af5 in db_stack_trace_print ()
#7  0xc04a41c1 in panic ()
#8  0xc054bf83 in trap ()
#9  0xc010cd98 in calltrap ()
#10 0xc0495ae0 in turnstile_block ()
#11 0xc047dafb in mutex_vector_enter ()
#12 0xc047a858 in lockmgr ()
#13 0xc041d9b9 in ufs_ihashins ()
#14 0xc03fd7f6 in ffs_vget ()
#15 0xc041fd3e in ufs_lookup ()
#16 0xc04e506c in VOP_LOOKUP ()
#17 0xc04d438c in lookup ()
#18 0xc04d4a3d in namei ()
#19 0xc04dc037 in do_sys_stat ()
#20 0xc04dc103 in sys___stat30 ()
#21 0xc054b8f1 in syscall ()
#22 0xc010051d in syscall1 ()

and the panic message:

uvm_fault(0xd1827940, 0, 1) -> 0xe
fatal page fault in supervisor mode
trap type 6 code 0 eip c0495ae0 cs 8 eflags 210286 cr2 0 ilevel 6
panic: trap
Begin traceback...
uvm_fault(0xd1827940, 0, 1) -> 0xe
fatal page fault in supervisor mode
trap type 6 code 0 eip c0543271 cs 8 eflags 210246 cr2 0 ilevel 6
panic: trap

I'm at a loss to guess what's being left behind that could cause this.
At the time of this crash, the USB device wasn't even physically
connected to the system.  Any suggestions on what I should be looking
for?

                --Steve Bellovin, http://www.cs.columbia.edu/~smb



Home | Main Index | Thread Index | Old Index