[00:32:48] --- Russ has left: Disconnected
[00:35:36] --- Born Fool has become available
[01:18:20] --- Born Fool has left
[01:57:52] --- kaj has become available
[01:59:16] --- Simon Wilkinson has become available
[02:59:27] --- kaj has left
[03:06:25] --- kaj has become available
[03:16:41] --- kaj has left
[09:57:01] --- mattjsm has become available
[10:19:08] --- Russ has become available
[12:07:35] --- mattjsm has left
[15:16:30] --- deason has become available
[15:16:43] --- deason has left
[15:17:41] --- deason has become available
[16:19:56] --- kaduk@mit.edu/barnowl has become available
[16:31:26] --- Kevin Sumner has become available
[16:50:49] <kaduk@mit.edu/barnowl> jhutz: I assume at this point that the next step is to stop staring at
the coredump and reproduce while a tcpdump is running?
[18:22:11] --- cclausen has become available
[18:52:20] <kaduk@mit.edu/barnowl> Oh, hey, I can set debug.witness.kdb=1 and enter the debugger on LORs.
Better stack trace for that rx LOR coming up (I hope) ...
[22:29:44] <kaduk@mit.edu/barnowl> Hm, this one will be fun.  Page fault in kernel mode, virtual address
0x72.  We take the fault in afs_GetDCache, afs_dcache.c:1774 :
   1772         for (index = afs_dchashTbl[i]; index != NULLIDX;) {
   1773             if (afs_indexUnique[index] == avc->f.fid.Fid.Unique) {
   1774                 tdc = afs_GetDSlot(index, NULL);
   1775                 ReleaseReadLock(&tdc->tlock);
afs_chunkops.h:#define  afs_GetDSlot(slot, adc) (*(afs_cacheType->GetDSlot))(slot, adc)
This is memcache, so
(kgdb) p afs_MemCacheOps
$6 = {open = 0xffffff800095adf0 <afs_MemCacheOpen>,
  truncate = 0xffffff800095ae70 <afs_MemCacheTruncate>,
  fread = 0xffffff800095b330 <afs_MemReadBlk>,
  fwrite = 0xffffff800095baf0 <afs_MemWriteBlk>,
  close = 0xffffff800095abf0 <afs_MemCacheClose>,
  vread = 0xffffff80009787f0 <afs_MemRead>,
  vwrite = 0xffffff800097f020 <afs_MemWrite>,
  GetDSlot = 0xffffff800094cee0 <afs_MemGetDSlot>,
  GetVolSlot = 0xffffff800097fdb0 <afs_MemGetVolSlot>,
  HandleLink = 0xffffff800097d860 <afs_MemHandleLink>}

I am confuzzled.
[22:36:00] <kaduk@mit.edu/barnowl> I'm also slightly confused why kgdb thinks that there's no symbol 'i'
in the current context ...
[23:13:03] <kaduk@mit.edu/barnowl> Fatal trap 12: page fault while in kernel mode
cpuid = 0; apic id = 00
fault virtual address   = 0x72
fault code              = supervisor read data, page not present
instruction pointer     = 0x20:0xffffff800094f7d9
stack pointer           = 0x28:0xffffff801fddd7a0
frame pointer           = 0x28:0xffffffff
code segment            = base 0x0, limit 0xfffff, type 0x1b
                        = DPL 0, pres 1, long 1, def32 0, gran 1
processor eflags        = interrupt enabled, resume, IOPL = 0
current process         = 1327 (ls)


geofft notes that 0xffffff800094f7d9 (instruction pointer) appears to
be misaligned.
(kgdb) x/i 0xffffff800094f7d9
0xffffff800094f7d9 <afs_GetDCache+841>: movzwl 0x4a(%rax),%eax
(kgdb) x/10i (0xffffff800094f7d9-5)
0xffffff800094f7d4 <afs_GetDCache+836>: push   %rax
0xffffff800094f7d5 <afs_GetDCache+837>: cmp    %cl,0xffffffffffffff89(%rcx)
0xffffff800094f7d8 <afs_GetDCache+840>: movl   $0x834a40b7,(%rdi)
0xffffff800094f7de <afs_GetDCache+846>: callq  0xffffff7fc11a5de4
0xffffff800094f7e3 <afs_GetDCache+851>: mov    %ax,0x4a(%r15)
0xffffff800094f7e8 <afs_GetDCache+856>:    jne    0xffffff800094f7f5 <afs_GetDCache+869>
0xffffff800094f7ea <afs_GetDCache+858>: cmpb   $0x0,0x48(%r15)
0xffffff800094f7ef <afs_GetDCache+863>:    jne    0xffffff8000950080 <afs_GetDCache+3056>
0xffffff800094f7f5 <afs_GetDCache+869>: mov    %gs:0x0,%rax
0xffffff800094f7fe <afs_GetDCache+878>: mov    0x8(%rax),%rdx
[23:15:16] --- Born Fool has become available
[23:20:38] --- deason has left
[23:35:13] --- Russ has left: Disconnected