fatal kernel trap

[Marcel Moolenaar]

On Feb 7, 2012, at 9:46 AM, Anton Shterenlikht wrote:
>> db> bt
>> Tracing pid 3832 tid 100130 td 0xe000000012122000
>> cpu_set_upcall(0xe000000012120cf0, 0xe000000012122000, 0xa0000000f8913780, 0xa0000000f8913550) at cpu_set_upcall+0x190
>> create_thread(0xe000000012122000, 0x0, 0x14064bac0, 0x140804800, 0x7ffffffff3bfe000, 0xc000000, 0x14008c200, 0x140804800) at create_thread+0x1c0
>> kern_thr_new(0xe000000012122000, 0xa0000000f88f3330, 0x9ffc0000004363d0) at kern_thr_new+0x100
>> sys_thr_new(0xe000000012122000, 0xa0000000f88f34e8, 0x9ffc0000008ccbf0, 0x48d) at sys_thr_new+0xa0
>> syscall(0xe000000012112d50, 0xa0000000f88f33a8, 0x14080442c, 0xe000000012122000, 0x0, 0x0, 0x9ffc0000008c80a0, 0x8) at syscall+0x550
>> epc_syscall_return() at epc_syscall_return
>> db> 
> 
> Marcel, these panics make the system unusable on
> r231087, r231124. I'll try to roll back 1000 or
> so and try again. Let me know if you want
> any more info on this panic.

This looks like a regression caused by changes to either
libthr or MI kernel code, for which no MD ia64 code
changes were made, though needed. The backtrace is too
regular...

I'll take a look at it. If you have a simple trigger
case, let me know.

FYI,

-- 
Marcel Moolenaar
xcllnt at mac.com