On Fri, Nov 30, 2012 at 04:42:45AM -0800, sig6247 wrote: > On Fri, 30 Nov 2012 01:29:44 +0200, Konstantin Belousov <kostikbel_at_gmail.com> wrote: > > > Please apply the patch below and obtain the backtrace of the double fault > > panic again. I will commit the patch later. > > Thanks for the patch. > > WARNING: WITNESS option enabled, expect reduced performance. > Trying to mount root from zfs:zroot []... > > Fatal double fault: > eip = 0xc0e93e07 > esp = 0xc86bffbc > ebp = 0xc86c0032 > cpuid = 1; apic id = 01 > panic: double fault > cpuid = 1 > KDB: enter: panic > [ thread pid 1 tid 100002 ] > Stopped at kdb_enter+0x3d: movl $0,kdb_why > db> bt > Tracing pid 1 tid 100002 td 0xc89efbc0 > kdb_enter(c1065960,c1065960,c10b903b,c139f438,aef01785,...) at kdb_enter+0x3d/frame 0xc139f3f0 > panic(c10b903b,1,1,1,c86c0032,...) at panic+0x14b/frame 0xc139f42c > dblfault_handler() at dblfault_handler+0xab/frame 0xc139f42c > --- trap 0x17, eip = 0xc0e93e07, esp = 0xc86bffbc, ebp = 0xc86c0032 --- > __qdivrem(0,aba80000,c10d,98600000,68c119,...) at __qdivrem+0x197/frame 0xc86c0032 > db> Hm, this is not very useful. Although the panic is again caused by the stack overflow, most likely (please also include the output of the "show thread" from ddb), it is at different place, and probably at the leaf function. Can you try some more times, so that we could see 'big' backtrace ?
This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:40:32 UTC