dump hangs when initiated from kdb

[Bruce Edge]

When lkcd is called from kdb, the dump process hangs:

Entering kdb (current=0x9d534000, pid 883) on processor 2 Oops: divide error
due to oops @ 0xc2d16edd
eax = 0x0000000a ebx = 0x00000000 ecx = 0x0000000a edx = 0x00000000
esi = 0x9db926e0 edi = 0x00000000 esp = 0x9d535ef0 eip = 0xc2d16edd
ebp = 0x9d535f08 xss = 0x00000018 xcs = 0x00000010 eflags = 0x00010246
xds = 0xb3e70018 xes = 0x00000018 origeax = 0xffffffff &regs = 0x9d535ebc
Forcing dump (if configured)
LKCD entered from KDB
Dumping from interrupt handler !
Uncertain scenario - but will try my best
                 
dump: Dumping to device 0x1602 [ide1(22,2)] on CPU 2 ...
dump: Compression value is 0x0, Writing dump header     <-no more output after this

Before I fixed the lkcd patch to apply correctly on top of a kdb patched kernel, the 
dump was being called from lkcd directly, bypassing kdb.
This yielded a truncated dump image, however the resultant analysis file was still 
useable.

Admittedly, this is being called from within an interrupt handle, but it worked better 
bypassing kdb. 

-Bruce


Use http://oss.sgi.com/ecartis to modify your settings or to unsubscribe.