* Insan Praja SW <[EMAIL PROTECTED]> [2008-06-24 18:32]:
> Hi Misc@,
> I currently caught a kernel panic that says:
> uvm_fault(0xffffffff 80b7b0e0, 0x0, 0, 1) -> e
> kernel : page fault trap, code=0
> Stopped at pf_state_tree_RB_REMOVE_COLOR + 0x1C0: cmpl $0x1,0x40(%rsi)
> ddb {0}> trace
this problem has been reported by a few people, but so far we're unable
to track it down or even reproduce. it would help enourmously if we
knew WHEN this was introduced. so if someone who can reproduce this can
compile kernels going backwards day by day (cvs -D) and then ideally even
spot the commit that introduced it, that would help a LOT. yes, it is a
lot of work :(
in short, it seems some element of the pf state table (which is an RB
tree, pf_state_tree) gets freed or overwritten before being removed
from the RB tree, or something tries to remove it before it was
inserted. Ryan and I have been reading the code up and down without
being able to spot such a case yet.
--
Henning Brauer, [EMAIL PROTECTED], [EMAIL PROTECTED]
BS Web Services, http://bsws.de
Full-Service ISP - Secure Hosting, Mail and DNS Services
Dedicated Servers, Rootservers, Application Hosting - Hamburg & Amsterdam
