Subject: Re: anyone seen a crash in pmap_emulate_reference?
To: None <firstname.lastname@example.org>
From: Matthew Jacob <email@example.com>
Date: 05/03/1997 19:20:54
>On Fri, 2 May 1997, Matthew Jacob wrote:
>> I've had my multia crash a couple of times in pmap_emulate_reference.
>> Kernel's of last week or so origin.
>I'm seeing lots of crashes, almost all VM or pmap related. I'm
>working on further analysis of the problem, but it's slow going.
On alpha alone, or other architectures?
Also, odd other things are happening. I got at least one pmap_emulate_reference
whine when doing a *copyout* from msgbuf to user space today. Why it
would get a fault on the K0SEG address is a puzzlement to me- it's
mapped by the prom, not 'managed' by VM, so how could there be fault
I got one 'pmap_page_protect' panic back on 4/18.
Also, we have a variety of cc1 core dumps. I've now had it happen
twice, but I have no clue from the cc1 core files why- it isn't
an unresolved stack fault.
I've done a bit of modifying of alpha/trap.c to print out info
about a process killed of with SIGSEGV- I may commit this sooner
or later as it's useful (IMHO), and only on for DIAGNOSTIC. I just
tested it with a little 'eatstack' program.