Re: Ultrasparc III+ kernel panic

Takeshi Nakayama a écrit :
>>>> BERTRAND_Joël <[email protected]> wrote
>
>> text_access_fault: pc=5ab4bce8 va=5ab4a000
>> kernel trap 64: +fast instruction access MMU miss
>> stopped in pid 14306.1 (netstat) at 5ab4bce8:
>
> I wondered why netstat or pstat (in your past mail) are related to
> OFW, and I found that sysctl CPU_BOOTED_KERNEL uses OFW calls.
>
> I just avoid OFW call from these system commands in the change
> below, so please update your kernel and see what's going to happen.
>
>    http://mail-index.netbsd.org/source-changes/2015/03/15/msg063978.html

	Hello,

	I have tried your patch. Running kernel runs with ddb.onpanic=1 and 
ddb.commandonenter="mach stack". On new kernel, I have rebuilt all 
NetBSD and started pkg_rolling-replace -u. Load average was between 2 
and 3 and kernel crashes after 50 hours of uptime.

	Kernel doesn't enter in ddb (or try to enter in ddb and aborts with a 
deadlock). On console, I only have :
1 tt=30 tstate=4411001503 tpc=0x1001488 tnpc=0x100148c
2 tt=30 tstate=4482000601 tpc=0x12e1da0 tnpc=0x12e1da4
Stopped in pid 16402.1 (sh) at  1010418:		nop
Window 0 frame64 0xe004ff50 locals, ins:
10c89f980 8000000019a30532 116 0 1cd3468 0 0 0
0 0 0 0 0 0 ffffffffffff9341=sp 0=pc:0
Window 1 frame64 0xffffffffffff9b40 locals, ins:

no more information...

	Best regards,

	JKB