Re: Ultrasparc III+ kernel panic

Martin Husemann a écrit :
> On Mon, Mar 09, 2015 at 11:44:47AM +0100, BERTRAND Joël wrote:
>> 	I have built last saturday and I'm building a 7.99.6 (cvs update from this
>> morning).
>
> I expect more usable backtraces from this one.
>
> Martin
>

	Hello,

	I'm not sure they are more usable...

Mar 10 22:14:30 legendre /netbsd: cpu1: data fault: pc=f00092e0 
rpc=103b3d5e0 ad
dr=fffffff52c8a2000
Mar 10 22:14:30 legendre /netbsd: text_access_fault: pc=5db0fcd8 va=5db0e000
Mar 10 22:14:30 legendre /netbsd: kernel trap 30: data access exception
Mar 10 22:14:30 legendre /netbsd: k
Mar 10 22:14:30 legendre /netbsd: cpu1: rebooting
Mar 10 22:14:30 legendre /netbsd:

	ddb only returns :

cpu1: data fault: pc=f00092e0 rpc=103b3d5e0 addr=fffffff52c8a2000
text_access_fault: pc=5db0fcd8 va=5db0e000
kernel trap 30: data access exception
Stopped in pid 9084.1 (pstat) at f00092e0:	stx %l0, [%g0 + %g4]
tstate	82000603
pc	f00092e0
npc	f00092e4
ipl	f
y	0
g0	0
g1	1
g2	f0000000
g3	fff78000
g4	fffffff52c8a3af8
...

	I only have register trace. ddb crashs when it try to print backtrace.

	On this server, I have replaced bridge0 interface by regular ethernet 
interface (hme0) and an old 3Com switch. I have noticed that uptime is 
greater when I disable bridged interface.

	Regards,

	JKB