Re: PCU panic on V240

[Sujit K M <sjt.kar%gmail.com@localhost>]

On Thu, Dec 5, 2013 at 4:04 PM, J. Hannken-Illjes
<hannken%eis.cs.tu-bs.de@localhost> wrote:
> Running -current on a Sparc V240 I get a diagnostic assertion running
> the script attached:

I don't understand, if you are running a script, you should know what the script
does. I didn't understand what the script does? Could you please let us know
what the script does?

> panic: kernel diagnostic assertion "!cpu_intr_p()" failed: file 
> "/work/build/src/sys/kern/subr_xcall.c", line 351
> Stopped in pid 0.20 (system) at netbsd:cpu_Debugger+0x4:        nop
> db{1}> bt
> db{1}> call stackdump
> Frame pointer is at 0x2b6d6ea41
> Call traceback:
>  netbsd:db_command+0x94(ff00, 0, ffffffffffffffff, 2b6d6f428, 1c19c00, 69) fp 
> = 2b6d6eb71
>  netbsd:db_command_loop+0x110(1c16c80, 1c16cb8, 0, 0, 6, 73) fp = 2b6d6ecb1
>  netbsd:db_trap+0x100(10ef028, 0, 14, 1c6feb9, 0, 1000000) fp = 2b6d6ed91
>  netbsd:kdb_trap+0x134(101, 0, 0, 0, 10d06a0, 1857de0) fp = 2b6d6ee51
>  netbsd:trap+0x398(101, 2b6d6f950, 0, fefefefefefefeff, ffffffffffffffff, 
> ff000000) fp = 2b6d6ef01
>  netbsd:1008d2c+0(2b6d6f950, 101, 10ef020, 1d0006, 2b6d6fb08, 73) fp = 
> 2b6d6f0a1
>  netbsd:vpanic+0x16c(1868da0, 1d32fa8, 18168c0, 2b6d6fc78, 100a080, 
> 2b6d58000) fp = 2b6d6f281
>  netbsd:kern_assert+0x34(18168c0, 2b6d6fc78, 1624520, 1624524, 28, 14c8e80) 
> fp = 2b6d6f331

I would suggest you look for this particular line of code in the debugger and
analyze why the script is failing?

>  netbsd:xc__highpri_intr+0xa8(18168c0, 18168f8, 1820218, 188ce90, 15f, 
> 1c93800) fp = 2b6d6f3f1

My suggestion here is the the interrupt could not be satisfied,
because of priority?

>  netbsd:softint_dispatch+0x104(0, 1c95580, 1d2a290, 0, 18533d8, 1853ce0) fp = 
> 2b6d6f4a1
>  netbsd:softint_fastintr+0x8c(1856f60, 4, 1856f38, 18168c0, 18168f8, 1856d48) 
> fp = 2b6d6f571
>  netbsd:softint_dispatch+0x248(109428880, 4, 1857000, 18168c0, 18168f8, 
> 1857de0) fp = 2b6d6f621
>  netbsd:100030+0(f005eab8, 113800, 113c00, 1118e0, 111d78, 111858) fp = 
> fedc9651

Same here.

> 0
>
> This is the high priority xcall from subr_pcu.c and
> curcpu()->ci_idepth is zero.

Again go to the line number and analyse, my possible suggestion is null pointer.

-- Sujit K M

blog(http://kmsujit.blogspot.com/)