[parisc-linux] Re:[parisc-linux-cvs] linux-2.6 kyle

[Joel Soete]

> On Mon, 2006-08-07 at 19:04 +0200, Joel Soete wrote:
> > > On Mon, 2006-08-07 at 12:32 +0200, Joel Soete wrote:
> > > > Oops my bad: I forgot the raw data, here there are
> > > 
> > > Thanks, that's still not quite enough to say what they were doing at the
> > > time.
> > > 
> > > cpu0 ioaq is 10104c90 10104c94
> > > cpu1 ioaq is 1024a1a4 1024a18c
> > > 
> > > Could you do me the disassembly from the entry symbol down into each of
> > > these so I can try to figure out where they're stuck? (Your prior
> > > disassembly ends just a bit short).
> > > 
> > No pb:
> > CPU#0 is runing:
> > as from System.map:
> 
> Oh, that's actually not quite what I thought it was doing.  Both CPU0
> and 1 are spinning trying to acquire the rw->lock.  That means something
> else has it held for write.  How many CPUs does this system have?
This is a D380 2-way pa8000:
processor       : 0
cpu family      : PA-RISC 2.0
cpu             : PA8000 (PCX-U)
cpu MHz         : 180.000000
model           : 9000/820/D380
model name      : UL proc 2-way U180 1M/1M
hversion        : 0x00005aa0
sversion        : 0x00000491
I-cache         : 1024 KB
D-cache         : 1024 KB (WB, direct-mapped)
ITLB entries    : 96
DTLB entries    : 96 - shared with ITLB
BTLB            : not supported
bogomips        : 359.62
software id     : 2005347373

with 256Mb of ram and 512Mb of swap.

> And what were the rest of them doing?
> 
It was runing a kernel 2.6.18-rc3-pa5 32bit smp (scsi drives ncr53c720 doesn't
yet allow me to try 64bit) and the most its job at the hang moment was a
kernel's compile loop:
#while true ; do nice -n -4 make -j 2 clean ; nice -n -4 make -j 2 oldconfig;
nice -n -4 make -j 2 vmlinux 2>&1 | tee -a /var/logs/k-loop ; done

(A top showing me well at this moment the 2 cpu busy by cc1)

(and no stree loop on fs).

That said the pb can occure (with excately the same kernel and same test)
after few hours or after few days?

and this time, according to backtrace:
D380 32bit smp:
Backtrace:
  [<101560d8>] softlockup_tick+0xd4/0x148
  [<1013be90>] update_process_times+0x3c/0x88
  [<101091dc>] timer_interrupt+0xd0/0x1b0
  [<10156534>] handle_IRQ_event+0x5c/0xa4
  [<10156614>] __do_IRQ+0x98/0x1d0
  [<10109960>] do_cpu_irq_mask+0xdc/0x194
  [<1010c068>] intr_return+0x0/0x1c

N4k 64bit smp:
BUG: soft lockup detected on CPU#0!
Backtrace:
  [<00000000101122b0>] dump_stack+0x18/0x28
  [<0000000010171b50>] softlockup_tick+0x128/0x158
  [<00000000101518f0>] run_local_timers+0x28/0x38
  [<0000000010152660>] update_process_times+0x58/0xd8
  [<000000001011cb98>] smp_do_timer+0x70/0x80
  [<00000000101134cc>] timer_interrupt+0xdc/0x1e0
  [<0000000010171cf4>] handle_IRQ_event+0x74/0xd0
  [<0000000010171e0c>] __do_IRQ+0xbc/0x268
  [<0000000010113e04>] do_cpu_irq_mask+0x114/0x1e0
  [<0000000010104074>] intr_return+0x0/0x1c 

this is very similar to n4k hanging pb?

but unfortunately, even with CONFIG_DEBUG_SPINLOCK & CONFIG_DEBUG_RWLOCK check
(=yes), I didn't reach to grab more info ;-( 

Thanks,
    Joel

> James
> 
> 
> 

----------
Club Scarlet : Tout le monde gagne! Si vous devenez aujourd'hui Scarlet One grace a un client existant de Scarlet, vous recevez tous les deux un cadeau d'une valeur de 50 euros! Surfez vite sur http://www.clubscarlet.be