Re: BUG: spinlock recursion on CPU#0, scsi_eh_3/737

[Chris Studholme <cvs@cs.utoronto.ca>]

[-- Attachment #1: Type: text/plain, Size: 1677 bytes --]

Hi Jeff,

With 2.6.14-git5, the BUG doesn't happen, but I still have errors that 
are equally bad.  First I get:

ATA: abnormal status 0xFF on port 0xF880239C
ata4: translated ATA stat/err 0xff/00 to SCSI SK/ASC/ASCQ 0xb/47/00
ata4: status=0xff { Busy }

followed by a series of messages:

ata4: command timeout
ATA: abnormal status 0xFF on port 0xF880239C
ata4: translated ATA stat/err 0xff/00 to SCSI SK/ASC/ASCQ 0xb/47/00
ata4: status=0xff { Busy }
sd 3:0:0:0: SCSI error: return code = 0x8000002
sda: Current: sense key=0xb
    ASC=0x47 ASCQ=0x0
end_request: I/O error, dev sda, sector 9523967

continuing for sectors: 9523967 9523975 9523983 9523991 9523999 9524007
   9524015 9524023 9524031 ... (every 8th sector)

At this point the machine is pretty much useless.  I cannot login.  I
can get a shell from a mutt process that was running before I started
the test, but I can't su to root and cat /proc/mdstat hangs.  All I
could do is reboot from my serial console using break-s (emergency sync)
followed by break-b.

And here's another datapoint.  Sometimes when I boot, my second
processor fails to initialize and I just reboot again to get it started. 
This happened today but I left the machine up with just a single
processor running.  The tar|gzip test completed without any failures in
this case (tried it twice).

Chris.


On Monday, November  7, Jeff Garzik wrote:
> Chris Studholme wrote:
> >Hi,
> >
> >I'm having the fillowing problem.
> >
> >
> >[1.] One line summary of the problem:
> >
> >BUG: spinlock recursion on CPU#0, scsi_eh_3/737
> 
> Can you verify that 2.6.14-git5+ fixes it?
> 
> 	Jeff
> 
> 

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 189 bytes --]