SCSI regression: sym53c8xxx_2 (2.6test11 vs. 2.6final)

SCSI regression: sym53c8xxx_2 (2.6test11 vs. 2.6final)

[Tom Winkler]

Hello,

First of all: I'm not subscribed to the list so please cc me on replies.

I wanted to report a problem concerning 2.6.0 final with my Tekram DC390U2W
SCSI controller using the sym53c8xx_2 driver. The same machine was running
without problems with 2.6test11 and the exactly same .config file.

Booting 2.6.0 final results in the following messages:

[...]
sym0: <895> rev 0x1 at pci 0000:02:0a.0 irq 22
sym0: Tekram NVRAM, ID 7, Fast-40, LVD, parity checking
sym0: SCSI BUS has been reset.
scsi0: sym-2.1.18b
sym0:0:0: ABORT operation started.
sym0:0:0: ABORT oberation timed-out.
sym0:0:0: DEVICE RESET operation started.
sym0:0:0: DEVICE RESET operation timed-out.
sym0:0:0: BUS RESET operation started.
sym0:0:0: BUS RESET operation timed-out.
sym0:0:0: HOST RESET operation started.
sym0: SCSI BUS has been reset.

After the last line the machine is totally frozen and has to be reset. I'm
really wondering what is causing this since there were not many changes in
drivers/scsi from test11 to final.
Any idea what might be going wrong? test11 still boots and works like a charm
so I don't think that the hardware died over night.

If you need any further information about the specific setup please drop me a
line.

cat /proc/pci    http://www.wnk.at/tmp/pci.txt
.config    http://www.wnk.at/tmp/config_2.6.txt


Thanks,
--
Tom Winkler
e-mail: tom@qwws.net

RE: SCSI regression: sym53c8xxx_2 (2.6test11 vs. 2.6final)

[Tony Battersby]

> Booting 2.6.0 final results in the following messages:
>
> [...]
> sym0: <895> rev 0x1 at pci 0000:02:0a.0 irq 22
> sym0: Tekram NVRAM, ID 7, Fast-40, LVD, parity checking
> sym0: SCSI BUS has been reset.
> scsi0: sym-2.1.18b
> sym0:0:0: ABORT operation started.
> sym0:0:0: ABORT oberation timed-out.
> sym0:0:0: DEVICE RESET operation started.
> sym0:0:0: DEVICE RESET operation timed-out.
> sym0:0:0: BUS RESET operation started.
> sym0:0:0: BUS RESET operation timed-out.
> sym0:0:0: HOST RESET operation started.
> sym0: SCSI BUS has been reset.

I've seen this under 2.4.x when the machine is not processing interrupts
correctly.  I know nothing of 2.6, but it's a good bet that there is
something that is preventing the SCSI interrupt from being processed.
Under 2.4 you can verify this by checking the interrupt count from
/proc/interrupts.  Under 2.4, I would say to try booting with "noapic".
I think 2.6 has different boot options (I remember seeing something
about ACPI and interrupts).  Maybe someone else can help out here.

Anthony J. Battersby
Cybernetics

Re: SCSI regression: sym53c8xxx_2 (2.6test11 vs. 2.6final)

[Matthias Andree]

Tom Winkler writes:

> First of all: I'm not subscribed to the list so please cc me on replies.

[x] done

> I wanted to report a problem concerning 2.6.0 final with my Tekram DC390U2W
> SCSI controller using the sym53c8xx_2 driver. The same machine was running
> without problems with 2.6test11 and the exactly same .config file.
>
> Booting 2.6.0 final results in the following messages:
>
> [...]
> sym0: <895> rev 0x1 at pci 0000:02:0a.0 irq 22
> sym0: Tekram NVRAM, ID 7, Fast-40, LVD, parity checking
> sym0: SCSI BUS has been reset.
> scsi0: sym-2.1.18b
> sym0:0:0: ABORT operation started.
> sym0:0:0: ABORT oberation timed-out.

FWIW, my Symbios 53C875 based Tekram DC-390F is fine with that driver
with Linux 2.6.0 test11 and final, and AFAICS, the driver hasn't changed
between these versions.

Does running "make clean" or saving the .config and "make distclean"
before the build of 2.6.0 final help?

-- 
Matthias Andree

Encrypt your mail: my GnuPG key ID is 0x052E7D95