Debian Sparc4 goes crazy (SCSI error)

[Daniel Smith]

On Wed, Sep 18, 2002 at 01:58:09PM +1000 or thereabouts, Antti.Roppola at brs.gov.au wrote:

All very well behaved here. Almost identical config. I've got a HME SBus
card as my ethernet interface.
> 
> Has anyone one had issues with Debian Sparc going nuts?
> (Potato with a stock kernel)

Which stock kernel? 2.2.19|20 Is there a Debian 2.4 Sparc?

> 
> 	esp0: DMA error a4400202
> 	esp: Resetting scsi bus
> 
> repeatedly to console before refusing to listen to me at all.
> Reboot, fsck (messy) and it's been running OK since.

Ick. Migrate to ext3 if you can. It's designed to deal with drives that
behave in unfriendly ways.

> The only reference I have found on Usenet or Google was
> SCSI termination on an IPX, but since it was working before,
> and it's an SCA disk, I can't see it being the case here.

Doesn't look like SCSI chain. The only place 'DMA error' is mentioned in
the driver is ..

        if (sbus_readl(esp->dregs + DMA_CSR) & DMA_HNDL_ERROR) {
                /* A DMA gate array error.  Here we must
                 * be seeing one of two things.  Either the
                 * virtual to physical address translation
                 * on the SBUS could not occur, else the
                 * translation it did get pointed to a bogus
                 * page.  Ho hum...
                 */
       ==>	ESPLOG(("esp%d: DMA error %08x\n", esp->esp_id,
                        sbus_readl(esp->dregs + DMA_CSR)));

                /* DMA gate array itself must be reset to clear the
                 * error condition.
                 */
                esp_reset_dma(esp);

                what_next = do_reset_bus;
                goto state_machine;
        }

..and it does pretty much what the driver says it's doing, ie trying to
reset the SCSI bus. As to whether it's succeeding *dunno*.

Maybe you might want to capture your logs on another machine to track if
a previous oops had occured.

Other than that, what was the uptime pre splat?

Daniel