Re: Trouble with Symbios Logic 53c810 on XL300 with Kernel v2.6
On Wed, May 11, 2005 at 06:14:30PM +0200, Jerome Warnier wrote:
> [..]
>
> > My onboard SCSI controller (in an Alpha XL300) is:
> >
> > 0000:00:09.0 SCSI storage controller: LSI Logic / Symbios Logic 53c810
> > (rev 11)
> > Flags: bus master, medium devsel, latency 32, IRQ 28
> > I/O ports at 9000 [size=256]
> > Memory at 0000000002231000 (32-bit, non-prefetchable) [size=256]
> If I remember correctly, there are several kernel modules supporting
> those cards in 2.6. Did you already try the other?
Thanks, but I've already started with kernel hacking;) So I've figured
also out what is producing this bug. Unfortunately I can't offer a well
formed patch for it.
Ok, let's jump into some details. The problem is the narrow/wide mode.
The disks with problems are 68-pin disks, the cable is 50-pin narrow and
the working devices on the bus are also 50-pin narrow devices.
Unfortunately the 68-pin disks have no jumpers to force them into narrow
mode:(
And so they reach the "test width" part in the
function spi_dv_device_internal() in the file
drivers/scsi/scsi_transport_spi.c which I've copied here
/* test width */
if (i->f->set_width && sdev->wdtr) {
i->f->set_width(sdev, 1);
if (!spi_dv_device_compare_inquiry(sreq, buffer,
buffer + len,
DV_LOOPS)) {
SPI_PRINTK(sdev, KERN_ERR, "Wide Transfers
Fail\n");
i->f->set_width(sdev, 0);
}
}
Both 68-pin disks will be set to _wide_ "1" with set_width(sdev, 1) and
they don't reach the clearing part set_width(sdev, 0) where they will be
set to _narrow_ "0" again. So those 68-pin disks are running in wide
mode all the time on my narrow bus. This should explain my trouble. But I
still don't understand why this happens?
In the debian kernel patch collection is also a patch named
drivers-scsi-sym53c8xx_revert.dpatch.bz2 which does no longer apply :(
It could be possible that it also solves it. But my machine needs so
muuuuuch tiiiiiiime for compiling. I haven't had the time until now to
test it and some other things I want to...
martin
Reply to: