Re: 2.6.0-t9 libata: Promise FastTrak S150 TX4 not working

[Jeff Garzik]

On Mon, Oct 27, 2003 at 04:15:05PM +0100, Oliver M. Bolzer wrote:
> Several weeks ago, I tested without success the 2.4.22-bk32-libata1
> patch with my Promise FastTrak S150 TX4 controller. Seeing that
> 2.6.0-test9 merged libata, I tryed again, without success.
> 
> 4 drives are connected to the controller and are configured as 4
> seperate arrays (the thing doesn't have a non-RAID-mode) that are
> working with Promise's partially-binary-only ft3xx driver on 2.4.22.
> 
> Below are the contents of /proc/pci on the machine and dmesg output
> with ATA_DEBUG and ATA_VERBOSE_DEBUG enabled. As it seemed that it
> could not soft-reset the ports, I also compiled the driver without
> the ATA_FLAG_SRST flag for my controller. It looks like the driver
> can recognize the arrays this way but then the kernel enters some
> kind of endless-loop after DMA timeouts. The dmesg output without
> ATA_FLAG_SRST ist also included below. 
> 
> In both cases, the actual error seems to be
> ATA: abnormal status 0x80 on port 0xF881321C
> 
> As I got 8 identical machines with only 3 currenly in beta-use, I'm
> ready to test any kind of patches to help improve the driver.

Wow, this is an excellent bug report.

Yes, it looks like I will need to use a Promise-specific reset
mechanism, rather than one of the three generic mechanisms:
EXECUTE DEVICE DIAGNOSTIC (no flags), ATA software reset
(ATA_FLAG_SRST), and Serial ATA COMRESET (ATA_FLAG_SATA_RESET).

I would prefer to switch both VIA and Promise drivers over to
ATA_FLAG_SATA_RESET.

	Jeff



-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/