Re: [patch] scsi/libata: correct bug for ULi M5281

[Jeff Garzik <jgarzik@pobox.com>]

Clear.Zhang@uli.com.tw wrote:
> Hi, Jeff
> 
> The sata_uli patch for m5281 looks fine,but some of M5281 chip's PHY
> is not stable sometime, so it need a workaround to correct the bug.
> 
> patch_M5281workaround is just for it.
> 
> The patch is applied to kernel 2.6.9. Pleases apply to new kernels, thanks
> a lot.
> 
> Signed-off-by: Clear Zhang <Clear.Zhang@uli.com.tw>

I'm afraid I must reject this patch.  We do not wish to put
hardware-specific workarounds in libata-core.c or libata-scsi.c.

I would suggest either
(a) finding a way to add the workaround to sata_uli.c, or
(b) finding the "root cause" of the problem, and determine why the
libata code is insufficient.

Two suggestions I have while investigating your problem:
1) Replacing the ->phy_reset() hook may allow you to solve the problem
in sata_uli.c.
2) It is entirely possible that we need to add additional error handling
code to libata, including hooks back into the low-level driver (such as
sata_uli.c).

Regards,

	Jeff