Bill Rugolsky Jr. wrote:
> On Fri, Mar 17, 2006 at 07:56:53PM -0500, Jeff Garzik wrote:
> 
>>Could I get you to diff against the attached version?
> 
> 
> Certainly.
> 
> I took the opportunity to modify my source file to fix a pair of typos in
> the pci_device_id changes, and make cosmetic changes by switching to enums
> rather than #defines, removing unnecessary braces around single statements,
> and fixing whitespace.  That reduced the noise considerably, and made the
> NV_ADMA_CTL_READ_NON_COHERENT difference manifest.

OK, can you try the attached sata_nv.c?  Does it perform to the level 
that yours does?

It should contain all the functional changes in your patch, while 
ignoring all the noise such as

+	if (host->host_desc->host_type == ADMA)
+		return nv_adma_port_start(ap);
+	else
+		return ata_port_start(ap);

and the hotplug return code changes.

I also intentionally excluded the READ_NON_COHERENT and 
ap->ops->check_status() paranoia function call, to see if those actually 
solve some problems.

	Jeff