Re: ioremap_uc() followed by set_memory_wc() - burrying MTRR

On Wed, Apr 22, 2015 at 12:10 PM, Doug Ledford [off-list ref] wrote:

On Wed, 2015-04-22 at 21:05 +0200, Luis R. Rodriguez wrote:

quoted

quoted

quoted

I'd also love to remove the driver if it turns out there are actually
no users. qib substantially replaces it except for a few very old
cards.

To be precise, the split is that ipath powers the old HTX bus cards that
only work in AMD systems,

Do those systems have PAT support? CAn anyone check if PAT is enabled
if booted on a recent kernel?

I don't have one of these systems any more.  The *only* one I ever had
was a monster IBM box...I can't even find a reference to it any more.

Um, yeah if its so rare then I think the compromise proposed might
make sense, specially since folks were even *considering* seriously
removing this device driver. I'll send some patches to propose the
strategy explained to require booting with pat disabled.

 Luis