On Mon, Aug 12, 2013 at 08:58:37PM +0200, Mark Kettenis wrote:
> > From: Egbert Eich <[email protected]>
> > Date: Mon, 12 Aug 2013 18:24:47 +0200
> >
> > DACDelay() used to be defined differently on platforms
> > that don't have legacy VGA support.
> >
> > This distinction was removed with
> >
> > commit 6d9efdce0d06df6b85f0681bea306c0b1e851502
> > Author: Adam Jackson <[email protected]>
> > Date: Tue Sep 20 18:12:29 2011 -0400
> >
> > vgahw: Port to pciaccess IO space routines
> >
> > Reviewed-by: Jeremy Huddleston <[email protected]>
> > Tested-by: Jeremy Huddleston <[email protected]>
> > Signed-off-by: Adam Jackson <[email protected]>
> > Reviewed-by: Jamey Sharp <[email protected]>
> >
> > which has produced crashes at least on PowerPC.
>
> Which driver uses DACDelay(), but doesn't actually need working VGA
> legacy support?
Mark, this is exactly the question that I've asked the reporter
when he told me that this code was needed. According to him the
hardware is nv on ppc64. Apparently DACDelay() causes an MC when
the memory that provides access to the legacy PIO address of ST01
is read as there is no response.
I've got the same feeling that this code is papering over an issue
that sits elsewhere. But there's nothing immediately obvious I can
find in this driver. Due to lack of hardware I'm unable to do
further investigations myself. So the only thing I can do here is
to post the patch here and relay back that my concerns are shared
by others.
Cheers,
Egbert.
_______________________________________________
[email protected]: X.Org development
Archives: http://lists.x.org/archives/xorg-devel
Info: http://lists.x.org/mailman/listinfo/xorg-devel
