[lkml]   [2012]   [Aug]   [13]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
SubjectRe: [PATCH] x86, pci: Fix all early PCI scans to check the vendor ID first
On Sat, 2012-08-11 at 03:43 -0700, Andi Kleen wrote:
> On Fri, Aug 10, 2012 at 04:57:02PM -0700, H. Peter Anvin wrote:
> > On 08/09/2012 03:34 PM, Betty Dall wrote:
> > >
> > > I thought this should be a break instead of a continue since the code
> > > does a break if the class is 0xffffffff. If the function does not have a
> > > valid VENDOR_ID, then the remaining function numbers do not have to be
> > > scanned because functions are required to be implemented in order (no
> > > skipping a function number.)
> > >
> >
> > Is that true? This is certainly not true in PCI in general: there is
> > required to be a function 0, but there is no guarantee that functions
> > 1-7 don't have gaps.
> These scans are for special known devices, presumably true for them.
> Anwyays if you don't like it please use v1 of the patch.
> -Andi

I checked the PCI specification, and Peter is right that function
numbers can be sparse. Please go with version 1 of the patch, as Andi
suggested. I will follow up by looking at why the three scans are not
consistent and send a patch, if appropriate. The scans could be improved
by stopping the function scan if function 0 does not exist because
function 0 is required, and if it is not there then none of the other
functions will be implemented.


 \ /
  Last update: 2012-08-14 00:42    [W:0.058 / U:3.896 seconds]
©2003-2018 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site