svn commit: r219737 - head/sys/dev/pci

[John Baldwin]

On Friday, March 18, 2011 8:48:28 am Nathan Whitehorn wrote:
> On 03/18/11 07:13, John Baldwin wrote:
> > Author: jhb
> > Date: Fri Mar 18 12:13:04 2011
> > New Revision: 219737
> > URL: http://svn.freebsd.org/changeset/base/219737
> >
> > Log:
> >    Fix a few issues with HyperTransport devices and MSI interrupts:
> >    - Always enable the HyperTransport MSI mapping window for HyperTransport
> >      to PCI bridges (these show up as HyperTransport slave devices).
> >      The mapping windows in PCI-PCI bridges are enabled by existing code
> >      in the PCI-PCI bridge driver as MSI requests propagate up the device
> >      tree, but Host-PCI bridges don't really show up in that tree.
> >    - If the PCI device at domain 0 bus 0 slot 0 function 0 is not a
> >      HyperTransport device, then blacklist MSI on any other HT devices in
> >      the system.  Linux has a similar quirk.
> >
> 
> I think this last change only works on x86 systems. My powerpc desktop 
> has hypertransport and MSI, but PCI bus 0 (the PCI-E bus with the 
> graphics card in it) is not connected over hypertransport and has 
> nothing to do with it. Also, the root host->HT bridge doesn't show up in 
> PCI space at all.

Hmm, that is likely true.  Rereading the Linux code, this last quirk only
applies to Nvidia chipsets.  Bah, I also botched it, so I'll remove that part
for now.  Only the first fix was needed for the PR.

-- 
John Baldwin