Re: [PATCH] pci-quirks: disable MSI on RS400-200 and RS480, take #2

[Jay Cliburn <jacliburn@bellsouth.net>]

Jeff Garzik wrote:
> Chuck Ebbert wrote:
> > Tejun Heo wrote:
> > > > FWIW several distros have turned off MSI by default and added
> > > > a "pci=msi" option to enable it.
> > > Yeah, it seem to cause a lot of problems on certain chips but I think
> > > the correct path is to add PCI quirks for those.  Most MSI problems I've
> > > seen are on these ATI chips.  Do you happen to know any other?
> >
> > We had devices that didn't do MSI right, e.g. forcedeth and others I
> > can't recall now.
>
> AFAIK that's a broken diagnosis.   It's the system, not the device, that 
> is problematic.

In the case of the Attansic L1 ethernet driver, here's what we see:

chipset			kernel arch	MSI functionality
===============		===========	=================
Intel 945G/ICH7		x86_64		yes
Intel 945G/ICH7		i386		yes
Via K8T890		x86_64		yes
Via K8T890		i386		no

I still don't know why, but we get a flood of APIC errors after starting the 
atl1 driver on a Via K8T890 board (Asus M2V, for example) under a 32-bit kernel, 
and *only* under a 32-bit kernel.

http://lkml.org/lkml/2007/4/8/68

Supporting files and such at ftp://ftp.hogchain.net/pub/linux/m2v/apic-problem

Any hints heartily welcomed...

Jay

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/