Re: [PATCH 7/7] bootirqquirk= parameter to enable bootirq quirks for additional chips

[Olaf Dabrunz <od@suse.de>]

On 03-Jun-08, Thomas Gleixner wrote:
> On Mon, 2 Jun 2008, Olaf Dabrunz wrote:
> > From: Olaf Dabrunz <od@suse.de>
> > 
> > The existing bootirq quirks and the reroute workaround may work for other chips
> > where we could not test them. This parameter allows users to apply these to
> > other chips without the need to re-build the kernel.
> > 
> > This patch was conceived simultaneously by Stefan Assmann, Daniel Gollub and
> > Olaf Dabrunz. The implementation is the author's.
> > 
> >     bootirqquirk=0x<vendor>,0x<device>,<quirk_type>
> > 
> >     - quirk type  1 - 32 selects an IRQ reroute algorithm for devices
> >       connected to that PCI bridge (currently only algorithm "1" is
> >       implemented),
> > 
> >     - quirk type 33 -  x applies one of the known quirks to the PCI
> >       device, currently these:
> > 
> >         33 -> quirk_disable_intel_boot_interrupt
> >         34 -> quirk_disable_broadcom_boot_interrupt
> >         35 -> quirk_disable_amd_8111_boot_interrupt
> >         36 -> quirk_disable_amd_813x_boot_interrupt
> >         37 -> quirk_disable_amd_sb700s_boot_interrupt
> 
> Oh no. This can be used as an debug aid when the need arises, but we
> dont want to add this to the kernel.

Yes, I agree. We should rather patch in additional chips as needed.
This is more a helper functionality for early testing.

Regards,

-- 
Olaf Dabrunz (od/odabrunz), SUSE Linux Products GmbH, Nürnberg