[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] Re: [Xen-devel] [Intel-gfx] [RFC][PATCH] gpu:drm:i915:intel_detect_pch: back to check devfn instead of check class type
> From: Daniel Vetter > Sent: Monday, July 07, 2014 11:40 AM > > On Mon, Jul 07, 2014 at 07:58:30PM +0200, Paolo Bonzini wrote: > > Il 07/07/2014 19:54, Daniel Vetter ha scritto: > > >On Mon, Jul 07, 2014 at 04:57:45PM +0200, Paolo Bonzini wrote: > > >>Il 07/07/2014 16:49, Daniel Vetter ha scritto: > > >>>So the correct fix to forward intel gpus to guests is indeed to somehow > > >>>fake the pch pci ids since the driver really needs them. Gross design, > > >>>but > > >>>that's how the hardware works. > > >> > > >>A way that could work for virtualization is this: if you find the card > > >>has a > > >>magic subsystem vendor id, fetch the subsystem device id and use _that_ > as > > >>the PCH device id. > > >> > > >>Would that work for you? > > > > > >I guess for quemu it also depends upon what windows does since we can't > > >change that. If we care about that part. Another consideration is > > >supporting older kernels, if that's possible at all. > > > > Yes, but right now it's more important to get something that's not too gross > > for the future, for both Linux and Windows. Hacks for existing guests can > > be done separately, especially since they might differ between Linux (check > > ISA bridge) and Windows (check 1f.0). > > Well old Linux also checked 1f.0, so kinda the same really. As long as > 1f.0 is an isa bridge. Wrt Windows I don't really expect them to change > this (they're probably more focuesed on the windows hypervisor or whatever). discussion is also on-going with Windows driver folks. Add Allen here. > > In the end if the approach is ok for quemu and isn't much worse than what > we currently have I don't mind at all about the i915.ko code. I just want > to avoid flip-flopping around on the hack du jour like we seem to do just > now. > -Daniel actually I'm curious whether it's still necessary to __detect__ PCH. Could we assume a 1:1 mapping between GPU and PCH, e.g. BDW already hard code the knowledge: } else if (IS_BROADWELL(dev)) { dev_priv->pch_type = PCH_LPT; dev_priv->pch_id = INTEL_PCH_LPT_LP_DEVICE_ID_TYPE; DRM_DEBUG_KMS("This is Broadwell, assuming " "LynxPoint LP PCH\n"); Or if there is real usage on non-fixed mapping (not majority), could it be a better option to have fixed mapping as a fallback instead of leaving as PCH_NONE? Then even when Qemu doesn't provide a special tweaked PCH, the majority case just works. Thanks Kevin _______________________________________________ Xen-devel mailing list Xen-devel@xxxxxxxxxxxxx http://lists.xen.org/xen-devel
|
Lists.xenproject.org is hosted with RackSpace, monitoring our |