[Intel-gfx] [RFC] kill pI830->memory_manager checks

[Magnus Kessler]

On Wednesday 13 May 2009 01:25:29 Jesse Barnes wrote:
> On Tue, 12 May 2009 17:08:41 -0700
>
> Eric Anholt <eric at anholt.net> wrote:
> > On Tue, 2009-05-12 at 17:02 -0700, Jesse Barnes wrote:
> > > This patch gets rid of the ugly KMS hack that messed with videoRam
> > > by removing all of the pI830->memory_manager checks in favor of
> > > what we're really looking for: GEM.  Needs some review though; in
> > > some places we might just be looking for any bufmgr, not just GEM...
> >
> > I like it a lot, except for:
> > > --
> > > Jesse Barnes, Intel Open Source Technology Center
> > >
> > > diff --git a/src/i830.h b/src/i830.h
> > > index 0d8726c..68bc0a5 100644
> > > _
> > > diff --git a/src/i830_driver.c b/src/i830_driver.c
> > > index aaf5a20..ef90123 100644
> > > --- a/src/i830_driver.c
> > > +++ b/src/i830_driver.c
> > > @@ -2556,12 +2557,7 @@ I830ScreenInit(int scrnIndex, ScreenPtr
> > > pScreen, int argc, char **argv) #endif
> > >     }
> > >
> > > -   if (pI830->use_drm_mode) {
> > > -       pI830->stolen_size = 0;
> > > -       pScrn->videoRam = ~0UL / KB(1);
> > > -   } else {
> > > -       I830AdjustMemory(pScreen);
> > > -   }
> > > +   I830AdjustMemory(pScreen);
> >
> > calling AGP functions in KMS mode sounds like failure.
>
> Yes it does... suck.  Peter can you try it with that call protected by
> a !pI830->use_drm_mode check?

I tested the patch with the proposed change to protect the call to 
I830AdjustMemory(), and it works fine for me on 64-bit Linux with KMS.

Many thanks,

Magnus

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 198 bytes
Desc: This is a digitally signed message part.
URL: <http://lists.freedesktop.org/archives/intel-gfx/attachments/20090513/99b44e7e/attachment.sig>