On Wed, 2003-08-06 at 12:58, Michel Dänzer wrote:
>
> The patch was flawed (it would free map even if it was NULL), try this
> one instead.
Still broken. :( The maplist entry needs to be unlinked before freeing
it.
If this still doesn't cut it, I hope you get the idea and fix it
properly. :)
-
On Wed, 2003-08-06 at 13:14, Michel Dänzer wrote:
> On Wed, 2003-08-06 at 12:58, Michel Dänzer wrote:
> >
> > The patch was flawed (it would free map even if it was NULL), try this
> > one instead.
>
> Still broken. :( The maplist entry needs to be unlinked before freeing
> it.
>
> If this still
On Wed, 2003-08-06 at 12:51, Michel Dänzer wrote:
>
> Does this patch fix it?
The patch was flawed (it would free map even if it was NULL), try this
one instead.
--
Earthling Michel Dänzer \ Debian (powerpc), XFree86 and DRI developer
Software libre enthusiast \ http://svcs.affero.net/
On Mon, 2003-08-04 at 17:51, Benjamin Herrenschmidt wrote:
>
> It's actually dying in drm_ioremapfree() called by radeon_takedown(),
> This is a uninorth based PowerMac, so it's using the vmap mecanism
> (cant_use_aperture is 1). It's accessing a freed object (apparently,
> a pointer set to 6b6b
Ok, this is more infos about what I reported earlier (it was DRM crash
on VT switch & suspend).
Actually, the problem happens when closing the DRM files, which I had
happening with VT switch & suspend because I was use DRIReInit patch.
Without this (now that DRM suspend is merged), I only see this