On 04/07/2014 12:56 PM, Daniel Vetter wrote: > On Mon, Apr 7, 2014 at 5:19 PM, Ian Romanick <idr at freedesktop.org> wrote: >> On 04/05/2014 02:44 AM, Daniel Vetter wrote: >>> ttm_bo_unref unconditionally calls kref_put on it's argument, so the >>> thing can't be NULL without already causing Oopses. >> >> Doesn't this mean the NULL check is in the wrong place (rather than the >> NULL check should be removed)? > > Afaics chasing callchains it's a bug to call this with NULL pointer > and no one really should be doing it. Like David Herrmann said it's > sometimes useful if unref/free functions automatically cope with NULL, > but ttm buffers don't seem to be of this kind. So consistency with > other ttm drivers seems better, same with all the gem_free_object > callbacks.
That's fair. I'm convinced. Patches 1, 3, and 5 are also Reviewed-by: Ian Romanick <ian.d.romanick at intel.com> > -Daniel
