[Intel-gfx] [PATCH] drm/i915: Ensure dbuf state is defined

[Kahola, Mika]

> -----Original Message-----
> From: Ville Syrjälä <ville.syrjala at linux.intel.com>
> Sent: Wednesday, October 5, 2022 2:29 PM
> To: Kahola, Mika <mika.kahola at intel.com>
> Cc: intel-gfx at lists.freedesktop.org
> Subject: Re: [Intel-gfx] [PATCH] drm/i915: Ensure dbuf state is defined
> 
> On Wed, Oct 05, 2022 at 02:11:58PM +0300, Mika Kahola wrote:
> > Ensure that the new dbuf state is not null. If so, throw an error,
> > discontinue to allocate ddb and return.
> >
> > Signed-off-by: Mika Kahola <mika.kahola at intel.com>
> > ---
> >  drivers/gpu/drm/i915/display/skl_watermark.c | 3 +++
> >  1 file changed, 3 insertions(+)
> >
> > diff --git a/drivers/gpu/drm/i915/display/skl_watermark.c
> > b/drivers/gpu/drm/i915/display/skl_watermark.c
> > index d58e667016e4..4ecf3b43d6d2 100644
> > --- a/drivers/gpu/drm/i915/display/skl_watermark.c
> > +++ b/drivers/gpu/drm/i915/display/skl_watermark.c
> > @@ -638,6 +638,9 @@ skl_crtc_allocate_ddb(struct intel_atomic_state
> *state, struct intel_crtc *crtc)
> >  	u32 start, end;
> >  	int ret;
> >
> > +	if (IS_ERR(new_dbuf_state))
> > +		return PTR_ERR(new_dbuf_state);
> > +
> 
> You can't get here without the dbuf_state. Also even if we didn't have one
> intel_atomic_get_new_dbuf_state() certainly shouldn't give us an error pointer
> (you either get a valid pointer or NULL).
Ok, so we can get a null here. Should we add a check for that one?

> 
> >  	if (new_dbuf_state->weight[pipe] == 0) {
> >  		skl_ddb_entry_init(&new_dbuf_state->ddb[pipe], 0, 0);
> >  		goto out;
> > --
> > 2.34.1
> 
> --
> Ville Syrjälä
> Intel