> > > + /* Start in a signaled state so that even non-submitted batches > > > + * (those that have no draw/clear) can be waited upon. > > > + */ > > > > When would this happen? Presumably if a batch does nothing whatsoever, > > it doesn't make sense to wait on it. > > Was just simpler to have all batch fences containing a syncobjs on > which we can call WAIT_SYNCOBJ even no real fence is attached to it. > > The other option being to set fence->signaled to true when a batch with > no draw/clear is submitted, and then skip those entries when we build > the array passed to the waitsyncobj() func.
I'm wondering the latter option would be cleaner, rather than having extra DRM objects floating around? > > > #include "pan_resource.h" > > > > > > + /* Batch that created this fence object. Will become NULL at > > > batch > > > + * submission time. This field is mainly here to know whether the > > > + * batch has been flushed or not. > > > + */ > > > + struct panfrost_batch *batch; > > > > Could this be replaced by just a `bool flushed`, or do we actually use > > the value in a later patch? > > I actually use the value to flush deps when they're not already flushed. Fair enough. _______________________________________________ mesa-dev mailing list mesa-dev@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/mesa-dev