> The original code has a loop over num_renames, so if two renames > were in the list in order like > > 8 renames to 4 and 4 renames to 2, then 8 would rename to 2. > however if 4 renamed to 2 then 8 renamed to 4 (not sure if this could > happen), it was different. > > My initial patch fails because it only does one transition per slot, > so if 8 renames to 4, then nothing renames it later to 2.
I see, well, with my patch will not fail for this reason, since a valid rename will always assign the final register index when evaluating the mapping (see 4/6), so if it will accepted than it can (and should) be merged with your rename_temp_registers method. best, Gert _______________________________________________ mesa-dev mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/mesa-dev
