On Mon, Jan 16, 2017 at 09:26:53AM +0100, Eric Botcazou wrote: > rs6000 (Fix reload failures in 64-bit mode): > https://gcc.gnu.org/ml/gcc-patches/2016-11/msg00568.html
The one thing I find questionable about this patch is that it will set mem_alias_set for the resulting MEMs to the TOC alias set. Is that correct for memory not in the TOC? My guess is that in this case it is OK, because you're handling constants forced to memory. By giving them a non-zero alias set you're saying that they won't alias other mem accesses, which should be true. I'm commenting because when you first posted the patch, Segher asked me whether I had any concerns and I replied with the first paragraph or something like it. So I'm probably partly to blame for the delay in review. The second paragraph is me rethinking that concern. -- Alan Modra Australia Development Lab, IBM
