On 10/01/2015 04:00 AM, Richard Biener wrote:
Apart from what Jakub said - this keeps the list non-recycled for example
after DCE if that doesnt call cleanup_cfg. Likewise after passes that call
cleanup_cfg manually. It also doesn't get called after IPA transform
passes (which would require calling on each function).
To at least catch those passes returning 0 (do nothing) I'd place the
call into execute_todo instead, unconditionally on flags.
As I suspected, the effect of doing this is small. Just under 1%
increase in reused SSA_NAMEs (using an older version of GCC as input
test files).
And just for future reference, the recycled vs allocated ratio is just
under 40%.
@@ -607,6 +626,7 @@ unsigned int
pass_release_ssa_names::execute (function *fun)
{
unsigned i, j;
+ flush_ssaname_freelist ();
which would make this redundant as well. I suppose it would be
interesting to see some before/after
statistics of the release_ssa_names pass. I expect the number of
holes removed to increase, hopefully
not too much (esp. important for analysis passes using sbitmaps of SSA names).
Interestingly enough, flushing the pending freelist in execute_todo
results in marginally fewer holes being found (and it's very marginal --
.04%). Presumably that's because we actually do a better job at
recycling and thus there's fewer holes to be found when
pass_release_ssa_names eventually gets called.
Anyway, I'll spin that up. Hopefully it'll complete before the day is over.
jeff
