[Bug tree-optimization/86010] [7/8/9 Regression] redundant memset with smaller size not eliminated
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=86010 Richard Biener changed: What|Removed |Added Target Milestone|7.4 |7.5
[Bug tree-optimization/86010] [7/8/9 Regression] redundant memset with smaller size not eliminated
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=86010 --- Comment #9 from Jeffrey A. Law --- We don't merge the two memset calls in the second example. It's of arguable value IMHO.
[Bug tree-optimization/86010] [7/8/9 Regression] redundant memset with smaller size not eliminated
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=86010 --- Comment #8 from Martin Sebor --- I believe the fix is behind the regression reported in pr87011.
[Bug tree-optimization/86010] [7/8/9 Regression] redundant memset with smaller size not eliminated
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=86010 Jakub Jelinek changed: What|Removed |Added CC||jakub at gcc dot gnu.org --- Comment #7 from Jakub Jelinek --- So fixed for 9+?
[Bug tree-optimization/86010] [7/8/9 Regression] redundant memset with smaller size not eliminated
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=86010 --- Comment #6 from Jeffrey A. Law --- Author: law Date: Tue Jul 17 23:54:10 2018 New Revision: 262841 URL: https://gcc.gnu.org/viewcvs?rev=262841=gcc=rev Log: PR tree-optimization/86010 * tree-ssa-dse.c (compute_trims): Fix typo/thinko. Modified: trunk/gcc/ChangeLog trunk/gcc/tree-ssa-dse.c
[Bug tree-optimization/86010] [7/8/9 Regression] redundant memset with smaller size not eliminated
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=86010 --- Comment #5 from Jeffrey A. Law --- Author: law Date: Fri Jul 6 03:43:42 2018 New Revision: 262464 URL: https://gcc.gnu.org/viewcvs?rev=262464=gcc=rev Log: PR tree-optimization/86010 * tree-ssa-dse.c (compute_trims): More aggressively trim at both the head and tail of mem* and str* calls. Modified: trunk/gcc/ChangeLog trunk/gcc/tree-ssa-dse.c
[Bug tree-optimization/86010] [7/8/9 Regression] redundant memset with smaller size not eliminated
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=86010 --- Comment #4 from Jeffrey A. Law --- Actually, alignment considerations are really just for head trimming. When trimming the tail we can be more aggressive as residuals are usually handled reasonably efficiently.
[Bug tree-optimization/86010] [7/8/9 Regression] redundant memset with smaller size not eliminated
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=86010
--- Comment #3 from Jeffrey A. Law ---
It's not an off by 1 error, but a conscious decision to not pass odd addresses
into the mem* routines, which are typically doing to emit crappy code for badly
aligned data.
We correctly see that 7 bytes of data are live and that we only need a single
byte write for the first memset. One could certainly argue that in the case
were a single byte (or even perhaps up to 3 bytes) are all that's left that we
should do a full trim because the first mem* routine should end up open-coded.
If I hack something like that together we end up with:
MEM[(void *) + 7B] = 0;
__builtin_memset (, 0, 7);
f ();
a ={v} {CLOBBER};
return;
THe fact that the two memset calls use the same value is a testsuite quirk that
would allow this to be further optimized. I'm skeptical that happens often
enough in practice to be worth optimizing. If someone were to want to
optimize that better, I'd think the strlen pass is probably the right place.
[Bug tree-optimization/86010] [7/8/9 Regression] redundant memset with smaller size not eliminated
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=86010 Richard Biener changed: What|Removed |Added Priority|P3 |P2 Status|UNCONFIRMED |NEW Last reconfirmed||2018-06-01 CC||law at gcc dot gnu.org Known to work||4.8.5 Target Milestone|--- |7.4 Summary|[7/8 Regression] redundant |[7/8/9 Regression] |memset with smaller size|redundant memset with |not eliminated |smaller size not eliminated Ever confirmed|0 |1 --- Comment #2 from Richard Biener --- This is the new byte-tracking in DSE I presume which isn't able to prune optimally (off-by-one?) or do the better thing, namely removing the later memset instead of the earlier. With GCC 4.8 we managed to combine the memsets at the RTL level.
