https://gcc.gnu.org/bugzilla/show_bug.cgi?id=78057
--- Comment #8 from uros at gcc dot gnu.org ---
Author: uros
Date: Wed Oct 26 15:13:39 2016
New Revision: 241556
URL: https://gcc.gnu.org/viewcvs?rev=241556=gcc=rev
Log:
Backport from mainline
2016-10-21 Jakub Jelinek
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=78057
Jakub Jelinek changed:
What|Removed |Added
Status|ASSIGNED|RESOLVED
Resolution|---
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=78057
--- Comment #6 from Jakub Jelinek ---
Author: jakub
Date: Fri Oct 21 15:39:25 2016
New Revision: 241411
URL: https://gcc.gnu.org/viewcvs?rev=241411=gcc=rev
Log:
PR target/78057
* config/i386/i386.c: Include fold-const-call.h,
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=78057
--- Comment #5 from Jakub Jelinek ---
Created attachment 39863
--> https://gcc.gnu.org/bugzilla/attachment.cgi?id=39863=edit
gcc7-i386-fold.patch
And incremental patch to add a bunch of BMI/BMI2/TBM intrinsics folding.
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=78057
Jakub Jelinek changed:
What|Removed |Added
Attachment #39855|0 |1
is obsolete|
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=78057
--- Comment #3 from Jakub Jelinek ---
I've been concerned about something like:
int a = ((int (*) (void)) __builtin_clz) ();
int b = ((int (*) (float, double)) __builtin_clz) (0.5f, 1.0);
but that is apparently rejected. Perhaps builtins
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=78057
--- Comment #2 from Uroš Bizjak ---
+ case IX86_BUILTIN_TZCNT16:
+ case IX86_BUILTIN_TZCNT32:
+ case IX86_BUILTIN_TZCNT64:
+ if (n_args == 1 && TREE_CODE (args[0]) == INTEGER_CST)
Maybe:
...
case
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=78057
Jakub Jelinek changed:
What|Removed |Added
Status|UNCONFIRMED |ASSIGNED
Last reconfirmed|
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=78057
Uroš Bizjak changed:
What|Removed |Added
Target||x86
Target Milestone|---