re PR rtl-optimization/36438 (gcc ICE compiling code with mmx builtin)
authorUros Bizjak <uros@gcc.gnu.org>
Fri, 6 Jun 2008 15:04:51 +0000 (17:04 +0200)
committerUros Bizjak <uros@gcc.gnu.org>
Fri, 6 Jun 2008 15:04:51 +0000 (17:04 +0200)
commit88057dc868335e8f59eb0798d605c29de154de6d
tree213585bbf8dfd4c6d9d250a5e3d911115c057f6d
parent8b34d6c2bfcf7e70d3ab0909cb73ba347f47160e
re PR rtl-optimization/36438 (gcc ICE compiling code with mmx builtin)

PR rtl-optimization/36438
* cse.c (fold_rtx) [ASHIFT, LSHIFTRT, ASHIFTRT]: Break out early
for vector shifts with constant scalar shift operands.

From-SVN: r136486
gcc/ChangeLog
gcc/cse.c
gcc/testsuite/gcc.target/i386/pr36438.c [new file with mode: 0644]