| commit | 66592e95db41216e6fb1d76a831f8efa6bcf0d6f | |
| author | Alan Modra <amodra@gmail.com> | |
| Wed, 17 Aug 2016 22:41:22 +0000 (18 08:11 +0930) | ||
| committer | Alan Modra <amodra@gcc.gnu.org> | |
| Wed, 17 Aug 2016 22:41:22 +0000 (18 08:11 +0930) | ||
| tree | f81e4c1cdce3465669f486e5c83c18bd770bc7ee | treesnapshot (tar.gz zip) |
| parent | 6fbf26cc575cd5d50acc63a0d935744399230435 | commitdiff |
[RELOAD] Don't assume subreg mem address is ok
This patch fixes a case where reload blindly assumes a subreg mem is
OK if its address has been partially reloaded by legitimize_reload_address.
PR rtl-optimization/72771
* reload.c (find_reloads): Don't assume that a subreg mem is OK
when find_reloads_toplev returns address_reloaded==-1.
(alternative_allows_const_pool_ref): Update comment.
testsuite/
* gcc.c-torture/compile/pr72771.c: New.
From-SVN: r239549
This patch fixes a case where reload blindly assumes a subreg mem is
OK if its address has been partially reloaded by legitimize_reload_address.
PR rtl-optimization/72771
* reload.c (find_reloads): Don't assume that a subreg mem is OK
when find_reloads_toplev returns address_reloaded==-1.
(alternative_allows_const_pool_ref): Update comment.
testsuite/
* gcc.c-torture/compile/pr72771.c: New.
From-SVN: r239549
| gcc/ChangeLog | diffblobblamehistory | |
| gcc/reload.c | diffblobblamehistory | |
| gcc/testsuite/ChangeLog | diffblobblamehistory | |
| gcc/testsuite/gcc.c-torture/compile/pr72771.c | [new file with mode: 0644] | blob |