Move lowering of vector shifts from v/s to v/v from gimple to rtl.
commit83a28c112ebae02d188c8f18778dec8e8c5cdbe3
authorrth <rth@138bc75d-0d04-0410-961f-82ee72b054a4>
Fri, 14 Oct 2011 03:25:58 +0000 (14 03:25 +0000)
committerrth <rth@138bc75d-0d04-0410-961f-82ee72b054a4>
Fri, 14 Oct 2011 03:25:58 +0000 (14 03:25 +0000)
tree59dd3e79664e96b03e31703448c40d89aff4dd85
parent6ae3cabe2278e4d0cbccd9feb872f92efe763ea2
Move lowering of vector shifts from v/s to v/v from gimple to rtl.

This allows other rtl expanders to rely on shifts of vector by scalar.

git-svn-id: svn+ssh://gcc.gnu.org/svn/gcc/trunk@179956 138bc75d-0d04-0410-961f-82ee72b054a4
gcc/ChangeLog
gcc/optabs.c
gcc/testsuite/ChangeLog
gcc/testsuite/gcc.dg/vect/vec-scal-opt.c
gcc/testsuite/gcc.dg/vect/vec-scal-opt1.c
gcc/testsuite/gcc.dg/vect/vec-scal-opt2.c
gcc/testsuite/lib/target-supports.exp
gcc/tree-vect-generic.c