diff options
author | Richard Sandiford <richard.sandiford@linaro.org> | 2018-01-09 14:30:27 +0000 |
---|---|---|
committer | Richard Sandiford <rsandifo@gcc.gnu.org> | 2018-01-09 14:30:27 +0000 |
commit | 7ad429a4de1197b4b032da0981a2c91ac5d0dd06 (patch) | |
tree | f961b0ea1d30ee5ba25759c42016f2ef9c9ed9dd /gcc/fortran/expr.c | |
parent | 82c066f587fe99cc3d5cd0d685582f0272d3a40f (diff) | |
download | gcc-7ad429a4de1197b4b032da0981a2c91ac5d0dd06.zip gcc-7ad429a4de1197b4b032da0981a2c91ac5d0dd06.tar.gz gcc-7ad429a4de1197b4b032da0981a2c91ac5d0dd06.tar.bz2 |
Fix permute handling when vectorising scatters
As mentioned in https://gcc.gnu.org/ml/gcc-patches/2017-11/msg01575.html ,
the scatter handling in vectorizable_store seems to be dead code at the
moment. Enabling it with the vect_analyze_data_ref_access part of
that patch triggered an ICE in the avx512f-scatter-*.c tests (which
previously didn't use scatters). The problem was that the NARROW
and WIDEN handling uses permute_vec_elements to marshal the inputs,
and permute_vec_elements expected the lhs of the stmt to be an SSA_NAME,
which of course it isn't for stores.
This patch makes permute_vec_elements create a fresh variable in this case.
2018-01-09 Richard Sandiford <richard.sandiford@linaro.org>
gcc/
* tree-vect-stmts.c (permute_vec_elements): Create a fresh variable
if the destination isn't an SSA_NAME.
From-SVN: r256383
Diffstat (limited to 'gcc/fortran/expr.c')
0 files changed, 0 insertions, 0 deletions