On 8/30/23 04:13, Manolis Tsamis wrote:
The existing implementation of need_cmov_or_rewire and
noce_convert_multiple_sets_1 assumes that sets are either REG or SUBREG.
This commit enchances them so they can handle/rewire arbitrary set statements.
To do that a new helper struct noce_multiple_sets_info is introduced which is
used by noce_convert_multiple_sets and its helper functions. This results in
cleaner function signatures, improved efficientcy (a number of vecs and hash
set/map are replaced with a single vec of struct) and simplicity.
gcc/ChangeLog:
* ifcvt.cc (need_cmov_or_rewire): Renamed init_noce_multiple_sets_info.
(init_noce_multiple_sets_info): Initialize noce_multiple_sets_info.
(noce_convert_multiple_sets_1): Use noce_multiple_sets_info and handle
rewiring of multiple registers.
(noce_convert_multiple_sets): Updated to use noce_multiple_sets_info.
* ifcvt.h (struct noce_multiple_sets_info): Introduce new struct
noce_multiple_sets_info to store info for noce_convert_multiple_sets.
gcc/testsuite/ChangeLog:
* gcc.target/aarch64/ifcvt_multiple_sets_rewire.c: New test.
So this seems like (in theory) it could move forward independently. The
handling of arbitrary statements code wouldn't be exercised yet, but
that's OK IMHO as I don't think anyone is fundamentally against trying
to handle additional kinds of statements.
So my suggestion would be to bootstrap & regression test this
independently. AFAICT this should have no functional change if it were
to go in on its own. Note the testsuite entry might not be applicable
if this were to go in on its own and would need to roll into another
patch in the series.
Jeff
