This is the mail archive of the
mailing list for the GCC project.
r242168 - in /branches/ARM/sve-branch/gcc: comb...
- From: rsandifo at gcc dot gnu dot org
- To: gcc-cvs at gcc dot gnu dot org
- Date: Fri, 11 Nov 2016 17:12:04 -0000
- Subject: r242168 - in /branches/ARM/sve-branch/gcc: comb...
Date: Fri Nov 11 17:12:03 2016
New Revision: 242168
Make more use of paradoxical_subreg_p
This patch makes more use of the existing paradoxical_subreg_p
predicate and also adds a version that operates on outer and
Some of the affected tests were based on GET_MODE_SIZE rather than
GET_MODE_PRECISION and so could change the result for modes that
have the same size but different precisions. In each case the
change should be a no-op or an improvement. (Specifically, the
big-endian byte offset correction will be 0 for paradoxical subregs
whose byte sizes are the same.)
The patch changes the branch taken for full subregs in
simplify_subreg, but the new form matches the commentary more
closely. Both branches should be equally good in that situation.
The patch also removes a redundant "GET_CODE (x) == SUBREG"
check from move_plus_up. By this point we've already checked
that it's a SUBREG and extracted the SUBREG_REG.