On 5/18/21 11:30 AM, Segher Boessenkool wrote:
On Tue, May 18, 2021 at 09:09:07AM -0500, Bill Schmidt wrote:
Long ago we were forced to make some small ABI breaks to correct errors
in the implementation, and we added warning messages for the changes
from GCC 4.9 to GCC 5. Enough time has passed that these are now just
irritants, so let's remove them. Also clean up associated macros using
rs6000_special_adjust_field_align_p, which has been always returning
false for a long time.
One tiny thing:
--- a/gcc/config/rs6000/sysv4.h
+++ b/gcc/config/rs6000/sysv4.h
@@ -325,8 +325,7 @@ do {
\
/* An expression for the alignment of a structure field FIELD if the
alignment computed in the usual way is COMPUTED. */
#define ADJUST_FIELD_ALIGN(FIELD, TYPE, COMPUTED) \
- (rs6000_special_adjust_field_align_p ((TYPE), (COMPUTED)) \
- ? 128 : COMPUTED)
+ (COMPUTED)
You should be able to just not define the macro at all, but this may be
clearer. Some of the "which of these headers is used for what" is not
clear to me even, so :-)
Okay for trunk. Thanks!
Will you also remove the -Wno-psabi options in testcases that do not
need it now? There are more than 200 that use it, but some may be for
other ABI changes / implementation fixes. This isn't high priority at
all of course, we have much bigger anachronisms in the testsuite. But
someone who is bored could just delete them all and see what breaks,
perhaps :-)
Thanks for the quick review! I'll open an internal issue so that
somebody can look at this in the future. We need to do careful testing
on AIX and probably Darwin before removing any that are non-obvious...
Bill
Segher
