This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

fix target/18932


This pr is indeed due to my patch for 17025.

The patch for 17025 strengthened the tests for which CCmodes are
valid for particular sets of inputs.  This exposed a latent bug
where the input CCmode was not preserved across a number of splits
and peepholes.

I would have noticed this in my original patch had I bootstrapped
with i586 instead of i686.

Fixed by using match_operator and match_operand to capture the
mode from the input, and match_dup and match_op_dup to copy it to
the output.

Tested with {i386,i585,i686}-linux.


r~

Attachment: z-33.gz
Description: GNU Zip compressed data

Attachment: z-34.gz
Description: GNU Zip compressed data

Attachment: z-40.gz
Description: GNU Zip compressed data


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]