This is the mail archive of the gcc@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]

sh le fp arg passing


So, I have a question...  Why do this:

Excerpt from cvs log gcc/config/sh/sh.h:

revision 1.23
date: 1998/11/23 08:50:40;  author: amylaar;  state: Exp;  lines: +314 -73
        Add SH4 support:
...
        (FUNCTION_ARG): Little endian adjustment for SH4 SFmode.

Excerpt from cvs diff -u -r1.22 -r1.23 sh.h:

@@ -883,7 +1039,9 @@
   ((PASS_IN_REG_P ((CUM), (MODE), (TYPE))                              \
     && ((NAMED) || TARGET_SH3E || ! current_function_varargs))         \
    ? gen_rtx (REG, (MODE),                                             \
-             (BASE_ARG_REG (MODE) + ROUND_REG ((CUM), (MODE)))) \
+             ((BASE_ARG_REG (MODE) + ROUND_REG ((CUM), (MODE)))        \
+              ^ ((MODE) == SFmode && TARGET_SH4                        \
+                 && TARGET_LITTLE_ENDIAN != 0)))                       \
    : 0)

is it compelling?  Do we really want to do this?  Why?  Ignorant minds
want to know!


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