This is the mail archive of the
mailing list for the GCC project.
Re: [patch RFC] SH: Use FRAME_GROWS_DOWNWARD
Joern RENNECKE <email@example.com> wrote:
> I'm currently working on a merge from mainline (yesterday 12:00 UTC)
> into the sh-elf-4_1-branch. I also see a failure of stdarg-2.c for
> sh-elf/-m4/-ml at "-O3 -fomit-frame-pointer" and "-O3 -g"
> (-m4 -mb not tested).
> f12 returns an incorrect result, but not when you re-run the program.
> I suspect that the va_list has not been set up before it is used.
> (foo is inlined into f12).
> There is also a failure of execute/20000603-1.c at -O2. Do you see
> that too? f is miscompiled: in sched1, the DFmode store is moved after
> the load.
Yes. I see
FAIL: gcc.c-torture/execute/20000603-1.c execution, -O2
for sh-elf/-m4 and sh-lf/-m4/-ml. For sh4-unknown-linux-gnu,
those 3 tests
gcc.c-torture/execute/20000603-1.c execution, -O2
gcc.c-torture/execute/stdarg-2.c execution, -O3 -fomit-frame-pointer
gcc.c-torture/execute/stdarg-2.c execution, -O3 -g
started to fail on mainline Jul 16.