[Bug target/119765] FAIL: gcc.target/i386/amd64-abi-9.c on Windows

2025-04-13 Thread nightstrike at gmail dot com via Gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=119765

--- Comment #6 from nightstrike  ---
Patch posted:
https://gcc.gnu.org/pipermail/gcc-patches/2025-April/680856.html

[Bug target/119765] FAIL: gcc.target/i386/amd64-abi-9.c on Windows

2025-04-13 Thread peter0x44 at disroot dot org via Gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=119765

Peter Damianov  changed:

   What|Removed |Added

 CC||peter0x44 at disroot dot org

--- Comment #5 from Peter Damianov  ---
putting __attribute__((__sysv_abi__)) on foo should make this test pass even on
x86_64-w64-mingw32.

https://gcc.godbolt.org/z/dr8ae4r6c
I couldn't run the testsuite, but just purely judging from the assembly here it
looks okay.

As a basic fix that would work, but extending the test to cover msabi (assuming
there aren't any existing ones) would also be a good idea.

[Bug target/119765] FAIL: gcc.target/i386/amd64-abi-9.c on Windows

2025-04-13 Thread nightstrike at gmail dot com via Gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=119765

--- Comment #4 from nightstrike  ---
Ah, I think I understand better. It's not the AMD64 ABI saying to pass the
register count in EAX, it's the SysV or whatever ABI saying to do it, and the
Windows ABI says to do something else. So really instead of skipping the test,
we can just test that varargs are handled right on Windows.

[Bug target/119765] FAIL: gcc.target/i386/amd64-abi-9.c on Windows

2025-04-13 Thread nightstrike at gmail dot com via Gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=119765

--- Comment #3 from nightstrike  ---
Is this a case of "skip windows", or a case of "skip targets with/without
feature X"?

Also, I'm a little surprised, as based on the name of the test, I would think
it was testing a feature of the cpu type, not the operating system.

[Bug target/119765] FAIL: gcc.target/i386/amd64-abi-9.c on Windows

2025-04-13 Thread pinskia at gcc dot gnu.org via Gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=119765

--- Comment #2 from Andrew Pinski  ---
Yes the testcase should be skipped for the windows x86_64 which never sets eax
for varargs.

[Bug target/119765] FAIL: gcc.target/i386/amd64-abi-9.c on Windows

2025-04-12 Thread pinskia at gcc dot gnu.org via Gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=119765

--- Comment #1 from Andrew Pinski  ---
I suspect it is disable the test for windows abi. It is testing varargs setting
of eax for the number of use registers passed.