[Bug testsuite/68886] FAIL: gcc.c-torture/execute/stkalign.c execution test
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=68886 Alan Modra changed: What|Removed |Added Status|NEW |RESOLVED Resolution|--- |FIXED --- Comment #5 from Alan Modra --- Fixed with https://gcc.gnu.org/ml/gcc-cvs/2016-02/msg00385.html
[Bug testsuite/68886] FAIL: gcc.c-torture/execute/stkalign.c execution test
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=68886 Alan Modra changed: What|Removed |Added CC||amodra at gmail dot com --- Comment #4 from Alan Modra --- Created attachment 37626 --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=37626=edit fix the testcase From the git commit log Function (or more narrow) scope static variables (as well as others not placed on the stack) should also not have any effect on the stack alignment. but that is not exactly what the testcase is testing.. So, make it test just that, by using two functions, identical except for the non-stack var alignment.
[Bug testsuite/68886] FAIL: gcc.c-torture/execute/stkalign.c execution test
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=68886 Eric Botcazou changed: What|Removed |Added Target|hppa-*-*, powerpc*-*-* |hppa-*-*, powerpc*-*-*, ||sparc64-*-* CC||ebotcazou at gcc dot gnu.org --- Comment #3 from Eric Botcazou --- Likewise on SPARC 64-bit. And there isn't a single comment in the test...
[Bug testsuite/68886] FAIL: gcc.c-torture/execute/stkalign.c execution test
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=68886 Andrew Pinski changed: What|Removed |Added Status|UNCONFIRMED |NEW Last reconfirmed||2016-01-06 Summary|[6 regression] FAIL:|FAIL: |gcc.c-torture/execute/stkal |gcc.c-torture/execute/stkal |ign.c execution test|ign.c execution test Ever confirmed|0 |1 --- Comment #2 from Andrew Pinski --- If the stack size is just happens to be a multiple of 64 and the incoming stack to test was 64 byte aligned, then the testcase will always fail. This is what is happening on PowerPC and HPPA. It is just on accident it passes. The testcase should either be removed or changed. Changing the alignment to 128 or 256 might allow more targets to pass but not always pass on every target so removing it will be best. This is a new test and so it can't be a regression.