This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: egcs-19980508 powerpc-ibm-aix4.1.4.0 results (failures with -p/-pg)
David Edelsohn writes:
| >>>>> "Kaveh R Ghazi" writes:
|
| Kaveh> NOTE: There were several extra c-torture execute failures when using -p/-pg.
|
| Note: all of these extra failures involve testcases that use the
| GCC nested function extension. I don't know the details, but these cases
| cause the AIX profiler to perform some form of illegal operation -- maybe
| it gets confused about the stack or encounters an unexpected address in
| link register from a nested call.
|
| Whatever, this really isn't a GCC bug and isn't an AIX bug because
| AIX cannot be expected to support GCC extensions in this type of instance.
| If you want to keep reporting these, fine; but understand that they never
| will be fixed. I would prefer if they could be listed as expected
| failures when -p or -pg is used so that real regressions are not obscured
| by the many known failure modes.
Well GCC probably need to save r11 across profile calls (or switch
profiling before/after the prologue).