https://gcc.gnu.org/bugzilla/show_bug.cgi?id=96756
rsandifo at gcc dot gnu.org <rsandifo at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Last reconfirmed| |2020-08-28
Ever confirmed|0 |1
Severity|normal |enhancement
Resolution|INVALID |---
Status|RESOLVED |NEW
CC| |rsandifo at gcc dot gnu.org
--- Comment #2 from rsandifo at gcc dot gnu.org <rsandifo at gcc dot gnu.org>
---
I think it's unclear to what extent this should work with current GCC,
but I think the PR is a valid enhancement request. The ACLE certainly
defines routines for reading the Q flag, and while GCC doesn't yet
implement them, the facility should presumably be made available
at some point.
Note that the test is derived from ones in the GCC testsuite and is
the kind of thing that we already test. So this is not a newly-
invented requirement. It's more that we fluked the test in some
cases due to other flaws that Huawei are fixing.
Also note that the use of volatile was copied from arm-neon-ref.h
in the GCC testsuite. I agree it's redundant, but the volatile
actually qualifies the return type rather than the function type
(and so gets a warning via -Wignored-qualifiers). The noreturn
syntax is instead __attribute__((volatile)). Either way though,
it's not the reporter's mistake.
