On Wed, Oct 29, 2025 at 4:38 PM H.J. Lu <[email protected]> wrote:
>
> On Tue, Oct 28, 2025 at 3:55 PM H.J. Lu <[email protected]> wrote:
> >
> > On Tue, Oct 28, 2025 at 12:07 PM <[email protected]> wrote:
> > >
> > > Dear contributor,
> > >
> > > Our automatic CI has detected problems related to your patch(es). Please 
> > > find some details below.
> > >
> > > In gcc_check master-aarch64, after:
> > >   | commit gcc-16-4667-gdcf69bdcd49
> > >   | Author: H.J. Lu <[email protected]>
> > >   | Date:   Sun Oct 26 08:42:20 2025 +0800
> > >   |
> > >   |     c: Try the type with the previous function attributes
> > >   |
> > >   |     When there are 2 conflicting function declarations, try the new 
> > > type
> > >   |     with the previous TYPE_ATTRIBUTES if the current declaration has 
> > > no
> > >   |     TYPE_ATTRIBUTES to support
> > >   | ... 44 lines of the commit log omitted.
> > >
> > > Produces 19 regressions:
> > >   |
> > >   | regressions.sum:
> > >   | Running gcc:gcc.target/aarch64/sme/aarch64-sme.exp ...
> > >   | FAIL: gcc.target/aarch64/sme/streaming_mode_1.c (test for errors, 
> > > line 10)
> > >   | FAIL: gcc.target/aarch64/sme/streaming_mode_1.c (test for errors, 
> > > line 121)
> > >   | FAIL: gcc.target/aarch64/sme/streaming_mode_1.c (test for errors, 
> > > line 16)
> > >   | FAIL: gcc.target/aarch64/sme/streaming_mode_1.c (test for errors, 
> > > line 30)
> > >   | ... and 15 more
> > >
> >
> > After
> >
> > commit dcf69bdcd49bccd901bfb01db7c15530e9a70dc0
> > Author: H.J. Lu <[email protected]>
> > Date:   Sun Oct 26 08:42:20 2025 +0800
> >
> >     c: Try the type with the previous function attributes
> >
> > gcc no longer issues an error for:
> >
> > void sc_c () [[arm::streaming_compatible]];
> > void sc_c () {}
> >
> > Instead, the previous type attributes are applied to the current function
> > definition.  The resulting function definition is compatible with the
> > previous declaration.
> >
> > PR c/122427
> > * gcc.target/aarch64/sme/streaming_mode_1.c: Remove dg-error.
> >
>
> Here is the v2 patch with the gcc.target/aarch64/sme/za_state_1.c change.

This patch is wrong. The testcases should/need be rejected by both the
C and C++ front-ends for these attributes on the function types.


Thanks,
Andrew Pinski


>
>
> --
> H.J.

Reply via email to