Re: [PATCH] PowerPC, fix gdb.base/retval-large-struct.exp

[Carl Love <cel@us.ibm.com>]

Ulrich:

On Thu, 2022-11-17 at 12:57 +0000, Ulrich Weigand wrote:
> > Currently the test gdb.base/retval-large-struct.exp is not able to
> > access the return buffer address for the data being returned by the
> > function.  The functionality needed to obtain the value of r3 which
> > holds the return buffer address on entry to the function was
> > recently
> > added in a new gdb method.  This new method can be used by this
> > test to
> > allow it to correctly access the data being returned by the
> > function.
> 
> Are there any more test cases that should get -fvar-tracking now?
> 
> In the long run, it might be good to work with GCC developers to see
> if we can't have entry value DWARF records emitted (at least for
> simple cases like the return buffer address) even without variable
> tracking ...

At the moment, this is the only one I know of.  The number of test
failures on Power is getting relatively short. There are about 15ish
tests that are currently failing on Power 10.  I have the list and have
been adding notes for each test as to what the issue is or is fixed by
a patch under development.  I have a couple of patches that have been
posted and approved and a couple that I am working on still.  Let me
refresh the list and my notes and I can send it out so you can see
where we are at with the regression test clean up.

                           Carl