Re: [PATCH V2] Powerpc: Update expected floating point output for gdb.arch/altivec-regs.exp and gdb.arch/vsx-regs.exp

[Ulrich Weigand <Ulrich.Weigand@de.ibm.com>]

Carl Love <cel@us.ibm.com> wrote:

>PowerPC: Update expected floating point output for gdb.arch/altivec-
>regs.exp and gdb.arch/vsx-regs.exp

This patch is OK.

As additional comment, I'm wondering whether the display of vector
registers via the union we're using is actually useful when in
hexadecimal mode.  The intent of the union is that it should be easy to
interpret the register contents in various ways, in particular both as
integer and as floating-point value, as we may not know what type is
currently being held in the register.

However, with the new way of printing floating-point values when /x is
in effect, it seems that the contents of v4_float will now always be
identical to v4_int32, and similar for the other floating-point union
members.

So maybe it would be better to always show the floating-point members
as actual floating-point, even when the rest of the union is printed
via /x.  However, that may be a more invasive change, so I think you
should go ahead and commit the patch as-is now, to improve the
testsuite results.

Bye,
Ulrich