[PATCH] x86, sse5: fix protX with immediate operand

H.J. Lu hjl.tools@gmail.com
Wed Apr 15 12:48:00 GMT 2009


On Wed, Apr 15, 2009 at 3:16 AM, Jan Beulich <jbeulich@novell.com> wrote:
> Fix opcode generation for SSE5's protX opcodes, which had their opcode
> extension fields set to zero instead of None.
>
> While for pshaX the extension field also was incorrectly set to zero,
> other code overwrote the bad value later. For consistency, this is also
> being fixed here.
>
> I'd suspect that for all the pmacsXX this should also be corrected, and

See

http://sourceware.org/bugzilla/show_bug.cgi?id=6873

Can you also fix it?

> possibly fixing this in the table would allow removing the need to
> later overwrite these values.
>
> gas/testsuite/
> 2009-04-15  Jan Beulich  <jbeulich@novell.com>
>
>        * gas/i386/x86-64-sse5.s: Add test of protd.
>        * gas/i386/x86-64-sse5.d: Adjust expectations to match input.
>
> opcodes/
> 2009-04-15  Jan Beulich  <jbeulich@novell.com>
>
>        * i386-opc.tbl (protb, protw, protd, protq): Set opcode
>        extension to None.
>        (pshab, pshaw, pshad, pshaq): Likewise.
>        * i386-tbl.h: Re-generate.
>

Thanks.


-- 
H.J.



More information about the Binutils mailing list