On Mon, Nov 5, 2012 at 10:56 AM, Richard Sandiford
<rdsandif...@googlemail.com> wrote:
> Uros Bizjak <ubiz...@gmail.com> writes:
>> On Mon, Nov 5, 2012 at 9:30 AM, Richard Sandiford
>> <rdsandif...@googlemail.com> wrote:
>>> gcc/
>>> PR target/55204
>>> * config/i386/i386.c (ix86_address_subreg_operand): Remove stack
>>> pointer check.
>>> (print_reg): Use true_regnum rather than REGNO.
>>> (ix86_print_operand_address): Remove SUBREG handling.
>>
>> The patch is OK for mainline and 4.7, if it passes H.J.'s tests with
>> -maddress-mode={short,long} on x32.
>>
>>> + unsigned int regno = true_regnum (x);
>>
>> I'd rather see the declaration at the beginning of the function.
>
> OK, thanks, applied to both with that change.
>
> Richard
On 4.7 branch, this failed to build x32 run-time library:
/tmp/ccID3vr6.s: Assembler messages:
/tmp/ccID3vr6.s:2788: Error: bad register name `%rr15'
There is an extra 'r' in register name.
--
H.J.
