On Tue, Apr 24, 2012 at 12:27 PM, Maciej W. Rozycki
<macro@codesourcery.com> wrote:
> On Fri, 20 Apr 2012, Pinski, Andrew wrote:
>
>> Here is the updated patch with the style fixes and one extra change as I
>> noticed itype_op (inst) was being called a few times in mips32_next_pc,
>> I merged all of them into one variable.
>
>  Thanks for doing this, I meant to do such a change as the next step.
> However I'd prefer functionally separate changes to be made as separate
> commits, so please split this change into two, first that eliminates the
> repetitive itype_op (inst) operations, and second that adds your new
> feature.
>
>>  And renamed is_octeon_bit_op to
>> is_octeon_bbit_op since the instructions are named bbit and not bit.
>
>  Thanks, that looks reasonable to me.  Please also take into account my
> previous comment about function documentation that Joel has been kind
> enough to reiterate.
>
>  OK with these changes.

These are the three patches I committed in the end.  I also added a
testcase for testing bbit.

Thanks,
Andrew Pinski

Patch 1:
ChangeLog:
* mips-tdep.c (mips32_next_pc): Consolidate calls to itype_op.

Patch 2:
ChangeLog:
* mips-tdep.c (mips32_next_pc): Fix line spacing of the comment
before the function.

Patch 3:
ChangeLog:
* mips-tdep.c (is_octeon): New function.
(is_octeon_bbit_op): New function.
(mips32_next_pc): Handle Octeon's bbit instructions.
(mips32_instruction_has_delay_slot): Likewise.

testsuite/ChangeLog:
* gdb.arch/mips-octeon-bbit.c: New file.
* gdb.arch/mips-octeon-bbit.exp: New Test.