Re: Clang and X86-EFlags (was Re: [PATCH] usbhid: Fix lockdep unannotated irqs-off warning)

[Sedat Dilek]

On Tue, Apr 24, 2018 at 5:22 PM, Linus Torvalds
<torvalds@xxxxxxxxxxxxxxxxxxxx> wrote:
> On Tue, Apr 24, 2018 at 6:28 AM, Sedat Dilek <sedat.dilek@xxxxxxxxx> wrote:
>>
>> $ objdump -S clang-eflag.o
>>
>> Does this now look good?
>
> Looks fine to me. The instruction choice is still pretty odd:
>
>   1a:   f6 c3 ff                test   $0xff,%bl
>   1d:   74 02                   je     21 <bar+0x21>
>
> that "test   $0xff,%bl" is a rather odd way of testing the byte for zero, since
>
>   1a:   84 db                test   %bl,%bl
>   1c:   74 02                   je     20 <bar+0x20>
>
> would have been a byte shorter and is the canonical way on x86 to test
> a register against zero.
>
> But that's just a "looks odd to somebody who is used to x86 asm", not
> a bug or anything really noticeable.
>
>                   Linus

Thanks for the quick reply and the confirmation.

Personally, I do not speak x86/asm.
I hope CCed LLVM folks know on how to improve things.

- Sedat -