Re: Compiler optimizing variables in inline assembly

[Cody Rigney <codyrigney92@gmail.com>]

I added all the registers and "memory" to clobber and it worked!
Since I think it's more efficient to specify which memory is changing,
I'm going to give the "=m" (*a) a shot.  So even though you
dereference the pointer, gcc will know the length of the memory(say 2
bytes) that changes?  Or do you have to specify each index of the
memory(e.g. char a[2]; ..... asm ... "=m" (*a), "=m" (*(a+1)))?

On Fri, Feb 21, 2014 at 4:53 AM, Andrew Haley <aph@redhat.com> wrote:
> On 02/20/2014 07:29 PM, Cody Rigney wrote:
>> On Thu, Feb 20, 2014 at 4:14 AM, Andrew Haley <aph@redhat.com> wrote:
>>> Hi,
>>>
>>> On 02/19/2014 07:04 PM, Cody Rigney wrote:
>>>> I'm trying to add NEON optimizations to OpenCV's LK optical flow.  See
>>>> link below.
>>>> https://github.com/Itseez/opencv/blob/2.4/modules/video/src/lkpyramid.cpp
>>>>
>>>> The gcc version could vary since this is an open source project, but
>>>> the one I'm currently using is 4.8.1. The target architecture is ARMv7
>>>> w/ NEON. The processor I'm testing on is an ARM
>>>> Cortex-A15(big.LITTLE).
>>>>
>>>> The problem is, in release mode (where optimizations are set) it does
>>>> not work properly. However, in debug mode, it works fine. I tracked
>>>> down a specific variable(FLT_SCALE) that was being optimized out and
>>>> made it volatile and that part worked fine after that. However, I'm
>>>> still having incorrect behavior from some other optimization.
>>>
>>> Forget about using volatile here.  That's just wrong.
>>>
>>> You have to mark your inputs, outputs, and clobbers correctly.
>>>
>> That makes sense.  In this case, the input parameters are actually
>> memory addresses.  So how would I do an output or clobber that would
>> tell the compiler that the memory at those addresses will change?
>
> You can use a memory operand as an output, as in "=m"(*a) or simply
> add "memory" to the clobber list.  And you must add all clobbered
> registers to the clobber list.  Then it should work.
>
> Andrew.
>
>