On 10 March 2016 at 19:52, Yvan Roux <yvan.r...@linaro.org> wrote:
> On 10 March 2016 at 19:42, Edward Nevill <edward.nev...@linaro.org> wrote:
>>> /usr/local/linare-gcc-5.2/bin/gcc -S -O3 -march=armv8-a+lse test.c
>>>
>>> add_int:
>>> ldaddal w0, w0, [x1]
>>> add w2, w0, w0
>>> mov w0, w2
>>> ret
>>
>> Am I going mad, or does this just return the contents of the memory
>> location * 2.
>>
>> ldaddal w0, w0, [x1]
>>
>> Returns the original contents of [x1] in w0.
>>
>> add w2, w0, w0
>>
>> doubles it.
>>
>> mov w0, w2
>>
>> returns it.
>>
>> I think __sync_add_and_fetch should return the updated contents.
>>
>> Here is the test C code again.
>>
>> int add_int(int add_value, int *dest)
>> {
>> return __sync_add_and_fetch(dest, add_value);
>> }
>
> Hmm, it is not the code I get with our latest release candidate (the
> release should be out next week), it gives the code:
>
> add_int:
> ldaddal w0, w3, [x1]
> add w2, w3, w0
> mov w0, w2
> ret
Ed, just for info it was fixed by Andrew in trunk and backported in
our branch end of December.
https://gcc.gnu.org/ml/gcc-patches/2015-12/msg01962.html
>> Regards,
>> Ed.
_______________________________________________
linaro-toolchain mailing list
linaro-toolchain@lists.linaro.org
https://lists.linaro.org/mailman/listinfo/linaro-toolchain
