[PATCH v2] nptl: optimize cancelstate and canceltype changing

[Richard Henderson]

On 06/15/2014 11:56 PM, Alexander Fyodorov wrote:
> If CAS is atomic, then "a" will equal 2. But if CAS is implemented via spinlock, the write from CPU0 could happen in between the load and the store that comprise CAS, and "a" will equal 1 which is wrong.

Ah, I mis-read the original patch, wherein I thought you were declaring a local
lock for the write.  Sorry for the noise.


r~