On 04/05/2013 02:04 PM, Bjorn Lindgren wrote: > On Fri, 5 Apr 2013, Andrew Haley wrote: > >>> Well, that was what I wanted to find out, do I need error handling if the >>> event of a colliction (two cores exectuing CMPXCHG within the same clock >>> cycles), the instruction will be automaticly retried until it succeeds? >> >> Correct. >> >>> Is that true for the x86 asm implementations of __atomic_*_fetch / >>> __atomic_fetch_* calls too, that they are retried? >> >> Which x86 asm implementations do you mean? We're talking about >> GCC here, and it has builtins. > > I was refering to the internal GCC implementation of the builtin > functions: > > __atomic_add_fetch > __atomic_sub_fetch > __atomic_and_fetch > __atomic_xor_fetch > etc.. > > I'd assume there is a separate assembler implementaion for each > processor architecture (x86, SPARC, MIPS, ARM, etc). Sometimes there isn't: they're usually builtins. And they all block until they succeed. Andrew.
