On Wed, Feb 27, 2019 at 11:41:31AM -0800, Linus Torvalds wrote: > On Wed, Feb 27, 2019 at 9:57 AM Nadav Amit <na...@vmware.com> wrote: > > > > I’ll have a look at some specific function assembly, but overall, the “+m” > > approach might prevent even more code optimizations than the “volatile” one. > > Ok, that being the case, let's forget that patch. > > I still wonder about the added volatiles to the xadd/cmpxchg cases, > which already had the "memory" clobber which should make the volatile > immaterial..
That was mostly me being OCD style consistent; but also notice that the atomic ops also often have volatile even though they have a memory clobber.