[lkml]   [2011]   [Feb]   [17]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
SubjectRe: [PATCH 0/2] jump label: 2.6.38 updates
On Thu, 2011-02-17 at 11:13 -0500, Mathieu Desnoyers wrote:

> >
> > > lfoo = foo;
> >
> > IOW, will that smp_mb() really make lfoo read the new foo in memory? If
> > foo happens to still be in cache and no coherency has been performed to
> > flush it, would it just simply read foo straight from the cache?
> If we were to deploy the Linux kernel on an architecture without
> coherent caches, I think smp_mb() should imply a cacheline invalidation,
> otherwise we completely mess up the order of data writes vs their
> observability from each invididual core POV.

Um but this thread is not about non-coherent caches. It's about a HW
that happens to do something stupid with ll/sc. That is, everything
deals with the cache except ll/sc which skips it.

Although, this was more or less answered in another email. That is, the
cache on this HW is not really coherent but all the CPUs just seem to
share the same cache. Thus a invalidate of the cache line affects all
CPUs which makes my question moot.

-- Steve

 \ /
  Last update: 2011-02-17 21:11    [W:0.084 / U:14.880 seconds]
©2003-2018 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site