Messages in this thread | | | Date | Wed, 15 Oct 2008 11:29:32 -0700 (PDT) | From | Linus Torvalds <> | Subject | Re: [rfc] SLOB memory ordering issue |
| |
On Thu, 16 Oct 2008, Nick Piggin wrote: > > When I said "I'd really hate to add a branch to the slab fastpath", it > wasn't a tacit acknowlegement that the barrier is the only way to go, > if it sounded that way. > > I meant: I'd *really* hate to add a branch to the slab fastpath :)
Well, quite frankly, your choice of subject line and whole point of argument may have confused me.
You started out - and continue to - make this sound like it's a SLAB/SLOB/SLUB issue. It's not.
I agree there is quite likely memory ordering issues - possibly old ones, but quite possibly also ones that have just happened fairly recently as we've done more unlocked lookups - and all I've ever disagreed with is how you seem to have mixed this up with the allocator.
And I still don't understand why you even _mention_ the slab fastpath. It seems totally immaterial.
Linus
| |