Messages in this thread | | | From | Nick Piggin <> | Subject | Re: [RFC PATCH] Implement slub fastpath with sequence number | Date | Wed, 12 Mar 2008 10:13:20 +1100 |
| |
On Wednesday 12 March 2008 01:45, Pekka Enberg wrote: > Hi Nick, > > On Tue, Mar 11, 2008 at 11:41 AM, Nick Piggin <nickpiggin@yahoo.com.au> wrote: > > Wow. I applaud the effort to micro optimise things ;) > > > > But I hope this doesn't get merged until macro-regressions in SLUB > > are verified to be fixed. It's pretty clear that SLUB's problem is > > not fastpath performance, so I think this would be premature > > optimisation. > > What regressions are you referring to? The SLAB_HWCACHE_ALIGN > regression patch you sent is being merged. What else?
The oracle/tpcc one I don't know if it has been fixed?
> And FWIW, I don't like the patch because it makes the code very hairy. > But I don't see why we shouldn't merge SLUB fast-path optimizations if > they're clean and you have the numbers to show it's a gain even if > there are other remaining regressions.
I'm talking about this patch specifically though. It makes it much harder to work with.
| |