[lkml]   [2008]   [Nov]   [6]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
SubjectRe: [patch 5/7] x86_64: Support for cpu ops
Ingo Molnar wrote:
> * Christoph Lameter <> wrote:
>> On Thu, 6 Nov 2008, Ingo Molnar wrote:
>>> hm, what happened to the rebase-PDA-to-percpu-area optimization
>>> patches you guys were working on? I remember there was some
>>> binutils flakiness - weird crashes and things like that. Did you
>>> ever manage to stabilize it? It would be sad if only 32-bit could
>>> take advantage of the optimized ops.
>> I thought that was in your tree? I saw a conflict in -next with the
>> zero based stuff a couple of weeks ago. Mike is working on that
> No, what's in tip/core/percpu is not the PDA patches:
> f8d90d9: percpu: zero based percpu build error on s390
> cfcfdff: Merge branch 'linus' into core/percpu
> d379497: Zero based percpu: infrastructure to rebase the per cpu area to zero
> b3a0cb4: x86: extend percpu ops to 64 bit
> But it's not actually utilized on x86. AFAICS you guys never came back
> with working patches for that (tip/x86/percpu is empty currently), and
> now i see something related on lkml on a separate track not Cc:-ed to
> the x86 folks so i thought i'd ask whether more coordination is
> desired here.
> So ... what's the merge plan here? I like your fundamental idea, it's
> a nice improvement in a couple of areas and i'd like to help out make
> it happen. Also, the new per-cpu allocator would be nice for the
> sparseirq code.
> Ingo

Sorry, this was on my plate but the 4096 cpus if far more critical to get
released and available. As soon as that's finally done, I can get back to
the pda/zero-based changes.


 \ /
  Last update: 2008-11-06 16:45    [W:0.067 / U:2.412 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site