Re: 2.6.16-rc6-mm2

[Keith Owens <kaos@sgi.com>]

Andi Kleen (on 12 Jun 2006 14:14:09 +0200) wrote:
>Ingo Molnar <mingo@elte.hu> writes:
>
>> * Ingo Molnar <mingo@elte.hu> wrote:
>> 
>> > 
>> > * Christoph Lameter <clameter@sgi.com> wrote:
>> > 
>> > > Sorry that patch was still against mm1. Here is a fixed up version 
>> > > that applies cleanly against mm2:
>> > 
>> > i have applied both patches you sent in this thread but it still 
>> > triggers tons of messages:
>> 
>> > trying to fix it i realized that i'd have to touch tons of 
>> > architectures, which all duplicate this piece of code:
>> 
>> below is an updated patch that includes fixups for i386 - but the real 
>> fix should be to properly reduce the per-arch local.h footprint to the 
>> bare minimum possible, and to do this fix on the asm-generic headers.
>
>I think an even better approach would be to use local_save_flags() / 
>local_restore_flags () and then use a normal increment and get
>rid of smp_processor_id completely.
>
>I've never seen any evidence that the complex and bloated code generated by 
>this is any better that just enabling/disabling interrupts. 
>
>In the x86 world P4 has costly cli/sti, but I wouldn't expect
>that problem to be very widespread.

I have previously suggested a lightweight solution that pins a process
to a cpu for the short duration that it accesses cpu local variables.
http://lkml.org/lkml/2005/12/16/2.  Time to revistit that option?