[PATCH 0 of 9] x86/smp function calls: convert x86 tlb flushes to use function calls [POST 2]

!MAILaRCHIVE_VOTE_RePLACE
Previous message: [thread] [date] [author]
Next message: [thread] [date] [author]
To: Ingo Molnar <mingo@...>
Cc: LKML <linux-kernel@...>, <x86@...>, Andi Kleen <andi@...>, Nick Piggin <nickpiggin@...>, Jens Axboe <jens.axboe@...>
Date: Monday, August 18, 2008 - 2:23 pm

This series:
 - adds a simple debugfs profiling entry for cross-cpu tlb flushes
 - converts them to using smp_call_function_mask
 - unifies 32 and 64-bit tlb flushes
 - converts smp_call_function to using multiple queues (using the now
   freed vectors)
 - allows config-time adjustment of the number of queues
 - adds a kernel parameter to disable multi-queue in case it causes
   problems

The main concern is whether using smp_call_function adds an
unacceptible performance hit to cross-cpu tlb flushes.  My limited
measurements show a ~35% regression in latency for a particular flush;
it would be interesting to try this on a wider range of hardware.  I
gather the effect tlb flush performance is very application specific
as well, but I'm not sure what benchmarks show what effects.

Trading off agains the latency of a given flush, the smp_function_call
mechanism allows multiple requests to be queued, and so may improve
throughput on a system-wide basis.

So, I'd like people to try this out and see what performance effects it
has.

Thanks,
	J


--
Previous message: [thread] [date] [author]
Next message: [thread] [date] [author]

Messages in current thread:
[PATCH 0 of 9] x86/smp function calls: convert x86 tlb flush..., Jeremy Fitzhardinge, (Mon Aug 18, 2:23 pm)
Re: [PATCH 0 of 9] x86/smp function calls: convert x86 tlb f..., Jeremy Fitzhardinge, (Tue Aug 19, 3:44 am)
Re: [PATCH 0 of 9] x86/smp function calls: convert x86 tlb f..., Jeremy Fitzhardinge, (Tue Aug 19, 4:04 am)
Re: [PATCH 0 of 9] x86/smp function calls: convert x86 tlb f..., Jeremy Fitzhardinge, (Tue Aug 19, 1:37 am)
Re: [PATCH 0 of 9] x86/smp function calls: convert x86 tlb f..., Jeremy Fitzhardinge, (Tue Aug 19, 2:18 am)
Re: [PATCH 0 of 9] x86/smp function calls: convert x86 tlb f..., Jeremy Fitzhardinge, (Tue Aug 19, 10:58 am)
Re: [PATCH 0 of 9] x86/smp function calls: convert x86 tlb f..., Jeremy Fitzhardinge, (Tue Aug 19, 10:58 am)
[PATCH 6 of 9] smp_function_call: add multiple queues for sc..., Jeremy Fitzhardinge, (Mon Aug 18, 2:23 pm)
[PATCH 7 of 9] x86: add multiple smp_call_function queues, Jeremy Fitzhardinge, (Mon Aug 18, 2:23 pm)
[PATCH 2 of 9] x86-32: use smp_call_function_mask for SMP TL..., Jeremy Fitzhardinge, (Mon Aug 18, 2:23 pm)
[PATCH 9 of 9] smp function calls: add kernel parameter to d..., Jeremy Fitzhardinge, (Mon Aug 18, 2:23 pm)
[PATCH 1 of 9] x86: put tlb_flush_others() stats in debugfs, Jeremy Fitzhardinge, (Mon Aug 18, 2:23 pm)
[PATCH 8 of 9] x86: make number of smp_call_function queues ..., Jeremy Fitzhardinge, (Mon Aug 18, 2:23 pm)
[PATCH 4 of 9] x86: make tlb_32|64 closer, Jeremy Fitzhardinge, (Mon Aug 18, 2:23 pm)
[PATCH 3 of 9] x86-64: use smp_call_function_mask for SMP TL..., Jeremy Fitzhardinge, (Mon Aug 18, 2:23 pm)
[PATCH 5 of 9] x86: unify tlb.c, Jeremy Fitzhardinge, (Mon Aug 18, 2:23 pm)