Re: [PATCH 0/3] TLB flush multiple pages per IPI v5

[Ingo Molnar]

* Dave Hansen <dave.hansen@xxxxxxxxx> wrote:

> On 06/08/2015 12:52 PM, Ingo Molnar wrote:
> > A CR3 driven TLB flush takes less time than a single INVLPG (!):
> > 
> >    [    0.389028] x86/fpu: Cost of: __flush_tlb()               fn            :    96 cycles
> >    [    0.405885] x86/fpu: Cost of: __flush_tlb_one()           fn            :   260 cycles
> >    [    0.414302] x86/fpu: Cost of: __flush_tlb_range()         fn            :   404 cycles
> 
> How was that measured, btw?  Are these instructions running in a loop?

Yes - see the x86 benchmarking patch in the big FPU submission for an earlier 
version.

> Does __flush_tlb_one() include the tracepoint?

No tracing overhead.

> (From the commit I referenced) This was (probably) using a different method than 
> you did, but "FULL" below is __flush_tlb() while "1" is __flush_tlb_one().  The 
> "cycles" includes some overhead from the tracing:
> 
> >       FULL:   2.20%   2.20% avg cycles:  2283 cycles/page: xxxx samples: 23960
> >          1:  56.92%  59.12% avg cycles:  1276 cycles/page: 1276 samples: 620895
> 
> So it looks like we've got some discrepancy, either from the test methodology or 
> the CPU.  All of the code and my methodology are in the commit.  Could you share 
> yours?

Yes, you can reproduce it by applying this patch from the FPU series:

  Subject: [PATCH 207/208] x86/fpu: Add FPU performance measurement subsystem

(you were Cc:-ed to it, so it should be in your inbox.)

I've got a more advanced version meanwhile, will post it in the next couple of 
days or so.

> > it's true that a full flush has hidden costs not measured above, because it has 
> > knock-on effects (because it drops non-global TLB entries), but it's not _that_ 
> > bad due to:
> > 
> >   - there almost always being a L1 or L2 cache miss when a TLB miss occurs,
> >     which latency can be overlaid
> > 
> >   - global bit being held for kernel entries
> > 
> >   - user-space with high memory pressure trashing through TLBs typically
> > 
> > ... and especially with caches and Intel's historically phenomenally low TLB 
> > refill latency it's difficult to measure the effects of local TLB refills, let 
> > alone measure it in any macro benchmark.
> 
> All that you're saying there is that you need to consider how TLB misses act in 
> _practice_ and not just measure worst-case or theoretical TLB miss cost.  I 
> completely agree with that.

So I'm saying considerably more than that: I consider it likely that a full TLB 
flush is not nearly as costly as assumed, for the three reasons outlined above.

It might even be a performance win in Mel's benchmark - although possibly not 
measurable within measurement noise levels.

Thanks,

	Ingo
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/