Re: [RFC][PATCH] fs: optimize inotify/fsnotify code for unwatched files

[Paul E. McKenney]

On Fri, Jun 19, 2015 at 05:39:11PM -0700, Dave Hansen wrote:
> On 06/19/2015 04:33 PM, Andi Kleen wrote:
> >> > I *think* we can avoid taking the srcu_read_lock() for the
> >> > common case where there are no actual marks on the file
> >> > being modified *or* the vfsmount.
> > What is so expensive in it? Just the memory barrier in it?
> 
> The profiling doesn't hit on the mfence directly, but I assume that the
> overhead is coming from there.  The "mov    0x8(%rdi),%rcx" is identical
> before and after the barrier, but it appears much more expensive
> _after_.  That makes no sense unless the barrier is the thing causing it.

OK, one thing to try is to simply delete the memory barrier.  The
resulting code will be unsafe, but will probably run well enough to
get benchmark results.  If it is the memory barrier, you should of
course get increased throughput.

							Thanx, Paul

> Here's how the annotation mode of 'perf top' breaks it down:
> 
> >        â    ffffffff810fb480 <load0>:
> >        â      nop
> >        â      mov    (%rdi),%rax
> >   0.58 â      push   %rbp
> >        â      incl   %gs:0x7ef0f488(%rip)
> >   1.73 â      mov    %rsp,%rbp
> >        â      and    $0x1,%eax
> >        â      movslq %eax,%rdx
> >   0.58 â      mov    0x8(%rdi),%rcx
> >        â      incq   %gs:(%rcx,%rdx,8)
> >        â      mfence
> >  69.94 â      add    $0x2,%rdx
> >   7.51 â      mov    0x8(%rdi),%rcx
> >   4.05 â      incq   %gs:(%rcx,%rdx,8)
> >  13.87 â      decl   %gs:0x7ef0f45f(%rip)
> >        â      pop    %rbp
> >   1.73 â    â retq
> >                                                          
> 

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
Please read the FAQ at  http://www.tux.org/lkml/