Re: [PATCH] proc: faster /proc/*/status

From	Joe Perches <joe@perches.com>
Newsgroups	linux.kernel
Subject	Re: [PATCH] proc: faster /proc/*/status
Date	2016-08-07 19:40 +0200
Message-ID	<s3Adk-6L0-3@gated-at.bofh.it> (permalink)
References	<s3hDI-2QM-3@gated-at.bofh.it> <s3mN3-6xA-5@gated-at.bofh.it> <s3s65-1ub-5@gated-at.bofh.it> <s3zAB-6iR-1@gated-at.bofh.it>
Organization	linux.* mail to news gateway

Show all headers | View raw

On Sun, 2016-08-07 at 09:59 -0700, Andi Kleen wrote:
> > And then what? Parsing format string is still be there.
> Perhaps could have a fast path for simple cases.
> > 
> > This is first line of profile of the first function (format_decode)
> > 
> >        │     static noinline_for_stack
> >        │     int format_decode(const char *fmt, struct printf_spec *spec)
> >        │     {
> >  10.38 │       push   %rbp			<===
> >   1.07 │       mov    %rsp,%rbp
> >   1.09 │       push   %r12
> >   4.51 │       mov    %rsi,%r12
> >   1.40 │       push   %rbx
> >   1.86 │       mov    %rdi,%rbx
> >        │       sub    $0x8,%rsp
> > 
> > It is so bloated that gcc needs to be asked to not screw up with stack
> > size.
> What happens when you drop all the noinlines for this? I assume
> this would alread make it faster. And now that we have bigger
> stacks we can likely tolerate it.

%pV recurses through these code paths.
I believe the maximum current recursion depth is 3.

Thread

[PATCH] proc: faster /proc/*/status Alexey Dobriyan <adobriyan@gmail.com> - 2016-08-06 23:50 +0200
  Re: [PATCH] proc: faster /proc/*/status Joe Perches <joe@perches.com> - 2016-08-07 00:50 +0200
    [PATCH] seq/proc: Modify seq_put_decimal_[u]ll to take a const char *, not char Joe Perches <joe@perches.com> - 2016-08-09 05:10 +0200
  Re: [PATCH] proc: faster /proc/*/status Andi Kleen <andi@firstfloor.org> - 2016-08-07 05:20 +0200
    Re: [PATCH] proc: faster /proc/*/status Alexey Dobriyan <adobriyan@gmail.com> - 2016-08-07 11:00 +0200
      Re: [PATCH] proc: faster /proc/*/status Andi Kleen <andi@firstfloor.org> - 2016-08-07 19:00 +0200
        Re: [PATCH] proc: faster /proc/*/status Joe Perches <joe@perches.com> - 2016-08-07 19:40 +0200
          Re: [PATCH] proc: faster /proc/*/status Andi Kleen <andi@firstfloor.org> - 2016-08-07 19:50 +0200
            Re: [PATCH] proc: faster /proc/*/status Joe Perches <joe@perches.com> - 2016-08-07 23:30 +0200
              Re: [PATCH] proc: faster /proc/*/status Andrew Morton <akpm@linux-foundation.org> - 2016-08-10 01:30 +0200
                [RFC PATCH] meminfo: Break apart a very long seq_printf with #ifdefs Joe Perches <joe@perches.com> - 2016-08-10 20:30 +0200

csiph-web