From: Andrew Morton <akpm@linux-foundation.org>
To: Christoph Lameter <clameter@sgi.com>
Cc: Hugh Dickins <hugh@veritas.com>, Nick Piggin <npiggin@suse.de>,
dgc@sgi.com, linux-kernel@vger.kernel.org
Subject: Re: [PATCH 2/2] Optimize compound_head() by avoiding a shared page flag
Date: Sat, 7 Apr 2007 15:51:48 -0700 [thread overview]
Message-ID: <20070407155148.94da92e8.akpm@linux-foundation.org> (raw)
In-Reply-To: <Pine.LNX.4.64.0704071509550.31503@schroedinger.engr.sgi.com>
On Sat, 7 Apr 2007 15:16:17 -0700 (PDT) Christoph Lameter <clameter@sgi.com> wrote:
> On Fri, 6 Apr 2007, Andrew Morton wrote:
>
> > Did you investigate
> >
> > static inline int page_tail(struct page *page)
> > {
> > return ((page->flags & (PG_compound|PG_tail)) == (PG_compound|PG_tail));
> > }
>
> The usual test_bit that we are using there uses a volatile reference
> so these wont be combined if I check them separately.
>
> A working example of the above would be much uglier:
>
> static inline int page_tail(struct page *page)
> {
> return ((page->flags & ((1L << PG_compound)|(1L << PG_tail))) ==
> ((1L << PG_compound)|(1L << PG_tail)));
> }
>
> May be this can be cleaned up somehow.
It might generate better code to do
unsigned long compound;
compound = page->flags & (1 << PG_compound);
if (PG_compound > PG_tail)
return compound & (page->flags << (PG_compound - PG_tail));
else
return compound & (page->flags << (PG_tail - PG_compound));
ie: get the PG_compound flag into `compound', then bitwise-and that with
the PG_tail flag, after shifting it into PG_compound' slot. The return
value will be zero if either bit is clear, (1<<PG_compound) if both are
set. The `if (PG_compound > PG_tail)' will be swallowed by the compiler.
The compiler should turn it all into
(page->flags & N) & (page->flags << M)
Which may or may not be better than (page->flags & N == N), dunno.
Probably not - if the compiler's any good it won't save a branch, I
suspect.
Which is all a ton of fun, but this subversion of the architecture's
freedom to use volatile, memory barriers etc is a worry. We do the same in
page_alloc.c, of course...
next prev parent reply other threads:[~2007-04-07 22:51 UTC|newest]
Thread overview: 19+ messages / expand[flat|nested] mbox.gz Atom feed top
2007-04-05 22:36 [PATCH 1/2] Make page->private usable in compound pages V1 Christoph Lameter
2007-04-05 22:36 ` [PATCH 2/2] Optimize compound_head() by avoiding a shared page flag Christoph Lameter
2007-04-07 5:23 ` Andrew Morton
2007-04-07 22:16 ` Christoph Lameter
2007-04-07 22:51 ` Andrew Morton [this message]
2007-04-08 0:21 ` Christoph Lameter
2007-04-08 1:25 ` Andrew Morton
2007-04-08 1:32 ` Christoph Lameter
2007-04-08 1:48 ` Andrew Morton
2007-04-09 17:22 ` Christoph Lameter
2007-04-09 18:09 ` Christoph Lameter
2007-04-09 18:45 ` Andrew Morton
2007-04-09 18:49 ` Christoph Lameter
2007-04-09 22:05 ` Christoph Lameter
2007-04-05 23:43 ` [PATCH 1/2] Make page->private usable in compound pages V1 Andrew Morton
2007-04-06 17:01 ` Christoph Lameter
2007-04-06 20:04 ` Andrew Morton
2007-04-06 20:28 ` Christoph Lameter
2007-04-06 19:46 ` Christoph Lameter
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20070407155148.94da92e8.akpm@linux-foundation.org \
--to=akpm@linux-foundation.org \
--cc=clameter@sgi.com \
--cc=dgc@sgi.com \
--cc=hugh@veritas.com \
--cc=linux-kernel@vger.kernel.org \
--cc=npiggin@suse.de \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox