From: Paolo Bonzini <pbonzini@redhat.com>
To: Peter Lieven <pl@dlhnet.de>
Cc: Orit Wasserman <owasserm@redhat.com>,
"qemu-devel@nongnu.org" <qemu-devel@nongnu.org>,
Corentin Chary <corentin.chary@gmail.com>
Subject: Re: [Qemu-devel] [RFC] find_next_bit optimizations
Date: Mon, 11 Mar 2013 15:14:46 +0100 [thread overview]
Message-ID: <513DE6D6.9000105@redhat.com> (raw)
In-Reply-To: <513DDFA3.1020308@dlhnet.de>
Il 11/03/2013 14:44, Peter Lieven ha scritto:
> Hi,
>
> I ever since had a few VMs which are very hard to migrate because of a
> lot of memory I/O. I found that finding the next dirty bit
> seemed to be one of the culprits (apart from removing locking which
> Paolo is working on).
>
> I have to following proposal which seems to help a lot in my case. Just
> wanted to have some feedback.
> I applied the same unrolling idea like in buffer_is_zero().
>
> Peter
>
> --- a/util/bitops.c
> +++ b/util/bitops.c
> @@ -24,12 +24,13 @@ unsigned long find_next_bit(const unsigned long
> *addr, unsigned long size,
> const unsigned long *p = addr + BITOP_WORD(offset);
> unsigned long result = offset & ~(BITS_PER_LONG-1);
> unsigned long tmp;
> + unsigned long d0,d1,d2,d3;
>
> if (offset >= size) {
> return size;
> }
> size -= result;
> - offset %= BITS_PER_LONG;
> + offset &= (BITS_PER_LONG-1);
> if (offset) {
> tmp = *(p++);
> tmp &= (~0UL << offset);
> @@ -43,6 +44,18 @@ unsigned long find_next_bit(const unsigned long
> *addr, unsigned long size,
> result += BITS_PER_LONG;
> }
> while (size & ~(BITS_PER_LONG-1)) {
> + while (!(size & (4*BITS_PER_LONG-1))) {
This really means
if (!(size & (4*BITS_PER_LONG-1))) {
while (1) {
...
}
}
because the subtraction will not change the result of the "while" loop
condition.
What you want is probably "while (size & ~(4*BITS_PER_LONG-1))", which
in turn means "while (size >= 4*BITS_PER_LONG).
Please change both while loops to use a ">=" condition, it's easier to read.
Paolo
> + d0 = *p;
> + d1 = *(p+1);
> + d2 = *(p+2);
> + d3 = *(p+3);
> + if (d0 || d1 || d2 || d3) {
> + break;
> + }
> + p+=4;
> + result += 4*BITS_PER_LONG;
> + size -= 4*BITS_PER_LONG;
> + }
> if ((tmp = *(p++))) {
> goto found_middle;
> }
next prev parent reply other threads:[~2013-03-11 14:14 UTC|newest]
Thread overview: 22+ messages / expand[flat|nested] mbox.gz Atom feed top
2013-03-11 13:44 [Qemu-devel] [RFC] find_next_bit optimizations Peter Lieven
2013-03-11 14:04 ` Peter Maydell
2013-03-11 14:14 ` Paolo Bonzini [this message]
2013-03-11 14:22 ` Peter Lieven
2013-03-11 14:29 ` Peter Lieven
2013-03-11 14:35 ` Paolo Bonzini
2013-03-11 15:24 ` Peter Lieven
2013-03-11 15:25 ` Peter Maydell
2013-03-11 15:29 ` Paolo Bonzini
2013-03-11 15:37 ` Peter Lieven
2013-03-11 15:58 ` Paolo Bonzini
2013-03-11 17:06 ` ronnie sahlberg
2013-03-11 17:07 ` Paolo Bonzini
2013-03-11 18:20 ` Peter Lieven
2013-03-12 7:32 ` [Qemu-devel] [PATCH] bitops: unroll while loop in find_next_bit() Peter Lieven
2013-03-11 15:37 ` [Qemu-devel] [RFC] find_next_bit optimizations Peter Maydell
2013-03-11 15:41 ` Peter Lieven
2013-03-11 15:42 ` Paolo Bonzini
2013-03-11 15:48 ` Peter Lieven
2013-03-12 8:35 ` Stefan Hajnoczi
2013-03-12 8:41 ` Peter Lieven
2013-03-12 15:12 ` Stefan Hajnoczi
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=513DE6D6.9000105@redhat.com \
--to=pbonzini@redhat.com \
--cc=corentin.chary@gmail.com \
--cc=owasserm@redhat.com \
--cc=pl@dlhnet.de \
--cc=qemu-devel@nongnu.org \
/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;
as well as URLs for NNTP newsgroup(s).