From: "Martin J. Bligh" <mbligh@aracnet.com>
To: "Juan M. de la Torre" <jmtorre@gmx.net>, linux-kernel@vger.kernel.org
Subject: Re: Possible bug in __alloc_pages() ?
Date: Tue, 17 Sep 2002 15:02:15 -0700 [thread overview]
Message-ID: <132900000.1032300135@flay> (raw)
In-Reply-To: <20020917214804.GA891@apocalipsis>
This is deliberate, and is to discourage fallback. I'm not desperately fond
of the method, but I'm told it's not an accidental typo / bug.
M.
--On Tuesday, September 17, 2002 23:48:04 +0200 "Juan M. de la Torre" <jmtorre@gmx.net> wrote:
>
> Hi, this code appears at the beggining of __page_alloc() (kernel 2.4.19):
>
> min = 1UL << order;
> for (;;) {
> zone_t *z = *(zone++);
> if (!z)
> break;
>
> min += z->pages_low;
> if (z->free_pages > min) {
> page = rmqueue(z, order);
> if (page)
> return page;
> }
> }
>
> AFAIK, what this code does is to try to alloc the requested pages from
> the first zone in a zone_list (passed as argument) which have enought free
> pages.
>
> A zone is considered to have enought free pages if z->free_pages is greater
> than (number_of_requested_pages + z->pages_low).
>
> In the loop shown, the first iteration is OK, but in the second iteration
> (which only occurs if the first zone in the zone_list hasn't enought free
> pages) the zone will only be considered to have enought free pages if
> z->free_pages is greater that (number_of_requested_pages + z->pages_low
> + PREV_ZONE->pages_low).
>
> I think this is a bug, but i'm not sure (i'm not a VM hacker).
>
> If it is a bug, there are other two loops in the same function which
> are buggy.
>
>
> Possible patch:
>
> --- linux/mm/page_alloc.c.orig Tue Sep 17 23:45:02 2002
> +++ linux/mm/page_alloc.c Tue Sep 17 23:46:45 2002
> @@ -330,8 +330,7 @@
> if (!z)
> break;
>
> - min += z->pages_low;
> - if (z->free_pages > min) {
> + if (z->free_pages > min + z->pages_low) {
> page = rmqueue(z, order);
> if (page)
> return page;
> @@ -354,8 +353,8 @@
> local_min = z->pages_min;
> if (!(gfp_mask & __GFP_WAIT))
> local_min >>= 2;
> - min += local_min;
> - if (z->free_pages > min) {
> +
> + if (z->free_pages > min + local_min) {
> page = rmqueue(z, order);
> if (page)
> return page;
> @@ -394,8 +393,7 @@
> if (!z)
> break;
>
> - min += z->pages_min;
> - if (z->free_pages > min) {
> + if (z->free_pages > min + z->pages_min) {
> page = rmqueue(z, order);
> if (page)
> return page;
>
> Regards,
> Juanma
>
> --
> /jm
>
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at http://www.tux.org/lkml/
>
>
prev parent reply other threads:[~2002-09-17 22:00 UTC|newest]
Thread overview: 2+ messages / expand[flat|nested] mbox.gz Atom feed top
2002-09-17 21:48 Possible bug in __alloc_pages() ? Juan M. de la Torre
2002-09-17 22:02 ` Martin J. Bligh [this message]
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=132900000.1032300135@flay \
--to=mbligh@aracnet.com \
--cc=jmtorre@gmx.net \
--cc=linux-kernel@vger.kernel.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).