From: William Lee Irwin III <wli@holomorphy.com>
To: linux-kernel@vger.kernel.org
Subject: Re: [RFC] tree-based bootmem
Date: Wed, 21 Nov 2001 08:20:45 -0800 [thread overview]
Message-ID: <20011121082045.A17332@holomorphy.com> (raw)
In-Reply-To: <20011117011415.B1180@holomorphy.com>
In-Reply-To: <20011117011415.B1180@holomorphy.com>; from wli@holomorphy.com on Sat, Nov 17, 2001 at 01:14:15AM -0800
On Sat, Nov 17, 2001 at 01:14:15AM -0800, William Lee Irwin III wrote:
> This is a repost including some corrections of a bootmem allocator that
> tracks ranges explicitly, and uses segment trees to assist in searching
> for available memory. Perhaps it is even a new version. Some prior
> reports indicated mail headers were munged, preventing replies and some
> people from seeing it at all.
Some more corrections are needed, and many thanks to Russell King for
assisting me in finding them, and for providing access to a system in
order to debug these issues.
(1) prevent integer overflow in FEASIBLE()
(2) prevent integer overflow in ENDS_ABOVE()
(3) test for !segment_contains_point(optimum, goal)
in __alloc_bootmem_core() as an additional condition under
which the the goal should be discounted as a possible starting
address for the returned interval. The symptom seen without
the test is an interval that wraps around ULONG_MAX
This patch tested on ARM.
Cheers,
Bill
--- linux/mm/bootmem.c Sun Nov 18 23:42:26 2001
+++ linux-arm/mm/bootmem.c Wed Nov 21 07:58:25 2001
@@ -462,14 +440,18 @@
* that is not sufficient because of alignment constraints.
*/
-#define FEASIBLE(seg, len, align) \
- ((RND_UP(segment_start(seg), align) + (len) - 1) <= segment_end(seg))
+#define FEASIBLE(seg, len, align) \
+( \
+ (segment_end(seg) >= RND_UP(segment_start(seg), align)) \
+ && \
+ ((segment_end(seg) - RND_UP(segment_start(seg), align)) > (len))\
+)
#define STARTS_BELOW(seg,goal,align,len) \
(RND_UP(segment_start(seg), align) <= (goal))
#define ENDS_ABOVE(seg, goal, align, len) \
- (segment_end(seg) > ((goal) + (len)))
+ ((segment_end(seg) > (goal)) && ((segment_end(seg) - (goal)) > (len)))
#define GOAL_WITHIN(seg,goal,align,len) \
(STARTS_BELOW(seg,goal,align,len) && ENDS_ABOVE(seg,goal,align,len))
@@ -635,7 +608,9 @@
segment_set_endpoints(&reserved, goal, goal + length - 1);
- if(!segment_contains(optimum, &reserved))
+ if(!segment_contains_point(optimum, goal)
+ || !segment_contains(optimum, &reserved))
+
segment_set_endpoints(&reserved,
RND_UP(segment_start(optimum), align),
RND_UP(segment_start(optimum),align)+length-1);
next prev parent reply other threads:[~2001-11-21 16:21 UTC|newest]
Thread overview: 14+ messages / expand[flat|nested] mbox.gz Atom feed top
2001-11-17 9:14 [RFC] tree-based bootmem William Lee Irwin III
[not found] ` <20011118001657.A467@ucw.cz>
2001-11-18 0:01 ` William Lee Irwin III
2001-11-19 8:08 ` Robert Love
2001-11-26 21:02 ` Martin Mares
2001-11-21 16:20 ` William Lee Irwin III [this message]
2001-11-23 9:30 ` William Lee Irwin III
2001-11-28 9:04 ` Chris Wright
2001-11-28 12:40 ` William Lee Irwin III
2001-11-29 0:33 ` Chris Wright
2001-11-29 1:23 ` William Lee Irwin III
2001-11-30 2:29 ` Robert Love
2001-12-02 6:59 ` William Lee Irwin III
2001-12-02 7:08 ` Anton Blanchard
[not found] <20011118005819.3762.qmail@science.horizon.com>
2001-11-18 1:24 ` William Lee Irwin III
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=20011121082045.A17332@holomorphy.com \
--to=wli@holomorphy.com \
--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