From: Nigel Cunningham <ncunningham@linuxmail.org>
To: Nick Piggin <nickpiggin@yahoo.com.au>
Cc: Bernard Blackham <bernard@blackham.com.au>,
Linux Memory Management <linux-mm@kvack.org>
Subject: Re: Odd kswapd behaviour after suspending in 2.6.11-rc1
Date: Sat, 15 Jan 2005 21:48:35 +1100 [thread overview]
Message-ID: <1105786115.13918.9.camel@desktop.cunninghams> (raw)
In-Reply-To: <41E8F313.4030102@yahoo.com.au>
Hi Nick.
On Sat, 2005-01-15 at 21:40, Nick Piggin wrote:
> Nigel Cunningham wrote:
> > Hi Nick and Bernard.
> >
> > On Sat, 2005-01-15 at 21:16, Nick Piggin wrote:
> >
> >>OK I think the problem is due to swsusp allocating a very large
> >>chunk of memory before suspending. After resuming, kswapd is more
> >>or less in the same state and tries a bit too hard to free things.
> >
> >
> > I'm not sure about this theory. The normal case will be that all
> > allocations (maybe one or two order 1 or order 2 allocations if I've
> > forgotten something) are order 0 and processes are thawed after we've
> > freed all the memory we were using. Could that still trigger kswapd?
> >
>
> I've seen try to do order 8 allocations or something almost as
> ridiculous. Atomic too.
I believe you. But Bernard and I are dealing with Suspend2.
> Well, correction, I've seen _reports_. Never tried swsusp myself.
:>
> I don't think a few order 0 and 1 allocations would do any harm
> because otherwise every man and his dog would be having problems.
Yes. Suspend2 does allocate a large number of zero order allocations for
submitting I/O, but again, they're all freed prior to thawing frozen
processes.
> >
> >>Thanks for the report... I'll come up with something for you to try
> >>in the next day or so.
> >
> >
> > I'm flying to America on Monday, but I'll try to keep up with the
> > progress in this and do anything I can to help.
> >
>
> It is basically a problem with one of my patches. I should be able
> to fix it (although fixing swsusp would be nice too :) ).
:> Nevertheless, if there's something suspend2 related I should fix...
Nigel
--
Nigel Cunningham
Software Engineer, Canberra, Australia
http://www.cyclades.com
Ph: +61 (2) 6292 8028 Mob: +61 (417) 100 574
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"aart@kvack.org"> aart@kvack.org </a>
next prev parent reply other threads:[~2005-01-15 10:48 UTC|newest]
Thread overview: 16+ messages / expand[flat|nested] mbox.gz Atom feed top
2005-01-13 6:14 Bernard Blackham
2005-01-13 6:26 ` Nick Piggin
2005-01-13 8:56 ` Bernard Blackham
2005-01-13 10:14 ` Bernard Blackham
2005-01-15 10:16 ` Nick Piggin
2005-01-15 10:34 ` Nigel Cunningham
2005-01-15 10:40 ` Nick Piggin
2005-01-15 10:48 ` Nigel Cunningham [this message]
2005-01-15 11:01 ` Nick Piggin
2005-01-15 11:32 ` Nigel Cunningham
2005-01-15 12:40 ` Bernard Blackham
2005-01-15 12:53 ` Bernard Blackham
2005-01-16 3:55 ` Nick Piggin
2005-01-16 4:36 ` Bernard Blackham
2005-01-16 4:56 ` Nigel Cunningham
2005-01-16 5:02 ` Nick Piggin
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=1105786115.13918.9.camel@desktop.cunninghams \
--to=ncunningham@linuxmail.org \
--cc=bernard@blackham.com.au \
--cc=linux-mm@kvack.org \
--cc=nickpiggin@yahoo.com.au \
/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