xfs
[Top] [All Lists]

Re: patch to give kmalloc a chance (RFC)

To: Andrew Morton <akpm@xxxxxxxxx>
Subject: Re: patch to give kmalloc a chance (RFC)
From: Chris Wedgwood <cw@xxxxxxxx>
Date: Fri, 27 Sep 2002 17:06:57 -0700
Cc: Steve Lord <lord@xxxxxxx>, Ravi Wijayaratne <ravi_wija@xxxxxxxxx>, linux-xfs@xxxxxxxxxxx, Eric Sandeen <sandeen@xxxxxxx>, riel@xxxxxxxxxxxxxxxxxxxxxxxxx
In-reply-to: <3D94E12C.F0D0FE60@digeo.com>
References: <20020927215041.77738.qmail@web40614.mail.yahoo.com> <1033163688.10533.79.camel@jen.americas.sgi.com> <3D94E12C.F0D0FE60@digeo.com>
Sender: linux-xfs-bounce@xxxxxxxxxxx
User-agent: Mutt/1.4i
On Fri, Sep 27, 2002 at 03:52:28PM -0700, Andrew Morton wrote:

> For a NOFS allocation, simply retrying is a crude-but-effective fix.
> We have kicked kswapd and kswapd _does_ have __GFP_FS.

I wondered about this... and sleeping briefly isn't a bad if it
penalizes the memory hog processes (sure, this isn't guaranteed but
one hopes it happens more often than not).

> So things work out, as long as the filesystem doesn't hold locks
> which would prevent kswapd from being able to reclaim some memory.

What about swapping to a file on the fs?  Might this not be such a
lock?


  --cw


<Prev in Thread] Current Thread [Next in Thread>