[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <17631.1185956833@redhat.com>
Date: Wed, 01 Aug 2007 09:27:13 +0100
From: David Howells <dhowells@...hat.com>
To: Peter Zijlstra <peterz@...radead.org>
Cc: torvalds@...l.org, akpm@...l.org, steved@...hat.com,
trond.myklebust@....uio.no, linux-fsdevel@...r.kernel.org,
linux-cachefs@...hat.com, nfsv4@...ux-nfs.org,
linux-kernel@...r.kernel.org
Subject: Re: [PATCH 02/14] FS-Cache: Recruit a couple of page flags for cache management
Peter Zijlstra <peterz@...radead.org> wrote:
> Not sure its a good idea to overload page_has_private() with an
> overloadable page-flag. What if some future FS wants to use
> PG_owner_priv_2 for other purposes?
All that it means is that releasepage() and co will get called if a page is to
be released or invalidated that has that bit set. I think that's something a
future FS could probably live with.
However, I do have to trigger a call to releasepage() and co *somehow*.
> Obviously filesystems cannot use these two page-flags if they want to be
> compatible with FS-cache, but need all filesystems be?
What do you mean? That's why I went for the PG_owner_priv_* approach rather
than just naming the bits unto FS-Cache directly.
> (also, ouch! - 2 pageflags)
Yeah. The consequence of having things asynchronous is that you have to find
signalling mechanisms to synchronise around the asynchronicity:-/
Furthermore, it occurs to me that I can't use PG_private or page->private to
store this information because I want to make isofs use caching, and those two
pieces of information are owned by the buffering code.
David
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Powered by blists - more mailing lists