[PATCH] CacheFS - general filesystem cache

From: David Howells (dhowells_at_redhat.com)
Date: 08/27/04

  • Next message: Wouter Van Hemel: "Re: kernel 2.6.8 pwc patches and counterpatches" 
    To: torvalds@osdl.org, akpm@osdl.org
Date:	Fri, 27 Aug 2004 01:23:03 +0100

    Hi Linus, Andrew,

    I've packaged my generic filesystem cache filesystem into patches and also
    produced patches for my AFS filesystem to use it. Work is also in progress to
    alter the NFS client use this interface too, and I think the ISO9660
    filesystem could also benefit.

    I've put a tarball of them on my Red Hat webpage because one of them is rather
    large. If you grab:

            http://people.redhat.com/~dhowells/cachefs/cachefs-patches.tar.bz2

    You'll find the following patches contained therein:

        (1) cachefs-nullfileptr-2681mm4.diff

            Fix a bug introduced into do_generic_mapping_read() by which a file
            pointer becomes required. I had arranged things so that it was
            optional so that I could call it directly on an inode.

        (2) cachefs-rastateinit-2681mm4.diff

            Revert the change made to file_ra_state_init() that removed the symbol
            export that I'd put in, and also make it fully initialise the state
            again.

            This was also done so that do_generic_mapping_read() could be called
            directly on an inode.

        (3) cachefs-fs-2681mm4.diff

            The cachefs filesystem source and interface header file.

        (4) cachefs-doc-2681mm4.diff

            Documentation about using cachefs.

        (5) cachefs-wppageop-2681mm4.diff

            Add a notification operation to the address space that gets invoked
            when a read-only protected page is about to be made writable. I've
            shown this to Rik van Riel who seems happy with it.

            Using a combination of (5) and (6), I can have cachefs write a netfs
            page into the cache in the background whilst the netfs readpage
            completes and passes the now unlocked page back to userspace.

            Previously, the netfs readpage function had to be synchronous with
            respect to writing the page to disc. Now it can use this callback and
            prepare_write() to synchronise upon the page being modified in the
            page cache before it gets DMA'd to cachefs's disc block.

            David Woodhouse is also interested in using this to do lazy
            reservations in JFFS2 shared writable mapping support.

        (6) cachefs-fspagesync-2681mm4.diff

            This adds another page flag for intra-filesystem synchronisation, to
            operate in a manner similar to PG_writeback.

            I've made my AFS client use this to indicate that a page has now been
            written to the backing cache and can now be safely modified.

        (7) cachefs-afs-2681mm4.diff

            This changes my AFS client so that it can make use of cachefs.

    Signed-Off-By: David Howells <dhowells@redhat.com>

    I've also put a utility on my webpage for dumping the cachefs update journal
    contents:

            http://people.redhat.com/~dhowells/cachefs/dump-journal.c

    It just needs the fs/cachefs/cachefs-int.h header file placing in the same
    directory to compile.

    David
    -
    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/

  • Next message: Wouter Van Hemel: "Re: kernel 2.6.8 pwc patches and counterpatches"

    Relevant Pages

    • Re: Ext4 and the "30 second window of death"
      ... I use Ext4 as my filesystem. ... Fedora will have the patches so that applications that do ... I've been using ext4 with kde 4.2.1 ...
      (Linux-Kernel)
    • Re: [patch] mlock-as-nonroot revisted
      ... >> by these patches. ... Hugetlb file creation through the filesystem never calls ... 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/ ...
      (Linux-Kernel)
    • Re: [PATCH 1/4] vfs: adds the S_PRIVATE flag and adds use to security
      ... >> infrastructure to allow the filesystem to perform its own access control. ... I'll assume that the other three patches are unchanged. ... send the line "unsubscribe linux-kernel" in ...
      (Linux-Kernel)
    • [PATCH 00/25] Read-only bind mounts
      ... are a few merge conflicts with the audit git tree, ... The first four patches here are very simple cleanups, ... Why do we need r/o bind mounts? ... This feature allows a read-only view into a read-write filesystem. ...
      (Linux-Kernel)
    • [PATCH 0/13] Filesystem in Userspace
      ... Please consider applying the following patches, which add Filesystem ... 11 - file data invalidate operation ... send the line "unsubscribe linux-kernel" in ...
      (Linux-Kernel)