[lkml]   [2010]   [Nov]   [21]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
    SubjectRe: CFQ and dm-crypt
    >> On 11/03/10 04:23, Jeff Moyer wrote:
    >>>>> CFQ io scheduler relies on using task_struct current to determine which
    >>>>> process makes the io request. On the other hand, some dm modules (such
    >>>>> as dm-crypt) use separate threads for doing io. As CFQ sees only these
    >>>>> threads, it provides a very poor performance in such a case.
    >>>>> IMHO the correct solution for this would be to store, for every io
    >>>>> request, the process that initiated it (and preserve this information
    >>>>> while the request is processed by device mapper). Would that be feasible?
    >>> Sure. Try the attached patch (still an rfc) and let us know how it
    >>> goes. In my environment, it speed up multiple concurrent buffered
    >>> readers. I wasn't able to do a full analysis via blktrace as 2.6.37-rc1
    >>> seems to have broken blktrace support on my system.
    >> Thanks for the patch. Unfortunately, I got a kernel panic quite soon
    >> after booting the patched kernel. I was not able to reproduce the
    >> panic in a virtual machine, so I had to manually note the backtrace,
    >> thus I apologize that it's incomplete:
    > Hi, Richard,
    > I have another patch for you to try. This one holds up pretty well in
    > my testing using a dm-crypt target. Without the patch, I see no
    > priority separation:


    I am sorry for a late reply (it took me some time to try the patch on
    my main machine). Thank you for your work; unfortunately, I'm still
    getting a panic at early boot stages: The BUG_ON assert in
    cfq-iosched.c:cic_free_func failed, with a call trace as follows:


    I applied your patch at Linus's git tree (commit
    b86db4744230c94e480de56f1b7f31117edbf193), it applied cleanly.

    I tried a little to find the source of the problem, but without
    success so far. However, I think that calling put_io_context in
    bio_endio only may introduce memory leaks under certain situations
    (although this is probably not related to the panic I get):

    In blk-core.c:blk_rq_prep_clone, __bio_clone allocates the io_context,
    but if subsequent allocations fail, bio_free in free_and_out does not
    free it.

    Similarly, in dm-crypt.c:crypt_alloc_buffer, bio_put does not
    free the io_context.

    Or have I misunderstood something?


     \ /
      Last update: 2010-11-21 21:19    [W:0.022 / U:45.196 seconds]
    ©2003-2017 Jasper Spaans. hosted at Digital OceanAdvertise on this site