lkml.org 
[lkml]   [1998]   [Oct]   [15]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
    /
    Date
    From
    SubjectRe: sb16 duplex fix
    Hannu Savolainen wrote:
    >
    > On Wed, 14 Oct 1998, Alan Cox wrote:
    >
    > > I would agree with this, mmap() on sound is a neat hack but it is very very
    > > questionable if it is needed or useful. Its really a quake compatibility
    > > feature 8)
    > I second this opinion.
    >
    > It looks like sound devices in future will have more and more local
    > buffer memory or FIFO space which makes using mmap() impossible, difficult
    > or at least rather inpractical. I don't see any reason to support this
    > feature in future if somebody gets the Quake folks to implement a non
    > mmap() version.
    >
    > This feature will disappear from the OSS programmer's guide in near
    > future.

    I'd like to convert a few applications that now have their own sound
    driver to use the standard sound driver. However, currently I can't, as
    a few features are missing, namely the following:

    - "true queue length": the app should be able to query the true queue
    length,
    including on-card FIFO buffers etc. (and possibly including codec
    filter
    delay)

    - the application should be able to query and eventually set (within
    bounds)
    the wakeup latency (this is related to the fragment size)

    - I need the possibility to switch between input and output (and vice
    versa)
    as fast as possible on half duplex cards, that means I need a
    possibility
    to suppress lengthy autocalibration cycles when switching. Not
    necessarily
    a new API element, could be implicit when read follows write and vice
    versa.

    - It would be nice if it was possible to tell the sound driver something
    like
    "only wake me up when there are at least N samples (or bytes) in the
    receive queue"
    and "only wake me up if there are more than N samples free in the
    transmit queue
    or the transmit queue contains less than M samples".

    Can we agree on an API for this?

    Tom

    -
    To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
    the body of a message to majordomo@vger.rutgers.edu
    Please read the FAQ at http://www.tux.org/lkml/

    \
     
     \ /
      Last update: 2005-03-22 13:45    [W:0.022 / U:0.292 seconds]
    ©2003-2016 Jasper Spaans. hosted at Digital OceanAdvertise on this site