Re: [PARPORT] More libieee questions

From: Tim Waugh (
Date: Tue May 27 2003 - 07:23:47 EDT

  • Next message: Peter Asemann: "[PARPORT] ppdev questions again"

    On Tue, May 27, 2003 at 12:29:09PM +0200, Peter Asemann wrote:

    > 1. If "disallow method ppdev" is set - will there be any difference
    > regarding functionality, superuser rights requirements, speed?

    I think there are still some protocols that are not implemented inside
    libieee1284 but are implemented in ppdev: so those protocols won't be
    available when 'disallow method ppdev' is set. (I might be wrong;
    they might all be implemented now.)

    One feature of the ppdev method is that you only need to have
    read/write access to /dev/parport0 to communicate on that port. One
    other access method which has this nice non-root feature is /dev/port,
    which libieee1284 will try. So if you are not root, and don't want to
    use ppdev, better make sure you have access to /dev/port. (Of course,
    that gives access to all I/O ports on the machine.)

    But /dev/port is quite a lot slower than the ioperm method (needs

    Regarding speed: it depends what method will be used in its place, and
    what type of protocol you are using. If you were using hardware
    accelerated ECP, then yes it will be quite a lot slower to avoid

    > 2. If libieee detects a "software (epp, ecp...) mode, which "software" is
    > meant then? Is that software some driver libieee makes use of (ppdev?), or
    > does libieee itself implement some protocols?

    libieee1284 itself implements some protocols, and so does ppdev. When
    CAP1284_{ECP,EPP}SWE is set it means that libieee1284 has a full
    implementation of that mode, or knows that the method it will use
    does. (For EPP at the moment libieee1284 itself only does data
    transfers and not addresses.)

    There isn't a way for the library to ask ppdev what modes it
    implements in software, so it doesn't say anything about it.


    -- To unsubscribe, send mail to: --
    -- with the single word "unsubscribe" in the body of the message. --

    This archive was generated by hypermail 2b29 : Tue May 27 2003 - 07:28:14 EDT