[PARPORT] [Fwd: ECP slower in 2.4.19 than in 2.4.10]

From: Philip Blundell (Philip.Blundell@pobox.com)
Date: Sun Jun 16 2002 - 07:15:36 EDT

  • Next message: Keith Bainbridge: "[PARPORT] To: "linux-parport@torque.net" <linux-parport@torque.net>"

    Can anybody help this man?

    p.


    attached mail follows:


    Hi,
    I'm currently developing an external hardware that receives data via ECP
    from a Compaq Armada 300 notebook. I don't know for sure whether the
    parallel port is handled by the i82371 or the SMSC-Super-IO, but guessing
    from the transfer rates (below 500k/s) it's the SMSC.

    Now the problem: With 2.4.10, I got around 350kbyte/s with ECP, the
    low-pulses of the strobe signal were at around 500ns. Now with 2.4.19-pre10,
    I get only about 250kbyte/s, the low duration increased to about 750ns. The
    250kbyte/s are too slow for my (realtime) device, so I have a problem.
    Switching back to the old kernel solves this issue...

    Do you know what has changed between 2.4.10 and 2.4.19 that can cause this
    slowdown? I've searched the parport_pc-code for timing related stuff, but
    found nothing useful. Can it be caused by some changed "tuning" of the
    PCI-ISA-bridge?

    The detection output when loading parport_pc with specified io, irq and dma
    looks like this:
    parport0: PC-style at 0x378 (0x778), irq 7<7>0x378: FIFO is 16 bytes
    0x378: writeIntrThreshold is 15
    0x378: readIntrThreshold is 15
    0x378: PWord is 8 bits
    0x378: Interrupts are ISA-Pulses
    0x378: ECP port cfgA=0x10 cfgB=0x00
    0x378: ECP settings irq=<none or set by other means> dma=<none or set by
    other means>
    , dma 3 [PCSPP,TRISTATE,COMPAT,EPP,ECP,DMA]

    Thank you for any hint in tracking this problem...

    -- 
             Georg Acher, acher@in.tum.de         
             http://www.in.tum.de/~acher/
              "Oh no, not again !" The bowl of petunias          
    

    -- To unsubscribe, send mail to: linux-parport-request@torque.net -- -- with the single word "unsubscribe" in the body of the message. --



    This archive was generated by hypermail 2b29 : Sun Jun 16 2002 - 07:22:51 EDT