[PARPORT] Parport sharing code on old 486 laptop

From: Ian Hinder (ian.hinder@cantab.net)
Date: Thu Aug 23 2001 - 14:26:15 EDT

  • Next message: Dave Strauss: "Re: [PARPORT] ECP mode transfers in 2.4.x kernels"


    I've been trying to get an old 486 laptop and my Pentium III to
    communicate over PLIP. I've got it successfully working with a 2.4.2
    kernel on the PIII and a 2.0.3 kernel on the laptop. The laptop kernel
    had PLIP statically linked and lp removed so there weren't any conflicts.
    This works fine, and I can ping the laptop from the PIII, getting about 6
    ms reply time; flood ping gives a few dropped packets initially (about 6)
    then no more; I can do 'ping -i 0.01 <host>' from the PIII and it
    doesn't drop any packets. The PLIP code in the laptop is from a kernel
    before the parallel port sharing code was introduced as far as I can tell.

    HOWEVER, I'd like to run a more recent kernel on the laptop. I have tried
    2.2.19 and 2.4.5 (both from Slackware 8.0) and found that I could no
    longer compile the PLIP support statically into the kernel as it couldn't
    find the IRQ to use etc. So I separated them out as modules. The kernels
    had module support enabled, plip enabled, parport and parport_pc
    enabled. The kernels boot, and I do

    insmod parport.o
    insmod parport_pc.o io=0x378 irq=7
    insmod plip.o
    ifconfig plip0 pointopoint up
    route add -net netmask dev plip0

    on the laptop. Now, the connection *seems* ok, since I can do a normal
    "ping <otherhost>", but I get about 20 ms this time, rather than 6 ms as
    before with the non parport-sharing code. However, when I do a flood ping
    or try ping -i 0.01, I get a zillion lost packets. NFS barely works at
    all, whereas before it worked perfectly and gave me about 30 k/s I think.

    My guess is that the parport sharing code is causing some sort of
    bottleneck which is only showing up on the slow 486. When the laptop is
    using the 2.0.3 PLIP code (no parport sharing), it communicates fine with
    the PIII which is using the more modern PLIP (sharing) code.

    Additonal information: The kernels compiled for the 486 are extremely
    minimal; everything I can get away with has been disabled. PCI is
    disabled, but I'm not sure this is relevent. This laptop is curious
    because it will hang on kernel boot if you use the newer IDE drivers; I
    had to tell the kernel to use the older ones for it to boot properly.

    Any ideas? If this fails I may have to resort to PPP over the serial
    port; something I do not relish the thought of!

    Ian Hinder

    -- 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 : Thu Aug 23 2001 - 14:29:06 EDT