Re: Keyboard is frozen on boot of 2.3.41

From: Jeremy Fitzhardinge (jeremy@goop.org)
Date: Sat Jan 29 2000 - 17:33:13 EST

  • Next message: Alan Cox: "Re: autofs v4, nfs v3, 2.2.15 ?"

    On 29-Jan-00 Linus Torvalds wrote:
    >
    >
    > On Sat, 29 Jan 2000, Jeremy Fitzhardinge wrote:
    >>
    >> I've had a problem with the keyboard/trackpad freezing with USB enabled on
    >> my
    >> Sharp Actius A100 laptop.
    >>
    >> The problem is the PS/2 mouse driver is most unhappy about sharing its
    >> interrupt with anything else. It wants IRQ 12, and if the USB controller
    >> also
    >> gets IRQ 12 the keyboard (and mouse) locks up as soon as you touch the
    >> mouse.
    >
    > ok, this just confirms my theory.
    >
    > Mouse, USB and PCMCIA can easily all be on irq12 - it may not be the most
    > common setup, but it is definitely _one_ common setup (another is having
    > USB and PCMCIA share irq9).
    >
    > And yes, USB and PCMCIA are the only drivers that are _likely_ to share
    > the interrupt, so this also explains why problems like this are new. The
    > keyboard/mouse driver didn't commonly use to care because it historically
    > didn't get many "spurious" interrupts.
    >
    > This might actually be an old lockup problem - there have always been
    > reports of keyboards dying. It might just have been explained only now..
    >
    > I don't see anything actually _wrong_ in the keyboard interrupt handler,
    > though. That worries me. What we do to the mouse hardware is pretty much
    > the same whether the mouse device is open or not, and the only real
    > difference in opening the mouse is that it makes us save the events...
    >
    > Oh. We do some mouse initialization at open too. That might be a problem,
    > and that _does_ make a difference unlike the actual irq path.
    >
    > In fact, in "aux_open()", could you people who see a problem try to
    > disable the toshiba4030cdt workaround? Just #if 0 the send_data() code
    > that sends a KBD_CMD_ENABLE thing to the keyboard, and see if that makes
    > any difference? The shared interrupt thing may be a red herring, just a
    > timing difference.

    Didn't help. What's even more odd, there's no interrupts recorded:

    : abulafia:pts/0; cat /proc/interrupts
               CPU0
      0: 11745 XT-PIC timer
      1: 116 XT-PIC keyboard
      2: 0 XT-PIC cascade
      3: 243 XT-PIC pcnet_cs
      5: 1 XT-PIC Crystal audio controller
      8: 1 XT-PIC rtc
     12: 0 XT-PIC usb-uhci, Ricoh Co Ltd RL5c475, PS/2 Mouse
     13: 1 XT-PIC fpu
     14: 1758 XT-PIC ide0
    NMI: 0
    ERR: 0

    Keyboard was functional when I booted the system and logged in. As soon as I
    touched the mouse, the keyboard and mouse locked up. I can log in fine over
    the net.

            J

    -
    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/



    This archive was generated by hypermail 2b29 : Sat Jan 29 2000 - 21:52:29 EST