problems waking up from S3

From: Juha Jäykkä (juhaj_at_iki.fi)
Date: 11/30/04

  • Next message: Noah Durell: "Re: CUPS Server Error" 
    Date: Tue, 30 Nov 2004 16:26:59 +0200
To: debian-user@lists.debian.org

    First, some background. I have two boxes, that I want to put into ACPI S3
    when they're not in use. They are running kernels 2.6.8 and 2.6.9 (custom,
    not pre-packaged Debian), with all the relevant ACPI options turned on (or
    as modules and loaded). Suspending works fine on both machines. Waking up
    creates some problems. Let me now introduce the systems.

    The other is a VIA EPIA C3 -based box and the other an elderly ASUS A7V
    with an Athlon (thoroughbred) CPU.[1] On resume, both systems have
    problems with input devices. The EPIA's keyboard (PS/2) dies and A7V's
    mouse dies. The mouse is not a problem: reloading the USB modules fixes it
    and I presume reloading the keyboard modules would fix the keyboard, but I
    haven't tried that for two reasons. First, I do not use the keyboard
    anyway (in normal use, I use an IR RCU, naturally). Second, the hard disc
    does not restart, so S3 is useless anyway. This hard disc failure is the
    main problem, it happens on BOTH SYSTEMS.[2]

    The symptoms

    # echo -n mem > /sys/power/state
    [system goes to sleep]
    [resume with power button]
    [system resumes]
    [fix the mouse/keyboard with rmmod/insmod]
    # ls -l
    <hang>

    I attached the lines kernel wrote into syslog until it no longer could
    write to it (write-cache full, I presume). Anyway, nothing disc-related
    works afterwards.

    Note a strange time-discontinuity at Nov 30 15:27:11. It jumps an hour and
    11 minutes forward for 3 lines and then returns to 15:27:12! This *ALWAYS*
    happens, just before a device reset, so there must be something
    conspicious going on.

    I also tried to use hdparm -Y to put the disc to sleep and this seems to
    work *once*, but only once. With a second sleep-cycle, the results are
    exactly the same as with S3. I doubt this is a disc- or controller-related
    since both the EPIA and the A7V have different BIOSes, discs, chipsets and
    controllers, aged ~three years apart. I can dig out those details if
    necessary, but at the moment I cannot (due to the discs being frozen and
    lspci not in cache).

    Any ideas would be appreciated - as would be a cc directly to me since I'm
    not on the list (any more). Oh, the A7V system is sid, EPIA is sarge.
    Thanks.

    [1] Anyone trying to suggest I buy silenced cases for them, forget it. I'd
    rather keep the cases and use S3 than buy stuff, since silent cases cost a
    lot, the HDD in the Athlon box is noisy and the Athlon box is going to be
    relocated soon (6 months or so) anyway. Lastly, the fanless EPIA ITX cases
    which can accommodate a Technotrend DVB-C card are (as far as I know)
    non-existent.

    [2] But not on my IBM Thinkpad T40, though it has its own problems: it has
    no DMA on its hard disc, it had with kernel <2.6.4, but never since. I
    haven't investigated this too much yet, but perhaps some of you know what
    to do? I haven't yet tested ide0=dma parameter to kernel (or ide.ko) since
    I never boot the machine, just put it to S4. A working S4 is also the
    reason I gave up 2.6.3 - S4 is (in a laptop) more important than DMA.


    --
                     -----------------------------------------------
                    | Juha Jäykkä, juolja@utu.fi |
                    | Laboratory of Theoretical Physics |
                    | Department of Physics, University of Turku |
                    | home: http://www.utu.fi/~juolja/ |
                     ----------------------------------------------- 

    


    


    -- 
To UNSUBSCRIBE, email to debian-user-REQUEST@lists.debian.org 
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
  • Next message: Noah Durell: "Re: CUPS Server Error"