lkml.org 
[lkml]   [2000]   [Sep]   [22]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
    /
    Date
    From
    SubjectRe: Reproducable hard locks in 2.2.17. IDE related

    Why don't you try my patches ontop and thin see what happens.
    Then you are claiming heavy loads.

    Go grab the smart-suite off of source-forge and see if you crapped out the
    drives.


    On Fri, 22 Sep 2000, Carsten Lang wrote:

    > Hi Alan,
    >
    > > Uncorrectable error is the drive.
    > >
    >
    > i don't think that we can blame the disks!
    > I have 3 different drives, which produce this error after 24 hrs under heavy load.
    >
    > I'm very sure, that the whole IDE-stuff is broken under heavy load conditions, but i
    > can't find out, what happens exactly in this case.

    Yes and if you do not use the latest drivers that are backported, you can
    not recover.

    > > > Sep 21 23:55:44 fs1 kernel: hdb: timeout waiting for DMA
    > > > Sep 21 23:55:44 fs1 kernel: hdb: irq timeout: status=0xd0 { Busy }
    > > > Sep 21 23:55:44 fs1 kernel: hda: DMA disabled

    This is the following problem

    hda: Maxtor 71626 AP, 1554MB w/128kB Cache, CHS=789/64/63, DMA
    hdb: Maxtor 91021U2, 9641MB w/512kB Cache, CHS=1229/255/63, (U)DMA
    hdc: Maxtor 91021U2, 9641MB w/512kB Cache, CHS=19590/16/63, (U)DMA

    We do not set the master drive to be the slower one.

    DMA mode 2 and UDMA mode 2 on a via core in master/slave order can get
    flaky.

    > Perhaps some timeouts are not set correctly or the specs include some errors, but
    > it is very strange behaviour, that with kernel 2.0.36 there was never such a problem
    > with heavy loaded systems and after switching to 2.2.16 / 2.2.17 the system dies
    > after a short time!!!

    > I would like to help if i can, but i'm no kernel developer; all i can do is to provide
    > a very heavy loaded machine (video-sequencing) with three different drives.
    > I hope they haven't been destroyed by the driver, but i'm sure the drivers behaviour
    > let them think they have large amounts of bad blocks.

    Dude if your drives are toast because of what I did and not the stock
    kernel, I will besure that the OEM replaces them.

    If you have junk hardware and under an powered system and high loads cause
    the spindle to slow, this is a problem that was created by not match you
    hardware power needs.

    > BTW: The drives were new! And checked... And ok with 2.0.38!!!!

    Do they do DMA correctly?
    2.0.38 does not handle heavy IO like a patched 2.2 or 2.4, because it
    can't.

    Cheers,

    Andre Hedrick
    The Linux ATA/IDE guy



    -
    To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
    the body of a message to majordomo@vger.kernel.org
    Please read the FAQ at http://www.tux.org/lkml/

    \
     
     \ /
      Last update: 2005-03-22 12:38    [W:2.919 / U:0.036 seconds]
    ©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site