Re: scanimage -L crashes SuSE 6.4

From: Douglas Gilbert (dgilbert@interlog.com)
Date: Sun Apr 16 2000 - 09:37:23 PDT

  • Next message: Oliver Rauch: "Re: linotype-hell is just that!"

    Bernhard Scheffold wrote:
    >
    > On 11-Apr-00 abel deuring wrote:
    > > Bernhard Scheffold wrote:
    > >
    > >
    > >>
    > >> I have gotten rid of the crash by going back to sane-1.0.1 :-(
    > >>
    > >> So the problem doesn't seem to be in the scsi drivers.
    > >>
    > >
    > > There seems to be a complex combination of reasons... Yesterday evening,
    > > I played a little bit with the aha1542 driver, and it seems that the
    > > driver detects a corrupted scatter-gather list (== fragmented buffer)
    > > for the SCSI data. The aha1542 works fine, if the maximum required
    > > buffer size is 32 kB. As soon as the buffer size is increased to 32769
    > > bytes, I get the kernel crash,
    > >
    > > Therefore, you might try to configure sane-1.0.2 with the option
    > > --enable-scsibuffersize=32768
    >
    > That gets rid of the crash. Thank you very much! I've only got another problem
    > with using xsane as a Gimp-plugin: Gimp displays a checkerboard instead of the
    > scanned image data. :-(

    There is a new version of the linux sg driver (sg 2.1.37) on:
    http://www.torque.net/sg

    that addresses this problem (which is a disagreement between sg and
    the aha1542 driver when command length is 0 and scatter gather is
    being used (i.e. when default size > 32 KB)). It fixed the problem
    on my hardware and perhaps those trying it could report if there
    are any problems.

    Since linux kernel 2.2.15 is imminent, this fix probably won't get
    in until lk 2.2.16 .

    Doug Gilbert

    --
    Source code, list archive, and docs: http://www.mostang.com/sane/
    To unsubscribe: echo unsubscribe sane-devel | mail majordomo@mostang.com
    



    This archive was generated by hypermail 2b29 : Sun Apr 16 2000 - 09:35:59 PDT