Gujin 2.7 hangs during probing

Benny
2010-02-03
2013-02-08
  • Benny

    Benny - 2010-02-03

    Reporting here a private conversation with Etienne for everybody's convenience.
    My son's (pretty old) desktop is hanging during boot sequence. Below the instructions received by Etienne to create some logging traces and his suggestions after the logs readings:

    Etienne:
    For that kind of problem, my usual method is to run dbgdisk.exe and dbgfs.exe
    found in standard-2.7.tar.gz from a DOS bootable USB created by:
    http://www.4shared.com/file/62979959/e3572fe7/HPUSBFW.html
    Both create the file DBG (so rename in between running dbgdisk
    and dbgfs) when running under DOS and exiting Gujin by Control-D,
    if you send me the two files I will probably find the problem.
    Else there is ways to display a lot of debug information on the screen
    or on the serial port, but it will be more difficult for me to get this
    information (photo of the screen).
    Can you tell me exactly what is the last message displayed,
    is "probing filesystem" printed?

    -->> After log sending:
    Etienne:
    think I know your problem, or in fact your two problems:

    1- Your BIOS reacts badly when you ask to read a sector after the end of the drive,
       you can see that by all the messages:
    DISK_RWsector read disk EBIOS 0x80 8 sector at 0 (of partition 2)
    EXCEPTION/timeout 5 second RECEIVED!
    you have in the logs. In the case of a disk read over the limit,
    your BIOS should just answer an error code and not try
    indefinitely to read that sector. It is fortunate that Gujin has
    a timeout there and that this code works.

    2- Your hard disk powers up with a size limited below 137 Gbytes because long time ago
       a lot of BIOSes did not implement the IDE LBA48 extension and could not boot with
       a disk bigger than 137 Gb. Your BIOS do not extend the size from 268,435,455
    sectors to 312,581,807 sectors, and because Gujin do not probe
    the IDE interface (in that log) Gujin do not extend it neither
    before probing the BIOS disk interface.
    I am not sure if you have the "IDE probing" checkbox ticked or
    not in your real bootloader, it may be different when running
    from DOS like for this log.
    When installed as a main bootloader, if IDE probing checkbox
    is ticked, Gujin do first probe the IDE interface and if the
    disk size is limited, it will remove that limit if and only if
    some partition are defined over that limit. If it does not
    behave like this there is a bug - but last time I checked Gujin
    did handle that perfectly. The Linux kernel will also behave
    the same way, removing the HPA limit when a partition is
    located over that limit.

    Now for the solution: I hope your hard disk is a 5 inch and not
    a 2.5 inch drive. On the back of the drive near the power plug
    of that hard drive, you probably have (usually) 4 small plugs,
    and an explanation of how to put connectors on those plugs on
    the sticker on the top of the drive itself.
    Some of those small plugs will select if the hard disk shall
    behave as a master or a slave, or as "cable select". Usually
    on a PC less than 7 years old you select "cable select".
    Some other combinations will (probably) enable or disable
    a "4096" cylinder limit or a "137 Gb" limit, you have to
    unplug/replug those small black clips like the drawing on the
    sticker (top of the drive) to disable that limit.
    Take care not to enable a "power-up in standby" mode if your
    hard disk has that capacity.
    I would advise you to draw the initial configuration with a
    marker on the top of the drive itself, if that is the first time
    you play with those switches: you will be able to go back
    to the initial configuration at any time in case of problem.

    I know that this solution needs you to remove your hard drive
    from the PC, but it will probably solve all your problems at
    once. After removing and re-inserting a hard drive on a PC,
    always check that the nearby CDROM drive, floppy drive and other
    possible hard drives have their plugs fully fitted…

    Note that I would have liked this kind of messages to be on
    the forum, a lot of people probably have the same problem,
    and I will not answer that kind of long message to everyone

    DONE
    THANKS Etienne

    Benny

     
  • Etienne LORRAIN

    Etienne LORRAIN - 2010-02-04

    Thanks for the summary. By the way, did it work?

    Etienne

     
  • Benny

    Benny - 2010-02-07

    sorry being late… My (20Y) son is studying for examinations and I did not access that pc yet….(actually he doesn't trust me)
    I managed at least to identify the disk on board ST3160021A and find the manual on the web:
    http://ec1.images-amazon.com/media/i3d/01/A/man-migrate/MANUAL000021069.pdf

    I will execute the mission  when my son will be out….

    benny

     
  • Etienne LORRAIN

    Etienne LORRAIN - 2010-02-08

    Looking at page 22 of this doc, the jumper limiting the size of the disk is the closest to the power plug - so you do not need to remove the disk if you can access that jumper.
    But the problem I have is that your disk is limited to 268435455 sectors, i.e. 127 Gb, it full size is 312581807 sectors: the doc says the size limit is for 32 Gb so the jumper is probably already not present.
    I found a problem in Gujin, problem I introduced when testing a hitachi drive (you cannot remove the effects of the size limiting jumper on those drives), that broke other hard disk brands to remove the limit - will be fixed in next version…
    It seems that your disk has a HPA set to 128 Gb and this HPA is restored at power-on because the hard disk has the "keep setup over reset" flag set (see command hdparm).
    As of being late, you are not alone!

     
  • Benny

    Benny - 2010-02-13

    While my son was watching the TV AND I was home I could open it 10 minutes ago: confirmed, only the first jumper is present (Master mode).

     
  • Etienne LORRAIN

    Etienne LORRAIN - 2010-02-22

    It seems that simply this PC BIOS is not able to read hard disks bigger then 132 Gb, and Gujin times out after 5 seconds after the BIOS request. It should not be a fatal error, and after 40 seconds (few reads in Gujin-2.7) or 10 seconds in the later version, the disk analysis should finish correctly. To reduce the time waited on this PC, I would advise to disable BIOS/EBIOS probing, only keeping Gujin IDE probing.

     

Log in to post a comment.

Get latest updates about Open Source Projects, Conferences and News.

Sign up for the SourceForge newsletter:

JavaScript is required for this form.





No, thanks