Jump to content
Compatible Support Forums

martouf

Members
  • Content count

    335
  • Joined

  • Last visited

    Never

Posts posted by martouf


  1. hmm..

     

    [size:4][tt]$ dc

    705822926 705817760 - p

    5166[/tt][/color] timer interrupts processed.

     

    Not that many other interrupts between the two interrupt count samples.

    Guess this machine is networked via ndiswrapper, because eth0 didn't have any interrupts.

     

    what version of ndiswrapper are you using?

     

    where is the display driver? it should be shown hooked into one of the interrupt lines.

     

    thought you said the machine runs slowly when X is running and not

    so badly otherwise...

     

    would you provide the output of [size:4][tt]free[/tt][/color] ?

    we'll be able to see how the system memory (RAM + swap) is configured.


  2. with two DOS (fat|vfat) formatted floppies in hand (best if full formatted

    by the destination) --

     

    disassembly at the Linux host begins with:

     

    [size:4][tt]dd if=nameofthedriver.sys of=a.bin bs=1024 count=1000

    dd if=nameofthedriver.sys of=b.bin bs=1024 skip=1000[/tt][/color]

     

     

    reassembly at the Win98 destination will involve something like:

     

    [size:4][tt]copy a.bin /b + b.bin /b nameofthedriver.sys /b[/tt][/color]

    (it might work without all the '/b' params designating the 'binary-ness'

    of the files. the paranoia about proper byte handling is worth the extra

    typing, though)

     

     

    easy enough, yes?


  3. so let me see if I've got this right..

     

    Your Win98 box has a static IP address (DHCP disabled) and

    your WinXP box has both DHCP and Connection Sharing enabled.

     

    Does the device at IP 192.168.1.1 (which is also the DHCP server)

    know the Win98 box has a static address? to ensure it won't ever

    issue IP 192.168.1.104 ? (you can't have to boxes with the same IP)


  4. doing your homework pays off and right now it really shows!

     

    /boot will contain the bootloader [lilo/grub] and the kernel(s) [which includes

    the initial ramdisk (initrd), kernel symbol table, kernel config, and the kernel itself]

     

    I'd make /boot about 80MB.

     

    The rest of your disk layout plan is just fine as is. Wouldn't change it.

     

    Keep in mind that because you have already used 2 of the 4 'primary' partitions

    on HD2, that the new partitions must all three be part of an 'extended'

    partition encompassing all 8GB you have set aside.

     

    Also keep in mind the system boots off of HD1 and because you want to dual boot

    you need to ensure you install the bootloader in the MBR of HD1.

     

    clear?

     


  5. when building xine, many times it is necessary to install

    the *-devel version of a library package because the xine

    configure script is expecting to find a special configuration

    script provided by the *-devel package.

     

    in other words, try installing the libpng-devel package

    and try again..

     


  6. did you by any chance actually make the WinXP emergency recovery disks (ERD)?

    umm, before you modified the disk partitions with Partition Magic?

     

    I'm thinking your partition table was changed and the disk geometry it now

    has is not what it had when WinXP could run.. autochk is a vital and necessary

    part of WinXP start up. If it's 'missing', then something is quite wrong.

     

    You did the right thing to use Partition Magic to set up the new disk layout,

    but I'm thinking the Ubuntu installation updated the partition table.

    That would have unfortunately undone all of your efforts to preserve your

    disk geometry.

     

    Did you have the option to 'use existing partition' when setting up Ubuntu?

    ..and you made sure to select it?

     


  7. hello again .. I see you still are reporting the same problem with ymessenger.

     

    Perhaps you would care to read all of the responses you have already received

    in your previous two threads regarding this problem?

     

    Please read this one first:

    http://linuxcompatible.org/thread31288-1.html

     

    and here is the second prior thread, but I'm afraid it only poses a question

    to you which has as yet not been answered:

    http://www.linuxcompatible.org/thread31167-1.html

     

    In fact, if you care to add a response regarding your problem, it probably

    would be best to add it to Thread#31288 above..

     

    Important question for you:

    Have you tried uninstalling the 'phantom' ymessenger-1.0.4-1 package

    with "rpm -e ymessenger-1.0.4-1"?

     

     

    Also, is there a reason you don't provide a subject title when you create

    a new thread? You seem to like using the default - "Untitled thread".

     

     

    cheers!


  8. you don't appear to have a boot problem.

    you seem to have a WinXP start-up problem.

     

    can you boot WinXP in Safe mode?

     

    you'll need to type F8 just after selecting either Windows grub entry

    and before the WinXP "green caterpillar" appears. Choose "Safe Mode".

     

    The equivalent registry keys for the ones mentioned in the referenced

    Microsoft KB317189 are:

     

    (Note that in WinXP the 'hive' loading and unloading is unnecessary)

     

    1. for finding the current control set (HKLM == HKEY_LOCAL_MACHINE)

    - HKLM\SYSTEM\Select

    - check value of Current

     

    2. for finding pending file rename operations

    - HKLM\SYSTEM\ControlSetNNN\Control\Session Manager\FileRenameOperations

    - (where NNN is the three digit leading zero decimal value of 'Current' above)

     

    3. for finding the RunOnce and RunOnceEx entries

    - HKLM\SOFTWARE\Microsoft\Windows\CurrentVersion\RunOnce

    - HKLM\SOFTWARE\Microsoft\Windows\CurrentVersion\RunOnceEx

     

    It seems a software upgrade or OS patch install of yours has gone awry.

     

    Standard warnings about editing the Windows Registry apply, but then

    your system doesn't start up WinXP properly, now does it?

    Can't make it too much worse... ;(


  9. can anyone confirm you obtain the same result if you download the kernel

    update rpm and then run the 'rpm -Uhv' manually?

     

    For those of you who got the update via up2date, the kernel update rpm

    should be on your disk already somewhere in /var (I forget exactly):

    'find /var -name kernel\*.rpm'

     


  10. saving a copy of an MBR is really easy..

     

    (as root)

    dd if=/dev/hda of=extraspecialoriginalMBR.bin bs=512 count=1

    [if == input file, of == output file, bs == block size]

     

    (check contents with)

    od -tx1 -Ax extraspecialoriginalMBR.bin

    [you'll observe: first byte == 'eb', last two bytes == '55', 'aa']

     

    (check identity with)

    strings extraspecialoriginalMBR.bin

     

     

    got a thumb drive and a LiveCD for disaster recovery handy? wink

     

     

    as for drive name/identity - it all has to do with the order the disks

    are found as the initialization probes sequentially through each interface.

    They are assigned names in the order they are discovered by the probes.

    The probe order isn't affected by the boot order, though.

     

     

    cheers!

     


  11. Windows being the stick-in-the-mud that it is, you must make the WinME

    drive the master (making it "C:" in Windows and hda in Linux).

     

    Before someone posts a rebuttal: it's a must unless you're willing to start

    over with installing WinME from scratch on the slave ("D:" aka hdb).

     

    Danleff has been trying to point you in the right direction - if you want to have

    WinME on the master and Linux on the slave and dual boot, then you'll

    have to ensure that grub installs itself in the MBR of hda (aka "C:").

     

    I realize as I write this the multiple naming is probably confusing.

    Sorry.. hda [Linux] == hd0 [grub] == "C:" [WinME] <-- does that help?

     


  12. short answer: sneakernet, localareanet or another OS.

     

    sneakernet ::= carry files to the linux host on removable media

    (thumb drive, CD, Zip)

     

    localareanet ::= grab files on another machine and share them over a 'room area' network.

     

    another OS ::= boot another OS (WinXP, Win2k, etc) on the same machine

    and grab the files you want, then boot Linux and mount the Win filesystem.

     

    you've got your choice of new kernel- and kernel-doc- rpms here:

    http://download.fedora.redhat.com/pub/fedora/linux/core/updates/3/i386/

     

    check for other rpms you'll need to get using "rpm -q --requires -p name-of-kernel.rpm"

    (look for lines like "package-name = version.string" in the output)

     

    once you've got the collection of rpms together: "rpm -Uhv *.rpm"

     


  13. I took another look at the ATrpms collection, and maybe you'd like to give

    this a try:

     

    - first determine the precise kernel version you've got: "uname -r"

    - then fetch the appropriate madwifi-kmdl-`uname -r` ..

     

    for instance, if "uname -r" == 2.6.9-1.724 and you have a single processor

    Pentium 4 system, then you'll want to fetch

    -> madwifi-kmdl-2.6.9-1.724_FC3-0.9.4.12-15.rhfc3.at.i686.rpm

     

    check your processor type with "uname -p" (probably "i686" ) and

    note 'FC3smp' only applies to a multiprocessor system.

     

    In the other thread, I didn't mean to say one should pick the first rpm

    in the ATrpms list. I meant only that one shouldn't pick either of the last two

    in the list. smile

×