Re: Oops with 4.9 beta linuxboot ISO and custom-compiled on IBM 8183

Pierre Bourgin <[email protected]> wrote on 05/04/2011 05:46:07 PM:

> Le 02/05/2011 19:04, Timothy J Massey a écrit :
> > But it "functioned" (failed) exactly the same. However, like I said, 
the
> > 4.8 CD works correctly.
> 
> It's rather strange.
> what kind of hard drive do you have ? SATA or IDE/PATA ?

PATA.  The system has both interfaces, but from the factory it includes a 
PATA drive.

> I remember had to deal with ATA/SATA/AHCI BIOS settings, in order to 
> choose the way the IDE/SATA controller had to be "announced" by the 
BIOS.
> 
> Which settings the BIOS offers you related to it ?

Very few:
        Parallel ATA:  Primary/Disabled
        Serial ATA:  Enabled/Disabled
        Native Mode Operation:  Automatic/Serial ATA

> Can you try the different settings around it ?

With a Parallel ATA drive:
Parallel ATA: Primary; Serial ATA: Enabled; Native Mode Operation: 
Automatic = FAIL
Parallel ATA: Primary; Serial ATA: Enabled; Native Mode Operation: Serial 
ATA = FAIL
Parallel ATA: Primary; Serial ATA: Disabled; (Native Mode Operation 
disappears) = FAIL

For the fun of it, I decided to replace the PATA drive with a SATA drive 
in the system to see what happens.  I still can't disable the Parallel 
ATA:  my CD-ROM is attached there.

With a Serial ATA drive, there are fewer options:
Parallel ATA: Primary; Serial ATA: Enabled; Native Mode Operation: 
Automatic = FAIL
Parallel ATA: Primary; Serial ATA: Enabled; Native Mode Operation: Serial 
ATA = SUCCESS

So, it seems that there is a problem with whatever module is being used 
for talking to hard drives via Parallel ATA (even with SATA, when it's 
emulating a parallel ATA), but not (and only not) with native SATA hard 
drives.  I wonder why it didn't choke when still dealing with the PATA 
CD-ROM?

Affordably replacing the drives on several dozen old (but still needed) 
machines is not practical, so switching to SATA-only on these machines is 
not really a solution.  I will have to find a better solution.


So, I started experimenting with IDE/ATA modules.  First, I modified 
linuxboot/state1/etc/master and added a "read PAUSE" right after the line 
that contains "*** Excellent.  Now, let's see about mass storage 
controllers...".  I could then load whatever modules I wanted by hand.  I 
then rebuilt and burned the CD.

Once I did this, it paused before the storage modules were loaded.  I left 
it paused and switched to console 2.  I then did "modprobe ide-generic 
probe_mask=0x1f".  My PATA hard drive and CD-ROM were recognized!  Yay!  I 
then retured to console 1 and pressed <ENTER> to continue with the master 
script.  However, it continued to load the rest of the storage modules and 
crashed once again on "piix".  (Much debugging effort here snipped, but 
note that ata_piix also crashes in the same way.)

So, I then rebooted and before pressing <ENTER> (on my pause before 
loading storage modules) I switched to console 2 and deleted /lib/modules/ 
... /ata/ata_piix.ko and /lib/modules/ ... /ide/piix.ko.  I then did a 
manual "modprobe ide-generic probe_mask=0x1f" and returned to console 1 
and pressed <ENTER> to continue the master script.

Success.  I was able to continue all the way through the nt5x script and 
into the first reboot (the text-mode Windows kernel-based installer).  I 
didn't wait for that to finish, but I'm reasonably confident that this 
will work properly.

So, it seems that this system is having a problem with ata_piix.ko and 
piix.ko.  I don't know if this is just an Unattended boot CD thing or a 
generic kernel thing.  I am downloading some recent Linux live CD's 
(Ubuntu 11.04 and Fedora 15 beta) to see if they will work properly, and 
with what storage modules.


For now, it seems I have two workarounds:  1) Use the Unattended 4.8 boot 
CD.  2) Modify my compiled Unattended CD to delete the piix.ko and 
ata_piix.ko and get it to load the ide-generic with the proper probe_mask 
parameter (without the parameter it doesn't work).

Any suggestions on how best to do that within the Unattended linuxboot 
build process?  I can hack things terribly (I'd probably just throw the 
commands that I'm typing by hand into the master script right where I put 
the pause), but is there a better, cleaner way to do this?

Timothy J. Massey

 
Out of the Box Solutions, Inc. 
Creative IT Solutions Made Simple!
http://www.OutOfTheBoxSolutions.com
[email protected] 
 
22108 Harper Ave.
St. Clair Shores, MI 48080
Office: (800)750-4OBS (4627)
Cell: (586)945-8796

------------------------------------------------------------------------------
WhatsUp Gold - Download Free Network Management Software
The most intuitive, comprehensive, and cost-effective network 
management toolset available today.  Delivers lowest initial 
acquisition cost and overall TCO of any competing solution.
http://p.sf.net/sfu/whatsupgold-sd

_______________________________________________
unattended-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/unattended-devel