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

Juanjo Pablos <[email protected]> wrote on 05/05/2011 02:41:29 PM:

> > 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?
> 
> A few hours ago I upload 2.6.38.5 here:
> https://sourceforge.net/projects/unattended/files/unattended/49rc5/

No joy: same failure.  So I get the same failure with kernels compiled by 
both you and I, and with various different kernel versions from 2.6.35 to 
2.6.38.

Also for the record:  Ubuntu 11.04 (Kernel 2.6.38.8) loads just fine on 
this system.  Doing a "dmesg | grep ata" tells me that this is using 
libata version 3.0 and ata_piix version 2.13, same as all of the 
Unattended kernels I've tried.  Fedora 15 Beta also boots correctly: 
kernel 2.6.38.2-9, libata 3.0 and ata_piix 2.13.

So this does not seem to be a generic kernel bug.  Ubuntu and Fedora 
kernels work fine with libata and ata_piix.  It has something to do with 
the way the Unattended kernel is doing it.


To continue moving forward, I have hacked linuxboot/misc/master to delete 
ata_piix.ko and piix.ko, and manually modprobe ide-generic probe_mask=0x1f 
(at the same point where I previously put in the pause).  This works 
properly on this machine.

Now to test it on other machines.  The first machine I tested it on was an 
IBM 8215.  Guess what?  Same kernel oops, but *different* module!  This 
time it's the Intel Ethernet driver (e1000) that crashes.  So, this 
doesn't even seem to be related to IDE:  I haven't *got* to the point of 
loading storage modules!

As a sanity check, I booted the new Unattended 4.9rc5 disc on this machine 
(the 8215):  same Oops.  Here's what I get from dmesg (unmodified 
Unattended 4.9rc5):

e1000e: Intel(R) PRO/1000 Network Driver - 1.2.20-k2
e1000e: Copyright(c) 1999 - 2011 Intel Corporation.
e1000e 0000:02:00.0: Disabling ASPM  L1
BUG: unable to handle kernel paging request at ffff9000
IP: [<c0113136>] ioapic_write_entry+0x3e/0x80
*pde = 003b2067 *pte = 00000000
Oops: 0002 [#1]
last sysfs file: /sys/devices/pci0000:00/0000:00:1c.0/0000:02:00.0/class
Modules linked in: e1000e(+) usbhid ehci_hcd uhci_hcd usbcore


So, it's not just the kernel compiled by me, but yours too.

I am now officially out of ideas.  There is something fundamentally wrong 
with the current family of Unattended kernels that is not solely related 
to mainline kernels (Fedora and Ubuntu kernels work fine with very similar 
versions) that is preventing these machines from booting.

Any other ideas of where we can look and what we can try?  I don't think 
this is going to be fixed by simply choosing different kernel versions. 
There's something about the way the kernels are being built or maybe the 
order of modules loaded that is killing these machines.

I'm open to suggestions...

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