Hello readers,

I've posted this to linux-kernel because I thought that it was more
appropriate, but it seemingly has got lost in the noise.

I'll repost here, hoping for a clue. It's probably not specific to TSL
though. I also realise that's a pretty dirty hack over TSL 2.2.
Unfortunately I can't give TSL 3.0 a try on this box: it will be used
for full production work in a few days and our production standard O/S
is TSL 2.2. Kernel 2.6 was tried out to improve NFS client performance
(it's client of a NetApp NAS). No kind of "support" on this problem
expected of course, but just in case that rings a bell, here it is:

This is an IBM xSeries dual-P3 1.0Ghz server box with an IBM ServeRAID
4Mx controller and 3 * IBM 18 Gb SCSI hotswap disks.
Base O/S is Trustix 2.2, manually updated with a 2.6.13 kernel built
from source and a mkinitrd-4.1.18-2 RPM
stolen from FC3 because TSL 2.2's mkinitrd can't handle .ko modules.

After upgrading to kernel 2.6.13, the machine stops at boot during
initrd processing. After much head-scratching I've figured out that
the "mkdevices /dev" command in the "init" script of the initrd was
looping.
Hand-editing the init script like this:

echo "Loading ips.ko module"
insmod /lib/ips.ko
sleep 10 <-------- this line added
echo Creating block devices
mkdevices /dev
and rebuilding the initrd makes the problem go away. So it's really
a timing issue. Everything is fine too if I build the kernel with
ips and scsi modules statically linked.

When failing, console output goes like (approx., typed by hand):

Loading scsi_mod.ko module
SCSI subsystem initialized
Loading sd_mod.ko module
Loading ips.ko module
input: AT Translated Set 2 keyboard on isa0060/serio0
scsi0 : IBM PCI ServeRAID 7.12.02 Build 761 <ServeRAID 4Mx>
Vendor: IBM Model: SERVERAID Rev: 1.00
Type: Direct-Access ANSI SCSI revision: 02
SCSI device sda: 71094272 512-byte hdwr sectors (36400 MB)
sda: got wrong page
sda: assuming drive cache: write through
SCSI device sda: 71094272 512-byte hdwr sectors (36400 MB)
sda: got wrong page
sda: assuming drive cache: write through
sda: sda1 sda2 sda3 sda4 < sda5 sda6 sda7 sda8 >
Attached scsi disk sda at scsi0, channel 0, id 0, lun 0
Vendor: IBM Model: SERVERAID Rev: 1.00
Type: Processor ANSI SCSI revision: 02
Vendor: IBM Model: CaVv3 S2 Rev: 0
Type: Processor ANSI SCSI revision: 02
Creating block devices
input: ImExPS/2 Generic Explorer Mouse on isa0060/serio1

(stops here... nothing happens anymore, system is *not* hung
though, typed characters echo, Ctrl-Alt-Del reboots)

Can anyone provide a hint here? I'm really stuck.
Of course I have two possible work-arounds, but I'd really like to
understand.

Greets,
_Alain_
_______________________________________________
tsl-discuss mailing list
[email protected]
http://lists.trustix.org/mailman/listinfo/tsl-discuss

Reply via email to