FS#5253 - mkinitcpio: cpio based image loads some modules in reversed order than initrd

Attached to Project: Arch Linux
Opened by vladimir (prhlava) - Saturday, 19 August 2006, 18:29 GMT
Task Type Bug Report
Category System
Status Closed
Assigned To No-one
Architecture not specified
Severity Low
Priority Normal
Reported Version 0.7.2 Gimmick
Due in Version Undecided
Due Date Undecided
Percent Complete 100%
Votes 0
Private No

Details

on a server with two disk controllers (sil3114 - sata, and adaptec 29160N - scsi):
initrd: adaptec starts first, sil3114 second
initcpio: sil3114 starts first, adaptec second

(using the same 2.6.17-ARCH kernel in both cases)

the order of /etc/mkinitcpio.conf:HOOKS is scsi, then sata (i tried with and without autodetect).

i have a raid5 array on the sil3114 controller (initialised by /etc/rc.sysinit), it is working under initrd and not under initcpio. i could (maybe) solve this by building my own kernel (if the order of kernel device initialisation is the same as of the initrd, but it would be nice if i could use the supplied kernel).
This task depends upon

Closed by  Tobias Powalowski (tpowa)
Tuesday, 22 August 2006, 06:14 GMT
Reason for closing:  Not a bug
Comment by vladimir (prhlava) - Saturday, 19 August 2006, 19:11 GMT
forgot to add, the reason that the software raid does not work with initcpio is that the scsi disk names are different (due to order the modules are initialised) with initcpio than initrd.
Comment by Tobias Powalowski (tpowa) - Sunday, 20 August 2006, 06:51 GMT
you can easily solve this issue by specifying the correct order in MODULES= in mkinitcpio.conf,
or other option would be add a udev rule to initcpio image that loads it correct.
Comment by vladimir (prhlava) - Monday, 21 August 2006, 21:15 GMT
thank you for the suggestion, the MODULES= did the trick. obviously, i should have read the mkinitcpio.conf with greater attention...

Loading...