Bug #2557

stock 3.4.1 kernel halts during booting if dm and dm_target_crypt are loaded and RAID controller is present

Added by phma about 1 year ago.

Status:NewStart date:05/12/2013
Priority:NormalDue date:
Assignee:-% Done:

0%

Category:-
Target version:-

Description

Two computers: darner and zyxomma. Darner is 32-bit running 3.5.0. Zyxomma is 64-bit running 3.4.1 copied from the thumb drive. Both have an encrypted partition.

Darner's loader.conf:
#vfs.root.mountfrom="hammer:serno/Y2P0C2BE.s1d"
vfs.root.mountfrom="hammer:serno/WD-WCC2ECF38674.s1d"
kern.emergency_intr_enable=1
fred_disable=NO
dm_load=YES
linux_load=YES
#dsched.policy.ad0=fq
#dsched.policy.ad4=fq
dsched.policy.ad0=noop
dsched.policy.ad4=noop
snd_ich_load=YES
hw.snd.pcm0.vchans=4
dm_target_crypt_load=YES
vfs.hammer.skip_redo=1

Zyxomma's loader.conf:
vfs.root.mountfrom="hammer:serno/131061401278.s1d"
#dm_load=YES
#dm_target_crypt_load=YES

If the dm lines are not commented out, and I boot it in verbose mode, I get the following:
FQ scheduler policy version 1.1 loaded
wdog: In-kernel automatic watchdog reset enabled
wlan: <802.11 Link Layer>
kbd: new array size 4
kbd1 at kbdmux0
crypto: <crypto core>
No policy for md0 specified, or policy not found
disk scheduler: set policy of md0 to noop
md0: Malloc disk
md0: invalid primary partition table: no magic
hpt27xx: RocketRAID 27xx controller driver v1.0 (Apr 24 2013 20:01:00)
hptrr: RocketRAID 17xx/2xxx SATA controller driver v1.2

If I comment out the lines, it boots normally, but I have to load the dm module manually before starting cryptdisks, and dntpd doesn't work. I had to mount the crypt partition and run dntpd in debug mode to get it to work.

Also available in: Atom PDF