Ticket #2100 (closed defect: fixed in core)

Opened 12 months ago

Last modified 11 months ago

126: doesn't fully boot

Reported by: elisa42 Owned by: mrmoku
Priority: major Milestone: shr-2012.07
Component: SHR Image Version: SHR-core
Keywords: Cc:

Description

booting stops with:

[ 3.750000] systemd[1]: Failed to create root cgroup hierarchy: Invalid argument
[ 3.760000] systemd[1]: Failed to allocate manager object: Invalid argument

Attachments

OM-GTA02_20131031_220507.jpg Download (1.9 MB) - added by jama 12 months ago.

Change History

comment:1 Changed 12 months ago by elisa42

on gta02

comment:2 Changed 12 months ago by jama

Can you please try with image from not yet closed feed 127?

 http://build.shr-project.org/shr-core-staging/127/images/om-gta02/

comment:3 Changed 12 months ago by elisa42

Same happens with 127

comment:4 Changed 12 months ago by jama

I've tested it on om-gta04 with 127  http://build.shr-project.org/shr-core-staging/127/images/om-gta04/ and it booted, just E17 wasn't started (only plain Xorg without Xsession).

Changed 12 months ago by jama

comment:6 Changed 12 months ago by jama

Corresponding block from my own image..

[    2.455000] mmc1: host does not support reading read-only switch. assuming write-enable.
[    2.470000] glamo-mci glamo-mci.0: powered (vdd = 15) clk: 399kHz div=0 (req: 400kHz). Bus width=0
[    2.480000] glamo-mci glamo-mci.0: powered (vdd = 15) clk: 16373kHz div=0 (req: 17000kHz). Bus width=0
[    2.490000] glamo-mci glamo-mci.0: powered (vdd = 15) clk: 16373kHz div=0 (req: 17000kHz). Bus width=2
[    2.495000] mmc1: new high speed SDHC card at address aaaa
[    2.500000] mmcblk0: mmc1:aaaa SU08G 7.60 GiB 
[    2.510000]  mmcblk0: p1 p2 p3 p4
[    2.605000] async_waiting @ 1
[    2.610000] async_continuing @ 1 after 6 usec
[    2.635000] EXT2-fs (mmcblk0p1): warning: mounting unchecked fs, running e2fsck is recommended
[    2.650000] VFS: Mounted root (ext2 filesystem) on device 179:1.
[    2.660000] devtmpfs: mounted
[    2.665000] async_waiting @ 1
[    2.670000] async_continuing @ 1 after 6 usec
[    2.675000] Freeing init memory: 124K
[    2.690000] Failed to execute /usr/sbin/bootchartd.  Attempting defaults...
[    3.775000] systemd[1]: systemd 206 running in system mode. (+PAM +LIBWRAP -AUDIT -SELINUX +IMA +SYSVINIT -LIBCRYPTSETUP -GCRYPT +ACL +XZ)
[    3.880000] calling  init_autofs4_fs+0x0/0x20 [autofs4] @ 1
[    3.885000] initcall init_autofs4_fs+0x0/0x20 [autofs4] returned 0 after 726 usecs
[    3.890000] systemd[1]: Inserted module 'autofs4'
[    4.050000] calling  inet6_init+0x0/0x274 [ipv6] @ 1
[    4.055000] NET: Registered protocol family 10
[    4.060000] initcall inet6_init+0x0/0x274 [ipv6] returned 0 after 7285 usecs
[    4.065000] systemd[1]: Inserted module 'ipv6'
[    4.075000] systemd[1]: Set hostname to <gjama>.
[    4.600000] pcf50633 0-0073: usb curlim to 1000 mA
[    4.965000] systemd[1]: Cannot add dependency job for unit uuidd.socket, ignoring: Unit uuidd.socket failed to load: No such file or directory. See system logs and 'systemctl status uuidd.socket' for details.
[    4.975000] systemd[1]: Cannot add dependency job for unit xinput-calibrator.service, ignoring: Unit xinput-calibrator.service failed to load: No such file or directory. See system logs and 'systemctl status xinput-calibrator.service' for details.
[    4.985000] systemd[1]: Expecting device dev-ttySAC2.device...
[    4.990000] systemd[1]: Starting Forward Password Requests to Wall Directory Watch.
[    4.995000] systemd[1]: Started Forward Password Requests to Wall Directory Watch.
[    5.000000] systemd[1]: Starting Syslog Socket.
[    5.005000] systemd[1]: Socket service syslog.service not loaded, refusing.
[    5.010000] systemd[1]: Failed to listen on Syslog Socket.
[    5.015000] systemd[1]: Starting Remote File Systems.
[    5.020000] systemd[1]: Reached target Remote File Systems.
[    5.030000] systemd[1]: Starting Delayed Shutdown Socket.
[    5.035000] systemd[1]: Listening on Delayed Shutdown Socket.
[    5.040000] systemd[1]: Starting /dev/initctl Compatibility Named Pipe.
[    5.045000] systemd[1]: Listening on /dev/initctl Compatibility Named Pipe.
[    5.050000] systemd[1]: Starting Root Slice.
[    5.055000] systemd[1]: Created slice Root Slice.
[    5.060000] systemd[1]: Starting User and Session Slice.
[    5.065000] systemd[1]: Created slice User and Session Slice.
[    5.075000] systemd[1]: Starting System Slice.

comment:7 Changed 12 months ago by jama

Sigh, I've completely forgotten that in my builds I'm using newer kernel:
-rw-r--r-- 1 root root 1481328 Nov 1 15:21 uImage-2.6.39.4
-rw-r--r-- 1 root root 1536480 Sep 17 22:43 uImage-3.2.18-my

Now it makes sense, I'm going to backport
 http://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=676db4af043014e852f67ba0349dae0071bd11f3
referenced in:
 https://wiki.ubuntu.com/systemd#Kernel_requirements

comment:8 Changed 12 months ago by jama

Ah we have that one already.

comment:9 Changed 12 months ago by jama

Systemd docs were updated to require 3.0 kernel with cgroup.procs support
 http://cgit.freedesktop.org/systemd/systemd/commit/README?id=e946948eff517e895b287d0fd8c6d069ab9bbbb9

To backport cgroup.procs we would need to cherry-pick following commits
8f3ff20862cfcb85500a2bb55ee64622bd59fd0c
102a775e3647628727ae83a9a6abf0564c3ca7cb
72a8cb30d10d4041c455a7054607a7d519167c87
c378369d8b4fa516ff2b1e79c3eded4e0e955ebb
be367d09927023d081f9199665c8500f69f14d22
f780bdb7c1c73009cb57adcf99ef50027d80bf3c
74a1166dfe1135dcc168d35fa5261aa7e087011b
d846687d7f84e45f23ecf3846dbb43312a1206dd

They don't apply cleanly, but looks doable, but I'm not sure if it's worth it, I'm using kernel 3.2 on gta02 for long time (since January 2012 or something like that), it's not without issues, but time spent on fixing 3.2 would be more valuable than spending it on hacking systemd support to 2.6.39 (until some other change will require even newer kernel again). What do you think?

comment:10 Changed 12 months ago by Thamos0815

Why do we need new systemd? I would postpone systemd update a while to support dos efforts to make a solid release of shr.
This is in relation to the issues with kernel 3.2, if they are not much, i would give it a try. Backporting is only an option, if we relly need new systemd in a solid shr-release and the new kernel makes too much troubles.

comment:11 Changed 12 months ago by jama

It's part of newer oe-core to which we upgraded lately, in order to get older systemd back I would need to add it into our layer and downgrade it in feeds.

comment:12 Changed 11 months ago by jama

Sorry for delay, my current plan is to downgrade to older systemd in current builds (based on dora release) and upgrade to 3.2 in master branch (for next release) - so that we don't drag old systemd forever.

comment:13 Changed 11 months ago by jama

Older systemd included in staging 128, tested on om-gta02 that it now boots OK.

comment:14 Changed 11 months ago by elisa42

  • Status changed from new to closed
  • Resolution set to fixed in core

works here too, thanks

comment:15 Changed 11 months ago by jama

BTW:
3.12 is now default kernel for om-gta04 in next release (shr branch)
3.2 for om-gta02 in next release (which doesn't have downgraded systemd)

More testing from would be great, I wonder if I should revive  http://build.shr-project.org/shr-kernels/ and upload newer kernels there first, what do you think?

comment:16 Changed 11 months ago by elisa42

I tested 3.2 a few days ago in om-gta02 and it has even more problems than 2.6.39. 2.6.34 is still the best choice for om-gta02. Unfortunatly not supported with systemd.

I'd really like to do some kernelwork as a more uptodate and stable kernel for om-gta02 would realy contribute to the usability of om-gta02 with shr. But I fear it's a going to be alot of work and I'll need to learn alot for that.

Testing kernel wouldn't really help until someone is doing work on them.

3.12 sounds great for om-gta04.

Note: See TracTickets for help on using tickets.