Bug 203909 – sys-apps/hal-0.5.10 - hald startup takes long time

• Product: Gentoo Linux
• Status: RESOLVED WORKSFORME

[Description Imre Péntek 2007-12-31 17:14:57 UTC]

it takes about one minute for hald to start up when booting my gentoo. I've seen several distros with the same hald issue, but several others (and several other versions of the same distros) starting up hald without a delay.

Reproducible: Always

Steps to Reproduce:
boot your gentoo
 * starting hardware abstraction layer daemon
at this point I have to wait about one minute to see:
 * starting hardware abstraction layer daemon          [OK]

[Comment 1 Jakub Moc (RETIRED) 2008-01-02 10:08:15 UTC]

Without info like emerge --info output, HAL version or relevant part of dmesg, nothing we could debug or fix.

[Comment 2 Imre Péntek 2008-01-22 22:42:19 UTC]

I now use hal-0.5.10

[Comment 3 Imre Péntek 2008-01-22 22:43:54 UTC]

Created attachment 141617 [details]
emerge --info

[Comment 4 Imre Péntek 2008-01-22 22:46:17 UTC]

Created attachment 141619 [details]
dmesg output

[Comment 5 Imre Péntek 2008-01-23 14:30:20 UTC]

I replaced my kernel, and noticed that the same applies for sys-kernel/genkernel-3.4.9 with default config
Linux boborjan 2.6.23-gentoo-r5 #1 SMP Wed Jan 23 13:52:24 CET 2008 x86_64 Mobile AMD Athlon(tm) 64 Processor 3000+ AuthenticAMD GNU/Linux

After that, I also replaced my hal several times, and found that same applies for the following hal versions:
boborjan ~ # grep hal /etc/portage/package.mask 
=sys-apps/hal-0.5.10
=sys-apps/hal-0.5.9.1-r3
=sys-apps/hal-0.5.9.1-r2
=sys-apps/hal-0.5.9.1-r1
=sys-apps/hal-0.5.9-r1
boborjan ~ # 

but sys-apps/hal-0.5.7.1-r5 starts up quickly.

[Comment 6 Imre Péntek 2008-02-17 03:55:10 UTC]

Created attachment 143720 [details]
hald startup log snippet

HALD_VERBOSE="yes"
/var/log/messages snippet
disabling battery module solves the issue.

[Comment 7 Kai Krakow 2008-07-18 20:29:33 UTC]

This happens to me when dbus is not ready fast enough. Inserting a "sleep 3" in /etc/init.d/dbus after the line starting dbus fixes this for me. I'll post a separate bug for this because it only hit me after switching to baselayout-2 and parallel_startup=yes.

[Comment 8 Imre Péntek 2008-07-19 08:00:28 UTC]

I noticed far ago, that the bug isn't present when my kernel doesn't have battery.o (or something like that). Otherwise, dmesg will say my battery voltage is less than the half of the design voltage. btw my battery is in very bad shape if usable at all.

[Comment 9 Tony Vroon (RETIRED) 2009-02-04 19:56:23 UTC]

I must say that I am completely unable to reproduce the HAL/d-bus parallel start interaction on a dual quad-core Opteron 2254 system with 32GB RAM (if that isn't fast enough, I'm not sure what is...)
If this is still an issue with HAL 0.5.11-r7, could you please open a (new) bug and assign it directly to me. Sorry for the lack of response to your bug report.

[Comment 10 B.Rathmann 2009-05-27 19:42:28 UTC]

Created attachment 192635 [details]
dmsg part with HALD_VERBOSE=true

Just a short snippet from /var/log/messages to prove the delay

[Comment 11 B.Rathmann 2009-05-27 19:43:36 UTC]

Created attachment 192636 [details]
Current Kernel Config (2.6.27-gentoo-r8

[Comment 12 B.Rathmann 2009-05-27 19:44:34 UTC]

Created attachment 192637 [details]
output of emerge --info

[Comment 13 B.Rathmann 2009-05-27 19:47:49 UTC]

(In reply to comment #9)
> If this is still an issue with HAL 0.5.11-r7, could you please open a (new) bug
> and assign it directly to me. Sorry for the lack of response to your bug
> report.
> 
Since updating to kernel 2.6.27-gentoo-r8 and hald 0.5.11-r8 I am seeing this kind of delays (30-60 seconds) too. Boot sequence prints "Starting Hardware Abstraction layer..." and just waits. After ~60s it resumes work as expected.

I added a log snippet, my kernel config and emerge --info. If you have any additional questions, just ask me.