Gentoo Websites Logo
Go to: Gentoo Home Documentation Forums Lists Bugs Planet Store Wiki Get Gentoo!
Bug 120038 - sys-apps/hal-0.5.5.1-r3 seg fault
Summary: sys-apps/hal-0.5.5.1-r3 seg fault
Status: RESOLVED NEEDINFO
Alias: None
Product: Gentoo Linux
Classification: Unclassified
Component: [OLD] Core system (show other bugs)
Hardware: x86 Linux
: High normal (vote)
Assignee: Project Gentopia
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2006-01-23 04:31 UTC by Jedi
Modified: 2006-05-17 16:01 UTC (History)
1 user (show)

See Also:
Package list:
Runtime testing required: ---


Attachments
hal-0.5.5.1-r3.ebuild.patch (hal-0.5.5.1-r3.ebuild.patch,364 bytes, patch)
2006-01-24 04:27 UTC, Dmitry S. Kulyabov
Details | Diff

Note You need to log in before you can comment on or make changes to this bug.
Description Jedi 2006-01-23 04:31:57 UTC
upgrading to new hal, the service fail to start.
executing as root:

hald  --verbose=yes  --daemon=no

...
13:20:51.289 [I] hald.c:89: Added device to GDL; udi=/org/freedesktop/Hal/devices/platform_i8042_i8042_Kbd_Port
13:20:51.289 [I] classdev.c:1182: class_add: subsys=input sysfs_path=/sys/class/input/input0 dev= physdev=0x080cf2d8
13:20:51.291 [E] util.c:731: Couldn't spawn 'hald-probe-input' err=Failed to execute child process "hald-probe-input" (No such file or directory)!
Segmentation fault

the "hald-probe-input" is 
-rwxr-xr-x  1 root root 6784 Jan 23 13:18 /usr/libexec/hald-probe-input

the problem still exists if I add --retain-privileges option or add  export PATH=$PATH:/usr/libexec/
Comment 1 Jedi 2006-01-23 04:35:18 UTC
if it is linked like
ln -s /usr/libexec/hald-probe-input /sbin/hald-probe-input

works but this operation must be done for all
hal.hotplug           hald-addon-storage    hald-probe-pc-floppy  hald-probe-storage
hald-addon-acpi       hald-addon-usb-csr    hald-probe-printer    hald-probe-volume
hald-addon-hid-ups    hald-probe-hiddev     hald-probe-serial
hald-addon-pmu        hald-probe-input      hald-probe-smbios

it seems like the ebuild mismatch the target directory for those files
Comment 2 Dmitry S. Kulyabov 2006-01-24 04:27:03 UTC
Created attachment 77986 [details, diff]
hal-0.5.5.1-r3.ebuild.patch

This is dirty workaround.
Comment 3 Jedi 2006-01-24 12:47:56 UTC
thanks for the patch.

now hald go on up to the following error

..
21:40:51.821 [I] acpi.c:1037: Add callouts completed udi=/org/freedesktop/Hal/devices/acpi_C13921:40:51.821 [I] hald.c:89: Added device to GDL; udi=/org/freedesktop/Hal/devices/acpi_C139
21:40:51.821 [I] hald.c:653: Device probing completed
21:40:51.821 [I] hald_dbus.c:3063: entering
21:40:51.821 [E] hald_dbus.c:3070: dbus_bus_get(): Failed to connect to socket /var/run/dbus/system_bus_socket: No such file or directory
21:40:51.821 [I] util.c:1262: Killing helper with pid 1676
21:40:51.821 [I] util.c:1262: Killing helper with pid 1671
21:40:51.821 [I] util.c:1262: Killing helper with pid 1661

dbus service is started and ok, the problem is that the virtual path of the socket is:

/var/lib/dbus/system_bus_socket
and not
/var/run/dbus/system_bus_socket


maybe would be better mask this version of hal.

can I safely try a newer version?
Comment 4 Jedi 2006-01-25 01:32:15 UTC
I tried with 0.5.6, same behavior..
Comment 5 Jedi 2006-01-26 00:05:47 UTC
no solutions?
for let work hal again should I downgrade the whole gnome !
Comment 6 Dmitry S. Kulyabov 2006-01-26 02:52:11 UTC
> ..
> 21:40:51.821 [I] acpi.c:1037: Add callouts completed
> udi=/org/freedesktop/Hal/devices/acpi_C13921:40:51.821 [I] hald.c:89: Added
> device to GDL; udi=/org/freedesktop/Hal/devices/acpi_C139
> 21:40:51.821 [I] hald.c:653: Device probing completed
> 21:40:51.821 [I] hald_dbus.c:3063: entering
> 21:40:51.821 [E] hald_dbus.c:3070: dbus_bus_get(): Failed to connect to socket
> /var/run/dbus/system_bus_socket: No such file or directory
> 21:40:51.821 [I] util.c:1262: Killing helper with pid 1676
> 21:40:51.821 [I] util.c:1262: Killing helper with pid 1671
> 21:40:51.821 [I] util.c:1262: Killing helper with pid 1661

My solution (maybe wrong, but it's working fo me):
emerge =dbus-0.60*
emerge hal ivman pmount 
/etc/init.d/dbus restart
Comment 7 Jedi 2006-01-26 12:11:48 UTC
can you attach your /etc/dbus-1/system.conf
and the log of hald  --verbose=yes  --daemon=no after you stopped the hald service

what I cannot understand is why in your system exec files in /usr/libexec works.
And why dbus creates the socket in a path that is not compatible with hal: who is the guilty dbus or hal?
I never touched dbus config files, I remember only that upgrading to new dbus and hal the some configurations are changed and I take all the new ones

thanks

P.S.
I already have installed
Calculating dependencies ...done!
[ebuild   R   ] sys-apps/hal-0.5.5.1-r3  +acpi -debug +doc* -pam_console +pcmcia (-selinux) 0 kB
[ebuild   R   ] sys-apps/ivman-0.6.8  -debug 0 kB
[ebuild   R   ] sys-apps/pmount-0.9.6  +crypt 0 kB
[ebuild   R   ] sys-apps/dbus-0.60-r3  +X -debug +doc -gcj +gtk -mono +python +qt (-selinux) +xml2 0 kB
Comment 8 Jedi 2006-01-27 03:48:30 UTC
one guilty is found:

a version of hal-0.5.5.1-r3.ebuild
creates in /etc/dbus-1/system.conf the listen tag with value
/var/lib/dbus/system_bus_socket
instead of
/var/run/dbus/system_bus_socket
using the  --with-system-socket flag by build

the current one is correct, but seems not update correctly the configuration:
i.e. I modified manually /etc/dbus-1/system.conf, then emerge dbus
the emerge works fine but not creates any /etc/dbus-1/._cfg0000_system.conf 

I suppose is a bug.

now remains the second question: the /usr/libexec in hal ....

feedbacks are wellcome :-)
Comment 9 Jedi 2006-01-30 01:40:02 UTC
I cannot understand: is a bug or not?

They are answers on the questions of above?
Comment 10 Doug Goldstein (RETIRED) gentoo-dev 2006-03-19 07:04:58 UTC
Please test again with hal-0.5.7. Something is borked in your configs.
Comment 11 Steev Klimaszewski (RETIRED) gentoo-dev 2006-05-17 16:01:42 UTC
No response from the submitter for 2 months