determining HAL usefulness

cach0rr0 · Posted: Wed Dec 16, 2009 10:44 am Post subject: determining HAL usefulness

hi folks

SO...I'm wondering if there's any way to determine what all hal is actually doing.

As in, I'd like to look and see what if anything wouldn't be functioning properly without hal in the picture - I am not aware of a way to do this offhand, or well, not one that doesn't require a fair bit of pain.

If it's not doing anything for me, it's considerable bloat.
If it is doing anything for me, I'd be keen to know just what, as I could then make a determination as to whether I could do the same without hal relatively easily, or if it'd be too painful.

I did find HALD_VERBOSE="yes" on the wiki, haven't tried it yet, but wondered if anyone knew of a relatively painless way to check out what hal is doing and see if it's even needed.

thanks in advance

aCOSwt · Posted: Wed Dec 16, 2009 10:59 am Post subject:

Hello,

As you certainly know this, HAL stands for Hardware Abstraction Layer.
Being said that I do not know how things are achieved under Linux, the only thing I can tell you about its utility refers to FreeBSD.
Under FreeBSD, hal is compulsory in order to enable KDE / GNOME automounting of removable devices.
It could well be the same under Linux.
http://docs.kde.org/stable/en/kdebase-runtime/userguide/multimedia.html

EDIT : I additionally remember that the last installation I did of the nVidia driver for FreeBSD also requested the hal daemon to run (I never investigated what for)

cach0rr0 · Posted: Wed Dec 16, 2009 11:19 am Post subject:

Just as a test i nuked it from the default runlevel, rebooted, and keyboard/mouse were frozen (as many others have experienced without these running)

looking at equery hasuse hal && equery hasuse dbus, not a ton that used this, but the key one that struck me was xinit

Also noticed enough important init scripts needed either hald or dbus, not as straightforward as I'd hoped.

SO it's looking like basically I'm not going to be able to test this without rebuilding a bunch of stuff - and since my primary aim is cutting down the time taken to boot my system (currently down to 32s thanks largely to a recent migration to OpenRC/baselayout-2), given that everything is working pretty damn nicely right now I'm not sure if it's worth going down the road of having to rebuild a bunch of stuff only to find out it doesn't work, and have to revert.