The udev guys should hear about hwdetect

Gullible Jones · 2006-04-28 19:29:05

I think [edit]tpowa[/edit] ought to send that script to the folks working on udev... Despite not loading modules in parallel, it was faster than uevents, and I think udev could benefit from it.

Then again, I don't really know much about how uevents works, so for all I know this idea might be pretty dumb.

Edit: if "No" please explain why. If you feel you need to explain one of the two other poll options, feel free.

Dusty · 2006-04-28 21:42:10

Why iphitus?????

trapdoor · 2006-04-28 22:06:11

hwd deinetely was faster and detected more modules

Dusty · 2006-04-28 22:44:30

FTR, hwd != hwdetect

iphitus · 2006-04-28 22:49:02

Dusty wrote:

Why iphitus?????

because I'm cool.

Dusty · 2006-04-28 22:57:26

I didn't mean to imply that you weren't, but I figure you have cooler things to do with your time too. Better to assign a minion. ;-)

Dusty

Gullible Jones · 2006-04-29 00:57:16

I thought iphitus wrote it? Or was it Phrakture? :oops:

T-Dawg · 2006-04-29 01:01:02

Gullible Jones wrote:

I thought iphitus wrote it? Or was it Phrakture? :oops:

tpowa

Dusty · 2006-04-29 01:23:58

The best thing to do here, I suppose, is to incorporate whatever coolness supposedly makes hwdetect better than udev into udev and then submit a patch to the udev devs.

Dusty

brain0 · 2006-04-29 10:12:23

hwdetect was slower than udev. And (despite the parallel module loading), they do the same thing. I am very happy with udev's detection mechanism.

dtw · 2006-04-29 14:28:25

Udev loads ten times less spurious modules on my system - hwdetect should stay dead and buried, bloat that it was.

Gullible Jones · 2006-04-29 16:26:53

Penguin wrote:

Gullible Jones wrote:
I thought iphitus wrote it? Or was it Phrakture? :oops:
tpowa

Shit. *bangs head on keyboard*

gradgrind · 2006-04-30 14:17:05

I don't have any great problems with the latest udev stuff in normal usage. It seems a little slower than the hwd(etect) method(s), but it still within acceptable bounds.

However, I have been playing around with live CDs (larch) and been getting quite ridiculously long uevents times (nearly 2 mins with udev-091-2). A few weeks ago with an older version it was about 30 seconds - which I already found rather long.
Does anyone have any idea what could be causing this? It suggests to me that maybe udev still has a few rough edges.

The main points of functional difference between normal systems and live CDs are: hd vs cd, unionfs, squashfs. But other device/module schemes don't degrade so dramatically under these circumstances.

Maybe someone is aware of another live CD using the latest udev stuff - one which isn't so sluggish?

jondkent · 2006-05-15 12:44:35

Always think that been shown another way is a "good thing", as it can sometimes lead you in supprising directions

Jon

Arch Linux

#1 2006-04-28 19:29:05

The udev guys should hear about hwdetect

#2 2006-04-28 21:42:10

Re: The udev guys should hear about hwdetect

#3 2006-04-28 22:06:11

Re: The udev guys should hear about hwdetect

#4 2006-04-28 22:44:30

Re: The udev guys should hear about hwdetect

#5 2006-04-28 22:49:02

Re: The udev guys should hear about hwdetect

#6 2006-04-28 22:57:26

Re: The udev guys should hear about hwdetect

#7 2006-04-29 00:57:16

Re: The udev guys should hear about hwdetect

#8 2006-04-29 01:01:02

Re: The udev guys should hear about hwdetect

#9 2006-04-29 01:23:58

Re: The udev guys should hear about hwdetect

#10 2006-04-29 10:12:23

Re: The udev guys should hear about hwdetect

#11 2006-04-29 14:28:25

Re: The udev guys should hear about hwdetect

#12 2006-04-29 16:26:53

Re: The udev guys should hear about hwdetect

#13 2006-04-30 14:17:05

Re: The udev guys should hear about hwdetect

#14 2006-05-15 12:44:35

Re: The udev guys should hear about hwdetect

Board footer