Skip to contentSkip to conversation

Remove devfs from 2.6.12-git

PATCH David Brownell 3 participants 3 messages Jun 23, 2005

lore
Share & tools
lore
#0 lore raw

Quoth Mike Bell:

Quoted 6 lines
Quoted 3 lines

Also, no disto uses devfs only (gentoo is close, but offers users udev and a static /dev also.)

It breaks a lot of my embedded setups which have read-only storage only and thus need /dev on devfs or tmpfs.

I'd agree that embedded setups are the ones that have been slowest to switch over, for various reasons. One of them is that many LKML folk ignore embedded systems issues; "just PC class or better". Another has been that the basic hotplug scripts never worked well with "ash", and who's going to want to ship "bash" and friends? :)

Those problems seem resolved with 2.6.12 and current modutils and udev. Leaving basically an "upgrade your userspace" requirement.

Quoted 2 lines

and thus need /dev on devfs or tmpfs. With early-userspace-udev-on-tmpfs being - in my experience - still unready.

Hmm, could you explain why you think udev-on-ramfs/tmpfs/... is still unready? And what it's "unready" for? It can work fine without needing to hack early userspace and initramfs, by the way. :)

I recently submitted patches to make buildroot support it. They're mostly merged [1] but adding a simple patch fixes up the remaining glitches. OpenEmbedded has supported it for a while too.

In short, I think udev-on-ramfs is simple enough to set up on 2.6.12 based embedded configs, and with "modprobe -q $MODALIAS" phasing in, it seems to me [2] that hotplug-with-udev [3] can do most of what folk have used devfs to achieve. I've quilted buildroot so that it works that way by default for me; it's turnkey, with no problems.

Of course, 2.6.13 will be better with cardmgr finally gone. :)

- Dave

[1] http://bugs.busybox.net/view.php?id=290
[2] http://marc.theaimsgroup.com/?l=linux-hotplug-devel&m=111903617808816&w=2
[3] http://marc.theaimsgroup.com/?l=linux-hotplug-devel&m=111903647518255&w=2

#1 lore raw
Quoted 3 lines · David Brownell

I'd agree that embedded setups are the ones that have been slowest to switch over, for various reasons. One of them is that many LKML folk ignore embedded systems issues; "just PC class or better".

I dunno about that -- while maybe the average LKMLer doesn't actively worry so much about embedded issues, I've found many of them are very friendly and helpful in getting patches for embedded/small-system functionality cleaned up and merged into the mainstream kernel.

-miles

Signature · 2 lines
The car has become... an article of dress without which we feel uncertain,
unclad, and incomplete.  [Marshall McLuhan, Understanding Media, 1964]
#2 lore raw
In reply to David Brownell
Quoted 3 lines · David Brownell

I'd agree that embedded setups are the ones that have been slowest to switch over, for various reasons. One of them is that many LKML folk ignore embedded systems issues; "just PC class or better".

I don't think that's a fair comment - however I never read any of those udev/devfs flamewars. As the lead developer of an embedded architecture, I find that many people here do not take that attitude.

More the problem is that many embedded folk aren't here to air their *technical* arguments because they can't keep up with the lkml traffic level.

Consequently, lkml folk don't get the feedback from embedded folk, and hence probably why udev doesn't/didn't work with ash.

Signature · 3 lines
Russell King
 Linux kernel    2.6 ARM Linux   - http://www.arm.linux.org.uk/
 maintainer of:  2.6 Serial core