Gentoo Forums
Gentoo Forums
Gentoo Forums
Quick Search: in
Got udev working, now what?!
View unanswered posts
View posts from last 24 hours

Goto page Previous  1, 2, 3 ... 8, 9, 10 ... 28, 29, 30  Next  
Reply to topic    Gentoo Forums Forum Index Other Things Gentoo
View previous topic :: View next topic  
Author Message
PrakashP
Veteran
Veteran


Joined: 27 Oct 2003
Posts: 1249
Location: C.C.A.A., Germania

PostPosted: Sun Feb 15, 2004 4:43 pm    Post subject: Reply with quote

[OT]
Yes, I guess I am that one. I am still trying to find out what between 2.6.2-rc1 and 2.6.2-rc2 makes nvidia binary causing lock-ups and otr flooding of dmesg with badness in bla bla bla (quite famous now at nvnews). I hope either I find that bugger or nvidie fixes their driver sooner than later.

Ross' patch works quite nicely, btw. :)
[/OT]
Back to top
View user's profile Send private message
Moled
l33t
l33t


Joined: 09 Jul 2003
Posts: 635

PostPosted: Sun Feb 15, 2004 6:08 pm    Post subject: Reply with quote

just tried 17 on my new T41

rm -r'd the contents of /dev
rebooted and everything was automagically generated
all good so far
I have gone with a totally modular system here
first test, load the loop module

loop0,loop1 etc etc all are generated
unloading the module, they are all removed as well \o/

I have a usb stick, so I gave that a go

on top of the standard usb modules, it requires usb-storage + scsi core + scsi disc support

I plug it in -> usb storage + scsi support are loaded, but not scsi disc support

remove it and plug it in again, and this time the other module is loaded , but I guess thats more of a hotplug issue
udev generated sda + sda1 all ok after that

more on the "what next" issue, I currently have /dev/sda1 mounted as supermount, which nautilus produces a nice icon on the gnome desktop

any ideas on how to get this to only show up when the stick is plugged in?
Back to top
View user's profile Send private message
Moonlight-Flower
Tux's lil' helper
Tux's lil' helper


Joined: 15 Sep 2003
Posts: 135

PostPosted: Tue Feb 17, 2004 9:29 am    Post subject: Reply with quote

Hrm. I've got udev installed, and apparently, everything I need is working without a hitch (I guess I'm lucky...)

However, it seems to be creating a lot of devices in /dev, and most of them aren't necessary (or worse, don't exist... do I really have that many harddisks? ^_^)

Is there any way I can clean it up a bit?
Back to top
View user's profile Send private message
dsd
Developer
Developer


Joined: 30 Mar 2003
Posts: 2162
Location: nr London

PostPosted: Tue Feb 17, 2004 10:32 am    Post subject: Reply with quote

sounds like udev isnt working at all, and you just have the "static /dev" layout that everybody used before devfs came along.
udev will only create nodes for hard disks (and other devices) that actually exist.
_________________
http://dev.gentoo.org/~dsd
Back to top
View user's profile Send private message
deepthought
Guru
Guru


Joined: 04 Apr 2003
Posts: 321
Location: icbm://5131''N:0710''E

PostPosted: Tue Feb 17, 2004 10:59 am    Post subject: Reply with quote

I just emerged udev (16-r1 seems to be stable in portage) and did the following:

1. setup /etc/udev/udev.rules (removed devfs compatibility, changed mousedev symlink and the like)
2. removed useless /lib/udev-state crap and stupid (de-)population lines in /sbin/rc and /etc/init.d/halt.sh
3. emerged latest nvidia stuff (5536-r1) and added nvidia to /etc/modules.autoload.d/kernel-2.6, as xfree doesn't seem to be able to load the nvidia module automatically.
4. crossed fingers and rebooted

And... <snare roll/> WFM(tm).

Still, I have a few questions:

1. How can I get rid of 64 (!) tty device node symlinks? Shouldn't stuff under /dev/vc/ suffice? Removing the relevant SYMLINK part in /etc/udev/udev.rules breaks horribly and renders the system unusable (no console). Any ideas?

2. I am running hotplug in the default runlevel, not boot. Still, everything works fine. Is this arbitrary?

3. Can I unmerge devfsd (PLEASE!)?

Regards,
Alexander
_________________
Out of loyalty to its disregarded comrades, this message feels free to ignore the reader.
Registered Linux User #317705
Back to top
View user's profile Send private message
dsd
Developer
Developer


Joined: 30 Mar 2003
Posts: 2162
Location: nr London

PostPosted: Tue Feb 17, 2004 3:15 pm    Post subject: Reply with quote

interesting that you get the same problem as me relating to nvidia/X - i think i figured it out last night, see here for my interpretation of the problem...

Quote:
2. I am running hotplug in the default runlevel, not boot. Still, everything works fine. Is this arbitrary?

it should go in boot really. it may depend on how you have your kernel configure - which items are modules.

Quote:
3. Can I unmerge devfsd (PLEASE!)?

I have done so and it causes no problems. The only thing is that it wants to be re-merged when you do an "emerge -up world".
_________________
http://dev.gentoo.org/~dsd
Back to top
View user's profile Send private message
mani_
n00b
n00b


Joined: 17 Feb 2004
Posts: 8

PostPosted: Tue Feb 17, 2004 3:21 pm    Post subject: Reply with quote

hi

i replaced devfs with udev :)

but i've got one problem: it doesn't create the nodes for my framebuffer (/dev/fb0 and /dev/fb/0)

framebuffer works despite of that :?:

bootsplash reports:
Code:
open /dev/fb/0 and /dev/fb0: No such file or directory
Back to top
View user's profile Send private message
PrakashP
Veteran
Veteran


Joined: 27 Oct 2003
Posts: 1249
Location: C.C.A.A., Germania

PostPosted: Tue Feb 17, 2004 4:59 pm    Post subject: Reply with quote

Either use newest mm sources or search for Greg's fb sysfs patches on lkml and patch yourself.
Back to top
View user's profile Send private message
deepthought
Guru
Guru


Joined: 04 Apr 2003
Posts: 321
Location: icbm://5131''N:0710''E

PostPosted: Tue Feb 17, 2004 5:27 pm    Post subject: Reply with quote

dsd wrote:
interesting that you get the same problem as me relating to nvidia/X - i think i figured it out last night, see here for my interpretation of the problem...

Hm, it's a bit strange and modules.autoload doesn't seem to be the *clean*solution, but since udev is still unstable (and the workaround isn't that messy), I'll stay with it until udev calms down a bit.

dsd wrote:
it should go in boot really. it may depend on how you have your kernel configure - which items are modules.

Yes, many people told me to do so, but (at least on my box) it doesn't make a single difference... So the question is: Are there real good arguments for doing so or is it just the usual "do it because it really should be done and everybody else does so, too" FUD thingy? :evil:

dsd wrote:

I have done so and it causes no problems. The only thing is that it wants to be re-merged when you do an "emerge -up world".

I filed a bug on this. Hopefully, this issue is adressed ASAP.

Regards,
Alexander
_________________
Out of loyalty to its disregarded comrades, this message feels free to ignore the reader.
Registered Linux User #317705
Back to top
View user's profile Send private message
dsd
Developer
Developer


Joined: 30 Mar 2003
Posts: 2162
Location: nr London

PostPosted: Wed Feb 18, 2004 10:37 am    Post subject: Reply with quote

Decibels, any chance you could update your UDEV Primer webpage?

i was asked to write a "how to install udev" page to accompany my "writing udev rules" info, but i dont see the point considering that you have already put the effort in.

here's my suggestions:
- version 017 is in portage now, masked. its worth getting that instead of 016 as 017 fixes some bugs that people keep running into on 016..
- mention that a newish kernel must be used: mm-sources newer than 2.6.1-mm3, gentoo-dev-sources newer than 2.6.1-gentoo, development sources 2.6.3 or newer.
- mention that devices.tar.bz2 sort of goes against what udev tries to achieve and was only implemented by gentoo to work around earlier problems with udev which have mostly gone
- also mention that restoring from devices.tar.bz2 causes some problems, e.g. restoring a static /dev each time for people that install udev when devfs was not running.
- show which lines to comment out to prevent devices.tar.bz2 being used
- remove the stuff about making alsa nodes because this is done automatically with new udev and new kernel.
- talk about how people can optionally return to devfs on 2.4 kernels by leaving devfsd merged and booting with gentoo=noudev
- talk about how people can optionally return to devfs on 2.6 kernels by compiling devfs in as normal and booting with gentoo=noudev. when booting normally in "udev-mode" they will have to use devfs=nomount
_________________
http://dev.gentoo.org/~dsd
Back to top
View user's profile Send private message
Decibels
Veteran
Veteran


Joined: 16 Aug 2002
Posts: 1598
Location: U.S.A.

PostPosted: Wed Feb 18, 2004 3:07 pm    Post subject: Reply with quote

I thought about putting information about not using the devices.tar.bz2 in there before, but that was when we first tried it. Guess it would be a good idea to add it now.

Haven't been doing much with it lately, but have been keeping up with the thread, so guess I am up to the job. Or if you really wanted to do it, I don't mind if you copy the web page source.
Code:
 - remove the stuff about making alsa nodes because this is done automatically with new udev and new kernel.

Believe that is the same thing with the nvidia drivers also isn't it. Had it creating them in local.start until saw this. Took them out after and worked.
Code:
*nvidia-kernel-1.0.5336-r1 (01 Feb 2004)
                                                                               
  01 Feb 2004; Martin Schlemmer <azarah@gentoo.org>
  nvidia-kernel-1.0.5336-r1.ebuild,
  files/1.0.5336/NVIDIA_kernel-1.0-5336-basic-sysfs-support.patch,
  files/1.0.5336/Makefile:
  Add basic sysfs support for 2.6 and udev.  Patch donated by myself.


Probably just remove the whole device creation script now. Maybe leave the part in about how to do it for others, if they want to create a device or two, but don't need a whole script.

Code:
- also mention that restoring from devices.tar.bz2 causes some problems, e.g. restoring a static /dev each time for people that install udev when devfs was not running.


Do you mean if they undo the comments in the edited scripts and go back to using devices.tar.bz2 ? If they comment them out, the tarball isn't an issue. If they need a device that isn't being created by udev, guess they could just keep using the tarball. Or if like the nvidia driver issue I had earlier and it isn't needed during initial boot like a block device, they could just temp add creating it to local.start.

There was another issue that might need mentioned.
Seemed that some devices like console and null weren't there if someone started from scratch and didn't use the tarball at first. Know sounds weird, because if not using the tarball, then it doesn't get saved, so wouldn't be there at boot next time. But I ran into that once and so did someone else.
Created the two devices and rebooted fine the next time. Guess going to have to make a new install to check this one out again.

Well, unless you tell me you want to do it, I have print out your suggestions. Will look over a few others that people have given me and start working on that this weekend.
_________________
Support bacteria – they’re the only culture some people have.”

– Steven Wright
Back to top
View user's profile Send private message
dsd
Developer
Developer


Joined: 30 Mar 2003
Posts: 2162
Location: nr London

PostPosted: Wed Feb 18, 2004 4:10 pm    Post subject: Reply with quote

Decibels wrote:

Code:
- also mention that restoring from devices.tar.bz2 causes some problems, e.g. restoring a static /dev each time for people that install udev when devfs was not running.



no, ive spotted a few people fall into a common trap. their installation method looks something like:
1. compile/install new kernel, without devfs
2. reboot into new kernel [at this point they have no devfs or udev, so they have a static /dev with hundreds of devices]
3. update baselayout and install udev
4. reboot [on shutdown, their _static_ /dev layout is saved into devices.tar.bz2]
5. boot up into new kernel with udev [devices.tar.bz2 is then restored]
6. look in /dev and have a heart attack ("i dont have that many hard disks!!")
7. post to gentoo forums asking why udev has created so many nodes


as for nvidia, did you get it working "as normal"?

i.e. boot up into udev-based system (where devices.tar.bz2 is not used)
login as user
[at this stage nvidia module isnt loaded]
startx
[X loads nvidia module and starts up]

just asking because i have run into issues here. i think i know the problem, but i do not know how to address it.
_________________
http://dev.gentoo.org/~dsd
Back to top
View user's profile Send private message
Decibels
Veteran
Veteran


Joined: 16 Aug 2002
Posts: 1598
Location: U.S.A.

PostPosted: Wed Feb 18, 2004 4:56 pm    Post subject: Reply with quote

Damn, thought I had udev install on this partition. Changed the scripts and rebooted and failed. Said didn't have the nvidia devices. Didn't have the one Martin said he changed written down and just emerge the lastest stable. But didn't work on reboot, so didn't reboot again. So could write this.

Then found system said I didn't even have udev installed??? Who knows, been doing some funky stuff on here lately. This system was rsynced from the one I was originally running udev on. Who knows, but it is running pure udev at this moment.

Then found had old kde-3.1.4 written in /etc/rc.conf, so couldn't get X started. Had to change that. Back up again, but know when reboot going to loose the nvidia devices, unless upgrade to *nvidia-kernel-1.0.5336-r1 .

Answer to your question: Am I sure, ya, did it once. Remember taking the mknod lines out of local.start that put in.

Need to get some sleep, so will test that out tonight. Pretty sure worked.
But, even if does. What to do about people that don't want to install that ebuild. Guess they can just add the lines to local.start. If works tonight will add a note about that in the Primer and what to do to get X.

Long story, but does show how easy it is to go from devfs to udev. Had it fixed in about 5-10 min once figured out what was going on and fixed booboo's.

dsd wrote:

1. compile/install new kernel, without devfs
2. reboot into new kernel [at this point they have no devfs or udev, so they have a static /dev with hundreds of devices]
3. update baselayout and install udev


So, might be a good idea to say if you install a new kernel without devfs, make sure you install udev, update baselayout before you reboot, and edit the proper scripts listed if want a pure udev system (skip the edit if you don't). Not sure that it matters if the kernel using at the time when installing a new one has devfs.
Apparently on this kernel I didn't have udev installed or devfs compiled in (I checked while emerging nvidia-kernel). Except for the rc.conf kde-3.1.4 line and nvidia dev didn't have any other issues.
OH, but wait. Brain fart. I was using this kernel, not installing a new one.
After I get most of it updated, will install a new kernel and see if there are any issues.

**Later: Well, guess back to recreating the nvidia devices. Noticed something weird also, tried adding to udev.rules and got bunch of alsa errors at reboot and X still didn't come up. Will mess more with it later. Will work on the primer for now and add nvidia mknod back to the local.start.
_________________
Support bacteria – they’re the only culture some people have.”

– Steven Wright
Back to top
View user's profile Send private message
Decibels
Veteran
Veteran


Joined: 16 Aug 2002
Posts: 1598
Location: U.S.A.

PostPosted: Thu Feb 19, 2004 3:12 am    Post subject: Reply with quote

dsd wrote:

- mention that a newish kernel must be used: mm-sources newer than 2.6.1-mm3, gentoo-dev-sources newer than 2.6.1-gentoo, development sources 2.6.3 or newer.

- talk about how people can optionally return to devfs on 2.4 kernels by leaving devfsd merged and booting with gentoo=noudev


Haven't used udev on a 2.4 system. Is is possible, kinda conflict with first statement. Or do you mean, if they have an alternate 2.4 kernel and using udev on a 2.6 . If so, haven't noticed any issues. Before when had a 2.4 kernel also (again, that was before), I just rebooted to it. It just used devfs like normal. Has the scripts changed where actually have to add 'gentoo=noudev' on a 2.4 kernel, if also have a 2.6 with udev installed?

Also: has anyone recently install a new system with udev. Do you still manually have to create the /sys folder in root?

*** Most of the change to the UDEV Primer have been made. Still have to add a few items, look over for omissions,... Hopefully it is good, cause have to change a head gasket and replace broken timing belt this weekend. So take a look.
_________________
Support bacteria – they’re the only culture some people have.”

– Steven Wright
Back to top
View user's profile Send private message
jeffrice
Tux's lil' helper
Tux's lil' helper


Joined: 25 Jun 2003
Posts: 89
Location: New York, USA

PostPosted: Thu Feb 19, 2004 6:38 pm    Post subject: Reply with quote

I am missing something? I have gentoo-dev-sources 2.6.3 but I can't find the proc_kcore or ramfs options under pseudo file systems. I made sure I had the option for developmental components selected...

Jeff
Back to top
View user's profile Send private message
phlashback
Tux's lil' helper
Tux's lil' helper


Joined: 27 Jul 2002
Posts: 142
Location: Normal

PostPosted: Thu Feb 19, 2004 11:10 pm    Post subject: Reply with quote

I think that proc_kcore and ramfs are there by default... you can allways look at your .config, to be sure :)

btw... udev is working awsom thanks so much for the guide

Decibels if you are ever in the centrial illinois area I ow you a beer and pizza :)
_________________
The terrible thing about hell is that when you're there you can't even tell as you move through this life you love so you could be there and not even know
Back to top
View user's profile Send private message
Decibels
Veteran
Veteran


Joined: 16 Aug 2002
Posts: 1598
Location: U.S.A.

PostPosted: Fri Feb 20, 2004 12:11 am    Post subject: Reply with quote

I live right across the river, little south and have family in central Illinois. I'll take a stuffed crust pizza and maybe bring some of my home-brewed lager. :)

As far as proc_kcore and ramfs. Must be the way they are doing the kernel config display. Yes, if you look in .config they are there, but guess not giving you a choice to disable them so you don't screw up your system. I am going to change the text to a pic to better demonstrate what needs to be compiled.

Thanks for catching that. That slipped by me. Will update shortly.
Almost have the new pics ready and looks like isp webpages are down (probably for maint. but they never tell you that). Will get them up when see server back up.
_________________
Support bacteria – they’re the only culture some people have.”

– Steven Wright
Back to top
View user's profile Send private message
dedeaux
Apprentice
Apprentice


Joined: 19 Jun 2002
Posts: 183
Location: Nairobi, Kenya

PostPosted: Fri Feb 20, 2004 5:29 am    Post subject: please elaborate Reply with quote

deepthought wrote:
I just emerged udev (16-r1 seems to be stable in portage) and did the following:

1. setup /etc/udev/udev.rules (removed devfs compatibility, changed mousedev symlink and the like)
2. removed useless /lib/udev-state crap and stupid (de-)population lines in /sbin/rc and /etc/init.d/halt.sh
3. emerged latest nvidia stuff (5536-r1) and added nvidia to /etc/modules.autoload.d/kernel-2.6, as xfree doesn't seem to be able to load the nvidia module automatically.
4. crossed fingers and rebooted

And... <snare roll/> WFM(tm).

Still, I have a few questions:

1. How can I get rid of 64 (!) tty device node symlinks? Shouldn't stuff under /dev/vc/ suffice? Removing the relevant SYMLINK part in /etc/udev/udev.rules breaks horribly and renders the system unusable (no console). Any ideas?

2. I am running hotplug in the default runlevel, not boot. Still, everything works fine. Is this arbitrary?

3. Can I unmerge devfsd (PLEASE!)?

Regards,
Alexander


Can you elaborate on your steps 1 and 2 please..... what exactly did you remove from where....

Thanks!
Back to top
View user's profile Send private message
floam
Veteran
Veteran


Joined: 27 Oct 2002
Posts: 1067
Location: Vancouver, WA USA

PostPosted: Fri Feb 20, 2004 5:50 am    Post subject: Reply with quote

He took out the devices tarball. It's gross and I hope Gentoo guys get rid of it soon or make it optional. You should be able to have a 100% dynamic /dev/.
_________________
Think about your breathing.
http://floam.sh.nu
Back to top
View user's profile Send private message
IdleVoid
n00b
n00b


Joined: 20 Feb 2004
Posts: 1
Location: Denmark

PostPosted: Fri Feb 20, 2004 10:40 am    Post subject: Reply with quote

Hi Everyone...

I got a problem with udev..or the 2,6 kernel itself.... dont know yet.

I get the following erro when booting my 2.6.3-mm1 kernel

Code:
RAMDISK : Compressed image found at block 0
VFS : Mounted root (ext2 filesystem) readonly
Freeing unused Kernel memory 184K freed
Red Hat nash version 3.5.7 starting
Mounting / proc filesystem
Creating block devices
 failed to create  /dev/hda


and that's it.....

I'm also btw using a custom build (using mkinitrd) initrd file.
Without it I get another error..something like Kernel Panic - missing root partition (Not at my linux computer at the moment..so going to check later)

the Hotplug daemon is is set the boot (rc-update add hotplug boot), and I've chcked that its not set to default allso.

I don't know what I'm missing or what I've done wrong.... anyone got any ideas what could be wrong ?

I compiled the kernel manually...using this command : make && make modules_install..... should I've have used genkernel instead ?
And it's a clean install of gentoo....

I also fo rtesting added to echo statements on the /sbin/rc script file...they didn't get printed out at all...shouldn't it be called at boot time ?....

the /dev/ folder is filled with device files....I guess that's from the gentoo installation... (stage 3)

I've set up my harddisk like this :
/dev/hda1 : is boot partion with size of 100MB and using ext2 filesystem
/dec/hda2 : is swap partition...1GB
/dev/hda3 : is /home partition' using reiserfs filesystem
/dev/hda4 : is /root partition using reiserfs filesystem

and I'm using Grub 0.9.4 following settings (all written from my memory)
Code:
kernel (hd0,0)/kernel-2.6.3-mm1 root=/dev/ram0 init=/linuxrc real_root=/dev/hda4 vga=788
initrd (hd0,0)/initrd-2.6.3-mm1


::EDIT : I made the initrd file using this command : mkinitrd --image-version /boot/initrd 2.6.3-mm1

Any help would be appreciated alot :)

/IdleVoid
Back to top
View user's profile Send private message
charlieg
Advocate
Advocate


Joined: 30 Jul 2002
Posts: 2149
Location: Manchester UK

PostPosted: Fri Feb 20, 2004 11:25 am    Post subject: Reply with quote

Looks like major updatees and bugfixes in udev 18.
_________________
Want Free games?
Free Gamer - open source games list & commentary

Open source web-enabled rich UI platform: Vexi
Back to top
View user's profile Send private message
Decibels
Veteran
Veteran


Joined: 16 Aug 2002
Posts: 1598
Location: U.S.A.

PostPosted: Fri Feb 20, 2004 2:39 pm    Post subject: Reply with quote

dedeaux haven't been able to get the nvidia-kernel working anymore either and had it working while back. Added it to problem devices on the Primer. From the error I get, looks like the new portage may not be removing the old one properly. If you look it makes the new one nvidia.ko instead of nvidia.o .

When I had it working was running probably 2.6.1-test11 and a portage or two ago. You can read the error on the Primer. Don't have time right now to do much more looking into it. You can just create the devices in local.start or manually until figured out.

IdleVoid wrote:
I also fo rtesting added to echo statements on the /sbin/rc script file...they didn't get printed out at all...shouldn't it be called at boot time ?....

Also wrote:
Can you elaborate on your steps 1 and 2 please..... what exactly did you remove from where....

Did you 'etc-update' to get the new baselayout in place? I would assume so, since you edited it, but right now that is all got.

I have added what scripts of baselayout to edit and where in the primer. Haven't done anything with udev.rules yet though.

Well, udev-018 sees when plug my scanner in now. But instead of seeing it as a disk now, see's it as a usbcam. Xsane works again.
Code:
Feb 20 09:36:58 decibels default.hotplug[22890]: invoke /etc/hotplug/usb.agent ()
Feb 20 09:36:58 decibels usb.agent[22892]: Setup usbcam for USB product 5da/b6/100
Feb 20 09:36:58 decibels usb.agent[22892]: Module setup usbcam for USB product 5da/b6/100

_________________
Support bacteria – they’re the only culture some people have.”

– Steven Wright
Back to top
View user's profile Send private message
Decibels
Veteran
Veteran


Joined: 16 Aug 2002
Posts: 1598
Location: U.S.A.

PostPosted: Fri Feb 20, 2004 5:12 pm    Post subject: Reply with quote

NVIDIA-KERNEL working now. Either portage isn't correctly removing the previous nvidia-kernel or it's is because I have /lib/modules added to my CONFIG_PROTECT line. Probably the latter.

Code:
The 2.6 Linux kernel also introduces the sysfs filesystem to provide a hierarchical
view of each system's device tree (and to prevent further overloading of the proc
filesystem). The 2.6 Linux kernel also introduces a new naming convention for loadable
kernel modules, using the .ko extension (kernel object) rather than the standard .o (object)
extension used for loadable kernel modules in all previous stable releases of the Linux kernel.


Previous driver ext. was '.o', new module is '.ko'. CONFIG_PROTECT normally allows me to replace the current nvidia.o in the kernel using only. Have multiple kernels and don't want to have to remerge each time change. Due to the extension change it doesn't, so both are there. I did try renaming it and didn't work, maybe deleting it might have?

Anyway, unmerged nvidia-kernel and nvidia-glx. Emerged nvidia-kernel and glx version 1.0.5336-r1. Restarted computer and got the error about /dev/nvidia* .... already there, cause added mknod to local.start. Removed lines and rebooted. Nvidia devices made. So does work, just can't have the old nvidia driver in there.

The only other errors got were about Alsa devices. This happened last time did this also. Think starting alsamixer again and redoing fixed that. Will check later today.

For the time being, follow above or if using lower version nvidia drivers add to local.start. Will update the Primer tonight.

Cheers.
_________________
Support bacteria – they’re the only culture some people have.”

– Steven Wright
Back to top
View user's profile Send private message
Odius
n00b
n00b


Joined: 28 Jul 2003
Posts: 74
Location: USA

PostPosted: Fri Feb 20, 2004 7:21 pm    Post subject: Reply with quote

How did you guys get Alsa working with udev and the 2.6 kernel? I am using the 2.6.3 gentoo-dev-sources and udev 016-r1. I compiled the right module (emu10k1) for my SB Audigy when I built my kernel. Hotplug loads it (although I'm not sure if it loads all the needed modules). alsa-lib is installed and I use alsamixergui to make sure the volume is high and nothing is muted. What am I missing? Do I still need the line " alias snd-card-0 snd-emu10k1" in /etc/modules.conf? This is my first try getting alsa to work in a 2.6.x kernel.

Here is what sound modules that are loading:
Code:
snd_emu10k1            95492  0
snd_rawmidi            21664  1 snd_emu10k1
snd_ac97_codec         60740  1 snd_emu10k1
snd_util_mem            3520  1 snd_emu10k1
snd_hwdep               7776  1 snd_emu10k1


And here are my /dev/snd devices:

Code:
crw-------    1 jason    audio    116,   0 Feb 18 19:21 controlC0
crw-------    1 jason    audio    116,   4 Feb 18 19:22 hwC0D0
crw-------    1 jason    audio    116,   8 Feb 18 19:22 midiC0D0
crw-------    1 jason    audio    116,   9 Feb 18 19:22 midiC0D1
crw-------    1 jason    audio    116,  24 Feb 18 19:22 pcmC0D0c
crw-------    1 jason    audio    116,  16 Feb 18 19:22 pcmC0D0p
crw-------    1 jason    audio    116,  25 Feb 18 19:22 pcmC0D1c
crw-------    1 jason    audio    116,  26 Feb 18 19:22 pcmC0D2c
crw-------    1 jason    audio    116,  19 Feb 18 19:22 pcmC0D3p
crw-------    1 jason    audio    116,   1 Feb 18 19:21 seq
crw-------    1 jason    audio    116,  33 Feb 18 19:21 timer


EDIT: Spelling mistake.


Last edited by Odius on Fri Feb 20, 2004 7:58 pm; edited 1 time in total
Back to top
View user's profile Send private message
Decibels
Veteran
Veteran


Joined: 16 Aug 2002
Posts: 1598
Location: U.S.A.

PostPosted: Fri Feb 20, 2004 7:30 pm    Post subject: Reply with quote

Okay, figured out the New Install problem:

If you installed fresh and are deciding to choose between devfs and udev, you will reboot after done (if chose udev) and get an error about:
Code:
WARNING: Unable to open an initial console.


It needs two devices before udev is doing it's job. Console & Null.

While doing a new install, if you pick udev. Do this:
Code:
cd /dev
mknod -m 660 console c 5 1
mknod -m 660 null c 1 3


Then you will be able to reboot into the new system with pure udev. You may or may not see this error if using the tarball - devices.tar.bz2

That is why people that are changing over don't see this error, cause console and null are there. Make a new system and will not have those two devices and udev isn't creating them in time.

Will update this and nvidia-kernel on primer tonight.

For now you need to have those two static devices in /dev or will not be able to boot.
_________________
Support bacteria – they’re the only culture some people have.”

– Steven Wright
Back to top
View user's profile Send private message
Display posts from previous:   
Reply to topic    Gentoo Forums Forum Index Other Things Gentoo All times are GMT
Goto page Previous  1, 2, 3 ... 8, 9, 10 ... 28, 29, 30  Next
Page 9 of 30

 
Jump to:  
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum