Udev Issue HLFS--Nothing seems to work just right...

Hello.

I've been trying to get help on this issue for quite some time. I just can't figure it out. Here are some of the issues I've been encountering:

1. If I pass vga=4 to the kernel to get more lines on the screen, invoking s2ram seems to reverse this. I actually won't be able to see the bottom 4-5 lines, so I won't be able to see what I am typing.

2. I needed to create an init script to start the udev daemon after the udev script because udev won't create any device nodes or load firmware otherwise. udevmonitor picks up hotplug events, but it won't create any device nodes unless udevd is running.

3. I have an IBM ThinkPad X41 Tablet that has a fingerprint reader that works over USB. I'm using bioapi and pam to get it to work, and it does work. However, lsusb doesn't show the device (and it's unusable) unless I suspend the laptop with s2ram and then wake it again.

I am using a modified Hardened LFS OS with udev 130 and Linux 2.6.24.7-grsec. Any help would be terrific.

Thanks,

Micah Henning

Quote:

---

Originally Posted by micchickenburger

1. If I pass vga=4 to the kernel to get more lines on the screen, invoking s2ram seems to reverse this. I actually won't be able to see the bottom 4-5 lines, so I won't be able to see what I am typing.

---

Well, we all know how much everyone loves to throw links at people on these forums... so let me add to it! haha
Get s2ram working
From that link, I gather that this problem is not only expected, but really not repairable. It sounds like it's a problem with the video card itself. There is a possible workaround presented in that article.

Quote:

---

Originally Posted by micchickenburger

2. I needed to create an init script to start the udev daemon after the udev script because udev won't create any device nodes or load firmware otherwise. udevmonitor picks up hotplug events, but it won't create any device nodes unless udevd is running.

---

I hate to give you tutorials on init script creating, but I don't know how to do it myself. So here are some links. If you've tried them already, then my bad for sending you there. lol
Managing Linux daemons with init scripts
Linux Startup Scripts

Hopefully those help.

Quote:

---

Originally Posted by micchickenburger

3. I have an IBM ThinkPad X41 Tablet that has a fingerprint reader that works over USB. I'm using bioapi and pam to get it to work, and it does work. However, lsusb doesn't show the device (and it's unusable) unless I suspend the laptop with s2ram and then wake it again.

---

This doesn't surprise me at all. The fingerprint reader has caused problems since it was first created. Even in Windows, which is what it was originally designed for, it has been extremely buggy and screwy. cnet and The Register were flooded with articles regarding the big screw up that the fingerprint reader was. So to get it to work in Linux will be even more difficult.

Again, I'm sorry if I wasn't much help. I just saw that you posted three days ago and that no one had responded. So I figured I'd at least try to give you something. Good luck, let me know how it turns out.

SkittleLinux18

Firstly, thank you for your response. While I did not find any related information from those links, I did find a solution.

I did have a problem with udev-130, actually. I followed the LFS instructions from their development page and noticed an interesting problem that has not yet been resolved: The udev rules installed by LFS are installed to /etc/udev/rules.d when instead they should be installed to /lib/udev/rules.d like all the rest of them are. This probably wasn't the cause of my problem but it did get rid of a bunch of annoying daemon messages.

The cause was actually an outdated udev init script. I was using Zack Winkles Udev cold-plugging script, version 00.01. I updated it to 00.02. It turns out that the old init script never starts the udev daemon. I had already figured that out, but I could never figure out how to create device nodes for already-discovered devices. Using "udevadm trigger" reloads those nodes by requesting all uevents from the kernel. Use "udevadm settle" to wait for the event queue to finish.

I'm not sure what problems others have with the fingerprint reader, but it works just fine on my tablet. The only issue I had was trying to get the driver to load on boot. This was resolved with the aforementioned udev init script update. On another note, the fingerprint reader won't be detected while the laptop is on the dock. I'm not sure why that is. Driver issue?

s2ram is a funky program and needs all sorts of crazy hacks to get framebuffer devices to work correctly. The help section from the s2ram program shows how I can keep the VBE state after resuming, but this actually prevents the screen from turning back on (the system doesn't halt, though). s2ram by itself works flawlessly on my machine without any hacks. I might need a hack to get this to work, though.

Once again, thanks a lot for your help!

Hey buddy, I'm really glad you got things working. I'm not quite sure what you mean but the laptop being on the dock, and therefore why the fingerprint reader doesn't work when such is the case. Could I trouble you to clarify that for me??

Everything else sounds like script and udev issues, just like you said. So I'm glad you are hunting down solutions. As I mentioned in my first response, I wasn't sure if I'd find anything that would be the miracle cure for you. My intention was just to put something out there that might help point you in the right direction, spark an idea, or maybe cause you to look at other things you may not have looked at yet. I know for me, sometimes I already know the answer, I'm just caught up on the wrong hypothesis for my problem.

Anyway, I'm glad you're making good progress.

Quote:

---

Originally Posted by micchickenburger

Firstly, thank you for your response. While I did not find any related information from those links, I did find a solution.

I did have a problem with udev-130, actually. I followed the LFS instructions from their development page and noticed an interesting problem that has not yet been resolved: The udev rules installed by LFS are installed to /etc/udev/rules.d when instead they should be installed to /lib/udev/rules.d like all the rest of them are. This probably wasn't the cause of my problem but it did get rid of a bunch of annoying daemon messages.

The cause was actually an outdated udev init script. I was using Zack Winkles Udev cold-plugging script, version 00.01. I updated it to 00.02. It turns out that the old init script never starts the udev daemon. I had already figured that out, but I could never figure out how to create device nodes for already-discovered devices. Using "udevadm trigger" reloads those nodes by requesting all uevents from the kernel. Use "udevadm settle" to wait for the event queue to finish.

I'm not sure what problems others have with the fingerprint reader, but it works just fine on my tablet. The only issue I had was trying to get the driver to load on boot. This was resolved with the aforementioned udev init script update. On another note, the fingerprint reader won't be detected while the laptop is on the dock. I'm not sure why that is. Driver issue?

s2ram is a funky program and needs all sorts of crazy hacks to get framebuffer devices to work correctly. The help section from the s2ram program shows how I can keep the VBE state after resuming, but this actually prevents the screen from turning back on (the system doesn't halt, though). s2ram by itself works flawlessly on my machine without any hacks. I might need a hack to get this to work, though.

Once again, thanks a lot for your help!

---

You did spark the idea that there was an issue with the udev init script. Thanks :-). I accidentally ran into the newer init script version through Google.

The problem with the fingerprint reader is that lsusb will only show the device when the laptop is not on the dock. If I boot my laptop on the dock or wake it from suspend mode while it is on the dock, the fingerprint reader will not be discovered. If I boot my laptop by itself or wake the laptop from suspend mode when it is not on the dock, the fingerprint reader will be discovered. It's not really an issue to me, but I am curious as to why that is.

Thanks for your support!

Well that is all good news. Still don't know what a "dock" is... or maybe I do, but just by another name??

I'm glad I helped!