USB 1-1: device not accepting address issue with Dell 9020's

[JGallo]

[george1421]

@jgallo two things

1. What version of fog are you running?
2. Is the firmware up to date on the 9020?
   2b. Is the target in uefi or legacy mode?

I can see if I can duplicate it in the AM with one of our systems. I should have a 9020 someplace.

[JGallo]

@george1421

• Version 64 on working branch
• Yes. It is A20.
• Target is set to legacy

[Sebastian Roth]

@jgallo said in USB 1-1: device not accepting address issue with Dell 9020's:

I’m not using a USB NIC

Do I get this right? You are not using any USB device for this PC to image? Maybe a USB keyboard/mouse?

TSC_Deadline error - firmware bug - please update microcode

Could be related in that way that the buggy firmware is actually causing the USB issue. But can’t say for sure yet. We need more information on the CPU! Please boot up the client to a debug task (capture or deploy doesn’t matter), run cat/proc/cpuinfo, take a picture and post here. Read up about this here: https://wiki.archlinux.org/index.php/microcode

[JGallo]

@sebastian-roth

It does it with and without the USB keyboard. Interestingly, these 9020’s come with PS2 connections. Not sure it makes a difference.

[Sebastian Roth]

@jgallo said in USB 1-1: device not accepting address issue with Dell 9020's:

It does it with and without the USB keyboard

Well that’s interesting. So no USB devices at all and you still see the USB error.

About the picture… Can you issue the command cat /proc/cpuinfo | head -15 so we have the important top part of the output. Sorry, didn’t think about low screen resolution in FOS on the first run. As well make sure there is as little reflection on the screen as possible so everything can be read easily.

[JGallo]

@sebastian-roth

Yup. Very weird.

[Sebastian Roth]

@JGallo I have to admit that though I have compiled the linux kernel many times I’ve never actually added CPU microcode. Reading up on this it turns out that you can obviously only add one microcode binary for Intel and one for AMD at once. So it’s impossible for us to add a specific CPU microcode for all the different CPUs out there.

I’d advise you to check out our wiki on how to build your own custom kernel and modify the initrd. It’s actually not too hard if you have a bit of experience with linux:
https://wiki.fogproject.org/wiki/index.php?title=Build_TomElliott_Kernel#Build_TomElliott_Kernel_for_FOG_0.33b_and_newer
https://wiki.fogproject.org/wiki/index.php?title=Modifying_the_Init_Image#Version_1.0.0_and_up

[Sebastian Roth]

@JGallo Forgot to tell you, the microcode you will find on Intel’s website: https://downloadcenter.intel.com/product/75043/Intel-Core-i5-4570-Processor-6M-Cache-up-to-3-60-GHz- (should be your CPU I think - maybe try using the CPU identifying tool to make sure)

[JGallo]

@sebastian-roth

Well sounds like fun of course. Most definitely going to begin looking into creating a custom kernel. In any case, would it be possible to silent the error messages on the kernel arguments LOL? I did that for ACPI errors. I mean those USB errors are really annoying because your trying to register a host and two three lines of USB errors occur and get in the way. Thank you for your help.

[Sebastian Roth]

@JGallo AFAIK the default loglevel in FOG is set to 4 (KERN_WARNING). You can try adding kernel parameter loglevel=3 (or even down to zero) in the host settings in the web UI.

[Sebastian Roth]

@JGallo Ah, sorry, got that wrong. Better change FOG_KERNEL_LOGLEVEL in FOG Configuration -> FOG Settings -> FOG Boot Settings… as this is always set anyway. I don’t know which setting would make it if you have loglevel=4 (from the general settings) plus loglevel=3 set in the host settings…

[JGallo]

@sebastian-roth

Sweet. The default setting for the loglevel is 4. Taking it down to 3 no longer produces those USB errors. Thank you. The computer does register and I can image but those USB errors were extremely annoying. Seems like this will suppress those messages and works just fine now.