TK1 u-boot Not Halting On Key Press

I have a situation whereby u-boot is not halting when key presses are made.

I have two TK1 modules - an old V1.0B module and a V1.1A module. The older module is with a customer and the newer module is on my desk. The older module is the one with the issue whereas the newer module is working fine.

Both have been previously updated with u-boot:

U-Boot 2016.11-2.7.2+g60021a4 (Apr 10 2017 - 07:55:49 +0200)

The very odd thing is, that when the V1.0B module boots through to Ubuntu, the key presses are accepted (although there are issues with the login not working, “authentication failure”) but fundamentally, it works.

Is there anything fundamentally different about this version of u-boot which would not work on the TK1 V1.0B?

Well, all those UART signals go through a bunch of level shifters. If anything in the pin muxing is amiss then something like that may happen I guess. But generally of course it should really be working in our latest released BSPs.

The strange thing is though, that occasionally it works and key presses are read by u-boot and the system can be halted. I will add though that even though the key presses can be ignored by u-boot, they ALWAYS register once booting has complete.

Having finally managed to get the customer to halt their u-boot on their V1.0B TK1, I’ve done a side-by-side comparison of the kernel arguments:

Kernel Args Comparison

Whereby my V1.1A TK1 is on the left and the V1,0B TK1 is on the right.

Is it reasonable to expect them to be so different or is it advisable to reset the V1.0B args to defaults using env default -a?

It looks like the older one till has the U-Boot environment of an older U-Boot version which ultimately means you did not carefully follow the following article on our developer website when updating e.g. don’t forget the env default -a; saveenv part.

Yep, that’s pretty much correct but in fairness, I didn’t realise that the customer had such an old TK1 and had foolishly assumed that they were using the latest equipment. Lesson learned. Whether it fixes the issue, I’m not sure yet but I will try as soon as the customer is awake and functional.

Thank you.

Yep, that’s pretty much correct but in fairness,

Sorry, I didn’t want to sound unfair or anything.

I didn’t realise that the customer had such an old TK1 and had foolishly assumed that they were using the latest equipment.

Yes, those early V1.0A sample modules are pretty much unfit for most realistic use cases.

Lesson learned. Whether it fixes the issue, I’m not sure yet but I will try as soon as the customer is awake and functional.

It really could also have something to do with that early sample hardware. At least we are not aware of any such issue on the current V1.1A hardware version.

Thank you.

It’s us having to thank you for the continued trust you put in our products.

I’m arranging for new hardware to be sent to the customer, just in case - this includes an up-to-date TK1 with our latest Linux deployment. Hopefully this will fix the issues seen if resetting the environment variables does not.

Your hardware is great and we’ve now integrated the TK1/T30’s into several of our mainstream products. Unfortunately it is my knowledge of Linux that at times has let us down and hence why I’ve had to ask such stupid and basic questions.

Thanks again.