Yes!’ Precisely! It actually IS the same kernel as Joao’s image; albeit with the DEOT splash screen. So that is good to hear that it seems that there could be some hardware discrepancies/compatibilities directly with the kernel. Although …
In this instance, when you mean software, are you referring directly and SPECIFICALLY to the OS, launcher, and OS run drivers, or is this more an umbrella term you have used fo differentiate that which is on the SD card, from the physical hardware? Ie, would you also say that the kernel is a part of the software that is the issue? This is important for a reliably isolating the problem. I thought I was getting close, but this puts me a step backwards.
If it is indeed the software (which I am defining as Debian + the Launcher, and NOT the Kernel/DTB) I would love to know exactly what your findings are, and what steps you have taken fo potentially resolve it. In particular, whether or not you have tried using either the underclocked (which is actually stock), or overclocked settings.
Saying it is “definitely” the software is an extremely decisive word to use, and one that can easily turn people off of trying the image if that is what they see. It will always be those who have problems who speak out the most; with those who don’t just going about their lives. It’s just unfortunate that only now, three releases into using the aforementioned kernels that we are getting an influx of reports of problems. These are the same kernels that have been in use since around June.
As mentioned above, the governed profile dynamically changes the clockspeed of the Gameshell, which seems to affect users who a) leave their gameshells on for extended periods idle, and b) those who use wifi for extended periods. I don’t actually use my gameshell this way; instead preferring to use the efficient suspend mode when I need to take a call, feed the pets, change trains etc. I never thought to test leaving it idle to see what happens.
Hmmm. The battery definitely seems like it’s on its way out. It would be worthwhile to get a new one anyway. Just make sure you don’t get one rated over 1500mAh - if it’s too good to be true, it’s probably dangerous to use.
I doubt it’s a cause, and heck, I don’t know if it would actually change anything, but at least in my field of work (music), people have been known to use an inline ferrite ring to reduce any buzzing/interference from a power source to audio equipment. Again, I doubt it would help, but perhaps give it a try. You can probably pull one off of an old mini USB digital camera cable. They normally have one on them.
Please do, and report your findings. While you’re at it, also attempt to replace the entire contents of the “boot” partition of the DEOT image with the contents of the stock 0.5 image. You should be able to see it as the drive that comes up when mounted on your computer.
I seriously doubt that it is a software related issue, and instead to do with the kernel. In fact, I believe it was in the 0.5 image that there was actually a test “5.4.6” that in the end was rejected in testing, no doubt due to similar problems to what people are experiencing. “5.3.6” became the standard kernel that initially brought us Lima support. I believe the kernel was made by @shell. See here:
@guu - do you have any memory of the there possibly being any problems with the 5.4.6 kernel that made it more viable to use the 5.3.6 kernel?
Either way, if the stock OS doesn’t produce any of the screen problems, don’t immediately dismiss it as being down to this custom image. It would be great to further development together if we properly narrowed down the root of the problems.
Currently I have got three CPI boards that I thoroughly test my OS on before publishing, and two sets of screens/arduinos/chassis etc. I would LOVE to get my hands on a nonfunctional board with the flashing screen problem. And even the exact card the image was flashed to. I’d be happy to do a 1 for 1 trade, although goodness knows how long that would take in the mail in this current climate.
It does mean a lot that so many people have been trying the image out, and testing it for and problems. I am extremely sorry for all the problems, and wish I could replicate them myself. Thanks for all being in essence, the testers. I would like to say that this image is stable and ready to be wrapped up, but that seems to only be the case for some people.
On a complete side note, if needed I can provide a video of my gameshell running using the 5.7 kernel, and not exhibiting any of the screen problems; just in case anyone thinks I’m telling porkie pies. I’ve posted videos in the past, and they have all been glitch free. Just let me know, and I upload a boot sequence, display of the about screen to indicate the kernel, and running a game of choice. But realistically I don’t thinks it’s necessary.