This chip is ESP32-C6 not ESP32

There's your error: you have something (the flash tool?) set up for an ESP32 and you're flashing an ESP32C6. Not sure how exactly you would change that in the GUI tool but if that's impossible, perhaps you can use the command-line tool for this?

The P4 doesn't have *any* RF subsystem, it's microcontroller-only. However, you should be able to pair it with any ESP32Cx or ESP32Hx chip, so once the C5 is out, you could pair it with that to get 5GHz.

Are the repeats the same? If so, they could be WiFi retransmits. WiFi packets are ACKed by the access point when received, and when the sender doesn't get that ACK, it'll think the packet is lost and will retransmit. This looks like the packet made it over OK but the ACK got lost.

What chip is this on? Any chance you can whittle down your code to the smallest project that still exhibits it and post that here?

What specific board hardware is this? If it's custom, do you have schematics / board layout files?

There's likely no AT code in that devboard by default, so it's no surprise that doesn't work. What specifically (OS, software) are you using to flash it?

The root cause is that something (likely the timer interrupt for the matrix display) is trying to execute code from flash, but WiFi needs to write to that flash so it is unavailable. Probably a mistake in the rgbmatrix library I guesss; it should be entirely in IRAM. If you decode that backtrace, yo...

I don't think that is possible. The ringbuffer memory contains semaphores, which in turn contain pointers to internal FreeRTOS structures. Those structures are not guaranteed to be the same or at the same position after every reboot, so the pointers probably point to wrong data.

Just wondering: is it actually a variable at an absolute address you want, or do you simply want a variable that does not get cleared after a reset so your firmware can simply 'pick up' the value again after a reboot? Because if it's the second, you're probably better off declaring it static with a ...

That sounds like a very noisy environment indeed. Perhaps some capacitors on the ADC pins can help a bit in cleaning up the signal? (Something like 100nF should be sufficient.) Another trick, especially for slower-moving signals, is to take lots of measures and average them to get rid of the noise.