I too have been getting mysterious reboots recently
previously the system had seemed quite stable, alas my flash card
appeared to die from excessive compiles :( and my rebuilt setup is
much less stable it may be that it does not like swap and compiling
on a USB drive
I don't get many errors mostly just silent lock-ups where
both the serial console and network connection just stop updating.
I did get one reasonable dump from the serial console
[19178.028964] Unhandled kernel unaligned access[#1]:
[19178.034362] CPU: 0 PID: 1672 Comm: screen Not tainted 3.18.3test1+ #1
[19178.041573] task: 8aca7080 ti: 8b498000 task.ti: 8b498000
[19178.047610] $ 0 : 00000000 00000001 775cb700 00000000
[19178.053548] $ 4 : 8b499d30 fffff461 80017764 00000000
[19178.059483] $ 8 : 10000c01 1000001e 00000000 00000000
[19178.065419] $12 : 00000000 0000b800 00000000 00000001
[19178.071353] $16 : 8b499d68 fffff461 80017764 8000000c
[19178.077288] $20 : 8b499d30 8aca7604 00470000 1000e25e
[19178.083222] $24 : 00000000 802b1fc4
[19178.089155] $28 : 8b498000 8b499b48 00470000 8000000c
[19178.095090] Hi : 00000000
[19178.098304] Lo : 00000138
[19178.101536] epc : 8001ac8c emulate_load_store_insn+0x29c/0x580
[19178.108254] Not tainted
mostly however I have been getting either nothing or 8bit mixed dumps
So I thought I would try the 4.15 (current) kernel.
Serial console and ethernet seem to work.
I have a set of /dev/mtd devices so nand may be working
It appears that Neither MMC or USB work on a vanilla kernel
I would prefer to keep the nand as a backup boot source since it seems
I have already killed a SD card.
I had thought that ehci and ohci were listed as working on
https://elinux.org/CI20_upstreamAny thoughts on getting MMC and USB up on 4.15?
Thank you,
Jim McMechan