#GPIO uart serial debugging of bad nvme boot

1 messages · Page 1 of 1 (latest)

elder veldt
#

I'll be happy to try any serial port suggestions you have, and may be able to throw a scope on the tx/rx lines as well

zenith bone
#

Should be 115200

elder veldt
#

That's what I had it at, same settings that worked on the usb serial. But nothing sensible came through. I'll see if I can grab a scope trace

zenith bone
#

Do you have GND connected?

elder veldt
#

ok, I don't think I actually got the new config set on the cm4. I ran the update-pieeprom.sh from a linux machine and thought it 'did something' but the 'did something' is the same if I have the yellow connected or not.

#

should I boot the yellow from a usb stick and then tyr to run the usbboot commands from a terminal on the cm4?

#

I did have gnd connected.

elder veldt
#

ok, booted from a usb raspbian image, and confirmed that the eeprom config is `pi@raspberrypi:~ $ rpi-eeprom-config
[all]
BOOT_UART=0
WAKE_ON_GPIO=1
POWER_OFF_ON_HALT=0

Disable bootloader updates from USB MSD & network.

ENABLE_SELF_UPDATE=0

SD, PCIe USB MSD, BCM2711 USB MSD, LOOP

BOOT_ORDER=0xf541`

#

vcgencmd bootloader_version gives
Jan 19 2021 12:00:11 version e5e540537282a28aa8ec29d9662d65e06d0d2e24 (release) timestamp 1611057611 update-time 1611057611 capabilities 0x0000001f

elder veldt
#

ok, double checking things, I ran around in circles where I see that I can't flash the eeprom from the cm4 itself, so I'll attempt the usb recovery approach again, following Jeff Geerling's page.

#

I switched the dip pins from uart to usb and connected the usb recovery jumper, plugged in the usb to the linux machine, and supplied power

#

I then ran the commands and I believe got the eeprom updated with boot uart =1 and a nvme boot option

#

switched the jumpers back and getting ready to boot with the gpios connected to my usb serial adapter

#

looking closely, I might have had the wrong pin for ground, so maybe that was the issue.

#

success!!!

#

poor silkscreening of the pins is to blame.

#

Failed to open device: 'sdcard' (cmd 371a0010 status 1fff0001) Boot mode: USB-MSD (04) order f256 USB xHC init failed Boot mode: NVME (06) order f25 VID 0xc0a9 MN CT500P2SSD8 NVME on MBR: 0x00000800, 65536 type: 0x0c MBR: 0x00000001, 2047 type: 0xee MBR: 0x00000000, 0 type: 0x00 MBR: 0x00000000, 0 type: 0x00 Trying partition: 0 type: 16 lba: 2048 oem: 'mkfs.fat' volume: ' V ^ ' rsc 4 fat-sectors 64 c-count 16343 c-size 4 root dir cluster 1 sectors 32 entries 512 FAT16 clusters 16343 Trying partition: 0 type: 16 lba: 2048 oem: 'mkfs.fat' volume: ' V ^ ' rsc 4 fat-sectors 64 c-count 16343 c-size 4 root dir cluster 1 sectors 32 entries 512 FAT16 clusters 16343 Read config.txt bytes 864 hnd 0x20 Read start4.elf bytes 2250656 hnd 0x1ea Read fixup4.dat bytes 5400 hnd 0x21 0x00b03140 0x00000000 0x00001fff MEM GPU: 32 ARM: 992 TOTAL: 1024 Firmware: c72ad6b26ff40c91ef776b847436094ee63fabee Oct 26 2022 11:09:05 Starting start4.elf @ 0xfec00200 partition 0 NVME off +

#

I wonder if maybe it just didn't have the nvme boot option in the eeprom, and the installer couldn't update it because of cm4 lite problems? but then why would samsung nvmes work....? maybe I can put the eeprom config back to the (not including nvme boot option) and try the installer image with a samsung drive??

elder veldt
#

Well, thanks for your help, and sorry for the apparent goose chase. It really seems like it was just the eeprom and I didn't successfully update it when I tried that first.

zenith bone