5B will not boot with the NVMe SSD connected

#21

NVMe you are going to be disappointed. You can you it as storage but if you want to boot from it it will not be reliable. Even with the “official” power supply.

You need a specific boot loader file to boot with it and armbian works to boot. EXCEPT. it will randomly not boot. Set up a cron every 5 min to reboot. If you make it past an hour without it failing you would be lucky. So basically it useless because you can’t depend on it. You can search my username history and find out all about it. Just save yourself the time, if you need to boot from nvme, abandon it.

1 Like
#22

I have similar problem. I use dell 130W usb power brick.

When trying to boot from NVME i can boot when i use image:

https://github.com/radxa/debos-radxa/releases/download/20221031-1045/rock-5b-ubuntu-focal-server-arm64-20221031-1328-gpt.img.xz

but when i try:

https://github.com/radxa-build/rock-5b/releases/download/b36/rock-5b_ubuntu_jammy_cli_b36.img.xz

It does nothing on screen and on serial i have:

01:32: [00]DDR Version V1.08 20220617
LPDDR4X, 2112MHz
channel[0] BW=16 Col=10 Bk=8 CS0 Row=16 CS1 Row=16 CS=2 Die BW=16 Size=2048MB
channel[1] BW=16 Col=10 Bk=8 CS0 Row=16 CS1 Row=16 CS=2 Die BW=16 Size=2048MB
channel[2] BW=16 Col=10 Bk=8 CS0 Row=16 CS1 Row=16 CS=2 Die BW=16 Size=2048MB
channel[3] BW=16 Col=10 Bk=8 CS0 Row=16 CS1 Row=16 CS=2 Die BW=16 Size=2048MB
Manufacturer ID:0x6 
CH0 RX Vref:30.7%, TX Vref:20.8%,19.8%
CH1 RX Vref:29.7%, TX Vref:21.8%,21.8%
CH2 RX Vref:31.7%, TX Vref:20.8%,19.8%
CH3 RX Vref:30.7%, TX Vref:21.8%,21.8%
change to F1: 528MHz
change to F2: 1068MHz
change to F3: 1560MHz
change to F0: 2112MHz
out
INFO:    Preloader serial: 2
NOTICE:  BL31: v2.3():v2.3-405-gb52c2eadd:derrick.huang
NOTICE:  BL31: Built : 11:23:47, Aug 15 2022
INFO:    spec: 0x1
INFO:    ext 32k is not valid
INFO:    GICv3 without legacy support detected.
INFO:    ARM GICv3 driver initialized in EL3
INFO:    system boots from cpu-hwid-0
INFO:    idle_st=0x21fff, pd_st=0x11fff9, repair_st=0xfff70001
INFO:    dfs DDR fsp_params[0].freq_mhz= 2112MHz
INFO:    dfs DDR fsp_params[1].freq_mhz= 528MHz
INFO:    dfs DDR fsp_params[2].freq_mhz= 1068MHz
INFO:    dfs DDR fsp_params[3].freq_mhz= 1560MHz
INFO:    BL31: Initialising Exception Handling Framework
INFO:    BL31: Initializing runtime services
WARNING: No OPTEE provided by BL2 boot loader, Booting device without OPTEE initialization. SMC`s destined for OPTEE will return SMC_UNK
ERROR:   Error initializing runtime service opteed_fast
INFO:    BL31: Preparing for EL3 exit to normal world
INFO:    Entry point address = 0x200000
INFO:    SPSR = 0x3c9
#23

I circled back to https://wiki.radxa.com/Rock5/install/spi and used debug botloader: https://dl.radxa.com/rock5/sw/images/loader/rock-5b/debug/rock-5b-spi-image-gbf47e81-20230607-debug.img and after that - images from radxa-build github started to work with NVME (tested on samsun 980 and some old intel 660).

Will try to find eMMC and check it would work.

#24

@Peter.Wang, thanks. No, no visible problem with the M.2 port pins, nor the pins on the SSD. The SSD tests OK. I put it into an Intel NUC and installed Ubuntu x86 on it, including formatting it, and it worked fine. I presume to request an RMA I just use the Contact Us form? Have done that now.

@natedog, that’s fine, I don’t need to boot from it. Just want to use it for storage. Happy to boot off the eMMC.

#25

@Peter.Wang I did in fact order one of those serial cables. Now I can’t get any legible output in minicom. I’m using dev/ttyUSB0 as the device and all the other settings from the page you shared. All I see in minicom is this:

���n[�;��������|?��У��ֲ������7ۿ�]�������5��~��=�����6�֢�l�|����l����׶�ߚ�            �登

The footer of the minicom UI says this:

CTRL-A Z for help | 115200 8N1 | NOR | Minicom 2.8 | VT102 | Offline | ttyUSB0
#26

You’re using baudrate 115200 so this output is reasonable.

#27

please use baudrate 1500000

#28

Done, but output is the same.

minicom is not picking up my ~/.minirc.radxa file, possibly because I still have to run it using sudo for it to work, but these other settings in the file seem to be what I’m getting anyway:

   pu bits             8
   pu parity           N
   pu stopbits         1
   pu rtscts           No
#29

can you show how you connect the serial cable to the rock5B ?

#30

Sure. images attached for the docs, and for mine.

mine
mine.jpg1575×1724 336 KB
docs.webp (56.3 KB)

#31

Yes, It’s right,nvme works fine on pc but not work for rock5B, and you still can’t get the serial log.
you can require a RMA.

#32

I had a similar problem when using a nvme drive although I wasn’t using the Radxa power supply. I have a Dell 130w power adaptor & Apple 96W both should be able to supply enough power for the board. With the nvme attached both resulted in continuous reboot, with dell resulting in a complete shutdown after a number of reboots. My nvme is rated 3.3v@2.8A , I suspect the nvme power draw is interfering with the USB PD negotiations. Booting from a SD card is fine with the Dell adaptor. Having spent a fair amount of time debugging this the symptoms are either uboot fails to complete or if it completes then the kernel fails. In my opinion this a bad board design by Radxa which they aren’t addressing and leaving it to the buyer to debug.

To resolve the issue use a non PD adaptor. I had a cheap 45w adapter with the USB A ports rated at 3.1A, using a usb A to C cable the board boots consistently with nvme drive attached on every power up or reboot. However 5@3.1A won’t be enough when the board is under considerable load. In the end as mentioned in other threads the best option is to use a DC to USB-C adapter. Fortunately I had a DC 12v@5A adapter to hand and stress tested it works under considerable load with the nvme attached. In my opinion to help address these issues I think Radxa should include a dc to usb c adapter with the board.

1 Like
#33

Safe mode means that debian in it’s default configuration will fail to boot if there’s a local drive in fstab that can’t be mounted, as the systemd target for Multi-User depends on mount to be successful. Nothing to be concerned about.

The way to avoid this is add the nofail parameter to the fstab line after auto(as in auto, nofail)

#34

Same issue.
Rock 5B boot with SD Card only if no SSD NMVe M2 plugged. When i plugged it, boot seems to fall in a infinite loop with the BLUE Led blinking.
@eliotstock, did you solve your issue with it ? how ?
I read everywhere that’s a possible power supply issue, but i use the official one (35W : https://fr.rs-online.com/web/p/alimentations-rock-sbc/2436356)
My SSD : Crucial P3 M2 2280 4To… seems already use without trouble by a user (ROCK 5B - list of supported NVME)

#35

I might be wrong but I have the dark memory that for this reason Armbian patched the u-boot. Mine boots Armbian fine from NVME with the official Radxa powersuply.

#36

Just tried with a Rasberry Pi official Power Supply (5V/3A) et it’s work well.
Eventually, Radxa Official Power Supply sucks, that’s the conclusion for my case.

1 Like
#37

I have an issue with NVMe trying to use DP Alt Mode from the USB-C port.

The NVMe actually works quite well for me when I use the Apple 20W adapter (9V).

But when I plugged this into a USB-C (Hyperdrive MacBook hub) hub to get video output on USB-C (I tried both a 65W PD Charger and the Apple 20W Charger), the light just slowly (every 1-2 seconds) alternately turning blue and green. It stays at 5V and does not trigger PD.

This issue does not happen on the microSD Card. I can use PD with DP Alt with Mouse and Keyboard all works well with an sd card.

#38

Same analysis, the official Radxa UBS-C PD not work with NMVe plugged. I tried the official USB-C Raspberry Pi 5 PD (27W), same issue. I should take a classic 5V/3A power supply (official RPi 4 PD) to success boot with NMVe SSD disk plugged.

There is a clear bug with the u-boot on Rock 5B card, where is the Radxa dev team to argue @radxa ?
When will we have got a official patch ?

#39

They’ve been trying (maybe) to fix it for over a year now. It’s not just u-boot, it’s a hardware issue.