I just updated two Orange Pi One boards and rebooted them to get to the new kernel only to have them boot loop.

 

I know, development, not supported.   I'm not asking for a solution, I'm just alerting people to the situation so they can avoid messing up their boxes.

 

Console log attached if you're curious.

orangepione.txt

Edited September 20, 20179 yr by willmore
Forgot to mention the purpose of the attached file.

Can you reproduce it on a fresh image or on other H3 boards (if you can test on any)? Which overlay options did you add? (i.e. full contents of /boot/armbianEnv.txt would be ideal)

Okay, one of them--that I was using for SPI testing--has:

verbosity=7
console=both
disp_mode=1920x1080p60
rootdev=UUID=1ab2a8dc-160b-474b-ab35-ba43e08d90ef
rootfstype=ext4
overlay_prefix=sun8i-h3
overlays=spi-spidev
param_spidev_spi_bus=0
param_spidev_max_freq=100000000 

The second Orange Pi One has the same config.  I'm going to pull the spi overlay and see if that changes anything.

 

I have an up to date Orange Pi PC which is H3 based and it rebooted fine after the recent update.  Would that be sufficient?  If not, can you point me to one you would like me to test?  The https://dl.armbian.com/orangepione/ directory doesn't seem to have anything new in it.

Taking out the last four line fixed it.  DTB format changes or something, maybe?

 

Let me know if you want any more testing, I'm here to help if I can be of use.

8 minutes ago, willmore said:

DTB format changes or something, maybe?

No, most likely some dtc or environment commands changes that were made recently. Now that we have a test case I will try to bisect the sources.

... and we have some rather strange results:

this:

fdt set /soc/${tmp_spi_path}/spiflash spi-max-frequency "<${param_spinor_max_freq}>"

is broken after this commit: http://git.denx.de/?p=u-boot.git;a=commit;h=7cbc12415dad707b94d755311126059e5cdc65e3

... and (hopefully) fixed: a717f1a

 

Funny how the problematic code here is identical to the "bad" example here.

ShallI update and retest?

9 hours ago, willmore said:

ShallI update and retest?

Yes, please. While this particular case was fixed, there may be similar issues somewhere else in the code, so more tests is better.

Okay, I put the changes back in and rebooted.  It works just fine.  I haven't verified functionality of SPI, though.  Kernel log says:

[    9.546118] spidev spi0.0: probing from DT

 

So, it looks like it's working.  Need to me do anything else?

7 hours ago, willmore said:

Need to me do anything else?

No, at least this seems to be fixed. In any case overlays functionality affects stable branches too, so it's not "development, not supported", so thanks for the bug report.

Thanks for pulling in the fix.  I'm glad I could help even if it's just me stumbling in the dark and running into things.