I was doing some testing and it looks latest u-boot is broken for Allwinner H3 (no boot), A20 (Banana no network), I assume A10 too. There was some RFC upstream, but I did not go further into details ... Shall we rather rollback to 2017.03?

1 minute ago, Igor said:

Allwinner H3 (no boot)

Which boards exactly? I remember only issues on Zero 256MB

 

1 minute ago, Igor said:

A20 (Banana no network)

Both kernels or only legacy? 100MBit EMAC, 100/1000 GMAC or both?

Opi One/Lite/PC+ , the latter gives: 

U-Boot SPL 2017.05-armbian (May 29 2017 - 19:24:06)
DRAM: 1024 MiB
Trying to boot from MMC1
MMC: no card present
mmc_init: -123, time 2
spl: mmc init failed with error: -123
SPL: failed to boot from all boot devices
### ERROR ### Please RESET the board ###

Banana PRO gmac, legacy kernel IIRC

1 minute ago, Igor said:

Trying to boot from MMC1 MMC: no card present

Huh? And older releases boot fine?

7 minutes ago, zador.blood.stained said:

Huh? And older releases boot fine?

 

Yes. I just boot PC+ with 2017.03 and it boots fine. Slim chances that it was not recognised before and it is now. 

 

Could be SD card issue, but now I have 2017.03 and it works. Will do more tests ... 

 

Edit: Banana PRO is real problem, reported by some user few days ago.

Orange Pi PC+ boots fine ... except boot logo is not working. GMAC at A20 than. Will check some H3 gmac if there are any troubles.
 

Spoiler

U-Boot SPL 2017.05-armbian (May 29 2017 - 20:22:28)
DRAM: 1024 MiB
Trying to boot from MMC1

U-Boot 2017.05-armbian (May 29 2017 - 20:22:28 +0200) Allwinner Technology
CPU:   Allwinner H3 (SUN8I 1680)
Model: Xunlong Orange Pi PC Plus
I2C:   ready
DRAM:  1 GiB
MMC:   SUNXI SD/MMC: 0, SUNXI SD/MMC: 1
*** Warning - bad CRC, using default environment
In:    serial
Out:   serial
Err:   serial
Net:   phy interface0
eth0: ethernet@1c30000
Hit any key to stop autoboot:  0 
38518 bytes read in 138 ms (272.5 KiB/s)
Unknown command 'bmp' - try 'help'
switch to partitions #0, OK
mmc0 is current device
Scanning mmc 0:1...
Found U-Boot script /boot/boot.scr
3565 bytes read in 186 ms (18.6 KiB/s)
## Executing script at 43100000
U-boot loaded from SD
Boot script loaded from mmc

 

 

 

On 29.05.2017 at 9:27 PM, Igor said:

GMAC at A20 than. Will check some H3 gmac if there are any troubles.

I remembered this: https://irclog.whitequark.org/linux-sunxi/2017-05-26#19471905

So GPIO related issues in u-boot would explain broken Ethernet when legacy kernel relies on powering GMAC from u-boot. Can you test SATA too?

 

This commit has a list of boards with SATA powered by GPIO: http://git.denx.de/?p=u-boot.git;a=commit;h=d7b560e6653b248a1920cbd2c5ac393d42e30b61

 

This commit has a list of boards with GMAC powered by GPIO: http://git.denx.de/?p=u-boot.git;a=commit;h=f5fd78860ab4dbb3bc14ead4f14433f04d6e5cea

 

I assume mainline kernels should be fine mostly if DT has proper config for those pins, and Ethernet can also be tested in u-boot without booting the kernel.

 

Edit:

GPIO works on H3. Can't check on A20

 

Edit 2:

May be retated to possible CONFIG_DM_GPIO differences between different SoCs - DM_GPIO pulls DM based sunxi_name_to_gpio() which doesn't look right to me.