07:09 <chewitt> tomeu: old spi instructions https://web.archive.org/web/20210422170113/https://docs.khadas.com/vim3/BootFromSpiFlash.html

07:09 <chewitt> but I think this is written for vendor u-boot

08:07 <narmstrong> the sf instructions should still work, but to enable the SPI boot some i2c commands should be executed manually, similar to the way we enable PCIe at https://u-boot.readthedocs.io/en/latest/board/amlogic/khadas-vim3.html#pcie-setup

08:09 <narmstrong> MCU regs are documented at https://dl.khadas.com/hardware/VIM3/MCU/VIM3_MCU_REG_EN.pdf

08:09 <narmstrong> BOOT_MODE is reg 0x22, 0 for SPI, 1 for eMMC

09:09 <tomeu> chewitt: narmstrong: we have found a difference in the power setup of these boards that could explain the situation

09:09 <tomeu> 2 work reliably and are on a different power adapter than the other 5 that work unreliably

14:19 <chewitt> ahh, power is definitely a candidate for causing weird issues

14:20 <chewitt> lower amp rating on the 5x unreliable ones?

15:20 <tomeu> it's quite weird, the unreliable ones are using Khadas adapter, and the FW reports to use the 12V/2A mode

15:20 <tomeu> the reliable ones uses some charger from Dell and the FW chooses the 9V/3A level

15:21 <tomeu> I run the board on my desk with a !PD charger that is 5V/1A (and it's reliable) :)

23:37 <rockosov> narmstrong: jbrunet mturquette khilman Hello guys! My name is Dmitry, I'm from SberDevices Kernel team. We are maintaining and developing Linux Kernel on our own consumer smart devices, which are all based on Amlogic SoCs.

23:37 <rockosov> I'm really happy to join linux-amlogic community, you did really hard work for upstreaming meson stuff to kernel and uboot. So cool honestly!

23:37 <rockosov> We have our internally developed kernel code for several Amlogic SoCs, so we want to share it with the community and help you to make meson upstream support more robust and widespread. Currently, we are preparing a few amount of A113L (aka A1) drivers. We will send them soon.

23:37 <rockosov> For now, I just sent the A1 clock peripherals and A1 PLL clock controllers drivers to the LKML, please look into it if possible at link https://lore.kernel.org/linux-amlogic/20221201225703.6507-1-ddrokosov@sberdevices.ru/. But I'm faced with one little problem. Hope you can explain me, how should I do it correctly.

23:37 <rockosov> I've found several versions of A1 clkc drivers from Jian Hu Amlogic engineer, the last one version was a couple years ago. I've reworked the last (v7) version and didn't make right choice, how do I have to send it to LKML. I thought about two options:

23:37 <rockosov> 1) Merge all my changes with Jian Hu patches and leave Jian Hu SoB

23:37 <rockosov> 2) Don't touch Jian Hu v7 patches and apply my changes above.

23:37 <rockosov> I have chosen the second option. I don't know, if it's the right way or not. Please correct me if possible and I will send the new patch series if it's required.