Re: reset-loop on Odroid U2: cros-ec communications failure

[Joost van Zwieten <joost@diskos.nl>]

On Fri, Jan 20, 2023 at 15:11, Simon Glass <sjg@chromium.org> wrote:
> Hi Joost,
> 
> On Fri, 20 Jan 2023 at 15:00, Joost van Zwieten <joost@diskos.nl> 
> wrote:
>> 
>> 
>> 
>>  On Fri, Jan 13, 2023 at 11:00, Simon Glass <sjg@chromium.org> wrote:
>>  > Hi Joost,
>>  >
>>  > On Mon, 9 Jan 2023 at 06:31, Joost van Zwieten <joost@diskos.nl>
>>  > wrote:
>>  >>
>>  >>  Dear maintainers,
>>  >>
>>  >>  U-Boot on my Odroid U2 fails to initialize as of commit
>>  >>  e44d7e73fe0d649693d8d0a110cd7632bc919273 (introduced after
>>  >> v2022.10). I
>>  >>  get the following output:
>>  >>
>>  >>      U-Boot 2022.10-00565-ge44d7e73fe-dirty (Jan 09 2023 - 
>> 01:04:19
>>  >>  +0100)
>>  >>
>>  >>      CPU: Exynos4412 @ 1 GHz
>>  >>      Model: Odroid based on Exynos4412
>>  >>      Type: u3
>>  >>      DRAM: 2 GiB
>>  >>      Core: 101 devices, 12 uclasses, devicetree: separate
>>  >>      LDO20@VDDQ_EMMC_1.8V: set 1800000 uV; enabling
>>  >>      LDO22@VDDQ_EMMC_2.8V: set 2800000 uV; enabling
>>  >>      LDO21@TFLASH_2.8V: set 2800000 uV; enabling
>>  >>      MMC: SAMSUNG SDHCI: 2, EXYNOS DWMMC: 0
>>  >>      Loading Environment from MMC... *** Warning - bad CRC, using
>>  >>  default environment
>>  >>
>>  >>      cros-ec communications failure -96
>>  >>
>>  >>      Please reset with Power+Refresh
>>  >>
>>  >>      Cannot init cros-ec device
>>  >>      resetting ...
>>  >>
>>  >>  This repeats indefinitely. I'm building U-Boot using the default
>>  >> config
>>  >>  for board `odroid` and boot the Odroid from an SD card with 
>> firmware
>>  >>  (`bl1`, `bl2` and `tzsw`) from Hardkernel's clone of the u-boot
>>  >>  repository [1]. There are no peripherals connected apart from 
>> the SD
>>  >>  card and the serial console. Please let me know if you need more
>>  >>  information.
>>  >
>>  > board_late_init() is presumably enabled with 
>> CONFIG_BOARD_LATE_INIT
>>  > and the code in board/samsung/common/board.c is being called.
>>  >
>>  > -96 is -EPFNOSUPPORT which means that the CROS_EC uclass is 
>> missing.
>>  >
>>  > So either enable CROS_EC so you have the uclass, or disable
>>  > BOARD_LATE_INIT so it doesn't matter.
>>  >
>>  >>
>>  >>  Best, Joost
>>  >>
>>  >>  [1]:
>>  >>
>>  >> 
>> https://github.com/hardkernel/u-boot/tree/33e05ffb159141b640571e91470172d83a2a1273/sd_fuse
>>  >>
>>  >>
>>  >
>>  > Regards,
>>  > Simon
>> 
>>  Hi Simon,
>> 
>>  Thanks. Both solutions work. With `CROS_EC` enabled
>>  `uclass_first_device_err(UCLASS_CROS_EC, &dev)` returns `ENODEV`. 
>> Since
>>  `board_late_init()` seems to do legitimate work for this board
>>  (defining the `mmcbootdev` environment variable) I think enabling
>>  `CROS_EC` is the way to go. Should this be part of the defconfig for
>>  odroid? If so, I can send a patch.
> 
> Yes that seems OK to me.
> 
> Regards,
> Simon

I've send a patch.

I'm still a bit puzzled by this problem, though. Is it correct that 
`uclass_first_device_err(UCLASS_CROS_EC, ...)` always returns 
`-EPFNOSUPPORT` when `CROS_EC` is disabled in the `.config`? If so, I 
don't see how U-Boot is going to function properly on any board that 
hits `board_late_init()` (board/samsung/common/board.c) with 
`BOARD_LATE_INIT=y` and `CROS_EC` undefined and perhaps the best 
solution is to run the CROS_EC related statements only if `CROS_EC` is 
actually defined. I'm not sure about the implications of this proposal 
further down the road, though.

Best, Joost