Re: [PATCH] mfd: cros ec: spi: Fix "in progress" error signaling

[Enric Balletbo Serra]

Dear all,

Cc'ing some more chromium folks.

2017-11-29 13:11 GMT+01:00 Lee Jones <lee.jones@xxxxxxxxxx>:
> On Wed, 27 Sep 2017, Shawn Nematbakhsh wrote:
>
>> For host commands that take a long time to process, cros ec can return
>> early by signaling a EC_RES_IN_PROGRESS result. The host must then poll
>> status with EC_CMD_GET_COMMS_STATUS until completion of the command.
>>
>> None of the above applies when data link errors are encountered. When
>> errors such as EC_SPI_PAST_END are encountered during command
>> transmission, it usually means the command was not received by the EC.
>> Treating such errors as if they were 'EC_RES_IN_PROGRESS' results is
>> almost always the wrong decision, and can result in host commands
>> silently being lost.
>>
>> Reported-and-tested-by: Jon Hunter <jonathanh@xxxxxxxxxx>
>> Signed-off-by: Shawn Nematbakhsh <shawnn@xxxxxxxxxxxx>
>> ---
>>  drivers/mfd/cros_ec_spi.c | 52 ++++++++++++++++++++++-------------------------
>>  1 file changed, 24 insertions(+), 28 deletions(-)
>
> Applied, thanks.
>

This patch is a bit old and is already applied but I would like to
discuss an issue that Tomeu found playing with kernelci and a Veyron
Jaq attached to our lab.

Seems that after this patch, the veyron jaq spits out lots of spi
tranfer error messages. See full log here [1]

 cros-ec-spi spi0.0: spi transfer failed: -121
 cros-ec-spi spi0.0: Command xfer error (err:-121)
 cros-ec-i2c-tunnel ff110000.spi:ec@0:i2c-tunnel: Error transferring
EC i2c message -121

The issue is random, not always happens, but when happens makes
cros-ec unusable. Reproduce the issue is easier if CONFIG_SMP is not
set.

What happens is that the master starts the transmission and the
cros-ec returns an spi error event (EC_SPI_RX_BAD_DATA  - data is
0xfb). The difference between after and before the patch is that now
the cros-ec does not recover, whereas without this patch after some
tries it succeeds (note that before the patch the affected code
returned -EAGAIN whereas now returns -EREMOTEIO)

I think that reverting this patch is NOT the solution, but I don't
have enough knowledge yet to understand why the cros-ec fails. I need
to look at the cros-ec firmware to understand better what is
happening, but meanwhile, thoughts? clues?

An important note is that I did not reproduce the issue on a Veyron
Minnie, even with CONFIG_SMP disabled.

[1] https://lava.collabora.co.uk/scheduler/job/1085493#L905

Best regards,
  Enric

> --
> Lee Jones
> Linaro STMicroelectronics Landing Team Lead
> Linaro.org â Open source software for ARM SoCs
> Follow Linaro: Facebook | Twitter | Blog