Den 30.06.2021 21.46, skrev Noralf Trønnes:
>
>
> Den 30.06.2021 21.13, skrev Linus Walleij:
>> This creates a macro wrapping mipi_dbi_command() such that we get
>> some explicit error reporting if something goes wrong.
>>
>> Cc: Noralf Trønnes <nor...@tronnes.org>
>> Suggested-by: Douglas Anderson <diand...@chromium.org>
>> Signed-off-by: Linus Walleij <linus.wall...@linaro.org>
>> ---
>> drivers/gpu/drm/panel/panel-samsung-db7430.c | 66 +++++++++++---------
>> 1 file changed, 36 insertions(+), 30 deletions(-)
>>
>> diff --git a/drivers/gpu/drm/panel/panel-samsung-db7430.c
>> b/drivers/gpu/drm/panel/panel-samsung-db7430.c
>> index fe58263bd9cd..c42d43ab6a4f 100644
>> --- a/drivers/gpu/drm/panel/panel-samsung-db7430.c
>> +++ b/drivers/gpu/drm/panel/panel-samsung-db7430.c
>> @@ -90,9 +90,17 @@ static inline struct db7430 *to_db7430(struct drm_panel
>> *panel)
>> return container_of(panel, struct db7430, panel);
>> }
>>
>> +#define db7430_command(db, cmd, seq...) \
>> +({ \
>> + struct mipi_dbi *dbi = &db->dbi; \
>> + int ret; \
>> + ret = mipi_dbi_command(dbi, cmd, seq); \
>> + if (ret) \
>> + dev_err(db->dev, "failure in writing command %#02x\n", cmd); \
>> +})
>> +
>
> I did a grep and there's only one mipi_dbi_command() caller that checks
> the return code: mipi_dbi_poweron_reset_conditional().
>
> Can you add the error reporting to mipi_dbi_command() instead? Instead
> of drivers adding their own similar macros.
>
> In that case I think you need to add a plain 'ret' at the end of the
> macro for it to return the error code.
>
I think the error reporting needs to ratelimited if added to
mipi_dbi_command() because worst case a blinking fbcon cursor will
trigger 10 errors a second for the framebuffer drivers.
Noralf.
