Re: [PATCH v2 03/21] of: Rename of_modalias_node()

From: Rob Herring
Date: Tue Mar 07 2023 - 19:19:10 EST


On Tue, Mar 07, 2023 at 05:53:41PM +0100, Miquel Raynal wrote:
> This helper does not produce a real modalias, but tries to get the
> "product" compatible part of the "vendor,product" compatibles only. It
> is far from creating a purely useful modalias string and does not seem
> to be used like that directly anyway, so let's try to give this helper a
> more meaningful name before moving there a real modalias helper (already
> existing under of/device.c).
>
> Also update the various documentations to refer to the strings as
> "aliases" rather than "modaliases" which has a real meaning in the Linux
> kernel.
>
> There is no functional change.
>
> Cc: Rafael J. Wysocki <rafael@xxxxxxxxxx>
> Cc: Len Brown <lenb@xxxxxxxxxx>
> Cc: Maarten Lankhorst <maarten.lankhorst@xxxxxxxxxxxxxxx>
> Cc: Maxime Ripard <mripard@xxxxxxxxxx>
> Cc: Thomas Zimmermann <tzimmermann@xxxxxxx>
> Cc: Sebastian Reichel <sre@xxxxxxxxxx>
> Cc: Wolfram Sang <wsa@xxxxxxxxxx>
> Cc: Mark Brown <broonie@xxxxxxxxxx>
> Signed-off-by: Miquel Raynal <miquel.raynal@xxxxxxxxxxx>
> ---
> drivers/acpi/bus.c | 7 ++++---
> drivers/gpu/drm/drm_mipi_dsi.c | 2 +-
> drivers/hsi/hsi_core.c | 2 +-

These should not have been using this function. The matching on just the
product was a relic from I2C and SPI which we don't want to propogate.
No clue why ACPI needed it...

If you respin or want to fixup while applying, can you add a kerneldoc
comment to not add new users of the function. Not that anyone will
follow that... :(

Reviewed-by: Rob Herring <robh@xxxxxxxxxx>

> drivers/i2c/busses/i2c-powermac.c | 2 +-
> drivers/i2c/i2c-core-of.c | 2 +-
> drivers/of/base.c | 15 ++++++++-------
> drivers/spi/spi.c | 4 ++--
> include/linux/of.h | 2 +-
> 8 files changed, 19 insertions(+), 17 deletions(-)