On 11/06/2025 at 01:05, Bartosz Golaszewski wrote:

On Tue, Jun 10, 2025 at 5:48 PM Vincent Mailhol
[off-list ref] wrote:

quoted

On 10/06/2025 at 23:05, Bartosz Golaszewski wrote:

quoted

On Tue, Jun 10, 2025 at 3:55 PM Vincent Mailhol
[off-list ref] wrote:

quoted

On 10/06/2025 at 21:37, Bartosz Golaszewski wrote:

quoted

From: Bartosz Golaszewski <redacted>

struct gpio_chip now has callbacks for setting line values that return
an integer, allowing to indicate failures. Convert the driver to using
them.

Signed-off-by: Bartosz Golaszewski <redacted>

                                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

This does not match the address with which you sent the patch: brgl@bgdev.pl

quoted

---
 drivers/net/can/spi/mcp251x.c | 16 ++++++++++------
 1 file changed, 10 insertions(+), 6 deletions(-)

diff --git a/drivers/net/can/spi/mcp251x.c b/drivers/net/can/spi/mcp251x.c
index ec5c64006a16f703bc816983765584c5f3ac76e8..7545497d14b46c6388f3976c2bf7b9a99e959c1e 100644
--- a/drivers/net/can/spi/mcp251x.c
+++ b/drivers/net/can/spi/mcp251x.c

@@ -530,8 +530,8 @@ static int mcp251x_gpio_get_multiple(struct gpio_chip *chip,
      return 0;
 }

-static void mcp251x_gpio_set(struct gpio_chip *chip, unsigned int offset,
-                          int value)
+static int mcp251x_gpio_set(struct gpio_chip *chip, unsigned int offset,
+                         int value)
 {
      struct mcp251x_priv *priv = gpiochip_get_data(chip);
      u8 mask, val;

@@ -545,9 +545,11 @@ static void mcp251x_gpio_set(struct gpio_chip *chip, unsigned int offset,

      priv->reg_bfpctrl &= ~mask;
      priv->reg_bfpctrl |= val;
+
+     return 0;

mcp251x_gpio_set() calls mcp251x_write_bits() which calls mcp251x_spi_write()
which can fail.

For this change to really make sense, the return value of mcp251x_spi_write()
should be propagated all the way around.

I don't know this code so I followed the example of the rest of the
codebase where the result of this function is never checked - even in
functions that do return values. I didn't know the reason for this and
so didn't want to break anything as I have no means of testing it.

The return value of mcp251x_spi_write() is used in mcp251x_hw_reset(). In other
locations, mcp251x_spi_write() is only used in functions which return void, so
obviously, the return value is not checked.

Wait, after a second look GPIO callbacks (including those that return
a value like request()) use mcp251x_write_bits() which has no return
value.

Yes. Read again my first message:

  mcp251x_gpio_set() calls mcp251x_write_bits() which calls mcp251x_spi_write()
  which can fail.

My point is that the grand father can fail.

It probably should propagate what mcp251x_spi_write() returns

Exactly what I asked for :)

but that's material for a different series.

Why? Are you going to do this other series?

If the answer is no, then please just do it here. Propagating the error in
mcp251x_write_bits() is a three line change. Am I asking for too much?

The goal of this one is to
use the new setters treewide and drop the old ones from struct
gpio_chip.

Yours sincerely,
Vincent Mailhol