spi: atmel: Do not cancel a transfer upon any signal

The intended move from wait_for_completion_*() to
wait_for_completion_interruptible_*() was to allow (very) long spi memory
transfers to be stopped upon user request instead of freezing the
machine forever as the timeout value could now be significantly bigger.

However, depending on the user logic, applications can receive many
signals for their own "internal" purpose and have nothing to do with the
requested kernel operations, hence interrupting spi transfers upon any
signal is probably not a wise choice. Instead, let's switch to
wait_for_completion_killable_*() to only catch the "important"
signals. This was likely the intended behavior anyway.

Fixes: e0205d6 ("spi: atmel: Prevent false timeouts on long transfers")
Cc: stable@vger.kernel.org
Reported-by: Ronald Wahl <ronald.wahl@raritan.com>
Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Link: https://lore.kernel.org/r/20231127095842.389631-1-miquel.raynal@bootlin.com
Signed-off-by: Mark Brown <broonie@kernel.org>

Author: miquelraynal

drivers/spi/spi-atmel.c

@@ -1336,8 +1336,8 @@ static int atmel_spi_one_transfer(struct spi_controller *host,
 		}
 
 		dma_timeout = msecs_to_jiffies(spi_controller_xfer_timeout(host, xfer));
-		ret_timeout = wait_for_completion_interruptible_timeout(&as->xfer_completion,
-									dma_timeout);
+		ret_timeout = wait_for_completion_killable_timeout(&as->xfer_completion,
+								   dma_timeout);
 		if (ret_timeout <= 0) {
 			dev_err(&spi->dev, "spi transfer %s\n",
 				!ret_timeout ? "timeout" : "canceled");