spi: Make error handling of gpiod_count() call cleaner

Each time we call spi_get_gpio_descs() the num_chipselect is overwritten
either by new value or by the old one. This is an extra operation in case
gpiod_count() returns an error. Besides that it slashes the error handling
of gpiod_count().

Refactor the code to make error handling of gpiod_count() call cleaner.

Note, that gpiod_count() never returns 0, take this into account as well.
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Link: https://lore.kernel.org/r/20210420164040.40055-1-andriy.shevchenko@linux.intel.comSigned-off-by: Mark Brown <broonie@kernel.org>

drivers/spi/spi.c

@@ -2551,13 +2551,14 @@ static int spi_get_gpio_descs(struct spi_controller *ctlr)
 	unsigned int num_cs_gpios = 0;
 
 	nb = gpiod_count(dev, "cs");
-	ctlr->num_chipselect = max_t(int, nb, ctlr->num_chipselect);
-
+	if (nb < 0) {
 		/* No GPIOs at all is fine, else return the error */
-	if (nb == 0 || nb == -ENOENT)
+		if (nb == -ENOENT)
 			return 0;
-	else if (nb < 0)
 		return nb;
+	}
+
+	ctlr->num_chipselect = max_t(int, nb, ctlr->num_chipselect);
 
 	cs = devm_kcalloc(dev, ctlr->num_chipselect, sizeof(*cs),
 			  GFP_KERNEL);