linux-kernel - [PATCH 5.2 49/66] Revert "gpio/spi: Fix spi-gpio regression on active high CS"

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Fri, 26 Jul 2019 17:24:48 +0200
From:   Greg Kroah-Hartman <gregkh@...uxfoundation.org>
To:     linux-kernel@...r.kernel.org
Cc:     Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        stable@...r.kernel.org, linux-spi@...r.kernel.org,
        Linus Walleij <linus.walleij@...aro.org>
Subject: [PATCH 5.2 49/66] Revert "gpio/spi: Fix spi-gpio regression on active high CS"

From: Linus Walleij <linus.walleij@...aro.org>

commit da7f134972f473053ea9d721a1d8397546476dc4 upstream.

This reverts commit fbbf145a0e0a0177e089c52275fbfa55763e7d1d.

It seems I was misguided in my fixup, which was working at the
time but did not work on the final v5.2.

The patch tried to avoid a quirk the gpiolib code not to treat
"spi-gpio" CS gpios "special" by enforcing them to be active
low, in the belief that since the "spi-gpio" driver was
parsing the device tree on its own, it did not care to inspect
the "spi-cs-high" attribute on the device nodes.

That's wrong. The SPI core was inspecting them inside the
of_spi_parse_dt() funtion and setting SPI_CS_HIGH on the
nodes, and the driver inspected this flag when driving the
line.

As of now, the core handles the GPIO and it will consistently
set the GPIO descriptor to 1 to enable CS, strictly requireing
the gpiolib to invert it. And the gpiolib should indeed
enforce active low on the CS line.

Device trees should of course put the right flag on the GPIO
handles, but it used to not matter. If we don't enforce active
low on "gpio-gpio" we may run into ABI backward compatibility
issues, so revert this.

Cc: linux-spi@...r.kernel.org
Cc: stable@...r.kernel.org
Signed-off-by: Linus Walleij <linus.walleij@...aro.org>
Link: https://lore.kernel.org/r/20190715204529.9539-1-linus.walleij@linaro.org
Signed-off-by: Greg Kroah-Hartman <gregkh@...uxfoundation.org>

---
 drivers/gpio/gpiolib-of.c |    9 +--------
 1 file changed, 1 insertion(+), 8 deletions(-)

--- a/drivers/gpio/gpiolib-of.c
+++ b/drivers/gpio/gpiolib-of.c
@@ -118,15 +118,8 @@ static void of_gpio_flags_quirks(struct
 	 * Legacy handling of SPI active high chip select. If we have a
 	 * property named "cs-gpios" we need to inspect the child node
 	 * to determine if the flags should have inverted semantics.
-	 *
-	 * This does not apply to an SPI device named "spi-gpio", because
-	 * these have traditionally obtained their own GPIOs by parsing
-	 * the device tree directly and did not respect any "spi-cs-high"
-	 * property on the SPI bus children.
 	 */
-	if (IS_ENABLED(CONFIG_SPI_MASTER) &&
-	    !strcmp(propname, "cs-gpios") &&
-	    !of_device_is_compatible(np, "spi-gpio") &&
+	if (IS_ENABLED(CONFIG_SPI_MASTER) && !strcmp(propname, "cs-gpios") &&
 	    of_property_read_bool(np, "cs-gpios")) {
 		struct device_node *child;
 		u32 cs;