Skip to contentSkip to conversation

serial: mctrl_gpio: Check if GPIO property exisits before requesting it

PATCH In review Stefan Roese 4 participants 8 messages Jun 20, 2019 – Jul 25, 2019

Subsystems (4)
lore
Share & tools
lore
#0 lore raw

PATCH v9 1/2 serial: mctrl_gpio: Check if GPIO property exisits before requesting it

This patch adds a check for the GPIOs property existence, before the GPIO is requested. This fixes an issue seen when the 8250 mctrl_gpio support is added (2nd patch in this patch series) on x86 platforms using ACPI.

Here Mika's comments from 2016-08-09:

"
I noticed that with v4.8-rc1 serial console of some of our Broxton systems does not work properly anymore. I'm able to see output but input does not work.

I bisected it down to commit 4ef03d328769eddbfeca1f1c958fdb181a69c341 ("tty/serial/8250: use mctrl_gpio helpers").

The reason why it fails is that in ACPI we do not have names for GPIOs (except when _DSD is used) so we use the "idx" to index into _CRS GPIO resources. Now mctrl_gpio_init_noauto() goes through a list of GPIOs calling devm_gpiod_get_index_optional() passing "idx" of 0 for each. The UART device in Broxton has following (simplified) ACPI description:

    Device (URT4)
    {
        ...
        Name (_CRS, ResourceTemplate () {
            GpioIo (Exclusive, PullDefault, 0x0000, 0x0000, IoRestrictionOutputOnly,
                    "\\_SB.GPO0", 0x00, ResourceConsumer)
            {
                0x003A
            }
            GpioIo (Exclusive, PullDefault, 0x0000, 0x0000, IoRestrictionOutputOnly,
                    "\\_SB.GPO0", 0x00, ResourceConsumer)
            {
                0x003D
            }
        })

In this case it finds the first GPIO (0x003A which happens to be RX pin for that UART), turns it into GPIO which then breaks input for the UART device. This also breaks systems with bluetooth connected to UART (those typically have some GPIOs in their _CRS).

Any ideas how to fix this?

We cannot just drop the _CRS index lookup fallback because that would break many existing machines out there so maybe we can limit this to only DT enabled machines. Or alternatively probe if the property first exists before trying to acquire the GPIOs (using device_property_present()).
"

This patch implements the fix suggested by Mika in his statement above.

Cc Mika Westerberg <[email protected]>
Cc Andy Shevchenko <[email protected]>
Cc Yegor Yefremov <[email protected]>
Cc Greg Kroah-Hartman <[email protected]>
Cc Giulio Benetti <[email protected]>

---
v9:
- Rebased on top of "tty-next", patch 2/3 dropped as its already applied

v8:
- Rebased on top of "tty-next"

v7:
- Include <linux/property.h> to fix compile breakage on OMAP

v6:
- No change

v5:
- Simplified the code a bit (Andy)
- Added gpio_str == NULL handling (Andy)

v4:
- Add missing free() calls (Johan)
- Added Mika's reviewed by tag
- Added Johan to Cc

v3:
- No change

v2:
- Include the problem description and analysis from Mika into the commit text, as suggested by Greg.

 drivers/tty/serial/serial_mctrl_gpio.c | 14 ++++++++++++++
 1 file changed, 14 insertions(+)
drivers/tty/serial/serial_mctrl_gpio.c +14 -0
index 39ed56214cd3..2b400189be91 100644
@@ -12,6 +12,7 @@
 #include <linux/termios.h>
 #include <linux/serial_core.h>
 #include <linux/module.h>
+#include <linux/property.h>
 
 #include "serial_mctrl_gpio.h"
 
@@ -116,6 +117,19 @@ struct mctrl_gpios *mctrl_gpio_init_noauto(struct device *dev, unsigned int idx)
 
 	for (i = 0; i < UART_GPIO_MAX; i++) {
 		enum gpiod_flags flags;
+		char *gpio_str;
+		bool present;
+
+		/* Check if GPIO property exists and continue if not */
+		gpio_str = kasprintf(GFP_KERNEL, "%s-gpios",
+				     mctrl_gpios_desc[i].name);
+		if (!gpio_str)
+			continue;
+
+		present = device_property_present(dev, gpio_str);
+		kfree(gpio_str);
+		if (!present)
+			continue;
 
 		if (mctrl_gpios_desc[i].dir_out)
 			flags = GPIOD_OUT_LOW;
Signature · 1 line
2.22.0
#1 lore raw

PATCH v9 2/2 tty/serial/8250: use mctrl_gpio helpers

From: Yegor Yefremov <[email protected]>

This patch permits the usage for GPIOs to control
the CTS/RTS/DTR/DSR/DCD/RI signals.

Changed by Stefan:
Only call mctrl_gpio_init(), if the device has no ACPI companion device to not break existing ACPI based systems. Also only use the mctrl_gpio_ functions when "gpios" is available.

Use MSR / MCR <-> TIOCM wrapper functions.

---
v9:
- Rebased on top of "tty-next", patch 2/3 dropped as its already applied

v8:
- Rebased on top of "tty-next"

v7:
- Change serial8250_do_get_mctrl() so that systems with a "mixed setup" (i.e. CTS controlled by UART but other status pins controlled by GPIO) are also supported again (Yegor)

v6:
- Use newly introduced TIOCM <-> MCR/MSR wrapper functions
- serial8250_in_MCR(): Don't save the already read MCR bits in TIOCM format but "or" them later to the GPIO MCR value
- Correctly use "!up->gpios" (Andy)
- Removed Mika's reviewed by tag (because of changes)

v5:
- Dropped a few "if (up->gpios)" checks, as the mctrl_gpio_foo() API handles gpios == NULL (return)
- 8250_omap: Changed "IS_ERR_OR_NULL(mctrl_gpio_to_gpiod(up->gpios, ...))" to "up->gpios == NULL", as mctrl_gpio_to_gpiod() does not handle gpios == NULL correctly.

v4:
- Added Mika's reviewed by tag
- Added Johan to Cc

v3:
- Only call mctrl_gpio_init(), if the device has no ACPI companion device to not break existing ACPI based systems, as suggested by Andy

v2:
- No change

Please note that this patch was already applied before [1]. And later reverted [2] because it introduced problems on some x86 based boards (ACPI GPIO related). Here a detailed description of the issue at that time:

https://lkml.org/lkml/2016/8/9/357
http://www.spinics.net/lists/linux-serial/msg23071.html

This is a re-send of the original patch that was applied at that time. With patch 1/2 from this series this issue should be fixed now (please note that I can't test it on such an x86 platform causing these problems).

Andy (or Mika), perhaps it would be possible for you to test this patch again, now with patch 1/2 of this series applied as well? That would be really helpful.

Thanks,
Stefan

[1] 4ef03d328769 ("tty/serial/8250: use mctrl_gpio helpers")
[2] 5db4f7f80d16 ("Revert "tty/serial/8250: use mctrl_gpio helpers"")

 .../devicetree/bindings/serial/8250.txt       | 19 ++++++++++++
 drivers/tty/serial/8250/8250.h                | 18 +++++++++++-
 drivers/tty/serial/8250/8250_core.c           | 17 +++++++++++
 drivers/tty/serial/8250/8250_omap.c           | 29 ++++++++++---------
 drivers/tty/serial/8250/8250_port.c           | 11 ++++++-
 drivers/tty/serial/8250/Kconfig               |  1 +
 include/linux/serial_8250.h                   |  1 +
 7 files changed, 81 insertions(+), 15 deletions(-)
Documentation/devicetree/bindings/serial/8250.txt +19 -0
index 3cba12f855b7..20d351f268ef 100644
@@ -53,6 +53,9 @@ Optional properties:
   programmable TX FIFO thresholds.
 - resets : phandle + reset specifier pairs
 - overrun-throttle-ms : how long to pause uart rx when input overrun is encountered.
+- {rts,cts,dtr,dsr,rng,dcd}-gpios: specify a GPIO for RTS/CTS/DTR/DSR/RI/DCD
+  line respectively. It will use specified GPIO instead of the peripheral
+  function pin for the UART feature. If unsure, don't specify this property.
 
 Note:
 * fsl,ns16550:
@@ -74,3 +77,19 @@ Example:
 		interrupts = <10>;
 		reg-shift = <2>;
 	};
+
+Example for OMAP UART using GPIO-based modem control signals:
+
+	uart4: serial@49042000 {
+		compatible = "ti,omap3-uart";
+		reg = <0x49042000 0x400>;
+		interrupts = <80>;
+		ti,hwmods = "uart4";
+		clock-frequency = <48000000>;
+		cts-gpios = <&gpio3 5 GPIO_ACTIVE_LOW>;
+		rts-gpios = <&gpio3 6 GPIO_ACTIVE_LOW>;
+		dtr-gpios = <&gpio1 12 GPIO_ACTIVE_LOW>;
+		dsr-gpios = <&gpio1 13 GPIO_ACTIVE_LOW>;
+		dcd-gpios = <&gpio1 14 GPIO_ACTIVE_LOW>;
+		rng-gpios = <&gpio1 15 GPIO_ACTIVE_LOW>;
+	};
drivers/tty/serial/8250/8250.h +17 -1
index 57db8c1689af..33ad9d6de532 100644
@@ -11,6 +11,8 @@
 #include <linux/serial_reg.h>
 #include <linux/dmaengine.h>
 
+#include "../serial_mctrl_gpio.h"
+
 struct uart_8250_dma {
 	int (*tx_dma)(struct uart_8250_port *p);
 	int (*rx_dma)(struct uart_8250_port *p);
@@ -214,11 +216,25 @@ static inline int serial8250_MSR_to_TIOCM(int msr)
 static inline void serial8250_out_MCR(struct uart_8250_port *up, int value)
 {
 	serial_out(up, UART_MCR, value);
+
+	if (up->gpios)
+		mctrl_gpio_set(up->gpios, serial8250_MCR_to_TIOCM(value));
 }
 
 static inline int serial8250_in_MCR(struct uart_8250_port *up)
 {
-	return serial_in(up, UART_MCR);
+	int mctrl;
+
+	mctrl = serial_in(up, UART_MCR);
+
+	if (up->gpios) {
+		unsigned int mctrl_gpio = 0;
+
+		mctrl_gpio = mctrl_gpio_get_outputs(up->gpios, &mctrl_gpio);
+		mctrl |= serial8250_TIOCM_to_MCR(mctrl_gpio);
+	}
+
+	return mctrl;
 }
 
 #if defined(__alpha__) && !defined(CONFIG_PCI)
drivers/tty/serial/8250/8250_core.c +17 -0
index e441221e04b9..a4470771005f 100644
@@ -14,6 +14,7 @@
  *	      serial8250_register_8250_port() ports
  */
 
+#include <linux/acpi.h>
 #include <linux/module.h>
 #include <linux/moduleparam.h>
 #include <linux/ioport.h>
@@ -982,6 +983,8 @@ int serial8250_register_8250_port(struct uart_8250_port *up)
 
 	uart = serial8250_find_match_or_unused(&up->port);
 	if (uart && uart->port.type != PORT_8250_CIR) {
+		struct mctrl_gpios *gpios;
+
 		if (uart->port.dev)
 			uart_remove_one_port(&serial8250_reg, &uart->port);
 
@@ -1016,6 +1019,20 @@ int serial8250_register_8250_port(struct uart_8250_port *up)
 		if (up->port.flags & UPF_FIXED_TYPE)
 			uart->port.type = up->port.type;
 
+		/*
+		 * Only call mctrl_gpio_init(), if the device has no ACPI
+		 * companion device
+		 */
+		if (!has_acpi_companion(uart->port.dev)) {
+			gpios = mctrl_gpio_init(&uart->port, 0);
+			if (IS_ERR(gpios)) {
+				if (PTR_ERR(gpios) != -ENOSYS)
+					return PTR_ERR(gpios);
+			} else {
+				uart->gpios = gpios;
+			}
+		}
+
 		serial8250_set_defaults(uart);
 
 		/* Possibly override default I/O functions.  */
drivers/tty/serial/8250/8250_omap.c +16 -13
index ed25cfc3be13..3ef65cbd2478 100644
@@ -141,18 +141,20 @@ static void omap8250_set_mctrl(struct uart_port *port, unsigned int mctrl)
 
 	serial8250_do_set_mctrl(port, mctrl);
 
-	/*
-	 * Turn off autoRTS if RTS is lowered and restore autoRTS setting
-	 * if RTS is raised
-	 */
-	lcr = serial_in(up, UART_LCR);
-	serial_out(up, UART_LCR, UART_LCR_CONF_MODE_B);
-	if ((mctrl & TIOCM_RTS) && (port->status & UPSTAT_AUTORTS))
-		priv->efr |= UART_EFR_RTS;
-	else
-		priv->efr &= ~UART_EFR_RTS;
-	serial_out(up, UART_EFR, priv->efr);
-	serial_out(up, UART_LCR, lcr);
+	if (!up->gpios) {
+		/*
+		 * Turn off autoRTS if RTS is lowered and restore autoRTS
+		 * setting if RTS is raised
+		 */
+		lcr = serial_in(up, UART_LCR);
+		serial_out(up, UART_LCR, UART_LCR_CONF_MODE_B);
+		if ((mctrl & TIOCM_RTS) && (port->status & UPSTAT_AUTORTS))
+			priv->efr |= UART_EFR_RTS;
+		else
+			priv->efr &= ~UART_EFR_RTS;
+		serial_out(up, UART_EFR, priv->efr);
+		serial_out(up, UART_LCR, lcr);
+	}
 }
 
 /*
@@ -453,7 +455,8 @@ static void omap_8250_set_termios(struct uart_port *port,
 	priv->efr = 0;
 	up->port.status &= ~(UPSTAT_AUTOCTS | UPSTAT_AUTORTS | UPSTAT_AUTOXOFF);
 
-	if (termios->c_cflag & CRTSCTS && up->port.flags & UPF_HARD_FLOW) {
+	if (termios->c_cflag & CRTSCTS && up->port.flags & UPF_HARD_FLOW &&
+	    !up->gpios) {
 		/* Enable AUTOCTS (autoRTS is enabled when RTS is raised) */
 		up->port.status |= UPSTAT_AUTOCTS | UPSTAT_AUTORTS;
 		priv->efr |= UART_EFR_CTS;
drivers/tty/serial/8250/8250_port.c +10 -1
index 6bd3f65ac967..71f54454a0f2 100644
@@ -1656,6 +1656,8 @@ static void serial8250_disable_ms(struct uart_port *port)
 	if (up->bugs & UART_BUG_NOMSR)
 		return;
 
+	mctrl_gpio_disable_ms(up->gpios);
+
 	up->ier &= ~UART_IER_MSI;
 	serial_port_out(port, UART_IER, up->ier);
 }
@@ -1668,6 +1670,8 @@ static void serial8250_enable_ms(struct uart_port *port)
 	if (up->bugs & UART_BUG_NOMSR)
 		return;
 
+	mctrl_gpio_enable_ms(up->gpios);
+
 	up->ier |= UART_IER_MSI;
 
 	serial8250_rpm_get(up);
@@ -1939,12 +1943,17 @@ unsigned int serial8250_do_get_mctrl(struct uart_port *port)
 {
 	struct uart_8250_port *up = up_to_u8250p(port);
 	unsigned int status;
+	unsigned int val;
 
 	serial8250_rpm_get(up);
 	status = serial8250_modem_status(up);
 	serial8250_rpm_put(up);
 
-	return serial8250_MSR_to_TIOCM(status);
+	val = serial8250_MSR_to_TIOCM(status);
+	if (up->gpios)
+		return mctrl_gpio_get(up->gpios, &val);
+
+	return val;
 }
 EXPORT_SYMBOL_GPL(serial8250_do_get_mctrl);
 
drivers/tty/serial/8250/Kconfig +1 -0
index 296115f6a4d8..509f6a3bb9ff 100644
@@ -8,6 +8,7 @@ config SERIAL_8250
 	tristate "8250/16550 and compatible serial support"
 	depends on !S390
 	select SERIAL_CORE
+	select SERIAL_MCTRL_GPIO if GPIOLIB
 	---help---
 	  This selects whether you want to include the driver for the standard
 	  serial ports.  The standard answer is Y.  People who might say N
include/linux/serial_8250.h +1 -0
index 5e0b59422a68..bb2bc99388ca 100644
@@ -110,6 +110,7 @@ struct uart_8250_port {
 						 *   if no_console_suspend
 						 */
 	unsigned char		probe;
+	struct mctrl_gpios	*gpios;
 #define UART_PROBE_RSA	(1 << 0)
 
 	/*
Signature · 1 line
2.22.0
#2 lore raw
In reply to Stefan Roese

CC gpio

This is now commit d99482673f950817 ("serial: mctrl_gpio: Check if GPIO property exisits before requesting it") in tty-next.

Quoted 107 lines · Stefan Roese

This patch adds a check for the GPIOs property existence, before the GPIO is requested. This fixes an issue seen when the 8250 mctrl_gpio support is added (2nd patch in this patch series) on x86 platforms using ACPI.

Here Mika's comments from 2016-08-09:

"
I noticed that with v4.8-rc1 serial console of some of our Broxton systems does not work properly anymore. I'm able to see output but input does not work.

I bisected it down to commit 4ef03d328769eddbfeca1f1c958fdb181a69c341 ("tty/serial/8250: use mctrl_gpio helpers").

The reason why it fails is that in ACPI we do not have names for GPIOs (except when _DSD is used) so we use the "idx" to index into _CRS GPIO resources. Now mctrl_gpio_init_noauto() goes through a list of GPIOs calling devm_gpiod_get_index_optional() passing "idx" of 0 for each. The UART device in Broxton has following (simplified) ACPI description:

    Device (URT4)
    {
        ...
        Name (_CRS, ResourceTemplate () {
            GpioIo (Exclusive, PullDefault, 0x0000, 0x0000, IoRestrictionOutputOnly,
                    "\\_SB.GPO0", 0x00, ResourceConsumer)
            {
                0x003A
            }
            GpioIo (Exclusive, PullDefault, 0x0000, 0x0000, IoRestrictionOutputOnly,
                    "\\_SB.GPO0", 0x00, ResourceConsumer)
            {
                0x003D
            }
        })

In this case it finds the first GPIO (0x003A which happens to be RX pin for that UART), turns it into GPIO which then breaks input for the UART device. This also breaks systems with bluetooth connected to UART (those typically have some GPIOs in their _CRS).

Any ideas how to fix this?

We cannot just drop the _CRS index lookup fallback because that would break many existing machines out there so maybe we can limit this to only DT enabled machines. Or alternatively probe if the property first exists before trying to acquire the GPIOs (using device_property_present()).
"

This patch implements the fix suggested by Mika in his statement above.

Cc Mika Westerberg <[email protected]>
Cc Andy Shevchenko <[email protected]>
Cc Yegor Yefremov <[email protected]>
Cc Greg Kroah-Hartman <[email protected]>
Cc Giulio Benetti <[email protected]>

---
v9:
- Rebased on top of "tty-next", patch 2/3 dropped as its already applied

v8:
- Rebased on top of "tty-next"

v7:
- Include <linux/property.h> to fix compile breakage on OMAP

v6:
- No change

v5:
- Simplified the code a bit (Andy)
- Added gpio_str == NULL handling (Andy)

v4:
- Add missing free() calls (Johan)
- Added Mika's reviewed by tag
- Added Johan to Cc

v3:
- No change

v2:
- Include the problem description and analysis from Mika into the commit text, as suggested by Greg.

 drivers/tty/serial/serial_mctrl_gpio.c | 14 ++++++++++++++
 1 file changed, 14 insertions(+)
drivers/tty/serial/serial_mctrl_gpio.c +1 -0
index 39ed56214cd3..2b400189be91 100644
@@ -12,6 +12,7 @@
 #include <linux/termios.h>
 #include <linux/serial_core.h>
 #include <linux/module.h>
+#include <linux/property.h>
 
 #include "serial_mctrl_gpio.h"
 

drivers/tty/serial/serial_mctrl_gpio.c struct mctrl_gpios *mctrl_gpio_init_noauto(struct device *dev, unsigned int idx) · line 125

Quoted 9 lines
drivers/tty/serial/serial_mctrl_gpio.c +6 -0
@@ -116,6 +117,19 @@ struct mctrl_gpios *mctrl_gpio_init_noauto(struct device *dev, unsigned int idx)
 
        for (i = 0; i < UART_GPIO_MAX; i++) {
                enum gpiod_flags flags;
+               char *gpio_str;
+               bool present;
+
+               /* Check if GPIO property exists and continue if not */
+               gpio_str = kasprintf(GFP_KERNEL, "%s-gpios",
+                                    mctrl_gpios_desc[i].name);

This will silently break DTBs using "(cts|dsr|dcd|rng|rts|dtr)-gpio" instead of "(cts|dsr|dcd|rng|rts|dtr)-gpios".

drivers/tty/serial/serial_mctrl_gpio.c struct mctrl_gpios *mctrl_gpio_init_noauto(struct device *dev, unsigned int idx) · line 135

Quoted 10 lines
drivers/tty/serial/serial_mctrl_gpio.c +7 -0
+               if (!gpio_str)
+                       continue;
+
+               present = device_property_present(dev, gpio_str);
+               kfree(gpio_str);
+               if (!present)
+                       continue;
 
                if (mctrl_gpios_desc[i].dir_out)
                        flags = GPIOD_OUT_LOW;

Gr{oetje,eeting}s,

Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- [email protected]

In personal conversations with technical people, I call myself a hacker. But when I'm talking to journalists I just say "programmer" or something like that.
-- Linus Torvalds

#3 lore raw
Quoted 114 lines · Geert Uytterhoeven

CC gpio

This is now commit d99482673f950817 ("serial: mctrl_gpio: Check if GPIO property exisits before requesting it") in tty-next.

Quoted 108 lines · Stefan Roese

This patch adds a check for the GPIOs property existence, before the GPIO is requested. This fixes an issue seen when the 8250 mctrl_gpio support is added (2nd patch in this patch series) on x86 platforms using ACPI.

Here Mika's comments from 2016-08-09:

"
I noticed that with v4.8-rc1 serial console of some of our Broxton systems does not work properly anymore. I'm able to see output but input does not work.

I bisected it down to commit 4ef03d328769eddbfeca1f1c958fdb181a69c341 ("tty/serial/8250: use mctrl_gpio helpers").

The reason why it fails is that in ACPI we do not have names for GPIOs (except when _DSD is used) so we use the "idx" to index into _CRS GPIO resources. Now mctrl_gpio_init_noauto() goes through a list of GPIOs calling devm_gpiod_get_index_optional() passing "idx" of 0 for each. The UART device in Broxton has following (simplified) ACPI description:

     Device (URT4)
     {
         ...
         Name (_CRS, ResourceTemplate () {
             GpioIo (Exclusive, PullDefault, 0x0000, 0x0000, IoRestrictionOutputOnly,
                     "\\_SB.GPO0", 0x00, ResourceConsumer)
             {
                 0x003A
             }
             GpioIo (Exclusive, PullDefault, 0x0000, 0x0000, IoRestrictionOutputOnly,
                     "\\_SB.GPO0", 0x00, ResourceConsumer)
             {
                 0x003D
             }
         })

In this case it finds the first GPIO (0x003A which happens to be RX pin for that UART), turns it into GPIO which then breaks input for the UART device. This also breaks systems with bluetooth connected to UART (those typically have some GPIOs in their _CRS).

Any ideas how to fix this?

We cannot just drop the _CRS index lookup fallback because that would break many existing machines out there so maybe we can limit this to only DT enabled machines. Or alternatively probe if the property first exists before trying to acquire the GPIOs (using device_property_present()).
"

This patch implements the fix suggested by Mika in his statement above.

Cc Mika Westerberg <[email protected]>
Cc Andy Shevchenko <[email protected]>
Cc Yegor Yefremov <[email protected]>
Cc Greg Kroah-Hartman <[email protected]>
Cc Giulio Benetti <[email protected]>

---
v9:
- Rebased on top of "tty-next", patch 2/3 dropped as its already applied

v8:
- Rebased on top of "tty-next"

v7:
- Include <linux/property.h> to fix compile breakage on OMAP

v6:
- No change

v5:
- Simplified the code a bit (Andy)
- Added gpio_str == NULL handling (Andy)

v4:
- Add missing free() calls (Johan)
- Added Mika's reviewed by tag
- Added Johan to Cc

v3:
- No change

v2:
- Include the problem description and analysis from Mika into the commit text, as suggested by Greg.

  drivers/tty/serial/serial_mctrl_gpio.c | 14 ++++++++++++++
  1 file changed, 14 insertions(+)
drivers/tty/serial/serial_mctrl_gpio.c +1 -0
index 39ed56214cd3..2b400189be91 100644
@@ -12,6 +12,7 @@
  #include <linux/termios.h>
  #include <linux/serial_core.h>
  #include <linux/module.h>
+#include <linux/property.h>
 
  #include "serial_mctrl_gpio.h"
 
@@ -116,6 +117,19 @@ struct mctrl_gpios *mctrl_gpio_init_noauto(struct device *dev, unsigned int idx)
 
Quoted 11 lines

drivers/tty/serial/serial_mctrl_gpio.c struct mctrl_gpios *mctrl_gpio_init_noauto(struct device *dev, unsigned int idx) · line 125

Quoted 8 lines
drivers/tty/serial/serial_mctrl_gpio.c +6 -0
         for (i = 0; i < UART_GPIO_MAX; i++) {
                 enum gpiod_flags flags;
+               char *gpio_str;
+               bool present;
+
+               /* Check if GPIO property exists and continue if not */
+               gpio_str = kasprintf(GFP_KERNEL, "%s-gpios",
+                                    mctrl_gpios_desc[i].name);

This will silently break DTBs using "(cts|dsr|dcd|rng|rts|dtr)-gpio" instead of "(cts|dsr|dcd|rng|rts|dtr)-gpios".

Should both options be supported ("cts-gpio" vs "cts-gpios")? Documentation/devicetree/bindings/serial/serial.txt only mentions the "-gpios" variant.

Thanks,
Stefan

#4 lore raw

Hi Stefan,

Quoted 74 lines · Stefan Roese
Quoted 61 lines · Geert Uytterhoeven
Quoted 60 lines · Stefan Roese

This patch adds a check for the GPIOs property existence, before the GPIO is requested. This fixes an issue seen when the 8250 mctrl_gpio support is added (2nd patch in this patch series) on x86 platforms using ACPI.

Here Mika's comments from 2016-08-09:

"
I noticed that with v4.8-rc1 serial console of some of our Broxton systems does not work properly anymore. I'm able to see output but input does not work.

I bisected it down to commit 4ef03d328769eddbfeca1f1c958fdb181a69c341 ("tty/serial/8250: use mctrl_gpio helpers").

The reason why it fails is that in ACPI we do not have names for GPIOs (except when _DSD is used) so we use the "idx" to index into _CRS GPIO resources. Now mctrl_gpio_init_noauto() goes through a list of GPIOs calling devm_gpiod_get_index_optional() passing "idx" of 0 for each. The UART device in Broxton has following (simplified) ACPI description:

     Device (URT4)
     {
         ...
         Name (_CRS, ResourceTemplate () {
             GpioIo (Exclusive, PullDefault, 0x0000, 0x0000, IoRestrictionOutputOnly,
                     "\\_SB.GPO0", 0x00, ResourceConsumer)
             {
                 0x003A
             }
             GpioIo (Exclusive, PullDefault, 0x0000, 0x0000, IoRestrictionOutputOnly,
                     "\\_SB.GPO0", 0x00, ResourceConsumer)
             {
                 0x003D
             }
         })

In this case it finds the first GPIO (0x003A which happens to be RX pin for that UART), turns it into GPIO which then breaks input for the UART device. This also breaks systems with bluetooth connected to UART (those typically have some GPIOs in their _CRS).

Any ideas how to fix this?

We cannot just drop the _CRS index lookup fallback because that would break many existing machines out there so maybe we can limit this to only DT enabled machines. Or alternatively probe if the property first exists before trying to acquire the GPIOs (using device_property_present()).
"

This patch implements the fix suggested by Mika in his statement above.

drivers/tty/serial/serial_mctrl_gpio.c +0 -0
@@ -116,6 +117,19 @@ struct mctrl_gpios *mctrl_gpio_init_noauto(struct device *dev, unsigned int idx)
 
Quoted 11 lines

drivers/tty/serial/serial_mctrl_gpio.c struct mctrl_gpios *mctrl_gpio_init_noauto(struct device *dev, unsigned int idx) · line 125

Quoted 8 lines
drivers/tty/serial/serial_mctrl_gpio.c +6 -0
         for (i = 0; i < UART_GPIO_MAX; i++) {
                 enum gpiod_flags flags;
+               char *gpio_str;
+               bool present;
+
+               /* Check if GPIO property exists and continue if not */
+               gpio_str = kasprintf(GFP_KERNEL, "%s-gpios",
+                                    mctrl_gpios_desc[i].name);

This will silently break DTBs using "(cts|dsr|dcd|rng|rts|dtr)-gpio" instead of "(cts|dsr|dcd|rng|rts|dtr)-gpios".

Quoted 3 lines

Should both options be supported ("cts-gpio" vs "cts-gpios")? Documentation/devicetree/bindings/serial/serial.txt only mentions the "-gpios" variant.

Well, the "-gpio" variant is deprecated, but still supported by devm_gpiod_get_index_optional(), and there are active users in upstream DTS files.

My main objection is (trying to) replicate the matching logic inside gpiolib.c, causing subtle semantic differences. And keeping it consistent, of course.

It would be nice if this could be fixed inside acpi_find_gpio(), so users don't need to be updated. There may be other subsystems where the difference between DT and ACPI may cause issues, unbeknownst.

Gr{oetje,eeting}s,

Geert

Signature · 5 lines
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- [email protected]

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds
#5 lore raw

Hi Geert,

Quoted 83 lines · Geert Uytterhoeven
Quoted 74 lines · Stefan Roese
Quoted 61 lines · Geert Uytterhoeven
Quoted 60 lines · Stefan Roese

This patch adds a check for the GPIOs property existence, before the GPIO is requested. This fixes an issue seen when the 8250 mctrl_gpio support is added (2nd patch in this patch series) on x86 platforms using ACPI.

Here Mika's comments from 2016-08-09:

"
I noticed that with v4.8-rc1 serial console of some of our Broxton systems does not work properly anymore. I'm able to see output but input does not work.

I bisected it down to commit 4ef03d328769eddbfeca1f1c958fdb181a69c341 ("tty/serial/8250: use mctrl_gpio helpers").

The reason why it fails is that in ACPI we do not have names for GPIOs (except when _DSD is used) so we use the "idx" to index into _CRS GPIO resources. Now mctrl_gpio_init_noauto() goes through a list of GPIOs calling devm_gpiod_get_index_optional() passing "idx" of 0 for each. The UART device in Broxton has following (simplified) ACPI description:

      Device (URT4)
      {
          ...
          Name (_CRS, ResourceTemplate () {
              GpioIo (Exclusive, PullDefault, 0x0000, 0x0000, IoRestrictionOutputOnly,
                      "\\_SB.GPO0", 0x00, ResourceConsumer)
              {
                  0x003A
              }
              GpioIo (Exclusive, PullDefault, 0x0000, 0x0000, IoRestrictionOutputOnly,
                      "\\_SB.GPO0", 0x00, ResourceConsumer)
              {
                  0x003D
              }
          })

In this case it finds the first GPIO (0x003A which happens to be RX pin for that UART), turns it into GPIO which then breaks input for the UART device. This also breaks systems with bluetooth connected to UART (those typically have some GPIOs in their _CRS).

Any ideas how to fix this?

We cannot just drop the _CRS index lookup fallback because that would break many existing machines out there so maybe we can limit this to only DT enabled machines. Or alternatively probe if the property first exists before trying to acquire the GPIOs (using device_property_present()).
"

This patch implements the fix suggested by Mika in his statement above.

drivers/tty/serial/serial_mctrl_gpio.c +0 -0
@@ -116,6 +117,19 @@ struct mctrl_gpios *mctrl_gpio_init_noauto(struct device *dev, unsigned int idx)
 
Quoted 11 lines

drivers/tty/serial/serial_mctrl_gpio.c struct mctrl_gpios *mctrl_gpio_init_noauto(struct device *dev, unsigned int idx) · line 125

Quoted 8 lines
drivers/tty/serial/serial_mctrl_gpio.c +6 -0
          for (i = 0; i < UART_GPIO_MAX; i++) {
                  enum gpiod_flags flags;
+               char *gpio_str;
+               bool present;
+
+               /* Check if GPIO property exists and continue if not */
+               gpio_str = kasprintf(GFP_KERNEL, "%s-gpios",
+                                    mctrl_gpios_desc[i].name);

This will silently break DTBs using "(cts|dsr|dcd|rng|rts|dtr)-gpio" instead of "(cts|dsr|dcd|rng|rts|dtr)-gpios".

Quoted 3 lines

Should both options be supported ("cts-gpio" vs "cts-gpios")? Documentation/devicetree/bindings/serial/serial.txt only mentions the "-gpios" variant.

Well, the "-gpio" variant is deprecated, but still supported by devm_gpiod_get_index_optional(), and there are active users in upstream DTS files.

Quoted 7 lines

My main objection is (trying to) replicate the matching logic inside gpiolib.c, causing subtle semantic differences. And keeping it consistent, of course.

It would be nice if this could be fixed inside acpi_find_gpio(), so users don't need to be updated. There may be other subsystems where the difference between DT and ACPI may cause issues, unbeknownst.

Sure, I can fix this. I would prefer to do this in a follow-up patch though, if nobody objects.

Thanks,
Stefan

#6 lore raw
Quoted 16 lines · Stefan Roese
Quoted 15 lines · Geert Uytterhoeven
Quoted 5 lines · Stefan Roese

On 24.06.19 10:42, Geert Uytterhoeven wrote:

Should both options be supported ("cts-gpio" vs "cts-gpios")? Documentation/devicetree/bindings/serial/serial.txt only mentions the "-gpios" variant.

Well, the "-gpio" variant is deprecated, but still supported by devm_gpiod_get_index_optional(), and there are active users in upstream DTS files.

My main objection is (trying to) replicate the matching logic inside gpiolib.c, causing subtle semantic differences. And keeping it consistent, of course.

Quoted 6 lines
Quoted 3 lines

It would be nice if this could be fixed inside acpi_find_gpio(), so users don't need to be updated. There may be other subsystems where the difference between DT and ACPI may cause issues, unbeknownst.

Sure, I can fix this. I would prefer to do this in a follow-up patch though, if nobody objects.

In case if you are going to do this, use approach from GPIO library. Perhaps, it may require to create something like for_each_gpio_suffix() helper.

Signature · 2 lines
With Best Regards,
Andy Shevchenko
#7 lore raw
In reply to Stefan Roese

Hello,

Quoted 108 lines · Stefan Roese

This patch adds a check for the GPIOs property existence, before the GPIO is requested. This fixes an issue seen when the 8250 mctrl_gpio support is added (2nd patch in this patch series) on x86 platforms using ACPI.

Here Mika's comments from 2016-08-09:

"
I noticed that with v4.8-rc1 serial console of some of our Broxton systems does not work properly anymore. I'm able to see output but input does not work.

I bisected it down to commit 4ef03d328769eddbfeca1f1c958fdb181a69c341 ("tty/serial/8250: use mctrl_gpio helpers").

The reason why it fails is that in ACPI we do not have names for GPIOs (except when _DSD is used) so we use the "idx" to index into _CRS GPIO resources. Now mctrl_gpio_init_noauto() goes through a list of GPIOs calling devm_gpiod_get_index_optional() passing "idx" of 0 for each. The UART device in Broxton has following (simplified) ACPI description:

    Device (URT4)
    {
        ...
        Name (_CRS, ResourceTemplate () {
            GpioIo (Exclusive, PullDefault, 0x0000, 0x0000, IoRestrictionOutputOnly,
                    "\\_SB.GPO0", 0x00, ResourceConsumer)
            {
                0x003A
            }
            GpioIo (Exclusive, PullDefault, 0x0000, 0x0000, IoRestrictionOutputOnly,
                    "\\_SB.GPO0", 0x00, ResourceConsumer)
            {
                0x003D
            }
        })

In this case it finds the first GPIO (0x003A which happens to be RX pin for that UART), turns it into GPIO which then breaks input for the UART device. This also breaks systems with bluetooth connected to UART (those typically have some GPIOs in their _CRS).

Any ideas how to fix this?

We cannot just drop the _CRS index lookup fallback because that would break many existing machines out there so maybe we can limit this to only DT enabled machines. Or alternatively probe if the property first exists before trying to acquire the GPIOs (using device_property_present()).
"

This patch implements the fix suggested by Mika in his statement above.

Cc Mika Westerberg <[email protected]>
Cc Andy Shevchenko <[email protected]>
Cc Yegor Yefremov <[email protected]>
Cc Greg Kroah-Hartman <[email protected]>
Cc Giulio Benetti <[email protected]>

---
v9:
- Rebased on top of "tty-next", patch 2/3 dropped as its already applied

v8:
- Rebased on top of "tty-next"

v7:
- Include <linux/property.h> to fix compile breakage on OMAP

v6:
- No change

v5:
- Simplified the code a bit (Andy)
- Added gpio_str == NULL handling (Andy)

v4:
- Add missing free() calls (Johan)
- Added Mika's reviewed by tag
- Added Johan to Cc

v3:
- No change

v2:
- Include the problem description and analysis from Mika into the commit text, as suggested by Greg.

 drivers/tty/serial/serial_mctrl_gpio.c | 14 ++++++++++++++
 1 file changed, 14 insertions(+)
drivers/tty/serial/serial_mctrl_gpio.c +1 -0
index 39ed56214cd3..2b400189be91 100644
@@ -12,6 +12,7 @@
 #include <linux/termios.h>
 #include <linux/serial_core.h>
 #include <linux/module.h>
+#include <linux/property.h>
 
 #include "serial_mctrl_gpio.h"
 
@@ -116,6 +117,19 @@ struct mctrl_gpios *mctrl_gpio_init_noauto(struct device *dev, unsigned int idx)
 

drivers/tty/serial/serial_mctrl_gpio.c struct mctrl_gpios *mctrl_gpio_init_noauto(struct device *dev, unsigned int idx) · line 127

Quoted 10 lines
drivers/tty/serial/serial_mctrl_gpio.c +8 -0
 	for (i = 0; i < UART_GPIO_MAX; i++) {
 		enum gpiod_flags flags;
+		char *gpio_str;
+		bool present;
+
+		/* Check if GPIO property exists and continue if not */
+		gpio_str = kasprintf(GFP_KERNEL, "%s-gpios",
+				     mctrl_gpios_desc[i].name);
+		if (!gpio_str)
+			continue;

Is it a good idea to handle a memory allocation problem with a continue?

drivers/tty/serial/serial_mctrl_gpio.c struct mctrl_gpios *mctrl_gpio_init_noauto(struct device *dev, unsigned int idx) · line 133

Quoted 5 lines
drivers/tty/serial/serial_mctrl_gpio.c +4 -0
+		present = device_property_present(dev, gpio_str);
+		kfree(gpio_str);
+		if (!present)
+			continue;
 

Assuming this fixes the situation on x86 this means that there device_property_present(...) returns false and devm_gpiod_get_index_optional returns an error code.

I wonder what the problem is here. Is CONFIG_GPIOLIB on in the failing situation?

I assume this will end in the usual discussion if gpiod_get_*_optional should return an error code if GPIOLIB is off.

Best regards
Uwe

Signature · 2 lines
Pengutronix e.K.                           | Uwe Kleine-König            |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |