Skip to contentSkip to conversation

mfd: cros ec: spi: Add delay for raising CS

PATCH Applied Thierry Reding 5 participants 10 messages Nov 18, 2013 – Nov 19, 2013

Subsystems (3)

v1 of 5 Applied Latest is v5 →

lore
Share & tools
lore
#0 lore raw

PATCH 1/3 mfd: cros ec: spi: Add delay for raising CS

From: Rhyland Klein <[email protected]>

The EC has specific timing it requires. Add support for an optional delay after raising CS to fix timing issues. This is configurable based on a DT property "google,cros-ec-spi-msg-delay".

If this property isn't set, then no delay will be added. However, if set it will cause a delay equal to the value passed to it to be inserted at the end of a transaction.

---
 Documentation/devicetree/bindings/mfd/cros-ec.txt |  4 +++
 drivers/mfd/cros_ec_spi.c                         | 30 +++++++++++++++++++++++
 2 files changed, 34 insertions(+)
Documentation/devicetree/bindings/mfd/cros-ec.txt +4 -0
index 5f229c5f6da9..fb3034a87ae0 100644
@@ -17,6 +17,10 @@ Required properties (SPI):
 - compatible: "google,cros-ec-spi"
 - reg: SPI chip select
 
+Optional properties (SPI):
+- google,cros-ec-spi-msg-delay: This property is how long of a delay, in usecs,
+  to use on the last transfer of a message, to force time between transactions.
+
 Required properties (LPC):
 - compatible: "google,cros-ec-lpc"
 - reg: List of (IO address, size) pairs defining the interface uses
drivers/mfd/cros_ec_spi.c +30 -0
index 367ccb58ecb1..d05349a7381e 100644
@@ -18,6 +18,7 @@
 #include <linux/module.h>
 #include <linux/mfd/cros_ec.h>
 #include <linux/mfd/cros_ec_commands.h>
+#include <linux/of.h>
 #include <linux/platform_device.h>
 #include <linux/slab.h>
 #include <linux/spi/spi.h>
@@ -61,10 +62,13 @@
  * @spi: SPI device we are connected to
  * @last_transfer_ns: time that we last finished a transfer, or 0 if there
  *	if no record
+ * @end_of_msg_delay: used to set the delay_usecs on the spi_transfer that
+ *      is sent when we want to turn off CS at the end of a transaction.
  */
 struct cros_ec_spi {
 	struct spi_device *spi;
 	s64 last_transfer_ns;
+	unsigned int end_of_msg_delay;
 };
 
 static void debug_packet(struct device *dev, const char *name, u8 *ptr,
@@ -235,6 +239,17 @@ static int cros_ec_command_spi_xfer(struct cros_ec_device *ec_dev,
 
 	/* turn off CS */
 	spi_message_init(&msg);
+
+	if (ec_spi->end_of_msg_delay) {
+		/*
+		 * Add delay for last transaction, to ensure the rising edge
+		 * doesn't come too soon after the end of the data.
+		 */
+		memset(&trans, '\0', sizeof(trans));
+		trans.delay_usecs = ec_spi->end_of_msg_delay;
+		spi_message_add_tail(&trans, &msg);
+	}
+
 	final_ret = spi_sync(ec_spi->spi, &msg);
 	ktime_get_ts(&ts);
 	ec_spi->last_transfer_ns = timespec_to_ns(&ts);
@@ -281,6 +296,17 @@ static int cros_ec_command_spi_xfer(struct cros_ec_device *ec_dev,
 	return 0;
 }
 
+static void cros_ec_probe_spi_dt(struct cros_ec_spi *ec_spi, struct device *dev)
+{
+	struct device_node *np = dev->of_node;
+	u32 val;
+	int ret;
+
+	ret = of_property_read_u32(np, "google,cros-ec-spi-msg-delay", &val);
+	if (!ret)
+		ec_spi->end_of_msg_delay = val;
+}
+
 static int cros_ec_probe_spi(struct spi_device *spi)
 {
 	struct device *dev = &spi->dev;
@@ -302,6 +328,10 @@ static int cros_ec_probe_spi(struct spi_device *spi)
 	if (!ec_dev)
 		return -ENOMEM;
 
+	/* Check for any DT properties */
+	if (IS_ENABLED(CONFIG_OF) && dev->of_node)
+		cros_ec_probe_spi_dt(ec_spi, dev);
+
 	spi_set_drvdata(spi, ec_dev);
 	ec_dev->name = "SPI";
 	ec_dev->dev = dev;
Signature · 1 line
1.8.4.2
#1 lore raw

PATCH 2/3 mfd: cros ec: spi: Increase EC transaction delay

From: Derek Basehore <[email protected]>

50 us is not a long enough delay between EC transactions. At least 70 us are needed for the 16 MHz STM32L part. Increase the delay to 200 us for an extra safety margin.

---
 drivers/mfd/cros_ec_spi.c | 7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)
drivers/mfd/cros_ec_spi.c +4 -3
index d05349a7381e..af5c2f6601cf 100644
@@ -51,10 +51,11 @@
 /*
   * Time between raising the SPI chip select (for the end of a
   * transaction) and dropping it again (for the next transaction).
-  * If we go too fast, the EC will miss the transaction. It seems
-  * that 50us is enough with the 16MHz STM32 EC.
+  * If we go too fast, the EC will miss the transaction. We know that we
+  * need at least 70 us with the 16 MHz STM32 EC, so go with 200 us to be
+  * safe.
   */
-#define EC_SPI_RECOVERY_TIME_NS	(50 * 1000)
+#define EC_SPI_RECOVERY_TIME_NS	(200 * 1000)
 
 /**
  * struct cros_ec_spi - information about a SPI-connected EC
Signature · 6 lines
1.8.4.2

--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
#2 lore raw

Re: [PATCH 2/3] mfd: cros ec: spi: Increase EC transaction delay

Quoted 6 lines · Thierry Reding

From: Derek Basehore <[email protected]>

50 us is not a long enough delay between EC transactions. At least 70 us are needed for the 16 MHz STM32L part. Increase the delay to 200 us for an extra safety margin.

Quoted 6 lines
---
 drivers/mfd/cros_ec_spi.c | 7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)

As I'm sure you've tested this and you didn't notice any new undue latency, I'll apply the patch, thanks.

Signature · 8 lines
Lee Jones
Linaro STMicroelectronics Landing Team Lead
Linaro.org │ Open source software for ARM SoCs
Follow Linaro: Facebook | Twitter | Blog
--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
#3 lore raw

Re: [PATCH 2/3] mfd: cros ec: spi: Increase EC transaction delay

We haven't had any issues with this patch and we've had it in our tree for a while.

I also stress tested the patch by repeatedly reading over the spi a lot.

#4 lore raw

Re: [PATCH 2/3] mfd: cros ec: spi: Increase EC transaction delay

Quoted 4 lines · dbasehore .

We haven't had any issues with this patch and we've had it in our tree for a while.

I also stress tested the patch by repeatedly reading over the spi a lot.

Okay, thanks for letting me know.

Signature · 8 lines
Lee Jones
Linaro STMicroelectronics Landing Team Lead
Linaro.org │ Open source software for ARM SoCs
Follow Linaro: Facebook | Twitter | Blog
--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
#5 lore raw

PATCH 3/3 mfd: cros ec: spi: Fix debug output

In reply to Thierry Reding

The dev_cont() symbol doesn't exist, so replace it with pr_cont(). While at it, also append a newline to the debug output to make it look nicer.

---
 drivers/mfd/cros_ec_spi.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)
drivers/mfd/cros_ec_spi.c +3 -1
index af5c2f6601cf..e3690fcb86ae 100644
@@ -80,7 +80,9 @@ static void debug_packet(struct device *dev, const char *name, u8 *ptr,
 
 	dev_dbg(dev, "%s: ", name);
 	for (i = 0; i < len; i++)
-		dev_cont(dev, " %02x", ptr[i]);
+		pr_cont(" %02x", ptr[i]);
+
+	pr_cont("\n");
 #endif
 }
 
Signature · 6 lines
1.8.4.2

--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Applied
#6 lore raw

Re: [PATCH 3/3] mfd: cros ec: spi: Fix debug output

Quoted 7 lines · Thierry Reding

The dev_cont() symbol doesn't exist, so replace it with pr_cont(). While at it, also append a newline to the debug output to make it look nicer.

---
 drivers/mfd/cros_ec_spi.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

Applied, thanks.

Signature · 8 lines
Lee Jones
Linaro STMicroelectronics Landing Team Lead
Linaro.org │ Open source software for ARM SoCs
Follow Linaro: Facebook | Twitter | Blog
--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
#7 lore raw
In reply to Thierry Reding
Quoted 28 lines · Thierry Reding

From: Rhyland Klein <[email protected]>

The EC has specific timing it requires. Add support for an optional delay after raising CS to fix timing issues. This is configurable based on a DT property "google,cros-ec-spi-msg-delay".

If this property isn't set, then no delay will be added. However, if set it will cause a delay equal to the value passed to it to be inserted at the end of a transaction.

---
 Documentation/devicetree/bindings/mfd/cros-ec.txt |  4 +++
 drivers/mfd/cros_ec_spi.c                         | 30 +++++++++++++++++++++++
 2 files changed, 34 insertions(+)
Documentation/devicetree/bindings/mfd/cros-ec.txt +0 -0
index 5f229c5f6da9..fb3034a87ae0 100644

Documentation/devicetree/bindings/mfd/cros-ec.txt Required properties (SPI): · line 23

Quoted 7 lines
Documentation/devicetree/bindings/mfd/cros-ec.txt +4 -0
@@ -17,6 +17,10 @@ Required properties (SPI):
 - compatible: "google,cros-ec-spi"
 - reg: SPI chip select
 
+Optional properties (SPI):
+- google,cros-ec-spi-msg-delay: This property is how long of a delay, in usecs,
+  to use on the last transfer of a message, to force time between transactions.
+

Lose the "This property is", that's obvious from the structure of the document.

It would be nice to have an explanation in the binding document as to _why_ you might want to do this (e.g. the HW expects the rising edge to come some number of uS after the data, if it comes too early it explodes).

Otherwise this looks fine to me.

Thanks,
Mark.
--
To unsubscribe from this list: send the line "unsubscribe devicetree" in the body of a message to [email protected] More majordomo info at http://vger.kernel.org/majordomo-info.html

#8 lore raw
Quoted 32 lines · Mark Rutland
Quoted 31 lines · Thierry Reding

From: Rhyland Klein <[email protected]>

The EC has specific timing it requires. Add support for an optional delay after raising CS to fix timing issues. This is configurable based on a DT property "google,cros-ec-spi-msg-delay".

If this property isn't set, then no delay will be added. However, if set it will cause a delay equal to the value passed to it to be inserted at the end of a transaction.

---
 Documentation/devicetree/bindings/mfd/cros-ec.txt |  4 +++
 drivers/mfd/cros_ec_spi.c                         | 30 +++++++++++++++++++++++
 2 files changed, 34 insertions(+)
Documentation/devicetree/bindings/mfd/cros-ec.txt +0 -0
index 5f229c5f6da9..fb3034a87ae0 100644
@@ -17,6 +17,10 @@ Required properties (SPI):
 - compatible: "google,cros-ec-spi"
 - reg: SPI chip select
 
Quoted 7 lines

Documentation/devicetree/bindings/mfd/cros-ec.txt Required properties (SPI): · line 23

Quoted 4 lines
Documentation/devicetree/bindings/mfd/cros-ec.txt +4 -0
+Optional properties (SPI):
+- google,cros-ec-spi-msg-delay: This property is how long of a delay, in usecs,
+  to use on the last transfer of a message, to force time between transactions.
+

Lose the "This property is", that's obvious from the structure of the document.

Will do.

Quoted 4 lines

It would be nice to have an explanation in the binding document as to _why_ you might want to do this (e.g. the HW expects the rising edge to come some number of uS after the data, if it comes too early it explodes).

From what I can tell, this might differ on the exact variant of the EC, but generally it seems that when the interval between two transactions isn't long enough the EC won't be able to respond properly in time and cause subsequent transactions to hang. Perhaps Rhyland, Bernie or Andrew are more familiar with the details and therefore can provide a better or more accurate explanation.

Quoted 1 line

Otherwise this looks fine to me.

Thanks,
Thierry

#9 lore raw
Quoted 50 lines · Thierry Reding

* PGP Signed by an unknown key

Quoted 32 lines · Mark Rutland
Quoted 31 lines · Thierry Reding

From: Rhyland Klein <[email protected]>

The EC has specific timing it requires. Add support for an optional delay after raising CS to fix timing issues. This is configurable based on a DT property "google,cros-ec-spi-msg-delay".

If this property isn't set, then no delay will be added. However, if set it will cause a delay equal to the value passed to it to be inserted at the end of a transaction.

---
 Documentation/devicetree/bindings/mfd/cros-ec.txt |  4 +++
 drivers/mfd/cros_ec_spi.c                         | 30 +++++++++++++++++++++++
 2 files changed, 34 insertions(+)
Documentation/devicetree/bindings/mfd/cros-ec.txt +0 -0
index 5f229c5f6da9..fb3034a87ae0 100644
@@ -17,6 +17,10 @@ Required properties (SPI):
 - compatible: "google,cros-ec-spi"
 - reg: SPI chip select
 
Quoted 7 lines

Documentation/devicetree/bindings/mfd/cros-ec.txt Required properties (SPI): · line 23

Quoted 4 lines
Documentation/devicetree/bindings/mfd/cros-ec.txt +4 -0
+Optional properties (SPI):
+- google,cros-ec-spi-msg-delay: This property is how long of a delay, in usecs,
+  to use on the last transfer of a message, to force time between transactions.
+

Lose the "This property is", that's obvious from the structure of the document.

Will do.

Quoted 4 lines

It would be nice to have an explanation in the binding document as to _why_ you might want to do this (e.g. the HW expects the rising edge to come some number of uS after the data, if it comes too early it explodes).

Quoted 6 lines

From what I can tell, this might differ on the exact variant of the EC, but generally it seems that when the interval between two transactions isn't long enough the EC won't be able to respond properly in time and cause subsequent transactions to hang. Perhaps Rhyland, Bernie or Andrew are more familiar with the details and therefore can provide a better or more accurate explanation.

I believe this explanation is correct. As I recall, I think adding this delay helped stabilize communication with the EC, where before sometimes transactions would time out.

-rhyland

Quoted 9 lines
Quoted 2 lines

Otherwise this looks fine to me.

Thanks,
Thierry

* Unknown Key
* 0x7F3EB3A1

Signature · 1 line
nvpublic