Skip to contentSkip to conversation

Add data hold delay support

PATCH Applied Przemyslaw Gaj 3 participants 15 messages Nov 14, 2019 – Nov 15, 2019

Subsystems (3)

v1 of 3 Applied Latest is v3 →

lore
Share & tools
lore

Przemyslaw Gaj

author maintainer
#0 lore raw

cover Add data hold delay support

Add THD_DEL (Data Hold Delay) support. After testing different scenarios, on various systems, sometimes there is a need to delay SDA_OUT propagation.

Adding support which allows to configure that delay using the device tree parameter.

Przemyslaw Gaj (3):
i3c: master: add data hold delay support
dt-bindings: i3c: Document data hold delay feature
MAINTAINERS: add myself as maintainer of Cadence I3C master controller
driver

 Documentation/devicetree/bindings/i3c/cdns,i3c-master.txt | 6 ++++++
 MAINTAINERS                                               | 6 ++++++
 drivers/i3c/master/i3c-master-cdns.c                      | 5 ++++-
 3 files changed, 16 insertions(+), 1 deletion(-)
Signature · 7 lines
2.14.0


_______________________________________________
linux-i3c mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-i3c
author maintainer
#1 lore raw

PATCH 1/3 i3c: master: add data hold delay support

This patch adds support for THD_DEL (Data Hold Delay) to Cadence I3C master constoller driver.

---
 drivers/i3c/master/i3c-master-cdns.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)
drivers/i3c/master/i3c-master-cdns.c +4 -1
index 10db0bf0655a..90ea98eef905 100644
@@ -60,6 +60,7 @@
 #define CTRL_HALT_EN			BIT(30)
 #define CTRL_MCS			BIT(29)
 #define CTRL_MCS_EN			BIT(28)
+#define CTRL_THD_DEL(x) 		(((x) << 24) & GENMASK(25, 24))
 #define CTRL_HJ_DISEC			BIT(8)
 #define CTRL_MST_ACK			BIT(7)
 #define CTRL_HJ_ACK			BIT(6)
@@ -1186,7 +1187,7 @@ static int cdns_i3c_master_bus_init(struct i3c_master_controller *m)
 	struct cdns_i3c_master *master = to_cdns_i3c_master(m);
 	unsigned long pres_step, sysclk_rate, max_i2cfreq;
 	struct i3c_bus *bus = i3c_master_get_bus(m);
-	u32 ctrl, prescl0, prescl1, pres, low;
+	u32 ctrl, prescl0, prescl1, pres, low, thd_del;
 	struct i3c_device_info info = { };
 	int ret, ncycles;
 
@@ -1264,6 +1265,8 @@ static int cdns_i3c_master_bus_init(struct i3c_master_controller *m)
 	 * We will issue ENTDAA afterwards from the threaded IRQ handler.
 	 */
 	ctrl |= CTRL_HJ_ACK | CTRL_HJ_DISEC | CTRL_HALT_EN | CTRL_MCS_EN;
+	if (!of_property_read_u32(m->dev.of_node, "thd_del", &thd_del))
+		ctrl |= CTRL_THD_DEL(thd_del);
 	writel(ctrl, master->regs + CTRL);
 
 	cdns_i3c_master_enable(master);
Signature · 7 lines
2.14.0


_______________________________________________
linux-i3c mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-i3c
#2 lore raw

RE: [PATCH 1/3] i3c: master: add data hold delay support

From: Przemyslaw Gaj <[email protected]>
Date: Thu, Nov 14, 2019 at 05:51:53

Quoted 40 lines

This patch adds support for THD_DEL (Data Hold Delay) to Cadence I3C master constoller driver.

---
 drivers/i3c/master/i3c-master-cdns.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)
drivers/i3c/master/i3c-master-cdns.c +4 -1
index 10db0bf0655a..90ea98eef905 100644
@@ -60,6 +60,7 @@
 #define CTRL_HALT_EN			BIT(30)
 #define CTRL_MCS			BIT(29)
 #define CTRL_MCS_EN			BIT(28)
+#define CTRL_THD_DEL(x) 		(((x) << 24) & GENMASK(25, 24))
 #define CTRL_HJ_DISEC			BIT(8)
 #define CTRL_MST_ACK			BIT(7)
 #define CTRL_HJ_ACK			BIT(6)
@@ -1186,7 +1187,7 @@ static int cdns_i3c_master_bus_init(struct i3c_master_controller *m)
 	struct cdns_i3c_master *master = to_cdns_i3c_master(m);
 	unsigned long pres_step, sysclk_rate, max_i2cfreq;
 	struct i3c_bus *bus = i3c_master_get_bus(m);
-	u32 ctrl, prescl0, prescl1, pres, low;
+	u32 ctrl, prescl0, prescl1, pres, low, thd_del;
 	struct i3c_device_info info = { };
 	int ret, ncycles;
 
@@ -1264,6 +1265,8 @@ static int cdns_i3c_master_bus_init(struct i3c_master_controller *m)
 	 * We will issue ENTDAA afterwards from the threaded IRQ handler.
 	 */
 	ctrl |= CTRL_HJ_ACK | CTRL_HJ_DISEC | CTRL_HALT_EN | CTRL_MCS_EN;
+	if (!of_property_read_u32(m->dev.of_node, "thd_del", &thd_del))
+		ctrl |= CTRL_THD_DEL(thd_del);
 	writel(ctrl, master->regs + CTRL);
 
 	cdns_i3c_master_enable(master);
Signature · 1 line
2.14.0

Please change the commit message so it reflects Cadence driver.

Best regards,
Vitor Soares

_______________________________________________
linux-i3c mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-i3c

author maintainer
#3 lore raw

PATCH 2/3 dt-bindings: i3c: Document data hold delay feature

In reply to Przemyslaw Gaj

Documenting THD_DEL (Data Hold Delay) feature of Cadence I3C master controller.

---
 Documentation/devicetree/bindings/i3c/cdns,i3c-master.txt | 6 ++++++
 1 file changed, 6 insertions(+)
Documentation/devicetree/bindings/i3c/cdns,i3c-master.txt +6 -0
index 1cf6182f888c..7d6349354390 100644
@@ -21,6 +21,12 @@ Documentation/devicetree/bindings/i3c/i3c.txt for more details):
 - i2c-scl-hz
 - i3c-scl-hz
 
+Optional properties, Cadence I3C master controller specific:
+
+- thd_del: Data Hold Delay. Sets data hold delay (i.e. SDA_OUT data
+	   line will change its value with extra delay that specified
+	   in this register).
+
 I3C device connected on the bus follow the generic description (see
 Documentation/devicetree/bindings/i3c/i3c.txt for more details).
 
Signature · 7 lines
2.14.0


_______________________________________________
linux-i3c mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-i3c
#4 lore raw

Re: [PATCH 2/3] dt-bindings: i3c: Document data hold delay feature

Quoted 13 lines · On Thu, 14 Nov 2019 06:51:54 +0100 Przemyslaw Gaj <[email protected]> wrote:

Documenting THD_DEL (Data Hold Delay) feature of Cadence I3C master controller.

---
 Documentation/devicetree/bindings/i3c/cdns,i3c-master.txt | 6 ++++++
 1 file changed, 6 insertions(+)
Documentation/devicetree/bindings/i3c/cdns,i3c-master.txt +0 -0
index 1cf6182f888c..7d6349354390 100644

Documentation/devicetree/bindings/i3c/cdns,i3c-master.txt Documentation/devicetree/bindings/i3c/i3c.txt for more details): · line 28

Quoted 8 lines
Documentation/devicetree/bindings/i3c/cdns,i3c-master.txt +5 -0
@@ -21,6 +21,12 @@ Documentation/devicetree/bindings/i3c/i3c.txt for more details):
 - i2c-scl-hz
 - i3c-scl-hz
 
+Optional properties, Cadence I3C master controller specific:
+
+- thd_del: Data Hold Delay. Sets data hold delay (i.e. SDA_OUT data
+	   line will change its value with extra delay that specified
+	   in this register).

If it's a Cadence specific property, it should be prefixed with 'cdns,':

- cdns,thd-delay

Quick question about this delay, is it related to the IP integration in a SoC or is it board specific? In former case, I'd recommend attaching this piece of information to the compatible and have one compatible per SoC.

Documentation/devicetree/bindings/i3c/cdns,i3c-master.txt Documentation/devicetree/bindings/i3c/i3c.txt for more details): · line 32

Quoted 4 lines
Documentation/devicetree/bindings/i3c/cdns,i3c-master.txt +1 -0
+
 I3C device connected on the bus follow the generic description (see
 Documentation/devicetree/bindings/i3c/i3c.txt for more details).
 

_______________________________________________
linux-i3c mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-i3c

#5 lore raw

Re: [PATCH 2/3] dt-bindings: i3c: Document data hold delay feature

Quoted 27 lines · On Thu, 14 Nov 2019 10:15:49 +0100 Boris Brezillon…
Quoted 13 lines · On Thu, 14 Nov 2019 06:51:54 +0100 Przemyslaw Gaj <[email protected]> wrote:

Documenting THD_DEL (Data Hold Delay) feature of Cadence I3C master controller.

---
 Documentation/devicetree/bindings/i3c/cdns,i3c-master.txt | 6 ++++++
 1 file changed, 6 insertions(+)
Documentation/devicetree/bindings/i3c/cdns,i3c-master.txt +0 -0
index 1cf6182f888c..7d6349354390 100644

Documentation/devicetree/bindings/i3c/cdns,i3c-master.txt Documentation/devicetree/bindings/i3c/i3c.txt for more details): · line 28

Quoted 8 lines
Documentation/devicetree/bindings/i3c/cdns,i3c-master.txt +5 -0
@@ -21,6 +21,12 @@ Documentation/devicetree/bindings/i3c/i3c.txt for more details):
 - i2c-scl-hz
 - i3c-scl-hz
 
+Optional properties, Cadence I3C master controller specific:
+
+- thd_del: Data Hold Delay. Sets data hold delay (i.e. SDA_OUT data
+	   line will change its value with extra delay that specified
+	   in this register).  

If it's a Cadence specific property, it should be prefixed with 'cdns,':

Quoted 1 line

- cdns,thd-delay

Oh, and you need to specify the unit. Given the code, I suspect it's in clk-cycles, which is not great, since you probably want the delay to always be the same no matter the frequency of the clk driving the I3C master block.

Quoted 11 lines

Quick question about this delay, is it related to the IP integration in a SoC or is it board specific? In former case, I'd recommend attaching this piece of information to the compatible and have one compatible per SoC.

Quoted 4 lines
Documentation/devicetree/bindings/i3c/cdns,i3c-master.txt +1 -0
+
 I3C device connected on the bus follow the generic description (see
 Documentation/devicetree/bindings/i3c/i3c.txt for more details).
 

_______________________________________________
linux-i3c mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-i3c

author maintainer
#6 lore raw

Re: [PATCH 2/3] dt-bindings: i3c: Document data hold delay feature

Quoted 31 lines · Boris Brezillon
Quoted 27 lines · On Thu, 14 Nov 2019 10:15:49 +0100 Boris Brezillon…
Quoted 13 lines · On Thu, 14 Nov 2019 06:51:54 +0100 Przemyslaw Gaj <[email protected]> wrote:

Documenting THD_DEL (Data Hold Delay) feature of Cadence I3C master controller.

---
 Documentation/devicetree/bindings/i3c/cdns,i3c-master.txt | 6 ++++++
 1 file changed, 6 insertions(+)
Documentation/devicetree/bindings/i3c/cdns,i3c-master.txt +0 -0
index 1cf6182f888c..7d6349354390 100644

Documentation/devicetree/bindings/i3c/cdns,i3c-master.txt Documentation/devicetree/bindings/i3c/i3c.txt for more details): · line 28

Quoted 8 lines
Documentation/devicetree/bindings/i3c/cdns,i3c-master.txt +5 -0
@@ -21,6 +21,12 @@ Documentation/devicetree/bindings/i3c/i3c.txt for more details):
 - i2c-scl-hz
 - i3c-scl-hz
 
+Optional properties, Cadence I3C master controller specific:
+
+- thd_del: Data Hold Delay. Sets data hold delay (i.e. SDA_OUT data
+	   line will change its value with extra delay that specified
+	   in this register).  

If it's a Cadence specific property, it should be prefixed with 'cdns,':

Quoted 1 line
Quoted 1 line

- cdns,thd-delay

Ok.

Quoted 6 lines

Oh, and you need to specify the unit. Given the code, I suspect it's in clk-cycles, which is not great, since you probably want the delay to always be the same no matter the frequency of the clk driving the I3C master block.

Actually, this is encoded value. 3 means no delay, 2 - 1x clk delay, 1 - 2x clk dealy, 0 - 3x clk delay. I should mention about that in the documentation.

Quoted 6 lines
Quoted 6 lines

Quick question about this delay, is it related to the IP integration in a SoC or is it board specific? In former case, I'd recommend attaching this piece of information to the compatible and have one compatible per SoC.

This is spec requirement, slave shouldn't see SDA change before SCL. It is possible to achive this requirement during SoC physical design. If this is not the case, you can achive this using thd_del functionality. For now this is generic driver for our controller. So the question is, what should I do? Should I add default value for existing compatible and wait for different SoCs compatibility?

Quoted 6 lines
Quoted 5 lines
Quoted 4 lines
Documentation/devicetree/bindings/i3c/cdns,i3c-master.txt +1 -0
+
 I3C device connected on the bus follow the generic description (see
 Documentation/devicetree/bindings/i3c/i3c.txt for more details).
 
Signature · 7 lines
-- 
Przemyslaw Gaj

_______________________________________________
linux-i3c mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-i3c
#7 lore raw

Re: [PATCH 2/3] dt-bindings: i3c: Document data hold delay feature

Quoted 43 lines · On Thu, 14 Nov 2019 11:54:32 +0100 Przemyslaw Gaj <[email protected]> wrote:
Quoted 31 lines · Boris Brezillon
Quoted 27 lines · On Thu, 14 Nov 2019 10:15:49 +0100 Boris Brezillon…
Quoted 13 lines · On Thu, 14 Nov 2019 06:51:54 +0100 Przemyslaw Gaj <[email protected]> wrote:

Documenting THD_DEL (Data Hold Delay) feature of Cadence I3C master controller.

---
 Documentation/devicetree/bindings/i3c/cdns,i3c-master.txt | 6 ++++++
 1 file changed, 6 insertions(+)
Documentation/devicetree/bindings/i3c/cdns,i3c-master.txt +0 -0
index 1cf6182f888c..7d6349354390 100644

Documentation/devicetree/bindings/i3c/cdns,i3c-master.txt Documentation/devicetree/bindings/i3c/i3c.txt for more details): · line 28

Quoted 8 lines
Documentation/devicetree/bindings/i3c/cdns,i3c-master.txt +5 -0
@@ -21,6 +21,12 @@ Documentation/devicetree/bindings/i3c/i3c.txt for more details):
 - i2c-scl-hz
 - i3c-scl-hz
 
+Optional properties, Cadence I3C master controller specific:
+
+- thd_del: Data Hold Delay. Sets data hold delay (i.e. SDA_OUT data
+	   line will change its value with extra delay that specified
+	   in this register).    

If it's a Cadence specific property, it should be prefixed with 'cdns,':

Quoted 1 line
Quoted 1 line

- cdns,thd-delay

Ok.

Quoted 6 lines

Oh, and you need to specify the unit. Given the code, I suspect it's in clk-cycles, which is not great, since you probably want the delay to always be the same no matter the frequency of the clk driving the I3C master block.

Quoted 3 lines

Actually, this is encoded value. 3 means no delay, 2 - 1x clk delay, 1 - 2x clk dealy, 0 - 3x clk delay. I should mention about that in the documentation.

By clk you mean SCL or the clock driving the I3C master logic (which is likely to run at a higher freq)?

Quoted 8 lines
Quoted 7 lines
Quoted 7 lines

Quick question about this delay, is it related to the IP integration in a SoC or is it board specific? In former case, I'd recommend attaching this piece of information to the compatible and have one compatible per SoC.

Quoted 6 lines

This is spec requirement, slave shouldn't see SDA change before SCL. It is possible to achive this requirement during SoC physical design. If this is not the case, you can achive this using thd_del functionality. For now this is generic driver for our controller. So the question is, what should I do? Should I add default value for existing compatible and wait for different SoCs compatibility?

Yes, exactly.

_______________________________________________
linux-i3c mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-i3c

author maintainer
#8 lore raw

Re: [PATCH 2/3] dt-bindings: i3c: Document data hold delay feature

Quoted 49 lines · Boris Brezillon

EXTERNAL MAIL

Quoted 43 lines · On Thu, 14 Nov 2019 11:54:32 +0100 Przemyslaw Gaj <[email protected]> wrote:
Quoted 31 lines · Boris Brezillon
Quoted 27 lines · On Thu, 14 Nov 2019 10:15:49 +0100 Boris Brezillon…
Quoted 13 lines · On Thu, 14 Nov 2019 06:51:54 +0100 Przemyslaw Gaj <[email protected]> wrote:

Documenting THD_DEL (Data Hold Delay) feature of Cadence I3C master controller.

---
 Documentation/devicetree/bindings/i3c/cdns,i3c-master.txt | 6 ++++++
 1 file changed, 6 insertions(+)
Documentation/devicetree/bindings/i3c/cdns,i3c-master.txt +0 -0
index 1cf6182f888c..7d6349354390 100644

Documentation/devicetree/bindings/i3c/cdns,i3c-master.txt Documentation/devicetree/bindings/i3c/i3c.txt for more details): · line 28

Quoted 8 lines
Documentation/devicetree/bindings/i3c/cdns,i3c-master.txt +5 -0
@@ -21,6 +21,12 @@ Documentation/devicetree/bindings/i3c/i3c.txt for more details):
 - i2c-scl-hz
 - i3c-scl-hz
 
+Optional properties, Cadence I3C master controller specific:
+
+- thd_del: Data Hold Delay. Sets data hold delay (i.e. SDA_OUT data
+	   line will change its value with extra delay that specified
+	   in this register).    

If it's a Cadence specific property, it should be prefixed with 'cdns,':

Quoted 1 line
Quoted 1 line

- cdns,thd-delay

Ok.

Quoted 6 lines

Oh, and you need to specify the unit. Given the code, I suspect it's in clk-cycles, which is not great, since you probably want the delay to always be the same no matter the frequency of the clk driving the I3C master block.

Quoted 6 lines
Quoted 3 lines

Actually, this is encoded value. 3 means no delay, 2 - 1x clk delay, 1 - 2x clk dealy, 0 - 3x clk delay. I should mention about that in the documentation.

By clk you mean SCL or the clock driving the I3C master logic (which is likely to run at a higher freq)?

Yes, this is the clock driving I3C master logic.

Quoted 9 lines
Quoted 9 lines
Quoted 8 lines
Quoted 8 lines

Quick question about this delay, is it related to the IP integration in a SoC or is it board specific? In former case, I'd recommend attaching this piece of information to the compatible and have one compatible per SoC.

Quoted 8 lines
Quoted 6 lines

This is spec requirement, slave shouldn't see SDA change before SCL. It is possible to achive this requirement during SoC physical design. If this is not the case, you can achive this using thd_del functionality. For now this is generic driver for our controller. So the question is, what should I do? Should I add default value for existing compatible and wait for different SoCs compatibility?

Yes, exactly.

Ok. In that case I don't need dt binding.

Signature · 7 lines
-- 
Przemyslaw Gaj

_______________________________________________
linux-i3c mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-i3c
author maintainer
#9 lore raw

PATCH 3/3 MAINTAINERS: add myself as maintainer of Cadence I3C master controller driver

In reply to Przemyslaw Gaj

As discussed with Boris Brezillon - I'm adding myself as the maintainer.

---
 MAINTAINERS | 6 ++++++
 1 file changed, 6 insertions(+)
MAINTAINERS +6 -0
index c4c532c70b86..afdce16d2be5 100644
@@ -7821,6 +7821,12 @@ S:	Maintained
 F:	Documentation/devicetree/bindings/i3c/snps,dw-i3c-master.txt
 F:	drivers/i3c/master/dw*
 
+I3C DRIVER FOR CADENCE I3C MASTER IP
+M:      Przemysław Gaj <[email protected]>
+S:      Maintained
+F:      Documentation/devicetree/bindings/i3c/cdns,i3c-master.txt
+F:      drivers/i3c/master/i3c-master-cdns.c
+
 IA64 (Itanium) PLATFORM
 M:	Tony Luck <[email protected]>
 M:	Fenghua Yu <[email protected]>
Signature · 7 lines
2.14.0


_______________________________________________
linux-i3c mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-i3c
#10 lore raw

Re: [PATCH 3/3] MAINTAINERS: add myself as maintainer of Cadence I3C master controller driver

Quoted 1 line · On Thu, 14 Nov 2019 06:51:55 +0100 Przemyslaw Gaj <[email protected]> wrote:

As discussed with Boris Brezillon - I'm adding myself as the maintainer.

Queued to i3c/next.

Quoted 13 lines
---
 MAINTAINERS | 6 ++++++
 1 file changed, 6 insertions(+)
MAINTAINERS +0 -0
index c4c532c70b86..afdce16d2be5 100644
@@ -7821,6 +7821,12 @@ S:	Maintained
 F:	Documentation/devicetree/bindings/i3c/snps,dw-i3c-master.txt
 F:	drivers/i3c/master/dw*

MAINTAINERS S: Maintained · line 7832

Quoted 10 lines
MAINTAINERS +6 -0
 
+I3C DRIVER FOR CADENCE I3C MASTER IP
+M:      Przemysław Gaj <[email protected]>
+S:      Maintained
+F:      Documentation/devicetree/bindings/i3c/cdns,i3c-master.txt
+F:      drivers/i3c/master/i3c-master-cdns.c
+
 IA64 (Itanium) PLATFORM
 M:	Tony Luck <[email protected]>
 M:	Fenghua Yu <[email protected]>

_______________________________________________
linux-i3c mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-i3c

#11 lore raw
In reply to Przemyslaw Gaj

Hi,

From: Przemyslaw Gaj <[email protected]>
Date: Thu, Nov 14, 2019 at 05:51:52

Quoted 12 lines

Add THD_DEL (Data Hold Delay) support. After testing different scenarios, on various systems, sometimes there is a need to delay SDA_OUT propagation.

Adding support which allows to configure that delay using the device tree parameter.

Przemyslaw Gaj (3):
i3c: master: add data hold delay support
dt-bindings: i3c: Document data hold delay feature
MAINTAINERS: add myself as maintainer of Cadence I3C master controller
driver

Quoted 7 lines
 Documentation/devicetree/bindings/i3c/cdns,i3c-master.txt | 6 ++++++
 MAINTAINERS                                               | 6 ++++++
 drivers/i3c/master/i3c-master-cdns.c                      | 5 ++++-
 3 files changed, 16 insertions(+), 1 deletion(-)
Signature · 1 line
2.14.0

I'm trying to understand this implementation and the use case but I'm not understanding.

Could you please elaborate?

Best regards,
Vitor Soares

_______________________________________________
linux-i3c mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-i3c

author maintainer
#12 lore raw

Hi Vitor,

Quoted 23 lines · Vitor Soares

Hi,

From: Przemyslaw Gaj <[email protected]>
Date: Thu, Nov 14, 2019 at 05:51:52

Quoted 17 lines

Add THD_DEL (Data Hold Delay) support. After testing different scenarios, on various systems, sometimes there is a need to delay SDA_OUT propagation.

Adding support which allows to configure that delay using the device tree parameter.

Przemyslaw Gaj (3):
i3c: master: add data hold delay support
dt-bindings: i3c: Document data hold delay feature
MAINTAINERS: add myself as maintainer of Cadence I3C master controller
driver

 Documentation/devicetree/bindings/i3c/cdns,i3c-master.txt | 6 ++++++
 MAINTAINERS                                               | 6 ++++++
 drivers/i3c/master/i3c-master-cdns.c                      | 5 ++++-
 3 files changed, 16 insertions(+), 1 deletion(-)
Quoted 7 lines
Quoted 2 lines
Signature · 1 line
2.14.0

I'm trying to understand this implementation and the use case but I'm not understanding.

Could you please elaborate?

There is new controller functionality which allows to meet SDA to SCL propagation time. Baisically, you can delay SDA propagation using this functionality to meet I3C MIPI spec timing. This is very useful for some SoC designs.

From Spec:
Table 75 of MIPI I3C Specification 1.0 (page 142) defines non-zero minimal tHD_PP timing on master output (Fig 65). This setting allows to meet this timing on master's soc outputs, regardless of PCB balancing.

Quoted 3 lines

Best regards,
Vitor Soares

Signature · 7 lines
-- 
Przemyslaw Gaj

_______________________________________________
linux-i3c mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-i3c
#13 lore raw

Hi,

From: Przemyslaw Gaj <[email protected]>
Date: Thu, Nov 14, 2019 at 12:48:56

Quoted 6 lines

There is new controller functionality which allows to meet SDA to SCL propagation time. Baisically, you can delay SDA propagation using this functionality to meet I3C MIPI spec timing. This is very useful for some SoC designs.

Quoted 12 lines

From Spec:
Table 75 of MIPI I3C Specification 1.0 (page 142) defines non-zero minimal tHD_PP timing on master output (Fig 65). This setting allows to meet this timing on master's soc outputs, regardless of PCB balancing.

Quoted 3 lines

Best regards,
Vitor Soares

Signature · 2 lines
-- 
Przemyslaw Gaj

Thanks for pointing me the right place to read about it. I would suggest you to put that info in the patch as well.

Best regards,
Vitor Soares
_______________________________________________
linux-i3c mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-i3c

#14 lore raw
Quoted 19 lines · On Thu, 14 Nov 2019 13:31:33 +0000 Vitor Soares <[email protected]>…

Hi,

From: Przemyslaw Gaj <[email protected]>
Date: Thu, Nov 14, 2019 at 12:48:56

Quoted 15 lines

There is new controller functionality which allows to meet SDA to SCL propagation time. Baisically, you can delay SDA propagation using this functionality to meet I3C MIPI spec timing. This is very useful for some SoC designs.

From Spec:
Table 75 of MIPI I3C Specification 1.0 (page 142) defines non-zero minimal tHD_PP timing on master output (Fig 65). This setting allows to meet this timing on master's soc outputs, regardless of PCB balancing.

Quoted 3 lines

Best regards,
Vitor Soares

Quoted 6 lines
Quoted 3 lines
Signature · 2 lines
-- 
Przemyslaw Gaj  

Thanks for pointing me the right place to read about it. I would suggest you to put that info in the patch as well.

I agree. The spec details should be mentioned in the commit message and in the code as well (in a comment).

_______________________________________________
linux-i3c mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-i3c