Skip to contentSkip to conversation

RZN1 UART DMA support

PATCH In review Miquel Raynal 4 participants 25 messages Mar 10, 2022 – Mar 16, 2022

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

cover RZN1 UART DMA support

Hello,

Support for the RZN1 DMA engine allows us adapt a little bit the 8250 DW UART driver with to bring DMA support for this SoC.

This short series applies on top of the series bringing RZN1 DMA support, currently on its v4, see [1]. Technically speaking, only the DT patch needs to be applied after [1]. The other patches can come in at any moment, because if no "dmas" property is provided in the DT, DMA support will simply be ignored.

[1] https://lore.kernel.org/dmaengine/[email protected]/T/#mce6fec36e16dca560ab18935c273fcaf794a1cc4

Thanks,
Miquèl

Miquel Raynal (2):
serial: 8250_dw: Provide the RZN1 CPR register value
ARM: dts: r9a06g032: Fill the UART DMA properties

Phil Edworthy (5):
serial: 8250_dma: Use ->tx_dma function pointer to start next DMA
serial: 8250_dw: Move the per-device structure
serial: 8250_dw: Use a fallback CPR value if not synthesized
serial: 8250_dw: Add a dma_capable bit to the platform data
serial: 8250_dw: Add support for RZ/N1 DMA

 arch/arm/boot/dts/r9a06g032.dtsi     |  15 ++++
 drivers/tty/serial/8250/8250_dma.c   |   4 +-
 drivers/tty/serial/8250/8250_dw.c    | 119 +++++++++++++++++++++++----
 drivers/tty/serial/8250/8250_dwlib.c |  17 +++-
 drivers/tty/serial/8250/8250_dwlib.h |  22 +++++
 5 files changed, 155 insertions(+), 22 deletions(-)
Signature · 1 line
2.27.0
#1 lore raw

PATCH 1/7 serial: 8250_dma: Use ->tx_dma function pointer to start next DMA

From: Phil Edworthy <[email protected]>

The 8250 driver is quite flexible. Regarding DMA handling, there is the possibility to either use the default helper (serial8250_tx_dma()) or call a specific function. Only the omap and brcm implementation do that. In both cases, they don't use the serial8250_tx_dma() helper at all.

As we are going to write a new DMA handling function for the RZ/N1 SoCs which will use the serial8250_tx_dma() implementation (preceded by a couple of register writes), we need the ->tx_dma() pointer to link to our own function, but within the __dma_tx_complete() helper we also need to call our own implementation instead of the default one directly.

In order to do that, let's call ->tx_dma() instead of serial8250_tx_dma() from __dma_tx_complete().

---
 drivers/tty/serial/8250/8250_dma.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/tty/serial/8250/8250_dma.c +1 -1
index 890fa7ddaa7f..a0563f2341ac 100644
@@ -33,7 +33,7 @@ static void __dma_tx_complete(void *param)
 	if (uart_circ_chars_pending(xmit) < WAKEUP_CHARS)
 		uart_write_wakeup(&p->port);
 
-	ret = serial8250_tx_dma(p);
+	ret = dma->tx_dma(p);
 	if (ret)
 		serial8250_set_THRI(p);
 
Signature · 1 line
2.27.0
#2 lore raw

Re: [PATCH 1/7] serial: 8250_dma: Use ->tx_dma function pointer to start next DMA

Quoted 8 lines · Miquel Raynal

From: Phil Edworthy <[email protected]>

The 8250 driver is quite flexible. Regarding DMA handling, there is the possibility to either use the default helper (serial8250_tx_dma()) or call a specific function. Only the omap and brcm implementation do that. In both cases, they don't use the serial8250_tx_dma() helper at all.

Quoted 8 lines

As we are going to write a new DMA handling function for the RZ/N1 SoCs which will use the serial8250_tx_dma() implementation (preceded by a couple of register writes), we need the ->tx_dma() pointer to link to our own function, but within the __dma_tx_complete() helper we also need to call our own implementation instead of the default one directly.

In order to do that, let's call ->tx_dma() instead of serial8250_tx_dma() from __dma_tx_complete().

In 8250 driver the pattern is to give the generic function "do" name and then call it in a wrapper:

	if (->foo())
		return ->foo();

	return serial8250_do_foo();

There are plenty of examples in that driver.

Signature · 2 lines
With Best Regards,
Andy Shevchenko
#3 lore raw

PATCH 2/7 serial: 8250_dw: Move the per-device structure

In reply to Miquel Raynal

From: Phil Edworthy <[email protected]>

This structure needs to be reused from dwlib, so let's move it into a shared header. There is no functional change.

[[email protected]: Extracted from a bigger change]
---
 drivers/tty/serial/8250/8250_dw.c    | 16 ----------------
 drivers/tty/serial/8250/8250_dwlib.h | 16 ++++++++++++++++
 2 files changed, 16 insertions(+), 16 deletions(-)
drivers/tty/serial/8250/8250_dw.c +0 -16
index 1769808031c5..ee7562a9ec76 100644
@@ -37,22 +37,6 @@
 /* DesignWare specific register fields */
 #define DW_UART_MCR_SIRE		BIT(6)
 
-struct dw8250_data {
-	struct dw8250_port_data	data;
-
-	u8			usr_reg;
-	int			msr_mask_on;
-	int			msr_mask_off;
-	struct clk		*clk;
-	struct clk		*pclk;
-	struct notifier_block	clk_notifier;
-	struct work_struct	clk_work;
-	struct reset_control	*rst;
-
-	unsigned int		skip_autocfg:1;
-	unsigned int		uart_16550_compatible:1;
-};
-
 static inline struct dw8250_data *to_dw8250_data(struct dw8250_port_data *data)
 {
 	return container_of(data, struct dw8250_data, data);
drivers/tty/serial/8250/8250_dwlib.h +16 -0
index 83d528e5cc21..ef63eaf7e598 100644
@@ -16,5 +16,21 @@ struct dw8250_port_data {
 	u8			dlf_size;
 };
 
+struct dw8250_data {
+	struct dw8250_port_data	data;
+
+	u8			usr_reg;
+	int			msr_mask_on;
+	int			msr_mask_off;
+	struct clk		*clk;
+	struct clk		*pclk;
+	struct notifier_block	clk_notifier;
+	struct work_struct	clk_work;
+	struct reset_control	*rst;
+
+	unsigned int		skip_autocfg:1;
+	unsigned int		uart_16550_compatible:1;
+};
+
 void dw8250_do_set_termios(struct uart_port *p, struct ktermios *termios, struct ktermios *old);
 void dw8250_setup_port(struct uart_port *p);
Signature · 1 line
2.27.0
#4 lore raw

Re: [PATCH 2/7] serial: 8250_dw: Move the per-device structure

Quoted 4 lines · Miquel Raynal

From: Phil Edworthy <[email protected]>

This structure needs to be reused from dwlib, so let's move it into a shared header. There is no functional change.

...

drivers/tty/serial/8250/8250_dwlib.h struct dw8250_port_data { · line 22

Quoted 4 lines
drivers/tty/serial/8250/8250_dwlib.h +3 -0
+struct dw8250_data {
+	struct dw8250_port_data	data;
 
+	u8			usr_reg;

I believe types.h is already in the inclusion block.

drivers/tty/serial/8250/8250_dwlib.h struct dw8250_port_data { · line 26

Quoted 5 lines
drivers/tty/serial/8250/8250_dwlib.h +4 -0
+	int			msr_mask_on;
+	int			msr_mask_off;
 
+	struct clk		*clk;
+	struct clk		*pclk;

Do you need forward declarations?

drivers/tty/serial/8250/8250_dwlib.h struct dw8250_port_data { · line 28

Quoted 2 lines
drivers/tty/serial/8250/8250_dwlib.h +2 -0
+	struct notifier_block	clk_notifier;
+	struct work_struct	clk_work;

I haven't seen the change in the inclusion block.
Don't you miss necessary headers?

drivers/tty/serial/8250/8250_dwlib.h struct dw8250_port_data { · line 29

Quoted 1 line
drivers/tty/serial/8250/8250_dwlib.h +1 -0
+	struct reset_control	*rst;

Do you need forward declaration?

Quoted 3 lines
drivers/tty/serial/8250/8250_dwlib.h +3 -0
+	unsigned int		skip_autocfg:1;
+	unsigned int		uart_16550_compatible:1;
+};
Signature · 2 lines
With Best Regards,
Andy Shevchenko
#5 lore raw

PATCH 3/7 serial: 8250_dw: Use a fallback CPR value if not synthesized

In reply to Miquel Raynal

From: Phil Edworthy <[email protected]>

This UART controller can be synthesized without the CPR register. In that case, let's use the platform information to provide a CPR value.

---
 drivers/tty/serial/8250/8250_dwlib.c | 10 ++++++++--
 drivers/tty/serial/8250/8250_dwlib.h |  4 ++++
 2 files changed, 12 insertions(+), 2 deletions(-)
drivers/tty/serial/8250/8250_dwlib.c +8 -2
index 622d3b0d89e7..5cf298c5a0f9 100644
@@ -3,6 +3,7 @@
 
 #include <linux/bitops.h>
 #include <linux/device.h>
+#include <linux/of_device.h>
 #include <linux/io.h>
 #include <linux/kernel.h>
 #include <linux/serial_8250.h>
@@ -90,6 +91,7 @@ EXPORT_SYMBOL_GPL(dw8250_do_set_termios);
 void dw8250_setup_port(struct uart_port *p)
 {
 	struct uart_8250_port *up = up_to_u8250p(p);
+	const struct dw8250_platform_data *plat = of_device_get_match_data(up->port.dev);
 	u32 reg;
 
 	/*
@@ -116,8 +118,12 @@ void dw8250_setup_port(struct uart_port *p)
 	}
 
 	reg = dw8250_readl_ext(p, DW_UART_CPR);
-	if (!reg)
-		return;
+	if (!reg) {
+		if (!plat)
+			return;
+
+		reg = plat->cpr;
+	}
 
 	/* Select the type based on FIFO */
 	if (reg & DW_UART_CPR_FIFO_MODE) {
drivers/tty/serial/8250/8250_dwlib.h +4 -0
index ef63eaf7e598..ffce2744a28e 100644
@@ -16,6 +16,10 @@ struct dw8250_port_data {
 	u8			dlf_size;
 };
 
+struct dw8250_platform_data {
+	u32 cpr;
+};
+
 struct dw8250_data {
 	struct dw8250_port_data	data;
 
Signature · 1 line
2.27.0
#6 lore raw

Re: [PATCH 3/7] serial: 8250_dw: Use a fallback CPR value if not synthesized

Quoted 4 lines · Miquel Raynal

From: Phil Edworthy <[email protected]>

This UART controller can be synthesized without the CPR register. In that case, let's use the platform information to provide a CPR value.

...

Review

Quoted 3 lines
(quoted diff) +2 -0
+#include <linux/of_device.h>
 
+	const struct dw8250_platform_data *plat = of_device_get_match_data(up->port.dev);

No. Please use device property APIs.

Signature · 2 lines
With Best Regards,
Andy Shevchenko
#7 lore raw

Re: [PATCH 3/7] serial: 8250_dw: Use a fallback CPR value if not synthesized

Hi Andy,

[email protected] wrote on Thu, 10 Mar 2022 20:02:41 +0200:

Quoted 6 lines · Miquel Raynal
Quoted 4 lines

From: Phil Edworthy <[email protected]>

This UART controller can be synthesized without the CPR register. In that case, let's use the platform information to provide a CPR value.

Quoted 7 lines

...

Review

Quoted 3 lines
(quoted diff) +2 -0
+#include <linux/of_device.h>  
 
+	const struct dw8250_platform_data *plat = of_device_get_match_data(up->port.dev);  

No. Please use device property APIs.

Are you suggesting the use of CPR DT property? If yes, what is the point if there is one CPR per SoC? I would argue that DT description is already quite complex and supporting one or another register is up to the OS as long as we can map CPR registers to SoCs?

TBH Phil initially introduced a DT property which I turned into a pdata entry because when I can avoid playing with the bindings, I generally do so.

Thanks,
Miquèl

#8 lore raw

Re: [PATCH 3/7] serial: 8250_dw: Use a fallback CPR value if not synthesized

Quoted 4 lines · Miquel Raynal

[email protected] wrote on Thu, 10 Mar 2022 20:02:41 +0200:

Quoted 2 lines · Miquel Raynal
Quoted 1 line

From: Phil Edworthy <[email protected]>

...

Quoted 10 lines
Quoted 5 lines

Review

Quoted 3 lines
(quoted diff) +2 -0
+#include <linux/of_device.h>  
 
+	const struct dw8250_platform_data *plat = of_device_get_match_data(up->port.dev);  

No. Please use device property APIs.

Are you suggesting the use of CPR DT property? If yes, what is the point if there is one CPR per SoC? I would argue that DT description is already quite complex and supporting one or another register is up to the OS as long as we can map CPR registers to SoCs?

I'm suggesting to use device property APIs, I'm not talking about ABI. In this case instead of above you may simply do

#include <linux/property.h>

	const struct dw8250_platform_data *plat = device_get_match_data(up->port.dev);
Signature · 2 lines
With Best Regards,
Andy Shevchenko
#9 lore raw

PATCH 4/7 serial: 8250_dw: Provide the RZN1 CPR register value

In reply to Miquel Raynal

On the RZN1 SoC, the DW serial IP has been synthesized without CPR support. In order to continue being able to parse the hardware capabilities, provide the equivalent register value as platform data.

---
 drivers/tty/serial/8250/8250_dw.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)
drivers/tty/serial/8250/8250_dw.c +5 -1
index ee7562a9ec76..1f7a423d6ef2 100644
@@ -677,11 +677,15 @@ static const struct dev_pm_ops dw8250_pm_ops = {
 	SET_RUNTIME_PM_OPS(dw8250_runtime_suspend, dw8250_runtime_resume, NULL)
 };
 
+static const struct dw8250_platform_data rzn1_pdata = {
+	.cpr = 0x00012f32,
+};
+
 static const struct of_device_id dw8250_of_match[] = {
 	{ .compatible = "snps,dw-apb-uart" },
 	{ .compatible = "cavium,octeon-3860-uart" },
 	{ .compatible = "marvell,armada-38x-uart" },
-	{ .compatible = "renesas,rzn1-uart" },
+	{ .compatible = "renesas,rzn1-uart", .data = &rzn1_pdata },
 	{ .compatible = "starfive,jh7100-uart" },
 	{ /* Sentinel */ }
 };
Signature · 1 line
2.27.0
#10 lore raw

PATCH 5/7 serial: 8250_dw: Add a dma_capable bit to the platform data

In reply to Miquel Raynal

From: Phil Edworthy <[email protected]>

The CPR register can give the information whether the IP is DMA capable or not. Let's extract this information and use it to discriminate when the DMA can be hooked up or not.

---
 drivers/tty/serial/8250/8250_dw.c    | 4 ++--
 drivers/tty/serial/8250/8250_dwlib.c | 7 +++++--
 drivers/tty/serial/8250/8250_dwlib.h | 1 +
 3 files changed, 8 insertions(+), 4 deletions(-)
drivers/tty/serial/8250/8250_dw.c +2 -2
index 1f7a423d6ef2..c0f54284bc70 100644
@@ -556,8 +556,8 @@ static int dw8250_probe(struct platform_device *pdev)
 	if (!data->skip_autocfg)
 		dw8250_setup_port(p);
 
-	/* If we have a valid fifosize, try hooking up DMA */
-	if (p->fifosize) {
+	/* If we have a valid fifosize and DMA support, try hooking up DMA */
+	if (p->fifosize && data->dma_capable) {
 		data->data.dma.rxconf.src_maxburst = p->fifosize / 4;
 		data->data.dma.txconf.dst_maxburst = p->fifosize / 4;
 		up->dma = &data->data.dma;
drivers/tty/serial/8250/8250_dwlib.c +5 -2
index 5cf298c5a0f9..5ec7c12ed117 100644
@@ -92,6 +92,8 @@ void dw8250_setup_port(struct uart_port *p)
 {
 	struct uart_8250_port *up = up_to_u8250p(p);
 	const struct dw8250_platform_data *plat = of_device_get_match_data(up->port.dev);
+	struct dw8250_port_data *d = p->private_data;
+	struct dw8250_data *data = container_of(d, struct dw8250_data, data);
 	u32 reg;
 
 	/*
@@ -110,8 +112,6 @@ void dw8250_setup_port(struct uart_port *p)
 	dw8250_writel_ext(p, DW_UART_DLF, 0);
 
 	if (reg) {
-		struct dw8250_port_data *d = p->private_data;
-
 		d->dlf_size = fls(reg);
 		p->get_divisor = dw8250_get_divisor;
 		p->set_divisor = dw8250_set_divisor;
@@ -138,5 +138,8 @@ void dw8250_setup_port(struct uart_port *p)
 
 	if (reg & DW_UART_CPR_SIR_MODE)
 		up->capabilities |= UART_CAP_IRDA;
+
+	if (reg & DW_UART_CPR_DMA_EXTRA)
+		data->dma_capable = 1;
 }
 EXPORT_SYMBOL_GPL(dw8250_setup_port);
drivers/tty/serial/8250/8250_dwlib.h +1 -0
index ffce2744a28e..900b2321aff0 100644
@@ -34,6 +34,7 @@ struct dw8250_data {
 
 	unsigned int		skip_autocfg:1;
 	unsigned int		uart_16550_compatible:1;
+	unsigned int		dma_capable:1;
 };
 
 void dw8250_do_set_termios(struct uart_port *p, struct ktermios *termios, struct ktermios *old);
Signature · 1 line
2.27.0
#11 lore raw

Re: [PATCH 5/7] serial: 8250_dw: Add a dma_capable bit to the platform data

Quoted 5 lines · Miquel Raynal

From: Phil Edworthy <[email protected]>

The CPR register can give the information whether the IP is DMA capable or not. Let's extract this information and use it to discriminate when the DMA can be hooked up or not.

...

drivers/tty/serial/8250/8250_dwlib.c void dw8250_setup_port(struct uart_port *p) · line 143

Quoted 5 lines
drivers/tty/serial/8250/8250_dwlib.c +4 -0
+	/* If we have a valid fifosize and DMA support, try hooking up DMA */
+	if (p->fifosize && data->dma_capable) {
 
+	if (reg & DW_UART_CPR_DMA_EXTRA)
+		data->dma_capable = 1;

How many designs will be broken by this change?

...

drivers/tty/serial/8250/8250_dwlib.h struct dw8250_data { · line 37

Quoted 1 line
drivers/tty/serial/8250/8250_dwlib.h +1 -0
+	unsigned int		dma_capable:1;

Note, we use up->dma == NULL for no-DMA, no additional flag is needed. Just make sure that for your platform you enable DMA by filling that.

Signature · 2 lines
With Best Regards,
Andy Shevchenko
#12 lore raw

Re: [PATCH 5/7] serial: 8250_dw: Add a dma_capable bit to the platform data

Hi Andy,

[email protected] wrote on Thu, 10 Mar 2022 20:06:25 +0200:

Quoted 9 lines · Miquel Raynal
Quoted 5 lines

From: Phil Edworthy <[email protected]>

The CPR register can give the information whether the IP is DMA capable or not. Let's extract this information and use it to discriminate when the DMA can be hooked up or not.

...

Quoted 7 lines

drivers/tty/serial/8250/8250_dwlib.c void dw8250_setup_port(struct uart_port *p) · line 143

Quoted 5 lines
drivers/tty/serial/8250/8250_dwlib.c +4 -0
+	/* If we have a valid fifosize and DMA support, try hooking up DMA */
+	if (p->fifosize && data->dma_capable) {  
 
+	if (reg & DW_UART_CPR_DMA_EXTRA)
+		data->dma_capable = 1;  

How many designs will be broken by this change?

My understanding was that CPR registers where always synthesized until now even though it was not mandatory and that the RZN1 SoC was the first one to not embed it. My hope was that people using this driver would have brought "external" CPR support earlier if they needed it, but I understand this assumption might be wrong.

Anyway, I also hesitated to do something more custom for the RZN1 I'll try something else.

Quoted 7 lines

...

drivers/tty/serial/8250/8250_dwlib.h struct dw8250_data { · line 37

Quoted 1 line
drivers/tty/serial/8250/8250_dwlib.h +1 -0
+	unsigned int		dma_capable:1;  

Note, we use up->dma == NULL for no-DMA, no additional flag is needed. Just make sure that for your platform you enable DMA by filling that.

dma_capable is just a capability the SoC has. It was discovered at probe time and should be saved to know, later, if DMA can be hooked up or not. At the time we look at the CPR register we don't yet have DMA fields populated so its too early to set up->dma to NULL.

Thanks,
Miquèl

#13 lore raw

Re: [PATCH 5/7] serial: 8250_dw: Add a dma_capable bit to the platform data

Quoted 4 lines · Miquel Raynal

[email protected] wrote on Thu, 10 Mar 2022 20:06:25 +0200:

Quoted 2 lines · Miquel Raynal
Quoted 1 line

From: Phil Edworthy <[email protected]>

...

Quoted 8 lines
Quoted 7 lines

drivers/tty/serial/8250/8250_dwlib.c void dw8250_setup_port(struct uart_port *p) · line 143

Quoted 5 lines
drivers/tty/serial/8250/8250_dwlib.c +4 -0
+	/* If we have a valid fifosize and DMA support, try hooking up DMA */
+	if (p->fifosize && data->dma_capable) {  
 
+	if (reg & DW_UART_CPR_DMA_EXTRA)
+		data->dma_capable = 1;  

How many designs will be broken by this change?

Quoted 8 lines

My understanding was that CPR registers where always synthesized until now even though it was not mandatory and that the RZN1 SoC was the first one to not embed it. My hope was that people using this driver would have brought "external" CPR support earlier if they needed it, but I understand this assumption might be wrong.

Anyway, I also hesitated to do something more custom for the RZN1 I'll try something else.

My point is that you have put this requirement to the above conditionals. Have you checked all supported platforms that announce CPR that that bit is set when DMA is indeed in use?

AFAIK on Intel SoCs the native UART DMA signalling is not used, the integration between DMA and UART is done differently because it requires more signals to connect. It might be that I misread the documentation and this is not the case and we indeed set that bit as well.

Also, what to do with the platforms that have no CPR but support DMA currently?

...

Quoted 9 lines
Quoted 4 lines

drivers/tty/serial/8250/8250_dwlib.h struct dw8250_data { · line 37

Quoted 1 line
drivers/tty/serial/8250/8250_dwlib.h +1 -0
+	unsigned int		dma_capable:1;  

Note, we use up->dma == NULL for no-DMA, no additional flag is needed. Just make sure that for your platform you enable DMA by filling that.

dma_capable is just a capability the SoC has. It was discovered at probe time and should be saved to know, later, if DMA can be hooked up or not. At the time we look at the CPR register we don't yet have DMA fields populated so its too early to set up->dma to NULL.

'up->dma == NULL' check will tell you that, no?

Signature · 2 lines
With Best Regards,
Andy Shevchenko
#14 lore raw

PATCH 6/7 serial: 8250_dw: Add support for RZ/N1 DMA

In reply to Miquel Raynal

From: Phil Edworthy <[email protected]>

The Renesas RZ/N1 devices have a modified Synopsys DW UART. The modifications are mostly related to the DMA handlnig, and so this patch adds support for DMA.

The RZ/N1 UART must be used with the peripheral as the flow controller. This means the DMA length should also be programmed into UART registers.

Aside from this, there are some points to note about DMA burst sizes. First, DMA must not remove all of the data from the rx FIFO. Otherwise, we do not get a 'character timeout' interrupt, and so do not know that we should push data up the serial stack. Therefore, we have the rx threshold for generating an interrupt set to half the FIFO depth (this is the default for 16550A), and set the DMA burst size when reading the FIFO to a quarter of the FIFO depth.

Second, when transmitting data using DMA, the burst size must be limited to 1 byte to handle then case when transmitting just 1 byte. Otherwise the DMA doesn't complete the burst, and nothing happens.

---
 drivers/tty/serial/8250/8250_dma.c   |  2 +
 drivers/tty/serial/8250/8250_dw.c    | 97 ++++++++++++++++++++++++++++
 drivers/tty/serial/8250/8250_dwlib.h |  1 +
 3 files changed, 100 insertions(+)
drivers/tty/serial/8250/8250_dma.c +2 -0
index a0563f2341ac..0858c0b988a2 100644
@@ -105,6 +105,7 @@ int serial8250_tx_dma(struct uart_8250_port *p)
 	dma->tx_err = 1;
 	return ret;
 }
+EXPORT_SYMBOL_GPL(serial8250_tx_dma);
 
 int serial8250_rx_dma(struct uart_8250_port *p)
 {
@@ -130,6 +131,7 @@ int serial8250_rx_dma(struct uart_8250_port *p)
 
 	return 0;
 }
+EXPORT_SYMBOL_GPL(serial8250_rx_dma);
 
 void serial8250_rx_dma_flush(struct uart_8250_port *p)
 {
drivers/tty/serial/8250/8250_dw.c +97 -0

133 lines not on this page · raw message

drivers/tty/serial/8250/8250_dwlib.h +1 -0
index 900b2321aff0..014005c170e4 100644
@@ -35,6 +35,7 @@ struct dw8250_data {
 	unsigned int		skip_autocfg:1;
 	unsigned int		uart_16550_compatible:1;
 	unsigned int		dma_capable:1;
+	unsigned int		is_rzn1:1;
 };
 
 void dw8250_do_set_termios(struct uart_port *p, struct ktermios *termios, struct ktermios *old);
Signature · 1 line
2.27.0
#15 lore raw

Re: [PATCH 6/7] serial: 8250_dw: Add support for RZ/N1 DMA

Quoted 8 lines · Miquel Raynal

From: Phil Edworthy <[email protected]>

The Renesas RZ/N1 devices have a modified Synopsys DW UART. The modifications are mostly related to the DMA handlnig, and so this patch adds support for DMA.

The RZ/N1 UART must be used with the peripheral as the flow controller.

(1)

Quoted 2 lines

This means the DMA length should also be programmed into UART registers.

(2)

Hmm... DMA controller vs. Peripheral flow control is about signalling on the HW level on who starts the transaction. This is programmed in the DMA controller device driver. Is it what you do in DesignWare DMA patch series?

Ah, I see now, you set fc here.

But still it's not clear how (2) and (1) are related.

Quoted 11 lines

Aside from this, there are some points to note about DMA burst sizes. First, DMA must not remove all of the data from the rx FIFO. Otherwise, we do not get a 'character timeout' interrupt, and so do not know that we should push data up the serial stack. Therefore, we have the rx threshold for generating an interrupt set to half the FIFO depth (this is the default for 16550A), and set the DMA burst size when reading the FIFO to a quarter of the FIFO depth.

Second, when transmitting data using DMA, the burst size must be limited to 1 byte to handle then case when transmitting just 1 byte. Otherwise the DMA doesn't complete the burst, and nothing happens.

...

drivers/tty/serial/8250/8250_dw.c line 42

Quoted 3 lines
drivers/tty/serial/8250/8250_dw.c +3 -0
+/* Offsets for the Renesas RZ/N1 DesignWare specific registers */
+/* DMA Software Ack */
+#define RZN1_UART_DMASA			0xa8

Is it specific to Renesas? IIRC it's Synopsys DesignWare register, makes sense to use appropriate prefix or no prefix.

...

drivers/tty/serial/8250/8250_dw.c line 51

Quoted 3 lines
drivers/tty/serial/8250/8250_dw.c +3 -0
+#define RZN1_UART_xDMACR_1_WORD_BURST	0
+#define RZN1_UART_xDMACR_4_WORD_BURST	BIT(1)
+#define RZN1_UART_xDMACR_8_WORD_BURST	(BIT(1) | BIT(2))

This looks like incorrect use of BIT() macro. Please, use plain decimal integers. Something like

	1	(0 << 1)
	4	(1 << 1)
	8	(3 << 1)

If I'm mistaken, describe the meaning of each bit there.

...

drivers/tty/serial/8250/8250_dw.c static unsigned int dw8250_serial_in32be(struct uart_port *p, int offset) · line 240

Quoted 7 lines
drivers/tty/serial/8250/8250_dw.c +6 -0
+static void rzn1_8250_handle_irq(struct uart_port *port, unsigned int iir)
+{
+	struct uart_8250_port *up = up_to_u8250p(port);
+	struct uart_8250_dma *dma = up->dma;
+	unsigned char status;
 
+	if (up->dma && dma->rx_running) {

With

	if (!)up->dma && dma->rx_running))
		return;

maybe easier to read the rest.

Quoted 8 lines
drivers/tty/serial/8250/8250_dw.c +8 -0
+		status = port->serial_in(port, UART_LSR);
+		if (status & (UART_LSR_DR | UART_LSR_BI)) {
+			/* Stop the DMA transfer */
+			writel(0, port->membase + RZN1_UART_RDMACR);
+			writel(1, port->membase + RZN1_UART_DMASA);
+		}
+	}
+}

...

drivers/tty/serial/8250/8250_dw.c static int dw8250_handle_irq(struct uart_port *p) · line 279

Quoted 2 lines
drivers/tty/serial/8250/8250_dw.c +2 -0
+	if (d->is_rzn1 && ((iir & 0x3f) == UART_IIR_RX_TIMEOUT))
+		rzn1_8250_handle_irq(p, iir);

A few years ago it was a discussion about broken timeout on some platforms with Synopsys DesignWare UART + DMA. Can it be that this is actually required for all of them that uses same combination of IPs?

...

drivers/tty/serial/8250/8250_dw.c static bool dw8250_idma_filter(struct dma_chan *chan, void *param) · line 406

Quoted 4 lines
drivers/tty/serial/8250/8250_dw.c +3 -0
+static u32 rzn1_get_dmacr_burst(int max_burst)
+{
 
+	u32 val = 0;

Redundant assignment and variable itself. Use return statements directly.

Quoted 9 lines
drivers/tty/serial/8250/8250_dw.c +9 -0
+	if (max_burst >= 8)
+		val = RZN1_UART_xDMACR_8_WORD_BURST;
+	else if (max_burst >= 4)
+		val = RZN1_UART_xDMACR_4_WORD_BURST;
+	else
+		val = RZN1_UART_xDMACR_1_WORD_BURST;
+
+	return val;
+}

...

Quoted 9 lines
drivers/tty/serial/8250/8250_dw.c +9 -0
+static int rzn1_dw8250_tx_dma(struct uart_8250_port *p)
+{
+	struct uart_port		*up = &p->port;
+	struct uart_8250_dma		*dma = p->dma;
+	struct circ_buf			*xmit = &p->port.state->xmit;
+	int tx_size;
+	u32 val;
+
+	if (uart_tx_stopped(&p->port) || dma->tx_running ||

drivers/tty/serial/8250/8250_dw.c static bool dw8250_idma_filter(struct dma_chan *chan, void *param) · line 436

Quoted 10 lines
drivers/tty/serial/8250/8250_dw.c +9 -0
+	    uart_circ_empty(xmit))
+		return 0;
+
+	tx_size = CIRC_CNT_TO_END(xmit->head, xmit->tail, UART_XMIT_SIZE);
 
+	writel(0, up->membase + RZN1_UART_TDMACR);
+	val = rzn1_get_dmacr_burst(dma->txconf.dst_maxburst);
+	val |= tx_size << RZN1_UART_xDMACR_BLK_SZ_OFFSET;
+	val |= RZN1_UART_xDMACR_DMA_EN;
+	writel(val, up->membase + RZN1_UART_TDMACR);

Can this be added as a callback to the serial8250_tx_dma()? Ditto for Rx counterpart.

Quoted 2 lines
(quoted diff) +2 -0
+	return serial8250_tx_dma(p);
+}

...

drivers/tty/serial/8250/8250_dw.c static int dw8250_probe(struct platform_device *pdev) · line 592

Quoted 1 line
drivers/tty/serial/8250/8250_dw.c +1 -0
+	data->is_rzn1 = of_device_is_compatible(dev->of_node, "renesas,rzn1-uart");

Device property API.

Quoted 2 lines
drivers/tty/serial/8250/8250_dw.c +0 -0
 	/* Always ask for fixed clock rate from a property. */
 	device_property_read_u32(dev, "clock-frequency", &p->uartclk);
Signature · 2 lines
With Best Regards,
Andy Shevchenko
#16 lore raw

Re: [PATCH 6/7] serial: 8250_dw: Add support for RZ/N1 DMA

Hi Andy,

[email protected] wrote on Thu, 10 Mar 2022 20:25:52 +0200:

Quoted 17 lines · Miquel Raynal
Quoted 8 lines

From: Phil Edworthy <[email protected]>

The Renesas RZ/N1 devices have a modified Synopsys DW UART. The modifications are mostly related to the DMA handlnig, and so this patch adds support for DMA.

The RZ/N1 UART must be used with the peripheral as the flow controller.

(1)

Quoted 2 lines

This means the DMA length should also be programmed into UART registers.

(2)

Quoted 7 lines

Hmm... DMA controller vs. Peripheral flow control is about signalling on the HW level on who starts the transaction. This is programmed in the DMA controller device driver. Is it what you do in DesignWare DMA patch series?

Ah, I see now, you set fc here.

But still it's not clear how (2) and (1) are related.

Both come from the system manual:
(1) table 11.45 "Flow Control Combinations" states that using UART with DMA requires setting the DMA in the peripheral flow controller mode regardless of the direction.
(2) chapter 11.6.1.3 "Basic Interface Definitions" explains that the burst size in the above case must be configured in the peripheral's register DEST/SRC_BURST_SIZE.

Quoted 12 lines
Quoted 11 lines

Aside from this, there are some points to note about DMA burst sizes. First, DMA must not remove all of the data from the rx FIFO. Otherwise, we do not get a 'character timeout' interrupt, and so do not know that we should push data up the serial stack. Therefore, we have the rx threshold for generating an interrupt set to half the FIFO depth (this is the default for 16550A), and set the DMA burst size when reading the FIFO to a quarter of the FIFO depth.

Second, when transmitting data using DMA, the burst size must be limited to 1 byte to handle then case when transmitting just 1 byte. Otherwise the DMA doesn't complete the burst, and nothing happens.

Quoted 8 lines

...

drivers/tty/serial/8250/8250_dw.c line 42

Quoted 3 lines
drivers/tty/serial/8250/8250_dw.c +3 -0
+/* Offsets for the Renesas RZ/N1 DesignWare specific registers */
+/* DMA Software Ack */
+#define RZN1_UART_DMASA			0xa8  

Is it specific to Renesas? IIRC it's Synopsys DesignWare register, makes sense to use appropriate prefix or no prefix.

I have no idea, I can use a more common prefix.

Quoted 10 lines

...

drivers/tty/serial/8250/8250_dw.c line 51

Quoted 3 lines
drivers/tty/serial/8250/8250_dw.c +3 -0
+#define RZN1_UART_xDMACR_1_WORD_BURST	0
+#define RZN1_UART_xDMACR_4_WORD_BURST	BIT(1)
+#define RZN1_UART_xDMACR_8_WORD_BURST	(BIT(1) | BIT(2))  

This looks like incorrect use of BIT() macro. Please, use plain decimal integers. Something like

Quoted 5 lines
	1	(0 << 1)
	4	(1 << 1)
	8	(3 << 1)

If I'm mistaken, describe the meaning of each bit there.

Matter of taste, I believe, whatever.

Quoted 29 lines

...

drivers/tty/serial/8250/8250_dw.c static unsigned int dw8250_serial_in32be(struct uart_port *p, int offset) · line 240

Quoted 7 lines
drivers/tty/serial/8250/8250_dw.c +6 -0
+static void rzn1_8250_handle_irq(struct uart_port *port, unsigned int iir)
+{
+	struct uart_8250_port *up = up_to_u8250p(port);
+	struct uart_8250_dma *dma = up->dma;
+	unsigned char status;  
 
+	if (up->dma && dma->rx_running) {  

With

	if (!)up->dma && dma->rx_running))
		return;

maybe easier to read the rest.

Quoted 8 lines
drivers/tty/serial/8250/8250_dw.c +8 -0
+		status = port->serial_in(port, UART_LSR);
+		if (status & (UART_LSR_DR | UART_LSR_BI)) {
+			/* Stop the DMA transfer */
+			writel(0, port->membase + RZN1_UART_RDMACR);
+			writel(1, port->membase + RZN1_UART_DMASA);
+		}
+	}
+}  

...

Quoted 6 lines

drivers/tty/serial/8250/8250_dw.c static int dw8250_handle_irq(struct uart_port *p) · line 279

Quoted 2 lines
drivers/tty/serial/8250/8250_dw.c +2 -0
+	if (d->is_rzn1 && ((iir & 0x3f) == UART_IIR_RX_TIMEOUT))
+		rzn1_8250_handle_irq(p, iir);  

A few years ago it was a discussion about broken timeout on some platforms with Synopsys DesignWare UART + DMA. Can it be that this is actually required for all of them that uses same combination of IPs?

I am not sure because I went through the fix for this issue and for me there are two different things

Quoted 36 lines

...

drivers/tty/serial/8250/8250_dw.c static bool dw8250_idma_filter(struct dma_chan *chan, void *param) · line 406

Quoted 4 lines
drivers/tty/serial/8250/8250_dw.c +3 -0
+static u32 rzn1_get_dmacr_burst(int max_burst)
+{  
 
+	u32 val = 0;  

Redundant assignment and variable itself. Use return statements directly.

Quoted 9 lines
drivers/tty/serial/8250/8250_dw.c +9 -0
+	if (max_burst >= 8)
+		val = RZN1_UART_xDMACR_8_WORD_BURST;
+	else if (max_burst >= 4)
+		val = RZN1_UART_xDMACR_4_WORD_BURST;
+	else
+		val = RZN1_UART_xDMACR_1_WORD_BURST;
+
+	return val;
+}  

...

Quoted 13 lines
drivers/tty/serial/8250/8250_dw.c +13 -0
+static int rzn1_dw8250_tx_dma(struct uart_8250_port *p)
+{
+	struct uart_port		*up = &p->port;
+	struct uart_8250_dma		*dma = p->dma;
+	struct circ_buf			*xmit = &p->port.state->xmit;
+	int tx_size;
+	u32 val;
+
+	if (uart_tx_stopped(&p->port) || dma->tx_running ||
+	    uart_circ_empty(xmit))
+		return 0;
+
+	tx_size = CIRC_CNT_TO_END(xmit->head, xmit->tail, UART_XMIT_SIZE);  
Quoted 8 lines

drivers/tty/serial/8250/8250_dw.c static bool dw8250_idma_filter(struct dma_chan *chan, void *param) · line 436

Quoted 5 lines
drivers/tty/serial/8250/8250_dw.c +5 -0
+	writel(0, up->membase + RZN1_UART_TDMACR);
+	val = rzn1_get_dmacr_burst(dma->txconf.dst_maxburst);
+	val |= tx_size << RZN1_UART_xDMACR_BLK_SZ_OFFSET;
+	val |= RZN1_UART_xDMACR_DMA_EN;
+	writel(val, up->membase + RZN1_UART_TDMACR);  

Can this be added as a callback to the serial8250_tx_dma()? Ditto for Rx counterpart.

Fair enough.

Quoted 9 lines
Quoted 3 lines
(quoted diff) +2 -0
+	return serial8250_tx_dma(p);
+}  

...

drivers/tty/serial/8250/8250_dw.c static int dw8250_probe(struct platform_device *pdev) · line 592

Quoted 1 line
drivers/tty/serial/8250/8250_dw.c +1 -0
+	data->is_rzn1 = of_device_is_compatible(dev->of_node, "renesas,rzn1-uart");  

Device property API.

I'm not sure to get what you mean here again. The Information is in the device tree, the compatible string already gives us what we need, why would we need a device property? (or perhaps I misunderstand what "device property API" means)

Quoted 4 lines

drivers/tty/serial/8250/8250_dw.c static int dw8250_probe(struct platform_device *pdev) · line 595

Quoted 3 lines
drivers/tty/serial/8250/8250_dw.c +0 -0
 	/* Always ask for fixed clock rate from a property. */
 	device_property_read_u32(dev, "clock-frequency", &p->uartclk);  

Thanks,
Miquèl

#17 lore raw

Re: [PATCH 6/7] serial: 8250_dw: Add support for RZ/N1 DMA

Quoted 3 lines · Miquel Raynal

[email protected] wrote on Thu, 10 Mar 2022 20:25:52 +0200:

Quoted 1 line · Miquel Raynal

...

Quoted 8 lines
Quoted 6 lines

drivers/tty/serial/8250/8250_dw.c line 42

Quoted 3 lines
drivers/tty/serial/8250/8250_dw.c +3 -0
+/* Offsets for the Renesas RZ/N1 DesignWare specific registers */
+/* DMA Software Ack */
+#define RZN1_UART_DMASA			0xa8  

Is it specific to Renesas? IIRC it's Synopsys DesignWare register, makes sense to use appropriate prefix or no prefix.

I have no idea, I can use a more common prefix.

It's a register described in Synopsys DesignWare specification. It's not related to Renesas IP integration.

...

Quoted 7 lines
Quoted 7 lines

drivers/tty/serial/8250/8250_dw.c line 51

Quoted 3 lines
drivers/tty/serial/8250/8250_dw.c +3 -0
+#define RZN1_UART_xDMACR_1_WORD_BURST	0
+#define RZN1_UART_xDMACR_4_WORD_BURST	BIT(1)
+#define RZN1_UART_xDMACR_8_WORD_BURST	(BIT(1) | BIT(2))  

This looks like incorrect use of BIT() macro. Please, use plain decimal integers. Something like

Quoted 7 lines
Quoted 5 lines
	1	(0 << 1)
	4	(1 << 1)
	8	(3 << 1)

If I'm mistaken, describe the meaning of each bit there.

Matter of taste, I believe, whatever.

Actually no, when one uses BIT() masks it implies that in the datasheet each bit is meaningful. So, please clarify this and update accordingly.

...

Quoted 8 lines
Quoted 3 lines

drivers/tty/serial/8250/8250_dw.c static int dw8250_probe(struct platform_device *pdev) · line 592

Quoted 1 line
drivers/tty/serial/8250/8250_dw.c +1 -0
+	data->is_rzn1 = of_device_is_compatible(dev->of_node, "renesas,rzn1-uart");

Device property API.

I'm not sure to get what you mean here again. The Information is in the device tree, the compatible string already gives us what we need, why would we need a device property? (or perhaps I misunderstand what "device property API" means)

Use non-OF APIs, which usually starts with device_property_ or fwnode_. But Geert already suggested something better.

Quoted 2 lines
Quoted 2 lines
Quoted 2 lines
drivers/tty/serial/8250/8250_dw.c +0 -0
 	/* Always ask for fixed clock rate from a property. */
 	device_property_read_u32(dev, "clock-frequency", &p->uartclk);
Signature · 2 lines
With Best Regards,
Andy Shevchenko
#18 lore raw

Re: [PATCH 6/7] serial: 8250_dw: Add support for RZ/N1 DMA

In reply to Miquel Raynal

Hi Miquel,

Quoted 18 lines · Miquel Raynal

From: Phil Edworthy <[email protected]>

The Renesas RZ/N1 devices have a modified Synopsys DW UART. The modifications are mostly related to the DMA handlnig, and so this patch adds support for DMA.

The RZ/N1 UART must be used with the peripheral as the flow controller. This means the DMA length should also be programmed into UART registers.

Aside from this, there are some points to note about DMA burst sizes. First, DMA must not remove all of the data from the rx FIFO. Otherwise, we do not get a 'character timeout' interrupt, and so do not know that we should push data up the serial stack. Therefore, we have the rx threshold for generating an interrupt set to half the FIFO depth (this is the default for 16550A), and set the DMA burst size when reading the FIFO to a quarter of the FIFO depth.

Quoted 6 lines

Second, when transmitting data using DMA, the burst size must be limited to 1 byte to handle then case when transmitting just 1 byte. Otherwise the DMA doesn't complete the burst, and nothing happens.

Thanks for your patch!

drivers/tty/serial/8250/8250_dma.c static int dw8250_probe(struct platform_device *pdev) · line 592

Quoted 8 lines
drivers/tty/serial/8250/8250_dma.c +1 -0
@@ -501,6 +589,8 @@ static int dw8250_probe(struct platform_device *pdev)
                data->msr_mask_off |= UART_MSR_TERI;
        }
 
+       data->is_rzn1 = of_device_is_compatible(dev->of_node, "renesas,rzn1-uart");

Explicit checks for compatible values are frowned upon if you have a match table.
Please handle this through of_device.data, cfr. the various quirks. Please rename "is_rzn1" to something that describes the feature.

drivers/tty/serial/8250/8250_dma.c static int dw8250_probe(struct platform_device *pdev) · line 595

Quoted 3 lines
drivers/tty/serial/8250/8250_dma.c +1 -0
+
        /* Always ask for fixed clock rate from a property. */
        device_property_read_u32(dev, "clock-frequency", &p->uartclk);

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

#19 lore raw

Re: [PATCH 6/7] serial: 8250_dw: Add support for RZ/N1 DMA

Hi Miquel,

CC esmil

Quoted 38 lines · Geert Uytterhoeven

Hi Miquel,

Quoted 18 lines · Miquel Raynal

From: Phil Edworthy <[email protected]>

The Renesas RZ/N1 devices have a modified Synopsys DW UART. The modifications are mostly related to the DMA handlnig, and so this patch adds support for DMA.

The RZ/N1 UART must be used with the peripheral as the flow controller. This means the DMA length should also be programmed into UART registers.

Aside from this, there are some points to note about DMA burst sizes. First, DMA must not remove all of the data from the rx FIFO. Otherwise, we do not get a 'character timeout' interrupt, and so do not know that we should push data up the serial stack. Therefore, we have the rx threshold for generating an interrupt set to half the FIFO depth (this is the default for 16550A), and set the DMA burst size when reading the FIFO to a quarter of the FIFO depth.

Quoted 6 lines

Second, when transmitting data using DMA, the burst size must be limited to 1 byte to handle then case when transmitting just 1 byte. Otherwise the DMA doesn't complete the burst, and nothing happens.

Thanks for your patch!

Quoted 7 lines
drivers/tty/serial/8250/8250_dma.c +0 -0
@@ -501,6 +589,8 @@ static int dw8250_probe(struct platform_device *pdev)
                data->msr_mask_off |= UART_MSR_TERI;
        }
 
Quoted 5 lines

drivers/tty/serial/8250/8250_dma.c static int dw8250_probe(struct platform_device *pdev) · line 592

Quoted 1 line
drivers/tty/serial/8250/8250_dma.c +1 -0
+       data->is_rzn1 = of_device_is_compatible(dev->of_node, "renesas,rzn1-uart");

Explicit checks for compatible values are frowned upon if you have a match table.
Please handle this through of_device.data, cfr. the various quirks.

Oops, these are not yet upstream, but present in my tree due to including support for StarLight, cfr.
https://github.com/esmil/linux/commits/visionfive/drivers/tty/serial/8250/8250_dw.c

But you do already have:

+ { .compatible = "renesas,rzn1-uart", .data = &rzn1_pdata },

since "[PATCH 4/7] serial: 8250_dw: Provide the RZN1 CPR register value".

Quoted 5 lines

Please rename "is_rzn1" to something that describes the feature.

Quoted 3 lines
drivers/tty/serial/8250/8250_dma.c +1 -0
+
        /* Always ask for fixed clock rate from a property. */
        device_property_read_u32(dev, "clock-frequency", &p->uartclk);

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

#20 lore raw

Re: [PATCH 6/7] serial: 8250_dw: Add support for RZ/N1 DMA

Hi Geert,

[email protected] wrote on Fri, 11 Mar 2022 10:51:53 +0100:

Quoted 18 lines

Hi Miquel,

CC esmil

Quoted 7 lines
Quoted 7 lines
drivers/tty/serial/8250/8250_dma.c +0 -0
@@ -501,6 +589,8 @@ static int dw8250_probe(struct platform_device *pdev)
                data->msr_mask_off |= UART_MSR_TERI;
        }
 
Quoted 5 lines

drivers/tty/serial/8250/8250_dma.c static int dw8250_probe(struct platform_device *pdev) · line 592

Quoted 1 line
drivers/tty/serial/8250/8250_dma.c +1 -0
+       data->is_rzn1 = of_device_is_compatible(dev->of_node, "renesas,rzn1-uart");  

Explicit checks for compatible values are frowned upon if you have a match table.
Please handle this through of_device.data, cfr. the various quirks.

Quoted 3 lines

Oops, these are not yet upstream, but present in my tree due to including support for StarLight, cfr.
https://github.com/esmil/linux/commits/visionfive/drivers/tty/serial/8250/8250_dw.c

Oh thanks for pointing it! Too bad that these quirks were not introduced inside a wider structure, I think it's always a must even if there is only one parameter there. Anyway, I'll introduce a wider specific structure and use it.

Quoted 8 lines

But you do already have:

+ { .compatible = "renesas,rzn1-uart", .data = &rzn1_pdata },

since "[PATCH 4/7] serial: 8250_dw: Provide the RZN1 CPR register value".

Quoted 2 lines

Please rename "is_rzn1" to something that describes the feature.

Quoted 7 lines
Quoted 3 lines

drivers/tty/serial/8250/8250_dma.c static int dw8250_probe(struct platform_device *pdev) · line 595

Quoted 3 lines
drivers/tty/serial/8250/8250_dma.c +1 -0
+
        /* Always ask for fixed clock rate from a property. */
        device_property_read_u32(dev, "clock-frequency", &p->uartclk);  

Gr{oetje,eeting}s,

Geert

Thanks,
Miquèl

#21 lore raw

PATCH 1/1 serial: 8250_dw: Use device tree match data

..rather than multiple calls to of_device_is_compatible().

---

Hi Miquel,

Quoted 18 lines
Quoted 4 lines
Quoted 4 lines

drivers/tty/serial/8250/8250_dma.c =20

Quoted 2 lines
drivers/tty/serial/8250/8250_dma.c =20 renamed from drivers/tty/serial/8250/8250_dma.c +0 -0

=20

drivers/tty/serial/8250/8250_dma.c =20 static int dw8250_probe(struct platform_device *p=

Quoted 1 line
drivers/tty/serial/8250/8250_dma.c =20 renamed from drivers/tty/serial/8250/8250_dma.c +0 -0
@@ -501,6 +589,8 @@ static int dw8250_probe(struct platform_device *p=

dev)

Quoted 4 lines
Quoted 4 lines

drivers/tty/serial/8250/8250_dma.c =20 static int dw8250_probe(struct platform_device *p= · line 592

Quoted 4 lines
drivers/tty/serial/8250/8250_dma.c =20 renamed from drivers/tty/serial/8250/8250_dma.c +1 -0
                data->msr_mask_off |=3D UART_MSR_TERI;
        }
 
+       data->is_rzn1 =3D of_device_is_compatible(dev->of_node, "rene=

sas,rzn1-uart"); =20

Quoted 8 lines
Quoted 4 lines

Explicit checks for compatible values are frowned upon if you have a match table.
Please handle this through of_device.data, cfr. the various quirks. =20

=20
Oops, these are not yet upstream, but present in my tree due to including support for StarLight, cfr.
https://github.com/esmil/linux/commits/visionfive/drivers/tty/serial/8250=

Quoted 6 lines

/8250_dw.c

Oh thanks for pointing it! Too bad that these quirks were not introduced inside a wider structure, I think it's always a must even if there is only one parameter there. Anyway, I'll introduce a wider specific structure and use it.

For reference this is the patch I wrote for the StarFive JH7100 tree. Feel free to use it or do something better as you see fit.

/Emil

 drivers/tty/serial/8250/8250_dw.c | 18 ++++++++++++------
 1 file changed, 12 insertions(+), 6 deletions(-)
drivers/tty/serial/8250/8250_dw.c +12 -6
index 1769808031c5..f564a019a7be 100644
@@ -37,6 +37,11 @@
 /* DesignWare specific register fields */
 #define DW_UART_MCR_SIRE		BIT(6)
 
+/* Quirks */
+#define DW_UART_QUIRK_OCTEON		BIT(0)
+#define DW_UART_QUIRK_ARMADA_38X	BIT(1)
+#define DW_UART_QUIRK_SKIP_SET_RATE	BIT(2)
+
 struct dw8250_data {
 	struct dw8250_port_data	data;
 
@@ -389,6 +394,7 @@ static void dw8250_quirks(struct uart_port *p, struct dw8250_data *data)
 	struct device_node *np = p->dev->of_node;
 
 	if (np) {
+		unsigned long quirks = (unsigned long)of_device_get_match_data(p->dev);
 		int id;
 
 		/* get index of serial line, if found in DT aliases */
@@ -396,7 +402,7 @@ static void dw8250_quirks(struct uart_port *p, struct dw8250_data *data)
 		if (id >= 0)
 			p->line = id;
 #ifdef CONFIG_64BIT
-		if (of_device_is_compatible(np, "cavium,octeon-3860-uart")) {
+		if (quirks & DW_UART_QUIRK_OCTEON) {
 			p->serial_in = dw8250_serial_inq;
 			p->serial_out = dw8250_serial_outq;
 			p->flags = UPF_SKIP_TEST | UPF_SHARE_IRQ | UPF_FIXED_TYPE;
@@ -412,9 +418,9 @@ static void dw8250_quirks(struct uart_port *p, struct dw8250_data *data)
 			p->serial_out = dw8250_serial_out32be;
 		}
 
-		if (of_device_is_compatible(np, "marvell,armada-38x-uart"))
+		if (quirks & DW_UART_QUIRK_ARMADA_38X)
 			p->serial_out = dw8250_serial_out38x;
-		if (of_device_is_compatible(np, "starfive,jh7100-uart"))
+		if (quirks & DW_UART_QUIRK_SKIP_SET_RATE)
 			p->set_termios = dw8250_do_set_termios;
 
 	} else if (acpi_dev_present("APMC0D08", NULL, -1)) {
@@ -695,10 +701,10 @@ static const struct dev_pm_ops dw8250_pm_ops = {
 
 static const struct of_device_id dw8250_of_match[] = {
 	{ .compatible = "snps,dw-apb-uart" },
-	{ .compatible = "cavium,octeon-3860-uart" },
-	{ .compatible = "marvell,armada-38x-uart" },
+	{ .compatible = "cavium,octeon-3860-uart", .data = (void *)DW_UART_QUIRK_OCTEON },
+	{ .compatible = "marvell,armada-38x-uart", .data = (void *)DW_UART_QUIRK_ARMADA_38X },
 	{ .compatible = "renesas,rzn1-uart" },
-	{ .compatible = "starfive,jh7100-uart" },
+	{ .compatible = "starfive,jh7100-uart",    .data = (void *)DW_UART_QUIRK_SKIP_SET_RATE },
 	{ /* Sentinel */ }
 };
 MODULE_DEVICE_TABLE(of, dw8250_of_match);
Signature · 1 line
2.35.1
#22 lore raw

Re: [PATCH] serial: 8250_dw: Use device tree match data

drivers/tty/serial/8250/8250_dw.c static void dw8250_quirks(struct uart_port *p, struct dw8250_data *data) · line 397

Quoted 7 lines · Emil Renner Berthing

..rather than multiple calls to of_device_is_compatible().

For reference this is the patch I wrote for the StarFive JH7100 tree. Feel free to use it or do something better as you see fit.

drivers/tty/serial/8250/8250_dw.c +1 -0
 	if (np) {
+		unsigned long quirks = (unsigned long)of_device_get_match_data(p->dev);

It can be done outside of the np check with device property APIs in use. Also it needs to use (uintptr_t) for better coverage.

		unsigned long quirks = (uintptr_t)device_get_match_data(p->dev);

Or use data structure as driver_data.

		const struct ... *data = device_get_match_data(p->dev);
Signature · 2 lines
With Best Regards,
Andy Shevchenko
#23 lore raw

Re: [PATCH] serial: 8250_dw: Use device tree match data

In reply to Emil Renner Berthing

Hi Emil,

[email protected] wrote on Fri, 11 Mar 2022 15:48:14 +0100:

Quoted 27 lines

..rather than multiple calls to of_device_is_compatible().

---

Hi Miquel,

Quoted 20 lines
Quoted 4 lines
Quoted 4 lines

drivers/tty/serial/8250/8250_dma.c =20

Quoted 2 lines
drivers/tty/serial/8250/8250_dma.c =20 renamed from drivers/tty/serial/8250/8250_dma.c +0 -0

=20

drivers/tty/serial/8250/8250_dma.c =20 static int dw8250_probe(struct platform_device *p=

Quoted 1 line
drivers/tty/serial/8250/8250_dma.c =20 renamed from drivers/tty/serial/8250/8250_dma.c +0 -0
@@ -501,6 +589,8 @@ static int dw8250_probe(struct platform_device *p=  

dev)

Quoted 4 lines
Quoted 4 lines

drivers/tty/serial/8250/8250_dma.c =20 static int dw8250_probe(struct platform_device *p= · line 592

Quoted 4 lines
drivers/tty/serial/8250/8250_dma.c =20 renamed from drivers/tty/serial/8250/8250_dma.c +1 -0
                data->msr_mask_off |=3D UART_MSR_TERI;
        }
 
+       data->is_rzn1 =3D of_device_is_compatible(dev->of_node, "rene=  

sas,rzn1-uart"); =20

Quoted 8 lines
Quoted 4 lines

Explicit checks for compatible values are frowned upon if you have a match table.
Please handle this through of_device.data, cfr. the various quirks. =20

=20
Oops, these are not yet upstream, but present in my tree due to including support for StarLight, cfr.
https://github.com/esmil/linux/commits/visionfive/drivers/tty/serial/8250=

/8250_dw.c

Quoted 7 lines
Quoted 4 lines

Oh thanks for pointing it! Too bad that these quirks were not introduced inside a wider structure, I think it's always a must even if there is only one parameter there. Anyway, I'll introduce a wider specific structure and use it.

For reference this is the patch I wrote for the StarFive JH7100 tree. Feel free to use it or do something better as you see fit.

Thanks for the pointers, I've fetched the three 8250_dw patches from your tree directly, and will build on top of them!

Cheers,
Miquèl

Quoted 55 lines

/Emil

 drivers/tty/serial/8250/8250_dw.c | 18 ++++++++++++------
 1 file changed, 12 insertions(+), 6 deletions(-)
drivers/tty/serial/8250/8250_dw.c +9 -3
index 1769808031c5..f564a019a7be 100644
@@ -37,6 +37,11 @@
 /* DesignWare specific register fields */
 #define DW_UART_MCR_SIRE		BIT(6)
 
+/* Quirks */
+#define DW_UART_QUIRK_OCTEON		BIT(0)
+#define DW_UART_QUIRK_ARMADA_38X	BIT(1)
+#define DW_UART_QUIRK_SKIP_SET_RATE	BIT(2)
+
 struct dw8250_data {
 	struct dw8250_port_data	data;
 
@@ -389,6 +394,7 @@ static void dw8250_quirks(struct uart_port *p, struct dw8250_data *data)
 	struct device_node *np = p->dev->of_node;
 
 	if (np) {
+		unsigned long quirks = (unsigned long)of_device_get_match_data(p->dev);
 		int id;
 
 		/* get index of serial line, if found in DT aliases */
@@ -396,7 +402,7 @@ static void dw8250_quirks(struct uart_port *p, struct dw8250_data *data)
 		if (id >= 0)
 			p->line = id;
 #ifdef CONFIG_64BIT
-		if (of_device_is_compatible(np, "cavium,octeon-3860-uart")) {
+		if (quirks & DW_UART_QUIRK_OCTEON) {
 			p->serial_in = dw8250_serial_inq;
 			p->serial_out = dw8250_serial_outq;
 			p->flags = UPF_SKIP_TEST | UPF_SHARE_IRQ | UPF_FIXED_TYPE;
@@ -412,9 +418,9 @@ static void dw8250_quirks(struct uart_port *p, struct dw8250_data *data)
 			p->serial_out = dw8250_serial_out32be;
 		}
 
-		if (of_device_is_compatible(np, "marvell,armada-38x-uart"))
+		if (quirks & DW_UART_QUIRK_ARMADA_38X)
 			p->serial_out = dw8250_serial_out38x;
-		if (of_device_is_compatible(np, "starfive,jh7100-uart"))
+		if (quirks & DW_UART_QUIRK_SKIP_SET_RATE)
 			p->set_termios = dw8250_do_set_termios;
 
 	} else if (acpi_dev_present("APMC0D08", NULL, -1)) {
@@ -695,10 +701,10 @@ static const struct dev_pm_ops dw8250_pm_ops = {
 
 static const struct of_device_id dw8250_of_match[] = {
 	{ .compatible = "snps,dw-apb-uart" },

drivers/tty/serial/8250/8250_dw.c static const struct dev_pm_ops dw8250_pm_ops = { · line 710

Quoted 10 lines
drivers/tty/serial/8250/8250_dw.c +3 -3
-	{ .compatible = "cavium,octeon-3860-uart" },
-	{ .compatible = "marvell,armada-38x-uart" },
+	{ .compatible = "cavium,octeon-3860-uart", .data = (void *)DW_UART_QUIRK_OCTEON },
+	{ .compatible = "marvell,armada-38x-uart", .data = (void *)DW_UART_QUIRK_ARMADA_38X },
 	{ .compatible = "renesas,rzn1-uart" },
-	{ .compatible = "starfive,jh7100-uart" },
+	{ .compatible = "starfive,jh7100-uart",    .data = (void *)DW_UART_QUIRK_SKIP_SET_RATE },
 	{ /* Sentinel */ }
 };
 MODULE_DEVICE_TABLE(of, dw8250_of_match);

Thanks,
Miquèl

#24 lore raw

PATCH 7/7 ARM: dts: r9a06g032: Fill the UART DMA properties

In reply to Miquel Raynal

UART 0 to 2 do not have DMA support, while UART 3 to 7 do.

Fill the "dmas" and "dma-names" properties for each of these nodes.

Please mind that these nodes go through the dmamux node which will redirect the requests to the right DMA controller. The first 4 cells of the "dmas" properties will be transferred as-is to the DMA controllers. The last 2 cells are consumed by the dmamux. Which means cell 0 and 4 are almost redundant, one giving the controller request ID and the other the dmamux channel which is a 1:1 translation of the request IDs, shifted by 16 when pointing to the second DMA controller.

---
 arch/arm/boot/dts/r9a06g032.dtsi | 15 +++++++++++++++
 1 file changed, 15 insertions(+)
arch/arm/boot/dts/r9a06g032.dtsi +15 -0
index 804f2d6f416f..aa447e2622e0 100644
@@ -128,6 +128,9 @@ uart3: serial@50000000 {
 			reg-io-width = <4>;
 			clocks = <&sysctrl R9A06G032_CLK_UART3>, <&sysctrl R9A06G032_HCLK_UART3>;
 			clock-names = "baudclk", "apb_pclk";
+			dmas =  <&dmamux 0 0 0 0 0 1>,
+				<&dmamux 1 0 0 0 1 1>;
+			dma-names = "rx", "tx";
 			status = "disabled";
 		};
 
@@ -139,6 +142,9 @@ uart4: serial@50001000 {
 			reg-io-width = <4>;
 			clocks = <&sysctrl R9A06G032_CLK_UART4>, <&sysctrl R9A06G032_HCLK_UART4>;
 			clock-names = "baudclk", "apb_pclk";
+			dmas =  <&dmamux 2 0 0 0 2 1>,
+				<&dmamux 3 0 0 0 3 1>;
+			dma-names = "rx", "tx";
 			status = "disabled";
 		};
 
@@ -150,6 +156,9 @@ uart5: serial@50002000 {
 			reg-io-width = <4>;
 			clocks = <&sysctrl R9A06G032_CLK_UART5>, <&sysctrl R9A06G032_HCLK_UART5>;
 			clock-names = "baudclk", "apb_pclk";
+			dmas =  <&dmamux 4 0 0 0 4 1>,
+				<&dmamux 5 0 0 0 5 1>;
+			dma-names = "rx", "tx";
 			status = "disabled";
 		};
 
@@ -161,6 +170,9 @@ uart6: serial@50003000 {
 			reg-io-width = <4>;
 			clocks = <&sysctrl R9A06G032_CLK_UART6>, <&sysctrl R9A06G032_HCLK_UART6>;
 			clock-names = "baudclk", "apb_pclk";
+			dmas =  <&dmamux 6 0 0 0 6 1>,
+				<&dmamux 7 0 0 0 7 1>;
+			dma-names = "rx", "tx";
 			status = "disabled";
 		};
 
@@ -172,6 +184,9 @@ uart7: serial@50004000 {
 			reg-io-width = <4>;
 			clocks = <&sysctrl R9A06G032_CLK_UART7>, <&sysctrl R9A06G032_HCLK_UART7>;
 			clock-names = "baudclk", "apb_pclk";
+			dmas =  <&dmamux 4 0 0 0 20 1>,
+				<&dmamux 5 0 0 0 21 1>;
+			dma-names = "rx", "tx";
 			status = "disabled";
 		};
 
Signature · 1 line
2.27.0