Skip to contentSkip to conversation

tty/serial: at91: add support to FIFOs

PATCH Superseded Cyrille Pitchen 3 participants 12 messages Jun 11, 2015 – Jun 29, 2015

Subsystems (5)

v2 of 4 Superseded Latest is v4 →

lore
Share & tools
lore
#0 lore raw

v2 cover tty/serial: at91: add support to FIFOs

ChangeLog

v2:
- remove "atmel,rts-high-threshold" and "atmel,rts-low-threshold" from new DT properties. For now these two thresholds are set once for all during the probe but a later patch might allow to configure them at run time.
- reword the commit message of the DT property patch to better explain why we have chosen to introduce the new property "atmel,fifo-size".
- add a missing chunk in the FIFO patch: as the commit message explains, we need to use 8bit accesses when dealing with the Transmit/Receive Holding Registers.

v1:
This series of patches add support to FIFOs which will be introduced with Atmel sama5d2x SoC.

FIFOs allow to reduce the number of I/O access. Indeed depending on the data size and the USART mode, FIFOs work in either single or multiple data. For multiple data, up to 4 data can be written into the Transmit Holding Register or read from the Receive Holding Register in a single word access.

Also the RX FIFO allows to reduce the risk of receive overrun and data loss, especially when the DMA controller is not used.

Finally, when the hardware handshake mode is selected, the RTS line can now be controlled by two thresholds on the RX FIFO. When FIFO level (the number of data available to be read) crosses the high threshold, the RTS line is set to high level telling the peer that it should stop sending new data. Data are read from the RX FIFO and when the FIFO level crosses the low threshold, the RTS is set back to low level telling the remote peer that it can send data again.

This new feature resolves a long time hardware limitation where the RTS line was directly controlled by a PDC signal. There is no equivalent of that signal for the DMA controller so for all SoCs without PDC for USART there was no mean to control the RTS line properly: once the hardware handshake mode selected, the RTS line remains high level. Next SoCs with FIFOs will be able to use the hardware handshake mode thanks to the RX FIFO.

Cyrille Pitchen (4):
ARM: at91/dt: add a new DT property to support FIFOs on Atmel USARTs tty/serial: at91: fix some macro definitions to fit coding style tty/serial: at91: add support to FIFOs
tty/serial: at91: use 32bit writes into TX FIFO when DMA is enabled

 .../devicetree/bindings/serial/atmel-usart.txt     |   3 +
 drivers/tty/serial/atmel_serial.c                  | 193 +++++++++++++----
 include/linux/atmel_serial.h                       | 240 ++++++++++++---------
 3 files changed, 293 insertions(+), 143 deletions(-)
Signature · 1 line
1.8.2.2
#1 lore raw

PATCH v2 1/4 ARM: at91/dt: add a new DT property to support FIFOs on Atmel USARTs

This patch adds a new DT property, "atmel,fifo-size", to enable and set the maximum number of data the RX and TX FIFOs can store on FIFO capable USARTs.

Please be aware that the VERSION register can not be used to guess the size of FIFOs. Indeed, for a given hardware version, the USARTs can be integrated on Atmel SoCs with different FIFO sizes. Also the "atmel,fifo-size" property is optional as older USARTs don't embed FIFO at all.

Besides, the FIFO size can not be read or guessed from other registers:
When designing the FIFO feature, no dedicated registers were added to store this size. Unsed spaces in the I/O register range are limited and better reserved for future usages. Instead, the FIFO size of each peripheral is documented in the programmer datasheet.

Finally, on a given SoC, there can be several instances of USART with different FIFO sizes. This explain why we'd rather use a dedicated DT property than use the "compatible" property.

---
 Documentation/devicetree/bindings/serial/atmel-usart.txt | 3 +++
 1 file changed, 3 insertions(+)
Documentation/devicetree/bindings/serial/atmel-usart.txt +3 -0
index 90787aa..e6e6142 100644
@@ -22,6 +22,8 @@ Optional properties:
 		memory peripheral interface and USART DMA channel ID, FIFO configuration.
 		Refer to dma.txt and atmel-dma.txt for details.
 	- dma-names: "rx" for RX channel, "tx" for TX channel.
+- atmel,fifo-size: maximum number of data the RX and TX FIFOs can store for FIFO
+  capable USARTs.
 
 <chip> compatible description:
 - at91rm9200:  legacy USART support
@@ -57,4 +59,5 @@ Example:
 		dmas = <&dma0 2 0x3>,
 		       <&dma0 2 0x204>;
 		dma-names = "tx", "rx";
+		atmel,fifo-size = <32>;
 	};
Signature · 1 line
1.8.2.2
Acked-by on v2 1/4
#2 lore raw

Re: [PATCH linux-next v2 1/4] ARM: at91/dt: add a new DT property to support FIFOs on Atmel USARTs

Quoted 16 lines · Cyrille Pitchen

This patch adds a new DT property, "atmel,fifo-size", to enable and set the maximum number of data the RX and TX FIFOs can store on FIFO capable USARTs.

Please be aware that the VERSION register can not be used to guess the size of FIFOs. Indeed, for a given hardware version, the USARTs can be integrated on Atmel SoCs with different FIFO sizes. Also the "atmel,fifo-size" property is optional as older USARTs don't embed FIFO at all.

Besides, the FIFO size can not be read or guessed from other registers:
When designing the FIFO feature, no dedicated registers were added to store this size. Unsed spaces in the I/O register range are limited and better reserved for future usages. Instead, the FIFO size of each peripheral is documented in the programmer datasheet.

Quoted 5 lines

Finally, on a given SoC, there can be several instances of USART with different FIFO sizes. This explain why we'd rather use a dedicated DT property than use the "compatible" property.

Signature · 3 lines
Alexandre Belloni, Free Electrons
Embedded Linux, Kernel and Android engineering
http://free-electrons.com
Acked-by on v2 1/4
#3 lore raw

Re: [PATCH linux-next v2 1/4] ARM: at91/dt: add a new DT property to support FIFOs on Atmel USARTs

In reply to Cyrille Pitchen

Le 11/06/2015 18:20, Cyrille Pitchen a ?crit :

Quoted 16 lines

This patch adds a new DT property, "atmel,fifo-size", to enable and set the maximum number of data the RX and TX FIFOs can store on FIFO capable USARTs.

Please be aware that the VERSION register can not be used to guess the size of FIFOs. Indeed, for a given hardware version, the USARTs can be integrated on Atmel SoCs with different FIFO sizes. Also the "atmel,fifo-size" property is optional as older USARTs don't embed FIFO at all.

Besides, the FIFO size can not be read or guessed from other registers:
When designing the FIFO feature, no dedicated registers were added to store this size. Unsed spaces in the I/O register range are limited and better reserved for future usages. Instead, the FIFO size of each peripheral is documented in the programmer datasheet.

Quoted 5 lines

Finally, on a given SoC, there can be several instances of USART with different FIFO sizes. This explain why we'd rather use a dedicated DT property than use the "compatible" property.

Quoted 24 lines
---
 Documentation/devicetree/bindings/serial/atmel-usart.txt | 3 +++
 1 file changed, 3 insertions(+)
Documentation/devicetree/bindings/serial/atmel-usart.txt +3 -0
index 90787aa..e6e6142 100644
@@ -22,6 +22,8 @@ Optional properties:
 		memory peripheral interface and USART DMA channel ID, FIFO configuration.
 		Refer to dma.txt and atmel-dma.txt for details.
 	- dma-names: "rx" for RX channel, "tx" for TX channel.
+- atmel,fifo-size: maximum number of data the RX and TX FIFOs can store for FIFO
+  capable USARTs.
 
 <chip> compatible description:
 - at91rm9200:  legacy USART support
@@ -57,4 +59,5 @@ Example:
 		dmas = <&dma0 2 0x3>,
 		       <&dma0 2 0x204>;
 		dma-names = "tx", "rx";
+		atmel,fifo-size = <32>;
 	};
Signature · 1 line
Nicolas Ferre
#4 lore raw

PATCH v2 2/4 tty/serial: at91: fix some macro definitions to fit coding style

In reply to Cyrille Pitchen

This patch updates macro definitions in atmel_serial.h to fit the 80 column rule.

Please note that some deprecated comments such as "[AT91SAM9261 only]" are removed as the corresponding bits also exist in some later chips.

The patch also fix macro definitions in atmel_serial.c to replace (port,v) by (port, v).

---
 drivers/tty/serial/atmel_serial.c |  52 +++++-----
 include/linux/atmel_serial.h      | 204 +++++++++++++++++++-------------------
 2 files changed, 128 insertions(+), 128 deletions(-)
drivers/tty/serial/atmel_serial.c +26 -26
index 2a8f528..112e74b 100644
@@ -89,35 +89,35 @@ static void atmel_stop_rx(struct uart_port *port);
 #define ATMEL_ISR_PASS_LIMIT	256
 
 /* UART registers. CR is write-only, hence no GET macro */
-#define UART_PUT_CR(port,v)	__raw_writel(v, (port)->membase + ATMEL_US_CR)
-#define UART_GET_MR(port)	__raw_readl((port)->membase + ATMEL_US_MR)
-#define UART_PUT_MR(port,v)	__raw_writel(v, (port)->membase + ATMEL_US_MR)
-#define UART_PUT_IER(port,v)	__raw_writel(v, (port)->membase + ATMEL_US_IER)
-#define UART_PUT_IDR(port,v)	__raw_writel(v, (port)->membase + ATMEL_US_IDR)
-#define UART_GET_IMR(port)	__raw_readl((port)->membase + ATMEL_US_IMR)
-#define UART_GET_CSR(port)	__raw_readl((port)->membase + ATMEL_US_CSR)
-#define UART_GET_CHAR(port)	__raw_readl((port)->membase + ATMEL_US_RHR)
-#define UART_PUT_CHAR(port,v)	__raw_writel(v, (port)->membase + ATMEL_US_THR)
-#define UART_GET_BRGR(port)	__raw_readl((port)->membase + ATMEL_US_BRGR)
-#define UART_PUT_BRGR(port,v)	__raw_writel(v, (port)->membase + ATMEL_US_BRGR)
-#define UART_PUT_RTOR(port,v)	__raw_writel(v, (port)->membase + ATMEL_US_RTOR)
-#define UART_PUT_TTGR(port, v)	__raw_writel(v, (port)->membase + ATMEL_US_TTGR)
-#define UART_GET_IP_NAME(port)	__raw_readl((port)->membase + ATMEL_US_NAME)
-#define UART_GET_IP_VERSION(port) __raw_readl((port)->membase + ATMEL_US_VERSION)
+#define UART_PUT_CR(port, v)   __raw_writel(v, (port)->membase + ATMEL_US_CR)
+#define UART_GET_MR(port)      __raw_readl((port)->membase + ATMEL_US_MR)
+#define UART_PUT_MR(port, v)   __raw_writel(v, (port)->membase + ATMEL_US_MR)
+#define UART_PUT_IER(port, v)  __raw_writel(v, (port)->membase + ATMEL_US_IER)
+#define UART_PUT_IDR(port, v)  __raw_writel(v, (port)->membase + ATMEL_US_IDR)
+#define UART_GET_IMR(port)     __raw_readl((port)->membase + ATMEL_US_IMR)
+#define UART_GET_CSR(port)     __raw_readl((port)->membase + ATMEL_US_CSR)
+#define UART_GET_CHAR(port)    __raw_readl((port)->membase + ATMEL_US_RHR)
+#define UART_PUT_CHAR(port, v) __raw_writel(v, (port)->membase + ATMEL_US_THR)
+#define UART_GET_BRGR(port)    __raw_readl((port)->membase + ATMEL_US_BRGR)
+#define UART_PUT_BRGR(port, v) __raw_writel(v, (port)->membase + ATMEL_US_BRGR)
+#define UART_PUT_RTOR(port, v) __raw_writel(v, (port)->membase + ATMEL_US_RTOR)
+#define UART_PUT_TTGR(port, v) __raw_writel(v, (port)->membase + ATMEL_US_TTGR)
+#define UART_GET_IP_NAME(port) __raw_readl((port)->membase + ATMEL_US_NAME)
+#define UART_GET_IP_VERS(port) __raw_readl((port)->membase + ATMEL_US_VERSION)
 
  /* PDC registers */
-#define UART_PUT_PTCR(port,v)	__raw_writel(v, (port)->membase + ATMEL_PDC_PTCR)
-#define UART_GET_PTSR(port)	__raw_readl((port)->membase + ATMEL_PDC_PTSR)
+#define UART_PUT_PTCR(port, v) __raw_writel(v, (port)->membase + ATMEL_PDC_PTCR)
+#define UART_GET_PTSR(port)    __raw_readl((port)->membase + ATMEL_PDC_PTSR)
 
-#define UART_PUT_RPR(port,v)	__raw_writel(v, (port)->membase + ATMEL_PDC_RPR)
-#define UART_GET_RPR(port)	__raw_readl((port)->membase + ATMEL_PDC_RPR)
-#define UART_PUT_RCR(port,v)	__raw_writel(v, (port)->membase + ATMEL_PDC_RCR)
-#define UART_PUT_RNPR(port,v)	__raw_writel(v, (port)->membase + ATMEL_PDC_RNPR)
-#define UART_PUT_RNCR(port,v)	__raw_writel(v, (port)->membase + ATMEL_PDC_RNCR)
+#define UART_PUT_RPR(port, v)  __raw_writel(v, (port)->membase + ATMEL_PDC_RPR)
+#define UART_GET_RPR(port)     __raw_readl((port)->membase + ATMEL_PDC_RPR)
+#define UART_PUT_RCR(port, v)  __raw_writel(v, (port)->membase + ATMEL_PDC_RCR)
+#define UART_PUT_RNPR(port, v) __raw_writel(v, (port)->membase + ATMEL_PDC_RNPR)
+#define UART_PUT_RNCR(port, v) __raw_writel(v, (port)->membase + ATMEL_PDC_RNCR)
 
-#define UART_PUT_TPR(port,v)	__raw_writel(v, (port)->membase + ATMEL_PDC_TPR)
-#define UART_PUT_TCR(port,v)	__raw_writel(v, (port)->membase + ATMEL_PDC_TCR)
-#define UART_GET_TCR(port)	__raw_readl((port)->membase + ATMEL_PDC_TCR)
+#define UART_PUT_TPR(port, v)  __raw_writel(v, (port)->membase + ATMEL_PDC_TPR)
+#define UART_PUT_TCR(port, v)  __raw_writel(v, (port)->membase + ATMEL_PDC_TCR)
+#define UART_GET_TCR(port)     __raw_readl((port)->membase + ATMEL_PDC_TCR)
 
 struct atmel_dma_buffer {
 	unsigned char	*buf;
@@ -1684,7 +1684,7 @@ static void atmel_get_ip_name(struct uart_port *port)
 		atmel_port->is_usart = false;
 	} else {
 		/* fallback for older SoCs: use version field */
-		version = UART_GET_IP_VERSION(port);
+		version = UART_GET_IP_VERS(port);
 		switch (version) {
 		case 0x302:
 		case 0x10213:
include/linux/atmel_serial.h +102 -102

217 lines not on this page · raw message

Signature · 1 line
1.8.2.2
Acked-by on v2 2/4
#5 lore raw

Re: [PATCH linux-next v2 2/4] tty/serial: at91: fix some macro definitions to fit coding style

Quoted 10 lines · Cyrille Pitchen

This patch updates macro definitions in atmel_serial.h to fit the 80 column rule.

Please note that some deprecated comments such as "[AT91SAM9261 only]" are removed as the corresponding bits also exist in some later chips.

The patch also fix macro definitions in atmel_serial.c to replace (port,v) by (port, v).

Signature · 3 lines
Alexandre Belloni, Free Electrons
Embedded Linux, Kernel and Android engineering
http://free-electrons.com
Acked-by on v2 2/4
#6 lore raw

Re: [PATCH linux-next v2 2/4] tty/serial: at91: fix some macro definitions to fit coding style

In reply to Cyrille Pitchen

Le 11/06/2015 18:20, Cyrille Pitchen a ?crit :

Quoted 10 lines

This patch updates macro definitions in atmel_serial.h to fit the 80 column rule.

Please note that some deprecated comments such as "[AT91SAM9261 only]" are removed as the corresponding bits also exist in some later chips.

The patch also fix macro definitions in atmel_serial.c to replace (port,v) by (port, v).

Quoted 302 lines
---
 drivers/tty/serial/atmel_serial.c |  52 +++++-----
 include/linux/atmel_serial.h      | 204 +++++++++++++++++++-------------------
 2 files changed, 128 insertions(+), 128 deletions(-)
drivers/tty/serial/atmel_serial.c +26 -26
index 2a8f528..112e74b 100644
@@ -89,35 +89,35 @@ static void atmel_stop_rx(struct uart_port *port);
 #define ATMEL_ISR_PASS_LIMIT	256
 
 /* UART registers. CR is write-only, hence no GET macro */
-#define UART_PUT_CR(port,v)	__raw_writel(v, (port)->membase + ATMEL_US_CR)
-#define UART_GET_MR(port)	__raw_readl((port)->membase + ATMEL_US_MR)
-#define UART_PUT_MR(port,v)	__raw_writel(v, (port)->membase + ATMEL_US_MR)
-#define UART_PUT_IER(port,v)	__raw_writel(v, (port)->membase + ATMEL_US_IER)
-#define UART_PUT_IDR(port,v)	__raw_writel(v, (port)->membase + ATMEL_US_IDR)
-#define UART_GET_IMR(port)	__raw_readl((port)->membase + ATMEL_US_IMR)
-#define UART_GET_CSR(port)	__raw_readl((port)->membase + ATMEL_US_CSR)
-#define UART_GET_CHAR(port)	__raw_readl((port)->membase + ATMEL_US_RHR)
-#define UART_PUT_CHAR(port,v)	__raw_writel(v, (port)->membase + ATMEL_US_THR)
-#define UART_GET_BRGR(port)	__raw_readl((port)->membase + ATMEL_US_BRGR)
-#define UART_PUT_BRGR(port,v)	__raw_writel(v, (port)->membase + ATMEL_US_BRGR)
-#define UART_PUT_RTOR(port,v)	__raw_writel(v, (port)->membase + ATMEL_US_RTOR)
-#define UART_PUT_TTGR(port, v)	__raw_writel(v, (port)->membase + ATMEL_US_TTGR)
-#define UART_GET_IP_NAME(port)	__raw_readl((port)->membase + ATMEL_US_NAME)
-#define UART_GET_IP_VERSION(port) __raw_readl((port)->membase + ATMEL_US_VERSION)
+#define UART_PUT_CR(port, v)   __raw_writel(v, (port)->membase + ATMEL_US_CR)
+#define UART_GET_MR(port)      __raw_readl((port)->membase + ATMEL_US_MR)
+#define UART_PUT_MR(port, v)   __raw_writel(v, (port)->membase + ATMEL_US_MR)
+#define UART_PUT_IER(port, v)  __raw_writel(v, (port)->membase + ATMEL_US_IER)
+#define UART_PUT_IDR(port, v)  __raw_writel(v, (port)->membase + ATMEL_US_IDR)
+#define UART_GET_IMR(port)     __raw_readl((port)->membase + ATMEL_US_IMR)
+#define UART_GET_CSR(port)     __raw_readl((port)->membase + ATMEL_US_CSR)
+#define UART_GET_CHAR(port)    __raw_readl((port)->membase + ATMEL_US_RHR)
+#define UART_PUT_CHAR(port, v) __raw_writel(v, (port)->membase + ATMEL_US_THR)
+#define UART_GET_BRGR(port)    __raw_readl((port)->membase + ATMEL_US_BRGR)
+#define UART_PUT_BRGR(port, v) __raw_writel(v, (port)->membase + ATMEL_US_BRGR)
+#define UART_PUT_RTOR(port, v) __raw_writel(v, (port)->membase + ATMEL_US_RTOR)
+#define UART_PUT_TTGR(port, v) __raw_writel(v, (port)->membase + ATMEL_US_TTGR)
+#define UART_GET_IP_NAME(port) __raw_readl((port)->membase + ATMEL_US_NAME)
+#define UART_GET_IP_VERS(port) __raw_readl((port)->membase + ATMEL_US_VERSION)
 
  /* PDC registers */
-#define UART_PUT_PTCR(port,v)	__raw_writel(v, (port)->membase + ATMEL_PDC_PTCR)
-#define UART_GET_PTSR(port)	__raw_readl((port)->membase + ATMEL_PDC_PTSR)
+#define UART_PUT_PTCR(port, v) __raw_writel(v, (port)->membase + ATMEL_PDC_PTCR)
+#define UART_GET_PTSR(port)    __raw_readl((port)->membase + ATMEL_PDC_PTSR)
 
-#define UART_PUT_RPR(port,v)	__raw_writel(v, (port)->membase + ATMEL_PDC_RPR)
-#define UART_GET_RPR(port)	__raw_readl((port)->membase + ATMEL_PDC_RPR)
-#define UART_PUT_RCR(port,v)	__raw_writel(v, (port)->membase + ATMEL_PDC_RCR)
-#define UART_PUT_RNPR(port,v)	__raw_writel(v, (port)->membase + ATMEL_PDC_RNPR)
-#define UART_PUT_RNCR(port,v)	__raw_writel(v, (port)->membase + ATMEL_PDC_RNCR)
+#define UART_PUT_RPR(port, v)  __raw_writel(v, (port)->membase + ATMEL_PDC_RPR)
+#define UART_GET_RPR(port)     __raw_readl((port)->membase + ATMEL_PDC_RPR)
+#define UART_PUT_RCR(port, v)  __raw_writel(v, (port)->membase + ATMEL_PDC_RCR)
+#define UART_PUT_RNPR(port, v) __raw_writel(v, (port)->membase + ATMEL_PDC_RNPR)
+#define UART_PUT_RNCR(port, v) __raw_writel(v, (port)->membase + ATMEL_PDC_RNCR)
 
-#define UART_PUT_TPR(port,v)	__raw_writel(v, (port)->membase + ATMEL_PDC_TPR)
-#define UART_PUT_TCR(port,v)	__raw_writel(v, (port)->membase + ATMEL_PDC_TCR)
-#define UART_GET_TCR(port)	__raw_readl((port)->membase + ATMEL_PDC_TCR)
+#define UART_PUT_TPR(port, v)  __raw_writel(v, (port)->membase + ATMEL_PDC_TPR)
+#define UART_PUT_TCR(port, v)  __raw_writel(v, (port)->membase + ATMEL_PDC_TCR)
+#define UART_GET_TCR(port)     __raw_readl((port)->membase + ATMEL_PDC_TCR)
 
 struct atmel_dma_buffer {
 	unsigned char	*buf;
@@ -1684,7 +1684,7 @@ static void atmel_get_ip_name(struct uart_port *port)
 		atmel_port->is_usart = false;
 	} else {
 		/* fallback for older SoCs: use version field */
-		version = UART_GET_IP_VERSION(port);
+		version = UART_GET_IP_VERS(port);
 		switch (version) {
 		case 0x302:
 		case 0x10213:
include/linux/atmel_serial.h +102 -102
index 00beddf..c384c21 100644
@@ -16,115 +16,115 @@
 #ifndef ATMEL_SERIAL_H
 #define ATMEL_SERIAL_H
 
-#define ATMEL_US_CR		0x00			/* Control Register */
-#define		ATMEL_US_RSTRX		(1 <<  2)		/* Reset Receiver */
-#define		ATMEL_US_RSTTX		(1 <<  3)		/* Reset Transmitter */
-#define		ATMEL_US_RXEN		(1 <<  4)		/* Receiver Enable */
-#define		ATMEL_US_RXDIS		(1 <<  5)		/* Receiver Disable */
-#define		ATMEL_US_TXEN		(1 <<  6)		/* Transmitter Enable */
-#define		ATMEL_US_TXDIS		(1 <<  7)		/* Transmitter Disable */
-#define		ATMEL_US_RSTSTA		(1 <<  8)		/* Reset Status Bits */
-#define		ATMEL_US_STTBRK		(1 <<  9)		/* Start Break */
-#define		ATMEL_US_STPBRK		(1 << 10)		/* Stop Break */
-#define		ATMEL_US_STTTO		(1 << 11)		/* Start Time-out */
-#define		ATMEL_US_SENDA		(1 << 12)		/* Send Address */
-#define		ATMEL_US_RSTIT		(1 << 13)		/* Reset Iterations */
-#define		ATMEL_US_RSTNACK	(1 << 14)		/* Reset Non Acknowledge */
-#define		ATMEL_US_RETTO		(1 << 15)		/* Rearm Time-out */
-#define		ATMEL_US_DTREN		(1 << 16)		/* Data Terminal Ready Enable [AT91RM9200 only] */
-#define		ATMEL_US_DTRDIS		(1 << 17)		/* Data Terminal Ready Disable [AT91RM9200 only] */
-#define		ATMEL_US_RTSEN		(1 << 18)		/* Request To Send Enable */
-#define		ATMEL_US_RTSDIS		(1 << 19)		/* Request To Send Disable */
+#define ATMEL_US_CR		0x00	/* Control Register */
+#define	ATMEL_US_RSTRX		BIT(2)	/* Reset Receiver */
+#define	ATMEL_US_RSTTX		BIT(3)	/* Reset Transmitter */
+#define	ATMEL_US_RXEN		BIT(4)	/* Receiver Enable */
+#define	ATMEL_US_RXDIS		BIT(5)	/* Receiver Disable */
+#define	ATMEL_US_TXEN		BIT(6)	/* Transmitter Enable */
+#define	ATMEL_US_TXDIS		BIT(7)	/* Transmitter Disable */
+#define	ATMEL_US_RSTSTA		BIT(8)	/* Reset Status Bits */
+#define	ATMEL_US_STTBRK		BIT(9)	/* Start Break */
+#define	ATMEL_US_STPBRK		BIT(10)	/* Stop Break */
+#define	ATMEL_US_STTTO		BIT(11)	/* Start Time-out */
+#define	ATMEL_US_SENDA		BIT(12)	/* Send Address */
+#define	ATMEL_US_RSTIT		BIT(13)	/* Reset Iterations */
+#define	ATMEL_US_RSTNACK	BIT(14)	/* Reset Non Acknowledge */
+#define	ATMEL_US_RETTO		BIT(15)	/* Rearm Time-out */
+#define	ATMEL_US_DTREN		BIT(16)	/* Data Terminal Ready Enable */
+#define	ATMEL_US_DTRDIS		BIT(17)	/* Data Terminal Ready Disable */
+#define	ATMEL_US_RTSEN		BIT(18)	/* Request To Send Enable */
+#define	ATMEL_US_RTSDIS		BIT(19)	/* Request To Send Disable */
 
-#define ATMEL_US_MR		0x04			/* Mode Register */
-#define		ATMEL_US_USMODE		(0xf <<  0)		/* Mode of the USART */
-#define			ATMEL_US_USMODE_NORMAL		0
-#define			ATMEL_US_USMODE_RS485		1
-#define			ATMEL_US_USMODE_HWHS		2
-#define			ATMEL_US_USMODE_MODEM		3
-#define			ATMEL_US_USMODE_ISO7816_T0	4
-#define			ATMEL_US_USMODE_ISO7816_T1	6
-#define			ATMEL_US_USMODE_IRDA		8
-#define		ATMEL_US_USCLKS		(3   <<  4)		/* Clock Selection */
-#define			ATMEL_US_USCLKS_MCK		(0 <<  4)
-#define			ATMEL_US_USCLKS_MCK_DIV8	(1 <<  4)
-#define			ATMEL_US_USCLKS_SCK		(3 <<  4)
-#define		ATMEL_US_CHRL		(3   <<  6)		/* Character Length */
-#define			ATMEL_US_CHRL_5			(0 <<  6)
-#define			ATMEL_US_CHRL_6			(1 <<  6)
-#define			ATMEL_US_CHRL_7			(2 <<  6)
-#define			ATMEL_US_CHRL_8			(3 <<  6)
-#define		ATMEL_US_SYNC		(1 <<  8)		/* Synchronous Mode Select */
-#define		ATMEL_US_PAR		(7 <<  9)		/* Parity Type */
-#define			ATMEL_US_PAR_EVEN		(0 <<  9)
-#define			ATMEL_US_PAR_ODD		(1 <<  9)
-#define			ATMEL_US_PAR_SPACE		(2 <<  9)
-#define			ATMEL_US_PAR_MARK		(3 <<  9)
-#define			ATMEL_US_PAR_NONE		(4 <<  9)
-#define			ATMEL_US_PAR_MULTI_DROP		(6 <<  9)
-#define		ATMEL_US_NBSTOP		(3 << 12)		/* Number of Stop Bits */
-#define			ATMEL_US_NBSTOP_1		(0 << 12)
-#define			ATMEL_US_NBSTOP_1_5		(1 << 12)
-#define			ATMEL_US_NBSTOP_2		(2 << 12)
-#define		ATMEL_US_CHMODE		(3 << 14)		/* Channel Mode */
-#define			ATMEL_US_CHMODE_NORMAL		(0 << 14)
-#define			ATMEL_US_CHMODE_ECHO		(1 << 14)
-#define			ATMEL_US_CHMODE_LOC_LOOP	(2 << 14)
-#define			ATMEL_US_CHMODE_REM_LOOP	(3 << 14)
-#define		ATMEL_US_MSBF		(1 << 16)		/* Bit Order */
-#define		ATMEL_US_MODE9		(1 << 17)		/* 9-bit Character Length */
-#define		ATMEL_US_CLKO		(1 << 18)		/* Clock Output Select */
-#define		ATMEL_US_OVER		(1 << 19)		/* Oversampling Mode */
-#define		ATMEL_US_INACK		(1 << 20)		/* Inhibit Non Acknowledge */
-#define		ATMEL_US_DSNACK		(1 << 21)		/* Disable Successive NACK */
-#define		ATMEL_US_MAX_ITER	(7 << 24)		/* Max Iterations */
-#define		ATMEL_US_FILTER		(1 << 28)		/* Infrared Receive Line Filter */
+#define ATMEL_US_MR		0x04	/* Mode Register */
+#define	ATMEL_US_USMODE		GENMASK(3, 0)	/* Mode of the USART */
+#define		ATMEL_US_USMODE_NORMAL		0
+#define		ATMEL_US_USMODE_RS485		1
+#define		ATMEL_US_USMODE_HWHS		2
+#define		ATMEL_US_USMODE_MODEM		3
+#define		ATMEL_US_USMODE_ISO7816_T0	4
+#define		ATMEL_US_USMODE_ISO7816_T1	6
+#define		ATMEL_US_USMODE_IRDA		8
+#define	ATMEL_US_USCLKS		GENMASK(5, 4)	/* Clock Selection */
+#define		ATMEL_US_USCLKS_MCK		(0 <<  4)
+#define		ATMEL_US_USCLKS_MCK_DIV8	(1 <<  4)
+#define		ATMEL_US_USCLKS_SCK		(3 <<  4)
+#define	ATMEL_US_CHRL		GENMASK(7, 6)	/* Character Length */
+#define		ATMEL_US_CHRL_5			(0 <<  6)
+#define		ATMEL_US_CHRL_6			(1 <<  6)
+#define		ATMEL_US_CHRL_7			(2 <<  6)
+#define		ATMEL_US_CHRL_8			(3 <<  6)
+#define	ATMEL_US_SYNC		BIT(8)		/* Synchronous Mode Select */
+#define	ATMEL_US_PAR		GENMASK(11, 9)	/* Parity Type */
+#define		ATMEL_US_PAR_EVEN		(0 <<  9)
+#define		ATMEL_US_PAR_ODD		(1 <<  9)
+#define		ATMEL_US_PAR_SPACE		(2 <<  9)
+#define		ATMEL_US_PAR_MARK		(3 <<  9)
+#define		ATMEL_US_PAR_NONE		(4 <<  9)
+#define		ATMEL_US_PAR_MULTI_DROP		(6 <<  9)
+#define	ATMEL_US_NBSTOP		GENMASK(13, 12)	/* Number of Stop Bits */
+#define		ATMEL_US_NBSTOP_1		(0 << 12)
+#define		ATMEL_US_NBSTOP_1_5		(1 << 12)
+#define		ATMEL_US_NBSTOP_2		(2 << 12)
+#define	ATMEL_US_CHMODE		GENMASK(15, 14)	/* Channel Mode */
+#define		ATMEL_US_CHMODE_NORMAL		(0 << 14)
+#define		ATMEL_US_CHMODE_ECHO		(1 << 14)
+#define		ATMEL_US_CHMODE_LOC_LOOP	(2 << 14)
+#define		ATMEL_US_CHMODE_REM_LOOP	(3 << 14)
+#define	ATMEL_US_MSBF		BIT(16)	/* Bit Order */
+#define	ATMEL_US_MODE9		BIT(17)	/* 9-bit Character Length */
+#define	ATMEL_US_CLKO		BIT(18)	/* Clock Output Select */
+#define	ATMEL_US_OVER		BIT(19)	/* Oversampling Mode */
+#define	ATMEL_US_INACK		BIT(20)	/* Inhibit Non Acknowledge */
+#define	ATMEL_US_DSNACK		BIT(21)	/* Disable Successive NACK */
+#define	ATMEL_US_MAX_ITER	GENMASK(26, 24)	/* Max Iterations */
+#define	ATMEL_US_FILTER		BIT(28)	/* Infrared Receive Line Filter */
 
-#define ATMEL_US_IER		0x08			/* Interrupt Enable Register */
-#define		ATMEL_US_RXRDY		(1 <<  0)		/* Receiver Ready */
-#define		ATMEL_US_TXRDY		(1 <<  1)		/* Transmitter Ready */
-#define		ATMEL_US_RXBRK		(1 <<  2)		/* Break Received / End of Break */
-#define		ATMEL_US_ENDRX		(1 <<  3)		/* End of Receiver Transfer */
-#define		ATMEL_US_ENDTX		(1 <<  4)		/* End of Transmitter Transfer */
-#define		ATMEL_US_OVRE		(1 <<  5)		/* Overrun Error */
-#define		ATMEL_US_FRAME		(1 <<  6)		/* Framing Error */
-#define		ATMEL_US_PARE		(1 <<  7)		/* Parity Error */
-#define		ATMEL_US_TIMEOUT	(1 <<  8)		/* Receiver Time-out */
-#define		ATMEL_US_TXEMPTY	(1 <<  9)		/* Transmitter Empty */
-#define		ATMEL_US_ITERATION	(1 << 10)		/* Max number of Repetitions Reached */
-#define		ATMEL_US_TXBUFE		(1 << 11)		/* Transmission Buffer Empty */
-#define		ATMEL_US_RXBUFF		(1 << 12)		/* Reception Buffer Full */
-#define		ATMEL_US_NACK		(1 << 13)		/* Non Acknowledge */
-#define		ATMEL_US_RIIC		(1 << 16)		/* Ring Indicator Input Change [AT91RM9200 only] */
-#define		ATMEL_US_DSRIC		(1 << 17)		/* Data Set Ready Input Change [AT91RM9200 only] */
-#define		ATMEL_US_DCDIC		(1 << 18)		/* Data Carrier Detect Input Change [AT91RM9200 only] */
-#define		ATMEL_US_CTSIC		(1 << 19)		/* Clear to Send Input Change */
-#define		ATMEL_US_RI		(1 << 20)		/* RI */
-#define		ATMEL_US_DSR		(1 << 21)		/* DSR */
-#define		ATMEL_US_DCD		(1 << 22)		/* DCD */
-#define		ATMEL_US_CTS		(1 << 23)		/* CTS */
+#define ATMEL_US_IER		0x08	/* Interrupt Enable Register */
+#define	ATMEL_US_RXRDY		BIT(0)	/* Receiver Ready */
+#define	ATMEL_US_TXRDY		BIT(1)	/* Transmitter Ready */
+#define	ATMEL_US_RXBRK		BIT(2)	/* Break Received / End of Break */
+#define	ATMEL_US_ENDRX		BIT(3)	/* End of Receiver Transfer */
+#define	ATMEL_US_ENDTX		BIT(4)	/* End of Transmitter Transfer */
+#define	ATMEL_US_OVRE		BIT(5)	/* Overrun Error */
+#define	ATMEL_US_FRAME		BIT(6)	/* Framing Error */
+#define	ATMEL_US_PARE		BIT(7)	/* Parity Error */
+#define	ATMEL_US_TIMEOUT	BIT(8)	/* Receiver Time-out */
+#define	ATMEL_US_TXEMPTY	BIT(9)	/* Transmitter Empty */
+#define	ATMEL_US_ITERATION	BIT(10)	/* Max number of Repetitions Reached */
+#define	ATMEL_US_TXBUFE		BIT(11)	/* Transmission Buffer Empty */
+#define	ATMEL_US_RXBUFF		BIT(12)	/* Reception Buffer Full */
+#define	ATMEL_US_NACK		BIT(13)	/* Non Acknowledge */
+#define	ATMEL_US_RIIC		BIT(16)	/* Ring Indicator Input Change */
+#define	ATMEL_US_DSRIC		BIT(17)	/* Data Set Ready Input Change */
+#define	ATMEL_US_DCDIC		BIT(18)	/* Data Carrier Detect Input Change */
+#define	ATMEL_US_CTSIC		BIT(19)	/* Clear to Send Input Change */
+#define	ATMEL_US_RI		BIT(20)	/* RI */
+#define	ATMEL_US_DSR		BIT(21)	/* DSR */
+#define	ATMEL_US_DCD		BIT(22)	/* DCD */
+#define	ATMEL_US_CTS		BIT(23)	/* CTS */
 
-#define ATMEL_US_IDR		0x0c			/* Interrupt Disable Register */
-#define ATMEL_US_IMR		0x10			/* Interrupt Mask Register */
-#define ATMEL_US_CSR		0x14			/* Channel Status Register */
-#define ATMEL_US_RHR		0x18			/* Receiver Holding Register */
-#define ATMEL_US_THR		0x1c			/* Transmitter Holding Register */
-#define		ATMEL_US_SYNH		(1 << 15)		/* Transmit/Receive Sync [AT91SAM9261 only] */
+#define ATMEL_US_IDR		0x0c	/* Interrupt Disable Register */
+#define ATMEL_US_IMR		0x10	/* Interrupt Mask Register */
+#define ATMEL_US_CSR		0x14	/* Channel Status Register */
+#define ATMEL_US_RHR		0x18	/* Receiver Holding Register */
+#define ATMEL_US_THR		0x1c	/* Transmitter Holding Register */
+#define	ATMEL_US_SYNH		BIT(15)	/* Transmit/Receive Sync */
 
-#define ATMEL_US_BRGR		0x20			/* Baud Rate Generator Register */
-#define		ATMEL_US_CD		(0xffff << 0)		/* Clock Divider */
+#define ATMEL_US_BRGR		0x20	/* Baud Rate Generator Register */
+#define	ATMEL_US_CD		GENMASK(15, 0)	/* Clock Divider */
 
-#define ATMEL_US_RTOR		0x24			/* Receiver Time-out Register */
-#define		ATMEL_US_TO		(0xffff << 0)		/* Time-out Value */
+#define ATMEL_US_RTOR		0x24	/* Receiver Time-out Register */
+#define	ATMEL_US_TO		GENMASK(15, 0)	/* Time-out Value */
 
-#define ATMEL_US_TTGR		0x28			/* Transmitter Timeguard Register */
-#define		ATMEL_US_TG		(0xff << 0)		/* Timeguard Value */
+#define ATMEL_US_TTGR		0x28	/* Transmitter Timeguard Register */
+#define	ATMEL_US_TG		GENMASK(7, 0)	/* Timeguard Value */
 
-#define ATMEL_US_FIDI		0x40			/* FI DI Ratio Register */
-#define ATMEL_US_NER		0x44			/* Number of Errors Register */
-#define ATMEL_US_IF		0x4c			/* IrDA Filter Register */
+#define ATMEL_US_FIDI		0x40	/* FI DI Ratio Register */
+#define ATMEL_US_NER		0x44	/* Number of Errors Register */
+#define ATMEL_US_IF		0x4c	/* IrDA Filter Register */
 
-#define ATMEL_US_NAME		0xf0			/* Ip Name */
-#define ATMEL_US_VERSION	0xfc			/* Ip Version */
+#define ATMEL_US_NAME		0xf0	/* Ip Name */
+#define ATMEL_US_VERSION	0xfc	/* Ip Version */
 
 #endif
Signature · 1 line
Nicolas Ferre
#7 lore raw

PATCH v2 3/4 tty/serial: at91: add support to FIFOs

In reply to Cyrille Pitchen

Depending on the hardware, TX and RX FIFOs may be available. The RX FIFO can avoid receive overruns, especially when DMA transfers are not used to read data from the Receive Holding Register. For heavy system load, The CPU is likely not be able to fetch data fast enough from the RHR.

In addition, the RX FIFO can supersede the DMA/PDC to control the RTS line when the Hardware Handshaking mode is enabled. Two thresholds are to be set for that purpose:
- When the number of data in the RX FIFO crosses and becomes lower than or equal to the low threshold, the RTS line is set to low
level: the remote peer is requested to send data.
- When the number of data in the RX FIFO crosses and becomes greater than or equal to the high threshold, the RTS line is set to high
level: the remote peer should stop sending new data.
- low threshold <= high threshold
Once these two thresholds are set properly, this new feature is enabled by setting the FIFO RTS Control bit of the FIFO Mode Register.

FIFOs also introduce a new multiple data mode: the USART works either in multiple data mode or in single data (legacy) mode.

If MODE9 bit is set into the Mode Register or if USMODE is set to either LIN_MASTER, LIN_SLAVE or LON_MODE, FIFOs operate in single data mode. Otherwise, they operate in multiple data mode.

In this new multiple data mode, accesses to the Receive Holding Register or Transmit Holding Register slightly change.

Since this driver implements neither the 9bit data feature (MODE9 bit set into the Mode Register) nor LIN modes, the USART works in multiple data mode whenever FIFOs are available and enabled. We also assume that data are 8bit wide.

In single data mode, 32bit access CAN be used to read a single data from RHR or write a single data into THR.
However in multiple data mode, a 32bit access to RHR now allows us to read four consecutive data from RX FIFO. Also a 32bit access to THR now allows to write four consecutive data into TX FIFO. So we MUST use 8bit access whenever only one data have to be read/written at a time.

---
 drivers/tty/serial/atmel_serial.c | 79 ++++++++++++++++++++++++++++++++++++++-
 include/linux/atmel_serial.h      | 36 ++++++++++++++++++
 2 files changed, 113 insertions(+), 2 deletions(-)
drivers/tty/serial/atmel_serial.c +77 -2
index 112e74b..6767570 100644
@@ -96,8 +96,8 @@ static void atmel_stop_rx(struct uart_port *port);
 #define UART_PUT_IDR(port, v)  __raw_writel(v, (port)->membase + ATMEL_US_IDR)
 #define UART_GET_IMR(port)     __raw_readl((port)->membase + ATMEL_US_IMR)
 #define UART_GET_CSR(port)     __raw_readl((port)->membase + ATMEL_US_CSR)
-#define UART_GET_CHAR(port)    __raw_readl((port)->membase + ATMEL_US_RHR)
-#define UART_PUT_CHAR(port, v) __raw_writel(v, (port)->membase + ATMEL_US_THR)
+#define UART_GET_CHAR(port)    __raw_readb((port)->membase + ATMEL_US_RHR)
+#define UART_PUT_CHAR(port, v) __raw_writeb(v, (port)->membase + ATMEL_US_THR)
 #define UART_GET_BRGR(port)    __raw_readl((port)->membase + ATMEL_US_BRGR)
 #define UART_PUT_BRGR(port, v) __raw_writel(v, (port)->membase + ATMEL_US_BRGR)
 #define UART_PUT_RTOR(port, v) __raw_writel(v, (port)->membase + ATMEL_US_RTOR)
@@ -119,6 +119,16 @@ static void atmel_stop_rx(struct uart_port *port);
 #define UART_PUT_TCR(port, v)  __raw_writel(v, (port)->membase + ATMEL_PDC_TCR)
 #define UART_GET_TCR(port)     __raw_readl((port)->membase + ATMEL_PDC_TCR)
 
+/* FIFO registers */
+#define UART_PUT_FMR(port, v)  __raw_writel(v, (port)->membase + ATMEL_US_FMR)
+#define UART_GET_FMR(port)     __raw_readl((port)->membase + ATMEL_US_FMR)
+#define UART_GET_FESR(port)    __raw_readl((port)->membase + ATMEL_US_FESR)
+#define UART_PUT_FIER(port, v) __raw_writel(v, (port)->membase + ATMEL_US_FIER)
+#define UART_PUT_FIDR(port, v) __raw_writel(v, (port)->membase + ATMEL_US_FIDR)
+#define UART_GET_FIMR(port)    __raw_readl((port)->membase + ATMEL_US_FIMR)
+#define UART_GET_FLR(port)     __raw_readl((port)->membase + ATMEL_US_FLR)
+
+
 struct atmel_dma_buffer {
 	unsigned char	*buf;
 	dma_addr_t	dma_addr;
@@ -172,6 +182,9 @@ struct atmel_uart_port {
 	struct mctrl_gpios	*gpios;
 	int			gpio_irq[UART_GPIO_MAX];
 	unsigned int		tx_done_mask;
+	u32			fifo_size;
+	u32			rts_high;
+	u32			rts_low;
 	bool			ms_irq_enabled;
 	bool			is_usart;	/* usart or uart */
 	struct timer_list	uart_timer;	/* uart timer */
@@ -1797,6 +1810,29 @@ static int atmel_startup(struct uart_port *port)
 			atmel_set_ops(port);
 	}
 
+	/*
+	 * Enable FIFO when available
+	 */
+	if (atmel_port->fifo_size) {
+		unsigned int txrdym = ATMEL_US_ONE_DATA;
+		unsigned int rxrdym = ATMEL_US_ONE_DATA;
+		unsigned int fmr;
+
+		UART_PUT_CR(port,
+			    ATMEL_US_FIFOEN |
+			    ATMEL_US_RXFCLR |
+			    ATMEL_US_TXFLCLR);
+
+		fmr = ATMEL_US_TXRDYM(txrdym) | ATMEL_US_RXRDYM(rxrdym);
+		if (atmel_port->rts_high &&
+		    atmel_port->rts_low)
+			fmr |=	ATMEL_US_FRTSC |
+				ATMEL_US_RXFTHRES(atmel_port->rts_high) |
+				ATMEL_US_RXFTHRES2(atmel_port->rts_low);
+
+		UART_PUT_FMR(port, fmr);
+	}
+
 	/* Save current CSR for comparison in atmel_tasklet_func() */
 	atmel_port->irq_status_prev = atmel_get_lines_status(port);
 	atmel_port->irq_status = atmel_port->irq_status_prev;
@@ -2599,6 +2635,44 @@ static int atmel_init_gpios(struct atmel_uart_port *p, struct device *dev)
 	return 0;
 }
 
+static void atmel_serial_probe_fifos(struct atmel_uart_port *port,
+				     struct platform_device *pdev)
+{
+	port->fifo_size = 0;
+	port->rts_low = 0;
+	port->rts_high = 0;
+
+	if (of_property_read_u32(pdev->dev.of_node,
+				 "atmel,fifo-size",
+				 &port->fifo_size) ||
+	    !port->fifo_size)
+		return;
+
+	if (port->fifo_size < 8 || port->fifo_size > 32) {
+		port->fifo_size = 0;
+		dev_err(&pdev->dev, "Invalid FIFO size\n");
+		return;
+	}
+
+	/*
+	 * 0 <= rts_low <= rts_high <= fifo_size
+	 * Once their CTS line asserted by the remote peer, some x86 UARTs tend
+	 * to flush their internal TX FIFO, commonly up to 8 data, before
+	 * actually stopping to send new data. So we try to set the RTS High
+	 * Threshold to a raisonable high value respecting this 8 data empirical
+	 * rule when possible.
+	 */
+	port->rts_high = max_t(int, port->fifo_size >> 1, port->fifo_size - 8);
+	port->rts_low  = max_t(int, port->fifo_size >> 2, port->fifo_size - 16);
+
+	dev_info(&pdev->dev, "Using FIFO (%u data)\n",
+		 port->fifo_size);
+	dev_info(&pdev->dev, "RTS High Threshold : %2u data\n",
+		 port->rts_high);
+	dev_info(&pdev->dev, "RTS Low Threshold  : %2u data\n",
+		 port->rts_low);
+}
+
 static int atmel_serial_probe(struct platform_device *pdev)
 {
 	struct atmel_uart_port *port;
@@ -2635,6 +2709,7 @@ static int atmel_serial_probe(struct platform_device *pdev)
 	port = &atmel_ports[ret];
 	port->backup_imr = 0;
 	port->uart.line = ret;
+	atmel_serial_probe_fifos(port, pdev);
 
 	spin_lock_init(&port->lock_suspended);
 
include/linux/atmel_serial.h +36 -0
index c384c21..ee696d7 100644
@@ -35,6 +35,11 @@
 #define	ATMEL_US_DTRDIS		BIT(17)	/* Data Terminal Ready Disable */
 #define	ATMEL_US_RTSEN		BIT(18)	/* Request To Send Enable */
 #define	ATMEL_US_RTSDIS		BIT(19)	/* Request To Send Disable */
+#define	ATMEL_US_TXFCLR		BIT(24)	/* Transmit FIFO Clear */
+#define	ATMEL_US_RXFCLR		BIT(25)	/* Receive FIFO Clear */
+#define	ATMEL_US_TXFLCLR	BIT(26)	/* Transmit FIFO Lock Clear */
+#define	ATMEL_US_FIFOEN		BIT(30)	/* FIFO enable */
+#define	ATMEL_US_FIFODIS	BIT(31)	/* FIFO disable */
 
 #define ATMEL_US_MR		0x04	/* Mode Register */
 #define	ATMEL_US_USMODE		GENMASK(3, 0)	/* Mode of the USART */
@@ -124,6 +129,37 @@
 #define ATMEL_US_NER		0x44	/* Number of Errors Register */
 #define ATMEL_US_IF		0x4c	/* IrDA Filter Register */
 
+#define ATMEL_US_CMPR		0x90	/* Comparaison Register */
+#define ATMEL_US_FMR		0xa0	/* FIFO Mode Register */
+#define	ATMEL_US_TXRDYM(data)	(((data) & 0x3) << 0)	/* TX Ready Mode */
+#define	ATMEL_US_RXRDYM(data)	(((data) & 0x3) << 4)	/* RX Ready Mode */
+#define		ATMEL_US_ONE_DATA	0x0
+#define		ATMEL_US_TWO_DATA	0x1
+#define		ATMEL_US_FOUR_DATA	0x2
+#define	ATMEL_US_FRTSC		BIT(7)	/* FIFO RTS pin Control */
+#define	ATMEL_US_TXFTHRES(thr)	(((thr) & 0x3f) << 8)	/* TX FIFO Threshold */
+#define	ATMEL_US_RXFTHRES(thr)	(((thr) & 0x3f) << 16)	/* RX FIFO Threshold */
+#define	ATMEL_US_RXFTHRES2(thr)	(((thr) & 0x3f) << 24)	/* RX FIFO Threshold2 */
+
+#define ATMEL_US_FLR		0xa4	/* FIFO Level Register */
+#define	ATMEL_US_TXFL(reg)	(((reg) >> 0) & 0x3f)	/* TX FIFO Level */
+#define	ATMEL_US_RXFL(reg)	(((reg) >> 16) & 0x3f)	/* RX FIFO Level */
+
+#define ATMEL_US_FIER		0xa8	/* FIFO Interrupt Enable Register */
+#define ATMEL_US_FIDR		0xac	/* FIFO Interrupt Disable Register */
+#define ATMEL_US_FIMR		0xb0	/* FIFO Interrupt Mask Register */
+#define ATMEL_US_FESR		0xb4	/* FIFO Event Status Register */
+#define	ATMEL_US_TXFEF		BIT(0)	/* Transmit FIFO Empty Flag */
+#define	ATMEL_US_TXFFF		BIT(1)	/* Transmit FIFO Full Flag */
+#define	ATMEL_US_TXFTHF		BIT(2)	/* Transmit FIFO Threshold Flag */
+#define	ATMEL_US_RXFEF		BIT(3)	/* Receive FIFO Empty Flag */
+#define	ATMEL_US_RXFFF		BIT(4)	/* Receive FIFO Full Flag */
+#define	ATMEL_US_RXFTHF		BIT(5)	/* Receive FIFO Threshold Flag */
+#define	ATMEL_US_TXFPTEF	BIT(6)	/* Transmit FIFO Pointer Error Flag */
+#define	ATMEL_US_RXFPTEF	BIT(7)	/* Receive FIFO Pointer Error Flag */
+#define	ATMEL_US_TXFLOCK	BIT(8)	/* Transmit FIFO Lock (FESR only) */
+#define	ATMEL_US_RXFTHF2	BIT(9)	/* Receive FIFO Threshold Flag 2 */
+
 #define ATMEL_US_NAME		0xf0	/* Ip Name */
 #define ATMEL_US_VERSION	0xfc	/* Ip Version */
 
Signature · 1 line
1.8.2.2
#8 lore raw

Hi,

Quoted 2 lines · Cyrille Pitchen

Seems good to me, a few comments:

drivers/tty/serial/atmel_serial.c static int atmel_init_gpios(struct atmel_uart_port *p, struct device *dev) · line 2651

Quoted 1 line
drivers/tty/serial/atmel_serial.c +1 -0
+	if (port->fifo_size < 8 || port->fifo_size > 32) {

I'm not sure why you limit the size to 32. At some point, a new SoC may be release with a bigger FIFO then we'll have to change the code instead of just putting the appropriate value in the device tree.

drivers/tty/serial/atmel_serial.c static int atmel_init_gpios(struct atmel_uart_port *p, struct device *dev) · line 2662

Quoted 11 lines
drivers/tty/serial/atmel_serial.c +11 -0
+		port->fifo_size = 0;
+		dev_err(&pdev->dev, "Invalid FIFO size\n");
+		return;
+	}
+
+	/*
+	 * 0 <= rts_low <= rts_high <= fifo_size
+	 * Once their CTS line asserted by the remote peer, some x86 UARTs tend
+	 * to flush their internal TX FIFO, commonly up to 8 data, before
+	 * actually stopping to send new data. So we try to set the RTS High
+	 * Threshold to a raisonable high value respecting this 8 data empirical

reasonably --^

Signature · 3 lines
Alexandre Belloni, Free Electrons
Embedded Linux, Kernel and Android engineering
http://free-electrons.com
#9 lore raw
In reply to Cyrille Pitchen

Le 11/06/2015 18:20, Cyrille Pitchen a ?crit :

Quoted 98 lines

Depending on the hardware, TX and RX FIFOs may be available. The RX FIFO can avoid receive overruns, especially when DMA transfers are not used to read data from the Receive Holding Register. For heavy system load, The CPU is likely not be able to fetch data fast enough from the RHR.

In addition, the RX FIFO can supersede the DMA/PDC to control the RTS line when the Hardware Handshaking mode is enabled. Two thresholds are to be set for that purpose:
- When the number of data in the RX FIFO crosses and becomes lower than or equal to the low threshold, the RTS line is set to low
level: the remote peer is requested to send data.
- When the number of data in the RX FIFO crosses and becomes greater than or equal to the high threshold, the RTS line is set to high
level: the remote peer should stop sending new data.
- low threshold <= high threshold
Once these two thresholds are set properly, this new feature is enabled by setting the FIFO RTS Control bit of the FIFO Mode Register.

FIFOs also introduce a new multiple data mode: the USART works either in multiple data mode or in single data (legacy) mode.

If MODE9 bit is set into the Mode Register or if USMODE is set to either LIN_MASTER, LIN_SLAVE or LON_MODE, FIFOs operate in single data mode. Otherwise, they operate in multiple data mode.

In this new multiple data mode, accesses to the Receive Holding Register or Transmit Holding Register slightly change.

Since this driver implements neither the 9bit data feature (MODE9 bit set into the Mode Register) nor LIN modes, the USART works in multiple data mode whenever FIFOs are available and enabled. We also assume that data are 8bit wide.

In single data mode, 32bit access CAN be used to read a single data from RHR or write a single data into THR.
However in multiple data mode, a 32bit access to RHR now allows us to read four consecutive data from RX FIFO. Also a 32bit access to THR now allows to write four consecutive data into TX FIFO. So we MUST use 8bit access whenever only one data have to be read/written at a time.

---
 drivers/tty/serial/atmel_serial.c | 79 ++++++++++++++++++++++++++++++++++++++-
 include/linux/atmel_serial.h      | 36 ++++++++++++++++++
 2 files changed, 113 insertions(+), 2 deletions(-)
drivers/tty/serial/atmel_serial.c +19 -2
index 112e74b..6767570 100644
@@ -96,8 +96,8 @@ static void atmel_stop_rx(struct uart_port *port);
 #define UART_PUT_IDR(port, v)  __raw_writel(v, (port)->membase + ATMEL_US_IDR)
 #define UART_GET_IMR(port)     __raw_readl((port)->membase + ATMEL_US_IMR)
 #define UART_GET_CSR(port)     __raw_readl((port)->membase + ATMEL_US_CSR)
-#define UART_GET_CHAR(port)    __raw_readl((port)->membase + ATMEL_US_RHR)
-#define UART_PUT_CHAR(port, v) __raw_writel(v, (port)->membase + ATMEL_US_THR)
+#define UART_GET_CHAR(port)    __raw_readb((port)->membase + ATMEL_US_RHR)
+#define UART_PUT_CHAR(port, v) __raw_writeb(v, (port)->membase + ATMEL_US_THR)
 #define UART_GET_BRGR(port)    __raw_readl((port)->membase + ATMEL_US_BRGR)
 #define UART_PUT_BRGR(port, v) __raw_writel(v, (port)->membase + ATMEL_US_BRGR)
 #define UART_PUT_RTOR(port, v) __raw_writel(v, (port)->membase + ATMEL_US_RTOR)
@@ -119,6 +119,16 @@ static void atmel_stop_rx(struct uart_port *port);
 #define UART_PUT_TCR(port, v)  __raw_writel(v, (port)->membase + ATMEL_PDC_TCR)
 #define UART_GET_TCR(port)     __raw_readl((port)->membase + ATMEL_PDC_TCR)
 
+/* FIFO registers */
+#define UART_PUT_FMR(port, v)  __raw_writel(v, (port)->membase + ATMEL_US_FMR)
+#define UART_GET_FMR(port)     __raw_readl((port)->membase + ATMEL_US_FMR)
+#define UART_GET_FESR(port)    __raw_readl((port)->membase + ATMEL_US_FESR)
+#define UART_PUT_FIER(port, v) __raw_writel(v, (port)->membase + ATMEL_US_FIER)
+#define UART_PUT_FIDR(port, v) __raw_writel(v, (port)->membase + ATMEL_US_FIDR)
+#define UART_GET_FIMR(port)    __raw_readl((port)->membase + ATMEL_US_FIMR)
+#define UART_GET_FLR(port)     __raw_readl((port)->membase + ATMEL_US_FLR)
+
+
 struct atmel_dma_buffer {
 	unsigned char	*buf;
 	dma_addr_t	dma_addr;
@@ -172,6 +182,9 @@ struct atmel_uart_port {
 	struct mctrl_gpios	*gpios;
 	int			gpio_irq[UART_GPIO_MAX];
 	unsigned int		tx_done_mask;
+	u32			fifo_size;
+	u32			rts_high;
+	u32			rts_low;
 	bool			ms_irq_enabled;
 	bool			is_usart;	/* usart or uart */
 	struct timer_list	uart_timer;	/* uart timer */
@@ -1797,6 +1810,29 @@ static int atmel_startup(struct uart_port *port)
 			atmel_set_ops(port);
 	}
 
+	/*
+	 * Enable FIFO when available
+	 */
+	if (atmel_port->fifo_size) {

drivers/tty/serial/atmel_serial.c static int atmel_startup(struct uart_port *port) · line 1826

Quoted 10 lines
drivers/tty/serial/atmel_serial.c +10 -0
+		unsigned int txrdym = ATMEL_US_ONE_DATA;
+		unsigned int rxrdym = ATMEL_US_ONE_DATA;
+		unsigned int fmr;
+
+		UART_PUT_CR(port,
+			    ATMEL_US_FIFOEN |
+			    ATMEL_US_RXFCLR |
+			    ATMEL_US_TXFLCLR);
+
+		fmr = ATMEL_US_TXRDYM(txrdym) | ATMEL_US_RXRDYM(rxrdym);

If above is the only place where we use txrdym and rxrdym, why not use directly the defined constants here?

Quoted 17 lines
drivers/tty/serial/atmel_serial.c +10 -0
+		if (atmel_port->rts_high &&
+		    atmel_port->rts_low)
+			fmr |=	ATMEL_US_FRTSC |
+				ATMEL_US_RXFTHRES(atmel_port->rts_high) |
+				ATMEL_US_RXFTHRES2(atmel_port->rts_low);
+
+		UART_PUT_FMR(port, fmr);
+	}
+
 	/* Save current CSR for comparison in atmel_tasklet_func() */
 	atmel_port->irq_status_prev = atmel_get_lines_status(port);
 	atmel_port->irq_status = atmel_port->irq_status_prev;
@@ -2599,6 +2635,44 @@ static int atmel_init_gpios(struct atmel_uart_port *p, struct device *dev)
 	return 0;
 }
 
+static void atmel_serial_probe_fifos(struct atmel_uart_port *port,

drivers/tty/serial/atmel_serial.c static int atmel_init_gpios(struct atmel_uart_port *p, struct device *dev) · line 2648

Quoted 10 lines
drivers/tty/serial/atmel_serial.c +10 -0
+				     struct platform_device *pdev)
+{
+	port->fifo_size = 0;
+	port->rts_low = 0;
+	port->rts_high = 0;
+
+	if (of_property_read_u32(pdev->dev.of_node,
+				 "atmel,fifo-size",
+				 &port->fifo_size) ||
+	    !port->fifo_size)

I understand, but here, please split the test in 2: it's far more readable.

drivers/tty/serial/atmel_serial.c static int atmel_init_gpios(struct atmel_uart_port *p, struct device *dev) · line 2651

Quoted 3 lines
drivers/tty/serial/atmel_serial.c +3 -0
+		return;
+
+	if (port->fifo_size < 8 || port->fifo_size > 32) {

Can you define constants for these 2 values and use them here?

Quoted 5 lines
drivers/tty/serial/atmel_serial.c +5 -0
+		port->fifo_size = 0;
+		dev_err(&pdev->dev, "Invalid FIFO size\n");
+		return;
+	}
+

drivers/tty/serial/atmel_serial.c static int atmel_init_gpios(struct atmel_uart_port *p, struct device *dev) · line 2666

Quoted 10 lines
drivers/tty/serial/atmel_serial.c +10 -0
+	/*
+	 * 0 <= rts_low <= rts_high <= fifo_size
+	 * Once their CTS line asserted by the remote peer, some x86 UARTs tend
+	 * to flush their internal TX FIFO, commonly up to 8 data, before
+	 * actually stopping to send new data. So we try to set the RTS High
+	 * Threshold to a raisonable high value respecting this 8 data empirical
+	 * rule when possible.
+	 */
+	port->rts_high = max_t(int, port->fifo_size >> 1, port->fifo_size - 8);
+	port->rts_low  = max_t(int, port->fifo_size >> 2, port->fifo_size - 16);

Same here: I'd prefer constants with meaningful names.

drivers/tty/serial/atmel_serial.c static int atmel_init_gpios(struct atmel_uart_port *p, struct device *dev) · line 2669

Quoted 3 lines
drivers/tty/serial/atmel_serial.c +3 -0
+
+	dev_info(&pdev->dev, "Using FIFO (%u data)\n",
+		 port->fifo_size);

Okay for this information.

drivers/tty/serial/atmel_serial.c static int atmel_init_gpios(struct atmel_uart_port *p, struct device *dev) · line 2673

Quoted 4 lines
drivers/tty/serial/atmel_serial.c +4 -0
+	dev_info(&pdev->dev, "RTS High Threshold : %2u data\n",
+		 port->rts_high);
+	dev_info(&pdev->dev, "RTS Low Threshold  : %2u data\n",
+		 port->rts_low);

Aren't these two other dev_info() better suited as dev_dbg()?

Quoted 56 lines
drivers/tty/serial/atmel_serial.c +3 -0
+}
+
 static int atmel_serial_probe(struct platform_device *pdev)
 {
 	struct atmel_uart_port *port;
@@ -2635,6 +2709,7 @@ static int atmel_serial_probe(struct platform_device *pdev)
 	port = &atmel_ports[ret];
 	port->backup_imr = 0;
 	port->uart.line = ret;
+	atmel_serial_probe_fifos(port, pdev);
 
 	spin_lock_init(&port->lock_suspended);
 
include/linux/atmel_serial.h +28 -0
index c384c21..ee696d7 100644
@@ -35,6 +35,11 @@
 #define	ATMEL_US_DTRDIS		BIT(17)	/* Data Terminal Ready Disable */
 #define	ATMEL_US_RTSEN		BIT(18)	/* Request To Send Enable */
 #define	ATMEL_US_RTSDIS		BIT(19)	/* Request To Send Disable */
+#define	ATMEL_US_TXFCLR		BIT(24)	/* Transmit FIFO Clear */
+#define	ATMEL_US_RXFCLR		BIT(25)	/* Receive FIFO Clear */
+#define	ATMEL_US_TXFLCLR	BIT(26)	/* Transmit FIFO Lock Clear */
+#define	ATMEL_US_FIFOEN		BIT(30)	/* FIFO enable */
+#define	ATMEL_US_FIFODIS	BIT(31)	/* FIFO disable */
 
 #define ATMEL_US_MR		0x04	/* Mode Register */
 #define	ATMEL_US_USMODE		GENMASK(3, 0)	/* Mode of the USART */
@@ -124,6 +129,37 @@
 #define ATMEL_US_NER		0x44	/* Number of Errors Register */
 #define ATMEL_US_IF		0x4c	/* IrDA Filter Register */
 
+#define ATMEL_US_CMPR		0x90	/* Comparaison Register */
+#define ATMEL_US_FMR		0xa0	/* FIFO Mode Register */
+#define	ATMEL_US_TXRDYM(data)	(((data) & 0x3) << 0)	/* TX Ready Mode */
+#define	ATMEL_US_RXRDYM(data)	(((data) & 0x3) << 4)	/* RX Ready Mode */
+#define		ATMEL_US_ONE_DATA	0x0
+#define		ATMEL_US_TWO_DATA	0x1
+#define		ATMEL_US_FOUR_DATA	0x2
+#define	ATMEL_US_FRTSC		BIT(7)	/* FIFO RTS pin Control */
+#define	ATMEL_US_TXFTHRES(thr)	(((thr) & 0x3f) << 8)	/* TX FIFO Threshold */
+#define	ATMEL_US_RXFTHRES(thr)	(((thr) & 0x3f) << 16)	/* RX FIFO Threshold */
+#define	ATMEL_US_RXFTHRES2(thr)	(((thr) & 0x3f) << 24)	/* RX FIFO Threshold2 */
+
+#define ATMEL_US_FLR		0xa4	/* FIFO Level Register */
+#define	ATMEL_US_TXFL(reg)	(((reg) >> 0) & 0x3f)	/* TX FIFO Level */
+#define	ATMEL_US_RXFL(reg)	(((reg) >> 16) & 0x3f)	/* RX FIFO Level */
+
+#define ATMEL_US_FIER		0xa8	/* FIFO Interrupt Enable Register */
+#define ATMEL_US_FIDR		0xac	/* FIFO Interrupt Disable Register */
+#define ATMEL_US_FIMR		0xb0	/* FIFO Interrupt Mask Register */
+#define ATMEL_US_FESR		0xb4	/* FIFO Event Status Register */
+#define	ATMEL_US_TXFEF		BIT(0)	/* Transmit FIFO Empty Flag */
+#define	ATMEL_US_TXFFF		BIT(1)	/* Transmit FIFO Full Flag */
+#define	ATMEL_US_TXFTHF		BIT(2)	/* Transmit FIFO Threshold Flag */

include/linux/atmel_serial.h line 164

Quoted 10 lines
include/linux/atmel_serial.h +8 -0
+#define	ATMEL_US_RXFEF		BIT(3)	/* Receive FIFO Empty Flag */
+#define	ATMEL_US_RXFFF		BIT(4)	/* Receive FIFO Full Flag */
+#define	ATMEL_US_RXFTHF		BIT(5)	/* Receive FIFO Threshold Flag */
+#define	ATMEL_US_TXFPTEF	BIT(6)	/* Transmit FIFO Pointer Error Flag */
+#define	ATMEL_US_RXFPTEF	BIT(7)	/* Receive FIFO Pointer Error Flag */
+#define	ATMEL_US_TXFLOCK	BIT(8)	/* Transmit FIFO Lock (FESR only) */
+#define	ATMEL_US_RXFTHF2	BIT(9)	/* Receive FIFO Threshold Flag 2 */
+
 #define ATMEL_US_NAME		0xf0	/* Ip Name */
 #define ATMEL_US_VERSION	0xfc	/* Ip Version */

Otherwise it look good.

Thanks, bye,

Signature · 1 line
Nicolas Ferre
#10 lore raw

PATCH v2 4/4 tty/serial: at91: use 32bit writes into TX FIFO when DMA is enabled

In reply to Cyrille Pitchen

For now this improvement is only used with TX DMA transfers. The data width must be set properly when configuring the DMA controller. Also the FIFO configuration must be set to match the DMA transfer data width:
TXRDYM (Transmitter Ready Mode) and RXRDYM (Receiver Ready Mode) must be set into the FIFO Mode Register. These values are used by the USART to trigger the DMA controller. In single data mode they are not used and should be reset to 0.
So the TXRDYM bits are changed to FOUR_DATA; then USART triggers the DMA controller when at least 4 data can be written into the TX FIFO througth the THR. On the other hand the RXRDYM bits are left unchanged to ONE_DATA.

Atmel eXtended DMA controller allows us to set a different data width for each part of a scatter-gather transfer. So when calling dmaengine_slave_config() to configure the TX path, we just need to set dst_addr_width to the maximum data width. Then DMA writes into THR are split into up to two parts. The first part carries the first data to be sent and has a length equal to the greatest multiple of 4 (bytes) lower than or equal to the total length of the TX DMA transfer. The second part carries the trailing data (up to 3 bytes). The first part is written by the DMA into THR using 32 bit accesses, whereas 8bit accesses are used for the second part.

---
 drivers/tty/serial/atmel_serial.c | 66 ++++++++++++++++++++++++++++++---------
 1 file changed, 51 insertions(+), 15 deletions(-)
drivers/tty/serial/atmel_serial.c +51 -15
index 6767570..270bb28 100644
@@ -176,6 +176,7 @@ struct atmel_uart_port {
 	unsigned int		irq_status;
 	unsigned int		irq_status_prev;
 	unsigned int		status_change;
+	unsigned int		tx_len;
 
 	struct circ_buf		rx_ring;
 
@@ -743,10 +744,10 @@ static void atmel_complete_tx_dma(void *arg)
 
 	if (chan)
 		dmaengine_terminate_all(chan);
-	xmit->tail += sg_dma_len(&atmel_port->sg_tx);
+	xmit->tail += atmel_port->tx_len;
 	xmit->tail &= UART_XMIT_SIZE - 1;
 
-	port->icount.tx += sg_dma_len(&atmel_port->sg_tx);
+	port->icount.tx += atmel_port->tx_len;
 
 	spin_lock_irq(&atmel_port->lock_tx);
 	async_tx_ack(atmel_port->desc_tx);
@@ -794,7 +795,9 @@ static void atmel_tx_dma(struct uart_port *port)
 	struct circ_buf *xmit = &port->state->xmit;
 	struct dma_chan *chan = atmel_port->chan_tx;
 	struct dma_async_tx_descriptor *desc;
-	struct scatterlist *sg = &atmel_port->sg_tx;
+	struct scatterlist sgl[2], *sg, *sg_tx = &atmel_port->sg_tx;
+	unsigned int tx_len, part1_len, part2_len, sg_len;
+	dma_addr_t phys_addr;
 
 	/* Make sure we have an idle channel */
 	if (atmel_port->desc_tx != NULL)
@@ -810,18 +813,46 @@ static void atmel_tx_dma(struct uart_port *port)
 		 * Take the port lock to get a
 		 * consistent xmit buffer state.
 		 */
-		sg->offset = xmit->tail & (UART_XMIT_SIZE - 1);
-		sg_dma_address(sg) = (sg_dma_address(sg) &
-					~(UART_XMIT_SIZE - 1))
-					+ sg->offset;
-		sg_dma_len(sg) = CIRC_CNT_TO_END(xmit->head,
-						xmit->tail,
-						UART_XMIT_SIZE);
-		BUG_ON(!sg_dma_len(sg));
+		tx_len = CIRC_CNT_TO_END(xmit->head,
+					 xmit->tail,
+					 UART_XMIT_SIZE);
+
+		if (atmel_port->fifo_size) {
+			/* multi data mode */
+			part1_len = (tx_len & ~0x3); /* DWORD access */
+			part2_len = (tx_len & 0x3); /* BYTE access */
+		} else {
+			/* single data (legacy) mode */
+			part1_len = 0;
+			part2_len = tx_len; /* BYTE access only */
+		}
+
+		sg_init_table(sgl, 2);
+		sg_len = 0;
+		phys_addr = sg_dma_address(sg_tx) + xmit->tail;
+		if (part1_len) {
+			sg = &sgl[sg_len++];
+			sg_dma_address(sg) = phys_addr;
+			sg_dma_len(sg) = part1_len;
+
+			phys_addr += part1_len;
+		}
+
+		if (part2_len) {
+			sg = &sgl[sg_len++];
+			sg_dma_address(sg) = phys_addr;
+			sg_dma_len(sg) = part2_len;
+		}
+
+		/*
+		 * save tx_len so atmel_complete_tx_dma() will increase
+		 * xmit->tail correctly
+		 */
+		atmel_port->tx_len = tx_len;
 
 		desc = dmaengine_prep_slave_sg(chan,
-					       sg,
-					       1,
+					       sgl,
+					       sg_len,
 					       DMA_MEM_TO_DEV,
 					       DMA_PREP_INTERRUPT |
 					       DMA_CTRL_ACK);
@@ -830,7 +861,7 @@ static void atmel_tx_dma(struct uart_port *port)
 			return;
 		}
 
-		dma_sync_sg_for_device(port->dev, sg, 1, DMA_TO_DEVICE);
+		dma_sync_sg_for_device(port->dev, sg_tx, 1, DMA_TO_DEVICE);
 
 		atmel_port->desc_tx = desc;
 		desc->callback = atmel_complete_tx_dma;
@@ -890,7 +921,9 @@ static int atmel_prepare_tx_dma(struct uart_port *port)
 	/* Configure the slave DMA */
 	memset(&config, 0, sizeof(config));
 	config.direction = DMA_MEM_TO_DEV;
-	config.dst_addr_width = DMA_SLAVE_BUSWIDTH_1_BYTE;
+	config.dst_addr_width = (atmel_port->fifo_size) ?
+				DMA_SLAVE_BUSWIDTH_4_BYTES :
+				DMA_SLAVE_BUSWIDTH_1_BYTE;
 	config.dst_addr = port->mapbase + ATMEL_US_THR;
 	config.dst_maxburst = 1;
 
@@ -1823,6 +1856,9 @@ static int atmel_startup(struct uart_port *port)
 			    ATMEL_US_RXFCLR |
 			    ATMEL_US_TXFLCLR);
 
+		if (atmel_use_dma_tx(port))
+			txrdym = ATMEL_US_FOUR_DATA;
+
 		fmr = ATMEL_US_TXRDYM(txrdym) | ATMEL_US_RXRDYM(rxrdym);
 		if (atmel_port->rts_high &&
 		    atmel_port->rts_low)
Signature · 1 line
1.8.2.2
Acked-by on v2 4/4
#11 lore raw

Re: [PATCH linux-next v2 4/4] tty/serial: at91: use 32bit writes into TX FIFO when DMA is enabled

Le 11/06/2015 18:20, Cyrille Pitchen a ?crit :

Quoted 143 lines

For now this improvement is only used with TX DMA transfers. The data width must be set properly when configuring the DMA controller. Also the FIFO configuration must be set to match the DMA transfer data width:
TXRDYM (Transmitter Ready Mode) and RXRDYM (Receiver Ready Mode) must be set into the FIFO Mode Register. These values are used by the USART to trigger the DMA controller. In single data mode they are not used and should be reset to 0.
So the TXRDYM bits are changed to FOUR_DATA; then USART triggers the DMA controller when at least 4 data can be written into the TX FIFO througth the THR. On the other hand the RXRDYM bits are left unchanged to ONE_DATA.

Atmel eXtended DMA controller allows us to set a different data width for each part of a scatter-gather transfer. So when calling dmaengine_slave_config() to configure the TX path, we just need to set dst_addr_width to the maximum data width. Then DMA writes into THR are split into up to two parts. The first part carries the first data to be sent and has a length equal to the greatest multiple of 4 (bytes) lower than or equal to the total length of the TX DMA transfer. The second part carries the trailing data (up to 3 bytes). The first part is written by the DMA into THR using 32 bit accesses, whereas 8bit accesses are used for the second part.

---
 drivers/tty/serial/atmel_serial.c | 66 ++++++++++++++++++++++++++++++---------
 1 file changed, 51 insertions(+), 15 deletions(-)
drivers/tty/serial/atmel_serial.c +48 -15
index 6767570..270bb28 100644
@@ -176,6 +176,7 @@ struct atmel_uart_port {
 	unsigned int		irq_status;
 	unsigned int		irq_status_prev;
 	unsigned int		status_change;
+	unsigned int		tx_len;
 
 	struct circ_buf		rx_ring;
 
@@ -743,10 +744,10 @@ static void atmel_complete_tx_dma(void *arg)
 
 	if (chan)
 		dmaengine_terminate_all(chan);
-	xmit->tail += sg_dma_len(&atmel_port->sg_tx);
+	xmit->tail += atmel_port->tx_len;
 	xmit->tail &= UART_XMIT_SIZE - 1;
 
-	port->icount.tx += sg_dma_len(&atmel_port->sg_tx);
+	port->icount.tx += atmel_port->tx_len;
 
 	spin_lock_irq(&atmel_port->lock_tx);
 	async_tx_ack(atmel_port->desc_tx);
@@ -794,7 +795,9 @@ static void atmel_tx_dma(struct uart_port *port)
 	struct circ_buf *xmit = &port->state->xmit;
 	struct dma_chan *chan = atmel_port->chan_tx;
 	struct dma_async_tx_descriptor *desc;
-	struct scatterlist *sg = &atmel_port->sg_tx;
+	struct scatterlist sgl[2], *sg, *sg_tx = &atmel_port->sg_tx;
+	unsigned int tx_len, part1_len, part2_len, sg_len;
+	dma_addr_t phys_addr;
 
 	/* Make sure we have an idle channel */
 	if (atmel_port->desc_tx != NULL)
@@ -810,18 +813,46 @@ static void atmel_tx_dma(struct uart_port *port)
 		 * Take the port lock to get a
 		 * consistent xmit buffer state.
 		 */
-		sg->offset = xmit->tail & (UART_XMIT_SIZE - 1);
-		sg_dma_address(sg) = (sg_dma_address(sg) &
-					~(UART_XMIT_SIZE - 1))
-					+ sg->offset;
-		sg_dma_len(sg) = CIRC_CNT_TO_END(xmit->head,
-						xmit->tail,
-						UART_XMIT_SIZE);
-		BUG_ON(!sg_dma_len(sg));
+		tx_len = CIRC_CNT_TO_END(xmit->head,
+					 xmit->tail,
+					 UART_XMIT_SIZE);
+
+		if (atmel_port->fifo_size) {
+			/* multi data mode */
+			part1_len = (tx_len & ~0x3); /* DWORD access */
+			part2_len = (tx_len & 0x3); /* BYTE access */
+		} else {
+			/* single data (legacy) mode */
+			part1_len = 0;
+			part2_len = tx_len; /* BYTE access only */
+		}
+
+		sg_init_table(sgl, 2);
+		sg_len = 0;
+		phys_addr = sg_dma_address(sg_tx) + xmit->tail;
+		if (part1_len) {
+			sg = &sgl[sg_len++];
+			sg_dma_address(sg) = phys_addr;
+			sg_dma_len(sg) = part1_len;
+
+			phys_addr += part1_len;
+		}
+
+		if (part2_len) {
+			sg = &sgl[sg_len++];
+			sg_dma_address(sg) = phys_addr;
+			sg_dma_len(sg) = part2_len;
+		}
+
+		/*
+		 * save tx_len so atmel_complete_tx_dma() will increase
+		 * xmit->tail correctly
+		 */
+		atmel_port->tx_len = tx_len;
 
 		desc = dmaengine_prep_slave_sg(chan,
-					       sg,
-					       1,
+					       sgl,
+					       sg_len,
 					       DMA_MEM_TO_DEV,
 					       DMA_PREP_INTERRUPT |
 					       DMA_CTRL_ACK);
@@ -830,7 +861,7 @@ static void atmel_tx_dma(struct uart_port *port)
 			return;
 		}
 
-		dma_sync_sg_for_device(port->dev, sg, 1, DMA_TO_DEVICE);
+		dma_sync_sg_for_device(port->dev, sg_tx, 1, DMA_TO_DEVICE);
 
 		atmel_port->desc_tx = desc;
 		desc->callback = atmel_complete_tx_dma;
@@ -890,7 +921,9 @@ static int atmel_prepare_tx_dma(struct uart_port *port)
 	/* Configure the slave DMA */
 	memset(&config, 0, sizeof(config));
 	config.direction = DMA_MEM_TO_DEV;
-	config.dst_addr_width = DMA_SLAVE_BUSWIDTH_1_BYTE;
+	config.dst_addr_width = (atmel_port->fifo_size) ?
+				DMA_SLAVE_BUSWIDTH_4_BYTES :
+				DMA_SLAVE_BUSWIDTH_1_BYTE;
 	config.dst_addr = port->mapbase + ATMEL_US_THR;
 	config.dst_maxburst = 1;
 

drivers/tty/serial/atmel_serial.c static int atmel_startup(struct uart_port *port) · line 1861

Quoted 6 lines
drivers/tty/serial/atmel_serial.c +3 -0
@@ -1823,6 +1856,9 @@ static int atmel_startup(struct uart_port *port)
 			    ATMEL_US_RXFCLR |
 			    ATMEL_US_TXFLCLR);
 
+		if (atmel_use_dma_tx(port))
+			txrdym = ATMEL_US_FOUR_DATA;
+

Ok, I see now why you used a variable for txrdym: no problem to keep it like you did then.

drivers/tty/serial/atmel_serial.c static int atmel_startup(struct uart_port *port) · line 1864

Quoted 4 lines
drivers/tty/serial/atmel_serial.c +0 -0
 		fmr = ATMEL_US_TXRDYM(txrdym) | ATMEL_US_RXRDYM(rxrdym);
 		if (atmel_port->rts_high &&
 		    atmel_port->rts_low)

Thanks!

Signature · 1 line
Nicolas Ferre