[PATCH V2 0/7] can: mcp251x: fix and optimize driver

[PATCH V2 0/7] can: mcp251x: fix and optimize driver

[Marc Kleine-Budde]

Moin,

this series of patches improves the mcp251x driver. It first fixes the
local_softirq_pending problem. Then the amount of SPI transfers is reduced
in order to optimise the driver.

This series has been tested with a mcp2515 on i.MX35.

Changes since V1:
- Fix broken encoding in S-o-b

Please review, test and consider to apply.

regards, Marc

---

The following changes since commit cd2638a86c7b90e77ce623c09de2a26177f2a5c1:
  Carolyn Wyborny (1):
        igb: add check for fiber/serdes devices to igb_set_spd_dplx;

are available in the git repository at:

  git://git.pengutronix.de/git/mkl/linux-2.6.git can/mcp251x-for-net-next

Marc Kleine-Budde (4):
      can: mcp251x: fix NOHZ local_softirq_pending 08 warning
      can: mcp251x: write intf only when needed
      can: mcp251x: define helper functions mcp251x_is_2510, mcp251x_is_2515
      can: mcp251x: optimize 2515, rx int gets cleared automatically

Sascha Hauer (3):
      can: mcp251x: increase rx_errors on overflow, not only rx_over_errors
      can: mcp251x: allow to read two registers in one spi transfer
      can: mcp251x: read-modify-write eflag only when needed

 drivers/net/can/mcp251x.c |   77 +++++++++++++++++++++++++++++++++++----------
 1 files changed, 60 insertions(+), 17 deletions(-)

[PATCH 1/7] can: mcp251x: fix NOHZ local_softirq_pending 08 warning

[Marc Kleine-Budde]

This patch replaces netif_rx() with netif_rx_ni() which has to be used
from the threaded interrupt i.e. process context context.

Thanks to Christian Pellegrin for pointing at the right fix:
481a8199142c050b72bff8a1956a49fd0a75bbe0 by Oliver Hartkopp.

Signed-off-by: Marc Kleine-Budde <mkl-bIcnvbaLZ9MEGnE8C9+IrQ@public.gmane.org>
---
 drivers/net/can/mcp251x.c |    4 ++--
 1 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/net/can/mcp251x.c b/drivers/net/can/mcp251x.c
index b11a0cb..c06e023 100644
--- a/drivers/net/can/mcp251x.c
+++ b/drivers/net/can/mcp251x.c
@@ -451,7 +451,7 @@ static void mcp251x_hw_rx(struct spi_device *spi, int buf_idx)
 
 	priv->net->stats.rx_packets++;
 	priv->net->stats.rx_bytes += frame->can_dlc;
-	netif_rx(skb);
+	netif_rx_ni(skb);
 }
 
 static void mcp251x_hw_sleep(struct spi_device *spi)
@@ -676,7 +676,7 @@ static void mcp251x_error_skb(struct net_device *net, int can_id, int data1)
 	if (skb) {
 		frame->can_id = can_id;
 		frame->data[1] = data1;
-		netif_rx(skb);
+		netif_rx_ni(skb);
 	} else {
 		dev_err(&net->dev,
 			"cannot allocate error skb\n");
-- 
1.7.0.4

[PATCH 2/7] can: mcp251x: increase rx_errors on overflow, not only rx_over_errors

[Marc Kleine-Budde]

From: Sascha Hauer <s.hauer-bIcnvbaLZ9MEGnE8C9+IrQ@public.gmane.org>

Signed-off-by: Sascha Hauer <s.hauer-bIcnvbaLZ9MEGnE8C9+IrQ@public.gmane.org>
Signed-off-by: Marc Kleine-Budde <mkl-bIcnvbaLZ9MEGnE8C9+IrQ@public.gmane.org>
---
 drivers/net/can/mcp251x.c |    8 ++++++--
 1 files changed, 6 insertions(+), 2 deletions(-)

diff --git a/drivers/net/can/mcp251x.c b/drivers/net/can/mcp251x.c
index c06e023..fdea752 100644
--- a/drivers/net/can/mcp251x.c
+++ b/drivers/net/can/mcp251x.c
@@ -816,10 +816,14 @@ static irqreturn_t mcp251x_can_ist(int irq, void *dev_id)
 		if (intf & CANINTF_ERRIF) {
 			/* Handle overflow counters */
 			if (eflag & (EFLG_RX0OVR | EFLG_RX1OVR)) {
-				if (eflag & EFLG_RX0OVR)
+				if (eflag & EFLG_RX0OVR) {
 					net->stats.rx_over_errors++;
-				if (eflag & EFLG_RX1OVR)
+					net->stats.rx_errors++;
+				}
+				if (eflag & EFLG_RX1OVR) {
 					net->stats.rx_over_errors++;
+					net->stats.rx_errors++;
+				}
 				can_id |= CAN_ERR_CRTL;
 				data1 |= CAN_ERR_CRTL_RX_OVERFLOW;
 			}
-- 
1.7.0.4

[PATCH 3/7] can: mcp251x: allow to read two registers in one spi transfer

[Marc Kleine-Budde]

From: Sascha Hauer <s.hauer@pengutronix.de>

This patch bases on work done earlier by David Jander.

Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Acked-by: David Jander <david@protonic.nl>
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
---
 drivers/net/can/mcp251x.c |   20 +++++++++++++++++---
 1 files changed, 17 insertions(+), 3 deletions(-)

diff --git a/drivers/net/can/mcp251x.c b/drivers/net/can/mcp251x.c
index fdea752..9b3466a 100644
--- a/drivers/net/can/mcp251x.c
+++ b/drivers/net/can/mcp251x.c
@@ -319,6 +319,20 @@ static u8 mcp251x_read_reg(struct spi_device *spi, uint8_t reg)
 	return val;
 }
 
+static void mcp251x_read_2regs(struct spi_device *spi, uint8_t reg,
+		uint8_t *v1, uint8_t *v2)
+{
+	struct mcp251x_priv *priv = dev_get_drvdata(&spi->dev);
+
+	priv->spi_tx_buf[0] = INSTRUCTION_READ;
+	priv->spi_tx_buf[1] = reg;
+
+	mcp251x_spi_trans(spi, 4);
+
+	*v1 = priv->spi_rx_buf[2];
+	*v2 = priv->spi_rx_buf[3];
+}
+
 static void mcp251x_write_reg(struct spi_device *spi, u8 reg, uint8_t val)
 {
 	struct mcp251x_priv *priv = dev_get_drvdata(&spi->dev);
@@ -754,10 +768,11 @@ static irqreturn_t mcp251x_can_ist(int irq, void *dev_id)
 	mutex_lock(&priv->mcp_lock);
 	while (!priv->force_quit) {
 		enum can_state new_state;
-		u8 intf = mcp251x_read_reg(spi, CANINTF);
-		u8 eflag;
+		u8 intf, eflag;
 		int can_id = 0, data1 = 0;
 
+		mcp251x_read_2regs(spi, CANINTF, &intf, &eflag);
+
 		if (intf & CANINTF_RX0IF) {
 			mcp251x_hw_rx(spi, 0);
 			/* Free one buffer ASAP */
@@ -770,7 +785,6 @@ static irqreturn_t mcp251x_can_ist(int irq, void *dev_id)
 
 		mcp251x_write_bits(spi, CANINTF, intf, 0x00);
 
-		eflag = mcp251x_read_reg(spi, EFLG);
 		mcp251x_write_reg(spi, EFLG, 0x00);
 
 		/* Update can state */
-- 
1.7.0.4

_______________________________________________
Socketcan-core mailing list
Socketcan-core@lists.berlios.de
https://lists.berlios.de/mailman/listinfo/socketcan-core

[PATCH 4/7] can: mcp251x: read-modify-write eflag only when needed

[Marc Kleine-Budde]

From: Sascha Hauer <s.hauer-bIcnvbaLZ9MEGnE8C9+IrQ@public.gmane.org>

Use read-modify-write instead of a simple write to change the register
contents, to close existing the race window between the original manual
read and write.

Signed-off-by: Sascha Hauer <s.hauer-bIcnvbaLZ9MEGnE8C9+IrQ@public.gmane.org>
Signed-off-by: Marc Kleine-Budde <mkl-bIcnvbaLZ9MEGnE8C9+IrQ@public.gmane.org>
---
 drivers/net/can/mcp251x.c |    3 ++-
 1 files changed, 2 insertions(+), 1 deletions(-)

diff --git a/drivers/net/can/mcp251x.c b/drivers/net/can/mcp251x.c
index 9b3466a..7e2f951 100644
--- a/drivers/net/can/mcp251x.c
+++ b/drivers/net/can/mcp251x.c
@@ -785,7 +785,8 @@ static irqreturn_t mcp251x_can_ist(int irq, void *dev_id)
 
 		mcp251x_write_bits(spi, CANINTF, intf, 0x00);
 
-		mcp251x_write_reg(spi, EFLG, 0x00);
+		if (eflag)
+			mcp251x_write_bits(spi, EFLG, eflag, 0x00);
 
 		/* Update can state */
 		if (eflag & EFLG_TXBO) {
-- 
1.7.0.4

[PATCH 5/7] can: mcp251x: write intf only when needed

[Marc Kleine-Budde]

This patch introduces a variable "clear_intf" that hold the bits that
should be cleared. Only read-modify-write register if "clear_intf"
is set.

Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
---
 drivers/net/can/mcp251x.c |   15 +++++++++++++--
 1 files changed, 13 insertions(+), 2 deletions(-)

diff --git a/drivers/net/can/mcp251x.c b/drivers/net/can/mcp251x.c
index 7e2f951..f5e2edd 100644
--- a/drivers/net/can/mcp251x.c
+++ b/drivers/net/can/mcp251x.c
@@ -125,6 +125,8 @@
 #  define CANINTF_TX0IF 0x04
 #  define CANINTF_RX1IF 0x02
 #  define CANINTF_RX0IF 0x01
+#  define CANINTF_ERR_TX \
+	(CANINTF_ERRIF | CANINTF_TX2IF | CANINTF_TX1IF | CANINTF_TX0IF)
 #define EFLG	      0x2d
 #  define EFLG_EWARN	0x01
 #  define EFLG_RXWAR	0x02
@@ -769,10 +771,12 @@ static irqreturn_t mcp251x_can_ist(int irq, void *dev_id)
 	while (!priv->force_quit) {
 		enum can_state new_state;
 		u8 intf, eflag;
+		u8 clear_intf = 0;
 		int can_id = 0, data1 = 0;
 
 		mcp251x_read_2regs(spi, CANINTF, &intf, &eflag);
 
+		/* receive buffer 0 */
 		if (intf & CANINTF_RX0IF) {
 			mcp251x_hw_rx(spi, 0);
 			/* Free one buffer ASAP */
@@ -780,10 +784,17 @@ static irqreturn_t mcp251x_can_ist(int irq, void *dev_id)
 					   0x00);
 		}
 
-		if (intf & CANINTF_RX1IF)
+		/* receive buffer 1 */
+		if (intf & CANINTF_RX1IF) {
 			mcp251x_hw_rx(spi, 1);
+			clear_intf |= CANINTF_RX1IF;
+		}
 
-		mcp251x_write_bits(spi, CANINTF, intf, 0x00);
+		/* any error or tx interrupt we need to clear? */
+		if (intf & CANINTF_ERR_TX)
+			clear_intf |= intf & CANINTF_ERR_TX;
+		if (clear_intf)
+			mcp251x_write_bits(spi, CANINTF, clear_intf, 0x00);
 
 		if (eflag)
 			mcp251x_write_bits(spi, EFLG, eflag, 0x00);
-- 
1.7.0.4

[PATCH 6/7] can: mcp251x: define helper functions mcp251x_is_2510, mcp251x_is_2515

[Marc Kleine-Budde]

Signed-off-by: Marc Kleine-Budde <mkl-bIcnvbaLZ9MEGnE8C9+IrQ@public.gmane.org>
---
 drivers/net/can/mcp251x.c |   16 ++++++++++++----
 1 files changed, 12 insertions(+), 4 deletions(-)

diff --git a/drivers/net/can/mcp251x.c b/drivers/net/can/mcp251x.c
index f5e2edd..28281f9 100644
--- a/drivers/net/can/mcp251x.c
+++ b/drivers/net/can/mcp251x.c
@@ -252,6 +252,16 @@ struct mcp251x_priv {
 	int restart_tx;
 };
 
+#define MCP251X_IS(_model) \
+static inline int mcp251x_is_##_model(struct spi_device *spi) \
+{ \
+	struct mcp251x_platform_data *pdata = spi->dev.platform_data; \
+	return pdata->model == CAN_MCP251X_MCP##_model; \
+}
+
+MCP251X_IS(2510);
+MCP251X_IS(2515);
+
 static void mcp251x_clean(struct net_device *net)
 {
 	struct mcp251x_priv *priv = netdev_priv(net);
@@ -362,10 +372,9 @@ static void mcp251x_write_bits(struct spi_device *spi, u8 reg,
 static void mcp251x_hw_tx_frame(struct spi_device *spi, u8 *buf,
 				int len, int tx_buf_idx)
 {
-	struct mcp251x_platform_data *pdata = spi->dev.platform_data;
 	struct mcp251x_priv *priv = dev_get_drvdata(&spi->dev);
 
-	if (pdata->model == CAN_MCP251X_MCP2510) {
+	if (mcp251x_is_2510(spi)) {
 		int i;
 
 		for (i = 1; i < TXBDAT_OFF + len; i++)
@@ -408,9 +417,8 @@ static void mcp251x_hw_rx_frame(struct spi_device *spi, u8 *buf,
 				int buf_idx)
 {
 	struct mcp251x_priv *priv = dev_get_drvdata(&spi->dev);
-	struct mcp251x_platform_data *pdata = spi->dev.platform_data;
 
-	if (pdata->model == CAN_MCP251X_MCP2510) {
+	if (mcp251x_is_2510(spi)) {
 		int i, len;
 
 		for (i = 1; i < RXBDAT_OFF; i++)
-- 
1.7.0.4

[PATCH 7/7] can: mcp251x: optimize 2515, rx int gets cleared automatically

[Marc Kleine-Budde]

Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
---
 drivers/net/can/mcp251x.c |   13 +++++++++----
 1 files changed, 9 insertions(+), 4 deletions(-)

diff --git a/drivers/net/can/mcp251x.c b/drivers/net/can/mcp251x.c
index 28281f9..a42c3fa 100644
--- a/drivers/net/can/mcp251x.c
+++ b/drivers/net/can/mcp251x.c
@@ -787,15 +787,20 @@ static irqreturn_t mcp251x_can_ist(int irq, void *dev_id)
 		/* receive buffer 0 */
 		if (intf & CANINTF_RX0IF) {
 			mcp251x_hw_rx(spi, 0);
-			/* Free one buffer ASAP */
-			mcp251x_write_bits(spi, CANINTF, intf & CANINTF_RX0IF,
-					   0x00);
+			/*
+			 * Free one buffer ASAP
+			 * (The MCP2515 does this automatically.)
+			 */
+			if (mcp251x_is_2510(spi))
+				mcp251x_write_bits(spi, CANINTF, CANINTF_RX0IF, 0x00);
 		}
 
 		/* receive buffer 1 */
 		if (intf & CANINTF_RX1IF) {
 			mcp251x_hw_rx(spi, 1);
-			clear_intf |= CANINTF_RX1IF;
+			/* the MCP2515 does this automatically */
+			if (mcp251x_is_2510(spi))
+				clear_intf |= CANINTF_RX1IF;
 		}
 
 		/* any error or tx interrupt we need to clear? */
-- 
1.7.0.4

Re: [PATCH V2 0/7] can: mcp251x: fix and optimize driver

[Wolfgang Grandegger]

On 10/15/2010 12:49 PM, Marc Kleine-Budde wrote:
> Moin,
> 
> this series of patches improves the mcp251x driver. It first fixes the
> local_softirq_pending problem. Then the amount of SPI transfers is reduced
> in order to optimise the driver.
> 
> This series has been tested with a mcp2515 on i.MX35.
> 
> Changes since V1:
> - Fix broken encoding in S-o-b
> 
> Please review, test and consider to apply.

You can add to all patches my

Acked-by: Wolfgang Grandegger <wg-5Yr1BZd7O62+XT7JhA+gdA@public.gmane.org>

Wolfgang.

Re: [PATCH V2 0/7] can: mcp251x: fix and optimize driver

[Marc Kleine-Budde]

[-- Attachment #1.1: Type: text/plain, Size: 510 bytes --]

On 10/17/2010 04:29 PM, Wolfgang Grandegger wrote:
> Acked-by: Wolfgang Grandegger <wg-5Yr1BZd7O62+XT7JhA+gdA@public.gmane.org>

tnx, I'm going to repost the series with you Ack and ask David to pull
it to net-next-2.6.

Marc

-- 
Pengutronix e.K.                  | Marc Kleine-Budde           |
Industrial Linux Solutions        | Phone: +49-231-2826-924     |
Vertretung West/Dortmund          | Fax:   +49-5121-206917-5555 |
Amtsgericht Hildesheim, HRA 2686  | http://www.pengutronix.de   |


[-- Attachment #1.2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 262 bytes --]

[-- Attachment #2: Type: text/plain, Size: 188 bytes --]

_______________________________________________
Socketcan-core mailing list
Socketcan-core-0fE9KPoRgkgATYTw5x5z8w@public.gmane.org
https://lists.berlios.de/mailman/listinfo/socketcan-core