[net-next,v4,01/15] net: dsa: vsc73xx: use read_poll_timeout instead delay loop

Message ID 20240213220331.239031-2-paweldembicki@gmail.com
State New
Headers
Series net: dsa: vsc73xx: Make vsc73xx usable |

Commit Message

Pawel Dembicki Feb. 13, 2024, 10:03 p.m. UTC
  This commit switches delay loop to read_poll_timeout macro during
Arbiter empty check in adjust link function.

As Russel King suggested:

"This [change] avoids the issue that on the last iteration, the code reads
the register, test it, find the condition that's being waiting for is
false, _then_ waits and end up printing the error message - that last
wait is rather useless, and as the arbiter state isn't checked after
waiting, it could be that we had success during the last wait."

It also remove one short msleep delay.

Suggested-by: Russell King <linux@armlinux.org.uk>
Signed-off-by: Pawel Dembicki <paweldembicki@gmail.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
---
v4:
  - Resend patch
v3:
  - Add "Reviewed-by" to commit message only
v2:
  - introduced patch

 drivers/net/dsa/vitesse-vsc73xx-core.c | 25 +++++++++++--------------
 1 file changed, 11 insertions(+), 14 deletions(-)
  

Comments

Linus Walleij Feb. 13, 2024, 11:07 p.m. UTC | #1
On Tue, Feb 13, 2024 at 11:04 PM Pawel Dembicki <paweldembicki@gmail.com> wrote:


> This commit switches delay loop to read_poll_timeout macro during
> Arbiter empty check in adjust link function.
>
> As Russel King suggested:
>
> "This [change] avoids the issue that on the last iteration, the code reads
> the register, test it, find the condition that's being waiting for is
> false, _then_ waits and end up printing the error message - that last
> wait is rather useless, and as the arbiter state isn't checked after
> waiting, it could be that we had success during the last wait."
>
> It also remove one short msleep delay.
>
> Suggested-by: Russell King <linux@armlinux.org.uk>
> Signed-off-by: Pawel Dembicki <paweldembicki@gmail.com>
> Reviewed-by: Andrew Lunn <andrew@lunn.ch>

Reviewed-by: Linus Walleij <linus.walleij@linaro.org>

Yours,
Linus Walleij
  
Florian Fainelli Feb. 13, 2024, 11:16 p.m. UTC | #2
On 2/13/24 14:03, Pawel Dembicki wrote:
> This commit switches delay loop to read_poll_timeout macro during
> Arbiter empty check in adjust link function.
> 
> As Russel King suggested:
> 
> "This [change] avoids the issue that on the last iteration, the code reads
> the register, test it, find the condition that's being waiting for is
> false, _then_ waits and end up printing the error message - that last
> wait is rather useless, and as the arbiter state isn't checked after
> waiting, it could be that we had success during the last wait."
> 
> It also remove one short msleep delay.
> 
> Suggested-by: Russell King <linux@armlinux.org.uk>
> Signed-off-by: Pawel Dembicki <paweldembicki@gmail.com>
> Reviewed-by: Andrew Lunn <andrew@lunn.ch>

Reviewed-by: Florian Fainelli <florian.fainelli@broadcom.com>
  
Vladimir Oltean Feb. 14, 2024, 11:46 p.m. UTC | #3
On Tue, Feb 13, 2024 at 11:03:14PM +0100, Pawel Dembicki wrote:
> This commit switches delay loop to read_poll_timeout macro during
> Arbiter empty check in adjust link function.

Replace "This commit does X" with imperative mood: "Switch the delay
loop during the Arbiter empty check from vsc73xx_adjust_link() to use
read_poll_timeout(). Functionally, one msleep() call is eliminated at
the end of the loop, in the timeout case".

> 
> As Russel King suggested:

s/Russel/Russell/

> 
> "This [change] avoids the issue that on the last iteration, the code reads
> the register, test it, find the condition that's being waiting for is

s/test/tests/
s/find/finds/

> false, _then_ waits and end up printing the error message - that last
> wait is rather useless, and as the arbiter state isn't checked after
> waiting, it could be that we had success during the last wait."
> 
> It also remove one short msleep delay.

Apart from the fact that there's a grammatical mistake in this phrase
("it remove" -> "it removes"), it's also a bit redundant, since
Russell's explanation above implies this is what would happen. Anyway,
I've suggested a replacement for it in the first paragraph, the one
describing the change.

> Suggested-by: Russell King <linux@armlinux.org.uk>
> Signed-off-by: Pawel Dembicki <paweldembicki@gmail.com>
> Reviewed-by: Andrew Lunn <andrew@lunn.ch>
> ---
> v4:
>   - Resend patch
> v3:
>   - Add "Reviewed-by" to commit message only
> v2:
>   - introduced patch
> 
>  drivers/net/dsa/vitesse-vsc73xx-core.c | 25 +++++++++++--------------
>  1 file changed, 11 insertions(+), 14 deletions(-)
> 
> diff --git a/drivers/net/dsa/vitesse-vsc73xx-core.c b/drivers/net/dsa/vitesse-vsc73xx-core.c
> index ae70eac3be28..8b2219404601 100644
> --- a/drivers/net/dsa/vitesse-vsc73xx-core.c
> +++ b/drivers/net/dsa/vitesse-vsc73xx-core.c
> @@ -779,7 +779,7 @@ static void vsc73xx_adjust_link(struct dsa_switch *ds, int port,
>  	 * after a PHY or the CPU port comes up or down.
>  	 */
>  	if (!phydev->link) {
> -		int maxloop = 10;
> +		int ret, err;
>  
>  		dev_dbg(vsc->dev, "port %d: went down\n",
>  			port);
> @@ -794,19 +794,16 @@ static void vsc73xx_adjust_link(struct dsa_switch *ds, int port,
>  				    VSC73XX_ARBDISC, BIT(port), BIT(port));
>  
>  		/* Wait until queue is empty */
> -		vsc73xx_read(vsc, VSC73XX_BLOCK_ARBITER, 0,
> -			     VSC73XX_ARBEMPTY, &val);
> -		while (!(val & BIT(port))) {
> -			msleep(1);
> -			vsc73xx_read(vsc, VSC73XX_BLOCK_ARBITER, 0,
> -				     VSC73XX_ARBEMPTY, &val);
> -			if (--maxloop == 0) {
> -				dev_err(vsc->dev,
> -					"timeout waiting for block arbiter\n");
> -				/* Continue anyway */
> -				break;
> -			}
> -		}
> +		ret = read_poll_timeout(vsc73xx_read, err,
> +					err < 0 || (val & BIT(port)),
> +					1000, 10000, false,

Some #defines for 1000 and 10000 please (VSC73XX_ARBITER_SLEEP_US,
VSC73XX_ARBITER_TIMEOUT_US)?

> +					vsc, VSC73XX_BLOCK_ARBITER, 0,
> +					VSC73XX_ARBEMPTY, &val);
> +		if (ret)
> +			dev_err(vsc->dev,
> +				"timeout waiting for block arbiter\n");
> +		else if (err < 0)
> +			dev_err(vsc->dev, "error reading arbiter\n");
>  
>  		/* Put this port into reset */
>  		vsc73xx_write(vsc, VSC73XX_BLOCK_MAC, port, VSC73XX_MAC_CFG,
> -- 
> 2.34.1
>
  

Patch

diff --git a/drivers/net/dsa/vitesse-vsc73xx-core.c b/drivers/net/dsa/vitesse-vsc73xx-core.c
index ae70eac3be28..8b2219404601 100644
--- a/drivers/net/dsa/vitesse-vsc73xx-core.c
+++ b/drivers/net/dsa/vitesse-vsc73xx-core.c
@@ -779,7 +779,7 @@  static void vsc73xx_adjust_link(struct dsa_switch *ds, int port,
 	 * after a PHY or the CPU port comes up or down.
 	 */
 	if (!phydev->link) {
-		int maxloop = 10;
+		int ret, err;
 
 		dev_dbg(vsc->dev, "port %d: went down\n",
 			port);
@@ -794,19 +794,16 @@  static void vsc73xx_adjust_link(struct dsa_switch *ds, int port,
 				    VSC73XX_ARBDISC, BIT(port), BIT(port));
 
 		/* Wait until queue is empty */
-		vsc73xx_read(vsc, VSC73XX_BLOCK_ARBITER, 0,
-			     VSC73XX_ARBEMPTY, &val);
-		while (!(val & BIT(port))) {
-			msleep(1);
-			vsc73xx_read(vsc, VSC73XX_BLOCK_ARBITER, 0,
-				     VSC73XX_ARBEMPTY, &val);
-			if (--maxloop == 0) {
-				dev_err(vsc->dev,
-					"timeout waiting for block arbiter\n");
-				/* Continue anyway */
-				break;
-			}
-		}
+		ret = read_poll_timeout(vsc73xx_read, err,
+					err < 0 || (val & BIT(port)),
+					1000, 10000, false,
+					vsc, VSC73XX_BLOCK_ARBITER, 0,
+					VSC73XX_ARBEMPTY, &val);
+		if (ret)
+			dev_err(vsc->dev,
+				"timeout waiting for block arbiter\n");
+		else if (err < 0)
+			dev_err(vsc->dev, "error reading arbiter\n");
 
 		/* Put this port into reset */
 		vsc73xx_write(vsc, VSC73XX_BLOCK_MAC, port, VSC73XX_MAC_CFG,