2024-05-28 21:33:42

by Tristram.Ha

[permalink] [raw]
Subject: [PATCH net] net: dsa: microchip: fix KSZ9477 set_ageing_time function

From: Tristram Ha <[email protected]>

The aging count is not a simple 11-bit value but comprises a 3-bit
multiplier and an 8-bit second count. The code tries to find a set of
values with result close to the specifying value.

Note LAN937X has similar operation but provides an option to use
millisecond instead of second so there will be a separate fix in the
future.

Fixes: 2c119d9982b1 ("net: dsa: microchip: add the support for set_ageing_time")
Signed-off-by: Tristram Ha <[email protected]>
---
drivers/net/dsa/microchip/ksz9477.c | 64 +++++++++++++++++++++----
drivers/net/dsa/microchip/ksz9477_reg.h | 1 -
2 files changed, 54 insertions(+), 11 deletions(-)

diff --git a/drivers/net/dsa/microchip/ksz9477.c b/drivers/net/dsa/microchip/ksz9477.c
index f8ad7833f5d9..1af11aee3119 100644
--- a/drivers/net/dsa/microchip/ksz9477.c
+++ b/drivers/net/dsa/microchip/ksz9477.c
@@ -1099,26 +1099,70 @@ void ksz9477_get_caps(struct ksz_device *dev, int port,
int ksz9477_set_ageing_time(struct ksz_device *dev, unsigned int msecs)
{
u32 secs = msecs / 1000;
+ u8 first, last, mult, i;
+ int min, ret;
+ int diff[8];
u8 value;
u8 data;
- int ret;

- value = FIELD_GET(SW_AGE_PERIOD_7_0_M, secs);
+ /* The aging timer comprises a 3-bit multiplier and an 8-bit second
+ * value. Either of them cannot be zero. The maximum timer is then
+ * 7 * 255 = 1785.
+ */
+ if (!secs)
+ secs = 1;

- ret = ksz_write8(dev, REG_SW_LUE_CTRL_3, value);
+ ret = ksz_read8(dev, REG_SW_LUE_CTRL_0, &value);
if (ret < 0)
return ret;

- data = FIELD_GET(SW_AGE_PERIOD_10_8_M, secs);
+ /* Check whether there is need to update the multiplier. */
+ mult = FIELD_GET(SW_AGE_CNT_M, value);
+ if (mult > 0) {
+ /* Try to use the same multiplier already in the register. */
+ min = secs / mult;
+ if (min <= 0xff && min * mult == secs)
+ return ksz_write8(dev, REG_SW_LUE_CTRL_3, min);
+ }

- ret = ksz_read8(dev, REG_SW_LUE_CTRL_0, &value);
- if (ret < 0)
- return ret;
+ /* Return error if too large. */
+ if (secs > 7 * 0xff)
+ return -EINVAL;
+
+ /* Find out which combination of multiplier * value results in a timer
+ * value close to the specified timer value.
+ */
+ first = (secs + 0xfe) / 0xff;
+ for (i = first; i <= 7; i++) {
+ min = secs / i;
+ diff[i] = secs - i * min;
+ if (!diff[i]) {
+ i++;
+ break;
+ }
+ }
+
+ last = i;
+ min = 0xff;
+ for (i = last - 1; i >= first; i--) {
+ if (diff[i] < min) {
+ data = i;
+ min = diff[i];
+ }
+ if (!min)
+ break;
+ }

- value &= ~SW_AGE_CNT_M;
- value |= FIELD_PREP(SW_AGE_CNT_M, data);
+ if (mult != data) {
+ value &= ~SW_AGE_CNT_M;
+ value |= FIELD_PREP(SW_AGE_CNT_M, data);
+ ret = ksz_write8(dev, REG_SW_LUE_CTRL_0, value);
+ if (ret)
+ return ret;
+ }

- return ksz_write8(dev, REG_SW_LUE_CTRL_0, value);
+ value = secs / data;
+ return ksz_write8(dev, REG_SW_LUE_CTRL_3, value);
}

void ksz9477_port_queue_split(struct ksz_device *dev, int port)
diff --git a/drivers/net/dsa/microchip/ksz9477_reg.h b/drivers/net/dsa/microchip/ksz9477_reg.h
index f3a205ee483f..0c55a540f20d 100644
--- a/drivers/net/dsa/microchip/ksz9477_reg.h
+++ b/drivers/net/dsa/microchip/ksz9477_reg.h
@@ -171,7 +171,6 @@
#define SW_DROP_INVALID_VID BIT(6)
#define SW_AGE_CNT_M GENMASK(5, 3)
#define SW_AGE_CNT_S 3
-#define SW_AGE_PERIOD_10_8_M GENMASK(10, 8)
#define SW_RESV_MCAST_ENABLE BIT(2)
#define SW_HASH_OPTION_M 0x03
#define SW_HASH_OPTION_CRC 1
--
2.34.1



2024-05-30 09:13:21

by Paolo Abeni

[permalink] [raw]
Subject: Re: [PATCH net] net: dsa: microchip: fix KSZ9477 set_ageing_time function

On Tue, 2024-05-28 at 14:36 -0700, [email protected] wrote:
> From: Tristram Ha <[email protected]>
>
> The aging count is not a simple 11-bit value but comprises a 3-bit
> multiplier and an 8-bit second count. The code tries to find a set of
> values with result close to the specifying value.
>
> Note LAN937X has similar operation but provides an option to use
> millisecond instead of second so there will be a separate fix in the
> future.
>
> Fixes: 2c119d9982b1 ("net: dsa: microchip: add the support for set_ageing_time")
> Signed-off-by: Tristram Ha <[email protected]>
> ---
> drivers/net/dsa/microchip/ksz9477.c | 64 +++++++++++++++++++++----
> drivers/net/dsa/microchip/ksz9477_reg.h | 1 -
> 2 files changed, 54 insertions(+), 11 deletions(-)
>
> diff --git a/drivers/net/dsa/microchip/ksz9477.c b/drivers/net/dsa/microchip/ksz9477.c
> index f8ad7833f5d9..1af11aee3119 100644
> --- a/drivers/net/dsa/microchip/ksz9477.c
> +++ b/drivers/net/dsa/microchip/ksz9477.c
> @@ -1099,26 +1099,70 @@ void ksz9477_get_caps(struct ksz_device *dev, int port,
> int ksz9477_set_ageing_time(struct ksz_device *dev, unsigned int msecs)
> {
> u32 secs = msecs / 1000;
> + u8 first, last, mult, i;
> + int min, ret;
> + int diff[8];
> u8 value;
> u8 data;
> - int ret;
>
> - value = FIELD_GET(SW_AGE_PERIOD_7_0_M, secs);
> + /* The aging timer comprises a 3-bit multiplier and an 8-bit second
> + * value. Either of them cannot be zero. The maximum timer is then
> + * 7 * 255 = 1785.
> + */
> + if (!secs)
> + secs = 1;
>
> - ret = ksz_write8(dev, REG_SW_LUE_CTRL_3, value);
> + ret = ksz_read8(dev, REG_SW_LUE_CTRL_0, &value);
> if (ret < 0)
> return ret;
>
> - data = FIELD_GET(SW_AGE_PERIOD_10_8_M, secs);
> + /* Check whether there is need to update the multiplier. */
> + mult = FIELD_GET(SW_AGE_CNT_M, value);
> + if (mult > 0) {
> + /* Try to use the same multiplier already in the register. */
> + min = secs / mult;
> + if (min <= 0xff && min * mult == secs)
> + return ksz_write8(dev, REG_SW_LUE_CTRL_3, min);
> + }
>
> - ret = ksz_read8(dev, REG_SW_LUE_CTRL_0, &value);
> - if (ret < 0)
> - return ret;
> + /* Return error if too large. */
> + if (secs > 7 * 0xff)
> + return -EINVAL;
> +
> + /* Find out which combination of multiplier * value results in a timer
> + * value close to the specified timer value.
> + */
> + first = (secs + 0xfe) / 0xff;
> + for (i = first; i <= 7; i++) {
> + min = secs / i;
> + diff[i] = secs - i * min;
> + if (!diff[i]) {
> + i++;
> + break;
> + }
> + }
> +
> + last = i;
> + min = 0xff;
> + for (i = last - 1; i >= first; i--) {
> + if (diff[i] < min) {
> + data = i;
> + min = diff[i];
> + }
> + if (!min)
> + break;
> + }

Is the additional accuracy worthy the added complexity WRT:

mult = DIV_ROUND_UP(secs, 0xff);

?

Paolo


2024-05-30 23:13:46

by Tristram.Ha

[permalink] [raw]
Subject: RE: [PATCH net] net: dsa: microchip: fix KSZ9477 set_ageing_time function

> Subject: Re: [PATCH net] net: dsa: microchip: fix KSZ9477 set_ageing_time function
>
> EXTERNAL EMAIL: Do not click links or open attachments unless you know the content
> is safe
>
> On Tue, 2024-05-28 at 14:36 -0700, [email protected] wrote:
> > From: Tristram Ha <[email protected]>
> >
> > The aging count is not a simple 11-bit value but comprises a 3-bit
> > multiplier and an 8-bit second count. The code tries to find a set of
> > values with result close to the specifying value.
> >
> > Note LAN937X has similar operation but provides an option to use
> > millisecond instead of second so there will be a separate fix in the
> > future.
> >
> > Fixes: 2c119d9982b1 ("net: dsa: microchip: add the support for set_ageing_time")
> > Signed-off-by: Tristram Ha <[email protected]>
> > ---
> > drivers/net/dsa/microchip/ksz9477.c | 64 +++++++++++++++++++++----
> > drivers/net/dsa/microchip/ksz9477_reg.h | 1 -
> > 2 files changed, 54 insertions(+), 11 deletions(-)
> >
> > diff --git a/drivers/net/dsa/microchip/ksz9477.c
> b/drivers/net/dsa/microchip/ksz9477.c
> > index f8ad7833f5d9..1af11aee3119 100644
> > --- a/drivers/net/dsa/microchip/ksz9477.c
> > +++ b/drivers/net/dsa/microchip/ksz9477.c
> > @@ -1099,26 +1099,70 @@ void ksz9477_get_caps(struct ksz_device *dev, int
> port,
> > int ksz9477_set_ageing_time(struct ksz_device *dev, unsigned int msecs)
> > {
> > u32 secs = msecs / 1000;
> > + u8 first, last, mult, i;
> > + int min, ret;
> > + int diff[8];
> > u8 value;
> > u8 data;
> > - int ret;
> >
> > - value = FIELD_GET(SW_AGE_PERIOD_7_0_M, secs);
> > + /* The aging timer comprises a 3-bit multiplier and an 8-bit second
> > + * value. Either of them cannot be zero. The maximum timer is then
> > + * 7 * 255 = 1785.
> > + */
> > + if (!secs)
> > + secs = 1;
> >
> > - ret = ksz_write8(dev, REG_SW_LUE_CTRL_3, value);
> > + ret = ksz_read8(dev, REG_SW_LUE_CTRL_0, &value);
> > if (ret < 0)
> > return ret;
> >
> > - data = FIELD_GET(SW_AGE_PERIOD_10_8_M, secs);
> > + /* Check whether there is need to update the multiplier. */
> > + mult = FIELD_GET(SW_AGE_CNT_M, value);
> > + if (mult > 0) {
> > + /* Try to use the same multiplier already in the register. */
> > + min = secs / mult;
> > + if (min <= 0xff && min * mult == secs)
> > + return ksz_write8(dev, REG_SW_LUE_CTRL_3, min);
> > + }
> >
> > - ret = ksz_read8(dev, REG_SW_LUE_CTRL_0, &value);
> > - if (ret < 0)
> > - return ret;
> > + /* Return error if too large. */
> > + if (secs > 7 * 0xff)
> > + return -EINVAL;
> > +
> > + /* Find out which combination of multiplier * value results in a timer
> > + * value close to the specified timer value.
> > + */
> > + first = (secs + 0xfe) / 0xff;
> > + for (i = first; i <= 7; i++) {
> > + min = secs / i;
> > + diff[i] = secs - i * min;
> > + if (!diff[i]) {
> > + i++;
> > + break;
> > + }
> > + }
> > +
> > + last = i;
> > + min = 0xff;
> > + for (i = last - 1; i >= first; i--) {
> > + if (diff[i] < min) {
> > + data = i;
> > + min = diff[i];
> > + }
> > + if (!min)
> > + break;
> > + }
>
> Is the additional accuracy worthy the added complexity WRT:
>
> mult = DIV_ROUND_UP(secs, 0xff);
>
> ?

I do not know much accuracy is expected of this function. I do not know
whether users can easily specify the amount, but the default is 3 seconds
which is the same as the hardware default where the multiplier is 4 and
the count is 75. So most of the time the rest of the code will not be
executed.

2024-05-31 00:22:03

by Andrew Lunn

[permalink] [raw]
Subject: Re: [PATCH net] net: dsa: microchip: fix KSZ9477 set_ageing_time function

> > Is the additional accuracy worthy the added complexity WRT:
> >
> > mult = DIV_ROUND_UP(secs, 0xff);
> >
> > ?
>
> I do not know much accuracy is expected of this function. I do not know
> whether users can easily specify the amount, but the default is 3 seconds
> which is the same as the hardware default where the multiplier is 4 and
> the count is 75. So most of the time the rest of the code will not be
> executed.

Are you sure it is 3 seconds?

#define BR_DEFAULT_AGEING_TIME (300 * HZ)

Fast ageing, after a topology change, is i think 15 seconds.

Andrew