Re: [PATCH stable 4.9 v2] net: dsa: bcm_sf2: force pause link settings

[Florian Fainelli <f.fainelli@gmail.com>]

On 7/6/22 13:31, Vladimir Oltean wrote:
> Hi Florian,
> 
> On Wed, Jul 06, 2022 at 12:24:54PM -0700, Florian Fainelli wrote:
>> From: Doug Berger <opendmb@gmail.com>
>>
>> commit 7c97bc0128b2eecc703106112679a69d446d1a12 upstream
>>
>> The pause settings reported by the PHY should also be applied to the
>> GMII port status override otherwise the switch will not generate pause
>> frames towards the link partner despite the advertisement saying
>> otherwise.
>>
>> Fixes: 246d7f773c13 ("net: dsa: add Broadcom SF2 switch driver")
>> Signed-off-by: Doug Berger <opendmb@gmail.com>
>> Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
>> Link: https://lore.kernel.org/r/20220623030204.1966851-1-f.fainelli@gmail.com
>> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
>> ---
>> Changes in v2:
>>
>> - use both local and remote advertisement to determine when to apply
>>    flow control settings
>>
>>   drivers/net/dsa/bcm_sf2.c | 16 ++++++++++++++++
>>   1 file changed, 16 insertions(+)
>>
>> diff --git a/drivers/net/dsa/bcm_sf2.c b/drivers/net/dsa/bcm_sf2.c
>> index 40b3adf7ad99..562b5eb23d90 100644
>> --- a/drivers/net/dsa/bcm_sf2.c
>> +++ b/drivers/net/dsa/bcm_sf2.c
>> @@ -600,7 +600,9 @@ static void bcm_sf2_sw_adjust_link(struct dsa_switch *ds, int port,
>>   	struct bcm_sf2_priv *priv = bcm_sf2_to_priv(ds);
>>   	struct ethtool_eee *p = &priv->port_sts[port].eee;
>>   	u32 id_mode_dis = 0, port_mode;
>> +	u16 lcl_adv = 0, rmt_adv = 0;
>>   	const char *str = NULL;
>> +	u8 flowctrl = 0;
>>   	u32 reg;
>>   
>>   	switch (phydev->interface) {
>> @@ -667,10 +669,24 @@ static void bcm_sf2_sw_adjust_link(struct dsa_switch *ds, int port,
>>   		break;
>>   	}
>>   
>> +	if (phydev->pause)
>> +		rmt_adv = LPA_PAUSE_CAP;
>> +	if (phydev->asym_pause)
>> +		rmt_adv |= LPA_PAUSE_ASYM;
>> +	if (phydev->advertising & ADVERTISED_Pause)
>> +		lcl_adv = ADVERTISE_PAUSE_CAP;
>> +	if (phydev->advertising & ADVERTISED_Asym_Pause)
>> +		lcl_adv |= ADVERTISE_PAUSE_ASYM;
>> +	flowctrl = mii_resolve_flowctrl_fdx(lcl_adv, rmt_adv);
> 
> IEEE 802.3 says "The PAUSE function shall be enabled according to Table
> 28B–3 only if the highest common denominator is a full duplex technology."

Indeed, sorry about that, so I suppose the incremental patch would do, 
since we do not support changing pause frame auto-negotiation settings 
in 4.9:

diff --git a/drivers/net/dsa/bcm_sf2.c b/drivers/net/dsa/bcm_sf2.c
index 562b5eb23d90..f3d61f2bb0f7 100644
--- a/drivers/net/dsa/bcm_sf2.c
+++ b/drivers/net/dsa/bcm_sf2.c
@@ -669,15 +669,18 @@ static void bcm_sf2_sw_adjust_link(struct 
dsa_switch *ds, int port,
                 break;
         }

-       if (phydev->pause)
-               rmt_adv = LPA_PAUSE_CAP;
-       if (phydev->asym_pause)
-               rmt_adv |= LPA_PAUSE_ASYM;
-       if (phydev->advertising & ADVERTISED_Pause)
-               lcl_adv = ADVERTISE_PAUSE_CAP;
-       if (phydev->advertising & ADVERTISED_Asym_Pause)
-               lcl_adv |= ADVERTISE_PAUSE_ASYM;
-       flowctrl = mii_resolve_flowctrl_fdx(lcl_adv, rmt_adv);
+       if (phydev->duplex == DUPLEX_FULL &&
+           phydev->autoneg == AUTONEG_ENABLE) {
+               if (phydev->pause)
+                       rmt_adv = LPA_PAUSE_CAP;
+               if (phydev->asym_pause)
+                       rmt_adv |= LPA_PAUSE_ASYM;
+               if (phydev->advertising & ADVERTISED_Pause)
+                       lcl_adv = ADVERTISE_PAUSE_CAP;
+               if (phydev->advertising & ADVERTISED_Asym_Pause)
+                       lcl_adv |= ADVERTISE_PAUSE_ASYM;
+               flowctrl = mii_resolve_flowctrl_fdx(lcl_adv, rmt_adv);
+       }

         if (phydev->link)
                 reg |= LINK_STS;
-- 
Florian