[PATCH iwl-next 0/2] Add link_down_events ethtool stat to ixgbe and ice

[PATCH iwl-next 0/2] Add link_down_events ethtool stat to ixgbe and ice

[Martyna Szapar-Mudlaw]

This series introduces a new ethtool statistic, link_down_events,
to both the ixgbe and ice drivers. The purpose of this counter is
to track the number of times the network link transitions from up to
down.
This statistic can help diagnose issues related to link stability,
such as port flapping or unexpected link drops. It is exposed via
ethtool.
Patch 1 adds this functionality to the ice driver, while patch 2 adds
the same support to the ixge driver.


Martyna Szapar-Mudlaw (2):
  ice: add link_down_events statistic
  ixgbe: add link_down_events statistic

 drivers/net/ethernet/intel/ice/ice.h             | 1 +
 drivers/net/ethernet/intel/ice/ice_ethtool.c     | 1 +
 drivers/net/ethernet/intel/ice/ice_main.c        | 3 +++
 drivers/net/ethernet/intel/ixgbe/ixgbe.h         | 1 +
 drivers/net/ethernet/intel/ixgbe/ixgbe_ethtool.c | 1 +
 drivers/net/ethernet/intel/ixgbe/ixgbe_main.c    | 2 ++
 6 files changed, 9 insertions(+)

-- 
2.47.0

[PATCH iwl-next 1/2] ice: add link_down_events statistic

[Martyna Szapar-Mudlaw]

Introduce a new ethtool statistic to ice driver, `link_down_events`,
to track the number of times the link transitions from up to down.
This counter can help diagnose issues related to link stability,
such as port flapping or unexpected link drops.

The counter increments when a link-down event occurs and is exposed
via ethtool stats as `link_down_events.nic`.

Reviewed-by: Michal Kubiak <michal.kubiak@intel.com>
Signed-off-by: Martyna Szapar-Mudlaw <martyna.szapar-mudlaw@linux.intel.com>
---
 drivers/net/ethernet/intel/ice/ice.h         | 1 +
 drivers/net/ethernet/intel/ice/ice_ethtool.c | 1 +
 drivers/net/ethernet/intel/ice/ice_main.c    | 3 +++
 3 files changed, 5 insertions(+)

diff --git a/drivers/net/ethernet/intel/ice/ice.h b/drivers/net/ethernet/intel/ice/ice.h
index 7200d6042590..6304104d1900 100644
--- a/drivers/net/ethernet/intel/ice/ice.h
+++ b/drivers/net/ethernet/intel/ice/ice.h
@@ -621,6 +621,7 @@ struct ice_pf {
 	u16 globr_count;	/* Global reset count */
 	u16 empr_count;		/* EMP reset count */
 	u16 pfr_count;		/* PF reset count */
+	u32 link_down_events;
 
 	u8 wol_ena : 1;		/* software state of WoL */
 	u32 wakeup_reason;	/* last wakeup reason */
diff --git a/drivers/net/ethernet/intel/ice/ice_ethtool.c b/drivers/net/ethernet/intel/ice/ice_ethtool.c
index b0805704834d..7bad0113aa88 100644
--- a/drivers/net/ethernet/intel/ice/ice_ethtool.c
+++ b/drivers/net/ethernet/intel/ice/ice_ethtool.c
@@ -137,6 +137,7 @@ static const struct ice_stats ice_gstrings_pf_stats[] = {
 	ICE_PF_STAT("mac_remote_faults.nic", stats.mac_remote_faults),
 	ICE_PF_STAT("fdir_sb_match.nic", stats.fd_sb_match),
 	ICE_PF_STAT("fdir_sb_status.nic", stats.fd_sb_status),
+	ICE_PF_STAT("link_down_events.nic", link_down_events),
 	ICE_PF_STAT("tx_hwtstamp_skipped", ptp.tx_hwtstamp_skipped),
 	ICE_PF_STAT("tx_hwtstamp_timeouts", ptp.tx_hwtstamp_timeouts),
 	ICE_PF_STAT("tx_hwtstamp_flushed", ptp.tx_hwtstamp_flushed),
diff --git a/drivers/net/ethernet/intel/ice/ice_main.c b/drivers/net/ethernet/intel/ice/ice_main.c
index a03e1819e6d5..d68dd2a3f4a6 100644
--- a/drivers/net/ethernet/intel/ice/ice_main.c
+++ b/drivers/net/ethernet/intel/ice/ice_main.c
@@ -1144,6 +1144,9 @@ ice_link_event(struct ice_pf *pf, struct ice_port_info *pi, bool link_up,
 	if (link_up == old_link && link_speed == old_link_speed)
 		return 0;
 
+	if (!link_up && old_link)
+		pf->link_down_events++;
+
 	ice_ptp_link_change(pf, link_up);
 
 	if (ice_is_dcb_active(pf)) {
-- 
2.47.0

[PATCH iwl-next 2/2] ixgbe: add link_down_events statistic

[Martyna Szapar-Mudlaw]

Introduce a new ethtool statistic to ixgbe driver, `link_down_events`,
to track the number of times the link transitions from up to down.
This counter can help diagnose issues related to link stability,
such as port flapping or unexpected link drops.

The counter increments when a link-down event occurs and is exposed
via ethtool stats as `link_down_events`.

Reviewed-by: Aleksandr Loktionov <aleksandr.loktionov@intel.com>
Signed-off-by: Martyna Szapar-Mudlaw <martyna.szapar-mudlaw@linux.intel.com>
---
 drivers/net/ethernet/intel/ixgbe/ixgbe.h         | 1 +
 drivers/net/ethernet/intel/ixgbe/ixgbe_ethtool.c | 1 +
 drivers/net/ethernet/intel/ixgbe/ixgbe_main.c    | 2 ++
 3 files changed, 4 insertions(+)

diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe.h b/drivers/net/ethernet/intel/ixgbe/ixgbe.h
index e6a380d4929b..7a8b4b6053c7 100644
--- a/drivers/net/ethernet/intel/ixgbe/ixgbe.h
+++ b/drivers/net/ethernet/intel/ixgbe/ixgbe.h
@@ -743,6 +743,7 @@ struct ixgbe_adapter {
 	bool link_up;
 	unsigned long sfp_poll_time;
 	unsigned long link_check_timeout;
+	u32 link_down_events;
 
 	struct timer_list service_timer;
 	struct work_struct service_task;
diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_ethtool.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_ethtool.c
index f03925c1f521..ea1d2c2390f1 100644
--- a/drivers/net/ethernet/intel/ixgbe/ixgbe_ethtool.c
+++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_ethtool.c
@@ -91,6 +91,7 @@ static const struct ixgbe_stats ixgbe_gstrings_stats[] = {
 	{"rx_hwtstamp_cleared", IXGBE_STAT(rx_hwtstamp_cleared)},
 	{"tx_ipsec", IXGBE_STAT(tx_ipsec)},
 	{"rx_ipsec", IXGBE_STAT(rx_ipsec)},
+	{"link_down_events", IXGBE_STAT(link_down_events)},
 #ifdef IXGBE_FCOE
 	{"fcoe_bad_fccrc", IXGBE_STAT(stats.fccrc)},
 	{"rx_fcoe_dropped", IXGBE_STAT(stats.fcoerpdc)},
diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
index 467f81239e12..cb5c782817fa 100644
--- a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
+++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
@@ -7986,6 +7986,8 @@ static void ixgbe_watchdog_link_is_down(struct ixgbe_adapter *adapter)
 	if (!netif_carrier_ok(netdev))
 		return;
 
+	adapter->link_down_events++;
+
 	/* poll for SFP+ cable when link is down */
 	if (ixgbe_is_sfp(hw) && hw->mac.type == ixgbe_mac_82598EB)
 		adapter->flags2 |= IXGBE_FLAG2_SEARCH_FOR_SFP;
-- 
2.47.0

Re: [PATCH iwl-next 1/2] ice: add link_down_events statistic

[Ido Schimmel]

On Wed, Apr 09, 2025 at 01:36:23PM +0200, Martyna Szapar-Mudlaw wrote:
> Introduce a new ethtool statistic to ice driver, `link_down_events`,
> to track the number of times the link transitions from up to down.
> This counter can help diagnose issues related to link stability,
> such as port flapping or unexpected link drops.
> 
> The counter increments when a link-down event occurs and is exposed
> via ethtool stats as `link_down_events.nic`.

Are you aware of commit 9a0f830f8026 ("ethtool: linkstate: add a
statistic for PHY down events")?

Better to report this via the generic counter than a driver-specific
one.

Re: [Intel-wired-lan] [PATCH iwl-next 1/2] ice: add link_down_events statistic

[Paul Menzel]

Dear Martyna,


Thank you for your patch.

Am 09.04.25 um 13:36 schrieb Martyna Szapar-Mudlaw:
> Introduce a new ethtool statistic to ice driver, `link_down_events`,
> to track the number of times the link transitions from up to down.
> This counter can help diagnose issues related to link stability,
> such as port flapping or unexpected link drops.
> 
> The counter increments when a link-down event occurs and is exposed
> via ethtool stats as `link_down_events.nic`.

It’d be great if you pasted an example output.

> Reviewed-by: Michal Kubiak <michal.kubiak@intel.com>
> Signed-off-by: Martyna Szapar-Mudlaw <martyna.szapar-mudlaw@linux.intel.com>
> ---
>   drivers/net/ethernet/intel/ice/ice.h         | 1 +
>   drivers/net/ethernet/intel/ice/ice_ethtool.c | 1 +
>   drivers/net/ethernet/intel/ice/ice_main.c    | 3 +++
>   3 files changed, 5 insertions(+)
> 
> diff --git a/drivers/net/ethernet/intel/ice/ice.h b/drivers/net/ethernet/intel/ice/ice.h
> index 7200d6042590..6304104d1900 100644
> --- a/drivers/net/ethernet/intel/ice/ice.h
> +++ b/drivers/net/ethernet/intel/ice/ice.h
> @@ -621,6 +621,7 @@ struct ice_pf {
>   	u16 globr_count;	/* Global reset count */
>   	u16 empr_count;		/* EMP reset count */
>   	u16 pfr_count;		/* PF reset count */
> +	u32 link_down_events;

Why not u16?

>   
>   	u8 wol_ena : 1;		/* software state of WoL */
>   	u32 wakeup_reason;	/* last wakeup reason */
> diff --git a/drivers/net/ethernet/intel/ice/ice_ethtool.c b/drivers/net/ethernet/intel/ice/ice_ethtool.c
> index b0805704834d..7bad0113aa88 100644
> --- a/drivers/net/ethernet/intel/ice/ice_ethtool.c
> +++ b/drivers/net/ethernet/intel/ice/ice_ethtool.c
> @@ -137,6 +137,7 @@ static const struct ice_stats ice_gstrings_pf_stats[] = {
>   	ICE_PF_STAT("mac_remote_faults.nic", stats.mac_remote_faults),
>   	ICE_PF_STAT("fdir_sb_match.nic", stats.fd_sb_match),
>   	ICE_PF_STAT("fdir_sb_status.nic", stats.fd_sb_status),
> +	ICE_PF_STAT("link_down_events.nic", link_down_events),
>   	ICE_PF_STAT("tx_hwtstamp_skipped", ptp.tx_hwtstamp_skipped),
>   	ICE_PF_STAT("tx_hwtstamp_timeouts", ptp.tx_hwtstamp_timeouts),
>   	ICE_PF_STAT("tx_hwtstamp_flushed", ptp.tx_hwtstamp_flushed),
> diff --git a/drivers/net/ethernet/intel/ice/ice_main.c b/drivers/net/ethernet/intel/ice/ice_main.c
> index a03e1819e6d5..d68dd2a3f4a6 100644
> --- a/drivers/net/ethernet/intel/ice/ice_main.c
> +++ b/drivers/net/ethernet/intel/ice/ice_main.c
> @@ -1144,6 +1144,9 @@ ice_link_event(struct ice_pf *pf, struct ice_port_info *pi, bool link_up,
>   	if (link_up == old_link && link_speed == old_link_speed)
>   		return 0;
>   
> +	if (!link_up && old_link)
> +		pf->link_down_events++;
> +
>   	ice_ptp_link_change(pf, link_up);
>   
>   	if (ice_is_dcb_active(pf)) {

The diff looks good.


Kind regards,

Paul

Re: [PATCH iwl-next 1/2] ice: add link_down_events statistic

[Szapar-Mudlaw, Martyna]

On 4/9/2025 2:06 PM, Ido Schimmel wrote:
> On Wed, Apr 09, 2025 at 01:36:23PM +0200, Martyna Szapar-Mudlaw wrote:
>> Introduce a new ethtool statistic to ice driver, `link_down_events`,
>> to track the number of times the link transitions from up to down.
>> This counter can help diagnose issues related to link stability,
>> such as port flapping or unexpected link drops.
>>
>> The counter increments when a link-down event occurs and is exposed
>> via ethtool stats as `link_down_events.nic`.
> 
> Are you aware of commit 9a0f830f8026 ("ethtool: linkstate: add a
> statistic for PHY down events")?
> 
> Better to report this via the generic counter than a driver-specific
> one.
> 

Right, thank you for pointing this, I just submitted v2 where ethtool 
get_link_ext_stats() interface is used.

Re: [Intel-wired-lan] [PATCH iwl-next 1/2] ice: add link_down_events statistic

[Szapar-Mudlaw, Martyna]

On 4/9/2025 2:08 PM, Paul Menzel wrote:
> Dear Martyna,
> 
> 
> Thank you for your patch.
> 
> Am 09.04.25 um 13:36 schrieb Martyna Szapar-Mudlaw:
>> Introduce a new ethtool statistic to ice driver, `link_down_events`,
>> to track the number of times the link transitions from up to down.
>> This counter can help diagnose issues related to link stability,
>> such as port flapping or unexpected link drops.
>>
>> The counter increments when a link-down event occurs and is exposed
>> via ethtool stats as `link_down_events.nic`.
> 
> It’d be great if you pasted an example output.

In v2 (which I just submitted) the generic ethtool statistic is used for 
this, instead of driver specific, so I guess no need to paste the 
example output now.

> 
>> Reviewed-by: Michal Kubiak <michal.kubiak@intel.com>
>> Signed-off-by: Martyna Szapar-Mudlaw <martyna.szapar- 
>> mudlaw@linux.intel.com>
>> ---
>>   drivers/net/ethernet/intel/ice/ice.h         | 1 +
>>   drivers/net/ethernet/intel/ice/ice_ethtool.c | 1 +
>>   drivers/net/ethernet/intel/ice/ice_main.c    | 3 +++
>>   3 files changed, 5 insertions(+)
>>
>> diff --git a/drivers/net/ethernet/intel/ice/ice.h b/drivers/net/ 
>> ethernet/intel/ice/ice.h
>> index 7200d6042590..6304104d1900 100644
>> --- a/drivers/net/ethernet/intel/ice/ice.h
>> +++ b/drivers/net/ethernet/intel/ice/ice.h
>> @@ -621,6 +621,7 @@ struct ice_pf {
>>       u16 globr_count;    /* Global reset count */
>>       u16 empr_count;        /* EMP reset count */
>>       u16 pfr_count;        /* PF reset count */
>> +    u32 link_down_events;
> 
> Why not u16?

So now using u32 instead of u16 is more justified, as the v2 uses the 
generic ethtool stat, where this value is also u32 :)

> 
>>       u8 wol_ena : 1;        /* software state of WoL */
>>       u32 wakeup_reason;    /* last wakeup reason */
>> diff --git a/drivers/net/ethernet/intel/ice/ice_ethtool.c b/drivers/ 
>> net/ethernet/intel/ice/ice_ethtool.c
>> index b0805704834d..7bad0113aa88 100644
>> --- a/drivers/net/ethernet/intel/ice/ice_ethtool.c
>> +++ b/drivers/net/ethernet/intel/ice/ice_ethtool.c
>> @@ -137,6 +137,7 @@ static const struct ice_stats 
>> ice_gstrings_pf_stats[] = {
>>       ICE_PF_STAT("mac_remote_faults.nic", stats.mac_remote_faults),
>>       ICE_PF_STAT("fdir_sb_match.nic", stats.fd_sb_match),
>>       ICE_PF_STAT("fdir_sb_status.nic", stats.fd_sb_status),
>> +    ICE_PF_STAT("link_down_events.nic", link_down_events),
>>       ICE_PF_STAT("tx_hwtstamp_skipped", ptp.tx_hwtstamp_skipped),
>>       ICE_PF_STAT("tx_hwtstamp_timeouts", ptp.tx_hwtstamp_timeouts),
>>       ICE_PF_STAT("tx_hwtstamp_flushed", ptp.tx_hwtstamp_flushed),
>> diff --git a/drivers/net/ethernet/intel/ice/ice_main.c b/drivers/net/ 
>> ethernet/intel/ice/ice_main.c
>> index a03e1819e6d5..d68dd2a3f4a6 100644
>> --- a/drivers/net/ethernet/intel/ice/ice_main.c
>> +++ b/drivers/net/ethernet/intel/ice/ice_main.c
>> @@ -1144,6 +1144,9 @@ ice_link_event(struct ice_pf *pf, struct 
>> ice_port_info *pi, bool link_up,
>>       if (link_up == old_link && link_speed == old_link_speed)
>>           return 0;
>> +    if (!link_up && old_link)
>> +        pf->link_down_events++;
>> +
>>       ice_ptp_link_change(pf, link_up);
>>       if (ice_is_dcb_active(pf)) {
> 
> The diff looks good.

Thank you for the review,
Martyna
> 
> 
> Kind regards,
> 
> Paul
> 
> 

Re: [Intel-wired-lan] [PATCH iwl-next 1/2] ice: add link_down_events statistic

[Paul Menzel]

Dear Martyna,


Thank you for your reply.

Am 14.04.25 um 15:12 schrieb Szapar-Mudlaw, Martyna:

> On 4/9/2025 2:08 PM, Paul Menzel wrote:

>> Am 09.04.25 um 13:36 schrieb Martyna Szapar-Mudlaw:
>>> Introduce a new ethtool statistic to ice driver, `link_down_events`,
>>> to track the number of times the link transitions from up to down.
>>> This counter can help diagnose issues related to link stability,
>>> such as port flapping or unexpected link drops.
>>>
>>> The counter increments when a link-down event occurs and is exposed
>>> via ethtool stats as `link_down_events.nic`.
>>
>> It’d be great if you pasted an example output.
> 
> In v2 (which I just submitted) the generic ethtool statistic is used for 
> this, instead of driver specific, so I guess no need to paste the 
> example output now.

I think it’s always good also as a reference how to test the patch. I 
just saw your v3. Should you resent, it’d be great if you added the 
example output.

[…]


Kind regards,

Paul

Re: [Intel-wired-lan] [PATCH iwl-next 1/2] ice: add link_down_events statistic

[Szapar-Mudlaw, Martyna]

On 5/12/2025 11:27 AM, Paul Menzel wrote:
> Dear Martyna,
> 
> 
> Thank you for your reply.
> 
> Am 14.04.25 um 15:12 schrieb Szapar-Mudlaw, Martyna:
> 
>> On 4/9/2025 2:08 PM, Paul Menzel wrote:
> 
>>> Am 09.04.25 um 13:36 schrieb Martyna Szapar-Mudlaw:
>>>> Introduce a new ethtool statistic to ice driver, `link_down_events`,
>>>> to track the number of times the link transitions from up to down.
>>>> This counter can help diagnose issues related to link stability,
>>>> such as port flapping or unexpected link drops.
>>>>
>>>> The counter increments when a link-down event occurs and is exposed
>>>> via ethtool stats as `link_down_events.nic`.
>>>
>>> It’d be great if you pasted an example output.
>>
>> In v2 (which I just submitted) the generic ethtool statistic is used 
>> for this, instead of driver specific, so I guess no need to paste the 
>> example output now.
> 
> I think it’s always good also as a reference how to test the patch. I 
> just saw your v3. Should you resent, it’d be great if you added the 
> example output.
> 
> […]

Done in v4 updated cover letter.

> 
> 
> Kind regards,
> 
> Paul