[PATCH net] igb: extend PTP timestamp adjustments to i211

[PATCH net] igb: extend PTP timestamp adjustments to i211

[Tony Nguyen]

From: Oleksij Rempel <o.rempel@pengutronix.de>

The i211 requires the same PTP timestamp adjustments as the i210,
according to its datasheet. To ensure consistent timestamping across
different platforms, this change extends the existing adjustments to
include the i211.

The adjustment result are tested and comparable for i210 and i211 based
systems.

Fixes: 3f544d2a4d5c ("igb: adjust PTP timestamps for Tx/Rx latency")
Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de>
Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
Tested-by: Pucha Himasekhar Reddy <himasekharx.reddy.pucha@intel.com> (A Contingent worker at Intel)
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
---
 drivers/net/ethernet/intel/igb/igb_ptp.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/drivers/net/ethernet/intel/igb/igb_ptp.c b/drivers/net/ethernet/intel/igb/igb_ptp.c
index 319c544b9f04..f94570556120 100644
--- a/drivers/net/ethernet/intel/igb/igb_ptp.c
+++ b/drivers/net/ethernet/intel/igb/igb_ptp.c
@@ -957,7 +957,7 @@ static void igb_ptp_tx_hwtstamp(struct igb_adapter *adapter)
 
 	igb_ptp_systim_to_hwtstamp(adapter, &shhwtstamps, regval);
 	/* adjust timestamp for the TX latency based on link speed */
-	if (adapter->hw.mac.type == e1000_i210) {
+	if (hw->mac.type == e1000_i210 || hw->mac.type == e1000_i211) {
 		switch (adapter->link_speed) {
 		case SPEED_10:
 			adjust = IGB_I210_TX_LATENCY_10;
@@ -1003,6 +1003,7 @@ int igb_ptp_rx_pktstamp(struct igb_q_vector *q_vector, void *va,
 			ktime_t *timestamp)
 {
 	struct igb_adapter *adapter = q_vector->adapter;
+	struct e1000_hw *hw = &adapter->hw;
 	struct skb_shared_hwtstamps ts;
 	__le64 *regval = (__le64 *)va;
 	int adjust = 0;
@@ -1022,7 +1023,7 @@ int igb_ptp_rx_pktstamp(struct igb_q_vector *q_vector, void *va,
 	igb_ptp_systim_to_hwtstamp(adapter, &ts, le64_to_cpu(regval[1]));
 
 	/* adjust timestamp for the RX latency based on link speed */
-	if (adapter->hw.mac.type == e1000_i210) {
+	if (hw->mac.type == e1000_i210 || hw->mac.type == e1000_i211) {
 		switch (adapter->link_speed) {
 		case SPEED_10:
 			adjust = IGB_I210_RX_LATENCY_10;
-- 
2.41.0

Re: [PATCH net] igb: extend PTP timestamp adjustments to i211

[Jiri Pirko]

Tue, Feb 27, 2024 at 07:49:41PM CET, anthony.l.nguyen@intel.com wrote:
>From: Oleksij Rempel <o.rempel@pengutronix.de>
>
>The i211 requires the same PTP timestamp adjustments as the i210,
>according to its datasheet. To ensure consistent timestamping across
>different platforms, this change extends the existing adjustments to
>include the i211.
>
>The adjustment result are tested and comparable for i210 and i211 based
>systems.
>
>Fixes: 3f544d2a4d5c ("igb: adjust PTP timestamps for Tx/Rx latency")

IIUC, you are just extending the timestamp adjusting to another HW, not
actually fixing any error, don't you? In that case, I don't see why not
to rather target net-next and avoid "Fixes" tag. Or do I misunderstand
this?


>Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de>
>Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
>Tested-by: Pucha Himasekhar Reddy <himasekharx.reddy.pucha@intel.com> (A Contingent worker at Intel)
>Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
>---
> drivers/net/ethernet/intel/igb/igb_ptp.c | 5 +++--
> 1 file changed, 3 insertions(+), 2 deletions(-)
>
>diff --git a/drivers/net/ethernet/intel/igb/igb_ptp.c b/drivers/net/ethernet/intel/igb/igb_ptp.c
>index 319c544b9f04..f94570556120 100644
>--- a/drivers/net/ethernet/intel/igb/igb_ptp.c
>+++ b/drivers/net/ethernet/intel/igb/igb_ptp.c
>@@ -957,7 +957,7 @@ static void igb_ptp_tx_hwtstamp(struct igb_adapter *adapter)
> 
> 	igb_ptp_systim_to_hwtstamp(adapter, &shhwtstamps, regval);
> 	/* adjust timestamp for the TX latency based on link speed */
>-	if (adapter->hw.mac.type == e1000_i210) {
>+	if (hw->mac.type == e1000_i210 || hw->mac.type == e1000_i211) {
> 		switch (adapter->link_speed) {
> 		case SPEED_10:
> 			adjust = IGB_I210_TX_LATENCY_10;
>@@ -1003,6 +1003,7 @@ int igb_ptp_rx_pktstamp(struct igb_q_vector *q_vector, void *va,
> 			ktime_t *timestamp)
> {
> 	struct igb_adapter *adapter = q_vector->adapter;
>+	struct e1000_hw *hw = &adapter->hw;
> 	struct skb_shared_hwtstamps ts;
> 	__le64 *regval = (__le64 *)va;
> 	int adjust = 0;
>@@ -1022,7 +1023,7 @@ int igb_ptp_rx_pktstamp(struct igb_q_vector *q_vector, void *va,
> 	igb_ptp_systim_to_hwtstamp(adapter, &ts, le64_to_cpu(regval[1]));
> 
> 	/* adjust timestamp for the RX latency based on link speed */
>-	if (adapter->hw.mac.type == e1000_i210) {
>+	if (hw->mac.type == e1000_i210 || hw->mac.type == e1000_i211) {
> 		switch (adapter->link_speed) {
> 		case SPEED_10:
> 			adjust = IGB_I210_RX_LATENCY_10;
>-- 
>2.41.0
>
>

Re: [PATCH net] igb: extend PTP timestamp adjustments to i211

[Oleksij Rempel]

On Wed, Feb 28, 2024 at 10:37:56AM +0100, Jiri Pirko wrote:
> Tue, Feb 27, 2024 at 07:49:41PM CET, anthony.l.nguyen@intel.com wrote:
> >From: Oleksij Rempel <o.rempel@pengutronix.de>
> >
> >The i211 requires the same PTP timestamp adjustments as the i210,
> >according to its datasheet. To ensure consistent timestamping across
> >different platforms, this change extends the existing adjustments to
> >include the i211.
> >
> >The adjustment result are tested and comparable for i210 and i211 based
> >systems.
> >
> >Fixes: 3f544d2a4d5c ("igb: adjust PTP timestamps for Tx/Rx latency")
> 
> IIUC, you are just extending the timestamp adjusting to another HW, not
> actually fixing any error, don't you? In that case, I don't see why not
> to rather target net-next and avoid "Fixes" tag. Or do I misunderstand
> this?

From my perspective, it was an error, since two nearly identical systems
with only one difference (one used i210 other i211) showed different PTP
measurements. So, it would be nice if distributions would include this
fix. On other hand, I'm ok with what ever maintainer would decide how
to handle this patch.

Regards,
Oleksij
-- 
Pengutronix e.K.                           |                             |
Steuerwalder Str. 21                       | http://www.pengutronix.de/  |
31137 Hildesheim, Germany                  | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |

RE: [PATCH net] igb: extend PTP timestamp adjustments to i211

[Keller, Jacob E]

> -----Original Message-----
> From: Oleksij Rempel <o.rempel@pengutronix.de>
> Sent: Wednesday, February 28, 2024 4:28 AM
> To: Jiri Pirko <jiri@resnulli.us>
> Cc: Nguyen, Anthony L <anthony.l.nguyen@intel.com>; davem@davemloft.net;
> kuba@kernel.org; pabeni@redhat.com; edumazet@google.com;
> netdev@vger.kernel.org; richardcochran@gmail.com; nathan.sullivan@ni.com;
> Keller, Jacob E <jacob.e.keller@intel.com>; Pucha, HimasekharX Reddy
> <himasekharx.reddy.pucha@intel.com>
> Subject: Re: [PATCH net] igb: extend PTP timestamp adjustments to i211
> 
> On Wed, Feb 28, 2024 at 10:37:56AM +0100, Jiri Pirko wrote:
> > Tue, Feb 27, 2024 at 07:49:41PM CET, anthony.l.nguyen@intel.com wrote:
> > >From: Oleksij Rempel <o.rempel@pengutronix.de>
> > >
> > >The i211 requires the same PTP timestamp adjustments as the i210,
> > >according to its datasheet. To ensure consistent timestamping across
> > >different platforms, this change extends the existing adjustments to
> > >include the i211.
> > >
> > >The adjustment result are tested and comparable for i210 and i211 based
> > >systems.
> > >
> > >Fixes: 3f544d2a4d5c ("igb: adjust PTP timestamps for Tx/Rx latency")
> >
> > IIUC, you are just extending the timestamp adjusting to another HW, not
> > actually fixing any error, don't you? In that case, I don't see why not
> > to rather target net-next and avoid "Fixes" tag. Or do I misunderstand
> > this?
> 
> From my perspective, it was an error, since two nearly identical systems
> with only one difference (one used i210 other i211) showed different PTP
> measurements. So, it would be nice if distributions would include this
> fix. On other hand, I'm ok with what ever maintainer would decide how
> to handle this patch.
> 
> Regards,
> Oleksij

Without this, the i211 doesn't apply the Tx/Rx latency adjustments, so the timestamps would be less accurate than if the corrections are applied. On the one hand I guess this is a "feature" and the lack of a feature isn't a bug, so maybe its not viewed as a bug fix then.

Another interpretation is that lacking those corrections is a bug which this patch fixes.

Thanks,
Jake

> --
> Pengutronix e.K.                           |                             |
> Steuerwalder Str. 21                       | http://www.pengutronix.de/  |
> 31137 Hildesheim, Germany                  | Phone: +49-5121-206917-0    |
> Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |

Re: [PATCH net] igb: extend PTP timestamp adjustments to i211

[Jakub Kicinski]

On Wed, 28 Feb 2024 17:43:03 +0000 Keller, Jacob E wrote:
> Without this, the i211 doesn't apply the Tx/Rx latency adjustments,
> so the timestamps would be less accurate than if the corrections are
> applied. On the one hand I guess this is a "feature" and the lack of
> a feature isn't a bug, so maybe its not viewed as a bug fix then.

I tossed a coin and it said.... "whatever, man". Such sass, tsk tsk.

Re: [PATCH net] igb: extend PTP timestamp adjustments to i211

[patchwork-bot+netdevbpf]

Hello:

This patch was applied to netdev/net.git (main)
by Jakub Kicinski <kuba@kernel.org>:

On Tue, 27 Feb 2024 10:49:41 -0800 you wrote:
> From: Oleksij Rempel <o.rempel@pengutronix.de>
> 
> The i211 requires the same PTP timestamp adjustments as the i210,
> according to its datasheet. To ensure consistent timestamping across
> different platforms, this change extends the existing adjustments to
> include the i211.
> 
> [...]

Here is the summary with links:
  - [net] igb: extend PTP timestamp adjustments to i211
    https://git.kernel.org/netdev/net/c/0bb7b09392eb

You are awesome, thank you!
-- 
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html

Re: [PATCH net] igb: extend PTP timestamp adjustments to i211

[Jiri Pirko]

Wed, Feb 28, 2024 at 06:43:03PM CET, jacob.e.keller@intel.com wrote:
>
>
>> -----Original Message-----
>> From: Oleksij Rempel <o.rempel@pengutronix.de>
>> Sent: Wednesday, February 28, 2024 4:28 AM
>> To: Jiri Pirko <jiri@resnulli.us>
>> Cc: Nguyen, Anthony L <anthony.l.nguyen@intel.com>; davem@davemloft.net;
>> kuba@kernel.org; pabeni@redhat.com; edumazet@google.com;
>> netdev@vger.kernel.org; richardcochran@gmail.com; nathan.sullivan@ni.com;
>> Keller, Jacob E <jacob.e.keller@intel.com>; Pucha, HimasekharX Reddy
>> <himasekharx.reddy.pucha@intel.com>
>> Subject: Re: [PATCH net] igb: extend PTP timestamp adjustments to i211
>> 
>> On Wed, Feb 28, 2024 at 10:37:56AM +0100, Jiri Pirko wrote:
>> > Tue, Feb 27, 2024 at 07:49:41PM CET, anthony.l.nguyen@intel.com wrote:
>> > >From: Oleksij Rempel <o.rempel@pengutronix.de>
>> > >
>> > >The i211 requires the same PTP timestamp adjustments as the i210,
>> > >according to its datasheet. To ensure consistent timestamping across
>> > >different platforms, this change extends the existing adjustments to
>> > >include the i211.
>> > >
>> > >The adjustment result are tested and comparable for i210 and i211 based
>> > >systems.
>> > >
>> > >Fixes: 3f544d2a4d5c ("igb: adjust PTP timestamps for Tx/Rx latency")
>> >
>> > IIUC, you are just extending the timestamp adjusting to another HW, not
>> > actually fixing any error, don't you? In that case, I don't see why not
>> > to rather target net-next and avoid "Fixes" tag. Or do I misunderstand
>> > this?
>> 
>> From my perspective, it was an error, since two nearly identical systems
>> with only one difference (one used i210 other i211) showed different PTP
>> measurements. So, it would be nice if distributions would include this
>> fix. On other hand, I'm ok with what ever maintainer would decide how
>> to handle this patch.
>> 
>> Regards,
>> Oleksij
>
>Without this, the i211 doesn't apply the Tx/Rx latency adjustments, so the timestamps would be less accurate than if the corrections are applied. On the one hand I guess this is a "feature" and the lack of a feature isn't a bug, so maybe its not viewed as a bug fix then.

The behaviour of i211 is the same as it always was. I mean, 3f544d2a4d5c
didn't cause any regression. From that perspective, it is clearly a
feature.

I know that in netdev we are taking the meaning of "Fixes" quite on the
edge often, but I think this is off the cliff :)


>
>Another interpretation is that lacking those corrections is a bug which this patch fixes.
>
>Thanks,
>Jake
>
>> --
>> Pengutronix e.K.                           |                             |
>> Steuerwalder Str. 21                       | http://www.pengutronix.de/  |
>> 31137 Hildesheim, Germany                  | Phone: +49-5121-206917-0    |
>> Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |

RE: [PATCH net] igb: extend PTP timestamp adjustments to i211

[Keller, Jacob E]

> -----Original Message-----
> From: Jakub Kicinski <kuba@kernel.org>
> Sent: Wednesday, February 28, 2024 8:21 PM
> To: Keller, Jacob E <jacob.e.keller@intel.com>
> Cc: Oleksij Rempel <o.rempel@pengutronix.de>; Jiri Pirko <jiri@resnulli.us>;
> Nguyen, Anthony L <anthony.l.nguyen@intel.com>; davem@davemloft.net;
> pabeni@redhat.com; edumazet@google.com; netdev@vger.kernel.org;
> richardcochran@gmail.com; nathan.sullivan@ni.com; Pucha, HimasekharX Reddy
> <himasekharx.reddy.pucha@intel.com>
> Subject: Re: [PATCH net] igb: extend PTP timestamp adjustments to i211
> 
> On Wed, 28 Feb 2024 17:43:03 +0000 Keller, Jacob E wrote:
> > Without this, the i211 doesn't apply the Tx/Rx latency adjustments,
> > so the timestamps would be less accurate than if the corrections are
> > applied. On the one hand I guess this is a "feature" and the lack of
> > a feature isn't a bug, so maybe its not viewed as a bug fix then.
> 
> I tossed a coin and it said.... "whatever, man". Such sass, tsk tsk.

I might have not had enough caffiene when writing that email... 😊

Re: [PATCH net] igb: extend PTP timestamp adjustments to i211

[Jakub Kicinski]

On Thu, 29 Feb 2024 20:09:03 +0000 Keller, Jacob E wrote:
> > On Wed, 28 Feb 2024 17:43:03 +0000 Keller, Jacob E wrote:  
> > > Without this, the i211 doesn't apply the Tx/Rx latency adjustments,
> > > so the timestamps would be less accurate than if the corrections are
> > > applied. On the one hand I guess this is a "feature" and the lack of
> > > a feature isn't a bug, so maybe its not viewed as a bug fix then.  
> > 
> > I tossed a coin and it said.... "whatever, man". Such sass, tsk tsk.  
> 
> I might have not had enough caffiene when writing that email... 😊

Sorry, my bad, I was barely awake. What I meant to convey is that
it's borderline, so discussing how people feel is not a good use of
anyone's time.

It's a 3 LoC change to include one chip in what's done for another chip,
it has been well tested, and has user impact. It's fine.