[PATCH iwl-next v2 0/5] ice: add standard stats

[PATCH iwl-next v2 0/5] ice: add standard stats

[Jesse Brandeburg]

The main point of this series is to implement the standard stats for the
ice driver, but also add a related documentation fix and finish the
series off with a cleanup that removes a bunch of code.

Changelog:
v2: address review comments on 1/5 (Jakub) regarding backticks,
    fix email address for one reviewer in commit message
    pick up some reviewed-bys from the list
v1: https://lore.kernel.org/netdev/20240604221327.299184-1-jesse.brandeburg@intel.com/

Jesse Brandeburg (5):
  net: docs: add missing features that can have stats
  ice: implement ethtool standard stats
  ice: add tracking of good transmit timestamps
  ice: implement transmit hardware timestamp statistics
  ice: refactor to use helpers

 Documentation/networking/statistics.rst       |   4 +-
 drivers/net/ethernet/intel/ice/ice_ethtool.c  | 138 ++++++++++++++----
 .../net/ethernet/intel/ice/ice_flex_pipe.c    |   8 +-
 drivers/net/ethernet/intel/ice/ice_lag.c      |   5 +-
 drivers/net/ethernet/intel/ice/ice_main.c     |  10 +-
 drivers/net/ethernet/intel/ice/ice_ptp.c      |   9 ++
 drivers/net/ethernet/intel/ice/ice_ptp.h      |   1 +
 drivers/net/ethernet/intel/ice/ice_sriov.c    |   3 +-
 drivers/net/ethernet/intel/ice/ice_type.h     |   1 +
 9 files changed, 132 insertions(+), 47 deletions(-)


base-commit: 95cd03f32a1680f693b291da920ab5d3f9d8c5c1
-- 
2.43.0

[PATCH iwl-next v2 1/5] net: docs: add missing features that can have stats

[Jesse Brandeburg]

While trying to figure out ethtool -I | --include-statistics, I noticed
some docs got missed when implementing commit 0e9c127729be ("ethtool:
add interface to read Tx hardware timestamping statistics").

Fix up the docs to match the kernel code, and while there, sort them in
alphabetical order.

Cc: Rahul Rameshbabu <rrameshbabu@nvidia.com>
Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
Reviewed-by: Rahul Rameshbabu <rrameshbabu@nvidia.com>
Reviewed-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
---
v2: fix up a mixup of backticks and quotes (Jakub)

I didn't add a Fixes: tag because this is not an urgent kind of fix that
should require backports.
---
 Documentation/networking/statistics.rst | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/Documentation/networking/statistics.rst b/Documentation/networking/statistics.rst
index 75e017dfa825..06f01c0cd65a 100644
--- a/Documentation/networking/statistics.rst
+++ b/Documentation/networking/statistics.rst
@@ -184,9 +184,11 @@ Protocol-related statistics can be requested in get commands by setting
 the `ETHTOOL_FLAG_STATS` flag in `ETHTOOL_A_HEADER_FLAGS`. Currently
 statistics are supported in the following commands:
 
-  - `ETHTOOL_MSG_PAUSE_GET`
   - `ETHTOOL_MSG_FEC_GET`
+  - `ETHTOOL_MSG_LINKSTATE_GET`
   - `ETHTOOL_MSG_MM_GET`
+  - `ETHTOOL_MSG_PAUSE_GET`
+  - `ETHTOOL_MSG_TSINFO_GET`
 
 debugfs
 -------
-- 
2.43.0

Re: [PATCH iwl-next v2 1/5] net: docs: add missing features that can have stats

[Simon Horman]

On Thu, Jun 06, 2024 at 03:46:55PM -0700, Jesse Brandeburg wrote:
> While trying to figure out ethtool -I | --include-statistics, I noticed
> some docs got missed when implementing commit 0e9c127729be ("ethtool:
> add interface to read Tx hardware timestamping statistics").
> 
> Fix up the docs to match the kernel code, and while there, sort them in
> alphabetical order.
> 
> Cc: Rahul Rameshbabu <rrameshbabu@nvidia.com>
> Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
> Reviewed-by: Rahul Rameshbabu <rrameshbabu@nvidia.com>
> Reviewed-by: Jakub Kicinski <kuba@kernel.org>
> Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>

Reviewed-by: Simon Horman <horms@kernel.org>

[PATCH iwl-next v2 2/5] ice: implement ethtool standard stats

[Jesse Brandeburg]

Add support for MAC/pause/RMON stats. This enables reporting hardware
statistics in a common way via:

ethtool -S eth0 --all-groups
and
ethtool --include-statistics --show-pause eth0

While doing so, add support for one new stat, receive length error
(RLEC), which is extremely unlikely to happen since most L2 frames have
a type/length field specifying a "type", and raw ethernet frames aren't
used much any longer.

NOTE: I didn't implement Ctrl aka control frame stats because the
hardware doesn't seem to implement support.

Reviewed-by: Marcin Szycik <marcin.szycik@linux.intel.com>
Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
Reviewed-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
---
Example output:

Standard stats for ens785f1np1:
eth-mac-FramesTransmittedOK: 296
eth-mac-FramesReceivedOK: 339
eth-mac-FrameCheckSequenceErrors: 0
eth-mac-OctetsTransmittedOK: 70410
eth-mac-OctetsReceivedOK: 51414
eth-mac-MulticastFramesXmittedOK: 124
eth-mac-BroadcastFramesXmittedOK: 1
eth-mac-MulticastFramesReceivedOK: 164
eth-mac-BroadcastFramesReceivedOK: 3
eth-mac-InRangeLengthErrors: 0
eth-mac-FrameTooLongErrors: 0
rmon-etherStatsUndersizePkts: 0
rmon-etherStatsOversizePkts: 0
rmon-etherStatsFragments: 0
rmon-etherStatsJabbers: 0
rx-rmon-etherStatsPkts64Octets: 31
rx-rmon-etherStatsPkts65to127Octets: 265
rx-rmon-etherStatsPkts128to255Octets: 26
rx-rmon-etherStatsPkts256to511Octets: 7
rx-rmon-etherStatsPkts512to1023Octets: 1
rx-rmon-etherStatsPkts1024to1522Octets: 6
rx-rmon-etherStatsPkts1523to9522Octets: 3
tx-rmon-etherStatsPkts64Octets: 25
tx-rmon-etherStatsPkts65to127Octets: 255
tx-rmon-etherStatsPkts128to255Octets: 1
tx-rmon-etherStatsPkts256to511Octets: 2
tx-rmon-etherStatsPkts512to1023Octets: 1
tx-rmon-etherStatsPkts1024to1522Octets: 1
tx-rmon-etherStatsPkts1523to9522Octets: 11
and
Pause parameters for ens785f0np0:
Autonegotiate:  on
RX:             off
TX:             off
RX negotiated: off
TX negotiated: off
Statistics:
  tx_pause_frames: 0
  rx_pause_frames: 0
---
 drivers/net/ethernet/intel/ice/ice_ethtool.c | 78 ++++++++++++++++++++
 drivers/net/ethernet/intel/ice/ice_main.c    |  3 +
 drivers/net/ethernet/intel/ice/ice_type.h    |  1 +
 3 files changed, 82 insertions(+)

diff --git a/drivers/net/ethernet/intel/ice/ice_ethtool.c b/drivers/net/ethernet/intel/ice/ice_ethtool.c
index 62c8205fceba..6f0a857f55c9 100644
--- a/drivers/net/ethernet/intel/ice/ice_ethtool.c
+++ b/drivers/net/ethernet/intel/ice/ice_ethtool.c
@@ -4282,6 +4282,81 @@ ice_get_module_eeprom(struct net_device *netdev,
 	return 0;
 }
 
+static void ice_get_eth_mac_stats(struct net_device *netdev,
+				  struct ethtool_eth_mac_stats *mac_stats)
+{
+	struct ice_pf *pf = ice_netdev_to_pf(netdev);
+	struct ice_hw_port_stats *ps = &pf->stats;
+
+	mac_stats->FramesTransmittedOK = ps->eth.tx_unicast +
+					 ps->eth.tx_multicast +
+					 ps->eth.tx_broadcast;
+	mac_stats->FramesReceivedOK = ps->eth.rx_unicast +
+				      ps->eth.rx_multicast +
+				      ps->eth.rx_broadcast;
+	mac_stats->FrameCheckSequenceErrors = ps->crc_errors;
+	mac_stats->OctetsTransmittedOK = ps->eth.tx_bytes;
+	mac_stats->OctetsReceivedOK = ps->eth.rx_bytes;
+	mac_stats->MulticastFramesXmittedOK = ps->eth.tx_multicast;
+	mac_stats->BroadcastFramesXmittedOK = ps->eth.tx_broadcast;
+	mac_stats->MulticastFramesReceivedOK = ps->eth.rx_multicast;
+	mac_stats->BroadcastFramesReceivedOK = ps->eth.rx_broadcast;
+	mac_stats->InRangeLengthErrors = ps->rx_len_errors;
+	mac_stats->FrameTooLongErrors = ps->rx_oversize;
+}
+
+static void ice_get_pause_stats(struct net_device *netdev,
+				struct ethtool_pause_stats *pause_stats)
+{
+	struct ice_pf *pf = ice_netdev_to_pf(netdev);
+	struct ice_hw_port_stats *ps = &pf->stats;
+
+	pause_stats->tx_pause_frames = ps->link_xon_tx + ps->link_xoff_tx;
+	pause_stats->rx_pause_frames = ps->link_xon_rx + ps->link_xoff_rx;
+}
+
+static const struct ethtool_rmon_hist_range ice_rmon_ranges[] = {
+	{    0,    64 },
+	{   65,   127 },
+	{  128,   255 },
+	{  256,   511 },
+	{  512,  1023 },
+	{ 1024,  1522 },
+	{ 1523,  9522 },
+	{}
+};
+
+static void ice_get_rmon_stats(struct net_device *netdev,
+			       struct ethtool_rmon_stats *rmon,
+			       const struct ethtool_rmon_hist_range **ranges)
+{
+	struct ice_pf *pf = ice_netdev_to_pf(netdev);
+	struct ice_hw_port_stats *ps = &pf->stats;
+
+	rmon->undersize_pkts	= ps->rx_undersize;
+	rmon->oversize_pkts	= ps->rx_oversize;
+	rmon->fragments		= ps->rx_fragments;
+	rmon->jabbers		= ps->rx_jabber;
+
+	rmon->hist[0]		= ps->rx_size_64;
+	rmon->hist[1]		= ps->rx_size_127;
+	rmon->hist[2]		= ps->rx_size_255;
+	rmon->hist[3]		= ps->rx_size_511;
+	rmon->hist[4]		= ps->rx_size_1023;
+	rmon->hist[5]		= ps->rx_size_1522;
+	rmon->hist[6]		= ps->rx_size_big;
+
+	rmon->hist_tx[0]	= ps->tx_size_64;
+	rmon->hist_tx[1]	= ps->tx_size_127;
+	rmon->hist_tx[2]	= ps->tx_size_255;
+	rmon->hist_tx[3]	= ps->tx_size_511;
+	rmon->hist_tx[4]	= ps->tx_size_1023;
+	rmon->hist_tx[5]	= ps->tx_size_1522;
+	rmon->hist_tx[6]	= ps->tx_size_big;
+
+	*ranges = ice_rmon_ranges;
+}
+
 static const struct ethtool_ops ice_ethtool_ops = {
 	.cap_rss_ctx_supported  = true,
 	.supported_coalesce_params = ETHTOOL_COALESCE_USECS |
@@ -4329,6 +4404,9 @@ static const struct ethtool_ops ice_ethtool_ops = {
 	.set_fecparam		= ice_set_fecparam,
 	.get_module_info	= ice_get_module_info,
 	.get_module_eeprom	= ice_get_module_eeprom,
+	.get_eth_mac_stats	= ice_get_eth_mac_stats,
+	.get_pause_stats	= ice_get_pause_stats,
+	.get_rmon_stats		= ice_get_rmon_stats,
 };
 
 static const struct ethtool_ops ice_ethtool_safe_mode_ops = {
diff --git a/drivers/net/ethernet/intel/ice/ice_main.c b/drivers/net/ethernet/intel/ice/ice_main.c
index f60c022f7960..4db3a6056f41 100644
--- a/drivers/net/ethernet/intel/ice/ice_main.c
+++ b/drivers/net/ethernet/intel/ice/ice_main.c
@@ -7034,6 +7034,9 @@ void ice_update_pf_stats(struct ice_pf *pf)
 			  &prev_ps->mac_remote_faults,
 			  &cur_ps->mac_remote_faults);
 
+	ice_stat_update32(hw, GLPRT_RLEC(port), pf->stat_prev_loaded,
+			  &prev_ps->rx_len_errors, &cur_ps->rx_len_errors);
+
 	ice_stat_update32(hw, GLPRT_RUC(port), pf->stat_prev_loaded,
 			  &prev_ps->rx_undersize, &cur_ps->rx_undersize);
 
diff --git a/drivers/net/ethernet/intel/ice/ice_type.h b/drivers/net/ethernet/intel/ice/ice_type.h
index f0796a93f428..b156d01196e2 100644
--- a/drivers/net/ethernet/intel/ice/ice_type.h
+++ b/drivers/net/ethernet/intel/ice/ice_type.h
@@ -1007,6 +1007,7 @@ struct ice_hw_port_stats {
 	u64 error_bytes;		/* errbc */
 	u64 mac_local_faults;		/* mlfc */
 	u64 mac_remote_faults;		/* mrfc */
+	u64 rx_len_errors;		/* rlec */
 	u64 link_xon_rx;		/* lxonrxc */
 	u64 link_xoff_rx;		/* lxoffrxc */
 	u64 link_xon_tx;		/* lxontxc */
-- 
2.43.0

Re: [PATCH iwl-next v2 2/5] ice: implement ethtool standard stats

[Simon Horman]

On Thu, Jun 06, 2024 at 03:46:56PM -0700, Jesse Brandeburg wrote:
> Add support for MAC/pause/RMON stats. This enables reporting hardware
> statistics in a common way via:
> 
> ethtool -S eth0 --all-groups
> and
> ethtool --include-statistics --show-pause eth0
> 
> While doing so, add support for one new stat, receive length error
> (RLEC), which is extremely unlikely to happen since most L2 frames have
> a type/length field specifying a "type", and raw ethernet frames aren't
> used much any longer.
> 
> NOTE: I didn't implement Ctrl aka control frame stats because the
> hardware doesn't seem to implement support.
> 
> Reviewed-by: Marcin Szycik <marcin.szycik@linux.intel.com>
> Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
> Reviewed-by: Jakub Kicinski <kuba@kernel.org>
> Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>

Reviewed-by: Simon Horman <horms@kernel.org>

[PATCH iwl-next v2 3/5] ice: add tracking of good transmit timestamps

[Jesse Brandeburg]

As a pre-requisite to implementing timestamp statistics, start tracking
successful PTP timestamps. There already existed a trace event, but
add a counter as well so it can be displayed by the next patch.

Good count is a u64 as it is much more likely to be incremented. The
existing error stats are all u32 as before, and are less likely so will
wrap less.

Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
Reviewed-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
---
 drivers/net/ethernet/intel/ice/ice_ptp.c | 9 +++++++++
 drivers/net/ethernet/intel/ice/ice_ptp.h | 1 +
 2 files changed, 10 insertions(+)

diff --git a/drivers/net/ethernet/intel/ice/ice_ptp.c b/drivers/net/ethernet/intel/ice/ice_ptp.c
index 0f17fc1181d2..ff0ba81d0694 100644
--- a/drivers/net/ethernet/intel/ice/ice_ptp.c
+++ b/drivers/net/ethernet/intel/ice/ice_ptp.c
@@ -628,6 +628,9 @@ void ice_ptp_complete_tx_single_tstamp(struct ice_ptp_tx *tx)
 	if (tstamp) {
 		shhwtstamps.hwtstamp = ns_to_ktime(tstamp);
 		ice_trace(tx_tstamp_complete, skb, idx);
+
+		/* Count the number of Tx timestamps that succeeded */
+		pf->ptp.tx_hwtstamp_good++;
 	}
 
 	skb_tstamp_tx(skb, &shhwtstamps);
@@ -686,6 +689,7 @@ static void ice_ptp_process_tx_tstamp(struct ice_ptp_tx *tx)
 {
 	struct ice_ptp_port *ptp_port;
 	unsigned long flags;
+	u32 tstamp_good = 0;
 	struct ice_pf *pf;
 	struct ice_hw *hw;
 	u64 tstamp_ready;
@@ -786,11 +790,16 @@ static void ice_ptp_process_tx_tstamp(struct ice_ptp_tx *tx)
 		if (tstamp) {
 			shhwtstamps.hwtstamp = ns_to_ktime(tstamp);
 			ice_trace(tx_tstamp_complete, skb, idx);
+
+			/* Count the number of Tx timestamps that succeeded */
+			tstamp_good++;
 		}
 
 		skb_tstamp_tx(skb, &shhwtstamps);
 		dev_kfree_skb_any(skb);
 	}
+
+	pf->ptp.tx_hwtstamp_good += tstamp_good;
 }
 
 /**
diff --git a/drivers/net/ethernet/intel/ice/ice_ptp.h b/drivers/net/ethernet/intel/ice/ice_ptp.h
index 3af20025043a..2b15f2b58789 100644
--- a/drivers/net/ethernet/intel/ice/ice_ptp.h
+++ b/drivers/net/ethernet/intel/ice/ice_ptp.h
@@ -253,6 +253,7 @@ struct ice_ptp {
 	struct ptp_clock *clock;
 	struct hwtstamp_config tstamp_config;
 	u64 reset_time;
+	u64 tx_hwtstamp_good;
 	u32 tx_hwtstamp_skipped;
 	u32 tx_hwtstamp_timeouts;
 	u32 tx_hwtstamp_flushed;
-- 
2.43.0

Re: [PATCH iwl-next v2 3/5] ice: add tracking of good transmit timestamps

[Simon Horman]

On Thu, Jun 06, 2024 at 03:46:57PM -0700, Jesse Brandeburg wrote:
> As a pre-requisite to implementing timestamp statistics, start tracking
> successful PTP timestamps. There already existed a trace event, but
> add a counter as well so it can be displayed by the next patch.
> 
> Good count is a u64 as it is much more likely to be incremented. The
> existing error stats are all u32 as before, and are less likely so will
> wrap less.
> 
> Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
> Reviewed-by: Jakub Kicinski <kuba@kernel.org>
> Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>

Hi Jesse,

The minor nit below notwithstanding, this looks good for me.

Reviewed-by: Simon Horman <horms@kernel.org>

...

> diff --git a/drivers/net/ethernet/intel/ice/ice_ptp.h b/drivers/net/ethernet/intel/ice/ice_ptp.h
> index 3af20025043a..2b15f2b58789 100644
> --- a/drivers/net/ethernet/intel/ice/ice_ptp.h
> +++ b/drivers/net/ethernet/intel/ice/ice_ptp.h
> @@ -253,6 +253,7 @@ struct ice_ptp {
>  	struct ptp_clock *clock;
>  	struct hwtstamp_config tstamp_config;
>  	u64 reset_time;
> +	u64 tx_hwtstamp_good;

nit: There should be an entry for @tx_hwtstamp_good added to the Kernel doc
     for this structure.

Also, not strictly related to this patch, but related to Kernel doc.
It would be very nice, IMHO, if some work could be done to add
Return: sections to Kernel docs for the ice and moreover Intel Wired
Ethernet drivers. These are flagged by kernel-doc -none -Wall,
which was recently enabled for NIPA. And there are a lot of them.

>  	u32 tx_hwtstamp_skipped;
>  	u32 tx_hwtstamp_timeouts;
>  	u32 tx_hwtstamp_flushed;
> -- 
> 2.43.0
> 
> 

[PATCH iwl-next v2 4/5] ice: implement transmit hardware timestamp statistics

[Jesse Brandeburg]

The kernel now has common statistics for transmit timestamps, so
implement them in the ice driver.

use via
ethtool -I -T eth0

Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
Reviewed-by: Jakub Kicinski <kuba@kernel.org>
Reviewed-by: Hariprasad Kelam <hkelam@marvell.com>
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
---

$ sudo ethtool -I -T eth0
Time stamping parameters for eth0:
Capabilities:
        hardware-transmit
        software-transmit
        hardware-receive
        software-receive
        software-system-clock
        hardware-raw-clock
PTP Hardware Clock: 0
Hardware Transmit Timestamp Modes:
        off
        on
Hardware Receive Filter Modes:
        none
        all
Statistics:
  tx_pkts: 17
  tx_lost: 0
  tx_err: 0
---
 drivers/net/ethernet/intel/ice/ice_ethtool.c | 18 ++++++++++++++++++
 1 file changed, 18 insertions(+)

diff --git a/drivers/net/ethernet/intel/ice/ice_ethtool.c b/drivers/net/ethernet/intel/ice/ice_ethtool.c
index 6f0a857f55c9..97a7a0632a1d 100644
--- a/drivers/net/ethernet/intel/ice/ice_ethtool.c
+++ b/drivers/net/ethernet/intel/ice/ice_ethtool.c
@@ -4357,6 +4357,23 @@ static void ice_get_rmon_stats(struct net_device *netdev,
 	*ranges = ice_rmon_ranges;
 }
 
+/* ice_get_ts_stats - provide timestamping stats
+ * @netdev: the netdevice pointer from ethtool
+ * @ts_stats: the ethtool data structure to fill in
+ */
+static void ice_get_ts_stats(struct net_device *netdev,
+			     struct ethtool_ts_stats *ts_stats)
+{
+	struct ice_pf *pf = ice_netdev_to_pf(netdev);
+	struct ice_ptp *ptp = &pf->ptp;
+
+	ts_stats->pkts = ptp->tx_hwtstamp_good;
+	ts_stats->err = ptp->tx_hwtstamp_skipped +
+			ptp->tx_hwtstamp_flushed +
+			ptp->tx_hwtstamp_discarded;
+	ts_stats->lost = ptp->tx_hwtstamp_timeouts;
+}
+
 static const struct ethtool_ops ice_ethtool_ops = {
 	.cap_rss_ctx_supported  = true,
 	.supported_coalesce_params = ETHTOOL_COALESCE_USECS |
@@ -4407,6 +4424,7 @@ static const struct ethtool_ops ice_ethtool_ops = {
 	.get_eth_mac_stats	= ice_get_eth_mac_stats,
 	.get_pause_stats	= ice_get_pause_stats,
 	.get_rmon_stats		= ice_get_rmon_stats,
+	.get_ts_stats		= ice_get_ts_stats,
 };
 
 static const struct ethtool_ops ice_ethtool_safe_mode_ops = {
-- 
2.43.0

Re: [PATCH iwl-next v2 4/5] ice: implement transmit hardware timestamp statistics

[Simon Horman]

On Thu, Jun 06, 2024 at 03:46:58PM -0700, Jesse Brandeburg wrote:
> The kernel now has common statistics for transmit timestamps, so
> implement them in the ice driver.
> 
> use via
> ethtool -I -T eth0
> 
> Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
> Reviewed-by: Jakub Kicinski <kuba@kernel.org>
> Reviewed-by: Hariprasad Kelam <hkelam@marvell.com>
> Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>

Reviewed-by: Simon Horman <horms@kernel.org>

[PATCH iwl-next v2 5/5] ice: refactor to use helpers

[Jesse Brandeburg]

Use the ice_netdev_to_pf() helper in more places and remove a bunch of
boilerplate code. Not every instance could be replaced due to use of the
netdev_priv() output or the vsi variable within a bunch of functions.

Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
Reviewed-by: Przemek Kitszel <przemyslaw.kitszel@intel.com>
Reviewed-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
---
 drivers/net/ethernet/intel/ice/ice_ethtool.c  | 42 ++++++-------------
 .../net/ethernet/intel/ice/ice_flex_pipe.c    |  8 +---
 drivers/net/ethernet/intel/ice/ice_lag.c      |  5 +--
 drivers/net/ethernet/intel/ice/ice_main.c     |  7 +---
 drivers/net/ethernet/intel/ice/ice_sriov.c    |  3 +-
 5 files changed, 19 insertions(+), 46 deletions(-)

diff --git a/drivers/net/ethernet/intel/ice/ice_ethtool.c b/drivers/net/ethernet/intel/ice/ice_ethtool.c
index 97a7a0632a1d..2d307e7d9863 100644
--- a/drivers/net/ethernet/intel/ice/ice_ethtool.c
+++ b/drivers/net/ethernet/intel/ice/ice_ethtool.c
@@ -469,8 +469,7 @@ static int ice_get_regs_len(struct net_device __always_unused *netdev)
 static void
 ice_get_regs(struct net_device *netdev, struct ethtool_regs *regs, void *p)
 {
-	struct ice_netdev_priv *np = netdev_priv(netdev);
-	struct ice_pf *pf = np->vsi->back;
+	struct ice_pf *pf = ice_netdev_to_pf(netdev);
 	struct ice_hw *hw = &pf->hw;
 	u32 *regs_buf = (u32 *)p;
 	unsigned int i;
@@ -483,8 +482,7 @@ ice_get_regs(struct net_device *netdev, struct ethtool_regs *regs, void *p)
 
 static u32 ice_get_msglevel(struct net_device *netdev)
 {
-	struct ice_netdev_priv *np = netdev_priv(netdev);
-	struct ice_pf *pf = np->vsi->back;
+	struct ice_pf *pf = ice_netdev_to_pf(netdev);
 
 #ifndef CONFIG_DYNAMIC_DEBUG
 	if (pf->hw.debug_mask)
@@ -497,8 +495,7 @@ static u32 ice_get_msglevel(struct net_device *netdev)
 
 static void ice_set_msglevel(struct net_device *netdev, u32 data)
 {
-	struct ice_netdev_priv *np = netdev_priv(netdev);
-	struct ice_pf *pf = np->vsi->back;
+	struct ice_pf *pf = ice_netdev_to_pf(netdev);
 
 #ifndef CONFIG_DYNAMIC_DEBUG
 	if (ICE_DBG_USER & data)
@@ -512,8 +509,7 @@ static void ice_set_msglevel(struct net_device *netdev, u32 data)
 
 static int ice_get_eeprom_len(struct net_device *netdev)
 {
-	struct ice_netdev_priv *np = netdev_priv(netdev);
-	struct ice_pf *pf = np->vsi->back;
+	struct ice_pf *pf = ice_netdev_to_pf(netdev);
 
 	return (int)pf->hw.flash.flash_size;
 }
@@ -522,9 +518,7 @@ static int
 ice_get_eeprom(struct net_device *netdev, struct ethtool_eeprom *eeprom,
 	       u8 *bytes)
 {
-	struct ice_netdev_priv *np = netdev_priv(netdev);
-	struct ice_vsi *vsi = np->vsi;
-	struct ice_pf *pf = vsi->back;
+	struct ice_pf *pf = ice_netdev_to_pf(netdev);
 	struct ice_hw *hw = &pf->hw;
 	struct device *dev;
 	int ret;
@@ -623,8 +617,7 @@ static u64 ice_link_test(struct net_device *netdev)
  */
 static u64 ice_eeprom_test(struct net_device *netdev)
 {
-	struct ice_netdev_priv *np = netdev_priv(netdev);
-	struct ice_pf *pf = np->vsi->back;
+	struct ice_pf *pf = ice_netdev_to_pf(netdev);
 
 	netdev_info(netdev, "EEPROM test\n");
 	return !!(ice_nvm_validate_checksum(&pf->hw));
@@ -938,9 +931,8 @@ static int ice_lbtest_receive_frames(struct ice_rx_ring *rx_ring)
  */
 static u64 ice_loopback_test(struct net_device *netdev)
 {
-	struct ice_netdev_priv *np = netdev_priv(netdev);
-	struct ice_vsi *orig_vsi = np->vsi, *test_vsi;
-	struct ice_pf *pf = orig_vsi->back;
+	struct ice_pf *pf = ice_netdev_to_pf(netdev);
+	struct ice_vsi *test_vsi;
 	u8 *tx_frame __free(kfree) = NULL;
 	u8 broadcast[ETH_ALEN], ret = 0;
 	int num_frames, valid_frames;
@@ -1029,8 +1021,7 @@ static u64 ice_loopback_test(struct net_device *netdev)
  */
 static u64 ice_intr_test(struct net_device *netdev)
 {
-	struct ice_netdev_priv *np = netdev_priv(netdev);
-	struct ice_pf *pf = np->vsi->back;
+	struct ice_pf *pf = ice_netdev_to_pf(netdev);
 	u16 swic_old = pf->sw_int_count;
 
 	netdev_info(netdev, "interrupt test\n");
@@ -1058,9 +1049,8 @@ static void
 ice_self_test(struct net_device *netdev, struct ethtool_test *eth_test,
 	      u64 *data)
 {
-	struct ice_netdev_priv *np = netdev_priv(netdev);
+	struct ice_pf *pf = ice_netdev_to_pf(netdev);
 	bool if_running = netif_running(netdev);
-	struct ice_pf *pf = np->vsi->back;
 	struct device *dev;
 
 	dev = ice_pf_to_dev(pf);
@@ -1384,9 +1374,7 @@ static int ice_nway_reset(struct net_device *netdev)
  */
 static u32 ice_get_priv_flags(struct net_device *netdev)
 {
-	struct ice_netdev_priv *np = netdev_priv(netdev);
-	struct ice_vsi *vsi = np->vsi;
-	struct ice_pf *pf = vsi->back;
+	struct ice_pf *pf = ice_netdev_to_pf(netdev);
 	u32 i, ret_flags = 0;
 
 	for (i = 0; i < ICE_PRIV_FLAG_ARRAY_SIZE; i++) {
@@ -4128,9 +4116,7 @@ static int
 ice_get_module_info(struct net_device *netdev,
 		    struct ethtool_modinfo *modinfo)
 {
-	struct ice_netdev_priv *np = netdev_priv(netdev);
-	struct ice_vsi *vsi = np->vsi;
-	struct ice_pf *pf = vsi->back;
+	struct ice_pf *pf = ice_netdev_to_pf(netdev);
 	struct ice_hw *hw = &pf->hw;
 	u8 sff8472_comp = 0;
 	u8 sff8472_swap = 0;
@@ -4202,12 +4188,10 @@ static int
 ice_get_module_eeprom(struct net_device *netdev,
 		      struct ethtool_eeprom *ee, u8 *data)
 {
-	struct ice_netdev_priv *np = netdev_priv(netdev);
+	struct ice_pf *pf = ice_netdev_to_pf(netdev);
 #define SFF_READ_BLOCK_SIZE 8
 	u8 value[SFF_READ_BLOCK_SIZE] = { 0 };
 	u8 addr = ICE_I2C_EEPROM_DEV_ADDR;
-	struct ice_vsi *vsi = np->vsi;
-	struct ice_pf *pf = vsi->back;
 	struct ice_hw *hw = &pf->hw;
 	bool is_sfp = false;
 	unsigned int i, j;
diff --git a/drivers/net/ethernet/intel/ice/ice_flex_pipe.c b/drivers/net/ethernet/intel/ice/ice_flex_pipe.c
index 20d5db88c99f..4c322bed716c 100644
--- a/drivers/net/ethernet/intel/ice/ice_flex_pipe.c
+++ b/drivers/net/ethernet/intel/ice/ice_flex_pipe.c
@@ -574,9 +574,7 @@ ice_destroy_tunnel(struct ice_hw *hw, u16 index, enum ice_tunnel_type type,
 int ice_udp_tunnel_set_port(struct net_device *netdev, unsigned int table,
 			    unsigned int idx, struct udp_tunnel_info *ti)
 {
-	struct ice_netdev_priv *np = netdev_priv(netdev);
-	struct ice_vsi *vsi = np->vsi;
-	struct ice_pf *pf = vsi->back;
+	struct ice_pf *pf = ice_netdev_to_pf(netdev);
 	enum ice_tunnel_type tnl_type;
 	int status;
 	u16 index;
@@ -598,9 +596,7 @@ int ice_udp_tunnel_set_port(struct net_device *netdev, unsigned int table,
 int ice_udp_tunnel_unset_port(struct net_device *netdev, unsigned int table,
 			      unsigned int idx, struct udp_tunnel_info *ti)
 {
-	struct ice_netdev_priv *np = netdev_priv(netdev);
-	struct ice_vsi *vsi = np->vsi;
-	struct ice_pf *pf = vsi->back;
+	struct ice_pf *pf = ice_netdev_to_pf(netdev);
 	enum ice_tunnel_type tnl_type;
 	int status;
 
diff --git a/drivers/net/ethernet/intel/ice/ice_lag.c b/drivers/net/ethernet/intel/ice/ice_lag.c
index 1ccb572ce285..cdb0e59aeb26 100644
--- a/drivers/net/ethernet/intel/ice/ice_lag.c
+++ b/drivers/net/ethernet/intel/ice/ice_lag.c
@@ -1640,11 +1640,8 @@ static void ice_lag_chk_disabled_bond(struct ice_lag *lag, void *ptr)
  */
 static void ice_lag_disable_sriov_bond(struct ice_lag *lag)
 {
-	struct ice_netdev_priv *np;
-	struct ice_pf *pf;
+	struct ice_pf *pf = ice_netdev_to_pf(lag->netdev);
 
-	np = netdev_priv(lag->netdev);
-	pf = np->vsi->back;
 	ice_clear_feature_support(pf, ICE_F_SRIOV_LAG);
 }
 
diff --git a/drivers/net/ethernet/intel/ice/ice_main.c b/drivers/net/ethernet/intel/ice/ice_main.c
index 4db3a6056f41..9d852b169ead 100644
--- a/drivers/net/ethernet/intel/ice/ice_main.c
+++ b/drivers/net/ethernet/intel/ice/ice_main.c
@@ -7798,8 +7798,7 @@ static int ice_change_mtu(struct net_device *netdev, int new_mtu)
  */
 static int ice_eth_ioctl(struct net_device *netdev, struct ifreq *ifr, int cmd)
 {
-	struct ice_netdev_priv *np = netdev_priv(netdev);
-	struct ice_pf *pf = np->vsi->back;
+	struct ice_pf *pf = ice_netdev_to_pf(netdev);
 
 	switch (cmd) {
 	case SIOCGHWTSTAMP:
@@ -8027,9 +8026,7 @@ static int
 ice_bridge_getlink(struct sk_buff *skb, u32 pid, u32 seq,
 		   struct net_device *dev, u32 filter_mask, int nlflags)
 {
-	struct ice_netdev_priv *np = netdev_priv(dev);
-	struct ice_vsi *vsi = np->vsi;
-	struct ice_pf *pf = vsi->back;
+	struct ice_pf *pf = ice_netdev_to_pf(dev);
 	u16 bmode;
 
 	bmode = pf->first_sw->bridge_mode;
diff --git a/drivers/net/ethernet/intel/ice/ice_sriov.c b/drivers/net/ethernet/intel/ice/ice_sriov.c
index 067712f4923f..adcc2f967bab 100644
--- a/drivers/net/ethernet/intel/ice/ice_sriov.c
+++ b/drivers/net/ethernet/intel/ice/ice_sriov.c
@@ -1317,8 +1317,7 @@ ice_vf_lan_overflow_event(struct ice_pf *pf, struct ice_rq_event_info *event)
  */
 int ice_set_vf_spoofchk(struct net_device *netdev, int vf_id, bool ena)
 {
-	struct ice_netdev_priv *np = netdev_priv(netdev);
-	struct ice_pf *pf = np->vsi->back;
+	struct ice_pf *pf = ice_netdev_to_pf(netdev);
 	struct ice_vsi *vf_vsi;
 	struct device *dev;
 	struct ice_vf *vf;
-- 
2.43.0

Re: [PATCH iwl-next v2 5/5] ice: refactor to use helpers

[Simon Horman]

On Thu, Jun 06, 2024 at 03:46:59PM -0700, Jesse Brandeburg wrote:
> Use the ice_netdev_to_pf() helper in more places and remove a bunch of
> boilerplate code. Not every instance could be replaced due to use of the
> netdev_priv() output or the vsi variable within a bunch of functions.
> 
> Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
> Reviewed-by: Przemek Kitszel <przemyslaw.kitszel@intel.com>
> Reviewed-by: Jakub Kicinski <kuba@kernel.org>
> Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>

Hi Jesse,

The minor nit below notwithstanding, this looks good to me.

Reviewed-by: Simon Horman <horms@kernel.org>

...

> diff --git a/drivers/net/ethernet/intel/ice/ice_ethtool.c b/drivers/net/ethernet/intel/ice/ice_ethtool.c

...

> @@ -938,9 +931,8 @@ static int ice_lbtest_receive_frames(struct ice_rx_ring *rx_ring)
>   */
>  static u64 ice_loopback_test(struct net_device *netdev)
>  {
> -	struct ice_netdev_priv *np = netdev_priv(netdev);
> -	struct ice_vsi *orig_vsi = np->vsi, *test_vsi;
> -	struct ice_pf *pf = orig_vsi->back;
> +	struct ice_pf *pf = ice_netdev_to_pf(netdev);
> +	struct ice_vsi *test_vsi;
>  	u8 *tx_frame __free(kfree) = NULL;
>  	u8 broadcast[ETH_ALEN], ret = 0;
>  	int num_frames, valid_frames;

nit: If you end up respinning for some other reason,
     please consider arranging things reverse xmas tree order here.

...

RE: [Intel-wired-lan] [PATCH iwl-next v2 0/5] ice: add standard stats

[Pucha, HimasekharX Reddy]

> -----Original Message-----
> From: Intel-wired-lan <intel-wired-lan-bounces@osuosl.org> On Behalf Of Jesse Brandeburg
> Sent: Friday, June 7, 2024 4:17 AM
> To: netdev@vger.kernel.org; intel-wired-lan@lists.osuosl.org
> Cc: linux-doc@vger.kernel.org; corbet@lwn.net
> Subject: [Intel-wired-lan] [PATCH iwl-next v2 0/5] ice: add standard stats
>
> The main point of this series is to implement the standard stats for the ice driver, but also add a related documentation fix and finish the series off with a cleanup that removes a bunch of code.
>
> Changelog:
> v2: address review comments on 1/5 (Jakub) regarding backticks,
>     fix email address for one reviewer in commit message
>     pick up some reviewed-bys from the list
> v1: https://lore.kernel.org/netdev/20240604221327.299184-1-jesse.brandeburg@intel.com/
>
> Jesse Brandeburg (5):
>   net: docs: add missing features that can have stats
>   ice: implement ethtool standard stats
>   ice: add tracking of good transmit timestamps
>   ice: implement transmit hardware timestamp statistics
>   ice: refactor to use helpers
>
>  Documentation/networking/statistics.rst       |   4 +-
>  drivers/net/ethernet/intel/ice/ice_ethtool.c  | 138 ++++++++++++++----
>  .../net/ethernet/intel/ice/ice_flex_pipe.c    |   8 +-
>  drivers/net/ethernet/intel/ice/ice_lag.c      |   5 +-
>  drivers/net/ethernet/intel/ice/ice_main.c     |  10 +-
>  drivers/net/ethernet/intel/ice/ice_ptp.c      |   9 ++
>  drivers/net/ethernet/intel/ice/ice_ptp.h      |   1 +
>  drivers/net/ethernet/intel/ice/ice_sriov.c    |   3 +-
>  drivers/net/ethernet/intel/ice/ice_type.h     |   1 +
>  9 files changed, 132 insertions(+), 47 deletions(-)
>
>
> base-commit: 95cd03f32a1680f693b291da920ab5d3f9d8c5c1
> --
> 2.43.0
>

Not able to see stats with this series of patches.

$ sudo ethtool -I -T eth0
Time stamping parameters for eth0:
Capabilities:
        hardware-transmit
        software-transmit
        hardware-receive
        software-receive
        software-system-clock
        hardware-raw-clock
PTP Hardware Clock: 0
Hardware Transmit Timestamp Modes:
        off
        on
Hardware Receive Filter Modes:
        none
        all