[PATCH net-next 0/3] bnxt: add rx discards stats for oom and netpool

[PATCH net-next 0/3] bnxt: add rx discards stats for oom and netpool

[Jakub Kicinski]

Drivers should avoid silently dropping frames. This set adds two
stats for previously unaccounted events to bnxt - packets dropped
due to allocation failures and packets dropped during emergency
ring polling.

Jakub Kicinski (3):
  bnxt: reorder logic in bnxt_get_stats64()
  bnxt: count packets discarded because of netpoll
  bnxt: count discards due to memory allocation errors

 drivers/net/ethernet/broadcom/bnxt/bnxt.c     | 54 ++++++++++++++-----
 drivers/net/ethernet/broadcom/bnxt/bnxt.h     |  3 ++
 .../net/ethernet/broadcom/bnxt/bnxt_ethtool.c |  7 +++
 3 files changed, 51 insertions(+), 13 deletions(-)

-- 
2.31.1

[PATCH net-next 1/3] bnxt: reorder logic in bnxt_get_stats64()

[Jakub Kicinski]

Saved ring stats and port stats are completely disjoint.
We can reorder getting the old stats and collecting new
stats from rings with reading port stats.

We can also use bnxt_add_prev_stats() instead of doing
a struct assignment.

With that we can use the same code for closed and open
device, next commits will add more stats at the end
of bnxt_get_stats64().

Signed-off-by: Jakub Kicinski <kuba@kernel.org>
---
 drivers/net/ethernet/broadcom/bnxt/bnxt.c | 15 +++++++--------
 1 file changed, 7 insertions(+), 8 deletions(-)

diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt.c b/drivers/net/ethernet/broadcom/bnxt/bnxt.c
index ee66d410c82c..d39449e7b236 100644
--- a/drivers/net/ethernet/broadcom/bnxt/bnxt.c
+++ b/drivers/net/ethernet/broadcom/bnxt/bnxt.c
@@ -10673,14 +10673,8 @@ bnxt_get_stats64(struct net_device *dev, struct rtnl_link_stats64 *stats)
 	 * we check the BNXT_STATE_OPEN flag.
 	 */
 	smp_mb__after_atomic();
-	if (!test_bit(BNXT_STATE_OPEN, &bp->state)) {
-		clear_bit(BNXT_STATE_READ_STATS, &bp->state);
-		*stats = bp->net_stats_prev;
-		return;
-	}
-
-	bnxt_get_ring_stats(bp, stats);
-	bnxt_add_prev_stats(bp, stats);
+	if (!test_bit(BNXT_STATE_OPEN, &bp->state))
+		goto skip_current;
 
 	if (bp->flags & BNXT_FLAG_PORT_STATS) {
 		u64 *rx = bp->port_stats.sw_stats;
@@ -10704,6 +10698,11 @@ bnxt_get_stats64(struct net_device *dev, struct rtnl_link_stats64 *stats)
 			BNXT_GET_TX_PORT_STATS64(tx, tx_fifo_underruns);
 		stats->tx_errors = BNXT_GET_TX_PORT_STATS64(tx, tx_err);
 	}
+
+	bnxt_get_ring_stats(bp, stats);
+skip_current:
+	bnxt_add_prev_stats(bp, stats);
+
 	clear_bit(BNXT_STATE_READ_STATS, &bp->state);
 }
 
-- 
2.31.1

[PATCH net-next 2/3] bnxt: count packets discarded because of netpoll

[Jakub Kicinski]

bnxt may discard packets if Rx completions are consumed
in an attempt to let netpoll make progress. It should be
exteremely rare in practice but nonetheless such events
should be counted.

Since completion ring memory is allocated dynamically use
a similar scheme to what is done for HW stats to save them.

Report the stats in rx_dropped and per-netdev ethtool
counter. Chances that users care which ring dropped are
very low.

Signed-off-by: Jakub Kicinski <kuba@kernel.org>
---
 drivers/net/ethernet/broadcom/bnxt/bnxt.c     | 31 ++++++++++++++-----
 drivers/net/ethernet/broadcom/bnxt/bnxt.h     |  2 ++
 .../net/ethernet/broadcom/bnxt/bnxt_ethtool.c |  6 ++++
 3 files changed, 32 insertions(+), 7 deletions(-)

diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt.c b/drivers/net/ethernet/broadcom/bnxt/bnxt.c
index d39449e7b236..d12a9052388f 100644
--- a/drivers/net/ethernet/broadcom/bnxt/bnxt.c
+++ b/drivers/net/ethernet/broadcom/bnxt/bnxt.c
@@ -2003,6 +2003,7 @@ static int bnxt_force_rx_discard(struct bnxt *bp,
 	struct rx_cmp *rxcmp;
 	u16 cp_cons;
 	u8 cmp_type;
+	int ret;
 
 	cp_cons = RING_CMP(tmp_raw_cons);
 	rxcmp = (struct rx_cmp *)
@@ -2031,7 +2032,10 @@ static int bnxt_force_rx_discard(struct bnxt *bp,
 		tpa_end1->rx_tpa_end_cmp_errors_v2 |=
 			cpu_to_le32(RX_TPA_END_CMP_ERRORS);
 	}
-	return bnxt_rx_pkt(bp, cpr, raw_cons, event);
+	ret = bnxt_rx_pkt(bp, cpr, raw_cons, event);
+	if (ret != -EBUSY)
+		cpr->sw_stats.rx.rx_netpoll_discards += 1;
+	return ret;
 }
 
 u32 bnxt_fw_health_readl(struct bnxt *bp, int reg_idx)
@@ -10441,7 +10445,8 @@ static bool bnxt_drv_busy(struct bnxt *bp)
 }
 
 static void bnxt_get_ring_stats(struct bnxt *bp,
-				struct rtnl_link_stats64 *stats);
+				struct rtnl_link_stats64 *stats,
+				struct bnxt_sw_stats *bsw_stats);
 
 static void __bnxt_close_nic(struct bnxt *bp, bool irq_re_init,
 			     bool link_re_init)
@@ -10470,7 +10475,8 @@ static void __bnxt_close_nic(struct bnxt *bp, bool irq_re_init,
 
 	/* Save ring stats before shutdown */
 	if (bp->bnapi && irq_re_init)
-		bnxt_get_ring_stats(bp, &bp->net_stats_prev);
+		bnxt_get_ring_stats(bp, &bp->net_stats_prev,
+				    &bp->sw_stats_prev);
 	if (irq_re_init) {
 		bnxt_free_irq(bp);
 		bnxt_del_napi(bp);
@@ -10615,7 +10621,8 @@ static int bnxt_ioctl(struct net_device *dev, struct ifreq *ifr, int cmd)
 }
 
 static void bnxt_get_ring_stats(struct bnxt *bp,
-				struct rtnl_link_stats64 *stats)
+				struct rtnl_link_stats64 *stats,
+				struct bnxt_sw_stats *bsw_stats)
 {
 	int i;
 
@@ -10646,11 +10653,15 @@ static void bnxt_get_ring_stats(struct bnxt *bp,
 		stats->multicast += BNXT_GET_RING_STATS64(sw, rx_mcast_pkts);
 
 		stats->tx_dropped += BNXT_GET_RING_STATS64(sw, tx_error_pkts);
+
+		bsw_stats->rx.rx_netpoll_discards +=
+			cpr->sw_stats.rx.rx_netpoll_discards;
 	}
 }
 
 static void bnxt_add_prev_stats(struct bnxt *bp,
-				struct rtnl_link_stats64 *stats)
+				struct rtnl_link_stats64 *stats,
+				struct bnxt_sw_stats *bsw_stats)
 {
 	struct rtnl_link_stats64 *prev_stats = &bp->net_stats_prev;
 
@@ -10661,11 +10672,15 @@ static void bnxt_add_prev_stats(struct bnxt *bp,
 	stats->rx_missed_errors += prev_stats->rx_missed_errors;
 	stats->multicast += prev_stats->multicast;
 	stats->tx_dropped += prev_stats->tx_dropped;
+
+	bsw_stats->rx.rx_netpoll_discards +=
+		bp->sw_stats_prev.rx.rx_netpoll_discards;
 }
 
 static void
 bnxt_get_stats64(struct net_device *dev, struct rtnl_link_stats64 *stats)
 {
+	struct bnxt_sw_stats bsw_stats = {};
 	struct bnxt *bp = netdev_priv(dev);
 
 	set_bit(BNXT_STATE_READ_STATS, &bp->state);
@@ -10699,9 +10714,11 @@ bnxt_get_stats64(struct net_device *dev, struct rtnl_link_stats64 *stats)
 		stats->tx_errors = BNXT_GET_TX_PORT_STATS64(tx, tx_err);
 	}
 
-	bnxt_get_ring_stats(bp, stats);
+	bnxt_get_ring_stats(bp, stats, &bsw_stats);
 skip_current:
-	bnxt_add_prev_stats(bp, stats);
+	bnxt_add_prev_stats(bp, stats, &bsw_stats);
+
+	stats->rx_dropped += bsw_stats.rx.rx_netpoll_discards;
 
 	clear_bit(BNXT_STATE_READ_STATS, &bp->state);
 }
diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt.h b/drivers/net/ethernet/broadcom/bnxt/bnxt.h
index 7b989b6e4f6e..5c2e9a06e959 100644
--- a/drivers/net/ethernet/broadcom/bnxt/bnxt.h
+++ b/drivers/net/ethernet/broadcom/bnxt/bnxt.h
@@ -939,6 +939,7 @@ struct bnxt_rx_sw_stats {
 	u64			rx_l4_csum_errors;
 	u64			rx_resets;
 	u64			rx_buf_errors;
+	u64			rx_netpoll_discards;
 };
 
 struct bnxt_cmn_sw_stats {
@@ -1917,6 +1918,7 @@ struct bnxt {
 	dma_addr_t		hwrm_cmd_kong_resp_dma_addr;
 
 	struct rtnl_link_stats64	net_stats_prev;
+	struct bnxt_sw_stats	sw_stats_prev;
 	struct bnxt_stats_mem	port_stats;
 	struct bnxt_stats_mem	rx_port_stats_ext;
 	struct bnxt_stats_mem	tx_port_stats_ext;
diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c b/drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c
index 9f8c72d95228..25f1327aedb6 100644
--- a/drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c
+++ b/drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c
@@ -307,6 +307,7 @@ static const char * const bnxt_cmn_sw_stats_str[] = {
 enum {
 	RX_TOTAL_DISCARDS,
 	TX_TOTAL_DISCARDS,
+	RX_NETPOLL_DISCARDS,
 };
 
 static struct {
@@ -315,6 +316,7 @@ static struct {
 } bnxt_sw_func_stats[] = {
 	{0, "rx_total_discard_pkts"},
 	{0, "tx_total_discard_pkts"},
+	{0, "rx_netpoll_discards"},
 };
 
 #define NUM_RING_RX_SW_STATS		ARRAY_SIZE(bnxt_rx_sw_stats_str)
@@ -561,6 +563,8 @@ static void bnxt_get_ethtool_stats(struct net_device *dev,
 
 	for (i = 0; i < BNXT_NUM_SW_FUNC_STATS; i++)
 		bnxt_sw_func_stats[i].counter = 0;
+	bnxt_sw_func_stats[RX_NETPOLL_DISCARDS].counter =
+		bp->sw_stats_prev.rx.rx_netpoll_discards;
 
 	tpa_stats = bnxt_get_num_tpa_ring_stats(bp);
 	for (i = 0; i < bp->cp_nr_rings; i++) {
@@ -603,6 +607,8 @@ static void bnxt_get_ethtool_stats(struct net_device *dev,
 			BNXT_GET_RING_STATS64(sw_stats, rx_discard_pkts);
 		bnxt_sw_func_stats[TX_TOTAL_DISCARDS].counter +=
 			BNXT_GET_RING_STATS64(sw_stats, tx_discard_pkts);
+		bnxt_sw_func_stats[RX_NETPOLL_DISCARDS].counter +=
+			cpr->sw_stats.rx.rx_netpoll_discards;
 	}
 
 	for (i = 0; i < BNXT_NUM_SW_FUNC_STATS; i++, j++)
-- 
2.31.1

[PATCH net-next 3/3] bnxt: count discards due to memory allocation errors

[Jakub Kicinski]

Count packets dropped due to buffer or skb allocation errors.
Report as part of rx_dropped, and per-queue in ethtool
(retaining only the former across down/up cycles).

Signed-off-by: Jakub Kicinski <kuba@kernel.org>
---
 drivers/net/ethernet/broadcom/bnxt/bnxt.c         | 14 +++++++++++++-
 drivers/net/ethernet/broadcom/bnxt/bnxt.h         |  1 +
 drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c |  1 +
 3 files changed, 15 insertions(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt.c b/drivers/net/ethernet/broadcom/bnxt/bnxt.c
index d12a9052388f..bdc5eb42f55b 100644
--- a/drivers/net/ethernet/broadcom/bnxt/bnxt.c
+++ b/drivers/net/ethernet/broadcom/bnxt/bnxt.c
@@ -1651,6 +1651,7 @@ static inline struct sk_buff *bnxt_tpa_end(struct bnxt *bp,
 		skb = bnxt_copy_skb(bnapi, data_ptr, len, mapping);
 		if (!skb) {
 			bnxt_abort_tpa(cpr, idx, agg_bufs);
+			cpr->sw_stats.rx.rx_oom_discards += 1;
 			return NULL;
 		}
 	} else {
@@ -1660,6 +1661,7 @@ static inline struct sk_buff *bnxt_tpa_end(struct bnxt *bp,
 		new_data = __bnxt_alloc_rx_data(bp, &new_mapping, GFP_ATOMIC);
 		if (!new_data) {
 			bnxt_abort_tpa(cpr, idx, agg_bufs);
+			cpr->sw_stats.rx.rx_oom_discards += 1;
 			return NULL;
 		}
 
@@ -1675,6 +1677,7 @@ static inline struct sk_buff *bnxt_tpa_end(struct bnxt *bp,
 		if (!skb) {
 			kfree(data);
 			bnxt_abort_tpa(cpr, idx, agg_bufs);
+			cpr->sw_stats.rx.rx_oom_discards += 1;
 			return NULL;
 		}
 		skb_reserve(skb, bp->rx_offset);
@@ -1685,6 +1688,7 @@ static inline struct sk_buff *bnxt_tpa_end(struct bnxt *bp,
 		skb = bnxt_rx_pages(bp, cpr, skb, idx, agg_bufs, true);
 		if (!skb) {
 			/* Page reuse already handled by bnxt_rx_pages(). */
+			cpr->sw_stats.rx.rx_oom_discards += 1;
 			return NULL;
 		}
 	}
@@ -1888,6 +1892,7 @@ static int bnxt_rx_pkt(struct bnxt *bp, struct bnxt_cp_ring_info *cpr,
 			if (agg_bufs)
 				bnxt_reuse_rx_agg_bufs(cpr, cp_cons, 0,
 						       agg_bufs, false);
+			cpr->sw_stats.rx.rx_oom_discards += 1;
 			rc = -ENOMEM;
 			goto next_rx;
 		}
@@ -1901,6 +1906,7 @@ static int bnxt_rx_pkt(struct bnxt *bp, struct bnxt_cp_ring_info *cpr,
 		skb = bp->rx_skb_func(bp, rxr, cons, data, data_ptr, dma_addr,
 				      payload | len);
 		if (!skb) {
+			cpr->sw_stats.rx.rx_oom_discards += 1;
 			rc = -ENOMEM;
 			goto next_rx;
 		}
@@ -1909,6 +1915,7 @@ static int bnxt_rx_pkt(struct bnxt *bp, struct bnxt_cp_ring_info *cpr,
 	if (agg_bufs) {
 		skb = bnxt_rx_pages(bp, cpr, skb, cp_cons, agg_bufs, false);
 		if (!skb) {
+			cpr->sw_stats.rx.rx_oom_discards += 1;
 			rc = -ENOMEM;
 			goto next_rx;
 		}
@@ -10656,6 +10663,8 @@ static void bnxt_get_ring_stats(struct bnxt *bp,
 
 		bsw_stats->rx.rx_netpoll_discards +=
 			cpr->sw_stats.rx.rx_netpoll_discards;
+		bsw_stats->rx.rx_oom_discards +=
+			cpr->sw_stats.rx.rx_oom_discards;
 	}
 }
 
@@ -10675,6 +10684,7 @@ static void bnxt_add_prev_stats(struct bnxt *bp,
 
 	bsw_stats->rx.rx_netpoll_discards +=
 		bp->sw_stats_prev.rx.rx_netpoll_discards;
+	bsw_stats->rx.rx_oom_discards += bp->sw_stats_prev.rx.rx_oom_discards;
 }
 
 static void
@@ -10718,7 +10728,9 @@ bnxt_get_stats64(struct net_device *dev, struct rtnl_link_stats64 *stats)
 skip_current:
 	bnxt_add_prev_stats(bp, stats, &bsw_stats);
 
-	stats->rx_dropped += bsw_stats.rx.rx_netpoll_discards;
+	stats->rx_dropped +=
+		bsw_stats.rx.rx_netpoll_discards +
+		bsw_stats.rx.rx_oom_discards;
 
 	clear_bit(BNXT_STATE_READ_STATS, &bp->state);
 }
diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt.h b/drivers/net/ethernet/broadcom/bnxt/bnxt.h
index 5c2e9a06e959..2f37f03b7e2d 100644
--- a/drivers/net/ethernet/broadcom/bnxt/bnxt.h
+++ b/drivers/net/ethernet/broadcom/bnxt/bnxt.h
@@ -939,6 +939,7 @@ struct bnxt_rx_sw_stats {
 	u64			rx_l4_csum_errors;
 	u64			rx_resets;
 	u64			rx_buf_errors;
+	u64			rx_oom_discards;
 	u64			rx_netpoll_discards;
 };
 
diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c b/drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c
index 25f1327aedb6..f8a28021389b 100644
--- a/drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c
+++ b/drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c
@@ -188,6 +188,7 @@ static const char * const bnxt_rx_sw_stats_str[] = {
 	"rx_l4_csum_errors",
 	"rx_resets",
 	"rx_buf_errors",
+	"rx_oom_discards",
 };
 
 static const char * const bnxt_cmn_sw_stats_str[] = {
-- 
2.31.1

Re: [PATCH net-next 3/3] bnxt: count discards due to memory allocation errors

[Vladimir Oltean]

Hi Jakub,

On Wed, Aug 25, 2021 at 04:18:30PM -0700, Jakub Kicinski wrote:
> Count packets dropped due to buffer or skb allocation errors.
> Report as part of rx_dropped, and per-queue in ethtool
> (retaining only the former across down/up cycles).
> 
> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
> ---
>  drivers/net/ethernet/broadcom/bnxt/bnxt.c         | 14 +++++++++++++-
>  drivers/net/ethernet/broadcom/bnxt/bnxt.h         |  1 +
>  drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c |  1 +
>  3 files changed, 15 insertions(+), 1 deletion(-)
> diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c b/drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c
> index 25f1327aedb6..f8a28021389b 100644
> --- a/drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c
> +++ b/drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c
> @@ -188,6 +188,7 @@ static const char * const bnxt_rx_sw_stats_str[] = {
>  	"rx_l4_csum_errors",
>  	"rx_resets",
>  	"rx_buf_errors",
> +	"rx_oom_discards",

'Could you consider adding "driver" stats under RTM_GETSTATS,
or a similar new structured interface over ethtool?

Looks like the statistic in question has pretty clear semantics,
and may be more broadly useful.'

https://patchwork.ozlabs.org/project/netdev/patch/20201017213611.2557565-2-vladimir.oltean@nxp.com/

>  };
>  
>  static const char * const bnxt_cmn_sw_stats_str[] = {
> -- 
> 2.31.1
> 

Re: [PATCH net-next 3/3] bnxt: count discards due to memory allocation errors

[Jakub Kicinski]

On Thu, 26 Aug 2021 03:22:57 +0300 Vladimir Oltean wrote:
> 'Could you consider adding "driver" stats under RTM_GETSTATS,
> or a similar new structured interface over ethtool?
> 
> Looks like the statistic in question has pretty clear semantics,
> and may be more broadly useful.'

It's commonly reported per ring, I need for make a home for these 
first by adding that damn netlink queue API. It's my next project.

I can drop the ethtool stat from this patch if you have a strong
preference.

Re: [PATCH net-next 3/3] bnxt: count discards due to memory allocation errors

[Vladimir Oltean]

On Wed, Aug 25, 2021 at 05:35:37PM -0700, Jakub Kicinski wrote:
> On Thu, 26 Aug 2021 03:22:57 +0300 Vladimir Oltean wrote:
> > 'Could you consider adding "driver" stats under RTM_GETSTATS,
> > or a similar new structured interface over ethtool?
> >
> > Looks like the statistic in question has pretty clear semantics,
> > and may be more broadly useful.'
>
> It's commonly reported per ring, I need for make a home for these
> first by adding that damn netlink queue API. It's my next project.
>
> I can drop the ethtool stat from this patch if you have a strong
> preference.

I don't have any strong preference, far from it. What would you do if
you were reviewing somebody else's patch which made the same change?

Re: [PATCH net-next 3/3] bnxt: count discards due to memory allocation errors

[Jakub Kicinski]

On Thu, 26 Aug 2021 03:42:08 +0300 Vladimir Oltean wrote:
> On Wed, Aug 25, 2021 at 05:35:37PM -0700, Jakub Kicinski wrote:
> > On Thu, 26 Aug 2021 03:22:57 +0300 Vladimir Oltean wrote:  
> > > 'Could you consider adding "driver" stats under RTM_GETSTATS,
> > > or a similar new structured interface over ethtool?
> > >
> > > Looks like the statistic in question has pretty clear semantics,
> > > and may be more broadly useful.'  
> >
> > It's commonly reported per ring, I need for make a home for these
> > first by adding that damn netlink queue API. It's my next project.
> >
> > I can drop the ethtool stat from this patch if you have a strong
> > preference.  
> 
> I don't have any strong preference, far from it. What would you do if
> you were reviewing somebody else's patch which made the same change?

If someone else posted this patch I'd probably not complain, as I said
there is no well suited API, and my knee jerk expectation was it should
be reported in the per-queue API which doesn't exist.

When you'd seem me complain is when drivers expose in -S stats which
have proper APIs or when higher layer/common code is trying to piggy
back on -S instead of creating its own structured interface.

I don't see value in tracking this particular statistic in production
settings, maybe that's also affecting my judgment here. But since
that's the case I'll just drop it.


If you have any feedback on my suggestions, reviews, comments etc.
please do share on- or off-list at any time. No need to wait a year
until I post a vaguely similar patch ;)

Re: [PATCH net-next 3/3] bnxt: count discards due to memory allocation errors

[Vladimir Oltean]

On Wed, Aug 25, 2021 at 06:44:51PM -0700, Jakub Kicinski wrote:
> On Thu, 26 Aug 2021 03:42:08 +0300 Vladimir Oltean wrote:
> > On Wed, Aug 25, 2021 at 05:35:37PM -0700, Jakub Kicinski wrote:
> > > On Thu, 26 Aug 2021 03:22:57 +0300 Vladimir Oltean wrote:
> > > > 'Could you consider adding "driver" stats under RTM_GETSTATS,
> > > > or a similar new structured interface over ethtool?
> > > >
> > > > Looks like the statistic in question has pretty clear semantics,
> > > > and may be more broadly useful.'
> > >
> > > It's commonly reported per ring, I need for make a home for these
> > > first by adding that damn netlink queue API. It's my next project.
> > >
> > > I can drop the ethtool stat from this patch if you have a strong
> > > preference.
> >
> > I don't have any strong preference, far from it. What would you do if
> > you were reviewing somebody else's patch which made the same change?
>
> If someone else posted this patch I'd probably not complain, as I said
> there is no well suited API, and my knee jerk expectation was it should
> be reported in the per-queue API which doesn't exist.
>
> When you'd seem me complain is when drivers expose in -S stats which
> have proper APIs or when higher layer/common code is trying to piggy
> back on -S instead of creating its own structured interface.
>
> I don't see value in tracking this particular statistic in production
> settings, maybe that's also affecting my judgment here. But since
> that's the case I'll just drop it.
>
>
> If you have any feedback on my suggestions, reviews, comments etc.
> please do share on- or off-list at any time. No need to wait a year
> until I post a vaguely similar patch ;)

I don't know why you get the impression that "I waited a year until you
posted a vaguely similar patch". I am not following you, it just happens
that I was online and reading netdev when you posted this change now.
From the experience of threads that I directly participated in (and this
is why I dug up a DSA thread from a year ago, that was the one I could
find the quickest, again I am not watching your footsteps but
statistically speaking, it would be unlikely for the threads I
participated in to be the only ones where you've said this), you do seem
to tell people to try and use more "generic" and "structured" methods of
statistics reporting as opposed to putting everything in the plain
"ethtool -S", even if those methods don't exist or don't work for that
particular driver and would require major rework (like ndo_get_stats64
which is non-sleepable).

The 'driver stats under RTM_GETSTATS' was a direct quote exactly for
this reason. Now if this rx_oom_discards counter would be better expressed
as a generic 'driver counter' or a 'per-queue counter', none of which exist,
I don't know/don't care. I do wonder sometimes if you think about what
is the people's reaction when you tell them that ethtool -S is not fine
and they should use a kernel interface which doesn't exist, and I was
just curious to see what would yours be.

To create a new kernel interface for statistics would need not only the
vision, but also the passion and dedication to stick to those patches.
People will generally lack the desire to do that, because for better or
worse, "ethtool -S" is the central place to diagnose interface-level
problems. You've also expressed this clearer than words can say by
sending a patch to extend an interface you don't like.

In fact, my message seems to have hit quite the wrong way. I did not
want you to drop the counter from ethtool -S, please keep it if you want
it, but to sway you towards a more relaxed attitude when reviewing
patches for new counters added through that interface. Heck, I would
even like to resubmit the ethtool -S realloc counters if they had any
chance of getting accepted, it's not as if I had any serious intention
of extending the statistics reporting interfaces for something that minor.