Skip to contentSkip to conversation

Expand byte counters in struct station_info

RFC In review Vladimir Kondratiev 3 participants 7 messages Jan 31, 2013 – Feb 4, 2013

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

Hi,

Now wifi drivers reports per-station info using struct station_info; and currently for the data counters it has:

	u32 rx_bytes;
	u32 tx_bytes;

while for device-wide statistics one can use ndo_get_stats64() to fill 64-bit counters in the struct rtnl_link_stats64, per-station statistics are 32-bit.

This becomes problematic with gigabit speeds now observed for .11ac and .11ad - counters overflown every few seconds.

I'd like to extend rx and tx byte counters to 64-bit.

What is better - expand existing fields in struct station_info as:

	u64 rx_bytes;
	u64 tx_bytes;
or add ne ones like:
	u64 rx_bytes64;
	u64 tx_bytes64;

Then, I'll add

	NL80211_STA_INFO_RX_BYTES64,
	NL80211_STA_INFO_TX_BYTES64,

to the enum nl80211_sta_info

Before doing patch, I'd like to hear comments.
Any consideration why is this not to be done or done differently?

Thanks, Vladimir

maintainer
#1 lore raw
Quoted 11 lines · Vladimir Kondratiev

Hi,

Now wifi drivers reports per-station info using struct station_info; and currently for the data counters it has:

	u32 rx_bytes;
	u32 tx_bytes;

while for device-wide statistics one can use ndo_get_stats64() to fill 64-bit counters in the struct rtnl_link_stats64, per-station statistics are 32-bit.

Quoted 8 lines

This becomes problematic with gigabit speeds now observed for .11ac and .11ad - counters overflown every few seconds.

I'd like to extend rx and tx byte counters to 64-bit.

What is better - expand existing fields in struct station_info as:

	u64 rx_bytes;
	u64 tx_bytes;

This.

Quoted 3 lines

or add ne ones like:

	u64 rx_bytes64;
	u64 tx_bytes64;

I don't see a reason to do this.

Quoted 8 lines

Then, I'll add

	NL80211_STA_INFO_RX_BYTES64,
	NL80211_STA_INFO_TX_BYTES64,

to the enum nl80211_sta_info

Before doing patch, I'd like to hear comments.
Any consideration why is this not to be done or done differently?

Sounds good to me.

Two points:

1) You should provide the RX_TX_BYTES attributes, but I think only if the value fits into 32 bits. That way, we don't report invalid information.

2a) Need to be careful in downstream drivers/mac80211, they should also be converted to use 64-bit counters when gathering the data. Otherwise, userspace might assume the value is actually 64-bit, when it rolled over in mac80211.

2b) An alternative to converting all the non-mac80211 drivers and mac80211 would be to add new station info flags: STATION_INFO_TX_BYTES64 and STATION_INFO_RX_BYTES64. You'd still fill the rx_bytes/tx_bytes values, but if the driver sets STATION_INFO_TX_BYTES you'd never use a 64-bit attribute in nl80211.

I think 2a is preferable, but it'd be a bunch of work to make sure to catch all drivers, and some devices might not actually have 64-bit counters (if the data comes from firmware)? So 2b might be a better choice.

johannes

#2 lore raw
Quoted 38 lines · Johannes Berg
Quoted 11 lines · Vladimir Kondratiev

Hi,

Now wifi drivers reports per-station info using struct station_info; and currently for the data counters it has:

	u32 rx_bytes;
	u32 tx_bytes;

while for device-wide statistics one can use ndo_get_stats64() to fill 64-bit counters in the struct rtnl_link_stats64, per-station statistics are 32-bit.

Quoted 8 lines

This becomes problematic with gigabit speeds now observed for .11ac and .11ad - counters overflown every few seconds.

I'd like to extend rx and tx byte counters to 64-bit.

What is better - expand existing fields in struct station_info as:

	u64 rx_bytes;
	u64 tx_bytes;

This.

Quoted 3 lines

or add ne ones like:

	u64 rx_bytes64;
	u64 tx_bytes64;

I don't see a reason to do this.

Quoted 8 lines

Then, I'll add

	NL80211_STA_INFO_RX_BYTES64,
	NL80211_STA_INFO_TX_BYTES64,

to the enum nl80211_sta_info

Before doing patch, I'd like to hear comments.
Any consideration why is this not to be done or done differently?

Quoted 7 lines

Sounds good to me.

Two points:

1) You should provide the RX_TX_BYTES attributes, but I think only if the value fits into 32 bits. That way, we don't report invalid information.

There are already a large amounts of '64-bit' counters reported by netlink that really wrap at 32-bits, so any solid user-space
code already has to deal with this by making assumptions (ie, if cur value is < last value, then it wrapped at 32-bits, because there is no way you actually missed polling a real 64-bit counter wrap).

So, you might not have to worry about this so much.

In my own work, my biggest problem is that stations reset their counters when they got ifdown and then back to ifup.

Thanks,
Ben

Signature · 2 lines
Ben Greear <[email protected]>
Candela Technologies Inc  http://www.candelatech.com
#3 lore raw
In reply to Johannes Berg
Quoted 57 lines · Johannes Berg
Quoted 11 lines · Vladimir Kondratiev

Hi,

Now wifi drivers reports per-station info using struct station_info; and currently for the data counters it has:
u32 rx_bytes;
u32 tx_bytes;

while for device-wide statistics one can use ndo_get_stats64() to fill 64-bit counters in the struct rtnl_link_stats64, per-station statistics are 32-bit.

Quoted 8 lines

This becomes problematic with gigabit speeds now observed for .11ac and .11ad - counters overflown every few seconds.

I'd like to extend rx and tx byte counters to 64-bit.

What is better - expand existing fields in struct station_info as:
u64 rx_bytes;
u64 tx_bytes;

This.

Quoted 3 lines

or add ne ones like:
u64 rx_bytes64;
u64 tx_bytes64;

I don't see a reason to do this.

Quoted 8 lines

Then, I'll add
NL80211_STA_INFO_RX_BYTES64,
NL80211_STA_INFO_TX_BYTES64,
to the enum nl80211_sta_info

Before doing patch, I'd like to hear comments.
Any consideration why is this not to be done or done differently?

Sounds good to me.

Two points:

1) You should provide the RX_TX_BYTES attributes, but I think only if the value fits into 32 bits. That way, we don't report invalid information.

2a) Need to be careful in downstream drivers/mac80211, they should also be converted to use 64-bit counters when gathering the data. Otherwise, userspace might assume the value is actually 64-bit, when it rolled over in mac80211.

2b) An alternative to converting all the non-mac80211 drivers and mac80211 would be to add new station info flags: STATION_INFO_TX_BYTES64 and STATION_INFO_RX_BYTES64. You'd still fill the rx_bytes/tx_bytes values, but if the driver sets STATION_INFO_TX_BYTES you'd never use a 64-bit attribute in nl80211.

Quoted 7 lines

I think 2a is preferable, but it'd be a bunch of work to make sure to catch all drivers, and some devices might not actually have 64-bit counters (if the data comes from firmware)? So 2b might be a better choice.

johannes

So, it may be like the following, patch 1 expands counters and patch 2 uses this 64-bit counters for ath6kl - only easy to find driver with 64-bit counters.

Comments?

Quoted 1 line

From e08b764685cc8887b36f97e2f7957498dbc48b60 Mon Sep 17 00:00:00 2001

From: Vladimir Kondratiev <[email protected]>
Date: Fri, 1 Feb 2013 16:01:35 +0200
Subject: [PATCH 1/2] wireless: expand per-station byte counters to 64bit

In per-station statistics, present 32bit counters are too small for practical purposes - with gigabit speeds, it get overlapped every few seconds.

Expand counters in the struct station_info to be 64-bit. Driver can still fill only 32-bit and indicate in @filled only bits like STATION_INFO_[TR]X_BYTES; in case driver provides full 64-bit counter, it should also set in @filled
bit STATION_INFO_[TR]RX_BYTES64

Netlink sends both 32-bit and 64-bit counters, if present, to not break user space.

---
 include/net/cfg80211.h       |   12 ++++++++----
 include/uapi/linux/nl80211.h |    4 ++++
 net/wireless/nl80211.c       |   10 +++++++++-
 3 files changed, 21 insertions(+), 5 deletions(-)
include/net/cfg80211.h +8 -4
index 36e076e..bcb6b70 100644
@@ -672,8 +672,10 @@ struct station_parameters {
  * @STATION_INFO_SIGNAL: @signal filled
  * @STATION_INFO_TX_BITRATE: @txrate fields are filled
  *  (tx_bitrate, tx_bitrate_flags and tx_bitrate_mcs)
- * @STATION_INFO_RX_PACKETS: @rx_packets filled
- * @STATION_INFO_TX_PACKETS: @tx_packets filled
+ * @STATION_INFO_RX_PACKETS: @rx_packets filled with 32-bit value
+ * @STATION_INFO_TX_PACKETS: @tx_packets filled with 32-bit value
+ * @STATION_INFO_RX_PACKETS64: @rx_packets filled with 64-bit value
+ * @STATION_INFO_TX_PACKETS64: @tx_packets filled with 64-bit value
  * @STATION_INFO_TX_RETRIES: @tx_retries filled
  * @STATION_INFO_TX_FAILED: @tx_failed filled
  * @STATION_INFO_RX_DROP_MISC: @rx_dropped_misc filled
@@ -714,6 +716,8 @@ enum station_info_flags {
 	STATION_INFO_LOCAL_PM		= 1<<21,
 	STATION_INFO_PEER_PM		= 1<<22,
 	STATION_INFO_NONPEER_PM		= 1<<23,
+	STATION_INFO_RX_BYTES64		= 1<<24,
+	STATION_INFO_TX_BYTES64		= 1<<25,
 };
 
 /**
@@ -835,8 +839,8 @@ struct station_info {
 	u32 filled;
 	u32 connected_time;
 	u32 inactive_time;
-	u32 rx_bytes;
-	u32 tx_bytes;
+	u64 rx_bytes;
+	u64 tx_bytes;
 	u16 llid;
 	u16 plid;
 	u8 plink_state;
include/uapi/linux/nl80211.h +4 -0
index 5b7dbc1..5c7d54c 100644
@@ -1851,6 +1851,8 @@ enum nl80211_sta_bss_param {
  * @NL80211_STA_INFO_INACTIVE_TIME: time since last activity (u32, msecs)
  * @NL80211_STA_INFO_RX_BYTES: total received bytes (u32, from this station)
  * @NL80211_STA_INFO_TX_BYTES: total transmitted bytes (u32, to this station)
+ * @NL80211_STA_INFO_RX_BYTES64: total received bytes (u64, from this station)
+ * @NL80211_STA_INFO_TX_BYTES64: total transmitted bytes (u64, to this station)
  * @NL80211_STA_INFO_SIGNAL: signal strength of last received PPDU (u8, dBm)
  * @NL80211_STA_INFO_TX_BITRATE: current unicast tx rate, nested attribute
  * 	containing info as possible, see &enum nl80211_rate_info
@@ -1903,6 +1905,8 @@ enum nl80211_sta_info {
 	NL80211_STA_INFO_LOCAL_PM,
 	NL80211_STA_INFO_PEER_PM,
 	NL80211_STA_INFO_NONPEER_PM,
+	NL80211_STA_INFO_RX_BYTES64,
+	NL80211_STA_INFO_TX_BYTES64,
 
 	/* keep last */
 	__NL80211_STA_INFO_AFTER_LAST,
net/wireless/nl80211.c +9 -1
index b5978ab..b3a9c8e 100644
@@ -3059,10 +3059,18 @@ static int nl80211_send_station(struct sk_buff *msg, u32 portid, u32 seq,
 		goto nla_put_failure;
 	if ((sinfo->filled & STATION_INFO_RX_BYTES) &&
 	    nla_put_u32(msg, NL80211_STA_INFO_RX_BYTES,
-			sinfo->rx_bytes))
+			(u32)sinfo->rx_bytes))
 		goto nla_put_failure;
 	if ((sinfo->filled & STATION_INFO_TX_BYTES) &&
 	    nla_put_u32(msg, NL80211_STA_INFO_TX_BYTES,
+			(u32)sinfo->tx_bytes))
+		goto nla_put_failure;
+	if ((sinfo->filled & STATION_INFO_RX_BYTES64) &&
+	    nla_put_u64(msg, NL80211_STA_INFO_RX_BYTES64,
+			sinfo->rx_bytes))
+		goto nla_put_failure;
+	if ((sinfo->filled & STATION_INFO_TX_BYTES64) &&
+	    nla_put_u32(msg, NL80211_STA_INFO_TX_BYTES64,
 			sinfo->tx_bytes))
 		goto nla_put_failure;
 	if ((sinfo->filled & STATION_INFO_LLID) &&
Signature · 1 line
1.7.10.4
Quoted 1 line

From ec3c391673d70078008213d6243881919c977b07 Mon Sep 17 00:00:00 2001

From: Vladimir Kondratiev <[email protected]>
Date: Fri, 1 Feb 2013 16:03:34 +0200
Subject: [PATCH 2/2] ath6kl: provide 64-bit per-station byte counters

Internally, 64-bit byte counters maintained for per-station statistics. Tell to the netlink that full 64-bit value provided

---
 drivers/net/wireless/ath/ath6kl/cfg80211.c |    6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)
drivers/net/wireless/ath/ath6kl/cfg80211.c +4 -2
index 4225cca..90b1a93 100644
@@ -1778,14 +1778,16 @@ static int ath6kl_get_station(struct wiphy *wiphy, struct net_device *dev,
 
 	if (vif->target_stats.rx_byte) {
 		sinfo->rx_bytes = vif->target_stats.rx_byte;
-		sinfo->filled |= STATION_INFO_RX_BYTES;
+		sinfo->filled |= STATION_INFO_RX_BYTES |
+				 STATION_INFO_RX_BYTES64;
 		sinfo->rx_packets = vif->target_stats.rx_pkt;
 		sinfo->filled |= STATION_INFO_RX_PACKETS;
 	}
 
 	if (vif->target_stats.tx_byte) {
 		sinfo->tx_bytes = vif->target_stats.tx_byte;
-		sinfo->filled |= STATION_INFO_TX_BYTES;
+		sinfo->filled |= STATION_INFO_TX_BYTES |
+				 STATION_INFO_TX_BYTES64;
 		sinfo->tx_packets = vif->target_stats.tx_pkt;
 		sinfo->filled |= STATION_INFO_TX_PACKETS;
 	}
Signature · 1 line
1.7.10.4
#4 lore raw

cover wireless: expand per-station byte counters to 64bit

In reply to Johannes Berg
Quoted 57 lines · Johannes Berg
Quoted 11 lines · Vladimir Kondratiev

Hi,

Now wifi drivers reports per-station info using struct station_info; and currently for the data counters it has:

	u32 rx_bytes;
	u32 tx_bytes;

while for device-wide statistics one can use ndo_get_stats64() to fill 64-bit counters in the struct rtnl_link_stats64, per-station statistics are 32-bit.

Quoted 8 lines

This becomes problematic with gigabit speeds now observed for .11ac and .11ad - counters overflown every few seconds.

I'd like to extend rx and tx byte counters to 64-bit.

What is better - expand existing fields in struct station_info as:

	u64 rx_bytes;
	u64 tx_bytes;

This.

Quoted 3 lines

or add ne ones like:

	u64 rx_bytes64;
	u64 tx_bytes64;

I don't see a reason to do this.

Quoted 8 lines

Then, I'll add

	NL80211_STA_INFO_RX_BYTES64,
	NL80211_STA_INFO_TX_BYTES64,

to the enum nl80211_sta_info

Before doing patch, I'd like to hear comments.
Any consideration why is this not to be done or done differently?

Sounds good to me.

Two points:

1) You should provide the RX_TX_BYTES attributes, but I think only if the value fits into 32 bits. That way, we don't report invalid information.

2a) Need to be careful in downstream drivers/mac80211, they should also be converted to use 64-bit counters when gathering the data. Otherwise, userspace might assume the value is actually 64-bit, when it rolled over in mac80211.

2b) An alternative to converting all the non-mac80211 drivers and mac80211 would be to add new station info flags: STATION_INFO_TX_BYTES64 and STATION_INFO_RX_BYTES64. You'd still fill the rx_bytes/tx_bytes values, but if the driver sets STATION_INFO_TX_BYTES you'd never use a 64-bit attribute in nl80211.

Quoted 7 lines

I think 2a is preferable, but it'd be a bunch of work to make sure to catch all drivers, and some devices might not actually have 64-bit counters (if the data comes from firmware)? So 2b might be a better choice.

johannes

Patches follows.

patch 1 expands counters and patch 2 uses this 64-bit counters for the ath6kl - only easy to find driver with 64-bit counters.

Quoted 1 line

From e08b764685cc8887b36f97e2f7957498dbc48b60 Mon Sep 17 00:00:00 2001

From: Vladimir Kondratiev <[email protected]>
Date: Fri, 1 Feb 2013 16:01:35 +0200
Subject: [PATCH 1/2] wireless: expand per-station byte counters to 64bit

In per-station statistics, present 32bit counters are too small for practical purposes - with gigabit speeds, it get overlapped every few seconds.

Expand counters in the struct station_info to be 64-bit. Driver can still fill only 32-bit and indicate in @filled only bits like STATION_INFO_[TR]X_BYTES; in case driver provides full 64-bit counter, it should also set in @filled
bit STATION_INFO_[TR]RX_BYTES64

Netlink sends both 32-bit and 64-bit counters, if present, to not break user space.

---
 include/net/cfg80211.h       |   12 ++++++++----
 include/uapi/linux/nl80211.h |    4 ++++
 net/wireless/nl80211.c       |   10 +++++++++-
 3 files changed, 21 insertions(+), 5 deletions(-)
include/net/cfg80211.h +8 -4
index 36e076e..bcb6b70 100644
@@ -672,8 +672,10 @@ struct station_parameters {
  * @STATION_INFO_SIGNAL: @signal filled
  * @STATION_INFO_TX_BITRATE: @txrate fields are filled
  *  (tx_bitrate, tx_bitrate_flags and tx_bitrate_mcs)
- * @STATION_INFO_RX_PACKETS: @rx_packets filled
- * @STATION_INFO_TX_PACKETS: @tx_packets filled
+ * @STATION_INFO_RX_PACKETS: @rx_packets filled with 32-bit value
+ * @STATION_INFO_TX_PACKETS: @tx_packets filled with 32-bit value
+ * @STATION_INFO_RX_PACKETS64: @rx_packets filled with 64-bit value
+ * @STATION_INFO_TX_PACKETS64: @tx_packets filled with 64-bit value
  * @STATION_INFO_TX_RETRIES: @tx_retries filled
  * @STATION_INFO_TX_FAILED: @tx_failed filled
  * @STATION_INFO_RX_DROP_MISC: @rx_dropped_misc filled
@@ -714,6 +716,8 @@ enum station_info_flags {
 	STATION_INFO_LOCAL_PM		= 1<<21,
 	STATION_INFO_PEER_PM		= 1<<22,
 	STATION_INFO_NONPEER_PM		= 1<<23,
+	STATION_INFO_RX_BYTES64		= 1<<24,
+	STATION_INFO_TX_BYTES64		= 1<<25,
 };
 
 /**
@@ -835,8 +839,8 @@ struct station_info {
 	u32 filled;
 	u32 connected_time;
 	u32 inactive_time;
-	u32 rx_bytes;
-	u32 tx_bytes;
+	u64 rx_bytes;
+	u64 tx_bytes;
 	u16 llid;
 	u16 plid;
 	u8 plink_state;
include/uapi/linux/nl80211.h +4 -0
index 5b7dbc1..5c7d54c 100644
@@ -1851,6 +1851,8 @@ enum nl80211_sta_bss_param {
  * @NL80211_STA_INFO_INACTIVE_TIME: time since last activity (u32, msecs)
  * @NL80211_STA_INFO_RX_BYTES: total received bytes (u32, from this station)
  * @NL80211_STA_INFO_TX_BYTES: total transmitted bytes (u32, to this station)
+ * @NL80211_STA_INFO_RX_BYTES64: total received bytes (u64, from this station)
+ * @NL80211_STA_INFO_TX_BYTES64: total transmitted bytes (u64, to this station)
  * @NL80211_STA_INFO_SIGNAL: signal strength of last received PPDU (u8, dBm)
  * @NL80211_STA_INFO_TX_BITRATE: current unicast tx rate, nested attribute
  * 	containing info as possible, see &enum nl80211_rate_info
@@ -1903,6 +1905,8 @@ enum nl80211_sta_info {
 	NL80211_STA_INFO_LOCAL_PM,
 	NL80211_STA_INFO_PEER_PM,
 	NL80211_STA_INFO_NONPEER_PM,
+	NL80211_STA_INFO_RX_BYTES64,
+	NL80211_STA_INFO_TX_BYTES64,
 
 	/* keep last */
 	__NL80211_STA_INFO_AFTER_LAST,
net/wireless/nl80211.c +9 -1
index b5978ab..b3a9c8e 100644
@@ -3059,10 +3059,18 @@ static int nl80211_send_station(struct sk_buff *msg, u32 portid, u32 seq,
 		goto nla_put_failure;
 	if ((sinfo->filled & STATION_INFO_RX_BYTES) &&
 	    nla_put_u32(msg, NL80211_STA_INFO_RX_BYTES,
-			sinfo->rx_bytes))
+			(u32)sinfo->rx_bytes))
 		goto nla_put_failure;
 	if ((sinfo->filled & STATION_INFO_TX_BYTES) &&
 	    nla_put_u32(msg, NL80211_STA_INFO_TX_BYTES,
+			(u32)sinfo->tx_bytes))
+		goto nla_put_failure;
+	if ((sinfo->filled & STATION_INFO_RX_BYTES64) &&
+	    nla_put_u64(msg, NL80211_STA_INFO_RX_BYTES64,
+			sinfo->rx_bytes))
+		goto nla_put_failure;
+	if ((sinfo->filled & STATION_INFO_TX_BYTES64) &&
+	    nla_put_u32(msg, NL80211_STA_INFO_TX_BYTES64,
 			sinfo->tx_bytes))
 		goto nla_put_failure;
 	if ((sinfo->filled & STATION_INFO_LLID) &&
Signature · 1 line
1.7.10.4
Quoted 1 line

From ec3c391673d70078008213d6243881919c977b07 Mon Sep 17 00:00:00 2001

From: Vladimir Kondratiev <[email protected]>
Date: Fri, 1 Feb 2013 16:03:34 +0200
Subject: [PATCH 2/2] ath6kl: provide 64-bit per-station byte counters

Internally, 64-bit byte counters maintained for per-station statistics. Tell to the netlink that full 64-bit value provided

---
 drivers/net/wireless/ath/ath6kl/cfg80211.c |    6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)
drivers/net/wireless/ath/ath6kl/cfg80211.c +4 -2
index 4225cca..90b1a93 100644
@@ -1778,14 +1778,16 @@ static int ath6kl_get_station(struct wiphy *wiphy, struct net_device *dev,
 
 	if (vif->target_stats.rx_byte) {
 		sinfo->rx_bytes = vif->target_stats.rx_byte;
-		sinfo->filled |= STATION_INFO_RX_BYTES;
+		sinfo->filled |= STATION_INFO_RX_BYTES |
+				 STATION_INFO_RX_BYTES64;
 		sinfo->rx_packets = vif->target_stats.rx_pkt;
 		sinfo->filled |= STATION_INFO_RX_PACKETS;
 	}
 
 	if (vif->target_stats.tx_byte) {
 		sinfo->tx_bytes = vif->target_stats.tx_byte;
-		sinfo->filled |= STATION_INFO_TX_BYTES;
+		sinfo->filled |= STATION_INFO_TX_BYTES |
+				 STATION_INFO_TX_BYTES64;
 		sinfo->tx_packets = vif->target_stats.tx_pkt;
 		sinfo->filled |= STATION_INFO_TX_PACKETS;
 	}
Signature · 1 line
1.7.10.4
maintainer
#5 lore raw

Re: [PATCH 0/2] wireless: expand per-station byte counters to 64bit

Quoted 1 line · Vladimir Kondratiev

Patches follows.

Can you send them as proper separate emails, I'm not going to pick up ath6kl one :)

johannes

#6 lore raw

Re: [PATCH 0/2] wireless: expand per-station byte counters to 64bit

Quoted 8 lines · Johannes Berg
Quoted 1 line · Vladimir Kondratiev

Patches follows.

Can you send them as proper separate emails, I'm not going to pick up ath6kl one :)

johannes

Sure, sent.