[PATCH net v2 0/2] netdevsim: psp: fix issues with stats collection

[PATCH net v2 0/2] netdevsim: psp: fix issues with stats collection

[Daniel Zahka]

It has come to my attention via a sashiko review of my net-next series
for aes-gcm in netdevsim [1] that there were preexisting issues with
netdevsim's implementation of psp statistics.

API usage issues:
1. not calling u64_stats_init() on the u64_stats_sync object during
   init
2. not serializing usage of the writer side API during stats update

Logical Bugs:
1. We were incrementing rx stats on the sending devices stats
   counters.

Fix the first set of issues by removing the u64_stats_t api entirely,
and keep track of stats with atomics. Fix the second issue by charging
events to the right netdevsim object.

[1]: https://sashiko.dev/#/patchset/20260508-nsim-psp-crypto-v1-0-4b50ed09b794%40gmail.com

  TAP version 13
  1..28
  ok 1 psp.data_basic_send_v0_ip4
  ok 2 psp.data_basic_send_v0_ip6
  ok 3 psp.data_basic_send_v1_ip4
  ok 4 psp.data_basic_send_v1_ip6
  ok 5 psp.data_basic_send_v2_ip4
  ok 6 psp.data_basic_send_v2_ip6
  ok 7 psp.data_basic_send_v3_ip4
  ok 8 psp.data_basic_send_v3_ip6
  ok 9 psp.data_mss_adjust_ip4
  ok 10 psp.data_mss_adjust_ip6
  ok 11 psp.dev_list_devices
  ok 12 psp.dev_get_device
  ok 13 psp.dev_get_device_bad
  ok 14 psp.dev_rotate
  ok 15 psp.dev_rotate_spi
  ok 16 psp.assoc_basic
  ok 17 psp.assoc_bad_dev
  ok 18 psp.assoc_sk_only_conn
  ok 19 psp.assoc_sk_only_mismatch
  ok 20 psp.assoc_sk_only_mismatch_tx
  ok 21 psp.assoc_sk_only_unconn
  ok 22 psp.assoc_version_mismatch
  ok 23 psp.assoc_twice
  ok 24 psp.data_send_bad_key
  ok 25 psp.data_send_disconnect
  ok 26 psp.data_stale_key
  ok 27 psp.removal_device_rx
  ok 28 psp.removal_device_bi
  # Totals: pass:28 fail:0 xfail:0 xpass:0 skip:0 error:0

Dump stats on both devs tx on one should match rx on other:
local dev:
id=5 ifindex=2 stats={'dev-id': 5, 'key-rotations': 0,
'stale-events': 0, 'rx-packets': 1226, 'rx-bytes': 39244,
'rx-auth-fail': 0, 'rx-error': 0, 'rx-bad': 0, 'tx-packets': 1931,
'tx-bytes': 2478908, 'tx-error': 0}

remote dev:
id=3 ifindex=2 stats={'dev-id': 3, 'key-rotations': 0, 'stale-events':
0, 'rx-packets': 1931, 'rx-bytes': 2478908, 'rx-auth-fail': 0,
'rx-error': 0, 'rx-bad': 0, 'tx-packets': 1226, 'tx-bytes': 39244,
'tx-error': 0}

Signed-off-by: Daniel Zahka <daniel.zahka@gmail.com>
---
Changes in v2:
- get rid of u64_stats_t and use atomics instead
- Link to v1: https://lore.kernel.org/r/20260513-fix-psp-stats-v1-0-cdb3174f252f@gmail.com

---
Daniel Zahka (2):
      netdevsim: psp: update rx stats on the peer netdevsim
      netdevsim: psp: use atomic64 for psp stats counters

 drivers/net/netdevsim/netdevsim.h | 10 ++++------
 drivers/net/netdevsim/psp.c       | 27 +++++++++++----------------
 2 files changed, 15 insertions(+), 22 deletions(-)
---
base-commit: 422b5233b607476ac7176bfa2a101b9a103d7653
change-id: 20260512-fix-psp-stats-e96c6d069d01

Best regards,
-- 
Daniel Zahka <daniel.zahka@gmail.com>

[PATCH net v2 1/2] netdevsim: psp: update rx stats on the peer netdevsim

[Daniel Zahka]

nsim_do_psp() handles both tx and rx psp processing in the sending
device's nsim_start_xmit() path. The existing code has a logical bug,
where we erroneously increment rx_bytes and rx_packets on the sending
devices stats, instead of the peer device.

Additionally, compute psp_len after psp_dev_encapsulate() and before
psp_dev_rcv(), which modifies the header region of the skb. The
existing calculation was actually correct, because psp_dev_rcv()
leaves skb_inner_transport_header pointing at the tcp header, but this
is fragile and confusing as there is no actual inner transport header
after psp_dev_rcv has removed udp encapsulation.

Fixes: 178f0763c5f3 ("netdevsim: implement psp device stats")
Assisted-by: Codex:gpt-5.5
Signed-off-by: Daniel Zahka <daniel.zahka@gmail.com>
---
 drivers/net/netdevsim/psp.c | 19 +++++++++++--------
 1 file changed, 11 insertions(+), 8 deletions(-)

diff --git a/drivers/net/netdevsim/psp.c b/drivers/net/netdevsim/psp.c
index 6936ecb8173e..afe58b21041a 100644
--- a/drivers/net/netdevsim/psp.c
+++ b/drivers/net/netdevsim/psp.c
@@ -22,6 +22,7 @@ nsim_do_psp(struct sk_buff *skb, struct netdevsim *ns,
 	struct psp_dev *peer_psd;
 	struct psp_assoc *pas;
 	struct net *net;
+	int psp_len;
 	void **ptr;
 
 	rcu_read_lock();
@@ -48,6 +49,12 @@ nsim_do_psp(struct sk_buff *skb, struct netdevsim *ns,
 		goto out_unlock;
 	}
 
+	psp_len = skb->len - skb_inner_transport_offset(skb);
+	u64_stats_update_begin(&ns->psp.syncp);
+	u64_stats_inc(&ns->psp.tx_packets);
+	u64_stats_add(&ns->psp.tx_bytes, psp_len);
+	u64_stats_update_end(&ns->psp.syncp);
+
 	/* Now pretend we just received this frame */
 	peer_psd = rcu_dereference(peer_ns->psp.dev);
 	if (peer_psd && peer_psd->config.versions & (1 << pas->version)) {
@@ -72,14 +79,10 @@ nsim_do_psp(struct sk_buff *skb, struct netdevsim *ns,
 		refcount_inc(&(*psp_ext)->refcnt);
 		skb->decrypted = 1;
 
-		u64_stats_update_begin(&ns->psp.syncp);
-		u64_stats_inc(&ns->psp.tx_packets);
-		u64_stats_inc(&ns->psp.rx_packets);
-		u64_stats_add(&ns->psp.tx_bytes,
-			      skb->len - skb_inner_transport_offset(skb));
-		u64_stats_add(&ns->psp.rx_bytes,
-			      skb->len - skb_inner_transport_offset(skb));
-		u64_stats_update_end(&ns->psp.syncp);
+		u64_stats_update_begin(&peer_ns->psp.syncp);
+		u64_stats_inc(&peer_ns->psp.rx_packets);
+		u64_stats_add(&peer_ns->psp.rx_bytes, psp_len);
+		u64_stats_update_end(&peer_ns->psp.syncp);
 	} else {
 		struct ipv6hdr *ip6h __maybe_unused;
 		struct iphdr *iph;

-- 
2.52.0

[PATCH net v2 2/2] netdevsim: psp: use atomic64 for psp stats counters

[Daniel Zahka]

The existing u64_stats_t-based psp counters had two preexisting api
usage bugs: u64_stats_init() was never called on the syncp object, and
the writer side of the u64_stats_update_begin()/end() api was not
serialized. Switch the counters to atomic64_t instead. Atomics need
no initialization and are inherently safe against concurrent writers,
eliminating both bugs at once.

Use atomic64_t rather than atomic_long_t so byte counters don't wrap
at 4 GiB on 32-bit builds.

Fixes: 178f0763c5f3 ("netdevsim: implement psp device stats")
Assisted-by: Codex:gpt-5.5
Signed-off-by: Daniel Zahka <daniel.zahka@gmail.com>
---
 drivers/net/netdevsim/netdevsim.h | 10 ++++------
 drivers/net/netdevsim/psp.c       | 24 ++++++++----------------
 2 files changed, 12 insertions(+), 22 deletions(-)

diff --git a/drivers/net/netdevsim/netdevsim.h b/drivers/net/netdevsim/netdevsim.h
index d909c4160ea1..4c9cc96dcec3 100644
--- a/drivers/net/netdevsim/netdevsim.h
+++ b/drivers/net/netdevsim/netdevsim.h
@@ -22,7 +22,6 @@
 #include <linux/list.h>
 #include <linux/netdevice.h>
 #include <linux/ptp_mock.h>
-#include <linux/u64_stats_sync.h>
 #include <net/devlink.h>
 #include <net/udp_tunnel.h>
 #include <net/xdp.h>
@@ -115,11 +114,10 @@ struct netdevsim {
 	int rq_reset_mode;
 
 	struct {
-		u64_stats_t rx_packets;
-		u64_stats_t rx_bytes;
-		u64_stats_t tx_packets;
-		u64_stats_t tx_bytes;
-		struct u64_stats_sync syncp;
+		atomic64_t rx_packets;
+		atomic64_t rx_bytes;
+		atomic64_t tx_packets;
+		atomic64_t tx_bytes;
 		struct psp_dev __rcu *dev;
 		struct dentry *rereg;
 		struct mutex rereg_lock;
diff --git a/drivers/net/netdevsim/psp.c b/drivers/net/netdevsim/psp.c
index afe58b21041a..1c415b3d6f1a 100644
--- a/drivers/net/netdevsim/psp.c
+++ b/drivers/net/netdevsim/psp.c
@@ -50,10 +50,8 @@ nsim_do_psp(struct sk_buff *skb, struct netdevsim *ns,
 	}
 
 	psp_len = skb->len - skb_inner_transport_offset(skb);
-	u64_stats_update_begin(&ns->psp.syncp);
-	u64_stats_inc(&ns->psp.tx_packets);
-	u64_stats_add(&ns->psp.tx_bytes, psp_len);
-	u64_stats_update_end(&ns->psp.syncp);
+	atomic64_inc(&ns->psp.tx_packets);
+	atomic64_add(psp_len, &ns->psp.tx_bytes);
 
 	/* Now pretend we just received this frame */
 	peer_psd = rcu_dereference(peer_ns->psp.dev);
@@ -79,10 +77,8 @@ nsim_do_psp(struct sk_buff *skb, struct netdevsim *ns,
 		refcount_inc(&(*psp_ext)->refcnt);
 		skb->decrypted = 1;
 
-		u64_stats_update_begin(&peer_ns->psp.syncp);
-		u64_stats_inc(&peer_ns->psp.rx_packets);
-		u64_stats_add(&peer_ns->psp.rx_bytes, psp_len);
-		u64_stats_update_end(&peer_ns->psp.syncp);
+		atomic64_inc(&peer_ns->psp.rx_packets);
+		atomic64_add(psp_len, &peer_ns->psp.rx_bytes);
 	} else {
 		struct ipv6hdr *ip6h __maybe_unused;
 		struct iphdr *iph;
@@ -180,20 +176,16 @@ static void nsim_assoc_del(struct psp_dev *psd, struct psp_assoc *pas)
 static void nsim_get_stats(struct psp_dev *psd, struct psp_dev_stats *stats)
 {
 	struct netdevsim *ns = psd->drv_priv;
-	unsigned int start;
 
 	/* WARNING: do *not* blindly zero stats in real drivers!
 	 * All required stats must be reported by the device!
 	 */
 	memset(stats, 0, sizeof(struct psp_dev_stats));
 
-	do {
-		start = u64_stats_fetch_begin(&ns->psp.syncp);
-		stats->rx_bytes = u64_stats_read(&ns->psp.rx_bytes);
-		stats->rx_packets = u64_stats_read(&ns->psp.rx_packets);
-		stats->tx_bytes = u64_stats_read(&ns->psp.tx_bytes);
-		stats->tx_packets = u64_stats_read(&ns->psp.tx_packets);
-	} while (u64_stats_fetch_retry(&ns->psp.syncp, start));
+	stats->rx_bytes = atomic64_read(&ns->psp.rx_bytes);
+	stats->rx_packets = atomic64_read(&ns->psp.rx_packets);
+	stats->tx_bytes = atomic64_read(&ns->psp.tx_bytes);
+	stats->tx_packets = atomic64_read(&ns->psp.tx_packets);
 }
 
 static struct psp_dev_ops nsim_psp_ops = {

-- 
2.52.0

Re: [PATCH net v2 0/2] netdevsim: psp: fix issues with stats collection

[patchwork-bot+netdevbpf]

Hello:

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

On Fri, 29 May 2026 06:15:26 -0700 you wrote:
> It has come to my attention via a sashiko review of my net-next series
> for aes-gcm in netdevsim [1] that there were preexisting issues with
> netdevsim's implementation of psp statistics.
> 
> API usage issues:
> 1. not calling u64_stats_init() on the u64_stats_sync object during
>    init
> 2. not serializing usage of the writer side API during stats update
> 
> [...]

Here is the summary with links:
  - [net,v2,1/2] netdevsim: psp: update rx stats on the peer netdevsim
    https://git.kernel.org/netdev/net-next/c/abaef7e966fd
  - [net,v2,2/2] netdevsim: psp: use atomic64 for psp stats counters
    https://git.kernel.org/netdev/net-next/c/163bea8010bd

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