Netdev List
 help / color / mirror / Atom feed
* [PATCH net v8 0/2] net: wangxun: protect ring accesses with RCU
@ 2026-11-18  8:13 Mengyuan Lou
  2026-11-18  8:13 ` [PATCH net v8 1/2] " Mengyuan Lou
                   ` (2 more replies)
  0 siblings, 3 replies; 4+ messages in thread
From: Mengyuan Lou @ 2026-11-18  8:13 UTC (permalink / raw)
  To: netdev
  Cc: jiawenwu, duanqiangwen, linglingzhang, andrew+netdev, davem,
	edumazet, kuba, pabeni, hramamurthy, jacob.e.keller,
	przemyslaw.kitszel, Mengyuan Lou

This series protects Wangxun ring-array accesses during interrupt-scheme
reconfiguration.

Patch 1 converts ring-pointer access to RCU and initializes reg_idx before
publishing each ring. Patch 2 pairs lockless queue-count reads and writes
so readers remain safe while queue vectors are torn down or rebuilt.

Changelog:
v8:
  - Split queue-count access updates into a separate patch.
  - Initialize reg_idx before ring publication while preserving RSS/VMDQ
    queue-to-register mapping.
  - Pair lockless queue-count reads and writes across setup, teardown,
    allocation-error handling, and VF queue setup.

v7: https://lore.kernel.org/all/20260928085443.14595-1-mengyuanlou@net-swift.com/

Mengyuan Lou (2):
  net: wangxun: protect ring accesses with RCU
  net: wangxun: access queue counts consistently

 drivers/net/ethernet/wangxun/libwx/wx_err.c   |  25 ++-
 .../net/ethernet/wangxun/libwx/wx_ethtool.c   |  10 +-
 drivers/net/ethernet/wangxun/libwx/wx_hw.c    |  90 +++++++---
 drivers/net/ethernet/wangxun/libwx/wx_lib.c   | 164 ++++++++----------
 drivers/net/ethernet/wangxun/libwx/wx_ptp.c   |  11 +-
 drivers/net/ethernet/wangxun/libwx/wx_type.h  |  12 +-
 .../net/ethernet/wangxun/libwx/wx_vf_common.c |   2 +-
 .../net/ethernet/wangxun/libwx/wx_vf_lib.c    |   7 +-
 .../net/ethernet/wangxun/ngbe/ngbe_ethtool.c  |   4 +-
 drivers/net/ethernet/wangxun/ngbe/ngbe_main.c |  11 +-
 .../net/ethernet/wangxun/ngbevf/ngbevf_main.c |   4 +-
 .../ethernet/wangxun/txgbe/txgbe_ethtool.c    |   6 +-
 .../net/ethernet/wangxun/txgbe/txgbe_fdir.c   |   2 +-
 .../net/ethernet/wangxun/txgbe/txgbe_main.c   |   4 +-
 .../ethernet/wangxun/txgbevf/txgbevf_main.c   |  10 +-
 15 files changed, 215 insertions(+), 147 deletions(-)

-- 
2.30.1


^ permalink raw reply	[flat|nested] 4+ messages in thread

* [PATCH net v8 1/2] net: wangxun: protect ring accesses with RCU
  2026-11-18  8:13 [PATCH net v8 0/2] net: wangxun: protect ring accesses with RCU Mengyuan Lou
@ 2026-11-18  8:13 ` Mengyuan Lou
  2026-11-18  8:13 ` [PATCH net v8 2/2] net: wangxun: access queue counts consistently Mengyuan Lou
  2026-11-18  8:18 ` [PATCH net v8 0/2] net: wangxun: protect ring accesses with RCU netdev-bot+sinfo
  2 siblings, 0 replies; 4+ messages in thread
From: Mengyuan Lou @ 2026-11-18  8:13 UTC (permalink / raw)
  To: netdev
  Cc: jiawenwu, duanqiangwen, linglingzhang, andrew+netdev, davem,
	edumazet, kuba, pabeni, hramamurthy, jacob.e.keller,
	przemyslaw.kitszel, Mengyuan Lou

In the Wangxun driver family, ring pointers in wx->rx_ring[] and
wx->tx_ring[] can be published, cleared, or replaced during interrupt
scheme teardown, channel reconfiguration (e.g. via ethtool -L), or TC
setup.

Various lockless readers, including stats gathering, PTP watchdogs,
and flow control checks, access per-queue ring structures without
formal RCU annotations. Although q_vector memory lifetime is managed
via kfree_rcu(), the ring array accesses themselves lack explicit RCU
dereference semantics and barriers.

Annotate the ring arrays and convert their readers and writers to RCU
primitives. Move the queue-to-register index calculation from the
post-allocation wx_cache_ring_rss()/wx_cache_ring_vmdq() pass into per-ring
allocation via wx_get_reg_idx(). Initialize each ring's reg_idx before
publishing its pointer so lockless readers cannot observe an uninitialized
value.
Use rtnl_dereference() in control paths protected by RTNL, and RCU
read-side critical sections for lockless readers.

The ring arrays are managed by wx_alloc_q_vector() and
wx_free_q_vector(), reached through wx_init_interrupt_scheme() and
wx_clear_interrupt_scheme(). The probe paths call
wx_init_interrupt_scheme() before the netdev is registered; the
allocation path only publishes rings with rcu_assign_pointer(), so no
RTNL is needed there. wxvf_resume() defers rebuilding the interrupt
scheme to wxvf_reopen_subtask(), which runs under RTNL. wxvf_suspend()
holds RTNL across wxvf_close() and wx_clear_interrupt_scheme(),
protecting control-path readers and serializing teardown against the
reset subtask. The service timer and work are drained before the PCI
device is disabled.

ngbe_dev_shutdown() keeps RTNL held across the WoL setup as well.
That matches the locking of the other wx_configure_rx() callers and
keeps the WoL configuration from racing with a channel reconfiguration,
which re-creates the interrupt scheme. wx_clear_interrupt_scheme() has
already zeroed the queue counts before wx_configure_rx() runs, so its
ring loops do not access the ring arrays on this shutdown path.

wx_xmit_frame() also needs to handle teardown racing with transmission:
wx_free_q_vectors() can zero wx->num_tx_queues while the netdev remains
registered, which could otherwise cause a modulo-by-zero, and the ring
slot can be NULL. Bail out on either condition and account the dropped
skb. Also account for the carrier-off path and skb_put_padto() failure;
the latter frees the skb on error, so only the counter can be updated on
that path.

Fixes: 3f703186113f ("net: libwx: Add irq flow functions")
Signed-off-by: Mengyuan Lou <mengyuanlou@net-swift.com>
---
 drivers/net/ethernet/wangxun/libwx/wx_err.c   |  20 ++-
 .../net/ethernet/wangxun/libwx/wx_ethtool.c   |  10 +-
 drivers/net/ethernet/wangxun/libwx/wx_hw.c    |  73 ++++++++--
 drivers/net/ethernet/wangxun/libwx/wx_lib.c   | 137 ++++++++----------
 drivers/net/ethernet/wangxun/libwx/wx_ptp.c   |   6 +-
 drivers/net/ethernet/wangxun/libwx/wx_type.h  |  12 +-
 .../net/ethernet/wangxun/libwx/wx_vf_common.c |   2 +-
 .../net/ethernet/wangxun/libwx/wx_vf_lib.c    |   7 +-
 .../net/ethernet/wangxun/ngbe/ngbe_ethtool.c  |   4 +-
 drivers/net/ethernet/wangxun/ngbe/ngbe_main.c |  11 +-
 .../ethernet/wangxun/txgbe/txgbe_ethtool.c    |   6 +-
 .../net/ethernet/wangxun/txgbe/txgbe_fdir.c   |   2 +-
 .../net/ethernet/wangxun/txgbe/txgbe_main.c   |   4 +-
 13 files changed, 174 insertions(+), 120 deletions(-)

diff --git a/drivers/net/ethernet/wangxun/libwx/wx_err.c b/drivers/net/ethernet/wangxun/libwx/wx_err.c
index b56fbdc959dec..8eadb4d3abcff 100644
--- a/drivers/net/ethernet/wangxun/libwx/wx_err.c
+++ b/drivers/net/ethernet/wangxun/libwx/wx_err.c
@@ -201,12 +201,18 @@ static bool wx_ring_tx_pending(struct wx *wx)
 {
 	int i;
 
+	rcu_read_lock();
 	for (i = 0; i < wx->num_tx_queues; i++) {
-		struct wx_ring *tx_ring = wx->tx_ring[i];
+		struct wx_ring *tx_ring = rcu_dereference(wx->tx_ring[i]);
 
-		if (tx_ring->next_to_use != tx_ring->next_to_clean)
+		if (!tx_ring)
+			continue;
+		if (tx_ring->next_to_use != tx_ring->next_to_clean) {
+			rcu_read_unlock();
 			return true;
+		}
 	}
+	rcu_read_unlock();
 
 	return false;
 }
@@ -264,8 +270,14 @@ static void wx_detect_tx_hang(struct wx *wx)
 
 	/* Force detection of hung controller */
 	if (netif_carrier_ok(wx->netdev)) {
-		for (i = 0; i < wx->num_tx_queues; i++)
-			set_bit(WX_TX_DETECT_HANG, wx->tx_ring[i]->state);
+		rcu_read_lock();
+		for (i = 0; i < wx->num_tx_queues; i++) {
+			struct wx_ring *tx_ring = rcu_dereference(wx->tx_ring[i]);
+
+			if (tx_ring)
+				set_bit(WX_TX_DETECT_HANG, tx_ring->state);
+		}
+		rcu_read_unlock();
 	}
 }
 
diff --git a/drivers/net/ethernet/wangxun/libwx/wx_ethtool.c b/drivers/net/ethernet/wangxun/libwx/wx_ethtool.c
index 940d2e59876c7..2307149b07fdb 100644
--- a/drivers/net/ethernet/wangxun/libwx/wx_ethtool.c
+++ b/drivers/net/ethernet/wangxun/libwx/wx_ethtool.c
@@ -154,8 +154,9 @@ void wx_get_ethtool_stats(struct net_device *netdev,
 		}
 	}
 
+	rcu_read_lock();
 	for (j = 0; j < netdev->num_tx_queues; j++) {
-		ring = wx->tx_ring[j];
+		ring = rcu_dereference(wx->tx_ring[j]);
 		if (!ring) {
 			data[i++] = 0;
 			data[i++] = 0;
@@ -170,7 +171,7 @@ void wx_get_ethtool_stats(struct net_device *netdev,
 		i += 2;
 	}
 	for (j = 0; j < WX_NUM_RX_QUEUES; j++) {
-		ring = wx->rx_ring[j];
+		ring = rcu_dereference(wx->rx_ring[j]);
 		if (!ring) {
 			data[i++] = 0;
 			data[i++] = 0;
@@ -184,6 +185,7 @@ void wx_get_ethtool_stats(struct net_device *netdev,
 		} while (u64_stats_fetch_retry(&ring->syncp, start));
 		i += 2;
 	}
+	rcu_read_unlock();
 }
 EXPORT_SYMBOL(wx_get_ethtool_stats);
 
@@ -805,9 +807,9 @@ static int wx_set_ringparam_vf(struct net_device *netdev,
 
 	if (!netif_running(wx->netdev)) {
 		for (i = 0; i < wx->num_tx_queues; i++)
-			wx->tx_ring[i]->count = new_tx_count;
+			rtnl_dereference(wx->tx_ring[i])->count = new_tx_count;
 		for (i = 0; i < wx->num_rx_queues; i++)
-			wx->rx_ring[i]->count = new_rx_count;
+			rtnl_dereference(wx->rx_ring[i])->count = new_rx_count;
 		wx->tx_ring_count = new_tx_count;
 		wx->rx_ring_count = new_rx_count;
 
diff --git a/drivers/net/ethernet/wangxun/libwx/wx_hw.c b/drivers/net/ethernet/wangxun/libwx/wx_hw.c
index 111aadf792089..24c539ce6acbe 100644
--- a/drivers/net/ethernet/wangxun/libwx/wx_hw.c
+++ b/drivers/net/ethernet/wangxun/libwx/wx_hw.c
@@ -1598,13 +1598,17 @@ static void wx_vlan_strip_control(struct wx *wx, bool enable)
 {
 	int i, j;
 
+	rcu_read_lock();
 	for (i = 0; i < wx->num_rx_queues; i++) {
-		struct wx_ring *ring = wx->rx_ring[i];
+		struct wx_ring *ring = rcu_dereference(wx->rx_ring[i]);
 
+		if (!ring)
+			continue;
 		j = ring->reg_idx;
 		wr32m(wx, WX_PX_RR_CFG(j), WX_PX_RR_CFG_VLAN,
 		      enable ? WX_PX_RR_CFG_VLAN : 0);
 	}
+	rcu_read_unlock();
 }
 
 static void wx_vlan_promisc_enable(struct wx *wx)
@@ -1797,7 +1801,7 @@ static void wx_set_rx_buffer_len(struct wx *wx)
 	 * the Base and Length of the Rx Descriptor Ring
 	 */
 	for (i = 0; i < wx->num_rx_queues; i++) {
-		rx_ring = wx->rx_ring[i];
+		rx_ring = rtnl_dereference(wx->rx_ring[i]);
 		rx_ring->rx_buf_len = WX_RXBUFFER_2K;
 #if (PAGE_SIZE < 8192)
 		if (test_bit(WX_FLAG_RSC_ENABLED, wx->flags))
@@ -2021,8 +2025,11 @@ static void wx_configure_tx(struct wx *wx)
 	      WX_TDM_CTL_TE, WX_TDM_CTL_TE);
 
 	/* Setup the HW Tx Head and Tail descriptor pointers */
-	for (i = 0; i < wx->num_tx_queues; i++)
-		wx_configure_tx_ring(wx, wx->tx_ring[i]);
+	for (i = 0; i < wx->num_tx_queues; i++) {
+		struct wx_ring *tx_ring = rtnl_dereference(wx->tx_ring[i]);
+
+		wx_configure_tx_ring(wx, tx_ring);
+	}
 
 	wr32m(wx, WX_TSC_BUF_AE, WX_TSC_BUF_AE_THR, 0x10);
 
@@ -2247,8 +2254,11 @@ void wx_configure_rx(struct wx *wx)
 	/* Setup the HW Rx Head and Tail Descriptor Pointers and
 	 * the Base and Length of the Rx Descriptor Ring
 	 */
-	for (i = 0; i < wx->num_rx_queues; i++)
-		wx_configure_rx_ring(wx, wx->rx_ring[i]);
+	for (i = 0; i < wx->num_rx_queues; i++) {
+		struct wx_ring *rx_ring = rtnl_dereference(wx->rx_ring[i]);
+
+		wx_configure_rx_ring(wx, rx_ring);
+	}
 
 	/* Enable all receives, disable security engine prior to block traffic */
 	ret = wx_disable_sec_rx_path(wx);
@@ -2838,13 +2848,25 @@ int wx_fc_enable(struct wx *wx, bool tx_pause, bool rx_pause)
 	 *  This allows us to avoid head of line blocking for security
 	 *  and performance reasons.
 	 */
+	rcu_read_lock();
 	if (wx->num_rx_queues > 1 && !tx_pause) {
-		for (i = 0; i < wx->num_rx_queues; i++)
-			wx_enable_rx_drop(wx, wx->rx_ring[i]);
+		for (i = 0; i < wx->num_rx_queues; i++) {
+			struct wx_ring *rx_ring = rcu_dereference(wx->rx_ring[i]);
+
+			if (!rx_ring)
+				continue;
+			wx_enable_rx_drop(wx, rx_ring);
+		}
 	} else {
-		for (i = 0; i < wx->num_rx_queues; i++)
-			wx_disable_rx_drop(wx, wx->rx_ring[i]);
+		for (i = 0; i < wx->num_rx_queues; i++) {
+			struct wx_ring *rx_ring = rcu_dereference(wx->rx_ring[i]);
+
+			if (!rx_ring)
+				continue;
+			wx_disable_rx_drop(wx, rx_ring);
+		}
 	}
+	rcu_read_unlock();
 
 	return 0;
 }
@@ -2871,8 +2893,15 @@ static void wx_update_xoff_rx_lfc(struct wx *wx)
 	if (!data)
 		return;
 
-	for (i = 0; i < wx->num_tx_queues; i++)
-		clear_bit(WX_HANG_CHECK_ARMED, wx->tx_ring[i]->state);
+	rcu_read_lock();
+	for (i = 0; i < wx->num_tx_queues; i++) {
+		struct wx_ring *tx_ring = rcu_dereference(wx->tx_ring[i]);
+
+		if (!tx_ring)
+			continue;
+		clear_bit(WX_HANG_CHECK_ARMED, tx_ring->state);
+	}
+	rcu_read_unlock();
 }
 
 /**
@@ -2894,10 +2923,13 @@ void wx_update_stats(struct wx *wx)
 
 	spin_lock(&wx->hw_stats_lock);
 
+	rcu_read_lock();
 	/* gather some stats to the wx struct that are per queue */
 	for (i = 0; i < wx->num_rx_queues; i++) {
-		struct wx_ring *rx_ring = wx->rx_ring[i];
+		struct wx_ring *rx_ring = rcu_dereference(wx->rx_ring[i]);
 
+		if (!rx_ring)
+			continue;
 		non_eop_descs += rx_ring->rx_stats.non_eop_descs;
 		alloc_rx_buff_failed += rx_ring->rx_stats.alloc_rx_buff_failed;
 		hw_csum_rx_good += rx_ring->rx_stats.csum_good_cnt;
@@ -2913,19 +2945,28 @@ void wx_update_stats(struct wx *wx)
 		u64 rsc_flush = 0;
 
 		for (i = 0; i < wx->num_rx_queues; i++) {
-			rsc_count += wx->rx_ring[i]->rx_stats.rsc_count;
-			rsc_flush += wx->rx_ring[i]->rx_stats.rsc_flush;
+			struct wx_ring *rx_ring = rcu_dereference(wx->rx_ring[i]);
+
+			if (!rx_ring)
+				continue;
+
+			rsc_count += rx_ring->rx_stats.rsc_count;
+			rsc_flush += rx_ring->rx_stats.rsc_flush;
 		}
 		wx->rsc_count = rsc_count;
 		wx->rsc_flush = rsc_flush;
 	}
 
 	for (i = 0; i < wx->num_tx_queues; i++) {
-		struct wx_ring *tx_ring = wx->tx_ring[i];
+		struct wx_ring *tx_ring = rcu_dereference(wx->tx_ring[i]);
+
+		if (!tx_ring)
+			continue;
 
 		restart_queue += tx_ring->tx_stats.restart_queue;
 		tx_busy += tx_ring->tx_stats.tx_busy;
 	}
+	rcu_read_unlock();
 	wx->restart_queue = restart_queue;
 	wx->tx_busy = tx_busy;
 
diff --git a/drivers/net/ethernet/wangxun/libwx/wx_lib.c b/drivers/net/ethernet/wangxun/libwx/wx_lib.c
index dcbf5811046e9..0fbaa585bb854 100644
--- a/drivers/net/ethernet/wangxun/libwx/wx_lib.c
+++ b/drivers/net/ethernet/wangxun/libwx/wx_lib.c
@@ -1721,23 +1721,35 @@ netdev_tx_t wx_xmit_frame(struct sk_buff *skb,
 	unsigned int r_idx = skb->queue_mapping;
 	struct wx *wx = netdev_priv(netdev);
 	struct wx_ring *tx_ring;
+	int num_tx_queues;
 
-	if (!netif_carrier_ok(netdev)) {
-		dev_kfree_skb_any(skb);
-		return NETDEV_TX_OK;
-	}
+	if (!netif_carrier_ok(netdev))
+		goto drop;
 
 	/* The minimum packet size for olinfo paylen is 17 so pad the skb
 	 * in order to meet this minimum size requirement.
 	 */
-	if (skb_put_padto(skb, 17))
+	if (skb_put_padto(skb, 17)) {
+		dev_core_stats_tx_dropped_inc(netdev);
 		return NETDEV_TX_OK;
+	}
+
+	num_tx_queues = wx->num_tx_queues;
+	if (unlikely(!num_tx_queues))
+		goto drop;
 
-	if (r_idx >= wx->num_tx_queues)
-		r_idx = r_idx % wx->num_tx_queues;
-	tx_ring = wx->tx_ring[r_idx];
+	if (r_idx >= num_tx_queues)
+		r_idx = r_idx % num_tx_queues;
+	tx_ring = rcu_dereference_bh(wx->tx_ring[r_idx]);
+	if (unlikely(!tx_ring))
+		goto drop;
 
 	return wx_xmit_frame_ring(skb, tx_ring);
+
+drop:
+	dev_core_stats_tx_dropped_inc(netdev);
+	dev_kfree_skb_any(skb);
+	return NETDEV_TX_OK;
 }
 EXPORT_SYMBOL(wx_xmit_frame);
 
@@ -2061,68 +2073,30 @@ static int wx_set_interrupt_capability(struct wx *wx)
 	return 0;
 }
 
-static bool wx_cache_ring_vmdq(struct wx *wx)
+static u16 wx_get_reg_idx(struct wx *wx, u16 queue_index, bool is_rx)
 {
 	struct wx_ring_feature *vmdq = &wx->ring_feature[RING_F_VMDQ];
 	struct wx_ring_feature *rss = &wx->ring_feature[RING_F_RSS];
 	u16 reg_idx;
 	int i;
 
-	/* only proceed if VMDq is enabled */
 	if (!test_bit(WX_FLAG_VMDQ_ENABLED, wx->flags))
-		return false;
+		return queue_index;
 
-	if (test_bit(WX_FLAG_MULTI_64_FUNC, wx->flags)) {
-		/* start at VMDq register offset for SR-IOV enabled setups */
-		reg_idx = vmdq->offset * __ALIGN_MASK(1, ~vmdq->mask);
-		for (i = 0; i < wx->num_rx_queues; i++, reg_idx++) {
-			/* If we are greater than indices move to next pool */
+	if (!test_bit(WX_FLAG_MULTI_64_FUNC, wx->flags))
+		return vmdq->offset + queue_index;
+
+	reg_idx = vmdq->offset * __ALIGN_MASK(1, ~vmdq->mask);
+	for (i = 0; i <= queue_index; i++, reg_idx++) {
+		if (is_rx) {
 			if ((reg_idx & ~vmdq->mask) >= rss->indices)
 				reg_idx = __ALIGN_MASK(reg_idx, ~vmdq->mask);
-			wx->rx_ring[i]->reg_idx = reg_idx;
-		}
-		reg_idx = vmdq->offset * __ALIGN_MASK(1, ~vmdq->mask);
-		for (i = 0; i < wx->num_tx_queues; i++, reg_idx++) {
-			/* If we are greater than indices move to next pool */
-			if ((reg_idx & rss->mask) >= rss->indices)
-				reg_idx = __ALIGN_MASK(reg_idx, ~vmdq->mask);
-			wx->tx_ring[i]->reg_idx = reg_idx;
+		} else if ((reg_idx & rss->mask) >= rss->indices) {
+			reg_idx = __ALIGN_MASK(reg_idx, ~vmdq->mask);
 		}
-	} else {
-		/* start at VMDq register offset for SR-IOV enabled setups */
-		reg_idx = vmdq->offset;
-		for (i = 0; i < wx->num_rx_queues; i++)
-			/* If we are greater than indices move to next pool */
-			wx->rx_ring[i]->reg_idx = reg_idx + i;
-
-		reg_idx = vmdq->offset;
-		for (i = 0; i < wx->num_tx_queues; i++)
-			/* If we are greater than indices move to next pool */
-			wx->tx_ring[i]->reg_idx = reg_idx + i;
 	}
 
-	return true;
-}
-
-/**
- * wx_cache_ring_rss - Descriptor ring to register mapping for RSS
- * @wx: board private structure to initialize
- *
- * Cache the descriptor ring offsets for RSS, ATR, FCoE, and SR-IOV.
- *
- **/
-static void wx_cache_ring_rss(struct wx *wx)
-{
-	u16 i;
-
-	if (wx_cache_ring_vmdq(wx))
-		return;
-
-	for (i = 0; i < wx->num_rx_queues; i++)
-		wx->rx_ring[i]->reg_idx = i;
-
-	for (i = 0; i < wx->num_tx_queues; i++)
-		wx->tx_ring[i]->reg_idx = i;
+	return reg_idx - 1;
 }
 
 static void wx_add_ring(struct wx_ring *ring, struct wx_ring_container *head)
@@ -2210,9 +2184,10 @@ static int wx_alloc_q_vector(struct wx *wx,
 		ring->count = wx->tx_ring_count;
 
 		ring->queue_index = txr_idx;
+		ring->reg_idx = wx_get_reg_idx(wx, txr_idx, false);
 
 		/* assign ring to wx */
-		wx->tx_ring[txr_idx] = ring;
+		rcu_assign_pointer(wx->tx_ring[txr_idx], ring);
 
 		/* update count and index */
 		txr_count--;
@@ -2236,9 +2211,10 @@ static int wx_alloc_q_vector(struct wx *wx,
 		/* apply Rx specific ring traits */
 		ring->count = wx->rx_ring_count;
 		ring->queue_index = rxr_idx;
+		ring->reg_idx = wx_get_reg_idx(wx, rxr_idx, true);
 
 		/* assign ring to wx */
-		wx->rx_ring[rxr_idx] = ring;
+		rcu_assign_pointer(wx->rx_ring[rxr_idx], ring);
 
 		/* update count and index */
 		rxr_count--;
@@ -2266,10 +2242,10 @@ static void wx_free_q_vector(struct wx *wx, int v_idx)
 	struct wx_ring *ring;
 
 	wx_for_each_ring(ring, q_vector->tx)
-		wx->tx_ring[ring->queue_index] = NULL;
+		rcu_assign_pointer(wx->tx_ring[ring->queue_index], NULL);
 
 	wx_for_each_ring(ring, q_vector->rx)
-		wx->rx_ring[ring->queue_index] = NULL;
+		rcu_assign_pointer(wx->rx_ring[ring->queue_index], NULL);
 
 	wx->q_vector[v_idx] = NULL;
 	netif_napi_del(&q_vector->napi);
@@ -2400,8 +2376,6 @@ int wx_init_interrupt_scheme(struct wx *wx)
 		return ret;
 	}
 
-	wx_cache_ring_rss(wx);
-
 	set_bit(WX_STATE_DOWN, wx->state);
 
 	return 0;
@@ -2680,7 +2654,7 @@ void wx_clean_all_rx_rings(struct wx *wx)
 	int i;
 
 	for (i = 0; i < wx->num_rx_queues; i++)
-		wx_clean_rx_ring(wx->rx_ring[i]);
+		wx_clean_rx_ring(rtnl_dereference(wx->rx_ring[i]));
 }
 EXPORT_SYMBOL(wx_clean_all_rx_rings);
 
@@ -2722,7 +2696,7 @@ static void wx_free_all_rx_resources(struct wx *wx)
 	int i;
 
 	for (i = 0; i < wx->num_rx_queues; i++)
-		wx_free_rx_resources(wx->rx_ring[i]);
+		wx_free_rx_resources(rtnl_dereference(wx->rx_ring[i]));
 }
 
 /**
@@ -2796,7 +2770,7 @@ void wx_clean_all_tx_rings(struct wx *wx)
 	int i;
 
 	for (i = 0; i < wx->num_tx_queues; i++)
-		wx_clean_tx_ring(wx->tx_ring[i]);
+		wx_clean_tx_ring(rtnl_dereference(wx->tx_ring[i]));
 }
 EXPORT_SYMBOL(wx_clean_all_tx_rings);
 
@@ -2844,7 +2818,7 @@ static void wx_free_all_tx_resources(struct wx *wx)
 	int i;
 
 	for (i = 0; i < wx->num_tx_queues; i++)
-		wx_free_tx_resources(wx->tx_ring[i]);
+		wx_free_tx_resources(rtnl_dereference(wx->tx_ring[i]));
 }
 
 void wx_free_resources(struct wx *wx)
@@ -2954,7 +2928,7 @@ static int wx_setup_all_rx_resources(struct wx *wx)
 	int i, err = 0;
 
 	for (i = 0; i < wx->num_rx_queues; i++) {
-		err = wx_setup_rx_resources(wx->rx_ring[i]);
+		err = wx_setup_rx_resources(rtnl_dereference(wx->rx_ring[i]));
 		if (!err)
 			continue;
 
@@ -2966,7 +2940,7 @@ static int wx_setup_all_rx_resources(struct wx *wx)
 err_setup_rx:
 	/* rewind the index freeing the rings as we go */
 	while (i--)
-		wx_free_rx_resources(wx->rx_ring[i]);
+		wx_free_rx_resources(rtnl_dereference(wx->rx_ring[i]));
 	return err;
 }
 
@@ -3057,7 +3031,7 @@ static int wx_setup_all_tx_resources(struct wx *wx)
 	int i, err = 0;
 
 	for (i = 0; i < wx->num_tx_queues; i++) {
-		err = wx_setup_tx_resources(wx->tx_ring[i]);
+		err = wx_setup_tx_resources(rtnl_dereference(wx->tx_ring[i]));
 		if (!err)
 			continue;
 
@@ -3069,7 +3043,7 @@ static int wx_setup_all_tx_resources(struct wx *wx)
 err_setup_tx:
 	/* rewind the index freeing the rings as we go */
 	while (i--)
-		wx_free_tx_resources(wx->tx_ring[i]);
+		wx_free_tx_resources(rtnl_dereference(wx->tx_ring[i]));
 	return err;
 }
 
@@ -3118,7 +3092,7 @@ void wx_get_stats64(struct net_device *netdev,
 
 	rcu_read_lock();
 	for (i = 0; i < wx->num_rx_queues; i++) {
-		struct wx_ring *ring = READ_ONCE(wx->rx_ring[i]);
+		struct wx_ring *ring = rcu_dereference(wx->rx_ring[i]);
 		u64 bytes, packets;
 		unsigned int start;
 
@@ -3134,7 +3108,7 @@ void wx_get_stats64(struct net_device *netdev,
 	}
 
 	for (i = 0; i < wx->num_tx_queues; i++) {
-		struct wx_ring *ring = READ_ONCE(wx->tx_ring[i]);
+		struct wx_ring *ring = rcu_dereference(wx->tx_ring[i]);
 		u64 bytes, packets;
 		unsigned int start;
 
@@ -3345,7 +3319,8 @@ int wx_set_ring(struct wx *wx, u32 new_tx_count,
 	 */
 	if (new_tx_count != wx->tx_ring_count) {
 		for (i = 0; i < wx->num_tx_queues; i++) {
-			memcpy(&temp_ring[i], wx->tx_ring[i],
+			memcpy(&temp_ring[i],
+			       rtnl_dereference(wx->tx_ring[i]),
 			       sizeof(struct wx_ring));
 
 			temp_ring[i].count = new_tx_count;
@@ -3361,9 +3336,11 @@ int wx_set_ring(struct wx *wx, u32 new_tx_count,
 		}
 
 		for (i = 0; i < wx->num_tx_queues; i++) {
-			wx_free_tx_resources(wx->tx_ring[i]);
+			struct wx_ring *tx_ring = rtnl_dereference(wx->tx_ring[i]);
 
-			memcpy(wx->tx_ring[i], &temp_ring[i],
+			wx_free_tx_resources(tx_ring);
+
+			memcpy(tx_ring, &temp_ring[i],
 			       sizeof(struct wx_ring));
 		}
 
@@ -3373,7 +3350,7 @@ int wx_set_ring(struct wx *wx, u32 new_tx_count,
 	/* Repeat the process for the Rx rings if needed */
 	if (new_rx_count != wx->rx_ring_count) {
 		for (i = 0; i < wx->num_rx_queues; i++) {
-			memcpy(&temp_ring[i], wx->rx_ring[i],
+			memcpy(&temp_ring[i], rtnl_dereference(wx->rx_ring[i]),
 			       sizeof(struct wx_ring));
 
 			temp_ring[i].count = new_rx_count;
@@ -3389,8 +3366,10 @@ int wx_set_ring(struct wx *wx, u32 new_tx_count,
 		}
 
 		for (i = 0; i < wx->num_rx_queues; i++) {
-			wx_free_rx_resources(wx->rx_ring[i]);
-			memcpy(wx->rx_ring[i], &temp_ring[i],
+			struct wx_ring *rx_ring = rtnl_dereference(wx->rx_ring[i]);
+
+			wx_free_rx_resources(rx_ring);
+			memcpy(rx_ring, &temp_ring[i],
 			       sizeof(struct wx_ring));
 		}
 
diff --git a/drivers/net/ethernet/wangxun/libwx/wx_ptp.c b/drivers/net/ethernet/wangxun/libwx/wx_ptp.c
index e2d60d6bcade9..3104a4d24474b 100644
--- a/drivers/net/ethernet/wangxun/libwx/wx_ptp.c
+++ b/drivers/net/ethernet/wangxun/libwx/wx_ptp.c
@@ -311,11 +311,15 @@ static void wx_ptp_rx_hang(struct wx *wx)
 
 	/* determine the most recent watchdog or rx_timestamp event */
 	rx_event = wx->last_rx_ptp_check;
+	rcu_read_lock();
 	for (n = 0; n < wx->num_rx_queues; n++) {
-		rx_ring = wx->rx_ring[n];
+		rx_ring = rcu_dereference(wx->rx_ring[n]);
+		if (!rx_ring)
+			continue;
 		if (time_after(rx_ring->last_rx_timestamp, rx_event))
 			rx_event = rx_ring->last_rx_timestamp;
 	}
+	rcu_read_unlock();
 
 	/* only need to read the high RXSTMP register to clear the lock */
 	if (time_is_before_jiffies(rx_event + 5 * HZ)) {
diff --git a/drivers/net/ethernet/wangxun/libwx/wx_type.h b/drivers/net/ethernet/wangxun/libwx/wx_type.h
index a7f3c48e9603c..fdac611bd7f65 100644
--- a/drivers/net/ethernet/wangxun/libwx/wx_type.h
+++ b/drivers/net/ethernet/wangxun/libwx/wx_type.h
@@ -1359,8 +1359,16 @@ struct wx {
 	u32 tx_ring_count;
 	u32 rx_ring_count;
 
-	struct wx_ring *tx_ring[64] ____cacheline_aligned_in_smp;
-	struct wx_ring *rx_ring[64];
+	/* Ring pointers are published and cleared by the interrupt scheme
+	 * setup and teardown paths with RCU pointer operations. Control-path
+	 * readers use rtnl_dereference() while holding RTNL. Lockless readers
+	 * (such as xmit, stats, and watchdog paths) must use rcu_dereference()
+	 * inside an RCU read-side critical section and NULL-check the result.
+	 * RCU protects pointer lifetime; fields initialized after publication
+	 * need their own ordering guarantees.
+	 */
+	struct wx_ring __rcu *tx_ring[64] ____cacheline_aligned_in_smp;
+	struct wx_ring __rcu *rx_ring[64];
 	struct wx_q_vector *q_vector[64];
 	int num_rx_pools;
 	int num_rx_queues_per_pool;
diff --git a/drivers/net/ethernet/wangxun/libwx/wx_vf_common.c b/drivers/net/ethernet/wangxun/libwx/wx_vf_common.c
index 26de78e9a69ef..61e138ac91784 100644
--- a/drivers/net/ethernet/wangxun/libwx/wx_vf_common.c
+++ b/drivers/net/ethernet/wangxun/libwx/wx_vf_common.c
@@ -205,7 +205,7 @@ static void wx_configure_rx_vf(struct wx *wx)
 	 * the Base and Length of the Rx Descriptor Ring
 	 */
 	for (i = 0; i < wx->num_rx_queues; i++) {
-		struct wx_ring *rx_ring = wx->rx_ring[i];
+		struct wx_ring *rx_ring = rtnl_dereference(wx->rx_ring[i]);
 #ifdef HAVE_SWIOTLB_SKIP_CPU_SYNC
 		wx_set_rx_buffer_len_vf(wx, rx_ring);
 #endif
diff --git a/drivers/net/ethernet/wangxun/libwx/wx_vf_lib.c b/drivers/net/ethernet/wangxun/libwx/wx_vf_lib.c
index 7325b475ee10c..0ccf1de4010dc 100644
--- a/drivers/net/ethernet/wangxun/libwx/wx_vf_lib.c
+++ b/drivers/net/ethernet/wangxun/libwx/wx_vf_lib.c
@@ -175,8 +175,11 @@ void wx_configure_tx_vf(struct wx *wx)
 	u32 i;
 
 	/* Setup the HW Tx Head and Tail descriptor pointers */
-	for (i = 0; i < wx->num_tx_queues; i++)
-		wx_configure_tx_ring_vf(wx, wx->tx_ring[i]);
+	for (i = 0; i < wx->num_tx_queues; i++) {
+		struct wx_ring *tx_ring = rtnl_dereference(wx->tx_ring[i]);
+
+		wx_configure_tx_ring_vf(wx, tx_ring);
+	}
 }
 
 static void wx_configure_srrctl_vf(struct wx *wx, struct wx_ring *ring,
diff --git a/drivers/net/ethernet/wangxun/ngbe/ngbe_ethtool.c b/drivers/net/ethernet/wangxun/ngbe/ngbe_ethtool.c
index 1960f7154151d..2997a93f786c4 100644
--- a/drivers/net/ethernet/wangxun/ngbe/ngbe_ethtool.c
+++ b/drivers/net/ethernet/wangxun/ngbe/ngbe_ethtool.c
@@ -37,9 +37,9 @@ static int ngbe_set_ringparam(struct net_device *netdev,
 
 	if (!netif_running(wx->netdev)) {
 		for (i = 0; i < wx->num_tx_queues; i++)
-			wx->tx_ring[i]->count = new_tx_count;
+			rtnl_dereference(wx->tx_ring[i])->count = new_tx_count;
 		for (i = 0; i < wx->num_rx_queues; i++)
-			wx->rx_ring[i]->count = new_rx_count;
+			rtnl_dereference(wx->rx_ring[i])->count = new_rx_count;
 		wx->tx_ring_count = new_tx_count;
 		wx->rx_ring_count = new_rx_count;
 
diff --git a/drivers/net/ethernet/wangxun/ngbe/ngbe_main.c b/drivers/net/ethernet/wangxun/ngbe/ngbe_main.c
index 0aea5a99a1e2c..bef7089d0e47e 100644
--- a/drivers/net/ethernet/wangxun/ngbe/ngbe_main.c
+++ b/drivers/net/ethernet/wangxun/ngbe/ngbe_main.c
@@ -406,7 +406,7 @@ static void ngbe_disable_device(struct wx *wx)
 	/* disable all enabled rx queues */
 	for (i = 0; i < wx->num_rx_queues; i++)
 		/* this call also flushes the previous write */
-		wx_disable_rx_queue(wx, wx->rx_ring[i]);
+		wx_disable_rx_queue(wx, rtnl_dereference(wx->rx_ring[i]));
 	/* disable receives */
 	wx_disable_rx(wx);
 	wx_napi_disable_all(wx);
@@ -422,7 +422,7 @@ static void ngbe_disable_device(struct wx *wx)
 	wx_irq_disable(wx);
 	/* disable transmits in the hardware now that interrupts are off */
 	for (i = 0; i < wx->num_tx_queues; i++) {
-		u8 reg_idx = wx->tx_ring[i]->reg_idx;
+		u8 reg_idx = rtnl_dereference(wx->tx_ring[i])->reg_idx;
 
 		wr32(wx, WX_PX_TR_CFG(reg_idx), WX_PX_TR_CFG_SWFLSH);
 	}
@@ -580,8 +580,11 @@ static void ngbe_dev_shutdown(struct pci_dev *pdev, bool *enable_wake)
 	if (netif_running(netdev))
 		ngbe_close(netdev);
 	wx_clear_interrupt_scheme(wx);
-	rtnl_unlock();
 
+	/* Keep RTNL held across the WoL setup so wx_configure_rx() runs under
+	 * the same lock as its other callers, and cannot race with a channel
+	 * reconfiguration.
+	 */
 	if (wufc) {
 		wx_set_rx_mode(netdev);
 		wx_configure_rx(wx);
@@ -589,6 +592,8 @@ static void ngbe_dev_shutdown(struct pci_dev *pdev, bool *enable_wake)
 	} else {
 		wr32(wx, WX_PSR_WKUP_CTL, 0);
 	}
+	rtnl_unlock();
+
 	pci_wake_from_d3(pdev, !!wufc);
 	*enable_wake = !!wufc;
 	wx_control_hw(wx, false);
diff --git a/drivers/net/ethernet/wangxun/txgbe/txgbe_ethtool.c b/drivers/net/ethernet/wangxun/txgbe/txgbe_ethtool.c
index 3e32aca728065..d2d43fec01572 100644
--- a/drivers/net/ethernet/wangxun/txgbe/txgbe_ethtool.c
+++ b/drivers/net/ethernet/wangxun/txgbe/txgbe_ethtool.c
@@ -61,9 +61,9 @@ static int txgbe_set_ringparam(struct net_device *netdev,
 
 	if (!netif_running(wx->netdev)) {
 		for (i = 0; i < wx->num_tx_queues; i++)
-			wx->tx_ring[i]->count = new_tx_count;
+			rtnl_dereference(wx->tx_ring[i])->count = new_tx_count;
 		for (i = 0; i < wx->num_rx_queues; i++)
-			wx->rx_ring[i]->count = new_rx_count;
+			rtnl_dereference(wx->rx_ring[i])->count = new_rx_count;
 		wx->tx_ring_count = new_tx_count;
 		wx->rx_ring_count = new_rx_count;
 
@@ -366,7 +366,7 @@ static int txgbe_add_ethtool_fdir_entry(struct txgbe *txgbe,
 
 		/* Map the ring onto the absolute queue index */
 		if (!vf)
-			queue = wx->rx_ring[ring]->reg_idx;
+			queue = rtnl_dereference(wx->rx_ring[ring])->reg_idx;
 		else
 			queue = ((vf - 1) * wx->num_rx_queues_per_pool) + ring;
 	}
diff --git a/drivers/net/ethernet/wangxun/txgbe/txgbe_fdir.c b/drivers/net/ethernet/wangxun/txgbe/txgbe_fdir.c
index 59a47532618c8..e0a18aae27859 100644
--- a/drivers/net/ethernet/wangxun/txgbe/txgbe_fdir.c
+++ b/drivers/net/ethernet/wangxun/txgbe/txgbe_fdir.c
@@ -606,7 +606,7 @@ static void txgbe_fdir_filter_restore(struct wx *wx)
 
 			/* Map the ring onto the absolute queue index */
 			if (!vf)
-				queue = wx->rx_ring[ring]->reg_idx;
+				queue = rtnl_dereference(wx->rx_ring[ring])->reg_idx;
 			else
 				queue = ((vf - 1) * wx->num_rx_queues_per_pool) + ring;
 		}
diff --git a/drivers/net/ethernet/wangxun/txgbe/txgbe_main.c b/drivers/net/ethernet/wangxun/txgbe/txgbe_main.c
index 184194200917b..3111409861590 100644
--- a/drivers/net/ethernet/wangxun/txgbe/txgbe_main.c
+++ b/drivers/net/ethernet/wangxun/txgbe/txgbe_main.c
@@ -239,7 +239,7 @@ static void txgbe_disable_device(struct wx *wx)
 	/* disable all enabled rx queues */
 	for (i = 0; i < wx->num_rx_queues; i++)
 		/* this call also flushes the previous write */
-		wx_disable_rx_queue(wx, wx->rx_ring[i]);
+		wx_disable_rx_queue(wx, rtnl_dereference(wx->rx_ring[i]));
 
 	netif_tx_stop_all_queues(netdev);
 	netif_tx_disable(netdev);
@@ -275,7 +275,7 @@ static void txgbe_disable_device(struct wx *wx)
 
 	/* disable transmits in the hardware now that interrupts are off */
 	for (i = 0; i < wx->num_tx_queues; i++) {
-		u8 reg_idx = wx->tx_ring[i]->reg_idx;
+		u8 reg_idx = rtnl_dereference(wx->tx_ring[i])->reg_idx;
 
 		wr32(wx, WX_PX_TR_CFG(reg_idx), WX_PX_TR_CFG_SWFLSH);
 	}
-- 
2.30.1


^ permalink raw reply related	[flat|nested] 4+ messages in thread

* [PATCH net v8 2/2] net: wangxun: access queue counts consistently
  2026-11-18  8:13 [PATCH net v8 0/2] net: wangxun: protect ring accesses with RCU Mengyuan Lou
  2026-11-18  8:13 ` [PATCH net v8 1/2] " Mengyuan Lou
@ 2026-11-18  8:13 ` Mengyuan Lou
  2026-11-18  8:18 ` [PATCH net v8 0/2] net: wangxun: protect ring accesses with RCU netdev-bot+sinfo
  2 siblings, 0 replies; 4+ messages in thread
From: Mengyuan Lou @ 2026-11-18  8:13 UTC (permalink / raw)
  To: netdev
  Cc: jiawenwu, duanqiangwen, linglingzhang, andrew+netdev, davem,
	edumazet, kuba, pabeni, hramamurthy, jacob.e.keller,
	przemyslaw.kitszel, Mengyuan Lou

Queue-vector teardown clears wx->num_tx_queues and
wx->num_rx_queues while the netdev can remain registered. Lockless
readers may use these counts as ring-array loop bounds while teardown
is clearing the ring slots. wx_xmit_frame() also uses the Tx count to
select a ring and must not divide by zero during teardown.

Use READ_ONCE() to take a single count snapshot in the lockless readers,
and use WRITE_ONCE() for queue-count updates during scheme setup,
teardown, and allocation-error handling, including the VF queue setup
callbacks. Keep the existing NULL checks so a reader that observed an
old non-zero count can safely skip ring slots that have since been
cleared. wx_xmit_frame() retains its zero-count guard before
queue-index modulo.

Fixes: 3f703186113f ("net: libwx: Add irq flow functions")
Signed-off-by: Mengyuan Lou <mengyuanlou@net-swift.com>
---
 drivers/net/ethernet/wangxun/libwx/wx_err.c   |  7 +++--
 drivers/net/ethernet/wangxun/libwx/wx_hw.c    | 23 ++++++++++-----
 drivers/net/ethernet/wangxun/libwx/wx_lib.c   | 29 ++++++++++---------
 drivers/net/ethernet/wangxun/libwx/wx_ptp.c   |  5 ++--
 .../net/ethernet/wangxun/ngbevf/ngbevf_main.c |  4 +--
 .../ethernet/wangxun/txgbevf/txgbevf_main.c   | 10 +++----
 6 files changed, 46 insertions(+), 32 deletions(-)

diff --git a/drivers/net/ethernet/wangxun/libwx/wx_err.c b/drivers/net/ethernet/wangxun/libwx/wx_err.c
index 8eadb4d3abcff..fa0f7c57090f5 100644
--- a/drivers/net/ethernet/wangxun/libwx/wx_err.c
+++ b/drivers/net/ethernet/wangxun/libwx/wx_err.c
@@ -199,10 +199,11 @@ EXPORT_SYMBOL(wx_init_err_task);
 
 static bool wx_ring_tx_pending(struct wx *wx)
 {
+	int num_tx_queues = READ_ONCE(wx->num_tx_queues);
 	int i;
 
 	rcu_read_lock();
-	for (i = 0; i < wx->num_tx_queues; i++) {
+	for (i = 0; i < num_tx_queues; i++) {
 		struct wx_ring *tx_ring = rcu_dereference(wx->tx_ring[i]);
 
 		if (!tx_ring)
@@ -261,6 +262,7 @@ static void wx_watchdog_flush_tx(struct wx *wx)
 
 static void wx_detect_tx_hang(struct wx *wx)
 {
+	int num_tx_queues;
 	int i;
 
 	/* If we're down or resetting, just bail */
@@ -270,8 +272,9 @@ static void wx_detect_tx_hang(struct wx *wx)
 
 	/* Force detection of hung controller */
 	if (netif_carrier_ok(wx->netdev)) {
+		num_tx_queues = READ_ONCE(wx->num_tx_queues);
 		rcu_read_lock();
-		for (i = 0; i < wx->num_tx_queues; i++) {
+		for (i = 0; i < num_tx_queues; i++) {
 			struct wx_ring *tx_ring = rcu_dereference(wx->tx_ring[i]);
 
 			if (tx_ring)
diff --git a/drivers/net/ethernet/wangxun/libwx/wx_hw.c b/drivers/net/ethernet/wangxun/libwx/wx_hw.c
index 24c539ce6acbe..c285c1de24ba9 100644
--- a/drivers/net/ethernet/wangxun/libwx/wx_hw.c
+++ b/drivers/net/ethernet/wangxun/libwx/wx_hw.c
@@ -1596,10 +1596,11 @@ EXPORT_SYMBOL(wx_enable_sec_rx_path);
 
 static void wx_vlan_strip_control(struct wx *wx, bool enable)
 {
+	int num_rx_queues = READ_ONCE(wx->num_rx_queues);
 	int i, j;
 
 	rcu_read_lock();
-	for (i = 0; i < wx->num_rx_queues; i++) {
+	for (i = 0; i < num_rx_queues; i++) {
 		struct wx_ring *ring = rcu_dereference(wx->rx_ring[i]);
 
 		if (!ring)
@@ -2777,6 +2778,7 @@ static void wx_disable_rx_drop(struct wx *wx, struct wx_ring *ring)
 
 int wx_fc_enable(struct wx *wx, bool tx_pause, bool rx_pause)
 {
+	int num_rx_queues = READ_ONCE(wx->num_rx_queues);
 	u16 pause_time = WX_DEFAULT_FCPAUSE;
 	u32 mflcn_reg, fccfg_reg, reg;
 	u32 fcrtl, fcrth;
@@ -2849,8 +2851,8 @@ int wx_fc_enable(struct wx *wx, bool tx_pause, bool rx_pause)
 	 *  and performance reasons.
 	 */
 	rcu_read_lock();
-	if (wx->num_rx_queues > 1 && !tx_pause) {
-		for (i = 0; i < wx->num_rx_queues; i++) {
+	if (num_rx_queues > 1 && !tx_pause) {
+		for (i = 0; i < num_rx_queues; i++) {
 			struct wx_ring *rx_ring = rcu_dereference(wx->rx_ring[i]);
 
 			if (!rx_ring)
@@ -2858,7 +2860,7 @@ int wx_fc_enable(struct wx *wx, bool tx_pause, bool rx_pause)
 			wx_enable_rx_drop(wx, rx_ring);
 		}
 	} else {
-		for (i = 0; i < wx->num_rx_queues; i++) {
+		for (i = 0; i < num_rx_queues; i++) {
 			struct wx_ring *rx_ring = rcu_dereference(wx->rx_ring[i]);
 
 			if (!rx_ring)
@@ -2875,6 +2877,7 @@ EXPORT_SYMBOL(wx_fc_enable);
 static void wx_update_xoff_rx_lfc(struct wx *wx)
 {
 	struct wx_hw_stats *hwstats = &wx->stats;
+	int num_tx_queues;
 	u64 data;
 	int i;
 
@@ -2893,8 +2896,9 @@ static void wx_update_xoff_rx_lfc(struct wx *wx)
 	if (!data)
 		return;
 
+	num_tx_queues = READ_ONCE(wx->num_tx_queues);
 	rcu_read_lock();
-	for (i = 0; i < wx->num_tx_queues; i++) {
+	for (i = 0; i < num_tx_queues; i++) {
 		struct wx_ring *tx_ring = rcu_dereference(wx->tx_ring[i]);
 
 		if (!tx_ring)
@@ -2915,6 +2919,7 @@ void wx_update_stats(struct wx *wx)
 	u64 non_eop_descs = 0, alloc_rx_buff_failed = 0;
 	u64 hw_csum_rx_good = 0, hw_csum_rx_error = 0;
 	u64 restart_queue = 0, tx_busy = 0;
+	int num_rx_queues, num_tx_queues;
 	u32 i;
 
 	if (test_bit(WX_STATE_DOWN, wx->state) ||
@@ -2923,9 +2928,11 @@ void wx_update_stats(struct wx *wx)
 
 	spin_lock(&wx->hw_stats_lock);
 
+	num_rx_queues = READ_ONCE(wx->num_rx_queues);
+	num_tx_queues = READ_ONCE(wx->num_tx_queues);
 	rcu_read_lock();
 	/* gather some stats to the wx struct that are per queue */
-	for (i = 0; i < wx->num_rx_queues; i++) {
+	for (i = 0; i < num_rx_queues; i++) {
 		struct wx_ring *rx_ring = rcu_dereference(wx->rx_ring[i]);
 
 		if (!rx_ring)
@@ -2944,7 +2951,7 @@ void wx_update_stats(struct wx *wx)
 		u64 rsc_count = 0;
 		u64 rsc_flush = 0;
 
-		for (i = 0; i < wx->num_rx_queues; i++) {
+		for (i = 0; i < num_rx_queues; i++) {
 			struct wx_ring *rx_ring = rcu_dereference(wx->rx_ring[i]);
 
 			if (!rx_ring)
@@ -2957,7 +2964,7 @@ void wx_update_stats(struct wx *wx)
 		wx->rsc_flush = rsc_flush;
 	}
 
-	for (i = 0; i < wx->num_tx_queues; i++) {
+	for (i = 0; i < num_tx_queues; i++) {
 		struct wx_ring *tx_ring = rcu_dereference(wx->tx_ring[i]);
 
 		if (!tx_ring)
diff --git a/drivers/net/ethernet/wangxun/libwx/wx_lib.c b/drivers/net/ethernet/wangxun/libwx/wx_lib.c
index 0fbaa585bb854..e931c57837578 100644
--- a/drivers/net/ethernet/wangxun/libwx/wx_lib.c
+++ b/drivers/net/ethernet/wangxun/libwx/wx_lib.c
@@ -1734,7 +1734,7 @@ netdev_tx_t wx_xmit_frame(struct sk_buff *skb,
 		return NETDEV_TX_OK;
 	}
 
-	num_tx_queues = wx->num_tx_queues;
+	num_tx_queues = READ_ONCE(wx->num_tx_queues);
 	if (unlikely(!num_tx_queues))
 		goto drop;
 
@@ -1897,8 +1897,8 @@ static bool wx_set_vmdq_queues(struct wx *wx)
 	wx->num_rx_pools = vmdq_i;
 	wx->num_rx_queues_per_pool = rss_i;
 
-	wx->num_rx_queues = vmdq_i * rss_i;
-	wx->num_tx_queues = vmdq_i * rss_i;
+	WRITE_ONCE(wx->num_rx_queues, vmdq_i * rss_i);
+	WRITE_ONCE(wx->num_tx_queues, vmdq_i * rss_i);
 
 	return true;
 }
@@ -1943,15 +1943,15 @@ static void wx_set_rss_queues(struct wx *wx)
 	}
 
 out:
-	wx->num_rx_queues = f->indices;
-	wx->num_tx_queues = f->indices;
+	WRITE_ONCE(wx->num_rx_queues, f->indices);
+	WRITE_ONCE(wx->num_tx_queues, f->indices);
 }
 
 static void wx_set_num_queues(struct wx *wx)
 {
 	/* Start with base case */
-	wx->num_rx_queues = 1;
-	wx->num_tx_queues = 1;
+	WRITE_ONCE(wx->num_rx_queues, 1);
+	WRITE_ONCE(wx->num_tx_queues, 1);
 	wx->queues_per_pool = 1;
 
 	if (wx_set_vmdq_queues(wx))
@@ -2288,8 +2288,8 @@ static int wx_alloc_q_vectors(struct wx *wx)
 	return 0;
 
 err_out:
-	wx->num_tx_queues = 0;
-	wx->num_rx_queues = 0;
+	WRITE_ONCE(wx->num_tx_queues, 0);
+	WRITE_ONCE(wx->num_rx_queues, 0);
 	wx->num_q_vectors = 0;
 
 	while (v_idx--)
@@ -2310,8 +2310,8 @@ static void wx_free_q_vectors(struct wx *wx)
 {
 	int v_idx = wx->num_q_vectors;
 
-	wx->num_tx_queues = 0;
-	wx->num_rx_queues = 0;
+	WRITE_ONCE(wx->num_tx_queues, 0);
+	WRITE_ONCE(wx->num_rx_queues, 0);
 	wx->num_q_vectors = 0;
 
 	while (v_idx--)
@@ -3085,13 +3085,16 @@ void wx_get_stats64(struct net_device *netdev,
 		    struct rtnl_link_stats64 *stats)
 {
 	struct wx *wx = netdev_priv(netdev);
+	int num_rx_queues, num_tx_queues;
 	struct wx_hw_stats *hwstats;
 	int i;
 
 	wx_update_stats(wx);
 
+	num_rx_queues = READ_ONCE(wx->num_rx_queues);
+	num_tx_queues = READ_ONCE(wx->num_tx_queues);
 	rcu_read_lock();
-	for (i = 0; i < wx->num_rx_queues; i++) {
+	for (i = 0; i < num_rx_queues; i++) {
 		struct wx_ring *ring = rcu_dereference(wx->rx_ring[i]);
 		u64 bytes, packets;
 		unsigned int start;
@@ -3107,7 +3110,7 @@ void wx_get_stats64(struct net_device *netdev,
 		}
 	}
 
-	for (i = 0; i < wx->num_tx_queues; i++) {
+	for (i = 0; i < num_tx_queues; i++) {
 		struct wx_ring *ring = rcu_dereference(wx->tx_ring[i]);
 		u64 bytes, packets;
 		unsigned int start;
diff --git a/drivers/net/ethernet/wangxun/libwx/wx_ptp.c b/drivers/net/ethernet/wangxun/libwx/wx_ptp.c
index 3104a4d24474b..9dec1698c6d33 100644
--- a/drivers/net/ethernet/wangxun/libwx/wx_ptp.c
+++ b/drivers/net/ethernet/wangxun/libwx/wx_ptp.c
@@ -296,8 +296,8 @@ static void wx_ptp_rx_hang(struct wx *wx)
 {
 	struct wx_ring *rx_ring;
 	unsigned long rx_event;
+	int num_rx_queues, n;
 	u32 tsyncrxctl;
-	int n;
 
 	tsyncrxctl = rd32(wx, WX_PSR_1588_CTL);
 
@@ -311,8 +311,9 @@ static void wx_ptp_rx_hang(struct wx *wx)
 
 	/* determine the most recent watchdog or rx_timestamp event */
 	rx_event = wx->last_rx_ptp_check;
+	num_rx_queues = READ_ONCE(wx->num_rx_queues);
 	rcu_read_lock();
-	for (n = 0; n < wx->num_rx_queues; n++) {
+	for (n = 0; n < num_rx_queues; n++) {
 		rx_ring = rcu_dereference(wx->rx_ring[n]);
 		if (!rx_ring)
 			continue;
diff --git a/drivers/net/ethernet/wangxun/ngbevf/ngbevf_main.c b/drivers/net/ethernet/wangxun/ngbevf/ngbevf_main.c
index d79cf8d8484f9..c40b64564d41f 100644
--- a/drivers/net/ethernet/wangxun/ngbevf/ngbevf_main.c
+++ b/drivers/net/ethernet/wangxun/ngbevf/ngbevf_main.c
@@ -53,8 +53,8 @@ static const struct net_device_ops ngbevf_netdev_ops = {
 static void ngbevf_set_num_queues(struct wx *wx)
 {
 	/* Start with base case */
-	wx->num_rx_queues = 1;
-	wx->num_tx_queues = 1;
+	WRITE_ONCE(wx->num_rx_queues, 1);
+	WRITE_ONCE(wx->num_tx_queues, 1);
 }
 
 static int ngbevf_sw_init(struct wx *wx)
diff --git a/drivers/net/ethernet/wangxun/txgbevf/txgbevf_main.c b/drivers/net/ethernet/wangxun/txgbevf/txgbevf_main.c
index 8b16b900820aa..f1784c0ed114d 100644
--- a/drivers/net/ethernet/wangxun/txgbevf/txgbevf_main.c
+++ b/drivers/net/ethernet/wangxun/txgbevf/txgbevf_main.c
@@ -53,8 +53,8 @@ static void txgbevf_set_num_queues(struct wx *wx)
 	int ret = 0;
 
 	/* Start with base case */
-	wx->num_rx_queues = 1;
-	wx->num_tx_queues = 1;
+	WRITE_ONCE(wx->num_rx_queues, 1);
+	WRITE_ONCE(wx->num_tx_queues, 1);
 
 	spin_lock_bh(&wx->mbx.mbx_lock);
 	/* fetch queue configuration from the PF */
@@ -66,15 +66,15 @@ static void txgbevf_set_num_queues(struct wx *wx)
 
 	/* we need as many queues as traffic classes */
 	if (num_tcs > 1) {
-		wx->num_rx_queues = num_tcs;
+		WRITE_ONCE(wx->num_rx_queues, num_tcs);
 	} else {
 		rss = min_t(u16, num_online_cpus(), TXGBEVF_MAX_RSS_NUM);
 		queue = min_t(u16, wx->mac.max_rx_queues, wx->mac.max_tx_queues);
 		rss = min_t(u16, queue, rss);
 
 		if (wx->vfinfo->vf_api >= wx_mbox_api_13) {
-			wx->num_rx_queues = rss;
-			wx->num_tx_queues = rss;
+			WRITE_ONCE(wx->num_rx_queues, rss);
+			WRITE_ONCE(wx->num_tx_queues, rss);
 		}
 	}
 }
-- 
2.30.1


^ permalink raw reply related	[flat|nested] 4+ messages in thread

* Re: [PATCH net v8 0/2] net: wangxun: protect ring accesses with RCU
  2026-11-18  8:13 [PATCH net v8 0/2] net: wangxun: protect ring accesses with RCU Mengyuan Lou
  2026-11-18  8:13 ` [PATCH net v8 1/2] " Mengyuan Lou
  2026-11-18  8:13 ` [PATCH net v8 2/2] net: wangxun: access queue counts consistently Mengyuan Lou
@ 2026-11-18  8:18 ` netdev-bot+sinfo
  2 siblings, 0 replies; 4+ messages in thread
From: netdev-bot+sinfo @ 2026-11-18  8:18 UTC (permalink / raw)
  To: Mengyuan Lou
  Cc: netdev, jiawenwu, duanqiangwen, linglingzhang, andrew+netdev,
	davem, edumazet, kuba, pabeni, hramamurthy, jacob.e.keller,
	przemyslaw.kitszel

Hi!

This is an automated message. This series looks like a fix, but its
commit messages seem to be missing some information:

 - How the issue was discovered, e.g. hit in production, hit during
   development, syzbot report, manual code inspection, LLM or static
   analysis tool scan.

 - Whether the issue was actually triggered, or is only theoretical
   (e.g. found by LLM and not proven at runtime). If it was triggered
   please include the symptoms, like the stack trace or kernel messages.

 - What hardware the change was tested on. For driver fixes please
   mention the device (and if relevant firmware version) used for
   testing, or say that the change was not tested on real hardware.

Please do not repost the series just to address the above. Instead,
reply to this email with the missing information, so that reviewers
can take it into account.

*IMPORTANT*: If you do repost (for other reasons) - please include
the relevant info in the cover letter or commit messages.
-- 
The evaluation is done by an LLM so it may be wrong, if you think
that is the case please reply and explain.

^ permalink raw reply	[flat|nested] 4+ messages in thread

end of thread, other threads:[~2026-11-18  8:18 UTC | newest]

Thread overview: 4+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2026-11-18  8:13 [PATCH net v8 0/2] net: wangxun: protect ring accesses with RCU Mengyuan Lou
2026-11-18  8:13 ` [PATCH net v8 1/2] " Mengyuan Lou
2026-11-18  8:13 ` [PATCH net v8 2/2] net: wangxun: access queue counts consistently Mengyuan Lou
2026-11-18  8:18 ` [PATCH net v8 0/2] net: wangxun: protect ring accesses with RCU netdev-bot+sinfo

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox