[PATCH net 0/5] mlx5 misc fixes 2026-03-05

[PATCH net 0/5] mlx5 misc fixes 2026-03-05

[Tariq Toukan]

Hi,

This patchset provides misc bug fixes from the team to the mlx5
core and Eth drivers.

Thanks,
Tariq.

Carolina Jubran (1):
  net/mlx5: Fix peer miss rules host disabled checks

Dragos Tatulea (2):
  net/mlx5e: RX, Fix XDP multi-buf frag counting for striding RQ
  net/mlx5e: RX, Fix XDP multi-buf frag counting for legacy RQ

Gal Pressman (1):
  net/mlx5e: Fix DMA FIFO desync on error CQE SQ recovery

Patrisious Haddad (1):
  net/mlx5: Fix crash when moving to switchdev mode

 .../mellanox/mlx5/core/en/reporter_tx.c       |  1 -
 .../mellanox/mlx5/core/en_accel/ipsec_fs.c    |  2 +-
 .../net/ethernet/mellanox/mlx5/core/en_rx.c   | 23 +++++++---------
 .../mellanox/mlx5/core/eswitch_offloads.c     | 27 +++++++++----------
 4 files changed, 22 insertions(+), 31 deletions(-)


base-commit: ae779bcb18cb0ef0da1402b9dd837e2084e23e27
-- 
2.44.0

[PATCH net 1/5] net/mlx5: Fix crash when moving to switchdev mode

[Tariq Toukan]

From: Patrisious Haddad <phaddad@nvidia.com>

When moving to switchdev mode when the device doesn't support IPsec,
we try to clean up the IPsec resources anyway which causes the crash
below, fix that by correctly checking for IPsec support before trying
to clean up its resources.

[27642.515799] WARNING: arch/x86/mm/fault.c:1276 at
do_user_addr_fault+0x18a/0x680, CPU#4: devlink/6490
[27642.517159] Modules linked in: xt_conntrack xt_MASQUERADE
ip6table_nat ip6table_filter ip6_tables iptable_nat nf_nat xt_addrtype
rpcsec_gss_krb5 auth_rpcgss oid_registry overlay mlx5_fwctl nfnetlink
zram zsmalloc mlx5_ib fuse rpcrdma rdma_ucm ib_uverbs ib_iser libiscsi
scsi_transport_iscsi ib_umad rdma_cm ib_ipoib iw_cm ib_cm mlx5_core
ib_core
[27642.521358] CPU: 4 UID: 0 PID: 6490 Comm: devlink Not tainted
6.19.0-rc5_for_upstream_min_debug_2026_01_14_16_47 #1 NONE
[27642.522923] Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS
rel-1.16.3-0-ga6ed6b701f0a-prebuilt.qemu.org 04/01/2014
[27642.524528] RIP: 0010:do_user_addr_fault+0x18a/0x680
[27642.525362] Code: ff 0f 84 75 03 00 00 48 89 ee 4c 89 e7 e8 5e b9 22
00 49 89 c0 48 85 c0 0f 84 a8 02 00 00 f7 c3 60 80 00 00 74 22 31 c9 eb
   ae <0f> 0b 48 83 c4 10 48 89 ea 48 89 de 4c 89 f7 5b 5d 41 5c 41 5d
41
[27642.528166] RSP: 0018:ffff88810770f6b8 EFLAGS: 00010046
[27642.529038] RAX: 0000000000000000 RBX: 0000000000000002 RCX:
ffff88810b980f00
[27642.530158] RDX: 00000000000000a0 RSI: 0000000000000002 RDI:
ffff88810770f728
[27642.531270] RBP: 00000000000000a0 R08: 0000000000000000 R09:
0000000000000000
[27642.532383] R10: 0000000000000000 R11: 0000000000000000 R12:
ffff888103f3c4c0
[27642.533499] R13: 0000000000000000 R14: ffff88810770f728 R15:
0000000000000000
[27642.534614] FS:  00007f197c741740(0000) GS:ffff88856a94c000(0000)
knlGS:0000000000000000
[27642.535915] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[27642.536858] CR2: 00000000000000a0 CR3: 000000011334c003 CR4:
0000000000172eb0
[27642.537982] Call Trace:
[27642.538466]  <TASK>
[27642.538907]  exc_page_fault+0x76/0x140
[27642.539583]  asm_exc_page_fault+0x22/0x30
[27642.540282] RIP: 0010:_raw_spin_lock_irqsave+0x10/0x30
[27642.541134] Code: 07 85 c0 75 11 ba ff 00 00 00 f0 0f b1 17 75 06 b8
01 00 00 00 c3 31 c0 c3 90 0f 1f 44 00 00 53 9c 5b fa 31 c0 ba 01 00 00
   00 <f0> 0f b1 17 75 05 48 89 d8 5b c3 89 c6 e8 7e 02 00 00 48 89 d8
      5b
[27642.543936] RSP: 0018:ffff88810770f7d8 EFLAGS: 00010046
[27642.544803] RAX: 0000000000000000 RBX: 0000000000000202 RCX:
ffff888113ad96d8
[27642.545916] RDX: 0000000000000001 RSI: ffff88810770f818 RDI:
00000000000000a0
[27642.547027] RBP: 0000000000000098 R08: 0000000000000400 R09:
ffff88810b980f00
[27642.548140] R10: 0000000000000001 R11: ffff888101845a80 R12:
00000000000000a8
[27642.549263] R13: ffffffffa02a9060 R14: 00000000000000a0 R15:
ffff8881130d8a40
[27642.550379]  complete_all+0x20/0x90
[27642.551010]  mlx5e_ipsec_disable_events+0xb6/0xf0 [mlx5_core]
[27642.552022]  mlx5e_nic_disable+0x12d/0x220 [mlx5_core]
[27642.552929]  mlx5e_detach_netdev+0x66/0xf0 [mlx5_core]
[27642.553822]  mlx5e_netdev_change_profile+0x5b/0x120 [mlx5_core]
[27642.554821]  mlx5e_vport_rep_load+0x419/0x590 [mlx5_core]
[27642.555757]  ? xa_load+0x53/0x90
[27642.556361]  __esw_offloads_load_rep+0x54/0x70 [mlx5_core]
[27642.557328]  mlx5_esw_offloads_rep_load+0x45/0xd0 [mlx5_core]
[27642.558320]  esw_offloads_enable+0xb4b/0xc90 [mlx5_core]
[27642.559247]  mlx5_eswitch_enable_locked+0x34e/0x4f0 [mlx5_core]
[27642.560257]  ? mlx5_rescan_drivers_locked+0x222/0x2d0 [mlx5_core]
[27642.561284]  mlx5_devlink_eswitch_mode_set+0x5ac/0x9c0 [mlx5_core]
[27642.562334]  ? devlink_rate_set_ops_supported+0x21/0x3a0
[27642.563220]  devlink_nl_eswitch_set_doit+0x67/0xe0
[27642.564026]  genl_family_rcv_msg_doit+0xe0/0x130
[27642.564816]  genl_rcv_msg+0x183/0x290
[27642.565466]  ? __devlink_nl_pre_doit.isra.0+0x160/0x160
[27642.566329]  ? devlink_nl_eswitch_get_doit+0x290/0x290
[27642.567181]  ? devlink_nl_pre_doit_parent_dev_optional+0x20/0x20
[27642.568147]  ? genl_family_rcv_msg_dumpit+0xf0/0xf0
[27642.568966]  netlink_rcv_skb+0x4b/0xf0
[27642.569629]  genl_rcv+0x24/0x40
[27642.570215]  netlink_unicast+0x255/0x380
[27642.570901]  ? __alloc_skb+0xfa/0x1e0
[27642.571560]  netlink_sendmsg+0x1f3/0x420
[27642.572249]  __sock_sendmsg+0x38/0x60
[27642.572911]  __sys_sendto+0x119/0x180
[27642.573561]  ? __sys_recvmsg+0x5c/0xb0
[27642.574227]  __x64_sys_sendto+0x20/0x30
[27642.574904]  do_syscall_64+0x55/0xc10
[27642.575554]  entry_SYSCALL_64_after_hwframe+0x4b/0x53
[27642.576391] RIP: 0033:0x7f197c85e807
[27642.577050] Code: c7 c0 ff ff ff ff eb be 66 2e 0f 1f 84 00 00 00 00
00 90 f3 0f 1e fa 80 3d 45 08 0d 00 00 41 89 ca 74 10 b8 2c 00 00 00 0f
   05 <48> 3d 00 f0 ff ff 77 69 c3 55 48 89 e5 53 48 83 ec 38 44 89 4d
      d0
[27642.579846] RSP: 002b:00007ffebd4e2248 EFLAGS: 00000202 ORIG_RAX:
000000000000002c
[27642.581082] RAX: ffffffffffffffda RBX: 000055cfcd9cd2a0 RCX:
00007f197c85e807
[27642.582200] RDX: 0000000000000038 RSI: 000055cfcd9cd490 RDI:
0000000000000003
[27642.583320] RBP: 00007ffebd4e2290 R08: 00007f197c942200 R09:
000000000000000c
[27642.584437] R10: 0000000000000000 R11: 0000000000000202 R12:
0000000000000000
[27642.585555] R13: 000055cfcd9cd490 R14: 00007ffebd4e45d1 R15:
000055cfcd9cd2a0
[27642.586671]  </TASK>
[27642.587121] ---[ end trace 0000000000000000 ]---
[27642.587910] BUG: kernel NULL pointer dereference, address:
00000000000000a0

Fixes: 664f76be38a1 ("net/mlx5: Fix IPsec cleanup over MPV device")
Signed-off-by: Patrisious Haddad <phaddad@nvidia.com>
Reviewed-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
---
 drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_fs.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_fs.c b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_fs.c
index 197a1c6930c0..329608c59313 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_fs.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_fs.c
@@ -2912,7 +2912,7 @@ void mlx5e_ipsec_disable_events(struct mlx5e_priv *priv)
 		goto out;
 
 	peer_priv = mlx5_devcom_get_next_peer_data(priv->devcom, &tmp);
-	if (peer_priv)
+	if (peer_priv && peer_priv->ipsec)
 		complete_all(&peer_priv->ipsec->comp);
 
 	mlx5_devcom_for_each_peer_end(priv->devcom);
-- 
2.44.0

[PATCH net 2/5] net/mlx5: Fix peer miss rules host disabled checks

[Tariq Toukan]

From: Carolina Jubran <cjubran@nvidia.com>

The check on mlx5_esw_host_functions_enabled(esw->dev) for adding VF
peer miss rules is incorrect. These rules match traffic from peer's VFs,
so the local device's host function status is irrelevant. Remove this
check to ensure peer VF traffic is properly handled regardless of local
host configuration.

Also fix the PF peer miss rule deletion to be symmetric with the add
path, so only attempt to delete the rule if it was actually created.

Fixes: 520369ef43a8 ("net/mlx5: Support disabling host PFs")
Signed-off-by: Carolina Jubran <cjubran@nvidia.com>
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
---
 .../mellanox/mlx5/core/eswitch_offloads.c     | 27 +++++++++----------
 1 file changed, 12 insertions(+), 15 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
index 1366f6e489bd..2f55ea3f8bf8 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
@@ -1241,21 +1241,17 @@ static int esw_add_fdb_peer_miss_rules(struct mlx5_eswitch *esw,
 		flows[peer_vport->index] = flow;
 	}
 
-	if (mlx5_esw_host_functions_enabled(esw->dev)) {
-		mlx5_esw_for_each_vf_vport(peer_esw, i, peer_vport,
-					   mlx5_core_max_vfs(peer_dev)) {
-			esw_set_peer_miss_rule_source_port(esw, peer_esw,
-							   spec,
-							   peer_vport->vport);
-
-			flow = mlx5_add_flow_rules(mlx5_eswitch_get_slow_fdb(esw),
-						   spec, &flow_act, &dest, 1);
-			if (IS_ERR(flow)) {
-				err = PTR_ERR(flow);
-				goto add_vf_flow_err;
-			}
-			flows[peer_vport->index] = flow;
+	mlx5_esw_for_each_vf_vport(peer_esw, i, peer_vport,
+				   mlx5_core_max_vfs(peer_dev)) {
+		esw_set_peer_miss_rule_source_port(esw, peer_esw, spec,
+						   peer_vport->vport);
+		flow = mlx5_add_flow_rules(mlx5_eswitch_get_slow_fdb(esw),
+					   spec, &flow_act, &dest, 1);
+		if (IS_ERR(flow)) {
+			err = PTR_ERR(flow);
+			goto add_vf_flow_err;
 		}
+		flows[peer_vport->index] = flow;
 	}
 
 	if (mlx5_core_ec_sriov_enabled(peer_dev)) {
@@ -1347,7 +1343,8 @@ static void esw_del_fdb_peer_miss_rules(struct mlx5_eswitch *esw,
 		mlx5_del_flow_rules(flows[peer_vport->index]);
 	}
 
-	if (mlx5_core_is_ecpf_esw_manager(peer_dev)) {
+	if (mlx5_core_is_ecpf_esw_manager(peer_dev) &&
+	    mlx5_esw_host_functions_enabled(peer_dev)) {
 		peer_vport = mlx5_eswitch_get_vport(peer_esw, MLX5_VPORT_PF);
 		mlx5_del_flow_rules(flows[peer_vport->index]);
 	}
-- 
2.44.0

[PATCH net 3/5] net/mlx5e: Fix DMA FIFO desync on error CQE SQ recovery

[Tariq Toukan]

From: Gal Pressman <gal@nvidia.com>

In case of a TX error CQE, a recovery flow is triggered,
mlx5e_reset_txqsq_cc_pc() resets dma_fifo_cc to 0 but not dma_fifo_pc,
desyncing the DMA FIFO producer and consumer.

After recovery, the producer pushes new DMA entries at the old
dma_fifo_pc, while the consumer reads from position 0.
This causes us to unmap stale DMA addresses from before the recovery.

The DMA FIFO is a purely software construct with no HW counterpart.
At the point of reset, all WQEs have been flushed so dma_fifo_cc is
already equal to dma_fifo_pc. There is no need to reset either counter,
similar to how skb_fifo pc/cc are untouched.

Remove the 'dma_fifo_cc = 0' reset.

This fixes the following WARNING:
    WARNING: CPU: 0 PID: 0 at drivers/iommu/dma-iommu.c:1240 iommu_dma_unmap_page+0x79/0x90
    Modules linked in: mlx5_vdpa vringh vdpa bonding mlx5_ib mlx5_vfio_pci ipip mlx5_fwctl tunnel4 mlx5_core ib_ipoib geneve ip6_gre ip_gre gre nf_tables ip6_tunnel rdma_ucm ib_uverbs ib_umad vfio_pci vfio_pci_core act_mirred act_skbedit act_vlan vhost_net vhost tap ip6table_mangle ip6table_nat ip6table_filter ip6_tables iptable_mangle cls_matchall nfnetlink_cttimeout act_gact cls_flower sch_ingress vhost_iotlb iptable_raw tunnel6 vfio_iommu_type1 vfio openvswitch nsh rpcsec_gss_krb5 auth_rpcgss oid_registry xt_conntrack xt_MASQUERADE nf_conntrack_netlink nfnetlink iptable_nat nf_nat xt_addrtype br_netfilter overlay zram zsmalloc rpcrdma ib_iser libiscsi scsi_transport_iscsi rdma_cm iw_cm ib_cm ib_core fuse [last unloaded: nf_tables]
    CPU: 0 UID: 0 PID: 0 Comm: swapper/0 Not tainted 6.13.0-rc5_for_upstream_min_debug_2024_12_30_21_33 #1
    Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS rel-1.13.0-0-gf21b5a4aeb02-prebuilt.qemu.org 04/01/2014
    RIP: 0010:iommu_dma_unmap_page+0x79/0x90
    Code: 2b 4d 3b 21 72 26 4d 3b 61 08 73 20 49 89 d8 44 89 f9 5b 4c 89 f2 4c 89 e6 48 89 ef 5d 41 5c 41 5d 41 5e 41 5f e9 c7 ae 9e ff <0f> 0b 5b 5d 41 5c 41 5d 41 5e 41 5f c3 66 2e 0f 1f 84 00 00 00 00
    Call Trace:
     <IRQ>
     ? __warn+0x7d/0x110
     ? iommu_dma_unmap_page+0x79/0x90
     ? report_bug+0x16d/0x180
     ? handle_bug+0x4f/0x90
     ? exc_invalid_op+0x14/0x70
     ? asm_exc_invalid_op+0x16/0x20
     ? iommu_dma_unmap_page+0x79/0x90
     ? iommu_dma_unmap_page+0x2e/0x90
     dma_unmap_page_attrs+0x10d/0x1b0
     mlx5e_tx_wi_dma_unmap+0xbe/0x120 [mlx5_core]
     mlx5e_poll_tx_cq+0x16d/0x690 [mlx5_core]
     mlx5e_napi_poll+0x8b/0xac0 [mlx5_core]
     __napi_poll+0x24/0x190
     net_rx_action+0x32a/0x3b0
     ? mlx5_eq_comp_int+0x7e/0x270 [mlx5_core]
     ? notifier_call_chain+0x35/0xa0
     handle_softirqs+0xc9/0x270
     irq_exit_rcu+0x71/0xd0
     common_interrupt+0x7f/0xa0
     </IRQ>
     <TASK>
     asm_common_interrupt+0x22/0x40

Fixes: db75373c91b0 ("net/mlx5e: Recover Send Queue (SQ) from error state")
Signed-off-by: Gal Pressman <gal@nvidia.com>
Reviewed-by: Dragos Tatulea <dtatulea@nvidia.com>
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
---
 drivers/net/ethernet/mellanox/mlx5/core/en/reporter_tx.c | 1 -
 1 file changed, 1 deletion(-)

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/reporter_tx.c b/drivers/net/ethernet/mellanox/mlx5/core/en/reporter_tx.c
index 60ba840e00fa..afdeb1b3d425 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/reporter_tx.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/reporter_tx.c
@@ -47,7 +47,6 @@ static void mlx5e_reset_txqsq_cc_pc(struct mlx5e_txqsq *sq)
 		  "SQ 0x%x: cc (0x%x) != pc (0x%x)\n",
 		  sq->sqn, sq->cc, sq->pc);
 	sq->cc = 0;
-	sq->dma_fifo_cc = 0;
 	sq->pc = 0;
 }
 
-- 
2.44.0

[PATCH net 4/5] net/mlx5e: RX, Fix XDP multi-buf frag counting for striding RQ

[Tariq Toukan]

From: Dragos Tatulea <dtatulea@nvidia.com>

XDP multi-buf programs can modify the layout of the XDP buffer when the
program calls bpf_xdp_pull_data() or bpf_xdp_adjust_tail(). The
referenced commit in the fixes tag corrected the assumption in the mlx5
driver that the XDP buffer layout doesn't change during a program
execution. However, this fix introduced another issue: the dropped
fragments still need to be counted on the driver side to avoid page
fragment reference counting issues.

The issue was discovered by the drivers/net/xdp.py selftest,
more specifically the test_xdp_native_tx_mb:
- The mlx5 driver allocates a page_pool page and initializes it with
  a frag counter of 64 (pp_ref_count=64) and the internal frag counter
  to 0.
- The test sends one packet with no payload.
- On RX (mlx5e_skb_from_cqe_mpwrq_nonlinear()), mlx5 configures the XDP
  buffer with the packet data starting in the first fragment which is the
  page mentioned above.
- The XDP program runs and calls bpf_xdp_pull_data() which moves the
  header into the linear part of the XDP buffer. As the packet doesn't
  contain more data, the program drops the tail fragment since it no
  longer contains any payload (pp_ref_count=63).
- mlx5 device skips counting this fragment. Internal frag counter
  remains 0.
- mlx5 releases all 64 fragments of the page but page pp_ref_count is
  63 => negative reference counting error.

Resulting splat during the test:

  WARNING: CPU: 0 PID: 188225 at ./include/net/page_pool/helpers.h:297 mlx5e_page_release_fragmented.isra.0+0xbd/0xe0 [mlx5_core]
  Modules linked in: [...]
  CPU: 0 UID: 0 PID: 188225 Comm: ip Not tainted 6.18.0-rc7_for_upstream_min_debug_2025_12_08_11_44 #1 NONE
  Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS rel-1.13.0-0-gf21b5a4aeb02-prebuilt.qemu.org 04/01/2014
  RIP: 0010:mlx5e_page_release_fragmented.isra.0+0xbd/0xe0 [mlx5_core]
  [...]
  Call Trace:
   <TASK>
   mlx5e_free_rx_mpwqe+0x20a/0x250 [mlx5_core]
   mlx5e_dealloc_rx_mpwqe+0x37/0xb0 [mlx5_core]
   mlx5e_free_rx_descs+0x11a/0x170 [mlx5_core]
   mlx5e_close_rq+0x78/0xa0 [mlx5_core]
   mlx5e_close_queues+0x46/0x2a0 [mlx5_core]
   mlx5e_close_channel+0x24/0x90 [mlx5_core]
   mlx5e_close_channels+0x5d/0xf0 [mlx5_core]
   mlx5e_safe_switch_params+0x2ec/0x380 [mlx5_core]
   mlx5e_change_mtu+0x11d/0x490 [mlx5_core]
   mlx5e_change_nic_mtu+0x19/0x30 [mlx5_core]
   netif_set_mtu_ext+0xfc/0x240
   do_setlink.isra.0+0x226/0x1100
   rtnl_newlink+0x7a9/0xba0
   rtnetlink_rcv_msg+0x220/0x3c0
   netlink_rcv_skb+0x4b/0xf0
   netlink_unicast+0x255/0x380
   netlink_sendmsg+0x1f3/0x420
   __sock_sendmsg+0x38/0x60
   ____sys_sendmsg+0x1e8/0x240
   ___sys_sendmsg+0x7c/0xb0
   [...]
   __sys_sendmsg+0x5f/0xb0
   do_syscall_64+0x55/0xc70

The problem applies for XDP_PASS as well which is handled in a different
code path in the driver.

This patch fixes the issue by doing page frag counting on all the
original XDP buffer fragments for all relevant XDP actions (XDP_TX ,
XDP_REDIRECT and XDP_PASS). This is basically reverting to the original
counting before the commit in the fixes tag.

As frag_page is still pointing to the original tail, the nr_frags
parameter to xdp_update_skb_frags_info() needs to be calculated
in a different way to reflect the new nr_frags.

Fixes: 87bcef158ac1 ("net/mlx5e: RX, Fix generating skb from non-linear xdp_buff for striding RQ")
Signed-off-by: Dragos Tatulea <dtatulea@nvidia.com>
Cc: Amery Hung <ameryhung@gmail.com>
Reviewed-by: Nimrod Oren <noren@nvidia.com>
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
---
 drivers/net/ethernet/mellanox/mlx5/core/en_rx.c | 12 +++++-------
 1 file changed, 5 insertions(+), 7 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c b/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c
index efcfcddab376..40e53a612989 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c
@@ -1957,14 +1957,13 @@ mlx5e_skb_from_cqe_mpwrq_nonlinear(struct mlx5e_rq *rq, struct mlx5e_mpw_info *w
 
 	if (prog) {
 		u8 nr_frags_free, old_nr_frags = sinfo->nr_frags;
+		u8 new_nr_frags;
 		u32 len;
 
 		if (mlx5e_xdp_handle(rq, prog, mxbuf)) {
 			if (__test_and_clear_bit(MLX5E_RQ_FLAG_XDP_XMIT, rq->flags)) {
 				struct mlx5e_frag_page *pfp;
 
-				frag_page -= old_nr_frags - sinfo->nr_frags;
-
 				for (pfp = head_page; pfp < frag_page; pfp++)
 					pfp->frags++;
 
@@ -1975,13 +1974,12 @@ mlx5e_skb_from_cqe_mpwrq_nonlinear(struct mlx5e_rq *rq, struct mlx5e_mpw_info *w
 			return NULL; /* page/packet was consumed by XDP */
 		}
 
-		nr_frags_free = old_nr_frags - sinfo->nr_frags;
-		if (unlikely(nr_frags_free)) {
-			frag_page -= nr_frags_free;
+		new_nr_frags = sinfo->nr_frags;
+		nr_frags_free = old_nr_frags - new_nr_frags;
+		if (unlikely(nr_frags_free))
 			truesize -= (nr_frags_free - 1) * PAGE_SIZE +
 				ALIGN(pg_consumed_bytes,
 				      BIT(rq->mpwqe.log_stride_sz));
-		}
 
 		len = mxbuf->xdp.data_end - mxbuf->xdp.data;
 
@@ -2003,7 +2001,7 @@ mlx5e_skb_from_cqe_mpwrq_nonlinear(struct mlx5e_rq *rq, struct mlx5e_mpw_info *w
 			struct mlx5e_frag_page *pagep;
 
 			/* sinfo->nr_frags is reset by build_skb, calculate again. */
-			xdp_update_skb_frags_info(skb, frag_page - head_page,
+			xdp_update_skb_frags_info(skb, new_nr_frags,
 						  sinfo->xdp_frags_size,
 						  truesize,
 						  xdp_buff_get_skb_flags(&mxbuf->xdp));
-- 
2.44.0

[PATCH net 5/5] net/mlx5e: RX, Fix XDP multi-buf frag counting for legacy RQ

[Tariq Toukan]

From: Dragos Tatulea <dtatulea@nvidia.com>

XDP multi-buf programs can modify the layout of the XDP buffer when the
program calls bpf_xdp_pull_data() or bpf_xdp_adjust_tail(). The
referenced commit in the fixes tag corrected the assumption in the mlx5
driver that the XDP buffer layout doesn't change during a program
execution. However, this fix introduced another issue: the dropped
fragments still need to be counted on the driver side to avoid page
fragment reference counting issues.

Such issue can be observed with the
test_xdp_native_adjst_tail_shrnk_data selftest when using a payload of
3600 and shrinking by 256 bytes (an upcoming selftest patch): the last
fragment gets released by the XDP code but doesn't get tracked by the
driver. This results in a negative pp_ref_count during page release and
the following splat:

  WARNING: include/net/page_pool/helpers.h:297 at mlx5e_page_release_fragmented.isra.0+0x4a/0x50 [mlx5_core], CPU#12: ip/3137
  Modules linked in: [...]
  CPU: 12 UID: 0 PID: 3137 Comm: ip Not tainted 6.19.0-rc3+ #12 NONE
  Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS rel-1.16.3-0-ga6ed6b701f0a-prebuilt.qemu.org 04/01/2014
  RIP: 0010:mlx5e_page_release_fragmented.isra.0+0x4a/0x50 [mlx5_core]
  [...]
  Call Trace:
   <TASK>
   mlx5e_dealloc_rx_wqe+0xcb/0x1a0 [mlx5_core]
   mlx5e_free_rx_descs+0x7f/0x110 [mlx5_core]
   mlx5e_close_rq+0x50/0x60 [mlx5_core]
   mlx5e_close_queues+0x36/0x2c0 [mlx5_core]
   mlx5e_close_channel+0x1c/0x50 [mlx5_core]
   mlx5e_close_channels+0x45/0x80 [mlx5_core]
   mlx5e_safe_switch_params+0x1a5/0x230 [mlx5_core]
   mlx5e_change_mtu+0xf3/0x2f0 [mlx5_core]
   netif_set_mtu_ext+0xf1/0x230
   do_setlink.isra.0+0x219/0x1180
   rtnl_newlink+0x79f/0xb60
   rtnetlink_rcv_msg+0x213/0x3a0
   netlink_rcv_skb+0x48/0xf0
   netlink_unicast+0x24a/0x350
   netlink_sendmsg+0x1ee/0x410
   __sock_sendmsg+0x38/0x60
   ____sys_sendmsg+0x232/0x280
   ___sys_sendmsg+0x78/0xb0
   __sys_sendmsg+0x5f/0xb0
   [...]
   do_syscall_64+0x57/0xc50

This patch fixes the issue by doing page frag counting on all the
original XDP buffer fragments for all relevant XDP actions (XDP_TX ,
XDP_REDIRECT and XDP_PASS). This is basically reverting to the original
counting before the commit in the fixes tag.

As frag_page is still pointing to the original tail, the nr_frags
parameter to xdp_update_skb_frags_info() needs to be calculated
in a different way to reflect the new nr_frags.

Fixes: afd5ba577c10 ("net/mlx5e: RX, Fix generating skb from non-linear xdp_buff for legacy RQ")
Signed-off-by: Dragos Tatulea <dtatulea@nvidia.com>
Cc: Amery Hung <ameryhung@gmail.com>
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
---
 drivers/net/ethernet/mellanox/mlx5/core/en_rx.c | 11 ++++-------
 1 file changed, 4 insertions(+), 7 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c b/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c
index 40e53a612989..268e20884757 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c
@@ -1589,6 +1589,7 @@ mlx5e_skb_from_cqe_nonlinear(struct mlx5e_rq *rq, struct mlx5e_wqe_frag_info *wi
 	struct skb_shared_info *sinfo;
 	u32 frag_consumed_bytes;
 	struct bpf_prog *prog;
+	u8 nr_frags_free = 0;
 	struct sk_buff *skb;
 	dma_addr_t addr;
 	u32 truesize;
@@ -1631,15 +1632,13 @@ mlx5e_skb_from_cqe_nonlinear(struct mlx5e_rq *rq, struct mlx5e_wqe_frag_info *wi
 
 	prog = rcu_dereference(rq->xdp_prog);
 	if (prog) {
-		u8 nr_frags_free, old_nr_frags = sinfo->nr_frags;
+		u8 old_nr_frags = sinfo->nr_frags;
 
 		if (mlx5e_xdp_handle(rq, prog, mxbuf)) {
 			if (__test_and_clear_bit(MLX5E_RQ_FLAG_XDP_XMIT,
 						 rq->flags)) {
 				struct mlx5e_wqe_frag_info *pwi;
 
-				wi -= old_nr_frags - sinfo->nr_frags;
-
 				for (pwi = head_wi; pwi < wi; pwi++)
 					pwi->frag_page->frags++;
 			}
@@ -1647,10 +1646,8 @@ mlx5e_skb_from_cqe_nonlinear(struct mlx5e_rq *rq, struct mlx5e_wqe_frag_info *wi
 		}
 
 		nr_frags_free = old_nr_frags - sinfo->nr_frags;
-		if (unlikely(nr_frags_free)) {
-			wi -= nr_frags_free;
+		if (unlikely(nr_frags_free))
 			truesize -= nr_frags_free * frag_info->frag_stride;
-		}
 	}
 
 	skb = mlx5e_build_linear_skb(
@@ -1666,7 +1663,7 @@ mlx5e_skb_from_cqe_nonlinear(struct mlx5e_rq *rq, struct mlx5e_wqe_frag_info *wi
 
 	if (xdp_buff_has_frags(&mxbuf->xdp)) {
 		/* sinfo->nr_frags is reset by build_skb, calculate again. */
-		xdp_update_skb_frags_info(skb, wi - head_wi - 1,
+		xdp_update_skb_frags_info(skb, wi - head_wi - nr_frags_free - 1,
 					  sinfo->xdp_frags_size, truesize,
 					  xdp_buff_get_skb_flags(&mxbuf->xdp));
 
-- 
2.44.0

Re: [PATCH net 5/5] net/mlx5e: RX, Fix XDP multi-buf frag counting for legacy RQ

[Amery Hung]

On Thu, Mar 5, 2026 at 6:27 AM Tariq Toukan <tariqt@nvidia.com> wrote:
>
> From: Dragos Tatulea <dtatulea@nvidia.com>
>
> XDP multi-buf programs can modify the layout of the XDP buffer when the
> program calls bpf_xdp_pull_data() or bpf_xdp_adjust_tail(). The
> referenced commit in the fixes tag corrected the assumption in the mlx5
> driver that the XDP buffer layout doesn't change during a program
> execution. However, this fix introduced another issue: the dropped
> fragments still need to be counted on the driver side to avoid page
> fragment reference counting issues.
>
> Such issue can be observed with the
> test_xdp_native_adjst_tail_shrnk_data selftest when using a payload of
> 3600 and shrinking by 256 bytes (an upcoming selftest patch): the last
> fragment gets released by the XDP code but doesn't get tracked by the
> driver. This results in a negative pp_ref_count during page release and
> the following splat:
>
>   WARNING: include/net/page_pool/helpers.h:297 at mlx5e_page_release_fragmented.isra.0+0x4a/0x50 [mlx5_core], CPU#12: ip/3137
>   Modules linked in: [...]
>   CPU: 12 UID: 0 PID: 3137 Comm: ip Not tainted 6.19.0-rc3+ #12 NONE
>   Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS rel-1.16.3-0-ga6ed6b701f0a-prebuilt.qemu.org 04/01/2014
>   RIP: 0010:mlx5e_page_release_fragmented.isra.0+0x4a/0x50 [mlx5_core]
>   [...]
>   Call Trace:
>    <TASK>
>    mlx5e_dealloc_rx_wqe+0xcb/0x1a0 [mlx5_core]
>    mlx5e_free_rx_descs+0x7f/0x110 [mlx5_core]
>    mlx5e_close_rq+0x50/0x60 [mlx5_core]
>    mlx5e_close_queues+0x36/0x2c0 [mlx5_core]
>    mlx5e_close_channel+0x1c/0x50 [mlx5_core]
>    mlx5e_close_channels+0x45/0x80 [mlx5_core]
>    mlx5e_safe_switch_params+0x1a5/0x230 [mlx5_core]
>    mlx5e_change_mtu+0xf3/0x2f0 [mlx5_core]
>    netif_set_mtu_ext+0xf1/0x230
>    do_setlink.isra.0+0x219/0x1180
>    rtnl_newlink+0x79f/0xb60
>    rtnetlink_rcv_msg+0x213/0x3a0
>    netlink_rcv_skb+0x48/0xf0
>    netlink_unicast+0x24a/0x350
>    netlink_sendmsg+0x1ee/0x410
>    __sock_sendmsg+0x38/0x60
>    ____sys_sendmsg+0x232/0x280
>    ___sys_sendmsg+0x78/0xb0
>    __sys_sendmsg+0x5f/0xb0
>    [...]
>    do_syscall_64+0x57/0xc50
>
> This patch fixes the issue by doing page frag counting on all the
> original XDP buffer fragments for all relevant XDP actions (XDP_TX ,
> XDP_REDIRECT and XDP_PASS). This is basically reverting to the original
> counting before the commit in the fixes tag.
>
> As frag_page is still pointing to the original tail, the nr_frags
> parameter to xdp_update_skb_frags_info() needs to be calculated
> in a different way to reflect the new nr_frags.

I see the error I made. Thanks for fixing it.

Reviewed-by: Amery Hung <ameryhung@gmail.com>

>
> Fixes: afd5ba577c10 ("net/mlx5e: RX, Fix generating skb from non-linear xdp_buff for legacy RQ")
> Signed-off-by: Dragos Tatulea <dtatulea@nvidia.com>
> Cc: Amery Hung <ameryhung@gmail.com>
> Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
> ---
>  drivers/net/ethernet/mellanox/mlx5/core/en_rx.c | 11 ++++-------
>  1 file changed, 4 insertions(+), 7 deletions(-)
>
> diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c b/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c
> index 40e53a612989..268e20884757 100644
> --- a/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c
> +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c
> @@ -1589,6 +1589,7 @@ mlx5e_skb_from_cqe_nonlinear(struct mlx5e_rq *rq, struct mlx5e_wqe_frag_info *wi
>         struct skb_shared_info *sinfo;
>         u32 frag_consumed_bytes;
>         struct bpf_prog *prog;
> +       u8 nr_frags_free = 0;
>         struct sk_buff *skb;
>         dma_addr_t addr;
>         u32 truesize;
> @@ -1631,15 +1632,13 @@ mlx5e_skb_from_cqe_nonlinear(struct mlx5e_rq *rq, struct mlx5e_wqe_frag_info *wi
>
>         prog = rcu_dereference(rq->xdp_prog);
>         if (prog) {
> -               u8 nr_frags_free, old_nr_frags = sinfo->nr_frags;
> +               u8 old_nr_frags = sinfo->nr_frags;
>
>                 if (mlx5e_xdp_handle(rq, prog, mxbuf)) {
>                         if (__test_and_clear_bit(MLX5E_RQ_FLAG_XDP_XMIT,
>                                                  rq->flags)) {
>                                 struct mlx5e_wqe_frag_info *pwi;
>
> -                               wi -= old_nr_frags - sinfo->nr_frags;
> -
>                                 for (pwi = head_wi; pwi < wi; pwi++)
>                                         pwi->frag_page->frags++;
>                         }
> @@ -1647,10 +1646,8 @@ mlx5e_skb_from_cqe_nonlinear(struct mlx5e_rq *rq, struct mlx5e_wqe_frag_info *wi
>                 }
>
>                 nr_frags_free = old_nr_frags - sinfo->nr_frags;
> -               if (unlikely(nr_frags_free)) {
> -                       wi -= nr_frags_free;
> +               if (unlikely(nr_frags_free))
>                         truesize -= nr_frags_free * frag_info->frag_stride;
> -               }
>         }
>
>         skb = mlx5e_build_linear_skb(
> @@ -1666,7 +1663,7 @@ mlx5e_skb_from_cqe_nonlinear(struct mlx5e_rq *rq, struct mlx5e_wqe_frag_info *wi
>
>         if (xdp_buff_has_frags(&mxbuf->xdp)) {
>                 /* sinfo->nr_frags is reset by build_skb, calculate again. */
> -               xdp_update_skb_frags_info(skb, wi - head_wi - 1,
> +               xdp_update_skb_frags_info(skb, wi - head_wi - nr_frags_free - 1,
>                                           sinfo->xdp_frags_size, truesize,
>                                           xdp_buff_get_skb_flags(&mxbuf->xdp));
>
> --
> 2.44.0
>

Re: [PATCH net 0/5] mlx5 misc fixes 2026-03-05

[patchwork-bot+netdevbpf]

Hello:

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

On Thu, 5 Mar 2026 16:26:29 +0200 you wrote:
> Hi,
> 
> This patchset provides misc bug fixes from the team to the mlx5
> core and Eth drivers.
> 
> Thanks,
> Tariq.
> 
> [...]

Here is the summary with links:
  - [net,1/5] net/mlx5: Fix crash when moving to switchdev mode
    https://git.kernel.org/netdev/net/c/24b2795f9683
  - [net,2/5] net/mlx5: Fix peer miss rules host disabled checks
    https://git.kernel.org/netdev/net/c/76324e4041c0
  - [net,3/5] net/mlx5e: Fix DMA FIFO desync on error CQE SQ recovery
    https://git.kernel.org/netdev/net/c/1633111d6905
  - [net,4/5] net/mlx5e: RX, Fix XDP multi-buf frag counting for striding RQ
    https://git.kernel.org/netdev/net/c/db25c42c2e1f
  - [net,5/5] net/mlx5e: RX, Fix XDP multi-buf frag counting for legacy RQ
    https://git.kernel.org/netdev/net/c/a6413e6f6c9d

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

Re: [PATCH net 4/5] net/mlx5e: RX, Fix XDP multi-buf frag counting for striding RQ: manual merge

[Matthieu Baerts]

Hi Tariq, Dragos,

+cc: linux-next

On 05/03/2026 15:26, Tariq Toukan wrote:
> From: Dragos Tatulea <dtatulea@nvidia.com>
> 
> XDP multi-buf programs can modify the layout of the XDP buffer when the
> program calls bpf_xdp_pull_data() or bpf_xdp_adjust_tail(). The
> referenced commit in the fixes tag corrected the assumption in the mlx5
> driver that the XDP buffer layout doesn't change during a program
> execution. However, this fix introduced another issue: the dropped
> fragments still need to be counted on the driver side to avoid page
> fragment reference counting issues.

FYI, we got a small conflict when merging 'net' in 'net-next' in the
MPTCP tree due to this patch applied in 'net':

  db25c42c2e1f ("net/mlx5e: RX, Fix XDP multi-buf frag counting for striding RQ")

and this one from 'net-next':

  dff1c3164a69 ("net/mlx5e: SHAMPO, Always calculate page size")

----- Generic Message -----
The best is to avoid conflicts between 'net' and 'net-next' trees but if
they cannot be avoided when preparing patches, a note about how to fix
them is much appreciated.

The conflict has been resolved on our side [1] and the resolution we
suggest is attached to this email. Please report any issues linked to
this conflict resolution as it might be used by others. If you worked on
the mentioned patches, don't hesitate to ACK this conflict resolution.
---------------------------

Rerere cache is available in [2].

[1] https://github.com/multipath-tcp/mptcp_net-next/commit/9cbb5f8a4a18
[2] https://github.com/multipath-tcp/mptcp-upstream-rr-cache/commit/0bbafdd

(...)

> diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c b/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c
> index efcfcddab376..40e53a612989 100644
> --- a/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c
> +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c

(...)

> @@ -1975,13 +1974,12 @@ mlx5e_skb_from_cqe_mpwrq_nonlinear(struct mlx5e_rq *rq, struct mlx5e_mpw_info *w
>  			return NULL; /* page/packet was consumed by XDP */
>  		}
>  
> -		nr_frags_free = old_nr_frags - sinfo->nr_frags;
> -		if (unlikely(nr_frags_free)) {
> -			frag_page -= nr_frags_free;
> +		new_nr_frags = sinfo->nr_frags;
> +		nr_frags_free = old_nr_frags - new_nr_frags;
> +		if (unlikely(nr_frags_free))
>  			truesize -= (nr_frags_free - 1) * PAGE_SIZE +

The conflict is in the context: this line above has been modified on
'net-next' (s/PAGE_SIZE/page_size/), while the ones above it have been
modified on 'net'.

Cheers,
Matt
-- 
Sponsored by the NGI0 Core fund.

---
diff --cc drivers/net/ethernet/mellanox/mlx5/core/en_rx.c
index 8fb57a4f36dd,268e20884757..f5c0e2a0ada9
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c
@@@ -1977,13 -1971,12 +1973,12 @@@ mlx5e_skb_from_cqe_mpwrq_nonlinear(stru
  			return NULL; /* page/packet was consumed by XDP */
  		}
  
- 		nr_frags_free = old_nr_frags - sinfo->nr_frags;
- 		if (unlikely(nr_frags_free)) {
- 			frag_page -= nr_frags_free;
+ 		new_nr_frags = sinfo->nr_frags;
+ 		nr_frags_free = old_nr_frags - new_nr_frags;
+ 		if (unlikely(nr_frags_free))
 -			truesize -= (nr_frags_free - 1) * PAGE_SIZE +
 +			truesize -= (nr_frags_free - 1) * page_size +
  				ALIGN(pg_consumed_bytes,
  				      BIT(rq->mpwqe.log_stride_sz));
- 		}
  
  		len = mxbuf->xdp.data_end - mxbuf->xdp.data;
  

Re: [PATCH net 4/5] net/mlx5e: RX, Fix XDP multi-buf frag counting for striding RQ: manual merge

[Dragos Tatulea]

Hi Matthieu,

On 09.03.26 13:52, Matthieu Baerts wrote:
> Hi Tariq, Dragos,
> 
> +cc: linux-next
> 
> On 05/03/2026 15:26, Tariq Toukan wrote:
>> From: Dragos Tatulea <dtatulea@nvidia.com>
>>
>> XDP multi-buf programs can modify the layout of the XDP buffer when the
>> program calls bpf_xdp_pull_data() or bpf_xdp_adjust_tail(). The
>> referenced commit in the fixes tag corrected the assumption in the mlx5
>> driver that the XDP buffer layout doesn't change during a program
>> execution. However, this fix introduced another issue: the dropped
>> fragments still need to be counted on the driver side to avoid page
>> fragment reference counting issues.
> 
> FYI, we got a small conflict when merging 'net' in 'net-next' in the
> MPTCP tree due to this patch applied in 'net':
> 
>   db25c42c2e1f ("net/mlx5e: RX, Fix XDP multi-buf frag counting for striding RQ")
> 
> and this one from 'net-next':
> 
>   dff1c3164a69 ("net/mlx5e: SHAMPO, Always calculate page size")
> 
> ----- Generic Message -----
> The best is to avoid conflicts between 'net' and 'net-next' trees but if
> they cannot be avoided when preparing patches, a note about how to fix
> them is much appreciated.
> 
Apologies for this. Will take note next time.

> The conflict has been resolved on our side [1] and the resolution we
> suggest is attached to this email. Please report any issues linked to
> this conflict resolution as it might be used by others. If you worked on
> the mentioned patches, don't hesitate to ACK this conflict resolution.
> ---------------------------
> 
> Rerere cache is available in [2].
> 
> [1] https://github.com/multipath-tcp/mptcp_net-next/commit/9cbb5f8a4a18
Conflict resolution from [1] seems good.

Acked-by: Dragos Tatulea <dtatulea@nvidia.com>

Thanks,
Dragos