[PATCH 6.1.y] net/mlx5e: Fix netif state handling

[Artem Dinaburg <artem@trailofbits.com>]

From: Shay Drory <shayd@nvidia.com>

[ Upstream commit 3d5918477f94e4c2f064567875c475468e264644 ]

mlx5e_suspend cleans resources only if netif_device_present() returns
true. However, mlx5e_resume changes the state of netif, via
mlx5e_nic_enable, only if reg_state == NETREG_REGISTERED.
In the below case, the above leads to NULL-ptr Oops[1] and memory
leaks:

mlx5e_probe
 _mlx5e_resume
  mlx5e_attach_netdev
   mlx5e_nic_enable  <-- netdev not reg, not calling netif_device_attach()
  register_netdev <-- failed for some reason.
ERROR_FLOW:
 _mlx5e_suspend <-- netif_device_present return false, resources aren't freed :(

Hence, clean resources in this case as well.

[1]
BUG: kernel NULL pointer dereference, address: 0000000000000000
PGD 0 P4D 0
Oops: 0010 [#1] SMP
CPU: 2 PID: 9345 Comm: test-ovs-ct-gen Not tainted 6.5.0_for_upstream_min_debug_2023_09_05_16_01 #1
Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS rel-1.13.0-0-gf21b5a4aeb02-prebuilt.qemu.org 04/01/2014
RIP: 0010:0x0
Code: Unable to access opcode bytes at0xffffffffffffffd6.
RSP: 0018:ffff888178aaf758 EFLAGS: 00010246
Call Trace:
 <TASK>
 ? __die+0x20/0x60
 ? page_fault_oops+0x14c/0x3c0
 ? exc_page_fault+0x75/0x140
 ? asm_exc_page_fault+0x22/0x30
 notifier_call_chain+0x35/0xb0
 blocking_notifier_call_chain+0x3d/0x60
 mlx5_blocking_notifier_call_chain+0x22/0x30 [mlx5_core]
 mlx5_core_uplink_netdev_event_replay+0x3e/0x60 [mlx5_core]
 mlx5_mdev_netdev_track+0x53/0x60 [mlx5_ib]
 mlx5_ib_roce_init+0xc3/0x340 [mlx5_ib]
 __mlx5_ib_add+0x34/0xd0 [mlx5_ib]
 mlx5r_probe+0xe1/0x210 [mlx5_ib]
 ? auxiliary_match_id+0x6a/0x90
 auxiliary_bus_probe+0x38/0x80
 ? driver_sysfs_add+0x51/0x80
 really_probe+0xc9/0x3e0
 ? driver_probe_device+0x90/0x90
 __driver_probe_device+0x80/0x160
 driver_probe_device+0x1e/0x90
 __device_attach_driver+0x7d/0x100
 bus_for_each_drv+0x80/0xd0
 __device_attach+0xbc/0x1f0
 bus_probe_device+0x86/0xa0
 device_add+0x637/0x840
 __auxiliary_device_add+0x3b/0xa0
 add_adev+0xc9/0x140 [mlx5_core]
 mlx5_rescan_drivers_locked+0x22a/0x310 [mlx5_core]
 mlx5_register_device+0x53/0xa0 [mlx5_core]
 mlx5_init_one_devl_locked+0x5c4/0x9c0 [mlx5_core]
 mlx5_init_one+0x3b/0x60 [mlx5_core]
 probe_one+0x44c/0x730 [mlx5_core]
 local_pci_probe+0x3e/0x90
 pci_device_probe+0xbf/0x210
 ? kernfs_create_link+0x5d/0xa0
 ? sysfs_do_create_link_sd+0x60/0xc0
 really_probe+0xc9/0x3e0
 ? driver_probe_device+0x90/0x90
 __driver_probe_device+0x80/0x160
 driver_probe_device+0x1e/0x90
 __device_attach_driver+0x7d/0x100
 bus_for_each_drv+0x80/0xd0
 __device_attach+0xbc/0x1f0
 pci_bus_add_device+0x54/0x80
 pci_iov_add_virtfn+0x2e6/0x320
 sriov_enable+0x208/0x420
 mlx5_core_sriov_configure+0x9e/0x200 [mlx5_core]
 sriov_numvfs_store+0xae/0x1a0
 kernfs_fop_write_iter+0x10c/0x1a0
 vfs_write+0x291/0x3c0
 ksys_write+0x5f/0xe0
 do_syscall_64+0x3d/0x90
 entry_SYSCALL_64_after_hwframe+0x46/0xb0
 CR2: 0000000000000000
 ---[ end trace 0000000000000000  ]---

[ Backport to 6.1.y: added an internal suspend helper so probe unwind
  cleans resources without changing the registered-device suspend path. ]

Fixes: 2c3b5beec46a ("net/mlx5e: More generic netdev management API")
Signed-off-by: Shay Drory <shayd@nvidia.com>
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Link: https://lore.kernel.org/r/20240509112951.590184-2-tariqt@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Assisted-by: LLM
Signed-off-by: Artem Dinaburg <artem@trailofbits.com>
---
Hi Greg, Sasha, and net maintainers,

I am working through the small CVE backports still missing from 6.1.y.
This one addresses CVE-2024-38608. It releases probe resources even when
register_netdev() failed before device-present state.

The corresponding 6.6.y backport is already in the 6.6.y stable queue.
The fix is already present in 6.12.y, 6.18.y, and 7.2.y, but not
in 6.1.y.
The target-specific adjustment is recorded in the bracketed note above.

Could you please queue it for 6.1.y?

CVE: CVE-2024-38608
Upstream: 3d5918477f94e4c2f064567875c475468e264644

AI assistance: An LLM helped identify, adapt, and validate this backport; I
reviewed the resulting code and validation evidence.

Thanks,
Artem Dinaburg

 drivers/net/ethernet/mellanox/mlx5/core/en_main.c | 13 +++++++++----
 1 file changed, 9 insertions(+), 4 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
index 2d0b57583ea35c..d26c36479ee915 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
@@ -5896,13 +5896,14 @@ static int mlx5e_resume(struct auxiliary_device *adev)
 	return 0;
 }
 
-static int mlx5e_suspend(struct auxiliary_device *adev, pm_message_t state)
+static int mlx5e_suspend_common(struct auxiliary_device *adev,
+				bool pre_netdev_reg)
 {
 	struct mlx5e_priv *priv = auxiliary_get_drvdata(adev);
 	struct net_device *netdev = priv->netdev;
 	struct mlx5_core_dev *mdev = priv->mdev;
 
-	if (!netif_device_present(netdev)) {
+	if (!pre_netdev_reg && !netif_device_present(netdev)) {
 		if (test_bit(MLX5E_STATE_DESTROYING, &priv->state))
 			mlx5e_destroy_mdev_resources(mdev);
 		return -ENODEV;
@@ -5913,6 +5914,11 @@ static int mlx5e_suspend(struct auxiliary_device *adev, pm_message_t state)
 	return 0;
 }
 
+static int mlx5e_suspend(struct auxiliary_device *adev, pm_message_t state)
+{
+	return mlx5e_suspend_common(adev, false);
+}
+
 static int mlx5e_probe(struct auxiliary_device *adev,
 		       const struct auxiliary_device_id *id)
 {
@@ -5920,7 +5926,6 @@ static int mlx5e_probe(struct auxiliary_device *adev,
 	const struct mlx5e_profile *profile = &mlx5e_nic_profile;
 	struct mlx5_core_dev *mdev = edev->mdev;
 	struct net_device *netdev;
-	pm_message_t state = {};
 	struct mlx5e_priv *priv;
 	int err;
 
@@ -5969,7 +5974,7 @@ static int mlx5e_probe(struct auxiliary_device *adev,
 	return 0;
 
 err_resume:
-	mlx5e_suspend(adev, state);
+	mlx5e_suspend_common(adev, true);
 err_profile_cleanup:
 	profile->cleanup(priv);
 err_devlink_cleanup:
-- 
2.39.5