[PATCH net v5] bnxt_en: Take RTNL before netif_close() when aborting FW reset

[Bryan Cochrane <bcochrane@google.com>]

netif_close() must be called with RTNL held in addition to the netdev
instance lock: __dev_close_many() asserts it, and the NETDEV_GOING_DOWN
and NETDEV_DOWN notifier chains it runs rely on it for serialization.

Commit 004b5008016a ("eth: bnxt: remove most dependencies on RTNL")
replaced rtnl_lock() with netdev_lock() in bnxt_fw_reset_task() and
bnxt_fw_reset(). That is sufficient for the paths that only tear the
NIC down internally with __bnxt_close_nic(), but the abort paths close
the netdev via bnxt_fw_reset_abort() -> netif_close(), and the error
path in bnxt_fw_reset() calls netif_close() directly. Those now run
without RTNL and trigger a WARN_ONCE() cascade when a firmware reset is
aborted:

  RTNL: assertion failed at net/core/dev.c (1740)
  WARNING: net/core/dev.c:1740 at __dev_close_many+0x1be/0x1f0
  Workqueue: bnxt_pf_wq bnxt_fw_reset_task [bnxt_en]
  Call Trace:
   netif_close_many+0xb5/0x190
   netif_close+0xa2/0xd0
   bnxt_fw_reset_task.cold+0x8d/0x213 [bnxt_en]
   process_one_work+0x19d/0x3a0
   worker_thread+0x1af/0x320
   kthread+0xe3/0x120

followed by the same assertion from call_netdevice_notifiers_info(),
fib_rules_event(), inetdev_event(), rtnl_fill_ifinfo(),
devlink_compat_phys_port_name_get() and ip_mc_down().

This was observed on a machine in real use with a BCM57504 (4 PFs,
FW 236.1.153.0) running v7.1.6: the firmware crashed, announced a
"Fatal firmware reset event" and then failed to come back within the
advertised max wait, so bnxt_fw_reset_task() aborted the reset on all
ports:

  bnxt_en 0000:c1:00.1 eno16905np1: Firmware reset aborted
  bnxt_en 0000:c1:00.1 eno16905np1: fw_health_status 0x448100

Take RTNL (before the netdev instance lock, matching the OPENING state
which already uses rtnl_trylock()) in the fw_reset_abort path of
bnxt_fw_reset_task() and route the POLL_VF abort through that same
path.

In bnxt_fw_reset(), only a PF can reach the bnxt_get_registered_vfs()
error path that calls netif_close(), so take RTNL up front via
bnxt_rtnl_lock_sp() when BNXT_PF(bp) is true and keep bnxt_lock_sp() on
VFs. That avoids dropping the instance lock mid-reset (which would
expose a transient BNXT_STATE_IN_FW_RESET and let a concurrent devlink
health recover set BNXT_STATE_IN_SP_TASK), avoids racing with
bnxt_sriov_configure() updating active_vfs outside RTNL, and keeps VFs
from contending on RTNL during recovery while PFs already run on the
single-threaded bnxt_pf_wq.

The bug was found by inspecting the kernel log of the affected machine
and the driver source; the fix was prepared with the help of an AI
coding assistant (Gemini based) and reviewed by the submitter. It is
build tested (allmodconfig with W=1, plus the same change on top of
v7.1.6 built as a module for the affected BCM57504 host). The aborted
reset path itself only runs when the firmware fails to return from a
reset, which could not be reproduced on demand.

Fixes: 004b5008016a ("eth: bnxt: remove most dependencies on RTNL")
Cc: stable@vger.kernel.org
Signed-off-by: Bryan Cochrane <bcochrane@google.com>
---
v5:
 - In bnxt_fw_reset(), take bnxt_rtnl_lock_sp() up front when BNXT_PF(bp)
   is true (keeping bnxt_lock_sp() on VFs) instead of dropping and
   reacquiring locks on the bnxt_get_registered_vfs() < 0 path, and drop
   the bnxt_in_sp_task() change from v4. Dropping netdev_lock mid-reset
   exposed a transient BNXT_STATE_IN_FW_RESET, while guarding
   clear_bit(BNXT_STATE_IN_SP_TASK) in bnxt_[rtnl_]lock_sp() risked a
   deadlock with devl_lock during devlink health recover (reported by the
   netdev Sashiko review). Using BNXT_PF(bp) rather than active_vfs avoids
   racing with bnxt_sriov_configure() while still keeping VFs off RTNL.
v4: https://lore.kernel.org/netdev/20261108222942.1233295-1-bcochrane@google.com/
 - Return to the synchronous close in bnxt_fw_reset() from v2, using
   netdev_unlock() + bnxt_rtnl_lock_sp() as suggested by Michael Chan,
   and guard BNXT_STATE_IN_SP_TASK in bnxt_[rtnl_]lock_sp() and
   bnxt_[rtnl_]unlock_sp() with current_work() == &bp->sp_task. Deferring
   the close to bnxt_fw_reset_task() in v3 still hung when entered via
   bnxt_fw_recover(), raced with bnxt_hwrm_if_change() clearing
   fw_reset_state, and marked the reporter SEVERITY_FATAL (reported by
   the netdev Sashiko review).
 - Omitted Michael's Reviewed-by and Stanislav's Acked-by from v2 due to
   the new bnxt_in_sp_task() check.
v3: https://lore.kernel.org/netdev/20261104045534.79783-1-bcochrane@google.com/
 - Do not drop and retake the locks in bnxt_fw_reset(). Route the
   bnxt_get_registered_vfs() failure through the bnxt_fw_reset_task()
   abort path instead; this also restarts the ULP after that abort. The
   v2 unlock/relock window let a concurrent devlink health recover set
   BNXT_STATE_IN_SP_TASK via bnxt_unlock_sp() and hang the close
   (reported by the netdev Sashiko review).
v2: https://lore.kernel.org/netdev/20261103042210.2984373-1-bcochrane@google.com/
 - Keep bnxt_lock_sp() in bnxt_fw_reset() and take RTNL only on the
   bnxt_get_registered_vfs() error path, to avoid RTNL contention when
   many VFs recover at the same time (Michael Chan).
 - Dropped Stanislav's Acked-by because of the change above.
v1: https://lore.kernel.org/netdev/20261029032832.1511180-1-bcochrane@google.com/
 drivers/net/ethernet/broadcom/bnxt/bnxt.c | 15 +++++++++++----
 1 file changed, 11 insertions(+), 4 deletions(-)

diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt.c b/drivers/net/ethernet/broadcom/bnxt/bnxt.c
index 33c4cacc8ffe..8b7ebf2eefaa 100644
--- a/drivers/net/ethernet/broadcom/bnxt/bnxt.c
+++ b/drivers/net/ethernet/broadcom/bnxt/bnxt.c
@@ -14838,7 +14838,10 @@ static int bnxt_get_registered_vfs(struct bnxt *bp)
 void bnxt_fw_reset(struct bnxt *bp)
 {
 	bnxt_ulp_stop(bp);
-	bnxt_lock_sp(bp);
+	if (BNXT_PF(bp))
+		bnxt_rtnl_lock_sp(bp);
+	else
+		bnxt_lock_sp(bp);
 	if (test_bit(BNXT_STATE_OPEN, &bp->state) &&
 	    !test_bit(BNXT_STATE_IN_FW_RESET, &bp->state)) {
 		struct bnxt_ptp_cfg *ptp = bp->ptp_cfg;
@@ -14884,7 +14887,10 @@ void bnxt_fw_reset(struct bnxt *bp)
 		bnxt_queue_fw_reset_work(bp, tmo);
 	}
 fw_reset_exit:
-	bnxt_unlock_sp(bp);
+	if (BNXT_PF(bp))
+		bnxt_rtnl_unlock_sp(bp);
+	else
+		bnxt_unlock_sp(bp);
 }
 
 static void bnxt_chk_missed_irq(struct bnxt *bp)
@@ -15547,9 +15553,8 @@ static void bnxt_fw_reset_task(struct work_struct *work)
 		bp->fw_reset_timestamp = jiffies;
 		netdev_lock(bp->dev);
 		if (test_bit(BNXT_STATE_ABORT_ERR, &bp->state)) {
-			bnxt_fw_reset_abort(bp, rc);
 			netdev_unlock(bp->dev);
-			goto ulp_start;
+			goto fw_reset_abort;
 		}
 		bnxt_fw_reset_close(bp);
 		if (bp->fw_cap & BNXT_FW_CAP_ERR_RECOVER_RELOAD) {
@@ -15682,9 +15687,11 @@ static void bnxt_fw_reset_task(struct work_struct *work)
 		netdev_err(bp->dev, "fw_health_status 0x%x\n", sts);
 	}
 fw_reset_abort:
+	rtnl_lock();
 	netdev_lock(bp->dev);
 	bnxt_fw_reset_abort(bp, rc);
 	netdev_unlock(bp->dev);
+	rtnl_unlock();
 ulp_start:
 	if (!rc)
 		bnxt_ulp_start(bp);

base-commit: a45dbf182137275f44ad7c8bc8acd23f7af6b2cf
-- 
2.56.0.506.g506a7eabc5-goog