[PATCH net v2] net: gro_cells: Reduce lock scope in gro_cell_poll

One GRO-cell device's NAPI callback can nest into the GRO-cell of
another device if the underlying device is also using GRO-cell.
This is the case for IPsec over vxlan.
These two GRO-cells are separate devices. From lockdep's point of view
it is the same because each device is sharing the same lock class and so
it reports a possible deadlock assuming one device is nesting into
itself.

Hold the bh_lock only while accessing gro_cell::napi_skbs in
gro_cell_poll(). This reduces the locking scope and avoids acquiring the
same lock class multiple times.

Fixes: 25718fdcbdd2 ("net: gro_cells: Use nested-BH locking for gro_cell")
Reported-by: Gal Pressman <redacted>
Closes: https://lore.kernel.org/all/66664116-edb8-48dc-ad72-d5223696dd19@nvidia.com/ (local)
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
---
v1…v2:
  - Drop the lock and reacquire it again in gro_cell_poll() instead
    providing the lock class. The addition lock class needs to be
    registered and unregistered. The latter must not have from the RCU
    callback. This looks simpler.
    Reported by Jakub.

 net/core/gro_cells.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/net/core/gro_cells.c b/net/core/gro_cells.c
index fd57b845de333..a725d21159a6f 100644
--- a/net/core/gro_cells.c
+++ b/net/core/gro_cells.c

@@ -60,9 +60,10 @@ static int gro_cell_poll(struct napi_struct *napi, int budget)
 	struct sk_buff *skb;
 	int work_done = 0;
 
-	__local_lock_nested_bh(&cell->bh_lock);
 	while (work_done < budget) {
+		__local_lock_nested_bh(&cell->bh_lock);
 		skb = __skb_dequeue(&cell->napi_skbs);
+		__local_unlock_nested_bh(&cell->bh_lock);
 		if (!skb)
 			break;
 		napi_gro_receive(napi, skb);

@@ -71,7 +72,6 @@ static int gro_cell_poll(struct napi_struct *napi, int budget)
 
 	if (work_done < budget)
 		napi_complete_done(napi, work_done);
-	__local_unlock_nested_bh(&cell->bh_lock);
 	return work_done;
 }
 

-- 
2.51.0