bpf, xdp: drop rcu_read_lock from bpf_prog_run_xdp and move to caller

After 326fe02 ("net/mlx4_en: protect ring->xdp_prog with rcu_read_lock"),
the rcu_read_lock() in bpf_prog_run_xdp() is superfluous, since callers
need to hold rcu_read_lock() already to make sure BPF program doesn't
get released in the background.

Thus, drop it from bpf_prog_run_xdp(), as it can otherwise be misleading.
Still keeping the bpf_prog_run_xdp() is useful as it allows for grepping
in XDP supported drivers and to keep the typecheck on the context intact.
For mlx4, this means we don't have a double rcu_read_lock() anymore. nfp can
just make use of bpf_prog_run_xdp(), too. For qede, just move rcu_read_lock()
out of the helper. When the driver gets atomic replace support, this will
move to call-sites eventually.

mlx5 needs actual fixing as it has the same issue as described already in
326fe02 ("net/mlx4_en: protect ring->xdp_prog with rcu_read_lock"),
that is, we're under RCU bh at this time, BPF programs are released via
call_rcu(), and call_rcu() != call_rcu_bh(), so we need to properly mark
read side as programs can get xchg()'ed in mlx5e_xdp_set() without queue
reset.

Fixes: 8699415 ("net/mlx5e: XDP fast RX drop bpf programs support")
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

drivers/net/ethernet/mellanox/mlx5/core/en_rx.c

@@ -737,10 +737,10 @@ static inline
 struct sk_buff *skb_from_cqe(struct mlx5e_rq *rq, struct mlx5_cqe64 *cqe,
 			     u16 wqe_counter, u32 cqe_bcnt)
 {
-	struct bpf_prog *xdp_prog = READ_ONCE(rq->xdp_prog);
 	struct mlx5e_dma_info *di;
 	struct sk_buff *skb;
 	void *va, *data;
+	bool consumed;
 
 	di             = &rq->dma_info[wqe_counter];
 	va             = page_address(di->page);
@@ -759,7 +759,11 @@ struct sk_buff *skb_from_cqe(struct mlx5e_rq *rq, struct mlx5_cqe64 *cqe,
 		return NULL;
 	}
 
-	if (mlx5e_xdp_handle(rq, xdp_prog, di, data, cqe_bcnt))
+	rcu_read_lock();
+	consumed = mlx5e_xdp_handle(rq, READ_ONCE(rq->xdp_prog), di, data,
+				    cqe_bcnt);
+	rcu_read_unlock();
+	if (consumed)
 		return NULL; /* page/packet was consumed by XDP */
 
 	skb = build_skb(va, RQ_PAGE_SIZE(rq));

drivers/net/ethernet/netronome/nfp/nfp_net_common.c

@@ -1518,7 +1518,7 @@ static int nfp_net_run_xdp(struct bpf_prog *prog, void *data, unsigned int len)
 	xdp.data = data;
 	xdp.data_end = data + len;
 
-	return BPF_PROG_RUN(prog, &xdp);
+	return bpf_prog_run_xdp(prog, &xdp);
 }
 
 /**

drivers/net/ethernet/qlogic/qede/qede_main.c

@@ -1497,7 +1497,14 @@ static bool qede_rx_xdp(struct qede_dev *edev,
 
 	xdp.data = page_address(bd->data) + cqe->placement_offset;
 	xdp.data_end = xdp.data + len;
+
+	/* Queues always have a full reset currently, so for the time
+	 * being until there's atomic program replace just mark read
+	 * side for map helpers.
+	 */
+	rcu_read_lock();
 	act = bpf_prog_run_xdp(prog, &xdp);
+	rcu_read_unlock();
 
 	if (act == XDP_PASS)
 		return true;

include/linux/filter.h

@@ -498,16 +498,16 @@ static inline u32 bpf_prog_run_clear_cb(const struct bpf_prog *prog,
 	return BPF_PROG_RUN(prog, skb);
 }
 
-static inline u32 bpf_prog_run_xdp(const struct bpf_prog *prog,
-				   struct xdp_buff *xdp)
+static __always_inline u32 bpf_prog_run_xdp(const struct bpf_prog *prog,
+					    struct xdp_buff *xdp)
 {
-	u32 ret;
-
-	rcu_read_lock();
-	ret = BPF_PROG_RUN(prog, xdp);
-	rcu_read_unlock();
-
-	return ret;
+	/* Caller needs to hold rcu_read_lock() (!), otherwise program
+	 * can be released while still running, or map elements could be
+	 * freed early while still having concurrent users. XDP fastpath
+	 * already takes rcu_read_lock() when fetching the program, so
+	 * it's not necessary here anymore.
+	 */
+	return BPF_PROG_RUN(prog, xdp);
 }
 
 static inline unsigned int bpf_prog_size(unsigned int proglen)