[bpf-next,v3,15/16] stmmac: remove rcu_read_lock() around XDP program invocation

Message ID 20210617212748.32456-16-toke@redhat.com
State New
Headers show
Series
  • [bpf-next,v3,01/16] rcu: Create an unrcu_pointer() to remove __rcu from a pointer
Related show

Commit Message

Toke Høiland-Jørgensen June 17, 2021, 9:27 p.m.
The stmmac driver has rcu_read_lock()/rcu_read_unlock() pairs around XDP
program invocations. However, the actual lifetime of the objects referred
by the XDP program invocation is longer, all the way through to the call to
xdp_do_flush(), making the scope of the rcu_read_lock() too small. This
turns out to be harmless because it all happens in a single NAPI poll
cycle (and thus under local_bh_disable()), but it makes the rcu_read_lock()
misleading.

Rather than extend the scope of the rcu_read_lock(), just get rid of it
entirely. With the addition of RCU annotations to the XDP_REDIRECT map
types that take bh execution into account, lockdep even understands this to
be safe, so there's really no reason to keep it around.

Cc: Giuseppe Cavallaro <peppe.cavallaro@st.com>
Cc: Alexandre Torgue <alexandre.torgue@foss.st.com>
Cc: Jose Abreu <joabreu@synopsys.com>
Signed-off-by: Toke Høiland-Jørgensen <toke@redhat.com>
---
 drivers/net/ethernet/stmicro/stmmac/stmmac_main.c | 13 +++++--------
 1 file changed, 5 insertions(+), 8 deletions(-)

Comments

Wong Vee Khee June 18, 2021, 9:47 a.m. | #1
On Thu, Jun 17, 2021 at 11:27:47PM +0200, Toke Høiland-Jørgensen wrote:
> The stmmac driver has rcu_read_lock()/rcu_read_unlock() pairs around XDP

> program invocations. However, the actual lifetime of the objects referred

> by the XDP program invocation is longer, all the way through to the call to

> xdp_do_flush(), making the scope of the rcu_read_lock() too small. This

> turns out to be harmless because it all happens in a single NAPI poll

> cycle (and thus under local_bh_disable()), but it makes the rcu_read_lock()

> misleading.

> 

> Rather than extend the scope of the rcu_read_lock(), just get rid of it

> entirely. With the addition of RCU annotations to the XDP_REDIRECT map

> types that take bh execution into account, lockdep even understands this to

> be safe, so there's really no reason to keep it around.

> 

> Cc: Giuseppe Cavallaro <peppe.cavallaro@st.com>

> Cc: Alexandre Torgue <alexandre.torgue@foss.st.com>

> Cc: Jose Abreu <joabreu@synopsys.com>

> Signed-off-by: Toke Høiland-Jørgensen <toke@redhat.com>


Acked-by: Wong Vee Khee <vee.khee.wong@linux.intel.com>

Tested-by: Song, Yoong Siang <yoong.siang.song@intel.com>


> ---

>  drivers/net/ethernet/stmicro/stmmac/stmmac_main.c | 13 +++++--------

>  1 file changed, 5 insertions(+), 8 deletions(-)

Patch

diff --git a/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c b/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c
index bf9fe25fed69..5dcc8a42abf9 100644
--- a/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c
+++ b/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c
@@ -4654,7 +4654,6 @@  static int stmmac_xdp_xmit_back(struct stmmac_priv *priv,
 	return res;
 }
 
-/* This function assumes rcu_read_lock() is held by the caller. */
 static int __stmmac_xdp_run_prog(struct stmmac_priv *priv,
 				 struct bpf_prog *prog,
 				 struct xdp_buff *xdp)
@@ -4662,6 +4661,9 @@  static int __stmmac_xdp_run_prog(struct stmmac_priv *priv,
 	u32 act;
 	int res;
 
+	/* This code is invoked within a single NAPI poll cycle and thus under
+	 * local_bh_disable(), which provides the needed RCU protection.
+	 */
 	act = bpf_prog_run_xdp(prog, xdp);
 	switch (act) {
 	case XDP_PASS:
@@ -4696,17 +4698,14 @@  static struct sk_buff *stmmac_xdp_run_prog(struct stmmac_priv *priv,
 	struct bpf_prog *prog;
 	int res;
 
-	rcu_read_lock();
-
 	prog = READ_ONCE(priv->xdp_prog);
 	if (!prog) {
 		res = STMMAC_XDP_PASS;
-		goto unlock;
+		goto out;
 	}
 
 	res = __stmmac_xdp_run_prog(priv, prog, xdp);
-unlock:
-	rcu_read_unlock();
+out:
 	return ERR_PTR(-res);
 }
 
@@ -4976,10 +4975,8 @@  static int stmmac_rx_zc(struct stmmac_priv *priv, int limit, u32 queue)
 		buf->xdp->data_end = buf->xdp->data + buf1_len;
 		xsk_buff_dma_sync_for_cpu(buf->xdp, rx_q->xsk_pool);
 
-		rcu_read_lock();
 		prog = READ_ONCE(priv->xdp_prog);
 		res = __stmmac_xdp_run_prog(priv, prog, buf->xdp);
-		rcu_read_unlock();
 
 		switch (res) {
 		case STMMAC_XDP_PASS: