[PATCH net v2 v2 2/3] macsec: use rcu_work to defer RX SA crypto cleanup out of softirq

[alexjlzheng]

From: Jinliang Zheng <alexjlzheng@xxxxxxxxxxx>

crypto_free_aead() can internally invoke vunmap() (e.g. via
dma_free_attrs() in hardware crypto drivers such as hisi_sec2).
vunmap() must not be called from softirq context, but free_rxsa()
is an RCU callback that runs in softirq, leading to a kernel crash:

  vunmap+0x4c/0x70
  __iommu_dma_free+0xd0/0x138
  dma_free_attrs+0xf4/0x100
  sec_aead_exit+0x64/0xb8 [hisi_sec2]
  crypto_destroy_tfm+0x98/0x110
  free_rxsa+0x28/0x50 [macsec]
  rcu_do_batch+0x184/0x460
  rcu_core+0xf4/0x1f8
  handle_softirqs+0x118/0x330

Use rcu_work to defer the cleanup to a workqueue. rcu_work dispatches
the worker asynchronously after the RCU grace period, so no thread
blocks waiting, and concurrent releases of multiple SAs naturally
share the same grace period.

Fixes: c09440f7dcb3 ("macsec: introduce IEEE 802.1AE driver")
Signed-off-by: Jinliang Zheng <alexjlzheng@xxxxxxxxxxx>
---
 drivers/net/macsec.c | 8 +++++---
 include/net/macsec.h | 3 ++-
 2 files changed, 7 insertions(+), 4 deletions(-)

diff --git a/drivers/net/macsec.c b/drivers/net/macsec.c
index ddb22473e701..e0862ecd23f2 100644
--- a/drivers/net/macsec.c
+++ b/drivers/net/macsec.c
@@ -176,9 +176,10 @@ static void macsec_rxsc_put(struct macsec_rx_sc *sc)
 		call_rcu(&sc->rcu_head, free_rx_sc_rcu);
 }
 
-static void free_rxsa(struct rcu_head *head)
+static void free_rxsa_work(struct work_struct *work)
 {
-	struct macsec_rx_sa *sa = container_of(head, struct macsec_rx_sa, rcu);
+	struct macsec_rx_sa *sa =
+		container_of(to_rcu_work(work), struct macsec_rx_sa, destroy_work);
 
 	crypto_free_aead(sa->key.tfm);
 	free_percpu(sa->stats);
@@ -188,7 +189,7 @@ static void free_rxsa(struct rcu_head *head)
 static void macsec_rxsa_put(struct macsec_rx_sa *sa)
 {
 	if (refcount_dec_and_test(&sa->refcnt))
-		call_rcu(&sa->rcu, free_rxsa);
+		queue_rcu_work(macsec_wq, &sa->destroy_work);
 }
 
 static struct macsec_tx_sa *macsec_txsa_get(struct macsec_tx_sa __rcu *ptr)
@@ -1409,6 +1410,7 @@ static int init_rx_sa(struct macsec_rx_sa *rx_sa, char *sak, int key_len,
 	rx_sa->next_pn = 1;
 	refcount_set(&rx_sa->refcnt, 1);
 	spin_lock_init(&rx_sa->lock);
+	INIT_RCU_WORK(&rx_sa->destroy_work, free_rxsa_work);
 
 	return 0;
 }
diff --git a/include/net/macsec.h b/include/net/macsec.h
index bc7de5b53e54..d6136debe1c8 100644
--- a/include/net/macsec.h
+++ b/include/net/macsec.h
@@ -9,6 +9,7 @@
 
 #include <linux/u64_stats_sync.h>
 #include <linux/if_vlan.h>
+#include <linux/workqueue.h>
 #include <uapi/linux/if_link.h>
 #include <uapi/linux/if_macsec.h>
 
@@ -136,7 +137,7 @@ struct macsec_rx_sa {
 	bool active;
 	struct macsec_rx_sa_stats __percpu *stats;
 	struct macsec_rx_sc *sc;
-	struct rcu_head rcu;
+	struct rcu_work destroy_work;
 };
 
 struct pcpu_rx_sc_stats {
-- 
2.39.3