[PATCH AUTOSEL 6.1 08/38] selftests/bpf: Verify copy_register_state() preserves parent/live fields

[PATCH AUTOSEL 6.1 08/38] selftests/bpf: Verify copy_register_state() preserves parent/live fields

[Sasha Levin]

From: Eduard Zingerman <eddyz87@gmail.com>

[ Upstream commit b9fa9bc839291020b362ab5392e5f18ba79657ac ]

A testcase to check that verifier.c:copy_register_state() preserves
register parentage chain and livness information.

Signed-off-by: Eduard Zingerman <eddyz87@gmail.com>
Link: https://lore.kernel.org/r/20230106142214.1040390-3-eddyz87@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 .../selftests/bpf/verifier/search_pruning.c   | 36 +++++++++++++++++++
 1 file changed, 36 insertions(+)

diff --git a/tools/testing/selftests/bpf/verifier/search_pruning.c b/tools/testing/selftests/bpf/verifier/search_pruning.c
index 68b14fdfebdb1..d63fd8991b03a 100644
--- a/tools/testing/selftests/bpf/verifier/search_pruning.c
+++ b/tools/testing/selftests/bpf/verifier/search_pruning.c
@@ -225,3 +225,39 @@
 	.result_unpriv = ACCEPT,
 	.insn_processed = 15,
 },
+/* The test performs a conditional 64-bit write to a stack location
+ * fp[-8], this is followed by an unconditional 8-bit write to fp[-8],
+ * then data is read from fp[-8]. This sequence is unsafe.
+ *
+ * The test would be mistakenly marked as safe w/o dst register parent
+ * preservation in verifier.c:copy_register_state() function.
+ *
+ * Note the usage of BPF_F_TEST_STATE_FREQ to force creation of the
+ * checkpoint state after conditional 64-bit assignment.
+ */
+{
+	"write tracking and register parent chain bug",
+	.insns = {
+	/* r6 = ktime_get_ns() */
+	BPF_EMIT_CALL(BPF_FUNC_ktime_get_ns),
+	BPF_MOV64_REG(BPF_REG_6, BPF_REG_0),
+	/* r0 = ktime_get_ns() */
+	BPF_EMIT_CALL(BPF_FUNC_ktime_get_ns),
+	/* if r0 > r6 goto +1 */
+	BPF_JMP_REG(BPF_JGT, BPF_REG_0, BPF_REG_6, 1),
+	/* *(u64 *)(r10 - 8) = 0xdeadbeef */
+	BPF_ST_MEM(BPF_DW, BPF_REG_FP, -8, 0xdeadbeef),
+	/* r1 = 42 */
+	BPF_MOV64_IMM(BPF_REG_1, 42),
+	/* *(u8 *)(r10 - 8) = r1 */
+	BPF_STX_MEM(BPF_B, BPF_REG_FP, BPF_REG_1, -8),
+	/* r2 = *(u64 *)(r10 - 8) */
+	BPF_LDX_MEM(BPF_DW, BPF_REG_2, BPF_REG_FP, -8),
+	/* exit(0) */
+	BPF_MOV64_IMM(BPF_REG_0, 0),
+	BPF_EXIT_INSN(),
+	},
+	.flags = BPF_F_TEST_STATE_FREQ,
+	.errstr = "invalid read from stack off -8+1 size 8",
+	.result = REJECT,
+},
-- 
2.39.0

[PATCH AUTOSEL 6.1 11/38] bpf, sockmap: Don't let sock_map_{close,destroy,unhash} call itself

[Sasha Levin]

From: Jakub Sitnicki <jakub@cloudflare.com>

[ Upstream commit 5b4a79ba65a1ab479903fff2e604865d229b70a9 ]

sock_map proto callbacks should never call themselves by design. Protect
against bugs like [1] and break out of the recursive loop to avoid a stack
overflow in favor of a resource leak.

[1] https://lore.kernel.org/all/00000000000073b14905ef2e7401@google.com/

Suggested-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Link: https://lore.kernel.org/r/20230113-sockmap-fix-v2-1-1e0ee7ac2f90@cloudflare.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 net/core/sock_map.c | 61 +++++++++++++++++++++++++--------------------
 1 file changed, 34 insertions(+), 27 deletions(-)

diff --git a/net/core/sock_map.c b/net/core/sock_map.c
index 22fa2c5bc6ec9..a68a7290a3b2b 100644
--- a/net/core/sock_map.c
+++ b/net/core/sock_map.c
@@ -1569,15 +1569,16 @@ void sock_map_unhash(struct sock *sk)
 	psock = sk_psock(sk);
 	if (unlikely(!psock)) {
 		rcu_read_unlock();
-		if (sk->sk_prot->unhash)
-			sk->sk_prot->unhash(sk);
-		return;
+		saved_unhash = READ_ONCE(sk->sk_prot)->unhash;
+	} else {
+		saved_unhash = psock->saved_unhash;
+		sock_map_remove_links(sk, psock);
+		rcu_read_unlock();
 	}
-
-	saved_unhash = psock->saved_unhash;
-	sock_map_remove_links(sk, psock);
-	rcu_read_unlock();
-	saved_unhash(sk);
+	if (WARN_ON_ONCE(saved_unhash == sock_map_unhash))
+		return;
+	if (saved_unhash)
+		saved_unhash(sk);
 }
 EXPORT_SYMBOL_GPL(sock_map_unhash);
 
@@ -1590,17 +1591,18 @@ void sock_map_destroy(struct sock *sk)
 	psock = sk_psock_get(sk);
 	if (unlikely(!psock)) {
 		rcu_read_unlock();
-		if (sk->sk_prot->destroy)
-			sk->sk_prot->destroy(sk);
-		return;
+		saved_destroy = READ_ONCE(sk->sk_prot)->destroy;
+	} else {
+		saved_destroy = psock->saved_destroy;
+		sock_map_remove_links(sk, psock);
+		rcu_read_unlock();
+		sk_psock_stop(psock);
+		sk_psock_put(sk, psock);
 	}
-
-	saved_destroy = psock->saved_destroy;
-	sock_map_remove_links(sk, psock);
-	rcu_read_unlock();
-	sk_psock_stop(psock);
-	sk_psock_put(sk, psock);
-	saved_destroy(sk);
+	if (WARN_ON_ONCE(saved_destroy == sock_map_destroy))
+		return;
+	if (saved_destroy)
+		saved_destroy(sk);
 }
 EXPORT_SYMBOL_GPL(sock_map_destroy);
 
@@ -1615,16 +1617,21 @@ void sock_map_close(struct sock *sk, long timeout)
 	if (unlikely(!psock)) {
 		rcu_read_unlock();
 		release_sock(sk);
-		return sk->sk_prot->close(sk, timeout);
+		saved_close = READ_ONCE(sk->sk_prot)->close;
+	} else {
+		saved_close = psock->saved_close;
+		sock_map_remove_links(sk, psock);
+		rcu_read_unlock();
+		sk_psock_stop(psock);
+		release_sock(sk);
+		cancel_work_sync(&psock->work);
+		sk_psock_put(sk, psock);
 	}
-
-	saved_close = psock->saved_close;
-	sock_map_remove_links(sk, psock);
-	rcu_read_unlock();
-	sk_psock_stop(psock);
-	release_sock(sk);
-	cancel_work_sync(&psock->work);
-	sk_psock_put(sk, psock);
+	/* Make sure we do not recurse. This is a bug.
+	 * Leak the socket instead of crashing on a stack overflow.
+	 */
+	if (WARN_ON_ONCE(saved_close == sock_map_close))
+		return;
 	saved_close(sk, timeout);
 }
 EXPORT_SYMBOL_GPL(sock_map_close);
-- 
2.39.0