Re: [PATCH v2 bpf-next 2/5] bpf: af_unix: Use batching algorithm in bpf unix iter.

Re: [PATCH v2 bpf-next 2/5] bpf: af_unix: Use batching algorithm in bpf unix iter.

[kernel test robot]

[-- Attachment #1: Type: text/plain, Size: 4336 bytes --]

CC: kbuild-all(a)lists.01.org
In-Reply-To: <20220113002849.4384-3-kuniyu@amazon.co.jp>
References: <20220113002849.4384-3-kuniyu@amazon.co.jp>
TO: Kuniyuki Iwashima <kuniyu@amazon.co.jp>
TO: "David S. Miller" <davem@davemloft.net>
CC: netdev(a)vger.kernel.org
TO: Jakub Kicinski <kuba@kernel.org>
TO: Alexei Starovoitov <ast@kernel.org>
TO: Daniel Borkmann <daniel@iogearbox.net>
TO: Andrii Nakryiko <andrii@kernel.org>
CC: Martin KaFai Lau <kafai@fb.com>
CC: Benjamin Herrenschmidt <benh@amazon.com>
CC: Kuniyuki Iwashima <kuniyu@amazon.co.jp>
CC: bpf(a)vger.kernel.org

Hi Kuniyuki,

Thank you for the patch! Perhaps something to improve:

[auto build test WARNING on bpf-next/master]

url:    https://github.com/0day-ci/linux/commits/Kuniyuki-Iwashima/bpf-Batching-iter-for-AF_UNIX-sockets/20220113-083151
base:   https://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf-next.git master
:::::: branch date: 17 hours ago
:::::: commit date: 17 hours ago
config: arm-randconfig-m031-20220113 (https://download.01.org/0day-ci/archive/20220114/202201140137.dztxZIjJ-lkp(a)intel.com/config)
compiler: arm-linux-gnueabi-gcc (GCC) 11.2.0

If you fix the issue, kindly add following tag as appropriate
Reported-by: kernel test robot <lkp@intel.com>
Reported-by: Dan Carpenter <dan.carpenter@oracle.com>

New smatch warnings:
net/unix/af_unix.c:3464 bpf_iter_unix_batch() warn: sleeping in atomic context

Old smatch warnings:
net/unix/af_unix.c:1586 unix_stream_connect() warn: variable dereferenced before check 'other' (see line 1469)
net/unix/af_unix.c:2652 manage_oob() warn: returning freed memory 'skb'

vim +3464 net/unix/af_unix.c

a8f4be17743b80 Kuniyuki Iwashima 2022-01-13  3436  
a8f4be17743b80 Kuniyuki Iwashima 2022-01-13  3437  static struct sock *bpf_iter_unix_batch(struct seq_file *seq,
a8f4be17743b80 Kuniyuki Iwashima 2022-01-13  3438  					loff_t *pos)
a8f4be17743b80 Kuniyuki Iwashima 2022-01-13  3439  {
a8f4be17743b80 Kuniyuki Iwashima 2022-01-13  3440  	struct bpf_unix_iter_state *iter = seq->private;
a8f4be17743b80 Kuniyuki Iwashima 2022-01-13  3441  	unsigned int expected;
a8f4be17743b80 Kuniyuki Iwashima 2022-01-13  3442  	bool resized = false;
a8f4be17743b80 Kuniyuki Iwashima 2022-01-13  3443  	struct sock *sk;
a8f4be17743b80 Kuniyuki Iwashima 2022-01-13  3444  
a8f4be17743b80 Kuniyuki Iwashima 2022-01-13  3445  	if (iter->st_bucket_done)
a8f4be17743b80 Kuniyuki Iwashima 2022-01-13  3446  		*pos = set_bucket_offset(get_bucket(*pos) + 1, 1);
a8f4be17743b80 Kuniyuki Iwashima 2022-01-13  3447  
a8f4be17743b80 Kuniyuki Iwashima 2022-01-13  3448  again:
a8f4be17743b80 Kuniyuki Iwashima 2022-01-13  3449  	/* Get a new batch */
a8f4be17743b80 Kuniyuki Iwashima 2022-01-13  3450  	iter->cur_sk = 0;
a8f4be17743b80 Kuniyuki Iwashima 2022-01-13  3451  	iter->end_sk = 0;
a8f4be17743b80 Kuniyuki Iwashima 2022-01-13  3452  
a8f4be17743b80 Kuniyuki Iwashima 2022-01-13  3453  	sk = unix_get_first(seq, pos);
a8f4be17743b80 Kuniyuki Iwashima 2022-01-13  3454  	if (!sk)
a8f4be17743b80 Kuniyuki Iwashima 2022-01-13  3455  		return NULL; /* Done */
a8f4be17743b80 Kuniyuki Iwashima 2022-01-13  3456  
a8f4be17743b80 Kuniyuki Iwashima 2022-01-13  3457  	expected = bpf_iter_unix_hold_batch(seq, sk);
a8f4be17743b80 Kuniyuki Iwashima 2022-01-13  3458  
a8f4be17743b80 Kuniyuki Iwashima 2022-01-13  3459  	if (iter->end_sk == expected) {
a8f4be17743b80 Kuniyuki Iwashima 2022-01-13  3460  		iter->st_bucket_done = true;
a8f4be17743b80 Kuniyuki Iwashima 2022-01-13  3461  		return sk;
a8f4be17743b80 Kuniyuki Iwashima 2022-01-13  3462  	}
a8f4be17743b80 Kuniyuki Iwashima 2022-01-13  3463  
a8f4be17743b80 Kuniyuki Iwashima 2022-01-13 @3464  	if (!resized && !bpf_iter_unix_realloc_batch(iter, expected * 3 / 2)) {
a8f4be17743b80 Kuniyuki Iwashima 2022-01-13  3465  		resized = true;
a8f4be17743b80 Kuniyuki Iwashima 2022-01-13  3466  		goto again;
a8f4be17743b80 Kuniyuki Iwashima 2022-01-13  3467  	}
a8f4be17743b80 Kuniyuki Iwashima 2022-01-13  3468  
a8f4be17743b80 Kuniyuki Iwashima 2022-01-13  3469  	return sk;
a8f4be17743b80 Kuniyuki Iwashima 2022-01-13  3470  }
a8f4be17743b80 Kuniyuki Iwashima 2022-01-13  3471  

---
0-DAY CI Kernel Test Service, Intel Corporation
https://lists.01.org/hyperkitty/list/kbuild-all(a)lists.01.org

[PATCH v2 bpf-next 0/5] bpf: Batching iter for AF_UNIX sockets.

[Kuniyuki Iwashima]

Last year the commit afd20b9290e1 ("af_unix: Replace the big lock with
small locks.") landed on bpf-next.  Now we can use a batching algorithm
for AF_UNIX bpf iter as TCP bpf iter.


Changelog:
- Add the 1st patch.
- Call unix_get_first() in .start()/.next() to always acquire a lock in
  each iteration in the 2nd patch.


Kuniyuki Iwashima (5):
  af_unix: Refactor unix_next_socket().
  bpf: af_unix: Use batching algorithm in bpf unix iter.
  bpf: Support bpf_(get|set)sockopt() in bpf unix iter.
  selftest/bpf: Test batching and bpf_(get|set)sockopt in bpf unix iter.
  selftest/bpf: Fix a stale comment.

 net/unix/af_unix.c                            | 250 ++++++++++++++++--
 .../bpf/prog_tests/bpf_iter_setsockopt_unix.c | 100 +++++++
 .../bpf/progs/bpf_iter_setsockopt_unix.c      |  60 +++++
 .../selftests/bpf/progs/bpf_iter_unix.c       |   2 +-
 .../selftests/bpf/progs/bpf_tracing_net.h     |   2 +
 5 files changed, 385 insertions(+), 29 deletions(-)
 create mode 100644 tools/testing/selftests/bpf/prog_tests/bpf_iter_setsockopt_unix.c
 create mode 100644 tools/testing/selftests/bpf/progs/bpf_iter_setsockopt_unix.c

-- 
2.30.2

[PATCH v2 bpf-next 2/5] bpf: af_unix: Use batching algorithm in bpf unix iter.

[Kuniyuki Iwashima]

The commit 04c7820b776f ("bpf: tcp: Bpf iter batching and lock_sock")
introduces the batching algorithm to iterate TCP sockets with more
consistency.

This patch uses the same algorithm to iterate AF_UNIX sockets.

Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.co.jp>
---
 net/unix/af_unix.c | 184 +++++++++++++++++++++++++++++++++++++++++++--
 1 file changed, 177 insertions(+), 7 deletions(-)

diff --git a/net/unix/af_unix.c b/net/unix/af_unix.c
index e1c4082accdb..d383d5f63b6b 100644
--- a/net/unix/af_unix.c
+++ b/net/unix/af_unix.c
@@ -3356,6 +3356,15 @@ static const struct seq_operations unix_seq_ops = {
 };
 
 #if IS_BUILTIN(CONFIG_UNIX) && defined(CONFIG_BPF_SYSCALL)
+struct bpf_unix_iter_state {
+	struct seq_net_private p;
+	unsigned int cur_sk;
+	unsigned int end_sk;
+	unsigned int max_sk;
+	struct sock **batch;
+	bool st_bucket_done;
+};
+
 struct bpf_iter__unix {
 	__bpf_md_ptr(struct bpf_iter_meta *, meta);
 	__bpf_md_ptr(struct unix_sock *, unix_sk);
@@ -3374,24 +3383,156 @@ static int unix_prog_seq_show(struct bpf_prog *prog, struct bpf_iter_meta *meta,
 	return bpf_iter_run_prog(prog, &ctx);
 }
 
+static int bpf_iter_unix_hold_batch(struct seq_file *seq, struct sock *start_sk)
+
+{
+	struct bpf_unix_iter_state *iter = seq->private;
+	unsigned int expected = 1;
+	struct sock *sk;
+
+	sock_hold(start_sk);
+	iter->batch[iter->end_sk++] = start_sk;
+
+	for (sk = sk_next(start_sk); sk; sk = sk_next(sk)) {
+		if (sock_net(sk) != seq_file_net(seq))
+			continue;
+
+		if (iter->end_sk < iter->max_sk) {
+			sock_hold(sk);
+			iter->batch[iter->end_sk++] = sk;
+		}
+
+		expected++;
+	}
+
+	spin_unlock(&unix_table_locks[start_sk->sk_hash]);
+
+	return expected;
+}
+
+static void bpf_iter_unix_put_batch(struct bpf_unix_iter_state *iter)
+{
+	while (iter->cur_sk < iter->end_sk)
+		sock_put(iter->batch[iter->cur_sk++]);
+}
+
+static int bpf_iter_unix_realloc_batch(struct bpf_unix_iter_state *iter,
+				       unsigned int new_batch_sz)
+{
+	struct sock **new_batch;
+
+	new_batch = kvmalloc(sizeof(*new_batch) * new_batch_sz,
+			     GFP_USER | __GFP_NOWARN);
+	if (!new_batch)
+		return -ENOMEM;
+
+	bpf_iter_unix_put_batch(iter);
+	kvfree(iter->batch);
+	iter->batch = new_batch;
+	iter->max_sk = new_batch_sz;
+
+	return 0;
+}
+
+static struct sock *bpf_iter_unix_batch(struct seq_file *seq,
+					loff_t *pos)
+{
+	struct bpf_unix_iter_state *iter = seq->private;
+	unsigned int expected;
+	bool resized = false;
+	struct sock *sk;
+
+	if (iter->st_bucket_done)
+		*pos = set_bucket_offset(get_bucket(*pos) + 1, 1);
+
+again:
+	/* Get a new batch */
+	iter->cur_sk = 0;
+	iter->end_sk = 0;
+
+	sk = unix_get_first(seq, pos);
+	if (!sk)
+		return NULL; /* Done */
+
+	expected = bpf_iter_unix_hold_batch(seq, sk);
+
+	if (iter->end_sk == expected) {
+		iter->st_bucket_done = true;
+		return sk;
+	}
+
+	if (!resized && !bpf_iter_unix_realloc_batch(iter, expected * 3 / 2)) {
+		resized = true;
+		goto again;
+	}
+
+	return sk;
+}
+
+static void *bpf_iter_unix_seq_start(struct seq_file *seq, loff_t *pos)
+{
+	if (!*pos)
+		return SEQ_START_TOKEN;
+
+	/* bpf iter does not support lseek, so it always
+	 * continue from where it was stop()-ped.
+	 */
+	return bpf_iter_unix_batch(seq, pos);
+}
+
+static void *bpf_iter_unix_seq_next(struct seq_file *seq, void *v, loff_t *pos)
+{
+	struct bpf_unix_iter_state *iter = seq->private;
+	struct sock *sk;
+
+	/* Whenever seq_next() is called, the iter->cur_sk is
+	 * done with seq_show(), so advance to the next sk in
+	 * the batch.
+	 */
+	if (iter->cur_sk < iter->end_sk)
+		sock_put(iter->batch[iter->cur_sk++]);
+
+	++*pos;
+
+	if (iter->cur_sk < iter->end_sk)
+		sk = iter->batch[iter->cur_sk];
+	else
+		sk = bpf_iter_unix_batch(seq, pos);
+
+	return sk;
+}
+
 static int bpf_iter_unix_seq_show(struct seq_file *seq, void *v)
 {
 	struct bpf_iter_meta meta;
 	struct bpf_prog *prog;
 	struct sock *sk = v;
 	uid_t uid;
+	bool slow;
+	int ret;
 
 	if (v == SEQ_START_TOKEN)
 		return 0;
 
+	slow = lock_sock_fast(sk);
+
+	if (unlikely(sk_unhashed(sk))) {
+		ret = SEQ_SKIP;
+		goto unlock;
+	}
+
 	uid = from_kuid_munged(seq_user_ns(seq), sock_i_uid(sk));
 	meta.seq = seq;
 	prog = bpf_iter_get_info(&meta, false);
-	return unix_prog_seq_show(prog, &meta, v, uid);
+	ret = unix_prog_seq_show(prog, &meta, v, uid);
+unlock:
+	unlock_sock_fast(sk, slow);
+	return ret;
 }
 
 static void bpf_iter_unix_seq_stop(struct seq_file *seq, void *v)
 {
+	struct bpf_unix_iter_state *iter = seq->private;
 	struct bpf_iter_meta meta;
 	struct bpf_prog *prog;
 
@@ -3402,12 +3543,13 @@ static void bpf_iter_unix_seq_stop(struct seq_file *seq, void *v)
 			(void)unix_prog_seq_show(prog, &meta, v, 0);
 	}
 
-	unix_seq_stop(seq, v);
+	if (iter->cur_sk < iter->end_sk)
+		bpf_iter_unix_put_batch(iter);
 }
 
 static const struct seq_operations bpf_iter_unix_seq_ops = {
-	.start	= unix_seq_start,
-	.next	= unix_seq_next,
+	.start	= bpf_iter_unix_seq_start,
+	.next	= bpf_iter_unix_seq_next,
 	.stop	= bpf_iter_unix_seq_stop,
 	.show	= bpf_iter_unix_seq_show,
 };
@@ -3456,11 +3598,39 @@ static struct pernet_operations unix_net_ops = {
 DEFINE_BPF_ITER_FUNC(unix, struct bpf_iter_meta *meta,
 		     struct unix_sock *unix_sk, uid_t uid)
 
+#define INIT_BATCH_SZ 16
+
+static int bpf_iter_init_unix(void *priv_data, struct bpf_iter_aux_info *aux)
+{
+	struct bpf_unix_iter_state *iter = priv_data;
+	int err;
+
+	err = bpf_iter_init_seq_net(priv_data, aux);
+	if (err)
+		return err;
+
+	err = bpf_iter_unix_realloc_batch(iter, INIT_BATCH_SZ);
+	if (err) {
+		bpf_iter_fini_seq_net(priv_data);
+		return err;
+	}
+
+	return 0;
+}
+
+static void bpf_iter_fini_unix(void *priv_data)
+{
+	struct bpf_unix_iter_state *iter = priv_data;
+
+	bpf_iter_fini_seq_net(priv_data);
+	kvfree(iter->batch);
+}
+
 static const struct bpf_iter_seq_info unix_seq_info = {
 	.seq_ops		= &bpf_iter_unix_seq_ops,
-	.init_seq_private	= bpf_iter_init_seq_net,
-	.fini_seq_private	= bpf_iter_fini_seq_net,
-	.seq_priv_size		= sizeof(struct seq_net_private),
+	.init_seq_private	= bpf_iter_init_unix,
+	.fini_seq_private	= bpf_iter_fini_unix,
+	.seq_priv_size		= sizeof(struct bpf_unix_iter_state),
 };
 
 static struct bpf_iter_reg unix_reg_info = {
-- 
2.30.2