| Message ID | 20220621171913.73401-6-kuniyu@amazon.com (mailing list archive) |
|---|---|
| State | Accepted |
| Commit | cf2f225e2653734e66e91c09e1cbe004bfd3d4a7 |
| Delegated to: | Netdev Maintainers |
| Headers | show |
| Series | af_unix: Introduce per-netns socket hash table. | expand |
Greeting,
FYI, we noticed the following commit (built with gcc-11):
commit: d66d39e4713c1c9e70965e2375b780ab9522dfb3 ("[PATCH v3 net-next 5/6] af_unix: Put a socket into a per-netns hash table.")
url: https://github.com/intel-lab-lkp/linux/commits/Kuniyuki-Iwashima/af_unix-Introduce-per-netns-socket-hash-table/20220622-012539
base: https://git.kernel.org/cgit/linux/kernel/git/davem/net-next.git 8720bd951b8e8515ffd995c7631790fdabaa9265
patch link: https://lore.kernel.org/netdev/20220621171913.73401-6-kuniyu@amazon.com
in testcase: ltp
version: ltp-x86_64-14c1f76-1_20220625
with following parameters:
test: net.rpc_tests
ucode: 0x21
test-description: The LTP testsuite contains a collection of tools for testing the Linux kernel and related features.
test-url: http://linux-test-project.github.io/
on test machine: 4 threads 1 sockets Intel(R) Core(TM) i3-3220 CPU @ 3.30GHz with 8G memory
caused below changes (please refer to attached dmesg/kmsg for entire log/backtrace):
please be noted we also observed other three tests failed upon this commit
while pass on parent:
39d14b6772921fa2 d66d39e4713c1c9e70965e2375b
---------------- ---------------------------
fail:runs %reproduction fail:runs
| | |
:6 100% 6:6 ltp.rpc_pmap_set.fail
:6 100% 6:6 ltp.rpc_pmap_unset.fail
:6 100% 6:6 ltp.rpc_registerrpc.fail
:6 100% 6:6 ltp.rpc_svc_register.fail
If you fix the issue, kindly add following tag
Reported-by: kernel test robot <oliver.sang@intel.com>
<<<test_start>>>
tag=rpc_pmap_set stime=1656693410
cmdline="rpc_test.sh -c rpc_pmap_set"
contacts=""
analysis=exit
<<<test_output>>>
rpc_test 1 TINFO: initialize 'lhost' 'ltp_ns_veth2' interface
rpc_test 1 TINFO: add local addr 10.0.0.2/24
rpc_test 1 TINFO: add local addr fd00:1:1:1::2/64
rpc_test 1 TINFO: initialize 'rhost' 'ltp_ns_veth1' interface
rpc_test 1 TINFO: add remote addr 10.0.0.1/24
rpc_test 1 TINFO: add remote addr fd00:1:1:1::1/64
rpc_test 1 TINFO: Network config (local -- remote):
rpc_test 1 TINFO: ltp_ns_veth2 -- ltp_ns_veth1
rpc_test 1 TINFO: 10.0.0.2/24 -- 10.0.0.1/24
rpc_test 1 TINFO: fd00:1:1:1::2/64 -- fd00:1:1:1::1/64
rpc_test 1 TINFO: timeout per run is 0h 5m 0s
rpc_test 1 TINFO: check registered RPC with rpcinfo
rpc_test 1 TINFO: registered RPC:
program vers proto port service
100000 4 tcp 111 portmapper
100000 3 tcp 111 portmapper
100000 2 tcp 111 portmapper
100000 4 udp 111 portmapper
100000 3 udp 111 portmapper
100000 2 udp 111 portmapper
100001 1 udp 44050 rstatd
100001 2 udp 44050 rstatd
100001 3 udp 44050 rstatd
100001 4 udp 44050 rstatd
100001 5 udp 44050 rstatd
100002 2 udp 35568 rusersd
100002 3 udp 35568 rusersd
2000333 10 udp 51134
2000333 10 tcp 63711
rpc_test 1 TINFO: using libtirpc: yes
rpc_test 1 TFAIL: rpc_pmap_set 10.0.0.2 536875000 failed unexpectedly
1
Summary:
passed 0
failed 1
broken 0
skipped 0
warnings 0
<<<execution_status>>>
initiation_status="ok"
duration=1 termination_type=exited termination_id=1 corefile=no
cutime=7 cstime=44
<<<test_end>>>
<<<test_start>>>
tag=rpc_pmap_unset stime=1656693411
cmdline="rpc_test.sh -c rpc_pmap_unset"
contacts=""
analysis=exit
<<<test_output>>>
rpc_test 1 TINFO: initialize 'lhost' 'ltp_ns_veth2' interface
rpc_test 1 TINFO: add local addr 10.0.0.2/24
rpc_test 1 TINFO: add local addr fd00:1:1:1::2/64
rpc_test 1 TINFO: initialize 'rhost' 'ltp_ns_veth1' interface
rpc_test 1 TINFO: add remote addr 10.0.0.1/24
rpc_test 1 TINFO: add remote addr fd00:1:1:1::1/64
rpc_test 1 TINFO: Network config (local -- remote):
rpc_test 1 TINFO: ltp_ns_veth2 -- ltp_ns_veth1
rpc_test 1 TINFO: 10.0.0.2/24 -- 10.0.0.1/24
rpc_test 1 TINFO: fd00:1:1:1::2/64 -- fd00:1:1:1::1/64
rpc_test 1 TINFO: timeout per run is 0h 5m 0s
rpc_test 1 TINFO: check registered RPC with rpcinfo
rpc_test 1 TINFO: registered RPC:
program vers proto port service
100000 4 tcp 111 portmapper
100000 3 tcp 111 portmapper
100000 2 tcp 111 portmapper
100000 4 udp 111 portmapper
100000 3 udp 111 portmapper
100000 2 udp 111 portmapper
100001 1 udp 44050 rstatd
100001 2 udp 44050 rstatd
100001 3 udp 44050 rstatd
100001 4 udp 44050 rstatd
100001 5 udp 44050 rstatd
100002 2 udp 35568 rusersd
100002 3 udp 35568 rusersd
2000333 10 udp 51134
2000333 10 tcp 63711
rpc_test 1 TINFO: using libtirpc: yes
rpc_test 1 TFAIL: rpc_pmap_unset 10.0.0.2 536875000 failed unexpectedly
1
Summary:
passed 0
failed 1
broken 0
skipped 0
warnings 0
<<<execution_status>>>
initiation_status="ok"
duration=0 termination_type=exited termination_id=1 corefile=no
cutime=8 cstime=44
<<<test_end>>>
...
<<<test_start>>>
tag=rpc_svc_register stime=1656693442
cmdline="rpc_test.sh -c rpc_svc_register"
contacts=""
analysis=exit
<<<test_output>>>
rpc_test 1 TINFO: initialize 'lhost' 'ltp_ns_veth2' interface
rpc_test 1 TINFO: add local addr 10.0.0.2/24
rpc_test 1 TINFO: add local addr fd00:1:1:1::2/64
rpc_test 1 TINFO: initialize 'rhost' 'ltp_ns_veth1' interface
rpc_test 1 TINFO: add remote addr 10.0.0.1/24
rpc_test 1 TINFO: add remote addr fd00:1:1:1::1/64
rpc_test 1 TINFO: Network config (local -- remote):
rpc_test 1 TINFO: ltp_ns_veth2 -- ltp_ns_veth1
rpc_test 1 TINFO: 10.0.0.2/24 -- 10.0.0.1/24
rpc_test 1 TINFO: fd00:1:1:1::2/64 -- fd00:1:1:1::1/64
rpc_test 1 TINFO: timeout per run is 0h 5m 0s
rpc_test 1 TINFO: check registered RPC with rpcinfo
rpc_test 1 TINFO: registered RPC:
program vers proto port service
100000 4 tcp 111 portmapper
100000 3 tcp 111 portmapper
100000 2 tcp 111 portmapper
100000 4 udp 111 portmapper
100000 3 udp 111 portmapper
100000 2 udp 111 portmapper
100001 1 udp 44050 rstatd
100001 2 udp 44050 rstatd
100001 3 udp 44050 rstatd
100001 4 udp 44050 rstatd
100001 5 udp 44050 rstatd
100002 2 udp 35568 rusersd
100002 3 udp 35568 rusersd
2000333 10 udp 51134
2000333 10 tcp 63711
rpc_test 1 TINFO: using libtirpc: yes
rpc_test 1 TFAIL: rpc_svc_register 10.0.0.2 536875000 failed unexpectedly
1
Summary:
passed 0
failed 1
broken 0
skipped 0
warnings 0
<<<execution_status>>>
initiation_status="ok"
duration=1 termination_type=exited termination_id=1 corefile=no
cutime=6 cstime=45
<<<test_end>>>
...
<<<test_start>>>
tag=rpc_registerrpc stime=1656693443
cmdline="rpc_test.sh -c rpc_registerrpc"
contacts=""
analysis=exit
<<<test_output>>>
rpc_test 1 TINFO: initialize 'lhost' 'ltp_ns_veth2' interface
rpc_test 1 TINFO: add local addr 10.0.0.2/24
rpc_test 1 TINFO: add local addr fd00:1:1:1::2/64
rpc_test 1 TINFO: initialize 'rhost' 'ltp_ns_veth1' interface
rpc_test 1 TINFO: add remote addr 10.0.0.1/24
rpc_test 1 TINFO: add remote addr fd00:1:1:1::1/64
rpc_test 1 TINFO: Network config (local -- remote):
rpc_test 1 TINFO: ltp_ns_veth2 -- ltp_ns_veth1
rpc_test 1 TINFO: 10.0.0.2/24 -- 10.0.0.1/24
rpc_test 1 TINFO: fd00:1:1:1::2/64 -- fd00:1:1:1::1/64
rpc_test 1 TINFO: timeout per run is 0h 5m 0s
rpc_test 1 TINFO: check registered RPC with rpcinfo
rpc_test 1 TINFO: registered RPC:
program vers proto port service
100000 4 tcp 111 portmapper
100000 3 tcp 111 portmapper
100000 2 tcp 111 portmapper
100000 4 udp 111 portmapper
100000 3 udp 111 portmapper
100000 2 udp 111 portmapper
100001 1 udp 44050 rstatd
100001 2 udp 44050 rstatd
100001 3 udp 44050 rstatd
100001 4 udp 44050 rstatd
100001 5 udp 44050 rstatd
100002 2 udp 35568 rusersd
100002 3 udp 35568 rusersd
2000333 10 udp 51134
2000333 10 tcp 63711
rpc_test 1 TINFO: using libtirpc: yes
rpc_test 1 TFAIL: rpc_registerrpc 10.0.0.2 536875000 failed unexpectedly
rpc_registerrpc: rpc_reg: couldn't register prog 536875000 vers 1 for udp
rpc_registerrpc: rpc_reg: couldn't register prog 536875000 vers 1 for udp6
rpc_registerrpc: rpc_reg: cant find suitable transport for udp
1
Summary:
passed 0
failed 1
broken 0
skipped 0
warnings 0
<<<execution_status>>>
initiation_status="ok"
duration=1 termination_type=exited termination_id=1 corefile=no
cutime=8 cstime=44
<<<test_end>>>
To reproduce:
git clone https://github.com/intel/lkp-tests.git
cd lkp-tests
sudo bin/lkp install job.yaml # job file is attached in this email
bin/lkp split-job --compatible job.yaml # generate the yaml file for lkp run
sudo bin/lkp run generated-yaml-file
# if come across any failure that blocks the test,
# please remove ~/.lkp and /lkp dir to run from a clean state.
diff --git a/include/net/af_unix.h b/include/net/af_unix.h index acb56e463db1..b1748c9b6db2 100644 --- a/include/net/af_unix.h +++ b/include/net/af_unix.h @@ -22,7 +22,6 @@ struct sock *unix_peer_get(struct sock *sk); extern unsigned int unix_tot_inflight; extern spinlock_t unix_table_locks[UNIX_HASH_SIZE]; -extern struct hlist_head unix_socket_table[UNIX_HASH_SIZE]; struct unix_address { refcount_t refcnt; diff --git a/net/unix/af_unix.c b/net/unix/af_unix.c index 79f8fc5cdce8..9d0b07235dbc 100644 --- a/net/unix/af_unix.c +++ b/net/unix/af_unix.c @@ -120,8 +120,6 @@ spinlock_t unix_table_locks[UNIX_HASH_SIZE]; EXPORT_SYMBOL_GPL(unix_table_locks); -struct hlist_head unix_socket_table[UNIX_HASH_SIZE]; -EXPORT_SYMBOL_GPL(unix_socket_table); static atomic_long_t unix_nr_socks; /* SMP locking strategy: @@ -308,20 +306,20 @@ static void __unix_remove_socket(struct sock *sk) sk_del_node_init(sk); } -static void __unix_insert_socket(struct sock *sk) +static void __unix_insert_socket(struct net *net, struct sock *sk) { DEBUG_NET_WARN_ON_ONCE(!sk_unhashed(sk)); - sk_add_node(sk, &unix_socket_table[sk->sk_hash]); + sk_add_node(sk, &net->unx.table.buckets[sk->sk_hash]); } -static void __unix_set_addr_hash(struct sock *sk, struct unix_address *addr, - unsigned int hash) +static void __unix_set_addr_hash(struct net *net, struct sock *sk, + struct unix_address *addr, unsigned int hash) { __unix_remove_socket(sk); smp_store_release(&unix_sk(sk)->addr, addr); sk->sk_hash = hash; - __unix_insert_socket(sk); + __unix_insert_socket(net, sk); } static void unix_remove_socket(struct net *net, struct sock *sk) @@ -337,7 +335,7 @@ static void unix_insert_unbound_socket(struct net *net, struct sock *sk) { spin_lock(&unix_table_locks[sk->sk_hash]); spin_lock(&net->unx.table.locks[sk->sk_hash]); - __unix_insert_socket(sk); + __unix_insert_socket(net, sk); spin_unlock(&net->unx.table.locks[sk->sk_hash]); spin_unlock(&unix_table_locks[sk->sk_hash]); } @@ -348,12 +346,9 @@ static struct sock *__unix_find_socket_byname(struct net *net, { struct sock *s; - sk_for_each(s, &unix_socket_table[hash]) { + sk_for_each(s, &net->unx.table.buckets[hash]) { struct unix_sock *u = unix_sk(s); - if (!net_eq(sock_net(s), net)) - continue; - if (u->addr->len == len && !memcmp(u->addr->name, sunname, len)) return s; @@ -384,7 +379,7 @@ static struct sock *unix_find_socket_byinode(struct net *net, struct inode *i) spin_lock(&unix_table_locks[hash]); spin_lock(&net->unx.table.locks[hash]); - sk_for_each(s, &unix_socket_table[hash]) { + sk_for_each(s, &net->unx.table.buckets[hash]) { struct dentry *dentry = unix_sk(s)->path.dentry; if (dentry && d_backing_inode(dentry) == i) { @@ -1140,7 +1135,7 @@ static int unix_autobind(struct sock *sk) goto retry; } - __unix_set_addr_hash(sk, addr, new_hash); + __unix_set_addr_hash(net, sk, addr, new_hash); unix_table_double_unlock(net, old_hash, new_hash); err = 0; @@ -1199,7 +1194,7 @@ static int unix_bind_bsd(struct sock *sk, struct sockaddr_un *sunaddr, unix_table_double_lock(net, old_hash, new_hash); u->path.mnt = mntget(parent.mnt); u->path.dentry = dget(dentry); - __unix_set_addr_hash(sk, addr, new_hash); + __unix_set_addr_hash(net, sk, addr, new_hash); unix_table_double_unlock(net, old_hash, new_hash); mutex_unlock(&u->bindlock); done_path_create(&parent, dentry); @@ -1246,7 +1241,7 @@ static int unix_bind_abstract(struct sock *sk, struct sockaddr_un *sunaddr, if (__unix_find_socket_byname(net, addr->name, addr->len, new_hash)) goto out_spin; - __unix_set_addr_hash(sk, addr, new_hash); + __unix_set_addr_hash(net, sk, addr, new_hash); unix_table_double_unlock(net, old_hash, new_hash); mutex_unlock(&u->bindlock); return 0; @@ -3239,12 +3234,11 @@ static struct sock *unix_from_bucket(struct seq_file *seq, loff_t *pos) { unsigned long offset = get_offset(*pos); unsigned long bucket = get_bucket(*pos); - struct sock *sk; unsigned long count = 0; + struct sock *sk; - for (sk = sk_head(&unix_socket_table[bucket]); sk; sk = sk_next(sk)) { - if (sock_net(sk) != seq_file_net(seq)) - continue; + for (sk = sk_head(&seq_file_net(seq)->unx.table.buckets[bucket]); + sk; sk = sk_next(sk)) { if (++count == offset) break; } @@ -3279,13 +3273,13 @@ static struct sock *unix_get_next(struct seq_file *seq, struct sock *sk, loff_t *pos) { unsigned long bucket = get_bucket(*pos); - struct net *net = seq_file_net(seq); - for (sk = sk_next(sk); sk; sk = sk_next(sk)) - if (sock_net(sk) == net) - return sk; + sk = sk_next(sk); + if (sk) + return sk; + - spin_unlock(&net->unx.table.locks[bucket]); + spin_unlock(&seq_file_net(seq)->unx.table.locks[bucket]); spin_unlock(&unix_table_locks[bucket]); *pos = set_bucket_offset(++bucket, 1); @@ -3406,7 +3400,6 @@ static int bpf_iter_unix_hold_batch(struct seq_file *seq, struct sock *start_sk) { struct bpf_unix_iter_state *iter = seq->private; - struct net *net = seq_file_net(seq); unsigned int expected = 1; struct sock *sk; @@ -3414,9 +3407,6 @@ static int bpf_iter_unix_hold_batch(struct seq_file *seq, struct sock *start_sk) iter->batch[iter->end_sk++] = start_sk; for (sk = sk_next(start_sk); sk; sk = sk_next(sk)) { - if (sock_net(sk) != net) - continue; - if (iter->end_sk < iter->max_sk) { sock_hold(sk); iter->batch[iter->end_sk++] = sk; @@ -3425,7 +3415,7 @@ static int bpf_iter_unix_hold_batch(struct seq_file *seq, struct sock *start_sk) expected++; } - spin_unlock(&net->unx.table.locks[start_sk->sk_hash]); + spin_unlock(&seq_file_net(seq)->unx.table.locks[start_sk->sk_hash]); spin_unlock(&unix_table_locks[start_sk->sk_hash]); return expected; diff --git a/net/unix/diag.c b/net/unix/diag.c index 7fc377435114..4d0f0ca6a1eb 100644 --- a/net/unix/diag.c +++ b/net/unix/diag.c @@ -210,9 +210,7 @@ static int unix_diag_dump(struct sk_buff *skb, struct netlink_callback *cb) num = 0; spin_lock(&unix_table_locks[slot]); spin_lock(&net->unx.table.locks[slot]); - sk_for_each(sk, &unix_socket_table[slot]) { - if (!net_eq(sock_net(sk), net)) - continue; + sk_for_each(sk, &net->unx.table.buckets[slot]) { if (num < s_num) goto next; if (!(req->udiag_states & (1 << sk->sk_state))) @@ -246,13 +244,14 @@ static struct sock *unix_lookup_by_ino(struct net *net, unsigned int ino) for (i = 0; i < UNIX_HASH_SIZE; i++) { spin_lock(&unix_table_locks[i]); spin_lock(&net->unx.table.locks[i]); - sk_for_each(sk, &unix_socket_table[i]) + sk_for_each(sk, &net->unx.table.buckets[i]) { if (ino == sock_i_ino(sk)) { sock_hold(sk); spin_unlock(&net->unx.table.locks[i]); spin_unlock(&unix_table_locks[i]); return sk; } + } spin_unlock(&net->unx.table.locks[i]); spin_unlock(&unix_table_locks[i]); } @@ -277,8 +276,6 @@ static int unix_diag_get_exact(struct sk_buff *in_skb, err = -ENOENT; if (sk == NULL) goto out_nosk; - if (!net_eq(sock_net(sk), net)) - goto out; err = sock_diag_check_cookie(sk, req->udiag_cookie); if (err)
This commit replaces the global hash table with a per-netns one and removes the global one. We now link a socket in each netns's hash table so we can save some netns comparisons when iterating through a hash bucket. Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com> --- include/net/af_unix.h | 1 - net/unix/af_unix.c | 50 +++++++++++++++++-------------------------- net/unix/diag.c | 9 +++----- 3 files changed, 23 insertions(+), 37 deletions(-)