BMC/Intel-BMC/linux.git/kernel/bpf/lpm_trie.c, branch dev-4.17

bpf: fix rcu lockdep warning for lpm_trie map_free callback

2018-02-22T20:29:12+00:00

Commit 9a3efb6b661f ("bpf: fix memory leak in lpm_trie map_free callback function") fixed a memory leak and removed unnecessary locks in map_free callback function. Unfortrunately, it introduced a lockdep warning. When lockdep checking is turned on, running tools/testing/selftests/bpf/test_lpm_map will have: [ 98.294321] ============================= [ 98.294807] WARNING: suspicious RCU usage [ 98.295359] 4.16.0-rc2+ #193 Not tainted [ 98.295907] ----------------------------- [ 98.296486] /home/yhs/work/bpf/kernel/bpf/lpm_trie.c:572 suspicious rcu_dereference_check() usage! [ 98.297657] [ 98.297657] other info that might help us debug this: [ 98.297657] [ 98.298663] [ 98.298663] rcu_scheduler_active = 2, debug_locks = 1 [ 98.299536] 2 locks held by kworker/2:1/54: [ 98.300152] #0: ((wq_completion)"events"){+.+.}, at: [<00000000196bc1f0>] process_one_work+0x157/0x5c0 [ 98.301381] #1: ((work_completion)(&map->work)){+.+.}, at: [<00000000196bc1f0>] process_one_work+0x157/0x5c0 Since actual trie tree removal happens only after no other accesses to the tree are possible, replacing rcu_dereference_protected(*slot, lockdep_is_held(&trie->lock)) with rcu_dereference_protected(*slot, 1) fixed the issue. Fixes: 9a3efb6b661f ("bpf: fix memory leak in lpm_trie map_free callback function") Reported-by: Eric Dumazet Suggested-by: Eric Dumazet Signed-off-by: Yonghong Song Reviewed-by: Eric Dumazet Acked-by: David S. Miller Signed-off-by: Daniel Borkmann

bpf: fix memory leak in lpm_trie map_free callback function

2018-02-14T03:15:16+00:00

There is a memory leak happening in lpm_trie map_free callback function trie_free. The trie structure itself does not get freed. Also, trie_free function did not do synchronize_rcu before freeing various data structures. This is incorrect as some rcu_read_lock region(s) for lookup, update, delete or get_next_key may not complete yet. The fix is to add synchronize_rcu in the beginning of trie_free. The useless spin_lock is removed from this function as well. Fixes: b95a5c4db09b ("bpf: add a longest prefix match trie map implementation") Reported-by: Mathieu Malaterre Reported-by: Alexei Starovoitov Tested-by: Mathieu Malaterre Signed-off-by: Yonghong Song Signed-off-by: Alexei Starovoitov

bpf: fix kernel page fault in lpm map trie_get_next_key

2018-01-27T01:06:22+00:00

Commit b471f2f1de8b ("bpf: implement MAP_GET_NEXT_KEY command for LPM_TRIE map") introduces a bug likes below: if (!rcu_dereference(trie->root)) return -ENOENT; if (!key || key->prefixlen > trie->max_prefixlen) { root = &trie->root; goto find_leftmost; } ...... find_leftmost: for (node = rcu_dereference(*root); node;) { In the code after label find_leftmost, it is assumed that *root should not be NULL, but it is not true as it is possbile trie->root is changed to NULL by an asynchronous delete operation. The issue is reported by syzbot and Eric Dumazet with the below error log: ...... kasan: CONFIG_KASAN_INLINE enabled kasan: GPF could be caused by NULL-ptr deref or user memory access general protection fault: 0000 [#1] SMP KASAN Dumping ftrace buffer: (ftrace buffer empty) Modules linked in: CPU: 1 PID: 8033 Comm: syz-executor3 Not tainted 4.15.0-rc8+ #4 Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011 RIP: 0010:trie_get_next_key+0x3c2/0xf10 kernel/bpf/lpm_trie.c:682 ...... This patch fixed the issue by use local rcu_dereferenced pointer instead of *(&trie->root) later on. Fixes: b471f2f1de8b ("bpf: implement MAP_GET_NEXT_KEY command or LPM_TRIE map") Reported-by: syzbot Reported-by: Eric Dumazet Signed-off-by: Yonghong Song Signed-off-by: Alexei Starovoitov

bpf: fix incorrect kmalloc usage in lpm_trie MAP_GET_NEXT_KEY rcu region

2018-01-23T16:33:57+00:00

In commit b471f2f1de8b ("bpf: implement MAP_GET_NEXT_KEY command for LPM_TRIE map"), the implemented MAP_GET_NEXT_KEY callback function is guarded with rcu read lock. In the function body, "kmalloc(size, GFP_USER | __GFP_NOWARN)" is used which may sleep and violate rcu read lock region requirements. This patch fixed the issue by using GFP_ATOMIC instead to avoid blocking kmalloc. Tested with CONFIG_DEBUG_ATOMIC_SLEEP=y as suggested by Eric Dumazet. Fixes: b471f2f1de8b ("bpf: implement MAP_GET_NEXT_KEY command for LPM_TRIE map") Signed-off-by: Yonghong Song Reported-by: syzbot Reviewed-by: Eric Dumazet Signed-off-by: Daniel Borkmann

bpf: implement MAP_GET_NEXT_KEY command for LPM_TRIE map

2018-01-19T22:26:41+00:00

Current LPM_TRIE map type does not implement MAP_GET_NEXT_KEY command. This command is handy when users want to enumerate keys. Otherwise, a different map which supports key enumeration may be required to store the keys. If the map data is sparse and all map data are to be deleted without closing file descriptor, using MAP_GET_NEXT_KEY to find all keys is much faster than enumerating all key space. This patch implements MAP_GET_NEXT_KEY command for LPM_TRIE map. If user provided key pointer is NULL or the key does not have an exact match in the trie, the first key will be returned. Otherwise, the next key will be returned. In this implemenation, key enumeration follows a postorder traversal of internal trie. More specific keys will be returned first than less specific ones, given a sequence of MAP_GET_NEXT_KEY syscalls. Signed-off-by: Yonghong Song Signed-off-by: Daniel Borkmann

bpf: add helper for copying attrs to struct bpf_map

2018-01-14T22:36:29+00:00

All map types reimplement the field-by-field copy of union bpf_attr members into struct bpf_map. Add a helper to perform this operation. Signed-off-by: Jakub Kicinski Reviewed-by: Quentin Monnet Acked-by: Alexei Starovoitov Signed-off-by: Daniel Borkmann

bpf: Add file mode configuration into bpf maps

2017-10-20T12:32:59+00:00

Introduce the map read/write flags to the eBPF syscalls that returns the map fd. The flags is used to set up the file mode when construct a new file descriptor for bpf maps. To not break the backward capability, the f_flags is set to O_RDWR if the flag passed by syscall is 0. Otherwise it should be O_RDONLY or O_WRONLY. When the userspace want to modify or read the map content, it will check the file mode to see if it is allowed to make the change. Signed-off-by: Chenbo Feng Acked-by: Alexei Starovoitov Acked-by: Daniel Borkmann Signed-off-by: David S. Miller

bpf: Optimize lpm trie delete

2017-09-25T21:37:54+00:00

Before the delete operator was added, this datastructure maintained an invariant that intermediate nodes were only present when necessary to build the tree. This patch updates the delete operation to reinstate that invariant by removing unnecessary intermediate nodes after a node is removed and thus keeping the tree structure at a minimal size. Suggested-by: Daniel Mack Signed-off-by: Craig Gallek Signed-off-by: David S. Miller

bpf: Implement map_delete_elem for BPF_MAP_TYPE_LPM_TRIE

2017-09-19T20:55:15+00:00

This is a simple non-recursive delete operation. It prunes paths of empty nodes in the tree, but it does not try to further compress the tree as nodes are removed. Signed-off-by: Craig Gallek Acked-by: Daniel Borkmann Signed-off-by: David S. Miller

bpf: Allow selecting numa node during map creation

2017-08-20T04:35:43+00:00

The current map creation API does not allow to provide the numa-node preference. The memory usually comes from where the map-creation-process is running. The performance is not ideal if the bpf_prog is known to always run in a numa node different from the map-creation-process. One of the use case is sharding on CPU to different LRU maps (i.e. an array of LRU maps). Here is the test result of map_perf_test on the INNER_LRU_HASH_PREALLOC test if we force the lru map used by CPU0 to be allocated from a remote numa node: [ The machine has 20 cores. CPU0-9 at node 0. CPU10-19 at node 1 ] ># taskset -c 10 ./map_perf_test 512 8 1260000 8000000 5:inner_lru_hash_map_perf pre-alloc 1628380 events per sec 4:inner_lru_hash_map_perf pre-alloc 1626396 events per sec 3:inner_lru_hash_map_perf pre-alloc 1626144 events per sec 6:inner_lru_hash_map_perf pre-alloc 1621657 events per sec 2:inner_lru_hash_map_perf pre-alloc 1621534 events per sec 1:inner_lru_hash_map_perf pre-alloc 1620292 events per sec 7:inner_lru_hash_map_perf pre-alloc 1613305 events per sec 0:inner_lru_hash_map_perf pre-alloc 1239150 events per sec #<<< After specifying numa node: ># taskset -c 10 ./map_perf_test 512 8 1260000 8000000 5:inner_lru_hash_map_perf pre-alloc 1629627 events per sec 3:inner_lru_hash_map_perf pre-alloc 1628057 events per sec 1:inner_lru_hash_map_perf pre-alloc 1623054 events per sec 6:inner_lru_hash_map_perf pre-alloc 1616033 events per sec 2:inner_lru_hash_map_perf pre-alloc 1614630 events per sec 4:inner_lru_hash_map_perf pre-alloc 1612651 events per sec 7:inner_lru_hash_map_perf pre-alloc 1609337 events per sec 0:inner_lru_hash_map_perf pre-alloc 1619340 events per sec #<<< This patch adds one field, numa_node, to the bpf_attr. Since numa node 0 is a valid node, a new flag BPF_F_NUMA_NODE is also added. The numa_node field is honored if and only if the BPF_F_NUMA_NODE flag is set. Numa node selection is not supported for percpu map. This patch does not change all the kmalloc. F.e. 'htab = kzalloc()' is not changed since the object is small enough to stay in the cache. Signed-off-by: Martin KaFai Lau Acked-by: Daniel Borkmann Acked-by: Alexei Starovoitov Signed-off-by: David S. Miller