From 0b4ef9d9d9ff1191d645de880ca144a98605df20 Mon Sep 17 00:00:00 2001 From: Yonghong Song Date: Thu, 21 Mar 2024 23:13:53 -0700 Subject: [PATCH] selftests/bpf: Fix flaky test btf_map_in_map/lookup_update [ Upstream commit 14bb1e8c8d4ad5d9d2febb7d19c70a3cf536e1e5 ] Recently, I frequently hit the following test failure: [root@arch-fb-vm1 bpf]# ./test_progs -n 33/1 test_lookup_update:PASS:skel_open 0 nsec [...] test_lookup_update:PASS:sync_rcu 0 nsec test_lookup_update:FAIL:map1_leak inner_map1 leaked! #33/1 btf_map_in_map/lookup_update:FAIL #33 btf_map_in_map:FAIL In the test, after map is closed and then after two rcu grace periods, it is assumed that map_id is not available to user space. But the above assumption cannot be guaranteed. After zero or one or two rcu grace periods in different siturations, the actual freeing-map-work is put into a workqueue. Later on, when the work is dequeued, the map will be actually freed. See bpf_map_put() in kernel/bpf/syscall.c. By using workqueue, there is no ganrantee that map will be actually freed after a couple of rcu grace periods. This patch removed such map leak detection and then the test can pass consistently. Signed-off-by: Yonghong Song Signed-off-by: Daniel Borkmann Link: https://lore.kernel.org/bpf/20240322061353.632136-1-yonghong.song@linux.dev Signed-off-by: Sasha Levin --- .../selftests/bpf/prog_tests/btf_map_in_map.c | 26 +------------------ 1 file changed, 1 insertion(+), 25 deletions(-) diff --git a/tools/testing/selftests/bpf/prog_tests/btf_map_in_map.c b/tools/testing/selftests/bpf/prog_tests/btf_map_in_map.c index eb90a6b8850d2f..f4d753185001af 100644 --- a/tools/testing/selftests/bpf/prog_tests/btf_map_in_map.c +++ b/tools/testing/selftests/bpf/prog_tests/btf_map_in_map.c @@ -25,7 +25,7 @@ static void test_lookup_update(void) int map1_fd, map2_fd, map3_fd, map4_fd, map5_fd, map1_id, map2_id; int outer_arr_fd, outer_hash_fd, outer_arr_dyn_fd; struct test_btf_map_in_map *skel; - int err, key = 0, val, i, fd; + int err, key = 0, val, i; skel = test_btf_map_in_map__open_and_load(); if (CHECK(!skel, "skel_open", "failed to open&load skeleton\n")) @@ -102,30 +102,6 @@ static void test_lookup_update(void) CHECK(map1_id == 0, "map1_id", "failed to get ID 1\n"); CHECK(map2_id == 0, "map2_id", "failed to get ID 2\n"); - test_btf_map_in_map__destroy(skel); - skel = NULL; - - /* we need to either wait for or force synchronize_rcu(), before - * checking for "still exists" condition, otherwise map could still be - * resolvable by ID, causing false positives. - * - * Older kernels (5.8 and earlier) freed map only after two - * synchronize_rcu()s, so trigger two, to be entirely sure. - */ - CHECK(kern_sync_rcu(), "sync_rcu", "failed\n"); - CHECK(kern_sync_rcu(), "sync_rcu", "failed\n"); - - fd = bpf_map_get_fd_by_id(map1_id); - if (CHECK(fd >= 0, "map1_leak", "inner_map1 leaked!\n")) { - close(fd); - goto cleanup; - } - fd = bpf_map_get_fd_by_id(map2_id); - if (CHECK(fd >= 0, "map2_leak", "inner_map2 leaked!\n")) { - close(fd); - goto cleanup; - } - cleanup: test_btf_map_in_map__destroy(skel); }