This series improves error propagation in cpumap and adds selftests that cover the failure cases.
Currently, failures returned from __cpu_map_entry_alloc() are ignored and always converted to -ENOMEM by cpu_map_update_elem(). This series ensures the correct error propagation and adds selftests.
Changes: v2: - send to bpf-next, not to bpf - drop Fixes: tag v1: https://lore.kernel.org/bpf/20251128160504.57844-1-enjuk@amazon.com/
Kohei Enju (2): bpf: cpumap: propagate underlying error in cpu_map_update_elem() selftests/bpf: add tests for attaching invalid fd
kernel/bpf/cpumap.c | 21 ++++++++++++------- .../bpf/prog_tests/xdp_cpumap_attach.c | 19 +++++++++++++++-- 2 files changed, 30 insertions(+), 10 deletions(-)
After commit 9216477449f3 ("bpf: cpumap: Add the possibility to attach an eBPF program to cpumap"), __cpu_map_entry_alloc() may fail with errors other than -ENOMEM, such as -EBADF or -EINVAL.
However, __cpu_map_entry_alloc() returns NULL on all failures, and cpu_map_update_elem() unconditionally converts this NULL into -ENOMEM. As a result, user space always receives -ENOMEM regardless of the actual underlying error.
Examples of unexpected behavior: - Nonexistent fd : -ENOMEM (should be -EBADF) - Non-BPF fd : -ENOMEM (should be -EINVAL) - Bad attach type : -ENOMEM (should be -EINVAL)
Change __cpu_map_entry_alloc() to return ERR_PTR(err) instead of NULL and have cpu_map_update_elem() propagate this error.
Signed-off-by: Kohei Enju enjuk@amazon.com --- kernel/bpf/cpumap.c | 21 +++++++++++++-------- 1 file changed, 13 insertions(+), 8 deletions(-)
diff --git a/kernel/bpf/cpumap.c b/kernel/bpf/cpumap.c index 703e5df1f4ef..04171fbc39cb 100644 --- a/kernel/bpf/cpumap.c +++ b/kernel/bpf/cpumap.c @@ -430,7 +430,7 @@ static struct bpf_cpu_map_entry * __cpu_map_entry_alloc(struct bpf_map *map, struct bpf_cpumap_val *value, u32 cpu) { - int numa, err, i, fd = value->bpf_prog.fd; + int numa, err = -ENOMEM, i, fd = value->bpf_prog.fd; gfp_t gfp = GFP_KERNEL | __GFP_NOWARN; struct bpf_cpu_map_entry *rcpu; struct xdp_bulk_queue *bq; @@ -440,7 +440,7 @@ __cpu_map_entry_alloc(struct bpf_map *map, struct bpf_cpumap_val *value,
rcpu = bpf_map_kmalloc_node(map, sizeof(*rcpu), gfp | __GFP_ZERO, numa); if (!rcpu) - return NULL; + return ERR_PTR(err);
/* Alloc percpu bulkq */ rcpu->bulkq = bpf_map_alloc_percpu(map, sizeof(*rcpu->bulkq), @@ -468,16 +468,21 @@ __cpu_map_entry_alloc(struct bpf_map *map, struct bpf_cpumap_val *value, rcpu->value.qsize = value->qsize; gro_init(&rcpu->gro);
- if (fd > 0 && __cpu_map_load_bpf_program(rcpu, map, fd)) - goto free_ptr_ring; + if (fd > 0) { + err = __cpu_map_load_bpf_program(rcpu, map, fd); + if (err) + goto free_ptr_ring; + }
/* Setup kthread */ init_completion(&rcpu->kthread_running); rcpu->kthread = kthread_create_on_node(cpu_map_kthread_run, rcpu, numa, "cpumap/%d/map:%d", cpu, map->id); - if (IS_ERR(rcpu->kthread)) + if (IS_ERR(rcpu->kthread)) { + err = PTR_ERR(rcpu->kthread); goto free_prog; + }
/* Make sure kthread runs on a single CPU */ kthread_bind(rcpu->kthread, cpu); @@ -503,7 +508,7 @@ __cpu_map_entry_alloc(struct bpf_map *map, struct bpf_cpumap_val *value, free_percpu(rcpu->bulkq); free_rcu: kfree(rcpu); - return NULL; + return ERR_PTR(err); }
static void __cpu_map_entry_free(struct work_struct *work) @@ -596,8 +601,8 @@ static long cpu_map_update_elem(struct bpf_map *map, void *key, void *value, } else { /* Updating qsize cause re-allocation of bpf_cpu_map_entry */ rcpu = __cpu_map_entry_alloc(map, &cpumap_value, key_cpu); - if (!rcpu) - return -ENOMEM; + if (IS_ERR(rcpu)) + return PTR_ERR(rcpu); } rcu_read_lock(); __cpu_map_entry_replace(cmap, key_cpu, rcpu);
Add test cases for situations where adding the following types of file descriptors to a cpumap entry should fail: - Non-BPF file descriptor (expect -EINVAL) - Nonexistent file descriptor (expect -EBADF)
Also tighten the assertion for the expected error when adding a non-BPF_XDP_CPUMAP program to a cpumap entry.
Signed-off-by: Kohei Enju enjuk@amazon.com --- .../bpf/prog_tests/xdp_cpumap_attach.c | 19 +++++++++++++++++-- 1 file changed, 17 insertions(+), 2 deletions(-)
diff --git a/tools/testing/selftests/bpf/prog_tests/xdp_cpumap_attach.c b/tools/testing/selftests/bpf/prog_tests/xdp_cpumap_attach.c index df27535995af..ad56e4370ce3 100644 --- a/tools/testing/selftests/bpf/prog_tests/xdp_cpumap_attach.c +++ b/tools/testing/selftests/bpf/prog_tests/xdp_cpumap_attach.c @@ -18,7 +18,7 @@ static void test_xdp_with_cpumap_helpers(void) struct bpf_cpumap_val val = { .qsize = 192, }; - int err, prog_fd, prog_redir_fd, map_fd; + int err, prog_fd, prog_redir_fd, map_fd, bad_fd; struct nstoken *nstoken = NULL; __u32 idx = 0;
@@ -79,7 +79,22 @@ static void test_xdp_with_cpumap_helpers(void) val.qsize = 192; val.bpf_prog.fd = bpf_program__fd(skel->progs.xdp_dummy_prog); err = bpf_map_update_elem(map_fd, &idx, &val, 0); - ASSERT_NEQ(err, 0, "Add non-BPF_XDP_CPUMAP program to cpumap entry"); + ASSERT_EQ(err, -EINVAL, "Add non-BPF_XDP_CPUMAP program to cpumap entry"); + + /* Try to attach non-BPF file descriptor */ + bad_fd = open("/dev/null", O_RDONLY); + ASSERT_GE(bad_fd, 0, "Open /dev/null for non-BPF fd"); + + val.bpf_prog.fd = bad_fd; + err = bpf_map_update_elem(map_fd, &idx, &val, 0); + ASSERT_EQ(err, -EINVAL, "Add non-BPF fd to cpumap entry"); + + /* Try to attach nonexistent file descriptor */ + err = close(bad_fd); + ASSERT_EQ(err, 0, "Close non-BPF fd for nonexistent fd"); + + err = bpf_map_update_elem(map_fd, &idx, &val, 0); + ASSERT_EQ(err, -EBADF, "Add nonexistent fd to cpumap entry");
/* Try to attach BPF_XDP program with frags to cpumap when we have * already loaded a BPF_XDP program on the map
Kohei Enju enjuk@amazon.com writes:
Add test cases for situations where adding the following types of file descriptors to a cpumap entry should fail:
- Non-BPF file descriptor (expect -EINVAL)
- Nonexistent file descriptor (expect -EBADF)
Also tighten the assertion for the expected error when adding a non-BPF_XDP_CPUMAP program to a cpumap entry.
Signed-off-by: Kohei Enju enjuk@amazon.com
Reviewed-by: Toke Høiland-Jørgensen toke@redhat.com
Kohei Enju enjuk@amazon.com writes:
After commit 9216477449f3 ("bpf: cpumap: Add the possibility to attach an eBPF program to cpumap"), __cpu_map_entry_alloc() may fail with errors other than -ENOMEM, such as -EBADF or -EINVAL.
However, __cpu_map_entry_alloc() returns NULL on all failures, and cpu_map_update_elem() unconditionally converts this NULL into -ENOMEM. As a result, user space always receives -ENOMEM regardless of the actual underlying error.
Examples of unexpected behavior:
- Nonexistent fd : -ENOMEM (should be -EBADF)
- Non-BPF fd : -ENOMEM (should be -EINVAL)
- Bad attach type : -ENOMEM (should be -EINVAL)
Change __cpu_map_entry_alloc() to return ERR_PTR(err) instead of NULL and have cpu_map_update_elem() propagate this error.
Signed-off-by: Kohei Enju enjuk@amazon.com
Reviewed-by: Toke Høiland-Jørgensen toke@redhat.com
linux-kselftest-mirror@lists.linaro.org
-
Kohei Enju -
Toke Høiland-Jørgensen