Skip to content

bpf: Copy per-CPU map value padding in copy_map_value_long()

Leon Hwang bpf linux-kselftest lkml 1 version Applied

Status from this message: This series was applied to bpf/bpf-next.git (master)

Get series with b4
b4 am [email protected]

Versions newest first

  1. v1 Latest2 patches 0 repliesApplied 2/2

Patches in v1 All 2 applied

#Subject and filesSizeRepliesApplied
1/2bpf: Copy per-CPU map value padding in copy_map_value_long()include/linux/bpf.h+2 -20 repliesApplied
2/2selftests/bpf: Verify no non-zeroed kernel heap memory exposuretools/testing/selftests/bpf/prog_tests/test_map_uninit.c tools/testing/selftests/bpf/progs/map_kptr.c+80 -00 repliesApplied

Reviews trailers per patch

No Reviewed-by, Acked-by or Tested-by trailers for v1 yet.

Read the v1 thread

Cover letter [PATCH bpf 0/2] bpf: Copy per-CPU map value padding in copy_map_value_long()

Sashiko reported [1]:

This is a pre-existing issue, but does iterating over per-CPU maps expose uninitialized kernel heap memory?

When working with per-CPU maps, temporary buffers are allocated using kmalloc without the __GFP_ZERO flag in functions like bpf_iter_init_array_map in kernel/bpf/arraymap.c:

kernel/bpf/arraymap.c:bpf_iter_init_array_map() {
    ...
    value_buf = kmalloc(buf_size, GFP_USER | __GFP_NOWARN);
    ...
}

This is also done in kernel/bpf/hashtab.c:bpf_iter_init_hash_map().

If the map contains a BTF record, bpf_obj_memcpy in include/linux/bpf.h explicitly stops at map->value_size instead of filling the entire rounded-up size:

include/linux/bpf.h:bpf_obj_memcpy() {
    ...
    memcpy(dst + curr_off, src + curr_off, size - curr_off);
    ...
}

This fails to overwrite the padding bytes up to round_up(map->value_size, 8).

[1] https://lore.kernel.org/bpf/[email protected]/

===
For example,

struct map_uninit_value {
	struct prog_test_ref_kfunc __kptr_untrusted *unref_ptr;
	__u32 data;
} __attribute__((packed));
struct {
	__uint(type, BPF_MAP_TYPE_PERCPU_ARRAY);
	__type(key, int);
	__type(value, struct map_uninit_value);
	__uint(max_entries, 1);
} pcpu_array SEC(".maps");

There are 4 padding bytes in the kernel percpu_array map elements.

When lookup element from 'pcpu_array' map, for each CPU, the 4 padding bytes memory allocated by syscall.c::map_lookup_elem():kvmalloc() could be exposed to user space.

Without the fix, the selftest could fail with:

test_map_uninit_mem_exposure:FAIL:zeroed tail bytes unexpected memory mismatch
actual:

	2B 2B 2B 2B
expected:
	00 00 00 00

Leon Hwang (2):
bpf: Copy per-CPU map value padding in copy_map_value_long() selftests/bpf: Verify no non-zeroed kernel heap memory exposure

 include/linux/bpf.h                           |  4 +-
 .../bpf/prog_tests/test_map_uninit.c          | 68 +++++++++++++++++++
 tools/testing/selftests/bpf/progs/map_kptr.c  | 12 ++++
 3 files changed, 82 insertions(+), 2 deletions(-)
 create mode 100644 tools/testing/selftests/bpf/prog_tests/test_map_uninit.c
Signature · 1 line
2.54.0