diff options
| author | Daniel Borkmann <daniel@iogearbox.net> | 2021-03-26 02:41:34 +0300 |
|---|---|---|
| committer | Daniel Borkmann <daniel@iogearbox.net> | 2021-03-26 02:46:33 +0300 |
| commit | 80847a71b270b172814dc7562f3eb59507fca61e (patch) | |
| tree | 97200006ffd0bc26939823b4c5624a612384e2dd /kernel | |
| parent | 241949e488f38a192f2359dbb21d80e08173eb60 (diff) | |
| download | linux-80847a71b270b172814dc7562f3eb59507fca61e.tar.xz | |
bpf: Undo ptr_to_map_key alu sanitation for now
Remove PTR_TO_MAP_KEY for the time being from being sanitized on pointer ALU
through sanitize_ptr_alu() mainly for 3 reasons:
1) It's currently unused and not available from unprivileged. However that by
itself is not yet a strong reason to drop the code.
2) Commit 69c087ba6225 ("bpf: Add bpf_for_each_map_elem() helper") implemented
the sanitation not fully correct in that unlike stack or map_value pointer
it doesn't probe whether the access to the map key /after/ the simulated ALU
operation is still in bounds. This means that the generated mask can truncate
the offset in the non-speculative domain whereas it should only truncate in
the speculative domain. The verifier should instead reject such program as
we do for other types.
3) Given the recent fixes from f232326f6966 ("bpf: Prohibit alu ops for pointer
types not defining ptr_limit"), 10d2bb2e6b1d ("bpf: Fix off-by-one for area
size in creating mask to left"), b5871dca250c ("bpf: Simplify alu_limit masking
for pointer arithmetic") as well as 1b1597e64e1a ("bpf: Add sanity check for
upper ptr_limit") the code changed quite a bit and the merge in efd13b71a3fa
broke the PTR_TO_MAP_KEY case due to an incorrect merge conflict.
Remove the relevant pieces for the time being and we can rework the PTR_TO_MAP_KEY
case once everything settles.
Fixes: efd13b71a3fa ("Merge git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net")
Fixes: 69c087ba6225 ("bpf: Add bpf_for_each_map_elem() helper")
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Diffstat (limited to 'kernel')
| -rw-r--r-- | kernel/bpf/verifier.c | 14 |
1 files changed, 0 insertions, 14 deletions
diff --git a/kernel/bpf/verifier.c b/kernel/bpf/verifier.c index 210169c25ead..85f9f842d15c 100644 --- a/kernel/bpf/verifier.c +++ b/kernel/bpf/verifier.c @@ -6062,19 +6062,6 @@ static int retrieve_ptr_limit(const struct bpf_reg_state *ptr_reg, else *ptr_limit = -off - 1; return *ptr_limit >= max ? -ERANGE : 0; - case PTR_TO_MAP_KEY: - /* Currently, this code is not exercised as the only use - * is bpf_for_each_map_elem() helper which requires - * bpf_capble. The code has been tested manually for - * future use. - */ - if (mask_to_left) { - *ptr_limit = ptr_reg->umax_value + ptr_reg->off; - } else { - off = ptr_reg->smin_value + ptr_reg->off; - *ptr_limit = ptr_reg->map_ptr->key_size - off; - } - return 0; case PTR_TO_MAP_VALUE: max = ptr_reg->map_ptr->value_size; if (mask_to_left) { @@ -6281,7 +6268,6 @@ static int adjust_ptr_min_max_vals(struct bpf_verifier_env *env, verbose(env, "R%d pointer arithmetic on %s prohibited\n", dst, reg_type_str[ptr_reg->type]); return -EACCES; - case PTR_TO_MAP_KEY: case PTR_TO_MAP_VALUE: if (!env->allow_ptr_leaks && !known && (smin_val < 0) != (smax_val < 0)) { verbose(env, "R%d has unknown scalar with mixed signed bounds, pointer arithmetic with it prohibited for !root\n", |