On 03/02/2019 12:18 AM, Song Liu wrote:
> On Fri, Mar 1, 2019 at 1:06 PM Daniel Borkmann <daniel@xxxxxxxxxxxxx> wrote:
>>
>> Marek reported that he saw an issue with the below snippet in that
>> timing measurements where off when loaded as unpriv while results
>> were reasonable when loaded as privileged:
>>
>> [...]
>> uint64_t a = bpf_ktime_get_ns();
>> uint64_t b = bpf_ktime_get_ns();
>> uint64_t delta = b - a;
>> if ((int64_t)delta > 0) {
>> [...]
>>
>> Turns out there is a bug where a corner case is missing in the fix
>> d3bd7413e0ca ("bpf: fix sanitation of alu op with pointer / scalar
>> type from different paths"), namely fixup_bpf_calls() only checks
>> whether aux has a non-zero alu_state, but it also needs to test for
>> the case of BPF_ALU_NON_POINTER since in both occasions we need to
>> skip the masking rewrite (as there is nothing to mask).
>>
>> Fixes: d3bd7413e0ca ("bpf: fix sanitation of alu op with pointer / scalar type from different paths")
>> Reported-by: Marek Majkowski <marek@xxxxxxxxxxxxxx>
>> Reported-by: Arthur Fabre <afabre@xxxxxxxxxxxxxx>
>> Signed-off-by: Daniel Borkmann <daniel@xxxxxxxxxxxxx>
>> Link: https://lore.kernel.org/netdev/CAJPywTJqP34cK20iLM5YmUMz9KXQOdu1-+BZrGMAGgLuBWz7fg@xxxxxxxxxxxxxx/T/
>> ---
>> [ Test case will be routed via bpf-next to avoid useless merge churn
>> due to test_verifier rework in bpf-next. ]
>>
>> kernel/bpf/verifier.c | 3 ++-
>> 1 file changed, 2 insertions(+), 1 deletion(-)
>>
>> diff --git a/kernel/bpf/verifier.c b/kernel/bpf/verifier.c
>> index 8f295b790297..5fcce2f4209d 100644
>> --- a/kernel/bpf/verifier.c
>> +++ b/kernel/bpf/verifier.c
>> @@ -6920,7 +6920,8 @@ static int fixup_bpf_calls(struct bpf_verifier_env *env)
>> u32 off_reg;
>>
>> aux = &env->insn_aux_data[i + delta];
>> - if (!aux->alu_state)
>> + if (!aux->alu_state ||
>> + aux->alu_state == BPF_ALU_NON_POINTER)
>
> alu_state is a bitmap. Shall we check "aux->alu_state &
> BPF_ALU_NON_POINTER" here?
The state in this case can only ever be BPF_ALU_NON_POINTER, any other
setting from sanitize_val_alu() would be a violation.
