Re: [PATCH bpf-next v2 13/18] bpf: Bound program stack use by a per-program limit

["Kumar Kartikeya Dwivedi" <memxor@gmail.com>]

On Thu Sep 24, 2026 at 11:13 AM CEST, bot+bpf-ci wrote:
>> diff --git a/kernel/bpf/liveness.c b/kernel/bpf/liveness.c
>> index 2f52315d8bdd..de83736509a8 100644
>> --- a/kernel/bpf/liveness.c
>> +++ b/kernel/bpf/liveness.c
>> @@ -36,9 +36,9 @@ enum {
>>   * relative index @i is at &bits[(i * FM_MASK_CNT + kind) * words]. A
>>   * half-slot at or past @words * BITS_PER_LONG is never read by this frame,
>>   * hence never live. An instruction that may read the whole frame, such as a
>> - * call passing a frame pointer to another subprog, widens the array to
>> - * FRAME_MAX_WORDS, so that the read cannot lose half-slots to a later
>> - * widening.
>> + * call passing a frame pointer to another subprog, widens the array to the
>> + * program's stack budget, the deepest an accepted program can reach, so
>> + * that the read cannot lose half-slots to a later widening.
>>   */
>>  struct frame_masks {
>>  	u32 words;
>> @@ -264,12 +264,16 @@ static int mark_stack_write(struct func_instance *instance, u32 frame, u32 insn_
>>
>>  /*
>>   * Mark every half-slot of @frame as possibly read by @insn_idx. This widens
>> - * the masks to the maximum width: a full read recorded at a narrower width
>> - * would leave the bits added by a later widening clear and lose part of it.
>> + * the masks to the program's stack budget: a full read recorded at a narrower
>> + * width would leave the bits added by a later widening clear and lose part of
>> + * it, and an access past the budget is rejected by the main pass later, so no
>> + * widening of an accepted program goes further.
>>   */
>> -static int mark_stack_read_all(struct func_instance *instance, u32 frame, u32 insn_idx)
>> +static int mark_stack_read_all(struct bpf_verifier_env *env, struct func_instance *instance,
>> +			       u32 frame, u32 insn_idx)
>>  {
>> -	return mark_stack_read(instance, frame, insn_idx, 0, FRAME_HALF_SPIS - 1);
>> +	return mark_stack_read(instance, frame, insn_idx, 0,
>> +			       env->stack_limit / BPF_HALF_REG_SIZE - 1);
>>  }
>
> The new comment on mark_stack_read_all() says "an access past the budget
> is rejected by the main pass later, so no widening of an accepted program
> goes further." Can that claim hold in all cases?
>
> bpf_compute_subprog_arg_access() records accesses on every instruction
> the CFG can reach, but do_check() only explores feasible paths. In
> check_cond_jmp_op(), when is_branch_taken() returns 0 or 1, the other
> branch is pushed only if !env->bypass_spec_v1.
>
> For a privileged load, a branch that is provably dead but not
> constant-folded is never verified. Looking at something like this:
>
>     call bpf_get_prandom_u32
>     r0 &= 15
>     if r0 > 20 goto L
>     r0 = 0
>     exit
>   L:
>     *(u64 *)(r10 - 1024) = r0
>
> The program is accepted with stack_limit == 512. The liveness pass
> still calls mark_stack_write() for half-slots 254..255, and
> widen_frame_masks() grows that frame past the budget, to 4 words.
>
> There is no functional impact. widen_frame_masks() keeps all old bits,
> and liveness is only queried from __clean_func_state() for half-slots
> below 2 * allocated_stack / BPF_REG_SIZE. allocated_stack only grows
> through grow_stack_state() after check_stack_access_within_bounds() has
> enforced -env->stack_limit. So bits past the budget are never read,
> and a read-all up to stack_limit / BPF_HALF_REG_SIZE - 1 covers every
> half-slot the verifier can query.
>
> Could the comments state that invariant (verifier states never hold slots
> past env->stack_limit), rather than claiming that accepted programs never
> record accesses past the budget? As written, a later change could rely on
> masks never being wider than the budget, for example to size a buffer from
> stack_limit instead of FRAME_MAX_WORDS, and that would be unsafe.
>

Reworded for v3: the comment now states the invariant that verifier states never
hold slots past the budget, and notes that unreachable code may still widen the
masks harmlessly.

>
> ---
> AI reviewed your patch. Please fix the bug or email reply why it's not a bug.
> See: https://github.com/kernel-patches/vmtest/blob/master/ci/claude/README.md
>
> CI run summary: https://github.com/kernel-patches/bpf/actions/runs/35976322553