Re: [PATCH bpf-next v1 4/6] mm: introduce BPF kfuncs to access memcg statistics and events

[Roman Gushchin <roman.gushchin@linux.dev>]

bot+bpf-ci@kernel.org writes:

>> diff --git a/mm/bpf_memcontrol.c b/mm/bpf_memcontrol.c
>> index 6d0d73bf0dd1..4d9d7d909f6c 100644
>> --- a/mm/bpf_memcontrol.c
>> +++ b/mm/bpf_memcontrol.c
>
> [ ... ]
>
>> +/**
>> + * bpf_mem_cgroup_vm_events - Read memory cgroup's vm event counter
>> + * @memcg: memory cgroup
>> + * @event: event id
>> + *
>> + * Allows to read memory cgroup event counters.
>> + */
>> +__bpf_kfunc unsigned long bpf_mem_cgroup_vm_events(struct mem_cgroup *memcg,
>> +						enum vm_event_item event)
>> +{
>> +	return memcg_events(memcg, event);
>> +}
>
> Should bpf_mem_cgroup_vm_events() validate the event parameter before
> passing it to memcg_events()? The call chain is:
>
>   bpf_mem_cgroup_vm_events() -> memcg_events() -> memcg_events_index()
>
> And memcg_events_index() does:
>
>   return mem_cgroup_events_index[idx];
>
> where mem_cgroup_events_index is an array of size NR_VM_EVENT_ITEMS. If a
> BPF program passes event >= NR_VM_EVENT_ITEMS, this would be an
> out-of-bounds array read.
>
> The sister function bpf_mem_cgroup_page_state() below validates its idx
> parameter before calling its callee:

It's a false positive: the check is enforced by the bpf verifier based
on type information.

>
>> +/**
>> + * bpf_mem_cgroup_page_state - Read memory cgroup's page state counter
>> + * @memcg: memory cgroup
>> + * @idx: counter idx
>> + *
>> + * Allows to read memory cgroup statistics. The output is in bytes.
>> + */
>> +__bpf_kfunc unsigned long bpf_mem_cgroup_page_state(struct mem_cgroup *memcg, int idx)
>> +{
>> +	if (idx < 0 || idx >= MEMCG_NR_STAT)
>> +		return (unsigned long)-1;
>> +
>> +	return memcg_page_state_output(memcg, idx);
>> +}
>
> Should bpf_mem_cgroup_vm_events() have a similar bounds check, perhaps:
>
>   if (event >= NR_VM_EVENT_ITEMS)
>       return (unsigned long)-1;
>
> before calling memcg_events()?

No. Here is required because idx is a simple int.