Hi,
On 12/4/2023 2:54 AM, Alexei Starovoitov wrote:
> On Thu, Nov 30, 2023 at 10:01:18PM +0800, Hou Tao wrote:
>> From: Hou Tao <houtao1@xxxxxxxxxx>
>>
>> When removing the inner map from the outer map, the inner map will be
>> freed after one RCU grace period and one RCU tasks trace grace
>> period, so it is certain that the bpf program, which may access the
>> inner map, has exited before the inner map is freed.
>>
>> However there is no need to wait for one RCU tasks trace grace period if
>> the outer map is only accessed by non-sleepable program. So adding
>> sleepable_refcnt in bpf_map and increasing sleepable_refcnt when adding
>> the outer map into env->used_maps for sleepable program. Considering the
>> max number of bpf program is INT_MAX - 1, atomic_t instead of atomic64_t
>> is used for sleepable_refcnt. When removing the inner map from the outer
>> map, using sleepable_refcnt to decide whether or not a RCU tasks trace
>> grace period is needed before freeing the inner map.
> Optimizing too soon as usual?
> I bet you saw that we use:
> atomic64_t refcnt
> for bpf_map, but you probably didn't dig into git history and
> missed commit 92117d8443bc ("bpf: fix refcnt overflow") ?
Yes. I didn't think it thoroughly. Although the max number of bpf
program INT_MAX - 1, but the allocation of id happens after the increase
of sleepable_refcnt, so the overflow is still possible. Will fix it in v5.
