On Thu, Nov 3, 2022 at 9:01 PM Alexei Starovoitov
<alexei.starovoitov@xxxxxxxxx> wrote:
>
> On Fri, Nov 04, 2022 at 12:39:55AM +0530, Kumar Kartikeya Dwivedi wrote:
> >
> > -enum bpf_kptr_type {
> > - BPF_KPTR_UNREF,
> > - BPF_KPTR_REF,
> > +enum btf_field_type {
> > + BPF_KPTR_UNREF = (1 << 2),
> > + BPF_KPTR_REF = (1 << 3),
> > + BPF_KPTR = BPF_KPTR_UNREF | BPF_KPTR_REF,
> > };
>
> ...
>
> > + for (i = 0; i < sizeof(map->record->field_mask) * 8; i++) {
> > + switch (map->record->field_mask & (1 << i)) {
> > + case 0:
> > + continue;
> > + case BPF_KPTR_UNREF:
> > + case BPF_KPTR_REF:
> > + if (map->map_type != BPF_MAP_TYPE_HASH &&
> > + map->map_type != BPF_MAP_TYPE_LRU_HASH &&
> > + map->map_type != BPF_MAP_TYPE_ARRAY &&
> > + map->map_type != BPF_MAP_TYPE_PERCPU_ARRAY) {
> > + ret = -EOPNOTSUPP;
> > + goto free_map_tab;
> > + }
> > + break;
> > + default:
> > + /* Fail if map_type checks are missing for a field type */
> > + ret = -EOPNOTSUPP;
> > + goto free_map_tab;
> > + }
>
> With this patch alone this is also wrong.
Actually this bit is probably fine. The bug is elsewhere.
The point below stands:
> And it breaks bisect.
> Please make sure to do a full vmtest.sh for every patch in the series.
