Re: [PATCH v5 bpf-next 2/5] libbpf: Add "map_extra" as a per-map-type extra flag

[Andrii Nakryiko <andrii.nakryiko@gmail.com>]

On Wed, Oct 27, 2021 at 11:39 AM Joanne Koong <joannekoong@fb.com> wrote:
>
> On 10/26/21 8:30 PM, Andrii Nakryiko wrote:
>
> >
> > On 10/22/21 3:02 PM, Joanne Koong wrote:
> >> This patch adds the libbpf infrastructure for supporting a
> >> per-map-type "map_extra" field, whose definition will be
> >> idiosyncratic depending on map type.
> >>
> >> For example, for the bloom filter map, the lower 4 bits of
> >> map_extra is used to denote the number of hash functions.
> >>
> >> Please note that until libbpf 1.0 is here, the
> >> "bpf_create_map_params" struct is used as a temporary
> >> means for propagating the map_extra field to the kernel.
> >>
> >> Signed-off-by: Joanne Koong <joannekoong@fb.com>
> >> ---
> >>   include/uapi/linux/bpf.h         |  1 +
> >>   tools/include/uapi/linux/bpf.h   |  1 +
> >>   tools/lib/bpf/bpf.c              | 27 ++++++++++++++++++++-
> >>   tools/lib/bpf/bpf_gen_internal.h |  2 +-
> >>   tools/lib/bpf/gen_loader.c       |  3 ++-
> >>   tools/lib/bpf/libbpf.c           | 41 ++++++++++++++++++++++++++++----
> >>   tools/lib/bpf/libbpf.h           |  3 +++
> >>   tools/lib/bpf/libbpf.map         |  2 ++
> >>   tools/lib/bpf/libbpf_internal.h  | 25 ++++++++++++++++++-
> >>   9 files changed, 96 insertions(+), 9 deletions(-)
> [...]
> >> diff --git a/tools/lib/bpf/libbpf.c b/tools/lib/bpf/libbpf.c
> >> index db6e48014839..751cfb9778dc 100644
> >> --- a/tools/lib/bpf/libbpf.c
> >> +++ b/tools/lib/bpf/libbpf.c
> >> @@ -400,6 +400,7 @@ struct bpf_map {
> >>       char *pin_path;
> >>       bool pinned;
> >>       bool reused;
> >> +    __u64 map_extra;
> >>   };
> >>     enum extern_type {
> >> @@ -2313,6 +2314,17 @@ int parse_btf_map_def(const char *map_name,
> >> struct btf *btf,
> >>               }
> >>               map_def->pinning = val;
> >>               map_def->parts |= MAP_DEF_PINNING;
> >> +        } else if (strcmp(name, "map_extra") == 0) {
> >> +            /*
> >> +             * TODO: When the BTF array supports __u64s, read into
> >> +             * map_def->map_extra directly.
> >> +             */
> >
> > Please drop the TODO comment. There are no plans to extend BTF arrays
> > to support __u64 sizes.
> >
> I see, I will remove this.
>
> If BTF arrays never support __u64 sizes, then people won't be able to
> define a map in libbpf
> that uses bits 33 - 64 of the map_extra field, correct? Or is there a
> workaround for them that
> I'm not seeing?

Right, they won't be able to do it declaratively (if we ever need
higher 32 bits). They will still be able to specify map_extra
programmatically through bpf_map__set_map_extra() and with low-level
map creation APIs. So not ideal, but there will be a work-around.

>
> >
> > [...]
> >