Hello,
On Mon, Oct 21, 2024 at 04:32:10PM -0700, Andrii Nakryiko wrote:
> On Thu, Oct 17, 2024 at 1:06 AM Namhyung Kim <namhyung@xxxxxxxxxx> wrote:
> >
> > Add a new open coded iterator for kmem_cache which can be called from a
> > BPF program like below. It doesn't take any argument and traverses all
> > kmem_cache entries.
> >
> > struct kmem_cache *pos;
> >
> > bpf_for_each(kmem_cache, pos) {
> > ...
> > }
> >
> > As it needs to grab slab_mutex, it should be called from sleepable BPF
> > programs only.
> >
> > Signed-off-by: Namhyung Kim <namhyung@xxxxxxxxxx>
> > ---
> > kernel/bpf/helpers.c | 3 ++
> > kernel/bpf/kmem_cache_iter.c | 87 ++++++++++++++++++++++++++++++++++++
> > 2 files changed, 90 insertions(+)
> >
> > diff --git a/kernel/bpf/helpers.c b/kernel/bpf/helpers.c
> > index 073e6f04f4d765ff..d1dfa4f335577914 100644
> > --- a/kernel/bpf/helpers.c
> > +++ b/kernel/bpf/helpers.c
> > @@ -3111,6 +3111,9 @@ BTF_ID_FLAGS(func, bpf_iter_bits_next, KF_ITER_NEXT | KF_RET_NULL)
> > BTF_ID_FLAGS(func, bpf_iter_bits_destroy, KF_ITER_DESTROY)
> > BTF_ID_FLAGS(func, bpf_copy_from_user_str, KF_SLEEPABLE)
> > BTF_ID_FLAGS(func, bpf_get_kmem_cache)
> > +BTF_ID_FLAGS(func, bpf_iter_kmem_cache_new, KF_ITER_NEW | KF_SLEEPABLE)
> > +BTF_ID_FLAGS(func, bpf_iter_kmem_cache_next, KF_ITER_NEXT | KF_RET_NULL | KF_SLEEPABLE)
> > +BTF_ID_FLAGS(func, bpf_iter_kmem_cache_destroy, KF_ITER_DESTROY | KF_SLEEPABLE)
>
> I'm curious. Having bpf_iter_kmem_cache_{new,next,destroy} functions,
> can we rewrite kmem_cache_iter_seq_next in terms of these ones, so
> that we have less duplication of iteration logic? Or there will be
> some locking concerns preventing this? (I haven't looked into the
> actual logic much, sorry, lazy question)
It should be fine with locking, I think there's a subtle difference
between seq interface and the open coded iterator. But I'll think about
how to reduce the duplication.
Thanks for your review!
Namhyung
