On Fri, Jun 26, 2020 at 2:45 AM Jakub Sitnicki <jakub@xxxxxxxxxxxxxx> wrote:
>
> On Thu, Jun 25, 2020 at 10:50 PM CEST, Andrii Nakryiko wrote:
> > On Thu, Jun 25, 2020 at 7:17 AM Jakub Sitnicki <jakub@xxxxxxxxxxxxxx> wrote:
> >>
> >> Prepare for having multi-prog attachments for new netns attach types by
> >> storing programs to run in a bpf_prog_array, which is well suited for
> >> iterating over programs and running them in sequence.
> >>
> >> After this change bpf(PROG_QUERY) may block to allocate memory in
> >> bpf_prog_array_copy_to_user() for collected program IDs. This forces a
> >> change in how we protect access to the attached program in the query
> >> callback. Because bpf_prog_array_copy_to_user() can sleep, we switch from
> >> an RCU read lock to holding a mutex that serializes updaters.
> >>
> >> Because we allow only one BPF flow_dissector program to be attached to
> >> netns at all times, the bpf_prog_array pointed by net->bpf.run_array is
> >> always either detached (null) or one element long.
> >>
> >> No functional changes intended.
> >>
> >> Signed-off-by: Jakub Sitnicki <jakub@xxxxxxxxxxxxxx>
> >> ---
> >
> > I wonder if instead of NULL prog_array, it's better to just use a
> > dummy empty (but allocated) array. Might help eliminate some of the
> > IFs, maybe even in the hot path.
>
> That was my initial approach, which I abandoned seeing that it leads to
> replacing NULL prog_array checks in flow_dissector with
> bpf_prog_array_is_empty() checks to determine which netns has a BPF
> program attached. So no IFs gone there.
>
> While on the hot path, where we run the program, we probably would still
> be left with an IF checking for empty prog_array to avoid building the
> context if no progs will RUN.
>
> The checks I'm referring to are on attach path, in
> flow_dissector_bpf_prog_attach_check(), and hot-path,
> __skb_flow_dissect().
>
Fair enough.
> >
> >
> >> include/net/netns/bpf.h | 5 +-
> >> kernel/bpf/net_namespace.c | 120 +++++++++++++++++++++++++------------
> >> net/core/flow_dissector.c | 19 +++---
> >> 3 files changed, 96 insertions(+), 48 deletions(-)
> >>
> >
> > [...]
> >
> >
> >>
> >> +/* Must be called with netns_bpf_mutex held. */
> >> +static int __netns_bpf_prog_query(const union bpf_attr *attr,
> >> + union bpf_attr __user *uattr,
> >> + struct net *net,
> >> + enum netns_bpf_attach_type type)
> >> +{
> >> + __u32 __user *prog_ids = u64_to_user_ptr(attr->query.prog_ids);
> >> + struct bpf_prog_array *run_array;
> >> + u32 prog_cnt = 0, flags = 0;
> >> +
> >> + run_array = rcu_dereference_protected(net->bpf.run_array[type],
> >> + lockdep_is_held(&netns_bpf_mutex));
> >> + if (run_array)
> >> + prog_cnt = bpf_prog_array_length(run_array);
> >> +
> >> + if (copy_to_user(&uattr->query.attach_flags, &flags, sizeof(flags)))
> >> + return -EFAULT;
> >> + if (copy_to_user(&uattr->query.prog_cnt, &prog_cnt, sizeof(prog_cnt)))
> >> + return -EFAULT;
> >> + if (!attr->query.prog_cnt || !prog_ids || !prog_cnt)
> >> + return 0;
> >> +
> >> + return bpf_prog_array_copy_to_user(run_array, prog_ids,
> >> + attr->query.prog_cnt);
> >
> > It doesn't seem like bpf_prog_array_copy_to_user can handle NULL run_array
>
> Correct. And we never invoke it when run_array is NULL because then
> prog_cnt == 0.
Oh, that !prog_cnt above, right.. it's easy to miss.
>
> >
> >> +}
> >> +
> >> int netns_bpf_prog_query(const union bpf_attr *attr,
> >> union bpf_attr __user *uattr)
> >> {
> >> - __u32 __user *prog_ids = u64_to_user_ptr(attr->query.prog_ids);
> >> - u32 prog_id, prog_cnt = 0, flags = 0;
> >> enum netns_bpf_attach_type type;
> >> - struct bpf_prog *attached;
> >> struct net *net;
> >> + int ret;
> >>
> >> if (attr->query.query_flags)
> >> return -EINVAL;
> >
> > [...]
