Message-ID: <CAJN39ogsaG0-nh0Q=_KUVCPaEhC0SHKhxR+jJJFJ6C_seEG+_g@mail.gmail.com>
Date:   Wed, 27 Mar 2019 13:44:58 -0700
From:   Brendan Gregg <bgregg@...flix.com>
To:     Javier Honduvilla Coto <javierhonduco@...com>
Cc:     Daniel Borkmann <daniel@...earbox.net>,
        "netdev@...r.kernel.org" <netdev@...r.kernel.org>,
        Yonghong Song <yhs@...com>, Kernel Team <Kernel-team@...com>
Subject: Re: [PATCH v4 bpf-next 1/3] bpf: add bpf_progenyof helper

On Wed, Mar 27, 2019 at 8:59 AM Javier Honduvilla Coto
<javierhonduco@...com> wrote:
>
> On Mon, Mar 25, 2019 at 03:17:03PM +0100, Daniel Borkmann wrote:
> > On 03/22/2019 11:38 PM, Javier Honduvilla Coto wrote:
> > > This patch adds the bpf_progenyof helper which receives a PID and returns
> > > 1 if the process currently being executed is in the process hierarchy
> > > including itself or 0 if not.
> > >
> > > This is very useful in tracing programs when we want to filter by a
> > > given PID and all the children it might spawn. The current workarounds
> > > most people implement for this purpose have issues:
> > >
> > > - Attaching to process spawning syscalls and dynamically add those PIDs
> > >   to some bpf map that would be used to filter is cumbersome and
> > > potentially racy.
> > > - Unrolling some loop to perform what this helper is doing consumes lots
> > >   of instructions. That and the impossibility to jump backwards makes it
> > > really hard to be correct in really large process chains.
> > >
> > > Signed-off-by: Javier Honduvilla Coto <javierhonduco@...com>
> > > ---
> > >  include/linux/bpf.h      |  1 +
> > >  include/uapi/linux/bpf.h | 10 +++++++++-
> > >  kernel/bpf/core.c        |  1 +
> > >  kernel/bpf/helpers.c     | 32 ++++++++++++++++++++++++++++++++
> > >  kernel/trace/bpf_trace.c |  2 ++
> > >  5 files changed, 45 insertions(+), 1 deletion(-)
> > >
> > > diff --git a/include/linux/bpf.h b/include/linux/bpf.h
> > > index f62897198844..bd0d2b38e7d5 100644
> > > --- a/include/linux/bpf.h
> > > +++ b/include/linux/bpf.h
> > > @@ -930,6 +930,7 @@ extern const struct bpf_func_proto bpf_sk_redirect_map_proto;
> > >  extern const struct bpf_func_proto bpf_spin_lock_proto;
> > >  extern const struct bpf_func_proto bpf_spin_unlock_proto;
> > >  extern const struct bpf_func_proto bpf_get_local_storage_proto;
> > > +extern const struct bpf_func_proto bpf_progenyof_proto;
> > >
> > >  /* Shared helpers among cBPF and eBPF. */
> > >  void bpf_user_rnd_init_once(void);
> > > diff --git a/include/uapi/linux/bpf.h b/include/uapi/linux/bpf.h
> > > index 3c04410137d9..cf54cc739bf4 100644
> > > --- a/include/uapi/linux/bpf.h
> > > +++ b/include/uapi/linux/bpf.h
> > > @@ -2463,6 +2463,13 @@ union bpf_attr {
> > >   *         Return
> > >   *                 0 if iph and th are a valid SYN cookie ACK, or a negative error
> > >   *                 otherwise.
> > > + * int bpf_progenyof(int pid)
> > > + * Description
> > > + *         This helper is useful in programs that want to filter events
> > > + *         happening to a pid of any of its descendants.
> > > + * Return
> > > + *         1 if the currently executing process' pid is in the process
> > > + *         hierarchy of the passed pid. 0 Otherwise.
> >
> > What about the -EINVAL?
>
> I think we can remove this, as Alexei told me this was not needed
> anymore (copied it from other helpers that have it).
>
> Should we remove that check in other patch for the other helpers
> that have it?
>
> >
> > >   */
> > >  #define __BPF_FUNC_MAPPER(FN)              \
> > >     FN(unspec),                     \
> > > @@ -2565,7 +2572,8 @@ union bpf_attr {
> > >     FN(skb_ecn_set_ce),             \
> > >     FN(get_listener_sock),          \
> > >     FN(skc_lookup_tcp),             \
> > > -   FN(tcp_check_syncookie),
> > > +   FN(tcp_check_syncookie),        \
> > > +   FN(progenyof),
> > >
> > >  /* integer value in 'imm' field of BPF_CALL instruction selects which helper
> > >   * function eBPF program intends to call
> > > diff --git a/kernel/bpf/core.c b/kernel/bpf/core.c
> > > index ff09d32a8a1b..437986497468 100644
> > > --- a/kernel/bpf/core.c
> > > +++ b/kernel/bpf/core.c
> > > @@ -2044,6 +2044,7 @@ const struct bpf_func_proto bpf_get_current_uid_gid_proto __weak;
> > >  const struct bpf_func_proto bpf_get_current_comm_proto __weak;
> > >  const struct bpf_func_proto bpf_get_current_cgroup_id_proto __weak;
> > >  const struct bpf_func_proto bpf_get_local_storage_proto __weak;
> > > +const struct bpf_func_proto bpf_progenyof_proto __weak;
> > >
> > >  const struct bpf_func_proto * __weak bpf_get_trace_printk_proto(void)
> > >  {
> > > diff --git a/kernel/bpf/helpers.c b/kernel/bpf/helpers.c
> > > index a411fc17d265..f093b35d1ba8 100644
> > > --- a/kernel/bpf/helpers.c
> > > +++ b/kernel/bpf/helpers.c
> > > @@ -18,6 +18,7 @@
> > >  #include <linux/sched.h>
> > >  #include <linux/uidgid.h>
> > >  #include <linux/filter.h>
> > > +#include <linux/init_task.h>
> > >
> > >  /* If kernel subsystem is allowing eBPF programs to call this function,
> > >   * inside its own verifier_ops->get_func_proto() callback it should return
> > > @@ -364,3 +365,34 @@ const struct bpf_func_proto bpf_get_local_storage_proto = {
> > >  };
> > >  #endif
> > >  #endif
> > > +
> > > +BPF_CALL_1(bpf_progenyof, int, pid)
> >
> > Nit: could we add a more descriptive helper name? What's progenyof? Also s/int/pid_t/?
> >
>
> It's true that "progeny" is not a very commonly used word :D. A coworker
> suggested "descendantof", what do you think? It's a bit difficult to
> convey "is true on the passed pid + on all the process under the
> hierarchy chain of that pid", so I will try rewording the docs so the
> semantics are very clear!

What about childof? I see the word child used more than anything:

$ man ps pgrep pidstat pstree top | grep child | wc -l
29
$ man ps pgrep pidstat pstree top | grep progeny | wc -l
0
$ man ps pgrep pidstat pstree top | grep ancestor | wc -l
2

Brendan
-- 
Brendan Gregg, Senior Performance Architect, Netflix

Powered by blists - more mailing lists