[RFC PATCH bpf-next] xdp: Add tracepoint on XDP program return

["Toke Høiland-Jørgensen" <toke@redhat.com>]

This adds a new tracepoint, xdp_prog_return, which is triggered at every
XDP program return. This was first discussed back in August[0] as a way to
hook XDP into the kernel drop_monitor framework, to have a one-stop place
to find all packet drops in the system.

Because trace/events/xdp.h includes filter.h, some ifdef guarding is needed
to be able to use the tracepoint from bpf_prog_run_xdp(). If anyone has any
ideas for how to improve on this, please to speak up. Sending this RFC
because of this issue, and to get some feedback from Ido on whether this
tracepoint has enough data for drop_monitor usage.

[0] https://lore.kernel.org/netdev/20190809125418.GB2931@splinter/

Cc: Ido Schimmel <idosch@idosch.org>
Signed-off-by: Toke Høiland-Jørgensen <toke@redhat.com>
---
 include/linux/filter.h     | 22 +++++++++++++++++--
 include/trace/events/xdp.h | 45 ++++++++++++++++++++++++++++++++++++++
 kernel/bpf/core.c          |  2 ++
 3 files changed, 67 insertions(+), 2 deletions(-)

diff --git a/include/linux/filter.h b/include/linux/filter.h
index 37ac7025031d..f5e79171902f 100644
--- a/include/linux/filter.h
+++ b/include/linux/filter.h
@@ -704,19 +704,37 @@ static inline u32 bpf_prog_run_clear_cb(const struct bpf_prog *prog,
 
 DECLARE_BPF_DISPATCHER(bpf_dispatcher_xdp)
 
+#if defined(_XDP_TRACE_DEF) || defined(_TRACE_XDP_H)
+static void call_trace_xdp_prog_return(const struct xdp_buff *xdp,
+				       const struct bpf_prog *prog,
+				       u32 act);
+#else
+#ifndef _CALL_TRACE_XDP
+#define _CALL_TRACE_XDP
+static inline void call_trace_xdp_prog_return(const struct xdp_buff *xdp,
+					      const struct bpf_prog *prog,
+					      u32 act) {}
+#endif
+#endif
+
 static __always_inline u32 bpf_prog_run_xdp(const struct bpf_prog *prog,
 					    struct xdp_buff *xdp)
 {
+	u32 ret;
+
 	/* Caller needs to hold rcu_read_lock() (!), otherwise program
 	 * can be released while still running, or map elements could be
 	 * freed early while still having concurrent users. XDP fastpath
 	 * already takes rcu_read_lock() when fetching the program, so
 	 * it's not necessary here anymore.
 	 */
-	return __BPF_PROG_RUN(prog, xdp,
-			      BPF_DISPATCHER_FUNC(bpf_dispatcher_xdp));
+	ret = __BPF_PROG_RUN(prog, xdp,
+			     BPF_DISPATCHER_FUNC(bpf_dispatcher_xdp));
+	call_trace_xdp_prog_return(xdp, prog, ret);
+	return ret;
 }
 
+
 void bpf_prog_change_xdp(struct bpf_prog *prev_prog, struct bpf_prog *prog);
 
 static inline u32 bpf_prog_insn_size(const struct bpf_prog *prog)
diff --git a/include/trace/events/xdp.h b/include/trace/events/xdp.h
index a7378bcd9928..e64f4221bd2e 100644
--- a/include/trace/events/xdp.h
+++ b/include/trace/events/xdp.h
@@ -50,6 +50,51 @@ TRACE_EVENT(xdp_exception,
 		  __entry->ifindex)
 );
 
+TRACE_EVENT(xdp_prog_return,
+
+	TP_PROTO(const struct xdp_buff *xdp,
+		 const struct bpf_prog *pr, u32 act),
+
+	TP_ARGS(xdp, pr, act),
+
+	TP_STRUCT__entry(
+		__field(int, prog_id)
+		__field(u32, act)
+		__field(int, ifindex)
+		__field(int, queue_index)
+		__field(const void *, data_addr)
+		__field(unsigned int, data_len)
+	),
+
+	TP_fast_assign(
+		__entry->prog_id	= pr->aux->id;
+		__entry->act		= act;
+		__entry->ifindex	= xdp->rxq->dev->ifindex;
+		__entry->queue_index	= xdp->rxq->queue_index;
+		__entry->data_addr	= xdp->data;
+		__entry->data_len	= (unsigned int)(xdp->data_end - xdp->data);
+	),
+
+	TP_printk("prog_id=%d action=%s ifindex=%d queue_index=%d data_addr=%p data_len=%u",
+		  __entry->prog_id,
+		  __print_symbolic(__entry->act, __XDP_ACT_SYM_TAB),
+		  __entry->ifindex,
+		  __entry->queue_index,
+		  __entry->data_addr,
+		  __entry->data_len)
+);
+
+#ifndef _CALL_TRACE_XDP
+#define _CALL_TRACE_XDP
+static inline void call_trace_xdp_prog_return(const struct xdp_buff *xdp,
+					      const struct bpf_prog *prog,
+					      u32 act)
+{
+	trace_xdp_prog_return(xdp, prog, act);
+}
+#endif
+
+
 TRACE_EVENT(xdp_bulk_tx,
 
 	TP_PROTO(const struct net_device *dev,
diff --git a/kernel/bpf/core.c b/kernel/bpf/core.c
index 2ff01a716128..a81d3b8d8e5c 100644
--- a/kernel/bpf/core.c
+++ b/kernel/bpf/core.c
@@ -17,6 +17,8 @@
  * Kris Katterjohn - Added many additional checks in bpf_check_classic()
  */
 
+#define _XDP_TRACE_DEF
+
 #include <uapi/linux/btf.h>
 #include <linux/filter.h>
 #include <linux/skbuff.h>
-- 
2.24.1