[PATCH net-next 0/3] BPF tracepoints

[PATCH net-next 0/3] BPF tracepoints

[Daniel Borkmann]

This set adds tracepoints to BPF for better introspection and
debugging. The first two patches are prerequisite for the actual
third patch that adds the tracepoints. I think the first two are
small and straight forward enough that they could ideally go via
net-next, but I'm also open to other suggestions on how to route
them in case that's not applicable (it would reduce potential
merge conflicts on BPF side, though). For details, please see
individual patches.

Thanks a lot!

Daniel Borkmann (3):
  trace: add variant without spacing in trace_print_hex_seq
  lib, traceevent: add PRINT_HEX_STR variant
  bpf: add initial bpf tracepoints

 drivers/net/ethernet/mellanox/mlx4/en_rx.c         |   3 +
 drivers/net/ethernet/mellanox/mlx5/core/en_rx.c    |  12 +-
 .../net/ethernet/netronome/nfp/nfp_net_common.c    |  15 +-
 drivers/net/ethernet/qlogic/qede/qede_fp.c         |   4 +
 drivers/net/virtio_net.c                           |  12 +-
 include/linux/bpf_trace.h                          |   7 +
 include/linux/trace_events.h                       |   3 +-
 include/trace/events/bpf.h                         | 347 +++++++++++++++++++++
 include/trace/events/xdp.h                         |  53 ++++
 include/trace/trace_events.h                       |   8 +-
 kernel/bpf/core.c                                  |   9 +
 kernel/bpf/inode.c                                 |  17 +-
 kernel/bpf/syscall.c                               |  19 +-
 kernel/trace/trace_output.c                        |   7 +-
 tools/lib/traceevent/event-parse.c                 |  34 +-
 tools/lib/traceevent/event-parse.h                 |   1 +
 .../perf/util/scripting-engines/trace-event-perl.c |   1 +
 .../util/scripting-engines/trace-event-python.c    |   1 +
 18 files changed, 530 insertions(+), 23 deletions(-)
 create mode 100644 include/linux/bpf_trace.h
 create mode 100644 include/trace/events/bpf.h
 create mode 100644 include/trace/events/xdp.h

-- 
1.9.3

[PATCH net-next 1/3] trace: add variant without spacing in trace_print_hex_seq

[Daniel Borkmann]

For upcoming tracepoint support for BPF, we want to dump the program's
tag. Format should be similar to __print_hex(), but without spacing.
Add a __print_hex_str() variant for exactly that purpose that reuses
trace_print_hex_seq().

Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 include/linux/trace_events.h | 3 ++-
 include/trace/trace_events.h | 8 +++++++-
 kernel/trace/trace_output.c  | 7 ++++---
 3 files changed, 13 insertions(+), 5 deletions(-)

diff --git a/include/linux/trace_events.h b/include/linux/trace_events.h
index be00761..cfa475a 100644
--- a/include/linux/trace_events.h
+++ b/include/linux/trace_events.h
@@ -33,7 +33,8 @@ const char *trace_print_bitmask_seq(struct trace_seq *p, void *bitmask_ptr,
 				    unsigned int bitmask_size);
 
 const char *trace_print_hex_seq(struct trace_seq *p,
-				const unsigned char *buf, int len);
+				const unsigned char *buf, int len,
+				bool spacing);
 
 const char *trace_print_array_seq(struct trace_seq *p,
 				   const void *buf, int count,
diff --git a/include/trace/trace_events.h b/include/trace/trace_events.h
index 467e12f..9f68462 100644
--- a/include/trace/trace_events.h
+++ b/include/trace/trace_events.h
@@ -297,7 +297,12 @@
 #endif
 
 #undef __print_hex
-#define __print_hex(buf, buf_len) trace_print_hex_seq(p, buf, buf_len)
+#define __print_hex(buf, buf_len)					\
+	trace_print_hex_seq(p, buf, buf_len, true)
+
+#undef __print_hex_str
+#define __print_hex_str(buf, buf_len)					\
+	trace_print_hex_seq(p, buf, buf_len, false)
 
 #undef __print_array
 #define __print_array(array, count, el_size)				\
@@ -711,6 +716,7 @@
 #undef __print_flags
 #undef __print_symbolic
 #undef __print_hex
+#undef __print_hex_str
 #undef __get_dynamic_array
 #undef __get_dynamic_array_len
 #undef __get_str
diff --git a/kernel/trace/trace_output.c b/kernel/trace/trace_output.c
index 5d33a73..30a144b1 100644
--- a/kernel/trace/trace_output.c
+++ b/kernel/trace/trace_output.c
@@ -163,14 +163,15 @@ enum print_line_t trace_print_printk_msg_only(struct trace_iterator *iter)
 EXPORT_SYMBOL_GPL(trace_print_bitmask_seq);
 
 const char *
-trace_print_hex_seq(struct trace_seq *p, const unsigned char *buf, int buf_len)
+trace_print_hex_seq(struct trace_seq *p, const unsigned char *buf, int buf_len,
+		    bool spacing)
 {
 	int i;
 	const char *ret = trace_seq_buffer_ptr(p);
 
 	for (i = 0; i < buf_len; i++)
-		trace_seq_printf(p, "%s%2.2x", i == 0 ? "" : " ", buf[i]);
-
+		trace_seq_printf(p, "%s%2.2x", !spacing || i == 0 ? "" : " ",
+				 buf[i]);
 	trace_seq_putc(p, 0);
 
 	return ret;
-- 
1.9.3

Re: [PATCH net-next 1/3] trace: add variant without spacing in trace_print_hex_seq

[Arnaldo Carvalho de Melo]

Em Wed, Jan 25, 2017 at 02:28:16AM +0100, Daniel Borkmann escreveu:
> For upcoming tracepoint support for BPF, we want to dump the program's
> tag. Format should be similar to __print_hex(), but without spacing.
> Add a __print_hex_str() variant for exactly that purpose that reuses
> trace_print_hex_seq().

Steven should be back to his side of the wall soon, will wait for his
Ack, ok?

- Arnaldo
 
> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
> Cc: Steven Rostedt <rostedt@goodmis.org>
> Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
> ---
>  include/linux/trace_events.h | 3 ++-
>  include/trace/trace_events.h | 8 +++++++-
>  kernel/trace/trace_output.c  | 7 ++++---
>  3 files changed, 13 insertions(+), 5 deletions(-)
> 
> diff --git a/include/linux/trace_events.h b/include/linux/trace_events.h
> index be00761..cfa475a 100644
> --- a/include/linux/trace_events.h
> +++ b/include/linux/trace_events.h
> @@ -33,7 +33,8 @@ const char *trace_print_bitmask_seq(struct trace_seq *p, void *bitmask_ptr,
>  				    unsigned int bitmask_size);
>  
>  const char *trace_print_hex_seq(struct trace_seq *p,
> -				const unsigned char *buf, int len);
> +				const unsigned char *buf, int len,
> +				bool spacing);
>  
>  const char *trace_print_array_seq(struct trace_seq *p,
>  				   const void *buf, int count,
> diff --git a/include/trace/trace_events.h b/include/trace/trace_events.h
> index 467e12f..9f68462 100644
> --- a/include/trace/trace_events.h
> +++ b/include/trace/trace_events.h
> @@ -297,7 +297,12 @@
>  #endif
>  
>  #undef __print_hex
> -#define __print_hex(buf, buf_len) trace_print_hex_seq(p, buf, buf_len)
> +#define __print_hex(buf, buf_len)					\
> +	trace_print_hex_seq(p, buf, buf_len, true)
> +
> +#undef __print_hex_str
> +#define __print_hex_str(buf, buf_len)					\
> +	trace_print_hex_seq(p, buf, buf_len, false)
>  
>  #undef __print_array
>  #define __print_array(array, count, el_size)				\
> @@ -711,6 +716,7 @@
>  #undef __print_flags
>  #undef __print_symbolic
>  #undef __print_hex
> +#undef __print_hex_str
>  #undef __get_dynamic_array
>  #undef __get_dynamic_array_len
>  #undef __get_str
> diff --git a/kernel/trace/trace_output.c b/kernel/trace/trace_output.c
> index 5d33a73..30a144b1 100644
> --- a/kernel/trace/trace_output.c
> +++ b/kernel/trace/trace_output.c
> @@ -163,14 +163,15 @@ enum print_line_t trace_print_printk_msg_only(struct trace_iterator *iter)
>  EXPORT_SYMBOL_GPL(trace_print_bitmask_seq);
>  
>  const char *
> -trace_print_hex_seq(struct trace_seq *p, const unsigned char *buf, int buf_len)
> +trace_print_hex_seq(struct trace_seq *p, const unsigned char *buf, int buf_len,
> +		    bool spacing)
>  {
>  	int i;
>  	const char *ret = trace_seq_buffer_ptr(p);
>  
>  	for (i = 0; i < buf_len; i++)
> -		trace_seq_printf(p, "%s%2.2x", i == 0 ? "" : " ", buf[i]);
> -
> +		trace_seq_printf(p, "%s%2.2x", !spacing || i == 0 ? "" : " ",
> +				 buf[i]);
>  	trace_seq_putc(p, 0);
>  
>  	return ret;
> -- 
> 1.9.3

Re: [PATCH net-next 1/3] trace: add variant without spacing in trace_print_hex_seq

[Daniel Borkmann]

On 01/26/2017 08:53 PM, Arnaldo Carvalho de Melo wrote:
> Em Wed, Jan 25, 2017 at 02:28:16AM +0100, Daniel Borkmann escreveu:
>> For upcoming tracepoint support for BPF, we want to dump the program's
>> tag. Format should be similar to __print_hex(), but without spacing.
>> Add a __print_hex_str() variant for exactly that purpose that reuses
>> trace_print_hex_seq().
>
> Steven should be back to his side of the wall soon, will wait for his
> Ack, ok?

Ok, seems this set got applied already to net-next in the meantime, so
if there are any objections on this, I will follow up with a patch of
course.

Thanks,
Daniel

> - Arnaldo

Re: [PATCH net-next 1/3] trace: add variant without spacing in trace_print_hex_seq

[Steven Rostedt]

On Wed, 25 Jan 2017 02:28:16 +0100
Daniel Borkmann <daniel@iogearbox.net> wrote:

> For upcoming tracepoint support for BPF, we want to dump the program's
> tag. Format should be similar to __print_hex(), but without spacing.
> Add a __print_hex_str() variant for exactly that purpose that reuses
> trace_print_hex_seq().
> 
> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
> Cc: Steven Rostedt <rostedt@goodmis.org>
> Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
> ---
>  include/linux/trace_events.h | 3 ++-
>  include/trace/trace_events.h | 8 +++++++-
>  kernel/trace/trace_output.c  | 7 ++++---
>  3 files changed, 13 insertions(+), 5 deletions(-)
> 
> diff --git a/include/linux/trace_events.h b/include/linux/trace_events.h
> index be00761..cfa475a 100644
> --- a/include/linux/trace_events.h
> +++ b/include/linux/trace_events.h
> @@ -33,7 +33,8 @@ const char *trace_print_bitmask_seq(struct trace_seq *p, void *bitmask_ptr,
>  				    unsigned int bitmask_size);
>  
>  const char *trace_print_hex_seq(struct trace_seq *p,
> -				const unsigned char *buf, int len);
> +				const unsigned char *buf, int len,
> +				bool spacing);

Hmm, "spacing" doesn't really mean much. What about the invert of it,
and have "concatenate"?

>  
>  const char *trace_print_array_seq(struct trace_seq *p,
>  				   const void *buf, int count,
> diff --git a/include/trace/trace_events.h b/include/trace/trace_events.h
> index 467e12f..9f68462 100644
> --- a/include/trace/trace_events.h
> +++ b/include/trace/trace_events.h
> @@ -297,7 +297,12 @@
>  #endif
>  
>  #undef __print_hex
> -#define __print_hex(buf, buf_len) trace_print_hex_seq(p, buf, buf_len)
> +#define __print_hex(buf, buf_len)					\
> +	trace_print_hex_seq(p, buf, buf_len, true)
> +
> +#undef __print_hex_str
> +#define __print_hex_str(buf, buf_len)					\
> +	trace_print_hex_seq(p, buf, buf_len, false)
>  
>  #undef __print_array
>  #define __print_array(array, count, el_size)				\
> @@ -711,6 +716,7 @@
>  #undef __print_flags
>  #undef __print_symbolic
>  #undef __print_hex
> +#undef __print_hex_str
>  #undef __get_dynamic_array
>  #undef __get_dynamic_array_len
>  #undef __get_str
> diff --git a/kernel/trace/trace_output.c b/kernel/trace/trace_output.c
> index 5d33a73..30a144b1 100644
> --- a/kernel/trace/trace_output.c
> +++ b/kernel/trace/trace_output.c
> @@ -163,14 +163,15 @@ enum print_line_t trace_print_printk_msg_only(struct trace_iterator *iter)
>  EXPORT_SYMBOL_GPL(trace_print_bitmask_seq);
>  

With the addition of this boolean parameter, this function shold
probably have a kernel doc header, that can explain the parameters.

-- Steve

>  const char *
> -trace_print_hex_seq(struct trace_seq *p, const unsigned char *buf, int buf_len)
> +trace_print_hex_seq(struct trace_seq *p, const unsigned char *buf, int buf_len,
> +		    bool spacing)
>  {
>  	int i;
>  	const char *ret = trace_seq_buffer_ptr(p);
>  
>  	for (i = 0; i < buf_len; i++)
> -		trace_seq_printf(p, "%s%2.2x", i == 0 ? "" : " ", buf[i]);
> -
> +		trace_seq_printf(p, "%s%2.2x", !spacing || i == 0 ? "" : " ",
> +				 buf[i]);
>  	trace_seq_putc(p, 0);
>  
>  	return ret;

Re: [PATCH net-next 1/3] trace: add variant without spacing in trace_print_hex_seq

[Daniel Borkmann]

On 01/30/2017 09:44 PM, Steven Rostedt wrote:
> On Wed, 25 Jan 2017 02:28:16 +0100
> Daniel Borkmann <daniel@iogearbox.net> wrote:
>
>> For upcoming tracepoint support for BPF, we want to dump the program's
>> tag. Format should be similar to __print_hex(), but without spacing.
>> Add a __print_hex_str() variant for exactly that purpose that reuses
>> trace_print_hex_seq().
>>
>> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
>> Cc: Steven Rostedt <rostedt@goodmis.org>
>> Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
>> ---
>>   include/linux/trace_events.h | 3 ++-
>>   include/trace/trace_events.h | 8 +++++++-
>>   kernel/trace/trace_output.c  | 7 ++++---
>>   3 files changed, 13 insertions(+), 5 deletions(-)
>>
>> diff --git a/include/linux/trace_events.h b/include/linux/trace_events.h
>> index be00761..cfa475a 100644
>> --- a/include/linux/trace_events.h
>> +++ b/include/linux/trace_events.h
>> @@ -33,7 +33,8 @@ const char *trace_print_bitmask_seq(struct trace_seq *p, void *bitmask_ptr,
>>   				    unsigned int bitmask_size);
>>
>>   const char *trace_print_hex_seq(struct trace_seq *p,
>> -				const unsigned char *buf, int len);
>> +				const unsigned char *buf, int len,
>> +				bool spacing);
>
> Hmm, "spacing" doesn't really mean much. What about the invert of it,
> and have "concatenate"?

Sure, I'm fine with that.

>>   const char *trace_print_array_seq(struct trace_seq *p,
>>   				   const void *buf, int count,
>> diff --git a/include/trace/trace_events.h b/include/trace/trace_events.h
>> index 467e12f..9f68462 100644
>> --- a/include/trace/trace_events.h
>> +++ b/include/trace/trace_events.h
>> @@ -297,7 +297,12 @@
>>   #endif
>>
>>   #undef __print_hex
>> -#define __print_hex(buf, buf_len) trace_print_hex_seq(p, buf, buf_len)
>> +#define __print_hex(buf, buf_len)					\
>> +	trace_print_hex_seq(p, buf, buf_len, true)
>> +
>> +#undef __print_hex_str
>> +#define __print_hex_str(buf, buf_len)					\
>> +	trace_print_hex_seq(p, buf, buf_len, false)
>>
>>   #undef __print_array
>>   #define __print_array(array, count, el_size)				\
>> @@ -711,6 +716,7 @@
>>   #undef __print_flags
>>   #undef __print_symbolic
>>   #undef __print_hex
>> +#undef __print_hex_str
>>   #undef __get_dynamic_array
>>   #undef __get_dynamic_array_len
>>   #undef __get_str
>> diff --git a/kernel/trace/trace_output.c b/kernel/trace/trace_output.c
>> index 5d33a73..30a144b1 100644
>> --- a/kernel/trace/trace_output.c
>> +++ b/kernel/trace/trace_output.c
>> @@ -163,14 +163,15 @@ enum print_line_t trace_print_printk_msg_only(struct trace_iterator *iter)
>>   EXPORT_SYMBOL_GPL(trace_print_bitmask_seq);
>
> With the addition of this boolean parameter, this function shold
> probably have a kernel doc header, that can explain the parameters.

Yeah, I can add that. Since the patch got already applied, I will send
a follow-up for adding the kdoc and for changing the bool logic/name as
concatenate.

Thanks for your review!
Daniel

[PATCH net-next 2/3] lib, traceevent: add PRINT_HEX_STR variant

[Daniel Borkmann]

Add support for the __print_hex_str() macro that was added for
tracing, so that user space tools such as perf can understand
it as well.

Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/lib/traceevent/event-parse.c                 | 34 ++++++++++++++++++++--
 tools/lib/traceevent/event-parse.h                 |  1 +
 .../perf/util/scripting-engines/trace-event-perl.c |  1 +
 .../util/scripting-engines/trace-event-python.c    |  1 +
 4 files changed, 34 insertions(+), 3 deletions(-)

diff --git a/tools/lib/traceevent/event-parse.c b/tools/lib/traceevent/event-parse.c
index 14a4f62..f2ea780 100644
--- a/tools/lib/traceevent/event-parse.c
+++ b/tools/lib/traceevent/event-parse.c
@@ -831,6 +831,7 @@ static void free_arg(struct print_arg *arg)
 		free_flag_sym(arg->symbol.symbols);
 		break;
 	case PRINT_HEX:
+	case PRINT_HEX_STR:
 		free_arg(arg->hex.field);
 		free_arg(arg->hex.size);
 		break;
@@ -2629,10 +2630,11 @@ static char *arg_eval (struct print_arg *arg)
 }
 
 static enum event_type
-process_hex(struct event_format *event, struct print_arg *arg, char **tok)
+process_hex_common(struct event_format *event, struct print_arg *arg,
+		   char **tok, enum print_arg_type type)
 {
 	memset(arg, 0, sizeof(*arg));
-	arg->type = PRINT_HEX;
+	arg->type = type;
 
 	if (alloc_and_process_delim(event, ",", &arg->hex.field))
 		goto out;
@@ -2651,6 +2653,19 @@ static char *arg_eval (struct print_arg *arg)
 }
 
 static enum event_type
+process_hex(struct event_format *event, struct print_arg *arg, char **tok)
+{
+	return process_hex_common(event, arg, tok, PRINT_HEX);
+}
+
+static enum event_type
+process_hex_str(struct event_format *event, struct print_arg *arg,
+		char **tok)
+{
+	return process_hex_common(event, arg, tok, PRINT_HEX_STR);
+}
+
+static enum event_type
 process_int_array(struct event_format *event, struct print_arg *arg, char **tok)
 {
 	memset(arg, 0, sizeof(*arg));
@@ -3009,6 +3024,10 @@ static void remove_func_handler(struct pevent *pevent, char *func_name)
 		free_token(token);
 		return process_hex(event, arg, tok);
 	}
+	if (strcmp(token, "__print_hex_str") == 0) {
+		free_token(token);
+		return process_hex_str(event, arg, tok);
+	}
 	if (strcmp(token, "__print_array") == 0) {
 		free_token(token);
 		return process_int_array(event, arg, tok);
@@ -3547,6 +3566,7 @@ struct event_format *
 	case PRINT_SYMBOL:
 	case PRINT_INT_ARRAY:
 	case PRINT_HEX:
+	case PRINT_HEX_STR:
 		break;
 	case PRINT_TYPE:
 		val = eval_num_arg(data, size, event, arg->typecast.item);
@@ -3962,6 +3982,7 @@ static void print_str_arg(struct trace_seq *s, void *data, int size,
 		}
 		break;
 	case PRINT_HEX:
+	case PRINT_HEX_STR:
 		if (arg->hex.field->type == PRINT_DYNAMIC_ARRAY) {
 			unsigned long offset;
 			offset = pevent_read_number(pevent,
@@ -3981,7 +4002,7 @@ static void print_str_arg(struct trace_seq *s, void *data, int size,
 		}
 		len = eval_num_arg(data, size, event, arg->hex.size);
 		for (i = 0; i < len; i++) {
-			if (i)
+			if (i && arg->type == PRINT_HEX)
 				trace_seq_putc(s, ' ');
 			trace_seq_printf(s, "%02x", hex[i]);
 		}
@@ -5727,6 +5748,13 @@ static void print_args(struct print_arg *args)
 		print_args(args->hex.size);
 		printf(")");
 		break;
+	case PRINT_HEX_STR:
+		printf("__print_hex_str(");
+		print_args(args->hex.field);
+		printf(", ");
+		print_args(args->hex.size);
+		printf(")");
+		break;
 	case PRINT_INT_ARRAY:
 		printf("__print_array(");
 		print_args(args->int_array.field);
diff --git a/tools/lib/traceevent/event-parse.h b/tools/lib/traceevent/event-parse.h
index 7aae746..74cecba 100644
--- a/tools/lib/traceevent/event-parse.h
+++ b/tools/lib/traceevent/event-parse.h
@@ -292,6 +292,7 @@ enum print_arg_type {
 	PRINT_FUNC,
 	PRINT_BITMASK,
 	PRINT_DYNAMIC_ARRAY_LEN,
+	PRINT_HEX_STR,
 };
 
 struct print_arg {
diff --git a/tools/perf/util/scripting-engines/trace-event-perl.c b/tools/perf/util/scripting-engines/trace-event-perl.c
index e55a132..e74adfb 100644
--- a/tools/perf/util/scripting-engines/trace-event-perl.c
+++ b/tools/perf/util/scripting-engines/trace-event-perl.c
@@ -217,6 +217,7 @@ static void define_event_symbols(struct event_format *event,
 				       cur_field_name);
 		break;
 	case PRINT_HEX:
+	case PRINT_HEX_STR:
 		define_event_symbols(event, ev_name, args->hex.field);
 		define_event_symbols(event, ev_name, args->hex.size);
 		break;
diff --git a/tools/perf/util/scripting-engines/trace-event-python.c b/tools/perf/util/scripting-engines/trace-event-python.c
index 089438d..581e0ef 100644
--- a/tools/perf/util/scripting-engines/trace-event-python.c
+++ b/tools/perf/util/scripting-engines/trace-event-python.c
@@ -236,6 +236,7 @@ static void define_event_symbols(struct event_format *event,
 			      cur_field_name);
 		break;
 	case PRINT_HEX:
+	case PRINT_HEX_STR:
 		define_event_symbols(event, ev_name, args->hex.field);
 		define_event_symbols(event, ev_name, args->hex.size);
 		break;
-- 
1.9.3

Re: [PATCH net-next 2/3] lib, traceevent: add PRINT_HEX_STR variant

[Steven Rostedt]

On Wed, 25 Jan 2017 02:28:17 +0100
Daniel Borkmann <daniel@iogearbox.net> wrote:

> Add support for the __print_hex_str() macro that was added for
> tracing, so that user space tools such as perf can understand
> it as well.
> 
> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
> Cc: Steven Rostedt <rostedt@goodmis.org>
> Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
> ---

I haven't tested it, but it looks fine with me:

Acked-by: Steven Rostedt (VMware) <rostedt@goodmis.org>

-- Steve

[PATCH net-next 3/3] bpf: add initial bpf tracepoints

[Daniel Borkmann]

This work adds a number of tracepoints to paths that are either
considered slow-path or exception-like states, where monitoring or
inspecting them would be desirable.

For bpf(2) syscall, tracepoints have been placed for main commands
when they succeed. In XDP case, tracepoint is for exceptions, that
is, f.e. on abnormal BPF program exit such as unknown or XDP_ABORTED
return code, or when error occurs during XDP_TX action and the packet
could not be forwarded.

Both have been split into separate event headers, and can be further
extended. Worst case, if they unexpectedly should get into our way in
future, they can also removed [1]. Of course, these tracepoints (like
any other) can be analyzed by eBPF itself, etc. Example output:

  # ./perf record -a -e bpf:* sleep 10
  # ./perf script
  sock_example  6197 [005]   283.980322:      bpf:bpf_map_create: map type=ARRAY ufd=4 key=4 val=8 max=256 flags=0
  sock_example  6197 [005]   283.980721:       bpf:bpf_prog_load: prog=a5ea8fa30ea6849c type=SOCKET_FILTER ufd=5
  sock_example  6197 [005]   283.988423:   bpf:bpf_prog_get_type: prog=a5ea8fa30ea6849c type=SOCKET_FILTER
  sock_example  6197 [005]   283.988443: bpf:bpf_map_lookup_elem: map type=ARRAY ufd=4 key=[06 00 00 00] val=[00 00 00 00 00 00 00 00]
  [...]
  sock_example  6197 [005]   288.990868: bpf:bpf_map_lookup_elem: map type=ARRAY ufd=4 key=[01 00 00 00] val=[14 00 00 00 00 00 00 00]
       swapper     0 [005]   289.338243:    bpf:bpf_prog_put_rcu: prog=a5ea8fa30ea6849c type=SOCKET_FILTER

  [1] https://lwn.net/Articles/705270/

Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Alexei Starovoitov <ast@kernel.org>
---
 drivers/net/ethernet/mellanox/mlx4/en_rx.c         |   3 +
 drivers/net/ethernet/mellanox/mlx5/core/en_rx.c    |  12 +-
 .../net/ethernet/netronome/nfp/nfp_net_common.c    |  15 +-
 drivers/net/ethernet/qlogic/qede/qede_fp.c         |   4 +
 drivers/net/virtio_net.c                           |  12 +-
 include/linux/bpf_trace.h                          |   7 +
 include/trace/events/bpf.h                         | 347 +++++++++++++++++++++
 include/trace/events/xdp.h                         |  53 ++++
 kernel/bpf/core.c                                  |   9 +
 kernel/bpf/inode.c                                 |  17 +-
 kernel/bpf/syscall.c                               |  19 +-
 11 files changed, 483 insertions(+), 15 deletions(-)
 create mode 100644 include/linux/bpf_trace.h
 create mode 100644 include/trace/events/bpf.h
 create mode 100644 include/trace/events/xdp.h

diff --git a/drivers/net/ethernet/mellanox/mlx4/en_rx.c b/drivers/net/ethernet/mellanox/mlx4/en_rx.c
index e362f99..f15ddba 100644
--- a/drivers/net/ethernet/mellanox/mlx4/en_rx.c
+++ b/drivers/net/ethernet/mellanox/mlx4/en_rx.c
@@ -33,6 +33,7 @@
 
 #include <net/busy_poll.h>
 #include <linux/bpf.h>
+#include <linux/bpf_trace.h>
 #include <linux/mlx4/cq.h>
 #include <linux/slab.h>
 #include <linux/mlx4/qp.h>
@@ -926,10 +927,12 @@ int mlx4_en_process_rx_cq(struct net_device *dev, struct mlx4_en_cq *cq, int bud
 							length, cq->ring,
 							&doorbell_pending)))
 					goto consumed;
+				trace_xdp_exception(dev, xdp_prog, act);
 				goto xdp_drop_no_cnt; /* Drop on xmit failure */
 			default:
 				bpf_warn_invalid_xdp_action(act);
 			case XDP_ABORTED:
+				trace_xdp_exception(dev, xdp_prog, act);
 			case XDP_DROP:
 				ring->xdp_drop++;
 xdp_drop_no_cnt:
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c b/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c
index 20f116f..b88297c 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c
@@ -33,6 +33,7 @@
 #include <linux/ip.h>
 #include <linux/ipv6.h>
 #include <linux/tcp.h>
+#include <linux/bpf_trace.h>
 #include <net/busy_poll.h>
 #include "en.h"
 #include "en_tc.h"
@@ -644,7 +645,7 @@ static inline void mlx5e_xmit_xdp_doorbell(struct mlx5e_sq *sq)
 	mlx5e_tx_notify_hw(sq, &wqe->ctrl, 0);
 }
 
-static inline void mlx5e_xmit_xdp_frame(struct mlx5e_rq *rq,
+static inline bool mlx5e_xmit_xdp_frame(struct mlx5e_rq *rq,
 					struct mlx5e_dma_info *di,
 					const struct xdp_buff *xdp)
 {
@@ -666,7 +667,7 @@ static inline void mlx5e_xmit_xdp_frame(struct mlx5e_rq *rq,
 		     MLX5E_SW2HW_MTU(rq->netdev->mtu) < dma_len)) {
 		rq->stats.xdp_drop++;
 		mlx5e_page_release(rq, di, true);
-		return;
+		return false;
 	}
 
 	if (unlikely(!mlx5e_sq_has_room_for(sq, MLX5E_XDP_TX_WQEBBS))) {
@@ -677,7 +678,7 @@ static inline void mlx5e_xmit_xdp_frame(struct mlx5e_rq *rq,
 		}
 		rq->stats.xdp_tx_full++;
 		mlx5e_page_release(rq, di, true);
-		return;
+		return false;
 	}
 
 	dma_len -= MLX5E_XDP_MIN_INLINE;
@@ -707,6 +708,7 @@ static inline void mlx5e_xmit_xdp_frame(struct mlx5e_rq *rq,
 
 	sq->db.xdp.doorbell = true;
 	rq->stats.xdp_tx++;
+	return true;
 }
 
 /* returns true if packet was consumed by xdp */
@@ -732,11 +734,13 @@ static inline int mlx5e_xdp_handle(struct mlx5e_rq *rq,
 		*len = xdp.data_end - xdp.data;
 		return false;
 	case XDP_TX:
-		mlx5e_xmit_xdp_frame(rq, di, &xdp);
+		if (unlikely(!mlx5e_xmit_xdp_frame(rq, di, &xdp)))
+			trace_xdp_exception(rq->netdev, prog, act);
 		return true;
 	default:
 		bpf_warn_invalid_xdp_action(act);
 	case XDP_ABORTED:
+		trace_xdp_exception(rq->netdev, prog, act);
 	case XDP_DROP:
 		rq->stats.xdp_drop++;
 		mlx5e_page_release(rq, di, true);
diff --git a/drivers/net/ethernet/netronome/nfp/nfp_net_common.c b/drivers/net/ethernet/netronome/nfp/nfp_net_common.c
index 67afd95..6ac43ab 100644
--- a/drivers/net/ethernet/netronome/nfp/nfp_net_common.c
+++ b/drivers/net/ethernet/netronome/nfp/nfp_net_common.c
@@ -42,6 +42,7 @@
  */
 
 #include <linux/bpf.h>
+#include <linux/bpf_trace.h>
 #include <linux/module.h>
 #include <linux/kernel.h>
 #include <linux/init.h>
@@ -1459,7 +1460,7 @@ static void nfp_net_set_hash(struct net_device *netdev, struct sk_buff *skb,
 		dev_kfree_skb_any(skb);
 }
 
-static void
+static bool
 nfp_net_tx_xdp_buf(struct nfp_net *nn, struct nfp_net_rx_ring *rx_ring,
 		   struct nfp_net_tx_ring *tx_ring,
 		   struct nfp_net_rx_buf *rxbuf, unsigned int pkt_off,
@@ -1473,13 +1474,13 @@ static void nfp_net_set_hash(struct net_device *netdev, struct sk_buff *skb,
 
 	if (unlikely(nfp_net_tx_full(tx_ring, 1))) {
 		nfp_net_rx_drop(rx_ring->r_vec, rx_ring, rxbuf, NULL);
-		return;
+		return false;
 	}
 
 	new_frag = nfp_net_napi_alloc_one(nn, DMA_BIDIRECTIONAL, &new_dma_addr);
 	if (unlikely(!new_frag)) {
 		nfp_net_rx_drop(rx_ring->r_vec, rx_ring, rxbuf, NULL);
-		return;
+		return false;
 	}
 	nfp_net_rx_give_one(rx_ring, new_frag, new_dma_addr);
 
@@ -1509,6 +1510,7 @@ static void nfp_net_set_hash(struct net_device *netdev, struct sk_buff *skb,
 
 	tx_ring->wr_p++;
 	tx_ring->wr_ptr_add++;
+	return true;
 }
 
 static int nfp_net_run_xdp(struct bpf_prog *prog, void *data, unsigned int len)
@@ -1613,12 +1615,15 @@ static int nfp_net_rx(struct nfp_net_rx_ring *rx_ring, int budget)
 			case XDP_PASS:
 				break;
 			case XDP_TX:
-				nfp_net_tx_xdp_buf(nn, rx_ring, tx_ring, rxbuf,
-						   pkt_off, pkt_len);
+				if (unlikely(!nfp_net_tx_xdp_buf(nn, rx_ring,
+								 tx_ring, rxbuf,
+								 pkt_off, pkt_len)))
+					trace_xdp_exception(nn->netdev, xdp_prog, act);
 				continue;
 			default:
 				bpf_warn_invalid_xdp_action(act);
 			case XDP_ABORTED:
+				trace_xdp_exception(nn->netdev, xdp_prog, act);
 			case XDP_DROP:
 				nfp_net_rx_give_one(rx_ring, rxbuf->frag,
 						    rxbuf->dma_addr);
diff --git a/drivers/net/ethernet/qlogic/qede/qede_fp.c b/drivers/net/ethernet/qlogic/qede/qede_fp.c
index 1a6ca48..445d4d2 100644
--- a/drivers/net/ethernet/qlogic/qede/qede_fp.c
+++ b/drivers/net/ethernet/qlogic/qede/qede_fp.c
@@ -32,6 +32,7 @@
 #include <linux/netdevice.h>
 #include <linux/etherdevice.h>
 #include <linux/skbuff.h>
+#include <linux/bpf_trace.h>
 #include <net/udp_tunnel.h>
 #include <linux/ip.h>
 #include <net/ipv6.h>
@@ -1016,6 +1017,7 @@ static bool qede_rx_xdp(struct qede_dev *edev,
 		/* We need the replacement buffer before transmit. */
 		if (qede_alloc_rx_buffer(rxq, true)) {
 			qede_recycle_rx_bd_ring(rxq, 1);
+			trace_xdp_exception(edev->ndev, prog, act);
 			return false;
 		}
 
@@ -1026,6 +1028,7 @@ static bool qede_rx_xdp(struct qede_dev *edev,
 			dma_unmap_page(rxq->dev, bd->mapping,
 				       PAGE_SIZE, DMA_BIDIRECTIONAL);
 			__free_page(bd->data);
+			trace_xdp_exception(edev->ndev, prog, act);
 		}
 
 		/* Regardless, we've consumed an Rx BD */
@@ -1035,6 +1038,7 @@ static bool qede_rx_xdp(struct qede_dev *edev,
 	default:
 		bpf_warn_invalid_xdp_action(act);
 	case XDP_ABORTED:
+		trace_xdp_exception(edev->ndev, prog, act);
 	case XDP_DROP:
 		qede_recycle_rx_bd_ring(rxq, cqe->bd_num);
 	}
diff --git a/drivers/net/virtio_net.c b/drivers/net/virtio_net.c
index 37db91d..f9bf948 100644
--- a/drivers/net/virtio_net.c
+++ b/drivers/net/virtio_net.c
@@ -23,6 +23,7 @@
 #include <linux/virtio.h>
 #include <linux/virtio_net.h>
 #include <linux/bpf.h>
+#include <linux/bpf_trace.h>
 #include <linux/scatterlist.h>
 #include <linux/if_vlan.h>
 #include <linux/slab.h>
@@ -330,7 +331,7 @@ static struct sk_buff *page_to_skb(struct virtnet_info *vi,
 	return skb;
 }
 
-static void virtnet_xdp_xmit(struct virtnet_info *vi,
+static bool virtnet_xdp_xmit(struct virtnet_info *vi,
 			     struct receive_queue *rq,
 			     struct send_queue *sq,
 			     struct xdp_buff *xdp,
@@ -382,10 +383,12 @@ static void virtnet_xdp_xmit(struct virtnet_info *vi,
 			put_page(page);
 		} else /* small buffer */
 			kfree_skb(data);
-		return; // On error abort to avoid unnecessary kick
+		/* On error abort to avoid unnecessary kick */
+		return false;
 	}
 
 	virtqueue_kick(sq->vq);
+	return true;
 }
 
 static u32 do_xdp_prog(struct virtnet_info *vi,
@@ -421,11 +424,14 @@ static u32 do_xdp_prog(struct virtnet_info *vi,
 			vi->xdp_queue_pairs +
 			smp_processor_id();
 		xdp.data = buf;
-		virtnet_xdp_xmit(vi, rq, &vi->sq[qp], &xdp, data);
+		if (unlikely(!virtnet_xdp_xmit(vi, rq, &vi->sq[qp], &xdp,
+					       data)))
+			trace_xdp_exception(vi->dev, xdp_prog, act);
 		return XDP_TX;
 	default:
 		bpf_warn_invalid_xdp_action(act);
 	case XDP_ABORTED:
+		trace_xdp_exception(vi->dev, xdp_prog, act);
 	case XDP_DROP:
 		return XDP_DROP;
 	}
diff --git a/include/linux/bpf_trace.h b/include/linux/bpf_trace.h
new file mode 100644
index 0000000..b22efbd
--- /dev/null
+++ b/include/linux/bpf_trace.h
@@ -0,0 +1,7 @@
+#ifndef __LINUX_BPF_TRACE_H__
+#define __LINUX_BPF_TRACE_H__
+
+#include <trace/events/bpf.h>
+#include <trace/events/xdp.h>
+
+#endif /* __LINUX_BPF_TRACE_H__ */
diff --git a/include/trace/events/bpf.h b/include/trace/events/bpf.h
new file mode 100644
index 0000000..c3a53fd
--- /dev/null
+++ b/include/trace/events/bpf.h
@@ -0,0 +1,347 @@
+#undef TRACE_SYSTEM
+#define TRACE_SYSTEM bpf
+
+#if !defined(_TRACE_BPF_H) || defined(TRACE_HEADER_MULTI_READ)
+#define _TRACE_BPF_H
+
+#include <linux/filter.h>
+#include <linux/bpf.h>
+#include <linux/fs.h>
+#include <linux/tracepoint.h>
+
+#define __PROG_TYPE_MAP(FN)	\
+	FN(SOCKET_FILTER)	\
+	FN(KPROBE)		\
+	FN(SCHED_CLS)		\
+	FN(SCHED_ACT)		\
+	FN(TRACEPOINT)		\
+	FN(XDP)			\
+	FN(PERF_EVENT)		\
+	FN(CGROUP_SKB)		\
+	FN(CGROUP_SOCK)		\
+	FN(LWT_IN)		\
+	FN(LWT_OUT)		\
+	FN(LWT_XMIT)
+
+#define __MAP_TYPE_MAP(FN)	\
+	FN(HASH)		\
+	FN(ARRAY)		\
+	FN(PROG_ARRAY)		\
+	FN(PERF_EVENT_ARRAY)	\
+	FN(PERCPU_HASH)		\
+	FN(PERCPU_ARRAY)	\
+	FN(STACK_TRACE)		\
+	FN(CGROUP_ARRAY)	\
+	FN(LRU_HASH)		\
+	FN(LRU_PERCPU_HASH)	\
+	FN(LPM_TRIE)
+
+#define __PROG_TYPE_TP_FN(x)	\
+	TRACE_DEFINE_ENUM(BPF_PROG_TYPE_##x);
+#define __PROG_TYPE_SYM_FN(x)	\
+	{ BPF_PROG_TYPE_##x, #x },
+#define __PROG_TYPE_SYM_TAB	\
+	__PROG_TYPE_MAP(__PROG_TYPE_SYM_FN) { -1, 0 }
+__PROG_TYPE_MAP(__PROG_TYPE_TP_FN)
+
+#define __MAP_TYPE_TP_FN(x)	\
+	TRACE_DEFINE_ENUM(BPF_MAP_TYPE_##x);
+#define __MAP_TYPE_SYM_FN(x)	\
+	{ BPF_MAP_TYPE_##x, #x },
+#define __MAP_TYPE_SYM_TAB	\
+	__MAP_TYPE_MAP(__MAP_TYPE_SYM_FN) { -1, 0 }
+__MAP_TYPE_MAP(__MAP_TYPE_TP_FN)
+
+DECLARE_EVENT_CLASS(bpf_prog_event,
+
+	TP_PROTO(const struct bpf_prog *prg),
+
+	TP_ARGS(prg),
+
+	TP_STRUCT__entry(
+		__array(u8, prog_tag, 8)
+		__field(u32, type)
+	),
+
+	TP_fast_assign(
+		BUILD_BUG_ON(sizeof(__entry->prog_tag) != sizeof(prg->tag));
+		memcpy(__entry->prog_tag, prg->tag, sizeof(prg->tag));
+		__entry->type = prg->type;
+	),
+
+	TP_printk("prog=%s type=%s",
+		  __print_hex_str(__entry->prog_tag, 8),
+		  __print_symbolic(__entry->type, __PROG_TYPE_SYM_TAB))
+);
+
+DEFINE_EVENT(bpf_prog_event, bpf_prog_get_type,
+
+	TP_PROTO(const struct bpf_prog *prg),
+
+	TP_ARGS(prg)
+);
+
+DEFINE_EVENT(bpf_prog_event, bpf_prog_put_rcu,
+
+	TP_PROTO(const struct bpf_prog *prg),
+
+	TP_ARGS(prg)
+);
+
+TRACE_EVENT(bpf_prog_load,
+
+	TP_PROTO(const struct bpf_prog *prg, int ufd),
+
+	TP_ARGS(prg, ufd),
+
+	TP_STRUCT__entry(
+		__array(u8, prog_tag, 8)
+		__field(u32, type)
+		__field(int, ufd)
+	),
+
+	TP_fast_assign(
+		BUILD_BUG_ON(sizeof(__entry->prog_tag) != sizeof(prg->tag));
+		memcpy(__entry->prog_tag, prg->tag, sizeof(prg->tag));
+		__entry->type = prg->type;
+		__entry->ufd  = ufd;
+	),
+
+	TP_printk("prog=%s type=%s ufd=%d",
+		  __print_hex_str(__entry->prog_tag, 8),
+		  __print_symbolic(__entry->type, __PROG_TYPE_SYM_TAB),
+		  __entry->ufd)
+);
+
+TRACE_EVENT(bpf_map_create,
+
+	TP_PROTO(const struct bpf_map *map, int ufd),
+
+	TP_ARGS(map, ufd),
+
+	TP_STRUCT__entry(
+		__field(u32, type)
+		__field(u32, size_key)
+		__field(u32, size_value)
+		__field(u32, max_entries)
+		__field(u32, flags)
+		__field(int, ufd)
+	),
+
+	TP_fast_assign(
+		__entry->type        = map->map_type;
+		__entry->size_key    = map->key_size;
+		__entry->size_value  = map->value_size;
+		__entry->max_entries = map->max_entries;
+		__entry->flags       = map->map_flags;
+		__entry->ufd         = ufd;
+	),
+
+	TP_printk("map type=%s ufd=%d key=%u val=%u max=%u flags=%x",
+		  __print_symbolic(__entry->type, __MAP_TYPE_SYM_TAB),
+		  __entry->ufd, __entry->size_key, __entry->size_value,
+		  __entry->max_entries, __entry->flags)
+);
+
+DECLARE_EVENT_CLASS(bpf_obj_prog,
+
+	TP_PROTO(const struct bpf_prog *prg, int ufd,
+		 const struct filename *pname),
+
+	TP_ARGS(prg, ufd, pname),
+
+	TP_STRUCT__entry(
+		__array(u8, prog_tag, 8)
+		__field(int, ufd)
+		__string(path, pname->name)
+	),
+
+	TP_fast_assign(
+		BUILD_BUG_ON(sizeof(__entry->prog_tag) != sizeof(prg->tag));
+		memcpy(__entry->prog_tag, prg->tag, sizeof(prg->tag));
+		__assign_str(path, pname->name);
+		__entry->ufd = ufd;
+	),
+
+	TP_printk("prog=%s path=%s ufd=%d",
+		  __print_hex_str(__entry->prog_tag, 8),
+		  __get_str(path), __entry->ufd)
+);
+
+DEFINE_EVENT(bpf_obj_prog, bpf_obj_pin_prog,
+
+	TP_PROTO(const struct bpf_prog *prg, int ufd,
+		 const struct filename *pname),
+
+	TP_ARGS(prg, ufd, pname)
+);
+
+DEFINE_EVENT(bpf_obj_prog, bpf_obj_get_prog,
+
+	TP_PROTO(const struct bpf_prog *prg, int ufd,
+		 const struct filename *pname),
+
+	TP_ARGS(prg, ufd, pname)
+);
+
+DECLARE_EVENT_CLASS(bpf_obj_map,
+
+	TP_PROTO(const struct bpf_map *map, int ufd,
+		 const struct filename *pname),
+
+	TP_ARGS(map, ufd, pname),
+
+	TP_STRUCT__entry(
+		__field(u32, type)
+		__field(int, ufd)
+		__string(path, pname->name)
+	),
+
+	TP_fast_assign(
+		__assign_str(path, pname->name);
+		__entry->type = map->map_type;
+		__entry->ufd  = ufd;
+	),
+
+	TP_printk("map type=%s ufd=%d path=%s",
+		  __print_symbolic(__entry->type, __MAP_TYPE_SYM_TAB),
+		  __entry->ufd, __get_str(path))
+);
+
+DEFINE_EVENT(bpf_obj_map, bpf_obj_pin_map,
+
+	TP_PROTO(const struct bpf_map *map, int ufd,
+		 const struct filename *pname),
+
+	TP_ARGS(map, ufd, pname)
+);
+
+DEFINE_EVENT(bpf_obj_map, bpf_obj_get_map,
+
+	TP_PROTO(const struct bpf_map *map, int ufd,
+		 const struct filename *pname),
+
+	TP_ARGS(map, ufd, pname)
+);
+
+DECLARE_EVENT_CLASS(bpf_map_keyval,
+
+	TP_PROTO(const struct bpf_map *map, int ufd,
+		 const void *key, const void *val),
+
+	TP_ARGS(map, ufd, key, val),
+
+	TP_STRUCT__entry(
+		__field(u32, type)
+		__field(u32, key_len)
+		__dynamic_array(u8, key, map->key_size)
+		__field(bool, key_trunc)
+		__field(u32, val_len)
+		__dynamic_array(u8, val, map->value_size)
+		__field(bool, val_trunc)
+		__field(int, ufd)
+	),
+
+	TP_fast_assign(
+		memcpy(__get_dynamic_array(key), key, map->key_size);
+		memcpy(__get_dynamic_array(val), val, map->value_size);
+		__entry->type      = map->map_type;
+		__entry->key_len   = min(map->key_size, 16U);
+		__entry->key_trunc = map->key_size != __entry->key_len;
+		__entry->val_len   = min(map->value_size, 16U);
+		__entry->val_trunc = map->value_size != __entry->val_len;
+		__entry->ufd       = ufd;
+	),
+
+	TP_printk("map type=%s ufd=%d key=[%s%s] val=[%s%s]",
+		  __print_symbolic(__entry->type, __MAP_TYPE_SYM_TAB),
+		  __entry->ufd,
+		  __print_hex(__get_dynamic_array(key), __entry->key_len),
+		  __entry->key_trunc ? " ..." : "",
+		  __print_hex(__get_dynamic_array(val), __entry->val_len),
+		  __entry->val_trunc ? " ..." : "")
+);
+
+DEFINE_EVENT(bpf_map_keyval, bpf_map_lookup_elem,
+
+	TP_PROTO(const struct bpf_map *map, int ufd,
+		 const void *key, const void *val),
+
+	TP_ARGS(map, ufd, key, val)
+);
+
+DEFINE_EVENT(bpf_map_keyval, bpf_map_update_elem,
+
+	TP_PROTO(const struct bpf_map *map, int ufd,
+		 const void *key, const void *val),
+
+	TP_ARGS(map, ufd, key, val)
+);
+
+TRACE_EVENT(bpf_map_delete_elem,
+
+	TP_PROTO(const struct bpf_map *map, int ufd,
+		 const void *key),
+
+	TP_ARGS(map, ufd, key),
+
+	TP_STRUCT__entry(
+		__field(u32, type)
+		__field(u32, key_len)
+		__dynamic_array(u8, key, map->key_size)
+		__field(bool, key_trunc)
+		__field(int, ufd)
+	),
+
+	TP_fast_assign(
+		memcpy(__get_dynamic_array(key), key, map->key_size);
+		__entry->type      = map->map_type;
+		__entry->key_len   = min(map->key_size, 16U);
+		__entry->key_trunc = map->key_size != __entry->key_len;
+		__entry->ufd       = ufd;
+	),
+
+	TP_printk("map type=%s ufd=%d key=[%s%s]",
+		  __print_symbolic(__entry->type, __MAP_TYPE_SYM_TAB),
+		  __entry->ufd,
+		  __print_hex(__get_dynamic_array(key), __entry->key_len),
+		  __entry->key_trunc ? " ..." : "")
+);
+
+TRACE_EVENT(bpf_map_next_key,
+
+	TP_PROTO(const struct bpf_map *map, int ufd,
+		 const void *key, const void *key_next),
+
+	TP_ARGS(map, ufd, key, key_next),
+
+	TP_STRUCT__entry(
+		__field(u32, type)
+		__field(u32, key_len)
+		__dynamic_array(u8, key, map->key_size)
+		__dynamic_array(u8, nxt, map->key_size)
+		__field(bool, key_trunc)
+		__field(int, ufd)
+	),
+
+	TP_fast_assign(
+		memcpy(__get_dynamic_array(key), key, map->key_size);
+		memcpy(__get_dynamic_array(nxt), key_next, map->key_size);
+		__entry->type      = map->map_type;
+		__entry->key_len   = min(map->key_size, 16U);
+		__entry->key_trunc = map->key_size != __entry->key_len;
+		__entry->ufd       = ufd;
+	),
+
+	TP_printk("map type=%s ufd=%d key=[%s%s] next=[%s%s]",
+		  __print_symbolic(__entry->type, __MAP_TYPE_SYM_TAB),
+		  __entry->ufd,
+		  __print_hex(__get_dynamic_array(key), __entry->key_len),
+		  __entry->key_trunc ? " ..." : "",
+		  __print_hex(__get_dynamic_array(nxt), __entry->key_len),
+		  __entry->key_trunc ? " ..." : "")
+);
+
+#endif /* _TRACE_BPF_H */
+
+#include <trace/define_trace.h>
diff --git a/include/trace/events/xdp.h b/include/trace/events/xdp.h
new file mode 100644
index 0000000..1b61357
--- /dev/null
+++ b/include/trace/events/xdp.h
@@ -0,0 +1,53 @@
+#undef TRACE_SYSTEM
+#define TRACE_SYSTEM xdp
+
+#if !defined(_TRACE_XDP_H) || defined(TRACE_HEADER_MULTI_READ)
+#define _TRACE_XDP_H
+
+#include <linux/netdevice.h>
+#include <linux/filter.h>
+#include <linux/tracepoint.h>
+
+#define __XDP_ACT_MAP(FN)	\
+	FN(ABORTED)		\
+	FN(DROP)		\
+	FN(PASS)		\
+	FN(TX)
+
+#define __XDP_ACT_TP_FN(x)	\
+	TRACE_DEFINE_ENUM(XDP_##x);
+#define __XDP_ACT_SYM_FN(x)	\
+	{ XDP_##x, #x },
+#define __XDP_ACT_SYM_TAB	\
+	__XDP_ACT_MAP(__XDP_ACT_SYM_FN) { -1, 0 }
+__XDP_ACT_MAP(__XDP_ACT_TP_FN)
+
+TRACE_EVENT(xdp_exception,
+
+	TP_PROTO(const struct net_device *dev,
+		 const struct bpf_prog *xdp, u32 act),
+
+	TP_ARGS(dev, xdp, act),
+
+	TP_STRUCT__entry(
+		__string(name, dev->name)
+		__array(u8, prog_tag, 8)
+		__field(u32, act)
+	),
+
+	TP_fast_assign(
+		BUILD_BUG_ON(sizeof(__entry->prog_tag) != sizeof(xdp->tag));
+		memcpy(__entry->prog_tag, xdp->tag, sizeof(xdp->tag));
+		__assign_str(name, dev->name);
+		__entry->act = act;
+	),
+
+	TP_printk("prog=%s device=%s action=%s",
+		  __print_hex_str(__entry->prog_tag, 8),
+		  __get_str(name),
+		  __print_symbolic(__entry->act, __XDP_ACT_SYM_TAB))
+);
+
+#endif /* _TRACE_XDP_H */
+
+#include <trace/define_trace.h>
diff --git a/kernel/bpf/core.c b/kernel/bpf/core.c
index 503d421..fddd76b 100644
--- a/kernel/bpf/core.c
+++ b/kernel/bpf/core.c
@@ -1173,3 +1173,12 @@ int __weak skb_copy_bits(const struct sk_buff *skb, int offset, void *to,
 {
 	return -EFAULT;
 }
+
+/* All definitions of tracepoints related to BPF. */
+#define CREATE_TRACE_POINTS
+#include <linux/bpf_trace.h>
+
+EXPORT_TRACEPOINT_SYMBOL_GPL(xdp_exception);
+
+EXPORT_TRACEPOINT_SYMBOL_GPL(bpf_prog_get_type);
+EXPORT_TRACEPOINT_SYMBOL_GPL(bpf_prog_put_rcu);
diff --git a/kernel/bpf/inode.c b/kernel/bpf/inode.c
index 0b030c9..fddcae8 100644
--- a/kernel/bpf/inode.c
+++ b/kernel/bpf/inode.c
@@ -21,6 +21,7 @@
 #include <linux/parser.h>
 #include <linux/filter.h>
 #include <linux/bpf.h>
+#include <linux/bpf_trace.h>
 
 enum bpf_type {
 	BPF_TYPE_UNSPEC	= 0,
@@ -281,6 +282,13 @@ int bpf_obj_pin_user(u32 ufd, const char __user *pathname)
 	ret = bpf_obj_do_pin(pname, raw, type);
 	if (ret != 0)
 		bpf_any_put(raw, type);
+	if ((trace_bpf_obj_pin_prog_enabled() ||
+	     trace_bpf_obj_pin_map_enabled()) && !ret) {
+		if (type == BPF_TYPE_PROG)
+			trace_bpf_obj_pin_prog(raw, ufd, pname);
+		if (type == BPF_TYPE_MAP)
+			trace_bpf_obj_pin_map(raw, ufd, pname);
+	}
 out:
 	putname(pname);
 	return ret;
@@ -342,8 +350,15 @@ int bpf_obj_get_user(const char __user *pathname)
 	else
 		goto out;
 
-	if (ret < 0)
+	if (ret < 0) {
 		bpf_any_put(raw, type);
+	} else if (trace_bpf_obj_get_prog_enabled() ||
+		   trace_bpf_obj_get_map_enabled()) {
+		if (type == BPF_TYPE_PROG)
+			trace_bpf_obj_get_prog(raw, ret, pname);
+		if (type == BPF_TYPE_MAP)
+			trace_bpf_obj_get_map(raw, ret, pname);
+	}
 out:
 	putname(pname);
 	return ret;
diff --git a/kernel/bpf/syscall.c b/kernel/bpf/syscall.c
index 1d6b29e..05ad086 100644
--- a/kernel/bpf/syscall.c
+++ b/kernel/bpf/syscall.c
@@ -10,6 +10,7 @@
  * General Public License for more details.
  */
 #include <linux/bpf.h>
+#include <linux/bpf_trace.h>
 #include <linux/syscalls.h>
 #include <linux/slab.h>
 #include <linux/anon_inodes.h>
@@ -215,6 +216,7 @@ static int map_create(union bpf_attr *attr)
 		/* failed to allocate fd */
 		goto free_map;
 
+	trace_bpf_map_create(map, err);
 	return err;
 
 free_map:
@@ -339,6 +341,7 @@ static int map_lookup_elem(union bpf_attr *attr)
 	if (copy_to_user(uvalue, value, value_size) != 0)
 		goto free_value;
 
+	trace_bpf_map_lookup_elem(map, ufd, key, value);
 	err = 0;
 
 free_value:
@@ -421,6 +424,8 @@ static int map_update_elem(union bpf_attr *attr)
 	__this_cpu_dec(bpf_prog_active);
 	preempt_enable();
 
+	if (!err)
+		trace_bpf_map_update_elem(map, ufd, key, value);
 free_value:
 	kfree(value);
 free_key:
@@ -466,6 +471,8 @@ static int map_delete_elem(union bpf_attr *attr)
 	__this_cpu_dec(bpf_prog_active);
 	preempt_enable();
 
+	if (!err)
+		trace_bpf_map_delete_elem(map, ufd, key);
 free_key:
 	kfree(key);
 err_put:
@@ -518,6 +525,7 @@ static int map_get_next_key(union bpf_attr *attr)
 	if (copy_to_user(unext_key, next_key, map->key_size) != 0)
 		goto free_next_key;
 
+	trace_bpf_map_next_key(map, ufd, key, next_key);
 	err = 0;
 
 free_next_key:
@@ -671,8 +679,10 @@ static void __bpf_prog_put_rcu(struct rcu_head *rcu)
 
 void bpf_prog_put(struct bpf_prog *prog)
 {
-	if (atomic_dec_and_test(&prog->aux->refcnt))
+	if (atomic_dec_and_test(&prog->aux->refcnt)) {
+		trace_bpf_prog_put_rcu(prog);
 		call_rcu(&prog->aux->rcu, __bpf_prog_put_rcu);
+	}
 }
 EXPORT_SYMBOL_GPL(bpf_prog_put);
 
@@ -781,7 +791,11 @@ struct bpf_prog *bpf_prog_get(u32 ufd)
 
 struct bpf_prog *bpf_prog_get_type(u32 ufd, enum bpf_prog_type type)
 {
-	return __bpf_prog_get(ufd, &type);
+	struct bpf_prog *prog = __bpf_prog_get(ufd, &type);
+
+	if (!IS_ERR(prog))
+		trace_bpf_prog_get_type(prog);
+	return prog;
 }
 EXPORT_SYMBOL_GPL(bpf_prog_get_type);
 
@@ -863,6 +877,7 @@ static int bpf_prog_load(union bpf_attr *attr)
 		/* failed to allocate fd */
 		goto free_used_maps;
 
+	trace_bpf_prog_load(prog, err);
 	return err;
 
 free_used_maps:
-- 
1.9.3

Re: [PATCH net-next 0/3] BPF tracepoints

[David Miller]

From: Daniel Borkmann <daniel@iogearbox.net>
Date: Wed, 25 Jan 2017 02:28:15 +0100

> This set adds tracepoints to BPF for better introspection and
> debugging. The first two patches are prerequisite for the actual
> third patch that adds the tracepoints. I think the first two are
> small and straight forward enough that they could ideally go via
> net-next, but I'm also open to other suggestions on how to route
> them in case that's not applicable (it would reduce potential
> merge conflicts on BPF side, though). For details, please see
> individual patches.

Looks great, series applied, thanks Daniel.