Skip to contentSkip to conversation

libperf: Add events to perf/event.h

PATCH Applied Jiri Olsa 4 participants 33 messages Aug 25, 2019 – Aug 27, 2019 perf events

lore
Share & tools
lore

Jiri Olsa

author reviewer
#0 lore raw

cover libperf: Add events to perf/event.h

hi,
as a preparation for sampling libperf interface, moving event definitions into the library header. Moving just the kernel non-AUX events now.

In order to keep libperf simple, we switch 'u64/u32/u16/u8' types used events to their generic '__u*' versions.

Perf added 'u*' types mainly to ease up printing __u64 values as stated in the linux/types.h comment:

/*
* We define u64 as uint64_t for every architecture
* so that we can print it with "%"PRIx64 without getting warnings. *
* typedef __u64 u64;
* typedef __s64 s64;
*/

Adding and using new PRI_lu64 and PRI_lx64 macros to be used for that. Using extra '_' to ease up the reading and differentiate them from standard PRI*64 macros.

It's also available in here:
git://git.kernel.org/pub/scm/linux/kernel/git/jolsa/perf.git
perf/fixes

thanks,
jirka

---
Jiri Olsa (12):
libperf: Add mmap_event to perf/event.h
libperf: Add mmap2_event to perf/event.h
libperf: Add comm_event to perf/event.h
libperf: Add namespaces_event to perf/event.h
libperf: Add fork_event to perf/event.h
libperf: Add lost_event to perf/event.h
libperf: Add lost_samples_event to perf/event.h
libperf: Add read_event to perf/event.h
libperf: Add throttle_event to perf/event.h
libperf: Add ksymbol_event to perf/event.h
libperf: Add bpf_event to perf/event.h
libperf: Add sample_event to perf/event.h

 tools/perf/builtin-sched.c          |   2 +-
 tools/perf/lib/include/perf/event.h | 112 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
 tools/perf/util/event.c             |  12 ++++++------
 tools/perf/util/event.h             | 104 +++-----------------------------------------------------------------------------------------------------
 tools/perf/util/evlist.c            |   2 +-
 tools/perf/util/evsel.c             |   8 ++++----
 tools/perf/util/machine.c           |   4 ++--
 tools/perf/util/python.c            |  14 +++++++-------
 tools/perf/util/session.c           |   8 ++++----
 9 files changed, 140 insertions(+), 126 deletions(-)
 create mode 100644 tools/perf/lib/include/perf/event.h
author reviewer
#1 lore raw

PATCH 1/12 libperf: Add mmap_event to perf/event.h

Moving mmap_event event definition into libperf's event.h header include.

In order to keep libperf simple, we switch 'u64/u32/u16/u8' types used events to their generic '__u*' versions.

Perf added 'u*' types mainly to ease up printing __u64 values as stated in the linux/types.h comment:

/*
* We define u64 as uint64_t for every architecture
* so that we can print it with "%"PRIx64 without getting warnings. *
* typedef __u64 u64;
* typedef __s64 s64;
*/

Adding and using new PRI_lu64 and PRI_lx64 macros to be used for that. Using extra '_' to ease up the reading and differentiate them from standard PRI*64 macros.

---
 tools/perf/lib/include/perf/event.h | 18 ++++++++++++++++++
 tools/perf/util/event.c             |  2 +-
 tools/perf/util/event.h             | 11 +++--------
 tools/perf/util/python.c            |  4 ++--
 4 files changed, 24 insertions(+), 11 deletions(-)
 create mode 100644 tools/perf/lib/include/perf/event.h
tools/perf/lib/include/perf/event.h new file +18 -0
new file mode 100644
index 000000000000..8eeba16e3ff7
@@ -0,0 +1,18 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#ifndef __LIBPERF_EVENT_H
+#define __LIBPERF_EVENT_H
+
+#include <linux/perf_event.h>
+#include <linux/types.h>
+#include <linux/limits.h>
+
+struct mmap_event {
+	struct perf_event_header header;
+	__u32 pid, tid;
+	__u64 start;
+	__u64 len;
+	__u64 pgoff;
+	char filename[PATH_MAX];
+};
+
+#endif /* __LIBPERF_EVENT_H */
tools/perf/util/event.c +1 -1
index 332edef8d394..43c86257e7fa 100644
@@ -1353,7 +1353,7 @@ int perf_event__process_bpf_event(struct perf_tool *tool __maybe_unused,
 
 size_t perf_event__fprintf_mmap(union perf_event *event, FILE *fp)
 {
-	return fprintf(fp, " %d/%d: [%#" PRIx64 "(%#" PRIx64 ") @ %#" PRIx64 "]: %c %s\n",
+	return fprintf(fp, " %d/%d: [%#" PRI_lx64 "(%#" PRI_lx64 ") @ %#" PRI_lx64 "]: %c %s\n",
 		       event->mmap.pid, event->mmap.tid, event->mmap.start,
 		       event->mmap.len, event->mmap.pgoff,
 		       (event->header.misc & PERF_RECORD_MISC_MMAP_DATA) ? 'r' : 'x',
tools/perf/util/event.h +3 -8
index 0e164e8ae28d..b244f9e77fa8 100644
@@ -7,19 +7,14 @@
 #include <linux/kernel.h>
 #include <linux/bpf.h>
 #include <linux/perf_event.h>
+#include <perf/event.h>
 
 #include "../perf.h"
 #include "build-id.h"
 #include "perf_regs.h"
 
-struct mmap_event {
-	struct perf_event_header header;
-	u32 pid, tid;
-	u64 start;
-	u64 len;
-	u64 pgoff;
-	char filename[PATH_MAX];
-};
+#define PRI_lu64 "l" PRIu64
+#define PRI_lx64 "l" PRIx64
 
 struct mmap2_event {
 	struct perf_event_header header;
tools/perf/util/python.c +2 -2
index 75ecc32a4427..55ff0c3182d6 100644
@@ -130,8 +130,8 @@ static PyObject *pyrf_mmap_event__repr(struct pyrf_event *pevent)
 	PyObject *ret;
 	char *s;
 
-	if (asprintf(&s, "{ type: mmap, pid: %u, tid: %u, start: %#" PRIx64 ", "
-			 "length: %#" PRIx64 ", offset: %#" PRIx64 ", "
+	if (asprintf(&s, "{ type: mmap, pid: %u, tid: %u, start: %#" PRI_lx64 ", "
+			 "length: %#" PRI_lx64 ", offset: %#" PRI_lx64 ", "
 			 "filename: %s }",
 		     pevent->event.mmap.pid, pevent->event.mmap.tid,
 		     pevent->event.mmap.start, pevent->event.mmap.len,
Signature · 1 line
2.21.0
merge bot Applied
#2 lore raw

[tip: perf/core] libperf: Add PERF_RECORD_MMAP 'struct mmap_event' to perf/event.h

The following commit has been merged into the perf/core branch of tip:

The following commit has been merged into the perf/core branch of tip:

Commit-ID:     1345e2ee87a83c758f336f03f7fb305bc5e24490
Gitweb:        https://git.kernel.org/tip/1345e2ee87a83c758f336f03f7fb305bc5e24490
Author:        Jiri Olsa <[email protected]>
AuthorDate:    Sun, 25 Aug 2019 20:17:41 +02:00
Committer:     Arnaldo Carvalho de Melo <[email protected]>
CommitterDate: Mon, 26 Aug 2019 19:38:04 -03:00

libperf: Add PERF_RECORD_MMAP 'struct mmap_event' to perf/event.h

Move the mmap_event event definition to libperf's event.h header include.

In order to keep libperf simple, we switch 'u64/u32/u16/u8' types used events to their generic '__u*' versions.

Perf added 'u*' types mainly to ease up printing __u64 values as stated in the linux/types.h comment:

/*
* We define u64 as uint64_t for every architecture
* so that we can print it with "%"PRIx64 without getting warnings. *
* typedef __u64 u64;
* typedef __s64 s64;
*/

Add and use new PRI_lu64 and PRI_lx64 macros for that. Use extra '_' to ease up reading and differentiate them from standard PRI*64 macros.

Committer notes:

Fixup the PRI_l[ux]64 macros on 32-bit arches, conditionally defining it with that extra 'l' modifier only on arches where __u64 is long long, leaving it aside on 32-bit arches.

---
 tools/perf/lib/include/perf/event.h | 18 ++++++++++++++++++
 tools/perf/util/event.c             |  2 +-
 tools/perf/util/event.h             | 22 ++++++++++++++--------
 tools/perf/util/python.c            |  4 ++--
 4 files changed, 35 insertions(+), 11 deletions(-)
 create mode 100644 tools/perf/lib/include/perf/event.h
tools/perf/lib/include/perf/event.h new file +18 -0
new file mode 100644
index 0000000..13fe15a
@@ -0,0 +1,18 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#ifndef __LIBPERF_EVENT_H
+#define __LIBPERF_EVENT_H
+
+#include <linux/perf_event.h>
+#include <linux/types.h>
+#include <linux/limits.h>
+
+struct mmap_event {
+	struct perf_event_header header;
+	__u32			 pid, tid;
+	__u64			 start;
+	__u64			 len;
+	__u64			 pgoff;
+	char			 filename[PATH_MAX];
+};
+
+#endif /* __LIBPERF_EVENT_H */
tools/perf/util/event.c +1 -1
index 332edef..43c8625 100644
@@ -1353,7 +1353,7 @@ int perf_event__process_bpf_event(struct perf_tool *tool __maybe_unused,
 
 size_t perf_event__fprintf_mmap(union perf_event *event, FILE *fp)
 {
-	return fprintf(fp, " %d/%d: [%#" PRIx64 "(%#" PRIx64 ") @ %#" PRIx64 "]: %c %s\n",
+	return fprintf(fp, " %d/%d: [%#" PRI_lx64 "(%#" PRI_lx64 ") @ %#" PRI_lx64 "]: %c %s\n",
 		       event->mmap.pid, event->mmap.tid, event->mmap.start,
 		       event->mmap.len, event->mmap.pgoff,
 		       (event->header.misc & PERF_RECORD_MISC_MMAP_DATA) ? 'r' : 'x',
tools/perf/util/event.h +14 -8
index 0e164e8..f43eff2 100644
@@ -7,19 +7,25 @@
 #include <linux/kernel.h>
 #include <linux/bpf.h>
 #include <linux/perf_event.h>
+#include <perf/event.h>
 
 #include "../perf.h"
 #include "build-id.h"
 #include "perf_regs.h"
 
-struct mmap_event {
-	struct perf_event_header header;
-	u32 pid, tid;
-	u64 start;
-	u64 len;
-	u64 pgoff;
-	char filename[PATH_MAX];
-};
+#ifdef __LP64__
+/*
+ * /usr/include/inttypes.h uses just 'lu' for PRIu64, but we end up defining
+ * __u64 as long long unsigned int, and then -Werror=format= kicks in and
+ * complains of the mismatched types, so use these two special extra PRI
+ * macros to overcome that.
+ */
+#define PRI_lu64 "l" PRIu64
+#define PRI_lx64 "l" PRIx64
+#else
+#define PRI_lu64 PRIu64
+#define PRI_lx64 PRIx64
+#endif
 
 struct mmap2_event {
 	struct perf_event_header header;
tools/perf/util/python.c +2 -2
index 75ecc32..55ff0c3 100644
@@ -130,8 +130,8 @@ static PyObject *pyrf_mmap_event__repr(struct pyrf_event *pevent)
 	PyObject *ret;
 	char *s;
 
-	if (asprintf(&s, "{ type: mmap, pid: %u, tid: %u, start: %#" PRIx64 ", "
-			 "length: %#" PRIx64 ", offset: %#" PRIx64 ", "
+	if (asprintf(&s, "{ type: mmap, pid: %u, tid: %u, start: %#" PRI_lx64 ", "
+			 "length: %#" PRI_lx64 ", offset: %#" PRI_lx64 ", "
 			 "filename: %s }",
 		     pevent->event.mmap.pid, pevent->event.mmap.tid,
 		     pevent->event.mmap.start, pevent->event.mmap.len,
author reviewer
#3 lore raw

PATCH 2/12 libperf: Add mmap2_event to perf/event.h

In reply to Jiri Olsa

Moving mmap2_event event definition into libperf's event.h header include.

In order to keep libperf simple, we switch 'u64/u32/u16/u8' types used events to their generic '__u*' versions.

Perf added 'u*' types mainly to ease up printing __u64 values as stated in the linux/types.h comment:

/*
* We define u64 as uint64_t for every architecture
* so that we can print it with "%"PRIx64 without getting warnings. *
* typedef __u64 u64;
* typedef __s64 s64;
*/

Adding and using new PRI_lu64 and PRI_lx64 macros to be used for that. Using extra '_' to ease up the reading and differentiate them from standard PRI*64 macros.

---
 tools/perf/lib/include/perf/event.h | 15 +++++++++++++++
 tools/perf/util/event.c             |  6 +++---
 tools/perf/util/event.h             | 15 ---------------
 3 files changed, 18 insertions(+), 18 deletions(-)
tools/perf/lib/include/perf/event.h +15 -0
index 8eeba16e3ff7..f1f5b0787782 100644
@@ -15,4 +15,19 @@ struct mmap_event {
 	char filename[PATH_MAX];
 };
 
+struct mmap2_event {
+	struct perf_event_header header;
+	__u32 pid, tid;
+	__u64 start;
+	__u64 len;
+	__u64 pgoff;
+	__u32 maj;
+	__u32 min;
+	__u64 ino;
+	__u64 ino_generation;
+	__u32 prot;
+	__u32 flags;
+	char filename[PATH_MAX];
+};
+
 #endif /* __LIBPERF_EVENT_H */
tools/perf/util/event.c +3 -3
index 43c86257e7fa..0954f980574f 100644
@@ -387,7 +387,7 @@ int perf_event__synthesize_mmap_events(struct perf_tool *tool,
 		strcpy(execname, "");
 
 		/* 00400000-0040c000 r-xp 00000000 fd:01 41038  /bin/cat */
-		n = sscanf(bf, "%"PRIx64"-%"PRIx64" %s %"PRIx64" %x:%x %u %[^\n]\n",
+		n = sscanf(bf, "%"PRI_lx64"-%"PRI_lx64" %s %"PRI_lx64" %x:%x %u %[^\n]\n",
 		       &event->mmap2.start, &event->mmap2.len, prot,
 		       &event->mmap2.pgoff, &event->mmap2.maj,
 		       &event->mmap2.min,
@@ -1362,8 +1362,8 @@ size_t perf_event__fprintf_mmap(union perf_event *event, FILE *fp)
 
 size_t perf_event__fprintf_mmap2(union perf_event *event, FILE *fp)
 {
-	return fprintf(fp, " %d/%d: [%#" PRIx64 "(%#" PRIx64 ") @ %#" PRIx64
-			   " %02x:%02x %"PRIu64" %"PRIu64"]: %c%c%c%c %s\n",
+	return fprintf(fp, " %d/%d: [%#" PRI_lx64 "(%#" PRI_lx64 ") @ %#" PRI_lx64
+			   " %02x:%02x %"PRI_lu64" %"PRI_lu64"]: %c%c%c%c %s\n",
 		       event->mmap2.pid, event->mmap2.tid, event->mmap2.start,
 		       event->mmap2.len, event->mmap2.pgoff, event->mmap2.maj,
 		       event->mmap2.min, event->mmap2.ino,
tools/perf/util/event.h +0 -15
index b244f9e77fa8..fdff9dcd7c2b 100644
@@ -16,21 +16,6 @@
 #define PRI_lu64 "l" PRIu64
 #define PRI_lx64 "l" PRIx64
 
-struct mmap2_event {
-	struct perf_event_header header;
-	u32 pid, tid;
-	u64 start;
-	u64 len;
-	u64 pgoff;
-	u32 maj;
-	u32 min;
-	u64 ino;
-	u64 ino_generation;
-	u32 prot;
-	u32 flags;
-	char filename[PATH_MAX];
-};
-
 struct comm_event {
 	struct perf_event_header header;
 	u32 pid, tid;
Signature · 1 line
2.21.0
author reviewer
#4 lore raw

PATCH 3/12 libperf: Add comm_event to perf/event.h

In reply to Jiri Olsa

Moving comm_event event definition into libperf's event.h header include.

In order to keep libperf simple, we switch 'u64/u32/u16/u8' types used events to their generic '__u*' versions.

---
 tools/perf/lib/include/perf/event.h | 6 ++++++
 tools/perf/util/event.h             | 6 ------
 2 files changed, 6 insertions(+), 6 deletions(-)
tools/perf/lib/include/perf/event.h +6 -0
index f1f5b0787782..11adfb1b1524 100644
@@ -30,4 +30,10 @@ struct mmap2_event {
 	char filename[PATH_MAX];
 };
 
+struct comm_event {
+	struct perf_event_header header;
+	__u32 pid, tid;
+	char comm[16];
+};
+
 #endif /* __LIBPERF_EVENT_H */
tools/perf/util/event.h +0 -6
index fdff9dcd7c2b..12bc9de3d539 100644
@@ -16,12 +16,6 @@
 #define PRI_lu64 "l" PRIu64
 #define PRI_lx64 "l" PRIx64
 
-struct comm_event {
-	struct perf_event_header header;
-	u32 pid, tid;
-	char comm[16];
-};
-
 struct namespaces_event {
 	struct perf_event_header header;
 	u32 pid, tid;
Signature · 1 line
2.21.0
merge bot Applied
#5 lore raw

[tip: perf/core] libperf: Add PERF_RECORD_COMM 'struct comm_event' to perf/event.h

The following commit has been merged into the perf/core branch of tip:

The following commit has been merged into the perf/core branch of tip:

Commit-ID:     002dda32a831b30d22a3620eb619a0d103e78e81
Gitweb:        https://git.kernel.org/tip/002dda32a831b30d22a3620eb619a0d103e78e81
Author:        Jiri Olsa <[email protected]>
AuthorDate:    Sun, 25 Aug 2019 20:17:43 +02:00
Committer:     Arnaldo Carvalho de Melo <[email protected]>
CommitterDate: Mon, 26 Aug 2019 19:39:09 -03:00

libperf: Add PERF_RECORD_COMM 'struct comm_event' to perf/event.h

Moving comm_event event definition into libperf's event.h header include.

In order to keep libperf simple, we switch 'u64/u32/u16/u8' types used events to their generic '__u*' versions.

---
 tools/perf/lib/include/perf/event.h | 6 ++++++
 tools/perf/util/event.h             | 6 ------
 2 files changed, 6 insertions(+), 6 deletions(-)
tools/perf/lib/include/perf/event.h +6 -0
index c82e0c2..3729a7d 100644
@@ -30,4 +30,10 @@ struct mmap2_event {
 	char			 filename[PATH_MAX];
 };
 
+struct comm_event {
+	struct perf_event_header header;
+	__u32			 pid, tid;
+	char			 comm[16];
+};
+
 #endif /* __LIBPERF_EVENT_H */
tools/perf/util/event.h +0 -6
index af252be..e8973a7 100644
@@ -27,12 +27,6 @@
 #define PRI_lx64 PRIx64
 #endif
 
-struct comm_event {
-	struct perf_event_header header;
-	u32 pid, tid;
-	char comm[16];
-};
-
 struct namespaces_event {
 	struct perf_event_header header;
 	u32 pid, tid;
author reviewer
#6 lore raw

PATCH 4/12 libperf: Add namespaces_event to perf/event.h

In reply to Jiri Olsa

Moving namespaces_event event definition into libperf's event.h header include.

In order to keep libperf simple, we switch 'u64/u32/u16/u8' types used events to their generic '__u*' versions.

---
 tools/perf/lib/include/perf/event.h | 7 +++++++
 tools/perf/util/event.h             | 7 -------
 2 files changed, 7 insertions(+), 7 deletions(-)
tools/perf/lib/include/perf/event.h +7 -0
index 11adfb1b1524..265bb01029f0 100644
@@ -36,4 +36,11 @@ struct comm_event {
 	char comm[16];
 };
 
+struct namespaces_event {
+	struct perf_event_header header;
+	__u32 pid, tid;
+	__u64 nr_namespaces;
+	struct perf_ns_link_info link_info[];
+};
+
 #endif /* __LIBPERF_EVENT_H */
tools/perf/util/event.h +0 -7
index 12bc9de3d539..e8c57c23aa24 100644
@@ -16,13 +16,6 @@
 #define PRI_lu64 "l" PRIu64
 #define PRI_lx64 "l" PRIx64
 
-struct namespaces_event {
-	struct perf_event_header header;
-	u32 pid, tid;
-	u64 nr_namespaces;
-	struct perf_ns_link_info link_info[];
-};
-
 struct fork_event {
 	struct perf_event_header header;
 	u32 pid, ppid;
Signature · 1 line
2.21.0
merge bot Applied
#7 lore raw

[tip: perf/core] libperf: Add PERF_RECORD_NAMESPACES 'struct namespaces_event' to perf/event.h

The following commit has been merged into the perf/core branch of tip:

The following commit has been merged into the perf/core branch of tip:

Commit-ID:     19d1765a3ed9a8f78d93909120f6d39398809f75
Gitweb:        https://git.kernel.org/tip/19d1765a3ed9a8f78d93909120f6d39398809f75
Author:        Jiri Olsa <[email protected]>
AuthorDate:    Sun, 25 Aug 2019 20:17:44 +02:00
Committer:     Arnaldo Carvalho de Melo <[email protected]>
CommitterDate: Mon, 26 Aug 2019 19:39:09 -03:00

libperf: Add PERF_RECORD_NAMESPACES 'struct namespaces_event' to perf/event.h

Move the namespaces_event event definition into libperf's event.h header include.

In order to keep libperf simple, we switch 'u64/u32/u16/u8' types used events to their generic '__u*' versions.

---
 tools/perf/lib/include/perf/event.h | 7 +++++++
 tools/perf/util/event.h             | 7 -------
 2 files changed, 7 insertions(+), 7 deletions(-)
tools/perf/lib/include/perf/event.h +7 -0
index 3729a7d..b90a8a2 100644
@@ -36,4 +36,11 @@ struct comm_event {
 	char			 comm[16];
 };
 
+struct namespaces_event {
+	struct perf_event_header header;
+	__u32			 pid, tid;
+	__u64			 nr_namespaces;
+	struct perf_ns_link_info link_info[];
+};
+
 #endif /* __LIBPERF_EVENT_H */
tools/perf/util/event.h +0 -7
index e8973a7..0d3ac4f 100644
@@ -27,13 +27,6 @@
 #define PRI_lx64 PRIx64
 #endif
 
-struct namespaces_event {
-	struct perf_event_header header;
-	u32 pid, tid;
-	u64 nr_namespaces;
-	struct perf_ns_link_info link_info[];
-};
-
 struct fork_event {
 	struct perf_event_header header;
 	u32 pid, ppid;
author reviewer
#8 lore raw

PATCH 5/12 libperf: Add fork_event to perf/event.h

In reply to Jiri Olsa

Moving fork_event event definition into libperf's event.h header include.

In order to keep libperf simple, we switch 'u64/u32/u16/u8' types used events to their generic '__u*' versions.

Perf added 'u*' types mainly to ease up printing __u64 values as stated in the linux/types.h comment:

/*
* We define u64 as uint64_t for every architecture
* so that we can print it with "%"PRIx64 without getting warnings. *
* typedef __u64 u64;
* typedef __s64 s64;
*/

Adding and using new PRI_lu64 and PRI_lx64 macros to be used for that. Using extra '_' to ease up the reading and differentiate them from standard PRI*64 macros.

---
 tools/perf/lib/include/perf/event.h | 7 +++++++
 tools/perf/util/event.h             | 7 -------
 tools/perf/util/python.c            | 2 +-
 3 files changed, 8 insertions(+), 8 deletions(-)
tools/perf/lib/include/perf/event.h +7 -0
index 265bb01029f0..7a34221abbba 100644
@@ -43,4 +43,11 @@ struct namespaces_event {
 	struct perf_ns_link_info link_info[];
 };
 
+struct fork_event {
+	struct perf_event_header header;
+	__u32 pid, ppid;
+	__u32 tid, ptid;
+	__u64 time;
+};
+
 #endif /* __LIBPERF_EVENT_H */
tools/perf/util/event.h +0 -7
index e8c57c23aa24..817f12643f11 100644
@@ -16,13 +16,6 @@
 #define PRI_lu64 "l" PRIu64
 #define PRI_lx64 "l" PRIx64
 
-struct fork_event {
-	struct perf_event_header header;
-	u32 pid, ppid;
-	u32 tid, ptid;
-	u64 time;
-};
-
 struct lost_event {
 	struct perf_event_header header;
 	u64 id;
tools/perf/util/python.c +1 -1
index 55ff0c3182d6..8bdadb24f6ba 100644
@@ -170,7 +170,7 @@ static PyMemberDef pyrf_task_event__members[] = {
 static PyObject *pyrf_task_event__repr(struct pyrf_event *pevent)
 {
 	return _PyUnicode_FromFormat("{ type: %s, pid: %u, ppid: %u, tid: %u, "
-				   "ptid: %u, time: %" PRIu64 "}",
+				   "ptid: %u, time: %" PRI_lu64 "}",
 				   pevent->event.header.type == PERF_RECORD_FORK ? "fork" : "exit",
 				   pevent->event.fork.pid,
 				   pevent->event.fork.ppid,
Signature · 1 line
2.21.0
merge bot Applied
#9 lore raw

[tip: perf/core] libperf: Add PERF_RECORD_FORK 'struct fork_event' to perf/event.h

The following commit has been merged into the perf/core branch of tip:

The following commit has been merged into the perf/core branch of tip:

Commit-ID:     bceb59b1f28d1ca812d816fd8f33c72b1a8378fb
Gitweb:        https://git.kernel.org/tip/bceb59b1f28d1ca812d816fd8f33c72b1a8378fb
Author:        Jiri Olsa <[email protected]>
AuthorDate:    Sun, 25 Aug 2019 20:17:45 +02:00
Committer:     Arnaldo Carvalho de Melo <[email protected]>
CommitterDate: Mon, 26 Aug 2019 19:39:09 -03:00

libperf: Add PERF_RECORD_FORK 'struct fork_event' to perf/event.h

Move the fork_event event definition into libperf's event.h header include.

In order to keep libperf simple, we switch 'u64/u32/u16/u8' types used events to their generic '__u*' versions.

Perf added 'u*' types mainly to ease up printing __u64 values as stated in the linux/types.h comment:

/*
* We define u64 as uint64_t for every architecture
* so that we can print it with "%"PRIx64 without getting warnings. *
* typedef __u64 u64;
* typedef __s64 s64;
*/

Add and use new PRI_lu64 and PRI_lx64 macros for that. Using extra '_' to ease up the reading and differentiate them from standard PRI*64 macros.

---
 tools/perf/lib/include/perf/event.h | 7 +++++++
 tools/perf/util/event.h             | 7 -------
 tools/perf/util/python.c            | 2 +-
 3 files changed, 8 insertions(+), 8 deletions(-)
tools/perf/lib/include/perf/event.h +7 -0
index b90a8a2..c7cae58 100644
@@ -43,4 +43,11 @@ struct namespaces_event {
 	struct perf_ns_link_info link_info[];
 };
 
+struct fork_event {
+	struct perf_event_header header;
+	__u32			 pid, ppid;
+	__u32			 tid, ptid;
+	__u64			 time;
+};
+
 #endif /* __LIBPERF_EVENT_H */
tools/perf/util/event.h +0 -7
index 0d3ac4f..38b258c 100644
@@ -27,13 +27,6 @@
 #define PRI_lx64 PRIx64
 #endif
 
-struct fork_event {
-	struct perf_event_header header;
-	u32 pid, ppid;
-	u32 tid, ptid;
-	u64 time;
-};
-
 struct lost_event {
 	struct perf_event_header header;
 	u64 id;
tools/perf/util/python.c +1 -1
index 55ff0c3..8bdadb2 100644
@@ -170,7 +170,7 @@ static PyMemberDef pyrf_task_event__members[] = {
 static PyObject *pyrf_task_event__repr(struct pyrf_event *pevent)
 {
 	return _PyUnicode_FromFormat("{ type: %s, pid: %u, ppid: %u, tid: %u, "
-				   "ptid: %u, time: %" PRIu64 "}",
+				   "ptid: %u, time: %" PRI_lu64 "}",
 				   pevent->event.header.type == PERF_RECORD_FORK ? "fork" : "exit",
 				   pevent->event.fork.pid,
 				   pevent->event.fork.ppid,
author reviewer
#10 lore raw

PATCH 6/12 libperf: Add lost_event to perf/event.h

In reply to Jiri Olsa

Moving lost_event event definition into libperf's event.h header include.

In order to keep libperf simple, we switch 'u64/u32/u16/u8' types used events to their generic '__u*' versions.

Perf added 'u*' types mainly to ease up printing __u64 values as stated in the linux/types.h comment:

/*
* We define u64 as uint64_t for every architecture
* so that we can print it with "%"PRIx64 without getting warnings. *
* typedef __u64 u64;
* typedef __s64 s64;
*/

Adding and using new PRI_lu64 and PRI_lx64 macros to be used for that. Using extra '_' to ease up the reading and differentiate them from standard PRI*64 macros.

---
 tools/perf/builtin-sched.c          | 2 +-
 tools/perf/lib/include/perf/event.h | 6 ++++++
 tools/perf/util/event.c             | 2 +-
 tools/perf/util/event.h             | 6 ------
 tools/perf/util/machine.c           | 2 +-
 tools/perf/util/python.c            | 4 ++--
 6 files changed, 11 insertions(+), 11 deletions(-)
tools/perf/builtin-sched.c +1 -1
index 0d6b4c3b1a51..025151dcb651 100644
@@ -2643,7 +2643,7 @@ static int process_lost(struct perf_tool *tool __maybe_unused,
 
 	timestamp__scnprintf_usec(sample->time, tstr, sizeof(tstr));
 	printf("%15s ", tstr);
-	printf("lost %" PRIu64 " events on cpu %d\n", event->lost.lost, sample->cpu);
+	printf("lost %" PRI_lu64 " events on cpu %d\n", event->lost.lost, sample->cpu);
 
 	return 0;
 }
tools/perf/lib/include/perf/event.h +6 -0
index 7a34221abbba..6c25d0adbe6d 100644
@@ -50,4 +50,10 @@ struct fork_event {
 	__u64 time;
 };
 
+struct lost_event {
+	struct perf_event_header header;
+	__u64 id;
+	__u64 lost;
+};
+
 #endif /* __LIBPERF_EVENT_H */
tools/perf/util/event.c +1 -1
index 0954f980574f..3bd9fc2a3ae8 100644
@@ -1480,7 +1480,7 @@ size_t perf_event__fprintf_switch(union perf_event *event, FILE *fp)
 
 static size_t perf_event__fprintf_lost(union perf_event *event, FILE *fp)
 {
-	return fprintf(fp, " lost %" PRIu64 "\n", event->lost.lost);
+	return fprintf(fp, " lost %" PRI_lu64 "\n", event->lost.lost);
 }
 
 size_t perf_event__fprintf_ksymbol(union perf_event *event, FILE *fp)
tools/perf/util/event.h +0 -6
index 817f12643f11..cc9cc38f0bf7 100644
@@ -16,12 +16,6 @@
 #define PRI_lu64 "l" PRIu64
 #define PRI_lx64 "l" PRIx64
 
-struct lost_event {
-	struct perf_event_header header;
-	u64 id;
-	u64 lost;
-};
-
 struct lost_samples_event {
 	struct perf_event_header header;
 	u64 lost;
tools/perf/util/machine.c +1 -1
index 5734460fc89e..f61349d1d0d4 100644
@@ -642,7 +642,7 @@ int machine__process_namespaces_event(struct machine *machine __maybe_unused,
 int machine__process_lost_event(struct machine *machine __maybe_unused,
 				union perf_event *event, struct perf_sample *sample __maybe_unused)
 {
-	dump_printf(": id:%" PRIu64 ": lost:%" PRIu64 "\n",
+	dump_printf(": id:%" PRI_lu64 ": lost:%" PRI_lu64 "\n",
 		    event->lost.id, event->lost.lost);
 	return 0;
 }
tools/perf/util/python.c +2 -2
index 8bdadb24f6ba..5be85f50cd1c 100644
@@ -263,8 +263,8 @@ static PyObject *pyrf_lost_event__repr(struct pyrf_event *pevent)
 	PyObject *ret;
 	char *s;
 
-	if (asprintf(&s, "{ type: lost, id: %#" PRIx64 ", "
-			 "lost: %#" PRIx64 " }",
+	if (asprintf(&s, "{ type: lost, id: %#" PRI_lx64 ", "
+			 "lost: %#" PRI_lx64 " }",
 		     pevent->event.lost.id, pevent->event.lost.lost) < 0) {
 		ret = PyErr_NoMemory();
 	} else {
Signature · 1 line
2.21.0
merge bot Applied
#11 lore raw

[tip: perf/core] libperf: Add PERF_RECORD_LOST 'struct lost_event' to perf/event.h

The following commit has been merged into the perf/core branch of tip:

The following commit has been merged into the perf/core branch of tip:

Commit-ID:     5290ed6955ebc481d5cd62f7175e8514931058bc
Gitweb:        https://git.kernel.org/tip/5290ed6955ebc481d5cd62f7175e8514931058bc
Author:        Jiri Olsa <[email protected]>
AuthorDate:    Sun, 25 Aug 2019 20:17:46 +02:00
Committer:     Arnaldo Carvalho de Melo <[email protected]>
CommitterDate: Mon, 26 Aug 2019 19:39:09 -03:00

libperf: Add PERF_RECORD_LOST 'struct lost_event' to perf/event.h

Move the lost_event event definition to libperf's event.h header include.

In order to keep libperf simple, we switch 'u64/u32/u16/u8' types used events to their generic '__u*' versions.

Perf added 'u*' types mainly to ease up printing __u64 values as stated in the linux/types.h comment:

/*
* We define u64 as uint64_t for every architecture
* so that we can print it with "%"PRIx64 without getting warnings. *
* typedef __u64 u64;
* typedef __s64 s64;
*/

Add and use new PRI_lu64 and PRI_lx64 macros for that. Use extra '_' to ease up the reading and differentiate them from standard PRI*64 macros.

---
 tools/perf/builtin-sched.c          | 2 +-
 tools/perf/lib/include/perf/event.h | 6 ++++++
 tools/perf/util/event.c             | 2 +-
 tools/perf/util/event.h             | 6 ------
 tools/perf/util/machine.c           | 2 +-
 tools/perf/util/python.c            | 4 ++--
 6 files changed, 11 insertions(+), 11 deletions(-)
tools/perf/builtin-sched.c +1 -1
index 0d6b4c3..025151d 100644
@@ -2643,7 +2643,7 @@ static int process_lost(struct perf_tool *tool __maybe_unused,
 
 	timestamp__scnprintf_usec(sample->time, tstr, sizeof(tstr));
 	printf("%15s ", tstr);
-	printf("lost %" PRIu64 " events on cpu %d\n", event->lost.lost, sample->cpu);
+	printf("lost %" PRI_lu64 " events on cpu %d\n", event->lost.lost, sample->cpu);
 
 	return 0;
 }
tools/perf/lib/include/perf/event.h +6 -0
index c7cae58..71045ea 100644
@@ -50,4 +50,10 @@ struct fork_event {
 	__u64			 time;
 };
 
+struct lost_event {
+	struct perf_event_header header;
+	__u64			 id;
+	__u64			 lost;
+};
+
 #endif /* __LIBPERF_EVENT_H */
tools/perf/util/event.c +1 -1
index 0954f98..3bd9fc2 100644
@@ -1480,7 +1480,7 @@ size_t perf_event__fprintf_switch(union perf_event *event, FILE *fp)
 
 static size_t perf_event__fprintf_lost(union perf_event *event, FILE *fp)
 {
-	return fprintf(fp, " lost %" PRIu64 "\n", event->lost.lost);
+	return fprintf(fp, " lost %" PRI_lu64 "\n", event->lost.lost);
 }
 
 size_t perf_event__fprintf_ksymbol(union perf_event *event, FILE *fp)
tools/perf/util/event.h +0 -6
index 38b258c..4a3f502 100644
@@ -27,12 +27,6 @@
 #define PRI_lx64 PRIx64
 #endif
 
-struct lost_event {
-	struct perf_event_header header;
-	u64 id;
-	u64 lost;
-};
-
 struct lost_samples_event {
 	struct perf_event_header header;
 	u64 lost;
tools/perf/util/machine.c +1 -1
index 47430af..1ad6e98 100644
@@ -645,7 +645,7 @@ int machine__process_namespaces_event(struct machine *machine __maybe_unused,
 int machine__process_lost_event(struct machine *machine __maybe_unused,
 				union perf_event *event, struct perf_sample *sample __maybe_unused)
 {
-	dump_printf(": id:%" PRIu64 ": lost:%" PRIu64 "\n",
+	dump_printf(": id:%" PRI_lu64 ": lost:%" PRI_lu64 "\n",
 		    event->lost.id, event->lost.lost);
 	return 0;
 }
tools/perf/util/python.c +2 -2
index 8bdadb2..5be85f5 100644
@@ -263,8 +263,8 @@ static PyObject *pyrf_lost_event__repr(struct pyrf_event *pevent)
 	PyObject *ret;
 	char *s;
 
-	if (asprintf(&s, "{ type: lost, id: %#" PRIx64 ", "
-			 "lost: %#" PRIx64 " }",
+	if (asprintf(&s, "{ type: lost, id: %#" PRI_lx64 ", "
+			 "lost: %#" PRI_lx64 " }",
 		     pevent->event.lost.id, pevent->event.lost.lost) < 0) {
 		ret = PyErr_NoMemory();
 	} else {
author reviewer
#12 lore raw

PATCH 7/12 libperf: Add lost_samples_event to perf/event.h

In reply to Jiri Olsa

Moving lost_samples_event event definition into libperf's event.h header include.

In order to keep libperf simple, we switch 'u64/u32/u16/u8' types used events to their generic '__u*' versions.

Perf added 'u*' types mainly to ease up printing __u64 values as stated in the linux/types.h comment:

/*
* We define u64 as uint64_t for every architecture
* so that we can print it with "%"PRIx64 without getting warnings. *
* typedef __u64 u64;
* typedef __s64 s64;
*/

Adding and using new PRI_lu64 and PRI_lx64 macros to be used for that. Using extra '_' to ease up the reading and differentiate them from standard PRI*64 macros.

---
 tools/perf/lib/include/perf/event.h | 5 +++++
 tools/perf/util/event.h             | 5 -----
 tools/perf/util/machine.c           | 2 +-
 3 files changed, 6 insertions(+), 6 deletions(-)
tools/perf/lib/include/perf/event.h +5 -0
index 6c25d0adbe6d..3bd2727ac5f9 100644
@@ -56,4 +56,9 @@ struct lost_event {
 	__u64 lost;
 };
 
+struct lost_samples_event {
+	struct perf_event_header header;
+	__u64 lost;
+};
+
 #endif /* __LIBPERF_EVENT_H */
tools/perf/util/event.h +0 -5
index cc9cc38f0bf7..1bf44b73fbbf 100644
@@ -16,11 +16,6 @@
 #define PRI_lu64 "l" PRIu64
 #define PRI_lx64 "l" PRIx64
 
-struct lost_samples_event {
-	struct perf_event_header header;
-	u64 lost;
-};
-
 /*
  * PERF_FORMAT_ENABLED | PERF_FORMAT_RUNNING | PERF_FORMAT_ID
  */
tools/perf/util/machine.c +1 -1
index f61349d1d0d4..9094e86e4db8 100644
@@ -650,7 +650,7 @@ int machine__process_lost_event(struct machine *machine __maybe_unused,
 int machine__process_lost_samples_event(struct machine *machine __maybe_unused,
 					union perf_event *event, struct perf_sample *sample)
 {
-	dump_printf(": id:%" PRIu64 ": lost samples :%" PRIu64 "\n",
+	dump_printf(": id:%" PRIu64 ": lost samples :%" PRI_lu64 "\n",
 		    sample->id, event->lost_samples.lost);
 	return 0;
 }
Signature · 1 line
2.21.0
merge bot Applied
#13 lore raw

[tip: perf/core] libperf: Add PERF_RECORD_LOST_SAMPLES 'struct lost_samples_event' to perf/event.h

The following commit has been merged into the perf/core branch of tip:

The following commit has been merged into the perf/core branch of tip:

Commit-ID:     a2e254d84172f7eb638261a83024d849f78c89e9
Gitweb:        https://git.kernel.org/tip/a2e254d84172f7eb638261a83024d849f78c89e9
Author:        Jiri Olsa <[email protected]>
AuthorDate:    Sun, 25 Aug 2019 20:17:47 +02:00
Committer:     Arnaldo Carvalho de Melo <[email protected]>
CommitterDate: Mon, 26 Aug 2019 19:39:10 -03:00

libperf: Add PERF_RECORD_LOST_SAMPLES 'struct lost_samples_event' to perf/event.h

Move the PERF_RECORD_LOST_SAMPLES event definition into libperf's event.h header include.

In order to keep libperf simple, we switch 'u64/u32/u16/u8' types used events to their generic '__u*' versions.

Perf added 'u*' types mainly to ease up printing __u64 values as stated in the linux/types.h comment:

/*
* We define u64 as uint64_t for every architecture
* so that we can print it with "%"PRIx64 without getting warnings. *
* typedef __u64 u64;
* typedef __s64 s64;
*/

Add and use new PRI_lu64 and PRI_lx64 macros for that. Use extra '_' to ease up the reading and differentiate them from standard PRI*64 macros.

---
 tools/perf/lib/include/perf/event.h | 5 +++++
 tools/perf/util/event.h             | 5 -----
 tools/perf/util/machine.c           | 2 +-
 3 files changed, 6 insertions(+), 6 deletions(-)
tools/perf/lib/include/perf/event.h +5 -0
index 71045ea..86a7795 100644
@@ -56,4 +56,9 @@ struct lost_event {
 	__u64			 lost;
 };
 
+struct lost_samples_event {
+	struct perf_event_header header;
+	__u64			 lost;
+};
+
 #endif /* __LIBPERF_EVENT_H */
tools/perf/util/event.h +0 -5
index 4a3f502..976a8f0 100644
@@ -27,11 +27,6 @@
 #define PRI_lx64 PRIx64
 #endif
 
-struct lost_samples_event {
-	struct perf_event_header header;
-	u64 lost;
-};
-
 /*
  * PERF_FORMAT_ENABLED | PERF_FORMAT_RUNNING | PERF_FORMAT_ID
  */
tools/perf/util/machine.c +1 -1
index 1ad6e98..823aaf7 100644
@@ -653,7 +653,7 @@ int machine__process_lost_event(struct machine *machine __maybe_unused,
 int machine__process_lost_samples_event(struct machine *machine __maybe_unused,
 					union perf_event *event, struct perf_sample *sample)
 {
-	dump_printf(": id:%" PRIu64 ": lost samples :%" PRIu64 "\n",
+	dump_printf(": id:%" PRIu64 ": lost samples :%" PRI_lu64 "\n",
 		    sample->id, event->lost_samples.lost);
 	return 0;
 }
author reviewer
#14 lore raw

PATCH 8/12 libperf: Add read_event to perf/event.h

In reply to Jiri Olsa

Moving read_event event definition into libperf's event.h header include.

In order to keep libperf simple, we switch 'u64/u32/u16/u8' types used events to their generic '__u*' versions.

Perf added 'u*' types mainly to ease up printing __u64 values as stated in the linux/types.h comment:

/*
* We define u64 as uint64_t for every architecture
* so that we can print it with "%"PRIx64 without getting warnings. *
* typedef __u64 u64;
* typedef __s64 s64;
*/

Adding and using new PRI_lu64 and PRI_lx64 macros to be used for that. Using extra '_' to ease up the reading and differentiate them from standard PRI*64 macros.

---
 tools/perf/lib/include/perf/event.h | 12 ++++++++++++
 tools/perf/util/event.h             | 12 ------------
 tools/perf/util/session.c           |  8 ++++----
 3 files changed, 16 insertions(+), 16 deletions(-)
tools/perf/lib/include/perf/event.h +12 -0
index 3bd2727ac5f9..a03dbe776a44 100644
@@ -61,4 +61,16 @@ struct lost_samples_event {
 	__u64 lost;
 };
 
+/*
+ * PERF_FORMAT_ENABLED | PERF_FORMAT_RUNNING | PERF_FORMAT_ID
+ */
+struct read_event {
+	struct perf_event_header header;
+	__u32 pid, tid;
+	__u64 value;
+	__u64 time_enabled;
+	__u64 time_running;
+	__u64 id;
+};
+
 #endif /* __LIBPERF_EVENT_H */
tools/perf/util/event.h +0 -12
index 1bf44b73fbbf..89872279a5af 100644
@@ -16,18 +16,6 @@
 #define PRI_lu64 "l" PRIu64
 #define PRI_lx64 "l" PRIx64
 
-/*
- * PERF_FORMAT_ENABLED | PERF_FORMAT_RUNNING | PERF_FORMAT_ID
- */
-struct read_event {
-	struct perf_event_header header;
-	u32 pid, tid;
-	u64 value;
-	u64 time_enabled;
-	u64 time_running;
-	u64 id;
-};
-
 struct throttle_event {
 	struct perf_event_header header;
 	u64 time;
tools/perf/util/session.c +4 -4
index 82e0438a9160..cb1d8dcd0c19 100644
@@ -1260,7 +1260,7 @@ static void dump_read(struct evsel *evsel, union perf_event *event)
 	if (!dump_trace)
 		return;
 
-	printf(": %d %d %s %" PRIu64 "\n", event->read.pid, event->read.tid,
+	printf(": %d %d %s %" PRI_lu64 "\n", event->read.pid, event->read.tid,
 	       perf_evsel__name(evsel),
 	       event->read.value);
 
@@ -1270,13 +1270,13 @@ static void dump_read(struct evsel *evsel, union perf_event *event)
 	read_format = evsel->core.attr.read_format;
 
 	if (read_format & PERF_FORMAT_TOTAL_TIME_ENABLED)
-		printf("... time enabled : %" PRIu64 "\n", read_event->time_enabled);
+		printf("... time enabled : %" PRI_lu64 "\n", read_event->time_enabled);
 
 	if (read_format & PERF_FORMAT_TOTAL_TIME_RUNNING)
-		printf("... time running : %" PRIu64 "\n", read_event->time_running);
+		printf("... time running : %" PRI_lu64 "\n", read_event->time_running);
 
 	if (read_format & PERF_FORMAT_ID)
-		printf("... id           : %" PRIu64 "\n", read_event->id);
+		printf("... id           : %" PRI_lu64 "\n", read_event->id);
 }
 
 static struct machine *machines__find_for_cpumode(struct machines *machines,
Signature · 1 line
2.21.0
merge bot Applied
#15 lore raw

[tip: perf/core] libperf: Add PERF_RECORD_READ 'struct read_event' to perf/event.h

The following commit has been merged into the perf/core branch of tip:

The following commit has been merged into the perf/core branch of tip:

Commit-ID:     213a6c1d20687d44acaa1cb4f77ce5bae4f1dd8f
Gitweb:        https://git.kernel.org/tip/213a6c1d20687d44acaa1cb4f77ce5bae4f1dd8f
Author:        Jiri Olsa <[email protected]>
AuthorDate:    Sun, 25 Aug 2019 20:17:48 +02:00
Committer:     Arnaldo Carvalho de Melo <[email protected]>
CommitterDate: Mon, 26 Aug 2019 19:39:10 -03:00

libperf: Add PERF_RECORD_READ 'struct read_event' to perf/event.h

Move the PERF_RECORD_READ event definition to libperf's event.h header include.

In order to keep libperf simple, we switch 'u64/u32/u16/u8' types used events to their generic '__u*' versions.

Perf added 'u*' types mainly to ease up printing __u64 values as stated in the linux/types.h comment:

/*
* We define u64 as uint64_t for every architecture
* so that we can print it with "%"PRIx64 without getting warnings. *
* typedef __u64 u64;
* typedef __s64 s64;
*/

Add and use new PRI_lu64 and PRI_lx64 macros for that. Use extra '_' to ease up the reading and differentiate them from standard PRI*64 macros.

---
 tools/perf/lib/include/perf/event.h | 12 ++++++++++++
 tools/perf/util/event.h             | 12 ------------
 tools/perf/util/session.c           |  8 ++++----
 3 files changed, 16 insertions(+), 16 deletions(-)
tools/perf/lib/include/perf/event.h +12 -0
index 86a7795..f183070 100644
@@ -61,4 +61,16 @@ struct lost_samples_event {
 	__u64			 lost;
 };
 
+/*
+ * PERF_FORMAT_ENABLED | PERF_FORMAT_RUNNING | PERF_FORMAT_ID
+ */
+struct read_event {
+	struct perf_event_header header;
+	__u32			 pid, tid;
+	__u64			 value;
+	__u64			 time_enabled;
+	__u64			 time_running;
+	__u64			 id;
+};
+
 #endif /* __LIBPERF_EVENT_H */
tools/perf/util/event.h +0 -12
index 976a8f0..008a283 100644
@@ -27,18 +27,6 @@
 #define PRI_lx64 PRIx64
 #endif
 
-/*
- * PERF_FORMAT_ENABLED | PERF_FORMAT_RUNNING | PERF_FORMAT_ID
- */
-struct read_event {
-	struct perf_event_header header;
-	u32 pid, tid;
-	u64 value;
-	u64 time_enabled;
-	u64 time_running;
-	u64 id;
-};
-
 struct throttle_event {
 	struct perf_event_header header;
 	u64 time;
tools/perf/util/session.c +4 -4
index 82e0438..cb1d8dc 100644
@@ -1260,7 +1260,7 @@ static void dump_read(struct evsel *evsel, union perf_event *event)
 	if (!dump_trace)
 		return;
 
-	printf(": %d %d %s %" PRIu64 "\n", event->read.pid, event->read.tid,
+	printf(": %d %d %s %" PRI_lu64 "\n", event->read.pid, event->read.tid,
 	       perf_evsel__name(evsel),
 	       event->read.value);
 
@@ -1270,13 +1270,13 @@ static void dump_read(struct evsel *evsel, union perf_event *event)
 	read_format = evsel->core.attr.read_format;
 
 	if (read_format & PERF_FORMAT_TOTAL_TIME_ENABLED)
-		printf("... time enabled : %" PRIu64 "\n", read_event->time_enabled);
+		printf("... time enabled : %" PRI_lu64 "\n", read_event->time_enabled);
 
 	if (read_format & PERF_FORMAT_TOTAL_TIME_RUNNING)
-		printf("... time running : %" PRIu64 "\n", read_event->time_running);
+		printf("... time running : %" PRI_lu64 "\n", read_event->time_running);
 
 	if (read_format & PERF_FORMAT_ID)
-		printf("... id           : %" PRIu64 "\n", read_event->id);
+		printf("... id           : %" PRI_lu64 "\n", read_event->id);
 }
 
 static struct machine *machines__find_for_cpumode(struct machines *machines,
author reviewer
#16 lore raw

PATCH 9/12 libperf: Add throttle_event to perf/event.h

In reply to Jiri Olsa

Moving throttle_event event definition into libperf's event.h header include.

In order to keep libperf simple, we switch 'u64/u32/u16/u8' types used events to their generic '__u*' versions.

Perf added 'u*' types mainly to ease up printing __u64 values as stated in the linux/types.h comment:

/*
* We define u64 as uint64_t for every architecture
* so that we can print it with "%"PRIx64 without getting warnings. *
* typedef __u64 u64;
* typedef __s64 s64;
*/

Adding and using new PRI_lu64 and PRI_lx64 macros to be used for that. Using extra '_' to ease up the reading and differentiate them from standard PRI*64 macros.

---
 tools/perf/lib/include/perf/event.h | 7 +++++++
 tools/perf/util/event.h             | 7 -------
 tools/perf/util/python.c            | 4 ++--
 3 files changed, 9 insertions(+), 9 deletions(-)
tools/perf/lib/include/perf/event.h +7 -0
index a03dbe776a44..80d0bc48baff 100644
@@ -73,4 +73,11 @@ struct read_event {
 	__u64 id;
 };
 
+struct throttle_event {
+	struct perf_event_header header;
+	__u64 time;
+	__u64 id;
+	__u64 stream_id;
+};
+
 #endif /* __LIBPERF_EVENT_H */
tools/perf/util/event.h +0 -7
index 89872279a5af..d03a70e2e6a4 100644
@@ -16,13 +16,6 @@
 #define PRI_lu64 "l" PRIu64
 #define PRI_lx64 "l" PRIx64
 
-struct throttle_event {
-	struct perf_event_header header;
-	u64 time;
-	u64 id;
-	u64 stream_id;
-};
-
 #ifndef KSYM_NAME_LEN
 #define KSYM_NAME_LEN 256
 #endif
tools/perf/util/python.c +2 -2
index 5be85f50cd1c..d21e270c7823 100644
@@ -233,8 +233,8 @@ static PyObject *pyrf_throttle_event__repr(struct pyrf_event *pevent)
 {
 	struct throttle_event *te = (struct throttle_event *)(&pevent->event.header + 1);
 
-	return _PyUnicode_FromFormat("{ type: %sthrottle, time: %" PRIu64 ", id: %" PRIu64
-				   ", stream_id: %" PRIu64 " }",
+	return _PyUnicode_FromFormat("{ type: %sthrottle, time: %" PRI_lu64 ", id: %" PRI_lu64
+				   ", stream_id: %" PRI_lu64 " }",
 				   pevent->event.header.type == PERF_RECORD_THROTTLE ? "" : "un",
 				   te->time, te->id, te->stream_id);
 }
Signature · 1 line
2.21.0
merge bot Applied
#17 lore raw

[tip: perf/core] libperf: Add PERF_RECORD_THROTTLE 'struct throttle_event' to perf/event.h

The following commit has been merged into the perf/core branch of tip:

The following commit has been merged into the perf/core branch of tip:

Commit-ID:     003c66fec28fea52825b60cad98af8cf11074d76
Gitweb:        https://git.kernel.org/tip/003c66fec28fea52825b60cad98af8cf11074d76
Author:        Jiri Olsa <[email protected]>
AuthorDate:    Sun, 25 Aug 2019 20:17:49 +02:00
Committer:     Arnaldo Carvalho de Melo <[email protected]>
CommitterDate: Mon, 26 Aug 2019 19:39:10 -03:00

libperf: Add PERF_RECORD_THROTTLE 'struct throttle_event' to perf/event.h

Move the PERF_RECORD_THROTTLE event definition into libperf's event.h header include.

In order to keep libperf simple, we switch 'u64/u32/u16/u8' types used events to their generic '__u*' versions.

Perf added 'u*' types mainly to ease up printing __u64 values as stated in the linux/types.h comment:

/*
* We define u64 as uint64_t for every architecture
* so that we can print it with "%"PRIx64 without getting warnings. *
* typedef __u64 u64;
* typedef __s64 s64;
*/

Add and use new PRI_lu64 and PRI_lx64 macros for that. Use extra '_' to ease up the reading and differentiate them from standard PRI*64 macros.

---
 tools/perf/lib/include/perf/event.h | 7 +++++++
 tools/perf/util/event.h             | 7 -------
 tools/perf/util/python.c            | 4 ++--
 3 files changed, 9 insertions(+), 9 deletions(-)
tools/perf/lib/include/perf/event.h +7 -0
index f183070..ef5ec66 100644
@@ -73,4 +73,11 @@ struct read_event {
 	__u64			 id;
 };
 
+struct throttle_event {
+	struct perf_event_header header;
+	__u64			 time;
+	__u64			 id;
+	__u64			 stream_id;
+};
+
 #endif /* __LIBPERF_EVENT_H */
tools/perf/util/event.h +0 -7
index 008a283..40020f5 100644
@@ -27,13 +27,6 @@
 #define PRI_lx64 PRIx64
 #endif
 
-struct throttle_event {
-	struct perf_event_header header;
-	u64 time;
-	u64 id;
-	u64 stream_id;
-};
-
 #ifndef KSYM_NAME_LEN
 #define KSYM_NAME_LEN 256
 #endif
tools/perf/util/python.c +2 -2
index 5be85f5..d21e270 100644
@@ -233,8 +233,8 @@ static PyObject *pyrf_throttle_event__repr(struct pyrf_event *pevent)
 {
 	struct throttle_event *te = (struct throttle_event *)(&pevent->event.header + 1);
 
-	return _PyUnicode_FromFormat("{ type: %sthrottle, time: %" PRIu64 ", id: %" PRIu64
-				   ", stream_id: %" PRIu64 " }",
+	return _PyUnicode_FromFormat("{ type: %sthrottle, time: %" PRI_lu64 ", id: %" PRI_lu64
+				   ", stream_id: %" PRI_lu64 " }",
 				   pevent->event.header.type == PERF_RECORD_THROTTLE ? "" : "un",
 				   te->time, te->id, te->stream_id);
 }
author reviewer
#18 lore raw

PATCH 10/12 libperf: Add ksymbol_event to perf/event.h

In reply to Jiri Olsa

Moving ksymbol_event event definition into libperf's event.h header include.

In order to keep libperf simple, we switch 'u64/u32/u16/u8' types used events to their generic '__u*' versions.

Perf added 'u*' types mainly to ease up printing __u64 values as stated in the linux/types.h comment:

/*
* We define u64 as uint64_t for every architecture
* so that we can print it with "%"PRIx64 without getting warnings. *
* typedef __u64 u64;
* typedef __s64 s64;
*/

Adding and using new PRI_lu64 and PRI_lx64 macros to be used for that. Using extra '_' to ease up the reading and differentiate them from standard PRI*64 macros.

---
 tools/perf/lib/include/perf/event.h | 13 +++++++++++++
 tools/perf/util/event.c             |  2 +-
 tools/perf/util/event.h             | 13 -------------
 3 files changed, 14 insertions(+), 14 deletions(-)
tools/perf/lib/include/perf/event.h +13 -0
index 80d0bc48baff..ecab3246913d 100644
@@ -80,4 +80,17 @@ struct throttle_event {
 	__u64 stream_id;
 };
 
+#ifndef KSYM_NAME_LEN
+#define KSYM_NAME_LEN 256
+#endif
+
+struct ksymbol_event {
+	struct perf_event_header header;
+	__u64 addr;
+	__u32 len;
+	__u16 ksym_type;
+	__u16 flags;
+	char name[KSYM_NAME_LEN];
+};
+
 #endif /* __LIBPERF_EVENT_H */
tools/perf/util/event.c +1 -1
index 3bd9fc2a3ae8..4447cd25e3f2 100644
@@ -1485,7 +1485,7 @@ static size_t perf_event__fprintf_lost(union perf_event *event, FILE *fp)
 
 size_t perf_event__fprintf_ksymbol(union perf_event *event, FILE *fp)
 {
-	return fprintf(fp, " addr %" PRIx64 " len %u type %u flags 0x%x name %s\n",
+	return fprintf(fp, " addr %" PRI_lx64 " len %u type %u flags 0x%x name %s\n",
 		       event->ksymbol_event.addr, event->ksymbol_event.len,
 		       event->ksymbol_event.ksym_type,
 		       event->ksymbol_event.flags, event->ksymbol_event.name);
tools/perf/util/event.h +0 -13
index d03a70e2e6a4..dbabde3a43a6 100644
@@ -16,19 +16,6 @@
 #define PRI_lu64 "l" PRIu64
 #define PRI_lx64 "l" PRIx64
 
-#ifndef KSYM_NAME_LEN
-#define KSYM_NAME_LEN 256
-#endif
-
-struct ksymbol_event {
-	struct perf_event_header header;
-	u64 addr;
-	u32 len;
-	u16 ksym_type;
-	u16 flags;
-	char name[KSYM_NAME_LEN];
-};
-
 struct bpf_event {
 	struct perf_event_header header;
 	u16 type;
Signature · 1 line
2.21.0
merge bot Applied
#19 lore raw

[tip: perf/core] libperf: Add PERF_RECORD_KSYMBOL 'struct ksymbol_event' to perf/event.h

The following commit has been merged into the perf/core branch of tip:

The following commit has been merged into the perf/core branch of tip:

Commit-ID:     f15e3c25a1b40794a0ef2647360afe873fe34f54
Gitweb:        https://git.kernel.org/tip/f15e3c25a1b40794a0ef2647360afe873fe34f54
Author:        Jiri Olsa <[email protected]>
AuthorDate:    Sun, 25 Aug 2019 20:17:50 +02:00
Committer:     Arnaldo Carvalho de Melo <[email protected]>
CommitterDate: Mon, 26 Aug 2019 19:39:10 -03:00

libperf: Add PERF_RECORD_KSYMBOL 'struct ksymbol_event' to perf/event.h

Move the PERF_RECORD_KSYMBOL event definition into libperf's event.h header include.

In order to keep libperf simple, we switch 'u64/u32/u16/u8' types used events to their generic '__u*' versions.

Perf added 'u*' types mainly to ease up printing __u64 values as stated in the linux/types.h comment:

/*
* We define u64 as uint64_t for every architecture
* so that we can print it with "%"PRIx64 without getting warnings. *
* typedef __u64 u64;
* typedef __s64 s64;
*/

Add and use new PRI_lu64 and PRI_lx64 macros for that. Use extra '_' to ease up the reading and differentiate them from standard PRI*64 macros.

---
 tools/perf/lib/include/perf/event.h | 13 +++++++++++++
 tools/perf/util/event.c             |  2 +-
 tools/perf/util/event.h             | 13 -------------
 3 files changed, 14 insertions(+), 14 deletions(-)
tools/perf/lib/include/perf/event.h +13 -0
index ef5ec66..8c36793 100644
@@ -80,4 +80,17 @@ struct throttle_event {
 	__u64			 stream_id;
 };
 
+#ifndef KSYM_NAME_LEN
+#define KSYM_NAME_LEN 256
+#endif
+
+struct ksymbol_event {
+	struct perf_event_header header;
+	__u64			 addr;
+	__u32			 len;
+	__u16			 ksym_type;
+	__u16			 flags;
+	char			 name[KSYM_NAME_LEN];
+};
+
 #endif /* __LIBPERF_EVENT_H */
tools/perf/util/event.c +1 -1
index 3bd9fc2..4447cd2 100644
@@ -1485,7 +1485,7 @@ static size_t perf_event__fprintf_lost(union perf_event *event, FILE *fp)
 
 size_t perf_event__fprintf_ksymbol(union perf_event *event, FILE *fp)
 {
-	return fprintf(fp, " addr %" PRIx64 " len %u type %u flags 0x%x name %s\n",
+	return fprintf(fp, " addr %" PRI_lx64 " len %u type %u flags 0x%x name %s\n",
 		       event->ksymbol_event.addr, event->ksymbol_event.len,
 		       event->ksymbol_event.ksym_type,
 		       event->ksymbol_event.flags, event->ksymbol_event.name);
tools/perf/util/event.h +0 -13
index 40020f5..c4eec1f 100644
@@ -27,19 +27,6 @@
 #define PRI_lx64 PRIx64
 #endif
 
-#ifndef KSYM_NAME_LEN
-#define KSYM_NAME_LEN 256
-#endif
-
-struct ksymbol_event {
-	struct perf_event_header header;
-	u64 addr;
-	u32 len;
-	u16 ksym_type;
-	u16 flags;
-	char name[KSYM_NAME_LEN];
-};
-
 struct bpf_event {
 	struct perf_event_header header;
 	u16 type;
author reviewer
#20 lore raw

PATCH 11/12 libperf: Add bpf_event to perf/event.h

In reply to Jiri Olsa

Moving lost_samples_event event definition into libperf's event.h header include.

In order to keep libperf simple, we switch 'u64/u32/u16/u8' types used events to their generic '__u*' versions.

---
 tools/perf/lib/include/perf/event.h | 11 +++++++++++
 tools/perf/util/event.h             | 10 ----------
 2 files changed, 11 insertions(+), 10 deletions(-)
tools/perf/lib/include/perf/event.h +11 -0
index ecab3246913d..74317bcf74bf 100644
@@ -5,6 +5,7 @@
 #include <linux/perf_event.h>
 #include <linux/types.h>
 #include <linux/limits.h>
+#include <linux/bpf.h>
 
 struct mmap_event {
 	struct perf_event_header header;
@@ -93,4 +94,14 @@ struct ksymbol_event {
 	char name[KSYM_NAME_LEN];
 };
 
+struct bpf_event {
+	struct perf_event_header header;
+	__u16 type;
+	__u16 flags;
+	__u32 id;
+
+	/* for bpf_prog types */
+	__u8 tag[BPF_TAG_SIZE];  // prog tag
+};
+
 #endif /* __LIBPERF_EVENT_H */
tools/perf/util/event.h +0 -10
index dbabde3a43a6..cd1d8d1007dc 100644
@@ -16,16 +16,6 @@
 #define PRI_lu64 "l" PRIu64
 #define PRI_lx64 "l" PRIx64
 
-struct bpf_event {
-	struct perf_event_header header;
-	u16 type;
-	u16 flags;
-	u32 id;
-
-	/* for bpf_prog types */
-	u8 tag[BPF_TAG_SIZE];  // prog tag
-};
-
 #define PERF_SAMPLE_MASK				\
 	(PERF_SAMPLE_IP | PERF_SAMPLE_TID |		\
 	 PERF_SAMPLE_TIME | PERF_SAMPLE_ADDR |		\
Signature · 1 line
2.21.0
merge bot Applied
#21 lore raw

[tip: perf/core] libperf: Add PERF_RECORD_BPF_EVENT 'struct bpf_event' to perf/event.h

The following commit has been merged into the perf/core branch of tip:

The following commit has been merged into the perf/core branch of tip:

Commit-ID:     b1b510142283c02991f48b27d399852364f7d89b
Gitweb:        https://git.kernel.org/tip/b1b510142283c02991f48b27d399852364f7d89b
Author:        Jiri Olsa <[email protected]>
AuthorDate:    Sun, 25 Aug 2019 20:17:51 +02:00
Committer:     Arnaldo Carvalho de Melo <[email protected]>
CommitterDate: Mon, 26 Aug 2019 19:39:10 -03:00

libperf: Add PERF_RECORD_BPF_EVENT 'struct bpf_event' to perf/event.h

Move the PERF_RECORD_BPF_EVENT event definition to libperf's event.h.

In order to keep libperf simple, we switch 'u64/u32/u16/u8' types used events to their generic '__u*' versions.

---
 tools/perf/lib/include/perf/event.h | 11 +++++++++++
 tools/perf/util/event.h             | 10 ----------
 2 files changed, 11 insertions(+), 10 deletions(-)
tools/perf/lib/include/perf/event.h +11 -0
index 8c36793..585c9d8 100644
@@ -5,6 +5,7 @@
 #include <linux/perf_event.h>
 #include <linux/types.h>
 #include <linux/limits.h>
+#include <linux/bpf.h>
 
 struct mmap_event {
 	struct perf_event_header header;
@@ -93,4 +94,14 @@ struct ksymbol_event {
 	char			 name[KSYM_NAME_LEN];
 };
 
+struct bpf_event {
+	struct perf_event_header header;
+	__u16			 type;
+	__u16			 flags;
+	__u32			 id;
+
+	/* for bpf_prog types */
+	__u8			 tag[BPF_TAG_SIZE];  // prog tag
+};
+
 #endif /* __LIBPERF_EVENT_H */
tools/perf/util/event.h +0 -10
index c4eec1f..091a069 100644
@@ -27,16 +27,6 @@
 #define PRI_lx64 PRIx64
 #endif
 
-struct bpf_event {
-	struct perf_event_header header;
-	u16 type;
-	u16 flags;
-	u32 id;
-
-	/* for bpf_prog types */
-	u8 tag[BPF_TAG_SIZE];  // prog tag
-};
-
 #define PERF_SAMPLE_MASK				\
 	(PERF_SAMPLE_IP | PERF_SAMPLE_TID |		\
 	 PERF_SAMPLE_TIME | PERF_SAMPLE_ADDR |		\
author reviewer
#22 lore raw

PATCH 12/12 libperf: Add sample_event to perf/event.h

In reply to Jiri Olsa

Moving sample_event event definition into libperf's event.h header include.

In order to keep libperf simple, we switch 'u64/u32/u16/u8' types used events to their generic '__u*' versions.

---
 tools/perf/lib/include/perf/event.h | 5 +++++
 tools/perf/util/event.h             | 5 -----
 tools/perf/util/evlist.c            | 2 +-
 tools/perf/util/evsel.c             | 8 ++++----
 4 files changed, 10 insertions(+), 10 deletions(-)
tools/perf/lib/include/perf/event.h +5 -0
index 74317bcf74bf..daa1924c1d11 100644
@@ -104,4 +104,9 @@ struct bpf_event {
 	__u8 tag[BPF_TAG_SIZE];  // prog tag
 };
 
+struct sample_event {
+	struct perf_event_header header;
+	__u64 array[];
+};
+
 #endif /* __LIBPERF_EVENT_H */
tools/perf/util/event.h +0 -5
index cd1d8d1007dc..37349ddb30f0 100644
@@ -26,11 +26,6 @@
 /* perf sample has 16 bits size limit */
 #define PERF_SAMPLE_MAX_SIZE (1 << 16)
 
-struct sample_event {
-	struct perf_event_header        header;
-	u64 array[];
-};
-
 struct regs_dump {
 	u64 abi;
 	u64 mask;
tools/perf/util/evlist.c +1 -1
index ff415680fe0a..47bc54111f57 100644
@@ -587,7 +587,7 @@ struct evsel *perf_evlist__id2evsel_strict(struct evlist *evlist,
 static int perf_evlist__event2id(struct evlist *evlist,
 				 union perf_event *event, u64 *id)
 {
-	const u64 *array = event->sample.array;
+	const __u64 *array = event->sample.array;
 	ssize_t n;
 
 	n = (event->header.size - sizeof(event->header)) >> 3;
tools/perf/util/evsel.c +4 -4
index e983e721beca..552bc4c839f4 100644
@@ -2008,7 +2008,7 @@ static int perf_evsel__parse_id_sample(const struct evsel *evsel,
 				       struct perf_sample *sample)
 {
 	u64 type = evsel->core.attr.sample_type;
-	const u64 *array = event->sample.array;
+	const __u64 *array = event->sample.array;
 	bool swapped = evsel->needs_swap;
 	union u64_swap u;
 
@@ -2098,7 +2098,7 @@ int perf_evsel__parse_sample(struct evsel *evsel, union perf_event *event,
 {
 	u64 type = evsel->core.attr.sample_type;
 	bool swapped = evsel->needs_swap;
-	const u64 *array;
+	const __u64 *array;
 	u16 max_size = event->header.size;
 	const void *endp = (void *)event + max_size;
 	u64 sz;
@@ -2377,7 +2377,7 @@ int perf_evsel__parse_sample_timestamp(struct evsel *evsel,
 				       u64 *timestamp)
 {
 	u64 type = evsel->core.attr.sample_type;
-	const u64 *array;
+	const __u64 *array;
 
 	if (!(type & PERF_SAMPLE_TIME))
 		return -1;
@@ -2528,7 +2528,7 @@ int perf_event__synthesize_sample(union perf_event *event, u64 type,
 				  u64 read_format,
 				  const struct perf_sample *sample)
 {
-	u64 *array;
+	__u64 *array;
 	size_t sz;
 	/*
 	 * used for cross-endian analysis. See git commit 65014ab3
Signature · 1 line
2.21.0
merge bot Applied
#23 lore raw

[tip: perf/core] libperf: Add PERF_RECORD_SAMPLE 'struct sample_event' to perf/event.h

The following commit has been merged into the perf/core branch of tip:

The following commit has been merged into the perf/core branch of tip:

Commit-ID:     b1fcd190bb3fc1234dca60390d171a4cc75b21b2
Gitweb:        https://git.kernel.org/tip/b1fcd190bb3fc1234dca60390d171a4cc75b21b2
Author:        Jiri Olsa <[email protected]>
AuthorDate:    Sun, 25 Aug 2019 20:17:52 +02:00
Committer:     Arnaldo Carvalho de Melo <[email protected]>
CommitterDate: Mon, 26 Aug 2019 19:39:10 -03:00

libperf: Add PERF_RECORD_SAMPLE 'struct sample_event' to perf/event.h

Move the PERF_RECORD_SAMPLE event definition to libperf's event.h header include.

In order to keep libperf simple, we switch 'u64/u32/u16/u8' types used events to their generic '__u*' versions.

---
 tools/perf/lib/include/perf/event.h | 5 +++++
 tools/perf/util/event.h             | 5 -----
 tools/perf/util/evlist.c            | 2 +-
 tools/perf/util/evsel.c             | 8 ++++----
 4 files changed, 10 insertions(+), 10 deletions(-)
tools/perf/lib/include/perf/event.h +5 -0
index 585c9d8..e768a2d 100644
@@ -104,4 +104,9 @@ struct bpf_event {
 	__u8			 tag[BPF_TAG_SIZE];  // prog tag
 };
 
+struct sample_event {
+	struct perf_event_header header;
+	__u64			 array[];
+};
+
 #endif /* __LIBPERF_EVENT_H */
tools/perf/util/event.h +0 -5
index 091a069..dee0ee5 100644
@@ -37,11 +37,6 @@
 /* perf sample has 16 bits size limit */
 #define PERF_SAMPLE_MAX_SIZE (1 << 16)
 
-struct sample_event {
-	struct perf_event_header        header;
-	u64 array[];
-};
-
 struct regs_dump {
 	u64 abi;
 	u64 mask;
tools/perf/util/evlist.c +1 -1
index ff41568..47bc541 100644
@@ -587,7 +587,7 @@ struct evsel *perf_evlist__id2evsel_strict(struct evlist *evlist,
 static int perf_evlist__event2id(struct evlist *evlist,
 				 union perf_event *event, u64 *id)
 {
-	const u64 *array = event->sample.array;
+	const __u64 *array = event->sample.array;
 	ssize_t n;
 
 	n = (event->header.size - sizeof(event->header)) >> 3;
tools/perf/util/evsel.c +4 -4
index 9fadd58..778262f 100644
@@ -2009,7 +2009,7 @@ static int perf_evsel__parse_id_sample(const struct evsel *evsel,
 				       struct perf_sample *sample)
 {
 	u64 type = evsel->core.attr.sample_type;
-	const u64 *array = event->sample.array;
+	const __u64 *array = event->sample.array;
 	bool swapped = evsel->needs_swap;
 	union u64_swap u;
 
@@ -2099,7 +2099,7 @@ int perf_evsel__parse_sample(struct evsel *evsel, union perf_event *event,
 {
 	u64 type = evsel->core.attr.sample_type;
 	bool swapped = evsel->needs_swap;
-	const u64 *array;
+	const __u64 *array;
 	u16 max_size = event->header.size;
 	const void *endp = (void *)event + max_size;
 	u64 sz;
@@ -2378,7 +2378,7 @@ int perf_evsel__parse_sample_timestamp(struct evsel *evsel,
 				       u64 *timestamp)
 {
 	u64 type = evsel->core.attr.sample_type;
-	const u64 *array;
+	const __u64 *array;
 
 	if (!(type & PERF_SAMPLE_TIME))
 		return -1;
@@ -2529,7 +2529,7 @@ int perf_event__synthesize_sample(union perf_event *event, u64 type,
 				  u64 read_format,
 				  const struct perf_sample *sample)
 {
-	u64 *array;
+	__u64 *array;
 	size_t sz;
 	/*
 	 * used for cross-endian analysis. See git commit 65014ab3
#24 lore raw
In reply to Jiri Olsa

Em Sun, Aug 25, 2019 at 08:17:40PM +0200, Jiri Olsa escreveu:

Quoted 16 lines

hi,
as a preparation for sampling libperf interface, moving event definitions into the library header. Moving just the kernel non-AUX events now.

In order to keep libperf simple, we switch 'u64/u32/u16/u8' types used events to their generic '__u*' versions.

Perf added 'u*' types mainly to ease up printing __u64 values as stated in the linux/types.h comment:

/*
* We define u64 as uint64_t for every architecture
* so that we can print it with "%"PRIx64 without getting warnings. *
* typedef __u64 u64;

Quoted 6 lines
   * typedef __s64 s64;
   */

Adding and using new PRI_lu64 and PRI_lx64 macros to be used for that. Using extra '_' to ease up the reading and differentiate them from standard PRI*64 macros.

I think we should take advantage of this moment to rename those structs to have the 'perf_record_' prefix on them, I guess we could even remove the _event from them, i.e.:

'struct mmap_event' becomes 'perf_record_mmap', as it is the description for the PERF_RECORD_MMAP meta-data event, are you ok with that?

I can go ahead and do it myself, updating each patch on this series to do that.

- Arnaldo

Quoted 34 lines

It's also available in here:
git://git.kernel.org/pub/scm/linux/kernel/git/jolsa/perf.git
perf/fixes

thanks,
jirka

---
Jiri Olsa (12):
libperf: Add mmap_event to perf/event.h
libperf: Add mmap2_event to perf/event.h
libperf: Add comm_event to perf/event.h
libperf: Add namespaces_event to perf/event.h
libperf: Add fork_event to perf/event.h
libperf: Add lost_event to perf/event.h
libperf: Add lost_samples_event to perf/event.h
libperf: Add read_event to perf/event.h
libperf: Add throttle_event to perf/event.h
libperf: Add ksymbol_event to perf/event.h
libperf: Add bpf_event to perf/event.h
libperf: Add sample_event to perf/event.h

 tools/perf/builtin-sched.c          |   2 +-
 tools/perf/lib/include/perf/event.h | 112 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
 tools/perf/util/event.c             |  12 ++++++------
 tools/perf/util/event.h             | 104 +++-----------------------------------------------------------------------------------------------------
 tools/perf/util/evlist.c            |   2 +-
 tools/perf/util/evsel.c             |   8 ++++----
 tools/perf/util/machine.c           |   4 ++--
 tools/perf/util/python.c            |  14 +++++++-------
 tools/perf/util/session.c           |   8 ++++----
 9 files changed, 140 insertions(+), 126 deletions(-)
 create mode 100644 tools/perf/lib/include/perf/event.h
Signature · 1 line
- Arnaldo
#25 lore raw
Quoted 24 lines · Arnaldo Carvalho de Melo

Em Sun, Aug 25, 2019 at 08:17:40PM +0200, Jiri Olsa escreveu:

Quoted 16 lines

hi,
as a preparation for sampling libperf interface, moving event definitions into the library header. Moving just the kernel non-AUX events now.

In order to keep libperf simple, we switch 'u64/u32/u16/u8' types used events to their generic '__u*' versions.

Perf added 'u*' types mainly to ease up printing __u64 values as stated in the linux/types.h comment:

/*
* We define u64 as uint64_t for every architecture
* so that we can print it with "%"PRIx64 without getting warnings. *
* typedef __u64 u64;

Quoted 6 lines
   * typedef __s64 s64;
   */

Adding and using new PRI_lu64 and PRI_lx64 macros to be used for that. Using extra '_' to ease up the reading and differentiate them from standard PRI*64 macros.

Quoted 6 lines

I think we should take advantage of this moment to rename those structs to have the 'perf_record_' prefix on them, I guess we could even remove the _event from them, i.e.:

'struct mmap_event' becomes 'perf_record_mmap', as it is the description for the PERF_RECORD_MMAP meta-data event, are you ok with that?

hum, not sure about loosing the '_event' here, but we are not public yet, so we can always change back ;-) I do like it'd follow the enum name

Quoted 2 lines

I can go ahead and do it myself, updating each patch on this series to do that.

sure, I thought we'd do it later, but feel free to do it, maybe in separate changes?

thanks,
jirka

#26 lore raw

Em Mon, Aug 26, 2019 at 06:47:34PM +0200, Jiri Olsa escreveu:

Quoted 29 lines · Arnaldo Carvalho de Melo
Quoted 28 lines

Em Sun, Aug 25, 2019 at 08:17:40PM +0200, Jiri Olsa escreveu:

Quoted 16 lines

hi,
as a preparation for sampling libperf interface, moving event definitions into the library header. Moving just the kernel non-AUX events now.

In order to keep libperf simple, we switch 'u64/u32/u16/u8' types used events to their generic '__u*' versions.

Perf added 'u*' types mainly to ease up printing __u64 values as stated in the linux/types.h comment:

/*
* We define u64 as uint64_t for every architecture
* so that we can print it with "%"PRIx64 without getting warnings. *
* typedef __u64 u64;

Quoted 6 lines
   * typedef __s64 s64;
   */

Adding and using new PRI_lu64 and PRI_lx64 macros to be used for that. Using extra '_' to ease up the reading and differentiate them from standard PRI*64 macros.

I think we should take advantage of this moment to rename those structs to have the 'perf_record_' prefix on them, I guess we could even remove the _event from them, i.e.:

Quoted 6 lines
Quoted 2 lines

'struct mmap_event' becomes 'perf_record_mmap', as it is the description for the PERF_RECORD_MMAP meta-data event, are you ok with that?

hum, not sure about loosing the '_event' here, but we are not public yet, so we can always change back ;-) I do like it'd follow the enum name

So I'm making the already exported to libperf to be renamed to have the same name as the PERF_RECORD_ enum they map to, just all lowercase.

Looks nice and also makes something exported by libperf to have a perf_ namespace prefix.

BTW: you forgot to move PERF_RECORD_CONTEXT_SWITCH :-)

Quoted 5 lines
Quoted 2 lines

I can go ahead and do it myself, updating each patch on this series to do that.

sure, I thought we'd do it later, but feel free to do it, maybe in separate changes?

I did it as a separate patch, one patch for all the PERF_RECORD_ already moved to libperf.

Also some other minor stuff, like having that perf_event.{bpf,ksymbol}_event renamed to play perf_event.{bpf,ksymbol}, like the other records. so to make this idiom more compact and less redundant:

    event->bpf_event

becomes:

    event->bpf

ditto for ksymbol_event.

- Arnaldo

#27 lore raw
In reply to Jiri Olsa

Em Sun, Aug 25, 2019 at 08:17:40PM +0200, Jiri Olsa escreveu:

Quoted 16 lines

hi,
as a preparation for sampling libperf interface, moving event definitions into the library header. Moving just the kernel non-AUX events now.

In order to keep libperf simple, we switch 'u64/u32/u16/u8' types used events to their generic '__u*' versions.

Perf added 'u*' types mainly to ease up printing __u64 values as stated in the linux/types.h comment:

/*
* We define u64 as uint64_t for every architecture
* so that we can print it with "%"PRIx64 without getting warnings. *
* typedef __u64 u64;

Quoted 6 lines
   * typedef __s64 s64;
   */

Adding and using new PRI_lu64 and PRI_lx64 macros to be used for that. Using extra '_' to ease up the reading and differentiate them from standard PRI*64 macros.

So, this is not building on android env:

builtin-sched.c: In function 'process_lost':
builtin-sched.c:2646:2: error: unknown conversion type character 'l' in format [-Werror=format=]
printf("lost %" PRI_lu64 " events on cpu %d\n", event->lost.lost, sample->cpu);
^
builtin-sched.c:2646:2: error: format '%d' expects argument of type 'int', but argument 2 has type '__u64' [-Werror=format=] builtin-sched.c:2646:2: error: too many arguments for format [-Werror=format-extra-args]
MKDIR /tmp/build/perf/util/

[perfbuilder@490c2c7bdaab ~]$ /opt/android-ndk-r12b//toolchains/arm-linux-androideabi-4.9/prebuilt/linux-x86_64/bin/arm-linux-androideabi-gcc -v
Using built-in specs.
COLLECT_GCC=/opt/android-ndk-r12b//toolchains/arm-linux-androideabi-4.9/prebuilt/linux-x86_64/bin/arm-linux-androideabi-gcc
COLLECT_LTO_WRAPPER=/opt/android-ndk-r12b/toolchains/arm-linux-androideabi-4.9/prebuilt/linux-x86_64/bin/../libexec/gcc/arm-linux-androideabi/4.9.x/lto-wrapper
Target: arm-linux-androideabi
Configured with: /usr/local/google/buildbot/src/android/gcc/toolchain/build/../gcc/gcc-4.9/configure --prefix=/tmp/59719db9ae19ff43aef46bbcb79596b6 --target=arm-linux-androideabi --host=x86_64-linux-gnu --build=x86_64-linux-gnu --with-gnu-as --with-gnu-ld --enable-languages=c,c++ --with-gmp=/buildbot/tmp/build/toolchain/temp-install --with-mpfr=/buildbot/tmp/build/toolchain/temp-install --with-mpc=/buildbot/tmp/build/toolchain/temp-install --with-cloog=/buildbot/tmp/build/toolchain/temp-install --with-isl=/buildbot/tmp/build/toolchain/temp-install --with-ppl=/buildbot/tmp/build/toolchain/temp-install --disable-ppl-version-check --disable-cloog-version-check --disable-isl-version-check --enable-cloog-backend=isl --with-host-libstdcxx='-static-libgcc -Wl,-Bstatic,-lstdc++,-Bdynamic -lm' --disable-libssp --enable-threads --disable-nls --disable-libmudflap --disable-libgomp --disable-libstdc__-v3 --disable-sjlj-exceptions --disable-shared --disable-tls --disable-libitm --with-float=soft --with-fpu=vfp --with-arch=armv5te --enable-target-optspace --enable-bionic-libs --enable-libatomic-ifuncs=no --enable-initfini-array --disable-nls --prefix=/tmp/59719db9ae19ff43aef46bbcb79596b6 --with-sysroot=/tmp/59719db9ae19ff43aef46bbcb79596b6/sysroot --with-binutils-version=2.25 --with-mpfr-version=3.1.1 --with-mpc-version=1.0.1 --with-gmp-version=5.0.5 --with-gcc-version=4.9 --with-gdb-version=none --with-gxx-include-dir=/tmp/59719db9ae19ff43aef46bbcb79596b6/include/c++/4.9.x --with-bugurl=http://source.android.com/source/report-bugs.html --enable-languages=c,c++ --disable-bootstrap --enable-plugins --enable-libgomp --enable-gnu-indirect-function --disable-libsanitizer --enable-gold --enable-threads --enable-eh-frame-hdr-for-static --enable-graphite=yes --with-isl-version=0.11.1 --with-cloog-version=0.18.0 --with-arch=armv5te --program-transform-name='s&^&arm-linux-androideabi-&' --enable-gold=default Thread model: posix
gcc version 4.9.x 20150123 (prerelease) (GCC)
[perfbuilder@490c2c7bdaab ~]$

It doesn't build on the r15b as well.

I'll investigate after lunch.

- Arnaldo

Quoted 34 lines

It's also available in here:
git://git.kernel.org/pub/scm/linux/kernel/git/jolsa/perf.git
perf/fixes

thanks,
jirka

---
Jiri Olsa (12):
libperf: Add mmap_event to perf/event.h
libperf: Add mmap2_event to perf/event.h
libperf: Add comm_event to perf/event.h
libperf: Add namespaces_event to perf/event.h
libperf: Add fork_event to perf/event.h
libperf: Add lost_event to perf/event.h
libperf: Add lost_samples_event to perf/event.h
libperf: Add read_event to perf/event.h
libperf: Add throttle_event to perf/event.h
libperf: Add ksymbol_event to perf/event.h
libperf: Add bpf_event to perf/event.h
libperf: Add sample_event to perf/event.h

 tools/perf/builtin-sched.c          |   2 +-
 tools/perf/lib/include/perf/event.h | 112 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
 tools/perf/util/event.c             |  12 ++++++------
 tools/perf/util/event.h             | 104 +++-----------------------------------------------------------------------------------------------------
 tools/perf/util/evlist.c            |   2 +-
 tools/perf/util/evsel.c             |   8 ++++----
 tools/perf/util/machine.c           |   4 ++--
 tools/perf/util/python.c            |  14 +++++++-------
 tools/perf/util/session.c           |   8 ++++----
 9 files changed, 140 insertions(+), 126 deletions(-)
 create mode 100644 tools/perf/lib/include/perf/event.h
Signature · 1 line
- Arnaldo
#28 lore raw

Em Mon, Aug 26, 2019 at 01:06:28PM -0300, Arnaldo Carvalho de Melo escreveu:

Quoted 45 lines

Em Sun, Aug 25, 2019 at 08:17:40PM +0200, Jiri Olsa escreveu:

Quoted 16 lines

hi,
as a preparation for sampling libperf interface, moving event definitions into the library header. Moving just the kernel non-AUX events now.

In order to keep libperf simple, we switch 'u64/u32/u16/u8' types used events to their generic '__u*' versions.

Perf added 'u*' types mainly to ease up printing __u64 values as stated in the linux/types.h comment:

/*
* We define u64 as uint64_t for every architecture
* so that we can print it with "%"PRIx64 without getting warnings. *
* typedef __u64 u64;

Quoted 6 lines
   * typedef __s64 s64;
   */

Adding and using new PRI_lu64 and PRI_lx64 macros to be used for that. Using extra '_' to ease up the reading and differentiate them from standard PRI*64 macros.

So, this is not building on android env:

builtin-sched.c: In function 'process_lost':
builtin-sched.c:2646:2: error: unknown conversion type character 'l' in format [-Werror=format=]
printf("lost %" PRI_lu64 " events on cpu %d\n", event->lost.lost, sample->cpu);
^
builtin-sched.c:2646:2: error: format '%d' expects argument of type 'int', but argument 2 has type '__u64' [-Werror=format=] builtin-sched.c:2646:2: error: too many arguments for format [-Werror=format-extra-args]
MKDIR /tmp/build/perf/util/

[perfbuilder@490c2c7bdaab ~]$ /opt/android-ndk-r12b//toolchains/arm-linux-androideabi-4.9/prebuilt/linux-x86_64/bin/arm-linux-androideabi-gcc -v
Using built-in specs.
COLLECT_GCC=/opt/android-ndk-r12b//toolchains/arm-linux-androideabi-4.9/prebuilt/linux-x86_64/bin/arm-linux-androideabi-gcc
COLLECT_LTO_WRAPPER=/opt/android-ndk-r12b/toolchains/arm-linux-androideabi-4.9/prebuilt/linux-x86_64/bin/../libexec/gcc/arm-linux-androideabi/4.9.x/lto-wrapper
Target: arm-linux-androideabi
Configured with: /usr/local/google/buildbot/src/android/gcc/toolchain/build/../gcc/gcc-4.9/configure --prefix=/tmp/59719db9ae19ff43aef46bbcb79596b6 --target=arm-linux-androideabi --host=x86_64-linux-gnu --build=x86_64-linux-gnu --with-gnu-as --with-gnu-ld --enable-languages=c,c++ --with-gmp=/buildbot/tmp/build/toolchain/temp-install --with-mpfr=/buildbot/tmp/build/toolchain/temp-install --with-mpc=/buildbot/tmp/build/toolchain/temp-install --with-cloog=/buildbot/tmp/build/toolchain/temp-install --with-isl=/buildbot/tmp/build/toolchain/temp-install --with-ppl=/buildbot/tmp/build/toolchain/temp-install --disable-ppl-version-check --disable-cloog-version-check --disable-isl-version-check --enable-cloog-backend=isl --with-host-libstdcxx='-static-libgcc -Wl,-Bstatic,-lstdc++,-Bdynamic -lm' --disable-libssp --enable-threads --disable-nls --disable-libmudflap --disable-libgomp --disable-libstdc__-v3 --disable-sjlj-exceptions --disable-shared --disable-tls --disable-libitm --with-float=soft --with-fpu=vfp --with-arch=armv5te --enable-target-optspace --enable-bionic-libs --enable-libatomic-ifuncs=no --enable-initfini-array --disable-nls --prefix=/tmp/59719db9ae19ff43aef46bbcb79596b6 --with-sysroot=/tmp/59719db9ae19ff43aef46bbcb79596b6/sysroot --with-binutils-version=2.25 --with-mpfr-version=3.1.1 --with-mpc-version=1.0.1 --with-gmp-version=5.0.5 --with-gcc-version=4.9 --with-gdb-version=none --with-gxx-include-dir=/tmp/59719db9ae19ff43aef46bbcb79596b6/include/c++/4.9.x --with-bugurl=http://source.android.com/source/report-bugs.html --enable-languages=c,c++ --disable-bootstrap --enable-plugins --enable-libgomp --enable-gnu-indirect-function --disable-libsanitizer --enable-gold --enable-threads --enable-eh-frame-hdr-for-static --enable-graphite=yes --with-isl-version=0.11.1 --with-cloog-version=0.18.0 --with-arch=armv5te --program-transform-name='s&^&arm-linux-androideabi-&' --enable-gold=default Thread model: posix
gcc version 4.9.x 20150123 (prerelease) (GCC)
[perfbuilder@490c2c7bdaab ~]$

Quoted 3 lines

It doesn't build on the r15b as well.

I'll investigate after lunch.

$ make $EXTRA_MAKE_ARGS ARCH=$ARCH CROSS_COMPILE=$CROSS_COMPILE EXTRA_CFLAGS="$EXTRA_CFLAGS" -C /git/perf/tools/perf O=/tmp/build/perf /tmp/build/perf/builtin-sched.i

We end up with one 'l' too many?

[perfbuilder@490c2c7bdaab ~]$ grep 'printf("lost' /tmp/build/perf/builtin-sched.i
printf("lost %" "l" "ll""u" " events on cpu %d\n", event->lost.lost, sample->cpu);
[perfbuilder@490c2c7bdaab ~]$

And if we do this on a fedora:30 x86_64:

$ make -C tools/perf O=/tmp/build/perf /tmp/build/perf/builtin-sched.i
[acme@quaco perf]$ grep -A4 'printf("lost' /tmp/build/perf/builtin-sched.i
 printf("lost %" "l" 
# 2646 "builtin-sched.c" 3 4
                "l" "u" 
# 2646 "builtin-sched.c"
                         " events on cpu %d\n", event->lost.lost, sample->cpu);
[acme@quaco perf]$

I.e. on 32-bit arches we shouldn't add that extra "l", right?

I bet the build for the mips/mipsel will fail too, lemme see... Yeah, both failed:

[root@quaco ~]# grep -m1 -A6 -- -Werror=format= dm.log/debian\:experimental-x-mips builtin-sched.c:2646:9: error: unknown conversion type character 'l' in format [-Werror=format=]

  printf("lost %" PRI_lu64 " events on cpu %d\n", event->lost.lost, sample->cpu);
         ^~~~~~~~
In file included from builtin-sched.c:31:
/usr/mips-linux-gnu/include/inttypes.h:47:28: note: format string is defined here
 #  define __PRI64_PREFIX "ll"
                            ^

[root@quaco ~]#

[root@quaco ~]# grep -m1 -A6 -- -Werror=format= dm.log/debian\:experimental-x-mipsel builtin-sched.c:2646:9: error: unknown conversion type character 'l' in format [-Werror=format=]

  printf("lost %" PRI_lu64 " events on cpu %d\n", event->lost.lost, sample->cpu);
         ^~~~~~~~
In file included from builtin-sched.c:31:
/usr/mipsel-linux-gnu/include/inttypes.h:47:28: note: format string is defined here
 #  define __PRI64_PREFIX "ll"
                            ^

[root@quaco ~]#

And also on a uclibc ARC arch container:

[root@quaco ~]# grep -m1 -A6 -- -Werror=format= dm.log/fedora\:24-x-ARC-uClibc builtin-sched.c:2646:9: error: unknown conversion type character 'l' in format [-Werror=format=]

  printf("lost %" PRI_lu64 " events on cpu %d\n", event->lost.lost, sample->cpu);
         ^~~~~~~~
In file included from builtin-sched.c:31:0:
/arc_gnu_2017.09-rc2_prebuilt_uclibc_le_arc700_linux_install/arc-snps-linux-uclibc/sysroot/usr/include/inttypes.h:47:28: note: format string is defined here
 #  define __PRI64_PREFIX "ll"
                            ^

[root@quaco ~]#

The _fix_ will come after lunch :)

- Arnaldo

#29 lore raw

On Mon, Aug 26, 2019 at 01:18:49PM -0300, Arnaldo Carvalho de Melo wrote:

SNIP

Quoted 15 lines

[perfbuilder@490c2c7bdaab ~]$ grep 'printf("lost' /tmp/build/perf/builtin-sched.i
printf("lost %" "l" "ll""u" " events on cpu %d\n", event->lost.lost, sample->cpu);
[perfbuilder@490c2c7bdaab ~]$

And if we do this on a fedora:30 x86_64:

$ make -C tools/perf O=/tmp/build/perf /tmp/build/perf/builtin-sched.i
[acme@quaco perf]$ grep -A4 'printf("lost' /tmp/build/perf/builtin-sched.i
 printf("lost %" "l" 
# 2646 "builtin-sched.c" 3 4
                "l" "u" 
# 2646 "builtin-sched.c"
                         " events on cpu %d\n", event->lost.lost, sample->cpu);
[acme@quaco perf]$
Quoted 1 line

I.e. on 32-bit arches we shouldn't add that extra "l", right?

hum, I guess we could #ifdef it 64/32 bits

Quoted 35 lines

I bet the build for the mips/mipsel will fail too, lemme see... Yeah, both failed:

[root@quaco ~]# grep -m1 -A6 -- -Werror=format= dm.log/debian\:experimental-x-mips builtin-sched.c:2646:9: error: unknown conversion type character 'l' in format [-Werror=format=]

  printf("lost %" PRI_lu64 " events on cpu %d\n", event->lost.lost, sample->cpu);
         ^~~~~~~~
In file included from builtin-sched.c:31:
/usr/mips-linux-gnu/include/inttypes.h:47:28: note: format string is defined here
 #  define __PRI64_PREFIX "ll"
                            ^

[root@quaco ~]#

[root@quaco ~]# grep -m1 -A6 -- -Werror=format= dm.log/debian\:experimental-x-mipsel builtin-sched.c:2646:9: error: unknown conversion type character 'l' in format [-Werror=format=]

  printf("lost %" PRI_lu64 " events on cpu %d\n", event->lost.lost, sample->cpu);
         ^~~~~~~~
In file included from builtin-sched.c:31:
/usr/mipsel-linux-gnu/include/inttypes.h:47:28: note: format string is defined here
 #  define __PRI64_PREFIX "ll"
                            ^

[root@quaco ~]#

And also on a uclibc ARC arch container:

[root@quaco ~]# grep -m1 -A6 -- -Werror=format= dm.log/fedora\:24-x-ARC-uClibc builtin-sched.c:2646:9: error: unknown conversion type character 'l' in format [-Werror=format=]

  printf("lost %" PRI_lu64 " events on cpu %d\n", event->lost.lost, sample->cpu);
         ^~~~~~~~
In file included from builtin-sched.c:31:0:
/arc_gnu_2017.09-rc2_prebuilt_uclibc_le_arc700_linux_install/arc-snps-linux-uclibc/sysroot/usr/include/inttypes.h:47:28: note: format string is defined here
 #  define __PRI64_PREFIX "ll"
                            ^
Quoted 3 lines

[root@quaco ~]#

The _fix_ will come after lunch :)

thanks ;-)

jirka

#30 lore raw

Em Mon, Aug 26, 2019 at 06:58:52PM +0200, Jiri Olsa escreveu:

Quoted 19 lines · Arnaldo Carvalho de Melo

SNIP

Quoted 15 lines

[perfbuilder@490c2c7bdaab ~]$ grep 'printf("lost' /tmp/build/perf/builtin-sched.i
printf("lost %" "l" "ll""u" " events on cpu %d\n", event->lost.lost, sample->cpu);
[perfbuilder@490c2c7bdaab ~]$

And if we do this on a fedora:30 x86_64:

$ make -C tools/perf O=/tmp/build/perf /tmp/build/perf/builtin-sched.i
[acme@quaco perf]$ grep -A4 'printf("lost' /tmp/build/perf/builtin-sched.i
 printf("lost %" "l" 
# 2646 "builtin-sched.c" 3 4
                "l" "u" 
# 2646 "builtin-sched.c"
                         " events on cpu %d\n", event->lost.lost, sample->cpu);
[acme@quaco perf]$
Quoted 3 lines
Quoted 1 line

I.e. on 32-bit arches we shouldn't add that extra "l", right?

hum, I guess we could #ifdef it 64/32 bits

I tried to figure out how to fix this better, but the int-ll64.h versus int-l64.h versus how __u64 is defined got me confused and I ended up with:

#if __WORDSIZE == 64
/*
* /usr/include/inttypes.h uses just 'lu' for PRIu64, but we end up defining
* __u64 as long long unsigned int, and then -Werror=format= kicks in and
* complains of the mismatched types, so use these two special extra PRI
* macros to overcome that.
*/
#define PRI_lu64 "l" PRIu64
#define PRI_lx64 "l" PRIx64
#else
#define PRI_lu64 PRIu64
#define PRI_lx64 PRIx64
#endif

Builds in all the containers I have, 32-bit, 64-bit, old gccs/clangs, new ones, uclibc, musl libc, glibc, etc

Quoted 42 lines
Quoted 35 lines

I bet the build for the mips/mipsel will fail too, lemme see... Yeah, both failed:

[root@quaco ~]# grep -m1 -A6 -- -Werror=format= dm.log/debian\:experimental-x-mips builtin-sched.c:2646:9: error: unknown conversion type character 'l' in format [-Werror=format=]

  printf("lost %" PRI_lu64 " events on cpu %d\n", event->lost.lost, sample->cpu);
         ^~~~~~~~
In file included from builtin-sched.c:31:
/usr/mips-linux-gnu/include/inttypes.h:47:28: note: format string is defined here
 #  define __PRI64_PREFIX "ll"
                            ^

[root@quaco ~]#

[root@quaco ~]# grep -m1 -A6 -- -Werror=format= dm.log/debian\:experimental-x-mipsel builtin-sched.c:2646:9: error: unknown conversion type character 'l' in format [-Werror=format=]

  printf("lost %" PRI_lu64 " events on cpu %d\n", event->lost.lost, sample->cpu);
         ^~~~~~~~
In file included from builtin-sched.c:31:
/usr/mipsel-linux-gnu/include/inttypes.h:47:28: note: format string is defined here
 #  define __PRI64_PREFIX "ll"
                            ^

[root@quaco ~]#

And also on a uclibc ARC arch container:

[root@quaco ~]# grep -m1 -A6 -- -Werror=format= dm.log/fedora\:24-x-ARC-uClibc builtin-sched.c:2646:9: error: unknown conversion type character 'l' in format [-Werror=format=]

  printf("lost %" PRI_lu64 " events on cpu %d\n", event->lost.lost, sample->cpu);
         ^~~~~~~~
In file included from builtin-sched.c:31:0:
/arc_gnu_2017.09-rc2_prebuilt_uclibc_le_arc700_linux_install/arc-snps-linux-uclibc/sysroot/usr/include/inttypes.h:47:28: note: format string is defined here
 #  define __PRI64_PREFIX "ll"
                            ^
Quoted 3 lines

[root@quaco ~]#

The _fix_ will come after lunch :)

thanks ;-)

jirka

Signature · 1 line
- Arnaldo
#31 lore raw

Em Mon, Aug 26, 2019 at 07:14:19PM -0300, Arnaldo Carvalho de Melo escreveu:

Quoted 22 lines

Em Mon, Aug 26, 2019 at 06:58:52PM +0200, Jiri Olsa escreveu:

Quoted 21 lines · Arnaldo Carvalho de Melo

SNIP

Quoted 16 lines

[perfbuilder@490c2c7bdaab ~]$ grep 'printf("lost' /tmp/build/perf/builtin-sched.i
printf("lost %" "l" "ll""u" " events on cpu %d\n", event->lost.lost, sample->cpu);
[perfbuilder@490c2c7bdaab ~]$

And if we do this on a fedora:30 x86_64:

$ make -C tools/perf O=/tmp/build/perf /tmp/build/perf/builtin-sched.i
[acme@quaco perf]$ grep -A4 'printf("lost' /tmp/build/perf/builtin-sched.i
 printf("lost %" "l" 
# 2646 "builtin-sched.c" 3 4
                "l" "u" 
# 2646 "builtin-sched.c"
                         " events on cpu %d\n", event->lost.lost, sample->cpu);
[acme@quaco perf]$

I.e. on 32-bit arches we shouldn't add that extra "l", right?

Quoted 7 lines
Quoted 1 line

hum, I guess we could #ifdef it 64/32 bits

I tried to figure out how to fix this better, but the int-ll64.h versus int-l64.h versus how __u64 is defined got me confused and I ended up with:

#if __WORDSIZE == 64

Make that:

#ifdef __LP64__ to build on Alpine/musl libc.

- Arnaldo

Quoted 62 lines

/*
* /usr/include/inttypes.h uses just 'lu' for PRIu64, but we end up defining
* __u64 as long long unsigned int, and then -Werror=format= kicks in and
* complains of the mismatched types, so use these two special extra PRI
* macros to overcome that.
*/
#define PRI_lu64 "l" PRIu64
#define PRI_lx64 "l" PRIx64
#else
#define PRI_lu64 PRIu64
#define PRI_lx64 PRIx64
#endif

Builds in all the containers I have, 32-bit, 64-bit, old gccs/clangs, new ones, uclibc, musl libc, glibc, etc

Quoted 42 lines
Quoted 35 lines

I bet the build for the mips/mipsel will fail too, lemme see... Yeah, both failed:

[root@quaco ~]# grep -m1 -A6 -- -Werror=format= dm.log/debian\:experimental-x-mips builtin-sched.c:2646:9: error: unknown conversion type character 'l' in format [-Werror=format=]

  printf("lost %" PRI_lu64 " events on cpu %d\n", event->lost.lost, sample->cpu);
         ^~~~~~~~
In file included from builtin-sched.c:31:
/usr/mips-linux-gnu/include/inttypes.h:47:28: note: format string is defined here
 #  define __PRI64_PREFIX "ll"
                            ^

[root@quaco ~]#

[root@quaco ~]# grep -m1 -A6 -- -Werror=format= dm.log/debian\:experimental-x-mipsel builtin-sched.c:2646:9: error: unknown conversion type character 'l' in format [-Werror=format=]

  printf("lost %" PRI_lu64 " events on cpu %d\n", event->lost.lost, sample->cpu);
         ^~~~~~~~
In file included from builtin-sched.c:31:
/usr/mipsel-linux-gnu/include/inttypes.h:47:28: note: format string is defined here
 #  define __PRI64_PREFIX "ll"
                            ^

[root@quaco ~]#

And also on a uclibc ARC arch container:

[root@quaco ~]# grep -m1 -A6 -- -Werror=format= dm.log/fedora\:24-x-ARC-uClibc builtin-sched.c:2646:9: error: unknown conversion type character 'l' in format [-Werror=format=]

  printf("lost %" PRI_lu64 " events on cpu %d\n", event->lost.lost, sample->cpu);
         ^~~~~~~~
In file included from builtin-sched.c:31:0:
/arc_gnu_2017.09-rc2_prebuilt_uclibc_le_arc700_linux_install/arc-snps-linux-uclibc/sysroot/usr/include/inttypes.h:47:28: note: format string is defined here
 #  define __PRI64_PREFIX "ll"
                            ^
Quoted 3 lines

[root@quaco ~]#

The _fix_ will come after lunch :)

thanks ;-)

jirka

Signature · 1 line
- Arnaldo
Signature · 1 line
- Arnaldo
#32 lore raw
Quoted 29 lines · Arnaldo Carvalho de Melo

Em Mon, Aug 26, 2019 at 07:14:19PM -0300, Arnaldo Carvalho de Melo escreveu:

Quoted 28 lines

Em Mon, Aug 26, 2019 at 06:58:52PM +0200, Jiri Olsa escreveu:

Quoted 19 lines · Arnaldo Carvalho de Melo

SNIP

Quoted 15 lines

[perfbuilder@490c2c7bdaab ~]$ grep 'printf("lost' /tmp/build/perf/builtin-sched.i
printf("lost %" "l" "ll""u" " events on cpu %d\n", event->lost.lost, sample->cpu);
[perfbuilder@490c2c7bdaab ~]$

And if we do this on a fedora:30 x86_64:

$ make -C tools/perf O=/tmp/build/perf /tmp/build/perf/builtin-sched.i
[acme@quaco perf]$ grep -A4 'printf("lost' /tmp/build/perf/builtin-sched.i
 printf("lost %" "l" 
# 2646 "builtin-sched.c" 3 4
                "l" "u" 
# 2646 "builtin-sched.c"
                         " events on cpu %d\n", event->lost.lost, sample->cpu);
[acme@quaco perf]$
Quoted 3 lines
Quoted 1 line

I.e. on 32-bit arches we shouldn't add that extra "l", right?

hum, I guess we could #ifdef it 64/32 bits

I tried to figure out how to fix this better, but the int-ll64.h versus int-l64.h versus how __u64 is defined got me confused and I ended up with:

Quoted 5 lines
Quoted 1 line

#if __WORDSIZE == 64

Make that:

#ifdef __LP64__ to build on Alpine/musl libc.

awesome, thanks ;-)

jirka