perf record: Support s390 random socket_id assignment

PATCH Applied Thomas Richter 3 participants 5 messages Jun 11, 2018 – Jun 26, 2018 perf events

Share & tools

Thomas Richter

author

2018-06-11 07:31 UTC #0 lore raw

PATCH 1/2 perf record: Support s390 random socket_id assignment

On s390 the socket identifier assigned to a CPU identifier is random and (depending on the configuration of the LPAR) may be higher than the CPU identifier. This is currently not supported.

Fix this by allowing arbitrary socket identifiers being assigned to CPU id.

Output before:

[root@p23lp27 perf]# ./perf report --header -I -v
...
socket_id number is too big.You may need to upgrade the perf tool. Error:
The perf.data file has no samples!
# ========

 # captured on    : Tue May 29 09:29:57 2018
 # header version : 1
 ...
 # Core ID and Socket ID information is not available
 ...
[root@p23lp27 perf]#

Output after:
[root@p23lp27 perf]# ./perf report --header -I -v
...
Error:
The perf.data file has no samples!
# ========

 # captured on    : Tue May 29 09:29:57 2018
 # header version : 1
 ...
 # CPU 0: Core ID 0, Socket ID 6
 # CPU 1: Core ID 1, Socket ID 3
 # CPU 2: Core ID -1, Socket ID -1
 ...
[root@p23lp27 perf]#

Signed-off-by Thomas Richter <[email protected]>

Reviewed-by Hendrik Brueckner <[email protected]>

---
 tools/perf/util/header.c | 10 +++++++++-
 1 file changed, 9 insertions(+), 1 deletion(-)

tools/perf/util/header.c +9 -1

index a8bff2178fbc..34e9d6b9d945 100644
@@ -2113,6 +2113,7 @@ static int process_cpu_topology(struct feat_fd *ff, void *data __maybe_unused)
 	int cpu_nr = ff->ph->env.nr_cpus_avail;
 	u64 size = 0;
 	struct perf_header *ph = ff->ph;
+	bool do_core_id_test = true;
 
 	ph->env.cpu = calloc(cpu_nr, sizeof(*ph->env.cpu));
 	if (!ph->env.cpu)

@@ -2167,6 +2168,13 @@ static int process_cpu_topology(struct feat_fd *ff, void *data __maybe_unused)
 		return 0;
 	}
 
+	/* On s390 the socket_id number is not related to the numbers of cpus.
+	 * The socket_id number might be higher than the numbers of cpus.
+	 * This depends on the configuration.
+	 */
+	if (ph->env.arch && !strncmp(ph->env.arch, "s390", 4))
+		do_core_id_test = false;
+
 	for (i = 0; i < (u32)cpu_nr; i++) {
 		if (do_read_u32(ff, &nr))
 			goto free_cpu;

@@ -2176,7 +2184,7 @@ static int process_cpu_topology(struct feat_fd *ff, void *data __maybe_unused)
 		if (do_read_u32(ff, &nr))
 			goto free_cpu;
 
-		if (nr != (u32)-1 && nr > (u32)cpu_nr) {
+		if (do_core_id_test && nr != (u32)-1 && nr > (u32)cpu_nr) {
 			pr_debug("socket_id number is too big."
 				 "You may need to upgrade the perf tool.\n");
 			goto free_cpu;

Signature · 1 line

2.14.3

Thomas Richter

author

2018-06-11 07:31 UTC #1 lore raw

PATCH 2/2 perf test: Fix test 39 Session topology for s390

On s390 this test case fails because the socket identifiction numbers assigned to the CPU are higher than the CPU identification numbers.

Fix this by adding the platform architecture into the perf data header flag information. This helps identifiing the test platform and handles s390 specifics in process_cpu_topology().

Output before:
[root@p23lp27 perf]# ./perf test -vvvvv -F 39
39: Session topology :
--- start ---
templ file: /tmp/perf-test-iUv755
socket_id number is too big.You may need to upgrade the perf tool. ---- end ----
Session topology: Skip
[root@p23lp27 perf]#

Output after:
[root@p23lp27 perf]# ./perf test -vvvvv -F 39
39: Session topology :
--- start ---
templ file: /tmp/perf-test-8X8VTs
CPU 0, core 0, socket 6
CPU 1, core 1, socket 3
---- end ----
Session topology: Ok
[root@p23lp27 perf]#

Fixes c84974ed9fb6 ("perf test: Add entry to test cpu topology")

Signed-off-by Thomas Richter <[email protected]>

Reviewed-by Hendrik Brueckner <[email protected]>

---
 tools/perf/tests/topology.c | 1 +
 1 file changed, 1 insertion(+)

tools/perf/tests/topology.c +1 -0

index 40e30a26b23c..9497d02f69e6 100644
@@ -45,6 +45,7 @@ static int session_write_header(char *path)
 
 	perf_header__set_feat(&session->header, HEADER_CPU_TOPOLOGY);
 	perf_header__set_feat(&session->header, HEADER_NRCPUS);
+	perf_header__set_feat(&session->header, HEADER_ARCH);
 
 	session->header.data_size += DATA_SIZE;

Signature · 1 line

2.14.3

tip-bot for Thomas Richter

merge bot

2018-06-26 06:52 UTC #2 lore raw

[tip:perf/urgent] perf test session topology: Fix test on s390

perf test session topology: Fix test on s390

Commit-ID:  b930e62ecd362843002bdf84c2940439822af321
Gitweb:     https://git.kernel.org/tip/b930e62ecd362843002bdf84c2940439822af321
Author:     Thomas Richter <[email protected]>
AuthorDate: Mon, 11 Jun 2018 09:31:53 +0200
Committer:  Arnaldo Carvalho de Melo <[email protected]>
CommitDate: Mon, 25 Jun 2018 11:59:35 -0300

perf test session topology: Fix test on s390

On s390 this test case fails because the socket identifiction numbers assigned to the CPU are higher than the CPU identification numbers.

F/ix this by adding the platform architecture into the perf data header flag information. This helps identifiing the test platform and handles s390 specifics in process_cpu_topology().

Before:

[root@p23lp27 perf]# perf test -vvvvv -F 39
39: Session topology :
--- start ---
templ file: /tmp/perf-test-iUv755
socket_id number is too big.You may need to upgrade the perf tool. ---- end ----
Session topology: Skip
[root@p23lp27 perf]#

After:

[root@p23lp27 perf]# perf test -vvvvv -F 39
39: Session topology :
--- start ---
templ file: /tmp/perf-test-8X8VTs
CPU 0, core 0, socket 6
CPU 1, core 1, socket 3
---- end ----
Session topology: Ok
[root@p23lp27 perf]#

Signed-off-by Thomas Richter <[email protected]>

Reviewed-by Hendrik Brueckner <[email protected]>

Cc Heiko Carstens <[email protected]>

Cc Martin Schwidefsky <[email protected]>

Fixes c84974ed9fb6 ("perf test: Add entry to test cpu topology")

Link http://lkml.kernel.org/r/[email protected] open here

Signed-off-by Arnaldo Carvalho de Melo <[email protected]>

---
 tools/perf/tests/topology.c | 1 +
 1 file changed, 1 insertion(+)

tools/perf/tests/topology.c +1 -0

index 40e30a26b23c..9497d02f69e6 100644
@@ -45,6 +45,7 @@ static int session_write_header(char *path)
 
 	perf_header__set_feat(&session->header, HEADER_CPU_TOPOLOGY);
 	perf_header__set_feat(&session->header, HEADER_NRCPUS);
+	perf_header__set_feat(&session->header, HEADER_ARCH);
 
 	session->header.data_size += DATA_SIZE;

Arnaldo Carvalho de Melo

maintainer Applied

2018-06-13 19:55 UTC #3 lore raw

In reply to Thomas Richter

Em Mon, Jun 11, 2018 at 09:31:52AM +0200, Thomas Richter escreveu:

Quoted 8 lines

On s390 the socket identifier assigned to a CPU identifier is random and (depending on the configuration of the LPAR) may be higher than the CPU identifier. This is currently not supported.
Fix this by allowing arbitrary socket identifiers being assigned to CPU id.
Output before:

Thanks, applied both patches in this series,

- Arnaldo

tip-bot for Thomas Richter

merge bot

2018-06-26 06:52 UTC #4 lore raw

In reply to Thomas Richter

perf record: Support s390 random socket_id assignment

Commit-ID:  01766229533f9bdb1144a41b4345c8c7286da7b4
Gitweb:     https://git.kernel.org/tip/01766229533f9bdb1144a41b4345c8c7286da7b4
Author:     Thomas Richter <[email protected]>
AuthorDate: Mon, 11 Jun 2018 09:31:52 +0200
Committer:  Arnaldo Carvalho de Melo <[email protected]>
CommitDate: Mon, 25 Jun 2018 11:59:35 -0300

perf record: Support s390 random socket_id assignment

On s390 the socket identifier assigned to a CPU identifier is random and (depending on the configuration of the LPAR) may be higher than the CPU identifier. This is currently not supported.

Fix this by allowing arbitrary socket identifiers being assigned to CPU id.

Output before:

[root@p23lp27 perf]# ./perf report --header -I -v
...
socket_id number is too big.You may need to upgrade the perf tool. Error:
The perf.data file has no samples!
# ========

  # captured on    : Tue May 29 09:29:57 2018
  # header version : 1
  ...
  # Core ID and Socket ID information is not available
  ...
  [root@p23lp27 perf]#

Output after:

[root@p23lp27 perf]# ./perf report --header -I -v
...
Error:
The perf.data file has no samples!
# ========

  # captured on    : Tue May 29 09:29:57 2018
  # header version : 1
  ...
  # CPU 0: Core ID 0, Socket ID 6
  # CPU 1: Core ID 1, Socket ID 3
  # CPU 2: Core ID -1, Socket ID -1
  ...
  [root@p23lp27 perf]#

Signed-off-by Thomas Richter <[email protected]>

Reviewed-by Hendrik Brueckner <[email protected]>

Cc Heiko Carstens <[email protected]>

Cc Martin Schwidefsky <[email protected]>

Link http://lkml.kernel.org/r/[email protected] open here

Signed-off-by Arnaldo Carvalho de Melo <[email protected]>

---
 tools/perf/util/header.c | 10 +++++++++-
 1 file changed, 9 insertions(+), 1 deletion(-)

tools/perf/util/header.c +9 -1

index 540cd2dcd3e7..59fcc790c865 100644
@@ -2129,6 +2129,7 @@ static int process_cpu_topology(struct feat_fd *ff, void *data __maybe_unused)
 	int cpu_nr = ff->ph->env.nr_cpus_avail;
 	u64 size = 0;
 	struct perf_header *ph = ff->ph;
+	bool do_core_id_test = true;
 
 	ph->env.cpu = calloc(cpu_nr, sizeof(*ph->env.cpu));
 	if (!ph->env.cpu)

@@ -2183,6 +2184,13 @@ static int process_cpu_topology(struct feat_fd *ff, void *data __maybe_unused)
 		return 0;
 	}
 
+	/* On s390 the socket_id number is not related to the numbers of cpus.
+	 * The socket_id number might be higher than the numbers of cpus.
+	 * This depends on the configuration.
+	 */
+	if (ph->env.arch && !strncmp(ph->env.arch, "s390", 4))
+		do_core_id_test = false;
+
 	for (i = 0; i < (u32)cpu_nr; i++) {
 		if (do_read_u32(ff, &nr))
 			goto free_cpu;

@@ -2192,7 +2200,7 @@ static int process_cpu_topology(struct feat_fd *ff, void *data __maybe_unused)
 		if (do_read_u32(ff, &nr))
 			goto free_cpu;
 
-		if (nr != (u32)-1 && nr > (u32)cpu_nr) {
+		if (do_core_id_test && nr != (u32)-1 && nr > (u32)cpu_nr) {
 			pr_debug("socket_id number is too big."
 				 "You may need to upgrade the perf tool.\n");
 			goto free_cpu;