Introduce perf check subcommand
v10 cover Introduce perf check subcommand
The Problem
===========
Currently the presence of a feature is checked with a combination of perf version --build-options and greps, such as:
perf version --build-options | grep " on .* HAVE_FEATURE"
This relies on the output of perf version, and is a common pattern in tests.
Proposed solution
=================
As suggested by contributors in:
https://lore.kernel.org/linux-perf-users/[email protected]/
Introduce a subcommand "perf check feature", with which scripts can test for presence of a feature or multiple features, such as:
perf check feature HAVE_LIBTRACEEVENT (feature macro)
or
perf check feature libtraceevent (feature name)
or
perf check feature LibTraceEvent (case-insensitive)
or
perf check feature libtraceevent,bpf (multiple features)
The usage of "perf version --build-options | grep" has been replaced in two tests, with "perf check feature" command
Also, to not duplicate the same feature list at multiple places, a new global 'supported_features' array has been introduced in builtin.h, so both commands 'perf check feature' and 'perf version --build-options' use the same array
'supported_features' feature is an array of 'struct feature_support', which also has the name of the feature, macro used to test it's presence, and a is_builtin member, which will be 0 if feature not built-in, and 1 if built-in
Architectures Tested
====================
* x86_64
* ppc64le
Commands ran for testing (Fedora & RHEL):
sudo dnf install -y libtraceevent-devel
make clean
make -j$(nproc)
./perf check feature libtraceevent,bpf; echo Return Code: $? ./perf check feature libtraceevent; echo Return Code: $?
sudo ./perf test -v "task-analyzer"
sudo ./perf test -v "probe libc's inet_pton & backtrace it with ping" sudo ./perf test -v "Use vfs_getname probe to get syscall args filenames"
sudo dnf remove -y libtraceevent-devel
make clean
make NO_LIBTRACEEVENT=1 -j$(nproc)
./perf check feature libtraceevent,bpf; echo Return Code: $? ./perf check feature libtraceevent; echo Return Code: $?
sudo ./perf test -v "task-analyzer"
sudo ./perf test -v "probe libc's inet_pton & backtrace it with ping" sudo ./perf test -v "Use vfs_getname probe to get syscall args filenames"
Git tree
========
Git tree with this patch series applied for testing:
https://github.com/adi-g15-ibm/linux/tree/perf-check-feature-v10
Changelog
=========
v10
+ patch #1: use 'strdup' instead of 'malloc+memcpy'
+ patch #1: replace '-q' with '--quiet' in doc
+ patch #1: add usage for perf check
V9
+ make 'feature' a subcommand instead of an option
+ make feature name/macro check case-insensitive
+ rename 'FEATURE_SUPPORT' as 'FEATURE_STATUS'
+ rebase on upstream perf-tools-next
V8
+ handle return value of 'malloc' in patch #1
+ fix error due to strncpy depending on source string's length
V7
+ modified patch #1 to fix compile issue, and add feature to allow multiple comma-separated features
V6
+ rebased to perf-tools-next/perf-tools-next
+ modified patch #1 to include FEATURE_SUPPORT("bpf_skeletons", HAVE_BPF_SKEL)
V5
+ invert return value of 'has_support', but return value of perf check --feature according to shell convention
V4
+ invert return value of perf check --feature
V3
+ simplified has_support code in builtin-check.c (patch #1)
+ modified patch #3 and patch #4 according to change in return value in patch #1
V2
+ improved the patch series with suggestions from Namhyung
+ fix incorrect return value, added -q option, and moved array definition to perf-check.c
V1
+ changed subcommand name to 'perf check --feature'
+ added documentation for perf check
+ support both macro (eg. HAVE_LIBTRACEEVENT), and name (eg. libtraceevent) as input to 'perf check --feature'
+ change subject and descriptions of all patch mentioning perf check instead of perf build
V0: Previous patch series: https://lore.kernel.org/linux-perf-users/[email protected]/
Aditya Gupta (3):
perf check: introduce check subcommand
perf version: update --build-options to use 'supported_features' array perf tests task_analyzer: use perf check for libtraceevent support
Athira Rajeev (1):
tools/perf/tests: Update probe_vfs_getname.sh script to use perf check
feature
tools/perf/Build | 1 + tools/perf/Documentation/perf-check.txt | 79 ++++++++ tools/perf/builtin-check.c | 179 ++++++++++++++++++ tools/perf/builtin-version.c | 43 +---- tools/perf/builtin.h | 17 ++ tools/perf/perf.c | 1 + .../perf/tests/shell/lib/probe_vfs_getname.sh | 4 +- .../shell/record+probe_libc_inet_pton.sh | 5 +- .../shell/record+script_probe_vfs_getname.sh | 5 +- tools/perf/tests/shell/test_task_analyzer.sh | 4 +- 10 files changed, 297 insertions(+), 41 deletions(-) create mode 100644 tools/perf/Documentation/perf-check.txt create mode 100644 tools/perf/builtin-check.c
Signature · 1 line
2.45.2
PATCH v10 1/4 perf check: introduce check subcommand
Currently the presence of a feature is checked with a combination of perf version --build-options and greps, such as:
perf version --build-options | grep " on .* HAVE_FEATURE"
Instead of this, introduce a subcommand "perf check feature", with which scripts can test for presence of a feature, such as:
perf check feature HAVE_FEATURE
'perf check feature' command is expected to have exit status of 0 if feature is built-in, and 1 if it's not built-in or if feature is not known.
Multiple features can also be passed as a comma-separated list, in which case the exit status will be 1 only if all of the passed features are built-in. For example, with below command, it will have exit status of 0 only if both libtraceevent and bpf are enabled, else 1 in all other cases
perf check feature libtraceevent,bpf
The arguments are case-insensitive.
An array 'supported_features' has also been introduced that can be used by other commands like 'perf version --build-options', so that new features can be added in one place, with the array
---
The entries in 'supported_features' array is just sorted list of features mentioned in 'library_status' function in builtin-version.c
---
--- tools/perf/Build | 1 + tools/perf/Documentation/perf-check.txt | 79 +++++++++++ tools/perf/builtin-check.c | 178 ++++++++++++++++++++++++ tools/perf/builtin.h | 17 +++ tools/perf/perf.c | 1 + 5 files changed, 276 insertions(+) create mode 100644 tools/perf/Documentation/perf-check.txt create mode 100644 tools/perf/builtin-check.c
tools/perf/Build +1 -0
@@ -1,5 +1,6 @@ perf-y += builtin-bench.o perf-y += builtin-annotate.o +perf-y += builtin-check.o perf-y += builtin-config.o perf-y += builtin-diff.o perf-y += builtin-evlist.o
tools/perf/Documentation/perf-check.txt new file +79 -0
79 lines not on this page · raw message
tools/perf/builtin-check.c new file +178 -0
178 lines not on this page · raw message
tools/perf/builtin.h +17 -0
@@ -2,6 +2,22 @@#ifndef BUILTIN_H #define BUILTIN_H +#include <stddef.h> +#include <linux/compiler.h> +#include <tools/config.h> + +struct feature_status { +const char *name; +const char *macro; +int is_builtin; +}; + +#define FEATURE_STATUS (name_, macro_) { \ + .name = name_, \ + .macro = #macro_, \ + .is_builtin = IS_BUILTIN(macro_) } + +extern struct feature_status supported_features[];struct cmdnames ;void list_common_cmds_help (void );
@@ -11,6 +27,7 @@ int cmd_annotate(int argc, const char **argv);int cmd_bench (int argc,const char **argv);int cmd_buildid_cache (int argc,const char **argv);int cmd_buildid_list (int argc,const char **argv); +int cmd_check (int argc,const char **argv);int cmd_config (int argc,const char **argv);int cmd_c2c (int argc,const char **argv);int cmd_diff (int argc,const char **argv);
tools/perf/perf.c +1 -0
@@ -52,6 +52,7 @@ static struct cmd_struct commands[] = { { "archive", NULL, 0 }, { "buildid-cache", cmd_buildid_cache, 0 }, { "buildid-list", cmd_buildid_list, 0 }, + { "check", cmd_check, 0 }, { "config", cmd_config, 0 }, { "c2c", cmd_c2c, 0 }, { "diff", cmd_diff, 0 },
Signature · 1 line
2.45.2
Namhyung Kim
maintainerRe: [PATCH v10 1/4] perf check: introduce check subcommand
Quoted 122 lines · Aditya Gupta
Currently the presence of a feature is checked with a combination of perf version --build-options and greps, such as:
perf version --build-options | grep " on .* HAVE_FEATURE"Instead of this, introduce a subcommand "perf check feature", with which scripts can test for presence of a feature, such as:
perf check feature
HAVE_FEATURE'perf check feature' command is expected to have exit status of 0 if feature is built-in, and 1 if it's not built-in or if feature is not known.
Multiple features can also be passed as a comma-separated list, in which case the exit status will be 1 only if all of the passed features are built-in. For example, with below command, it will have exit status of 0 only if both libtraceevent and bpf are enabled, else 1 in all other cases
perf check feature libtraceevent,bpf
The arguments are case-insensitive.
An array 'supported_features' has also been introduced that can be used by other commands like 'perf version --build-options', so that new features can be added in one place, with the arrayReviewed-by Athira Rajeev <[email protected]>Signed-off-by Aditya Gupta <[email protected]>---
The entries in 'supported_features' array is just sorted list of features mentioned in 'library_status' function inbuiltin-version.c
------ tools/perf/Build | 1 + tools/perf/Documentation/perf-check.txt | 79 +++++++++++ tools/perf/builtin-check.c | 178 ++++++++++++++++++++++++ tools/perf/builtin.h | 17 +++ tools/perf/perf.c | 1 + 5 files changed, 276 insertions(+) create mode 100644 tools/perf/Documentation/perf-check.txt create mode 100644 tools/perf/builtin-check.ctools/perf/Build +1
-0@@ -1,5 +1,6 @@ perf-y += builtin-bench.o perf-y += builtin-annotate.o +perf-y += builtin-check.o perf-y += builtin-config.o perf-y += builtin-diff.o perf-y += builtin-evlist.otools/perf/Documentation/perf-check.txt new file +63
-0@@ -0,0 +1,79 @@ +perf-check(1) +=============== + +NAME +---- +perf-check - check features in perf + +SYNOPSIS +-------- +[verse] +'perf check' [<options>] +'perf check' {feature <feature_list>} [<options>] + +DESCRIPTION +----------- +With no options given, the 'perf check' just prints the perf version +on the standard output. + +If the subcommand 'feature' is used, then status of feature is printed +on the standard output (unless '-q' is also passed), ie. whether it is +compiled-in/built-in or not. +Also, 'perf check feature' returns with exit status 0 if the feature +is built-in, otherwise returns with exit status 1. + +SUBCOMMANDS +----------- + +feature:: + + Print whether feature(s) is compiled-in or not, and also returns with an + exit status of 0, if passed feature(s) are compiled-in, else 1. + + It expects a feature list as an argument. There can be a single feature + name/macro, or multiple features can also be passed as a comma-separated + list, in which case the exit status will be 0 only if all of the passed + features are compiled-in. + + The feature names/macros are case-insensitive. + + Example Usage: + perf check feature libtraceevent + perf check feature HAVE_LIBTRACEEVENT + perf check feature libtraceevent,bpf + + Supported feature names/macro: + aio / HAVE_AIO_SUPPORT + bpf / HAVE_LIBBPF_SUPPORT + bpf_skeletons / HAVE_BPF_SKEL + debuginfod / HAVE_DEBUGINFOD_SUPPORT + dwarf / HAVE_DWARF_SUPPORT + dwarf_getlocations / HAVE_DWARF_GETLOCATIONS_SUPPORT + get_cpuid / HAVE_AUXTRACE_SUPPORT + numa_num_possible_cpus / HAVE_LIBNUMA_SUPPORT + libaudit / HAVE_LIBAUDIT_SUPPORT + libbfd / HAVE_LIBBFD_SUPPORT + libelf / HAVE_LIBELF_SUPPORT + libcrypto / HAVE_LIBCRYPTO_SUPPORT + libdw-dwarf-unwind / HAVE_DWARF_SUPPORT + libnuma / HAVE_LIBNUMA_SUPPORT + libperl / HAVE_LIBPERL_SUPPORT + libpfm4 / HAVE_LIBPFM + libpython / HAVE_LIBPYTHON_SUPPORT + libslang / HAVE_SLANG_SUPPORT
tools/perf/Documentation/perf-check.txt line 73
Quoted 10 lines
tools/perf/Documentation/perf-check.txt new file +10
-0+ libtraceevent / HAVE_LIBTRACEEVENT + libunwind / HAVE_LIBUNWIND_SUPPORT + lzma / HAVE_LZMA_SUPPORT + syscall_table / HAVE_SYSCALL_TABLE_SUPPORT + zlib / HAVE_ZLIB_SUPPORT + zstd / HAVE_ZSTD_SUPPORT + +OPTIONS +------- +--quiet::
Oh, I meant to have both -q and --quiet together.
Quoted 89 lines
tools/perf/Documentation/perf-check.txt new file +6
-0+ Do not print any messages or warnings + + This can be used along with subcommands such as 'perf check feature' + to hide unnecessary output in test scripts, eg. + 'perf check feature --quiet libtraceevent' +tools/perf/builtin-check.c new file +77
-0@@ -0,0 +1,178 @@ +// SPDX-License-Identifier: GPL-2.0 +#include "builtin.h" +#include "color.h" +#include "util/debug.h" +#include "util/header.h" +#include <tools/config.h> +#include <stdbool.h> +#include <stdio.h> +#include <string.h> +#include <subcmd/parse-options.h> + +static const char *const check_subcommands[] = {"feature" ,NULL }; +static struct option check_options[] = { +OPT_BOOLEAN ('q' ,"quiet" , &quiet,"do not show any warnings or messages" ), +OPT_END () +}; +static struct option check_feature_options[] = {OPT_END () }; + +static const char *const check_usage[] = { +"perf check [<subcommand>] [<options>]" , +NULL +}; +static const char *const check_feature_usage[] = { +"perf check feature <feature_list>" , +NULL +}; + +struct feature_status supported_features[] = { +FEATURE_STATUS ("aio" ,HAVE_AIO_SUPPORT ), +FEATURE_STATUS ("bpf" ,HAVE_LIBBPF_SUPPORT ), +FEATURE_STATUS ("bpf_skeletons" ,HAVE_BPF_SKEL ), +FEATURE_STATUS ("debuginfod" ,HAVE_DEBUGINFOD_SUPPORT ), +FEATURE_STATUS ("dwarf" ,HAVE_DWARF_SUPPORT ), +FEATURE_STATUS ("dwarf_getlocations" ,HAVE_DWARF_GETLOCATIONS_SUPPORT ), +FEATURE_STATUS ("dwarf-unwind-support" ,HAVE_DWARF_UNWIND_SUPPORT ), +FEATURE_STATUS ("get_cpuid" ,HAVE_AUXTRACE_SUPPORT ), +FEATURE_STATUS ("libaudit" ,HAVE_LIBAUDIT_SUPPORT ), +FEATURE_STATUS ("libbfd" ,HAVE_LIBBFD_SUPPORT ), +FEATURE_STATUS ("libcapstone" ,HAVE_LIBCAPSTONE_SUPPORT ), +FEATURE_STATUS ("libcrypto" ,HAVE_LIBCRYPTO_SUPPORT ), +FEATURE_STATUS ("libdw-dwarf-unwind" ,HAVE_DWARF_SUPPORT ), +FEATURE_STATUS ("libelf" ,HAVE_LIBELF_SUPPORT ), +FEATURE_STATUS ("libnuma" ,HAVE_LIBNUMA_SUPPORT ), +FEATURE_STATUS ("libopencsd" ,HAVE_CSTRACE_SUPPORT ), +FEATURE_STATUS ("libperl" ,HAVE_LIBPERL_SUPPORT ), +FEATURE_STATUS ("libpfm4" ,HAVE_LIBPFM ), +FEATURE_STATUS ("libpython" ,HAVE_LIBPYTHON_SUPPORT ), +FEATURE_STATUS ("libslang" ,HAVE_SLANG_SUPPORT ), +FEATURE_STATUS ("libtraceevent" ,HAVE_LIBTRACEEVENT ), +FEATURE_STATUS ("libunwind" ,HAVE_LIBUNWIND_SUPPORT ), +FEATURE_STATUS ("lzma" ,HAVE_LZMA_SUPPORT ), +FEATURE_STATUS ("numa_num_possible_cpus" ,HAVE_LIBNUMA_SUPPORT ), +FEATURE_STATUS ("syscall_table" ,HAVE_SYSCALL_TABLE_SUPPORT ), +FEATURE_STATUS ("zlib" ,HAVE_ZLIB_SUPPORT ), +FEATURE_STATUS ("zstd" ,HAVE_ZSTD_SUPPORT ), + +/* this should remain at end, to know the array end */ +FEATURE_STATUS (NULL , _) +}; + +static void on_off_print (const char *status) +{ +printf ("[ " ); + +if (!strcmp (status,"OFF" )) +color_fprintf (stdout,PERF_COLOR_RED ,"%-3s" , status); +else +color_fprintf (stdout,PERF_COLOR_GREEN ,"%-3s" , status); + +printf (" ]" ); +} + +/* Helper function to print status of a feature along with name/macro */ +static void status_print (const char *name,const char *macro, +const char *status) +{ +printf ("%22s: " , name);
tools/perf/builtin-check.c line 87
Quoted 10 lines
tools/perf/builtin-check.c new file +10
-0+on_off_print (status); +printf (" # %s\n" , macro); +} + +#define STATUS (feature) \ +do { \ + if (feature.is_builtin) \ + status_print(feature.name, feature.macro, "on"); \ + else \ + status_print(feature.name, feature.macro, "OFF"); \
It'd be nice if you could align the trailing backslashes.
Quoted 48 lines
tools/perf/builtin-check.c new file +48
-0+}while (0 ) + +/** +* check whether "feature" is built-in with perf +* +* returns: +* 0: NOT built-in or Feature not known +* 1: Built-in +*/ +static int has_support(const char *feature) +{ +for (int i =0 ; supported_features[i].name; ++i) { +if ((strcasecmp (feature, supported_features[i].name) ==0 ) || + (strcasecmp (feature, supported_features[i].macro) ==0 )) { +if (!quiet) +STATUS (supported_features[i]); +return supported_features[i].is_builtin; + } + } + +if (!quiet) +pr_err ("Feature not known: '%s'\n" , feature); + +return 0 ; +} + + +/** +* Usage: 'perf check feature <feature_list>' +* +* <feature_list> can be a single feature name/macro, or a comma-separated list +* of feature names/macros +* eg. argument can be "libtraceevent" or "libtraceevent,bpf" etc +* +* In case of a comma-separated list, feature_enabled will be 1, only if +* all features passed in the string are supported +* +* Note that argv will get modified +*/ +static int subcommand_feature (int argc,const char **argv) +{ +char *feature_list; +char *feature_name; +int feature_enabled; + + argc =parse_options (argc, argv, check_feature_options, + check_feature_usage,0 ); +
tools/perf/builtin-check.c line 145
Quoted 10 lines
tools/perf/builtin-check.c new file +10
-0+if (!argc) +usage_with_options (check_feature_usage, check_feature_options); + +if (argc >1 ) { +pr_err ("Too many arguments passed to 'perf check feature'\n" ); +return -1 ; + } + + feature_enabled =1 ; +/* feature_list is a non-const copy of 'argv[1]' */
You meant argv[0], right?
Quoted 42 lines
tools/perf/builtin-check.c new file +33
-0+ feature_list =strdup (argv[0 ]); +if (!feature_list) { +pr_err ("ERROR: failed to allocate memory for feature list\n" ); +return -1 ; + } + + feature_name =strtok (feature_list,"," ); + +while (feature_name) { + feature_enabled &=has_support (feature_name); + feature_name =strtok (NULL ,"," ); + } + +free (feature_list); + +return !feature_enabled; +} + +int cmd_check (int argc,const char **argv) +{ + argc =parse_options_subcommand (argc, argv, check_options, + check_subcommands, check_usage,0 ); + +if (!argc) +usage_with_options (check_usage, check_options); + +if (strcmp (argv[0 ],"feature" ) ==0 ) +return subcommand_feature (argc, argv); + +/* If no subcommand matched above, print usage help */ +usage_with_options (check_usage, check_options); +return 0 ; +}tools/perf/builtin.h +1
-0@@ -2,6 +2,22 @@#ifndef BUILTIN_H #define BUILTIN_H +#include <stddef.h>
tools/perf/builtin.h line 15
Quoted 10 lines
tools/perf/builtin.h +10
-0+#include <linux/compiler.h> +#include <tools/config.h> + +struct feature_status { +const char *name; +const char *macro; +int is_builtin; +}; + +#define FEATURE_STATUS (name_, macro_) { \
Align please.
Quoted 31 lines
tools/perf/builtin.h +6
-0+ .name = name_, \ + .macro = #macro_, \ + .is_builtin = IS_BUILTIN(macro_) } + +extern struct feature_status supported_features[];struct cmdnames ;void list_common_cmds_help (void );@@ -11,6 +27,7 @@ int cmd_annotate(int argc, const char **argv);int cmd_bench (int argc,const char **argv);int cmd_buildid_cache (int argc,const char **argv);int cmd_buildid_list (int argc,const char **argv); +int cmd_check (int argc,const char **argv);int cmd_config (int argc,const char **argv);int cmd_c2c (int argc,const char **argv);int cmd_diff (int argc,const char **argv);tools/perf/perf.c +1
-0@@ -52,6 +52,7 @@ static struct cmd_struct commands[] = { { "archive", NULL, 0 }, { "buildid-cache", cmd_buildid_cache, 0 }, { "buildid-list", cmd_buildid_list, 0 }, + { "check", cmd_check, 0 }, { "config", cmd_config, 0 }, { "c2c", cmd_c2c, 0 }, { "diff", cmd_diff, 0 },Signature · 1 line
2.45.2
Re: [PATCH v10 1/4] perf check: introduce check subcommand
Hi Namhyung,
Quoted 3 lines · Namhyung Kim
Quoted 1 line
+--quiet::
Oh, I meant to have both -q and --quiet together.
Oh okay, got it, will keep it like this in v12:
+-q:: +--quiet::
Quoted 10 lines
tools/perf/builtin-check.c line 87
Quoted 8 lines
tools/perf/builtin-check.c new file +6
-0+#define STATUS (feature) \ +do { \ + if (feature.is_builtin) \ + status_print(feature.name, feature.macro, "on"); \ + else \ + status_print(feature.name, feature.macro, "OFF"); \It'd be nice if you could align the trailing backslashes.
The alignment is getting messed up due to tabs I guess. I will keep tab width as 8 (tabstop=8) and align accordingly. Any other suggestions for ensuring alignment in future ?
Quoted 6 lines
tools/perf/builtin-check.c line 145
Quoted 4 lines
tools/perf/builtin-check.c new file +2
-0+ feature_enabled =1 ; +/* feature_list is a non-const copy of 'argv[1]' */ You meant argv[0], right?
Yes, and thanks for your patience pointing out these silly mistakes.
Quoted 5 lines
Quoted 3 lines
+#define
FEATURE_STATUS(name_, macro_){ \Align please.
Sure, will align.
Thanks,
Aditya Gupta
Quoted 33 lines
Quoted 33 lines
tools/perf/builtin.h +6
-0+ .name = name_, \ + .macro = #macro_, \ + .is_builtin = IS_BUILTIN(macro_) } + +extern struct feature_status supported_features[];struct cmdnames ;void list_common_cmds_help (void );@@ -11,6 +27,7 @@ int cmd_annotate(int argc, const char **argv);int cmd_bench (int argc,const char **argv);int cmd_buildid_cache (int argc,const char **argv);int cmd_buildid_list (int argc,const char **argv); +int cmd_check (int argc,const char **argv);int cmd_config (int argc,const char **argv);int cmd_c2c (int argc,const char **argv);int cmd_diff (int argc,const char **argv);tools/perf/perf.c +1
-0@@ -52,6 +52,7 @@ static struct cmd_struct commands[] = { { "archive", NULL, 0 }, { "buildid-cache", cmd_buildid_cache, 0 }, { "buildid-list", cmd_buildid_list, 0 }, + { "check", cmd_check, 0 }, { "config", cmd_config, 0 }, { "c2c", cmd_c2c, 0 }, { "diff", cmd_diff, 0 },Signature · 1 line
2.45.2
PATCH v10 2/4 perf version: update --build-options to use 'supported_features' array
In reply to Aditya GuptaNow that the feature list has been duplicated in a global 'supported_features' array, use that array instead of manually checking status of built-in features.
This helps in being consistent with commands such as 'perf check feature', so commands can use the same array, and any new feature can be added at one place, in the 'supported_features' array
--- tools/perf/builtin-version.c | 43 +++++++----------------------------- 1 file changed, 8 insertions(+), 35 deletions(-)
tools/perf/builtin-version.c +8 -35
@@ -46,45 +46,18 @@ static void status_print(const char *name, const char *macro,printf (" # %s\n" , macro); } -#define STATUS (__d, __m) \ -do { \ - if (IS_BUILTIN(__d)) \ - status_print(#__m, #__d, "on"); \ - else \ - status_print(#__m, #__d, "OFF"); \ +#define STATUS (feature) \ +do { \ + if (feature.is_builtin) \ + status_print(feature.name, feature.macro, "on"); \ + else \ + status_print(feature.name, feature.macro, "OFF"); \ } while (0)static void library_status (void ) { -STATUS (HAVE_DWARF_SUPPORT , dwarf); -STATUS (HAVE_DWARF_GETLOCATIONS_SUPPORT , dwarf_getlocations); -#ifndef HAVE_SYSCALL_TABLE_SUPPORT -STATUS (HAVE_LIBAUDIT_SUPPORT , libaudit); -#endif -STATUS (HAVE_SYSCALL_TABLE_SUPPORT , syscall_table); -STATUS (HAVE_LIBBFD_SUPPORT , libbfd); -STATUS (HAVE_DEBUGINFOD_SUPPORT , debuginfod); -STATUS (HAVE_LIBELF_SUPPORT , libelf); -STATUS (HAVE_LIBNUMA_SUPPORT , libnuma); -STATUS (HAVE_LIBNUMA_SUPPORT , numa_num_possible_cpus); -STATUS (HAVE_LIBPERL_SUPPORT , libperl); -STATUS (HAVE_LIBPYTHON_SUPPORT , libpython); -STATUS (HAVE_SLANG_SUPPORT , libslang); -STATUS (HAVE_LIBCRYPTO_SUPPORT , libcrypto); -STATUS (HAVE_LIBUNWIND_SUPPORT , libunwind); -STATUS (HAVE_DWARF_SUPPORT , libdw-dwarf-unwind); -STATUS (HAVE_LIBCAPSTONE_SUPPORT , libcapstone); -STATUS (HAVE_ZLIB_SUPPORT , zlib); -STATUS (HAVE_LZMA_SUPPORT , lzma); -STATUS (HAVE_AUXTRACE_SUPPORT , get_cpuid); -STATUS (HAVE_LIBBPF_SUPPORT , bpf); -STATUS (HAVE_AIO_SUPPORT , aio); -STATUS (HAVE_ZSTD_SUPPORT , zstd); -STATUS (HAVE_LIBPFM , libpfm4); -STATUS (HAVE_LIBTRACEEVENT , libtraceevent); -STATUS (HAVE_BPF_SKEL , bpf_skeletons); -STATUS (HAVE_DWARF_UNWIND_SUPPORT , dwarf-unwind-support); -STATUS (HAVE_CSTRACE_SUPPORT , libopencsd); +for (int i =0 ; supported_features[i].name; ++i) +STATUS (supported_features[i]); }int cmd_version (int argc,const char **argv)
Signature · 1 line
2.45.2
PATCH v10 3/4 perf tests task_analyzer: use perf check for libtraceevent support
In reply to Aditya GuptaCurrently we use output of 'perf version --build-options', to check whether perf was built with libtraceevent support.
Instead, use 'perf check feature libtraceevent' to check for libtraceevent support.
--- tools/perf/tests/shell/test_task_analyzer.sh | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)
tools/perf/tests/shell/test_task_analyzer.sh +2 -2
@@ -52,8 +52,8 @@ find_str_or_fail() { # check if perf is compiled with libtraceevent support skip_no_probe_record_support() { - perf version --build-options | grep -q " OFF .* HAVE_LIBTRACEEVENT" && return 2 - return 0 + perf check feature -q libtraceevent && return 0 + return 2 } prepare_perf_data() {
Signature · 1 line
2.45.2
PATCH v10 4/4 tools/perf/tests: Update probe_vfs_getname.sh script to use perf check feature
In reply to Aditya GuptaFrom: Athira Rajeev <[email protected]>
In probe_vfs_getname.sh, current we use "perf record --dry-run" to check for libtraceevent and skip the test if perf is not build with libtraceevent. Change the check to use "perf check feature" option
--- tools/perf/tests/shell/lib/probe_vfs_getname.sh | 4 ++-- tools/perf/tests/shell/record+probe_libc_inet_pton.sh | 5 ++++- tools/perf/tests/shell/record+script_probe_vfs_getname.sh | 5 ++++- 3 files changed, 10 insertions(+), 4 deletions(-)
tools/perf/tests/shell/lib/probe_vfs_getname.sh +2 -2
@@ -27,7 +27,7 @@ skip_if_no_debuginfo() { # check if perf is compiled with libtraceevent support skip_no_probe_record_support() { if [ $had_vfs_getname -eq 1 ] ; then - perf record --dry-run -e $1 2>&1 | grep "libtraceevent is necessary for tracepoint support" && return 2 - return 1 + perf check feature -q libtraceevent && return 1 + return 2 fi }
tools/perf/tests/shell/record+probe_libc_inet_pton.sh +4 -1
@@ -63,7 +63,10 @@ trace_libc_inet_pton_backtrace() { # Check presence of libtraceevent support to run perf record skip_no_probe_record_support "$event_name/$eventattr/" - [ $? -eq 2 ] && return 2 + if [ $? -eq 2 ]; then + echo "WARN: Skipping test trace_libc_inet_pton_backtrace. No libtraceevent support." + return 2 + fi perf record -e $event_name/$eventattr/ -o $perf_data ping -6 -c 1 ::1 > /dev/null 2>&1 # check if perf data file got created in above step.
tools/perf/tests/shell/record+script_probe_vfs_getname.sh +4 -1
@@ -21,7 +21,10 @@ record_open_file() { echo "Recording open file:" # Check presence of libtraceevent support to run perf record skip_no_probe_record_support "probe:vfs_getname*" - [ $? -eq 2 ] && return 2 + if [ $? -eq 2 ]; then + echo "WARN: Skipping test record_open_file. No libtraceevent support" + return 2 + fi perf record -o ${perfdata} -e probe:vfs_getname\* touch $file }
Signature · 1 line
2.45.2
Quoted 138 lines · Aditya Gupta
The Problem
===========Currently the presence of a feature is checked with a combination of perf version --build-options and greps, such as:
perf version --build-options | grep " on .* HAVE_FEATURE"This relies on the output of perf version, and is a common pattern in tests.
Proposed solution
=================As suggested by contributors in:
https://lore.kernel.org/linux-perf-users/[email protected]/Introduce a subcommand "perf check feature", with which scripts can test for presence of a feature or multiple features, such as:
perf check feature
HAVE_LIBTRACEEVENT(feature macro)or
perf check feature libtraceevent (feature name)
or
perf check feature LibTraceEvent (case-insensitive)
or
perf check feature libtraceevent,bpf (multiple features)
The usage of "perf version --build-options | grep" has been replaced in two tests, with "perf check feature" command
Also, to not duplicate the same feature list at multiple places, a new global '
supported_features' array has been introduced inbuiltin.h, so both commands 'perf check feature' and 'perf version --build-options' use the same array'
supported_features' feature is an array of 'structfeature_support', which also has the name of the feature, macro used to test it's presence, and ais_builtinmember, which will be 0 if feature not built-in, and 1 if built-inArchitectures Tested
====================
*x86_64
* ppc64leCommands ran for testing (Fedora & RHEL):
sudo dnf install -y libtraceevent-devel
make clean
make -j$(nproc)
./perf check feature libtraceevent,bpf; echo Return Code: $? ./perf check feature libtraceevent; echo Return Code: $?sudo ./perf test -v "task-analyzer"
sudo ./perf test -v "probe libc'sinet_pton& backtrace it with ping" sudo ./perf test -v "Usevfs_getnameprobe to get syscall args filenames"sudo dnf remove -y libtraceevent-devel
make clean
make NO_LIBTRACEEVENT=1 -j$(nproc)
./perf check feature libtraceevent,bpf; echo Return Code: $? ./perf check feature libtraceevent; echo Return Code: $?sudo ./perf test -v "task-analyzer"
sudo ./perf test -v "probe libc'sinet_pton& backtrace it with ping" sudo ./perf test -v "Usevfs_getnameprobe to get syscall args filenames"Git tree
========Git tree with this patch series applied for testing:
https://github.com/adi-g15-ibm/linux/tree/perf-check-feature-v10Changelog
=========v10
+ patch #1: use 'strdup' instead of 'malloc+memcpy'
+ patch #1: replace '-q' with '--quiet' in doc
+ patch #1: add usage for perf checkV9
+ make 'feature' a subcommand instead of an option
+ make feature name/macro check case-insensitive
+ rename 'FEATURE_SUPPORT' as 'FEATURE_STATUS'
+ rebase on upstream perf-tools-nextV8
+ handle return value of 'malloc' in patch #1
+ fix error due to strncpy depending on source string's lengthV7
+ modified patch #1 to fix compile issue, and add feature to allow multiple comma-separated featuresV6
+ rebased to perf-tools-next/perf-tools-next
+ modified patch #1 to includeFEATURE_SUPPORT("bpf_skeletons", HAVE_BPF_SKEL)V5
+ invert return value of 'has_support', but return value of perf check --feature according to shell conventionV4
+ invert return value of perf check --featureV3
+ simplifiedhas_supportcode inbuiltin-check.c(patch #1)
+ modified patch #3 and patch #4 according to change in return value in patch #1V2
+ improved the patch series with suggestions from Namhyung
+ fix incorrect return value, added -q option, and moved array definition toperf-check.cV1
+ changed subcommand name to 'perf check --feature'
+ added documentation for perf check
+ support both macro (eg.HAVE_LIBTRACEEVENT), and name (eg. libtraceevent) as input to 'perf check --feature'
+ change subject and descriptions of all patch mentioning perf check instead of perf buildV0: Previous patch series: https://lore.kernel.org/linux-perf-users/[email protected]/
Aditya Gupta (3):
perf check: introduce check subcommand
perf version: update --build-options to use 'supported_features' array perf teststask_analyzer: use perf check for libtraceevent supportAthira Rajeev (1):
tools/perf/tests: Updateprobe_vfs_getname.shscript to use perf check
feature
Quoted 14 lines
tools/perf/Build | 1 + tools/perf/Documentation/perf-check.txt | 79 ++++++++ tools/perf/builtin-check.c | 179 ++++++++++++++++++ tools/perf/builtin-version.c | 43 +---- tools/perf/builtin.h | 17 ++ tools/perf/perf.c | 1 + .../perf/tests/shell/lib/probe_vfs_getname.sh | 4 +- .../shell/record+probe_libc_inet_pton.sh | 5 +- .../shell/record+script_probe_vfs_getname.sh | 5 +- tools/perf/tests/shell/test_task_analyzer.sh | 4 +- 10 files changed, 297 insertions(+), 41 deletions(-) create mode 100644 tools/perf/Documentation/perf-check.txt create mode 100644 tools/perf/builtin-check.c
I have a running linux-next kernel on fedora 40.
I applied your patches onto linux-next head and got these results:
# git am ~/gupta1.patch Applying: perf check: introduce check subcommand .git/rebase-apply/patch:110: new blank line at EOF. + warning: 1 line adds whitespace errors. # git am ~/gupta2.patch
The other patches apply cleanly.
The compile of perf fails with this error:
# make
...
builtin-check.c: In function ‘cmd_check’:
builtin-check.c:167:44: error: passing argument 5 of
‘parse_options_subcommand’ discards ‘const’ qualifier
from pointer target type [-Werror=discarded-qualifiers]
167 | check_subcommands, check_usage, 0);
| ^~~~~~~~~~~
In file included from builtin-check.c:10:
/root/mirror-linux-next/tools/perf/libsubcmd/include/subcmd/
parse-options.h:180:45: note: expected ‘const char **’
but argument is of type ‘const char * const*’
180 | const char *usagestr[], int flags);
| ~~~~~~~~~~~~^~~~~~~~~~
CC builtin-timechart.o
CC builtin-sched.o
cc1: all warnings being treated as errors
make[3]: *** [/root/mirror-linux-next/tools/build/Makefile.build:106: builtin-check.o] Error 1
make[3]: *** Waiting for unfinished jobs....
make[2]: *** [Makefile.perf:728: perf-in.o] Error 2
make[1]: *** [Makefile.perf:265: sub-make] Error 2
make: *** [Makefile:70: all] Error 2
#Hope this helps...
Signature · 9 lines
Thomas Richter, Dept 3303, IBM s390 Linux Development, Boeblingen, Germany -- IBM Deutschland Research & Development GmbH Vorsitzender des Aufsichtsrats: Wolfgang Wendt Geschäftsführung: David Faller Sitz der Gesellschaft: Böblingen / Registergericht: Amtsgericht Stuttgart, HRB 243294
Hi Thomas,
Quoted 11 lines
<...snip...>
I have a running linux-next kernel on fedora 40.
I applied your patches onto linux-next head and got these results:# git am ~/gupta1.patch Applying: perf check: introduce check subcommand .git/rebase-apply/patch:110: new blank line at EOF. + warning: 1 line adds whitespace errors. # git am ~/gupta2.patch
Thanks for reporting this, I removed the blank line at EOF.
Quoted 27 lines
The other patches apply cleanly.
The compile of perf fails with this error:# make ... builtin-check.c: In function ‘cmd_check’: builtin-check.c:167:44: error: passing argument 5 of ‘parse_options_subcommand’ discards ‘const’ qualifier from pointer target type [-Werror=discarded-qualifiers] 167 | check_subcommands, check_usage, 0); | ^~~~~~~~~~~ In file included from builtin-check.c:10: /root/mirror-linux-next/tools/perf/libsubcmd/include/subcmd/ parse-options.h:180:45: note: expected ‘const char **’ but argument is of type ‘const char * const*’ 180 | const char *usagestr[], int flags); | ~~~~~~~~~~~~^~~~~~~~~~ CC builtin-timechart.o CC builtin-sched.o cc1: all warnings being treated as errors make[3]: *** [/root/mirror-linux-next/tools/build/Makefile.build:106: builtin-check.o] Error 1 make[3]: *** Waiting for unfinished jobs.... make[2]: *** [Makefile.perf:728: perf-in.o] Error 2 make[1]: *** [Makefile.perf:265: sub-make] Error 2 make: *** [Makefile:70: all] Error 2 #
Quoted 1 line
Hope this helps...
Thanks, fixed in v11. It occured due to me making last minute changes suggested by checkpatch to prefer 'const * const' instead of 'const *'. Should have tested it again, will take care in future.
- Aditya Gupta
Quoted 12 lines
Signature · 10 lines
-- Thomas Richter, Dept 3303, IBM s390 Linux Development, Boeblingen, Germany -- IBM Deutschland Research & Development GmbH Vorsitzender des Aufsichtsrats: Wolfgang Wendt Geschäftsführung: David Faller Sitz der Gesellschaft: Böblingen / Registergericht: Amtsgericht Stuttgart, HRB 243294