From 0e1bb304aed2810b78d6586bb89ebc7e1755d5d9 Mon Sep 17 00:00:00 2001 From: Ian Rogers Date: Fri, 1 Sep 2023 16:39:45 -0700 Subject: [PATCH 01/22] perf parse-events: Remove unnecessary __maybe_unused ANBZ: #24284 commit 6fcfe54d2c91925ec3054cf25e68064913ca7948 upstream. The parameter head_terms is always used in get_config_terms. Reviewed-by: James Clark Signed-off-by: Ian Rogers Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Ingo Molnar Cc: Jiri Olsa Cc: Kan Liang Cc: Mark Rutland Cc: Namhyung Kim Cc: Peter Zijlstra Cc: Rob Herring Link: https://lore.kernel.org/r/20230901233949.2930562-2-irogers@google.com Signed-off-by: Arnaldo Carvalho de Melo Signed-off-by: Fangyu Yu --- tools/perf/util/parse-events.c | 6 ++---- 1 file changed, 2 insertions(+), 4 deletions(-) diff --git a/tools/perf/util/parse-events.c b/tools/perf/util/parse-events.c index 65608a3cba81..e9e3623f3fed 100644 --- a/tools/perf/util/parse-events.c +++ b/tools/perf/util/parse-events.c @@ -34,8 +34,7 @@ #ifdef PARSER_DEBUG extern int parse_events_debug; #endif -static int get_config_terms(struct list_head *head_config, - struct list_head *head_terms __maybe_unused); +static int get_config_terms(struct list_head *head_config, struct list_head *head_terms); struct event_symbol event_symbols_hw[PERF_COUNT_HW_MAX] = { [PERF_COUNT_HW_CPU_CYCLES] = { @@ -1079,8 +1078,7 @@ static int config_attr(struct perf_event_attr *attr, return 0; } -static int get_config_terms(struct list_head *head_config, - struct list_head *head_terms __maybe_unused) +static int get_config_terms(struct list_head *head_config, struct list_head *head_terms) { #define ADD_CONFIG_TERM(__type, __weak) \ struct evsel_config_term *__t; \ -- Gitee From de027de0dfb806e890a0ca38d04403afac0523de Mon Sep 17 00:00:00 2001 From: Ian Rogers Date: Fri, 1 Sep 2023 16:39:46 -0700 Subject: [PATCH 02/22] perf parse-events: Tidy up str parameter ANBZ: #24284 commit 8f91662ef8be473fb025e170601b0dd75838f7d4 upstream. Add a const and rename str to event_name. Reviewed-by: James Clark Signed-off-by: Ian Rogers Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Ingo Molnar Cc: Jiri Olsa Cc: Kan Liang Cc: Mark Rutland Cc: Namhyung Kim Cc: Peter Zijlstra Cc: Rob Herring Link: https://lore.kernel.org/r/20230901233949.2930562-3-irogers@google.com Signed-off-by: Arnaldo Carvalho de Melo Signed-off-by: Fangyu Yu --- tools/perf/util/parse-events.c | 13 +++++++------ tools/perf/util/parse-events.h | 2 +- 2 files changed, 8 insertions(+), 7 deletions(-) diff --git a/tools/perf/util/parse-events.c b/tools/perf/util/parse-events.c index e9e3623f3fed..283c559a35b4 100644 --- a/tools/perf/util/parse-events.c +++ b/tools/perf/util/parse-events.c @@ -1482,7 +1482,7 @@ int parse_events_add_pmu(struct parse_events_state *parse_state, } int parse_events_multi_pmu_add(struct parse_events_state *parse_state, - char *str, struct list_head *head, + const char *event_name, struct list_head *head, struct list_head **listp, void *loc_) { struct parse_events_term *term; @@ -1502,7 +1502,8 @@ int parse_events_multi_pmu_add(struct parse_events_state *parse_state, INIT_LIST_HEAD(head); } - config = strdup(str); + + config = strdup(event_name); if (!config) goto out_err; @@ -1528,7 +1529,7 @@ int parse_events_multi_pmu_add(struct parse_events_state *parse_state, if (parse_events__filter_pmu(parse_state, pmu)) continue; - if (!perf_pmu__have_event(pmu, str)) + if (!perf_pmu__have_event(pmu, event_name)) continue; auto_merge_stats = perf_pmu__auto_merge_stats(pmu); @@ -1539,7 +1540,7 @@ int parse_events_multi_pmu_add(struct parse_events_state *parse_state, strbuf_init(&sb, /*hint=*/ 0); parse_events_term__to_strbuf(orig_head, &sb); - pr_debug("%s -> %s/%s/\n", str, pmu->name, sb.buf); + pr_debug("%s -> %s/%s/\n", event_name, pmu->name, sb.buf); strbuf_release(&sb); ok++; } @@ -1547,13 +1548,13 @@ int parse_events_multi_pmu_add(struct parse_events_state *parse_state, } if (parse_state->fake_pmu) { - if (!parse_events_add_pmu(parse_state, list, str, head, + if (!parse_events_add_pmu(parse_state, list, event_name, head, /*auto_merge_stats=*/true, loc)) { struct strbuf sb; strbuf_init(&sb, /*hint=*/ 0); parse_events_term__to_strbuf(head, &sb); - pr_debug("%s -> %s/%s/\n", str, "fake_pmu", sb.buf); + pr_debug("%s -> %s/%s/\n", event_name, "fake_pmu", sb.buf); strbuf_release(&sb); ok++; } diff --git a/tools/perf/util/parse-events.h b/tools/perf/util/parse-events.h index 594e5d2dc67f..36a67ef7b35a 100644 --- a/tools/perf/util/parse-events.h +++ b/tools/perf/util/parse-events.h @@ -217,7 +217,7 @@ struct evsel *parse_events__add_event(int idx, struct perf_event_attr *attr, struct perf_pmu *pmu); int parse_events_multi_pmu_add(struct parse_events_state *parse_state, - char *str, + const char *event_name, struct list_head *head_config, struct list_head **listp, void *loc); -- Gitee From 27da2a6e29818a55a547e0b7e879b7065c7005fc Mon Sep 17 00:00:00 2001 From: Ian Rogers Date: Fri, 1 Sep 2023 16:39:47 -0700 Subject: [PATCH 03/22] perf parse-events: Avoid enum casts ANBZ: #24284 commit 4163644818e95ea6b0afb3982b34c4d59ed50bb2 upstream. Add term_type to union of values returned by the lexer to avoid casts to and from an integer. Reviewed-by: James Clark Signed-off-by: Ian Rogers Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Ingo Molnar Cc: Jiri Olsa Cc: Kan Liang Cc: Mark Rutland Cc: Namhyung Kim Cc: Peter Zijlstra Cc: Rob Herring Link: https://lore.kernel.org/r/20230901233949.2930562-4-irogers@google.com Signed-off-by: Arnaldo Carvalho de Melo Signed-off-by: Fangyu Yu --- tools/perf/util/parse-events.l | 2 +- tools/perf/util/parse-events.y | 25 +++++++++++-------------- 2 files changed, 12 insertions(+), 15 deletions(-) diff --git a/tools/perf/util/parse-events.l b/tools/perf/util/parse-events.l index 4ef4b6f171a0..7bdf0565a92c 100644 --- a/tools/perf/util/parse-events.l +++ b/tools/perf/util/parse-events.l @@ -120,7 +120,7 @@ static int term(yyscan_t scanner, enum parse_events__term_type type) { YYSTYPE *yylval = parse_events_get_lval(scanner); - yylval->num = type; + yylval->term_type = type; return PE_TERM; } diff --git a/tools/perf/util/parse-events.y b/tools/perf/util/parse-events.y index c3a86ef4b7cf..720630202d4c 100644 --- a/tools/perf/util/parse-events.y +++ b/tools/perf/util/parse-events.y @@ -70,7 +70,7 @@ static void free_list_evsel(struct list_head* list_evsel) %type PE_VALUE_SYM_HW %type PE_VALUE_SYM_SW %type PE_VALUE_SYM_TOOL -%type PE_TERM +%type PE_TERM %type value_sym %type PE_RAW %type PE_NAME @@ -112,6 +112,7 @@ static void free_list_evsel(struct list_head* list_evsel) { char *str; u64 num; + enum parse_events__term_type term_type; struct list_head *list_evsel; struct list_head *list_terms; struct parse_events_term *term; @@ -777,8 +778,7 @@ PE_TERM_HW PE_TERM '=' name_or_raw { struct parse_events_term *term; - int err = parse_events_term__str(&term, (enum parse_events__term_type)$1, - /*config=*/NULL, $3, &@1, &@3); + int err = parse_events_term__str(&term, $1, /*config=*/NULL, $3, &@1, &@3); if (err) { free($3); @@ -790,8 +790,7 @@ PE_TERM '=' name_or_raw PE_TERM '=' PE_TERM_HW { struct parse_events_term *term; - int err = parse_events_term__str(&term, (enum parse_events__term_type)$1, - /*config=*/NULL, $3.str, &@1, &@3); + int err = parse_events_term__str(&term, $1, /*config=*/NULL, $3.str, &@1, &@3); if (err) { free($3.str); @@ -803,10 +802,7 @@ PE_TERM '=' PE_TERM_HW PE_TERM '=' PE_TERM { struct parse_events_term *term; - int err = parse_events_term__term(&term, - (enum parse_events__term_type)$1, - (enum parse_events__term_type)$3, - &@1, &@3); + int err = parse_events_term__term(&term, $1, $3, &@1, &@3); if (err) PE_ABORT(err); @@ -817,8 +813,9 @@ PE_TERM '=' PE_TERM PE_TERM '=' PE_VALUE { struct parse_events_term *term; - int err = parse_events_term__num(&term, (enum parse_events__term_type)$1, - /*config=*/NULL, $3, /*novalue=*/false, &@1, &@3); + int err = parse_events_term__num(&term, $1, + /*config=*/NULL, $3, /*novalue=*/false, + &@1, &@3); if (err) PE_ABORT(err); @@ -829,9 +826,9 @@ PE_TERM '=' PE_VALUE PE_TERM { struct parse_events_term *term; - int err = parse_events_term__num(&term, (enum parse_events__term_type)$1, - /*config=*/NULL, /*num=*/1, /*novalue=*/true, - &@1, /*loc_val=*/NULL); + int err = parse_events_term__num(&term, $1, + /*config=*/NULL, /*num=*/1, /*novalue=*/true, + &@1, /*loc_val=*/NULL); if (err) PE_ABORT(err); -- Gitee From 695ec04fe5ee7102e8698729e83efbb376f65c9f Mon Sep 17 00:00:00 2001 From: Ian Rogers Date: Fri, 1 Sep 2023 16:39:48 -0700 Subject: [PATCH 04/22] perf parse-events: Copy fewer term lists ANBZ: #24284 commit 727adeed06e82915841e121762eb329881ae0107 upstream. When trying to add events to multiple PMUs the term list is copied first as adding the event will rewrite the event's name term into the sysfs and/or json encoding terms (see perf_pmu__check_alias). Change the parse events add API so the passed in term list is const, then copy the list when modification is necessary. Reviewed-by: James Clark Signed-off-by: Ian Rogers Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Ingo Molnar Cc: Jiri Olsa Cc: Kan Liang Cc: Mark Rutland Cc: Namhyung Kim Cc: Peter Zijlstra Cc: Rob Herring Link: https://lore.kernel.org/r/20230901233949.2930562-5-irogers@google.com Signed-off-by: Arnaldo Carvalho de Melo Signed-off-by: Fangyu Yu --- tools/perf/util/parse-events.c | 108 ++++++++++++++++++--------------- tools/perf/util/parse-events.h | 7 +-- tools/perf/util/parse-events.y | 17 +----- 3 files changed, 65 insertions(+), 67 deletions(-) diff --git a/tools/perf/util/parse-events.c b/tools/perf/util/parse-events.c index 283c559a35b4..06a844bcce4a 100644 --- a/tools/perf/util/parse-events.c +++ b/tools/perf/util/parse-events.c @@ -35,6 +35,7 @@ extern int parse_events_debug; #endif static int get_config_terms(struct list_head *head_config, struct list_head *head_terms); +static int parse_events_terms__copy(const struct list_head *src, struct list_head *dest); struct event_symbol event_symbols_hw[PERF_COUNT_HW_MAX] = { [PERF_COUNT_HW_CPU_CYCLES] = { @@ -1367,7 +1368,7 @@ static bool config_term_percore(struct list_head *config_terms) int parse_events_add_pmu(struct parse_events_state *parse_state, struct list_head *list, const char *name, - struct list_head *head_config, + const struct list_head *const_head_terms, bool auto_merge_stats, void *loc_) { struct perf_event_attr attr; @@ -1377,6 +1378,7 @@ int parse_events_add_pmu(struct parse_events_state *parse_state, struct parse_events_error *err = parse_state->error; YYLTYPE *loc = loc_; LIST_HEAD(config_terms); + LIST_HEAD(head_terms); pmu = parse_state->fake_pmu ?: perf_pmus__find(name); @@ -1390,32 +1392,37 @@ int parse_events_add_pmu(struct parse_events_state *parse_state, return -EINVAL; } + if (const_head_terms) { + int ret = parse_events_terms__copy(const_head_terms, &head_terms); + + if (ret) + return ret; + } + if (verbose > 1) { struct strbuf sb; strbuf_init(&sb, /*hint=*/ 0); - if (pmu->selectable && !head_config) { + if (pmu->selectable && list_empty(&head_terms)) { strbuf_addf(&sb, "%s//", name); } else { strbuf_addf(&sb, "%s/", name); - parse_events_term__to_strbuf(head_config, &sb); + parse_events_term__to_strbuf(&head_terms, &sb); strbuf_addch(&sb, '/'); } fprintf(stderr, "Attempt to add: %s\n", sb.buf); strbuf_release(&sb); } - if (head_config) - fix_raw(head_config, pmu); + fix_raw(&head_terms, pmu); if (pmu->default_config) { - memcpy(&attr, pmu->default_config, - sizeof(struct perf_event_attr)); + memcpy(&attr, pmu->default_config, sizeof(struct perf_event_attr)); } else { memset(&attr, 0, sizeof(attr)); } attr.type = pmu->type; - if (!head_config) { + if (list_empty(&head_terms)) { evsel = __add_event(list, &parse_state->idx, &attr, /*init_attr=*/true, /*name=*/NULL, /*metric_id=*/NULL, pmu, @@ -1424,14 +1431,16 @@ int parse_events_add_pmu(struct parse_events_state *parse_state, return evsel ? 0 : -ENOMEM; } - if (!parse_state->fake_pmu && perf_pmu__check_alias(pmu, head_config, &info, err)) + if (!parse_state->fake_pmu && perf_pmu__check_alias(pmu, &head_terms, &info, err)) { + parse_events_terms__purge(&head_terms); return -EINVAL; + } if (verbose > 1) { struct strbuf sb; strbuf_init(&sb, /*hint=*/ 0); - parse_events_term__to_strbuf(head_config, &sb); + parse_events_term__to_strbuf(&head_terms, &sb); fprintf(stderr, "..after resolving event: %s/%s/\n", name, sb.buf); strbuf_release(&sb); } @@ -1440,39 +1449,52 @@ int parse_events_add_pmu(struct parse_events_state *parse_state, * Configure hardcoded terms first, no need to check * return value when called with fail == 0 ;) */ - if (config_attr(&attr, head_config, parse_state->error, config_term_pmu)) + if (config_attr(&attr, &head_terms, parse_state->error, config_term_pmu)) { + parse_events_terms__purge(&head_terms); return -EINVAL; + } - if (get_config_terms(head_config, &config_terms)) + if (get_config_terms(&head_terms, &config_terms)) { + parse_events_terms__purge(&head_terms); return -ENOMEM; + } /* * When using default config, record which bits of attr->config were * changed by the user. */ - if (pmu->default_config && get_config_chgs(pmu, head_config, &config_terms)) + if (pmu->default_config && get_config_chgs(pmu, &head_terms, &config_terms)) { + parse_events_terms__purge(&head_terms); return -ENOMEM; + } - if (!parse_state->fake_pmu && perf_pmu__config(pmu, &attr, head_config, parse_state->error)) { + if (!parse_state->fake_pmu && + perf_pmu__config(pmu, &attr, &head_terms, parse_state->error)) { free_config_terms(&config_terms); + parse_events_terms__purge(&head_terms); return -EINVAL; } evsel = __add_event(list, &parse_state->idx, &attr, /*init_attr=*/true, - get_config_name(head_config), - get_config_metric_id(head_config), pmu, + get_config_name(&head_terms), + get_config_metric_id(&head_terms), pmu, &config_terms, auto_merge_stats, /*cpu_list=*/NULL); - if (!evsel) + if (!evsel) { + parse_events_terms__purge(&head_terms); return -ENOMEM; + } if (evsel->name) evsel->use_config_name = true; evsel->percore = config_term_percore(&evsel->config_terms); - if (parse_state->fake_pmu) + if (parse_state->fake_pmu) { + parse_events_terms__purge(&head_terms); return 0; + } + parse_events_terms__purge(&head_terms); free((char *)evsel->unit); evsel->unit = strdup(info.unit); evsel->scale = info.scale; @@ -1482,25 +1504,25 @@ int parse_events_add_pmu(struct parse_events_state *parse_state, } int parse_events_multi_pmu_add(struct parse_events_state *parse_state, - const char *event_name, struct list_head *head, + const char *event_name, + const struct list_head *const_head_terms, struct list_head **listp, void *loc_) { struct parse_events_term *term; struct list_head *list = NULL; - struct list_head *orig_head = NULL; struct perf_pmu *pmu = NULL; YYLTYPE *loc = loc_; int ok = 0; const char *config; + LIST_HEAD(head_terms); *listp = NULL; - if (!head) { - head = malloc(sizeof(struct list_head)); - if (!head) - goto out_err; + if (const_head_terms) { + int ret = parse_events_terms__copy(const_head_terms, &head_terms); - INIT_LIST_HEAD(head); + if (ret) + return ret; } config = strdup(event_name); @@ -1514,7 +1536,7 @@ int parse_events_multi_pmu_add(struct parse_events_state *parse_state, zfree(&config); goto out_err; } - list_add_tail(&term->list, head); + list_add_tail(&term->list, &head_terms); /* Add it for all PMUs that support the alias */ list = malloc(sizeof(struct list_head)); @@ -1533,27 +1555,25 @@ int parse_events_multi_pmu_add(struct parse_events_state *parse_state, continue; auto_merge_stats = perf_pmu__auto_merge_stats(pmu); - parse_events_copy_term_list(head, &orig_head); if (!parse_events_add_pmu(parse_state, list, pmu->name, - orig_head, auto_merge_stats, loc)) { + &head_terms, auto_merge_stats, loc)) { struct strbuf sb; strbuf_init(&sb, /*hint=*/ 0); - parse_events_term__to_strbuf(orig_head, &sb); + parse_events_term__to_strbuf(&head_terms, &sb); pr_debug("%s -> %s/%s/\n", event_name, pmu->name, sb.buf); strbuf_release(&sb); ok++; } - parse_events_terms__delete(orig_head); } if (parse_state->fake_pmu) { - if (!parse_events_add_pmu(parse_state, list, event_name, head, + if (!parse_events_add_pmu(parse_state, list, event_name, &head_terms, /*auto_merge_stats=*/true, loc)) { struct strbuf sb; strbuf_init(&sb, /*hint=*/ 0); - parse_events_term__to_strbuf(head, &sb); + parse_events_term__to_strbuf(&head_terms, &sb); pr_debug("%s -> %s/%s/\n", event_name, "fake_pmu", sb.buf); strbuf_release(&sb); ok++; @@ -1561,12 +1581,12 @@ int parse_events_multi_pmu_add(struct parse_events_state *parse_state, } out_err: + parse_events_terms__purge(&head_terms); if (ok) *listp = list; else free(list); - parse_events_terms__delete(head); return ok ? 0 : -1; } @@ -2543,27 +2563,19 @@ void parse_events_term__delete(struct parse_events_term *term) free(term); } -int parse_events_copy_term_list(struct list_head *old, - struct list_head **new) +static int parse_events_terms__copy(const struct list_head *src, struct list_head *dest) { - struct parse_events_term *term, *n; - int ret; - - if (!old) { - *new = NULL; - return 0; - } + struct parse_events_term *term; - *new = malloc(sizeof(struct list_head)); - if (!*new) - return -ENOMEM; - INIT_LIST_HEAD(*new); + list_for_each_entry (term, src, list) { + struct parse_events_term *n; + int ret; - list_for_each_entry (term, old, list) { ret = parse_events_term__clone(&n, term); if (ret) return ret; - list_add_tail(&n->list, *new); + + list_add_tail(&n->list, dest); } return 0; } diff --git a/tools/perf/util/parse-events.h b/tools/perf/util/parse-events.h index 36a67ef7b35a..e6612856e881 100644 --- a/tools/perf/util/parse-events.h +++ b/tools/perf/util/parse-events.h @@ -209,7 +209,7 @@ int parse_events_add_breakpoint(struct parse_events_state *parse_state, struct list_head *head_config); int parse_events_add_pmu(struct parse_events_state *parse_state, struct list_head *list, const char *name, - struct list_head *head_config, + const struct list_head *const_head_terms, bool auto_merge_stats, void *loc); struct evsel *parse_events__add_event(int idx, struct perf_event_attr *attr, @@ -218,12 +218,9 @@ struct evsel *parse_events__add_event(int idx, struct perf_event_attr *attr, int parse_events_multi_pmu_add(struct parse_events_state *parse_state, const char *event_name, - struct list_head *head_config, + const struct list_head *head_terms, struct list_head **listp, void *loc); -int parse_events_copy_term_list(struct list_head *old, - struct list_head **new); - void parse_events__set_leader(char *name, struct list_head *list); void parse_events_update_lists(struct list_head *list_event, struct list_head *list_all); diff --git a/tools/perf/util/parse-events.y b/tools/perf/util/parse-events.y index 720630202d4c..d878a040c240 100644 --- a/tools/perf/util/parse-events.y +++ b/tools/perf/util/parse-events.y @@ -275,23 +275,18 @@ event_pmu: PE_NAME opt_pmu_config { struct parse_events_state *parse_state = _parse_state; - struct list_head *list = NULL, *orig_terms = NULL, *terms= NULL; + /* List of created evsels. */ + struct list_head *list = NULL; char *pattern = NULL; #define CLEANUP \ do { \ parse_events_terms__delete($2); \ - parse_events_terms__delete(orig_terms); \ free(list); \ free($1); \ free(pattern); \ } while(0) - if (parse_events_copy_term_list($2, &orig_terms)) { - CLEANUP; - YYNOMEM; - } - list = alloc_list(); if (!list) { CLEANUP; @@ -321,16 +316,11 @@ PE_NAME opt_pmu_config !perf_pmu__match(pattern, pmu->alias_name, $1)) { bool auto_merge_stats = perf_pmu__auto_merge_stats(pmu); - if (parse_events_copy_term_list(orig_terms, &terms)) { - CLEANUP; - YYNOMEM; - } - if (!parse_events_add_pmu(parse_state, list, pmu->name, terms, + if (!parse_events_add_pmu(parse_state, list, pmu->name, $2, auto_merge_stats, &@1)) { ok++; parse_state->wild_card_pmus = true; } - parse_events_terms__delete(terms); } } @@ -338,7 +328,6 @@ PE_NAME opt_pmu_config /* Failure to add, assume $1 is an event name. */ zfree(&list); ok = !parse_events_multi_pmu_add(parse_state, $1, $2, &list, &@1); - $2 = NULL; } if (!ok) { struct parse_events_error *error = parse_state->error; -- Gitee From a8d6a976d636a5858c677435255e0edf49f1eb8a Mon Sep 17 00:00:00 2001 From: Ian Rogers Date: Fri, 1 Sep 2023 16:39:49 -0700 Subject: [PATCH 05/22] perf parse-events: Introduce 'struct parse_events_terms' ANBZ: #24284 commit 0d3f0e6f94ef58d5532e23b6d153b0890cf0014c upstream. parse_events_terms() existed in function names but was passed a 'struct list_head'. As many parse_events functions take an evsel_config list as well as a parse_event_term list, and the naming head_terms and head_config is inconsistent, there's a potential to switch the lists and get errors. Introduce a 'struct parse_events_terms', that just wraps a list_head, to avoid this. Add the regular init/exit functions and transition the code to use them. Reviewed-by: James Clark Signed-off-by: Ian Rogers Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Ingo Molnar Cc: Jiri Olsa Cc: Kan Liang Cc: Mark Rutland Cc: Namhyung Kim Cc: Peter Zijlstra Cc: Rob Herring Link: https://lore.kernel.org/r/20230901233949.2930562-6-irogers@google.com Signed-off-by: Arnaldo Carvalho de Melo Signed-off-by: Fangyu Yu --- tools/perf/arch/x86/util/intel-pt.c | 15 +-- tools/perf/tests/parse-events.c | 12 +-- tools/perf/tests/pmu.c | 23 ++-- tools/perf/util/parse-events.c | 158 +++++++++++++++------------- tools/perf/util/parse-events.h | 29 +++-- tools/perf/util/parse-events.y | 12 +-- tools/perf/util/pmu.c | 49 ++++----- tools/perf/util/pmu.h | 6 +- 8 files changed, 160 insertions(+), 144 deletions(-) diff --git a/tools/perf/arch/x86/util/intel-pt.c b/tools/perf/arch/x86/util/intel-pt.c index dab540b74e87..7b47bbc7afc7 100644 --- a/tools/perf/arch/x86/util/intel-pt.c +++ b/tools/perf/arch/x86/util/intel-pt.c @@ -65,28 +65,23 @@ static int intel_pt_parse_terms_with_default(struct perf_pmu *pmu, const char *str, u64 *config) { - struct list_head *terms; + struct parse_events_terms terms; struct perf_event_attr attr = { .size = 0, }; int err; - terms = malloc(sizeof(struct list_head)); - if (!terms) - return -ENOMEM; - - INIT_LIST_HEAD(terms); - - err = parse_events_terms(terms, str, /*input=*/ NULL); + parse_events_terms__init(&terms); + err = parse_events_terms(&terms, str, /*input=*/ NULL); if (err) goto out_free; attr.config = *config; - err = perf_pmu__config_terms(pmu, &attr, terms, /*zero=*/true, /*err=*/NULL); + err = perf_pmu__config_terms(pmu, &attr, &terms, /*zero=*/true, /*err=*/NULL); if (err) goto out_free; *config = attr.config; out_free: - parse_events_terms__delete(terms); + parse_events_terms__exit(&terms); return err; } diff --git a/tools/perf/tests/parse-events.c b/tools/perf/tests/parse-events.c index 2b66ffba3bb0..f78be21a5999 100644 --- a/tools/perf/tests/parse-events.c +++ b/tools/perf/tests/parse-events.c @@ -771,12 +771,12 @@ static int test__checkevent_pmu_events_mix(struct evlist *evlist) return TEST_OK; } -static int test__checkterms_simple(struct list_head *terms) +static int test__checkterms_simple(struct parse_events_terms *terms) { struct parse_events_term *term; /* config=10 */ - term = list_entry(terms->next, struct parse_events_term, list); + term = list_entry(terms->terms.next, struct parse_events_term, list); TEST_ASSERT_VAL("wrong type term", term->type_term == PARSE_EVENTS__TERM_TYPE_CONFIG); TEST_ASSERT_VAL("wrong type val", @@ -2363,7 +2363,7 @@ static const struct evlist_test test__events_pmu[] = { struct terms_test { const char *str; - int (*check)(struct list_head *terms); + int (*check)(struct parse_events_terms *terms); }; static const struct terms_test test__terms[] = { @@ -2467,11 +2467,11 @@ static int test__events2(struct test_suite *test __maybe_unused, int subtest __m static int test_term(const struct terms_test *t) { - struct list_head terms; + struct parse_events_terms terms; int ret; - INIT_LIST_HEAD(&terms); + parse_events_terms__init(&terms); ret = parse_events_terms(&terms, t->str, /*input=*/ NULL); if (ret) { pr_debug("failed to parse terms '%s', err %d\n", @@ -2480,7 +2480,7 @@ static int test_term(const struct terms_test *t) } ret = t->check(&terms); - parse_events_terms__purge(&terms); + parse_events_terms__exit(&terms); return ret; } diff --git a/tools/perf/tests/pmu.c b/tools/perf/tests/pmu.c index eb60e5f66859..8f18127d876a 100644 --- a/tools/perf/tests/pmu.c +++ b/tools/perf/tests/pmu.c @@ -128,30 +128,35 @@ static int test_format_dir_put(char *dir) return system(buf); } -static struct list_head *test_terms_list(void) +static void add_test_terms(struct parse_events_terms *terms) { - static LIST_HEAD(terms); unsigned int i; - for (i = 0; i < ARRAY_SIZE(test_terms); i++) - list_add_tail(&test_terms[i].list, &terms); + for (i = 0; i < ARRAY_SIZE(test_terms); i++) { + struct parse_events_term *clone; - return &terms; + parse_events_term__clone(&clone, &test_terms[i]); + list_add_tail(&clone->list, &terms->terms); + } } static int test__pmu(struct test_suite *test __maybe_unused, int subtest __maybe_unused) { char dir[PATH_MAX]; char *format; - struct list_head *terms = test_terms_list(); + struct parse_events_terms terms; struct perf_event_attr attr; struct perf_pmu *pmu; int fd; int ret; + parse_events_terms__init(&terms); + add_test_terms(&terms); pmu = zalloc(sizeof(*pmu)); - if (!pmu) + if (!pmu) { + parse_events_terms__exit(&terms); return -ENOMEM; + } INIT_LIST_HEAD(&pmu->format); INIT_LIST_HEAD(&pmu->aliases); @@ -159,6 +164,7 @@ static int test__pmu(struct test_suite *test __maybe_unused, int subtest __maybe format = test_format_dir_get(dir, sizeof(dir)); if (!format) { free(pmu); + parse_events_terms__exit(&terms); return -EINVAL; } @@ -175,7 +181,7 @@ static int test__pmu(struct test_suite *test __maybe_unused, int subtest __maybe if (ret) goto out; - ret = perf_pmu__config_terms(pmu, &attr, terms, /*zero=*/false, /*err=*/NULL); + ret = perf_pmu__config_terms(pmu, &attr, &terms, /*zero=*/false, /*err=*/NULL); if (ret) goto out; @@ -191,6 +197,7 @@ static int test__pmu(struct test_suite *test __maybe_unused, int subtest __maybe out: test_format_dir_put(format); perf_pmu__delete(pmu); + parse_events_terms__exit(&terms); return ret; } diff --git a/tools/perf/util/parse-events.c b/tools/perf/util/parse-events.c index 06a844bcce4a..c56e07bd7dd6 100644 --- a/tools/perf/util/parse-events.c +++ b/tools/perf/util/parse-events.c @@ -34,8 +34,9 @@ #ifdef PARSER_DEBUG extern int parse_events_debug; #endif -static int get_config_terms(struct list_head *head_config, struct list_head *head_terms); -static int parse_events_terms__copy(const struct list_head *src, struct list_head *dest); +static int get_config_terms(struct parse_events_terms *head_config, struct list_head *head_terms); +static int parse_events_terms__copy(const struct parse_events_terms *src, + struct parse_events_terms *dest); struct event_symbol event_symbols_hw[PERF_COUNT_HW_MAX] = { [PERF_COUNT_HW_CPU_CYCLES] = { @@ -153,26 +154,27 @@ const char *event_type(int type) return "unknown"; } -static char *get_config_str(struct list_head *head_terms, enum parse_events__term_type type_term) +static char *get_config_str(struct parse_events_terms *head_terms, + enum parse_events__term_type type_term) { struct parse_events_term *term; if (!head_terms) return NULL; - list_for_each_entry(term, head_terms, list) + list_for_each_entry(term, &head_terms->terms, list) if (term->type_term == type_term) return term->val.str; return NULL; } -static char *get_config_metric_id(struct list_head *head_terms) +static char *get_config_metric_id(struct parse_events_terms *head_terms) { return get_config_str(head_terms, PARSE_EVENTS__TERM_TYPE_METRIC_ID); } -static char *get_config_name(struct list_head *head_terms) +static char *get_config_name(struct parse_events_terms *head_terms) { return get_config_str(head_terms, PARSE_EVENTS__TERM_TYPE_NAME); } @@ -188,11 +190,11 @@ static char *get_config_name(struct list_head *head_terms) * @config_terms: the list of terms that may contain a raw term. * @pmu: the PMU to scan for events from. */ -static void fix_raw(struct list_head *config_terms, struct perf_pmu *pmu) +static void fix_raw(struct parse_events_terms *config_terms, struct perf_pmu *pmu) { struct parse_events_term *term; - list_for_each_entry(term, config_terms, list) { + list_for_each_entry(term, &config_terms->terms, list) { u64 num; if (term->type_term != PARSE_EVENTS__TERM_TYPE_RAW) @@ -356,7 +358,7 @@ static int config_term_common(struct perf_event_attr *attr, struct parse_events_term *term, struct parse_events_error *err); static int config_attr(struct perf_event_attr *attr, - struct list_head *head, + struct parse_events_terms *head, struct parse_events_error *err, config_term_func_t config_term); @@ -442,7 +444,7 @@ bool parse_events__filter_pmu(const struct parse_events_state *parse_state, int parse_events_add_cache(struct list_head *list, int *idx, const char *name, struct parse_events_state *parse_state, - struct list_head *head_config) + struct parse_events_terms *head_config) { struct perf_pmu *pmu = NULL; bool found_supported = false; @@ -520,7 +522,7 @@ static void tracepoint_error(struct parse_events_error *e, int err, static int add_tracepoint(struct list_head *list, int *idx, const char *sys_name, const char *evt_name, struct parse_events_error *err, - struct list_head *head_config, void *loc_) + struct parse_events_terms *head_config, void *loc_) { YYLTYPE *loc = loc_; struct evsel *evsel = evsel__newtp_idx(sys_name, evt_name, (*idx)++); @@ -545,7 +547,7 @@ static int add_tracepoint(struct list_head *list, int *idx, static int add_tracepoint_multi_event(struct list_head *list, int *idx, const char *sys_name, const char *evt_name, struct parse_events_error *err, - struct list_head *head_config, YYLTYPE *loc) + struct parse_events_terms *head_config, YYLTYPE *loc) { char *evt_path; struct dirent *evt_ent; @@ -593,7 +595,7 @@ static int add_tracepoint_multi_event(struct list_head *list, int *idx, static int add_tracepoint_event(struct list_head *list, int *idx, const char *sys_name, const char *evt_name, struct parse_events_error *err, - struct list_head *head_config, YYLTYPE *loc) + struct parse_events_terms *head_config, YYLTYPE *loc) { return strpbrk(evt_name, "*?") ? add_tracepoint_multi_event(list, idx, sys_name, evt_name, @@ -605,7 +607,7 @@ static int add_tracepoint_event(struct list_head *list, int *idx, static int add_tracepoint_multi_sys(struct list_head *list, int *idx, const char *sys_name, const char *evt_name, struct parse_events_error *err, - struct list_head *head_config, YYLTYPE *loc) + struct parse_events_terms *head_config, YYLTYPE *loc) { struct dirent *events_ent; DIR *events_dir; @@ -680,7 +682,7 @@ do { \ int parse_events_add_breakpoint(struct parse_events_state *parse_state, struct list_head *list, u64 addr, char *type, u64 len, - struct list_head *head_config __maybe_unused) + struct parse_events_terms *head_config) { struct perf_event_attr attr; LIST_HEAD(config_terms); @@ -1066,20 +1068,20 @@ static int config_term_tracepoint(struct perf_event_attr *attr, #endif static int config_attr(struct perf_event_attr *attr, - struct list_head *head, + struct parse_events_terms *head, struct parse_events_error *err, config_term_func_t config_term) { struct parse_events_term *term; - list_for_each_entry(term, head, list) + list_for_each_entry(term, &head->terms, list) if (config_term(attr, term, err)) return -EINVAL; return 0; } -static int get_config_terms(struct list_head *head_config, struct list_head *head_terms) +static int get_config_terms(struct parse_events_terms *head_config, struct list_head *head_terms) { #define ADD_CONFIG_TERM(__type, __weak) \ struct evsel_config_term *__t; \ @@ -1112,7 +1114,7 @@ do { \ struct parse_events_term *term; - list_for_each_entry(term, head_config, list) { + list_for_each_entry(term, &head_config->terms, list) { switch (term->type_term) { case PARSE_EVENTS__TERM_TYPE_SAMPLE_PERIOD: ADD_CONFIG_TERM_VAL(PERIOD, period, term->val.num, term->weak); @@ -1193,14 +1195,14 @@ do { \ * Add EVSEL__CONFIG_TERM_CFG_CHG where cfg_chg will have a bit set for * each bit of attr->config that the user has changed. */ -static int get_config_chgs(struct perf_pmu *pmu, struct list_head *head_config, +static int get_config_chgs(struct perf_pmu *pmu, struct parse_events_terms *head_config, struct list_head *head_terms) { struct parse_events_term *term; u64 bits = 0; int type; - list_for_each_entry(term, head_config, list) { + list_for_each_entry(term, &head_config->terms, list) { switch (term->type_term) { case PARSE_EVENTS__TERM_TYPE_USER: type = perf_pmu__format_type(pmu, term->config); @@ -1250,7 +1252,7 @@ static int get_config_chgs(struct perf_pmu *pmu, struct list_head *head_config, int parse_events_add_tracepoint(struct list_head *list, int *idx, const char *sys, const char *event, struct parse_events_error *err, - struct list_head *head_config, void *loc_) + struct parse_events_terms *head_config, void *loc_) { YYLTYPE *loc = loc_; #ifdef HAVE_LIBTRACEEVENT @@ -1283,7 +1285,7 @@ int parse_events_add_tracepoint(struct list_head *list, int *idx, static int __parse_events_add_numeric(struct parse_events_state *parse_state, struct list_head *list, struct perf_pmu *pmu, u32 type, u32 extended_type, - u64 config, struct list_head *head_config) + u64 config, struct parse_events_terms *head_config) { struct perf_event_attr attr; LIST_HEAD(config_terms); @@ -1319,7 +1321,7 @@ static int __parse_events_add_numeric(struct parse_events_state *parse_state, int parse_events_add_numeric(struct parse_events_state *parse_state, struct list_head *list, u32 type, u64 config, - struct list_head *head_config, + struct parse_events_terms *head_config, bool wildcard) { struct perf_pmu *pmu = NULL; @@ -1368,7 +1370,7 @@ static bool config_term_percore(struct list_head *config_terms) int parse_events_add_pmu(struct parse_events_state *parse_state, struct list_head *list, const char *name, - const struct list_head *const_head_terms, + const struct parse_events_terms *const_parsed_terms, bool auto_merge_stats, void *loc_) { struct perf_event_attr attr; @@ -1378,7 +1380,7 @@ int parse_events_add_pmu(struct parse_events_state *parse_state, struct parse_events_error *err = parse_state->error; YYLTYPE *loc = loc_; LIST_HEAD(config_terms); - LIST_HEAD(head_terms); + struct parse_events_terms parsed_terms; pmu = parse_state->fake_pmu ?: perf_pmus__find(name); @@ -1392,8 +1394,9 @@ int parse_events_add_pmu(struct parse_events_state *parse_state, return -EINVAL; } - if (const_head_terms) { - int ret = parse_events_terms__copy(const_head_terms, &head_terms); + parse_events_terms__init(&parsed_terms); + if (const_parsed_terms) { + int ret = parse_events_terms__copy(const_parsed_terms, &parsed_terms); if (ret) return ret; @@ -1403,17 +1406,17 @@ int parse_events_add_pmu(struct parse_events_state *parse_state, struct strbuf sb; strbuf_init(&sb, /*hint=*/ 0); - if (pmu->selectable && list_empty(&head_terms)) { + if (pmu->selectable && list_empty(&parsed_terms.terms)) { strbuf_addf(&sb, "%s//", name); } else { strbuf_addf(&sb, "%s/", name); - parse_events_term__to_strbuf(&head_terms, &sb); + parse_events_terms__to_strbuf(&parsed_terms, &sb); strbuf_addch(&sb, '/'); } fprintf(stderr, "Attempt to add: %s\n", sb.buf); strbuf_release(&sb); } - fix_raw(&head_terms, pmu); + fix_raw(&parsed_terms, pmu); if (pmu->default_config) { memcpy(&attr, pmu->default_config, sizeof(struct perf_event_attr)); @@ -1422,7 +1425,7 @@ int parse_events_add_pmu(struct parse_events_state *parse_state, } attr.type = pmu->type; - if (list_empty(&head_terms)) { + if (list_empty(&parsed_terms.terms)) { evsel = __add_event(list, &parse_state->idx, &attr, /*init_attr=*/true, /*name=*/NULL, /*metric_id=*/NULL, pmu, @@ -1431,8 +1434,8 @@ int parse_events_add_pmu(struct parse_events_state *parse_state, return evsel ? 0 : -ENOMEM; } - if (!parse_state->fake_pmu && perf_pmu__check_alias(pmu, &head_terms, &info, err)) { - parse_events_terms__purge(&head_terms); + if (!parse_state->fake_pmu && perf_pmu__check_alias(pmu, &parsed_terms, &info, err)) { + parse_events_terms__exit(&parsed_terms); return -EINVAL; } @@ -1440,7 +1443,7 @@ int parse_events_add_pmu(struct parse_events_state *parse_state, struct strbuf sb; strbuf_init(&sb, /*hint=*/ 0); - parse_events_term__to_strbuf(&head_terms, &sb); + parse_events_terms__to_strbuf(&parsed_terms, &sb); fprintf(stderr, "..after resolving event: %s/%s/\n", name, sb.buf); strbuf_release(&sb); } @@ -1449,13 +1452,13 @@ int parse_events_add_pmu(struct parse_events_state *parse_state, * Configure hardcoded terms first, no need to check * return value when called with fail == 0 ;) */ - if (config_attr(&attr, &head_terms, parse_state->error, config_term_pmu)) { - parse_events_terms__purge(&head_terms); + if (config_attr(&attr, &parsed_terms, parse_state->error, config_term_pmu)) { + parse_events_terms__exit(&parsed_terms); return -EINVAL; } - if (get_config_terms(&head_terms, &config_terms)) { - parse_events_terms__purge(&head_terms); + if (get_config_terms(&parsed_terms, &config_terms)) { + parse_events_terms__exit(&parsed_terms); return -ENOMEM; } @@ -1463,24 +1466,24 @@ int parse_events_add_pmu(struct parse_events_state *parse_state, * When using default config, record which bits of attr->config were * changed by the user. */ - if (pmu->default_config && get_config_chgs(pmu, &head_terms, &config_terms)) { - parse_events_terms__purge(&head_terms); + if (pmu->default_config && get_config_chgs(pmu, &parsed_terms, &config_terms)) { + parse_events_terms__exit(&parsed_terms); return -ENOMEM; } if (!parse_state->fake_pmu && - perf_pmu__config(pmu, &attr, &head_terms, parse_state->error)) { + perf_pmu__config(pmu, &attr, &parsed_terms, parse_state->error)) { free_config_terms(&config_terms); - parse_events_terms__purge(&head_terms); + parse_events_terms__exit(&parsed_terms); return -EINVAL; } evsel = __add_event(list, &parse_state->idx, &attr, /*init_attr=*/true, - get_config_name(&head_terms), - get_config_metric_id(&head_terms), pmu, + get_config_name(&parsed_terms), + get_config_metric_id(&parsed_terms), pmu, &config_terms, auto_merge_stats, /*cpu_list=*/NULL); if (!evsel) { - parse_events_terms__purge(&head_terms); + parse_events_terms__exit(&parsed_terms); return -ENOMEM; } @@ -1490,11 +1493,11 @@ int parse_events_add_pmu(struct parse_events_state *parse_state, evsel->percore = config_term_percore(&evsel->config_terms); if (parse_state->fake_pmu) { - parse_events_terms__purge(&head_terms); + parse_events_terms__exit(&parsed_terms); return 0; } - parse_events_terms__purge(&head_terms); + parse_events_terms__exit(&parsed_terms); free((char *)evsel->unit); evsel->unit = strdup(info.unit); evsel->scale = info.scale; @@ -1505,7 +1508,7 @@ int parse_events_add_pmu(struct parse_events_state *parse_state, int parse_events_multi_pmu_add(struct parse_events_state *parse_state, const char *event_name, - const struct list_head *const_head_terms, + const struct parse_events_terms *const_parsed_terms, struct list_head **listp, void *loc_) { struct parse_events_term *term; @@ -1514,12 +1517,13 @@ int parse_events_multi_pmu_add(struct parse_events_state *parse_state, YYLTYPE *loc = loc_; int ok = 0; const char *config; - LIST_HEAD(head_terms); + struct parse_events_terms parsed_terms; *listp = NULL; - if (const_head_terms) { - int ret = parse_events_terms__copy(const_head_terms, &head_terms); + parse_events_terms__init(&parsed_terms); + if (const_parsed_terms) { + int ret = parse_events_terms__copy(const_parsed_terms, &parsed_terms); if (ret) return ret; @@ -1536,7 +1540,7 @@ int parse_events_multi_pmu_add(struct parse_events_state *parse_state, zfree(&config); goto out_err; } - list_add_tail(&term->list, &head_terms); + list_add_tail(&term->list, &parsed_terms.terms); /* Add it for all PMUs that support the alias */ list = malloc(sizeof(struct list_head)); @@ -1556,11 +1560,11 @@ int parse_events_multi_pmu_add(struct parse_events_state *parse_state, auto_merge_stats = perf_pmu__auto_merge_stats(pmu); if (!parse_events_add_pmu(parse_state, list, pmu->name, - &head_terms, auto_merge_stats, loc)) { + &parsed_terms, auto_merge_stats, loc)) { struct strbuf sb; strbuf_init(&sb, /*hint=*/ 0); - parse_events_term__to_strbuf(&head_terms, &sb); + parse_events_terms__to_strbuf(&parsed_terms, &sb); pr_debug("%s -> %s/%s/\n", event_name, pmu->name, sb.buf); strbuf_release(&sb); ok++; @@ -1568,12 +1572,12 @@ int parse_events_multi_pmu_add(struct parse_events_state *parse_state, } if (parse_state->fake_pmu) { - if (!parse_events_add_pmu(parse_state, list, event_name, &head_terms, + if (!parse_events_add_pmu(parse_state, list, event_name, &parsed_terms, /*auto_merge_stats=*/true, loc)) { struct strbuf sb; strbuf_init(&sb, /*hint=*/ 0); - parse_events_term__to_strbuf(&head_terms, &sb); + parse_events_terms__to_strbuf(&parsed_terms, &sb); pr_debug("%s -> %s/%s/\n", event_name, "fake_pmu", sb.buf); strbuf_release(&sb); ok++; @@ -1581,7 +1585,7 @@ int parse_events_multi_pmu_add(struct parse_events_state *parse_state, } out_err: - parse_events_terms__purge(&head_terms); + parse_events_terms__exit(&parsed_terms); if (ok) *listp = list; else @@ -1851,7 +1855,7 @@ static int parse_events__scanner(const char *str, /* * parse event config string, return a list of event terms. */ -int parse_events_terms(struct list_head *terms, const char *str, FILE *input) +int parse_events_terms(struct parse_events_terms *terms, const char *str, FILE *input) { struct parse_events_state parse_state = { .terms = NULL, @@ -1860,14 +1864,10 @@ int parse_events_terms(struct list_head *terms, const char *str, FILE *input) int ret; ret = parse_events__scanner(str, input, &parse_state); + if (!ret) + list_splice(&parse_state.terms->terms, &terms->terms); - if (!ret) { - list_splice(parse_state.terms, terms); - zfree(&parse_state.terms); - return 0; - } - - parse_events_terms__delete(parse_state.terms); + zfree(&parse_state.terms); return ret; } @@ -2563,11 +2563,12 @@ void parse_events_term__delete(struct parse_events_term *term) free(term); } -static int parse_events_terms__copy(const struct list_head *src, struct list_head *dest) +static int parse_events_terms__copy(const struct parse_events_terms *src, + struct parse_events_terms *dest) { struct parse_events_term *term; - list_for_each_entry (term, src, list) { + list_for_each_entry (term, &src->terms, list) { struct parse_events_term *n; int ret; @@ -2575,38 +2576,43 @@ static int parse_events_terms__copy(const struct list_head *src, struct list_hea if (ret) return ret; - list_add_tail(&n->list, dest); + list_add_tail(&n->list, &dest->terms); } return 0; } -void parse_events_terms__purge(struct list_head *terms) +void parse_events_terms__init(struct parse_events_terms *terms) +{ + INIT_LIST_HEAD(&terms->terms); +} + +void parse_events_terms__exit(struct parse_events_terms *terms) { struct parse_events_term *term, *h; - list_for_each_entry_safe(term, h, terms, list) { + list_for_each_entry_safe(term, h, &terms->terms, list) { list_del_init(&term->list); parse_events_term__delete(term); } } -void parse_events_terms__delete(struct list_head *terms) +void parse_events_terms__delete(struct parse_events_terms *terms) { if (!terms) return; - parse_events_terms__purge(terms); + parse_events_terms__exit(terms); free(terms); } -int parse_events_term__to_strbuf(struct list_head *term_list, struct strbuf *sb) +int parse_events_terms__to_strbuf(const struct parse_events_terms *terms, struct strbuf *sb) { struct parse_events_term *term; bool first = true; - if (!term_list) + if (!terms) return 0; - list_for_each_entry(term, term_list, list) { + list_for_each_entry(term, &terms->terms, list) { int ret; if (!first) { diff --git a/tools/perf/util/parse-events.h b/tools/perf/util/parse-events.h index e6612856e881..63c0a36a4bf1 100644 --- a/tools/perf/util/parse-events.h +++ b/tools/perf/util/parse-events.h @@ -44,7 +44,6 @@ static inline int parse_events(struct evlist *evlist, const char *str, int parse_event(struct evlist *evlist, const char *str); -int parse_events_terms(struct list_head *terms, const char *str, FILE *input); int parse_filter(const struct option *opt, const char *str, int unset); int exclude_perf(const struct option *opt, const char *arg, int unset); @@ -140,6 +139,11 @@ struct parse_events_error { char *first_help; }; +/* A wrapper around a list of terms for the sake of better type safety. */ +struct parse_events_terms { + struct list_head terms; +}; + struct parse_events_state { /* The list parsed events are placed on. */ struct list_head list; @@ -148,7 +152,7 @@ struct parse_events_state { /* Error information. */ struct parse_events_error *error; /* Holds returned terms for term parsing. */ - struct list_head *terms; + struct parse_events_terms *terms; /* Start token. */ int stoken; /* Special fake PMU marker for testing. */ @@ -181,35 +185,38 @@ int parse_events_term__term(struct parse_events_term **term, int parse_events_term__clone(struct parse_events_term **new, struct parse_events_term *term); void parse_events_term__delete(struct parse_events_term *term); -void parse_events_terms__delete(struct list_head *terms); -void parse_events_terms__purge(struct list_head *terms); -int parse_events_term__to_strbuf(struct list_head *term_list, struct strbuf *sb); + +void parse_events_terms__delete(struct parse_events_terms *terms); +void parse_events_terms__init(struct parse_events_terms *terms); +void parse_events_terms__exit(struct parse_events_terms *terms); +int parse_events_terms(struct parse_events_terms *terms, const char *str, FILE *input); +int parse_events_terms__to_strbuf(const struct parse_events_terms *terms, struct strbuf *sb); int parse_events__modifier_event(struct list_head *list, char *str, bool add); int parse_events__modifier_group(struct list_head *list, char *event_mod); int parse_events_name(struct list_head *list, const char *name); int parse_events_add_tracepoint(struct list_head *list, int *idx, const char *sys, const char *event, struct parse_events_error *error, - struct list_head *head_config, void *loc); + struct parse_events_terms *head_config, void *loc); int parse_events_add_numeric(struct parse_events_state *parse_state, struct list_head *list, u32 type, u64 config, - struct list_head *head_config, + struct parse_events_terms *head_config, bool wildcard); int parse_events_add_tool(struct parse_events_state *parse_state, struct list_head *list, int tool_event); int parse_events_add_cache(struct list_head *list, int *idx, const char *name, struct parse_events_state *parse_state, - struct list_head *head_config); + struct parse_events_terms *head_config); int parse_events__decode_legacy_cache(const char *name, int pmu_type, __u64 *config); int parse_events_add_breakpoint(struct parse_events_state *parse_state, struct list_head *list, u64 addr, char *type, u64 len, - struct list_head *head_config); + struct parse_events_terms *head_config); int parse_events_add_pmu(struct parse_events_state *parse_state, struct list_head *list, const char *name, - const struct list_head *const_head_terms, + const struct parse_events_terms *const_parsed_terms, bool auto_merge_stats, void *loc); struct evsel *parse_events__add_event(int idx, struct perf_event_attr *attr, @@ -218,7 +225,7 @@ struct evsel *parse_events__add_event(int idx, struct perf_event_attr *attr, int parse_events_multi_pmu_add(struct parse_events_state *parse_state, const char *event_name, - const struct list_head *head_terms, + const struct parse_events_terms *const_parsed_terms, struct list_head **listp, void *loc); void parse_events__set_leader(char *name, struct list_head *list); diff --git a/tools/perf/util/parse-events.y b/tools/perf/util/parse-events.y index d878a040c240..b46b80da5ea7 100644 --- a/tools/perf/util/parse-events.y +++ b/tools/perf/util/parse-events.y @@ -114,7 +114,7 @@ static void free_list_evsel(struct list_head* list_evsel) u64 num; enum parse_events__term_type term_type; struct list_head *list_evsel; - struct list_head *list_terms; + struct parse_events_terms *list_terms; struct parse_events_term *term; struct tracepoint_name { char *sys; @@ -645,26 +645,26 @@ start_terms: event_config event_config: event_config ',' event_term { - struct list_head *head = $1; + struct parse_events_terms *head = $1; struct parse_events_term *term = $3; if (!head) { parse_events_term__delete(term); YYABORT; } - list_add_tail(&term->list, head); + list_add_tail(&term->list, &head->terms); $$ = $1; } | event_term { - struct list_head *head = malloc(sizeof(*head)); + struct parse_events_terms *head = malloc(sizeof(*head)); struct parse_events_term *term = $1; if (!head) YYNOMEM; - INIT_LIST_HEAD(head); - list_add_tail(&term->list, head); + parse_events_terms__init(head); + list_add_tail(&term->list, &head->terms); $$ = head; } diff --git a/tools/perf/util/pmu.c b/tools/perf/util/pmu.c index 2587c4b463fa..43eac49522ae 100644 --- a/tools/perf/util/pmu.c +++ b/tools/perf/util/pmu.c @@ -65,7 +65,7 @@ struct perf_pmu_alias { */ char *topic; /** @terms: Owned list of the original parsed parameters. */ - struct list_head terms; + struct parse_events_terms terms; /** @list: List element of struct perf_pmu aliases. */ struct list_head list; /** @@ -417,7 +417,7 @@ static void perf_pmu_free_alias(struct perf_pmu_alias *newalias) zfree(&newalias->long_desc); zfree(&newalias->topic); zfree(&newalias->pmu_name); - parse_events_terms__purge(&newalias->terms); + parse_events_terms__exit(&newalias->terms); free(newalias); } @@ -518,7 +518,7 @@ static int update_alias(const struct pmu_event *pe, assign_str(pe->name, "topic", &data->alias->topic, pe->topic); data->alias->per_pkg = pe->perpkg; if (pe->event) { - parse_events_terms__purge(&data->alias->terms); + parse_events_terms__exit(&data->alias->terms); ret = parse_events_terms(&data->alias->terms, pe->event, /*input=*/NULL); } if (!ret && pe->unit) { @@ -558,7 +558,7 @@ static int perf_pmu__new_alias(struct perf_pmu *pmu, const char *name, if (!alias) return -ENOMEM; - INIT_LIST_HEAD(&alias->terms); + parse_events_terms__init(&alias->terms); alias->scale = 1.0; alias->unit[0] = '\0'; alias->per_pkg = perpkg; @@ -696,17 +696,17 @@ static int pmu_aliases_parse(struct perf_pmu *pmu) return 0; } -static int pmu_alias_terms(struct perf_pmu_alias *alias, - struct list_head *terms) +static int pmu_alias_terms(struct perf_pmu_alias *alias, struct list_head *terms) { struct parse_events_term *term, *cloned; - LIST_HEAD(list); - int ret; + struct parse_events_terms clone_terms; + + parse_events_terms__init(&clone_terms); + list_for_each_entry(term, &alias->terms.terms, list) { + int ret = parse_events_term__clone(&cloned, term); - list_for_each_entry(term, &alias->terms, list) { - ret = parse_events_term__clone(&cloned, term); if (ret) { - parse_events_terms__purge(&list); + parse_events_terms__exit(&clone_terms); return ret; } /* @@ -714,9 +714,10 @@ static int pmu_alias_terms(struct perf_pmu_alias *alias, * which we don't want for implicit terms in aliases. */ cloned->weak = true; - list_add_tail(&cloned->list, &list); + list_add_tail(&cloned->list, &clone_terms.terms); } - list_splice(&list, terms); + list_splice_init(&clone_terms.terms, terms); + parse_events_terms__exit(&clone_terms); return 0; } @@ -1257,12 +1258,12 @@ static __u64 pmu_format_max_value(const unsigned long *format) * in a config string) later on in the term list. */ static int pmu_resolve_param_term(struct parse_events_term *term, - struct list_head *head_terms, + struct parse_events_terms *head_terms, __u64 *value) { struct parse_events_term *t; - list_for_each_entry(t, head_terms, list) { + list_for_each_entry(t, &head_terms->terms, list) { if (t->type_val == PARSE_EVENTS__TERM_TYPE_NUM && t->config && !strcmp(t->config, term->config)) { t->used = true; @@ -1306,7 +1307,7 @@ static char *pmu_formats_string(struct list_head *formats) static int pmu_config_term(struct perf_pmu *pmu, struct perf_event_attr *attr, struct parse_events_term *term, - struct list_head *head_terms, + struct parse_events_terms *head_terms, bool zero, struct parse_events_error *err) { struct perf_pmu_format *format; @@ -1428,13 +1429,13 @@ static int pmu_config_term(struct perf_pmu *pmu, int perf_pmu__config_terms(struct perf_pmu *pmu, struct perf_event_attr *attr, - struct list_head *head_terms, + struct parse_events_terms *terms, bool zero, struct parse_events_error *err) { struct parse_events_term *term; - list_for_each_entry(term, head_terms, list) { - if (pmu_config_term(pmu, attr, term, head_terms, zero, err)) + list_for_each_entry(term, &terms->terms, list) { + if (pmu_config_term(pmu, attr, term, terms, zero, err)) return -EINVAL; } @@ -1447,7 +1448,7 @@ int perf_pmu__config_terms(struct perf_pmu *pmu, * 2) pmu format definitions - specified by pmu parameter */ int perf_pmu__config(struct perf_pmu *pmu, struct perf_event_attr *attr, - struct list_head *head_terms, + struct parse_events_terms *head_terms, struct parse_events_error *err) { bool zero = !!pmu->default_config; @@ -1541,7 +1542,7 @@ static int check_info_data(struct perf_pmu *pmu, * Find alias in the terms list and replace it with the terms * defined for the alias */ -int perf_pmu__check_alias(struct perf_pmu *pmu, struct list_head *head_terms, +int perf_pmu__check_alias(struct perf_pmu *pmu, struct parse_events_terms *head_terms, struct perf_pmu_info *info, struct parse_events_error *err) { struct parse_events_term *term, *h; @@ -1558,7 +1559,7 @@ int perf_pmu__check_alias(struct perf_pmu *pmu, struct list_head *head_terms, info->scale = 0.0; info->snapshot = false; - list_for_each_entry_safe(term, h, head_terms, list) { + list_for_each_entry_safe(term, h, &head_terms->terms, list) { alias = pmu_find_alias(pmu, term); if (!alias) continue; @@ -1704,7 +1705,7 @@ static char *format_alias(char *buf, int len, const struct perf_pmu *pmu, : (int)strlen(pmu->name); int used = snprintf(buf, len, "%.*s/%s", pmu_name_len, pmu->name, alias->name); - list_for_each_entry(term, &alias->terms, list) { + list_for_each_entry(term, &alias->terms.terms, list) { if (term->type_val == PARSE_EVENTS__TERM_TYPE_STR) used += snprintf(buf + used, sub_non_neg(len, used), ",%s=%s", term->config, @@ -1764,7 +1765,7 @@ int perf_pmu__for_each_event(struct perf_pmu *pmu, bool skip_duplicate_pmus, info.desc = event->desc; info.long_desc = event->long_desc; info.encoding_desc = buf + buf_used; - parse_events_term__to_strbuf(&event->terms, &sb); + parse_events_terms__to_strbuf(&event->terms, &sb); buf_used += snprintf(buf + buf_used, sizeof(buf) - buf_used, "%s/%s/", info.pmu_name, sb.buf) + 1; info.topic = event->topic; diff --git a/tools/perf/util/pmu.h b/tools/perf/util/pmu.h index 5a03c361cb04..e192010ad3c7 100644 --- a/tools/perf/util/pmu.h +++ b/tools/perf/util/pmu.h @@ -198,15 +198,15 @@ typedef int (*pmu_event_callback)(void *state, struct pmu_event_info *info); void pmu_add_sys_aliases(struct perf_pmu *pmu); int perf_pmu__config(struct perf_pmu *pmu, struct perf_event_attr *attr, - struct list_head *head_terms, + struct parse_events_terms *head_terms, struct parse_events_error *error); int perf_pmu__config_terms(struct perf_pmu *pmu, struct perf_event_attr *attr, - struct list_head *head_terms, + struct parse_events_terms *terms, bool zero, struct parse_events_error *error); __u64 perf_pmu__format_bits(struct perf_pmu *pmu, const char *name); int perf_pmu__format_type(struct perf_pmu *pmu, const char *name); -int perf_pmu__check_alias(struct perf_pmu *pmu, struct list_head *head_terms, +int perf_pmu__check_alias(struct perf_pmu *pmu, struct parse_events_terms *head_terms, struct perf_pmu_info *info, struct parse_events_error *err); int perf_pmu__find_event(struct perf_pmu *pmu, const char *event, void *state, pmu_event_callback cb); -- Gitee From 7cdf91bd485982d1d7a652c36522709d2f4f10d6 Mon Sep 17 00:00:00 2001 From: Ian Rogers Date: Sun, 24 Sep 2023 23:23:23 -0700 Subject: [PATCH 06/22] perf pmus: Make PMU alias name loading lazy ANBZ: #24284 commit b1f05622fef39dded385f9e360e859846c1ddaf1 upstream. PMU alias names were computed when the first perf_pmu is created, scanning all PMUs in event sources for a file called alias that generally doesn't exist. Switch to trying to load the file when all PMU related files are loaded in lookup. This would cause a PMU name lookup of an alias name to fail if no PMUs were loaded, so in that case all PMUs are loaded and the find repeated. The overhead is similar but in the (very) general case not all PMUs are scanned for the alias file. As the overhead occurs once per invocation it doesn't show in perf bench internals pmu-scan. On a tigerlake machine, the number of openat system calls for an event of cpu/cycles/ with perf stat reduces from 94 to 69 (ie 25 fewer openat calls). Signed-off-by: Ian Rogers Acked-by: Namhyung Kim Cc: Ravi Bangoria Cc: James Clark Cc: Leo Yan Cc: Kan Liang Link: https://lore.kernel.org/r/20230925062323.840799-1-irogers@google.com Signed-off-by: Namhyung Kim Signed-off-by: Fangyu Yu --- tools/perf/arch/x86/util/pmu.c | 139 --------------------------------- tools/perf/util/pmu.c | 39 ++++----- tools/perf/util/pmu.h | 2 - tools/perf/util/pmus.c | 10 +++ 4 files changed, 31 insertions(+), 159 deletions(-) diff --git a/tools/perf/arch/x86/util/pmu.c b/tools/perf/arch/x86/util/pmu.c index 0af256236466..7f71b86ed07c 100644 --- a/tools/perf/arch/x86/util/pmu.c +++ b/tools/perf/arch/x86/util/pmu.c @@ -17,15 +17,6 @@ #include "../../../util/pmus.h" #include "env.h" -struct pmu_alias { - char *name; - char *alias; - struct list_head list; -}; - -static LIST_HEAD(pmu_alias_name_list); -static bool cached_list; - struct perf_event_attr *perf_pmu__get_default_config(struct perf_pmu *pmu __maybe_unused) { #ifdef HAVE_AUXTRACE_SUPPORT @@ -41,136 +32,6 @@ struct perf_event_attr *perf_pmu__get_default_config(struct perf_pmu *pmu __mayb return NULL; } -static void pmu_alias__delete(struct pmu_alias *pmu_alias) -{ - if (!pmu_alias) - return; - - zfree(&pmu_alias->name); - zfree(&pmu_alias->alias); - free(pmu_alias); -} - -static struct pmu_alias *pmu_alias__new(char *name, char *alias) -{ - struct pmu_alias *pmu_alias = zalloc(sizeof(*pmu_alias)); - - if (pmu_alias) { - pmu_alias->name = strdup(name); - if (!pmu_alias->name) - goto out_delete; - - pmu_alias->alias = strdup(alias); - if (!pmu_alias->alias) - goto out_delete; - } - return pmu_alias; - -out_delete: - pmu_alias__delete(pmu_alias); - return NULL; -} - -static int setup_pmu_alias_list(void) -{ - int fd, dirfd; - DIR *dir; - struct dirent *dent; - struct pmu_alias *pmu_alias; - char buf[MAX_PMU_NAME_LEN]; - FILE *file; - int ret = -ENOMEM; - - dirfd = perf_pmu__event_source_devices_fd(); - if (dirfd < 0) - return -1; - - dir = fdopendir(dirfd); - if (!dir) - return -errno; - - while ((dent = readdir(dir))) { - if (!strcmp(dent->d_name, ".") || - !strcmp(dent->d_name, "..")) - continue; - - fd = perf_pmu__pathname_fd(dirfd, dent->d_name, "alias", O_RDONLY); - if (fd < 0) - continue; - - file = fdopen(fd, "r"); - if (!file) - continue; - - if (!fgets(buf, sizeof(buf), file)) { - fclose(file); - continue; - } - - fclose(file); - - /* Remove the last '\n' */ - buf[strlen(buf) - 1] = 0; - - pmu_alias = pmu_alias__new(dent->d_name, buf); - if (!pmu_alias) - goto close_dir; - - list_add_tail(&pmu_alias->list, &pmu_alias_name_list); - } - - ret = 0; - -close_dir: - closedir(dir); - return ret; -} - -static const char *__pmu_find_real_name(const char *name) -{ - struct pmu_alias *pmu_alias; - - list_for_each_entry(pmu_alias, &pmu_alias_name_list, list) { - if (!strcmp(name, pmu_alias->alias)) - return pmu_alias->name; - } - - return name; -} - -const char *pmu_find_real_name(const char *name) -{ - if (cached_list) - return __pmu_find_real_name(name); - - setup_pmu_alias_list(); - cached_list = true; - - return __pmu_find_real_name(name); -} - -static const char *__pmu_find_alias_name(const char *name) -{ - struct pmu_alias *pmu_alias; - - list_for_each_entry(pmu_alias, &pmu_alias_name_list, list) { - if (!strcmp(name, pmu_alias->name)) - return pmu_alias->alias; - } - return NULL; -} - -const char *pmu_find_alias_name(const char *name) -{ - if (cached_list) - return __pmu_find_alias_name(name); - - setup_pmu_alias_list(); - cached_list = true; - - return __pmu_find_alias_name(name); -} - int perf_pmus__num_mem_pmus(void) { /* AMD uses IBS OP pmu and not a core PMU for perf mem/c2c */ diff --git a/tools/perf/util/pmu.c b/tools/perf/util/pmu.c index 43eac49522ae..db41ddc72145 100644 --- a/tools/perf/util/pmu.c +++ b/tools/perf/util/pmu.c @@ -1006,16 +1006,27 @@ perf_pmu__get_default_config(struct perf_pmu *pmu __maybe_unused) return NULL; } -const char * __weak -pmu_find_real_name(const char *name) +static char *pmu_find_alias_name(struct perf_pmu *pmu, int dirfd) { - return name; -} + FILE *file = perf_pmu__open_file_at(pmu, dirfd, "alias"); + char *line = NULL; + size_t line_len = 0; + ssize_t ret; -const char * __weak -pmu_find_alias_name(const char *name __maybe_unused) -{ - return NULL; + if (!file) + return NULL; + + ret = getline(&line, &line_len, file); + if (ret < 0) { + fclose(file); + return NULL; + } + /* Remove trailing newline. */ + if (ret > 0 && line[ret - 1] == '\n') + line[--ret] = '\0'; + + fclose(file); + return line; } static int pmu_max_precise(int dirfd, struct perf_pmu *pmu) @@ -1026,12 +1037,10 @@ static int pmu_max_precise(int dirfd, struct perf_pmu *pmu) return max_precise; } -struct perf_pmu *perf_pmu__lookup(struct list_head *pmus, int dirfd, const char *lookup_name) +struct perf_pmu *perf_pmu__lookup(struct list_head *pmus, int dirfd, const char *name) { struct perf_pmu *pmu; __u32 type; - const char *name = pmu_find_real_name(lookup_name); - const char *alias_name; pmu = zalloc(sizeof(*pmu)); if (!pmu) @@ -1063,18 +1072,12 @@ struct perf_pmu *perf_pmu__lookup(struct list_head *pmus, int dirfd, const char pmu->is_core = is_pmu_core(name); pmu->cpus = pmu_cpumask(dirfd, name, pmu->is_core); - alias_name = pmu_find_alias_name(name); - if (alias_name) { - pmu->alias_name = strdup(alias_name); - if (!pmu->alias_name) - goto err; - } - pmu->type = type; pmu->is_uncore = pmu_is_uncore(dirfd, name); if (pmu->is_uncore) pmu->id = pmu_id(name); pmu->max_precise = pmu_max_precise(dirfd, pmu); + pmu->alias_name = pmu_find_alias_name(pmu, dirfd); pmu->events_table = perf_pmu__find_events_table(pmu); /* * Load the sys json events/aliases when loading the PMU as each event diff --git a/tools/perf/util/pmu.h b/tools/perf/util/pmu.h index e192010ad3c7..bfa95513e59d 100644 --- a/tools/perf/util/pmu.h +++ b/tools/perf/util/pmu.h @@ -260,8 +260,6 @@ void perf_pmu__warn_invalid_formats(struct perf_pmu *pmu); int perf_pmu__match(const char *pattern, const char *name, const char *tok); -const char *pmu_find_real_name(const char *name); -const char *pmu_find_alias_name(const char *name); double perf_pmu__cpu_slots_per_cycle(void); int perf_pmu__event_source_devices_scnprintf(char *pathname, size_t size); int perf_pmu__pathname_scnprintf(char *buf, size_t size, diff --git a/tools/perf/util/pmus.c b/tools/perf/util/pmus.c index dda5ba9c73fd..d4fed9456971 100644 --- a/tools/perf/util/pmus.c +++ b/tools/perf/util/pmus.c @@ -37,6 +37,8 @@ static LIST_HEAD(other_pmus); static bool read_sysfs_core_pmus; static bool read_sysfs_all_pmus; +static void pmu_read_sysfs(bool core_only); + int pmu_name_len_no_suffix(const char *str, unsigned long *num) { int orig_len, len; @@ -124,6 +126,14 @@ struct perf_pmu *perf_pmus__find(const char *name) pmu = perf_pmu__lookup(core_pmu ? &core_pmus : &other_pmus, dirfd, name); close(dirfd); + if (!pmu) { + /* + * Looking up an inidividual PMU failed. This may mean name is + * an alias, so read the PMUs from sysfs and try to find again. + */ + pmu_read_sysfs(core_pmu); + pmu = pmu_find(name); + } return pmu; } -- Gitee From 37dd86b623ae6ed7fd758354c24096598a87074e Mon Sep 17 00:00:00 2001 From: Ian Rogers Date: Thu, 12 Oct 2023 10:56:39 -0700 Subject: [PATCH 07/22] perf pmu: Rename perf_pmu__get_default_config to perf_pmu__arch_init ANBZ: #24284 commit aa61360155ac69003dc5c177f03a7f177a435286 upstream. Assign default_config as part of the init. perf_pmu__get_default_config was doing more than just getting the default config and so this is intended to better align with the code. Signed-off-by: Ian Rogers Cc: Ravi Bangoria Cc: James Clark Cc: Suzuki K Poulose Cc: Yang Jihong Cc: Will Deacon Cc: Leo Yan Cc: Mike Leach Cc: Jing Zhang Cc: Kajol Jain Cc: Thomas Richter Cc: Alexander Shishkin Cc: Kan Liang Cc: John Garry Cc: linux-arm-kernel@lists.infradead.org Cc: coresight@lists.linaro.org Link: https://lore.kernel.org/r/20231012175645.1849503-2-irogers@google.com Signed-off-by: Namhyung Kim Signed-off-by: Fangyu Yu --- tools/perf/arch/arm/util/pmu.c | 8 +++----- tools/perf/arch/s390/util/pmu.c | 3 +-- tools/perf/arch/x86/util/pmu.c | 5 ++--- tools/perf/util/pmu.c | 13 ++++++------- tools/perf/util/pmu.h | 2 +- 5 files changed, 13 insertions(+), 18 deletions(-) diff --git a/tools/perf/arch/arm/util/pmu.c b/tools/perf/arch/arm/util/pmu.c index a9623b128ece..d55d2b15f2e6 100644 --- a/tools/perf/arch/arm/util/pmu.c +++ b/tools/perf/arch/arm/util/pmu.c @@ -14,22 +14,20 @@ #include "../../../util/pmu.h" #include "../../../util/cs-etm.h" -struct perf_event_attr -*perf_pmu__get_default_config(struct perf_pmu *pmu __maybe_unused) +void perf_pmu__arch_init(struct perf_pmu *pmu __maybe_unused) { #ifdef HAVE_AUXTRACE_SUPPORT if (!strcmp(pmu->name, CORESIGHT_ETM_PMU_NAME)) { /* add ETM default config here */ pmu->selectable = true; - return cs_etm_get_default_config(pmu); + pmu->default_config = cs_etm_get_default_config(pmu); #if defined(__aarch64__) } else if (strstarts(pmu->name, ARM_SPE_PMU_NAME)) { - return arm_spe_pmu_default_config(pmu); + pmu->default_config = arm_spe_pmu_default_config(pmu); } else if (strstarts(pmu->name, HISI_PTT_PMU_NAME)) { pmu->selectable = true; #endif } #endif - return NULL; } diff --git a/tools/perf/arch/s390/util/pmu.c b/tools/perf/arch/s390/util/pmu.c index 11f03f32e3fd..886c30e001fa 100644 --- a/tools/perf/arch/s390/util/pmu.c +++ b/tools/perf/arch/s390/util/pmu.c @@ -13,11 +13,10 @@ #define S390_PMUPAI_EXT "pai_ext" #define S390_PMUCPUM_CF "cpum_cf" -struct perf_event_attr *perf_pmu__get_default_config(struct perf_pmu *pmu) +void perf_pmu__arch_init(struct perf_pmu *pmu) { if (!strcmp(pmu->name, S390_PMUPAI_CRYPTO) || !strcmp(pmu->name, S390_PMUPAI_EXT) || !strcmp(pmu->name, S390_PMUCPUM_CF)) pmu->selectable = true; - return NULL; } diff --git a/tools/perf/arch/x86/util/pmu.c b/tools/perf/arch/x86/util/pmu.c index 7f71b86ed07c..97f6aecb62cb 100644 --- a/tools/perf/arch/x86/util/pmu.c +++ b/tools/perf/arch/x86/util/pmu.c @@ -17,19 +17,18 @@ #include "../../../util/pmus.h" #include "env.h" -struct perf_event_attr *perf_pmu__get_default_config(struct perf_pmu *pmu __maybe_unused) +void perf_pmu__arch_init(struct perf_pmu *pmu __maybe_unused) { #ifdef HAVE_AUXTRACE_SUPPORT if (!strcmp(pmu->name, INTEL_PT_PMU_NAME)) { pmu->auxtrace = true; - return intel_pt_pmu_default_config(pmu); + pmu->default_config = intel_pt_pmu_default_config(pmu); } if (!strcmp(pmu->name, INTEL_BTS_PMU_NAME)) { pmu->auxtrace = true; pmu->selectable = true; } #endif - return NULL; } int perf_pmus__num_mem_pmus(void) diff --git a/tools/perf/util/pmu.c b/tools/perf/util/pmu.c index db41ddc72145..f9c11e243759 100644 --- a/tools/perf/util/pmu.c +++ b/tools/perf/util/pmu.c @@ -1000,12 +1000,6 @@ void pmu_add_sys_aliases(struct perf_pmu *pmu) pmu_for_each_sys_event(pmu_add_sys_aliases_iter_fn, pmu); } -struct perf_event_attr * __weak -perf_pmu__get_default_config(struct perf_pmu *pmu __maybe_unused) -{ - return NULL; -} - static char *pmu_find_alias_name(struct perf_pmu *pmu, int dirfd) { FILE *file = perf_pmu__open_file_at(pmu, dirfd, "alias"); @@ -1037,6 +1031,11 @@ static int pmu_max_precise(int dirfd, struct perf_pmu *pmu) return max_precise; } +void __weak +perf_pmu__arch_init(struct perf_pmu *pmu __maybe_unused) +{ +} + struct perf_pmu *perf_pmu__lookup(struct list_head *pmus, int dirfd, const char *name) { struct perf_pmu *pmu; @@ -1088,7 +1087,7 @@ struct perf_pmu *perf_pmu__lookup(struct list_head *pmus, int dirfd, const char pmu_add_sys_aliases(pmu); list_add_tail(&pmu->list, pmus); - pmu->default_config = perf_pmu__get_default_config(pmu); + perf_pmu__arch_init(pmu); return pmu; err: diff --git a/tools/perf/util/pmu.h b/tools/perf/util/pmu.h index bfa95513e59d..4be83dca20e1 100644 --- a/tools/perf/util/pmu.h +++ b/tools/perf/util/pmu.h @@ -240,7 +240,7 @@ bool perf_pmu__file_exists(struct perf_pmu *pmu, const char *name); int perf_pmu__test(void); -struct perf_event_attr *perf_pmu__get_default_config(struct perf_pmu *pmu); +void perf_pmu__arch_init(struct perf_pmu *pmu); void pmu_add_cpu_aliases_table(struct perf_pmu *pmu, const struct pmu_events_table *table); -- Gitee From 204739fdf17993b0fdafc3730eea46867706468e Mon Sep 17 00:00:00 2001 From: Ian Rogers Date: Thu, 12 Oct 2023 10:56:40 -0700 Subject: [PATCH 08/22] perf intel-pt: Move PMU initialization from default config code ANBZ: #24284 commit 461e3e636a26518d3dd9c0ae3aff32894ec00b26 upstream. Avoid setting PMU values in intel_pt_pmu_default_config, move to perf_pmu__arch_init. Signed-off-by: Ian Rogers Reviewed-by: Adrian Hunter Cc: Ravi Bangoria Cc: James Clark Cc: Suzuki K Poulose Cc: Yang Jihong Cc: Will Deacon Cc: Leo Yan Cc: Mike Leach Cc: Jing Zhang Cc: Kajol Jain Cc: Thomas Richter Cc: Kan Liang Cc: John Garry Cc: linux-arm-kernel@lists.infradead.org Cc: coresight@lists.linaro.org Link: https://lore.kernel.org/r/20231012175645.1849503-3-irogers@google.com Signed-off-by: Namhyung Kim Signed-off-by: Fangyu Yu --- tools/perf/arch/x86/util/intel-pt.c | 2 -- tools/perf/arch/x86/util/pmu.c | 1 + 2 files changed, 1 insertion(+), 2 deletions(-) diff --git a/tools/perf/arch/x86/util/intel-pt.c b/tools/perf/arch/x86/util/intel-pt.c index 7b47bbc7afc7..0583dadcc702 100644 --- a/tools/perf/arch/x86/util/intel-pt.c +++ b/tools/perf/arch/x86/util/intel-pt.c @@ -268,8 +268,6 @@ intel_pt_pmu_default_config(struct perf_pmu *intel_pt_pmu) attr->config = intel_pt_default_config(intel_pt_pmu); - intel_pt_pmu->selectable = true; - return attr; } diff --git a/tools/perf/arch/x86/util/pmu.c b/tools/perf/arch/x86/util/pmu.c index 97f6aecb62cb..6bb77acadb19 100644 --- a/tools/perf/arch/x86/util/pmu.c +++ b/tools/perf/arch/x86/util/pmu.c @@ -22,6 +22,7 @@ void perf_pmu__arch_init(struct perf_pmu *pmu __maybe_unused) #ifdef HAVE_AUXTRACE_SUPPORT if (!strcmp(pmu->name, INTEL_PT_PMU_NAME)) { pmu->auxtrace = true; + pmu->selectable = true; pmu->default_config = intel_pt_pmu_default_config(pmu); } if (!strcmp(pmu->name, INTEL_BTS_PMU_NAME)) { -- Gitee From b9d1df4b6b318f82fbd898ad8b1416b2f18be27d Mon Sep 17 00:00:00 2001 From: Ian Rogers Date: Thu, 12 Oct 2023 10:56:41 -0700 Subject: [PATCH 09/22] perf arm-spe: Move PMU initialization from default config code ANBZ: #24284 commit 672bd21390d38dfccbe52b9537555985d65446aa upstream. Avoid setting PMU values in arm_spe_pmu_default_config, move to perf_pmu__arch_init. Signed-off-by: Ian Rogers Reviewed-by: Adrian Hunter Tested-by: Leo Yan Cc: Ravi Bangoria Cc: James Clark Cc: Suzuki K Poulose Cc: Yang Jihong Cc: Will Deacon Cc: Mike Leach Cc: Jing Zhang Cc: Kajol Jain Cc: Thomas Richter Cc: Alexander Shishkin Cc: Kan Liang Cc: John Garry Cc: linux-arm-kernel@lists.infradead.org Cc: coresight@lists.linaro.org Link: https://lore.kernel.org/r/20231012175645.1849503-4-irogers@google.com Signed-off-by: Namhyung Kim Signed-off-by: Fangyu Yu --- tools/perf/arch/arm/util/pmu.c | 2 ++ tools/perf/arch/arm64/util/arm-spe.c | 3 --- 2 files changed, 2 insertions(+), 3 deletions(-) diff --git a/tools/perf/arch/arm/util/pmu.c b/tools/perf/arch/arm/util/pmu.c index d55d2b15f2e6..f25f68f84a94 100644 --- a/tools/perf/arch/arm/util/pmu.c +++ b/tools/perf/arch/arm/util/pmu.c @@ -23,6 +23,8 @@ void perf_pmu__arch_init(struct perf_pmu *pmu __maybe_unused) pmu->default_config = cs_etm_get_default_config(pmu); #if defined(__aarch64__) } else if (strstarts(pmu->name, ARM_SPE_PMU_NAME)) { + pmu->selectable = true; + pmu->is_uncore = false; pmu->default_config = arm_spe_pmu_default_config(pmu); } else if (strstarts(pmu->name, HISI_PTT_PMU_NAME)) { pmu->selectable = true; diff --git a/tools/perf/arch/arm64/util/arm-spe.c b/tools/perf/arch/arm64/util/arm-spe.c index 4e947814ea74..dbaf16e7fc2c 100644 --- a/tools/perf/arch/arm64/util/arm-spe.c +++ b/tools/perf/arch/arm64/util/arm-spe.c @@ -516,8 +516,5 @@ struct perf_event_attr attr->sample_period = 4096; } - arm_spe_pmu->selectable = true; - arm_spe_pmu->is_uncore = false; - return attr; } -- Gitee From 4912deabf4aa735aca01f4bda1c453dc6d618e38 Mon Sep 17 00:00:00 2001 From: Ian Rogers Date: Thu, 12 Oct 2023 10:56:42 -0700 Subject: [PATCH 10/22] perf pmu: Const-ify file APIs ANBZ: #24284 commit 3a42f4c796ce45a6b1d14975401deb2c457ae79d upstream. File APIs don't alter the struct pmu so allow const ones to be passed. Signed-off-by: Ian Rogers Reviewed-by: Adrian Hunter Cc: Ravi Bangoria Cc: James Clark Cc: Suzuki K Poulose Cc: Yang Jihong Cc: Will Deacon Cc: Leo Yan Cc: Mike Leach Cc: Jing Zhang Cc: Kajol Jain Cc: Thomas Richter Cc: Kan Liang Cc: John Garry Cc: linux-arm-kernel@lists.infradead.org Cc: coresight@lists.linaro.org Link: https://lore.kernel.org/r/20231012175645.1849503-5-irogers@google.com Signed-off-by: Namhyung Kim Signed-off-by: Fangyu Yu --- tools/perf/util/pmu.c | 12 ++++++------ tools/perf/util/pmu.h | 11 ++++++----- tools/perf/util/python.c | 2 +- 3 files changed, 13 insertions(+), 12 deletions(-) diff --git a/tools/perf/util/pmu.c b/tools/perf/util/pmu.c index f9c11e243759..bc18f196aa97 100644 --- a/tools/perf/util/pmu.c +++ b/tools/perf/util/pmu.c @@ -613,7 +613,7 @@ static int perf_pmu__new_alias(struct perf_pmu *pmu, const char *name, return 0; } -static inline bool pmu_alias_info_file(char *name) +static inline bool pmu_alias_info_file(const char *name) { size_t len; @@ -1833,7 +1833,7 @@ bool perf_pmu__is_software(const struct perf_pmu *pmu) return false; } -FILE *perf_pmu__open_file(struct perf_pmu *pmu, const char *name) +FILE *perf_pmu__open_file(const struct perf_pmu *pmu, const char *name) { char path[PATH_MAX]; @@ -1844,7 +1844,7 @@ FILE *perf_pmu__open_file(struct perf_pmu *pmu, const char *name) return fopen(path, "r"); } -FILE *perf_pmu__open_file_at(struct perf_pmu *pmu, int dirfd, const char *name) +FILE *perf_pmu__open_file_at(const struct perf_pmu *pmu, int dirfd, const char *name) { int fd; @@ -1855,7 +1855,7 @@ FILE *perf_pmu__open_file_at(struct perf_pmu *pmu, int dirfd, const char *name) return fdopen(fd, "r"); } -int perf_pmu__scan_file(struct perf_pmu *pmu, const char *name, const char *fmt, +int perf_pmu__scan_file(const struct perf_pmu *pmu, const char *name, const char *fmt, ...) { va_list args; @@ -1872,7 +1872,7 @@ int perf_pmu__scan_file(struct perf_pmu *pmu, const char *name, const char *fmt, return ret; } -int perf_pmu__scan_file_at(struct perf_pmu *pmu, int dirfd, const char *name, +int perf_pmu__scan_file_at(const struct perf_pmu *pmu, int dirfd, const char *name, const char *fmt, ...) { va_list args; @@ -1889,7 +1889,7 @@ int perf_pmu__scan_file_at(struct perf_pmu *pmu, int dirfd, const char *name, return ret; } -bool perf_pmu__file_exists(struct perf_pmu *pmu, const char *name) +bool perf_pmu__file_exists(const struct perf_pmu *pmu, const char *name) { char path[PATH_MAX]; diff --git a/tools/perf/util/pmu.h b/tools/perf/util/pmu.h index 4be83dca20e1..8e999a18c32e 100644 --- a/tools/perf/util/pmu.h +++ b/tools/perf/util/pmu.h @@ -229,14 +229,15 @@ bool pmu__name_match(const struct perf_pmu *pmu, const char *pmu_name); */ bool perf_pmu__is_software(const struct perf_pmu *pmu); -FILE *perf_pmu__open_file(struct perf_pmu *pmu, const char *name); -FILE *perf_pmu__open_file_at(struct perf_pmu *pmu, int dirfd, const char *name); +FILE *perf_pmu__open_file(const struct perf_pmu *pmu, const char *name); +FILE *perf_pmu__open_file_at(const struct perf_pmu *pmu, int dirfd, const char *name); -int perf_pmu__scan_file(struct perf_pmu *pmu, const char *name, const char *fmt, ...) __scanf(3, 4); -int perf_pmu__scan_file_at(struct perf_pmu *pmu, int dirfd, const char *name, +int perf_pmu__scan_file(const struct perf_pmu *pmu, const char *name, const char *fmt, ...) + __scanf(3, 4); +int perf_pmu__scan_file_at(const struct perf_pmu *pmu, int dirfd, const char *name, const char *fmt, ...) __scanf(4, 5); -bool perf_pmu__file_exists(struct perf_pmu *pmu, const char *name); +bool perf_pmu__file_exists(const struct perf_pmu *pmu, const char *name); int perf_pmu__test(void); diff --git a/tools/perf/util/python.c b/tools/perf/util/python.c index 06a1e09d7349..102e15510c5c 100644 --- a/tools/perf/util/python.c +++ b/tools/perf/util/python.c @@ -98,7 +98,7 @@ struct perf_pmu *evsel__find_pmu(const struct evsel *evsel __maybe_unused) return NULL; } -int perf_pmu__scan_file(struct perf_pmu *pmu, const char *name, const char *fmt, ...) +int perf_pmu__scan_file(const struct perf_pmu *pmu, const char *name, const char *fmt, ...) { return EOF; } -- Gitee From 36ee7fa41e7fe96e91cc8328f6c6acd4b977f354 Mon Sep 17 00:00:00 2001 From: Ian Rogers Date: Thu, 12 Oct 2023 10:56:43 -0700 Subject: [PATCH 11/22] perf pmu: Const-ify perf_pmu__config_terms ANBZ: #24284 commit 63883cb063846e9d0574038bb702a3a62a1ec046 upstream. Add const to related APIs, this is so they can be used to default initialize a perf_event_attr from a const pmu. Signed-off-by: Ian Rogers Reviewed-by: Adrian Hunter Cc: Ravi Bangoria Cc: James Clark Cc: Suzuki K Poulose Cc: Yang Jihong Cc: Will Deacon Cc: Leo Yan Cc: Mike Leach Cc: Jing Zhang Cc: Kajol Jain Cc: Thomas Richter Cc: Kan Liang Cc: John Garry Cc: linux-arm-kernel@lists.infradead.org Cc: coresight@lists.linaro.org Link: https://lore.kernel.org/r/20231012175645.1849503-6-irogers@google.com Signed-off-by: Namhyung Kim Signed-off-by: Fangyu Yu --- tools/perf/util/pmu.c | 10 +++++----- tools/perf/util/pmu.h | 2 +- 2 files changed, 6 insertions(+), 6 deletions(-) diff --git a/tools/perf/util/pmu.c b/tools/perf/util/pmu.c index bc18f196aa97..faf434119a75 100644 --- a/tools/perf/util/pmu.c +++ b/tools/perf/util/pmu.c @@ -168,7 +168,7 @@ static void __perf_pmu_format__load(struct perf_pmu_format *format, FILE *file) format->loaded = true; } -static void perf_pmu_format__load(struct perf_pmu *pmu, struct perf_pmu_format *format) +static void perf_pmu_format__load(const struct perf_pmu *pmu, struct perf_pmu_format *format) { char path[PATH_MAX]; FILE *file = NULL; @@ -1182,7 +1182,7 @@ void evsel__set_config_if_unset(struct perf_pmu *pmu, struct evsel *evsel, } static struct perf_pmu_format * -pmu_find_format(struct list_head *formats, const char *name) +pmu_find_format(const struct list_head *formats, const char *name) { struct perf_pmu_format *format; @@ -1280,7 +1280,7 @@ static int pmu_resolve_param_term(struct parse_events_term *term, return -1; } -static char *pmu_formats_string(struct list_head *formats) +static char *pmu_formats_string(const struct list_head *formats) { struct perf_pmu_format *format; char *str = NULL; @@ -1306,7 +1306,7 @@ static char *pmu_formats_string(struct list_head *formats) * Setup one of config[12] attr members based on the * user input data - term parameter. */ -static int pmu_config_term(struct perf_pmu *pmu, +static int pmu_config_term(const struct perf_pmu *pmu, struct perf_event_attr *attr, struct parse_events_term *term, struct parse_events_terms *head_terms, @@ -1429,7 +1429,7 @@ static int pmu_config_term(struct perf_pmu *pmu, return 0; } -int perf_pmu__config_terms(struct perf_pmu *pmu, +int perf_pmu__config_terms(const struct perf_pmu *pmu, struct perf_event_attr *attr, struct parse_events_terms *terms, bool zero, struct parse_events_error *err) diff --git a/tools/perf/util/pmu.h b/tools/perf/util/pmu.h index 8e999a18c32e..de6c3674ac60 100644 --- a/tools/perf/util/pmu.h +++ b/tools/perf/util/pmu.h @@ -200,7 +200,7 @@ void pmu_add_sys_aliases(struct perf_pmu *pmu); int perf_pmu__config(struct perf_pmu *pmu, struct perf_event_attr *attr, struct parse_events_terms *head_terms, struct parse_events_error *error); -int perf_pmu__config_terms(struct perf_pmu *pmu, +int perf_pmu__config_terms(const struct perf_pmu *pmu, struct perf_event_attr *attr, struct parse_events_terms *terms, bool zero, struct parse_events_error *error); -- Gitee From 55cd4ccf3c1c1255e478618279c6f8754ea596a1 Mon Sep 17 00:00:00 2001 From: Ian Rogers Date: Thu, 12 Oct 2023 10:56:44 -0700 Subject: [PATCH 12/22] perf pmu-events: Remember the perf_events_map for a PMU ANBZ: #24284 commit f20c15d13f017d4b46523990a0f3ba42025391a6 upstream. strcmp_cpuid_str performs regular expression comparisons and so per CPUID linear searches over the perf_events_map are expensive. Add a helper function called map_for_pmu that does the search but also caches the map specific to a PMU. As the PMU may differ, also cache the CPUID string so that PMUs with the same CPUID string don't require the linear search and regular expression comparisons. This speeds loading PMUs as the search is done once per PMU to find the appropriate tables. Signed-off-by: Ian Rogers Tested-by: Yang Jihong Cc: Ravi Bangoria Cc: James Clark Cc: Suzuki K Poulose Cc: Will Deacon Cc: Leo Yan Cc: Mike Leach Cc: Jing Zhang Cc: Kajol Jain Cc: Thomas Richter Cc: Kan Liang Cc: John Garry Cc: linux-arm-kernel@lists.infradead.org Cc: coresight@lists.linaro.org Link: https://lore.kernel.org/r/20231012175645.1849503-7-irogers@google.com Signed-off-by: Namhyung Kim Signed-off-by: Fangyu Yu --- tools/perf/pmu-events/jevents.py | 109 ++++++++++++++++++++----------- 1 file changed, 70 insertions(+), 39 deletions(-) diff --git a/tools/perf/pmu-events/jevents.py b/tools/perf/pmu-events/jevents.py index ae2bd49e8805..7bddad65f1da 100755 --- a/tools/perf/pmu-events/jevents.py +++ b/tools/perf/pmu-events/jevents.py @@ -976,68 +976,99 @@ int pmu_metrics_table__for_each_metric(const struct pmu_metrics_table *table, return 0; } -const struct pmu_events_table *perf_pmu__find_events_table(struct perf_pmu *pmu) +static const struct pmu_events_map *map_for_pmu(struct perf_pmu *pmu) { - const struct pmu_events_table *table = NULL; - char *cpuid = perf_pmu__getcpuid(pmu); + static struct { + const struct pmu_events_map *map; + struct perf_pmu *pmu; + } last_result; + static struct { + const struct pmu_events_map *map; + char *cpuid; + } last_map_search; + static bool has_last_result, has_last_map_search; + const struct pmu_events_map *map = NULL; + char *cpuid = NULL; size_t i; - /* on some platforms which uses cpus map, cpuid can be NULL for + if (has_last_result && last_result.pmu == pmu) + return last_result.map; + + cpuid = perf_pmu__getcpuid(pmu); + + /* + * On some platforms which uses cpus map, cpuid can be NULL for * PMUs other than CORE PMUs. */ if (!cpuid) - return NULL; + goto out_update_last_result; + + if (has_last_map_search && !strcmp(last_map_search.cpuid, cpuid)) { + map = last_map_search.map; + free(cpuid); + } else { + i = 0; + for (;;) { + map = &pmu_events_map[i++]; + + if (!map->arch) { + map = NULL; + break; + } + + if (!strcmp_cpuid_str(map->cpuid, cpuid)) + break; + } + free(last_map_search.cpuid); + last_map_search.cpuid = cpuid; + last_map_search.map = map; + has_last_map_search = true; + } +out_update_last_result: + last_result.pmu = pmu; + last_result.map = map; + has_last_result = true; + return map; +} - i = 0; - for (;;) { - const struct pmu_events_map *map = &pmu_events_map[i++]; - if (!map->arch) - break; +const struct pmu_events_table *perf_pmu__find_events_table(struct perf_pmu *pmu) +{ + const struct pmu_events_map *map = map_for_pmu(pmu); - if (!strcmp_cpuid_str(map->cpuid, cpuid)) { - table = &map->event_table; - break; - } - } - free(cpuid); - if (!pmu || !table) - return table; + if (!map) + return NULL; - for (i = 0; i < table->num_pmus; i++) { - const struct pmu_table_entry *table_pmu = &table->pmus[i]; + if (!pmu) + return &map->event_table; + + for (size_t i = 0; i < map->event_table.num_pmus; i++) { + const struct pmu_table_entry *table_pmu = &map->event_table.pmus[i]; const char *pmu_name = &big_c_string[table_pmu->pmu_name.offset]; if (pmu__name_match(pmu, pmu_name)) - return table; + return &map->event_table; } return NULL; } const struct pmu_metrics_table *perf_pmu__find_metrics_table(struct perf_pmu *pmu) { - const struct pmu_metrics_table *table = NULL; - char *cpuid = perf_pmu__getcpuid(pmu); - int i; + const struct pmu_events_map *map = map_for_pmu(pmu); - /* on some platforms which uses cpus map, cpuid can be NULL for - * PMUs other than CORE PMUs. - */ - if (!cpuid) + if (!map) return NULL; - i = 0; - for (;;) { - const struct pmu_events_map *map = &pmu_events_map[i++]; - if (!map->arch) - break; + if (!pmu) + return &map->metric_table; - if (!strcmp_cpuid_str(map->cpuid, cpuid)) { - table = &map->metric_table; - break; - } + for (size_t i = 0; i < map->metric_table.num_pmus; i++) { + const struct pmu_table_entry *table_pmu = &map->metric_table.pmus[i]; + const char *pmu_name = &big_c_string[table_pmu->pmu_name.offset]; + + if (pmu__name_match(pmu, pmu_name)) + return &map->metric_table; } - free(cpuid); - return table; + return NULL; } const struct pmu_events_table *find_core_events_table(const char *arch, const char *cpuid) -- Gitee From cb52fc75922c1d9e6d8de6fcf102df56e5e8426c Mon Sep 17 00:00:00 2001 From: Ian Rogers Date: Thu, 12 Oct 2023 10:56:45 -0700 Subject: [PATCH 13/22] perf pmu: Lazily compute default config ANBZ: #24284 commit 0197da7affab502cd6e25da616ad038b169a7a77 upstream. The default config is computed during creation of the PMU and may do things like scanning sysfs, when the PMU may just be used as part of scanning. Change default_config to perf_event_attr_init_default, a callback that is used when a default config needs initializing. This avoids holding onto the memory for a perf_event_attr and copying. On a tigerlake laptop running the pmu-scan benchmark: Before: Running 'internals/pmu-scan' benchmark: Computing performance of sysfs PMU event scan for 100 times Average core PMU scanning took: 28.780 usec (+- 0.503 usec) Average PMU scanning took: 283.480 usec (+- 18.471 usec) Number of openat syscalls: 30,227 After: Running 'internals/pmu-scan' benchmark: Computing performance of sysfs PMU event scan for 100 times Average core PMU scanning took: 27.880 usec (+- 0.169 usec) Average PMU scanning took: 245.260 usec (+- 15.758 usec) Number of openat syscalls: 28,914 Over 3 runs it is a nearly 12% reduction in execution time and a 4.3% of openat calls. Signed-off-by: Ian Rogers Reviewed-by: Adrian Hunter Cc: Ravi Bangoria Cc: James Clark Cc: Suzuki K Poulose Cc: Yang Jihong Cc: Will Deacon Cc: Leo Yan Cc: Mike Leach Cc: Jing Zhang Cc: Kajol Jain Cc: Thomas Richter Cc: Kan Liang Cc: John Garry Cc: linux-arm-kernel@lists.infradead.org Cc: coresight@lists.linaro.org Link: https://lore.kernel.org/r/20231012175645.1849503-8-irogers@google.com Signed-off-by: Namhyung Kim Signed-off-by: Fangyu Yu --- tools/perf/arch/arm/util/cs-etm.c | 13 ++------ tools/perf/arch/arm/util/pmu.c | 4 +-- tools/perf/arch/arm64/util/arm-spe.c | 45 ++++++++++++++-------------- tools/perf/arch/x86/util/intel-pt.c | 25 ++++++++-------- tools/perf/arch/x86/util/pmu.c | 2 +- tools/perf/util/arm-spe.h | 4 ++- tools/perf/util/cs-etm.h | 2 +- tools/perf/util/intel-pt.h | 3 +- tools/perf/util/parse-events.c | 12 ++++---- tools/perf/util/pmu.c | 3 +- tools/perf/util/pmu.h | 7 +++-- 11 files changed, 58 insertions(+), 62 deletions(-) diff --git a/tools/perf/arch/arm/util/cs-etm.c b/tools/perf/arch/arm/util/cs-etm.c index e37b10f3e280..14b8afabce3a 100644 --- a/tools/perf/arch/arm/util/cs-etm.c +++ b/tools/perf/arch/arm/util/cs-etm.c @@ -909,16 +909,9 @@ struct auxtrace_record *cs_etm_record_init(int *err) * (CFG_CHG and evsel__set_config_if_unset()). If no default is set then user * changes aren't tracked. */ -struct perf_event_attr * -cs_etm_get_default_config(struct perf_pmu *pmu __maybe_unused) +void +cs_etm_get_default_config(const struct perf_pmu *pmu __maybe_unused, + struct perf_event_attr *attr) { - struct perf_event_attr *attr; - - attr = zalloc(sizeof(struct perf_event_attr)); - if (!attr) - return NULL; - attr->sample_period = 1; - - return attr; } diff --git a/tools/perf/arch/arm/util/pmu.c b/tools/perf/arch/arm/util/pmu.c index f25f68f84a94..7f3af3b97f3b 100644 --- a/tools/perf/arch/arm/util/pmu.c +++ b/tools/perf/arch/arm/util/pmu.c @@ -20,12 +20,12 @@ void perf_pmu__arch_init(struct perf_pmu *pmu __maybe_unused) if (!strcmp(pmu->name, CORESIGHT_ETM_PMU_NAME)) { /* add ETM default config here */ pmu->selectable = true; - pmu->default_config = cs_etm_get_default_config(pmu); + pmu->perf_event_attr_init_default = cs_etm_get_default_config; #if defined(__aarch64__) } else if (strstarts(pmu->name, ARM_SPE_PMU_NAME)) { pmu->selectable = true; pmu->is_uncore = false; - pmu->default_config = arm_spe_pmu_default_config(pmu); + pmu->perf_event_attr_init_default = arm_spe_pmu_default_config; } else if (strstarts(pmu->name, HISI_PTT_PMU_NAME)) { pmu->selectable = true; #endif diff --git a/tools/perf/arch/arm64/util/arm-spe.c b/tools/perf/arch/arm64/util/arm-spe.c index dbaf16e7fc2c..0b52e67edb3b 100644 --- a/tools/perf/arch/arm64/util/arm-spe.c +++ b/tools/perf/arch/arm64/util/arm-spe.c @@ -113,6 +113,25 @@ arm_spe_snapshot_resolve_auxtrace_defaults(struct record_opts *opts, } } +static __u64 arm_spe_pmu__sample_period(const struct perf_pmu *arm_spe_pmu) +{ + static __u64 sample_period; + + if (sample_period) + return sample_period; + + /* + * If kernel driver doesn't advertise a minimum, + * use max allowable by PMSIDR_EL1.INTERVAL + */ + if (perf_pmu__scan_file(arm_spe_pmu, "caps/min_interval", "%llu", + &sample_period) != 1) { + pr_debug("arm_spe driver doesn't advertise a min. interval. Using 4096\n"); + sample_period = 4096; + } + return sample_period; +} + static int arm_spe_recording_options(struct auxtrace_record *itr, struct evlist *evlist, struct record_opts *opts) @@ -136,7 +155,7 @@ static int arm_spe_recording_options(struct auxtrace_record *itr, return -EINVAL; } evsel->core.attr.freq = 0; - evsel->core.attr.sample_period = arm_spe_pmu->default_config->sample_period; + evsel->core.attr.sample_period = arm_spe_pmu__sample_period(arm_spe_pmu); evsel->needs_auxtrace_mmap = true; arm_spe_evsel = evsel; opts->full_auxtrace = true; @@ -495,26 +514,8 @@ struct auxtrace_record *arm_spe_recording_init(int *err, return &sper->itr; } -struct perf_event_attr -*arm_spe_pmu_default_config(struct perf_pmu *arm_spe_pmu) +void +arm_spe_pmu_default_config(const struct perf_pmu *arm_spe_pmu, struct perf_event_attr *attr) { - struct perf_event_attr *attr; - - attr = zalloc(sizeof(struct perf_event_attr)); - if (!attr) { - pr_err("arm_spe default config cannot allocate a perf_event_attr\n"); - return NULL; - } - - /* - * If kernel driver doesn't advertise a minimum, - * use max allowable by PMSIDR_EL1.INTERVAL - */ - if (perf_pmu__scan_file(arm_spe_pmu, "caps/min_interval", "%llu", - &attr->sample_period) != 1) { - pr_debug("arm_spe driver doesn't advertise a min. interval. Using 4096\n"); - attr->sample_period = 4096; - } - - return attr; + attr->sample_period = arm_spe_pmu__sample_period(arm_spe_pmu); } diff --git a/tools/perf/arch/x86/util/intel-pt.c b/tools/perf/arch/x86/util/intel-pt.c index 0583dadcc702..4b710e875953 100644 --- a/tools/perf/arch/x86/util/intel-pt.c +++ b/tools/perf/arch/x86/util/intel-pt.c @@ -61,7 +61,7 @@ struct intel_pt_recording { size_t priv_size; }; -static int intel_pt_parse_terms_with_default(struct perf_pmu *pmu, +static int intel_pt_parse_terms_with_default(const struct perf_pmu *pmu, const char *str, u64 *config) { @@ -85,7 +85,7 @@ static int intel_pt_parse_terms_with_default(struct perf_pmu *pmu, return err; } -static int intel_pt_parse_terms(struct perf_pmu *pmu, const char *str, u64 *config) +static int intel_pt_parse_terms(const struct perf_pmu *pmu, const char *str, u64 *config) { *config = 0; return intel_pt_parse_terms_with_default(pmu, str, config); @@ -178,7 +178,7 @@ static int intel_pt_pick_bit(int bits, int target) return pick; } -static u64 intel_pt_default_config(struct perf_pmu *intel_pt_pmu) +static u64 intel_pt_default_config(const struct perf_pmu *intel_pt_pmu) { char buf[256]; int mtc, mtc_periods = 0, mtc_period; @@ -257,18 +257,17 @@ static int intel_pt_parse_snapshot_options(struct auxtrace_record *itr, return 0; } -struct perf_event_attr * -intel_pt_pmu_default_config(struct perf_pmu *intel_pt_pmu) +void intel_pt_pmu_default_config(const struct perf_pmu *intel_pt_pmu, + struct perf_event_attr *attr) { - struct perf_event_attr *attr; + static u64 config; + static bool initialized; - attr = zalloc(sizeof(struct perf_event_attr)); - if (!attr) - return NULL; - - attr->config = intel_pt_default_config(intel_pt_pmu); - - return attr; + if (!initialized) { + config = intel_pt_default_config(intel_pt_pmu); + initialized = true; + } + attr->config = config; } static const char *intel_pt_find_filter(struct evlist *evlist, diff --git a/tools/perf/arch/x86/util/pmu.c b/tools/perf/arch/x86/util/pmu.c index 6bb77acadb19..bd36c1c885b6 100644 --- a/tools/perf/arch/x86/util/pmu.c +++ b/tools/perf/arch/x86/util/pmu.c @@ -23,7 +23,7 @@ void perf_pmu__arch_init(struct perf_pmu *pmu __maybe_unused) if (!strcmp(pmu->name, INTEL_PT_PMU_NAME)) { pmu->auxtrace = true; pmu->selectable = true; - pmu->default_config = intel_pt_pmu_default_config(pmu); + pmu->perf_event_attr_init_default = intel_pt_pmu_default_config; } if (!strcmp(pmu->name, INTEL_BTS_PMU_NAME)) { pmu->auxtrace = true; diff --git a/tools/perf/util/arm-spe.h b/tools/perf/util/arm-spe.h index 98d3235781c3..4f4900c18f3e 100644 --- a/tools/perf/util/arm-spe.h +++ b/tools/perf/util/arm-spe.h @@ -27,5 +27,7 @@ struct auxtrace_record *arm_spe_recording_init(int *err, int arm_spe_process_auxtrace_info(union perf_event *event, struct perf_session *session); -struct perf_event_attr *arm_spe_pmu_default_config(struct perf_pmu *arm_spe_pmu); +void arm_spe_pmu_default_config(const struct perf_pmu *arm_spe_pmu, + struct perf_event_attr *attr); + #endif diff --git a/tools/perf/util/cs-etm.h b/tools/perf/util/cs-etm.h index 23e132ff05ea..a8caeea720aa 100644 --- a/tools/perf/util/cs-etm.h +++ b/tools/perf/util/cs-etm.h @@ -232,7 +232,7 @@ struct cs_etm_packet_queue { int cs_etm__process_auxtrace_info(union perf_event *event, struct perf_session *session); -struct perf_event_attr *cs_etm_get_default_config(struct perf_pmu *pmu); +void cs_etm_get_default_config(const struct perf_pmu *pmu, struct perf_event_attr *attr); enum cs_etm_pid_fmt { CS_ETM_PIDFMT_NONE, diff --git a/tools/perf/util/intel-pt.h b/tools/perf/util/intel-pt.h index c7d6068e3a6b..18fd0be52e6c 100644 --- a/tools/perf/util/intel-pt.h +++ b/tools/perf/util/intel-pt.h @@ -42,6 +42,7 @@ struct auxtrace_record *intel_pt_recording_init(int *err); int intel_pt_process_auxtrace_info(union perf_event *event, struct perf_session *session); -struct perf_event_attr *intel_pt_pmu_default_config(struct perf_pmu *pmu); +void intel_pt_pmu_default_config(const struct perf_pmu *intel_pt_pmu, + struct perf_event_attr *attr); #endif diff --git a/tools/perf/util/parse-events.c b/tools/perf/util/parse-events.c index c56e07bd7dd6..ea5579510b97 100644 --- a/tools/perf/util/parse-events.c +++ b/tools/perf/util/parse-events.c @@ -1418,11 +1418,10 @@ int parse_events_add_pmu(struct parse_events_state *parse_state, } fix_raw(&parsed_terms, pmu); - if (pmu->default_config) { - memcpy(&attr, pmu->default_config, sizeof(struct perf_event_attr)); - } else { - memset(&attr, 0, sizeof(attr)); - } + memset(&attr, 0, sizeof(attr)); + if (pmu->perf_event_attr_init_default) + pmu->perf_event_attr_init_default(pmu, &attr); + attr.type = pmu->type; if (list_empty(&parsed_terms.terms)) { @@ -1466,7 +1465,8 @@ int parse_events_add_pmu(struct parse_events_state *parse_state, * When using default config, record which bits of attr->config were * changed by the user. */ - if (pmu->default_config && get_config_chgs(pmu, &parsed_terms, &config_terms)) { + if (pmu->perf_event_attr_init_default && + get_config_chgs(pmu, &parsed_terms, &config_terms)) { parse_events_terms__exit(&parsed_terms); return -ENOMEM; } diff --git a/tools/perf/util/pmu.c b/tools/perf/util/pmu.c index faf434119a75..eeb166cd672b 100644 --- a/tools/perf/util/pmu.c +++ b/tools/perf/util/pmu.c @@ -1453,7 +1453,7 @@ int perf_pmu__config(struct perf_pmu *pmu, struct perf_event_attr *attr, struct parse_events_terms *head_terms, struct parse_events_error *err) { - bool zero = !!pmu->default_config; + bool zero = !!pmu->perf_event_attr_init_default; return perf_pmu__config_terms(pmu, attr, head_terms, zero, err); } @@ -2127,7 +2127,6 @@ void perf_pmu__delete(struct perf_pmu *pmu) perf_cpu_map__put(pmu->cpus); - zfree(&pmu->default_config); zfree(&pmu->name); zfree(&pmu->alias_name); zfree(&pmu->id); diff --git a/tools/perf/util/pmu.h b/tools/perf/util/pmu.h index de6c3674ac60..cae00f336e65 100644 --- a/tools/perf/util/pmu.h +++ b/tools/perf/util/pmu.h @@ -92,10 +92,11 @@ struct perf_pmu { */ int max_precise; /** - * @default_config: Optional default perf_event_attr determined in - * architecture specific code. + * @perf_event_attr_init_default: Optional function to default + * initialize PMU specific parts of the perf_event_attr. */ - struct perf_event_attr *default_config; + void (*perf_event_attr_init_default)(const struct perf_pmu *pmu, + struct perf_event_attr *attr); /** * @cpus: Empty or the contents of either of: * /bus/event_source/devices//cpumask. -- Gitee From c35cac57eac9b57ba20773465ed47b4964d0e92e Mon Sep 17 00:00:00 2001 From: Kan Liang Date: Wed, 25 Oct 2023 13:16:25 -0700 Subject: [PATCH 14/22] perf header: Support num and width of branch counters ANBZ: #24284 commit ac9cd7245fffa0fc053afce3b345469e5afa533a upstream. To support the branch counters feature, the information of the maximum number of supported counters and the width of the counters is exposed in the sysfs caps folder. The perf tool can use the information to parse the logged counters in each branch. Store the information in the perf_env for later usage. Reviewed-by: Ian Rogers Signed-off-by: Kan Liang Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Alexey Bayduraev Cc: Andi Kleen Cc: Ingo Molnar Cc: Jiri Olsa Cc: Mark Rutland Cc: Namhyung Kim Cc: Peter Zijlstra Cc: Stephane Eranian Cc: Tinghao Zhang Link: https://lore.kernel.org/r/20231025201626.3000228-7-kan.liang@linux.intel.com Signed-off-by: Arnaldo Carvalho de Melo Signed-off-by: Fangyu Yu --- tools/perf/util/env.h | 5 +++++ tools/perf/util/header.c | 18 +++++++++++++++--- 2 files changed, 20 insertions(+), 3 deletions(-) diff --git a/tools/perf/util/env.h b/tools/perf/util/env.h index bc2d0ef35199..b20655ff32d8 100644 --- a/tools/perf/util/env.h +++ b/tools/perf/util/env.h @@ -46,6 +46,9 @@ struct hybrid_node { struct pmu_caps { int nr_caps; unsigned int max_branches; + unsigned int br_cntr_nr; + unsigned int br_cntr_width; + char **caps; char *pmu_name; }; @@ -62,6 +65,8 @@ struct perf_env { unsigned long long total_mem; unsigned int msr_pmu_type; unsigned int max_branches; + unsigned int br_cntr_nr; + unsigned int br_cntr_width; int kernel_is_64_bit; int nr_cmdline; diff --git a/tools/perf/util/header.c b/tools/perf/util/header.c index 0b44176826bf..efbf591b8a60 100644 --- a/tools/perf/util/header.c +++ b/tools/perf/util/header.c @@ -3262,7 +3262,9 @@ static int process_compressed(struct feat_fd *ff, } static int __process_pmu_caps(struct feat_fd *ff, int *nr_caps, - char ***caps, unsigned int *max_branches) + char ***caps, unsigned int *max_branches, + unsigned int *br_cntr_nr, + unsigned int *br_cntr_width) { char *name, *value, *ptr; u32 nr_pmu_caps, i; @@ -3297,6 +3299,12 @@ static int __process_pmu_caps(struct feat_fd *ff, int *nr_caps, if (!strcmp(name, "branches")) *max_branches = atoi(value); + if (!strcmp(name, "branch_counter_nr")) + *br_cntr_nr = atoi(value); + + if (!strcmp(name, "branch_counter_width")) + *br_cntr_width = atoi(value); + free(value); free(name); } @@ -3321,7 +3329,9 @@ static int process_cpu_pmu_caps(struct feat_fd *ff, { int ret = __process_pmu_caps(ff, &ff->ph->env.nr_cpu_pmu_caps, &ff->ph->env.cpu_pmu_caps, - &ff->ph->env.max_branches); + &ff->ph->env.max_branches, + &ff->ph->env.br_cntr_nr, + &ff->ph->env.br_cntr_width); if (!ret && !ff->ph->env.cpu_pmu_caps) pr_debug("cpu pmu capabilities not available\n"); @@ -3350,7 +3360,9 @@ static int process_pmu_caps(struct feat_fd *ff, void *data __maybe_unused) for (i = 0; i < nr_pmu; i++) { ret = __process_pmu_caps(ff, &pmu_caps[i].nr_caps, &pmu_caps[i].caps, - &pmu_caps[i].max_branches); + &pmu_caps[i].max_branches, + &pmu_caps[i].br_cntr_nr, + &pmu_caps[i].br_cntr_width); if (ret) goto err; -- Gitee From aa923644966b684c514f97e3ebf8784f29d28782 Mon Sep 17 00:00:00 2001 From: Kan Liang Date: Wed, 25 Oct 2023 13:16:26 -0700 Subject: [PATCH 15/22] perf tools: Add branch counter knob ANBZ: #24284 commit 9fbb4b02302b0ae618303565025412070d32f85e upstream. Add a new branch filter, "counter", for the branch counter option. It is used to mark the events which should be logged in the branch. If it is applied with the -j option, the counters of all the events should be logged in the branch. If the legacy kernel doesn't support the new branch sample type, switching off the branch counter filter. The stored counter values in each branch are displayed right after the regular branch stack information via perf report -D. Usage examples: # perf record -e "{branch-instructions,branch-misses}:S" -j any,counter Only the first event, branch-instructions, collect the LBR. Both branch-instructions and branch-misses are marked as logged events. The occurrences information of them can be found in the branch stack extension space of each branch. # perf record -e "{cpu/branch-instructions,branch_type=any/,cpu/branch-misses,branch_type=counter/}" Only the first event, branch-instructions, collect the LBR. Only the branch-misses event is marked as a logged event. Committer notes: I noticed 'perf test "Sample parsing"' failing, reported to the list and Kan provided a patch that checks if the evsel has a leader and that evsel->evlist is set, the comment in the source code further explains it. Reviewed-by: Ian Rogers Signed-off-by: Kan Liang Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Alexey Bayduraev Cc: Andi Kleen Cc: Ingo Molnar Cc: Jiri Olsa Cc: Mark Rutland Cc: Namhyung Kim Cc: Peter Zijlstra Cc: Stephane Eranian Cc: Tinghao Zhang Link: https://lore.kernel.org/r/20231025201626.3000228-8-kan.liang@linux.intel.com Signed-off-by: Arnaldo Carvalho de Melo Signed-off-by: Fangyu Yu --- tools/perf/Documentation/perf-record.txt | 4 +++ tools/perf/util/evsel.c | 35 ++++++++++++++++++++++- tools/perf/util/evsel.h | 1 + tools/perf/util/parse-branch-options.c | 1 + tools/perf/util/perf_event_attr_fprintf.c | 1 + tools/perf/util/sample.h | 1 + tools/perf/util/session.c | 15 ++++++++-- 7 files changed, 55 insertions(+), 3 deletions(-) diff --git a/tools/perf/Documentation/perf-record.txt b/tools/perf/Documentation/perf-record.txt index 1889f66addf2..6015fdd08fb6 100644 --- a/tools/perf/Documentation/perf-record.txt +++ b/tools/perf/Documentation/perf-record.txt @@ -445,6 +445,10 @@ following filters are defined: 4th-Gen Xeon+ server), the save branch type is unconditionally enabled when the taken branch stack sampling is enabled. - priv: save privilege state during sampling in case binary is not available later + - counter: save occurrences of the event since the last branch entry. Currently, the + feature is only supported by a newer CPU, e.g., Intel Sierra Forest and + later platforms. An error out is expected if it's used on the unsupported + kernel or CPUs. + The option requires at least one branch type among any, any_call, any_ret, ind_call, cond. diff --git a/tools/perf/util/evsel.c b/tools/perf/util/evsel.c index 0ef2f41c5ce4..62c7bd12f066 100644 --- a/tools/perf/util/evsel.c +++ b/tools/perf/util/evsel.c @@ -1831,6 +1831,8 @@ static int __evsel__prepare_open(struct evsel *evsel, struct perf_cpu_map *cpus, static void evsel__disable_missing_features(struct evsel *evsel) { + if (perf_missing_features.branch_counters) + evsel->core.attr.branch_sample_type &= ~PERF_SAMPLE_BRANCH_COUNTERS; if (perf_missing_features.read_lost) evsel->core.attr.read_format &= ~PERF_FORMAT_LOST; if (perf_missing_features.weight_struct) { @@ -1884,7 +1886,12 @@ bool evsel__detect_missing_features(struct evsel *evsel) * Must probe features in the order they were added to the * perf_event_attr interface. */ - if (!perf_missing_features.read_lost && + if (!perf_missing_features.branch_counters && + (evsel->core.attr.branch_sample_type & PERF_SAMPLE_BRANCH_COUNTERS)) { + perf_missing_features.branch_counters = true; + pr_debug2("switching off branch counters support\n"); + return true; + } else if (!perf_missing_features.read_lost && (evsel->core.attr.read_format & PERF_FORMAT_LOST)) { perf_missing_features.read_lost = true; pr_debug2("switching off PERF_FORMAT_LOST support\n"); @@ -2344,6 +2351,22 @@ u64 evsel__bitfield_swap_branch_flags(u64 value) return new_val; } +static inline bool evsel__has_branch_counters(const struct evsel *evsel) +{ + struct evsel *cur, *leader = evsel__leader(evsel); + + /* The branch counters feature only supports group */ + if (!leader || !evsel->evlist) + return false; + + evlist__for_each_entry(evsel->evlist, cur) { + if ((leader == evsel__leader(cur)) && + (cur->core.attr.branch_sample_type & PERF_SAMPLE_BRANCH_COUNTERS)) + return true; + } + return false; +} + int evsel__parse_sample(struct evsel *evsel, union perf_event *event, struct perf_sample *data) { @@ -2576,6 +2599,16 @@ int evsel__parse_sample(struct evsel *evsel, union perf_event *event, OVERFLOW_CHECK(array, sz, max_size); array = (void *)array + sz; + + if (evsel__has_branch_counters(evsel)) { + OVERFLOW_CHECK_u64(array); + + data->branch_stack_cntr = (u64 *)array; + sz = data->branch_stack->nr * sizeof(u64); + + OVERFLOW_CHECK(array, sz, max_size); + array = (void *)array + sz; + } } if (type & PERF_SAMPLE_REGS_USER) { diff --git a/tools/perf/util/evsel.h b/tools/perf/util/evsel.h index 848534ec74fa..85f24c986392 100644 --- a/tools/perf/util/evsel.h +++ b/tools/perf/util/evsel.h @@ -191,6 +191,7 @@ struct perf_missing_features { bool code_page_size; bool weight_struct; bool read_lost; + bool branch_counters; }; extern struct perf_missing_features perf_missing_features; diff --git a/tools/perf/util/parse-branch-options.c b/tools/perf/util/parse-branch-options.c index fd67d204d720..f7f7aff3d85a 100644 --- a/tools/perf/util/parse-branch-options.c +++ b/tools/perf/util/parse-branch-options.c @@ -36,6 +36,7 @@ static const struct branch_mode branch_modes[] = { BRANCH_OPT("stack", PERF_SAMPLE_BRANCH_CALL_STACK), BRANCH_OPT("hw_index", PERF_SAMPLE_BRANCH_HW_INDEX), BRANCH_OPT("priv", PERF_SAMPLE_BRANCH_PRIV_SAVE), + BRANCH_OPT("counter", PERF_SAMPLE_BRANCH_COUNTERS), BRANCH_END }; diff --git a/tools/perf/util/perf_event_attr_fprintf.c b/tools/perf/util/perf_event_attr_fprintf.c index 1c1582688f03..59fbbba79697 100644 --- a/tools/perf/util/perf_event_attr_fprintf.c +++ b/tools/perf/util/perf_event_attr_fprintf.c @@ -57,6 +57,7 @@ static void __p_branch_sample_type(char *buf, size_t size, u64 value) bit_name(COND), bit_name(CALL_STACK), bit_name(IND_JUMP), bit_name(CALL), bit_name(NO_FLAGS), bit_name(NO_CYCLES), bit_name(TYPE_SAVE), bit_name(HW_INDEX), bit_name(PRIV_SAVE), + bit_name(COUNTERS), { .name = NULL, } }; #undef bit_name diff --git a/tools/perf/util/sample.h b/tools/perf/util/sample.h index c92ad0f51ecd..70b2c3135555 100644 --- a/tools/perf/util/sample.h +++ b/tools/perf/util/sample.h @@ -113,6 +113,7 @@ struct perf_sample { void *raw_data; struct ip_callchain *callchain; struct branch_stack *branch_stack; + u64 *branch_stack_cntr; struct regs_dump user_regs; struct regs_dump intr_regs; struct stack_dump user_stack; diff --git a/tools/perf/util/session.c b/tools/perf/util/session.c index e1edcc64f0ae..14c6ba18ecc4 100644 --- a/tools/perf/util/session.c +++ b/tools/perf/util/session.c @@ -1155,9 +1155,13 @@ static void callchain__printf(struct evsel *evsel, i, callchain->ips[i]); } -static void branch_stack__printf(struct perf_sample *sample, bool callstack) +static void branch_stack__printf(struct perf_sample *sample, + struct evsel *evsel) { struct branch_entry *entries = perf_sample__branch_entries(sample); + bool callstack = evsel__has_branch_callstack(evsel); + u64 *branch_stack_cntr = sample->branch_stack_cntr; + struct perf_env *env = evsel__env(evsel); uint64_t i; if (!callstack) { @@ -1199,6 +1203,13 @@ static void branch_stack__printf(struct perf_sample *sample, bool callstack) } } } + + if (branch_stack_cntr) { + printf("... branch stack counters: nr:%" PRIu64 " (counter width: %u max counter nr:%u)\n", + sample->branch_stack->nr, env->br_cntr_width, env->br_cntr_nr); + for (i = 0; i < sample->branch_stack->nr; i++) + printf("..... %2"PRIu64": %016" PRIx64 "\n", i, branch_stack_cntr[i]); + } } static void regs_dump__printf(u64 mask, u64 *regs, const char *arch) @@ -1360,7 +1371,7 @@ static void dump_sample(struct evsel *evsel, union perf_event *event, callchain__printf(evsel, sample); if (evsel__has_br_stack(evsel)) - branch_stack__printf(sample, evsel__has_branch_callstack(evsel)); + branch_stack__printf(sample, evsel); if (sample_type & PERF_SAMPLE_REGS_USER) regs_user__printf(sample, arch); -- Gitee From c47882a16690dfec952a10b007dc979516f557dd Mon Sep 17 00:00:00 2001 From: Kan Liang Date: Wed, 25 Oct 2023 13:16:24 -0700 Subject: [PATCH 16/22] tools headers UAPI: Sync include/uapi/linux/perf_event.h header with the kernel ANBZ: #24284 commit 76db7aab1fca6688ddf9f388157521c442e0ffb8 upstream. Sync the new sample type for the branch counters feature. Signed-off-by: Kan Liang Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Alexey Bayduraev Cc: Andi Kleen Cc: Ian Rogers Cc: Ingo Molnar Cc: Jiri Olsa Cc: Mark Rutland Cc: Namhyung Kim Cc: Peter Zijlstra Cc: Stephane Eranian Cc: Tinghao Zhang Link: https://lore.kernel.org/r/20231025201626.3000228-6-kan.liang@linux.intel.com Signed-off-by: Arnaldo Carvalho de Melo Signed-off-by: Fangyu Yu --- tools/include/uapi/linux/perf_event.h | 13 +++++++++++++ 1 file changed, 13 insertions(+) diff --git a/tools/include/uapi/linux/perf_event.h b/tools/include/uapi/linux/perf_event.h index 39c6a250dd1b..3a64499b0f5d 100644 --- a/tools/include/uapi/linux/perf_event.h +++ b/tools/include/uapi/linux/perf_event.h @@ -204,6 +204,8 @@ enum perf_branch_sample_type_shift { PERF_SAMPLE_BRANCH_PRIV_SAVE_SHIFT = 18, /* save privilege mode */ + PERF_SAMPLE_BRANCH_COUNTERS_SHIFT = 19, /* save occurrences of events on a branch */ + PERF_SAMPLE_BRANCH_MAX_SHIFT /* non-ABI */ }; @@ -235,6 +237,8 @@ enum perf_branch_sample_type { PERF_SAMPLE_BRANCH_PRIV_SAVE = 1U << PERF_SAMPLE_BRANCH_PRIV_SAVE_SHIFT, + PERF_SAMPLE_BRANCH_COUNTERS = 1U << PERF_SAMPLE_BRANCH_COUNTERS_SHIFT, + PERF_SAMPLE_BRANCH_MAX = 1U << PERF_SAMPLE_BRANCH_MAX_SHIFT, }; @@ -982,6 +986,12 @@ enum perf_event_type { * { u64 nr; * { u64 hw_idx; } && PERF_SAMPLE_BRANCH_HW_INDEX * { u64 from, to, flags } lbr[nr]; + * # + * # The format of the counters is decided by the + * # "branch_counter_nr" and "branch_counter_width", + * # which are defined in the ABI. + * # + * { u64 counters; } cntr[nr] && PERF_SAMPLE_BRANCH_COUNTERS * } && PERF_SAMPLE_BRANCH_STACK * * { u64 abi; # enum perf_sample_regs_abi @@ -1427,6 +1437,9 @@ struct perf_branch_entry { reserved:31; }; +/* Size of used info bits in struct perf_branch_entry */ +#define PERF_BRANCH_ENTRY_INFO_BITS_MAX 33 + union perf_sample_weight { __u64 full; #if defined(__LITTLE_ENDIAN_BITFIELD) -- Gitee From 40bedbe850f54c9be3a720f0502756d6ae6770a7 Mon Sep 17 00:00:00 2001 From: Shenlin Liang Date: Mon, 22 Apr 2024 08:08:32 +0000 Subject: [PATCH 17/22] RISCV: KVM: add tracepoints for entry and exit events ANBZ: #24284 commit 91195a90f1d1ae72a1a49681ee30118c9f7ab8c3 upstream. Like other architectures, RISCV KVM also needs to add these event tracepoints to count the number of times kvm guest entry/exit. Signed-off-by: Shenlin Liang Reviewed-by: Anup Patel Tested-by: Atish Patra Link: https://lore.kernel.org/r/20240422080833.8745-2-liangshenlin@eswincomputing.com Signed-off-by: Anup Patel Signed-off-by: Fangyu Yu --- arch/riscv/kvm/trace.h | 67 ++++++++++++++++++++++++++++++++++++++++++ arch/riscv/kvm/vcpu.c | 7 +++++ 2 files changed, 74 insertions(+) create mode 100644 arch/riscv/kvm/trace.h diff --git a/arch/riscv/kvm/trace.h b/arch/riscv/kvm/trace.h new file mode 100644 index 000000000000..3d54175d805c --- /dev/null +++ b/arch/riscv/kvm/trace.h @@ -0,0 +1,67 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Tracepoints for RISC-V KVM + * + * Copyright 2024 Beijing ESWIN Computing Technology Co., Ltd. + * + */ +#if !defined(_TRACE_KVM_H) || defined(TRACE_HEADER_MULTI_READ) +#define _TRACE_KVM_H + +#include + +#undef TRACE_SYSTEM +#define TRACE_SYSTEM kvm + +TRACE_EVENT(kvm_entry, + TP_PROTO(struct kvm_vcpu *vcpu), + TP_ARGS(vcpu), + + TP_STRUCT__entry( + __field(unsigned long, pc) + ), + + TP_fast_assign( + __entry->pc = vcpu->arch.guest_context.sepc; + ), + + TP_printk("PC: 0x016%lx", __entry->pc) +); + +TRACE_EVENT(kvm_exit, + TP_PROTO(struct kvm_cpu_trap *trap), + TP_ARGS(trap), + + TP_STRUCT__entry( + __field(unsigned long, sepc) + __field(unsigned long, scause) + __field(unsigned long, stval) + __field(unsigned long, htval) + __field(unsigned long, htinst) + ), + + TP_fast_assign( + __entry->sepc = trap->sepc; + __entry->scause = trap->scause; + __entry->stval = trap->stval; + __entry->htval = trap->htval; + __entry->htinst = trap->htinst; + ), + + TP_printk("SEPC:0x%lx, SCAUSE:0x%lx, STVAL:0x%lx, HTVAL:0x%lx, HTINST:0x%lx", + __entry->sepc, + __entry->scause, + __entry->stval, + __entry->htval, + __entry->htinst) +); + +#endif /* _TRACE_RSICV_KVM_H */ + +#undef TRACE_INCLUDE_PATH +#define TRACE_INCLUDE_PATH . +#undef TRACE_INCLUDE_FILE +#define TRACE_INCLUDE_FILE trace + +/* This part must be outside protection */ +#include diff --git a/arch/riscv/kvm/vcpu.c b/arch/riscv/kvm/vcpu.c index 14b166ae80db..0c0aee5b7632 100644 --- a/arch/riscv/kvm/vcpu.c +++ b/arch/riscv/kvm/vcpu.c @@ -21,6 +21,9 @@ #include #include +#define CREATE_TRACE_POINTS +#include "trace.h" + const struct _kvm_stats_desc kvm_vcpu_stats_desc[] = { KVM_GENERIC_VCPU_STATS(), STATS_DESC_COUNTER(VCPU, ecall_exit_stat), @@ -810,6 +813,8 @@ int kvm_arch_vcpu_ioctl_run(struct kvm_vcpu *vcpu) */ kvm_riscv_local_tlb_sanitize(vcpu); + trace_kvm_entry(vcpu); + guest_timing_enter_irqoff(); kvm_riscv_vcpu_enter_exit(vcpu); @@ -848,6 +853,8 @@ int kvm_arch_vcpu_ioctl_run(struct kvm_vcpu *vcpu) local_irq_enable(); + trace_kvm_exit(&trap); + preempt_enable(); kvm_vcpu_srcu_read_lock(vcpu); -- Gitee From 15e7cf3e0a621dfeea4dcbb3e3b3108087f144c6 Mon Sep 17 00:00:00 2001 From: Shenlin Liang Date: Mon, 22 Apr 2024 08:08:33 +0000 Subject: [PATCH 18/22] perf kvm/riscv: Port perf kvm stat to RISC-V ANBZ: #24284 commit da7b1b525e972b8c5b16640fa5b2ff2497b5c652 upstream. 'perf kvm stat report/record' generates a statistical analysis of KVM events and can be used to analyze guest exit reasons. "report" reports statistical analysis of guest exit events. To record kvm events on the host: # perf kvm stat record -a To report kvm VM EXIT events: # perf kvm stat report --event=vmexit Signed-off-by: Shenlin Liang Reviewed-by: Atish Patra Tested-by: Atish Patra Link: https://lore.kernel.org/r/20240422080833.8745-3-liangshenlin@eswincomputing.com Signed-off-by: Anup Patel Signed-off-by: Fangyu Yu --- tools/perf/arch/riscv/Makefile | 1 + tools/perf/arch/riscv/util/Build | 1 + tools/perf/arch/riscv/util/kvm-stat.c | 78 +++++++++++++++++++ .../arch/riscv/util/riscv_exception_types.h | 35 +++++++++ 4 files changed, 115 insertions(+) create mode 100644 tools/perf/arch/riscv/util/kvm-stat.c create mode 100644 tools/perf/arch/riscv/util/riscv_exception_types.h diff --git a/tools/perf/arch/riscv/Makefile b/tools/perf/arch/riscv/Makefile index a8d25d005207..90c3c476a242 100644 --- a/tools/perf/arch/riscv/Makefile +++ b/tools/perf/arch/riscv/Makefile @@ -3,3 +3,4 @@ PERF_HAVE_DWARF_REGS := 1 endif PERF_HAVE_ARCH_REGS_QUERY_REGISTER_OFFSET := 1 PERF_HAVE_JITDUMP := 1 +HAVE_KVM_STAT_SUPPORT := 1 diff --git a/tools/perf/arch/riscv/util/Build b/tools/perf/arch/riscv/util/Build index 603dbb5ae4dc..d72b04f8d32b 100644 --- a/tools/perf/arch/riscv/util/Build +++ b/tools/perf/arch/riscv/util/Build @@ -1,5 +1,6 @@ perf-y += perf_regs.o perf-y += header.o +perf-$(CONFIG_LIBTRACEEVENT) += kvm-stat.o perf-$(CONFIG_DWARF) += dwarf-regs.o perf-$(CONFIG_LIBDW_DWARF_UNWIND) += unwind-libdw.o diff --git a/tools/perf/arch/riscv/util/kvm-stat.c b/tools/perf/arch/riscv/util/kvm-stat.c new file mode 100644 index 000000000000..491aef449d1a --- /dev/null +++ b/tools/perf/arch/riscv/util/kvm-stat.c @@ -0,0 +1,78 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Arch specific functions for perf kvm stat. + * + * Copyright 2024 Beijing ESWIN Computing Technology Co., Ltd. + * + */ +#include +#include +#include "../../../util/evsel.h" +#include "../../../util/kvm-stat.h" +#include "riscv_exception_types.h" +#include "debug.h" + +define_exit_reasons_table(riscv_exit_reasons, kvm_riscv_exception_class); + +const char *vcpu_id_str = "id"; +const char *kvm_exit_reason = "scause"; +const char *kvm_entry_trace = "kvm:kvm_entry"; +const char *kvm_exit_trace = "kvm:kvm_exit"; + +const char *kvm_events_tp[] = { + "kvm:kvm_entry", + "kvm:kvm_exit", + NULL, +}; + +static void event_get_key(struct evsel *evsel, + struct perf_sample *sample, + struct event_key *key) +{ + key->info = 0; + key->key = evsel__intval(evsel, sample, kvm_exit_reason); + key->exit_reasons = riscv_exit_reasons; +} + +static bool event_begin(struct evsel *evsel, + struct perf_sample *sample __maybe_unused, + struct event_key *key __maybe_unused) +{ + return evsel__name_is(evsel, kvm_entry_trace); +} + +static bool event_end(struct evsel *evsel, + struct perf_sample *sample, + struct event_key *key) +{ + if (evsel__name_is(evsel, kvm_exit_trace)) { + event_get_key(evsel, sample, key); + return true; + } + return false; +} + +static struct kvm_events_ops exit_events = { + .is_begin_event = event_begin, + .is_end_event = event_end, + .decode_key = exit_event_decode_key, + .name = "VM-EXIT" +}; + +struct kvm_reg_events_ops kvm_reg_events_ops[] = { + { + .name = "vmexit", + .ops = &exit_events, + }, + { NULL, NULL }, +}; + +const char * const kvm_skip_events[] = { + NULL, +}; + +int cpu_isa_init(struct perf_kvm_stat *kvm, const char *cpuid __maybe_unused) +{ + kvm->exit_reasons_isa = "riscv64"; + return 0; +} diff --git a/tools/perf/arch/riscv/util/riscv_exception_types.h b/tools/perf/arch/riscv/util/riscv_exception_types.h new file mode 100644 index 000000000000..c49b8fa5e847 --- /dev/null +++ b/tools/perf/arch/riscv/util/riscv_exception_types.h @@ -0,0 +1,35 @@ +// SPDX-License-Identifier: GPL-2.0 +#ifndef ARCH_PERF_RISCV_EXCEPTION_TYPES_H +#define ARCH_PERF_RISCV_EXCEPTION_TYPES_H + +#define EXC_INST_MISALIGNED 0 +#define EXC_INST_ACCESS 1 +#define EXC_INST_ILLEGAL 2 +#define EXC_BREAKPOINT 3 +#define EXC_LOAD_MISALIGNED 4 +#define EXC_LOAD_ACCESS 5 +#define EXC_STORE_MISALIGNED 6 +#define EXC_STORE_ACCESS 7 +#define EXC_SYSCALL 8 +#define EXC_HYPERVISOR_SYSCALL 9 +#define EXC_SUPERVISOR_SYSCALL 10 +#define EXC_INST_PAGE_FAULT 12 +#define EXC_LOAD_PAGE_FAULT 13 +#define EXC_STORE_PAGE_FAULT 15 +#define EXC_INST_GUEST_PAGE_FAULT 20 +#define EXC_LOAD_GUEST_PAGE_FAULT 21 +#define EXC_VIRTUAL_INST_FAULT 22 +#define EXC_STORE_GUEST_PAGE_FAULT 23 + +#define EXC(x) {EXC_##x, #x } + +#define kvm_riscv_exception_class \ + EXC(INST_MISALIGNED), EXC(INST_ACCESS), EXC(INST_ILLEGAL), \ + EXC(BREAKPOINT), EXC(LOAD_MISALIGNED), EXC(LOAD_ACCESS), \ + EXC(STORE_MISALIGNED), EXC(STORE_ACCESS), EXC(SYSCALL), \ + EXC(HYPERVISOR_SYSCALL), EXC(SUPERVISOR_SYSCALL), \ + EXC(INST_PAGE_FAULT), EXC(LOAD_PAGE_FAULT), EXC(STORE_PAGE_FAULT), \ + EXC(INST_GUEST_PAGE_FAULT), EXC(LOAD_GUEST_PAGE_FAULT), \ + EXC(VIRTUAL_INST_FAULT), EXC(STORE_GUEST_PAGE_FAULT) + +#endif /* ARCH_PERF_RISCV_EXCEPTION_TYPES_H */ -- Gitee From d91791983104efdd793c92f796a8f4ee93968aba Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Thu, 22 May 2025 09:51:22 +0200 Subject: [PATCH 19/22] perf/uapi: Clean up a bit ANBZ: #24284 commit 44889ff67cee7b9ee2d305690ce7a5488b137a66 upstream. When applying a recent commit to the header I noticed that we have accumulated quite a bit of historic noise in this header, so do a bit of spring cleaning: - Define bitfields in a vertically aligned fashion, like perf_event_mmap_page::capabilities already does. This makes it easier to see the distribution and sizing of bits within a word, at a glance. The following is much more readable: __u64 cap_bit0 : 1, cap_bit0_is_deprecated : 1, cap_user_rdpmc : 1, cap_user_time : 1, cap_user_time_zero : 1, cap_user_time_short : 1, cap_____res : 58; Than: __u64 cap_bit0:1, cap_bit0_is_deprecated:1, cap_user_rdpmc:1, cap_user_time:1, cap_user_time_zero:1, cap_user_time_short:1, cap_____res:58; So convert all bitfield definitions from the latter style to the former style. - Fix typos and grammar - Fix capitalization - Remove whitespace noise - Harmonize the definitions of various generations and groups of PERF_MEM_ ABI values. - Vertically align all definitions and assignments to the same column (48), as the first definition (enum perf_type_id), throughout the entire header. - And in general make the code and comments to be more in sync with each other and to be more readable overall. No change in functionality. Copy the changes over to tools/include/uapi/linux/perf_event.h. Signed-off-by: Ingo Molnar Cc: Peter Zijlstra Cc: Arnaldo Carvalho de Melo Cc: Jiri Olsa Cc: Alexander Shishkin Cc: Mark Rutland Cc: Namhyung Kim Cc: Ian Rogers Link: https://lore.kernel.org/r/20250521221529.2547099-1-irogers@google.com Signed-off-by: Fangyu Yu --- include/uapi/linux/perf_event.h | 648 +++++++++++++------------- tools/include/uapi/linux/perf_event.h | 648 +++++++++++++------------- 2 files changed, 662 insertions(+), 634 deletions(-) diff --git a/include/uapi/linux/perf_event.h b/include/uapi/linux/perf_event.h index 3a64499b0f5d..365184d931ad 100644 --- a/include/uapi/linux/perf_event.h +++ b/include/uapi/linux/perf_event.h @@ -39,18 +39,21 @@ enum perf_type_id { /* * attr.config layout for type PERF_TYPE_HARDWARE and PERF_TYPE_HW_CACHE + * * PERF_TYPE_HARDWARE: 0xEEEEEEEE000000AA * AA: hardware event ID * EEEEEEEE: PMU type ID + * * PERF_TYPE_HW_CACHE: 0xEEEEEEEE00DDCCBB * BB: hardware cache ID * CC: hardware cache op ID * DD: hardware cache op result ID * EEEEEEEE: PMU type ID - * If the PMU type ID is 0, the PERF_TYPE_RAW will be applied. + * + * If the PMU type ID is 0, PERF_TYPE_RAW will be applied. */ -#define PERF_PMU_TYPE_SHIFT 32 -#define PERF_HW_EVENT_MASK 0xffffffff +#define PERF_PMU_TYPE_SHIFT 32 +#define PERF_HW_EVENT_MASK 0xffffffff /* * Generalized performance event event_id types, used by the @@ -112,7 +115,7 @@ enum perf_hw_cache_op_result_id { /* * Special "software" events provided by the kernel, even if the hardware * does not support performance events. These events measure various - * physical and sw events of the kernel (and allow the profiling of them as + * physical and SW events of the kernel (and allow the profiling of them as * well): */ enum perf_sw_ids { @@ -167,8 +170,9 @@ enum perf_event_sample_format { }; #define PERF_SAMPLE_WEIGHT_TYPE (PERF_SAMPLE_WEIGHT | PERF_SAMPLE_WEIGHT_STRUCT) + /* - * values to program into branch_sample_type when PERF_SAMPLE_BRANCH is set + * Values to program into branch_sample_type when PERF_SAMPLE_BRANCH is set. * * If the user does not pass priv level information via branch_sample_type, * the kernel uses the event's priv level. Branch and event priv levels do @@ -178,20 +182,20 @@ enum perf_event_sample_format { * of branches and therefore it supersedes all the other types. */ enum perf_branch_sample_type_shift { - PERF_SAMPLE_BRANCH_USER_SHIFT = 0, /* user branches */ - PERF_SAMPLE_BRANCH_KERNEL_SHIFT = 1, /* kernel branches */ - PERF_SAMPLE_BRANCH_HV_SHIFT = 2, /* hypervisor branches */ - - PERF_SAMPLE_BRANCH_ANY_SHIFT = 3, /* any branch types */ - PERF_SAMPLE_BRANCH_ANY_CALL_SHIFT = 4, /* any call branch */ - PERF_SAMPLE_BRANCH_ANY_RETURN_SHIFT = 5, /* any return branch */ - PERF_SAMPLE_BRANCH_IND_CALL_SHIFT = 6, /* indirect calls */ - PERF_SAMPLE_BRANCH_ABORT_TX_SHIFT = 7, /* transaction aborts */ - PERF_SAMPLE_BRANCH_IN_TX_SHIFT = 8, /* in transaction */ - PERF_SAMPLE_BRANCH_NO_TX_SHIFT = 9, /* not in transaction */ + PERF_SAMPLE_BRANCH_USER_SHIFT = 0, /* user branches */ + PERF_SAMPLE_BRANCH_KERNEL_SHIFT = 1, /* kernel branches */ + PERF_SAMPLE_BRANCH_HV_SHIFT = 2, /* hypervisor branches */ + + PERF_SAMPLE_BRANCH_ANY_SHIFT = 3, /* any branch types */ + PERF_SAMPLE_BRANCH_ANY_CALL_SHIFT = 4, /* any call branch */ + PERF_SAMPLE_BRANCH_ANY_RETURN_SHIFT = 5, /* any return branch */ + PERF_SAMPLE_BRANCH_IND_CALL_SHIFT = 6, /* indirect calls */ + PERF_SAMPLE_BRANCH_ABORT_TX_SHIFT = 7, /* transaction aborts */ + PERF_SAMPLE_BRANCH_IN_TX_SHIFT = 8, /* in transaction */ + PERF_SAMPLE_BRANCH_NO_TX_SHIFT = 9, /* not in transaction */ PERF_SAMPLE_BRANCH_COND_SHIFT = 10, /* conditional branches */ - PERF_SAMPLE_BRANCH_CALL_STACK_SHIFT = 11, /* call/ret stack */ + PERF_SAMPLE_BRANCH_CALL_STACK_SHIFT = 11, /* CALL/RET stack */ PERF_SAMPLE_BRANCH_IND_JUMP_SHIFT = 12, /* indirect jumps */ PERF_SAMPLE_BRANCH_CALL_SHIFT = 13, /* direct call */ @@ -210,96 +214,95 @@ enum perf_branch_sample_type_shift { }; enum perf_branch_sample_type { - PERF_SAMPLE_BRANCH_USER = 1U << PERF_SAMPLE_BRANCH_USER_SHIFT, - PERF_SAMPLE_BRANCH_KERNEL = 1U << PERF_SAMPLE_BRANCH_KERNEL_SHIFT, - PERF_SAMPLE_BRANCH_HV = 1U << PERF_SAMPLE_BRANCH_HV_SHIFT, + PERF_SAMPLE_BRANCH_USER = 1U << PERF_SAMPLE_BRANCH_USER_SHIFT, + PERF_SAMPLE_BRANCH_KERNEL = 1U << PERF_SAMPLE_BRANCH_KERNEL_SHIFT, + PERF_SAMPLE_BRANCH_HV = 1U << PERF_SAMPLE_BRANCH_HV_SHIFT, - PERF_SAMPLE_BRANCH_ANY = 1U << PERF_SAMPLE_BRANCH_ANY_SHIFT, - PERF_SAMPLE_BRANCH_ANY_CALL = 1U << PERF_SAMPLE_BRANCH_ANY_CALL_SHIFT, - PERF_SAMPLE_BRANCH_ANY_RETURN = 1U << PERF_SAMPLE_BRANCH_ANY_RETURN_SHIFT, - PERF_SAMPLE_BRANCH_IND_CALL = 1U << PERF_SAMPLE_BRANCH_IND_CALL_SHIFT, - PERF_SAMPLE_BRANCH_ABORT_TX = 1U << PERF_SAMPLE_BRANCH_ABORT_TX_SHIFT, - PERF_SAMPLE_BRANCH_IN_TX = 1U << PERF_SAMPLE_BRANCH_IN_TX_SHIFT, - PERF_SAMPLE_BRANCH_NO_TX = 1U << PERF_SAMPLE_BRANCH_NO_TX_SHIFT, - PERF_SAMPLE_BRANCH_COND = 1U << PERF_SAMPLE_BRANCH_COND_SHIFT, + PERF_SAMPLE_BRANCH_ANY = 1U << PERF_SAMPLE_BRANCH_ANY_SHIFT, + PERF_SAMPLE_BRANCH_ANY_CALL = 1U << PERF_SAMPLE_BRANCH_ANY_CALL_SHIFT, + PERF_SAMPLE_BRANCH_ANY_RETURN = 1U << PERF_SAMPLE_BRANCH_ANY_RETURN_SHIFT, + PERF_SAMPLE_BRANCH_IND_CALL = 1U << PERF_SAMPLE_BRANCH_IND_CALL_SHIFT, + PERF_SAMPLE_BRANCH_ABORT_TX = 1U << PERF_SAMPLE_BRANCH_ABORT_TX_SHIFT, + PERF_SAMPLE_BRANCH_IN_TX = 1U << PERF_SAMPLE_BRANCH_IN_TX_SHIFT, + PERF_SAMPLE_BRANCH_NO_TX = 1U << PERF_SAMPLE_BRANCH_NO_TX_SHIFT, + PERF_SAMPLE_BRANCH_COND = 1U << PERF_SAMPLE_BRANCH_COND_SHIFT, - PERF_SAMPLE_BRANCH_CALL_STACK = 1U << PERF_SAMPLE_BRANCH_CALL_STACK_SHIFT, - PERF_SAMPLE_BRANCH_IND_JUMP = 1U << PERF_SAMPLE_BRANCH_IND_JUMP_SHIFT, - PERF_SAMPLE_BRANCH_CALL = 1U << PERF_SAMPLE_BRANCH_CALL_SHIFT, + PERF_SAMPLE_BRANCH_CALL_STACK = 1U << PERF_SAMPLE_BRANCH_CALL_STACK_SHIFT, + PERF_SAMPLE_BRANCH_IND_JUMP = 1U << PERF_SAMPLE_BRANCH_IND_JUMP_SHIFT, + PERF_SAMPLE_BRANCH_CALL = 1U << PERF_SAMPLE_BRANCH_CALL_SHIFT, - PERF_SAMPLE_BRANCH_NO_FLAGS = 1U << PERF_SAMPLE_BRANCH_NO_FLAGS_SHIFT, - PERF_SAMPLE_BRANCH_NO_CYCLES = 1U << PERF_SAMPLE_BRANCH_NO_CYCLES_SHIFT, + PERF_SAMPLE_BRANCH_NO_FLAGS = 1U << PERF_SAMPLE_BRANCH_NO_FLAGS_SHIFT, + PERF_SAMPLE_BRANCH_NO_CYCLES = 1U << PERF_SAMPLE_BRANCH_NO_CYCLES_SHIFT, - PERF_SAMPLE_BRANCH_TYPE_SAVE = - 1U << PERF_SAMPLE_BRANCH_TYPE_SAVE_SHIFT, + PERF_SAMPLE_BRANCH_TYPE_SAVE = 1U << PERF_SAMPLE_BRANCH_TYPE_SAVE_SHIFT, - PERF_SAMPLE_BRANCH_HW_INDEX = 1U << PERF_SAMPLE_BRANCH_HW_INDEX_SHIFT, + PERF_SAMPLE_BRANCH_HW_INDEX = 1U << PERF_SAMPLE_BRANCH_HW_INDEX_SHIFT, - PERF_SAMPLE_BRANCH_PRIV_SAVE = 1U << PERF_SAMPLE_BRANCH_PRIV_SAVE_SHIFT, + PERF_SAMPLE_BRANCH_PRIV_SAVE = 1U << PERF_SAMPLE_BRANCH_PRIV_SAVE_SHIFT, - PERF_SAMPLE_BRANCH_COUNTERS = 1U << PERF_SAMPLE_BRANCH_COUNTERS_SHIFT, + PERF_SAMPLE_BRANCH_COUNTERS = 1U << PERF_SAMPLE_BRANCH_COUNTERS_SHIFT, - PERF_SAMPLE_BRANCH_MAX = 1U << PERF_SAMPLE_BRANCH_MAX_SHIFT, + PERF_SAMPLE_BRANCH_MAX = 1U << PERF_SAMPLE_BRANCH_MAX_SHIFT, }; /* - * Common flow change classification + * Common control flow change classifications: */ enum { - PERF_BR_UNKNOWN = 0, /* unknown */ - PERF_BR_COND = 1, /* conditional */ - PERF_BR_UNCOND = 2, /* unconditional */ - PERF_BR_IND = 3, /* indirect */ - PERF_BR_CALL = 4, /* function call */ - PERF_BR_IND_CALL = 5, /* indirect function call */ - PERF_BR_RET = 6, /* function return */ - PERF_BR_SYSCALL = 7, /* syscall */ - PERF_BR_SYSRET = 8, /* syscall return */ - PERF_BR_COND_CALL = 9, /* conditional function call */ - PERF_BR_COND_RET = 10, /* conditional function return */ - PERF_BR_ERET = 11, /* exception return */ - PERF_BR_IRQ = 12, /* irq */ - PERF_BR_SERROR = 13, /* system error */ - PERF_BR_NO_TX = 14, /* not in transaction */ - PERF_BR_EXTEND_ABI = 15, /* extend ABI */ + PERF_BR_UNKNOWN = 0, /* Unknown */ + PERF_BR_COND = 1, /* Conditional */ + PERF_BR_UNCOND = 2, /* Unconditional */ + PERF_BR_IND = 3, /* Indirect */ + PERF_BR_CALL = 4, /* Function call */ + PERF_BR_IND_CALL = 5, /* Indirect function call */ + PERF_BR_RET = 6, /* Function return */ + PERF_BR_SYSCALL = 7, /* Syscall */ + PERF_BR_SYSRET = 8, /* Syscall return */ + PERF_BR_COND_CALL = 9, /* Conditional function call */ + PERF_BR_COND_RET = 10, /* Conditional function return */ + PERF_BR_ERET = 11, /* Exception return */ + PERF_BR_IRQ = 12, /* IRQ */ + PERF_BR_SERROR = 13, /* System error */ + PERF_BR_NO_TX = 14, /* Not in transaction */ + PERF_BR_EXTEND_ABI = 15, /* Extend ABI */ PERF_BR_MAX, }; /* - * Common branch speculation outcome classification + * Common branch speculation outcome classifications: */ enum { - PERF_BR_SPEC_NA = 0, /* Not available */ - PERF_BR_SPEC_WRONG_PATH = 1, /* Speculative but on wrong path */ - PERF_BR_NON_SPEC_CORRECT_PATH = 2, /* Non-speculative but on correct path */ - PERF_BR_SPEC_CORRECT_PATH = 3, /* Speculative and on correct path */ + PERF_BR_SPEC_NA = 0, /* Not available */ + PERF_BR_SPEC_WRONG_PATH = 1, /* Speculative but on wrong path */ + PERF_BR_NON_SPEC_CORRECT_PATH = 2, /* Non-speculative but on correct path */ + PERF_BR_SPEC_CORRECT_PATH = 3, /* Speculative and on correct path */ PERF_BR_SPEC_MAX, }; enum { - PERF_BR_NEW_FAULT_ALGN = 0, /* Alignment fault */ - PERF_BR_NEW_FAULT_DATA = 1, /* Data fault */ - PERF_BR_NEW_FAULT_INST = 2, /* Inst fault */ - PERF_BR_NEW_ARCH_1 = 3, /* Architecture specific */ - PERF_BR_NEW_ARCH_2 = 4, /* Architecture specific */ - PERF_BR_NEW_ARCH_3 = 5, /* Architecture specific */ - PERF_BR_NEW_ARCH_4 = 6, /* Architecture specific */ - PERF_BR_NEW_ARCH_5 = 7, /* Architecture specific */ + PERF_BR_NEW_FAULT_ALGN = 0, /* Alignment fault */ + PERF_BR_NEW_FAULT_DATA = 1, /* Data fault */ + PERF_BR_NEW_FAULT_INST = 2, /* Inst fault */ + PERF_BR_NEW_ARCH_1 = 3, /* Architecture specific */ + PERF_BR_NEW_ARCH_2 = 4, /* Architecture specific */ + PERF_BR_NEW_ARCH_3 = 5, /* Architecture specific */ + PERF_BR_NEW_ARCH_4 = 6, /* Architecture specific */ + PERF_BR_NEW_ARCH_5 = 7, /* Architecture specific */ PERF_BR_NEW_MAX, }; enum { - PERF_BR_PRIV_UNKNOWN = 0, - PERF_BR_PRIV_USER = 1, - PERF_BR_PRIV_KERNEL = 2, - PERF_BR_PRIV_HV = 3, + PERF_BR_PRIV_UNKNOWN = 0, + PERF_BR_PRIV_USER = 1, + PERF_BR_PRIV_KERNEL = 2, + PERF_BR_PRIV_HV = 3, }; -#define PERF_BR_ARM64_FIQ PERF_BR_NEW_ARCH_1 -#define PERF_BR_ARM64_DEBUG_HALT PERF_BR_NEW_ARCH_2 -#define PERF_BR_ARM64_DEBUG_EXIT PERF_BR_NEW_ARCH_3 -#define PERF_BR_ARM64_DEBUG_INST PERF_BR_NEW_ARCH_4 -#define PERF_BR_ARM64_DEBUG_DATA PERF_BR_NEW_ARCH_5 +#define PERF_BR_ARM64_FIQ PERF_BR_NEW_ARCH_1 +#define PERF_BR_ARM64_DEBUG_HALT PERF_BR_NEW_ARCH_2 +#define PERF_BR_ARM64_DEBUG_EXIT PERF_BR_NEW_ARCH_3 +#define PERF_BR_ARM64_DEBUG_INST PERF_BR_NEW_ARCH_4 +#define PERF_BR_ARM64_DEBUG_DATA PERF_BR_NEW_ARCH_5 #define PERF_SAMPLE_BRANCH_PLM_ALL \ (PERF_SAMPLE_BRANCH_USER|\ @@ -310,9 +313,9 @@ enum { * Values to determine ABI of the registers dump. */ enum perf_sample_regs_abi { - PERF_SAMPLE_REGS_ABI_NONE = 0, - PERF_SAMPLE_REGS_ABI_32 = 1, - PERF_SAMPLE_REGS_ABI_64 = 2, + PERF_SAMPLE_REGS_ABI_NONE = 0, + PERF_SAMPLE_REGS_ABI_32 = 1, + PERF_SAMPLE_REGS_ABI_64 = 2, }; /* @@ -320,21 +323,21 @@ enum perf_sample_regs_abi { * abort events. Multiple bits can be set. */ enum { - PERF_TXN_ELISION = (1 << 0), /* From elision */ - PERF_TXN_TRANSACTION = (1 << 1), /* From transaction */ - PERF_TXN_SYNC = (1 << 2), /* Instruction is related */ - PERF_TXN_ASYNC = (1 << 3), /* Instruction not related */ - PERF_TXN_RETRY = (1 << 4), /* Retry possible */ - PERF_TXN_CONFLICT = (1 << 5), /* Conflict abort */ - PERF_TXN_CAPACITY_WRITE = (1 << 6), /* Capacity write abort */ - PERF_TXN_CAPACITY_READ = (1 << 7), /* Capacity read abort */ + PERF_TXN_ELISION = (1 << 0), /* From elision */ + PERF_TXN_TRANSACTION = (1 << 1), /* From transaction */ + PERF_TXN_SYNC = (1 << 2), /* Instruction is related */ + PERF_TXN_ASYNC = (1 << 3), /* Instruction is not related */ + PERF_TXN_RETRY = (1 << 4), /* Retry possible */ + PERF_TXN_CONFLICT = (1 << 5), /* Conflict abort */ + PERF_TXN_CAPACITY_WRITE = (1 << 6), /* Capacity write abort */ + PERF_TXN_CAPACITY_READ = (1 << 7), /* Capacity read abort */ - PERF_TXN_MAX = (1 << 8), /* non-ABI */ + PERF_TXN_MAX = (1 << 8), /* non-ABI */ - /* bits 32..63 are reserved for the abort code */ + /* Bits 32..63 are reserved for the abort code */ - PERF_TXN_ABORT_MASK = (0xffffffffULL << 32), - PERF_TXN_ABORT_SHIFT = 32, + PERF_TXN_ABORT_MASK = (0xffffffffULL << 32), + PERF_TXN_ABORT_SHIFT = 32, }; /* @@ -369,22 +372,22 @@ enum perf_event_read_format { PERF_FORMAT_MAX = 1U << 5, /* non-ABI */ }; -#define PERF_ATTR_SIZE_VER0 64 /* sizeof first published struct */ -#define PERF_ATTR_SIZE_VER1 72 /* add: config2 */ -#define PERF_ATTR_SIZE_VER2 80 /* add: branch_sample_type */ -#define PERF_ATTR_SIZE_VER3 96 /* add: sample_regs_user */ - /* add: sample_stack_user */ -#define PERF_ATTR_SIZE_VER4 104 /* add: sample_regs_intr */ -#define PERF_ATTR_SIZE_VER5 112 /* add: aux_watermark */ -#define PERF_ATTR_SIZE_VER6 120 /* add: aux_sample_size */ -#define PERF_ATTR_SIZE_VER7 128 /* add: sig_data */ -#define PERF_ATTR_SIZE_VER8 136 /* add: config3 */ +#define PERF_ATTR_SIZE_VER0 64 /* Size of first published 'struct perf_event_attr' */ +#define PERF_ATTR_SIZE_VER1 72 /* Add: config2 */ +#define PERF_ATTR_SIZE_VER2 80 /* Add: branch_sample_type */ +#define PERF_ATTR_SIZE_VER3 96 /* Add: sample_regs_user */ + /* Add: sample_stack_user */ +#define PERF_ATTR_SIZE_VER4 104 /* Add: sample_regs_intr */ +#define PERF_ATTR_SIZE_VER5 112 /* Add: aux_watermark */ +#define PERF_ATTR_SIZE_VER6 120 /* Add: aux_sample_size */ +#define PERF_ATTR_SIZE_VER7 128 /* Add: sig_data */ +#define PERF_ATTR_SIZE_VER8 136 /* Add: config3 */ /* - * Hardware event_id to monitor via a performance monitoring event: - * - * @sample_max_stack: Max number of frame pointers in a callchain, - * should be < /proc/sys/kernel/perf_event_max_stack + * 'struct perf_event_attr' contains various attributes that define + * a performance event - most of them hardware related configuration + * details, but also a lot of behavioral switches and values implemented + * by the kernel. */ struct perf_event_attr { @@ -394,7 +397,7 @@ struct perf_event_attr { __u32 type; /* - * Size of the attr structure, for fwd/bwd compat. + * Size of the attr structure, for forward/backwards compatibility. */ __u32 size; @@ -449,21 +452,21 @@ struct perf_event_attr { comm_exec : 1, /* flag comm events that are due to an exec */ use_clockid : 1, /* use @clockid for time fields */ context_switch : 1, /* context switch data */ - write_backward : 1, /* Write ring buffer from end to beginning */ + write_backward : 1, /* write ring buffer from end to beginning */ namespaces : 1, /* include namespaces data */ ksymbol : 1, /* include ksymbol events */ - bpf_event : 1, /* include bpf events */ + bpf_event : 1, /* include BPF events */ aux_output : 1, /* generate AUX records instead of events */ cgroup : 1, /* include cgroup events */ text_poke : 1, /* include text poke events */ - build_id : 1, /* use build id in mmap2 events */ + build_id : 1, /* use build ID in mmap2 events */ inherit_thread : 1, /* children only inherit if cloned with CLONE_THREAD */ remove_on_exec : 1, /* event is removed from task on exec */ sigtrap : 1, /* send synchronous SIGTRAP on event */ __reserved_1 : 26; union { - __u32 wakeup_events; /* wakeup every n events */ + __u32 wakeup_events; /* wake up every n events */ __u32 wakeup_watermark; /* bytes before wakeup */ }; @@ -472,13 +475,13 @@ struct perf_event_attr { __u64 bp_addr; __u64 kprobe_func; /* for perf_kprobe */ __u64 uprobe_path; /* for perf_uprobe */ - __u64 config1; /* extension of config */ + __u64 config1; /* extension of config */ }; union { __u64 bp_len; - __u64 kprobe_addr; /* when kprobe_func == NULL */ + __u64 kprobe_addr; /* when kprobe_func == NULL */ __u64 probe_offset; /* for perf_[k,u]probe */ - __u64 config2; /* extension of config1 */ + __u64 config2; /* extension of config1 */ }; __u64 branch_sample_type; /* enum perf_branch_sample_type */ @@ -508,7 +511,16 @@ struct perf_event_attr { * Wakeup watermark for AUX area */ __u32 aux_watermark; + + /* + * Max number of frame pointers in a callchain, should be + * lower than /proc/sys/kernel/perf_event_max_stack. + * + * Max number of entries of branch stack should be lower + * than the hardware limit. + */ __u16 sample_max_stack; + __u16 __reserved_2; __u32 aux_sample_size; __u32 __reserved_3; @@ -526,7 +538,7 @@ struct perf_event_attr { /* * Structure used by below PERF_EVENT_IOC_QUERY_BPF command - * to query bpf programs attached to the same perf tracepoint + * to query BPF programs attached to the same perf tracepoint * as the given perf event. */ struct perf_event_query_bpf { @@ -548,21 +560,21 @@ struct perf_event_query_bpf { /* * Ioctls that can be done on a perf event fd: */ -#define PERF_EVENT_IOC_ENABLE _IO ('$', 0) -#define PERF_EVENT_IOC_DISABLE _IO ('$', 1) -#define PERF_EVENT_IOC_REFRESH _IO ('$', 2) -#define PERF_EVENT_IOC_RESET _IO ('$', 3) -#define PERF_EVENT_IOC_PERIOD _IOW('$', 4, __u64) -#define PERF_EVENT_IOC_SET_OUTPUT _IO ('$', 5) -#define PERF_EVENT_IOC_SET_FILTER _IOW('$', 6, char *) -#define PERF_EVENT_IOC_ID _IOR('$', 7, __u64 *) -#define PERF_EVENT_IOC_SET_BPF _IOW('$', 8, __u32) -#define PERF_EVENT_IOC_PAUSE_OUTPUT _IOW('$', 9, __u32) +#define PERF_EVENT_IOC_ENABLE _IO ('$', 0) +#define PERF_EVENT_IOC_DISABLE _IO ('$', 1) +#define PERF_EVENT_IOC_REFRESH _IO ('$', 2) +#define PERF_EVENT_IOC_RESET _IO ('$', 3) +#define PERF_EVENT_IOC_PERIOD _IOW ('$', 4, __u64) +#define PERF_EVENT_IOC_SET_OUTPUT _IO ('$', 5) +#define PERF_EVENT_IOC_SET_FILTER _IOW ('$', 6, char *) +#define PERF_EVENT_IOC_ID _IOR ('$', 7, __u64 *) +#define PERF_EVENT_IOC_SET_BPF _IOW ('$', 8, __u32) +#define PERF_EVENT_IOC_PAUSE_OUTPUT _IOW ('$', 9, __u32) #define PERF_EVENT_IOC_QUERY_BPF _IOWR('$', 10, struct perf_event_query_bpf *) -#define PERF_EVENT_IOC_MODIFY_ATTRIBUTES _IOW('$', 11, struct perf_event_attr *) +#define PERF_EVENT_IOC_MODIFY_ATTRIBUTES _IOW ('$', 11, struct perf_event_attr *) enum perf_event_ioc_flags { - PERF_IOC_FLAG_GROUP = 1U << 0, + PERF_IOC_FLAG_GROUP = 1U << 0, }; /* @@ -573,7 +585,7 @@ struct perf_event_mmap_page { __u32 compat_version; /* lowest version this is compat with */ /* - * Bits needed to read the hw events in user-space. + * Bits needed to read the HW events in user-space. * * u32 seq, time_mult, time_shift, index, width; * u64 count, enabled, running; @@ -611,7 +623,7 @@ struct perf_event_mmap_page { __u32 index; /* hardware event identifier */ __s64 offset; /* add to hardware event value */ __u64 time_enabled; /* time event active */ - __u64 time_running; /* time event on cpu */ + __u64 time_running; /* time event on CPU */ union { __u64 capabilities; struct { @@ -639,7 +651,7 @@ struct perf_event_mmap_page { /* * If cap_usr_time the below fields can be used to compute the time - * delta since time_enabled (in ns) using rdtsc or similar. + * delta since time_enabled (in ns) using RDTSC or similar. * * u64 quot, rem; * u64 delta; @@ -712,7 +724,7 @@ struct perf_event_mmap_page { * after reading this value. * * When the mapping is PROT_WRITE the @data_tail value should be - * written by userspace to reflect the last read data, after issueing + * written by user-space to reflect the last read data, after issuing * an smp_mb() to separate the data read from the ->data_tail store. * In this case the kernel will not over-write unread data. * @@ -728,7 +740,7 @@ struct perf_event_mmap_page { /* * AUX area is defined by aux_{offset,size} fields that should be set - * by the userspace, so that + * by the user-space, so that * * aux_offset >= data_offset + data_size * @@ -802,7 +814,7 @@ struct perf_event_mmap_page { * Indicates that thread was preempted in TASK_RUNNING state. * * PERF_RECORD_MISC_MMAP_BUILD_ID: - * Indicates that mmap2 event carries build id data. + * Indicates that mmap2 event carries build ID data. */ #define PERF_RECORD_MISC_EXACT_IP (1 << 14) #define PERF_RECORD_MISC_SWITCH_OUT_PREEMPT (1 << 14) @@ -813,26 +825,26 @@ struct perf_event_mmap_page { #define PERF_RECORD_MISC_EXT_RESERVED (1 << 15) struct perf_event_header { - __u32 type; - __u16 misc; - __u16 size; + __u32 type; + __u16 misc; + __u16 size; }; struct perf_ns_link_info { - __u64 dev; - __u64 ino; + __u64 dev; + __u64 ino; }; enum { - NET_NS_INDEX = 0, - UTS_NS_INDEX = 1, - IPC_NS_INDEX = 2, - PID_NS_INDEX = 3, - USER_NS_INDEX = 4, - MNT_NS_INDEX = 5, - CGROUP_NS_INDEX = 6, - - NR_NAMESPACES, /* number of available namespaces */ + NET_NS_INDEX = 0, + UTS_NS_INDEX = 1, + IPC_NS_INDEX = 2, + PID_NS_INDEX = 3, + USER_NS_INDEX = 4, + MNT_NS_INDEX = 5, + CGROUP_NS_INDEX = 6, + + NR_NAMESPACES, /* number of available namespaces */ }; enum perf_event_type { @@ -848,11 +860,11 @@ enum perf_event_type { * optional fields being ignored. * * struct sample_id { - * { u32 pid, tid; } && PERF_SAMPLE_TID - * { u64 time; } && PERF_SAMPLE_TIME - * { u64 id; } && PERF_SAMPLE_ID - * { u64 stream_id;} && PERF_SAMPLE_STREAM_ID - * { u32 cpu, res; } && PERF_SAMPLE_CPU + * { u32 pid, tid; } && PERF_SAMPLE_TID + * { u64 time; } && PERF_SAMPLE_TIME + * { u64 id; } && PERF_SAMPLE_ID + * { u64 stream_id;} && PERF_SAMPLE_STREAM_ID + * { u32 cpu, res; } && PERF_SAMPLE_CPU * { u64 id; } && PERF_SAMPLE_IDENTIFIER * } && perf_event_attr::sample_id_all * @@ -863,7 +875,7 @@ enum perf_event_type { /* * The MMAP events record the PROT_EXEC mappings so that we can - * correlate userspace IPs to code. They have the following structure: + * correlate user-space IPs to code. They have the following structure: * * struct { * struct perf_event_header header; @@ -873,7 +885,7 @@ enum perf_event_type { * u64 len; * u64 pgoff; * char filename[]; - * struct sample_id sample_id; + * struct sample_id sample_id; * }; */ PERF_RECORD_MMAP = 1, @@ -883,7 +895,7 @@ enum perf_event_type { * struct perf_event_header header; * u64 id; * u64 lost; - * struct sample_id sample_id; + * struct sample_id sample_id; * }; */ PERF_RECORD_LOST = 2, @@ -894,7 +906,7 @@ enum perf_event_type { * * u32 pid, tid; * char comm[]; - * struct sample_id sample_id; + * struct sample_id sample_id; * }; */ PERF_RECORD_COMM = 3, @@ -905,7 +917,7 @@ enum perf_event_type { * u32 pid, ppid; * u32 tid, ptid; * u64 time; - * struct sample_id sample_id; + * struct sample_id sample_id; * }; */ PERF_RECORD_EXIT = 4, @@ -916,7 +928,7 @@ enum perf_event_type { * u64 time; * u64 id; * u64 stream_id; - * struct sample_id sample_id; + * struct sample_id sample_id; * }; */ PERF_RECORD_THROTTLE = 5, @@ -928,7 +940,7 @@ enum perf_event_type { * u32 pid, ppid; * u32 tid, ptid; * u64 time; - * struct sample_id sample_id; + * struct sample_id sample_id; * }; */ PERF_RECORD_FORK = 7, @@ -939,7 +951,7 @@ enum perf_event_type { * u32 pid, tid; * * struct read_format values; - * struct sample_id sample_id; + * struct sample_id sample_id; * }; */ PERF_RECORD_READ = 8, @@ -994,12 +1006,12 @@ enum perf_event_type { * { u64 counters; } cntr[nr] && PERF_SAMPLE_BRANCH_COUNTERS * } && PERF_SAMPLE_BRANCH_STACK * - * { u64 abi; # enum perf_sample_regs_abi - * u64 regs[weight(mask)]; } && PERF_SAMPLE_REGS_USER + * { u64 abi; # enum perf_sample_regs_abi + * u64 regs[weight(mask)]; } && PERF_SAMPLE_REGS_USER * - * { u64 size; - * char data[size]; - * u64 dyn_size; } && PERF_SAMPLE_STACK_USER + * { u64 size; + * char data[size]; + * u64 dyn_size; } && PERF_SAMPLE_STACK_USER * * { union perf_sample_weight * { @@ -1059,7 +1071,7 @@ enum perf_event_type { * }; * u32 prot, flags; * char filename[]; - * struct sample_id sample_id; + * struct sample_id sample_id; * }; */ PERF_RECORD_MMAP2 = 10, @@ -1068,12 +1080,12 @@ enum perf_event_type { * Records that new data landed in the AUX buffer part. * * struct { - * struct perf_event_header header; + * struct perf_event_header header; * - * u64 aux_offset; - * u64 aux_size; + * u64 aux_offset; + * u64 aux_size; * u64 flags; - * struct sample_id sample_id; + * struct sample_id sample_id; * }; */ PERF_RECORD_AUX = 11, @@ -1156,7 +1168,7 @@ enum perf_event_type { PERF_RECORD_KSYMBOL = 17, /* - * Record bpf events: + * Record BPF events: * enum perf_bpf_event_type { * PERF_BPF_EVENT_UNKNOWN = 0, * PERF_BPF_EVENT_PROG_LOAD = 1, @@ -1234,179 +1246,181 @@ enum perf_record_ksymbol_type { #define PERF_RECORD_KSYMBOL_FLAGS_UNREGISTER (1 << 0) enum perf_bpf_event_type { - PERF_BPF_EVENT_UNKNOWN = 0, - PERF_BPF_EVENT_PROG_LOAD = 1, - PERF_BPF_EVENT_PROG_UNLOAD = 2, - PERF_BPF_EVENT_MAX, /* non-ABI */ + PERF_BPF_EVENT_UNKNOWN = 0, + PERF_BPF_EVENT_PROG_LOAD = 1, + PERF_BPF_EVENT_PROG_UNLOAD = 2, + PERF_BPF_EVENT_MAX, /* non-ABI */ }; -#define PERF_MAX_STACK_DEPTH 127 -#define PERF_MAX_CONTEXTS_PER_STACK 8 +#define PERF_MAX_STACK_DEPTH 127 +#define PERF_MAX_CONTEXTS_PER_STACK 8 enum perf_callchain_context { - PERF_CONTEXT_HV = (__u64)-32, - PERF_CONTEXT_KERNEL = (__u64)-128, - PERF_CONTEXT_USER = (__u64)-512, + PERF_CONTEXT_HV = (__u64)-32, + PERF_CONTEXT_KERNEL = (__u64)-128, + PERF_CONTEXT_USER = (__u64)-512, - PERF_CONTEXT_GUEST = (__u64)-2048, - PERF_CONTEXT_GUEST_KERNEL = (__u64)-2176, - PERF_CONTEXT_GUEST_USER = (__u64)-2560, + PERF_CONTEXT_GUEST = (__u64)-2048, + PERF_CONTEXT_GUEST_KERNEL = (__u64)-2176, + PERF_CONTEXT_GUEST_USER = (__u64)-2560, - PERF_CONTEXT_MAX = (__u64)-4095, + PERF_CONTEXT_MAX = (__u64)-4095, }; /** * PERF_RECORD_AUX::flags bits */ -#define PERF_AUX_FLAG_TRUNCATED 0x01 /* record was truncated to fit */ -#define PERF_AUX_FLAG_OVERWRITE 0x02 /* snapshot from overwrite mode */ -#define PERF_AUX_FLAG_PARTIAL 0x04 /* record contains gaps */ -#define PERF_AUX_FLAG_COLLISION 0x08 /* sample collided with another */ +#define PERF_AUX_FLAG_TRUNCATED 0x0001 /* Record was truncated to fit */ +#define PERF_AUX_FLAG_OVERWRITE 0x0002 /* Snapshot from overwrite mode */ +#define PERF_AUX_FLAG_PARTIAL 0x0004 /* Record contains gaps */ +#define PERF_AUX_FLAG_COLLISION 0x0008 /* Sample collided with another */ #define PERF_AUX_FLAG_PMU_FORMAT_TYPE_MASK 0xff00 /* PMU specific trace format type */ /* CoreSight PMU AUX buffer formats */ -#define PERF_AUX_FLAG_CORESIGHT_FORMAT_CORESIGHT 0x0000 /* Default for backward compatibility */ -#define PERF_AUX_FLAG_CORESIGHT_FORMAT_RAW 0x0100 /* Raw format of the source */ +#define PERF_AUX_FLAG_CORESIGHT_FORMAT_CORESIGHT 0x0000 /* Default for backward compatibility */ +#define PERF_AUX_FLAG_CORESIGHT_FORMAT_RAW 0x0100 /* Raw format of the source */ -#define PERF_FLAG_FD_NO_GROUP (1UL << 0) -#define PERF_FLAG_FD_OUTPUT (1UL << 1) -#define PERF_FLAG_PID_CGROUP (1UL << 2) /* pid=cgroup id, per-cpu mode only */ -#define PERF_FLAG_FD_CLOEXEC (1UL << 3) /* O_CLOEXEC */ +#define PERF_FLAG_FD_NO_GROUP (1UL << 0) +#define PERF_FLAG_FD_OUTPUT (1UL << 1) +#define PERF_FLAG_PID_CGROUP (1UL << 2) /* pid=cgroup ID, per-CPU mode only */ +#define PERF_FLAG_FD_CLOEXEC (1UL << 3) /* O_CLOEXEC */ #if defined(__LITTLE_ENDIAN_BITFIELD) union perf_mem_data_src { __u64 val; struct { - __u64 mem_op:5, /* type of opcode */ - mem_lvl:14, /* memory hierarchy level */ - mem_snoop:5, /* snoop mode */ - mem_lock:2, /* lock instr */ - mem_dtlb:7, /* tlb access */ - mem_lvl_num:4, /* memory hierarchy level number */ - mem_remote:1, /* remote */ - mem_snoopx:2, /* snoop mode, ext */ - mem_blk:3, /* access blocked */ - mem_hops:3, /* hop level */ - mem_rsvd:18; + __u64 mem_op : 5, /* Type of opcode */ + mem_lvl : 14, /* Memory hierarchy level */ + mem_snoop : 5, /* Snoop mode */ + mem_lock : 2, /* Lock instr */ + mem_dtlb : 7, /* TLB access */ + mem_lvl_num : 4, /* Memory hierarchy level number */ + mem_remote : 1, /* Remote */ + mem_snoopx : 2, /* Snoop mode, ext */ + mem_blk : 3, /* Access blocked */ + mem_hops : 3, /* Hop level */ + mem_rsvd : 18; }; }; #elif defined(__BIG_ENDIAN_BITFIELD) union perf_mem_data_src { __u64 val; struct { - __u64 mem_rsvd:18, - mem_hops:3, /* hop level */ - mem_blk:3, /* access blocked */ - mem_snoopx:2, /* snoop mode, ext */ - mem_remote:1, /* remote */ - mem_lvl_num:4, /* memory hierarchy level number */ - mem_dtlb:7, /* tlb access */ - mem_lock:2, /* lock instr */ - mem_snoop:5, /* snoop mode */ - mem_lvl:14, /* memory hierarchy level */ - mem_op:5; /* type of opcode */ + __u64 mem_rsvd : 18, + mem_hops : 3, /* Hop level */ + mem_blk : 3, /* Access blocked */ + mem_snoopx : 2, /* Snoop mode, ext */ + mem_remote : 1, /* Remote */ + mem_lvl_num : 4, /* Memory hierarchy level number */ + mem_dtlb : 7, /* TLB access */ + mem_lock : 2, /* Lock instr */ + mem_snoop : 5, /* Snoop mode */ + mem_lvl : 14, /* Memory hierarchy level */ + mem_op : 5; /* Type of opcode */ }; }; #else -#error "Unknown endianness" +# error "Unknown endianness" #endif -/* type of opcode (load/store/prefetch,code) */ -#define PERF_MEM_OP_NA 0x01 /* not available */ -#define PERF_MEM_OP_LOAD 0x02 /* load instruction */ -#define PERF_MEM_OP_STORE 0x04 /* store instruction */ -#define PERF_MEM_OP_PFETCH 0x08 /* prefetch */ -#define PERF_MEM_OP_EXEC 0x10 /* code (execution) */ -#define PERF_MEM_OP_SHIFT 0 +/* Type of memory opcode: */ +#define PERF_MEM_OP_NA 0x0001 /* Not available */ +#define PERF_MEM_OP_LOAD 0x0002 /* Load instruction */ +#define PERF_MEM_OP_STORE 0x0004 /* Store instruction */ +#define PERF_MEM_OP_PFETCH 0x0008 /* Prefetch */ +#define PERF_MEM_OP_EXEC 0x0010 /* Code (execution) */ +#define PERF_MEM_OP_SHIFT 0 /* - * PERF_MEM_LVL_* namespace being depricated to some extent in the + * The PERF_MEM_LVL_* namespace is being deprecated to some extent in * favour of newer composite PERF_MEM_{LVLNUM_,REMOTE_,SNOOPX_} fields. - * Supporting this namespace inorder to not break defined ABIs. + * We support this namespace in order to not break defined ABIs. * - * memory hierarchy (memory level, hit or miss) + * Memory hierarchy (memory level, hit or miss) */ -#define PERF_MEM_LVL_NA 0x01 /* not available */ -#define PERF_MEM_LVL_HIT 0x02 /* hit level */ -#define PERF_MEM_LVL_MISS 0x04 /* miss level */ -#define PERF_MEM_LVL_L1 0x08 /* L1 */ -#define PERF_MEM_LVL_LFB 0x10 /* Line Fill Buffer */ -#define PERF_MEM_LVL_L2 0x20 /* L2 */ -#define PERF_MEM_LVL_L3 0x40 /* L3 */ -#define PERF_MEM_LVL_LOC_RAM 0x80 /* Local DRAM */ -#define PERF_MEM_LVL_REM_RAM1 0x100 /* Remote DRAM (1 hop) */ -#define PERF_MEM_LVL_REM_RAM2 0x200 /* Remote DRAM (2 hops) */ -#define PERF_MEM_LVL_REM_CCE1 0x400 /* Remote Cache (1 hop) */ -#define PERF_MEM_LVL_REM_CCE2 0x800 /* Remote Cache (2 hops) */ -#define PERF_MEM_LVL_IO 0x1000 /* I/O memory */ -#define PERF_MEM_LVL_UNC 0x2000 /* Uncached memory */ -#define PERF_MEM_LVL_SHIFT 5 - -#define PERF_MEM_REMOTE_REMOTE 0x01 /* Remote */ -#define PERF_MEM_REMOTE_SHIFT 37 - -#define PERF_MEM_LVLNUM_L1 0x01 /* L1 */ -#define PERF_MEM_LVLNUM_L2 0x02 /* L2 */ -#define PERF_MEM_LVLNUM_L3 0x03 /* L3 */ -#define PERF_MEM_LVLNUM_L4 0x04 /* L4 */ -/* 5-0x7 available */ -#define PERF_MEM_LVLNUM_UNC 0x08 /* Uncached */ -#define PERF_MEM_LVLNUM_CXL 0x09 /* CXL */ -#define PERF_MEM_LVLNUM_IO 0x0a /* I/O */ -#define PERF_MEM_LVLNUM_ANY_CACHE 0x0b /* Any cache */ -#define PERF_MEM_LVLNUM_LFB 0x0c /* LFB */ -#define PERF_MEM_LVLNUM_RAM 0x0d /* RAM */ -#define PERF_MEM_LVLNUM_PMEM 0x0e /* PMEM */ -#define PERF_MEM_LVLNUM_NA 0x0f /* N/A */ - -#define PERF_MEM_LVLNUM_SHIFT 33 - -/* snoop mode */ -#define PERF_MEM_SNOOP_NA 0x01 /* not available */ -#define PERF_MEM_SNOOP_NONE 0x02 /* no snoop */ -#define PERF_MEM_SNOOP_HIT 0x04 /* snoop hit */ -#define PERF_MEM_SNOOP_MISS 0x08 /* snoop miss */ -#define PERF_MEM_SNOOP_HITM 0x10 /* snoop hit modified */ -#define PERF_MEM_SNOOP_SHIFT 19 - -#define PERF_MEM_SNOOPX_FWD 0x01 /* forward */ -#define PERF_MEM_SNOOPX_PEER 0x02 /* xfer from peer */ -#define PERF_MEM_SNOOPX_SHIFT 38 - -/* locked instruction */ -#define PERF_MEM_LOCK_NA 0x01 /* not available */ -#define PERF_MEM_LOCK_LOCKED 0x02 /* locked transaction */ -#define PERF_MEM_LOCK_SHIFT 24 +#define PERF_MEM_LVL_NA 0x0001 /* Not available */ +#define PERF_MEM_LVL_HIT 0x0002 /* Hit level */ +#define PERF_MEM_LVL_MISS 0x0004 /* Miss level */ +#define PERF_MEM_LVL_L1 0x0008 /* L1 */ +#define PERF_MEM_LVL_LFB 0x0010 /* Line Fill Buffer */ +#define PERF_MEM_LVL_L2 0x0020 /* L2 */ +#define PERF_MEM_LVL_L3 0x0040 /* L3 */ +#define PERF_MEM_LVL_LOC_RAM 0x0080 /* Local DRAM */ +#define PERF_MEM_LVL_REM_RAM1 0x0100 /* Remote DRAM (1 hop) */ +#define PERF_MEM_LVL_REM_RAM2 0x0200 /* Remote DRAM (2 hops) */ +#define PERF_MEM_LVL_REM_CCE1 0x0400 /* Remote Cache (1 hop) */ +#define PERF_MEM_LVL_REM_CCE2 0x0800 /* Remote Cache (2 hops) */ +#define PERF_MEM_LVL_IO 0x1000 /* I/O memory */ +#define PERF_MEM_LVL_UNC 0x2000 /* Uncached memory */ +#define PERF_MEM_LVL_SHIFT 5 + +#define PERF_MEM_REMOTE_REMOTE 0x0001 /* Remote */ +#define PERF_MEM_REMOTE_SHIFT 37 + +#define PERF_MEM_LVLNUM_L1 0x0001 /* L1 */ +#define PERF_MEM_LVLNUM_L2 0x0002 /* L2 */ +#define PERF_MEM_LVLNUM_L3 0x0003 /* L3 */ +#define PERF_MEM_LVLNUM_L4 0x0004 /* L4 */ +#define PERF_MEM_LVLNUM_L2_MHB 0x0005 /* L2 Miss Handling Buffer */ +#define PERF_MEM_LVLNUM_MSC 0x0006 /* Memory-side Cache */ +/* 0x007 available */ +#define PERF_MEM_LVLNUM_UNC 0x0008 /* Uncached */ +#define PERF_MEM_LVLNUM_CXL 0x0009 /* CXL */ +#define PERF_MEM_LVLNUM_IO 0x000a /* I/O */ +#define PERF_MEM_LVLNUM_ANY_CACHE 0x000b /* Any cache */ +#define PERF_MEM_LVLNUM_LFB 0x000c /* LFB / L1 Miss Handling Buffer */ +#define PERF_MEM_LVLNUM_RAM 0x000d /* RAM */ +#define PERF_MEM_LVLNUM_PMEM 0x000e /* PMEM */ +#define PERF_MEM_LVLNUM_NA 0x000f /* N/A */ + +#define PERF_MEM_LVLNUM_SHIFT 33 + +/* Snoop mode */ +#define PERF_MEM_SNOOP_NA 0x0001 /* Not available */ +#define PERF_MEM_SNOOP_NONE 0x0002 /* No snoop */ +#define PERF_MEM_SNOOP_HIT 0x0004 /* Snoop hit */ +#define PERF_MEM_SNOOP_MISS 0x0008 /* Snoop miss */ +#define PERF_MEM_SNOOP_HITM 0x0010 /* Snoop hit modified */ +#define PERF_MEM_SNOOP_SHIFT 19 + +#define PERF_MEM_SNOOPX_FWD 0x0001 /* Forward */ +#define PERF_MEM_SNOOPX_PEER 0x0002 /* Transfer from peer */ +#define PERF_MEM_SNOOPX_SHIFT 38 + +/* Locked instruction */ +#define PERF_MEM_LOCK_NA 0x0001 /* Not available */ +#define PERF_MEM_LOCK_LOCKED 0x0002 /* Locked transaction */ +#define PERF_MEM_LOCK_SHIFT 24 /* TLB access */ -#define PERF_MEM_TLB_NA 0x01 /* not available */ -#define PERF_MEM_TLB_HIT 0x02 /* hit level */ -#define PERF_MEM_TLB_MISS 0x04 /* miss level */ -#define PERF_MEM_TLB_L1 0x08 /* L1 */ -#define PERF_MEM_TLB_L2 0x10 /* L2 */ -#define PERF_MEM_TLB_WK 0x20 /* Hardware Walker*/ -#define PERF_MEM_TLB_OS 0x40 /* OS fault handler */ -#define PERF_MEM_TLB_SHIFT 26 +#define PERF_MEM_TLB_NA 0x0001 /* Not available */ +#define PERF_MEM_TLB_HIT 0x0002 /* Hit level */ +#define PERF_MEM_TLB_MISS 0x0004 /* Miss level */ +#define PERF_MEM_TLB_L1 0x0008 /* L1 */ +#define PERF_MEM_TLB_L2 0x0010 /* L2 */ +#define PERF_MEM_TLB_WK 0x0020 /* Hardware Walker*/ +#define PERF_MEM_TLB_OS 0x0040 /* OS fault handler */ +#define PERF_MEM_TLB_SHIFT 26 /* Access blocked */ -#define PERF_MEM_BLK_NA 0x01 /* not available */ -#define PERF_MEM_BLK_DATA 0x02 /* data could not be forwarded */ -#define PERF_MEM_BLK_ADDR 0x04 /* address conflict */ -#define PERF_MEM_BLK_SHIFT 40 - -/* hop level */ -#define PERF_MEM_HOPS_0 0x01 /* remote core, same node */ -#define PERF_MEM_HOPS_1 0x02 /* remote node, same socket */ -#define PERF_MEM_HOPS_2 0x03 /* remote socket, same board */ -#define PERF_MEM_HOPS_3 0x04 /* remote board */ +#define PERF_MEM_BLK_NA 0x0001 /* Not available */ +#define PERF_MEM_BLK_DATA 0x0002 /* Data could not be forwarded */ +#define PERF_MEM_BLK_ADDR 0x0004 /* Address conflict */ +#define PERF_MEM_BLK_SHIFT 40 + +/* Hop level */ +#define PERF_MEM_HOPS_0 0x0001 /* Remote core, same node */ +#define PERF_MEM_HOPS_1 0x0002 /* Remote node, same socket */ +#define PERF_MEM_HOPS_2 0x0003 /* Remote socket, same board */ +#define PERF_MEM_HOPS_3 0x0004 /* Remote board */ /* 5-7 available */ -#define PERF_MEM_HOPS_SHIFT 43 +#define PERF_MEM_HOPS_SHIFT 43 #define PERF_MEM_S(a, s) \ (((__u64)PERF_MEM_##a##_##s) << PERF_MEM_##a##_SHIFT) /* - * single taken branch record layout: + * Layout of single taken branch records: * * from: source instruction (may not always be a branch insn) * to: branch target @@ -1425,37 +1439,37 @@ union perf_mem_data_src { struct perf_branch_entry { __u64 from; __u64 to; - __u64 mispred:1, /* target mispredicted */ - predicted:1,/* target predicted */ - in_tx:1, /* in transaction */ - abort:1, /* transaction abort */ - cycles:16, /* cycle count to last branch */ - type:4, /* branch type */ - spec:2, /* branch speculation info */ - new_type:4, /* additional branch type */ - priv:3, /* privilege level */ - reserved:31; + __u64 mispred : 1, /* target mispredicted */ + predicted : 1, /* target predicted */ + in_tx : 1, /* in transaction */ + abort : 1, /* transaction abort */ + cycles : 16, /* cycle count to last branch */ + type : 4, /* branch type */ + spec : 2, /* branch speculation info */ + new_type : 4, /* additional branch type */ + priv : 3, /* privilege level */ + reserved : 31; }; /* Size of used info bits in struct perf_branch_entry */ #define PERF_BRANCH_ENTRY_INFO_BITS_MAX 33 union perf_sample_weight { - __u64 full; + __u64 full; #if defined(__LITTLE_ENDIAN_BITFIELD) struct { - __u32 var1_dw; - __u16 var2_w; - __u16 var3_w; + __u32 var1_dw; + __u16 var2_w; + __u16 var3_w; }; #elif defined(__BIG_ENDIAN_BITFIELD) struct { - __u16 var3_w; - __u16 var2_w; - __u32 var1_dw; + __u16 var3_w; + __u16 var2_w; + __u32 var1_dw; }; #else -#error "Unknown endianness" +# error "Unknown endianness" #endif }; diff --git a/tools/include/uapi/linux/perf_event.h b/tools/include/uapi/linux/perf_event.h index 3a64499b0f5d..365184d931ad 100644 --- a/tools/include/uapi/linux/perf_event.h +++ b/tools/include/uapi/linux/perf_event.h @@ -39,18 +39,21 @@ enum perf_type_id { /* * attr.config layout for type PERF_TYPE_HARDWARE and PERF_TYPE_HW_CACHE + * * PERF_TYPE_HARDWARE: 0xEEEEEEEE000000AA * AA: hardware event ID * EEEEEEEE: PMU type ID + * * PERF_TYPE_HW_CACHE: 0xEEEEEEEE00DDCCBB * BB: hardware cache ID * CC: hardware cache op ID * DD: hardware cache op result ID * EEEEEEEE: PMU type ID - * If the PMU type ID is 0, the PERF_TYPE_RAW will be applied. + * + * If the PMU type ID is 0, PERF_TYPE_RAW will be applied. */ -#define PERF_PMU_TYPE_SHIFT 32 -#define PERF_HW_EVENT_MASK 0xffffffff +#define PERF_PMU_TYPE_SHIFT 32 +#define PERF_HW_EVENT_MASK 0xffffffff /* * Generalized performance event event_id types, used by the @@ -112,7 +115,7 @@ enum perf_hw_cache_op_result_id { /* * Special "software" events provided by the kernel, even if the hardware * does not support performance events. These events measure various - * physical and sw events of the kernel (and allow the profiling of them as + * physical and SW events of the kernel (and allow the profiling of them as * well): */ enum perf_sw_ids { @@ -167,8 +170,9 @@ enum perf_event_sample_format { }; #define PERF_SAMPLE_WEIGHT_TYPE (PERF_SAMPLE_WEIGHT | PERF_SAMPLE_WEIGHT_STRUCT) + /* - * values to program into branch_sample_type when PERF_SAMPLE_BRANCH is set + * Values to program into branch_sample_type when PERF_SAMPLE_BRANCH is set. * * If the user does not pass priv level information via branch_sample_type, * the kernel uses the event's priv level. Branch and event priv levels do @@ -178,20 +182,20 @@ enum perf_event_sample_format { * of branches and therefore it supersedes all the other types. */ enum perf_branch_sample_type_shift { - PERF_SAMPLE_BRANCH_USER_SHIFT = 0, /* user branches */ - PERF_SAMPLE_BRANCH_KERNEL_SHIFT = 1, /* kernel branches */ - PERF_SAMPLE_BRANCH_HV_SHIFT = 2, /* hypervisor branches */ - - PERF_SAMPLE_BRANCH_ANY_SHIFT = 3, /* any branch types */ - PERF_SAMPLE_BRANCH_ANY_CALL_SHIFT = 4, /* any call branch */ - PERF_SAMPLE_BRANCH_ANY_RETURN_SHIFT = 5, /* any return branch */ - PERF_SAMPLE_BRANCH_IND_CALL_SHIFT = 6, /* indirect calls */ - PERF_SAMPLE_BRANCH_ABORT_TX_SHIFT = 7, /* transaction aborts */ - PERF_SAMPLE_BRANCH_IN_TX_SHIFT = 8, /* in transaction */ - PERF_SAMPLE_BRANCH_NO_TX_SHIFT = 9, /* not in transaction */ + PERF_SAMPLE_BRANCH_USER_SHIFT = 0, /* user branches */ + PERF_SAMPLE_BRANCH_KERNEL_SHIFT = 1, /* kernel branches */ + PERF_SAMPLE_BRANCH_HV_SHIFT = 2, /* hypervisor branches */ + + PERF_SAMPLE_BRANCH_ANY_SHIFT = 3, /* any branch types */ + PERF_SAMPLE_BRANCH_ANY_CALL_SHIFT = 4, /* any call branch */ + PERF_SAMPLE_BRANCH_ANY_RETURN_SHIFT = 5, /* any return branch */ + PERF_SAMPLE_BRANCH_IND_CALL_SHIFT = 6, /* indirect calls */ + PERF_SAMPLE_BRANCH_ABORT_TX_SHIFT = 7, /* transaction aborts */ + PERF_SAMPLE_BRANCH_IN_TX_SHIFT = 8, /* in transaction */ + PERF_SAMPLE_BRANCH_NO_TX_SHIFT = 9, /* not in transaction */ PERF_SAMPLE_BRANCH_COND_SHIFT = 10, /* conditional branches */ - PERF_SAMPLE_BRANCH_CALL_STACK_SHIFT = 11, /* call/ret stack */ + PERF_SAMPLE_BRANCH_CALL_STACK_SHIFT = 11, /* CALL/RET stack */ PERF_SAMPLE_BRANCH_IND_JUMP_SHIFT = 12, /* indirect jumps */ PERF_SAMPLE_BRANCH_CALL_SHIFT = 13, /* direct call */ @@ -210,96 +214,95 @@ enum perf_branch_sample_type_shift { }; enum perf_branch_sample_type { - PERF_SAMPLE_BRANCH_USER = 1U << PERF_SAMPLE_BRANCH_USER_SHIFT, - PERF_SAMPLE_BRANCH_KERNEL = 1U << PERF_SAMPLE_BRANCH_KERNEL_SHIFT, - PERF_SAMPLE_BRANCH_HV = 1U << PERF_SAMPLE_BRANCH_HV_SHIFT, + PERF_SAMPLE_BRANCH_USER = 1U << PERF_SAMPLE_BRANCH_USER_SHIFT, + PERF_SAMPLE_BRANCH_KERNEL = 1U << PERF_SAMPLE_BRANCH_KERNEL_SHIFT, + PERF_SAMPLE_BRANCH_HV = 1U << PERF_SAMPLE_BRANCH_HV_SHIFT, - PERF_SAMPLE_BRANCH_ANY = 1U << PERF_SAMPLE_BRANCH_ANY_SHIFT, - PERF_SAMPLE_BRANCH_ANY_CALL = 1U << PERF_SAMPLE_BRANCH_ANY_CALL_SHIFT, - PERF_SAMPLE_BRANCH_ANY_RETURN = 1U << PERF_SAMPLE_BRANCH_ANY_RETURN_SHIFT, - PERF_SAMPLE_BRANCH_IND_CALL = 1U << PERF_SAMPLE_BRANCH_IND_CALL_SHIFT, - PERF_SAMPLE_BRANCH_ABORT_TX = 1U << PERF_SAMPLE_BRANCH_ABORT_TX_SHIFT, - PERF_SAMPLE_BRANCH_IN_TX = 1U << PERF_SAMPLE_BRANCH_IN_TX_SHIFT, - PERF_SAMPLE_BRANCH_NO_TX = 1U << PERF_SAMPLE_BRANCH_NO_TX_SHIFT, - PERF_SAMPLE_BRANCH_COND = 1U << PERF_SAMPLE_BRANCH_COND_SHIFT, + PERF_SAMPLE_BRANCH_ANY = 1U << PERF_SAMPLE_BRANCH_ANY_SHIFT, + PERF_SAMPLE_BRANCH_ANY_CALL = 1U << PERF_SAMPLE_BRANCH_ANY_CALL_SHIFT, + PERF_SAMPLE_BRANCH_ANY_RETURN = 1U << PERF_SAMPLE_BRANCH_ANY_RETURN_SHIFT, + PERF_SAMPLE_BRANCH_IND_CALL = 1U << PERF_SAMPLE_BRANCH_IND_CALL_SHIFT, + PERF_SAMPLE_BRANCH_ABORT_TX = 1U << PERF_SAMPLE_BRANCH_ABORT_TX_SHIFT, + PERF_SAMPLE_BRANCH_IN_TX = 1U << PERF_SAMPLE_BRANCH_IN_TX_SHIFT, + PERF_SAMPLE_BRANCH_NO_TX = 1U << PERF_SAMPLE_BRANCH_NO_TX_SHIFT, + PERF_SAMPLE_BRANCH_COND = 1U << PERF_SAMPLE_BRANCH_COND_SHIFT, - PERF_SAMPLE_BRANCH_CALL_STACK = 1U << PERF_SAMPLE_BRANCH_CALL_STACK_SHIFT, - PERF_SAMPLE_BRANCH_IND_JUMP = 1U << PERF_SAMPLE_BRANCH_IND_JUMP_SHIFT, - PERF_SAMPLE_BRANCH_CALL = 1U << PERF_SAMPLE_BRANCH_CALL_SHIFT, + PERF_SAMPLE_BRANCH_CALL_STACK = 1U << PERF_SAMPLE_BRANCH_CALL_STACK_SHIFT, + PERF_SAMPLE_BRANCH_IND_JUMP = 1U << PERF_SAMPLE_BRANCH_IND_JUMP_SHIFT, + PERF_SAMPLE_BRANCH_CALL = 1U << PERF_SAMPLE_BRANCH_CALL_SHIFT, - PERF_SAMPLE_BRANCH_NO_FLAGS = 1U << PERF_SAMPLE_BRANCH_NO_FLAGS_SHIFT, - PERF_SAMPLE_BRANCH_NO_CYCLES = 1U << PERF_SAMPLE_BRANCH_NO_CYCLES_SHIFT, + PERF_SAMPLE_BRANCH_NO_FLAGS = 1U << PERF_SAMPLE_BRANCH_NO_FLAGS_SHIFT, + PERF_SAMPLE_BRANCH_NO_CYCLES = 1U << PERF_SAMPLE_BRANCH_NO_CYCLES_SHIFT, - PERF_SAMPLE_BRANCH_TYPE_SAVE = - 1U << PERF_SAMPLE_BRANCH_TYPE_SAVE_SHIFT, + PERF_SAMPLE_BRANCH_TYPE_SAVE = 1U << PERF_SAMPLE_BRANCH_TYPE_SAVE_SHIFT, - PERF_SAMPLE_BRANCH_HW_INDEX = 1U << PERF_SAMPLE_BRANCH_HW_INDEX_SHIFT, + PERF_SAMPLE_BRANCH_HW_INDEX = 1U << PERF_SAMPLE_BRANCH_HW_INDEX_SHIFT, - PERF_SAMPLE_BRANCH_PRIV_SAVE = 1U << PERF_SAMPLE_BRANCH_PRIV_SAVE_SHIFT, + PERF_SAMPLE_BRANCH_PRIV_SAVE = 1U << PERF_SAMPLE_BRANCH_PRIV_SAVE_SHIFT, - PERF_SAMPLE_BRANCH_COUNTERS = 1U << PERF_SAMPLE_BRANCH_COUNTERS_SHIFT, + PERF_SAMPLE_BRANCH_COUNTERS = 1U << PERF_SAMPLE_BRANCH_COUNTERS_SHIFT, - PERF_SAMPLE_BRANCH_MAX = 1U << PERF_SAMPLE_BRANCH_MAX_SHIFT, + PERF_SAMPLE_BRANCH_MAX = 1U << PERF_SAMPLE_BRANCH_MAX_SHIFT, }; /* - * Common flow change classification + * Common control flow change classifications: */ enum { - PERF_BR_UNKNOWN = 0, /* unknown */ - PERF_BR_COND = 1, /* conditional */ - PERF_BR_UNCOND = 2, /* unconditional */ - PERF_BR_IND = 3, /* indirect */ - PERF_BR_CALL = 4, /* function call */ - PERF_BR_IND_CALL = 5, /* indirect function call */ - PERF_BR_RET = 6, /* function return */ - PERF_BR_SYSCALL = 7, /* syscall */ - PERF_BR_SYSRET = 8, /* syscall return */ - PERF_BR_COND_CALL = 9, /* conditional function call */ - PERF_BR_COND_RET = 10, /* conditional function return */ - PERF_BR_ERET = 11, /* exception return */ - PERF_BR_IRQ = 12, /* irq */ - PERF_BR_SERROR = 13, /* system error */ - PERF_BR_NO_TX = 14, /* not in transaction */ - PERF_BR_EXTEND_ABI = 15, /* extend ABI */ + PERF_BR_UNKNOWN = 0, /* Unknown */ + PERF_BR_COND = 1, /* Conditional */ + PERF_BR_UNCOND = 2, /* Unconditional */ + PERF_BR_IND = 3, /* Indirect */ + PERF_BR_CALL = 4, /* Function call */ + PERF_BR_IND_CALL = 5, /* Indirect function call */ + PERF_BR_RET = 6, /* Function return */ + PERF_BR_SYSCALL = 7, /* Syscall */ + PERF_BR_SYSRET = 8, /* Syscall return */ + PERF_BR_COND_CALL = 9, /* Conditional function call */ + PERF_BR_COND_RET = 10, /* Conditional function return */ + PERF_BR_ERET = 11, /* Exception return */ + PERF_BR_IRQ = 12, /* IRQ */ + PERF_BR_SERROR = 13, /* System error */ + PERF_BR_NO_TX = 14, /* Not in transaction */ + PERF_BR_EXTEND_ABI = 15, /* Extend ABI */ PERF_BR_MAX, }; /* - * Common branch speculation outcome classification + * Common branch speculation outcome classifications: */ enum { - PERF_BR_SPEC_NA = 0, /* Not available */ - PERF_BR_SPEC_WRONG_PATH = 1, /* Speculative but on wrong path */ - PERF_BR_NON_SPEC_CORRECT_PATH = 2, /* Non-speculative but on correct path */ - PERF_BR_SPEC_CORRECT_PATH = 3, /* Speculative and on correct path */ + PERF_BR_SPEC_NA = 0, /* Not available */ + PERF_BR_SPEC_WRONG_PATH = 1, /* Speculative but on wrong path */ + PERF_BR_NON_SPEC_CORRECT_PATH = 2, /* Non-speculative but on correct path */ + PERF_BR_SPEC_CORRECT_PATH = 3, /* Speculative and on correct path */ PERF_BR_SPEC_MAX, }; enum { - PERF_BR_NEW_FAULT_ALGN = 0, /* Alignment fault */ - PERF_BR_NEW_FAULT_DATA = 1, /* Data fault */ - PERF_BR_NEW_FAULT_INST = 2, /* Inst fault */ - PERF_BR_NEW_ARCH_1 = 3, /* Architecture specific */ - PERF_BR_NEW_ARCH_2 = 4, /* Architecture specific */ - PERF_BR_NEW_ARCH_3 = 5, /* Architecture specific */ - PERF_BR_NEW_ARCH_4 = 6, /* Architecture specific */ - PERF_BR_NEW_ARCH_5 = 7, /* Architecture specific */ + PERF_BR_NEW_FAULT_ALGN = 0, /* Alignment fault */ + PERF_BR_NEW_FAULT_DATA = 1, /* Data fault */ + PERF_BR_NEW_FAULT_INST = 2, /* Inst fault */ + PERF_BR_NEW_ARCH_1 = 3, /* Architecture specific */ + PERF_BR_NEW_ARCH_2 = 4, /* Architecture specific */ + PERF_BR_NEW_ARCH_3 = 5, /* Architecture specific */ + PERF_BR_NEW_ARCH_4 = 6, /* Architecture specific */ + PERF_BR_NEW_ARCH_5 = 7, /* Architecture specific */ PERF_BR_NEW_MAX, }; enum { - PERF_BR_PRIV_UNKNOWN = 0, - PERF_BR_PRIV_USER = 1, - PERF_BR_PRIV_KERNEL = 2, - PERF_BR_PRIV_HV = 3, + PERF_BR_PRIV_UNKNOWN = 0, + PERF_BR_PRIV_USER = 1, + PERF_BR_PRIV_KERNEL = 2, + PERF_BR_PRIV_HV = 3, }; -#define PERF_BR_ARM64_FIQ PERF_BR_NEW_ARCH_1 -#define PERF_BR_ARM64_DEBUG_HALT PERF_BR_NEW_ARCH_2 -#define PERF_BR_ARM64_DEBUG_EXIT PERF_BR_NEW_ARCH_3 -#define PERF_BR_ARM64_DEBUG_INST PERF_BR_NEW_ARCH_4 -#define PERF_BR_ARM64_DEBUG_DATA PERF_BR_NEW_ARCH_5 +#define PERF_BR_ARM64_FIQ PERF_BR_NEW_ARCH_1 +#define PERF_BR_ARM64_DEBUG_HALT PERF_BR_NEW_ARCH_2 +#define PERF_BR_ARM64_DEBUG_EXIT PERF_BR_NEW_ARCH_3 +#define PERF_BR_ARM64_DEBUG_INST PERF_BR_NEW_ARCH_4 +#define PERF_BR_ARM64_DEBUG_DATA PERF_BR_NEW_ARCH_5 #define PERF_SAMPLE_BRANCH_PLM_ALL \ (PERF_SAMPLE_BRANCH_USER|\ @@ -310,9 +313,9 @@ enum { * Values to determine ABI of the registers dump. */ enum perf_sample_regs_abi { - PERF_SAMPLE_REGS_ABI_NONE = 0, - PERF_SAMPLE_REGS_ABI_32 = 1, - PERF_SAMPLE_REGS_ABI_64 = 2, + PERF_SAMPLE_REGS_ABI_NONE = 0, + PERF_SAMPLE_REGS_ABI_32 = 1, + PERF_SAMPLE_REGS_ABI_64 = 2, }; /* @@ -320,21 +323,21 @@ enum perf_sample_regs_abi { * abort events. Multiple bits can be set. */ enum { - PERF_TXN_ELISION = (1 << 0), /* From elision */ - PERF_TXN_TRANSACTION = (1 << 1), /* From transaction */ - PERF_TXN_SYNC = (1 << 2), /* Instruction is related */ - PERF_TXN_ASYNC = (1 << 3), /* Instruction not related */ - PERF_TXN_RETRY = (1 << 4), /* Retry possible */ - PERF_TXN_CONFLICT = (1 << 5), /* Conflict abort */ - PERF_TXN_CAPACITY_WRITE = (1 << 6), /* Capacity write abort */ - PERF_TXN_CAPACITY_READ = (1 << 7), /* Capacity read abort */ + PERF_TXN_ELISION = (1 << 0), /* From elision */ + PERF_TXN_TRANSACTION = (1 << 1), /* From transaction */ + PERF_TXN_SYNC = (1 << 2), /* Instruction is related */ + PERF_TXN_ASYNC = (1 << 3), /* Instruction is not related */ + PERF_TXN_RETRY = (1 << 4), /* Retry possible */ + PERF_TXN_CONFLICT = (1 << 5), /* Conflict abort */ + PERF_TXN_CAPACITY_WRITE = (1 << 6), /* Capacity write abort */ + PERF_TXN_CAPACITY_READ = (1 << 7), /* Capacity read abort */ - PERF_TXN_MAX = (1 << 8), /* non-ABI */ + PERF_TXN_MAX = (1 << 8), /* non-ABI */ - /* bits 32..63 are reserved for the abort code */ + /* Bits 32..63 are reserved for the abort code */ - PERF_TXN_ABORT_MASK = (0xffffffffULL << 32), - PERF_TXN_ABORT_SHIFT = 32, + PERF_TXN_ABORT_MASK = (0xffffffffULL << 32), + PERF_TXN_ABORT_SHIFT = 32, }; /* @@ -369,22 +372,22 @@ enum perf_event_read_format { PERF_FORMAT_MAX = 1U << 5, /* non-ABI */ }; -#define PERF_ATTR_SIZE_VER0 64 /* sizeof first published struct */ -#define PERF_ATTR_SIZE_VER1 72 /* add: config2 */ -#define PERF_ATTR_SIZE_VER2 80 /* add: branch_sample_type */ -#define PERF_ATTR_SIZE_VER3 96 /* add: sample_regs_user */ - /* add: sample_stack_user */ -#define PERF_ATTR_SIZE_VER4 104 /* add: sample_regs_intr */ -#define PERF_ATTR_SIZE_VER5 112 /* add: aux_watermark */ -#define PERF_ATTR_SIZE_VER6 120 /* add: aux_sample_size */ -#define PERF_ATTR_SIZE_VER7 128 /* add: sig_data */ -#define PERF_ATTR_SIZE_VER8 136 /* add: config3 */ +#define PERF_ATTR_SIZE_VER0 64 /* Size of first published 'struct perf_event_attr' */ +#define PERF_ATTR_SIZE_VER1 72 /* Add: config2 */ +#define PERF_ATTR_SIZE_VER2 80 /* Add: branch_sample_type */ +#define PERF_ATTR_SIZE_VER3 96 /* Add: sample_regs_user */ + /* Add: sample_stack_user */ +#define PERF_ATTR_SIZE_VER4 104 /* Add: sample_regs_intr */ +#define PERF_ATTR_SIZE_VER5 112 /* Add: aux_watermark */ +#define PERF_ATTR_SIZE_VER6 120 /* Add: aux_sample_size */ +#define PERF_ATTR_SIZE_VER7 128 /* Add: sig_data */ +#define PERF_ATTR_SIZE_VER8 136 /* Add: config3 */ /* - * Hardware event_id to monitor via a performance monitoring event: - * - * @sample_max_stack: Max number of frame pointers in a callchain, - * should be < /proc/sys/kernel/perf_event_max_stack + * 'struct perf_event_attr' contains various attributes that define + * a performance event - most of them hardware related configuration + * details, but also a lot of behavioral switches and values implemented + * by the kernel. */ struct perf_event_attr { @@ -394,7 +397,7 @@ struct perf_event_attr { __u32 type; /* - * Size of the attr structure, for fwd/bwd compat. + * Size of the attr structure, for forward/backwards compatibility. */ __u32 size; @@ -449,21 +452,21 @@ struct perf_event_attr { comm_exec : 1, /* flag comm events that are due to an exec */ use_clockid : 1, /* use @clockid for time fields */ context_switch : 1, /* context switch data */ - write_backward : 1, /* Write ring buffer from end to beginning */ + write_backward : 1, /* write ring buffer from end to beginning */ namespaces : 1, /* include namespaces data */ ksymbol : 1, /* include ksymbol events */ - bpf_event : 1, /* include bpf events */ + bpf_event : 1, /* include BPF events */ aux_output : 1, /* generate AUX records instead of events */ cgroup : 1, /* include cgroup events */ text_poke : 1, /* include text poke events */ - build_id : 1, /* use build id in mmap2 events */ + build_id : 1, /* use build ID in mmap2 events */ inherit_thread : 1, /* children only inherit if cloned with CLONE_THREAD */ remove_on_exec : 1, /* event is removed from task on exec */ sigtrap : 1, /* send synchronous SIGTRAP on event */ __reserved_1 : 26; union { - __u32 wakeup_events; /* wakeup every n events */ + __u32 wakeup_events; /* wake up every n events */ __u32 wakeup_watermark; /* bytes before wakeup */ }; @@ -472,13 +475,13 @@ struct perf_event_attr { __u64 bp_addr; __u64 kprobe_func; /* for perf_kprobe */ __u64 uprobe_path; /* for perf_uprobe */ - __u64 config1; /* extension of config */ + __u64 config1; /* extension of config */ }; union { __u64 bp_len; - __u64 kprobe_addr; /* when kprobe_func == NULL */ + __u64 kprobe_addr; /* when kprobe_func == NULL */ __u64 probe_offset; /* for perf_[k,u]probe */ - __u64 config2; /* extension of config1 */ + __u64 config2; /* extension of config1 */ }; __u64 branch_sample_type; /* enum perf_branch_sample_type */ @@ -508,7 +511,16 @@ struct perf_event_attr { * Wakeup watermark for AUX area */ __u32 aux_watermark; + + /* + * Max number of frame pointers in a callchain, should be + * lower than /proc/sys/kernel/perf_event_max_stack. + * + * Max number of entries of branch stack should be lower + * than the hardware limit. + */ __u16 sample_max_stack; + __u16 __reserved_2; __u32 aux_sample_size; __u32 __reserved_3; @@ -526,7 +538,7 @@ struct perf_event_attr { /* * Structure used by below PERF_EVENT_IOC_QUERY_BPF command - * to query bpf programs attached to the same perf tracepoint + * to query BPF programs attached to the same perf tracepoint * as the given perf event. */ struct perf_event_query_bpf { @@ -548,21 +560,21 @@ struct perf_event_query_bpf { /* * Ioctls that can be done on a perf event fd: */ -#define PERF_EVENT_IOC_ENABLE _IO ('$', 0) -#define PERF_EVENT_IOC_DISABLE _IO ('$', 1) -#define PERF_EVENT_IOC_REFRESH _IO ('$', 2) -#define PERF_EVENT_IOC_RESET _IO ('$', 3) -#define PERF_EVENT_IOC_PERIOD _IOW('$', 4, __u64) -#define PERF_EVENT_IOC_SET_OUTPUT _IO ('$', 5) -#define PERF_EVENT_IOC_SET_FILTER _IOW('$', 6, char *) -#define PERF_EVENT_IOC_ID _IOR('$', 7, __u64 *) -#define PERF_EVENT_IOC_SET_BPF _IOW('$', 8, __u32) -#define PERF_EVENT_IOC_PAUSE_OUTPUT _IOW('$', 9, __u32) +#define PERF_EVENT_IOC_ENABLE _IO ('$', 0) +#define PERF_EVENT_IOC_DISABLE _IO ('$', 1) +#define PERF_EVENT_IOC_REFRESH _IO ('$', 2) +#define PERF_EVENT_IOC_RESET _IO ('$', 3) +#define PERF_EVENT_IOC_PERIOD _IOW ('$', 4, __u64) +#define PERF_EVENT_IOC_SET_OUTPUT _IO ('$', 5) +#define PERF_EVENT_IOC_SET_FILTER _IOW ('$', 6, char *) +#define PERF_EVENT_IOC_ID _IOR ('$', 7, __u64 *) +#define PERF_EVENT_IOC_SET_BPF _IOW ('$', 8, __u32) +#define PERF_EVENT_IOC_PAUSE_OUTPUT _IOW ('$', 9, __u32) #define PERF_EVENT_IOC_QUERY_BPF _IOWR('$', 10, struct perf_event_query_bpf *) -#define PERF_EVENT_IOC_MODIFY_ATTRIBUTES _IOW('$', 11, struct perf_event_attr *) +#define PERF_EVENT_IOC_MODIFY_ATTRIBUTES _IOW ('$', 11, struct perf_event_attr *) enum perf_event_ioc_flags { - PERF_IOC_FLAG_GROUP = 1U << 0, + PERF_IOC_FLAG_GROUP = 1U << 0, }; /* @@ -573,7 +585,7 @@ struct perf_event_mmap_page { __u32 compat_version; /* lowest version this is compat with */ /* - * Bits needed to read the hw events in user-space. + * Bits needed to read the HW events in user-space. * * u32 seq, time_mult, time_shift, index, width; * u64 count, enabled, running; @@ -611,7 +623,7 @@ struct perf_event_mmap_page { __u32 index; /* hardware event identifier */ __s64 offset; /* add to hardware event value */ __u64 time_enabled; /* time event active */ - __u64 time_running; /* time event on cpu */ + __u64 time_running; /* time event on CPU */ union { __u64 capabilities; struct { @@ -639,7 +651,7 @@ struct perf_event_mmap_page { /* * If cap_usr_time the below fields can be used to compute the time - * delta since time_enabled (in ns) using rdtsc or similar. + * delta since time_enabled (in ns) using RDTSC or similar. * * u64 quot, rem; * u64 delta; @@ -712,7 +724,7 @@ struct perf_event_mmap_page { * after reading this value. * * When the mapping is PROT_WRITE the @data_tail value should be - * written by userspace to reflect the last read data, after issueing + * written by user-space to reflect the last read data, after issuing * an smp_mb() to separate the data read from the ->data_tail store. * In this case the kernel will not over-write unread data. * @@ -728,7 +740,7 @@ struct perf_event_mmap_page { /* * AUX area is defined by aux_{offset,size} fields that should be set - * by the userspace, so that + * by the user-space, so that * * aux_offset >= data_offset + data_size * @@ -802,7 +814,7 @@ struct perf_event_mmap_page { * Indicates that thread was preempted in TASK_RUNNING state. * * PERF_RECORD_MISC_MMAP_BUILD_ID: - * Indicates that mmap2 event carries build id data. + * Indicates that mmap2 event carries build ID data. */ #define PERF_RECORD_MISC_EXACT_IP (1 << 14) #define PERF_RECORD_MISC_SWITCH_OUT_PREEMPT (1 << 14) @@ -813,26 +825,26 @@ struct perf_event_mmap_page { #define PERF_RECORD_MISC_EXT_RESERVED (1 << 15) struct perf_event_header { - __u32 type; - __u16 misc; - __u16 size; + __u32 type; + __u16 misc; + __u16 size; }; struct perf_ns_link_info { - __u64 dev; - __u64 ino; + __u64 dev; + __u64 ino; }; enum { - NET_NS_INDEX = 0, - UTS_NS_INDEX = 1, - IPC_NS_INDEX = 2, - PID_NS_INDEX = 3, - USER_NS_INDEX = 4, - MNT_NS_INDEX = 5, - CGROUP_NS_INDEX = 6, - - NR_NAMESPACES, /* number of available namespaces */ + NET_NS_INDEX = 0, + UTS_NS_INDEX = 1, + IPC_NS_INDEX = 2, + PID_NS_INDEX = 3, + USER_NS_INDEX = 4, + MNT_NS_INDEX = 5, + CGROUP_NS_INDEX = 6, + + NR_NAMESPACES, /* number of available namespaces */ }; enum perf_event_type { @@ -848,11 +860,11 @@ enum perf_event_type { * optional fields being ignored. * * struct sample_id { - * { u32 pid, tid; } && PERF_SAMPLE_TID - * { u64 time; } && PERF_SAMPLE_TIME - * { u64 id; } && PERF_SAMPLE_ID - * { u64 stream_id;} && PERF_SAMPLE_STREAM_ID - * { u32 cpu, res; } && PERF_SAMPLE_CPU + * { u32 pid, tid; } && PERF_SAMPLE_TID + * { u64 time; } && PERF_SAMPLE_TIME + * { u64 id; } && PERF_SAMPLE_ID + * { u64 stream_id;} && PERF_SAMPLE_STREAM_ID + * { u32 cpu, res; } && PERF_SAMPLE_CPU * { u64 id; } && PERF_SAMPLE_IDENTIFIER * } && perf_event_attr::sample_id_all * @@ -863,7 +875,7 @@ enum perf_event_type { /* * The MMAP events record the PROT_EXEC mappings so that we can - * correlate userspace IPs to code. They have the following structure: + * correlate user-space IPs to code. They have the following structure: * * struct { * struct perf_event_header header; @@ -873,7 +885,7 @@ enum perf_event_type { * u64 len; * u64 pgoff; * char filename[]; - * struct sample_id sample_id; + * struct sample_id sample_id; * }; */ PERF_RECORD_MMAP = 1, @@ -883,7 +895,7 @@ enum perf_event_type { * struct perf_event_header header; * u64 id; * u64 lost; - * struct sample_id sample_id; + * struct sample_id sample_id; * }; */ PERF_RECORD_LOST = 2, @@ -894,7 +906,7 @@ enum perf_event_type { * * u32 pid, tid; * char comm[]; - * struct sample_id sample_id; + * struct sample_id sample_id; * }; */ PERF_RECORD_COMM = 3, @@ -905,7 +917,7 @@ enum perf_event_type { * u32 pid, ppid; * u32 tid, ptid; * u64 time; - * struct sample_id sample_id; + * struct sample_id sample_id; * }; */ PERF_RECORD_EXIT = 4, @@ -916,7 +928,7 @@ enum perf_event_type { * u64 time; * u64 id; * u64 stream_id; - * struct sample_id sample_id; + * struct sample_id sample_id; * }; */ PERF_RECORD_THROTTLE = 5, @@ -928,7 +940,7 @@ enum perf_event_type { * u32 pid, ppid; * u32 tid, ptid; * u64 time; - * struct sample_id sample_id; + * struct sample_id sample_id; * }; */ PERF_RECORD_FORK = 7, @@ -939,7 +951,7 @@ enum perf_event_type { * u32 pid, tid; * * struct read_format values; - * struct sample_id sample_id; + * struct sample_id sample_id; * }; */ PERF_RECORD_READ = 8, @@ -994,12 +1006,12 @@ enum perf_event_type { * { u64 counters; } cntr[nr] && PERF_SAMPLE_BRANCH_COUNTERS * } && PERF_SAMPLE_BRANCH_STACK * - * { u64 abi; # enum perf_sample_regs_abi - * u64 regs[weight(mask)]; } && PERF_SAMPLE_REGS_USER + * { u64 abi; # enum perf_sample_regs_abi + * u64 regs[weight(mask)]; } && PERF_SAMPLE_REGS_USER * - * { u64 size; - * char data[size]; - * u64 dyn_size; } && PERF_SAMPLE_STACK_USER + * { u64 size; + * char data[size]; + * u64 dyn_size; } && PERF_SAMPLE_STACK_USER * * { union perf_sample_weight * { @@ -1059,7 +1071,7 @@ enum perf_event_type { * }; * u32 prot, flags; * char filename[]; - * struct sample_id sample_id; + * struct sample_id sample_id; * }; */ PERF_RECORD_MMAP2 = 10, @@ -1068,12 +1080,12 @@ enum perf_event_type { * Records that new data landed in the AUX buffer part. * * struct { - * struct perf_event_header header; + * struct perf_event_header header; * - * u64 aux_offset; - * u64 aux_size; + * u64 aux_offset; + * u64 aux_size; * u64 flags; - * struct sample_id sample_id; + * struct sample_id sample_id; * }; */ PERF_RECORD_AUX = 11, @@ -1156,7 +1168,7 @@ enum perf_event_type { PERF_RECORD_KSYMBOL = 17, /* - * Record bpf events: + * Record BPF events: * enum perf_bpf_event_type { * PERF_BPF_EVENT_UNKNOWN = 0, * PERF_BPF_EVENT_PROG_LOAD = 1, @@ -1234,179 +1246,181 @@ enum perf_record_ksymbol_type { #define PERF_RECORD_KSYMBOL_FLAGS_UNREGISTER (1 << 0) enum perf_bpf_event_type { - PERF_BPF_EVENT_UNKNOWN = 0, - PERF_BPF_EVENT_PROG_LOAD = 1, - PERF_BPF_EVENT_PROG_UNLOAD = 2, - PERF_BPF_EVENT_MAX, /* non-ABI */ + PERF_BPF_EVENT_UNKNOWN = 0, + PERF_BPF_EVENT_PROG_LOAD = 1, + PERF_BPF_EVENT_PROG_UNLOAD = 2, + PERF_BPF_EVENT_MAX, /* non-ABI */ }; -#define PERF_MAX_STACK_DEPTH 127 -#define PERF_MAX_CONTEXTS_PER_STACK 8 +#define PERF_MAX_STACK_DEPTH 127 +#define PERF_MAX_CONTEXTS_PER_STACK 8 enum perf_callchain_context { - PERF_CONTEXT_HV = (__u64)-32, - PERF_CONTEXT_KERNEL = (__u64)-128, - PERF_CONTEXT_USER = (__u64)-512, + PERF_CONTEXT_HV = (__u64)-32, + PERF_CONTEXT_KERNEL = (__u64)-128, + PERF_CONTEXT_USER = (__u64)-512, - PERF_CONTEXT_GUEST = (__u64)-2048, - PERF_CONTEXT_GUEST_KERNEL = (__u64)-2176, - PERF_CONTEXT_GUEST_USER = (__u64)-2560, + PERF_CONTEXT_GUEST = (__u64)-2048, + PERF_CONTEXT_GUEST_KERNEL = (__u64)-2176, + PERF_CONTEXT_GUEST_USER = (__u64)-2560, - PERF_CONTEXT_MAX = (__u64)-4095, + PERF_CONTEXT_MAX = (__u64)-4095, }; /** * PERF_RECORD_AUX::flags bits */ -#define PERF_AUX_FLAG_TRUNCATED 0x01 /* record was truncated to fit */ -#define PERF_AUX_FLAG_OVERWRITE 0x02 /* snapshot from overwrite mode */ -#define PERF_AUX_FLAG_PARTIAL 0x04 /* record contains gaps */ -#define PERF_AUX_FLAG_COLLISION 0x08 /* sample collided with another */ +#define PERF_AUX_FLAG_TRUNCATED 0x0001 /* Record was truncated to fit */ +#define PERF_AUX_FLAG_OVERWRITE 0x0002 /* Snapshot from overwrite mode */ +#define PERF_AUX_FLAG_PARTIAL 0x0004 /* Record contains gaps */ +#define PERF_AUX_FLAG_COLLISION 0x0008 /* Sample collided with another */ #define PERF_AUX_FLAG_PMU_FORMAT_TYPE_MASK 0xff00 /* PMU specific trace format type */ /* CoreSight PMU AUX buffer formats */ -#define PERF_AUX_FLAG_CORESIGHT_FORMAT_CORESIGHT 0x0000 /* Default for backward compatibility */ -#define PERF_AUX_FLAG_CORESIGHT_FORMAT_RAW 0x0100 /* Raw format of the source */ +#define PERF_AUX_FLAG_CORESIGHT_FORMAT_CORESIGHT 0x0000 /* Default for backward compatibility */ +#define PERF_AUX_FLAG_CORESIGHT_FORMAT_RAW 0x0100 /* Raw format of the source */ -#define PERF_FLAG_FD_NO_GROUP (1UL << 0) -#define PERF_FLAG_FD_OUTPUT (1UL << 1) -#define PERF_FLAG_PID_CGROUP (1UL << 2) /* pid=cgroup id, per-cpu mode only */ -#define PERF_FLAG_FD_CLOEXEC (1UL << 3) /* O_CLOEXEC */ +#define PERF_FLAG_FD_NO_GROUP (1UL << 0) +#define PERF_FLAG_FD_OUTPUT (1UL << 1) +#define PERF_FLAG_PID_CGROUP (1UL << 2) /* pid=cgroup ID, per-CPU mode only */ +#define PERF_FLAG_FD_CLOEXEC (1UL << 3) /* O_CLOEXEC */ #if defined(__LITTLE_ENDIAN_BITFIELD) union perf_mem_data_src { __u64 val; struct { - __u64 mem_op:5, /* type of opcode */ - mem_lvl:14, /* memory hierarchy level */ - mem_snoop:5, /* snoop mode */ - mem_lock:2, /* lock instr */ - mem_dtlb:7, /* tlb access */ - mem_lvl_num:4, /* memory hierarchy level number */ - mem_remote:1, /* remote */ - mem_snoopx:2, /* snoop mode, ext */ - mem_blk:3, /* access blocked */ - mem_hops:3, /* hop level */ - mem_rsvd:18; + __u64 mem_op : 5, /* Type of opcode */ + mem_lvl : 14, /* Memory hierarchy level */ + mem_snoop : 5, /* Snoop mode */ + mem_lock : 2, /* Lock instr */ + mem_dtlb : 7, /* TLB access */ + mem_lvl_num : 4, /* Memory hierarchy level number */ + mem_remote : 1, /* Remote */ + mem_snoopx : 2, /* Snoop mode, ext */ + mem_blk : 3, /* Access blocked */ + mem_hops : 3, /* Hop level */ + mem_rsvd : 18; }; }; #elif defined(__BIG_ENDIAN_BITFIELD) union perf_mem_data_src { __u64 val; struct { - __u64 mem_rsvd:18, - mem_hops:3, /* hop level */ - mem_blk:3, /* access blocked */ - mem_snoopx:2, /* snoop mode, ext */ - mem_remote:1, /* remote */ - mem_lvl_num:4, /* memory hierarchy level number */ - mem_dtlb:7, /* tlb access */ - mem_lock:2, /* lock instr */ - mem_snoop:5, /* snoop mode */ - mem_lvl:14, /* memory hierarchy level */ - mem_op:5; /* type of opcode */ + __u64 mem_rsvd : 18, + mem_hops : 3, /* Hop level */ + mem_blk : 3, /* Access blocked */ + mem_snoopx : 2, /* Snoop mode, ext */ + mem_remote : 1, /* Remote */ + mem_lvl_num : 4, /* Memory hierarchy level number */ + mem_dtlb : 7, /* TLB access */ + mem_lock : 2, /* Lock instr */ + mem_snoop : 5, /* Snoop mode */ + mem_lvl : 14, /* Memory hierarchy level */ + mem_op : 5; /* Type of opcode */ }; }; #else -#error "Unknown endianness" +# error "Unknown endianness" #endif -/* type of opcode (load/store/prefetch,code) */ -#define PERF_MEM_OP_NA 0x01 /* not available */ -#define PERF_MEM_OP_LOAD 0x02 /* load instruction */ -#define PERF_MEM_OP_STORE 0x04 /* store instruction */ -#define PERF_MEM_OP_PFETCH 0x08 /* prefetch */ -#define PERF_MEM_OP_EXEC 0x10 /* code (execution) */ -#define PERF_MEM_OP_SHIFT 0 +/* Type of memory opcode: */ +#define PERF_MEM_OP_NA 0x0001 /* Not available */ +#define PERF_MEM_OP_LOAD 0x0002 /* Load instruction */ +#define PERF_MEM_OP_STORE 0x0004 /* Store instruction */ +#define PERF_MEM_OP_PFETCH 0x0008 /* Prefetch */ +#define PERF_MEM_OP_EXEC 0x0010 /* Code (execution) */ +#define PERF_MEM_OP_SHIFT 0 /* - * PERF_MEM_LVL_* namespace being depricated to some extent in the + * The PERF_MEM_LVL_* namespace is being deprecated to some extent in * favour of newer composite PERF_MEM_{LVLNUM_,REMOTE_,SNOOPX_} fields. - * Supporting this namespace inorder to not break defined ABIs. + * We support this namespace in order to not break defined ABIs. * - * memory hierarchy (memory level, hit or miss) + * Memory hierarchy (memory level, hit or miss) */ -#define PERF_MEM_LVL_NA 0x01 /* not available */ -#define PERF_MEM_LVL_HIT 0x02 /* hit level */ -#define PERF_MEM_LVL_MISS 0x04 /* miss level */ -#define PERF_MEM_LVL_L1 0x08 /* L1 */ -#define PERF_MEM_LVL_LFB 0x10 /* Line Fill Buffer */ -#define PERF_MEM_LVL_L2 0x20 /* L2 */ -#define PERF_MEM_LVL_L3 0x40 /* L3 */ -#define PERF_MEM_LVL_LOC_RAM 0x80 /* Local DRAM */ -#define PERF_MEM_LVL_REM_RAM1 0x100 /* Remote DRAM (1 hop) */ -#define PERF_MEM_LVL_REM_RAM2 0x200 /* Remote DRAM (2 hops) */ -#define PERF_MEM_LVL_REM_CCE1 0x400 /* Remote Cache (1 hop) */ -#define PERF_MEM_LVL_REM_CCE2 0x800 /* Remote Cache (2 hops) */ -#define PERF_MEM_LVL_IO 0x1000 /* I/O memory */ -#define PERF_MEM_LVL_UNC 0x2000 /* Uncached memory */ -#define PERF_MEM_LVL_SHIFT 5 - -#define PERF_MEM_REMOTE_REMOTE 0x01 /* Remote */ -#define PERF_MEM_REMOTE_SHIFT 37 - -#define PERF_MEM_LVLNUM_L1 0x01 /* L1 */ -#define PERF_MEM_LVLNUM_L2 0x02 /* L2 */ -#define PERF_MEM_LVLNUM_L3 0x03 /* L3 */ -#define PERF_MEM_LVLNUM_L4 0x04 /* L4 */ -/* 5-0x7 available */ -#define PERF_MEM_LVLNUM_UNC 0x08 /* Uncached */ -#define PERF_MEM_LVLNUM_CXL 0x09 /* CXL */ -#define PERF_MEM_LVLNUM_IO 0x0a /* I/O */ -#define PERF_MEM_LVLNUM_ANY_CACHE 0x0b /* Any cache */ -#define PERF_MEM_LVLNUM_LFB 0x0c /* LFB */ -#define PERF_MEM_LVLNUM_RAM 0x0d /* RAM */ -#define PERF_MEM_LVLNUM_PMEM 0x0e /* PMEM */ -#define PERF_MEM_LVLNUM_NA 0x0f /* N/A */ - -#define PERF_MEM_LVLNUM_SHIFT 33 - -/* snoop mode */ -#define PERF_MEM_SNOOP_NA 0x01 /* not available */ -#define PERF_MEM_SNOOP_NONE 0x02 /* no snoop */ -#define PERF_MEM_SNOOP_HIT 0x04 /* snoop hit */ -#define PERF_MEM_SNOOP_MISS 0x08 /* snoop miss */ -#define PERF_MEM_SNOOP_HITM 0x10 /* snoop hit modified */ -#define PERF_MEM_SNOOP_SHIFT 19 - -#define PERF_MEM_SNOOPX_FWD 0x01 /* forward */ -#define PERF_MEM_SNOOPX_PEER 0x02 /* xfer from peer */ -#define PERF_MEM_SNOOPX_SHIFT 38 - -/* locked instruction */ -#define PERF_MEM_LOCK_NA 0x01 /* not available */ -#define PERF_MEM_LOCK_LOCKED 0x02 /* locked transaction */ -#define PERF_MEM_LOCK_SHIFT 24 +#define PERF_MEM_LVL_NA 0x0001 /* Not available */ +#define PERF_MEM_LVL_HIT 0x0002 /* Hit level */ +#define PERF_MEM_LVL_MISS 0x0004 /* Miss level */ +#define PERF_MEM_LVL_L1 0x0008 /* L1 */ +#define PERF_MEM_LVL_LFB 0x0010 /* Line Fill Buffer */ +#define PERF_MEM_LVL_L2 0x0020 /* L2 */ +#define PERF_MEM_LVL_L3 0x0040 /* L3 */ +#define PERF_MEM_LVL_LOC_RAM 0x0080 /* Local DRAM */ +#define PERF_MEM_LVL_REM_RAM1 0x0100 /* Remote DRAM (1 hop) */ +#define PERF_MEM_LVL_REM_RAM2 0x0200 /* Remote DRAM (2 hops) */ +#define PERF_MEM_LVL_REM_CCE1 0x0400 /* Remote Cache (1 hop) */ +#define PERF_MEM_LVL_REM_CCE2 0x0800 /* Remote Cache (2 hops) */ +#define PERF_MEM_LVL_IO 0x1000 /* I/O memory */ +#define PERF_MEM_LVL_UNC 0x2000 /* Uncached memory */ +#define PERF_MEM_LVL_SHIFT 5 + +#define PERF_MEM_REMOTE_REMOTE 0x0001 /* Remote */ +#define PERF_MEM_REMOTE_SHIFT 37 + +#define PERF_MEM_LVLNUM_L1 0x0001 /* L1 */ +#define PERF_MEM_LVLNUM_L2 0x0002 /* L2 */ +#define PERF_MEM_LVLNUM_L3 0x0003 /* L3 */ +#define PERF_MEM_LVLNUM_L4 0x0004 /* L4 */ +#define PERF_MEM_LVLNUM_L2_MHB 0x0005 /* L2 Miss Handling Buffer */ +#define PERF_MEM_LVLNUM_MSC 0x0006 /* Memory-side Cache */ +/* 0x007 available */ +#define PERF_MEM_LVLNUM_UNC 0x0008 /* Uncached */ +#define PERF_MEM_LVLNUM_CXL 0x0009 /* CXL */ +#define PERF_MEM_LVLNUM_IO 0x000a /* I/O */ +#define PERF_MEM_LVLNUM_ANY_CACHE 0x000b /* Any cache */ +#define PERF_MEM_LVLNUM_LFB 0x000c /* LFB / L1 Miss Handling Buffer */ +#define PERF_MEM_LVLNUM_RAM 0x000d /* RAM */ +#define PERF_MEM_LVLNUM_PMEM 0x000e /* PMEM */ +#define PERF_MEM_LVLNUM_NA 0x000f /* N/A */ + +#define PERF_MEM_LVLNUM_SHIFT 33 + +/* Snoop mode */ +#define PERF_MEM_SNOOP_NA 0x0001 /* Not available */ +#define PERF_MEM_SNOOP_NONE 0x0002 /* No snoop */ +#define PERF_MEM_SNOOP_HIT 0x0004 /* Snoop hit */ +#define PERF_MEM_SNOOP_MISS 0x0008 /* Snoop miss */ +#define PERF_MEM_SNOOP_HITM 0x0010 /* Snoop hit modified */ +#define PERF_MEM_SNOOP_SHIFT 19 + +#define PERF_MEM_SNOOPX_FWD 0x0001 /* Forward */ +#define PERF_MEM_SNOOPX_PEER 0x0002 /* Transfer from peer */ +#define PERF_MEM_SNOOPX_SHIFT 38 + +/* Locked instruction */ +#define PERF_MEM_LOCK_NA 0x0001 /* Not available */ +#define PERF_MEM_LOCK_LOCKED 0x0002 /* Locked transaction */ +#define PERF_MEM_LOCK_SHIFT 24 /* TLB access */ -#define PERF_MEM_TLB_NA 0x01 /* not available */ -#define PERF_MEM_TLB_HIT 0x02 /* hit level */ -#define PERF_MEM_TLB_MISS 0x04 /* miss level */ -#define PERF_MEM_TLB_L1 0x08 /* L1 */ -#define PERF_MEM_TLB_L2 0x10 /* L2 */ -#define PERF_MEM_TLB_WK 0x20 /* Hardware Walker*/ -#define PERF_MEM_TLB_OS 0x40 /* OS fault handler */ -#define PERF_MEM_TLB_SHIFT 26 +#define PERF_MEM_TLB_NA 0x0001 /* Not available */ +#define PERF_MEM_TLB_HIT 0x0002 /* Hit level */ +#define PERF_MEM_TLB_MISS 0x0004 /* Miss level */ +#define PERF_MEM_TLB_L1 0x0008 /* L1 */ +#define PERF_MEM_TLB_L2 0x0010 /* L2 */ +#define PERF_MEM_TLB_WK 0x0020 /* Hardware Walker*/ +#define PERF_MEM_TLB_OS 0x0040 /* OS fault handler */ +#define PERF_MEM_TLB_SHIFT 26 /* Access blocked */ -#define PERF_MEM_BLK_NA 0x01 /* not available */ -#define PERF_MEM_BLK_DATA 0x02 /* data could not be forwarded */ -#define PERF_MEM_BLK_ADDR 0x04 /* address conflict */ -#define PERF_MEM_BLK_SHIFT 40 - -/* hop level */ -#define PERF_MEM_HOPS_0 0x01 /* remote core, same node */ -#define PERF_MEM_HOPS_1 0x02 /* remote node, same socket */ -#define PERF_MEM_HOPS_2 0x03 /* remote socket, same board */ -#define PERF_MEM_HOPS_3 0x04 /* remote board */ +#define PERF_MEM_BLK_NA 0x0001 /* Not available */ +#define PERF_MEM_BLK_DATA 0x0002 /* Data could not be forwarded */ +#define PERF_MEM_BLK_ADDR 0x0004 /* Address conflict */ +#define PERF_MEM_BLK_SHIFT 40 + +/* Hop level */ +#define PERF_MEM_HOPS_0 0x0001 /* Remote core, same node */ +#define PERF_MEM_HOPS_1 0x0002 /* Remote node, same socket */ +#define PERF_MEM_HOPS_2 0x0003 /* Remote socket, same board */ +#define PERF_MEM_HOPS_3 0x0004 /* Remote board */ /* 5-7 available */ -#define PERF_MEM_HOPS_SHIFT 43 +#define PERF_MEM_HOPS_SHIFT 43 #define PERF_MEM_S(a, s) \ (((__u64)PERF_MEM_##a##_##s) << PERF_MEM_##a##_SHIFT) /* - * single taken branch record layout: + * Layout of single taken branch records: * * from: source instruction (may not always be a branch insn) * to: branch target @@ -1425,37 +1439,37 @@ union perf_mem_data_src { struct perf_branch_entry { __u64 from; __u64 to; - __u64 mispred:1, /* target mispredicted */ - predicted:1,/* target predicted */ - in_tx:1, /* in transaction */ - abort:1, /* transaction abort */ - cycles:16, /* cycle count to last branch */ - type:4, /* branch type */ - spec:2, /* branch speculation info */ - new_type:4, /* additional branch type */ - priv:3, /* privilege level */ - reserved:31; + __u64 mispred : 1, /* target mispredicted */ + predicted : 1, /* target predicted */ + in_tx : 1, /* in transaction */ + abort : 1, /* transaction abort */ + cycles : 16, /* cycle count to last branch */ + type : 4, /* branch type */ + spec : 2, /* branch speculation info */ + new_type : 4, /* additional branch type */ + priv : 3, /* privilege level */ + reserved : 31; }; /* Size of used info bits in struct perf_branch_entry */ #define PERF_BRANCH_ENTRY_INFO_BITS_MAX 33 union perf_sample_weight { - __u64 full; + __u64 full; #if defined(__LITTLE_ENDIAN_BITFIELD) struct { - __u32 var1_dw; - __u16 var2_w; - __u16 var3_w; + __u32 var1_dw; + __u16 var2_w; + __u16 var3_w; }; #elif defined(__BIG_ENDIAN_BITFIELD) struct { - __u16 var3_w; - __u16 var2_w; - __u32 var1_dw; + __u16 var3_w; + __u16 var2_w; + __u32 var1_dw; }; #else -#error "Unknown endianness" +# error "Unknown endianness" #endif }; -- Gitee From d3e9b9705cd967ea81131977c43d18320857b511 Mon Sep 17 00:00:00 2001 From: Emil Renner Berthing Date: Wed, 1 Nov 2023 11:32:59 -0700 Subject: [PATCH 20/22] riscv: Avoid unaligned access when relocating modules ANBZ: #24284 commit 8cbe0accc4a6ba7ed34812a1c7e1ba67e7f7b2a4 upstream. With the C-extension regular 32bit instructions are not necessarily aligned on 4-byte boundaries. RISC-V instructions are in fact an ordered list of 16bit little-endian "parcels", so access the instruction as such. This should also make the code work in case someone builds a big-endian RISC-V machine. Signed-off-by: Emil Renner Berthing Signed-off-by: Charlie Jenkins Link: https://lore.kernel.org/r/20231101-module_relocations-v9-1-8dfa3483c400@rivosinc.com Signed-off-by: Palmer Dabbelt Signed-off-by: Fangyu Yu --- arch/riscv/kernel/module.c | 157 +++++++++++++++++++------------------ 1 file changed, 81 insertions(+), 76 deletions(-) diff --git a/arch/riscv/kernel/module.c b/arch/riscv/kernel/module.c index df4f6fec5d17..26d5e2071690 100644 --- a/arch/riscv/kernel/module.c +++ b/arch/riscv/kernel/module.c @@ -27,68 +27,90 @@ static bool riscv_insn_valid_32bit_offset(ptrdiff_t val) #endif } -static int apply_r_riscv_32_rela(struct module *me, u32 *location, Elf_Addr v) +static int riscv_insn_rmw(void *location, u32 keep, u32 set) +{ + u16 *parcel = location; + u32 insn = (u32)le16_to_cpu(parcel[0]) | (u32)le16_to_cpu(parcel[1]) << 16; + + insn &= keep; + insn |= set; + + parcel[0] = cpu_to_le16(insn); + parcel[1] = cpu_to_le16(insn >> 16); + return 0; +} + +static int riscv_insn_rvc_rmw(void *location, u16 keep, u16 set) +{ + u16 *parcel = location; + u16 insn = le16_to_cpu(*parcel); + + insn &= keep; + insn |= set; + + *parcel = cpu_to_le16(insn); + return 0; +} + +static int apply_r_riscv_32_rela(struct module *me, void *location, Elf_Addr v) { if (v != (u32)v) { pr_err("%s: value %016llx out of range for 32-bit field\n", me->name, (long long)v); return -EINVAL; } - *location = v; + *(u32 *)location = v; return 0; } -static int apply_r_riscv_64_rela(struct module *me, u32 *location, Elf_Addr v) +static int apply_r_riscv_64_rela(struct module *me, void *location, Elf_Addr v) { *(u64 *)location = v; return 0; } -static int apply_r_riscv_branch_rela(struct module *me, u32 *location, +static int apply_r_riscv_branch_rela(struct module *me, void *location, Elf_Addr v) { - ptrdiff_t offset = (void *)v - (void *)location; + ptrdiff_t offset = (void *)v - location; u32 imm12 = (offset & 0x1000) << (31 - 12); u32 imm11 = (offset & 0x800) >> (11 - 7); u32 imm10_5 = (offset & 0x7e0) << (30 - 10); u32 imm4_1 = (offset & 0x1e) << (11 - 4); - *location = (*location & 0x1fff07f) | imm12 | imm11 | imm10_5 | imm4_1; - return 0; + return riscv_insn_rmw(location, 0x1fff07f, imm12 | imm11 | imm10_5 | imm4_1); } -static int apply_r_riscv_jal_rela(struct module *me, u32 *location, +static int apply_r_riscv_jal_rela(struct module *me, void *location, Elf_Addr v) { - ptrdiff_t offset = (void *)v - (void *)location; + ptrdiff_t offset = (void *)v - location; u32 imm20 = (offset & 0x100000) << (31 - 20); u32 imm19_12 = (offset & 0xff000); u32 imm11 = (offset & 0x800) << (20 - 11); u32 imm10_1 = (offset & 0x7fe) << (30 - 10); - *location = (*location & 0xfff) | imm20 | imm19_12 | imm11 | imm10_1; - return 0; + return riscv_insn_rmw(location, 0xfff, imm20 | imm19_12 | imm11 | imm10_1); } -static int apply_r_riscv_rvc_branch_rela(struct module *me, u32 *location, +static int apply_r_riscv_rvc_branch_rela(struct module *me, void *location, Elf_Addr v) { - ptrdiff_t offset = (void *)v - (void *)location; + ptrdiff_t offset = (void *)v - location; u16 imm8 = (offset & 0x100) << (12 - 8); u16 imm7_6 = (offset & 0xc0) >> (6 - 5); u16 imm5 = (offset & 0x20) >> (5 - 2); u16 imm4_3 = (offset & 0x18) << (12 - 5); u16 imm2_1 = (offset & 0x6) << (12 - 10); - *(u16 *)location = (*(u16 *)location & 0xe383) | - imm8 | imm7_6 | imm5 | imm4_3 | imm2_1; - return 0; + return riscv_insn_rvc_rmw(location, 0xe383, + imm8 | imm7_6 | imm5 | imm4_3 | imm2_1); } -static int apply_r_riscv_rvc_jump_rela(struct module *me, u32 *location, +static int apply_r_riscv_rvc_jump_rela(struct module *me, void *location, Elf_Addr v) { - ptrdiff_t offset = (void *)v - (void *)location; + ptrdiff_t offset = (void *)v - location; u16 imm11 = (offset & 0x800) << (12 - 11); u16 imm10 = (offset & 0x400) >> (10 - 8); u16 imm9_8 = (offset & 0x300) << (12 - 11); @@ -98,16 +120,14 @@ static int apply_r_riscv_rvc_jump_rela(struct module *me, u32 *location, u16 imm4 = (offset & 0x10) << (12 - 5); u16 imm3_1 = (offset & 0xe) << (12 - 10); - *(u16 *)location = (*(u16 *)location & 0xe003) | - imm11 | imm10 | imm9_8 | imm7 | imm6 | imm5 | imm4 | imm3_1; - return 0; + return riscv_insn_rvc_rmw(location, 0xe003, + imm11 | imm10 | imm9_8 | imm7 | imm6 | imm5 | imm4 | imm3_1); } -static int apply_r_riscv_pcrel_hi20_rela(struct module *me, u32 *location, +static int apply_r_riscv_pcrel_hi20_rela(struct module *me, void *location, Elf_Addr v) { - ptrdiff_t offset = (void *)v - (void *)location; - s32 hi20; + ptrdiff_t offset = (void *)v - location; if (!riscv_insn_valid_32bit_offset(offset)) { pr_err( @@ -116,23 +136,20 @@ static int apply_r_riscv_pcrel_hi20_rela(struct module *me, u32 *location, return -EINVAL; } - hi20 = (offset + 0x800) & 0xfffff000; - *location = (*location & 0xfff) | hi20; - return 0; + return riscv_insn_rmw(location, 0xfff, (offset + 0x800) & 0xfffff000); } -static int apply_r_riscv_pcrel_lo12_i_rela(struct module *me, u32 *location, +static int apply_r_riscv_pcrel_lo12_i_rela(struct module *me, void *location, Elf_Addr v) { /* * v is the lo12 value to fill. It is calculated before calling this * handler. */ - *location = (*location & 0xfffff) | ((v & 0xfff) << 20); - return 0; + return riscv_insn_rmw(location, 0xfffff, (v & 0xfff) << 20); } -static int apply_r_riscv_pcrel_lo12_s_rela(struct module *me, u32 *location, +static int apply_r_riscv_pcrel_lo12_s_rela(struct module *me, void *location, Elf_Addr v) { /* @@ -142,15 +159,12 @@ static int apply_r_riscv_pcrel_lo12_s_rela(struct module *me, u32 *location, u32 imm11_5 = (v & 0xfe0) << (31 - 11); u32 imm4_0 = (v & 0x1f) << (11 - 4); - *location = (*location & 0x1fff07f) | imm11_5 | imm4_0; - return 0; + return riscv_insn_rmw(location, 0x1fff07f, imm11_5 | imm4_0); } -static int apply_r_riscv_hi20_rela(struct module *me, u32 *location, +static int apply_r_riscv_hi20_rela(struct module *me, void *location, Elf_Addr v) { - s32 hi20; - if (IS_ENABLED(CONFIG_CMODEL_MEDLOW)) { pr_err( "%s: target %016llx can not be addressed by the 32-bit offset from PC = %p\n", @@ -158,22 +172,20 @@ static int apply_r_riscv_hi20_rela(struct module *me, u32 *location, return -EINVAL; } - hi20 = ((s32)v + 0x800) & 0xfffff000; - *location = (*location & 0xfff) | hi20; - return 0; + return riscv_insn_rmw(location, 0xfff, ((s32)v + 0x800) & 0xfffff000); } -static int apply_r_riscv_lo12_i_rela(struct module *me, u32 *location, +static int apply_r_riscv_lo12_i_rela(struct module *me, void *location, Elf_Addr v) { /* Skip medlow checking because of filtering by HI20 already */ s32 hi20 = ((s32)v + 0x800) & 0xfffff000; s32 lo12 = ((s32)v - hi20); - *location = (*location & 0xfffff) | ((lo12 & 0xfff) << 20); - return 0; + + return riscv_insn_rmw(location, 0xfffff, (lo12 & 0xfff) << 20); } -static int apply_r_riscv_lo12_s_rela(struct module *me, u32 *location, +static int apply_r_riscv_lo12_s_rela(struct module *me, void *location, Elf_Addr v) { /* Skip medlow checking because of filtering by HI20 already */ @@ -181,20 +193,18 @@ static int apply_r_riscv_lo12_s_rela(struct module *me, u32 *location, s32 lo12 = ((s32)v - hi20); u32 imm11_5 = (lo12 & 0xfe0) << (31 - 11); u32 imm4_0 = (lo12 & 0x1f) << (11 - 4); - *location = (*location & 0x1fff07f) | imm11_5 | imm4_0; - return 0; + + return riscv_insn_rmw(location, 0x1fff07f, imm11_5 | imm4_0); } -static int apply_r_riscv_got_hi20_rela(struct module *me, u32 *location, +static int apply_r_riscv_got_hi20_rela(struct module *me, void *location, Elf_Addr v) { - ptrdiff_t offset = (void *)v - (void *)location; - s32 hi20; + ptrdiff_t offset = (void *)v - location; /* Always emit the got entry */ if (IS_ENABLED(CONFIG_MODULE_SECTIONS)) { - offset = module_emit_got_entry(me, v); - offset = (void *)offset - (void *)location; + offset = (void *)module_emit_got_entry(me, v) - location; } else { pr_err( "%s: can not generate the GOT entry for symbol = %016llx from PC = %p\n", @@ -202,22 +212,19 @@ static int apply_r_riscv_got_hi20_rela(struct module *me, u32 *location, return -EINVAL; } - hi20 = (offset + 0x800) & 0xfffff000; - *location = (*location & 0xfff) | hi20; - return 0; + return riscv_insn_rmw(location, 0xfff, (offset + 0x800) & 0xfffff000); } -static int apply_r_riscv_call_plt_rela(struct module *me, u32 *location, +static int apply_r_riscv_call_plt_rela(struct module *me, void *location, Elf_Addr v) { - ptrdiff_t offset = (void *)v - (void *)location; + ptrdiff_t offset = (void *)v - location; u32 hi20, lo12; if (!riscv_insn_valid_32bit_offset(offset)) { /* Only emit the plt entry if offset over 32-bit range */ if (IS_ENABLED(CONFIG_MODULE_SECTIONS)) { - offset = module_emit_plt_entry(me, v); - offset = (void *)offset - (void *)location; + offset = (void *)module_emit_plt_entry(me, v) - location; } else { pr_err( "%s: target %016llx can not be addressed by the 32-bit offset from PC = %p\n", @@ -228,15 +235,14 @@ static int apply_r_riscv_call_plt_rela(struct module *me, u32 *location, hi20 = (offset + 0x800) & 0xfffff000; lo12 = (offset - hi20) & 0xfff; - *location = (*location & 0xfff) | hi20; - *(location + 1) = (*(location + 1) & 0xfffff) | (lo12 << 20); - return 0; + riscv_insn_rmw(location, 0xfff, hi20); + return riscv_insn_rmw(location + 4, 0xfffff, lo12 << 20); } -static int apply_r_riscv_call_rela(struct module *me, u32 *location, +static int apply_r_riscv_call_rela(struct module *me, void *location, Elf_Addr v) { - ptrdiff_t offset = (void *)v - (void *)location; + ptrdiff_t offset = (void *)v - location; u32 hi20, lo12; if (!riscv_insn_valid_32bit_offset(offset)) { @@ -248,18 +254,17 @@ static int apply_r_riscv_call_rela(struct module *me, u32 *location, hi20 = (offset + 0x800) & 0xfffff000; lo12 = (offset - hi20) & 0xfff; - *location = (*location & 0xfff) | hi20; - *(location + 1) = (*(location + 1) & 0xfffff) | (lo12 << 20); - return 0; + riscv_insn_rmw(location, 0xfff, hi20); + return riscv_insn_rmw(location + 4, 0xfffff, lo12 << 20); } -static int apply_r_riscv_relax_rela(struct module *me, u32 *location, +static int apply_r_riscv_relax_rela(struct module *me, void *location, Elf_Addr v) { return 0; } -static int apply_r_riscv_align_rela(struct module *me, u32 *location, +static int apply_r_riscv_align_rela(struct module *me, void *location, Elf_Addr v) { pr_err( @@ -268,49 +273,49 @@ static int apply_r_riscv_align_rela(struct module *me, u32 *location, return -EINVAL; } -static int apply_r_riscv_add16_rela(struct module *me, u32 *location, +static int apply_r_riscv_add16_rela(struct module *me, void *location, Elf_Addr v) { *(u16 *)location += (u16)v; return 0; } -static int apply_r_riscv_add32_rela(struct module *me, u32 *location, +static int apply_r_riscv_add32_rela(struct module *me, void *location, Elf_Addr v) { *(u32 *)location += (u32)v; return 0; } -static int apply_r_riscv_add64_rela(struct module *me, u32 *location, +static int apply_r_riscv_add64_rela(struct module *me, void *location, Elf_Addr v) { *(u64 *)location += (u64)v; return 0; } -static int apply_r_riscv_sub16_rela(struct module *me, u32 *location, +static int apply_r_riscv_sub16_rela(struct module *me, void *location, Elf_Addr v) { *(u16 *)location -= (u16)v; return 0; } -static int apply_r_riscv_sub32_rela(struct module *me, u32 *location, +static int apply_r_riscv_sub32_rela(struct module *me, void *location, Elf_Addr v) { *(u32 *)location -= (u32)v; return 0; } -static int apply_r_riscv_sub64_rela(struct module *me, u32 *location, +static int apply_r_riscv_sub64_rela(struct module *me, void *location, Elf_Addr v) { *(u64 *)location -= (u64)v; return 0; } -static int (*reloc_handlers_rela[]) (struct module *me, u32 *location, +static int (*reloc_handlers_rela[]) (struct module *me, void *location, Elf_Addr v) = { [R_RISCV_32] = apply_r_riscv_32_rela, [R_RISCV_64] = apply_r_riscv_64_rela, @@ -342,9 +347,9 @@ int apply_relocate_add(Elf_Shdr *sechdrs, const char *strtab, struct module *me) { Elf_Rela *rel = (void *) sechdrs[relsec].sh_addr; - int (*handler)(struct module *me, u32 *location, Elf_Addr v); + int (*handler)(struct module *me, void *location, Elf_Addr v); Elf_Sym *sym; - u32 *location; + void *location; unsigned int i, type; Elf_Addr v; int res; -- Gitee From d11c17f8a1c5d95b2feca91845e06b9bad6b21fe Mon Sep 17 00:00:00 2001 From: Charlie Jenkins Date: Wed, 1 Nov 2023 11:33:00 -0700 Subject: [PATCH 21/22] riscv: Add remaining module relocations ANBZ: #24284 commit 8fd6c5142395a106b63c8668e9f4a7106b6a0772 upstream. Add all final module relocations and add error logs explaining the ones that are not supported. Implement overflow checks for ADD/SUB/SET/ULEB128 relocations. Signed-off-by: Charlie Jenkins Link: https://lore.kernel.org/r/20231101-module_relocations-v9-2-8dfa3483c400@rivosinc.com Signed-off-by: Palmer Dabbelt Signed-off-by: Fangyu Yu --- arch/riscv/include/uapi/asm/elf.h | 5 +- arch/riscv/kernel/module.c | 448 ++++++++++++++++++++++++++++-- 2 files changed, 423 insertions(+), 30 deletions(-) diff --git a/arch/riscv/include/uapi/asm/elf.h b/arch/riscv/include/uapi/asm/elf.h index d696d6610231..11a71b8533d5 100644 --- a/arch/riscv/include/uapi/asm/elf.h +++ b/arch/riscv/include/uapi/asm/elf.h @@ -49,6 +49,7 @@ typedef union __riscv_fp_state elf_fpregset_t; #define R_RISCV_TLS_DTPREL64 9 #define R_RISCV_TLS_TPREL32 10 #define R_RISCV_TLS_TPREL64 11 +#define R_RISCV_IRELATIVE 58 /* Relocation types not used by the dynamic linker */ #define R_RISCV_BRANCH 16 @@ -81,7 +82,6 @@ typedef union __riscv_fp_state elf_fpregset_t; #define R_RISCV_ALIGN 43 #define R_RISCV_RVC_BRANCH 44 #define R_RISCV_RVC_JUMP 45 -#define R_RISCV_LUI 46 #define R_RISCV_GPREL_I 47 #define R_RISCV_GPREL_S 48 #define R_RISCV_TPREL_I 49 @@ -93,6 +93,9 @@ typedef union __riscv_fp_state elf_fpregset_t; #define R_RISCV_SET16 55 #define R_RISCV_SET32 56 #define R_RISCV_32_PCREL 57 +#define R_RISCV_PLT32 59 +#define R_RISCV_SET_ULEB128 60 +#define R_RISCV_SUB_ULEB128 61 #endif /* _UAPI_ASM_RISCV_ELF_H */ diff --git a/arch/riscv/kernel/module.c b/arch/riscv/kernel/module.c index 26d5e2071690..15ee4c6f76be 100644 --- a/arch/riscv/kernel/module.c +++ b/arch/riscv/kernel/module.c @@ -7,6 +7,9 @@ #include #include #include +#include +#include +#include #include #include #include @@ -14,6 +17,38 @@ #include #include +struct used_bucket { + struct list_head head; + struct hlist_head *bucket; +}; + +struct relocation_head { + struct hlist_node node; + struct list_head *rel_entry; + void *location; +}; + +struct relocation_entry { + struct list_head head; + Elf_Addr value; + unsigned int type; +}; + +struct relocation_handlers { + int (*reloc_handler)(struct module *me, void *location, Elf_Addr v); + int (*accumulate_handler)(struct module *me, void *location, + long buffer); +}; + +unsigned int initialize_relocation_hashtable(unsigned int num_relocations); +void process_accumulated_relocations(struct module *me); +int add_relocation_to_accumulate(struct module *me, int type, void *location, + unsigned int hashtable_bits, Elf_Addr v); + +struct hlist_head *relocation_hashtable; + +struct list_head used_buckets_list; + /* * The auipc+jalr instruction pair can reach any PC-relative offset * in the range [-2^31 - 2^11, 2^31 - 2^11) @@ -273,6 +308,12 @@ static int apply_r_riscv_align_rela(struct module *me, void *location, return -EINVAL; } +static int apply_r_riscv_add8_rela(struct module *me, void *location, Elf_Addr v) +{ + *(u8 *)location += (u8)v; + return 0; +} + static int apply_r_riscv_add16_rela(struct module *me, void *location, Elf_Addr v) { @@ -294,6 +335,12 @@ static int apply_r_riscv_add64_rela(struct module *me, void *location, return 0; } +static int apply_r_riscv_sub8_rela(struct module *me, void *location, Elf_Addr v) +{ + *(u8 *)location -= (u8)v; + return 0; +} + static int apply_r_riscv_sub16_rela(struct module *me, void *location, Elf_Addr v) { @@ -315,33 +362,369 @@ static int apply_r_riscv_sub64_rela(struct module *me, void *location, return 0; } -static int (*reloc_handlers_rela[]) (struct module *me, void *location, - Elf_Addr v) = { - [R_RISCV_32] = apply_r_riscv_32_rela, - [R_RISCV_64] = apply_r_riscv_64_rela, - [R_RISCV_BRANCH] = apply_r_riscv_branch_rela, - [R_RISCV_JAL] = apply_r_riscv_jal_rela, - [R_RISCV_RVC_BRANCH] = apply_r_riscv_rvc_branch_rela, - [R_RISCV_RVC_JUMP] = apply_r_riscv_rvc_jump_rela, - [R_RISCV_PCREL_HI20] = apply_r_riscv_pcrel_hi20_rela, - [R_RISCV_PCREL_LO12_I] = apply_r_riscv_pcrel_lo12_i_rela, - [R_RISCV_PCREL_LO12_S] = apply_r_riscv_pcrel_lo12_s_rela, - [R_RISCV_HI20] = apply_r_riscv_hi20_rela, - [R_RISCV_LO12_I] = apply_r_riscv_lo12_i_rela, - [R_RISCV_LO12_S] = apply_r_riscv_lo12_s_rela, - [R_RISCV_GOT_HI20] = apply_r_riscv_got_hi20_rela, - [R_RISCV_CALL_PLT] = apply_r_riscv_call_plt_rela, - [R_RISCV_CALL] = apply_r_riscv_call_rela, - [R_RISCV_RELAX] = apply_r_riscv_relax_rela, - [R_RISCV_ALIGN] = apply_r_riscv_align_rela, - [R_RISCV_ADD16] = apply_r_riscv_add16_rela, - [R_RISCV_ADD32] = apply_r_riscv_add32_rela, - [R_RISCV_ADD64] = apply_r_riscv_add64_rela, - [R_RISCV_SUB16] = apply_r_riscv_sub16_rela, - [R_RISCV_SUB32] = apply_r_riscv_sub32_rela, - [R_RISCV_SUB64] = apply_r_riscv_sub64_rela, +static int dynamic_linking_not_supported(struct module *me, void *location, + Elf_Addr v) +{ + pr_err("%s: Dynamic linking not supported in kernel modules PC = %p\n", + me->name, location); + return -EINVAL; +} + +static int tls_not_supported(struct module *me, void *location, Elf_Addr v) +{ + pr_err("%s: Thread local storage not supported in kernel modules PC = %p\n", + me->name, location); + return -EINVAL; +} + +static int apply_r_riscv_sub6_rela(struct module *me, void *location, Elf_Addr v) +{ + u8 *byte = location; + u8 value = v; + + *byte = (*byte - (value & 0x3f)) & 0x3f; + return 0; +} + +static int apply_r_riscv_set6_rela(struct module *me, void *location, Elf_Addr v) +{ + u8 *byte = location; + u8 value = v; + + *byte = (*byte & 0xc0) | (value & 0x3f); + return 0; +} + +static int apply_r_riscv_set8_rela(struct module *me, void *location, Elf_Addr v) +{ + *(u8 *)location = (u8)v; + return 0; +} + +static int apply_r_riscv_set16_rela(struct module *me, void *location, + Elf_Addr v) +{ + *(u16 *)location = (u16)v; + return 0; +} + +static int apply_r_riscv_set32_rela(struct module *me, void *location, + Elf_Addr v) +{ + *(u32 *)location = (u32)v; + return 0; +} + +static int apply_r_riscv_32_pcrel_rela(struct module *me, void *location, + Elf_Addr v) +{ + *(u32 *)location = v - (uintptr_t)location; + return 0; +} + +static int apply_r_riscv_plt32_rela(struct module *me, void *location, + Elf_Addr v) +{ + ptrdiff_t offset = (void *)v - location; + + if (!riscv_insn_valid_32bit_offset(offset)) { + /* Only emit the plt entry if offset over 32-bit range */ + if (IS_ENABLED(CONFIG_MODULE_SECTIONS)) { + offset = (void *)module_emit_plt_entry(me, v) - location; + } else { + pr_err("%s: target %016llx can not be addressed by the 32-bit offset from PC = %p\n", + me->name, (long long)v, location); + return -EINVAL; + } + } + + *(u32 *)location = (u32)offset; + return 0; +} + +static int apply_r_riscv_set_uleb128(struct module *me, void *location, Elf_Addr v) +{ + *(long *)location = v; + return 0; +} + +static int apply_r_riscv_sub_uleb128(struct module *me, void *location, Elf_Addr v) +{ + *(long *)location -= v; + return 0; +} + +static int apply_6_bit_accumulation(struct module *me, void *location, long buffer) +{ + u8 *byte = location; + u8 value = buffer; + + if (buffer > 0x3f) { + pr_err("%s: value %ld out of range for 6-bit relocation.\n", + me->name, buffer); + return -EINVAL; + } + + *byte = (*byte & 0xc0) | (value & 0x3f); + return 0; +} + +static int apply_8_bit_accumulation(struct module *me, void *location, long buffer) +{ + if (buffer > U8_MAX) { + pr_err("%s: value %ld out of range for 8-bit relocation.\n", + me->name, buffer); + return -EINVAL; + } + *(u8 *)location = (u8)buffer; + return 0; +} + +static int apply_16_bit_accumulation(struct module *me, void *location, long buffer) +{ + if (buffer > U16_MAX) { + pr_err("%s: value %ld out of range for 16-bit relocation.\n", + me->name, buffer); + return -EINVAL; + } + *(u16 *)location = (u16)buffer; + return 0; +} + +static int apply_32_bit_accumulation(struct module *me, void *location, long buffer) +{ + if (buffer > U32_MAX) { + pr_err("%s: value %ld out of range for 32-bit relocation.\n", + me->name, buffer); + return -EINVAL; + } + *(u32 *)location = (u32)buffer; + return 0; +} + +static int apply_64_bit_accumulation(struct module *me, void *location, long buffer) +{ + *(u64 *)location = (u64)buffer; + return 0; +} + +static int apply_uleb128_accumulation(struct module *me, void *location, long buffer) +{ + /* + * ULEB128 is a variable length encoding. Encode the buffer into + * the ULEB128 data format. + */ + u8 *p = location; + + while (buffer != 0) { + u8 value = buffer & 0x7f; + + buffer >>= 7; + value |= (!!buffer) << 7; + + *p++ = value; + } + return 0; +} + +/* + * Relocations defined in the riscv-elf-psabi-doc. + * This handles static linking only. + */ +static const struct relocation_handlers reloc_handlers[] = { + [R_RISCV_32] = { apply_r_riscv_32_rela }, + [R_RISCV_64] = { apply_r_riscv_64_rela }, + [R_RISCV_RELATIVE] = { dynamic_linking_not_supported }, + [R_RISCV_COPY] = { dynamic_linking_not_supported }, + [R_RISCV_JUMP_SLOT] = { dynamic_linking_not_supported }, + [R_RISCV_TLS_DTPMOD32] = { dynamic_linking_not_supported }, + [R_RISCV_TLS_DTPMOD64] = { dynamic_linking_not_supported }, + [R_RISCV_TLS_DTPREL32] = { dynamic_linking_not_supported }, + [R_RISCV_TLS_DTPREL64] = { dynamic_linking_not_supported }, + [R_RISCV_TLS_TPREL32] = { dynamic_linking_not_supported }, + [R_RISCV_TLS_TPREL64] = { dynamic_linking_not_supported }, + /* 12-15 undefined */ + [R_RISCV_BRANCH] = { apply_r_riscv_branch_rela }, + [R_RISCV_JAL] = { apply_r_riscv_jal_rela }, + [R_RISCV_CALL] = { apply_r_riscv_call_rela }, + [R_RISCV_CALL_PLT] = { apply_r_riscv_call_plt_rela }, + [R_RISCV_GOT_HI20] = { apply_r_riscv_got_hi20_rela }, + [R_RISCV_TLS_GOT_HI20] = { tls_not_supported }, + [R_RISCV_TLS_GD_HI20] = { tls_not_supported }, + [R_RISCV_PCREL_HI20] = { apply_r_riscv_pcrel_hi20_rela }, + [R_RISCV_PCREL_LO12_I] = { apply_r_riscv_pcrel_lo12_i_rela }, + [R_RISCV_PCREL_LO12_S] = { apply_r_riscv_pcrel_lo12_s_rela }, + [R_RISCV_HI20] = { apply_r_riscv_hi20_rela }, + [R_RISCV_LO12_I] = { apply_r_riscv_lo12_i_rela }, + [R_RISCV_LO12_S] = { apply_r_riscv_lo12_s_rela }, + [R_RISCV_TPREL_HI20] = { tls_not_supported }, + [R_RISCV_TPREL_LO12_I] = { tls_not_supported }, + [R_RISCV_TPREL_LO12_S] = { tls_not_supported }, + [R_RISCV_TPREL_ADD] = { tls_not_supported }, + [R_RISCV_ADD8] = { apply_r_riscv_add8_rela, apply_8_bit_accumulation }, + [R_RISCV_ADD16] = { apply_r_riscv_add16_rela, + apply_16_bit_accumulation }, + [R_RISCV_ADD32] = { apply_r_riscv_add32_rela, + apply_32_bit_accumulation }, + [R_RISCV_ADD64] = { apply_r_riscv_add64_rela, + apply_64_bit_accumulation }, + [R_RISCV_SUB8] = { apply_r_riscv_sub8_rela, apply_8_bit_accumulation }, + [R_RISCV_SUB16] = { apply_r_riscv_sub16_rela, + apply_16_bit_accumulation }, + [R_RISCV_SUB32] = { apply_r_riscv_sub32_rela, + apply_32_bit_accumulation }, + [R_RISCV_SUB64] = { apply_r_riscv_sub64_rela, + apply_64_bit_accumulation }, + /* 41-42 reserved for future standard use */ + [R_RISCV_ALIGN] = { apply_r_riscv_align_rela }, + [R_RISCV_RVC_BRANCH] = { apply_r_riscv_rvc_branch_rela }, + [R_RISCV_RVC_JUMP] = { apply_r_riscv_rvc_jump_rela }, + /* 46-50 reserved for future standard use */ + [R_RISCV_RELAX] = { apply_r_riscv_relax_rela }, + [R_RISCV_SUB6] = { apply_r_riscv_sub6_rela, apply_6_bit_accumulation }, + [R_RISCV_SET6] = { apply_r_riscv_set6_rela, apply_6_bit_accumulation }, + [R_RISCV_SET8] = { apply_r_riscv_set8_rela, apply_8_bit_accumulation }, + [R_RISCV_SET16] = { apply_r_riscv_set16_rela, + apply_16_bit_accumulation }, + [R_RISCV_SET32] = { apply_r_riscv_set32_rela, + apply_32_bit_accumulation }, + [R_RISCV_32_PCREL] = { apply_r_riscv_32_pcrel_rela }, + [R_RISCV_IRELATIVE] = { dynamic_linking_not_supported }, + [R_RISCV_PLT32] = { apply_r_riscv_plt32_rela }, + [R_RISCV_SET_ULEB128] = { apply_r_riscv_set_uleb128, + apply_uleb128_accumulation }, + [R_RISCV_SUB_ULEB128] = { apply_r_riscv_sub_uleb128, + apply_uleb128_accumulation }, + /* 62-191 reserved for future standard use */ + /* 192-255 nonstandard ABI extensions */ }; +void process_accumulated_relocations(struct module *me) +{ + /* + * Only ADD/SUB/SET/ULEB128 should end up here. + * + * Each bucket may have more than one relocation location. All + * relocations for a location are stored in a list in a bucket. + * + * Relocations are applied to a temp variable before being stored to the + * provided location to check for overflow. This also allows ULEB128 to + * properly decide how many entries are needed before storing to + * location. The final value is stored into location using the handler + * for the last relocation to an address. + * + * Three layers of indexing: + * - Each of the buckets in use + * - Groups of relocations in each bucket by location address + * - Each relocation entry for a location address + */ + struct used_bucket *bucket_iter; + struct relocation_head *rel_head_iter; + struct relocation_entry *rel_entry_iter; + int curr_type; + void *location; + long buffer; + + list_for_each_entry(bucket_iter, &used_buckets_list, head) { + hlist_for_each_entry(rel_head_iter, bucket_iter->bucket, node) { + buffer = 0; + location = rel_head_iter->location; + list_for_each_entry(rel_entry_iter, + rel_head_iter->rel_entry, head) { + curr_type = rel_entry_iter->type; + reloc_handlers[curr_type].reloc_handler( + me, &buffer, rel_entry_iter->value); + kfree(rel_entry_iter); + } + reloc_handlers[curr_type].accumulate_handler( + me, location, buffer); + kfree(rel_head_iter); + } + kfree(bucket_iter); + } + + kfree(relocation_hashtable); +} + +int add_relocation_to_accumulate(struct module *me, int type, void *location, + unsigned int hashtable_bits, Elf_Addr v) +{ + struct relocation_entry *entry; + struct relocation_head *rel_head; + struct hlist_head *current_head; + struct used_bucket *bucket; + unsigned long hash; + + entry = kmalloc(sizeof(*entry), GFP_KERNEL); + INIT_LIST_HEAD(&entry->head); + entry->type = type; + entry->value = v; + + hash = hash_min((uintptr_t)location, hashtable_bits); + + current_head = &relocation_hashtable[hash]; + + /* Find matching location (if any) */ + bool found = false; + struct relocation_head *rel_head_iter; + + hlist_for_each_entry(rel_head_iter, current_head, node) { + if (rel_head_iter->location == location) { + found = true; + rel_head = rel_head_iter; + break; + } + } + + if (!found) { + rel_head = kmalloc(sizeof(*rel_head), GFP_KERNEL); + rel_head->rel_entry = + kmalloc(sizeof(struct list_head), GFP_KERNEL); + INIT_LIST_HEAD(rel_head->rel_entry); + rel_head->location = location; + INIT_HLIST_NODE(&rel_head->node); + if (!current_head->first) { + bucket = + kmalloc(sizeof(struct used_bucket), GFP_KERNEL); + INIT_LIST_HEAD(&bucket->head); + bucket->bucket = current_head; + list_add(&bucket->head, &used_buckets_list); + } + hlist_add_head(&rel_head->node, current_head); + } + + /* Add relocation to head of discovered rel_head */ + list_add_tail(&entry->head, rel_head->rel_entry); + + return 0; +} + +unsigned int initialize_relocation_hashtable(unsigned int num_relocations) +{ + /* Can safely assume that bits is not greater than sizeof(long) */ + unsigned long hashtable_size = roundup_pow_of_two(num_relocations); + unsigned int hashtable_bits = ilog2(hashtable_size); + + /* + * Double size of hashtable if num_relocations * 1.25 is greater than + * hashtable_size. + */ + int should_double_size = ((num_relocations + (num_relocations >> 2)) > (hashtable_size)); + + hashtable_bits += should_double_size; + + hashtable_size <<= should_double_size; + + relocation_hashtable = kmalloc_array(hashtable_size, + sizeof(*relocation_hashtable), + GFP_KERNEL); + __hash_init(relocation_hashtable, hashtable_size); + + INIT_LIST_HEAD(&used_buckets_list); + + return hashtable_bits; +} + int apply_relocate_add(Elf_Shdr *sechdrs, const char *strtab, unsigned int symindex, unsigned int relsec, struct module *me) @@ -353,11 +736,13 @@ int apply_relocate_add(Elf_Shdr *sechdrs, const char *strtab, unsigned int i, type; Elf_Addr v; int res; + unsigned int num_relocations = sechdrs[relsec].sh_size / sizeof(*rel); + unsigned int hashtable_bits = initialize_relocation_hashtable(num_relocations); pr_debug("Applying relocate section %u to %u\n", relsec, sechdrs[relsec].sh_info); - for (i = 0; i < sechdrs[relsec].sh_size / sizeof(*rel); i++) { + for (i = 0; i < num_relocations; i++) { /* This is where to make the change */ location = (void *)sechdrs[sechdrs[relsec].sh_info].sh_addr + rel[i].r_offset; @@ -375,8 +760,8 @@ int apply_relocate_add(Elf_Shdr *sechdrs, const char *strtab, type = ELF_RISCV_R_TYPE(rel[i].r_info); - if (type < ARRAY_SIZE(reloc_handlers_rela)) - handler = reloc_handlers_rela[type]; + if (type < ARRAY_SIZE(reloc_handlers)) + handler = reloc_handlers[type].reloc_handler; else handler = NULL; @@ -432,11 +817,16 @@ int apply_relocate_add(Elf_Shdr *sechdrs, const char *strtab, } } - res = handler(me, location, v); + if (reloc_handlers[type].accumulate_handler) + res = add_relocation_to_accumulate(me, type, location, hashtable_bits, v); + else + res = handler(me, location, v); if (res) return res; } + process_accumulated_relocations(me); + return 0; } -- Gitee From 41d637bf07161dcc6ab259e979405bc7ee1b6077 Mon Sep 17 00:00:00 2001 From: Charlie Jenkins Date: Wed, 1 Nov 2023 11:33:01 -0700 Subject: [PATCH 22/22] riscv: Add tests for riscv module loading ANBZ: #24284 commit af71bc194916b10f9b394f9b14419d99700a5e67 upstream. Add test cases for the two main groups of relocations added: SUB and SET, along with uleb128. Signed-off-by: Charlie Jenkins Link: https://lore.kernel.org/r/20231101-module_relocations-v9-3-8dfa3483c400@rivosinc.com Signed-off-by: Palmer Dabbelt Signed-off-by: Fangyu Yu --- arch/riscv/Kconfig.debug | 1 + arch/riscv/kernel/Makefile | 1 + arch/riscv/kernel/tests/Kconfig.debug | 35 ++++++++ arch/riscv/kernel/tests/Makefile | 1 + arch/riscv/kernel/tests/module_test/Makefile | 15 ++++ .../module_test/test_module_linking_main.c | 88 +++++++++++++++++++ .../kernel/tests/module_test/test_set16.S | 23 +++++ .../kernel/tests/module_test/test_set32.S | 20 +++++ .../kernel/tests/module_test/test_set6.S | 23 +++++ .../kernel/tests/module_test/test_set8.S | 23 +++++ .../kernel/tests/module_test/test_sub16.S | 20 +++++ .../kernel/tests/module_test/test_sub32.S | 20 +++++ .../kernel/tests/module_test/test_sub6.S | 20 +++++ .../kernel/tests/module_test/test_sub64.S | 25 ++++++ .../kernel/tests/module_test/test_sub8.S | 20 +++++ .../kernel/tests/module_test/test_uleb128.S | 31 +++++++ 16 files changed, 366 insertions(+) create mode 100644 arch/riscv/kernel/tests/Kconfig.debug create mode 100644 arch/riscv/kernel/tests/Makefile create mode 100644 arch/riscv/kernel/tests/module_test/Makefile create mode 100644 arch/riscv/kernel/tests/module_test/test_module_linking_main.c create mode 100644 arch/riscv/kernel/tests/module_test/test_set16.S create mode 100644 arch/riscv/kernel/tests/module_test/test_set32.S create mode 100644 arch/riscv/kernel/tests/module_test/test_set6.S create mode 100644 arch/riscv/kernel/tests/module_test/test_set8.S create mode 100644 arch/riscv/kernel/tests/module_test/test_sub16.S create mode 100644 arch/riscv/kernel/tests/module_test/test_sub32.S create mode 100644 arch/riscv/kernel/tests/module_test/test_sub6.S create mode 100644 arch/riscv/kernel/tests/module_test/test_sub64.S create mode 100644 arch/riscv/kernel/tests/module_test/test_sub8.S create mode 100644 arch/riscv/kernel/tests/module_test/test_uleb128.S diff --git a/arch/riscv/Kconfig.debug b/arch/riscv/Kconfig.debug index e69de29bb2d1..eafe17ebf710 100644 --- a/arch/riscv/Kconfig.debug +++ b/arch/riscv/Kconfig.debug @@ -0,0 +1 @@ +source "arch/riscv/kernel/tests/Kconfig.debug" diff --git a/arch/riscv/kernel/Makefile b/arch/riscv/kernel/Makefile index 2ffd9bf4d911..7e7a7a83c969 100644 --- a/arch/riscv/kernel/Makefile +++ b/arch/riscv/kernel/Makefile @@ -60,6 +60,7 @@ obj-y += stacktrace.o obj-y += cacheinfo.o obj-y += patch.o obj-y += probes/ +obj-y += tests/ obj-$(CONFIG_MMU) += vdso.o vdso/ obj-$(CONFIG_RISCV_M_MODE) += traps_misaligned.o diff --git a/arch/riscv/kernel/tests/Kconfig.debug b/arch/riscv/kernel/tests/Kconfig.debug new file mode 100644 index 000000000000..5dba64e8e977 --- /dev/null +++ b/arch/riscv/kernel/tests/Kconfig.debug @@ -0,0 +1,35 @@ +# SPDX-License-Identifier: GPL-2.0-only +menu "arch/riscv/kernel Testing and Coverage" + +config AS_HAS_ULEB128 + def_bool $(as-instr,.reloc label$(comma) R_RISCV_SET_ULEB128$(comma) 127\n.reloc label$(comma) R_RISCV_SUB_ULEB128$(comma) 127\nlabel:\n.word 0) + +menuconfig RUNTIME_KERNEL_TESTING_MENU + bool "arch/riscv/kernel runtime Testing" + def_bool y + help + Enable riscv kernel runtime testing. + +if RUNTIME_KERNEL_TESTING_MENU + +config RISCV_MODULE_LINKING_KUNIT + bool "KUnit test riscv module linking at runtime" if !KUNIT_ALL_TESTS + depends on KUNIT + default KUNIT_ALL_TESTS + help + Enable this option to test riscv module linking at boot. This will + enable a module called "test_module_linking". + + KUnit tests run during boot and output the results to the debug log + in TAP format (http://testanything.org/). Only useful for kernel devs + running the KUnit test harness, and not intended for inclusion into a + production build. + + For more information on KUnit and unit tests in general please refer + to the KUnit documentation in Documentation/dev-tools/kunit/. + + If unsure, say N. + +endif # RUNTIME_TESTING_MENU + +endmenu # "arch/riscv/kernel runtime Testing" diff --git a/arch/riscv/kernel/tests/Makefile b/arch/riscv/kernel/tests/Makefile new file mode 100644 index 000000000000..7d6c76cffe20 --- /dev/null +++ b/arch/riscv/kernel/tests/Makefile @@ -0,0 +1 @@ +obj-$(CONFIG_RISCV_MODULE_LINKING_KUNIT) += module_test/ diff --git a/arch/riscv/kernel/tests/module_test/Makefile b/arch/riscv/kernel/tests/module_test/Makefile new file mode 100644 index 000000000000..d7a6fd8943de --- /dev/null +++ b/arch/riscv/kernel/tests/module_test/Makefile @@ -0,0 +1,15 @@ +obj-m += test_module_linking.o + +test_sub := test_sub6.o test_sub8.o test_sub16.o test_sub32.o test_sub64.o + +test_set := test_set6.o test_set8.o test_set16.o test_set32.o + +test_module_linking-objs += $(test_sub) + +test_module_linking-objs += $(test_set) + +ifeq ($(CONFIG_AS_HAS_ULEB128),y) +test_module_linking-objs += test_uleb128.o +endif + +test_module_linking-objs += test_module_linking_main.o diff --git a/arch/riscv/kernel/tests/module_test/test_module_linking_main.c b/arch/riscv/kernel/tests/module_test/test_module_linking_main.c new file mode 100644 index 000000000000..8df5fa5b834e --- /dev/null +++ b/arch/riscv/kernel/tests/module_test/test_module_linking_main.c @@ -0,0 +1,88 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (C) 2023 Rivos Inc. + */ + +#include +#include +#include +#include + +MODULE_LICENSE("GPL"); +MODULE_DESCRIPTION("Test module linking"); + +extern int test_set32(void); +extern int test_set16(void); +extern int test_set8(void); +extern int test_set6(void); +extern long test_sub64(void); +extern int test_sub32(void); +extern int test_sub16(void); +extern int test_sub8(void); +extern int test_sub6(void); + +#ifdef CONFIG_AS_HAS_ULEB128 +extern int test_uleb_basic(void); +extern int test_uleb_large(void); +#endif + +#define CHECK_EQ(lhs, rhs) KUNIT_ASSERT_EQ(test, lhs, rhs) + +void run_test_set(struct kunit *test); +void run_test_sub(struct kunit *test); +void run_test_uleb(struct kunit *test); + +void run_test_set(struct kunit *test) +{ + int val32 = test_set32(); + int val16 = test_set16(); + int val8 = test_set8(); + int val6 = test_set6(); + + CHECK_EQ(val32, 0); + CHECK_EQ(val16, 0); + CHECK_EQ(val8, 0); + CHECK_EQ(val6, 0); +} + +void run_test_sub(struct kunit *test) +{ + int val64 = test_sub64(); + int val32 = test_sub32(); + int val16 = test_sub16(); + int val8 = test_sub8(); + int val6 = test_sub6(); + + CHECK_EQ(val64, 0); + CHECK_EQ(val32, 0); + CHECK_EQ(val16, 0); + CHECK_EQ(val8, 0); + CHECK_EQ(val6, 0); +} + +#ifdef CONFIG_AS_HAS_ULEB128 +void run_test_uleb(struct kunit *test) +{ + int val_uleb = test_uleb_basic(); + int val_uleb2 = test_uleb_large(); + + CHECK_EQ(val_uleb, 0); + CHECK_EQ(val_uleb2, 0); +} +#endif + +static struct kunit_case __refdata riscv_module_linking_test_cases[] = { + KUNIT_CASE(run_test_set), + KUNIT_CASE(run_test_sub), +#ifdef CONFIG_AS_HAS_ULEB128 + KUNIT_CASE(run_test_uleb), +#endif + {} +}; + +static struct kunit_suite riscv_module_linking_test_suite = { + .name = "riscv_checksum", + .test_cases = riscv_module_linking_test_cases, +}; + +kunit_test_suites(&riscv_module_linking_test_suite); diff --git a/arch/riscv/kernel/tests/module_test/test_set16.S b/arch/riscv/kernel/tests/module_test/test_set16.S new file mode 100644 index 000000000000..2be0e441a12e --- /dev/null +++ b/arch/riscv/kernel/tests/module_test/test_set16.S @@ -0,0 +1,23 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * Copyright (C) 2023 Rivos Inc. + */ + +.text +.global test_set16 +test_set16: + lw a0, set16 + la t0, set16 +#ifdef CONFIG_32BIT + slli t0, t0, 16 + srli t0, t0, 16 +#else + slli t0, t0, 48 + srli t0, t0, 48 +#endif + sub a0, a0, t0 + ret +.data +set16: + .reloc set16, R_RISCV_SET16, set16 + .word 0 diff --git a/arch/riscv/kernel/tests/module_test/test_set32.S b/arch/riscv/kernel/tests/module_test/test_set32.S new file mode 100644 index 000000000000..de0444537e67 --- /dev/null +++ b/arch/riscv/kernel/tests/module_test/test_set32.S @@ -0,0 +1,20 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * Copyright (C) 2023 Rivos Inc. + */ + +.text +.global test_set32 +test_set32: + lw a0, set32 + la t0, set32 +#ifndef CONFIG_32BIT + slli t0, t0, 32 + srli t0, t0, 32 +#endif + sub a0, a0, t0 + ret +.data +set32: + .reloc set32, R_RISCV_SET32, set32 + .word 0 diff --git a/arch/riscv/kernel/tests/module_test/test_set6.S b/arch/riscv/kernel/tests/module_test/test_set6.S new file mode 100644 index 000000000000..c39ce4c219eb --- /dev/null +++ b/arch/riscv/kernel/tests/module_test/test_set6.S @@ -0,0 +1,23 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * Copyright (C) 2023 Rivos Inc. + */ + +.text +.global test_set6 +test_set6: + lw a0, set6 + la t0, set6 +#ifdef CONFIG_32BIT + slli t0, t0, 26 + srli t0, t0, 26 +#else + slli t0, t0, 58 + srli t0, t0, 58 +#endif + sub a0, a0, t0 + ret +.data +set6: + .reloc set6, R_RISCV_SET6, set6 + .word 0 diff --git a/arch/riscv/kernel/tests/module_test/test_set8.S b/arch/riscv/kernel/tests/module_test/test_set8.S new file mode 100644 index 000000000000..a656173f6f99 --- /dev/null +++ b/arch/riscv/kernel/tests/module_test/test_set8.S @@ -0,0 +1,23 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * Copyright (C) 2023 Rivos Inc. + */ + +.text +.global test_set8 +test_set8: + lw a0, set8 + la t0, set8 +#ifdef CONFIG_32BIT + slli t0, t0, 24 + srli t0, t0, 24 +#else + slli t0, t0, 56 + srli t0, t0, 56 +#endif + sub a0, a0, t0 + ret +.data +set8: + .reloc set8, R_RISCV_SET8, set8 + .word 0 diff --git a/arch/riscv/kernel/tests/module_test/test_sub16.S b/arch/riscv/kernel/tests/module_test/test_sub16.S new file mode 100644 index 000000000000..80f731d599ba --- /dev/null +++ b/arch/riscv/kernel/tests/module_test/test_sub16.S @@ -0,0 +1,20 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * Copyright (C) 2023 Rivos Inc. + */ + +.text +.global test_sub16 +test_sub16: + lh a0, sub16 + addi a0, a0, -32 + ret +first: + .space 32 +second: + +.data +sub16: + .reloc sub16, R_RISCV_ADD16, second + .reloc sub16, R_RISCV_SUB16, first + .half 0 diff --git a/arch/riscv/kernel/tests/module_test/test_sub32.S b/arch/riscv/kernel/tests/module_test/test_sub32.S new file mode 100644 index 000000000000..a341686e12df --- /dev/null +++ b/arch/riscv/kernel/tests/module_test/test_sub32.S @@ -0,0 +1,20 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * Copyright (C) 2023 Rivos Inc. + */ + +.text +.global test_sub32 +test_sub32: + lw a0, sub32 + addi a0, a0, -32 + ret +first: + .space 32 +second: + +.data +sub32: + .reloc sub32, R_RISCV_ADD32, second + .reloc sub32, R_RISCV_SUB32, first + .word 0 diff --git a/arch/riscv/kernel/tests/module_test/test_sub6.S b/arch/riscv/kernel/tests/module_test/test_sub6.S new file mode 100644 index 000000000000..e8b61c1ec527 --- /dev/null +++ b/arch/riscv/kernel/tests/module_test/test_sub6.S @@ -0,0 +1,20 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * Copyright (C) 2023 Rivos Inc. + */ + +.text +.global test_sub6 +test_sub6: + lb a0, sub6 + addi a0, a0, -32 + ret +first: + .space 32 +second: + +.data +sub6: + .reloc sub6, R_RISCV_SET6, second + .reloc sub6, R_RISCV_SUB6, first + .byte 0 diff --git a/arch/riscv/kernel/tests/module_test/test_sub64.S b/arch/riscv/kernel/tests/module_test/test_sub64.S new file mode 100644 index 000000000000..a59e8afa88fd --- /dev/null +++ b/arch/riscv/kernel/tests/module_test/test_sub64.S @@ -0,0 +1,25 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * Copyright (C) 2023 Rivos Inc. + */ + +.text +.global test_sub64 +test_sub64: +#ifdef CONFIG_32BIT + lw a0, sub64 +#else + ld a0, sub64 +#endif + addi a0, a0, -32 + ret +first: + .space 32 +second: + +.data +sub64: + .reloc sub64, R_RISCV_ADD64, second + .reloc sub64, R_RISCV_SUB64, first + .word 0 + .word 0 diff --git a/arch/riscv/kernel/tests/module_test/test_sub8.S b/arch/riscv/kernel/tests/module_test/test_sub8.S new file mode 100644 index 000000000000..ac5d0ec98de3 --- /dev/null +++ b/arch/riscv/kernel/tests/module_test/test_sub8.S @@ -0,0 +1,20 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * Copyright (C) 2023 Rivos Inc. + */ + +.text +.global test_sub8 +test_sub8: + lb a0, sub8 + addi a0, a0, -32 + ret +first: + .space 32 +second: + +.data +sub8: + .reloc sub8, R_RISCV_ADD8, second + .reloc sub8, R_RISCV_SUB8, first + .byte 0 diff --git a/arch/riscv/kernel/tests/module_test/test_uleb128.S b/arch/riscv/kernel/tests/module_test/test_uleb128.S new file mode 100644 index 000000000000..90f22049d553 --- /dev/null +++ b/arch/riscv/kernel/tests/module_test/test_uleb128.S @@ -0,0 +1,31 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * Copyright (C) 2023 Rivos Inc. + */ + +.text +.global test_uleb_basic +test_uleb_basic: + ld a0, second + addi a0, a0, -127 + ret + +.global test_uleb_large +test_uleb_large: + ld a0, fourth + addi a0, a0, -0x07e8 + ret + +.data +first: + .space 127 +second: + .reloc second, R_RISCV_SET_ULEB128, second + .reloc second, R_RISCV_SUB_ULEB128, first + .dword 0 +third: + .space 1000 +fourth: + .reloc fourth, R_RISCV_SET_ULEB128, fourth + .reloc fourth, R_RISCV_SUB_ULEB128, third + .dword 0 -- Gitee