diff options
35 files changed, 916 insertions, 420 deletions
diff --git a/arch/x86/kernel/cpu/perf_event.c b/arch/x86/kernel/cpu/perf_event.c index bd874302420..c1bbed1021d 100644 --- a/arch/x86/kernel/cpu/perf_event.c +++ b/arch/x86/kernel/cpu/perf_event.c @@ -2229,10 +2229,10 @@ validate_event(struct cpu_hw_events *cpuc, struct perf_event *event) { struct hw_perf_event fake_event = event->hw; - if (event->pmu != &pmu) + if (event->pmu && event->pmu != &pmu) return 0; - return x86_schedule_event(cpuc, &fake_event); + return x86_schedule_event(cpuc, &fake_event) >= 0; } static int validate_group(struct perf_event *event) diff --git a/kernel/perf_event.c b/kernel/perf_event.c index 9425c9600c8..35df94e344f 100644 --- a/kernel/perf_event.c +++ b/kernel/perf_event.c @@ -1831,7 +1831,7 @@ static int perf_event_read_group(struct perf_event *event, size = n * sizeof(u64); - if (copy_to_user(buf + size, values, size)) { + if (copy_to_user(buf + ret, values, size)) { ret = -EFAULT; goto unlock; } @@ -3914,7 +3914,7 @@ void perf_swevent_put_recursion_context(int rctx) { struct perf_cpu_context *cpuctx = &__get_cpu_var(perf_cpu_context); barrier(); - cpuctx->recursion[rctx]++; + cpuctx->recursion[rctx]--; put_cpu_var(perf_cpu_context); } EXPORT_SYMBOL_GPL(perf_swevent_put_recursion_context); diff --git a/tools/perf/.gitignore b/tools/perf/.gitignore index 0854f110bf7..fe08660ce0b 100644 --- a/tools/perf/.gitignore +++ b/tools/perf/.gitignore @@ -12,6 +12,7 @@ perf*.1 perf*.xml perf*.html common-cmds.h +perf.data tags TAGS cscope* diff --git a/tools/perf/Documentation/perf-kmem.txt b/tools/perf/Documentation/perf-kmem.txt new file mode 100644 index 00000000000..44b0ce35c28 --- /dev/null +++ b/tools/perf/Documentation/perf-kmem.txt @@ -0,0 +1,44 @@ +perf-kmem(1) +============== + +NAME +---- +perf-kmem - Tool to trace/measure kernel memory(slab) properties + +SYNOPSIS +-------- +[verse] +'perf kmem' {record} [<options>] + +DESCRIPTION +----------- +There's two variants of perf kmem: + + 'perf kmem record <command>' to record the kmem events + of an arbitrary workload. + + 'perf kmem' to report kernel memory statistics. + +OPTIONS +------- +-i <file>:: +--input=<file>:: + Select the input file (default: perf.data) + +--stat=<caller|alloc>:: + Select per callsite or per allocation statistics + +-s <key[,key2...]>:: +--sort=<key[,key2...]>:: + Sort the output (default: frag,hit,bytes) + +-l <num>:: +--line=<num>:: + Print n lines only + +--raw-ip:: + Print raw ip instead of symbol + +SEE ALSO +-------- +linkperf:perf-record[1] diff --git a/tools/perf/Makefile b/tools/perf/Makefile index 3ef6621bf6c..de37d492e10 100644 --- a/tools/perf/Makefile +++ b/tools/perf/Makefile @@ -369,6 +369,7 @@ LIB_H += util/sort.h LIB_H += util/hist.h LIB_H += util/thread.h LIB_H += util/data_map.h +LIB_H += util/process_events.h LIB_OBJS += util/abspath.o LIB_OBJS += util/alias.o @@ -411,6 +412,7 @@ LIB_OBJS += util/svghelper.o LIB_OBJS += util/sort.o LIB_OBJS += util/hist.o LIB_OBJS += util/data_map.o +LIB_OBJS += util/process_events.o BUILTIN_OBJS += builtin-annotate.o @@ -419,6 +421,7 @@ BUILTIN_OBJS += builtin-bench.o # Benchmark modules BUILTIN_OBJS += bench/sched-messaging.o BUILTIN_OBJS += bench/sched-pipe.o +BUILTIN_OBJS += bench/mem-memcpy.o BUILTIN_OBJS += builtin-help.o BUILTIN_OBJS += builtin-sched.o diff --git a/tools/perf/bench/bench.h b/tools/perf/bench/bench.h index 9fbd8d745fa..f7781c6267c 100644 --- a/tools/perf/bench/bench.h +++ b/tools/perf/bench/bench.h @@ -3,6 +3,7 @@ extern int bench_sched_messaging(int argc, const char **argv, const char *prefix); extern int bench_sched_pipe(int argc, const char **argv, const char *prefix); +extern int bench_mem_memcpy(int argc, const char **argv, const char *prefix __used); #define BENCH_FORMAT_DEFAULT_STR "default" #define BENCH_FORMAT_DEFAULT 0 diff --git a/tools/perf/bench/mem-memcpy.c b/tools/perf/bench/mem-memcpy.c new file mode 100644 index 00000000000..89773178e89 --- /dev/null +++ b/tools/perf/bench/mem-memcpy.c @@ -0,0 +1,193 @@ +/* + * mem-memcpy.c + * + * memcpy: Simple memory copy in various ways + * + * Written by Hitoshi Mitake <mitake@dcl.info.waseda.ac.jp> + */ +#include <ctype.h> + +#include "../perf.h" +#include "../util/util.h" +#include "../util/parse-options.h" +#include "../util/string.h" +#include "../util/header.h" +#include "bench.h" + +#include <stdio.h> +#include <stdlib.h> +#include <string.h> +#include <sys/time.h> +#include <errno.h> + +#define K 1024 + +static const char *length_str = "1MB"; +static const char *routine = "default"; +static int use_clock = 0; +static int clock_fd; + +static const struct option options[] = { + OPT_STRING('l', "length", &length_str, "1MB", + "Specify length of memory to copy. " + "available unit: B, MB, GB (upper and lower)"), + OPT_STRING('r', "routine", &routine, "default", + "Specify routine to copy"), + OPT_BOOLEAN('c', "clock", &use_clock, + "Use CPU clock for measuring"), + OPT_END() +}; + +struct routine { + const char *name; + const char *desc; + void * (*fn)(void *dst, const void *src, size_t len); +}; + +struct routine routines[] = { + { "default", + "Default memcpy() provided by glibc", + memcpy }, + { NULL, + NULL, + NULL } +}; + +static const char * const bench_mem_memcpy_usage[] = { + "perf bench mem memcpy <options>", + NULL +}; + +static struct perf_event_attr clock_attr = { + .type = PERF_TYPE_HARDWARE, + .config = PERF_COUNT_HW_CPU_CYCLES +}; + +static void init_clock(void) +{ + clock_fd = sys_perf_event_open(&clock_attr, getpid(), -1, -1, 0); + + if (clock_fd < 0 && errno == ENOSYS) + die("No CONFIG_PERF_EVENTS=y kernel support configured?\n"); + else + BUG_ON(clock_fd < 0); +} + +static u64 get_clock(void) +{ + int ret; + u64 clk; + + ret = read(clock_fd, &clk, sizeof(u64)); + BUG_ON(ret != sizeof(u64)); + + return clk; +} + +static double timeval2double(struct timeval *ts) +{ + return (double)ts->tv_sec + + (double)ts->tv_usec / (double)1000000; +} + +int bench_mem_memcpy(int argc, const char **argv, + const char *prefix __used) +{ + int i; + void *dst, *src; + size_t length; + double bps = 0.0; + struct timeval tv_start, tv_end, tv_diff; + u64 clock_start, clock_end, clock_diff; + + clock_start = clock_end = clock_diff = 0ULL; + argc = parse_options(argc, argv, options, + bench_mem_memcpy_usage, 0); + + tv_diff.tv_sec = 0; + tv_diff.tv_usec = 0; + length = (size_t)perf_atoll((char *)length_str); + + if ((s64)length <= 0) { + fprintf(stderr, "Invalid length:%s\n", length_str); + return 1; + } + + for (i = 0; routines[i].name; i++) { + if (!strcmp(routines[i].name, routine)) + break; + } + if (!routines[i].name) { + printf("Unknown routine:%s\n", routine); + printf("Available routines...\n"); + for (i = 0; routines[i].name; i++) { + printf("\t%s ... %s\n", + routines[i].name, routines[i].desc); + } + return 1; + } + + dst = zalloc(length); + if (!dst) + die("memory allocation failed - maybe length is too large?\n"); + + src = zalloc(length); + if (!src) + die("memory allocation failed - maybe length is too large?\n"); + + if (bench_format == BENCH_FORMAT_DEFAULT) { + printf("# Copying %s Bytes from %p to %p ...\n\n", + length_str, src, dst); + } + + if (use_clock) { + init_clock(); + clock_start = get_clock(); + } else { + BUG_ON(gettimeofday(&tv_start, NULL)); + } + + routines[i].fn(dst, src, length); + + if (use_clock) { + clock_end = get_clock(); + clock_diff = clock_end - clock_start; + } else { + BUG_ON(gettimeofday(&tv_end, NULL)); + timersub(&tv_end, &tv_start, &tv_diff); + bps = (double)((double)length / timeval2double(&tv_diff)); + } + + switch (bench_format) { + case BENCH_FORMAT_DEFAULT: + if (use_clock) { + printf(" %14lf Clock/Byte\n", + (double)clock_diff / (double)length); + } else { + if (bps < K) + printf(" %14lf B/Sec\n", bps); + else if (bps < K * K) + printf(" %14lfd KB/Sec\n", bps / 1024); + else if (bps < K * K * K) + printf(" %14lf MB/Sec\n", bps / 1024 / 1024); + else { + printf(" %14lf GB/Sec\n", + bps / 1024 / 1024 / 1024); + } + } + break; + case BENCH_FORMAT_SIMPLE: + if (use_clock) { + printf("%14lf\n", + (double)clock_diff / (double)length); + } else + printf("%lf\n", bps); + break; + default: + /* reaching this means there's some disaster: */ + die("unknown format: %d\n", bench_format); + break; + } + + return 0; +} diff --git a/tools/perf/builtin-annotate.c b/tools/perf/builtin-annotate.c index 6b13a1ecf1e..18ac5eaefc3 100644 --- a/tools/perf/builtin-annotate.c +++ b/tools/perf/builtin-annotate.c @@ -24,6 +24,7 @@ #include "util/thread.h" #include "util/sort.h" #include "util/hist.h" +#include "util/process_events.h" static char const *input_name = "perf.data"; @@ -33,11 +34,9 @@ static int input; static int full_paths; static int print_line; -static bool use_modules; static unsigned long page_size; static unsigned long mmap_window = 32; -const char *vmlinux_name; struct sym_hist { u64 sum; @@ -55,6 +54,11 @@ struct sym_priv { struct sym_ext *ext; }; +static struct symbol_conf symbol_conf = { + .priv_size = sizeof(struct sym_priv), + .try_vmlinux_path = true, +}; + static const char *sym_hist_filter; static int symbol_filter(struct map *map __used, struct symbol *sym) @@ -158,7 +162,7 @@ process_sample_event(event_t *event, unsigned long offset, unsigned long head) if (event->header.misc & PERF_RECORD_MISC_KERNEL) { level = 'k'; - sym = kernel_maps__find_symbol(ip, &map, symbol_filter); + sym = kernel_maps__find_function(ip, &map, symbol_filter); dump_printf(" ...... dso: %s\n", map ? map->dso->long_name : "<not found>"); } else if (event->header.misc & PERF_RECORD_MISC_USER) { @@ -167,7 +171,7 @@ process_sample_event(event_t *event, unsigned long offset, unsigned long head) if (map != NULL) { got_map: ip = map->map_ip(map, ip); - sym = map__find_symbol(map, ip, symbol_filter); + sym = map__find_function(map, ip, symbol_filter); } else { /* * If this is outside of all known maps, @@ -202,32 +206,6 @@ got_map: } static int -process_mmap_event(event_t *event, unsigned long offset, unsigned long head) -{ - struct map *map = map__new(&event->mmap, NULL, 0); - struct thread *thread = threads__findnew(event->mmap.pid); - - dump_printf("%p [%p]: PERF_RECORD_MMAP %d: [%p(%p) @ %p]: %s\n", - (void *)(offset + head), - (void *)(long)(event->header.size), - event->mmap.pid, - (void *)(long)event->mmap.start, - (void *)(long)event->mmap.len, - (void *)(long)event->mmap.pgoff, - event->mmap.filename); - - if (thread == NULL || map == NULL) { - dump_printf("problem processing PERF_RECORD_MMAP, skipping event.\n"); - return 0; - } - - thread__insert_map(thread, map); - total_mmap++; - - return 0; -} - -static int process_comm_event(event_t *event, unsigned long offset, unsigned long head) { struct thread *thread = threads__findnew(event->comm.pid); @@ -248,33 +226,6 @@ process_comm_event(event_t *event, unsigned long offset, unsigned long head) } static int -process_fork_event(event_t *event, unsigned long offset, unsigned long head) -{ - struct thread *thread = threads__findnew(event->fork.pid); - struct thread *parent = threads__findnew(event->fork.ppid); - - dump_printf("%p [%p]: PERF_RECORD_FORK: %d:%d\n", - (void *)(offset + head), - (void *)(long)(event->header.size), - event->fork.pid, event->fork.ppid); - - /* - * A thread clone will have the same PID for both - * parent and child. - */ - if (thread == parent) - return 0; - - if (!thread || !parent || thread__fork(thread, parent)) { - dump_printf("problem processing PERF_RECORD_FORK, skipping event.\n"); - return -1; - } - total_fork++; - - return 0; -} - -static int process_event(event_t *event, unsigned long offset, unsigned long head) { switch (event->header.type) { @@ -288,7 +239,7 @@ process_event(event_t *event, unsigned long offset, unsigned long head) return process_comm_event(event, offset, head); case PERF_RECORD_FORK: - return process_fork_event(event, offset, head); + return process_task_event(event, offset, head); /* * We dont process them right now but they are fine: */ @@ -638,11 +589,6 @@ static int __cmd_annotate(void) exit(0); } - if (kernel_maps__init(vmlinux_name, true, use_modules) < 0) { - pr_err("failed to create kernel maps for symbol resolution\b"); - return -1; - } - remap: buf = (char *)mmap(NULL, page_size * mmap_window, PROT_READ, MAP_SHARED, input, offset); @@ -743,8 +689,9 @@ static const struct option options[] = { "be more verbose (show symbol address, etc)"), OPT_BOOLEAN('D', "dump-raw-trace", &dump_trace, "dump raw trace in ASCII"), - OPT_STRING('k', "vmlinux", &vmlinux_name, "file", "vmlinux pathname"), - OPT_BOOLEAN('m', "modules", &use_modules, + OPT_STRING('k', "vmlinux", &symbol_conf.vmlinux_name, + "file", "vmlinux pathname"), + OPT_BOOLEAN('m', "modules", &symbol_conf.use_modules, "load module symbols - WARNING: use only with -k and LIVE kernel"), OPT_BOOLEAN('l', "print-line", &print_line, "print matching source lines (may be slow)"), @@ -770,7 +717,8 @@ static void setup_sorting(void) int cmd_annotate(int argc, const char **argv, const char *prefix __used) { - symbol__init(sizeof(struct sym_priv)); + if (symbol__init(&symbol_conf) < 0) + return -1; page_size = getpagesize(); diff --git a/tools/perf/builtin-bench.c b/tools/perf/builtin-bench.c index 90c39baae0d..e043eb83092 100644 --- a/tools/perf/builtin-bench.c +++ b/tools/perf/builtin-bench.c @@ -12,6 +12,7 @@ * * Available subsystem list: * sched ... scheduler and IPC mechanism + * mem ... memory access performance * */ @@ -43,6 +44,15 @@ static struct bench_suite sched_suites[] = { NULL } }; +static struct bench_suite mem_suites[] = { + { "memcpy", + "Simple memory copy in various ways", + bench_mem_memcpy }, + { NULL, + NULL, + NULL } +}; + struct bench_subsys { const char *name; const char *summary; @@ -53,9 +63,12 @@ static struct bench_subsys subsystems[] = { { "sched", "scheduler and IPC mechanism", sched_suites }, + { "mem", + "memory access performance", + mem_suites }, { NULL, NULL, - NULL } + NULL } }; static void dump_suites(int subsys_index) diff --git a/tools/perf/builtin-help.c b/tools/perf/builtin-help.c index 768f9c82631..9f810b17c25 100644 --- a/tools/perf/builtin-help.c +++ b/tools/perf/builtin-help.c @@ -179,7 +179,7 @@ static void add_man_viewer(const char *name) while (*p) p = &((*p)->next); - *p = calloc(1, (sizeof(**p) + len + 1)); + *p = zalloc(sizeof(**p) + len + 1); strncpy((*p)->name, name, len); } @@ -194,7 +194,7 @@ static void do_add_man_viewer_info(const char *name, size_t len, const char *value) { - struct man_viewer_info_list *new = calloc(1, sizeof(*new) + len + 1); + struct man_viewer_info_list *new = zalloc(sizeof(*new) + len + 1); strncpy(new->name, name, len); new->info = strdup(value); diff --git a/tools/perf/builtin-kmem.c b/tools/perf/builtin-kmem.c index 256d18fa047..35722fafc4d 100644 --- a/tools/perf/builtin-kmem.c +++ b/tools/perf/builtin-kmem.c @@ -26,26 +26,28 @@ static u64 sample_type; static int alloc_flag; static int caller_flag; -sort_fn_t alloc_sort_fn; -sort_fn_t caller_sort_fn; - static int alloc_lines = -1; static int caller_lines = -1; +static bool raw_ip; + +static char default_sort_order[] = "frag,hit,bytes"; + static char *cwd; static int cwdlen; +static int *cpunode_map; +static int max_cpu_num; + struct alloc_stat { - union { - struct { - char *name; - u64 call_site; - }; - u64 ptr; - }; + u64 call_site; + u64 ptr; u64 bytes_req; u64 bytes_alloc; u32 hit; + u32 pingpong; + + short alloc_cpu; struct rb_node node; }; @@ -56,12 +58,74 @@ static struct rb_root root_caller_stat; static struct rb_root root_caller_sorted; static unsigned long total_requested, total_allocated; +static unsigned long nr_allocs, nr_cross_allocs; struct raw_event_sample { u32 size; char data[0]; }; +#define PATH_SYS_NODE "/sys/devices/system/node" + +static void init_cpunode_map(void) +{ + FILE *fp; + int i; + + fp = fopen("/sys/devices/system/cpu/kernel_max", "r"); + if (!fp) { + max_cpu_num = 4096; + return; + } + + if (fscanf(fp, "%d", &max_cpu_num) < 1) + die("Failed to read 'kernel_max' from sysfs"); + max_cpu_num++; + + cpunode_map = calloc(max_cpu_num, sizeof(int)); + if (!cpunode_map) + die("calloc"); + for (i = 0; i < max_cpu_num; i++) + cpunode_map[i] = -1; + fclose(fp); +} + +static void setup_cpunode_map(void) +{ + struct dirent *dent1, *dent2; + DIR *dir1, *dir2; + unsigned int cpu, mem; + char buf[PATH_MAX]; + + init_cpunode_map(); + + dir1 = opendir(PATH_SYS_NODE); + if (!dir1) + return; + + while (true) { + dent1 = readdir(dir1); + if (!dent1) + break; + + if (sscanf(dent1->d_name, "node%u", &mem) < 1) + continue; + + snprintf(buf, PATH_MAX, "%s/%s", PATH_SYS_NODE, dent1->d_name); + dir2 = opendir(buf); + if (!dir2) + continue; + while (true) { + dent2 = readdir(dir2); + if (!dent2) + break; + if (sscanf(dent2->d_name, "cpu%u", &cpu) < 1) + continue; + cpunode_map[cpu] = mem; + } + } +} + static int process_comm_event(event_t *event, unsigned long offset, unsigned long head) { @@ -81,16 +145,13 @@ process_comm_event(event_t *event, unsigned long offset, unsigned long head) return 0; } -static void insert_alloc_stat(unsigned long ptr, - int bytes_req, int bytes_alloc) +static void insert_alloc_stat(unsigned long call_site, unsigned long ptr, + int bytes_req, int bytes_alloc, int cpu) { struct rb_node **node = &root_alloc_stat.rb_node; struct rb_node *parent = NULL; struct alloc_stat *data = NULL; - if (!alloc_flag) - return; - while (*node) { parent = *node; data = rb_entry(*node, struct alloc_stat, node); @@ -109,7 +170,10 @@ static void insert_alloc_stat(unsigned long ptr, data->bytes_alloc += bytes_req; } else { data = malloc(sizeof(*data)); + if (!data) + die("malloc"); data->ptr = ptr; + data->pingpong = 0; data->hit = 1; data->bytes_req = bytes_req; data->bytes_alloc = bytes_alloc; @@ -117,6 +181,8 @@ static void insert_alloc_stat(unsigned long ptr, rb_link_node(&data->node, parent, node); rb_insert_color(&data->node, &root_alloc_stat); } + data->call_site = call_site; + data->alloc_cpu = cpu; } static void insert_caller_stat(unsigned long call_site, @@ -126,9 +192,6 @@ static void insert_caller_stat(unsigned long call_site, struct rb_node *parent = NULL; struct alloc_stat *data = NULL; - if (!caller_flag) - return; - while (*node) { parent = *node; data = rb_entry(*node, struct alloc_stat, node); @@ -147,7 +210,10 @@ static void insert_caller_stat(unsigned long call_site, data->bytes_alloc += bytes_req; } else { data = malloc(sizeof(*data)); + if (!data) + die("malloc"); data->call_site = call_site; + data->pingpong = 0; data->hit = 1; data->bytes_req = bytes_req; data->bytes_alloc = bytes_alloc; @@ -159,34 +225,89 @@ static void insert_caller_stat(unsigned long call_site, static void process_alloc_event(struct raw_event_sample *raw, struct event *event, - int cpu __used, + int cpu, u64 timestamp __used, struct thread *thread __used, - int node __used) + int node) { unsigned long call_site; unsigned long ptr; int bytes_req; int bytes_alloc; + int node1, node2; ptr = raw_field_value(event, "ptr", raw->data); call_site = raw_field_value(event, "call_site", raw->data); bytes_req = raw_field_value(event, "bytes_req", raw->data); bytes_alloc = raw_field_value(event, "bytes_alloc", raw->data); - insert_alloc_stat(ptr, bytes_req, bytes_alloc); + insert_alloc_stat(call_site, ptr, bytes_req, bytes_alloc, cpu); insert_caller_stat(call_site, bytes_req, bytes_alloc); total_requested += bytes_req; total_allocated += bytes_alloc; + + if (node) { + node1 = cpunode_map[cpu]; + node2 = raw_field_value(event, "node", raw->data); + if (node1 != node2) + nr_cross_allocs++; + } + nr_allocs++; +} + +static int ptr_cmp(struct alloc_stat *, struct alloc_stat *); +static int callsite_cmp(struct alloc_stat *, struct alloc_stat *); + +static struct alloc_stat *search_alloc_stat(unsigned long ptr, + unsigned long call_site, + struct rb_root *root, + sort_fn_t sort_fn) +{ + struct rb_node *node = root->rb_node; + struct alloc_stat key = { .ptr = ptr, .call_site = call_site }; + + while (node) { + struct alloc_stat *data; + int cmp; + + data = rb_entry(node, struct alloc_stat, node); + + cmp = sort_fn(&key, data); + if (cmp < 0) + node = node->rb_left; + else if (cmp > 0) + node = node->rb_right; + else + return data; + } + return NULL; } -static void process_free_event(struct raw_event_sample *raw __used, - struct event *event __used, - int cpu __used, +static void process_free_event(struct raw_event_sample *raw, + struct event *event, + int cpu, u64 timestamp __used, struct thread *thread __used) { + unsigned long ptr; + struct alloc_stat *s_alloc, *s_caller; + + ptr = raw_field_value(event, "ptr", raw->data); + + s_alloc = search_alloc_stat(ptr, 0, &root_alloc_stat, ptr_cmp); + if (!s_alloc) + return; + + if (cpu != s_alloc->alloc_cpu) { + s_alloc->pingpong++; + + s_caller = search_alloc_stat(0, s_alloc->call_site, + &root_caller_stat, callsite_cmp); + assert(s_caller); + s_caller->pingpong++; + } + s_alloc->alloc_cpu = -1; } static void @@ -291,7 +412,7 @@ static int read_events(void) register_idle_thread(); register_perf_file_handler(&file_handler); - return mmap_dispatch_perf_file(&header, input_name, NULL, false, 0, 0, + return mmap_dispatch_perf_file(&header, input_name, 0, 0, &cwdlen, &cwd); } @@ -307,10 +428,10 @@ static void __print_result(struct rb_root *root, int n_lines, int is_caller) { struct rb_node *next; - printf("%.78s\n", graph_dotted_line); - printf("%-28s|", is_caller ? "Callsite": "Alloc Ptr"); - printf("Total_alloc/Per | Total_req/Per | Hit | Frag\n"); - printf("%.78s\n", graph_dotted_line); + printf("%.102s\n", graph_dotted_line); + printf(" %-34s |", is_caller ? "Callsite": "Alloc Ptr"); + printf(" Total_alloc/Per | Total_req/Per | Hit | Ping-pong | Frag\n"); + printf("%.102s\n", graph_dotted_line); next = rb_first(root); @@ -318,36 +439,39 @@ static void __print_result(struct rb_root *root, int n_lines, int is_caller) struct alloc_stat *data = rb_entry(next, struct alloc_stat, node); struct symbol *sym = NULL; - char bf[BUFSIZ]; + char buf[BUFSIZ]; u64 addr; if (is_caller) { addr = data->call_site; - sym = kernel_maps__find_symbol(addr, NULL, NULL); + if (!raw_ip) + sym = kernel_maps__find_function(addr, NULL, NULL); } else addr = data->ptr; if (sym != NULL) - snprintf(bf, sizeof(bf), "%s/%Lx", sym->name, + snprintf(buf, sizeof(buf), "%s+%Lx", sym->name, addr - sym->start); else - snprintf(bf, sizeof(bf), "%#Lx", addr); + snprintf(buf, sizeof(buf), "%#Lx", addr); + printf(" %-34s |", buf); - printf("%-28s|%8llu/%-6lu |%8llu/%-6lu|%6lu|%8.3f%%\n", - bf, (unsigned long long)data->bytes_alloc, + printf(" %9llu/%-5lu | %9llu/%-5lu | %6lu | %8lu | %6.3f%%\n", + (unsigned long long)data->bytes_alloc, (unsigned long)data->bytes_alloc / data->hit, (unsigned long long)data->bytes_req, (unsigned long)data->bytes_req / data->hit, (unsigned long)data->hit, + (unsigned long)data->pingpong, fragmentation(data->bytes_req, data->bytes_alloc)); next = rb_next(next); } if (n_lines == -1) - printf(" ... | ... | ... | ... | ... \n"); + printf(" ... | ... | ... | ... | ... | ... \n"); - printf(" ------------------------------------------------------------------------------\n"); + printf("%.102s\n", graph_dotted_line); } static void print_summary(void) @@ -359,6 +483,7 @@ static void print_summary(void) total_allocated - total_requested); printf("Internal fragmentation: %f%%\n", fragmentation(total_requested, total_allocated)); + printf("Cross CPU allocations: %lu/%lu\n", nr_cross_allocs, nr_allocs); } static void print_result(void) @@ -370,20 +495,34 @@ static void print_result(void) print_summary(); } +struct sort_dimension { + const char name[20]; + sort_fn_t cmp; + struct list_head list; +}; + +static LIST_HEAD(caller_sort); +static LIST_HEAD(alloc_sort); + static void sort_insert(struct rb_root *root, struct alloc_stat *data, - sort_fn_t sort_fn) + struct list_head *sort_list) { struct rb_node **new = &(root->rb_node); struct rb_node *parent = NULL; + struct sort_dimension *sort; while (*new) { struct alloc_stat *this; - int cmp; + int cmp = 0; this = rb_entry(*new, struct alloc_stat, node); parent = *new; - cmp = sort_fn(data, this); + list_for_each_entry(sort, sort_list, list) { + cmp = sort->cmp(data, this); + if (cmp) + break; + } if (cmp > 0) new = &((*new)->rb_left); @@ -396,7 +535,7 @@ static void sort_insert(struct rb_root *root, struct alloc_stat *data, } static void __sort_result(struct rb_root *root, struct rb_root *root_sorted, - sort_fn_t sort_fn) + struct list_head *sort_list) { struct rb_node *node; struct alloc_stat *data; @@ -408,14 +547,14 @@ static void __sort_result(struct rb_root *root, struct rb_root *root_sorted, rb_erase(node, root); data = rb_entry(node, struct alloc_stat, node); - sort_insert(root_sorted, data, sort_fn); + sort_insert(root_sorted, data, sort_list); } } static void sort_result(void) { - __sort_result(&root_alloc_stat, &root_alloc_sorted, alloc_sort_fn); - __sort_result(&root_caller_stat, &root_caller_sorted, caller_sort_fn); + __sort_result(&root_alloc_stat, &root_alloc_sorted, &alloc_sort); + __sort_result(&root_caller_stat, &root_caller_sorted, &caller_sort); } static int __cmd_kmem(void) @@ -433,7 +572,6 @@ static const char * const kmem_usage[] = { NULL }; - static int ptr_cmp(struct alloc_stat *l, struct alloc_stat *r) { if (l->ptr < r->ptr) @@ -443,6 +581,11 @@ static int ptr_cmp(struct alloc_stat *l, struct alloc_stat *r) return 0; } +static struct sort_dimension ptr_sort_dimension = { + .name = "ptr", + .cmp = ptr_cmp, +}; + static int callsite_cmp(struct alloc_stat *l, struct alloc_stat *r) { if (l->call_site < r->call_site) @@ -452,6 +595,11 @@ static int callsite_cmp(struct alloc_stat *l, struct alloc_stat *r) return 0; } +static struct sort_dimension callsite_sort_dimension = { + .name = "callsite", + .cmp = callsite_cmp, +}; + static int hit_cmp(struct alloc_stat *l, struct alloc_stat *r) { if (l->hit < r->hit) @@ -461,6 +609,11 @@ static int hit_cmp(struct alloc_stat *l, struct alloc_stat *r) return 0; } +static struct sort_dimension hit_sort_dimension = { + .name = "hit", + .cmp = hit_cmp, +}; + static int bytes_cmp(struct alloc_stat *l, struct alloc_stat *r) { if (l->bytes_alloc < r->bytes_alloc) @@ -470,6 +623,11 @@ static int bytes_cmp(struct alloc_stat *l, struct alloc_stat *r) return 0; } +static struct sort_dimension bytes_sort_dimension = { + .name = "bytes", + .cmp = bytes_cmp, +}; + static int frag_cmp(struct alloc_stat *l, struct alloc_stat *r) { double x, y; @@ -484,31 +642,88 @@ static int frag_cmp(struct alloc_stat *l, struct alloc_stat *r) return 0; } +static struct sort_dimension frag_sort_dimension = { + .name = "frag", + .cmp = frag_cmp, +}; + +static int pingpong_cmp(struct alloc_stat *l, struct alloc_stat *r) +{ + if (l->pingpong < r->pingpong) + return -1; + else if (l->pingpong > r->pingpong) + return 1; + return 0; +} + +static struct sort_dimension pingpong_sort_dimension = { + .name = "pingpong", + .cmp = pingpong_cmp, +}; + +static struct sort_dimension *avail_sorts[] = { + &ptr_sort_dimension, + &callsite_sort_dimension, + &hit_sort_dimension, + &bytes_sort_dimension, + &frag_sort_dimension, + &pingpong_sort_dimension, +}; + +#define NUM_AVAIL_SORTS \ + (int)(sizeof(avail_sorts) / sizeof(struct sort_dimension *)) + +static int sort_dimension__add(const char *tok, struct list_head *list) +{ + struct sort_dimension *sort; + int i; + + for (i = 0; i < NUM_AVAIL_SORTS; i++) { + if (!strcmp(avail_sorts[i]->name, tok)) { + sort = malloc(sizeof(*sort)); + if (!sort) + die("malloc"); + memcpy(sort, avail_sorts[i], sizeof(*sort)); + list_add_tail(&sort->list, list); + return 0; + } + } + + return -1; +} + +static int setup_sorting(struct list_head *sort_list, const char *arg) +{ + char *tok; + char *str = strdup(arg); + + if (!str) + die("strdup"); + + while (true) { + tok = strsep(&str, ","); + if (!tok) + break; + if (sort_dimension__add(tok, sort_list) < 0) { + error("Unknown --sort key: '%s'", tok); + return -1; + } + } + + free(str); + return 0; +} + static int parse_sort_opt(const struct option *opt __used, const char *arg, int unset __used) { - sort_fn_t sort_fn; - if (!arg) return -1; - if (strcmp(arg, "ptr") == 0) - sort_fn = ptr_cmp; - else if (strcmp(arg, "call_site") == 0) - sort_fn = callsite_cmp; - else if (strcmp(arg, "hit") == 0) - sort_fn = hit_cmp; - else if (strcmp(arg, "bytes") == 0) - sort_fn = bytes_cmp; - else if (strcmp(arg, "frag") == 0) - sort_fn = frag_cmp; - else - return -1; - if (caller_flag > alloc_flag) - caller_sort_fn = sort_fn; + return setup_sorting(&caller_sort, arg); else - alloc_sort_fn = sort_fn; + return setup_sorting(&alloc_sort, arg); return 0; } @@ -552,12 +767,13 @@ static const struct option kmem_options[] = { OPT_CALLBACK(0, "stat", NULL, "<alloc>|<caller>", "stat selector, Pass 'alloc' or 'caller'.", parse_stat_opt), - OPT_CALLBACK('s', "sort", NULL, "key", - "sort by key: ptr, call_site, hit, bytes, frag", + OPT_CALLBACK('s', "sort", NULL, "key[,key2...]", + "sort by keys: ptr, call_site, bytes, hit, pingpong, frag", parse_sort_opt), OPT_CALLBACK('l', "line", NULL, "num", "show n lins", parse_line_opt), + OPT_BOOLEAN(0, "raw-ip", &raw_ip, "show raw ip instead of symbol"), OPT_END() }; @@ -604,10 +820,12 @@ int cmd_kmem(int argc, const char **argv, const char *prefix __used) else if (argc) usage_with_options(kmem_usage, kmem_options); - if (!alloc_sort_fn) - alloc_sort_fn = bytes_cmp; - if (!caller_sort_fn) - caller_sort_fn = bytes_cmp; + if (list_empty(&caller_sort)) + setup_sorting(&caller_sort, default_sort_order); + if (list_empty(&alloc_sort)) + setup_sorting(&alloc_sort, default_sort_order); + + setup_cpunode_map(); return __cmd_kmem(); } diff --git a/tools/perf/builtin-probe.c b/tools/perf/builtin-probe.c index d78a3d94549..a2f6daf01ec 100644 --- a/tools/perf/builtin-probe.c +++ b/tools/perf/builtin-probe.c @@ -309,9 +309,9 @@ static int synthesize_probe_event(struct probe_point *pp) { char *buf; int i, len, ret; - pp->probes[0] = buf = (char *)calloc(MAX_CMDLEN, sizeof(char)); + pp->probes[0] = buf = zalloc(MAX_CMDLEN); if (!buf) - die("Failed to allocate memory by calloc."); + die("Failed to allocate memory by zalloc."); ret = snprintf(buf, MAX_CMDLEN, "%s+%d", pp->function, pp->offset); if (ret <= 0 || ret >= MAX_CMDLEN) goto error; diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c index fe474b7f8ad..e4b1004e76e 100644 --- a/tools/perf/builtin-report.c +++ b/tools/perf/builtin-report.c @@ -30,6 +30,7 @@ #include "util/thread.h" #include "util/sort.h" #include "util/hist.h" +#include "util/process_events.h" static char const *input_name = "perf.data"; @@ -38,7 +39,6 @@ static char *dso_list_str, *comm_list_str, *sym_list_str, static struct strlist *dso_list, *comm_list, *sym_list; static int force; -static bool use_modules; static int full_paths; static int show_nr_samples; @@ -52,15 +52,13 @@ static char *pretty_printing_style = default_pretty_printing_style; static int exclude_other = 1; static char callchain_default_opt[] = "fractal,0.5"; -const char *vmlinux_name; - -static char *cwd; -static int cwdlen; static struct perf_header *header; static u64 sample_type; +struct symbol_conf symbol_conf; + static size_t callchain__fprintf_left_margin(FILE *fp, int left_margin) @@ -450,14 +448,14 @@ got_map: * trick of looking in the whole kernel symbol list. */ if ((long long)ip < 0) - return kernel_maps__find_symbol(ip, mapp, NULL); + return kernel_maps__find_function(ip, mapp, NULL); } dump_printf(" ...... dso: %s\n", map ? map->dso->long_name : "<not found>"); dump_printf(" ...... map: %Lx -> %Lx\n", *ipp, ip); *ipp = ip; - return map ? map__find_symbol(map, ip, NULL) : NULL; + return map ? map__find_function(map, ip, NULL) : NULL; } static int call__match(struct symbol *sym) @@ -497,7 +495,7 @@ static struct symbol **resolve_callchain(struct thread *thread, case PERF_CONTEXT_HV: break; case PERF_CONTEXT_KERNEL: - sym = kernel_maps__find_symbol(ip, NULL, NULL); + sym = kernel_maps__find_function(ip, NULL, NULL); break; default: sym = resolve_symbol(thread, NULL, &ip); @@ -717,7 +715,7 @@ process_sample_event(event_t *event, unsigned long offset, unsigned long head) if (cpumode == PERF_RECORD_MISC_KERNEL) { level = 'k'; - sym = kernel_maps__find_symbol(ip, &map, NULL); + sym = kernel_maps__find_function(ip, &map, NULL); dump_printf(" ...... dso: %s\n", map ? map->dso->long_name : "<not found>"); } else if (cpumode == PERF_RECORD_MISC_USER) { @@ -751,33 +749,6 @@ process_sample_event(event_t *event, unsigned long offset, unsigned long head) } static int -process_mmap_event(event_t *event, unsigned long offset, unsigned long head) -{ - struct map *map = map__new(&event->mmap, cwd, cwdlen); - struct thread *thread = threads__findnew(event->mmap.pid); - - dump_printf("%p [%p]: PERF_RECORD_MMAP %d/%d: [%p(%p) @ %p]: %s\n", - (void *)(offset + head), - (void *)(long)(event->header.size), - event->mmap.pid, - event->mmap.tid, - (void *)(long)event->mmap.start, - (void *)(long)event->mmap.len, - (void *)(long)event->mmap.pgoff, - event->mmap.filename); - - if (thread == NULL || map == NULL) { - dump_printf("problem processing PERF_RECORD_MMAP, skipping event.\n"); - return 0; - } - - thread__insert_map(thread, map); - total_mmap++; - - return 0; -} - -static int process_comm_event(event_t *event, unsigned long offset, unsigned long head) { struct thread *thread = threads__findnew(event->comm.pid); @@ -798,38 +769,6 @@ process_comm_event(event_t *event, unsigned long offset, unsigned long head) } static int -process_task_event(event_t *event, unsigned long offset, unsigned long head) -{ - struct thread *thread = threads__findnew(event->fork.pid); - struct thread *parent = threads__findnew(event->fork.ppid); - - dump_printf("%p [%p]: PERF_RECORD_%s: (%d:%d):(%d:%d)\n", - (void *)(offset + head), - (void *)(long)(event->header.size), - event->header.type == PERF_RECORD_FORK ? "FORK" : "EXIT", - event->fork.pid, event->fork.tid, - event->fork.ppid, event->fork.ptid); - - /* - * A thread clone will have the same PID for both - * parent and child. - */ - if (thread == parent) - return 0; - - if (event->header.type == PERF_RECORD_EXIT) - return 0; - - if (!thread || !parent || thread__fork(thread, parent)) { - dump_printf("problem processing PERF_RECORD_FORK, skipping event.\n"); - return -1; - } - total_fork++; - - return 0; -} - -static int process_lost_event(event_t *event, unsigned long offset, unsigned long head) { dump_printf("%p [%p]: PERF_RECORD_LOST: id:%Ld: lost:%Ld\n", @@ -926,8 +865,7 @@ static int __cmd_report(void) register_perf_file_handler(&file_handler); - ret = mmap_dispatch_perf_file(&header, input_name, vmlinux_name, - !vmlinux_name, force, + ret = mmap_dispatch_perf_file(&header, input_name, force, full_paths, &cwdlen, &cwd); if (ret) return ret; @@ -1024,9 +962,10 @@ static const struct option options[] = { "be more verbose (show symbol address, etc)"), OPT_BOOLEAN('D', "dump-raw-trace", &dump_trace, "dump raw trace in ASCII"), - OPT_STRING('k', "vmlinux", &vmlinux_name, "file", "vmlinux pathname"), + OPT_STRING('k', "vmlinux", &symbol_conf.vmlinux_name, + "file", "vmlinux pathname"), OPT_BOOLEAN('f', "force", &force, "don't complain, do it"), - OPT_BOOLEAN('m', "modules", &use_modules, + OPT_BOOLEAN('m', "modules", &symbol_conf.use_modules, "load module symbols - WARNING: use only with -k and LIVE kernel"), OPT_BOOLEAN('n', "show-nr-samples", &show_nr_samples, "Show a column with the number of samples"), @@ -1096,7 +1035,8 @@ static void setup_list(struct strlist **list, const char *list_str, int cmd_report(int argc, const char **argv, const char *prefix __used) { - symbol__init(0); + if (symbol__init(&symbol_conf) < 0) + return -1; argc = parse_options(argc, argv, options, report_usage, 0); diff --git a/tools/perf/builtin-sched.c b/tools/perf/builtin-sched.c index 260f57a72ee..19eb708a706 100644 --- a/tools/perf/builtin-sched.c +++ b/tools/perf/builtin-sched.c @@ -225,7 +225,7 @@ static void calibrate_sleep_measurement_overhead(void) static struct sched_atom * get_new_event(struct task_desc *task, u64 timestamp) { - struct sched_atom *event = calloc(1, sizeof(*event)); + struct sched_atom *event = zalloc(sizeof(*event)); unsigned long idx = task->nr_events; size_t size; @@ -293,7 +293,7 @@ add_sched_event_wakeup(struct task_desc *task, u64 timestamp, return; } - wakee_event->wait_sem = calloc(1, sizeof(*wakee_event->wait_sem)); + wakee_event->wait_sem = zalloc(sizeof(*wakee_event->wait_sem)); sem_init(wakee_event->wait_sem, 0, 0); wakee_event->specific_wait = 1; event->wait_sem = wakee_event->wait_sem; @@ -323,7 +323,7 @@ static struct task_desc *register_pid(unsigned long pid, const char *comm) if (task) return task; - task = calloc(1, sizeof(*task)); + task = zalloc(sizeof(*task)); task->pid = pid; task->nr = nr_tasks; strcpy(task->comm, comm); @@ -962,9 +962,7 @@ __thread_latency_insert(struct rb_root *root, struct work_atoms *data, static void thread_atoms_insert(struct thread *thread) { - struct work_atoms *atoms; - - atoms = calloc(sizeof(*atoms), 1); + struct work_atoms *atoms = zalloc(sizeof(*atoms)); if (!atoms) die("No memory"); @@ -996,9 +994,7 @@ add_sched_out_event(struct work_atoms *atoms, char run_state, u64 timestamp) { - struct work_atom *atom; - - atom = calloc(sizeof(*atom), 1); + struct work_atom *atom = zalloc(sizeof(*atom)); if (!atom) die("Non memory"); @@ -1718,7 +1714,7 @@ static int read_events(void) register_idle_thread(); register_perf_file_handler(&file_handler); - return mmap_dispatch_perf_file(&header, input_name, NULL, false, 0, 0, + return mmap_dispatch_perf_file(&header, input_name, 0, 0, &cwdlen, &cwd); } diff --git a/tools/perf/builtin-top.c b/tools/perf/builtin-top.c index 6a5de90e9b8..ded6cf65ad9 100644 --- a/tools/perf/builtin-top.c +++ b/tools/perf/builtin-top.c @@ -79,7 +79,7 @@ static int dump_symtab = 0; static bool hide_kernel_symbols = false; static bool hide_user_symbols = false; static struct winsize winsize; -const char *vmlinux_name; +struct symbol_conf symbol_conf; /* * Source @@ -128,7 +128,7 @@ struct sym_entry { static inline struct symbol *sym_entry__symbol(struct sym_entry *self) { - return ((void *)self) + symbol__priv_size; + return ((void *)self) + symbol_conf.priv_size; } static void get_term_dimensions(struct winsize *ws) @@ -181,7 +181,7 @@ static void parse_source(struct sym_entry *syme) return; if (syme->src == NULL) { - syme->src = calloc(1, sizeof(*source)); + syme->src = zalloc(sizeof(*source)); if (syme->src == NULL) return; pthread_mutex_init(&syme->src->lock, NULL); @@ -451,9 +451,8 @@ static void print_sym_table(void) struct sym_entry *syme, *n; struct rb_root tmp = RB_ROOT; struct rb_node *nd; - int sym_width = 0, dso_width = 0; + int sym_width = 0, dso_width = 0, max_dso_width; const int win_width = winsize.ws_col - 1; - struct dso *unique_dso = NULL, *first_dso = NULL; samples = userspace_samples = 0; @@ -539,11 +538,6 @@ static void print_sym_table(void) (int)syme->snap_count < count_filter) continue; - if (first_dso == NULL) - unique_dso = first_dso = syme->map->dso; - else if (syme->map->dso != first_dso) - unique_dso = NULL; - if (syme->map->dso->long_name_len > dso_width) dso_width = syme->map->dso->long_name_len; @@ -553,14 +547,10 @@ static void print_sym_table(void) printed = 0; - if (unique_dso) - printf("DSO: %s\n", unique_dso->long_name); - else { - int max_dso_width = winsize.ws_col - sym_width - 29; - if (dso_width > max_dso_width) - dso_width = max_dso_width; - putchar('\n'); - } + max_dso_width = winsize.ws_col - sym_width - 29; + if (dso_width > max_dso_width) + dso_width = max_dso_width; + putchar('\n'); if (nr_counters == 1) printf(" samples pcnt"); else @@ -568,17 +558,13 @@ static void print_sym_table(void) if (verbose) printf(" RIP "); - printf(" %-*.*s", sym_width, sym_width, "function"); - if (!unique_dso) - printf(" DSO"); - putchar('\n'); + printf(" %-*.*s DSO\n", sym_width, sym_width, "function"); printf(" %s _______ _____", nr_counters == 1 ? " " : "______"); if (verbose) printf(" ________________"); printf(" %-*.*s", sym_width, sym_width, graph_line); - if (!unique_dso) - printf(" %-*.*s", dso_width, dso_width, graph_line); + printf(" %-*.*s", dso_width, dso_width, graph_line); puts("\n"); for (nd = rb_first(&tmp); nd; nd = rb_next(nd)) { @@ -603,12 +589,10 @@ static void print_sym_table(void) if (verbose) printf(" %016llx", sym->start); printf(" %-*.*s", sym_width, sym_width, sym->name); - if (!unique_dso) - printf(" %-*.*s", dso_width, dso_width, - dso_width >= syme->map->dso->long_name_len ? - syme->map->dso->long_name : - syme->map->dso->short_name); - printf("\n"); + printf(" %-*.*s\n", dso_width, dso_width, + dso_width >= syme->map->dso->long_name_len ? + syme->map->dso->long_name : + syme->map->dso->short_name); } } @@ -711,7 +695,7 @@ static void print_mapped_keys(void) fprintf(stdout, "\t[f] profile display filter (count). \t(%d)\n", count_filter); - if (vmlinux_name) { + if (symbol_conf.vmlinux_name) { fprintf(stdout, "\t[F] annotate display filter (percent). \t(%d%%)\n", sym_pcnt_filter); fprintf(stdout, "\t[s] annotate symbol. \t(%s)\n", name?: "NULL"); fprintf(stdout, "\t[S] stop annotation.\n"); @@ -748,7 +732,7 @@ static int key_mapped(int c) case 'F': case 's': case 'S': - return vmlinux_name ? 1 : 0; + return symbol_conf.vmlinux_name ? 1 : 0; default: break; } @@ -964,7 +948,7 @@ static void event__process_sample(const event_t *self, int counter) map = thread__find_map(thread, ip); if (map != NULL) { ip = map->map_ip(map, ip); - sym = map__find_symbol(map, ip, symbol_filter); + sym = map__find_function(map, ip, symbol_filter); if (sym == NULL) return; userspace_samples++; @@ -984,7 +968,7 @@ static void event__process_sample(const event_t *self, int counter) if (hide_kernel_symbols) return; - sym = kernel_maps__find_symbol(ip, &map, symbol_filter); + sym = kernel_maps__find_function(ip, &map, symbol_filter); if (sym == NULL) return; break; @@ -1277,7 +1261,8 @@ static const struct option options[] = { "system-wide collection from all CPUs"), OPT_INTEGER('C', "CPU", &profile_cpu, "CPU to profile on"), - OPT_STRING('k', "vmlinux", &vmlinux_name, "file", "vmlinux pathname"), + OPT_STRING('k', "vmlinux", &symbol_conf.vmlinux_name, + "file", "vmlinux pathname"), OPT_BOOLEAN('K', "hide_kernel_symbols", &hide_kernel_symbols, "hide kernel symbols"), OPT_INTEGER('m', "mmap-pages", &mmap_pages, @@ -1311,7 +1296,7 @@ static const struct option options[] = { int cmd_top(int argc, const char **argv, const char *prefix __used) { - int counter, err; + int counter; page_size = sysconf(_SC_PAGE_SIZE); @@ -1329,15 +1314,16 @@ int cmd_top(int argc, const char **argv, const char *prefix __used) if (!nr_counters) nr_counters = 1; - symbol__init(sizeof(struct sym_entry) + - (nr_counters + 1) * sizeof(unsigned long)); + symbol_conf.priv_size = (sizeof(struct sym_entry) + + (nr_counters + 1) * sizeof(unsigned long)); + if (symbol_conf.vmlinux_name == NULL) + symbol_conf.try_vmlinux_path = true; + if (symbol__init(&symbol_conf) < 0) + return -1; if (delay_secs < 1) delay_secs = 1; - err = kernel_maps__init(vmlinux_name, !vmlinux_name, true); - if (err < 0) - return err; parse_source(sym_filter_entry); /* diff --git a/tools/perf/builtin-trace.c b/tools/perf/builtin-trace.c index b71198e5dc1..75972fd073d 100644 --- a/tools/perf/builtin-trace.c +++ b/tools/perf/builtin-trace.c @@ -131,7 +131,7 @@ static int __cmd_trace(void) register_idle_thread(); register_perf_file_handler(&file_handler); - return mmap_dispatch_perf_file(&header, input_name, NULL, false, + return mmap_dispatch_perf_file(&header, input_name, 0, 0, &cwdlen, &cwd); } diff --git a/tools/perf/command-list.txt b/tools/perf/command-list.txt index d3a6e18e4a5..02b09ea17a3 100644 --- a/tools/perf/command-list.txt +++ b/tools/perf/command-list.txt @@ -14,3 +14,4 @@ perf-timechart mainporcelain common perf-top mainporcelain common perf-trace mainporcelain common perf-probe mainporcelain common +perf-kmem mainporcelain common diff --git a/tools/perf/util/data_map.c b/tools/perf/util/data_map.c index f318d19b256..b238462b898 100644 --- a/tools/perf/util/data_map.c +++ b/tools/perf/util/data_map.c @@ -101,8 +101,6 @@ out: int mmap_dispatch_perf_file(struct perf_header **pheader, const char *input_name, - const char *vmlinux_name, - bool try_vmlinux_path, int force, int full_paths, int *cwdlen, @@ -172,12 +170,6 @@ int mmap_dispatch_perf_file(struct perf_header **pheader, curr_handler->sample_type_check(sample_type) < 0) goto out_delete; - err = -ENOMEM; - if (kernel_maps__init(vmlinux_name, try_vmlinux_path, true) < 0) { - pr_err("failed to setup the kernel maps to resolve symbols\n"); - goto out_delete; - } - if (!full_paths) { if (getcwd(__cwd, sizeof(__cwd)) == NULL) { pr_err("failed to get the current directory\n"); diff --git a/tools/perf/util/data_map.h b/tools/perf/util/data_map.h index 3f0d21b3819..ae036ecd762 100644 --- a/tools/perf/util/data_map.h +++ b/tools/perf/util/data_map.h @@ -23,8 +23,6 @@ struct perf_file_handler { void register_perf_file_handler(struct perf_file_handler *handler); int mmap_dispatch_perf_file(struct perf_header **pheader, const char *input_name, - const char *vmlinux_name, - bool try_vmlinux_path, int force, int full_paths, int *cwdlen, diff --git a/tools/perf/util/event.h b/tools/perf/util/event.h index f1e39261265..882a9531db9 100644 --- a/tools/perf/util/event.h +++ b/tools/perf/util/event.h @@ -119,9 +119,10 @@ void map__delete(struct map *self); struct map *map__clone(struct map *self); int map__overlap(struct map *l, struct map *r); size_t map__fprintf(struct map *self, FILE *fp); -struct symbol *map__find_symbol(struct map *self, u64 ip, symbol_filter_t filter); -void map__fixup_start(struct map *self); -void map__fixup_end(struct map *self); +struct symbol *map__find_function(struct map *self, u64 ip, + symbol_filter_t filter); +void map__fixup_start(struct map *self, struct rb_root *symbols); +void map__fixup_end(struct map *self, struct rb_root *symbols); int event__synthesize_thread(pid_t pid, int (*process)(event_t *event)); void event__synthesize_threads(int (*process)(event_t *event)); diff --git a/tools/perf/util/header.c b/tools/perf/util/header.c index 1332f8ec04a..4b586569bb0 100644 --- a/tools/perf/util/header.c +++ b/tools/perf/util/header.c @@ -63,7 +63,7 @@ int perf_header_attr__add_id(struct perf_header_attr *self, u64 id) */ struct perf_header *perf_header__new(void) { - struct perf_header *self = calloc(sizeof(*self), 1); + struct perf_header *self = zalloc(sizeof(*self)); if (self != NULL) { self->size = 1; @@ -253,12 +253,6 @@ static int perf_header__adds_write(struct perf_header *self, int fd) buildid_sec = &feat_sec[idx++]; - /* - * Read the kernel buildid nad the list of loaded modules with - * its build_ids: - */ - kernel_maps__init(NULL, false, true); - /* Write build-ids */ buildid_sec->offset = lseek(fd, 0, SEEK_CUR); err = dsos__write_buildid_table(fd); diff --git a/tools/perf/util/include/asm/bug.h b/tools/perf/util/include/asm/bug.h new file mode 100644 index 00000000000..7fcc6810adc --- /dev/null +++ b/tools/perf/util/include/asm/bug.h @@ -0,0 +1,22 @@ +#ifndef _PERF_ASM_GENERIC_BUG_H +#define _PERF_ASM_GENERIC_BUG_H + +#define __WARN_printf(arg...) do { fprintf(stderr, arg); } while (0) + +#define WARN(condition, format...) ({ \ + int __ret_warn_on = !!(condition); \ + if (unlikely(__ret_warn_on)) \ + __WARN_printf(format); \ + unlikely(__ret_warn_on); \ +}) + +#define WARN_ONCE(condition, format...) ({ \ + static int __warned; \ + int __ret_warn_once = !!(condition); \ + \ + if (unlikely(__ret_warn_once)) \ + if (WARN(!__warned, format)) \ + __warned = 1; \ + unlikely(__ret_warn_once); \ +}) +#endif diff --git a/tools/perf/util/include/linux/bitops.h b/tools/perf/util/include/linux/bitops.h index ace57c36d1d..8d63116e943 100644 --- a/tools/perf/util/include/linux/bitops.h +++ b/tools/perf/util/include/linux/bitops.h @@ -7,6 +7,8 @@ #define CONFIG_GENERIC_FIND_FIRST_BIT #include "../../../../include/linux/bitops.h" +#undef __KERNEL__ + static inline void set_bit(int nr, unsigned long *addr) { addr[nr / BITS_PER_LONG] |= 1UL << (nr % BITS_PER_LONG); diff --git a/tools/perf/util/map.c b/tools/perf/util/map.c index 09412321a80..41c5c4a2001 100644 --- a/tools/perf/util/map.c +++ b/tools/perf/util/map.c @@ -80,18 +80,18 @@ void map__delete(struct map *self) free(self); } -void map__fixup_start(struct map *self) +void map__fixup_start(struct map *self, struct rb_root *symbols) { - struct rb_node *nd = rb_first(&self->dso->syms); + struct rb_node *nd = rb_first(symbols); if (nd != NULL) { struct symbol *sym = rb_entry(nd, struct symbol, rb_node); self->start = sym->start; } } -void map__fixup_end(struct map *self) +void map__fixup_end(struct map *self, struct rb_root *symbols) { - struct rb_node *nd = rb_last(&self->dso->syms); + struct rb_node *nd = rb_last(symbols); if (nd != NULL) { struct symbol *sym = rb_entry(nd, struct symbol, rb_node); self->end = sym->end; @@ -100,8 +100,8 @@ void map__fixup_end(struct map *self) #define DSO__DELETED "(deleted)" -struct symbol * -map__find_symbol(struct map *self, u64 ip, symbol_filter_t filter) +struct symbol *map__find_function(struct map *self, u64 ip, + symbol_filter_t filter) { if (!self->dso->loaded) { int nr = dso__load(self->dso, self, filter); @@ -136,7 +136,7 @@ map__find_symbol(struct map *self, u64 ip, symbol_filter_t filter) } } - return self->dso->find_symbol(self->dso, ip); + return self->dso->find_function(self->dso, ip); } struct map *map__clone(struct map *self) diff --git a/tools/perf/util/parse-events.c b/tools/perf/util/parse-events.c index 07002746927..9e5dbd66d34 100644 --- a/tools/perf/util/parse-events.c +++ b/tools/perf/util/parse-events.c @@ -197,7 +197,7 @@ struct tracepoint_path *tracepoint_id_to_path(u64 config) if (id == config) { closedir(evt_dir); closedir(sys_dir); - path = calloc(1, sizeof(path)); + path = zalloc(sizeof(path)); path->system = malloc(MAX_EVENT_LENGTH); if (!path->system) { free(path); diff --git a/tools/perf/util/process_event.c b/tools/perf/util/process_event.c new file mode 100644 index 00000000000..a970789581a --- /dev/null +++ b/tools/perf/util/process_event.c @@ -0,0 +1,53 @@ +#include "process_event.h" + +char *cwd; +int cwdlen; + +int +process_mmap_event(event_t *event, unsigned long offset, unsigned long head) +{ + struct map *map = map__new(&event->mmap, cwd, cwdlen); + struct thread *thread = threads__findnew(event->mmap.pid); + + dump_printf("%p [%p]: PERF_RECORD_MMAP %d/%d: [%p(%p) @ %p]: %s\n", + (void *)(offset + head), + (void *)(long)(event->header.size), + event->mmap.pid, + event->mmap.tid, + (void *)(long)event->mmap.start, + (void *)(long)event->mmap.len, + (void *)(long)event->mmap.pgoff, + event->mmap.filename); + + if (thread == NULL || map == NULL) { + dump_printf("problem processing PERF_RECORD_MMAP, skipping event.\n"); + return 0; + } + + thread__insert_map(thread, map); + total_mmap++; + + return 0; + +} + +int +process_comm_event(event_t *event, unsigned long offset, unsigned long head) +{ + struct thread *thread = threads__findnew(event->comm.pid); + + dump_printf("%p [%p]: PERF_RECORD_COMM: %s:%d\n", + (void *)(offset + head), + (void *)(long)(event->header.size), + event->comm.comm, event->comm.pid); + + if (thread == NULL || + thread__set_comm_adjust(thread, event->comm.comm)) { + dump_printf("problem processing PERF_RECORD_COMM, skipping event.\n"); + return -1; + } + total_comm++; + + return 0; +} + diff --git a/tools/perf/util/process_event.h b/tools/perf/util/process_event.h new file mode 100644 index 00000000000..6f68c69736c --- /dev/null +++ b/tools/perf/util/process_event.h @@ -0,0 +1,29 @@ +#ifndef __PROCESS_EVENT_H +#define __PROCESS_EVENT_H + +#include "../builtin.h" +#include "util.h" + +#include "color.h" +#include <linux/list.h> +#include "cache.h" +#include <linux/rbtree.h> +#include "symbol.h" +#include "string.h" + +#include "../perf.h" +#include "debug.h" + +#include "parse-options.h" +#include "parse-events.h" + +#include "thread.h" +#include "sort.h" +#include "hist.h" + +extern char *cwd; +extern int cwdlen; +extern int process_mmap_event(event_t *, unsigned long, unsigned long); +extern int process_comm_event(event_t *, unsigned long , unsigned long); + +#endif /* __PROCESS_H */ diff --git a/tools/perf/util/process_events.c b/tools/perf/util/process_events.c new file mode 100644 index 00000000000..a9204363efd --- /dev/null +++ b/tools/perf/util/process_events.c @@ -0,0 +1,64 @@ +#include "process_events.h" + +char *cwd; +int cwdlen; + +int +process_mmap_event(event_t *event, unsigned long offset, unsigned long head) +{ + struct map *map = map__new(&event->mmap, cwd, cwdlen); + struct thread *thread = threads__findnew(event->mmap.pid); + + dump_printf("%p [%p]: PERF_RECORD_MMAP %d/%d: [%p(%p) @ %p]: %s\n", + (void *)(offset + head), + (void *)(long)(event->header.size), + event->mmap.pid, + event->mmap.tid, + (void *)(long)event->mmap.start, + (void *)(long)event->mmap.len, + (void *)(long)event->mmap.pgoff, + event->mmap.filename); + + if (thread == NULL || map == NULL) { + dump_printf("problem processing PERF_RECORD_MMAP, skipping event.\n"); + return 0; + } + + thread__insert_map(thread, map); + total_mmap++; + + return 0; +} + +int +process_task_event(event_t *event, unsigned long offset, unsigned long head) +{ + struct thread *thread = threads__findnew(event->fork.pid); + struct thread *parent = threads__findnew(event->fork.ppid); + + dump_printf("%p [%p]: PERF_RECORD_%s: (%d:%d):(%d:%d)\n", + (void *)(offset + head), + (void *)(long)(event->header.size), + event->header.type == PERF_RECORD_FORK ? "FORK" : "EXIT", + event->fork.pid, event->fork.tid, + event->fork.ppid, event->fork.ptid); + + /* + * A thread clone will have the same PID for both + * parent and child. + */ + if (thread == parent) + return 0; + + if (event->header.type == PERF_RECORD_EXIT) + return 0; + + if (!thread || !parent || thread__fork(thread, parent)) { + dump_printf("problem processing PERF_RECORD_FORK, skipping event.\n"); + return -1; + } + total_fork++; + + return 0; +} + diff --git a/tools/perf/util/process_events.h b/tools/perf/util/process_events.h new file mode 100644 index 00000000000..73d092f8328 --- /dev/null +++ b/tools/perf/util/process_events.h @@ -0,0 +1,35 @@ +#ifndef __PROCESS_EVENTS_H +#define __PROCESS_EVENTS_H + +#include "../builtin.h" + +#include "util.h" +#include "color.h" +#include <linux/list.h> +#include "cache.h" +#include <linux/rbtree.h> +#include "symbol.h" +#include "string.h" +#include "callchain.h" +#include "strlist.h" +#include "values.h" + +#include "../perf.h" +#include "debug.h" +#include "header.h" + +#include "parse-options.h" +#include "parse-events.h" + +#include "data_map.h" +#include "thread.h" +#include "sort.h" +#include "hist.h" + +extern char *cwd; +extern int cwdlen; + +extern int process_mmap_event(event_t *, unsigned long , unsigned long); +extern int process_task_event(event_t *, unsigned long, unsigned long); + +#endif /* __PROCESS_EVENTS_H */ diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c index 44d81d5ae8c..4ed379b915f 100644 --- a/tools/perf/util/symbol.c +++ b/tools/perf/util/symbol.c @@ -6,6 +6,7 @@ #include "debug.h" +#include <asm/bug.h> #include <libelf.h> #include <gelf.h> #include <elf.h> @@ -37,11 +38,16 @@ unsigned int symbol__priv_size; static int vmlinux_path__nr_entries; static char **vmlinux_path; +static struct symbol_conf symbol_conf__defaults = { + .use_modules = true, + .try_vmlinux_path = true, +}; + static struct rb_root kernel_maps; -static void dso__fixup_sym_end(struct dso *self) +static void symbols__fixup_end(struct rb_root *self) { - struct rb_node *nd, *prevnd = rb_first(&self->syms); + struct rb_node *nd, *prevnd = rb_first(self); struct symbol *curr, *prev; if (prevnd == NULL) @@ -88,15 +94,14 @@ static void kernel_maps__fixup_end(void) static struct symbol *symbol__new(u64 start, u64 len, const char *name) { size_t namelen = strlen(name) + 1; - struct symbol *self = calloc(1, (symbol__priv_size + - sizeof(*self) + namelen)); - if (!self) + struct symbol *self = zalloc(symbol__priv_size + + sizeof(*self) + namelen); + if (self == NULL) return NULL; - if (symbol__priv_size) { - memset(self, 0, symbol__priv_size); + if (symbol__priv_size) self = ((void *)self) + symbol__priv_size; - } + self->start = start; self->end = len ? start + len - 1 : start; @@ -139,8 +144,8 @@ struct dso *dso__new(const char *name) strcpy(self->name, name); dso__set_long_name(self, self->name); self->short_name = self->name; - self->syms = RB_ROOT; - self->find_symbol = dso__find_symbol; + self->functions = RB_ROOT; + self->find_function = dso__find_function; self->slen_calculated = 0; self->origin = DSO__ORIG_NOT_FOUND; self->loaded = 0; @@ -150,22 +155,22 @@ struct dso *dso__new(const char *name) return self; } -static void dso__delete_symbols(struct dso *self) +static void symbols__delete(struct rb_root *self) { struct symbol *pos; - struct rb_node *next = rb_first(&self->syms); + struct rb_node *next = rb_first(self); while (next) { pos = rb_entry(next, struct symbol, rb_node); next = rb_next(&pos->rb_node); - rb_erase(&pos->rb_node, &self->syms); + rb_erase(&pos->rb_node, self); symbol__delete(pos); } } void dso__delete(struct dso *self) { - dso__delete_symbols(self); + symbols__delete(&self->functions); if (self->long_name != self->name) free(self->long_name); free(self); @@ -177,9 +182,9 @@ void dso__set_build_id(struct dso *self, void *build_id) self->has_build_id = 1; } -static void dso__insert_symbol(struct dso *self, struct symbol *sym) +static void symbols__insert(struct rb_root *self, struct symbol *sym) { - struct rb_node **p = &self->syms.rb_node; + struct rb_node **p = &self->rb_node; struct rb_node *parent = NULL; const u64 ip = sym->start; struct symbol *s; @@ -193,17 +198,17 @@ static void dso__insert_symbol(struct dso *self, struct symbol *sym) p = &(*p)->rb_right; } rb_link_node(&sym->rb_node, parent, p); - rb_insert_color(&sym->rb_node, &self->syms); + rb_insert_color(&sym->rb_node, self); } -struct symbol *dso__find_symbol(struct dso *self, u64 ip) +static struct symbol *symbols__find(struct rb_root *self, u64 ip) { struct rb_node *n; if (self == NULL) return NULL; - n = self->syms.rb_node; + n = self->rb_node; while (n) { struct symbol *s = rb_entry(n, struct symbol, rb_node); @@ -219,6 +224,11 @@ struct symbol *dso__find_symbol(struct dso *self, u64 ip) return NULL; } +struct symbol *dso__find_function(struct dso *self, u64 ip) +{ + return symbols__find(&self->functions, ip); +} + int build_id__sprintf(u8 *self, int len, char *bf) { char *bid = bf; @@ -248,9 +258,9 @@ size_t dso__fprintf(struct dso *self, FILE *fp) size_t ret = fprintf(fp, "dso: %s (", self->short_name); ret += dso__fprintf_buildid(self, fp); - ret += fprintf(fp, ")\n"); + ret += fprintf(fp, ")\nFunctions:\n"); - for (nd = rb_first(&self->syms); nd; nd = rb_next(nd)) { + for (nd = rb_first(&self->functions); nd; nd = rb_next(nd)) { struct symbol *pos = rb_entry(nd, struct symbol, rb_node); ret += symbol__fprintf(pos, fp); } @@ -315,7 +325,7 @@ static int kernel_maps__load_all_kallsyms(void) * kernel_maps__split_kallsyms, when we have split the * maps per module */ - dso__insert_symbol(kernel_map->dso, sym); + symbols__insert(&kernel_map->dso->functions, sym); } free(line); @@ -339,7 +349,7 @@ static int kernel_maps__split_kallsyms(symbol_filter_t filter) struct map *map = kernel_map; struct symbol *pos; int count = 0; - struct rb_node *next = rb_first(&kernel_map->dso->syms); + struct rb_node *next = rb_first(&kernel_map->dso->functions); int kernel_range = 0; while (next) { @@ -389,12 +399,13 @@ static int kernel_maps__split_kallsyms(symbol_filter_t filter) } if (filter && filter(map, pos)) { - rb_erase(&pos->rb_node, &kernel_map->dso->syms); + rb_erase(&pos->rb_node, &kernel_map->dso->functions); symbol__delete(pos); } else { if (map != kernel_map) { - rb_erase(&pos->rb_node, &kernel_map->dso->syms); - dso__insert_symbol(map->dso, pos); + rb_erase(&pos->rb_node, + &kernel_map->dso->functions); + symbols__insert(&map->dso->functions, pos); } count++; } @@ -409,7 +420,7 @@ static int kernel_maps__load_kallsyms(symbol_filter_t filter) if (kernel_maps__load_all_kallsyms()) return -1; - dso__fixup_sym_end(kernel_map->dso); + symbols__fixup_end(&kernel_map->dso->functions); kernel_map->dso->origin = DSO__ORIG_KERNEL; return kernel_maps__split_kallsyms(filter); @@ -480,7 +491,7 @@ static int dso__load_perf_map(struct dso *self, struct map *map, if (filter && filter(map, sym)) symbol__delete(sym); else { - dso__insert_symbol(self, sym); + symbols__insert(&self->functions, sym); nr_syms++; } } @@ -678,7 +689,7 @@ static int dso__synthesize_plt_symbols(struct dso *self, struct map *map, if (filter && filter(map, f)) symbol__delete(f); else { - dso__insert_symbol(self, f); + symbols__insert(&self->functions, f); ++nr; } } @@ -700,7 +711,7 @@ static int dso__synthesize_plt_symbols(struct dso *self, struct map *map, if (filter && filter(map, f)) symbol__delete(f); else { - dso__insert_symbol(self, f); + symbols__insert(&self->functions, f); ++nr; } } @@ -874,7 +885,7 @@ new_symbol: if (filter && filter(curr_map, f)) symbol__delete(f); else { - dso__insert_symbol(curr_dso, f); + symbols__insert(&curr_dso->functions, f); nr++; } } @@ -883,7 +894,7 @@ new_symbol: * For misannotated, zeroed, ASM function sizes. */ if (nr > 0) - dso__fixup_sym_end(self); + symbols__fixup_end(&self->functions); err = nr; out_elf_end: elf_end(elf); @@ -1155,8 +1166,8 @@ static void kernel_maps__insert(struct map *map) maps__insert(&kernel_maps, map); } -struct symbol *kernel_maps__find_symbol(u64 ip, struct map **mapp, - symbol_filter_t filter) +struct symbol *kernel_maps__find_function(u64 ip, struct map **mapp, + symbol_filter_t filter) { struct map *map = maps__find(&kernel_maps, ip); @@ -1165,8 +1176,10 @@ struct symbol *kernel_maps__find_symbol(u64 ip, struct map **mapp, if (map) { ip = map->map_ip(map, ip); - return map__find_symbol(map, ip, filter); - } + return map__find_function(map, ip, filter); + } else + WARN_ONCE(RB_EMPTY_ROOT(&kernel_maps), + "Empty kernel_maps, was symbol__init() called?\n"); return NULL; } @@ -1425,8 +1438,8 @@ do_kallsyms: if (err > 0) { out_fixup: - map__fixup_start(map); - map__fixup_end(map); + map__fixup_start(map, &map->dso->functions); + map__fixup_end(map, &map->dso->functions); } return err; @@ -1485,9 +1498,9 @@ size_t dsos__fprintf_buildid(FILE *fp) return ret; } -static int kernel_maps__create_kernel_map(const char *vmlinux_name) +static int kernel_maps__create_kernel_map(const struct symbol_conf *conf) { - struct dso *kernel = dso__new(vmlinux_name ?: "[kernel.kallsyms]"); + struct dso *kernel = dso__new(conf->vmlinux_name ?: "[kernel.kallsyms]"); if (kernel == NULL) return -1; @@ -1577,18 +1590,21 @@ out_fail: return -1; } -int kernel_maps__init(const char *vmlinux_name, bool try_vmlinux_path, - bool use_modules) +static int kernel_maps__init(const struct symbol_conf *conf) { - if (try_vmlinux_path && vmlinux_path__init() < 0) + const struct symbol_conf *pconf = conf ?: &symbol_conf__defaults; + + symbol__priv_size = pconf->priv_size; + + if (pconf->try_vmlinux_path && vmlinux_path__init() < 0) return -1; - if (kernel_maps__create_kernel_map(vmlinux_name) < 0) { + if (kernel_maps__create_kernel_map(pconf) < 0) { vmlinux_path__exit(); return -1; } - if (use_modules && kernel_maps__create_module_maps() < 0) + if (pconf->use_modules && kernel_maps__create_module_maps() < 0) pr_debug("Failed to load list of modules in use, " "continuing...\n"); /* @@ -1598,8 +1614,8 @@ int kernel_maps__init(const char *vmlinux_name, bool try_vmlinux_path, return 0; } -void symbol__init(unsigned int priv_size) +int symbol__init(struct symbol_conf *conf) { elf_version(EV_CURRENT); - symbol__priv_size = priv_size; + return kernel_maps__init(conf); } diff --git a/tools/perf/util/symbol.h b/tools/perf/util/symbol.h index 8c4d026e067..65846d0c5df 100644 --- a/tools/perf/util/symbol.h +++ b/tools/perf/util/symbol.h @@ -49,6 +49,13 @@ struct symbol { char name[0]; }; +struct symbol_conf { + unsigned short priv_size; + bool try_vmlinux_path, + use_modules; + const char *vmlinux_name; +}; + extern unsigned int symbol__priv_size; static inline void *symbol__priv(struct symbol *self) @@ -58,8 +65,8 @@ static inline void *symbol__priv(struct symbol *self) struct dso { struct list_head node; - struct rb_root syms; - struct symbol *(*find_symbol)(struct dso *, u64 ip); + struct rb_root functions; + struct symbol *(*find_function)(struct dso *, u64 ip); u8 adjust_symbols:1; u8 slen_calculated:1; u8 loaded:1; @@ -76,7 +83,7 @@ struct dso { struct dso *dso__new(const char *name); void dso__delete(struct dso *self); -struct symbol *dso__find_symbol(struct dso *self, u64 ip); +struct symbol *dso__find_function(struct dso *self, u64 ip); struct dso *dsos__findnew(const char *name); int dso__load(struct dso *self, struct map *map, symbol_filter_t filter); @@ -93,11 +100,9 @@ int sysfs__read_build_id(const char *filename, void *bf, size_t size); bool dsos__read_build_ids(void); int build_id__sprintf(u8 *self, int len, char *bf); -int kernel_maps__init(const char *vmlinux_name, bool try_vmlinux_path, - bool use_modules); size_t kernel_maps__fprintf(FILE *fp); -void symbol__init(unsigned int priv_size); +int symbol__init(struct symbol_conf *conf); extern struct list_head dsos; extern struct map *kernel_map; diff --git a/tools/perf/util/thread.c b/tools/perf/util/thread.c index 0f6d78c9863..1796625f778 100644 --- a/tools/perf/util/thread.c +++ b/tools/perf/util/thread.c @@ -11,7 +11,7 @@ static struct thread *last_match; static struct thread *thread__new(pid_t pid) { - struct thread *self = calloc(1, sizeof(*self)); + struct thread *self = zalloc(sizeof(*self)); if (self != NULL) { self->pid = pid; diff --git a/tools/perf/util/thread.h b/tools/perf/util/thread.h index e4b8d437725..74cba6487ed 100644 --- a/tools/perf/util/thread.h +++ b/tools/perf/util/thread.h @@ -26,8 +26,8 @@ size_t threads__fprintf(FILE *fp); void maps__insert(struct rb_root *maps, struct map *map); struct map *maps__find(struct rb_root *maps, u64 ip); -struct symbol *kernel_maps__find_symbol(const u64 ip, struct map **mapp, - symbol_filter_t filter); +struct symbol *kernel_maps__find_function(const u64 ip, struct map **mapp, + symbol_filter_t filter); struct map *kernel_maps__find_by_dso_name(const char *name); static inline struct map *thread__find_map(struct thread *self, u64 ip) diff --git a/tools/perf/util/util.h b/tools/perf/util/util.h index e1c623e0c99..c673d882588 100644 --- a/tools/perf/util/util.h +++ b/tools/perf/util/util.h @@ -290,17 +290,15 @@ static inline char *gitstrchrnul(const char *s, int c) * Wrappers: */ extern char *xstrdup(const char *str); -extern void *xmalloc(size_t size); +extern void *xmalloc(size_t size) __attribute__((weak)); extern void *xmemdupz(const void *data, size_t len); extern char *xstrndup(const char *str, size_t len); -extern void *xrealloc(void *ptr, size_t size); -extern void *xcalloc(size_t nmemb, size_t size); -extern void *xmmap(void *start, size_t length, int prot, int flags, int fd, off_t offset); -extern ssize_t xread(int fd, void *buf, size_t len); -extern ssize_t xwrite(int fd, const void *buf, size_t len); -extern int xdup(int fd); -extern FILE *xfdopen(int fd, const char *mode); -extern int xmkstemp(char *template); +extern void *xrealloc(void *ptr, size_t size) __attribute__((weak)); + +static inline void *zalloc(size_t size) +{ + return calloc(1, size); +} static inline size_t xsize_t(off_t len) { diff --git a/tools/perf/util/wrapper.c b/tools/perf/util/wrapper.c index 4574ac28396..bf44ca85d23 100644 --- a/tools/perf/util/wrapper.c +++ b/tools/perf/util/wrapper.c @@ -79,43 +79,12 @@ void *xrealloc(void *ptr, size_t size) return ret; } -void *xcalloc(size_t nmemb, size_t size) -{ - void *ret = calloc(nmemb, size); - if (!ret && (!nmemb || !size)) - ret = calloc(1, 1); - if (!ret) { - release_pack_memory(nmemb * size, -1); - ret = calloc(nmemb, size); - if (!ret && (!nmemb || !size)) - ret = calloc(1, 1); - if (!ret) - die("Out of memory, calloc failed"); - } - return ret; -} - -void *xmmap(void *start, size_t length, - int prot, int flags, int fd, off_t offset) -{ - void *ret = mmap(start, length, prot, flags, fd, offset); - if (ret == MAP_FAILED) { - if (!length) - return NULL; - release_pack_memory(length, fd); - ret = mmap(start, length, prot, flags, fd, offset); - if (ret == MAP_FAILED) - die("Out of memory? mmap failed: %s", strerror(errno)); - } - return ret; -} - /* * xread() is the same a read(), but it automatically restarts read() * operations with a recoverable error (EAGAIN and EINTR). xread() * DOES NOT GUARANTEE that "len" bytes is read even if the data is available. */ -ssize_t xread(int fd, void *buf, size_t len) +static ssize_t xread(int fd, void *buf, size_t len) { ssize_t nr; while (1) { @@ -131,7 +100,7 @@ ssize_t xread(int fd, void *buf, size_t len) * operations with a recoverable error (EAGAIN and EINTR). xwrite() DOES NOT * GUARANTEE that "len" bytes is written even if the operation is successful. */ -ssize_t xwrite(int fd, const void *buf, size_t len) +static ssize_t xwrite(int fd, const void *buf, size_t len) { ssize_t nr; while (1) { @@ -179,29 +148,3 @@ ssize_t write_in_full(int fd, const void *buf, size_t count) return total; } - -int xdup(int fd) -{ - int ret = dup(fd); - if (ret < 0) - die("dup failed: %s", strerror(errno)); - return ret; -} - -FILE *xfdopen(int fd, const char *mode) -{ - FILE *stream = fdopen(fd, mode); - if (stream == NULL) - die("Out of memory? fdopen failed: %s", strerror(errno)); - return stream; -} - -int xmkstemp(char *template) -{ - int fd; - - fd = mkstemp(template); - if (fd < 0) - die("Unable to create temporary file: %s", strerror(errno)); - return fd; -} |