static int opt_pid;
static int opt_userspace;
+static char *opt_cmd_name;
static int opt_kernel;
static char *opt_channel;
static int opt_domain;
enum {
OPT_HELP = 1,
+ OPT_USERSPACE,
};
static struct lttng_handle *handle;
/* longName, shortName, argInfo, argPtr, value, descrip, argDesc */
{"help", 'h', POPT_ARG_NONE, 0, OPT_HELP, 0, 0},
{"kernel", 'k', POPT_ARG_VAL, &opt_kernel, 1, 0, 0},
- {"userspace", 'u', POPT_ARG_VAL, &opt_userspace, 1, 0, 0},
+ {"userspace", 'u', POPT_ARG_STRING | POPT_ARGFLAG_OPTIONAL, &opt_cmd_name, OPT_USERSPACE, 0, 0},
{"pid", 'p', POPT_ARG_INT, &opt_pid, 0, 0, 0},
{"channel", 'c', POPT_ARG_STRING, &opt_channel, 0, 0, 0},
{"domain", 'd', POPT_ARG_VAL, &opt_domain, 1, 0, 0},
* On success, return an allocated string pointer to the proc cmdline.
* On error, return NULL.
*/
-#ifdef DISABLE
static char *get_cmdline_by_pid(pid_t pid)
{
int ret;
end:
return cmdline;
}
-#endif /* DISABLE */
+
+static
+const char *active_string(int value)
+{
+ switch (value) {
+ case 0: return " [inactive]";
+ case 1: return " [active]";
+ case -1: return "";
+ default: return NULL;
+ }
+}
+
+static
+const char *enabled_string(int value)
+{
+ switch (value) {
+ case 0: return " [disabled]";
+ case 1: return " [enabled]";
+ case -1: return "";
+ default: return NULL;
+ }
+}
+
+static
+const char *loglevel_string_pre(const char *loglevel)
+{
+ if (loglevel[0] == '\0') {
+ return "";
+ } else {
+ return " (loglevel: ";
+ }
+}
+
+static
+const char *loglevel_string_post(const char *loglevel)
+{
+ if (loglevel[0] == '\0') {
+ return "";
+ } else {
+ return ")";
+ }
+}
+
+/*
+ * Pretty print single event.
+ */
+static void print_events(struct lttng_event *event)
+{
+ switch (event->type) {
+ case LTTNG_EVENT_TRACEPOINT:
+ {
+ char ll_value[LTTNG_SYMBOL_NAME_LEN] = "";
+
+ if (event->loglevel[0] != '\0') {
+ int ret;
+
+ ret = snprintf(ll_value, LTTNG_SYMBOL_NAME_LEN,
+ " (%lld)", (long long) event->loglevel_value);
+ if (ret < 0)
+ ERR("snprintf error");
+ }
+ MSG("%s%s%s%s%s%s (type: tracepoint)%s", indent6,
+ event->name,
+ loglevel_string_pre(event->loglevel),
+ event->loglevel,
+ ll_value,
+ loglevel_string_post(event->loglevel),
+ enabled_string(event->enabled));
+ break;
+ }
+ case LTTNG_EVENT_PROBE:
+ MSG("%s%s (type: probe)%s", indent6,
+ event->name, enabled_string(event->enabled));
+ if (event->attr.probe.addr != 0) {
+ MSG("%saddr: 0x%" PRIx64, indent8, event->attr.probe.addr);
+ } else {
+ MSG("%soffset: 0x%" PRIx64, indent8, event->attr.probe.offset);
+ MSG("%ssymbol: %s", indent8, event->attr.probe.symbol_name);
+ }
+ break;
+ case LTTNG_EVENT_FUNCTION:
+ case LTTNG_EVENT_FUNCTION_ENTRY:
+ MSG("%s%s (type: function)%s", indent6,
+ event->name, enabled_string(event->enabled));
+ MSG("%ssymbol: \"%s\"", indent8, event->attr.ftrace.symbol_name);
+ break;
+ case LTTNG_EVENT_SYSCALL:
+ MSG("%s (type: syscall)%s", indent6,
+ enabled_string(event->enabled));
+ break;
+ case LTTNG_EVENT_NOOP:
+ MSG("%s (type: noop)%s", indent6,
+ enabled_string(event->enabled));
+ break;
+ case LTTNG_EVENT_TRACEPOINT_LOGLEVEL:
+ MSG("%s%s (type: tracepoint loglevel)%s", indent6,
+ event->name,
+ enabled_string(event->enabled));
+ break;
+ case LTTNG_EVENT_ALL:
+ /* We should never have "all" events in list. */
+ assert(0);
+ break;
+ }
+}
+
+/*
+ * Ask session daemon for all user space tracepoints available.
+ */
+static int list_ust_events(void)
+{
+ int i, size;
+ struct lttng_domain domain;
+ struct lttng_handle *handle;
+ struct lttng_event *event_list;
+ pid_t cur_pid = 0;
+
+ DBG("Getting UST tracing events");
+
+ domain.type = LTTNG_DOMAIN_UST;
+
+ handle = lttng_create_handle(NULL, &domain);
+ if (handle == NULL) {
+ goto error;
+ }
+
+ size = lttng_list_tracepoints(handle, &event_list);
+ if (size < 0) {
+ ERR("Unable to list UST events");
+ return size;
+ }
+
+ MSG("UST events:\n-------------");
+
+ if (size == 0) {
+ MSG("None");
+ }
+
+ for (i = 0; i < size; i++) {
+ if (cur_pid != event_list[i].pid) {
+ cur_pid = event_list[i].pid;
+ MSG("\nPID: %d - Name: %s", cur_pid, get_cmdline_by_pid(cur_pid));
+ }
+ print_events(&event_list[i]);
+ }
+
+ MSG("");
+
+ free(event_list);
+
+ return CMD_SUCCESS;
+
+error:
+ return -1;
+}
/*
* Ask for all trace events in the kernel and pretty print them.
static int list_kernel_events(void)
{
int i, size;
+ struct lttng_domain domain;
+ struct lttng_handle *handle;
struct lttng_event *event_list;
- DBG("Getting all tracing events");
+ DBG("Getting kernel tracing events");
+
+ domain.type = LTTNG_DOMAIN_KERNEL;
+
+ handle = lttng_create_handle(NULL, &domain);
+ if (handle == NULL) {
+ goto error;
+ }
size = lttng_list_tracepoints(handle, &event_list);
if (size < 0) {
MSG("Kernel events:\n-------------");
for (i = 0; i < size; i++) {
- MSG(" %s", event_list[i].name);
+ print_events(&event_list[i]);
}
+ MSG("");
+
free(event_list);
return CMD_SUCCESS;
+
+error:
+ return -1;
}
/*
MSG("\n%sEvents:", indent4);
if (count == 0) {
- MSG("%sNone", indent6);
+ MSG("%sNone\n", indent6);
goto end;
}
for (i = 0; i < count; i++) {
- switch (events[i].type) {
- case LTTNG_EVENT_TRACEPOINT:
- MSG("%s%s (type: tracepoint) [enabled: %d]", indent6,
- events[i].name, events[i].enabled);
- break;
- case LTTNG_EVENT_PROBE:
- MSG("%s%s (type: probe) [enabled: %d]", indent6,
- events[i].name, events[i].enabled);
- if (events[i].attr.probe.addr != 0) {
- MSG("%saddr: 0x%" PRIx64, indent8, events[i].attr.probe.addr);
- } else {
- MSG("%soffset: 0x%" PRIx64, indent8, events[i].attr.probe.offset);
- MSG("%ssymbol: %s", indent8, events[i].attr.probe.symbol_name);
- }
- break;
- case LTTNG_EVENT_FUNCTION:
- case LTTNG_EVENT_FUNCTION_ENTRY:
- MSG("%s%s (type: function) [enabled: %d]", indent6,
- events[i].name, events[i].enabled);
- MSG("%ssymbol: \"%s\"", indent8, events[i].attr.ftrace.symbol_name);
- break;
- case LTTNG_EVENT_SYSCALL:
- MSG("%s (type: syscall) [enabled: %d]", indent6,
- events[i].enabled);
- break;
- case LTTNG_EVENT_NOOP:
- MSG("%s (type: noop) [enabled: %d]", indent6,
- events[i].enabled);
- break;
- case LTTNG_EVENT_ALL:
- /* We should never have "all" events in list. */
- assert(0);
- break;
- }
+ print_events(&events[i]);
}
MSG("");
*/
static void print_channel(struct lttng_channel *channel)
{
- MSG("- %s (enabled: %d):\n", channel->name, channel->enabled);
+ MSG("- %s:%s\n", channel->name, enabled_string(channel->enabled));
MSG("%sAttributes:", indent4);
MSG("%soverwrite mode: %d", indent6, channel->attr.overwrite);
unsigned int chan_found = 0;
struct lttng_channel *channels = NULL;
- DBG("Listing channel(s) (%s)", channel_name);
+ DBG("Listing channel(s) (%s)", channel_name ? : "<all>");
count = lttng_list_channels(handle, &channels);
if (count < 0) {
if (session_name != NULL) {
if (strncmp(sessions[i].name, session_name, NAME_MAX) == 0) {
session_found = 1;
- MSG("Tracing session %s:", session_name);
+ MSG("Tracing session %s:%s", session_name, active_string(sessions[i].enabled));
MSG("%sTrace path: %s\n", indent4, sessions[i].path);
break;
}
continue;
}
- MSG(" %d) %s (%s)", i + 1, sessions[i].name, sessions[i].path);
+ MSG(" %d) %s (%s)%s", i + 1, sessions[i].name, sessions[i].path, active_string(sessions[i].enabled));
if (session_found) {
break;
switch (domains[i].type) {
case LTTNG_DOMAIN_KERNEL:
MSG(" - Kernel");
+ break;
+ case LTTNG_DOMAIN_UST:
+ MSG(" - UST global");
+ break;
default:
break;
}
int cmd_list(int argc, const char **argv)
{
int opt, i, ret = CMD_SUCCESS;
+ int nb_domain;
const char *session_name;
static poptContext pc;
struct lttng_domain domain;
case OPT_HELP:
usage(stderr);
goto end;
+ case OPT_USERSPACE:
+ opt_userspace = 1;
+ break;
default:
usage(stderr);
ret = CMD_UNDEFINED;
}
}
- if (opt_userspace || opt_pid != 0) {
- MSG("*** Userspace tracing not implemented ***\n");
+ if (opt_pid != 0) {
+ MSG("*** Userspace tracing not implemented for PID ***\n");
}
/* Get session name (trailing argument) */
session_name = poptGetArg(pc);
- DBG("Session name: %s", session_name);
+ DBG2("Session name: %s", session_name);
if (opt_kernel) {
domain.type = LTTNG_DOMAIN_KERNEL;
+ } else if (opt_userspace) {
+ DBG2("Listing userspace global domain");
+ domain.type = LTTNG_DOMAIN_UST;
}
handle = lttng_create_handle(session_name, &domain);
}
if (session_name == NULL) {
+ if (!opt_kernel && !opt_userspace) {
+ ret = list_sessions(NULL);
+ if (ret < 0) {
+ goto end;
+ }
+ }
if (opt_kernel) {
ret = list_kernel_events();
if (ret < 0) {
goto end;
}
- } else {
- ret = list_sessions(NULL);
+ }
+ if (opt_userspace) {
+ ret = list_ust_events();
if (ret < 0) {
goto end;
}
if (ret < 0) {
goto end;
}
- } else if (opt_userspace) {
- /* TODO: Userspace domain */
} else {
/* We want all domain(s) */
- ret = lttng_list_domains(handle, &domains);
- if (ret < 0) {
+ nb_domain = lttng_list_domains(handle, &domains);
+ if (nb_domain < 0) {
+ ret = nb_domain;
goto end;
}
- for (i = 0; i < ret; i++) {
+ for (i = 0; i < nb_domain; i++) {
switch (domains[i].type) {
case LTTNG_DOMAIN_KERNEL:
MSG("=== Domain: Kernel ===\n");
break;
+ case LTTNG_DOMAIN_UST:
+ MSG("=== Domain: UST global ===\n");
+ break;
default:
MSG("=== Domain: Unimplemented ===\n");
break;
projects / lttng-tools.git / blobdiff