From 7414db411919980bc4c9bab11d5d040b5dd7667c Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Wed, 23 Feb 2022 14:18:31 -0300 Subject: [PATCH 01/13] rtla: Fix systme -> system typo on man page Link: https://lkml.kernel.org/r/YhZsZxqk+IaFxorj@kernel.org Fixes: 496082df01bb08a4 ("rtla: Add rtla osnoise man page") Signed-off-by: Arnaldo Carvalho de Melo Acked-by: Daniel Bristot de Oliveira Signed-off-by: Steven Rostedt (Google) --- Documentation/tools/rtla/common_osnoise_description.rst | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/Documentation/tools/rtla/common_osnoise_description.rst b/Documentation/tools/rtla/common_osnoise_description.rst index 8973c5df888f..d5d61615b967 100644 --- a/Documentation/tools/rtla/common_osnoise_description.rst +++ b/Documentation/tools/rtla/common_osnoise_description.rst @@ -1,7 +1,7 @@ The **rtla osnoise** tool is an interface for the *osnoise* tracer. The *osnoise* tracer dispatches a kernel thread per-cpu. These threads read the time in a loop while with preemption, softirq and IRQs enabled, thus -allowing all the sources of operating systme noise during its execution. +allowing all the sources of operating system noise during its execution. The *osnoise*'s tracer threads take note of the delta between each time read, along with an interference counter of all sources of interference. At the end of each period, the *osnoise* tracer displays a summary of From ce33c845b030c9cf768370c951bc699470b09fa7 Mon Sep 17 00:00:00 2001 From: Daniel Bristot de Oliveira Date: Sun, 20 Feb 2022 23:49:57 +0100 Subject: [PATCH 02/13] tracing: Dump stacktrace trigger to the corresponding instance The stacktrace event trigger is not dumping the stacktrace to the instance where it was enabled, but to the global "instance." Use the private_data, pointing to the trigger file, to figure out the corresponding trace instance, and use it in the trigger action, like snapshot_trigger does. Link: https://lkml.kernel.org/r/afbb0b4f18ba92c276865bc97204d438473f4ebc.1645396236.git.bristot@kernel.org Cc: stable@vger.kernel.org Fixes: ae63b31e4d0e2 ("tracing: Separate out trace events from global variables") Reviewed-by: Tom Zanussi Tested-by: Tom Zanussi Signed-off-by: Daniel Bristot de Oliveira Signed-off-by: Steven Rostedt (Google) --- kernel/trace/trace_events_trigger.c | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-) diff --git a/kernel/trace/trace_events_trigger.c b/kernel/trace/trace_events_trigger.c index d00fee705f9c..e0d50c9577f3 100644 --- a/kernel/trace/trace_events_trigger.c +++ b/kernel/trace/trace_events_trigger.c @@ -1540,7 +1540,12 @@ stacktrace_trigger(struct event_trigger_data *data, struct trace_buffer *buffer, void *rec, struct ring_buffer_event *event) { - trace_dump_stack(STACK_SKIP); + struct trace_event_file *file = data->private_data; + + if (file) + __trace_stack(file->tr, tracing_gen_ctx(), STACK_SKIP); + else + trace_dump_stack(STACK_SKIP); } static void From 302e9edd54985f584cfc180098f3554774126969 Mon Sep 17 00:00:00 2001 From: "Steven Rostedt (Google)" Date: Wed, 23 Feb 2022 22:38:37 -0500 Subject: [PATCH 03/13] tracing: Have traceon and traceoff trigger honor the instance If a trigger is set on an event to disable or enable tracing within an instance, then tracing should be disabled or enabled in the instance and not at the top level, which is confusing to users. Link: https://lkml.kernel.org/r/20220223223837.14f94ec3@rorschach.local.home Cc: stable@vger.kernel.org Fixes: ae63b31e4d0e2 ("tracing: Separate out trace events from global variables") Tested-by: Daniel Bristot de Oliveira Reviewed-by: Tom Zanussi Signed-off-by: Steven Rostedt (Google) --- kernel/trace/trace_events_trigger.c | 52 +++++++++++++++++++++++++---- 1 file changed, 46 insertions(+), 6 deletions(-) diff --git a/kernel/trace/trace_events_trigger.c b/kernel/trace/trace_events_trigger.c index e0d50c9577f3..efe563140f27 100644 --- a/kernel/trace/trace_events_trigger.c +++ b/kernel/trace/trace_events_trigger.c @@ -1295,6 +1295,16 @@ traceon_trigger(struct event_trigger_data *data, struct trace_buffer *buffer, void *rec, struct ring_buffer_event *event) { + struct trace_event_file *file = data->private_data; + + if (file) { + if (tracer_tracing_is_on(file->tr)) + return; + + tracer_tracing_on(file->tr); + return; + } + if (tracing_is_on()) return; @@ -1306,8 +1316,15 @@ traceon_count_trigger(struct event_trigger_data *data, struct trace_buffer *buffer, void *rec, struct ring_buffer_event *event) { - if (tracing_is_on()) - return; + struct trace_event_file *file = data->private_data; + + if (file) { + if (tracer_tracing_is_on(file->tr)) + return; + } else { + if (tracing_is_on()) + return; + } if (!data->count) return; @@ -1315,7 +1332,10 @@ traceon_count_trigger(struct event_trigger_data *data, if (data->count != -1) (data->count)--; - tracing_on(); + if (file) + tracer_tracing_on(file->tr); + else + tracing_on(); } static void @@ -1323,6 +1343,16 @@ traceoff_trigger(struct event_trigger_data *data, struct trace_buffer *buffer, void *rec, struct ring_buffer_event *event) { + struct trace_event_file *file = data->private_data; + + if (file) { + if (!tracer_tracing_is_on(file->tr)) + return; + + tracer_tracing_off(file->tr); + return; + } + if (!tracing_is_on()) return; @@ -1334,8 +1364,15 @@ traceoff_count_trigger(struct event_trigger_data *data, struct trace_buffer *buffer, void *rec, struct ring_buffer_event *event) { - if (!tracing_is_on()) - return; + struct trace_event_file *file = data->private_data; + + if (file) { + if (!tracer_tracing_is_on(file->tr)) + return; + } else { + if (!tracing_is_on()) + return; + } if (!data->count) return; @@ -1343,7 +1380,10 @@ traceoff_count_trigger(struct event_trigger_data *data, if (data->count != -1) (data->count)--; - tracing_off(); + if (file) + tracer_tracing_off(file->tr); + else + tracing_off(); } static int From b61edd57740de5895f44f2ea417b164d9e1708bb Mon Sep 17 00:00:00 2001 From: "Steven Rostedt (Google)" Date: Fri, 18 Feb 2022 19:00:57 -0500 Subject: [PATCH 04/13] eprobes: Remove redundant event type information Currently, the event probes save the type of the event they are attached to when recording the event. For example: # echo 'e:switch sched/sched_switch prev_state=$prev_state prev_prio=$prev_prio next_pid=$next_pid next_prio=$next_prio' > dynamic_events # cat events/eprobes/switch/format name: switch ID: 1717 format: field:unsigned short common_type; offset:0; size:2; signed:0; field:unsigned char common_flags; offset:2; size:1; signed:0; field:unsigned char common_preempt_count; offset:3; size:1; signed:0; field:int common_pid; offset:4; size:4; signed:1; field:unsigned int __probe_type; offset:8; size:4; signed:0; field:u64 prev_state; offset:12; size:8; signed:0; field:u64 prev_prio; offset:20; size:8; signed:0; field:u64 next_pid; offset:28; size:8; signed:0; field:u64 next_prio; offset:36; size:8; signed:0; print fmt: "(%u) prev_state=0x%Lx prev_prio=0x%Lx next_pid=0x%Lx next_prio=0x%Lx", REC->__probe_type, REC->prev_state, REC->prev_prio, REC->next_pid, REC->next_prio The __probe_type adds 4 bytes to every event. One of the reasons for creating eprobes is to limit what is traced in an event to be able to limit what is written into the ring buffer. Having this redundant 4 bytes to every event takes away from this. The event that is recorded can be retrieved from the event probe itself, that is available when the trace is happening. For user space tools, it could simply read the dynamic_event file to find the event they are for. So there is really no reason to write this information into the ring buffer for every event. Link: https://lkml.kernel.org/r/20220218190057.2f5a19a8@gandalf.local.home Acked-by: Masami Hiramatsu Reviewed-by: Joel Fernandes Signed-off-by: Steven Rostedt (Google) --- kernel/trace/trace.h | 1 - kernel/trace/trace_eprobe.c | 16 +++++++--------- kernel/trace/trace_probe.c | 10 +++++----- kernel/trace/trace_probe.h | 1 - 4 files changed, 12 insertions(+), 16 deletions(-) diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index d038ddbf1bea..c5b09c31e077 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -136,7 +136,6 @@ struct kprobe_trace_entry_head { struct eprobe_trace_entry_head { struct trace_entry ent; - unsigned int type; }; struct kretprobe_trace_entry_head { diff --git a/kernel/trace/trace_eprobe.c b/kernel/trace/trace_eprobe.c index 191db32dec46..541aa13581b9 100644 --- a/kernel/trace/trace_eprobe.c +++ b/kernel/trace/trace_eprobe.c @@ -242,7 +242,6 @@ static int trace_eprobe_tp_arg_update(struct trace_eprobe *ep, int i) static int eprobe_event_define_fields(struct trace_event_call *event_call) { - int ret; struct eprobe_trace_entry_head field; struct trace_probe *tp; @@ -250,8 +249,6 @@ static int eprobe_event_define_fields(struct trace_event_call *event_call) if (WARN_ON_ONCE(!tp)) return -ENOENT; - DEFINE_FIELD(unsigned int, type, FIELD_STRING_TYPE, 0); - return traceprobe_define_arg_fields(event_call, sizeof(field), tp); } @@ -270,7 +267,9 @@ print_eprobe_event(struct trace_iterator *iter, int flags, struct trace_event_call *pevent; struct trace_event *probed_event; struct trace_seq *s = &iter->seq; + struct trace_eprobe *ep; struct trace_probe *tp; + unsigned int type; field = (struct eprobe_trace_entry_head *)iter->ent; tp = trace_probe_primary_from_call( @@ -278,15 +277,18 @@ print_eprobe_event(struct trace_iterator *iter, int flags, if (WARN_ON_ONCE(!tp)) goto out; + ep = container_of(tp, struct trace_eprobe, tp); + type = ep->event->event.type; + trace_seq_printf(s, "%s: (", trace_probe_name(tp)); - probed_event = ftrace_find_event(field->type); + probed_event = ftrace_find_event(type); if (probed_event) { pevent = container_of(probed_event, struct trace_event_call, event); trace_seq_printf(s, "%s.%s", pevent->class->system, trace_event_name(pevent)); } else { - trace_seq_printf(s, "%u", field->type); + trace_seq_printf(s, "%u", type); } trace_seq_putc(s, ')'); @@ -498,10 +500,6 @@ __eprobe_trace_func(struct eprobe_data *edata, void *rec) return; entry = fbuffer.entry = ring_buffer_event_data(fbuffer.event); - if (edata->ep->event) - entry->type = edata->ep->event->event.type; - else - entry->type = 0; store_trace_args(&entry[1], &edata->ep->tp, rec, sizeof(*entry), dsize); trace_event_buffer_commit(&fbuffer); diff --git a/kernel/trace/trace_probe.c b/kernel/trace/trace_probe.c index 73d90179b51b..80863c6508e5 100644 --- a/kernel/trace/trace_probe.c +++ b/kernel/trace/trace_probe.c @@ -871,15 +871,15 @@ static int __set_print_fmt(struct trace_probe *tp, char *buf, int len, switch (ptype) { case PROBE_PRINT_NORMAL: fmt = "(%lx)"; - arg = "REC->" FIELD_STRING_IP; + arg = ", REC->" FIELD_STRING_IP; break; case PROBE_PRINT_RETURN: fmt = "(%lx <- %lx)"; - arg = "REC->" FIELD_STRING_FUNC ", REC->" FIELD_STRING_RETIP; + arg = ", REC->" FIELD_STRING_FUNC ", REC->" FIELD_STRING_RETIP; break; case PROBE_PRINT_EVENT: - fmt = "(%u)"; - arg = "REC->" FIELD_STRING_TYPE; + fmt = ""; + arg = ""; break; default: WARN_ON_ONCE(1); @@ -903,7 +903,7 @@ static int __set_print_fmt(struct trace_probe *tp, char *buf, int len, parg->type->fmt); } - pos += snprintf(buf + pos, LEN_OR_ZERO, "\", %s", arg); + pos += snprintf(buf + pos, LEN_OR_ZERO, "\"%s", arg); for (i = 0; i < tp->nr_args; i++) { parg = tp->args + i; diff --git a/kernel/trace/trace_probe.h b/kernel/trace/trace_probe.h index 99e7a5df025e..92cc149af0fd 100644 --- a/kernel/trace/trace_probe.h +++ b/kernel/trace/trace_probe.h @@ -38,7 +38,6 @@ #define FIELD_STRING_IP "__probe_ip" #define FIELD_STRING_RETIP "__probe_ret_ip" #define FIELD_STRING_FUNC "__probe_func" -#define FIELD_STRING_TYPE "__probe_type" #undef DEFINE_FIELD #define DEFINE_FIELD(type, item, name, is_signed) \ From bc82c38a6933aab308387d4aca47e0a05de7b553 Mon Sep 17 00:00:00 2001 From: Christophe Leroy Date: Fri, 11 Feb 2022 08:10:18 +0100 Subject: [PATCH 05/13] tracing: Uninline trace_trigger_soft_disabled() partly On a powerpc32 build with CONFIG_CC_OPTIMISE_FOR_SIZE, the inline keyword is not honored and trace_trigger_soft_disabled() appears approx 50 times in vmlinux. Adding -Winline to the build, the following message appears: ./include/linux/trace_events.h:712:1: error: inlining failed in call to 'trace_trigger_soft_disabled': call is unlikely and code size would grow [-Werror=inline] That function is rather big for an inlined function: c003df60 : c003df60: 94 21 ff f0 stwu r1,-16(r1) c003df64: 7c 08 02 a6 mflr r0 c003df68: 90 01 00 14 stw r0,20(r1) c003df6c: bf c1 00 08 stmw r30,8(r1) c003df70: 83 e3 00 24 lwz r31,36(r3) c003df74: 73 e9 01 00 andi. r9,r31,256 c003df78: 41 82 00 10 beq c003df88 c003df7c: 38 60 00 00 li r3,0 c003df80: 39 61 00 10 addi r11,r1,16 c003df84: 4b fd 60 ac b c0014030 <_rest32gpr_30_x> c003df88: 73 e9 00 80 andi. r9,r31,128 c003df8c: 7c 7e 1b 78 mr r30,r3 c003df90: 41 a2 00 14 beq c003dfa4 c003df94: 38 c0 00 00 li r6,0 c003df98: 38 a0 00 00 li r5,0 c003df9c: 38 80 00 00 li r4,0 c003dfa0: 48 05 c5 f1 bl c009a590 c003dfa4: 73 e9 00 40 andi. r9,r31,64 c003dfa8: 40 82 00 28 bne c003dfd0 c003dfac: 73 ff 02 00 andi. r31,r31,512 c003dfb0: 41 82 ff cc beq c003df7c c003dfb4: 80 01 00 14 lwz r0,20(r1) c003dfb8: 83 e1 00 0c lwz r31,12(r1) c003dfbc: 7f c3 f3 78 mr r3,r30 c003dfc0: 83 c1 00 08 lwz r30,8(r1) c003dfc4: 7c 08 03 a6 mtlr r0 c003dfc8: 38 21 00 10 addi r1,r1,16 c003dfcc: 48 05 6f 6c b c0094f38 c003dfd0: 38 60 00 01 li r3,1 c003dfd4: 4b ff ff ac b c003df80 However it is located in a hot path so inlining it is important. But forcing inlining of the entire function by using __always_inline leads to increasing the text size by approx 20 kbytes. Instead, split the fonction in two parts, one part with the likely fast path, flagged __always_inline, and a second part out of line. With this change, on a powerpc32 with CONFIG_CC_OPTIMISE_FOR_SIZE vmlinux text increases by only 1,4 kbytes, which is partly compensated by a decrease of vmlinux data by 7 kbytes. On ppc64_defconfig which has CONFIG_CC_OPTIMISE_FOR_SPEED, this change reduces vmlinux text by more than 30 kbytes. Link: https://lkml.kernel.org/r/69ce0986a52d026d381d612801d978aa4f977460.1644563295.git.christophe.leroy@csgroup.eu Signed-off-by: Christophe Leroy Signed-off-by: Steven Rostedt (Google) --- include/linux/trace_events.h | 22 ++++++++++++---------- kernel/trace/trace_events_trigger.c | 14 ++++++++++++++ 2 files changed, 26 insertions(+), 10 deletions(-) diff --git a/include/linux/trace_events.h b/include/linux/trace_events.h index 70c069aef02c..dcea51fb60e2 100644 --- a/include/linux/trace_events.h +++ b/include/linux/trace_events.h @@ -699,6 +699,8 @@ event_triggers_post_call(struct trace_event_file *file, bool trace_event_ignore_this_pid(struct trace_event_file *trace_file); +bool __trace_trigger_soft_disabled(struct trace_event_file *file); + /** * trace_trigger_soft_disabled - do triggers and test if soft disabled * @file: The file pointer of the event to test @@ -708,20 +710,20 @@ bool trace_event_ignore_this_pid(struct trace_event_file *trace_file); * triggers that require testing the fields, it will return true, * otherwise false. */ -static inline bool +static __always_inline bool trace_trigger_soft_disabled(struct trace_event_file *file) { unsigned long eflags = file->flags; - if (!(eflags & EVENT_FILE_FL_TRIGGER_COND)) { - if (eflags & EVENT_FILE_FL_TRIGGER_MODE) - event_triggers_call(file, NULL, NULL, NULL); - if (eflags & EVENT_FILE_FL_SOFT_DISABLED) - return true; - if (eflags & EVENT_FILE_FL_PID_FILTER) - return trace_event_ignore_this_pid(file); - } - return false; + if (likely(!(eflags & (EVENT_FILE_FL_TRIGGER_MODE | + EVENT_FILE_FL_SOFT_DISABLED | + EVENT_FILE_FL_PID_FILTER)))) + return false; + + if (likely(eflags & EVENT_FILE_FL_TRIGGER_COND)) + return false; + + return __trace_trigger_soft_disabled(file); } #ifdef CONFIG_BPF_EVENTS diff --git a/kernel/trace/trace_events_trigger.c b/kernel/trace/trace_events_trigger.c index efe563140f27..7eb9d04f1c2e 100644 --- a/kernel/trace/trace_events_trigger.c +++ b/kernel/trace/trace_events_trigger.c @@ -84,6 +84,20 @@ event_triggers_call(struct trace_event_file *file, } EXPORT_SYMBOL_GPL(event_triggers_call); +bool __trace_trigger_soft_disabled(struct trace_event_file *file) +{ + unsigned long eflags = file->flags; + + if (eflags & EVENT_FILE_FL_TRIGGER_MODE) + event_triggers_call(file, NULL, NULL, NULL); + if (eflags & EVENT_FILE_FL_SOFT_DISABLED) + return true; + if (eflags & EVENT_FILE_FL_PID_FILTER) + return trace_event_ignore_this_pid(file); + return false; +} +EXPORT_SYMBOL_GPL(__trace_trigger_soft_disabled); + /** * event_triggers_post_call - Call 'post_triggers' for a trace event * @file: The trace_event_file associated with the event From 7acf3a127bb7c65ff39099afd78960e77b2ca5de Mon Sep 17 00:00:00 2001 From: Sven Schnelle Date: Mon, 14 Feb 2022 14:44:56 +0100 Subject: [PATCH 06/13] tracing: Ensure trace buffer is at least 4096 bytes large Booting the kernel with 'trace_buf_size=1' give a warning at boot during the ftrace selftests: [ 0.892809] Running postponed tracer tests: [ 0.892893] Testing tracer function: [ 0.901899] Callback from call_rcu_tasks_trace() invoked. [ 0.983829] Callback from call_rcu_tasks_rude() invoked. [ 1.072003] .. bad ring buffer .. corrupted trace buffer .. [ 1.091944] Callback from call_rcu_tasks() invoked. [ 1.097695] PASSED [ 1.097701] Testing dynamic ftrace: .. filter failed count=0 ..FAILED! [ 1.353474] ------------[ cut here ]------------ [ 1.353478] WARNING: CPU: 0 PID: 1 at kernel/trace/trace.c:1951 run_tracer_selftest+0x13c/0x1b0 Therefore enforce a minimum of 4096 bytes to make the selftest pass. Link: https://lkml.kernel.org/r/20220214134456.1751749-1-svens@linux.ibm.com Signed-off-by: Sven Schnelle Signed-off-by: Steven Rostedt (Google) --- kernel/trace/trace.c | 10 ++++++---- 1 file changed, 6 insertions(+), 4 deletions(-) diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 7c2578efde26..3050892d1812 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -1474,10 +1474,12 @@ static int __init set_buf_size(char *str) if (!str) return 0; buf_size = memparse(str, &str); - /* nr_entries can not be zero */ - if (buf_size == 0) - return 0; - trace_buf_size = buf_size; + /* + * nr_entries can not be zero and the startup + * tests require some buffer space. Therefore + * ensure we have at least 4096 bytes of buffer. + */ + trace_buf_size = max(4096UL, buf_size); return 1; } __setup("trace_buf_size=", set_buf_size); From ab2f993c01f261aa3eeb8842842ff38bff7806b6 Mon Sep 17 00:00:00 2001 From: Nathan Chancellor Date: Mon, 14 Feb 2022 12:28:47 -0700 Subject: [PATCH 07/13] ftrace: Remove unused ftrace_startup_enable() stub When building with clang + CONFIG_DYNAMIC_FTRACE=n + W=1, there is a warning: kernel/trace/ftrace.c:7194:20: error: unused function 'ftrace_startup_enable' [-Werror,-Wunused-function] static inline void ftrace_startup_enable(int command) { } ^ 1 error generated. Clang warns on instances of static inline functions in .c files with W=1 after commit 6863f5643dd7 ("kbuild: allow Clang to find unused static inline functions for W=1 build"). The ftrace_startup_enable() stub has been unused since commit e1effa0144a1 ("ftrace: Annotate the ops operation on update"), where its use outside of the CONFIG_DYNAMIC_TRACE section was replaced by ftrace_startup_all(). Remove it to resolve the warning. Link: https://lkml.kernel.org/r/20220214192847.488166-1-nathan@kernel.org Reported-by: kernel test robot Signed-off-by: Nathan Chancellor Signed-off-by: Steven Rostedt (Google) --- kernel/trace/ftrace.c | 1 - 1 file changed, 1 deletion(-) diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index f9feb197b2da..a4b462b6f944 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -7191,7 +7191,6 @@ static int __init ftrace_nodyn_init(void) core_initcall(ftrace_nodyn_init); static inline int ftrace_init_dyn_tracefs(struct dentry *d_tracer) { return 0; } -static inline void ftrace_startup_enable(int command) { } static inline void ftrace_startup_all(int command) { } # define ftrace_startup_sysctl() do { } while (0) From dd990352f01ee9a6c6eee152e5d11c021caccfe4 Mon Sep 17 00:00:00 2001 From: Daniel Bristot de Oliveira Date: Fri, 18 Feb 2022 16:17:38 +0100 Subject: [PATCH 08/13] tracing/osnoise: Make osnoise_main to sleep for microseconds osnoise's runtime and period are in the microseconds scale, but it is currently sleeping in the millisecond's scale. This behavior roots in the usage of hwlat as the skeleton for osnoise. Make osnoise to sleep in the microseconds scale. Also, move the sleep to a specialized function. Link: https://lkml.kernel.org/r/302aa6c7bdf2d131719b22901905e9da122a11b2.1645197336.git.bristot@kernel.org Cc: Ingo Molnar Signed-off-by: Daniel Bristot de Oliveira Signed-off-by: Steven Rostedt (Google) --- kernel/trace/trace_osnoise.c | 53 ++++++++++++++++++++++-------------- 1 file changed, 32 insertions(+), 21 deletions(-) diff --git a/kernel/trace/trace_osnoise.c b/kernel/trace/trace_osnoise.c index 870a08da5b48..cfddb30e65ab 100644 --- a/kernel/trace/trace_osnoise.c +++ b/kernel/trace/trace_osnoise.c @@ -1436,6 +1436,37 @@ out: static struct cpumask osnoise_cpumask; static struct cpumask save_cpumask; +/* + * osnoise_sleep - sleep until the next period + */ +static void osnoise_sleep(void) +{ + u64 interval; + ktime_t wake_time; + + mutex_lock(&interface_lock); + interval = osnoise_data.sample_period - osnoise_data.sample_runtime; + mutex_unlock(&interface_lock); + + /* + * differently from hwlat_detector, the osnoise tracer can run + * without a pause because preemption is on. + */ + if (!interval) { + /* Let synchronize_rcu_tasks() make progress */ + cond_resched_tasks_rcu_qs(); + return; + } + + wake_time = ktime_add_us(ktime_get(), interval); + __set_current_state(TASK_INTERRUPTIBLE); + + while (schedule_hrtimeout_range(&wake_time, 0, HRTIMER_MODE_ABS)) { + if (kthread_should_stop()) + break; + } +} + /* * osnoise_main - The osnoise detection kernel thread * @@ -1444,30 +1475,10 @@ static struct cpumask save_cpumask; */ static int osnoise_main(void *data) { - u64 interval; while (!kthread_should_stop()) { - run_osnoise(); - - mutex_lock(&interface_lock); - interval = osnoise_data.sample_period - osnoise_data.sample_runtime; - mutex_unlock(&interface_lock); - - do_div(interval, USEC_PER_MSEC); - - /* - * differently from hwlat_detector, the osnoise tracer can run - * without a pause because preemption is on. - */ - if (interval < 1) { - /* Let synchronize_rcu_tasks() make progress */ - cond_resched_tasks_rcu_qs(); - continue; - } - - if (msleep_interruptible(interval)) - break; + osnoise_sleep(); } return 0; From 851e99ebeec3f4a672bb5010cf1ece095acee447 Mon Sep 17 00:00:00 2001 From: "Steven Rostedt (Google)" Date: Fri, 25 Feb 2022 15:34:26 -0500 Subject: [PATCH 09/13] tracefs: Set the group ownership in apply_options() not parse_options() Al Viro brought it to my attention that the dentries may not be filled when the parse_options() is called, causing the call to set_gid() to possibly crash. It should only be called if parse_options() succeeds totally anyway. He suggested the logical place to do the update is in apply_options(). Link: https://lore.kernel.org/all/20220225165219.737025658@goodmis.org/ Link: https://lkml.kernel.org/r/20220225153426.1c4cab6b@gandalf.local.home Cc: stable@vger.kernel.org Acked-by: Al Viro Reported-by: Al Viro Fixes: 48b27b6b5191 ("tracefs: Set all files to the same group ownership as the mount option") Signed-off-by: Steven Rostedt (Google) --- fs/tracefs/inode.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/fs/tracefs/inode.c b/fs/tracefs/inode.c index bafc02bf8220..de7252715b12 100644 --- a/fs/tracefs/inode.c +++ b/fs/tracefs/inode.c @@ -264,7 +264,6 @@ static int tracefs_parse_options(char *data, struct tracefs_mount_opts *opts) if (!gid_valid(gid)) return -EINVAL; opts->gid = gid; - set_gid(tracefs_mount->mnt_root, gid); break; case Opt_mode: if (match_octal(&args[0], &option)) @@ -291,7 +290,9 @@ static int tracefs_apply_options(struct super_block *sb) inode->i_mode |= opts->mode; inode->i_uid = opts->uid; - inode->i_gid = opts->gid; + + /* Set all the group ids to the mount option */ + set_gid(sb->s_root, opts->gid); return 0; } From c5229a0bd47814770c895e94fbc97ad21819abfe Mon Sep 17 00:00:00 2001 From: Christophe Leroy Date: Mon, 20 Dec 2021 16:38:06 +0000 Subject: [PATCH 10/13] tracing: Fix selftest config check for function graph start up test CONFIG_DYNAMIC_FTRACE_WITH_DIRECT_CALLS is required to test direct tramp. Link: https://lkml.kernel.org/r/bdc7e594e13b0891c1d61bc8d56c94b1890eaed7.1640017960.git.christophe.leroy@csgroup.eu Signed-off-by: Christophe Leroy Signed-off-by: Steven Rostedt (Google) --- kernel/trace/trace_selftest.c | 6 ++---- 1 file changed, 2 insertions(+), 4 deletions(-) diff --git a/kernel/trace/trace_selftest.c b/kernel/trace/trace_selftest.c index afd937a46496..abcadbe933bb 100644 --- a/kernel/trace/trace_selftest.c +++ b/kernel/trace/trace_selftest.c @@ -784,9 +784,7 @@ static struct fgraph_ops fgraph_ops __initdata = { .retfunc = &trace_graph_return, }; -#if defined(CONFIG_DYNAMIC_FTRACE) && \ - defined(CONFIG_HAVE_DYNAMIC_FTRACE_WITH_ARGS) -#define TEST_DIRECT_TRAMP +#ifdef CONFIG_DYNAMIC_FTRACE_WITH_DIRECT_CALLS noinline __noclone static void trace_direct_tramp(void) { } #endif @@ -849,7 +847,7 @@ trace_selftest_startup_function_graph(struct tracer *trace, goto out; } -#ifdef TEST_DIRECT_TRAMP +#ifdef CONFIG_DYNAMIC_FTRACE_WITH_DIRECT_CALLS tracing_reset_online_cpus(&tr->array_buffer); set_graph_array(tr); From dd48f316a1216fa10f9ba26852457794417d9bc6 Mon Sep 17 00:00:00 2001 From: Daniel Bristot de Oliveira Date: Fri, 18 Feb 2022 18:57:07 +0100 Subject: [PATCH 11/13] rtla/hist: Make -E the short version of --entries Currently, --entries uses -e as the short version in the hist mode of timerlat and osnoise tools. But as -e is already used to enable events on trace sessions by other tools, thus let's keep it available for the same usage for all rtla tools. Make -E the short version of --entries for hist mode on all tools. Note: rtla was merged in this merge window, so rtla was not released yet. Link: https://lkml.kernel.org/r/5dbf0cbe7364d3a05e708926b41a097c59a02b1e.1645206561.git.bristot@kernel.org Cc: Jonathan Corbet Signed-off-by: Daniel Bristot de Oliveira Signed-off-by: Steven Rostedt (Google) --- Documentation/tools/rtla/common_hist_options.rst | 2 +- Documentation/tools/rtla/rtla-osnoise-hist.rst | 2 +- tools/tracing/rtla/src/osnoise_hist.c | 10 +++++----- tools/tracing/rtla/src/timerlat_hist.c | 10 +++++----- 4 files changed, 12 insertions(+), 12 deletions(-) diff --git a/Documentation/tools/rtla/common_hist_options.rst b/Documentation/tools/rtla/common_hist_options.rst index 0266cd08a6c9..df53ff835bfb 100644 --- a/Documentation/tools/rtla/common_hist_options.rst +++ b/Documentation/tools/rtla/common_hist_options.rst @@ -2,7 +2,7 @@ Set the histogram bucket size (default *1*). -**-e**, **--entries** *N* +**-E**, **--entries** *N* Set the number of entries of the histogram (default 256). diff --git a/Documentation/tools/rtla/rtla-osnoise-hist.rst b/Documentation/tools/rtla/rtla-osnoise-hist.rst index 52298ddd8701..f2e79d22c4c4 100644 --- a/Documentation/tools/rtla/rtla-osnoise-hist.rst +++ b/Documentation/tools/rtla/rtla-osnoise-hist.rst @@ -36,7 +36,7 @@ default). The reason for reducing the runtime is to avoid starving the **rtla** tool. The tool is also set to run for *one minute*. The output histogram is set to group outputs in buckets of *10us* and *25* entries:: - [root@f34 ~/]# rtla osnoise hist -P F:1 -c 0-11 -r 900000 -d 1M -b 10 -e 25 + [root@f34 ~/]# rtla osnoise hist -P F:1 -c 0-11 -r 900000 -d 1M -b 10 -E 25 # RTLA osnoise histogram # Time unit is microseconds (us) # Duration: 0 00:01:00 diff --git a/tools/tracing/rtla/src/osnoise_hist.c b/tools/tracing/rtla/src/osnoise_hist.c index 1f0b7fce55cf..52c053cc1789 100644 --- a/tools/tracing/rtla/src/osnoise_hist.c +++ b/tools/tracing/rtla/src/osnoise_hist.c @@ -426,7 +426,7 @@ static void osnoise_hist_usage(char *usage) static const char * const msg[] = { "", " usage: rtla osnoise hist [-h] [-D] [-d s] [-p us] [-r us] [-s us] [-S us] [-t[=file]] \\", - " [-c cpu-list] [-P priority] [-b N] [-e N] [--no-header] [--no-summary] \\", + " [-c cpu-list] [-P priority] [-b N] [-E N] [--no-header] [--no-summary] \\", " [--no-index] [--with-zeros]", "", " -h/--help: print this menu", @@ -439,7 +439,7 @@ static void osnoise_hist_usage(char *usage) " -D/--debug: print debug info", " -t/--trace[=file]: save the stopped trace to [file|osnoise_trace.txt]", " -b/--bucket-size N: set the histogram bucket size (default 1)", - " -e/--entries N: set the number of entries of the histogram (default 256)", + " -E/--entries N: set the number of entries of the histogram (default 256)", " --no-header: do not print header", " --no-summary: do not print summary", " --no-index: do not print index", @@ -486,7 +486,7 @@ static struct osnoise_hist_params while (1) { static struct option long_options[] = { {"bucket-size", required_argument, 0, 'b'}, - {"entries", required_argument, 0, 'e'}, + {"entries", required_argument, 0, 'E'}, {"cpus", required_argument, 0, 'c'}, {"debug", no_argument, 0, 'D'}, {"duration", required_argument, 0, 'd'}, @@ -507,7 +507,7 @@ static struct osnoise_hist_params /* getopt_long stores the option index here. */ int option_index = 0; - c = getopt_long(argc, argv, "c:b:d:e:Dhp:P:r:s:S:t::0123", + c = getopt_long(argc, argv, "c:b:d:E:Dhp:P:r:s:S:t::0123", long_options, &option_index); /* detect the end of the options. */ @@ -534,7 +534,7 @@ static struct osnoise_hist_params if (!params->duration) osnoise_hist_usage("Invalid -D duration\n"); break; - case 'e': + case 'E': params->entries = get_llong_from_str(optarg); if ((params->entries < 10) || (params->entries > 9999999)) osnoise_hist_usage("Entries must be > 10 and < 9999999\n"); diff --git a/tools/tracing/rtla/src/timerlat_hist.c b/tools/tracing/rtla/src/timerlat_hist.c index 436a799f9adf..237e1735afa7 100644 --- a/tools/tracing/rtla/src/timerlat_hist.c +++ b/tools/tracing/rtla/src/timerlat_hist.c @@ -429,7 +429,7 @@ static void timerlat_hist_usage(char *usage) char *msg[] = { "", " usage: [rtla] timerlat hist [-h] [-q] [-d s] [-D] [-n] [-p us] [-i us] [-T us] [-s us] [-t[=file]] \\", - " [-c cpu-list] [-P priority] [-e N] [-b N] [--no-irq] [--no-thread] [--no-header] [--no-summary] \\", + " [-c cpu-list] [-P priority] [-E N] [-b N] [--no-irq] [--no-thread] [--no-header] [--no-summary] \\", " [--no-index] [--with-zeros]", "", " -h/--help: print this menu", @@ -443,7 +443,7 @@ static void timerlat_hist_usage(char *usage) " -T/--trace[=file]: save the stopped trace to [file|timerlat_trace.txt]", " -n/--nano: display data in nanoseconds", " -b/--bucket-size N: set the histogram bucket size (default 1)", - " -e/--entries N: set the number of entries of the histogram (default 256)", + " -E/--entries N: set the number of entries of the histogram (default 256)", " --no-irq: ignore IRQ latencies", " --no-thread: ignore thread latencies", " --no-header: do not print header", @@ -494,7 +494,7 @@ static struct timerlat_hist_params {"cpus", required_argument, 0, 'c'}, {"bucket-size", required_argument, 0, 'b'}, {"debug", no_argument, 0, 'D'}, - {"entries", required_argument, 0, 'e'}, + {"entries", required_argument, 0, 'E'}, {"duration", required_argument, 0, 'd'}, {"help", no_argument, 0, 'h'}, {"irq", required_argument, 0, 'i'}, @@ -516,7 +516,7 @@ static struct timerlat_hist_params /* getopt_long stores the option index here. */ int option_index = 0; - c = getopt_long(argc, argv, "c:b:d:e:Dhi:np:P:s:t::T:012345", + c = getopt_long(argc, argv, "c:b:d:E:Dhi:np:P:s:t::T:012345", long_options, &option_index); /* detect the end of the options. */ @@ -543,7 +543,7 @@ static struct timerlat_hist_params if (!params->duration) timerlat_hist_usage("Invalid -D duration\n"); break; - case 'e': + case 'E': params->entries = get_llong_from_str(optarg); if ((params->entries < 10) || (params->entries > 9999999)) timerlat_hist_usage("Entries must be > 10 and < 9999999\n"); From 316f710172461c501f9b73f3b2fc7ce8aa5b84a5 Mon Sep 17 00:00:00 2001 From: Daniel Bristot de Oliveira Date: Fri, 18 Feb 2022 18:57:08 +0100 Subject: [PATCH 12/13] rtla/osnoise: Free params at the exit The variable that stores the parsed command line arguments are not being free()d at the rtla osnoise top exit path. Free params variable before exiting. Link: https://lkml.kernel.org/r/0be31d8259c7c53b98a39769d60cfeecd8421785.1645206561.git.bristot@kernel.org Fixes: 1eceb2fc2ca5 ("rtla/osnoise: Add osnoise top mode") Cc: Jonathan Corbet Signed-off-by: Daniel Bristot de Oliveira Signed-off-by: Steven Rostedt (Google) --- tools/tracing/rtla/src/osnoise_top.c | 1 + 1 file changed, 1 insertion(+) diff --git a/tools/tracing/rtla/src/osnoise_top.c b/tools/tracing/rtla/src/osnoise_top.c index c67dc28ef716..7af769b9c0de 100644 --- a/tools/tracing/rtla/src/osnoise_top.c +++ b/tools/tracing/rtla/src/osnoise_top.c @@ -573,6 +573,7 @@ out_top: osnoise_free_top(tool->data); osnoise_destroy_tool(record); osnoise_destroy_tool(tool); + free(params); out_exit: exit(return_value); } From 90f59ee41abf587ad4675a70434136c8707fdf4b Mon Sep 17 00:00:00 2001 From: Daniel Bristot de Oliveira Date: Fri, 18 Feb 2022 18:57:09 +0100 Subject: [PATCH 13/13] rtla/osnoise: Fix error message when failing to enable trace instance When a trace instance creation fails, tools are printing: Could not enable -> osnoiser <- tracer for tracing Print the actual (and correct) name of the tracer it fails to enable. Link: https://lkml.kernel.org/r/53ef0582605af91eca14b19dba9fc9febb95d4f9.1645206561.git.bristot@kernel.org Fixes: b1696371d865 ("rtla: Helper functions for rtla") Cc: Jonathan Corbet Signed-off-by: Daniel Bristot de Oliveira Signed-off-by: Steven Rostedt (Google) --- tools/tracing/rtla/src/osnoise.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tools/tracing/rtla/src/osnoise.c b/tools/tracing/rtla/src/osnoise.c index 5648f9252e58..e60f1862bad0 100644 --- a/tools/tracing/rtla/src/osnoise.c +++ b/tools/tracing/rtla/src/osnoise.c @@ -810,7 +810,7 @@ struct osnoise_tool *osnoise_init_trace_tool(char *tracer) retval = enable_tracer_by_name(trace->trace.inst, tracer); if (retval) { - err_msg("Could not enable osnoiser tracer for tracing\n"); + err_msg("Could not enable %s tracer for tracing\n", tracer); goto out_err; }