From dd17c8f72993f9461e9c19250e3f155d6d99df22 Mon Sep 17 00:00:00 2001 From: Rusty Russell Date: Thu, 29 Oct 2009 22:34:15 +0900 Subject: percpu: remove per_cpu__ prefix. Now that the return from alloc_percpu is compatible with the address of per-cpu vars, it makes sense to hand around the address of per-cpu variables. To make this sane, we remove the per_cpu__ prefix we used created to stop people accidentally using these vars directly. Now we have sparse, we can use that (next patch). tj: * Updated to convert stuff which were missed by or added after the original patch. * Kill per_cpu_var() macro. Signed-off-by: Rusty Russell Signed-off-by: Tejun Heo Reviewed-by: Christoph Lameter --- kernel/trace/trace.c | 6 +++--- kernel/trace/trace_functions_graph.c | 4 ++-- 2 files changed, 5 insertions(+), 5 deletions(-) (limited to 'kernel/trace') diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 85a5ed70b5b2..b808177af816 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -91,12 +91,12 @@ DEFINE_PER_CPU(int, ftrace_cpu_disabled); static inline void ftrace_disable_cpu(void) { preempt_disable(); - __this_cpu_inc(per_cpu_var(ftrace_cpu_disabled)); + __this_cpu_inc(ftrace_cpu_disabled); } static inline void ftrace_enable_cpu(void) { - __this_cpu_dec(per_cpu_var(ftrace_cpu_disabled)); + __this_cpu_dec(ftrace_cpu_disabled); preempt_enable(); } @@ -1085,7 +1085,7 @@ trace_function(struct trace_array *tr, struct ftrace_entry *entry; /* If we are reading the ring buffer, don't trace */ - if (unlikely(__this_cpu_read(per_cpu_var(ftrace_cpu_disabled)))) + if (unlikely(__this_cpu_read(ftrace_cpu_disabled))) return; event = trace_buffer_lock_reserve(buffer, TRACE_FN, sizeof(*entry), diff --git a/kernel/trace/trace_functions_graph.c b/kernel/trace/trace_functions_graph.c index 90a6daa10962..8614e3241ff8 100644 --- a/kernel/trace/trace_functions_graph.c +++ b/kernel/trace/trace_functions_graph.c @@ -176,7 +176,7 @@ static int __trace_graph_entry(struct trace_array *tr, struct ring_buffer *buffer = tr->buffer; struct ftrace_graph_ent_entry *entry; - if (unlikely(__this_cpu_read(per_cpu_var(ftrace_cpu_disabled)))) + if (unlikely(__this_cpu_read(ftrace_cpu_disabled))) return 0; event = trace_buffer_lock_reserve(buffer, TRACE_GRAPH_ENT, @@ -240,7 +240,7 @@ static void __trace_graph_return(struct trace_array *tr, struct ring_buffer *buffer = tr->buffer; struct ftrace_graph_ret_entry *entry; - if (unlikely(__this_cpu_read(per_cpu_var(ftrace_cpu_disabled)))) + if (unlikely(__this_cpu_read(ftrace_cpu_disabled))) return; event = trace_buffer_lock_reserve(buffer, TRACE_GRAPH_RET, -- cgit v1.2.3 From 07b139c8c81b97bbe55c68daf0cbeca8b1c609ca Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Mon, 21 Dec 2009 14:27:35 +0800 Subject: perf events: Remove CONFIG_EVENT_PROFILE Quoted from Ingo: | This reminds me - i think we should eliminate CONFIG_EVENT_PROFILE - | it's an unnecessary Kconfig complication. If both PERF_EVENTS and | EVENT_TRACING is enabled we should expose generic tracepoints. | | Nor is it limited to event 'profiling', so it has become a misnomer as | well. Signed-off-by: Li Zefan Cc: Frederic Weisbecker Cc: Steven Rostedt Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <4B2F1557.2050705@cn.fujitsu.com> Signed-off-by: Ingo Molnar --- kernel/trace/Makefile | 4 +++- kernel/trace/trace_events_filter.c | 4 ++-- kernel/trace/trace_kprobe.c | 14 +++++++------- kernel/trace/trace_syscalls.c | 5 ++--- 4 files changed, 14 insertions(+), 13 deletions(-) (limited to 'kernel/trace') diff --git a/kernel/trace/Makefile b/kernel/trace/Makefile index cd9ecd89ec77..d00c6fe23f54 100644 --- a/kernel/trace/Makefile +++ b/kernel/trace/Makefile @@ -51,7 +51,9 @@ endif obj-$(CONFIG_EVENT_TRACING) += trace_events.o obj-$(CONFIG_EVENT_TRACING) += trace_export.o obj-$(CONFIG_FTRACE_SYSCALLS) += trace_syscalls.o -obj-$(CONFIG_EVENT_PROFILE) += trace_event_profile.o +ifeq ($(CONFIG_PERF_EVENTS),y) +obj-$(CONFIG_EVENT_TRACING) += trace_event_profile.o +endif obj-$(CONFIG_EVENT_TRACING) += trace_events_filter.o obj-$(CONFIG_KPROBE_EVENT) += trace_kprobe.o obj-$(CONFIG_KSYM_TRACER) += trace_ksym.o diff --git a/kernel/trace/trace_events_filter.c b/kernel/trace/trace_events_filter.c index 50504cb228de..74563d7e102e 100644 --- a/kernel/trace/trace_events_filter.c +++ b/kernel/trace/trace_events_filter.c @@ -1360,7 +1360,7 @@ out_unlock: return err; } -#ifdef CONFIG_EVENT_PROFILE +#ifdef CONFIG_PERF_EVENTS void ftrace_profile_free_filter(struct perf_event *event) { @@ -1428,5 +1428,5 @@ out_unlock: return err; } -#endif /* CONFIG_EVENT_PROFILE */ +#endif /* CONFIG_PERF_EVENTS */ diff --git a/kernel/trace/trace_kprobe.c b/kernel/trace/trace_kprobe.c index 375f81a568dc..75d75dec226a 100644 --- a/kernel/trace/trace_kprobe.c +++ b/kernel/trace/trace_kprobe.c @@ -1249,7 +1249,7 @@ static int kretprobe_event_show_format(struct ftrace_event_call *call, ", REC->" FIELD_STRING_RETIP); } -#ifdef CONFIG_EVENT_PROFILE +#ifdef CONFIG_PERF_EVENTS /* Kprobe profile handler */ static __kprobes int kprobe_profile_func(struct kprobe *kp, @@ -1407,7 +1407,7 @@ static void probe_profile_disable(struct ftrace_event_call *call) disable_kprobe(&tp->rp.kp); } } -#endif /* CONFIG_EVENT_PROFILE */ +#endif /* CONFIG_PERF_EVENTS */ static __kprobes @@ -1417,10 +1417,10 @@ int kprobe_dispatcher(struct kprobe *kp, struct pt_regs *regs) if (tp->flags & TP_FLAG_TRACE) kprobe_trace_func(kp, regs); -#ifdef CONFIG_EVENT_PROFILE +#ifdef CONFIG_PERF_EVENTS if (tp->flags & TP_FLAG_PROFILE) kprobe_profile_func(kp, regs); -#endif /* CONFIG_EVENT_PROFILE */ +#endif return 0; /* We don't tweek kernel, so just return 0 */ } @@ -1431,10 +1431,10 @@ int kretprobe_dispatcher(struct kretprobe_instance *ri, struct pt_regs *regs) if (tp->flags & TP_FLAG_TRACE) kretprobe_trace_func(ri, regs); -#ifdef CONFIG_EVENT_PROFILE +#ifdef CONFIG_PERF_EVENTS if (tp->flags & TP_FLAG_PROFILE) kretprobe_profile_func(ri, regs); -#endif /* CONFIG_EVENT_PROFILE */ +#endif return 0; /* We don't tweek kernel, so just return 0 */ } @@ -1463,7 +1463,7 @@ static int register_probe_event(struct trace_probe *tp) call->regfunc = probe_event_enable; call->unregfunc = probe_event_disable; -#ifdef CONFIG_EVENT_PROFILE +#ifdef CONFIG_PERF_EVENTS call->profile_enable = probe_profile_enable; call->profile_disable = probe_profile_disable; #endif diff --git a/kernel/trace/trace_syscalls.c b/kernel/trace/trace_syscalls.c index 75289f372dd2..f694f66d75b0 100644 --- a/kernel/trace/trace_syscalls.c +++ b/kernel/trace/trace_syscalls.c @@ -421,7 +421,7 @@ int __init init_ftrace_syscalls(void) } core_initcall(init_ftrace_syscalls); -#ifdef CONFIG_EVENT_PROFILE +#ifdef CONFIG_PERF_EVENTS static DECLARE_BITMAP(enabled_prof_enter_syscalls, NR_syscalls); static DECLARE_BITMAP(enabled_prof_exit_syscalls, NR_syscalls); @@ -626,6 +626,5 @@ void prof_sysexit_disable(struct ftrace_event_call *call) mutex_unlock(&syscall_trace_lock); } -#endif - +#endif /* CONFIG_PERF_EVENTS */ -- cgit v1.2.3 From 79615760f380ec86cd58204744e774c33fab9211 Mon Sep 17 00:00:00 2001 From: Christoph Lameter Date: Tue, 5 Jan 2010 15:34:50 +0900 Subject: local_t: Move local.h include to ringbuffer.c and ring_buffer_benchmark.c ringbuffer*.c are the last users of local.h. Remove the include from modules.h and add it to ringbuffer files. Signed-off-by: Christoph Lameter Signed-off-by: Tejun Heo --- kernel/trace/ring_buffer.c | 1 + kernel/trace/ring_buffer_benchmark.c | 1 + 2 files changed, 2 insertions(+) (limited to 'kernel/trace') diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c index 2326b04c95c4..eb6c8988c31a 100644 --- a/kernel/trace/ring_buffer.c +++ b/kernel/trace/ring_buffer.c @@ -20,6 +20,7 @@ #include #include +#include #include "trace.h" /* diff --git a/kernel/trace/ring_buffer_benchmark.c b/kernel/trace/ring_buffer_benchmark.c index b2477caf09c2..df74c7982255 100644 --- a/kernel/trace/ring_buffer_benchmark.c +++ b/kernel/trace/ring_buffer_benchmark.c @@ -8,6 +8,7 @@ #include #include #include +#include struct rb_page { u64 ts; -- cgit v1.2.3 From 809826a389040e0ad9d646b587bccc0e34691afd Mon Sep 17 00:00:00 2001 From: Lai Jiangshan Date: Tue, 15 Dec 2009 15:39:34 +0800 Subject: tracing: Have __dynamic_array() define a field This is part of a patch set that removes the show_format method in the ftrace event macros. This patch set requires that all fields are added to the ftrace_event_call->fields. This patch changes __dynamic_array() to call trace_define_field() to include fields that use __dynamic_array(). Signed-off-by: Lai Jiangshan LKML-Reference: <4B273D36.8090100@cn.fujitsu.com> Signed-off-by: Steven Rostedt --- kernel/trace/trace_export.c | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-) (limited to 'kernel/trace') diff --git a/kernel/trace/trace_export.c b/kernel/trace/trace_export.c index d4fa5dc1ee4e..9978a4f40090 100644 --- a/kernel/trace/trace_export.c +++ b/kernel/trace/trace_export.c @@ -175,7 +175,12 @@ ftrace_format_##name(struct ftrace_event_call *unused, \ return ret; #undef __dynamic_array -#define __dynamic_array(type, item) +#define __dynamic_array(type, item) \ + ret = trace_define_field(event_call, #type, #item, \ + offsetof(typeof(field), item), \ + 0, is_signed_type(type), FILTER_OTHER);\ + if (ret) \ + return ret; #undef FTRACE_ENTRY #define FTRACE_ENTRY(name, struct_name, id, tstruct, print) \ -- cgit v1.2.3 From 509e760cd91c831983097ae174cb6c0b8c6c8e6b Mon Sep 17 00:00:00 2001 From: Lai Jiangshan Date: Tue, 15 Dec 2009 15:39:42 +0800 Subject: tracing: Add print_fmt field This is part of a patch set that removes the show_format method in the ftrace event macros. The print_fmt field is added to hold the string that shows the print_fmt in the event format files. This patch only adds the field but it is currently not used. Later patches will use this field to enable us to remove the show_format field and function. Signed-off-by: Lai Jiangshan LKML-Reference: <4B273D3E.2000704@cn.fujitsu.com> Signed-off-by: Steven Rostedt --- kernel/trace/trace_export.c | 7 +++++++ 1 file changed, 7 insertions(+) (limited to 'kernel/trace') diff --git a/kernel/trace/trace_export.c b/kernel/trace/trace_export.c index 9978a4f40090..95d14b640a66 100644 --- a/kernel/trace/trace_export.c +++ b/kernel/trace/trace_export.c @@ -203,6 +203,9 @@ static int ftrace_raw_init_event(struct ftrace_event_call *call) return 0; } +#undef __entry +#define __entry REC + #undef __field #define __field(type, item) @@ -218,6 +221,9 @@ static int ftrace_raw_init_event(struct ftrace_event_call *call) #undef __dynamic_array #define __dynamic_array(type, item) +#undef F_printk +#define F_printk(fmt, args...) #fmt ", " __stringify(args) + #undef FTRACE_ENTRY #define FTRACE_ENTRY(call, struct_name, type, tstruct, print) \ \ @@ -228,6 +234,7 @@ __attribute__((section("_ftrace_events"))) event_##call = { \ .id = type, \ .system = __stringify(TRACE_SYSTEM), \ .raw_init = ftrace_raw_init_event, \ + .print_fmt = print, \ .show_format = ftrace_format_##call, \ .define_fields = ftrace_define_fields_##call, \ }; \ -- cgit v1.2.3 From 50307a45f8515f6244e3b08e6b19824b9fbfe293 Mon Sep 17 00:00:00 2001 From: Lai Jiangshan Date: Tue, 15 Dec 2009 15:39:45 +0800 Subject: tracing/syscalls: Init print_fmt for syscall events This is part of a patch set that removes the show_format method in the ftrace event macros. Add the print_fmt initialization to the syscall events. The print_fmt is still not used, but will be in the follow up patches. Signed-off-by: Lai Jiangshan LKML-Reference: <4B273D41.609@cn.fujitsu.com> Signed-off-by: Steven Rostedt --- kernel/trace/trace_syscalls.c | 68 ++++++++++++++++++++++++++++++++++++++++++- 1 file changed, 67 insertions(+), 1 deletion(-) (limited to 'kernel/trace') diff --git a/kernel/trace/trace_syscalls.c b/kernel/trace/trace_syscalls.c index 75289f372dd2..1352b0a36fac 100644 --- a/kernel/trace/trace_syscalls.c +++ b/kernel/trace/trace_syscalls.c @@ -191,6 +191,67 @@ int syscall_enter_format(struct ftrace_event_call *call, struct trace_seq *s) return trace_seq_putc(s, '\n'); } +static +int __set_enter_print_fmt(struct syscall_metadata *entry, char *buf, int len) +{ + int i; + int pos = 0; + + /* When len=0, we just calculate the needed length */ +#define LEN_OR_ZERO (len ? len - pos : 0) + + pos += snprintf(buf + pos, LEN_OR_ZERO, "\""); + for (i = 0; i < entry->nb_args; i++) { + pos += snprintf(buf + pos, LEN_OR_ZERO, "%s: 0x%%0%zulx%s", + entry->args[i], sizeof(unsigned long), + i == entry->nb_args - 1 ? "" : ", "); + } + pos += snprintf(buf + pos, LEN_OR_ZERO, "\""); + + for (i = 0; i < entry->nb_args; i++) { + pos += snprintf(buf + pos, LEN_OR_ZERO, + ", ((unsigned long)(REC->%s))", entry->args[i]); + } + +#undef LEN_OR_ZERO + + /* return the length of print_fmt */ + return pos; +} + +static int set_syscall_print_fmt(struct ftrace_event_call *call) +{ + char *print_fmt; + int len; + struct syscall_metadata *entry = call->data; + + if (entry->enter_event != call) { + call->print_fmt = "\"0x%lx\", REC->ret"; + return 0; + } + + /* First: called with 0 length to calculate the needed length */ + len = __set_enter_print_fmt(entry, NULL, 0); + + print_fmt = kmalloc(len + 1, GFP_KERNEL); + if (!print_fmt) + return -ENOMEM; + + /* Second: actually write the @print_fmt */ + __set_enter_print_fmt(entry, print_fmt, len + 1); + call->print_fmt = print_fmt; + + return 0; +} + +static void free_syscall_print_fmt(struct ftrace_event_call *call) +{ + struct syscall_metadata *entry = call->data; + + if (entry->enter_event == call) + kfree(call->print_fmt); +} + int syscall_exit_format(struct ftrace_event_call *call, struct trace_seq *s) { int ret; @@ -386,9 +447,14 @@ int init_syscall_trace(struct ftrace_event_call *call) { int id; + if (set_syscall_print_fmt(call) < 0) + return -ENOMEM; + id = register_ftrace_event(call->event); - if (!id) + if (!id) { + free_syscall_print_fmt(call); return -ENODEV; + } call->id = id; INIT_LIST_HEAD(&call->fields); return 0; -- cgit v1.2.3 From a342a0280b981c130e32dbb94dbd3a57959c4d04 Mon Sep 17 00:00:00 2001 From: Lai Jiangshan Date: Tue, 15 Dec 2009 15:39:49 +0800 Subject: tracing/kprobes: Init print_fmt for kprobe events This is part of a patch set that removes the show_format method in the ftrace event macros. Add the print_fmt initialization to the kprobe events. The print_fmt is still not used, but will be in the follow up patches. Signed-off-by: Lai Jiangshan LKML-Reference: <4B273D45.3080100@cn.fujitsu.com> Acked-by: Masami Hiramatsu Signed-off-by: Steven Rostedt --- kernel/trace/trace_kprobe.c | 64 ++++++++++++++++++++++++++++++++++++++++++++- 1 file changed, 63 insertions(+), 1 deletion(-) (limited to 'kernel/trace') diff --git a/kernel/trace/trace_kprobe.c b/kernel/trace/trace_kprobe.c index 6ea90c0e2c96..147491dccead 100644 --- a/kernel/trace/trace_kprobe.c +++ b/kernel/trace/trace_kprobe.c @@ -1250,6 +1250,62 @@ static int kretprobe_event_show_format(struct ftrace_event_call *call, ", REC->" FIELD_STRING_RETIP); } +static int __set_print_fmt(struct trace_probe *tp, char *buf, int len) +{ + int i; + int pos = 0; + + const char *fmt, *arg; + + if (!probe_is_return(tp)) { + fmt = "(%lx)"; + arg = "REC->" FIELD_STRING_IP; + } else { + fmt = "(%lx <- %lx)"; + arg = "REC->" FIELD_STRING_FUNC ", REC->" FIELD_STRING_RETIP; + } + + /* When len=0, we just calculate the needed length */ +#define LEN_OR_ZERO (len ? len - pos : 0) + + pos += snprintf(buf + pos, LEN_OR_ZERO, "\"%s", fmt); + + for (i = 0; i < tp->nr_args; i++) { + pos += snprintf(buf + pos, LEN_OR_ZERO, " %s=%%lx", + tp->args[i].name); + } + + pos += snprintf(buf + pos, LEN_OR_ZERO, "\", %s", arg); + + for (i = 0; i < tp->nr_args; i++) { + pos += snprintf(buf + pos, LEN_OR_ZERO, ", REC->%s", + tp->args[i].name); + } + +#undef LEN_OR_ZERO + + /* return the length of print_fmt */ + return pos; +} + +static int set_print_fmt(struct trace_probe *tp) +{ + int len; + char *print_fmt; + + /* First: called with 0 length to calculate the needed length */ + len = __set_print_fmt(tp, NULL, 0); + print_fmt = kmalloc(len + 1, GFP_KERNEL); + if (!print_fmt) + return -ENOMEM; + + /* Second: actually write the @print_fmt */ + __set_print_fmt(tp, print_fmt, len + 1); + tp->call.print_fmt = print_fmt; + + return 0; +} + #ifdef CONFIG_EVENT_PROFILE /* Kprobe profile handler */ @@ -1456,10 +1512,14 @@ static int register_probe_event(struct trace_probe *tp) call->show_format = kprobe_event_show_format; call->define_fields = kprobe_event_define_fields; } + if (set_print_fmt(tp) < 0) + return -ENOMEM; call->event = &tp->event; call->id = register_ftrace_event(&tp->event); - if (!call->id) + if (!call->id) { + kfree(call->print_fmt); return -ENODEV; + } call->enabled = 0; call->regfunc = probe_event_enable; call->unregfunc = probe_event_disable; @@ -1472,6 +1532,7 @@ static int register_probe_event(struct trace_probe *tp) ret = trace_add_event_call(call); if (ret) { pr_info("Failed to register kprobe event: %s\n", call->name); + kfree(call->print_fmt); unregister_ftrace_event(&tp->event); } return ret; @@ -1481,6 +1542,7 @@ static void unregister_probe_event(struct trace_probe *tp) { /* tp->event is unregistered in trace_remove_event_call() */ trace_remove_event_call(&tp->call); + kfree(tp->call.print_fmt); } /* Make a debugfs interface for controling probe points */ -- cgit v1.2.3 From c7ef3a9004201bca90626db246a19dadd2c29c9b Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Mon, 28 Dec 2009 21:13:59 -0500 Subject: tracing: Have syscall tracing call its own init function In the clean up of having all events call one specific function, the syscall event init was changed to call this helper function. With the new print_fmt updates, the syscalls need to do special initializations. This patch converts the syscall events to call its own init function again. Cc: Lai Jiangshan Cc: Li Zefan Signed-off-by: Steven Rostedt --- kernel/trace/trace_syscalls.c | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-) (limited to 'kernel/trace') diff --git a/kernel/trace/trace_syscalls.c b/kernel/trace/trace_syscalls.c index 1352b0a36fac..a78e86349ecb 100644 --- a/kernel/trace/trace_syscalls.c +++ b/kernel/trace/trace_syscalls.c @@ -450,14 +450,14 @@ int init_syscall_trace(struct ftrace_event_call *call) if (set_syscall_print_fmt(call) < 0) return -ENOMEM; - id = register_ftrace_event(call->event); - if (!id) { + id = trace_event_raw_init(call); + + if (id < 0) { free_syscall_print_fmt(call); - return -ENODEV; + return id; } - call->id = id; - INIT_LIST_HEAD(&call->fields); - return 0; + + return id; } int __init init_ftrace_syscalls(void) -- cgit v1.2.3 From 5a65e956220efc2421e21ee56d6153fd5c533a95 Mon Sep 17 00:00:00 2001 From: Lai Jiangshan Date: Tue, 15 Dec 2009 15:39:53 +0800 Subject: tracing: Use defined fields and print_fmt to print formats The calls ftrace_format_##call() and ftrace_define_fields_##call() are almost duplicate in functionality. With the addition of the print_fmt in previous patches, these two functions can be merged into one. The trace_define_field() defines the fields and links them into the struct ftrace_event_call. The previous patches introduced the print_fmt field and this can now be used with the trace_define_field() to create the event format file fields and print_fmt field. The struct ftrace_event_call->fields are used to print the fields The struct ftrace_event_call->print_fmt is used to print the "print fmt: XXXXXXXXXXX" line. Signed-off-by: Lai Jiangshan LKML-Reference: <4B273D49.5000006@cn.fujitsu.com> Signed-off-by: Steven Rostedt --- kernel/trace/trace_events.c | 65 ++++++++++++++++++++++++++++----------------- 1 file changed, 41 insertions(+), 24 deletions(-) (limited to 'kernel/trace') diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c index 189b09baf4fb..250ec865d5f5 100644 --- a/kernel/trace/trace_events.c +++ b/kernel/trace/trace_events.c @@ -528,33 +528,16 @@ extern char *__bad_type_size(void); #type, "common_" #name, offsetof(typeof(field), name), \ sizeof(field.name), is_signed_type(type) -static int trace_write_header(struct trace_seq *s) -{ - struct trace_entry field; - - /* struct trace_entry */ - return trace_seq_printf(s, - "\tfield:%s %s;\toffset:%zu;\tsize:%zu;\tsigned:%u;\n" - "\tfield:%s %s;\toffset:%zu;\tsize:%zu;\tsigned:%u;\n" - "\tfield:%s %s;\toffset:%zu;\tsize:%zu;\tsigned:%u;\n" - "\tfield:%s %s;\toffset:%zu;\tsize:%zu;\tsigned:%u;\n" - "\tfield:%s %s;\toffset:%zu;\tsize:%zu;\tsigned:%u;\n" - "\n", - FIELD(unsigned short, type), - FIELD(unsigned char, flags), - FIELD(unsigned char, preempt_count), - FIELD(int, pid), - FIELD(int, lock_depth)); -} - static ssize_t event_format_read(struct file *filp, char __user *ubuf, size_t cnt, loff_t *ppos) { struct ftrace_event_call *call = filp->private_data; + struct ftrace_event_field *field; struct trace_seq *s; + int common_field_count = 5; char *buf; - int r; + int r = 0; if (*ppos) return 0; @@ -565,14 +548,48 @@ event_format_read(struct file *filp, char __user *ubuf, size_t cnt, trace_seq_init(s); - /* If any of the first writes fail, so will the show_format. */ - trace_seq_printf(s, "name: %s\n", call->name); trace_seq_printf(s, "ID: %d\n", call->id); trace_seq_printf(s, "format:\n"); - trace_write_header(s); - r = call->show_format(call, s); + list_for_each_entry_reverse(field, &call->fields, link) { + /* + * Smartly shows the array type(except dynamic array). + * Normal: + * field:TYPE VAR + * If TYPE := TYPE[LEN], it is shown: + * field:TYPE VAR[LEN] + */ + const char *array_descriptor = strchr(field->type, '['); + + if (!strncmp(field->type, "__data_loc", 10)) + array_descriptor = NULL; + + if (!array_descriptor) { + r = trace_seq_printf(s, "\tfield:%s %s;\toffset:%u;" + "\tsize:%u;\tsigned:%d;\n", + field->type, field->name, field->offset, + field->size, !!field->is_signed); + } else { + r = trace_seq_printf(s, "\tfield:%.*s %s%s;\toffset:%u;" + "\tsize:%u;\tsigned:%d;\n", + (int)(array_descriptor - field->type), + field->type, field->name, + array_descriptor, field->offset, + field->size, !!field->is_signed); + } + + if (--common_field_count == 0) + r = trace_seq_printf(s, "\n"); + + if (!r) + break; + } + + if (r) + r = trace_seq_printf(s, "\nprint fmt: %s\n", + call->print_fmt); + if (!r) { /* * ug! The format output is bigger than a PAGE!! -- cgit v1.2.3 From 0fa0edaf32b9a78b9854f1da98d4511a501089b0 Mon Sep 17 00:00:00 2001 From: Lai Jiangshan Date: Tue, 15 Dec 2009 15:39:57 +0800 Subject: tracing: Remove show_format and related macros from TRACE_EVENT The previous patches added the use of print_fmt string and changes the trace_define_field() function to also create the fields and format output for the event format files. text data bss dec hex filename 5857201 1355780 9336808 16549789 fc879d vmlinux 5884589 1351684 9337896 16574169 fce6d9 vmlinux-orig The above shows the size of the vmlinux after this patch set compared to the vmlinux-orig which is before the patch set. This saves us 27k on text, 1k on bss and adds just 4k of data. The total savings of 24k in size. Signed-off-by: Lai Jiangshan LKML-Reference: <4B273D4D.40604@cn.fujitsu.com> Acked-by: Masami Hiramatsu Signed-off-by: Steven Rostedt --- kernel/trace/trace_events.c | 12 ------- kernel/trace/trace_export.c | 73 ---------------------------------------- kernel/trace/trace_kprobe.c | 78 ------------------------------------------- kernel/trace/trace_syscalls.c | 66 ------------------------------------ 4 files changed, 229 deletions(-) (limited to 'kernel/trace') diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c index 250ec865d5f5..c2a3077b7353 100644 --- a/kernel/trace/trace_events.c +++ b/kernel/trace/trace_events.c @@ -520,14 +520,6 @@ out: return ret; } -extern char *__bad_type_size(void); - -#undef FIELD -#define FIELD(type, name) \ - sizeof(type) != sizeof(field.name) ? __bad_type_size() : \ - #type, "common_" #name, offsetof(typeof(field), name), \ - sizeof(field.name), is_signed_type(type) - static ssize_t event_format_read(struct file *filp, char __user *ubuf, size_t cnt, loff_t *ppos) @@ -965,10 +957,6 @@ event_create_dir(struct ftrace_event_call *call, struct dentry *d_events, filter); } - /* A trace may not want to export its format */ - if (!call->show_format) - return 0; - trace_create_file("format", 0444, call->dir, call, format); diff --git a/kernel/trace/trace_export.c b/kernel/trace/trace_export.c index 95d14b640a66..e091f64ba6ce 100644 --- a/kernel/trace/trace_export.c +++ b/kernel/trace/trace_export.c @@ -62,78 +62,6 @@ static void __always_unused ____ftrace_check_##name(void) \ #include "trace_entries.h" - -#undef __field -#define __field(type, item) \ - ret = trace_seq_printf(s, "\tfield:" #type " " #item ";\t" \ - "offset:%zu;\tsize:%zu;\tsigned:%u;\n", \ - offsetof(typeof(field), item), \ - sizeof(field.item), is_signed_type(type)); \ - if (!ret) \ - return 0; - -#undef __field_desc -#define __field_desc(type, container, item) \ - ret = trace_seq_printf(s, "\tfield:" #type " " #item ";\t" \ - "offset:%zu;\tsize:%zu;\tsigned:%u;\n", \ - offsetof(typeof(field), container.item), \ - sizeof(field.container.item), \ - is_signed_type(type)); \ - if (!ret) \ - return 0; - -#undef __array -#define __array(type, item, len) \ - ret = trace_seq_printf(s, "\tfield:" #type " " #item "[" #len "];\t" \ - "offset:%zu;\tsize:%zu;\tsigned:%u;\n", \ - offsetof(typeof(field), item), \ - sizeof(field.item), is_signed_type(type)); \ - if (!ret) \ - return 0; - -#undef __array_desc -#define __array_desc(type, container, item, len) \ - ret = trace_seq_printf(s, "\tfield:" #type " " #item "[" #len "];\t" \ - "offset:%zu;\tsize:%zu;\tsigned:%u;\n", \ - offsetof(typeof(field), container.item), \ - sizeof(field.container.item), \ - is_signed_type(type)); \ - if (!ret) \ - return 0; - -#undef __dynamic_array -#define __dynamic_array(type, item) \ - ret = trace_seq_printf(s, "\tfield:" #type " " #item ";\t" \ - "offset:%zu;\tsize:0;\tsigned:%u;\n", \ - offsetof(typeof(field), item), \ - is_signed_type(type)); \ - if (!ret) \ - return 0; - -#undef F_printk -#define F_printk(fmt, args...) "%s, %s\n", #fmt, __stringify(args) - -#undef __entry -#define __entry REC - -#undef FTRACE_ENTRY -#define FTRACE_ENTRY(name, struct_name, id, tstruct, print) \ -static int \ -ftrace_format_##name(struct ftrace_event_call *unused, \ - struct trace_seq *s) \ -{ \ - struct struct_name field __attribute__((unused)); \ - int ret = 0; \ - \ - tstruct; \ - \ - trace_seq_printf(s, "\nprint fmt: " print); \ - \ - return ret; \ -} - -#include "trace_entries.h" - #undef __field #define __field(type, item) \ ret = trace_define_field(event_call, #type, #item, \ @@ -235,7 +163,6 @@ __attribute__((section("_ftrace_events"))) event_##call = { \ .system = __stringify(TRACE_SYSTEM), \ .raw_init = ftrace_raw_init_event, \ .print_fmt = print, \ - .show_format = ftrace_format_##call, \ .define_fields = ftrace_define_fields_##call, \ }; \ diff --git a/kernel/trace/trace_kprobe.c b/kernel/trace/trace_kprobe.c index 147491dccead..c99029916c76 100644 --- a/kernel/trace/trace_kprobe.c +++ b/kernel/trace/trace_kprobe.c @@ -1174,82 +1174,6 @@ static int kretprobe_event_define_fields(struct ftrace_event_call *event_call) return 0; } -static int __probe_event_show_format(struct trace_seq *s, - struct trace_probe *tp, const char *fmt, - const char *arg) -{ - int i; - - /* Show format */ - if (!trace_seq_printf(s, "\nprint fmt: \"%s", fmt)) - return 0; - - for (i = 0; i < tp->nr_args; i++) - if (!trace_seq_printf(s, " %s=%%lx", tp->args[i].name)) - return 0; - - if (!trace_seq_printf(s, "\", %s", arg)) - return 0; - - for (i = 0; i < tp->nr_args; i++) - if (!trace_seq_printf(s, ", REC->%s", tp->args[i].name)) - return 0; - - return trace_seq_puts(s, "\n"); -} - -#undef SHOW_FIELD -#define SHOW_FIELD(type, item, name) \ - do { \ - ret = trace_seq_printf(s, "\tfield:" #type " %s;\t" \ - "offset:%u;\tsize:%u;\tsigned:%d;\n", name,\ - (unsigned int)offsetof(typeof(field), item),\ - (unsigned int)sizeof(type), \ - is_signed_type(type)); \ - if (!ret) \ - return 0; \ - } while (0) - -static int kprobe_event_show_format(struct ftrace_event_call *call, - struct trace_seq *s) -{ - struct kprobe_trace_entry field __attribute__((unused)); - int ret, i; - struct trace_probe *tp = (struct trace_probe *)call->data; - - SHOW_FIELD(unsigned long, ip, FIELD_STRING_IP); - SHOW_FIELD(int, nargs, FIELD_STRING_NARGS); - - /* Show fields */ - for (i = 0; i < tp->nr_args; i++) - SHOW_FIELD(unsigned long, args[i], tp->args[i].name); - trace_seq_puts(s, "\n"); - - return __probe_event_show_format(s, tp, "(%lx)", - "REC->" FIELD_STRING_IP); -} - -static int kretprobe_event_show_format(struct ftrace_event_call *call, - struct trace_seq *s) -{ - struct kretprobe_trace_entry field __attribute__((unused)); - int ret, i; - struct trace_probe *tp = (struct trace_probe *)call->data; - - SHOW_FIELD(unsigned long, func, FIELD_STRING_FUNC); - SHOW_FIELD(unsigned long, ret_ip, FIELD_STRING_RETIP); - SHOW_FIELD(int, nargs, FIELD_STRING_NARGS); - - /* Show fields */ - for (i = 0; i < tp->nr_args; i++) - SHOW_FIELD(unsigned long, args[i], tp->args[i].name); - trace_seq_puts(s, "\n"); - - return __probe_event_show_format(s, tp, "(%lx <- %lx)", - "REC->" FIELD_STRING_FUNC - ", REC->" FIELD_STRING_RETIP); -} - static int __set_print_fmt(struct trace_probe *tp, char *buf, int len) { int i; @@ -1504,12 +1428,10 @@ static int register_probe_event(struct trace_probe *tp) if (probe_is_return(tp)) { tp->event.trace = print_kretprobe_event; call->raw_init = probe_event_raw_init; - call->show_format = kretprobe_event_show_format; call->define_fields = kretprobe_event_define_fields; } else { tp->event.trace = print_kprobe_event; call->raw_init = probe_event_raw_init; - call->show_format = kprobe_event_show_format; call->define_fields = kprobe_event_define_fields; } if (set_print_fmt(tp) < 0) diff --git a/kernel/trace/trace_syscalls.c b/kernel/trace/trace_syscalls.c index a78e86349ecb..49cea70fbf6d 100644 --- a/kernel/trace/trace_syscalls.c +++ b/kernel/trace/trace_syscalls.c @@ -143,54 +143,6 @@ extern char *__bad_type_size(void); #type, #name, offsetof(typeof(trace), name), \ sizeof(trace.name), is_signed_type(type) -int syscall_enter_format(struct ftrace_event_call *call, struct trace_seq *s) -{ - int i; - int ret; - struct syscall_metadata *entry = call->data; - struct syscall_trace_enter trace; - int offset = offsetof(struct syscall_trace_enter, args); - - ret = trace_seq_printf(s, "\tfield:%s %s;\toffset:%zu;\tsize:%zu;" - "\tsigned:%u;\n", - SYSCALL_FIELD(int, nr)); - if (!ret) - return 0; - - for (i = 0; i < entry->nb_args; i++) { - ret = trace_seq_printf(s, "\tfield:%s %s;", entry->types[i], - entry->args[i]); - if (!ret) - return 0; - ret = trace_seq_printf(s, "\toffset:%d;\tsize:%zu;" - "\tsigned:%u;\n", offset, - sizeof(unsigned long), - is_signed_type(unsigned long)); - if (!ret) - return 0; - offset += sizeof(unsigned long); - } - - trace_seq_puts(s, "\nprint fmt: \""); - for (i = 0; i < entry->nb_args; i++) { - ret = trace_seq_printf(s, "%s: 0x%%0%zulx%s", entry->args[i], - sizeof(unsigned long), - i == entry->nb_args - 1 ? "" : ", "); - if (!ret) - return 0; - } - trace_seq_putc(s, '"'); - - for (i = 0; i < entry->nb_args; i++) { - ret = trace_seq_printf(s, ", ((unsigned long)(REC->%s))", - entry->args[i]); - if (!ret) - return 0; - } - - return trace_seq_putc(s, '\n'); -} - static int __set_enter_print_fmt(struct syscall_metadata *entry, char *buf, int len) { @@ -252,24 +204,6 @@ static void free_syscall_print_fmt(struct ftrace_event_call *call) kfree(call->print_fmt); } -int syscall_exit_format(struct ftrace_event_call *call, struct trace_seq *s) -{ - int ret; - struct syscall_trace_exit trace; - - ret = trace_seq_printf(s, - "\tfield:%s %s;\toffset:%zu;\tsize:%zu;" - "\tsigned:%u;\n" - "\tfield:%s %s;\toffset:%zu;\tsize:%zu;" - "\tsigned:%u;\n", - SYSCALL_FIELD(int, nr), - SYSCALL_FIELD(long, ret)); - if (!ret) - return 0; - - return trace_seq_printf(s, "\nprint fmt: \"0x%%lx\", REC->ret\n"); -} - int syscall_enter_define_fields(struct ftrace_event_call *call) { struct syscall_trace_enter trace; -- cgit v1.2.3 From 7e53bd42d14c75192b99674c40fcc359392da59d Mon Sep 17 00:00:00 2001 From: Lai Jiangshan Date: Wed, 6 Jan 2010 20:08:50 +0800 Subject: tracing: Consolidate protection of reader access to the ring buffer At the beginning, access to the ring buffer was fully serialized by trace_types_lock. Patch d7350c3f4569 gives more freedom to readers, and patch b04cc6b1f6 adds code to protect trace_pipe and cpu#/trace_pipe. But actually it is not enough, ring buffer readers are not always read-only, they may consume data. This patch makes accesses to trace, trace_pipe, trace_pipe_raw cpu#/trace, cpu#/trace_pipe and cpu#/trace_pipe_raw serialized. And removes tracing_reader_cpumask which is used to protect trace_pipe. Details: Ring buffer serializes readers, but it is low level protection. The validity of the events (which returns by ring_buffer_peek() ..etc) are not protected by ring buffer. The content of events may become garbage if we allow another process to consume these events concurrently: A) the page of the consumed events may become a normal page (not reader page) in ring buffer, and this page will be rewritten by the events producer. B) The page of the consumed events may become a page for splice_read, and this page will be returned to system. This patch adds trace_access_lock() and trace_access_unlock() primitives. These primitives allow multi process access to different cpu ring buffers concurrently. These primitives don't distinguish read-only and read-consume access. Multi read-only access is also serialized. And we don't use these primitives when we open files, we only use them when we read files. Signed-off-by: Lai Jiangshan LKML-Reference: <4B447D52.1050602@cn.fujitsu.com> Signed-off-by: Steven Rostedt --- kernel/trace/trace.c | 136 ++++++++++++++++++++++++++++++++++++--------------- 1 file changed, 97 insertions(+), 39 deletions(-) (limited to 'kernel/trace') diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 0df1b0f2cb9e..abdd333a0825 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -32,6 +32,7 @@ #include #include #include +#include #include #include #include @@ -102,9 +103,6 @@ static inline void ftrace_enable_cpu(void) static cpumask_var_t __read_mostly tracing_buffer_mask; -/* Define which cpu buffers are currently read in trace_pipe */ -static cpumask_var_t tracing_reader_cpumask; - #define for_each_tracing_cpu(cpu) \ for_each_cpu(cpu, tracing_buffer_mask) @@ -243,12 +241,91 @@ static struct tracer *current_trace __read_mostly; /* * trace_types_lock is used to protect the trace_types list. - * This lock is also used to keep user access serialized. - * Accesses from userspace will grab this lock while userspace - * activities happen inside the kernel. */ static DEFINE_MUTEX(trace_types_lock); +/* + * serialize the access of the ring buffer + * + * ring buffer serializes readers, but it is low level protection. + * The validity of the events (which returns by ring_buffer_peek() ..etc) + * are not protected by ring buffer. + * + * The content of events may become garbage if we allow other process consumes + * these events concurrently: + * A) the page of the consumed events may become a normal page + * (not reader page) in ring buffer, and this page will be rewrited + * by events producer. + * B) The page of the consumed events may become a page for splice_read, + * and this page will be returned to system. + * + * These primitives allow multi process access to different cpu ring buffer + * concurrently. + * + * These primitives don't distinguish read-only and read-consume access. + * Multi read-only access are also serialized. + */ + +#ifdef CONFIG_SMP +static DECLARE_RWSEM(all_cpu_access_lock); +static DEFINE_PER_CPU(struct mutex, cpu_access_lock); + +static inline void trace_access_lock(int cpu) +{ + if (cpu == TRACE_PIPE_ALL_CPU) { + /* gain it for accessing the whole ring buffer. */ + down_write(&all_cpu_access_lock); + } else { + /* gain it for accessing a cpu ring buffer. */ + + /* Firstly block other trace_access_lock(TRACE_PIPE_ALL_CPU). */ + down_read(&all_cpu_access_lock); + + /* Secondly block other access to this @cpu ring buffer. */ + mutex_lock(&per_cpu(cpu_access_lock, cpu)); + } +} + +static inline void trace_access_unlock(int cpu) +{ + if (cpu == TRACE_PIPE_ALL_CPU) { + up_write(&all_cpu_access_lock); + } else { + mutex_unlock(&per_cpu(cpu_access_lock, cpu)); + up_read(&all_cpu_access_lock); + } +} + +static inline void trace_access_lock_init(void) +{ + int cpu; + + for_each_possible_cpu(cpu) + mutex_init(&per_cpu(cpu_access_lock, cpu)); +} + +#else + +static DEFINE_MUTEX(access_lock); + +static inline void trace_access_lock(int cpu) +{ + (void)cpu; + mutex_lock(&access_lock); +} + +static inline void trace_access_unlock(int cpu) +{ + (void)cpu; + mutex_unlock(&access_lock); +} + +static inline void trace_access_lock_init(void) +{ +} + +#endif + /* trace_wait is a waitqueue for tasks blocked on trace_poll */ static DECLARE_WAIT_QUEUE_HEAD(trace_wait); @@ -1580,12 +1657,6 @@ static void tracing_iter_reset(struct trace_iterator *iter, int cpu) } /* - * No necessary locking here. The worst thing which can - * happen is loosing events consumed at the same time - * by a trace_pipe reader. - * Other than that, we don't risk to crash the ring buffer - * because it serializes the readers. - * * The current tracer is copied to avoid a global locking * all around. */ @@ -1640,12 +1711,16 @@ static void *s_start(struct seq_file *m, loff_t *pos) } trace_event_read_lock(); + trace_access_lock(cpu_file); return p; } static void s_stop(struct seq_file *m, void *p) { + struct trace_iterator *iter = m->private; + atomic_dec(&trace_record_cmdline_disabled); + trace_access_unlock(iter->cpu_file); trace_event_read_unlock(); } @@ -2836,22 +2911,6 @@ static int tracing_open_pipe(struct inode *inode, struct file *filp) mutex_lock(&trace_types_lock); - /* We only allow one reader per cpu */ - if (cpu_file == TRACE_PIPE_ALL_CPU) { - if (!cpumask_empty(tracing_reader_cpumask)) { - ret = -EBUSY; - goto out; - } - cpumask_setall(tracing_reader_cpumask); - } else { - if (!cpumask_test_cpu(cpu_file, tracing_reader_cpumask)) - cpumask_set_cpu(cpu_file, tracing_reader_cpumask); - else { - ret = -EBUSY; - goto out; - } - } - /* create a buffer to store the information to pass to userspace */ iter = kzalloc(sizeof(*iter), GFP_KERNEL); if (!iter) { @@ -2907,12 +2966,6 @@ static int tracing_release_pipe(struct inode *inode, struct file *file) mutex_lock(&trace_types_lock); - if (iter->cpu_file == TRACE_PIPE_ALL_CPU) - cpumask_clear(tracing_reader_cpumask); - else - cpumask_clear_cpu(iter->cpu_file, tracing_reader_cpumask); - - if (iter->trace->pipe_close) iter->trace->pipe_close(iter); @@ -3074,6 +3127,7 @@ waitagain: iter->pos = -1; trace_event_read_lock(); + trace_access_lock(iter->cpu_file); while (find_next_entry_inc(iter) != NULL) { enum print_line_t ret; int len = iter->seq.len; @@ -3090,6 +3144,7 @@ waitagain: if (iter->seq.len >= cnt) break; } + trace_access_unlock(iter->cpu_file); trace_event_read_unlock(); /* Now copy what we have to the user */ @@ -3215,6 +3270,7 @@ static ssize_t tracing_splice_read_pipe(struct file *filp, } trace_event_read_lock(); + trace_access_lock(iter->cpu_file); /* Fill as many pages as possible. */ for (i = 0, rem = len; i < PIPE_BUFFERS && rem; i++) { @@ -3238,6 +3294,7 @@ static ssize_t tracing_splice_read_pipe(struct file *filp, trace_seq_init(&iter->seq); } + trace_access_unlock(iter->cpu_file); trace_event_read_unlock(); mutex_unlock(&iter->mutex); @@ -3539,10 +3596,12 @@ tracing_buffers_read(struct file *filp, char __user *ubuf, info->read = 0; + trace_access_lock(info->cpu); ret = ring_buffer_read_page(info->tr->buffer, &info->spare, count, info->cpu, 0); + trace_access_unlock(info->cpu); if (ret < 0) return 0; @@ -3670,6 +3729,7 @@ tracing_buffers_splice_read(struct file *file, loff_t *ppos, len &= PAGE_MASK; } + trace_access_lock(info->cpu); entries = ring_buffer_entries_cpu(info->tr->buffer, info->cpu); for (i = 0; i < PIPE_BUFFERS && len && entries; i++, len -= PAGE_SIZE) { @@ -3717,6 +3777,7 @@ tracing_buffers_splice_read(struct file *file, loff_t *ppos, entries = ring_buffer_entries_cpu(info->tr->buffer, info->cpu); } + trace_access_unlock(info->cpu); spd.nr_pages = i; /* did we read anything? */ @@ -4153,6 +4214,8 @@ static __init int tracer_init_debugfs(void) struct dentry *d_tracer; int cpu; + trace_access_lock_init(); + d_tracer = tracing_init_dentry(); trace_create_file("tracing_enabled", 0644, d_tracer, @@ -4387,9 +4450,6 @@ __init static int tracer_alloc_buffers(void) if (!alloc_cpumask_var(&tracing_cpumask, GFP_KERNEL)) goto out_free_buffer_mask; - if (!zalloc_cpumask_var(&tracing_reader_cpumask, GFP_KERNEL)) - goto out_free_tracing_cpumask; - /* To save memory, keep the ring buffer size to its minimum */ if (ring_buffer_expanded) ring_buf_size = trace_buf_size; @@ -4447,8 +4507,6 @@ __init static int tracer_alloc_buffers(void) return 0; out_free_cpumask: - free_cpumask_var(tracing_reader_cpumask); -out_free_tracing_cpumask: free_cpumask_var(tracing_cpumask); out_free_buffer_mask: free_cpumask_var(tracing_buffer_mask); -- cgit v1.2.3 From d931369b74b3d6f2044f595af6f3dd074f65d9cf Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Wed, 6 Jan 2010 17:27:11 -0500 Subject: tracing: Add stack dump to trace_printk if stacktrace option is set If the ftrace stacktrace option is set, then add the stack dumps to trace_printk. Requested-by: Peter Zijlstra Signed-off-by: Steven Rostedt --- kernel/trace/trace.c | 8 ++++++-- 1 file changed, 6 insertions(+), 2 deletions(-) (limited to 'kernel/trace') diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index abdd333a0825..5314c90bbc83 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -1392,8 +1392,10 @@ int trace_vbprintk(unsigned long ip, const char *fmt, va_list args) entry->fmt = fmt; memcpy(entry->buf, trace_buf, sizeof(u32) * len); - if (!filter_check_discard(call, entry, buffer, event)) + if (!filter_check_discard(call, entry, buffer, event)) { ring_buffer_unlock_commit(buffer, event); + ftrace_trace_stack(buffer, flags, 6, pc); + } out_unlock: arch_spin_unlock(&trace_buf_lock); @@ -1466,8 +1468,10 @@ int trace_array_vprintk(struct trace_array *tr, memcpy(&entry->buf, trace_buf, len); entry->buf[len] = '\0'; - if (!filter_check_discard(call, entry, buffer, event)) + if (!filter_check_discard(call, entry, buffer, event)) { ring_buffer_unlock_commit(buffer, event); + ftrace_trace_stack(buffer, irq_flags, 6, pc); + } out_unlock: arch_spin_unlock(&trace_buf_lock); -- cgit v1.2.3 From 14640106f243a3b29944d7198569090fa6546f2d Mon Sep 17 00:00:00 2001 From: Masami Hiramatsu Date: Tue, 5 Jan 2010 17:46:48 -0500 Subject: tracing/kprobe: Drop function argument access syntax Drop function argument access syntax, because the function arguments depend on not only architecture but also compile-options and function API. And now, we have perf-probe for finding register/memory assigned to each argument. Signed-off-by: Masami Hiramatsu Cc: Frederic Weisbecker Cc: Arnaldo Carvalho de Melo Cc: systemtap Cc: DLE Cc: Frederic Weisbecker Cc: Steven Rostedt Cc: Roland McGrath Cc: Oleg Nesterov Cc: Mahesh Salgaonkar Cc: Benjamin Herrenschmidt Cc: Michael Neuling Cc: linuxppc-dev@ozlabs.org LKML-Reference: <20100105224648.19431.52309.stgit@dhcp-100-2-132.bos.redhat.com> Signed-off-by: Ingo Molnar --- kernel/trace/trace_kprobe.c | 18 +----------------- 1 file changed, 1 insertion(+), 17 deletions(-) (limited to 'kernel/trace') diff --git a/kernel/trace/trace_kprobe.c b/kernel/trace/trace_kprobe.c index 47f54ab57b68..7ac728ded964 100644 --- a/kernel/trace/trace_kprobe.c +++ b/kernel/trace/trace_kprobe.c @@ -91,11 +91,6 @@ static __kprobes unsigned long fetch_memory(struct pt_regs *regs, void *addr) return retval; } -static __kprobes unsigned long fetch_argument(struct pt_regs *regs, void *num) -{ - return regs_get_argument_nth(regs, (unsigned int)((unsigned long)num)); -} - static __kprobes unsigned long fetch_retvalue(struct pt_regs *regs, void *dummy) { @@ -231,9 +226,7 @@ static int probe_arg_string(char *buf, size_t n, struct fetch_func *ff) { int ret = -EINVAL; - if (ff->func == fetch_argument) - ret = snprintf(buf, n, "$arg%lu", (unsigned long)ff->data); - else if (ff->func == fetch_register) { + if (ff->func == fetch_register) { const char *name; name = regs_query_register_name((unsigned int)((long)ff->data)); ret = snprintf(buf, n, "%%%s", name); @@ -489,14 +482,6 @@ static int parse_probe_vars(char *arg, struct fetch_func *ff, int is_return) } } else ret = -EINVAL; - } else if (strncmp(arg, "arg", 3) == 0 && isdigit(arg[3])) { - ret = strict_strtoul(arg + 3, 10, ¶m); - if (ret || param > PARAM_MAX_ARGS) - ret = -EINVAL; - else { - ff->func = fetch_argument; - ff->data = (void *)param; - } } else ret = -EINVAL; return ret; @@ -611,7 +596,6 @@ static int create_trace_probe(int argc, char **argv) * - Add kprobe: p[:[GRP/]EVENT] KSYM[+OFFS]|KADDR [FETCHARGS] * - Add kretprobe: r[:[GRP/]EVENT] KSYM[+0] [FETCHARGS] * Fetch args: - * $argN : fetch Nth of function argument. (N:0-) * $retval : fetch return value * $stack : fetch stack address * $stackN : fetch Nth of stack (N:0-) -- cgit v1.2.3 From 24a53652e361321b09df5040711e69387344ce09 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Thu, 14 Jan 2010 08:23:15 +0100 Subject: tracing: Drop the tr check from the graph tracing path Each time we save a function entry from the function graph tracer, we check if the trace array is set, which is wasteful because it is set anyway before we start the tracer. All we need is to ensure we have good read and write orderings. When we set the trace array, we just need to guarantee it to be visible before starting tracing. Signed-off-by: Frederic Weisbecker Acked-by: Steven Rostedt Cc: Lai Jiangshan Cc: Paul E. McKenney LKML-Reference: <1263453795-7496-1-git-send-regression-fweisbec@gmail.com> Signed-off-by: Ingo Molnar --- kernel/trace/trace_functions_graph.c | 19 ++++++++++--------- 1 file changed, 10 insertions(+), 9 deletions(-) (limited to 'kernel/trace') diff --git a/kernel/trace/trace_functions_graph.c b/kernel/trace/trace_functions_graph.c index b1342c5d37cf..f2252296607c 100644 --- a/kernel/trace/trace_functions_graph.c +++ b/kernel/trace/trace_functions_graph.c @@ -212,9 +212,6 @@ int trace_graph_entry(struct ftrace_graph_ent *trace) int cpu; int pc; - if (unlikely(!tr)) - return 0; - if (!ftrace_trace_task(current)) return 0; @@ -287,11 +284,20 @@ void trace_graph_return(struct ftrace_graph_ret *trace) local_irq_restore(flags); } +void set_graph_array(struct trace_array *tr) +{ + graph_array = tr; + + /* Make graph_array visible before we start tracing */ + + smp_mb(); +} + static int graph_trace_init(struct trace_array *tr) { int ret; - graph_array = tr; + set_graph_array(tr); ret = register_ftrace_graph(&trace_graph_return, &trace_graph_entry); if (ret) @@ -301,11 +307,6 @@ static int graph_trace_init(struct trace_array *tr) return 0; } -void set_graph_array(struct trace_array *tr) -{ - graph_array = tr; -} - static void graph_trace_reset(struct trace_array *tr) { tracing_stop_cmdline_record(); -- cgit v1.2.3 From 231e36f4d2e63dd770db80b9f5113310c2bcfcfd Mon Sep 17 00:00:00 2001 From: Masami Hiramatsu Date: Thu, 14 Jan 2010 00:12:12 -0500 Subject: tracing/kprobe: Update kprobe tracing self test for new syntax Update kprobe tracing self test for new syntax (it supports deleting individual probes, and drops $argN support) and behavior change (new probes are disabled in default). This selftest includes the following checks: - Adding function-entry probe and return probe with arguments. - Enabling these probes. - Deleting it individually. Signed-off-by: Masami Hiramatsu Cc: systemtap Cc: DLE Cc: Steven Rostedt Cc: Frederic Weisbecker LKML-Reference: <20100114051211.7814.29436.stgit@localhost6.localdomain6> Signed-off-by: Ingo Molnar --- kernel/trace/trace_kprobe.c | 55 ++++++++++++++++++++++++++++++++++++++------- 1 file changed, 47 insertions(+), 8 deletions(-) (limited to 'kernel/trace') diff --git a/kernel/trace/trace_kprobe.c b/kernel/trace/trace_kprobe.c index 7ac728ded964..d6266cad6953 100644 --- a/kernel/trace/trace_kprobe.c +++ b/kernel/trace/trace_kprobe.c @@ -1507,28 +1507,67 @@ static int kprobe_trace_selftest_target(int a1, int a2, int a3, static __init int kprobe_trace_self_tests_init(void) { - int ret; + int ret, warn = 0; int (*target)(int, int, int, int, int, int); + struct trace_probe *tp; target = kprobe_trace_selftest_target; pr_info("Testing kprobe tracing: "); ret = command_trace_probe("p:testprobe kprobe_trace_selftest_target " - "$arg1 $arg2 $arg3 $arg4 $stack $stack0"); - if (WARN_ON_ONCE(ret)) - pr_warning("error enabling function entry\n"); + "$stack $stack0 +0($stack)"); + if (WARN_ON_ONCE(ret)) { + pr_warning("error on probing function entry.\n"); + warn++; + } else { + /* Enable trace point */ + tp = find_probe_event("testprobe", KPROBE_EVENT_SYSTEM); + if (WARN_ON_ONCE(tp == NULL)) { + pr_warning("error on getting new probe.\n"); + warn++; + } else + probe_event_enable(&tp->call); + } ret = command_trace_probe("r:testprobe2 kprobe_trace_selftest_target " "$retval"); - if (WARN_ON_ONCE(ret)) - pr_warning("error enabling function return\n"); + if (WARN_ON_ONCE(ret)) { + pr_warning("error on probing function return.\n"); + warn++; + } else { + /* Enable trace point */ + tp = find_probe_event("testprobe2", KPROBE_EVENT_SYSTEM); + if (WARN_ON_ONCE(tp == NULL)) { + pr_warning("error on getting new probe.\n"); + warn++; + } else + probe_event_enable(&tp->call); + } + + if (warn) + goto end; ret = target(1, 2, 3, 4, 5, 6); - cleanup_all_probes(); + ret = command_trace_probe("-:testprobe"); + if (WARN_ON_ONCE(ret)) { + pr_warning("error on deleting a probe.\n"); + warn++; + } + + ret = command_trace_probe("-:testprobe2"); + if (WARN_ON_ONCE(ret)) { + pr_warning("error on deleting a probe.\n"); + warn++; + } - pr_cont("OK\n"); +end: + cleanup_all_probes(); + if (warn) + pr_cont("NG: Some tests are failed. Please check them.\n"); + else + pr_cont("OK\n"); return 0; } -- cgit v1.2.3 From ea2c68a08fedb5053ba312d661e47df9f4d72411 Mon Sep 17 00:00:00 2001 From: Lai Jiangshan Date: Wed, 13 Jan 2010 19:38:30 +0800 Subject: tracing: Simplify test for function_graph tracing start point In the function graph tracer, a calling function is to be traced only when it is enabled through the set_graph_function file, or when it is nested in an enabled function. Current code uses TSK_TRACE_FL_GRAPH to test whether it is nested or not. Looking at the code, we can get this: (trace->depth > 0) <==> (TSK_TRACE_FL_GRAPH is set) trace->depth is more explicit to tell that it is nested. So we use trace->depth directly and simplify the code. No functionality is changed. TSK_TRACE_FL_GRAPH is not removed yet, it is left for future usage. Signed-off-by: Lai Jiangshan Cc: Ingo Molnar Cc: Steven Rostedt LKML-Reference: <4B4DB0B6.7040607@cn.fujitsu.com> Signed-off-by: Frederic Weisbecker --- kernel/trace/trace.h | 2 +- kernel/trace/trace_functions_graph.c | 8 ++------ 2 files changed, 3 insertions(+), 7 deletions(-) (limited to 'kernel/trace') diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index 4df6a77eb196..ce077fbbf552 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -504,7 +504,7 @@ static inline int ftrace_graph_addr(unsigned long addr) { int i; - if (!ftrace_graph_count || test_tsk_trace_graph(current)) + if (!ftrace_graph_count) return 1; for (i = 0; i < ftrace_graph_count; i++) { diff --git a/kernel/trace/trace_functions_graph.c b/kernel/trace/trace_functions_graph.c index f2252296607c..616b135c9eb9 100644 --- a/kernel/trace/trace_functions_graph.c +++ b/kernel/trace/trace_functions_graph.c @@ -215,7 +215,8 @@ int trace_graph_entry(struct ftrace_graph_ent *trace) if (!ftrace_trace_task(current)) return 0; - if (!ftrace_graph_addr(trace->func)) + /* trace it when it is-nested-in or is a function enabled. */ + if (!(trace->depth || ftrace_graph_addr(trace->func))) return 0; local_irq_save(flags); @@ -228,9 +229,6 @@ int trace_graph_entry(struct ftrace_graph_ent *trace) } else { ret = 0; } - /* Only do the atomic if it is not already set */ - if (!test_tsk_trace_graph(current)) - set_tsk_trace_graph(current); atomic_dec(&data->disabled); local_irq_restore(flags); @@ -278,8 +276,6 @@ void trace_graph_return(struct ftrace_graph_ret *trace) pc = preempt_count(); __trace_graph_return(tr, trace, flags, pc); } - if (!trace->depth) - clear_tsk_trace_graph(current); atomic_dec(&data->disabled); local_irq_restore(flags); } -- cgit v1.2.3 From 430ad5a600a83956749307b13257c464c3826b55 Mon Sep 17 00:00:00 2001 From: Xiao Guangrong Date: Thu, 28 Jan 2010 09:32:29 +0800 Subject: perf: Factorize trace events raw sample buffer operations Introduce ftrace_perf_buf_prepare() and ftrace_perf_buf_submit() to gather the common code that operates on raw events sampling buffer. This cleans up redundant code between regular trace events, syscall events and kprobe events. Changelog v1->v2: - Rename function name as per Masami and Frederic's suggestion - Add __kprobes for ftrace_perf_buf_prepare() and make ftrace_perf_buf_submit() inline as per Masami's suggestion - Export ftrace_perf_buf_prepare since modules will use it Signed-off-by: Xiao Guangrong Acked-by: Masami Hiramatsu Cc: Ingo Molnar Cc: Steven Rostedt Cc: Paul Mackerras Cc: Jason Baron Cc: Peter Zijlstra LKML-Reference: <4B60E92D.9000808@cn.fujitsu.com> Signed-off-by: Frederic Weisbecker --- kernel/trace/trace_event_profile.c | 52 ++++++++++++++++++++--- kernel/trace/trace_kprobe.c | 86 +++++--------------------------------- kernel/trace/trace_syscalls.c | 71 +++++-------------------------- 3 files changed, 67 insertions(+), 142 deletions(-) (limited to 'kernel/trace') diff --git a/kernel/trace/trace_event_profile.c b/kernel/trace/trace_event_profile.c index 9e25573242cf..f0d693005075 100644 --- a/kernel/trace/trace_event_profile.c +++ b/kernel/trace/trace_event_profile.c @@ -6,14 +6,12 @@ */ #include +#include #include "trace.h" -char *perf_trace_buf; -EXPORT_SYMBOL_GPL(perf_trace_buf); - -char *perf_trace_buf_nmi; -EXPORT_SYMBOL_GPL(perf_trace_buf_nmi); +static char *perf_trace_buf; +static char *perf_trace_buf_nmi; typedef typeof(char [FTRACE_MAX_PROFILE_SIZE]) perf_trace_t ; @@ -120,3 +118,47 @@ void ftrace_profile_disable(int event_id) } mutex_unlock(&event_mutex); } + +__kprobes void *ftrace_perf_buf_prepare(int size, unsigned short type, + int *rctxp, unsigned long *irq_flags) +{ + struct trace_entry *entry; + char *trace_buf, *raw_data; + int pc, cpu; + + pc = preempt_count(); + + /* Protect the per cpu buffer, begin the rcu read side */ + local_irq_save(*irq_flags); + + *rctxp = perf_swevent_get_recursion_context(); + if (*rctxp < 0) + goto err_recursion; + + cpu = smp_processor_id(); + + if (in_nmi()) + trace_buf = rcu_dereference(perf_trace_buf_nmi); + else + trace_buf = rcu_dereference(perf_trace_buf); + + if (!trace_buf) + goto err; + + raw_data = per_cpu_ptr(trace_buf, cpu); + + /* zero the dead bytes from align to not leak stack to user */ + *(u64 *)(&raw_data[size - sizeof(u64)]) = 0ULL; + + entry = (struct trace_entry *)raw_data; + tracing_generic_entry_update(entry, *irq_flags, pc); + entry->type = type; + + return raw_data; +err: + perf_swevent_put_recursion_context(*rctxp); +err_recursion: + local_irq_restore(*irq_flags); + return NULL; +} +EXPORT_SYMBOL_GPL(ftrace_perf_buf_prepare); diff --git a/kernel/trace/trace_kprobe.c b/kernel/trace/trace_kprobe.c index d6266cad6953..2e28ee36646f 100644 --- a/kernel/trace/trace_kprobe.c +++ b/kernel/trace/trace_kprobe.c @@ -1243,14 +1243,10 @@ static __kprobes int kprobe_profile_func(struct kprobe *kp, struct trace_probe *tp = container_of(kp, struct trace_probe, rp.kp); struct ftrace_event_call *call = &tp->call; struct kprobe_trace_entry *entry; - struct trace_entry *ent; - int size, __size, i, pc, __cpu; + int size, __size, i; unsigned long irq_flags; - char *trace_buf; - char *raw_data; int rctx; - pc = preempt_count(); __size = SIZEOF_KPROBE_TRACE_ENTRY(tp->nr_args); size = ALIGN(__size + sizeof(u32), sizeof(u64)); size -= sizeof(u32); @@ -1258,45 +1254,16 @@ static __kprobes int kprobe_profile_func(struct kprobe *kp, "profile buffer not large enough")) return 0; - /* - * Protect the non nmi buffer - * This also protects the rcu read side - */ - local_irq_save(irq_flags); - - rctx = perf_swevent_get_recursion_context(); - if (rctx < 0) - goto end_recursion; - - __cpu = smp_processor_id(); - - if (in_nmi()) - trace_buf = rcu_dereference(perf_trace_buf_nmi); - else - trace_buf = rcu_dereference(perf_trace_buf); - - if (!trace_buf) - goto end; - - raw_data = per_cpu_ptr(trace_buf, __cpu); - - /* Zero dead bytes from alignment to avoid buffer leak to userspace */ - *(u64 *)(&raw_data[size - sizeof(u64)]) = 0ULL; - entry = (struct kprobe_trace_entry *)raw_data; - ent = &entry->ent; + entry = ftrace_perf_buf_prepare(size, call->id, &rctx, &irq_flags); + if (!entry) + return 0; - tracing_generic_entry_update(ent, irq_flags, pc); - ent->type = call->id; entry->nargs = tp->nr_args; entry->ip = (unsigned long)kp->addr; for (i = 0; i < tp->nr_args; i++) entry->args[i] = call_fetch(&tp->args[i].fetch, regs); - perf_tp_event(call->id, entry->ip, 1, entry, size); -end: - perf_swevent_put_recursion_context(rctx); -end_recursion: - local_irq_restore(irq_flags); + ftrace_perf_buf_submit(entry, size, rctx, entry->ip, 1, irq_flags); return 0; } @@ -1308,14 +1275,10 @@ static __kprobes int kretprobe_profile_func(struct kretprobe_instance *ri, struct trace_probe *tp = container_of(ri->rp, struct trace_probe, rp); struct ftrace_event_call *call = &tp->call; struct kretprobe_trace_entry *entry; - struct trace_entry *ent; - int size, __size, i, pc, __cpu; + int size, __size, i; unsigned long irq_flags; - char *trace_buf; - char *raw_data; int rctx; - pc = preempt_count(); __size = SIZEOF_KRETPROBE_TRACE_ENTRY(tp->nr_args); size = ALIGN(__size + sizeof(u32), sizeof(u64)); size -= sizeof(u32); @@ -1323,46 +1286,17 @@ static __kprobes int kretprobe_profile_func(struct kretprobe_instance *ri, "profile buffer not large enough")) return 0; - /* - * Protect the non nmi buffer - * This also protects the rcu read side - */ - local_irq_save(irq_flags); - - rctx = perf_swevent_get_recursion_context(); - if (rctx < 0) - goto end_recursion; - - __cpu = smp_processor_id(); - - if (in_nmi()) - trace_buf = rcu_dereference(perf_trace_buf_nmi); - else - trace_buf = rcu_dereference(perf_trace_buf); - - if (!trace_buf) - goto end; - - raw_data = per_cpu_ptr(trace_buf, __cpu); - - /* Zero dead bytes from alignment to avoid buffer leak to userspace */ - *(u64 *)(&raw_data[size - sizeof(u64)]) = 0ULL; - entry = (struct kretprobe_trace_entry *)raw_data; - ent = &entry->ent; + entry = ftrace_perf_buf_prepare(size, call->id, &rctx, &irq_flags); + if (!entry) + return 0; - tracing_generic_entry_update(ent, irq_flags, pc); - ent->type = call->id; entry->nargs = tp->nr_args; entry->func = (unsigned long)tp->rp.kp.addr; entry->ret_ip = (unsigned long)ri->ret_addr; for (i = 0; i < tp->nr_args; i++) entry->args[i] = call_fetch(&tp->args[i].fetch, regs); - perf_tp_event(call->id, entry->ret_ip, 1, entry, size); -end: - perf_swevent_put_recursion_context(rctx); -end_recursion: - local_irq_restore(irq_flags); + ftrace_perf_buf_submit(entry, size, rctx, entry->ret_ip, 1, irq_flags); return 0; } diff --git a/kernel/trace/trace_syscalls.c b/kernel/trace/trace_syscalls.c index f694f66d75b0..4e332b9e449c 100644 --- a/kernel/trace/trace_syscalls.c +++ b/kernel/trace/trace_syscalls.c @@ -433,12 +433,9 @@ static void prof_syscall_enter(struct pt_regs *regs, long id) struct syscall_metadata *sys_data; struct syscall_trace_enter *rec; unsigned long flags; - char *trace_buf; - char *raw_data; int syscall_nr; int rctx; int size; - int cpu; syscall_nr = syscall_get_nr(current, regs); if (!test_bit(syscall_nr, enabled_prof_enter_syscalls)) @@ -457,37 +454,15 @@ static void prof_syscall_enter(struct pt_regs *regs, long id) "profile buffer not large enough")) return; - /* Protect the per cpu buffer, begin the rcu read side */ - local_irq_save(flags); - - rctx = perf_swevent_get_recursion_context(); - if (rctx < 0) - goto end_recursion; - - cpu = smp_processor_id(); - - trace_buf = rcu_dereference(perf_trace_buf); - - if (!trace_buf) - goto end; - - raw_data = per_cpu_ptr