* [PATCH v13 01/12] fprobe: Pass return address to the handlers
2023-05-26 4:17 [PATCH v13 00/12] tracing: Add fprobe/tracepoint events Masami Hiramatsu (Google)
@ 2023-05-26 4:17 ` Masami Hiramatsu (Google)
2023-05-26 4:18 ` [PATCH v13 02/12] tracing/probes: Avoid setting TPARG_FL_FENTRY and TPARG_FL_RETURN Masami Hiramatsu (Google)
` (10 subsequent siblings)
11 siblings, 0 replies; 17+ messages in thread
From: Masami Hiramatsu (Google) @ 2023-05-26 4:17 UTC (permalink / raw)
To: linux-trace-kernel
Cc: linux-kernel, Steven Rostedt, mhiramat, Florent Revest,
Mark Rutland, Will Deacon, Mathieu Desnoyers, Martin KaFai Lau,
bpf, Bagas Sanjaya
From: Masami Hiramatsu (Google) <mhiramat@kernel.org>
Pass return address as 'ret_ip' to the fprobe entry and return handlers
so that the fprobe user handler can get the reutrn address without
analyzing arch-dependent pt_regs.
Signed-off-by: Masami Hiramatsu (Google) <mhiramat@kernel.org>
---
Changes in v10:
- Add the reason why this requires in patch description.
---
include/linux/fprobe.h | 6 ++++--
include/linux/rethook.h | 2 +-
kernel/kprobes.c | 1 +
kernel/trace/bpf_trace.c | 6 ++++--
kernel/trace/fprobe.c | 6 +++---
kernel/trace/rethook.c | 3 ++-
lib/test_fprobe.c | 10 +++++++---
samples/fprobe/fprobe_example.c | 6 ++++--
8 files changed, 26 insertions(+), 14 deletions(-)
diff --git a/include/linux/fprobe.h b/include/linux/fprobe.h
index 47fefc7f363b..134f0f59ffa8 100644
--- a/include/linux/fprobe.h
+++ b/include/linux/fprobe.h
@@ -35,9 +35,11 @@ struct fprobe {
int nr_maxactive;
int (*entry_handler)(struct fprobe *fp, unsigned long entry_ip,
- struct pt_regs *regs, void *entry_data);
+ unsigned long ret_ip, struct pt_regs *regs,
+ void *entry_data);
void (*exit_handler)(struct fprobe *fp, unsigned long entry_ip,
- struct pt_regs *regs, void *entry_data);
+ unsigned long ret_ip, struct pt_regs *regs,
+ void *entry_data);
};
/* This fprobe is soft-disabled. */
diff --git a/include/linux/rethook.h b/include/linux/rethook.h
index c8ac1e5afcd1..fdf26cd0e742 100644
--- a/include/linux/rethook.h
+++ b/include/linux/rethook.h
@@ -14,7 +14,7 @@
struct rethook_node;
-typedef void (*rethook_handler_t) (struct rethook_node *, void *, struct pt_regs *);
+typedef void (*rethook_handler_t) (struct rethook_node *, void *, unsigned long, struct pt_regs *);
/**
* struct rethook - The rethook management data structure.
diff --git a/kernel/kprobes.c b/kernel/kprobes.c
index 00e177de91cc..ce13f1a35251 100644
--- a/kernel/kprobes.c
+++ b/kernel/kprobes.c
@@ -2127,6 +2127,7 @@ static int pre_handler_kretprobe(struct kprobe *p, struct pt_regs *regs)
NOKPROBE_SYMBOL(pre_handler_kretprobe);
static void kretprobe_rethook_handler(struct rethook_node *rh, void *data,
+ unsigned long ret_addr,
struct pt_regs *regs)
{
struct kretprobe *rp = (struct kretprobe *)data;
diff --git a/kernel/trace/bpf_trace.c b/kernel/trace/bpf_trace.c
index d804172b709c..c0a32118f08f 100644
--- a/kernel/trace/bpf_trace.c
+++ b/kernel/trace/bpf_trace.c
@@ -2646,7 +2646,8 @@ kprobe_multi_link_prog_run(struct bpf_kprobe_multi_link *link,
static int
kprobe_multi_link_handler(struct fprobe *fp, unsigned long fentry_ip,
- struct pt_regs *regs, void *data)
+ unsigned long ret_ip, struct pt_regs *regs,
+ void *data)
{
struct bpf_kprobe_multi_link *link;
@@ -2657,7 +2658,8 @@ kprobe_multi_link_handler(struct fprobe *fp, unsigned long fentry_ip,
static void
kprobe_multi_link_exit_handler(struct fprobe *fp, unsigned long fentry_ip,
- struct pt_regs *regs, void *data)
+ unsigned long ret_ip, struct pt_regs *regs,
+ void *data)
{
struct bpf_kprobe_multi_link *link;
diff --git a/kernel/trace/fprobe.c b/kernel/trace/fprobe.c
index 9abb3905bc8e..973bc664fcc1 100644
--- a/kernel/trace/fprobe.c
+++ b/kernel/trace/fprobe.c
@@ -52,7 +52,7 @@ static void fprobe_handler(unsigned long ip, unsigned long parent_ip,
}
if (fp->entry_handler)
- ret = fp->entry_handler(fp, ip, ftrace_get_regs(fregs), entry_data);
+ ret = fp->entry_handler(fp, ip, parent_ip, ftrace_get_regs(fregs), entry_data);
/* If entry_handler returns !0, nmissed is not counted. */
if (rh) {
@@ -81,7 +81,7 @@ static void fprobe_kprobe_handler(unsigned long ip, unsigned long parent_ip,
}
static void fprobe_exit_handler(struct rethook_node *rh, void *data,
- struct pt_regs *regs)
+ unsigned long ret_ip, struct pt_regs *regs)
{
struct fprobe *fp = (struct fprobe *)data;
struct fprobe_rethook_node *fpr;
@@ -91,7 +91,7 @@ static void fprobe_exit_handler(struct rethook_node *rh, void *data,
fpr = container_of(rh, struct fprobe_rethook_node, node);
- fp->exit_handler(fp, fpr->entry_ip, regs,
+ fp->exit_handler(fp, fpr->entry_ip, ret_ip, regs,
fp->entry_data_size ? (void *)fpr->data : NULL);
}
NOKPROBE_SYMBOL(fprobe_exit_handler);
diff --git a/kernel/trace/rethook.c b/kernel/trace/rethook.c
index 32c3dfdb4d6a..fc196e186737 100644
--- a/kernel/trace/rethook.c
+++ b/kernel/trace/rethook.c
@@ -301,7 +301,8 @@ unsigned long rethook_trampoline_handler(struct pt_regs *regs,
break;
handler = READ_ONCE(rhn->rethook->handler);
if (handler)
- handler(rhn, rhn->rethook->data, regs);
+ handler(rhn, rhn->rethook->data,
+ correct_ret_addr, regs);
if (first == node)
break;
diff --git a/lib/test_fprobe.c b/lib/test_fprobe.c
index 0fe5273e960b..ade7e3d93dac 100644
--- a/lib/test_fprobe.c
+++ b/lib/test_fprobe.c
@@ -39,7 +39,8 @@ static noinline u32 fprobe_selftest_nest_target(u32 value, u32 (*nest)(u32))
}
static notrace int fp_entry_handler(struct fprobe *fp, unsigned long ip,
- struct pt_regs *regs, void *data)
+ unsigned long ret_ip,
+ struct pt_regs *regs, void *data)
{
KUNIT_EXPECT_FALSE(current_test, preemptible());
/* This can be called on the fprobe_selftest_target and the fprobe_selftest_target2 */
@@ -57,6 +58,7 @@ static notrace int fp_entry_handler(struct fprobe *fp, unsigned long ip,
}
static notrace void fp_exit_handler(struct fprobe *fp, unsigned long ip,
+ unsigned long ret_ip,
struct pt_regs *regs, void *data)
{
unsigned long ret = regs_return_value(regs);
@@ -78,14 +80,16 @@ static notrace void fp_exit_handler(struct fprobe *fp, unsigned long ip,
}
static notrace int nest_entry_handler(struct fprobe *fp, unsigned long ip,
- struct pt_regs *regs, void *data)
+ unsigned long ret_ip,
+ struct pt_regs *regs, void *data)
{
KUNIT_EXPECT_FALSE(current_test, preemptible());
return 0;
}
static notrace void nest_exit_handler(struct fprobe *fp, unsigned long ip,
- struct pt_regs *regs, void *data)
+ unsigned long ret_ip,
+ struct pt_regs *regs, void *data)
{
KUNIT_EXPECT_FALSE(current_test, preemptible());
KUNIT_EXPECT_EQ(current_test, ip, target_nest_ip);
diff --git a/samples/fprobe/fprobe_example.c b/samples/fprobe/fprobe_example.c
index 4efc8feb6277..64e715e7ed11 100644
--- a/samples/fprobe/fprobe_example.c
+++ b/samples/fprobe/fprobe_example.c
@@ -49,6 +49,7 @@ static void show_backtrace(void)
}
static int sample_entry_handler(struct fprobe *fp, unsigned long ip,
+ unsigned long ret_ip,
struct pt_regs *regs, void *data)
{
if (use_trace)
@@ -65,10 +66,11 @@ static int sample_entry_handler(struct fprobe *fp, unsigned long ip,
return 0;
}
-static void sample_exit_handler(struct fprobe *fp, unsigned long ip, struct pt_regs *regs,
+static void sample_exit_handler(struct fprobe *fp, unsigned long ip,
+ unsigned long ret_ip, struct pt_regs *regs,
void *data)
{
- unsigned long rip = instruction_pointer(regs);
+ unsigned long rip = ret_ip;
if (use_trace)
/*
^ permalink raw reply related [flat|nested] 17+ messages in thread* [PATCH v13 02/12] tracing/probes: Avoid setting TPARG_FL_FENTRY and TPARG_FL_RETURN
2023-05-26 4:17 [PATCH v13 00/12] tracing: Add fprobe/tracepoint events Masami Hiramatsu (Google)
2023-05-26 4:17 ` [PATCH v13 01/12] fprobe: Pass return address to the handlers Masami Hiramatsu (Google)
@ 2023-05-26 4:18 ` Masami Hiramatsu (Google)
2023-05-26 4:18 ` [PATCH v13 03/12] tracing/probes: Add fprobe events for tracing function entry and exit Masami Hiramatsu (Google)
` (9 subsequent siblings)
11 siblings, 0 replies; 17+ messages in thread
From: Masami Hiramatsu (Google) @ 2023-05-26 4:18 UTC (permalink / raw)
To: linux-trace-kernel
Cc: linux-kernel, Steven Rostedt, mhiramat, Florent Revest,
Mark Rutland, Will Deacon, Mathieu Desnoyers, Martin KaFai Lau,
bpf, Bagas Sanjaya
From: Masami Hiramatsu (Google) <mhiramat@kernel.org>
When parsing a kprobe event, the return probe always sets both
TPARG_FL_RETURN and TPARG_FL_FENTRY, but this is not useful because
some fetchargs are only for return probe and some others only for
function entry. Make it obviously mutual exclusive.
Signed-off-by: Masami Hiramatsu (Google) <mhiramat@kernel.org>
---
kernel/trace/trace_kprobe.c | 2 +-
kernel/trace/trace_probe.h | 5 +++++
2 files changed, 6 insertions(+), 1 deletion(-)
diff --git a/kernel/trace/trace_kprobe.c b/kernel/trace/trace_kprobe.c
index 59cda19a9033..867ffb7ee31d 100644
--- a/kernel/trace/trace_kprobe.c
+++ b/kernel/trace/trace_kprobe.c
@@ -825,7 +825,7 @@ static int __trace_kprobe_create(int argc, const char *argv[])
if (is_return)
flags |= TPARG_FL_RETURN;
ret = kprobe_on_func_entry(NULL, symbol, offset);
- if (ret == 0)
+ if (ret == 0 && !is_return)
flags |= TPARG_FL_FENTRY;
/* Defer the ENOENT case until register kprobe */
if (ret == -EINVAL && is_return) {
diff --git a/kernel/trace/trace_probe.h b/kernel/trace/trace_probe.h
index ef8ed3b65d05..752a08fcf0be 100644
--- a/kernel/trace/trace_probe.h
+++ b/kernel/trace/trace_probe.h
@@ -357,6 +357,11 @@ int trace_probe_print_args(struct trace_seq *s, struct probe_arg *args, int nr_a
#define trace_probe_for_each_link_rcu(pos, tp) \
list_for_each_entry_rcu(pos, &(tp)->event->files, list)
+/*
+ * The flags used for parsing trace_probe arguments.
+ * TPARG_FL_RETURN, TPARG_FL_FENTRY and TPARG_FL_TPOINT are mutually exclusive.
+ * TPARG_FL_KERNEL and TPARG_FL_USER are also mutually exclusive.
+ */
#define TPARG_FL_RETURN BIT(0)
#define TPARG_FL_KERNEL BIT(1)
#define TPARG_FL_FENTRY BIT(2)
^ permalink raw reply related [flat|nested] 17+ messages in thread* [PATCH v13 03/12] tracing/probes: Add fprobe events for tracing function entry and exit.
2023-05-26 4:17 [PATCH v13 00/12] tracing: Add fprobe/tracepoint events Masami Hiramatsu (Google)
2023-05-26 4:17 ` [PATCH v13 01/12] fprobe: Pass return address to the handlers Masami Hiramatsu (Google)
2023-05-26 4:18 ` [PATCH v13 02/12] tracing/probes: Avoid setting TPARG_FL_FENTRY and TPARG_FL_RETURN Masami Hiramatsu (Google)
@ 2023-05-26 4:18 ` Masami Hiramatsu (Google)
2023-05-26 19:38 ` Florent Revest
2023-05-26 4:18 ` [PATCH v13 04/12] selftests/ftrace: Add fprobe related testcases Masami Hiramatsu (Google)
` (8 subsequent siblings)
11 siblings, 1 reply; 17+ messages in thread
From: Masami Hiramatsu (Google) @ 2023-05-26 4:18 UTC (permalink / raw)
To: linux-trace-kernel
Cc: linux-kernel, Steven Rostedt, mhiramat, Florent Revest,
Mark Rutland, Will Deacon, Mathieu Desnoyers, Martin KaFai Lau,
bpf, Bagas Sanjaya
From: Masami Hiramatsu (Google) <mhiramat@kernel.org>
Add fprobe events for tracing function entry and exit instead of kprobe
events. With this change, we can continue to trace function entry/exit
even if the CONFIG_KPROBES_ON_FTRACE is not available. Since
CONFIG_KPROBES_ON_FTRACE requires the CONFIG_DYNAMIC_FTRACE_WITH_REGS,
it is not available if the architecture only supports
CONFIG_DYNAMIC_FTRACE_WITH_ARGS. And that means kprobe events can not
probe function entry/exit effectively on such architecture.
But this can be solved if the dynamic events supports fprobe events.
The fprobe event is a new dynamic events which is only for the function
(symbol) entry and exit. This event accepts non register fetch arguments
so that user can trace the function arguments and return values.
The fprobe events syntax is here;
f[:[GRP/][EVENT]] FUNCTION [FETCHARGS]
f[MAXACTIVE][:[GRP/][EVENT]] FUNCTION%return [FETCHARGS]
E.g.
# echo 'f vfs_read $arg1' >> dynamic_events
# echo 'f vfs_read%return $retval' >> dynamic_events
# cat dynamic_events
f:fprobes/vfs_read__entry vfs_read arg1=$arg1
f:fprobes/vfs_read__exit vfs_read%return arg1=$retval
# echo 1 > events/fprobes/enable
# head -n 20 trace | tail
# TASK-PID CPU# ||||| TIMESTAMP FUNCTION
# | | | ||||| | |
sh-142 [005] ...1. 448.386420: vfs_read__entry: (vfs_read+0x4/0x340) arg1=0xffff888007f7c540
sh-142 [005] ..... 448.386436: vfs_read__exit: (ksys_read+0x75/0x100 <- vfs_read) arg1=0x1
sh-142 [005] ...1. 448.386451: vfs_read__entry: (vfs_read+0x4/0x340) arg1=0xffff888007f7c540
sh-142 [005] ..... 448.386458: vfs_read__exit: (ksys_read+0x75/0x100 <- vfs_read) arg1=0x1
sh-142 [005] ...1. 448.386469: vfs_read__entry: (vfs_read+0x4/0x340) arg1=0xffff888007f7c540
sh-142 [005] ..... 448.386476: vfs_read__exit: (ksys_read+0x75/0x100 <- vfs_read) arg1=0x1
sh-142 [005] ...1. 448.602073: vfs_read__entry: (vfs_read+0x4/0x340) arg1=0xffff888007f7c540
sh-142 [005] ..... 448.602089: vfs_read__exit: (ksys_read+0x75/0x100 <- vfs_read) arg1=0x1
Reported-by: kernel test robot <lkp@intel.com>
Link: https://lore.kernel.org/all/202302011530.7vm4O8Ro-lkp@intel.com/
Signed-off-by: Masami Hiramatsu (Google) <mhiramat@kernel.org>
---
Changes in v13:
- Fix to use TPARG_FL_FPROBE so that it rejects tracing registers correctly.
- Do NOT set both TPARG_FL_FENTRY and TPARG_FL_RETURN.
Changes in v10:
- Clean up the trace_fprobe_match().
- Remove FETCH_OP_REG case from process_fetch_insn()
- Use process_common_fetch_insn() for common FETCH_OPs.
- Use WARN_ON_ONCE() instead of WARN_ON() and exit if it hits.
Changes in v6:
- Drop bpf-prog call because it is not supported on fprobe events.
Changes in v5:
- Sort header files alphabetically.
- Fix typo.
- Change default event name with '__<place>' suffix.
- Change common error name MAXACT_NO_KPROBE to BAD_MAXACT_TYPE
- Fix to show README entry correctly
---
include/linux/fprobe.h | 5
include/linux/trace_events.h | 3
kernel/trace/Kconfig | 14
kernel/trace/Makefile | 1
kernel/trace/fprobe.c | 11
kernel/trace/trace.c | 8
kernel/trace/trace.h | 11
kernel/trace/trace_fprobe.c | 1053 ++++++++++++++++++++
kernel/trace/trace_kprobe.c | 2
kernel/trace/trace_probe.c | 4
kernel/trace/trace_probe.h | 3
.../ftrace/test.d/kprobe/kprobe_syntax_errors.tc | 2
12 files changed, 1109 insertions(+), 8 deletions(-)
create mode 100644 kernel/trace/trace_fprobe.c
diff --git a/include/linux/fprobe.h b/include/linux/fprobe.h
index 134f0f59ffa8..3e03758151f4 100644
--- a/include/linux/fprobe.h
+++ b/include/linux/fprobe.h
@@ -66,6 +66,7 @@ int register_fprobe(struct fprobe *fp, const char *filter, const char *notfilter
int register_fprobe_ips(struct fprobe *fp, unsigned long *addrs, int num);
int register_fprobe_syms(struct fprobe *fp, const char **syms, int num);
int unregister_fprobe(struct fprobe *fp);
+bool fprobe_is_registered(struct fprobe *fp);
#else
static inline int register_fprobe(struct fprobe *fp, const char *filter, const char *notfilter)
{
@@ -83,6 +84,10 @@ static inline int unregister_fprobe(struct fprobe *fp)
{
return -EOPNOTSUPP;
}
+static inline bool fprobe_is_registered(struct fprobe *fp)
+{
+ return false;
+}
#endif
/**
diff --git a/include/linux/trace_events.h b/include/linux/trace_events.h
index 0e373222a6df..fcdc1f0bed23 100644
--- a/include/linux/trace_events.h
+++ b/include/linux/trace_events.h
@@ -318,6 +318,7 @@ enum {
TRACE_EVENT_FL_KPROBE_BIT,
TRACE_EVENT_FL_UPROBE_BIT,
TRACE_EVENT_FL_EPROBE_BIT,
+ TRACE_EVENT_FL_FPROBE_BIT,
TRACE_EVENT_FL_CUSTOM_BIT,
};
@@ -332,6 +333,7 @@ enum {
* KPROBE - Event is a kprobe
* UPROBE - Event is a uprobe
* EPROBE - Event is an event probe
+ * FPROBE - Event is an function probe
* CUSTOM - Event is a custom event (to be attached to an exsiting tracepoint)
* This is set when the custom event has not been attached
* to a tracepoint yet, then it is cleared when it is.
@@ -346,6 +348,7 @@ enum {
TRACE_EVENT_FL_KPROBE = (1 << TRACE_EVENT_FL_KPROBE_BIT),
TRACE_EVENT_FL_UPROBE = (1 << TRACE_EVENT_FL_UPROBE_BIT),
TRACE_EVENT_FL_EPROBE = (1 << TRACE_EVENT_FL_EPROBE_BIT),
+ TRACE_EVENT_FL_FPROBE = (1 << TRACE_EVENT_FL_FPROBE_BIT),
TRACE_EVENT_FL_CUSTOM = (1 << TRACE_EVENT_FL_CUSTOM_BIT),
};
diff --git a/kernel/trace/Kconfig b/kernel/trace/Kconfig
index 8cf97fa4a4b3..8e10a9453c96 100644
--- a/kernel/trace/Kconfig
+++ b/kernel/trace/Kconfig
@@ -650,6 +650,20 @@ config BLK_DEV_IO_TRACE
If unsure, say N.
+config FPROBE_EVENTS
+ depends on FPROBE
+ depends on HAVE_REGS_AND_STACK_ACCESS_API
+ bool "Enable fprobe-based dynamic events"
+ select TRACING
+ select PROBE_EVENTS
+ select DYNAMIC_EVENTS
+ default y
+ help
+ This allows user to add tracing events on the function entry and
+ exit via ftrace interface. The syntax is same as the kprobe events
+ and the kprobe events on function entry and exit will be
+ transparently converted to this fprobe events.
+
config KPROBE_EVENTS
depends on KPROBES
depends on HAVE_REGS_AND_STACK_ACCESS_API
diff --git a/kernel/trace/Makefile b/kernel/trace/Makefile
index c6651e16b557..64b61f67a403 100644
--- a/kernel/trace/Makefile
+++ b/kernel/trace/Makefile
@@ -104,6 +104,7 @@ obj-$(CONFIG_BOOTTIME_TRACING) += trace_boot.o
obj-$(CONFIG_FTRACE_RECORD_RECURSION) += trace_recursion_record.o
obj-$(CONFIG_FPROBE) += fprobe.o
obj-$(CONFIG_RETHOOK) += rethook.o
+obj-$(CONFIG_FPROBE_EVENTS) += trace_fprobe.o
obj-$(CONFIG_TRACEPOINT_BENCHMARK) += trace_benchmark.o
obj-$(CONFIG_RV) += rv/
diff --git a/kernel/trace/fprobe.c b/kernel/trace/fprobe.c
index 973bc664fcc1..1cfd4edcd579 100644
--- a/kernel/trace/fprobe.c
+++ b/kernel/trace/fprobe.c
@@ -305,6 +305,14 @@ int register_fprobe_syms(struct fprobe *fp, const char **syms, int num)
}
EXPORT_SYMBOL_GPL(register_fprobe_syms);
+bool fprobe_is_registered(struct fprobe *fp)
+{
+ if (!fp || (fp->ops.saved_func != fprobe_handler &&
+ fp->ops.saved_func != fprobe_kprobe_handler))
+ return false;
+ return true;
+}
+
/**
* unregister_fprobe() - Unregister fprobe from ftrace
* @fp: A fprobe data structure to be unregistered.
@@ -317,8 +325,7 @@ int unregister_fprobe(struct fprobe *fp)
{
int ret;
- if (!fp || (fp->ops.saved_func != fprobe_handler &&
- fp->ops.saved_func != fprobe_kprobe_handler))
+ if (!fprobe_is_registered(fp))
return -EINVAL;
/*
diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c
index 427da2341bf0..d595cef93122 100644
--- a/kernel/trace/trace.c
+++ b/kernel/trace/trace.c
@@ -5644,10 +5644,16 @@ static const char readme_msg[] =
" uprobe_events\t\t- Create/append/remove/show the userspace dynamic events\n"
"\t\t\t Write into this file to define/undefine new trace events.\n"
#endif
-#if defined(CONFIG_KPROBE_EVENTS) || defined(CONFIG_UPROBE_EVENTS)
+#if defined(CONFIG_KPROBE_EVENTS) || defined(CONFIG_UPROBE_EVENTS) || \
+ defined(CONFIG_FPROBE_EVENTS)
"\t accepts: event-definitions (one definition per line)\n"
+#if defined(CONFIG_KPROBE_EVENTS) || defined(CONFIG_UPROBE_EVENTS)
"\t Format: p[:[<group>/][<event>]] <place> [<args>]\n"
"\t r[maxactive][:[<group>/][<event>]] <place> [<args>]\n"
+#endif
+#ifdef CONFIG_FPROBE_EVENTS
+ "\t f[:[<group>/][<event>]] <func-name>[%return] [<args>]\n"
+#endif
#ifdef CONFIG_HIST_TRIGGERS
"\t s:[synthetic/]<event> <field> [<field>]\n"
#endif
diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h
index 79bdefe9261b..b5ab5479f9e3 100644
--- a/kernel/trace/trace.h
+++ b/kernel/trace/trace.h
@@ -148,6 +148,17 @@ struct kretprobe_trace_entry_head {
unsigned long ret_ip;
};
+struct fentry_trace_entry_head {
+ struct trace_entry ent;
+ unsigned long ip;
+};
+
+struct fexit_trace_entry_head {
+ struct trace_entry ent;
+ unsigned long func;
+ unsigned long ret_ip;
+};
+
#define TRACE_BUF_SIZE 1024
struct trace_array;
diff --git a/kernel/trace/trace_fprobe.c b/kernel/trace/trace_fprobe.c
new file mode 100644
index 000000000000..48dbbc72b7dd
--- /dev/null
+++ b/kernel/trace/trace_fprobe.c
@@ -0,0 +1,1053 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Fprobe-based tracing events
+ * Copyright (C) 2022 Google LLC.
+ */
+#define pr_fmt(fmt) "trace_fprobe: " fmt
+
+#include <linux/fprobe.h>
+#include <linux/module.h>
+#include <linux/rculist.h>
+#include <linux/security.h>
+#include <linux/uaccess.h>
+
+#include "trace_dynevent.h"
+#include "trace_probe.h"
+#include "trace_probe_kernel.h"
+#include "trace_probe_tmpl.h"
+
+#define FPROBE_EVENT_SYSTEM "fprobes"
+#define RETHOOK_MAXACTIVE_MAX 4096
+
+static int trace_fprobe_create(const char *raw_command);
+static int trace_fprobe_show(struct seq_file *m, struct dyn_event *ev);
+static int trace_fprobe_release(struct dyn_event *ev);
+static bool trace_fprobe_is_busy(struct dyn_event *ev);
+static bool trace_fprobe_match(const char *system, const char *event,
+ int argc, const char **argv, struct dyn_event *ev);
+
+static struct dyn_event_operations trace_fprobe_ops = {
+ .create = trace_fprobe_create,
+ .show = trace_fprobe_show,
+ .is_busy = trace_fprobe_is_busy,
+ .free = trace_fprobe_release,
+ .match = trace_fprobe_match,
+};
+
+/*
+ * Fprobe event core functions
+ */
+struct trace_fprobe {
+ struct dyn_event devent;
+ struct fprobe fp;
+ const char *symbol;
+ struct trace_probe tp;
+};
+
+static bool is_trace_fprobe(struct dyn_event *ev)
+{
+ return ev->ops == &trace_fprobe_ops;
+}
+
+static struct trace_fprobe *to_trace_fprobe(struct dyn_event *ev)
+{
+ return container_of(ev, struct trace_fprobe, devent);
+}
+
+/**
+ * for_each_trace_fprobe - iterate over the trace_fprobe list
+ * @pos: the struct trace_fprobe * for each entry
+ * @dpos: the struct dyn_event * to use as a loop cursor
+ */
+#define for_each_trace_fprobe(pos, dpos) \
+ for_each_dyn_event(dpos) \
+ if (is_trace_fprobe(dpos) && (pos = to_trace_fprobe(dpos)))
+
+static bool trace_fprobe_is_return(struct trace_fprobe *tf)
+{
+ return tf->fp.exit_handler != NULL;
+}
+
+static const char *trace_fprobe_symbol(struct trace_fprobe *tf)
+{
+ return tf->symbol ? tf->symbol : "unknown";
+}
+
+static bool trace_fprobe_is_busy(struct dyn_event *ev)
+{
+ struct trace_fprobe *tf = to_trace_fprobe(ev);
+
+ return trace_probe_is_enabled(&tf->tp);
+}
+
+static bool trace_fprobe_match_command_head(struct trace_fprobe *tf,
+ int argc, const char **argv)
+{
+ char buf[MAX_ARGSTR_LEN + 1];
+
+ if (!argc)
+ return true;
+
+ snprintf(buf, sizeof(buf), "%s", trace_fprobe_symbol(tf));
+ if (strcmp(buf, argv[0]))
+ return false;
+ argc--; argv++;
+
+ return trace_probe_match_command_args(&tf->tp, argc, argv);
+}
+
+static bool trace_fprobe_match(const char *system, const char *event,
+ int argc, const char **argv, struct dyn_event *ev)
+{
+ struct trace_fprobe *tf = to_trace_fprobe(ev);
+
+ if (event[0] != '\0' && strcmp(trace_probe_name(&tf->tp), event))
+ return false;
+
+ if (system && strcmp(trace_probe_group_name(&tf->tp), system))
+ return false;
+
+ return trace_fprobe_match_command_head(tf, argc, argv);
+}
+
+static bool trace_fprobe_is_registered(struct trace_fprobe *tf)
+{
+ return fprobe_is_registered(&tf->fp);
+}
+
+/*
+ * Note that we don't verify the fetch_insn code, since it does not come
+ * from user space.
+ */
+static int
+process_fetch_insn(struct fetch_insn *code, void *rec, void *dest,
+ void *base)
+{
+ struct pt_regs *regs = rec;
+ unsigned long val;
+ int ret;
+
+retry:
+ /* 1st stage: get value from context */
+ switch (code->op) {
+ case FETCH_OP_STACK:
+ val = regs_get_kernel_stack_nth(regs, code->param);
+ break;
+ case FETCH_OP_STACKP:
+ val = kernel_stack_pointer(regs);
+ break;
+ case FETCH_OP_RETVAL:
+ val = regs_return_value(regs);
+ break;
+#ifdef CONFIG_HAVE_FUNCTION_ARG_ACCESS_API
+ case FETCH_OP_ARG:
+ val = regs_get_kernel_argument(regs, code->param);
+ break;
+#endif
+ case FETCH_NOP_SYMBOL: /* Ignore a place holder */
+ code++;
+ goto retry;
+ default:
+ ret = process_common_fetch_insn(code, &val);
+ if (ret < 0)
+ return ret;
+ }
+ code++;
+
+ return process_fetch_insn_bottom(code, val, dest, base);
+}
+NOKPROBE_SYMBOL(process_fetch_insn)
+
+/* function entry handler */
+static nokprobe_inline void
+__fentry_trace_func(struct trace_fprobe *tf, unsigned long entry_ip,
+ struct pt_regs *regs,
+ struct trace_event_file *trace_file)
+{
+ struct fentry_trace_entry_head *entry;
+ struct trace_event_call *call = trace_probe_event_call(&tf->tp);
+ struct trace_event_buffer fbuffer;
+ int dsize;
+
+ if (WARN_ON_ONCE(call != trace_file->event_call))
+ return;
+
+ if (trace_trigger_soft_disabled(trace_file))
+ return;
+
+ dsize = __get_data_size(&tf->tp, regs);
+
+ entry = trace_event_buffer_reserve(&fbuffer, trace_file,
+ sizeof(*entry) + tf->tp.size + dsize);
+ if (!entry)
+ return;
+
+ fbuffer.regs = regs;
+ entry = fbuffer.entry = ring_buffer_event_data(fbuffer.event);
+ entry->ip = entry_ip;
+ store_trace_args(&entry[1], &tf->tp, regs, sizeof(*entry), dsize);
+
+ trace_event_buffer_commit(&fbuffer);
+}
+
+static void
+fentry_trace_func(struct trace_fprobe *tf, unsigned long entry_ip,
+ struct pt_regs *regs)
+{
+ struct event_file_link *link;
+
+ trace_probe_for_each_link_rcu(link, &tf->tp)
+ __fentry_trace_func(tf, entry_ip, regs, link->file);
+}
+NOKPROBE_SYMBOL(fentry_trace_func);
+
+/* Kretprobe handler */
+static nokprobe_inline void
+__fexit_trace_func(struct trace_fprobe *tf, unsigned long entry_ip,
+ unsigned long ret_ip, struct pt_regs *regs,
+ struct trace_event_file *trace_file)
+{
+ struct fexit_trace_entry_head *entry;
+ struct trace_event_buffer fbuffer;
+ struct trace_event_call *call = trace_probe_event_call(&tf->tp);
+ int dsize;
+
+ if (WARN_ON_ONCE(call != trace_file->event_call))
+ return;
+
+ if (trace_trigger_soft_disabled(trace_file))
+ return;
+
+ dsize = __get_data_size(&tf->tp, regs);
+
+ entry = trace_event_buffer_reserve(&fbuffer, trace_file,
+ sizeof(*entry) + tf->tp.size + dsize);
+ if (!entry)
+ return;
+
+ fbuffer.regs = regs;
+ entry = fbuffer.entry = ring_buffer_event_data(fbuffer.event);
+ entry->func = entry_ip;
+ entry->ret_ip = ret_ip;
+ store_trace_args(&entry[1], &tf->tp, regs, sizeof(*entry), dsize);
+
+ trace_event_buffer_commit(&fbuffer);
+}
+
+static void
+fexit_trace_func(struct trace_fprobe *tf, unsigned long entry_ip,
+ unsigned long ret_ip, struct pt_regs *regs)
+{
+ struct event_file_link *link;
+
+ trace_probe_for_each_link_rcu(link, &tf->tp)
+ __fexit_trace_func(tf, entry_ip, ret_ip, regs, link->file);
+}
+NOKPROBE_SYMBOL(fexit_trace_func);
+
+#ifdef CONFIG_PERF_EVENTS
+
+static int fentry_perf_func(struct trace_fprobe *tf, unsigned long entry_ip,
+ struct pt_regs *regs)
+{
+ struct trace_event_call *call = trace_probe_event_call(&tf->tp);
+ struct fentry_trace_entry_head *entry;
+ struct hlist_head *head;
+ int size, __size, dsize;
+ int rctx;
+
+ head = this_cpu_ptr(call->perf_events);
+ if (hlist_empty(head))
+ return 0;
+
+ dsize = __get_data_size(&tf->tp, regs);
+ __size = sizeof(*entry) + tf->tp.size + dsize;
+ size = ALIGN(__size + sizeof(u32), sizeof(u64));
+ size -= sizeof(u32);
+
+ entry = perf_trace_buf_alloc(size, NULL, &rctx);
+ if (!entry)
+ return 0;
+
+ entry->ip = entry_ip;
+ memset(&entry[1], 0, dsize);
+ store_trace_args(&entry[1], &tf->tp, regs, sizeof(*entry), dsize);
+ perf_trace_buf_submit(entry, size, rctx, call->event.type, 1, regs,
+ head, NULL);
+ return 0;
+}
+NOKPROBE_SYMBOL(fentry_perf_func);
+
+static void
+fexit_perf_func(struct trace_fprobe *tf, unsigned long entry_ip,
+ unsigned long ret_ip, struct pt_regs *regs)
+{
+ struct trace_event_call *call = trace_probe_event_call(&tf->tp);
+ struct fexit_trace_entry_head *entry;
+ struct hlist_head *head;
+ int size, __size, dsize;
+ int rctx;
+
+ head = this_cpu_ptr(call->perf_events);
+ if (hlist_empty(head))
+ return;
+
+ dsize = __get_data_size(&tf->tp, regs);
+ __size = sizeof(*entry) + tf->tp.size + dsize;
+ size = ALIGN(__size + sizeof(u32), sizeof(u64));
+ size -= sizeof(u32);
+
+ entry = perf_trace_buf_alloc(size, NULL, &rctx);
+ if (!entry)
+ return;
+
+ entry->func = entry_ip;
+ entry->ret_ip = ret_ip;
+ store_trace_args(&entry[1], &tf->tp, regs, sizeof(*entry), dsize);
+ perf_trace_buf_submit(entry, size, rctx, call->event.type, 1, regs,
+ head, NULL);
+}
+NOKPROBE_SYMBOL(fexit_perf_func);
+#endif /* CONFIG_PERF_EVENTS */
+
+static int fentry_dispatcher(struct fprobe *fp, unsigned long entry_ip,
+ unsigned long ret_ip, struct pt_regs *regs,
+ void *entry_data)
+{
+ struct trace_fprobe *tf = container_of(fp, struct trace_fprobe, fp);
+ int ret = 0;
+
+ if (trace_probe_test_flag(&tf->tp, TP_FLAG_TRACE))
+ fentry_trace_func(tf, entry_ip, regs);
+#ifdef CONFIG_PERF_EVENTS
+ if (trace_probe_test_flag(&tf->tp, TP_FLAG_PROFILE))
+ ret = fentry_perf_func(tf, entry_ip, regs);
+#endif
+ return ret;
+}
+NOKPROBE_SYMBOL(fentry_dispatcher);
+
+static void fexit_dispatcher(struct fprobe *fp, unsigned long entry_ip,
+ unsigned long ret_ip, struct pt_regs *regs,
+ void *entry_data)
+{
+ struct trace_fprobe *tf = container_of(fp, struct trace_fprobe, fp);
+
+ if (trace_probe_test_flag(&tf->tp, TP_FLAG_TRACE))
+ fexit_trace_func(tf, entry_ip, ret_ip, regs);
+#ifdef CONFIG_PERF_EVENTS
+ if (trace_probe_test_flag(&tf->tp, TP_FLAG_PROFILE))
+ fexit_perf_func(tf, entry_ip, ret_ip, regs);
+#endif
+}
+NOKPROBE_SYMBOL(fexit_dispatcher);
+
+static void free_trace_fprobe(struct trace_fprobe *tf)
+{
+ if (tf) {
+ trace_probe_cleanup(&tf->tp);
+ kfree(tf->symbol);
+ kfree(tf);
+ }
+}
+
+/*
+ * Allocate new trace_probe and initialize it (including fprobe).
+ */
+static struct trace_fprobe *alloc_trace_fprobe(const char *group,
+ const char *event,
+ const char *symbol,
+ int maxactive,
+ int nargs, bool is_return)
+{
+ struct trace_fprobe *tf;
+ int ret = -ENOMEM;
+
+ tf = kzalloc(struct_size(tf, tp.args, nargs), GFP_KERNEL);
+ if (!tf)
+ return ERR_PTR(ret);
+
+ tf->symbol = kstrdup(symbol, GFP_KERNEL);
+ if (!tf->symbol)
+ goto error;
+
+ if (is_return)
+ tf->fp.exit_handler = fexit_dispatcher;
+ else
+ tf->fp.entry_handler = fentry_dispatcher;
+
+ tf->fp.nr_maxactive = maxactive;
+
+ ret = trace_probe_init(&tf->tp, event, group, false);
+ if (ret < 0)
+ goto error;
+
+ dyn_event_init(&tf->devent, &trace_fprobe_ops);
+ return tf;
+error:
+ free_trace_fprobe(tf);
+ return ERR_PTR(ret);
+}
+
+static struct trace_fprobe *find_trace_fprobe(const char *event,
+ const char *group)
+{
+ struct dyn_event *pos;
+ struct trace_fprobe *tf;
+
+ for_each_trace_fprobe(tf, pos)
+ if (strcmp(trace_probe_name(&tf->tp), event) == 0 &&
+ strcmp(trace_probe_group_name(&tf->tp), group) == 0)
+ return tf;
+ return NULL;
+}
+
+static inline int __enable_trace_fprobe(struct trace_fprobe *tf)
+{
+ if (trace_fprobe_is_registered(tf))
+ enable_fprobe(&tf->fp);
+
+ return 0;
+}
+
+static void __disable_trace_fprobe(struct trace_probe *tp)
+{
+ struct trace_fprobe *tf;
+
+ list_for_each_entry(tf, trace_probe_probe_list(tp), tp.list) {
+ if (!trace_fprobe_is_registered(tf))
+ continue;
+ disable_fprobe(&tf->fp);
+ }
+}
+
+/*
+ * Enable trace_probe
+ * if the file is NULL, enable "perf" handler, or enable "trace" handler.
+ */
+static int enable_trace_fprobe(struct trace_event_call *call,
+ struct trace_event_file *file)
+{
+ struct trace_probe *tp;
+ struct trace_fprobe *tf;
+ bool enabled;
+ int ret = 0;
+
+ tp = trace_probe_primary_from_call(call);
+ if (WARN_ON_ONCE(!tp))
+ return -ENODEV;
+ enabled = trace_probe_is_enabled(tp);
+
+ /* This also changes "enabled" state */
+ if (file) {
+ ret = trace_probe_add_file(tp, file);
+ if (ret)
+ return ret;
+ } else
+ trace_probe_set_flag(tp, TP_FLAG_PROFILE);
+
+ if (!enabled) {
+ list_for_each_entry(tf, trace_probe_probe_list(tp), tp.list) {
+ /* TODO: check the fprobe is gone */
+ __enable_trace_fprobe(tf);
+ }
+ }
+
+ return 0;
+}
+
+/*
+ * Disable trace_probe
+ * if the file is NULL, disable "perf" handler, or disable "trace" handler.
+ */
+static int disable_trace_fprobe(struct trace_event_call *call,
+ struct trace_event_file *file)
+{
+ struct trace_probe *tp;
+
+ tp = trace_probe_primary_from_call(call);
+ if (WARN_ON_ONCE(!tp))
+ return -ENODEV;
+
+ if (file) {
+ if (!trace_probe_get_file_link(tp, file))
+ return -ENOENT;
+ if (!trace_probe_has_single_file(tp))
+ goto out;
+ trace_probe_clear_flag(tp, TP_FLAG_TRACE);
+ } else
+ trace_probe_clear_flag(tp, TP_FLAG_PROFILE);
+
+ if (!trace_probe_is_enabled(tp))
+ __disable_trace_fprobe(tp);
+
+ out:
+ if (file)
+ /*
+ * Synchronization is done in below function. For perf event,
+ * file == NULL and perf_trace_event_unreg() calls
+ * tracepoint_synchronize_unregister() to ensure synchronize
+ * event. We don't need to care about it.
+ */
+ trace_probe_remove_file(tp, file);
+
+ return 0;
+}
+
+/* Event entry printers */
+static enum print_line_t
+print_fentry_event(struct trace_iterator *iter, int flags,
+ struct trace_event *event)
+{
+ struct fentry_trace_entry_head *field;
+ struct trace_seq *s = &iter->seq;
+ struct trace_probe *tp;
+
+ field = (struct fentry_trace_entry_head *)iter->ent;
+ tp = trace_probe_primary_from_call(
+ container_of(event, struct trace_event_call, event));
+ if (WARN_ON_ONCE(!tp))
+ goto out;
+
+ trace_seq_printf(s, "%s: (", trace_probe_name(tp));
+
+ if (!seq_print_ip_sym(s, field->ip, flags | TRACE_ITER_SYM_OFFSET))
+ goto out;
+
+ trace_seq_putc(s, ')');
+
+ if (trace_probe_print_args(s, tp->args, tp->nr_args,
+ (u8 *)&field[1], field) < 0)
+ goto out;
+
+ trace_seq_putc(s, '\n');
+ out:
+ return trace_handle_return(s);
+}
+
+static enum print_line_t
+print_fexit_event(struct trace_iterator *iter, int flags,
+ struct trace_event *event)
+{
+ struct fexit_trace_entry_head *field;
+ struct trace_seq *s = &iter->seq;
+ struct trace_probe *tp;
+
+ field = (struct fexit_trace_entry_head *)iter->ent;
+ tp = trace_probe_primary_from_call(
+ container_of(event, struct trace_event_call, event));
+ if (WARN_ON_ONCE(!tp))
+ goto out;
+
+ trace_seq_printf(s, "%s: (", trace_probe_name(tp));
+
+ if (!seq_print_ip_sym(s, field->ret_ip, flags | TRACE_ITER_SYM_OFFSET))
+ goto out;
+
+ trace_seq_puts(s, " <- ");
+
+ if (!seq_print_ip_sym(s, field->func, flags & ~TRACE_ITER_SYM_OFFSET))
+ goto out;
+
+ trace_seq_putc(s, ')');
+
+ if (trace_probe_print_args(s, tp->args, tp->nr_args,
+ (u8 *)&field[1], field) < 0)
+ goto out;
+
+ trace_seq_putc(s, '\n');
+
+ out:
+ return trace_handle_return(s);
+}
+
+static int fentry_event_define_fields(struct trace_event_call *event_call)
+{
+ int ret;
+ struct fentry_trace_entry_head field;
+ struct trace_probe *tp;
+
+ tp = trace_probe_primary_from_call(event_call);
+ if (WARN_ON_ONCE(!tp))
+ return -ENOENT;
+
+ DEFINE_FIELD(unsigned long, ip, FIELD_STRING_IP, 0);
+
+ return traceprobe_define_arg_fields(event_call, sizeof(field), tp);
+}
+
+static int fexit_event_define_fields(struct trace_event_call *event_call)
+{
+ int ret;
+ struct fexit_trace_entry_head field;
+ struct trace_probe *tp;
+
+ tp = trace_probe_primary_from_call(event_call);
+ if (WARN_ON_ONCE(!tp))
+ return -ENOENT;
+
+ DEFINE_FIELD(unsigned long, func, FIELD_STRING_FUNC, 0);
+ DEFINE_FIELD(unsigned long, ret_ip, FIELD_STRING_RETIP, 0);
+
+ return traceprobe_define_arg_fields(event_call, sizeof(field), tp);
+}
+
+static struct trace_event_functions fentry_funcs = {
+ .trace = print_fentry_event
+};
+
+static struct trace_event_functions fexit_funcs = {
+ .trace = print_fexit_event
+};
+
+static struct trace_event_fields fentry_fields_array[] = {
+ { .type = TRACE_FUNCTION_TYPE,
+ .define_fields = fentry_event_define_fields },
+ {}
+};
+
+static struct trace_event_fields fexit_fields_array[] = {
+ { .type = TRACE_FUNCTION_TYPE,
+ .define_fields = fexit_event_define_fields },
+ {}
+};
+
+static int fprobe_register(struct trace_event_call *event,
+ enum trace_reg type, void *data);
+
+static inline void init_trace_event_call(struct trace_fprobe *tf)
+{
+ struct trace_event_call *call = trace_probe_event_call(&tf->tp);
+
+ if (trace_fprobe_is_return(tf)) {
+ call->event.funcs = &fexit_funcs;
+ call->class->fields_array = fexit_fields_array;
+ } else {
+ call->event.funcs = &fentry_funcs;
+ call->class->fields_array = fentry_fields_array;
+ }
+
+ call->flags = TRACE_EVENT_FL_FPROBE;
+ call->class->reg = fprobe_register;
+}
+
+static int register_fprobe_event(struct trace_fprobe *tf)
+{
+ init_trace_event_call(tf);
+
+ return trace_probe_register_event_call(&tf->tp);
+}
+
+static int unregister_fprobe_event(struct trace_fprobe *tf)
+{
+ return trace_probe_unregister_event_call(&tf->tp);
+}
+
+/* Internal register function - just handle fprobe and flags */
+static int __register_trace_fprobe(struct trace_fprobe *tf)
+{
+ int i, ret;
+
+ /* Should we need new LOCKDOWN flag for fprobe? */
+ ret = security_locked_down(LOCKDOWN_KPROBES);
+ if (ret)
+ return ret;
+
+ if (trace_fprobe_is_registered(tf))
+ return -EINVAL;
+
+ for (i = 0; i < tf->tp.nr_args; i++) {
+ ret = traceprobe_update_arg(&tf->tp.args[i]);
+ if (ret)
+ return ret;
+ }
+
+ /* Set/clear disabled flag according to tp->flag */
+ if (trace_probe_is_enabled(&tf->tp))
+ tf->fp.flags &= ~FPROBE_FL_DISABLED;
+ else
+ tf->fp.flags |= FPROBE_FL_DISABLED;
+
+ /* TODO: handle filter, nofilter or symbol list */
+ return register_fprobe(&tf->fp, tf->symbol, NULL);
+}
+
+/* Internal unregister function - just handle fprobe and flags */
+static void __unregister_trace_fprobe(struct trace_fprobe *tf)
+{
+ if (trace_fprobe_is_registered(tf)) {
+ unregister_fprobe(&tf->fp);
+ memset(&tf->fp, 0, sizeof(tf->fp));
+ }
+}
+
+/* TODO: make this trace_*probe common function */
+/* Unregister a trace_probe and probe_event */
+static int unregister_trace_fprobe(struct trace_fprobe *tf)
+{
+ /* If other probes are on the event, just unregister fprobe */
+ if (trace_probe_has_sibling(&tf->tp))
+ goto unreg;
+
+ /* Enabled event can not be unregistered */
+ if (trace_probe_is_enabled(&tf->tp))
+ return -EBUSY;
+
+ /* If there's a reference to the dynamic event */
+ if (trace_event_dyn_busy(trace_probe_event_call(&tf->tp)))
+ return -EBUSY;
+
+ /* Will fail if probe is being used by ftrace or perf */
+ if (unregister_fprobe_event(tf))
+ return -EBUSY;
+
+unreg:
+ __unregister_trace_fprobe(tf);
+ dyn_event_remove(&tf->devent);
+ trace_probe_unlink(&tf->tp);
+
+ return 0;
+}
+
+static bool trace_fprobe_has_same_fprobe(struct trace_fprobe *orig,
+ struct trace_fprobe *comp)
+{
+ struct trace_probe_event *tpe = orig->tp.event;
+ int i;
+
+ list_for_each_entry(orig, &tpe->probes, tp.list) {
+ if (strcmp(trace_fprobe_symbol(orig),
+ trace_fprobe_symbol(comp)))
+ continue;
+
+ /*
+ * trace_probe_compare_arg_type() ensured that nr_args and
+ * each argument name and type are same. Let's compare comm.
+ */
+ for (i = 0; i < orig->tp.nr_args; i++) {
+ if (strcmp(orig->tp.args[i].comm,
+ comp->tp.args[i].comm))
+ break;
+ }
+
+ if (i == orig->tp.nr_args)
+ return true;
+ }
+
+ return false;
+}
+
+static int append_trace_fprobe(struct trace_fprobe *tf, struct trace_fprobe *to)
+{
+ int ret;
+
+ if (trace_fprobe_is_return(tf) != trace_fprobe_is_return(to)) {
+ trace_probe_log_set_index(0);
+ trace_probe_log_err(0, DIFF_PROBE_TYPE);
+ return -EEXIST;
+ }
+ ret = trace_probe_compare_arg_type(&tf->tp, &to->tp);
+ if (ret) {
+ /* Note that argument starts index = 2 */
+ trace_probe_log_set_index(ret + 1);
+ trace_probe_log_err(0, DIFF_ARG_TYPE);
+ return -EEXIST;
+ }
+ if (trace_fprobe_has_same_fprobe(to, tf)) {
+ trace_probe_log_set_index(0);
+ trace_probe_log_err(0, SAME_PROBE);
+ return -EEXIST;
+ }
+
+ /* Append to existing event */
+ ret = trace_probe_append(&tf->tp, &to->tp);
+ if (ret)
+ return ret;
+
+ ret = __register_trace_fprobe(tf);
+ if (ret)
+ trace_probe_unlink(&tf->tp);
+ else
+ dyn_event_add(&tf->devent, trace_probe_event_call(&tf->tp));
+
+ return ret;
+}
+
+/* Register a trace_probe and probe_event */
+static int register_trace_fprobe(struct trace_fprobe *tf)
+{
+ struct trace_fprobe *old_tf;
+ int ret;
+
+ mutex_lock(&event_mutex);
+
+ old_tf = find_trace_fprobe(trace_probe_name(&tf->tp),
+ trace_probe_group_name(&tf->tp));
+ if (old_tf) {
+ ret = append_trace_fprobe(tf, old_tf);
+ goto end;
+ }
+
+ /* Register new event */
+ ret = register_fprobe_event(tf);
+ if (ret) {
+ if (ret == -EEXIST) {
+ trace_probe_log_set_index(0);
+ trace_probe_log_err(0, EVENT_EXIST);
+ } else
+ pr_warn("Failed to register probe event(%d)\n", ret);
+ goto end;
+ }
+
+ /* Register fprobe */
+ ret = __register_trace_fprobe(tf);
+ if (ret < 0)
+ unregister_fprobe_event(tf);
+ else
+ dyn_event_add(&tf->devent, trace_probe_event_call(&tf->tp));
+
+end:
+ mutex_unlock(&event_mutex);
+ return ret;
+}
+
+static int __trace_fprobe_create(int argc, const char *argv[])
+{
+ /*
+ * Argument syntax:
+ * - Add fentry probe:
+ * f[:[GRP/][EVENT]] [MOD:]KSYM [FETCHARGS]
+ * - Add fexit probe:
+ * f[N][:[GRP/][EVENT]] [MOD:]KSYM%return [FETCHARGS]
+ *
+ * Fetch args:
+ * $retval : fetch return value
+ * $stack : fetch stack address
+ * $stackN : fetch Nth entry of stack (N:0-)
+ * $argN : fetch Nth argument (N:1-)
+ * $comm : fetch current task comm
+ * @ADDR : fetch memory at ADDR (ADDR should be in kernel)
+ * @SYM[+|-offs] : fetch memory at SYM +|- offs (SYM is a data symbol)
+ * Dereferencing memory fetch:
+ * +|-offs(ARG) : fetch memory at ARG +|- offs address.
+ * Alias name of args:
+ * NAME=FETCHARG : set NAME as alias of FETCHARG.
+ * Type of args:
+ * FETCHARG:TYPE : use TYPE instead of unsigned long.
+ */
+ struct trace_fprobe *tf = NULL;
+ int i, len, ret = 0;
+ bool is_return = false;
+ char *symbol = NULL, *tmp = NULL;
+ const char *event = NULL, *group = FPROBE_EVENT_SYSTEM;
+ int maxactive = 0;
+ char buf[MAX_EVENT_NAME_LEN];
+ char gbuf[MAX_EVENT_NAME_LEN];
+ unsigned int flags = TPARG_FL_KERNEL | TPARG_FL_FPROBE;
+
+ if (argv[0][0] != 'f' || argc < 2)
+ return -ECANCELED;
+
+ trace_probe_log_init("trace_fprobe", argc, argv);
+
+ event = strchr(&argv[0][1], ':');
+ if (event)
+ event++;
+
+ if (isdigit(argv[0][1])) {
+ if (event)
+ len = event - &argv[0][1] - 1;
+ else
+ len = strlen(&argv[0][1]);
+ if (len > MAX_EVENT_NAME_LEN - 1) {
+ trace_probe_log_err(1, BAD_MAXACT);
+ goto parse_error;
+ }
+ memcpy(buf, &argv[0][1], len);
+ buf[len] = '\0';
+ ret = kstrtouint(buf, 0, &maxactive);
+ if (ret || !maxactive) {
+ trace_probe_log_err(1, BAD_MAXACT);
+ goto parse_error;
+ }
+ /* fprobe rethook instances are iterated over via a list. The
+ * maximum should stay reasonable.
+ */
+ if (maxactive > RETHOOK_MAXACTIVE_MAX) {
+ trace_probe_log_err(1, MAXACT_TOO_BIG);
+ goto parse_error;
+ }
+ }
+
+ trace_probe_log_set_index(1);
+
+ /* a symbol specified */
+ symbol = kstrdup(argv[1], GFP_KERNEL);
+ if (!symbol)
+ return -ENOMEM;
+
+ tmp = strchr(symbol, '%');
+ if (tmp) {
+ if (!strcmp(tmp, "%return")) {
+ *tmp = '\0';
+ is_return = true;
+ } else {
+ trace_probe_log_err(tmp - symbol, BAD_ADDR_SUFFIX);
+ goto parse_error;
+ }
+ }
+ if (!is_return && maxactive) {
+ trace_probe_log_set_index(0);
+ trace_probe_log_err(1, BAD_MAXACT_TYPE);
+ goto parse_error;
+ }
+
+ if (is_return)
+ flags |= TPARG_FL_RETURN;
+ else
+ flags |= TPARG_FL_FENTRY;
+
+ trace_probe_log_set_index(0);
+ if (event) {
+ ret = traceprobe_parse_event_name(&event, &group, gbuf,
+ event - argv[0]);
+ if (ret)
+ goto parse_error;
+ }
+
+ if (!event) {
+ /* Make a new event name */
+ snprintf(buf, MAX_EVENT_NAME_LEN, "%s__%s", symbol,
+ is_return ? "exit" : "entry");
+ sanitize_event_name(buf);
+ event = buf;
+ }
+
+ /* setup a probe */
+ tf = alloc_trace_fprobe(group, event, symbol, maxactive,
+ argc - 2, is_return);
+ if (IS_ERR(tf)) {
+ ret = PTR_ERR(tf);
+ /* This must return -ENOMEM, else there is a bug */
+ WARN_ON_ONCE(ret != -ENOMEM);
+ goto out; /* We know tf is not allocated */
+ }
+ argc -= 2; argv += 2;
+
+ /* parse arguments */
+ for (i = 0; i < argc && i < MAX_TRACE_ARGS; i++) {
+ trace_probe_log_set_index(i + 2);
+ ret = traceprobe_parse_probe_arg(&tf->tp, i, argv[i], flags);
+ if (ret)
+ goto error; /* This can be -ENOMEM */
+ }
+
+ ret = traceprobe_set_print_fmt(&tf->tp,
+ is_return ? PROBE_PRINT_RETURN : PROBE_PRINT_NORMAL);
+ if (ret < 0)
+ goto error;
+
+ ret = register_trace_fprobe(tf);
+ if (ret) {
+ trace_probe_log_set_index(1);
+ if (ret == -EILSEQ)
+ trace_probe_log_err(0, BAD_INSN_BNDRY);
+ else if (ret == -ENOENT)
+ trace_probe_log_err(0, BAD_PROBE_ADDR);
+ else if (ret != -ENOMEM && ret != -EEXIST)
+ trace_probe_log_err(0, FAIL_REG_PROBE);
+ goto error;
+ }
+
+out:
+ trace_probe_log_clear();
+ kfree(symbol);
+ return ret;
+
+parse_error:
+ ret = -EINVAL;
+error:
+ free_trace_fprobe(tf);
+ goto out;
+}
+
+static int trace_fprobe_create(const char *raw_command)
+{
+ return trace_probe_create(raw_command, __trace_fprobe_create);
+}
+
+static int trace_fprobe_release(struct dyn_event *ev)
+{
+ struct trace_fprobe *tf = to_trace_fprobe(ev);
+ int ret = unregister_trace_fprobe(tf);
+
+ if (!ret)
+ free_trace_fprobe(tf);
+ return ret;
+}
+
+static int trace_fprobe_show(struct seq_file *m, struct dyn_event *ev)
+{
+ struct trace_fprobe *tf = to_trace_fprobe(ev);
+ int i;
+
+ seq_putc(m, 'f');
+ if (trace_fprobe_is_return(tf) && tf->fp.nr_maxactive)
+ seq_printf(m, "%d", tf->fp.nr_maxactive);
+ seq_printf(m, ":%s/%s", trace_probe_group_name(&tf->tp),
+ trace_probe_name(&tf->tp));
+
+ seq_printf(m, " %s%s", trace_fprobe_symbol(tf),
+ trace_fprobe_is_return(tf) ? "%return" : "");
+
+ for (i = 0; i < tf->tp.nr_args; i++)
+ seq_printf(m, " %s=%s", tf->tp.args[i].name, tf->tp.args[i].comm);
+ seq_putc(m, '\n');
+
+ return 0;
+}
+
+/*
+ * called by perf_trace_init() or __ftrace_set_clr_event() under event_mutex.
+ */
+static int fprobe_register(struct trace_event_call *event,
+ enum trace_reg type, void *data)
+{
+ struct trace_event_file *file = data;
+
+ switch (type) {
+ case TRACE_REG_REGISTER:
+ return enable_trace_fprobe(event, file);
+ case TRACE_REG_UNREGISTER:
+ return disable_trace_fprobe(event, file);
+
+#ifdef CONFIG_PERF_EVENTS
+ case TRACE_REG_PERF_REGISTER:
+ return enable_trace_fprobe(event, NULL);
+ case TRACE_REG_PERF_UNREGISTER:
+ return disable_trace_fprobe(event, NULL);
+ case TRACE_REG_PERF_OPEN:
+ case TRACE_REG_PERF_CLOSE:
+ case TRACE_REG_PERF_ADD:
+ case TRACE_REG_PERF_DEL:
+ return 0;
+#endif
+ }
+ return 0;
+}
+
+/*
+ * Register dynevent at core_initcall. This allows kernel to setup fprobe
+ * events in postcore_initcall without tracefs.
+ */
+static __init int init_fprobe_trace_early(void)
+{
+ int ret;
+
+ ret = dyn_event_register(&trace_fprobe_ops);
+ if (ret)
+ return ret;
+
+ return 0;
+}
+core_initcall(init_fprobe_trace_early);
diff --git a/kernel/trace/trace_kprobe.c b/kernel/trace/trace_kprobe.c
index 867ffb7ee31d..b7a4409674b3 100644
--- a/kernel/trace/trace_kprobe.c
+++ b/kernel/trace/trace_kprobe.c
@@ -764,7 +764,7 @@ static int __trace_kprobe_create(int argc, const char *argv[])
if (isdigit(argv[0][1])) {
if (!is_return) {
- trace_probe_log_err(1, MAXACT_NO_KPROBE);
+ trace_probe_log_err(1, BAD_MAXACT_TYPE);
goto parse_error;
}
if (event)
diff --git a/kernel/trace/trace_probe.c b/kernel/trace/trace_probe.c
index 20d0c4a97633..8646b097d56c 100644
--- a/kernel/trace/trace_probe.c
+++ b/kernel/trace/trace_probe.c
@@ -393,8 +393,8 @@ parse_probe_arg(char *arg, const struct fetch_type *type,
break;
case '%': /* named register */
- if (flags & TPARG_FL_TPOINT) {
- /* eprobes do not handle registers */
+ if (flags & (TPARG_FL_TPOINT | TPARG_FL_FPROBE)) {
+ /* eprobe and fprobe do not handle registers */
trace_probe_log_err(offs, BAD_VAR);
break;
}
diff --git a/kernel/trace/trace_probe.h b/kernel/trace/trace_probe.h
index 752a08fcf0be..cf146a7b7f81 100644
--- a/kernel/trace/trace_probe.h
+++ b/kernel/trace/trace_probe.h
@@ -367,6 +367,7 @@ int trace_probe_print_args(struct trace_seq *s, struct probe_arg *args, int nr_a
#define TPARG_FL_FENTRY BIT(2)
#define TPARG_FL_TPOINT BIT(3)
#define TPARG_FL_USER BIT(4)
+#define TPARG_FL_FPROBE BIT(5)
#define TPARG_FL_MASK GENMASK(4, 0)
extern int traceprobe_parse_probe_arg(struct trace_probe *tp, int i,
@@ -409,7 +410,7 @@ extern int traceprobe_define_arg_fields(struct trace_event_call *event_call,
C(REFCNT_OPEN_BRACE, "Reference counter brace is not closed"), \
C(BAD_REFCNT_SUFFIX, "Reference counter has wrong suffix"), \
C(BAD_UPROBE_OFFS, "Invalid uprobe offset"), \
- C(MAXACT_NO_KPROBE, "Maxactive is not for kprobe"), \
+ C(BAD_MAXACT_TYPE, "Maxactive is only for function exit"), \
C(BAD_MAXACT, "Invalid maxactive number"), \
C(MAXACT_TOO_BIG, "Maxactive is too big"), \
C(BAD_PROBE_ADDR, "Invalid probed address or symbol"), \
diff --git a/tools/testing/selftests/ftrace/test.d/kprobe/kprobe_syntax_errors.tc b/tools/testing/selftests/ftrace/test.d/kprobe/kprobe_syntax_errors.tc
index 9e85d3019ff0..97c08867490a 100644
--- a/tools/testing/selftests/ftrace/test.d/kprobe/kprobe_syntax_errors.tc
+++ b/tools/testing/selftests/ftrace/test.d/kprobe/kprobe_syntax_errors.tc
@@ -8,7 +8,7 @@ check_error() { # command-with-error-pos-by-^
}
if grep -q 'r\[maxactive\]' README; then
-check_error 'p^100 vfs_read' # MAXACT_NO_KPROBE
+check_error 'p^100 vfs_read' # BAD_MAXACT_TYPE
check_error 'r^1a111 vfs_read' # BAD_MAXACT
check_error 'r^100000 vfs_read' # MAXACT_TOO_BIG
fi
^ permalink raw reply related [flat|nested] 17+ messages in thread* Re: [PATCH v13 03/12] tracing/probes: Add fprobe events for tracing function entry and exit.
2023-05-26 4:18 ` [PATCH v13 03/12] tracing/probes: Add fprobe events for tracing function entry and exit Masami Hiramatsu (Google)
@ 2023-05-26 19:38 ` Florent Revest
2023-05-29 3:16 ` Masami Hiramatsu
0 siblings, 1 reply; 17+ messages in thread
From: Florent Revest @ 2023-05-26 19:38 UTC (permalink / raw)
To: Masami Hiramatsu (Google)
Cc: linux-trace-kernel, linux-kernel, Steven Rostedt, Mark Rutland,
Will Deacon, Mathieu Desnoyers, Martin KaFai Lau, bpf,
Bagas Sanjaya
On Fri, May 26, 2023 at 6:18 AM Masami Hiramatsu (Google)
<mhiramat@kernel.org> wrote:
>
> [...] Since
> CONFIG_KPROBES_ON_FTRACE requires the CONFIG_DYNAMIC_FTRACE_WITH_REGS,
> it is not available if the architecture only supports
> CONFIG_DYNAMIC_FTRACE_WITH_ARGS. And that means kprobe events can not
> probe function entry/exit effectively on such architecture.
> But this can be solved if the dynamic events supports fprobe events.
Currently CONFIG_FPROBE also requires CONFIG_DYNAMIC_FTRACE_WITH_REGS
so iiuc this will only be true when we'll have migrated fprobe to use
ftrace_regs instead of pt_regs right ?
We discussed having fprobe use ftrace_regs instead of pt_regs in the
past and I even had a proof of concept branch at one point but this
patch seems to make this transition quite a bit harder. Have you tried
to make fprobe work on ftrace_regs on top of this patch ?
> diff --git a/kernel/trace/trace_fprobe.c b/kernel/trace/trace_fprobe.c
> new file mode 100644
> index 000000000000..48dbbc72b7dd
> --- /dev/null
> +++ b/kernel/trace/trace_fprobe.c
> @@ -0,0 +1,1053 @@
> +// SPDX-License-Identifier: GPL-2.0
> +/*
> + * Fprobe-based tracing events
> + * Copyright (C) 2022 Google LLC.
> + */
> +#define pr_fmt(fmt) "trace_fprobe: " fmt
> +
> +#include <linux/fprobe.h>
> +#include <linux/module.h>
> +#include <linux/rculist.h>
> +#include <linux/security.h>
> +#include <linux/uaccess.h>
> +
> +#include "trace_dynevent.h"
> +#include "trace_probe.h"
> +#include "trace_probe_kernel.h"
> +#include "trace_probe_tmpl.h"
> +
> +#define FPROBE_EVENT_SYSTEM "fprobes"
> +#define RETHOOK_MAXACTIVE_MAX 4096
> +
> +static int trace_fprobe_create(const char *raw_command);
> +static int trace_fprobe_show(struct seq_file *m, struct dyn_event *ev);
> +static int trace_fprobe_release(struct dyn_event *ev);
> +static bool trace_fprobe_is_busy(struct dyn_event *ev);
> +static bool trace_fprobe_match(const char *system, const char *event,
> + int argc, const char **argv, struct dyn_event *ev);
> +
> +static struct dyn_event_operations trace_fprobe_ops = {
> + .create = trace_fprobe_create,
> + .show = trace_fprobe_show,
> + .is_busy = trace_fprobe_is_busy,
> + .free = trace_fprobe_release,
> + .match = trace_fprobe_match,
> +};
> +
> +/*
> + * Fprobe event core functions
> + */
> +struct trace_fprobe {
> + struct dyn_event devent;
> + struct fprobe fp;
> + const char *symbol;
> + struct trace_probe tp;
> +};
> +
> +static bool is_trace_fprobe(struct dyn_event *ev)
> +{
> + return ev->ops == &trace_fprobe_ops;
> +}
> +
> +static struct trace_fprobe *to_trace_fprobe(struct dyn_event *ev)
> +{
> + return container_of(ev, struct trace_fprobe, devent);
> +}
> +
> +/**
> + * for_each_trace_fprobe - iterate over the trace_fprobe list
> + * @pos: the struct trace_fprobe * for each entry
> + * @dpos: the struct dyn_event * to use as a loop cursor
> + */
> +#define for_each_trace_fprobe(pos, dpos) \
> + for_each_dyn_event(dpos) \
> + if (is_trace_fprobe(dpos) && (pos = to_trace_fprobe(dpos)))
> +
> +static bool trace_fprobe_is_return(struct trace_fprobe *tf)
> +{
> + return tf->fp.exit_handler != NULL;
> +}
> +
> +static const char *trace_fprobe_symbol(struct trace_fprobe *tf)
> +{
> + return tf->symbol ? tf->symbol : "unknown";
> +}
> +
> +static bool trace_fprobe_is_busy(struct dyn_event *ev)
> +{
> + struct trace_fprobe *tf = to_trace_fprobe(ev);
> +
> + return trace_probe_is_enabled(&tf->tp);
> +}
> +
> +static bool trace_fprobe_match_command_head(struct trace_fprobe *tf,
> + int argc, const char **argv)
> +{
> + char buf[MAX_ARGSTR_LEN + 1];
> +
> + if (!argc)
> + return true;
> +
> + snprintf(buf, sizeof(buf), "%s", trace_fprobe_symbol(tf));
> + if (strcmp(buf, argv[0]))
> + return false;
> + argc--; argv++;
> +
> + return trace_probe_match_command_args(&tf->tp, argc, argv);
> +}
> +
> +static bool trace_fprobe_match(const char *system, const char *event,
> + int argc, const char **argv, struct dyn_event *ev)
> +{
> + struct trace_fprobe *tf = to_trace_fprobe(ev);
> +
> + if (event[0] != '\0' && strcmp(trace_probe_name(&tf->tp), event))
> + return false;
> +
> + if (system && strcmp(trace_probe_group_name(&tf->tp), system))
> + return false;
> +
> + return trace_fprobe_match_command_head(tf, argc, argv);
> +}
> +
> +static bool trace_fprobe_is_registered(struct trace_fprobe *tf)
> +{
> + return fprobe_is_registered(&tf->fp);
> +}
> +
> +/*
> + * Note that we don't verify the fetch_insn code, since it does not come
> + * from user space.
> + */
> +static int
> +process_fetch_insn(struct fetch_insn *code, void *rec, void *dest,
> + void *base)
> +{
> + struct pt_regs *regs = rec;
I gave it a try this week and it was mostly a matter of replacing
pt_regs with ftrace_regs in this file. Like here for example. Not too
bad so far.
> + unsigned long val;
> + int ret;
> +
> +retry:
> + /* 1st stage: get value from context */
> + switch (code->op) {
> + case FETCH_OP_STACK:
> + val = regs_get_kernel_stack_nth(regs, code->param);
This does not have a ftrace_regs equivalent at the moment. I suppose
we could introduce one without too much effort so that's probably ok.
> + break;
> + case FETCH_OP_STACKP:
> + val = kernel_stack_pointer(regs);
> + break;
> + case FETCH_OP_RETVAL:
> + val = regs_return_value(regs);
> + break;
> +#ifdef CONFIG_HAVE_FUNCTION_ARG_ACCESS_API
> + case FETCH_OP_ARG:
> + val = regs_get_kernel_argument(regs, code->param);
> + break;
> +#endif
> + case FETCH_NOP_SYMBOL: /* Ignore a place holder */
> + code++;
> + goto retry;
> + default:
> + ret = process_common_fetch_insn(code, &val);
> + if (ret < 0)
> + return ret;
> + }
> + code++;
> +
> + return process_fetch_insn_bottom(code, val, dest, base);
> +}
> +NOKPROBE_SYMBOL(process_fetch_insn)
> +
> +/* function entry handler */
> +static nokprobe_inline void
> +__fentry_trace_func(struct trace_fprobe *tf, unsigned long entry_ip,
> + struct pt_regs *regs,
> + struct trace_event_file *trace_file)
> +{
> + struct fentry_trace_entry_head *entry;
> + struct trace_event_call *call = trace_probe_event_call(&tf->tp);
> + struct trace_event_buffer fbuffer;
> + int dsize;
> +
> + if (WARN_ON_ONCE(call != trace_file->event_call))
> + return;
> +
> + if (trace_trigger_soft_disabled(trace_file))
> + return;
> +
> + dsize = __get_data_size(&tf->tp, regs);
> +
> + entry = trace_event_buffer_reserve(&fbuffer, trace_file,
> + sizeof(*entry) + tf->tp.size + dsize);
> + if (!entry)
> + return;
> +
> + fbuffer.regs = regs;
> + entry = fbuffer.entry = ring_buffer_event_data(fbuffer.event);
> + entry->ip = entry_ip;
> + store_trace_args(&entry[1], &tf->tp, regs, sizeof(*entry), dsize);
> +
> + trace_event_buffer_commit(&fbuffer);
> +}
> +
> +static void
> +fentry_trace_func(struct trace_fprobe *tf, unsigned long entry_ip,
> + struct pt_regs *regs)
> +{
> + struct event_file_link *link;
> +
> + trace_probe_for_each_link_rcu(link, &tf->tp)
> + __fentry_trace_func(tf, entry_ip, regs, link->file);
> +}
> +NOKPROBE_SYMBOL(fentry_trace_func);
> +
> +/* Kretprobe handler */
> +static nokprobe_inline void
> +__fexit_trace_func(struct trace_fprobe *tf, unsigned long entry_ip,
> + unsigned long ret_ip, struct pt_regs *regs,
> + struct trace_event_file *trace_file)
> +{
> + struct fexit_trace_entry_head *entry;
> + struct trace_event_buffer fbuffer;
> + struct trace_event_call *call = trace_probe_event_call(&tf->tp);
> + int dsize;
> +
> + if (WARN_ON_ONCE(call != trace_file->event_call))
> + return;
> +
> + if (trace_trigger_soft_disabled(trace_file))
> + return;
> +
> + dsize = __get_data_size(&tf->tp, regs);
> +
> + entry = trace_event_buffer_reserve(&fbuffer, trace_file,
> + sizeof(*entry) + tf->tp.size + dsize);
> + if (!entry)
> + return;
> +
> + fbuffer.regs = regs;
> + entry = fbuffer.entry = ring_buffer_event_data(fbuffer.event);
> + entry->func = entry_ip;
> + entry->ret_ip = ret_ip;
> + store_trace_args(&entry[1], &tf->tp, regs, sizeof(*entry), dsize);
> +
> + trace_event_buffer_commit(&fbuffer);
> +}
> +
> +static void
> +fexit_trace_func(struct trace_fprobe *tf, unsigned long entry_ip,
> + unsigned long ret_ip, struct pt_regs *regs)
> +{
> + struct event_file_link *link;
> +
> + trace_probe_for_each_link_rcu(link, &tf->tp)
> + __fexit_trace_func(tf, entry_ip, ret_ip, regs, link->file);
> +}
> +NOKPROBE_SYMBOL(fexit_trace_func);
> +
> +#ifdef CONFIG_PERF_EVENTS
> +
> +static int fentry_perf_func(struct trace_fprobe *tf, unsigned long entry_ip,
> + struct pt_regs *regs)
> +{
> + struct trace_event_call *call = trace_probe_event_call(&tf->tp);
> + struct fentry_trace_entry_head *entry;
> + struct hlist_head *head;
> + int size, __size, dsize;
> + int rctx;
> +
> + head = this_cpu_ptr(call->perf_events);
> + if (hlist_empty(head))
> + return 0;
> +
> + dsize = __get_data_size(&tf->tp, regs);
> + __size = sizeof(*entry) + tf->tp.size + dsize;
> + size = ALIGN(__size + sizeof(u32), sizeof(u64));
> + size -= sizeof(u32);
> +
> + entry = perf_trace_buf_alloc(size, NULL, &rctx);
> + if (!entry)
> + return 0;
> +
> + entry->ip = entry_ip;
> + memset(&entry[1], 0, dsize);
> + store_trace_args(&entry[1], &tf->tp, regs, sizeof(*entry), dsize);
> + perf_trace_buf_submit(entry, size, rctx, call->event.type, 1, regs,
However, that call concerns me. Perf requires a pt_regs pointer here
and it expects certain specific fields of that pt_regs to be set (the
exact requirements don't seem to be explicitly stated anywhere).
For example, on arm64 (the architecture without
CONFIG_DYNAMIC_FTRACE_WITH_REGS on which I'd like to have fprobe...
:)) perf calls the user_mode(regs) macro which expects the pstate
register to be set in pt_regs. However, pstate is invalid outside of
an exception entry so ftrace_regs on arm64 does not have a pstate
field at all.
If we migrate fprobe to ftrace_regs and try to construct a sparse
pt_regs out of a ftrace_regs here, it wouldn't be enough to just copy
all the registers we know from ftrace_regs into the pt_regs: we would
also need to make up the pstate register with the knowledge that it
has to be set in certain way specifically to please perf... Arch code
wouldn't only have to provide a
"expand_ftrace_regs_into_sparse_pt_regs" macro but also a
"invent_registers_for_perf" similar to the current
perf_arch_fetch_caller_regs macro. This seems rough...
It sounds to me like we should have avoided the use of sparse and
"made up" pt_regs a long time back and the more users of fprobe that
expect pt_regs we add, the more difficult we make it to make fprobe
work on CONFIG_DYNAMIC_FTRACE_WITH_ARGS.
> + head, NULL);
> + return 0;
> +}
> +NOKPROBE_SYMBOL(fentry_perf_func);
> +
> +static void
> +fexit_perf_func(struct trace_fprobe *tf, unsigned long entry_ip,
> + unsigned long ret_ip, struct pt_regs *regs)
> +{
> + struct trace_event_call *call = trace_probe_event_call(&tf->tp);
> + struct fexit_trace_entry_head *entry;
> + struct hlist_head *head;
> + int size, __size, dsize;
> + int rctx;
> +
> + head = this_cpu_ptr(call->perf_events);
> + if (hlist_empty(head))
> + return;
> +
> + dsize = __get_data_size(&tf->tp, regs);
> + __size = sizeof(*entry) + tf->tp.size + dsize;
> + size = ALIGN(__size + sizeof(u32), sizeof(u64));
> + size -= sizeof(u32);
> +
> + entry = perf_trace_buf_alloc(size, NULL, &rctx);
> + if (!entry)
> + return;
> +
> + entry->func = entry_ip;
> + entry->ret_ip = ret_ip;
> + store_trace_args(&entry[1], &tf->tp, regs, sizeof(*entry), dsize);
> + perf_trace_buf_submit(entry, size, rctx, call->event.type, 1, regs,
> + head, NULL);
> +}
> +NOKPROBE_SYMBOL(fexit_perf_func);
> +#endif /* CONFIG_PERF_EVENTS */
> +
> +static int fentry_dispatcher(struct fprobe *fp, unsigned long entry_ip,
> + unsigned long ret_ip, struct pt_regs *regs,
> + void *entry_data)
> +{
> + struct trace_fprobe *tf = container_of(fp, struct trace_fprobe, fp);
> + int ret = 0;
> +
> + if (trace_probe_test_flag(&tf->tp, TP_FLAG_TRACE))
> + fentry_trace_func(tf, entry_ip, regs);
> +#ifdef CONFIG_PERF_EVENTS
> + if (trace_probe_test_flag(&tf->tp, TP_FLAG_PROFILE))
> + ret = fentry_perf_func(tf, entry_ip, regs);
> +#endif
> + return ret;
> +}
> +NOKPROBE_SYMBOL(fentry_dispatcher);
> +
> +static void fexit_dispatcher(struct fprobe *fp, unsigned long entry_ip,
> + unsigned long ret_ip, struct pt_regs *regs,
> + void *entry_data)
> +{
> + struct trace_fprobe *tf = container_of(fp, struct trace_fprobe, fp);
> +
> + if (trace_probe_test_flag(&tf->tp, TP_FLAG_TRACE))
> + fexit_trace_func(tf, entry_ip, ret_ip, regs);
> +#ifdef CONFIG_PERF_EVENTS
> + if (trace_probe_test_flag(&tf->tp, TP_FLAG_PROFILE))
> + fexit_perf_func(tf, entry_ip, ret_ip, regs);
> +#endif
> +}
> +NOKPROBE_SYMBOL(fexit_dispatcher);
^ permalink raw reply [flat|nested] 17+ messages in thread* Re: [PATCH v13 03/12] tracing/probes: Add fprobe events for tracing function entry and exit.
2023-05-26 19:38 ` Florent Revest
@ 2023-05-29 3:16 ` Masami Hiramatsu
0 siblings, 0 replies; 17+ messages in thread
From: Masami Hiramatsu @ 2023-05-29 3:16 UTC (permalink / raw)
To: Florent Revest
Cc: linux-trace-kernel, linux-kernel, Steven Rostedt, Mark Rutland,
Will Deacon, Mathieu Desnoyers, Martin KaFai Lau, bpf,
Bagas Sanjaya
On Fri, 26 May 2023 21:38:44 +0200
Florent Revest <revest@chromium.org> wrote:
> On Fri, May 26, 2023 at 6:18 AM Masami Hiramatsu (Google)
> <mhiramat@kernel.org> wrote:
> >
> > [...] Since
> > CONFIG_KPROBES_ON_FTRACE requires the CONFIG_DYNAMIC_FTRACE_WITH_REGS,
> > it is not available if the architecture only supports
> > CONFIG_DYNAMIC_FTRACE_WITH_ARGS. And that means kprobe events can not
> > probe function entry/exit effectively on such architecture.
> > But this can be solved if the dynamic events supports fprobe events.
>
> Currently CONFIG_FPROBE also requires CONFIG_DYNAMIC_FTRACE_WITH_REGS
> so iiuc this will only be true when we'll have migrated fprobe to use
> ftrace_regs instead of pt_regs right ?
Sorry for confusion, yes, that's right. Currently does, but I will remove
that.
>
> We discussed having fprobe use ftrace_regs instead of pt_regs in the
> past and I even had a proof of concept branch at one point but this
> patch seems to make this transition quite a bit harder. Have you tried
> to make fprobe work on ftrace_regs on top of this patch ?
No, not yet, but taht should not be so hard. Let me try.
Thank you!
>
> > diff --git a/kernel/trace/trace_fprobe.c b/kernel/trace/trace_fprobe.c
> > new file mode 100644
> > index 000000000000..48dbbc72b7dd
> > --- /dev/null
> > +++ b/kernel/trace/trace_fprobe.c
> > @@ -0,0 +1,1053 @@
> > +// SPDX-License-Identifier: GPL-2.0
> > +/*
> > + * Fprobe-based tracing events
> > + * Copyright (C) 2022 Google LLC.
> > + */
> > +#define pr_fmt(fmt) "trace_fprobe: " fmt
> > +
> > +#include <linux/fprobe.h>
> > +#include <linux/module.h>
> > +#include <linux/rculist.h>
> > +#include <linux/security.h>
> > +#include <linux/uaccess.h>
> > +
> > +#include "trace_dynevent.h"
> > +#include "trace_probe.h"
> > +#include "trace_probe_kernel.h"
> > +#include "trace_probe_tmpl.h"
> > +
> > +#define FPROBE_EVENT_SYSTEM "fprobes"
> > +#define RETHOOK_MAXACTIVE_MAX 4096
> > +
> > +static int trace_fprobe_create(const char *raw_command);
> > +static int trace_fprobe_show(struct seq_file *m, struct dyn_event *ev);
> > +static int trace_fprobe_release(struct dyn_event *ev);
> > +static bool trace_fprobe_is_busy(struct dyn_event *ev);
> > +static bool trace_fprobe_match(const char *system, const char *event,
> > + int argc, const char **argv, struct dyn_event *ev);
> > +
> > +static struct dyn_event_operations trace_fprobe_ops = {
> > + .create = trace_fprobe_create,
> > + .show = trace_fprobe_show,
> > + .is_busy = trace_fprobe_is_busy,
> > + .free = trace_fprobe_release,
> > + .match = trace_fprobe_match,
> > +};
> > +
> > +/*
> > + * Fprobe event core functions
> > + */
> > +struct trace_fprobe {
> > + struct dyn_event devent;
> > + struct fprobe fp;
> > + const char *symbol;
> > + struct trace_probe tp;
> > +};
> > +
> > +static bool is_trace_fprobe(struct dyn_event *ev)
> > +{
> > + return ev->ops == &trace_fprobe_ops;
> > +}
> > +
> > +static struct trace_fprobe *to_trace_fprobe(struct dyn_event *ev)
> > +{
> > + return container_of(ev, struct trace_fprobe, devent);
> > +}
> > +
> > +/**
> > + * for_each_trace_fprobe - iterate over the trace_fprobe list
> > + * @pos: the struct trace_fprobe * for each entry
> > + * @dpos: the struct dyn_event * to use as a loop cursor
> > + */
> > +#define for_each_trace_fprobe(pos, dpos) \
> > + for_each_dyn_event(dpos) \
> > + if (is_trace_fprobe(dpos) && (pos = to_trace_fprobe(dpos)))
> > +
> > +static bool trace_fprobe_is_return(struct trace_fprobe *tf)
> > +{
> > + return tf->fp.exit_handler != NULL;
> > +}
> > +
> > +static const char *trace_fprobe_symbol(struct trace_fprobe *tf)
> > +{
> > + return tf->symbol ? tf->symbol : "unknown";
> > +}
> > +
> > +static bool trace_fprobe_is_busy(struct dyn_event *ev)
> > +{
> > + struct trace_fprobe *tf = to_trace_fprobe(ev);
> > +
> > + return trace_probe_is_enabled(&tf->tp);
> > +}
> > +
> > +static bool trace_fprobe_match_command_head(struct trace_fprobe *tf,
> > + int argc, const char **argv)
> > +{
> > + char buf[MAX_ARGSTR_LEN + 1];
> > +
> > + if (!argc)
> > + return true;
> > +
> > + snprintf(buf, sizeof(buf), "%s", trace_fprobe_symbol(tf));
> > + if (strcmp(buf, argv[0]))
> > + return false;
> > + argc--; argv++;
> > +
> > + return trace_probe_match_command_args(&tf->tp, argc, argv);
> > +}
> > +
> > +static bool trace_fprobe_match(const char *system, const char *event,
> > + int argc, const char **argv, struct dyn_event *ev)
> > +{
> > + struct trace_fprobe *tf = to_trace_fprobe(ev);
> > +
> > + if (event[0] != '\0' && strcmp(trace_probe_name(&tf->tp), event))
> > + return false;
> > +
> > + if (system && strcmp(trace_probe_group_name(&tf->tp), system))
> > + return false;
> > +
> > + return trace_fprobe_match_command_head(tf, argc, argv);
> > +}
> > +
> > +static bool trace_fprobe_is_registered(struct trace_fprobe *tf)
> > +{
> > + return fprobe_is_registered(&tf->fp);
> > +}
> > +
> > +/*
> > + * Note that we don't verify the fetch_insn code, since it does not come
> > + * from user space.
> > + */
> > +static int
> > +process_fetch_insn(struct fetch_insn *code, void *rec, void *dest,
> > + void *base)
> > +{
> > + struct pt_regs *regs = rec;
>
> I gave it a try this week and it was mostly a matter of replacing
> pt_regs with ftrace_regs in this file. Like here for example. Not too
> bad so far.
>
> > + unsigned long val;
> > + int ret;
> > +
> > +retry:
> > + /* 1st stage: get value from context */
> > + switch (code->op) {
> > + case FETCH_OP_STACK:
> > + val = regs_get_kernel_stack_nth(regs, code->param);
>
> This does not have a ftrace_regs equivalent at the moment. I suppose
> we could introduce one without too much effort so that's probably ok.
>
> > + break;
> > + case FETCH_OP_STACKP:
> > + val = kernel_stack_pointer(regs);
> > + break;
> > + case FETCH_OP_RETVAL:
> > + val = regs_return_value(regs);
> > + break;
> > +#ifdef CONFIG_HAVE_FUNCTION_ARG_ACCESS_API
> > + case FETCH_OP_ARG:
> > + val = regs_get_kernel_argument(regs, code->param);
> > + break;
> > +#endif
> > + case FETCH_NOP_SYMBOL: /* Ignore a place holder */
> > + code++;
> > + goto retry;
> > + default:
> > + ret = process_common_fetch_insn(code, &val);
> > + if (ret < 0)
> > + return ret;
> > + }
> > + code++;
> > +
> > + return process_fetch_insn_bottom(code, val, dest, base);
> > +}
> > +NOKPROBE_SYMBOL(process_fetch_insn)
> > +
> > +/* function entry handler */
> > +static nokprobe_inline void
> > +__fentry_trace_func(struct trace_fprobe *tf, unsigned long entry_ip,
> > + struct pt_regs *regs,
> > + struct trace_event_file *trace_file)
> > +{
> > + struct fentry_trace_entry_head *entry;
> > + struct trace_event_call *call = trace_probe_event_call(&tf->tp);
> > + struct trace_event_buffer fbuffer;
> > + int dsize;
> > +
> > + if (WARN_ON_ONCE(call != trace_file->event_call))
> > + return;
> > +
> > + if (trace_trigger_soft_disabled(trace_file))
> > + return;
> > +
> > + dsize = __get_data_size(&tf->tp, regs);
> > +
> > + entry = trace_event_buffer_reserve(&fbuffer, trace_file,
> > + sizeof(*entry) + tf->tp.size + dsize);
> > + if (!entry)
> > + return;
> > +
> > + fbuffer.regs = regs;
> > + entry = fbuffer.entry = ring_buffer_event_data(fbuffer.event);
> > + entry->ip = entry_ip;
> > + store_trace_args(&entry[1], &tf->tp, regs, sizeof(*entry), dsize);
> > +
> > + trace_event_buffer_commit(&fbuffer);
> > +}
> > +
> > +static void
> > +fentry_trace_func(struct trace_fprobe *tf, unsigned long entry_ip,
> > + struct pt_regs *regs)
> > +{
> > + struct event_file_link *link;
> > +
> > + trace_probe_for_each_link_rcu(link, &tf->tp)
> > + __fentry_trace_func(tf, entry_ip, regs, link->file);
> > +}
> > +NOKPROBE_SYMBOL(fentry_trace_func);
> > +
> > +/* Kretprobe handler */
> > +static nokprobe_inline void
> > +__fexit_trace_func(struct trace_fprobe *tf, unsigned long entry_ip,
> > + unsigned long ret_ip, struct pt_regs *regs,
> > + struct trace_event_file *trace_file)
> > +{
> > + struct fexit_trace_entry_head *entry;
> > + struct trace_event_buffer fbuffer;
> > + struct trace_event_call *call = trace_probe_event_call(&tf->tp);
> > + int dsize;
> > +
> > + if (WARN_ON_ONCE(call != trace_file->event_call))
> > + return;
> > +
> > + if (trace_trigger_soft_disabled(trace_file))
> > + return;
> > +
> > + dsize = __get_data_size(&tf->tp, regs);
> > +
> > + entry = trace_event_buffer_reserve(&fbuffer, trace_file,
> > + sizeof(*entry) + tf->tp.size + dsize);
> > + if (!entry)
> > + return;
> > +
> > + fbuffer.regs = regs;
> > + entry = fbuffer.entry = ring_buffer_event_data(fbuffer.event);
> > + entry->func = entry_ip;
> > + entry->ret_ip = ret_ip;
> > + store_trace_args(&entry[1], &tf->tp, regs, sizeof(*entry), dsize);
> > +
> > + trace_event_buffer_commit(&fbuffer);
> > +}
> > +
> > +static void
> > +fexit_trace_func(struct trace_fprobe *tf, unsigned long entry_ip,
> > + unsigned long ret_ip, struct pt_regs *regs)
> > +{
> > + struct event_file_link *link;
> > +
> > + trace_probe_for_each_link_rcu(link, &tf->tp)
> > + __fexit_trace_func(tf, entry_ip, ret_ip, regs, link->file);
> > +}
> > +NOKPROBE_SYMBOL(fexit_trace_func);
> > +
> > +#ifdef CONFIG_PERF_EVENTS
> > +
> > +static int fentry_perf_func(struct trace_fprobe *tf, unsigned long entry_ip,
> > + struct pt_regs *regs)
> > +{
> > + struct trace_event_call *call = trace_probe_event_call(&tf->tp);
> > + struct fentry_trace_entry_head *entry;
> > + struct hlist_head *head;
> > + int size, __size, dsize;
> > + int rctx;
> > +
> > + head = this_cpu_ptr(call->perf_events);
> > + if (hlist_empty(head))
> > + return 0;
> > +
> > + dsize = __get_data_size(&tf->tp, regs);
> > + __size = sizeof(*entry) + tf->tp.size + dsize;
> > + size = ALIGN(__size + sizeof(u32), sizeof(u64));
> > + size -= sizeof(u32);
> > +
> > + entry = perf_trace_buf_alloc(size, NULL, &rctx);
> > + if (!entry)
> > + return 0;
> > +
> > + entry->ip = entry_ip;
> > + memset(&entry[1], 0, dsize);
> > + store_trace_args(&entry[1], &tf->tp, regs, sizeof(*entry), dsize);
> > + perf_trace_buf_submit(entry, size, rctx, call->event.type, 1, regs,
>
> However, that call concerns me. Perf requires a pt_regs pointer here
> and it expects certain specific fields of that pt_regs to be set (the
> exact requirements don't seem to be explicitly stated anywhere).
>
> For example, on arm64 (the architecture without
> CONFIG_DYNAMIC_FTRACE_WITH_REGS on which I'd like to have fprobe...
> :)) perf calls the user_mode(regs) macro which expects the pstate
> register to be set in pt_regs. However, pstate is invalid outside of
> an exception entry so ftrace_regs on arm64 does not have a pstate
> field at all.
>
> If we migrate fprobe to ftrace_regs and try to construct a sparse
> pt_regs out of a ftrace_regs here, it wouldn't be enough to just copy
> all the registers we know from ftrace_regs into the pt_regs: we would
> also need to make up the pstate register with the knowledge that it
> has to be set in certain way specifically to please perf... Arch code
> wouldn't only have to provide a
> "expand_ftrace_regs_into_sparse_pt_regs" macro but also a
> "invent_registers_for_perf" similar to the current
> perf_arch_fetch_caller_regs macro. This seems rough...
>
> It sounds to me like we should have avoided the use of sparse and
> "made up" pt_regs a long time back and the more users of fprobe that
> expect pt_regs we add, the more difficult we make it to make fprobe
> work on CONFIG_DYNAMIC_FTRACE_WITH_ARGS.
>
> > + head, NULL);
> > + return 0;
> > +}
> > +NOKPROBE_SYMBOL(fentry_perf_func);
> > +
> > +static void
> > +fexit_perf_func(struct trace_fprobe *tf, unsigned long entry_ip,
> > + unsigned long ret_ip, struct pt_regs *regs)
> > +{
> > + struct trace_event_call *call = trace_probe_event_call(&tf->tp);
> > + struct fexit_trace_entry_head *entry;
> > + struct hlist_head *head;
> > + int size, __size, dsize;
> > + int rctx;
> > +
> > + head = this_cpu_ptr(call->perf_events);
> > + if (hlist_empty(head))
> > + return;
> > +
> > + dsize = __get_data_size(&tf->tp, regs);
> > + __size = sizeof(*entry) + tf->tp.size + dsize;
> > + size = ALIGN(__size + sizeof(u32), sizeof(u64));
> > + size -= sizeof(u32);
> > +
> > + entry = perf_trace_buf_alloc(size, NULL, &rctx);
> > + if (!entry)
> > + return;
> > +
> > + entry->func = entry_ip;
> > + entry->ret_ip = ret_ip;
> > + store_trace_args(&entry[1], &tf->tp, regs, sizeof(*entry), dsize);
> > + perf_trace_buf_submit(entry, size, rctx, call->event.type, 1, regs,
> > + head, NULL);
> > +}
> > +NOKPROBE_SYMBOL(fexit_perf_func);
> > +#endif /* CONFIG_PERF_EVENTS */
> > +
> > +static int fentry_dispatcher(struct fprobe *fp, unsigned long entry_ip,
> > + unsigned long ret_ip, struct pt_regs *regs,
> > + void *entry_data)
> > +{
> > + struct trace_fprobe *tf = container_of(fp, struct trace_fprobe, fp);
> > + int ret = 0;
> > +
> > + if (trace_probe_test_flag(&tf->tp, TP_FLAG_TRACE))
> > + fentry_trace_func(tf, entry_ip, regs);
> > +#ifdef CONFIG_PERF_EVENTS
> > + if (trace_probe_test_flag(&tf->tp, TP_FLAG_PROFILE))
> > + ret = fentry_perf_func(tf, entry_ip, regs);
> > +#endif
> > + return ret;
> > +}
> > +NOKPROBE_SYMBOL(fentry_dispatcher);
> > +
> > +static void fexit_dispatcher(struct fprobe *fp, unsigned long entry_ip,
> > + unsigned long ret_ip, struct pt_regs *regs,
> > + void *entry_data)
> > +{
> > + struct trace_fprobe *tf = container_of(fp, struct trace_fprobe, fp);
> > +
> > + if (trace_probe_test_flag(&tf->tp, TP_FLAG_TRACE))
> > + fexit_trace_func(tf, entry_ip, ret_ip, regs);
> > +#ifdef CONFIG_PERF_EVENTS
> > + if (trace_probe_test_flag(&tf->tp, TP_FLAG_PROFILE))
> > + fexit_perf_func(tf, entry_ip, ret_ip, regs);
> > +#endif
> > +}
> > +NOKPROBE_SYMBOL(fexit_dispatcher);
--
Masami Hiramatsu (Google) <mhiramat@kernel.org>
^ permalink raw reply [flat|nested] 17+ messages in thread
* [PATCH v13 04/12] selftests/ftrace: Add fprobe related testcases
2023-05-26 4:17 [PATCH v13 00/12] tracing: Add fprobe/tracepoint events Masami Hiramatsu (Google)
` (2 preceding siblings ...)
2023-05-26 4:18 ` [PATCH v13 03/12] tracing/probes: Add fprobe events for tracing function entry and exit Masami Hiramatsu (Google)
@ 2023-05-26 4:18 ` Masami Hiramatsu (Google)
2023-05-26 4:18 ` [PATCH v13 05/12] tracing/probes: Add tracepoint support on fprobe_events Masami Hiramatsu (Google)
` (7 subsequent siblings)
11 siblings, 0 replies; 17+ messages in thread
From: Masami Hiramatsu (Google) @ 2023-05-26 4:18 UTC (permalink / raw)
To: linux-trace-kernel
Cc: linux-kernel, Steven Rostedt, mhiramat, Florent Revest,
Mark Rutland, Will Deacon, Mathieu Desnoyers, Martin KaFai Lau,
bpf, Bagas Sanjaya
From: Masami Hiramatsu (Google) <mhiramat@kernel.org>
Add syntax error testcase and add-remove testcase for fprobe events.
This ensures that the fprobe events can be added/removed and parser
handles syntax errors correctly.
Signed-off-by: Masami Hiramatsu (Google) <mhiramat@kernel.org>
---
Changes in v13:
- Fix to pass correct register name.
---
.../ftrace/test.d/dynevent/add_remove_fprobe.tc | 26 +++++
.../ftrace/test.d/dynevent/fprobe_syntax_errors.tc | 97 ++++++++++++++++++++
2 files changed, 123 insertions(+)
create mode 100644 tools/testing/selftests/ftrace/test.d/dynevent/add_remove_fprobe.tc
create mode 100644 tools/testing/selftests/ftrace/test.d/dynevent/fprobe_syntax_errors.tc
diff --git a/tools/testing/selftests/ftrace/test.d/dynevent/add_remove_fprobe.tc b/tools/testing/selftests/ftrace/test.d/dynevent/add_remove_fprobe.tc
new file mode 100644
index 000000000000..53e0d5671687
--- /dev/null
+++ b/tools/testing/selftests/ftrace/test.d/dynevent/add_remove_fprobe.tc
@@ -0,0 +1,26 @@
+#!/bin/sh
+# SPDX-License-Identifier: GPL-2.0
+# description: Generic dynamic event - add/remove fprobe events
+# requires: dynamic_events "f[:[<group>/][<event>]] <func-name>[%return] [<args>]": README
+
+echo 0 > events/enable
+echo > dynamic_events
+
+PLACE=$FUNCTION_FORK
+
+echo "f:myevent1 $PLACE" >> dynamic_events
+echo "f:myevent2 $PLACE%return" >> dynamic_events
+
+grep -q myevent1 dynamic_events
+grep -q myevent2 dynamic_events
+test -d events/fprobes/myevent1
+test -d events/fprobes/myevent2
+
+echo "-:myevent2" >> dynamic_events
+
+grep -q myevent1 dynamic_events
+! grep -q myevent2 dynamic_events
+
+echo > dynamic_events
+
+clear_trace
diff --git a/tools/testing/selftests/ftrace/test.d/dynevent/fprobe_syntax_errors.tc b/tools/testing/selftests/ftrace/test.d/dynevent/fprobe_syntax_errors.tc
new file mode 100644
index 000000000000..48e7a1a6aae3
--- /dev/null
+++ b/tools/testing/selftests/ftrace/test.d/dynevent/fprobe_syntax_errors.tc
@@ -0,0 +1,97 @@
+#!/bin/sh
+# SPDX-License-Identifier: GPL-2.0
+# description: Fprobe event parser error log check
+# requires: dynamic_events "f[:[<group>/][<event>]] <func-name>[%return] [<args>]": README
+
+check_error() { # command-with-error-pos-by-^
+ ftrace_errlog_check 'trace_fprobe' "$1" 'dynamic_events'
+}
+
+case `uname -m` in
+x86_64|i[3456]86)
+ REG=%ax ;;
+aarch64)
+ REG=%x0 ;;
+*)
+ REG=%r0 ;;
+esac
+
+check_error 'f^100 vfs_read' # MAXACT_NO_KPROBE
+check_error 'f^1a111 vfs_read' # BAD_MAXACT
+check_error 'f^100000 vfs_read' # MAXACT_TOO_BIG
+
+check_error 'f ^non_exist_func' # BAD_PROBE_ADDR (enoent)
+check_error 'f ^vfs_read+10' # BAD_PROBE_ADDR
+check_error 'f:^/bar vfs_read' # NO_GROUP_NAME
+check_error 'f:^12345678901234567890123456789012345678901234567890123456789012345/bar vfs_read' # GROUP_TOO_LONG
+
+check_error 'f:^foo.1/bar vfs_read' # BAD_GROUP_NAME
+check_error 'f:^ vfs_read' # NO_EVENT_NAME
+check_error 'f:foo/^12345678901234567890123456789012345678901234567890123456789012345 vfs_read' # EVENT_TOO_LONG
+check_error 'f:foo/^bar.1 vfs_read' # BAD_EVENT_NAME
+
+check_error 'f vfs_read ^$retval' # RETVAL_ON_PROBE
+check_error 'f vfs_read ^$stack10000' # BAD_STACK_NUM
+
+check_error 'f vfs_read ^$arg10000' # BAD_ARG_NUM
+
+check_error 'f vfs_read ^$none_var' # BAD_VAR
+check_error 'f vfs_read ^'$REG # BAD_VAR
+
+check_error 'f vfs_read ^@12345678abcde' # BAD_MEM_ADDR
+check_error 'f vfs_read ^@+10' # FILE_ON_KPROBE
+
+grep -q "imm-value" README && \
+check_error 'f vfs_read arg1=\^x' # BAD_IMM
+grep -q "imm-string" README && \
+check_error 'f vfs_read arg1=\"abcd^' # IMMSTR_NO_CLOSE
+
+check_error 'f vfs_read ^+0@0)' # DEREF_NEED_BRACE
+check_error 'f vfs_read ^+0ab1(@0)' # BAD_DEREF_OFFS
+check_error 'f vfs_read +0(+0(@0^)' # DEREF_OPEN_BRACE
+
+if grep -A1 "fetcharg:" README | grep -q '\$comm' ; then
+check_error 'f vfs_read +0(^$comm)' # COMM_CANT_DEREF
+fi
+
+check_error 'f vfs_read ^&1' # BAD_FETCH_ARG
+
+
+# We've introduced this limitation with array support
+if grep -q ' <type>\\\[<array-size>\\\]' README; then
+check_error 'f vfs_read +0(^+0(+0(+0(+0(+0(+0(+0(+0(+0(+0(+0(+0(+0(@0))))))))))))))' # TOO_MANY_OPS?
+check_error 'f vfs_read +0(@11):u8[10^' # ARRAY_NO_CLOSE
+check_error 'f vfs_read +0(@11):u8[10]^a' # BAD_ARRAY_SUFFIX
+check_error 'f vfs_read +0(@11):u8[^10a]' # BAD_ARRAY_NUM
+check_error 'f vfs_read +0(@11):u8[^256]' # ARRAY_TOO_BIG
+fi
+
+check_error 'f vfs_read @11:^unknown_type' # BAD_TYPE
+check_error 'f vfs_read $stack0:^string' # BAD_STRING
+check_error 'f vfs_read @11:^b10@a/16' # BAD_BITFIELD
+
+check_error 'f vfs_read ^arg123456789012345678901234567890=@11' # ARG_NAME_TOO_LOG
+check_error 'f vfs_read ^=@11' # NO_ARG_NAME
+check_error 'f vfs_read ^var.1=@11' # BAD_ARG_NAME
+check_error 'f vfs_read var1=@11 ^var1=@12' # USED_ARG_NAME
+check_error 'f vfs_read ^+1234567(+1234567(+1234567(+1234567(+1234567(+1234567(@1234))))))' # ARG_TOO_LONG
+check_error 'f vfs_read arg1=^' # NO_ARG_BODY
+
+
+# multiprobe errors
+if grep -q "Create/append/" README && grep -q "imm-value" README; then
+echo "f:fprobes/testevent $FUNCTION_FORK" > dynamic_events
+check_error '^f:fprobes/testevent do_exit%return' # DIFF_PROBE_TYPE
+
+# Explicitly use printf "%s" to not interpret \1
+printf "%s" "f:fprobes/testevent $FUNCTION_FORK abcd=\\1" > dynamic_events
+check_error "f:fprobes/testevent $FUNCTION_FORK ^bcd=\\1" # DIFF_ARG_TYPE
+check_error "f:fprobes/testevent $FUNCTION_FORK ^abcd=\\1:u8" # DIFF_ARG_TYPE
+check_error "f:fprobes/testevent $FUNCTION_FORK ^abcd=\\\"foo\"" # DIFF_ARG_TYPE
+check_error "^f:fprobes/testevent $FUNCTION_FORK abcd=\\1" # SAME_PROBE
+fi
+
+# %return suffix errors
+check_error 'f vfs_read^%hoge' # BAD_ADDR_SUFFIX
+
+exit 0
^ permalink raw reply related [flat|nested] 17+ messages in thread* [PATCH v13 05/12] tracing/probes: Add tracepoint support on fprobe_events
2023-05-26 4:17 [PATCH v13 00/12] tracing: Add fprobe/tracepoint events Masami Hiramatsu (Google)
` (3 preceding siblings ...)
2023-05-26 4:18 ` [PATCH v13 04/12] selftests/ftrace: Add fprobe related testcases Masami Hiramatsu (Google)
@ 2023-05-26 4:18 ` Masami Hiramatsu (Google)
2023-05-26 4:18 ` [PATCH v13 06/12] tracing/probes: Move event parameter fetching code to common parser Masami Hiramatsu (Google)
` (6 subsequent siblings)
11 siblings, 0 replies; 17+ messages in thread
From: Masami Hiramatsu (Google) @ 2023-05-26 4:18 UTC (permalink / raw)
To: linux-trace-kernel
Cc: linux-kernel, Steven Rostedt, mhiramat, Florent Revest,
Mark Rutland, Will Deacon, Mathieu Desnoyers, Martin KaFai Lau,
bpf, Bagas Sanjaya
From: Masami Hiramatsu (Google) <mhiramat@kernel.org>
Allow fprobe_events to trace raw tracepoints so that user can trace
tracepoints which don't have traceevent wrappers. This new event is
always available if the fprobe_events is enabled (thus no kconfig),
because the fprobe_events depends on the trace-event and traceporint.
e.g.
# echo 't sched_overutilized_tp' >> dynamic_events
# echo 't 9p_client_req' >> dynamic_events
# cat dynamic_events
t:tracepoints/sched_overutilized_tp sched_overutilized_tp
t:tracepoints/_9p_client_req 9p_client_req
The event name is based on the tracepoint name, but if it is started
with digit character, an underscore '_' will be added.
NOTE: to avoid further confusion, this renames TPARG_FL_TPOINT to
TPARG_FL_TEVENT because this flag is used for eprobe (trace-event probe).
And reuse TPARG_FL_TPOINT for this raw tracepoint probe.
Reported-by: kernel test robot <lkp@intel.com>
Link: https://lore.kernel.org/oe-kbuild-all/202305020453.afTJ3VVp-lkp@intel.com/
Signed-off-by: Masami Hiramatsu (Google) <mhiramat@kernel.org>
---
Changes in v13:
- Add a comment about new flags.
Changes in v9.1:
- Fix !CONFIG_MODULES case.
Changes in v6:
- Update description
- Make find_tracepoint() static.
- Fix some checkpatches warnings.
---
include/linux/tracepoint-defs.h | 1
include/linux/tracepoint.h | 5 +
kernel/trace/trace.c | 1
kernel/trace/trace_eprobe.c | 2 -
kernel/trace/trace_fprobe.c | 134 +++++++++++++++++++++++++++++++++++++--
kernel/trace/trace_probe.c | 15 +++-
kernel/trace/trace_probe.h | 15 +++-
7 files changed, 157 insertions(+), 16 deletions(-)
diff --git a/include/linux/tracepoint-defs.h b/include/linux/tracepoint-defs.h
index e7c2276be33e..4dc4955f0fbf 100644
--- a/include/linux/tracepoint-defs.h
+++ b/include/linux/tracepoint-defs.h
@@ -35,6 +35,7 @@ struct tracepoint {
struct static_call_key *static_call_key;
void *static_call_tramp;
void *iterator;
+ void *probestub;
int (*regfunc)(void);
void (*unregfunc)(void);
struct tracepoint_func __rcu *funcs;
diff --git a/include/linux/tracepoint.h b/include/linux/tracepoint.h
index 6811e43c1b5c..88c0ba623ee6 100644
--- a/include/linux/tracepoint.h
+++ b/include/linux/tracepoint.h
@@ -303,6 +303,7 @@ static inline struct tracepoint *tracepoint_ptr_deref(tracepoint_ptr_t *p)
__section("__tracepoints_strings") = #_name; \
extern struct static_call_key STATIC_CALL_KEY(tp_func_##_name); \
int __traceiter_##_name(void *__data, proto); \
+ void __probestub_##_name(void *__data, proto); \
struct tracepoint __tracepoint_##_name __used \
__section("__tracepoints") = { \
.name = __tpstrtab_##_name, \
@@ -310,6 +311,7 @@ static inline struct tracepoint *tracepoint_ptr_deref(tracepoint_ptr_t *p)
.static_call_key = &STATIC_CALL_KEY(tp_func_##_name), \
.static_call_tramp = STATIC_CALL_TRAMP_ADDR(tp_func_##_name), \
.iterator = &__traceiter_##_name, \
+ .probestub = &__probestub_##_name, \
.regfunc = _reg, \
.unregfunc = _unreg, \
.funcs = NULL }; \
@@ -330,6 +332,9 @@ static inline struct tracepoint *tracepoint_ptr_deref(tracepoint_ptr_t *p)
} \
return 0; \
} \
+ void __probestub_##_name(void *__data, proto) \
+ { \
+ } \
DEFINE_STATIC_CALL(tp_func_##_name, __traceiter_##_name);
#define DEFINE_TRACE(name, proto, args) \
diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c
index d595cef93122..abcc2ca615e8 100644
--- a/kernel/trace/trace.c
+++ b/kernel/trace/trace.c
@@ -5653,6 +5653,7 @@ static const char readme_msg[] =
#endif
#ifdef CONFIG_FPROBE_EVENTS
"\t f[:[<group>/][<event>]] <func-name>[%return] [<args>]\n"
+ "\t t[:[<group>/][<event>]] <tracepoint> [<args>]\n"
#endif
#ifdef CONFIG_HIST_TRIGGERS
"\t s:[synthetic/]<event> <field> [<field>]\n"
diff --git a/kernel/trace/trace_eprobe.c b/kernel/trace/trace_eprobe.c
index 67e854979d53..fd64cd5d5745 100644
--- a/kernel/trace/trace_eprobe.c
+++ b/kernel/trace/trace_eprobe.c
@@ -817,7 +817,7 @@ find_and_get_event(const char *system, const char *event_name)
static int trace_eprobe_tp_update_arg(struct trace_eprobe *ep, const char *argv[], int i)
{
- unsigned int flags = TPARG_FL_KERNEL | TPARG_FL_TPOINT;
+ unsigned int flags = TPARG_FL_KERNEL | TPARG_FL_TEVENT;
int ret;
ret = traceprobe_parse_probe_arg(&ep->tp, i, argv[i], flags);
diff --git a/kernel/trace/trace_fprobe.c b/kernel/trace/trace_fprobe.c
index 48dbbc72b7dd..aa71ccb4205c 100644
--- a/kernel/trace/trace_fprobe.c
+++ b/kernel/trace/trace_fprobe.c
@@ -9,6 +9,7 @@
#include <linux/module.h>
#include <linux/rculist.h>
#include <linux/security.h>
+#include <linux/tracepoint.h>
#include <linux/uaccess.h>
#include "trace_dynevent.h"
@@ -17,6 +18,7 @@
#include "trace_probe_tmpl.h"
#define FPROBE_EVENT_SYSTEM "fprobes"
+#define TRACEPOINT_EVENT_SYSTEM "tracepoints"
#define RETHOOK_MAXACTIVE_MAX 4096
static int trace_fprobe_create(const char *raw_command);
@@ -41,6 +43,8 @@ struct trace_fprobe {
struct dyn_event devent;
struct fprobe fp;
const char *symbol;
+ struct tracepoint *tpoint;
+ struct module *mod;
struct trace_probe tp;
};
@@ -68,6 +72,11 @@ static bool trace_fprobe_is_return(struct trace_fprobe *tf)
return tf->fp.exit_handler != NULL;
}
+static bool trace_fprobe_is_tracepoint(struct trace_fprobe *tf)
+{
+ return tf->tpoint != NULL;
+}
+
static const char *trace_fprobe_symbol(struct trace_fprobe *tf)
{
return tf->symbol ? tf->symbol : "unknown";
@@ -668,6 +677,21 @@ static int __register_trace_fprobe(struct trace_fprobe *tf)
else
tf->fp.flags |= FPROBE_FL_DISABLED;
+ if (trace_fprobe_is_tracepoint(tf)) {
+ struct tracepoint *tpoint = tf->tpoint;
+ unsigned long ip = (unsigned long)tpoint->probestub;
+ /*
+ * Here, we do 2 steps to enable fprobe on a tracepoint.
+ * At first, put __probestub_##TP function on the tracepoint
+ * and put a fprobe on the stub function.
+ */
+ ret = tracepoint_probe_register_prio_may_exist(tpoint,
+ tpoint->probestub, NULL, 0);
+ if (ret < 0)
+ return ret;
+ return register_fprobe_ips(&tf->fp, &ip, 1);
+ }
+
/* TODO: handle filter, nofilter or symbol list */
return register_fprobe(&tf->fp, tf->symbol, NULL);
}
@@ -678,6 +702,12 @@ static void __unregister_trace_fprobe(struct trace_fprobe *tf)
if (trace_fprobe_is_registered(tf)) {
unregister_fprobe(&tf->fp);
memset(&tf->fp, 0, sizeof(tf->fp));
+ if (trace_fprobe_is_tracepoint(tf)) {
+ tracepoint_probe_unregister(tf->tpoint,
+ tf->tpoint->probestub, NULL);
+ tf->tpoint = NULL;
+ tf->mod = NULL;
+ }
}
}
@@ -741,7 +771,8 @@ static int append_trace_fprobe(struct trace_fprobe *tf, struct trace_fprobe *to)
{
int ret;
- if (trace_fprobe_is_return(tf) != trace_fprobe_is_return(to)) {
+ if (trace_fprobe_is_return(tf) != trace_fprobe_is_return(to) ||
+ trace_fprobe_is_tracepoint(tf) != trace_fprobe_is_tracepoint(to)) {
trace_probe_log_set_index(0);
trace_probe_log_err(0, DIFF_PROBE_TYPE);
return -EEXIST;
@@ -811,6 +842,60 @@ static int register_trace_fprobe(struct trace_fprobe *tf)
return ret;
}
+#ifdef CONFIG_MODULES
+static int __tracepoint_probe_module_cb(struct notifier_block *self,
+ unsigned long val, void *data)
+{
+ struct tp_module *tp_mod = data;
+ struct trace_fprobe *tf;
+ struct dyn_event *pos;
+
+ if (val != MODULE_STATE_GOING)
+ return NOTIFY_DONE;
+
+ mutex_lock(&event_mutex);
+ for_each_trace_fprobe(tf, pos) {
+ if (tp_mod->mod == tf->mod) {
+ tracepoint_probe_unregister(tf->tpoint,
+ tf->tpoint->probestub, NULL);
+ tf->tpoint = NULL;
+ tf->mod = NULL;
+ }
+ }
+ mutex_unlock(&event_mutex);
+
+ return NOTIFY_DONE;
+}
+
+static struct notifier_block tracepoint_module_nb = {
+ .notifier_call = __tracepoint_probe_module_cb,
+};
+#endif /* CONFIG_MODULES */
+
+struct __find_tracepoint_cb_data {
+ const char *tp_name;
+ struct tracepoint *tpoint;
+};
+
+static void __find_tracepoint_cb(struct tracepoint *tp, void *priv)
+{
+ struct __find_tracepoint_cb_data *data = priv;
+
+ if (!data->tpoint && !strcmp(data->tp_name, tp->name))
+ data->tpoint = tp;
+}
+
+static struct tracepoint *find_tracepoint(const char *tp_name)
+{
+ struct __find_tracepoint_cb_data data = {
+ .tp_name = tp_name,
+ };
+
+ for_each_kernel_tracepoint(__find_tracepoint_cb, &data);
+
+ return data.tpoint;
+}
+
static int __trace_fprobe_create(int argc, const char *argv[])
{
/*
@@ -819,6 +904,8 @@ static int __trace_fprobe_create(int argc, const char *argv[])
* f[:[GRP/][EVENT]] [MOD:]KSYM [FETCHARGS]
* - Add fexit probe:
* f[N][:[GRP/][EVENT]] [MOD:]KSYM%return [FETCHARGS]
+ * - Add tracepoint probe:
+ * t[:[GRP/][EVENT]] TRACEPOINT [FETCHARGS]
*
* Fetch args:
* $retval : fetch return value
@@ -844,10 +931,16 @@ static int __trace_fprobe_create(int argc, const char *argv[])
char buf[MAX_EVENT_NAME_LEN];
char gbuf[MAX_EVENT_NAME_LEN];
unsigned int flags = TPARG_FL_KERNEL | TPARG_FL_FPROBE;
+ bool is_tracepoint = false;
- if (argv[0][0] != 'f' || argc < 2)
+ if ((argv[0][0] != 'f' && argv[0][0] != 't') || argc < 2)
return -ECANCELED;
+ if (argv[0][0] == 't') {
+ is_tracepoint = true;
+ group = TRACEPOINT_EVENT_SYSTEM;
+ }
+
trace_probe_log_init("trace_fprobe", argc, argv);
event = strchr(&argv[0][1], ':');
@@ -881,14 +974,14 @@ static int __trace_fprobe_create(int argc, const char *argv[])
trace_probe_log_set_index(1);
- /* a symbol specified */
+ /* a symbol(or tracepoint) must be specified */
symbol = kstrdup(argv[1], GFP_KERNEL);
if (!symbol)
return -ENOMEM;
tmp = strchr(symbol, '%');
if (tmp) {
- if (!strcmp(tmp, "%return")) {
+ if (!is_tracepoint && !strcmp(tmp, "%return")) {
*tmp = '\0';
is_return = true;
} else {
@@ -907,6 +1000,9 @@ static int __trace_fprobe_create(int argc, const char *argv[])
else
flags |= TPARG_FL_FENTRY;
+ if (is_tracepoint)
+ flags |= TPARG_FL_TPOINT;
+
trace_probe_log_set_index(0);
if (event) {
ret = traceprobe_parse_event_name(&event, &group, gbuf,
@@ -917,8 +1013,11 @@ static int __trace_fprobe_create(int argc, const char *argv[])
if (!event) {
/* Make a new event name */
- snprintf(buf, MAX_EVENT_NAME_LEN, "%s__%s", symbol,
- is_return ? "exit" : "entry");
+ if (is_tracepoint)
+ strscpy(buf, symbol, MAX_EVENT_NAME_LEN);
+ else
+ snprintf(buf, MAX_EVENT_NAME_LEN, "%s__%s", symbol,
+ is_return ? "exit" : "entry");
sanitize_event_name(buf);
event = buf;
}
@@ -932,6 +1031,18 @@ static int __trace_fprobe_create(int argc, const char *argv[])
WARN_ON_ONCE(ret != -ENOMEM);
goto out; /* We know tf is not allocated */
}
+
+ if (is_tracepoint) {
+ tf->tpoint = find_tracepoint(tf->symbol);
+ if (!tf->tpoint) {
+ trace_probe_log_set_index(1);
+ trace_probe_log_err(0, NO_TRACEPOINT);
+ goto parse_error;
+ }
+ tf->mod = __module_text_address(
+ (unsigned long)tf->tpoint->probestub);
+ }
+
argc -= 2; argv += 2;
/* parse arguments */
@@ -991,7 +1102,10 @@ static int trace_fprobe_show(struct seq_file *m, struct dyn_event *ev)
struct trace_fprobe *tf = to_trace_fprobe(ev);
int i;
- seq_putc(m, 'f');
+ if (trace_fprobe_is_tracepoint(tf))
+ seq_putc(m, 't');
+ else
+ seq_putc(m, 'f');
if (trace_fprobe_is_return(tf) && tf->fp.nr_maxactive)
seq_printf(m, "%d", tf->fp.nr_maxactive);
seq_printf(m, ":%s/%s", trace_probe_group_name(&tf->tp),
@@ -1048,6 +1162,12 @@ static __init int init_fprobe_trace_early(void)
if (ret)
return ret;
+#ifdef CONFIG_MODULES
+ ret = register_tracepoint_module_notifier(&tracepoint_module_nb);
+ if (ret)
+ return ret;
+#endif
+
return 0;
}
core_initcall(init_fprobe_trace_early);
diff --git a/kernel/trace/trace_probe.c b/kernel/trace/trace_probe.c
index 8646b097d56c..dffbed2ddae9 100644
--- a/kernel/trace/trace_probe.c
+++ b/kernel/trace/trace_probe.c
@@ -292,7 +292,7 @@ static int parse_probe_vars(char *arg, const struct fetch_type *t,
int ret = 0;
int len;
- if (flags & TPARG_FL_TPOINT) {
+ if (flags & TPARG_FL_TEVENT) {
if (code->data)
return -EFAULT;
code->data = kstrdup(arg, GFP_KERNEL);
@@ -326,8 +326,7 @@ static int parse_probe_vars(char *arg, const struct fetch_type *t,
} else if (strcmp(arg, "comm") == 0 || strcmp(arg, "COMM") == 0) {
code->op = FETCH_OP_COMM;
#ifdef CONFIG_HAVE_FUNCTION_ARG_ACCESS_API
- } else if (((flags & TPARG_FL_MASK) ==
- (TPARG_FL_KERNEL | TPARG_FL_FENTRY)) &&
+ } else if (tparg_is_function_entry(flags) &&
(len = str_has_prefix(arg, "arg"))) {
ret = kstrtoul(arg + len, 10, ¶m);
if (ret) {
@@ -338,6 +337,12 @@ static int parse_probe_vars(char *arg, const struct fetch_type *t,
}
code->op = FETCH_OP_ARG;
code->param = (unsigned int)param - 1;
+ /*
+ * The tracepoint probe will probe a stub function, and the
+ * first parameter of the stub is a dummy and should be ignored.
+ */
+ if (flags & TPARG_FL_TPOINT)
+ code->param++;
#endif
} else
goto inval_var;
@@ -393,7 +398,7 @@ parse_probe_arg(char *arg, const struct fetch_type *type,
break;
case '%': /* named register */
- if (flags & (TPARG_FL_TPOINT | TPARG_FL_FPROBE)) {
+ if (flags & (TPARG_FL_TEVENT | TPARG_FL_FPROBE)) {
/* eprobe and fprobe do not handle registers */
trace_probe_log_err(offs, BAD_VAR);
break;
@@ -633,7 +638,7 @@ static int traceprobe_parse_probe_arg_body(const char *argv, ssize_t *size,
* Since $comm and immediate string can not be dereferenced,
* we can find those by strcmp. But ignore for eprobes.
*/
- if (!(flags & TPARG_FL_TPOINT) &&
+ if (!(flags & TPARG_FL_TEVENT) &&
(strcmp(arg, "$comm") == 0 || strcmp(arg, "$COMM") == 0 ||
strncmp(arg, "\\\"", 2) == 0)) {
/* The type of $comm must be "string", and not an array. */
diff --git a/kernel/trace/trace_probe.h b/kernel/trace/trace_probe.h
index cf146a7b7f81..7a0703c00a7d 100644
--- a/kernel/trace/trace_probe.h
+++ b/kernel/trace/trace_probe.h
@@ -359,16 +359,24 @@ int trace_probe_print_args(struct trace_seq *s, struct probe_arg *args, int nr_a
/*
* The flags used for parsing trace_probe arguments.
- * TPARG_FL_RETURN, TPARG_FL_FENTRY and TPARG_FL_TPOINT are mutually exclusive.
+ * TPARG_FL_RETURN, TPARG_FL_FENTRY and TPARG_FL_TEVENT are mutually exclusive.
* TPARG_FL_KERNEL and TPARG_FL_USER are also mutually exclusive.
+ * TPARG_FL_FPROBE and TPARG_FL_TPOINT are optional but it should be with
+ * TPARG_FL_KERNEL.
*/
#define TPARG_FL_RETURN BIT(0)
#define TPARG_FL_KERNEL BIT(1)
#define TPARG_FL_FENTRY BIT(2)
-#define TPARG_FL_TPOINT BIT(3)
+#define TPARG_FL_TEVENT BIT(3)
#define TPARG_FL_USER BIT(4)
#define TPARG_FL_FPROBE BIT(5)
-#define TPARG_FL_MASK GENMASK(4, 0)
+#define TPARG_FL_TPOINT BIT(6)
+#define TPARG_FL_LOC_MASK GENMASK(4, 0)
+
+static inline bool tparg_is_function_entry(unsigned int flags)
+{
+ return (flags & TPARG_FL_LOC_MASK) == (TPARG_FL_KERNEL | TPARG_FL_FENTRY);
+}
extern int traceprobe_parse_probe_arg(struct trace_probe *tp, int i,
const char *argv, unsigned int flags);
@@ -415,6 +423,7 @@ extern int traceprobe_define_arg_fields(struct trace_event_call *event_call,
C(MAXACT_TOO_BIG, "Maxactive is too big"), \
C(BAD_PROBE_ADDR, "Invalid probed address or symbol"), \
C(BAD_RETPROBE, "Retprobe address must be an function entry"), \
+ C(NO_TRACEPOINT, "Tracepoint is not found"), \
C(BAD_ADDR_SUFFIX, "Invalid probed address suffix"), \
C(NO_GROUP_NAME, "Group name is not specified"), \
C(GROUP_TOO_LONG, "Group name is too long"), \
^ permalink raw reply related [flat|nested] 17+ messages in thread* [PATCH v13 06/12] tracing/probes: Move event parameter fetching code to common parser
2023-05-26 4:17 [PATCH v13 00/12] tracing: Add fprobe/tracepoint events Masami Hiramatsu (Google)
` (4 preceding siblings ...)
2023-05-26 4:18 ` [PATCH v13 05/12] tracing/probes: Add tracepoint support on fprobe_events Masami Hiramatsu (Google)
@ 2023-05-26 4:18 ` Masami Hiramatsu (Google)
2023-05-26 4:19 ` [PATCH v13 07/12] tracing/probes: Support function parameters if BTF is available Masami Hiramatsu (Google)
` (5 subsequent siblings)
11 siblings, 0 replies; 17+ messages in thread
From: Masami Hiramatsu (Google) @ 2023-05-26 4:18 UTC (permalink / raw)
To: linux-trace-kernel
Cc: linux-kernel, Steven Rostedt, mhiramat, Florent Revest,
Mark Rutland, Will Deacon, Mathieu Desnoyers, Martin KaFai Lau,
bpf, Bagas Sanjaya
From: Masami Hiramatsu (Google) <mhiramat@kernel.org>
Move trace event parameter fetching code to common parser in
trace_probe.c. This simplifies eprobe's trace-event variable fetching
code by introducing a parse context data structure.
Signed-off-by: Masami Hiramatsu (Google) <mhiramat@kernel.org>
---
Changes in v10:
- Cleanup code to handle errors consistency way.
---
kernel/trace/trace_eprobe.c | 44 +--------
kernel/trace/trace_fprobe.c | 4 +
kernel/trace/trace_kprobe.c | 4 +
kernel/trace/trace_probe.c | 218 ++++++++++++++++++++++++++-----------------
kernel/trace/trace_probe.h | 9 ++
kernel/trace/trace_uprobe.c | 8 +-
6 files changed, 155 insertions(+), 132 deletions(-)
diff --git a/kernel/trace/trace_eprobe.c b/kernel/trace/trace_eprobe.c
index fd64cd5d5745..cb0077ba2b49 100644
--- a/kernel/trace/trace_eprobe.c
+++ b/kernel/trace/trace_eprobe.c
@@ -227,37 +227,6 @@ static struct trace_eprobe *alloc_event_probe(const char *group,
return ERR_PTR(ret);
}
-static int trace_eprobe_tp_arg_update(struct trace_eprobe *ep, int i)
-{
- struct probe_arg *parg = &ep->tp.args[i];
- struct ftrace_event_field *field;
- struct list_head *head;
- int ret = -ENOENT;
-
- head = trace_get_fields(ep->event);
- list_for_each_entry(field, head, link) {
- if (!strcmp(parg->code->data, field->name)) {
- kfree(parg->code->data);
- parg->code->data = field;
- return 0;
- }
- }
-
- /*
- * Argument not found on event. But allow for comm and COMM
- * to be used to get the current->comm.
- */
- if (strcmp(parg->code->data, "COMM") == 0 ||
- strcmp(parg->code->data, "comm") == 0) {
- parg->code->op = FETCH_OP_COMM;
- ret = 0;
- }
-
- kfree(parg->code->data);
- parg->code->data = NULL;
- return ret;
-}
-
static int eprobe_event_define_fields(struct trace_event_call *event_call)
{
struct eprobe_trace_entry_head field;
@@ -817,19 +786,16 @@ find_and_get_event(const char *system, const char *event_name)
static int trace_eprobe_tp_update_arg(struct trace_eprobe *ep, const char *argv[], int i)
{
- unsigned int flags = TPARG_FL_KERNEL | TPARG_FL_TEVENT;
+ struct traceprobe_parse_context ctx = {
+ .event = ep->event,
+ .flags = TPARG_FL_KERNEL | TPARG_FL_TEVENT,
+ };
int ret;
- ret = traceprobe_parse_probe_arg(&ep->tp, i, argv[i], flags);
+ ret = traceprobe_parse_probe_arg(&ep->tp, i, argv[i], &ctx);
if (ret)
return ret;
- if (ep->tp.args[i].code->op == FETCH_OP_TP_ARG) {
- ret = trace_eprobe_tp_arg_update(ep, i);
- if (ret)
- trace_probe_log_err(0, BAD_ATTACH_ARG);
- }
-
/* Handle symbols "@" */
if (!ret)
ret = traceprobe_update_arg(&ep->tp.args[i]);
diff --git a/kernel/trace/trace_fprobe.c b/kernel/trace/trace_fprobe.c
index aa71ccb4205c..7d144e4a3fb6 100644
--- a/kernel/trace/trace_fprobe.c
+++ b/kernel/trace/trace_fprobe.c
@@ -1047,8 +1047,10 @@ static int __trace_fprobe_create(int argc, const char *argv[])
/* parse arguments */
for (i = 0; i < argc && i < MAX_TRACE_ARGS; i++) {
+ struct traceprobe_parse_context ctx = { .flags = flags };
+
trace_probe_log_set_index(i + 2);
- ret = traceprobe_parse_probe_arg(&tf->tp, i, argv[i], flags);
+ ret = traceprobe_parse_probe_arg(&tf->tp, i, argv[i], &ctx);
if (ret)
goto error; /* This can be -ENOMEM */
}
diff --git a/kernel/trace/trace_kprobe.c b/kernel/trace/trace_kprobe.c
index b7a4409674b3..1a3497719ada 100644
--- a/kernel/trace/trace_kprobe.c
+++ b/kernel/trace/trace_kprobe.c
@@ -867,8 +867,10 @@ static int __trace_kprobe_create(int argc, const char *argv[])
/* parse arguments */
for (i = 0; i < argc && i < MAX_TRACE_ARGS; i++) {
+ struct traceprobe_parse_context ctx = { .flags = flags };
+
trace_probe_log_set_index(i + 2);
- ret = traceprobe_parse_probe_arg(&tk->tp, i, argv[i], flags);
+ ret = traceprobe_parse_probe_arg(&tk->tp, i, argv[i], &ctx);
if (ret)
goto error; /* This can be -ENOMEM */
}
diff --git a/kernel/trace/trace_probe.c b/kernel/trace/trace_probe.c
index dffbed2ddae9..30eeac5e9b61 100644
--- a/kernel/trace/trace_probe.c
+++ b/kernel/trace/trace_probe.c
@@ -283,74 +283,114 @@ int traceprobe_parse_event_name(const char **pevent, const char **pgroup,
return 0;
}
+static int parse_trace_event_arg(char *arg, struct fetch_insn *code,
+ struct traceprobe_parse_context *ctx)
+{
+ struct ftrace_event_field *field;
+ struct list_head *head;
+
+ head = trace_get_fields(ctx->event);
+ list_for_each_entry(field, head, link) {
+ if (!strcmp(arg, field->name)) {
+ code->op = FETCH_OP_TP_ARG;
+ code->data = field;
+ return 0;
+ }
+ }
+ return -ENOENT;
+}
+
#define PARAM_MAX_STACK (THREAD_SIZE / sizeof(unsigned long))
static int parse_probe_vars(char *arg, const struct fetch_type *t,
- struct fetch_insn *code, unsigned int flags, int offs)
+ struct fetch_insn *code,
+ struct traceprobe_parse_context *ctx)
{
unsigned long param;
+ int err = TP_ERR_BAD_VAR;
int ret = 0;
int len;
- if (flags & TPARG_FL_TEVENT) {
+ if (ctx->flags & TPARG_FL_TEVENT) {
if (code->data)
return -EFAULT;
- code->data = kstrdup(arg, GFP_KERNEL);
- if (!code->data)
- return -ENOMEM;
- code->op = FETCH_OP_TP_ARG;
- } else if (strcmp(arg, "retval") == 0) {
- if (flags & TPARG_FL_RETURN) {
+ ret = parse_trace_event_arg(arg, code, ctx);
+ if (!ret)
+ return 0;
+ if (strcmp(arg, "comm") == 0 || strcmp(arg, "COMM") == 0) {
+ code->op = FETCH_OP_COMM;
+ return 0;
+ }
+ /* backward compatibility */
+ ctx->offset = 0;
+ goto inval;
+ }
+
+ if (strcmp(arg, "retval") == 0) {
+ if (ctx->flags & TPARG_FL_RETURN) {
code->op = FETCH_OP_RETVAL;
- } else {
- trace_probe_log_err(offs, RETVAL_ON_PROBE);
- ret = -EINVAL;
+ return 0;
}
- } else if ((len = str_has_prefix(arg, "stack"))) {
+ err = TP_ERR_RETVAL_ON_PROBE;
+ goto inval;
+ }
+
+ len = str_has_prefix(arg, "stack");
+ if (len) {
+
if (arg[len] == '\0') {
code->op = FETCH_OP_STACKP;
- } else if (isdigit(arg[len])) {
+ return 0;
+ }
+
+ if (isdigit(arg[len])) {
ret = kstrtoul(arg + len, 10, ¶m);
- if (ret) {
- goto inval_var;
- } else if ((flags & TPARG_FL_KERNEL) &&
- param > PARAM_MAX_STACK) {
- trace_probe_log_err(offs, BAD_STACK_NUM);
- ret = -EINVAL;
- } else {
- code->op = FETCH_OP_STACK;
- code->param = (unsigned int)param;
+ if (ret)
+ goto inval;
+
+ if ((ctx->flags & TPARG_FL_KERNEL) &&
+ param > PARAM_MAX_STACK) {
+ err = TP_ERR_BAD_STACK_NUM;
+ goto inval;
}
- } else
- goto inval_var;
- } else if (strcmp(arg, "comm") == 0 || strcmp(arg, "COMM") == 0) {
+ code->op = FETCH_OP_STACK;
+ code->param = (unsigned int)param;
+ return 0;
+ }
+ goto inval;
+ }
+
+ if (strcmp(arg, "comm") == 0 || strcmp(arg, "COMM") == 0) {
code->op = FETCH_OP_COMM;
+ return 0;
+ }
+
#ifdef CONFIG_HAVE_FUNCTION_ARG_ACCESS_API
- } else if (tparg_is_function_entry(flags) &&
- (len = str_has_prefix(arg, "arg"))) {
+ len = str_has_prefix(arg, "arg");
+ if (len && tparg_is_function_entry(ctx->flags)) {
ret = kstrtoul(arg + len, 10, ¶m);
- if (ret) {
- goto inval_var;
- } else if (!param || param > PARAM_MAX_STACK) {
- trace_probe_log_err(offs, BAD_ARG_NUM);
- return -EINVAL;
+ if (ret)
+ goto inval;
+
+ if (!param || param > PARAM_MAX_STACK) {
+ err = TP_ERR_BAD_ARG_NUM;
+ goto inval;
}
+
code->op = FETCH_OP_ARG;
code->param = (unsigned int)param - 1;
/*
* The tracepoint probe will probe a stub function, and the
* first parameter of the stub is a dummy and should be ignored.
*/
- if (flags & TPARG_FL_TPOINT)
+ if (ctx->flags & TPARG_FL_TPOINT)
code->param++;
+ return 0;
+ }
#endif
- } else
- goto inval_var;
- return ret;
-
-inval_var:
- trace_probe_log_err(offs, BAD_VAR);
+inval:
+ __trace_probe_log_err(ctx->offset, err);
return -EINVAL;
}
@@ -383,7 +423,7 @@ static int __parse_imm_string(char *str, char **pbuf, int offs)
static int
parse_probe_arg(char *arg, const struct fetch_type *type,
struct fetch_insn **pcode, struct fetch_insn *end,
- unsigned int flags, int offs)
+ struct traceprobe_parse_context *ctx)
{
struct fetch_insn *code = *pcode;
unsigned long param;
@@ -394,13 +434,13 @@ parse_probe_arg(char *arg, const struct fetch_type *type,
switch (arg[0]) {
case '$':
- ret = parse_probe_vars(arg + 1, type, code, flags, offs);
+ ret = parse_probe_vars(arg + 1, type, code, ctx);
break;
case '%': /* named register */
- if (flags & (TPARG_FL_TEVENT | TPARG_FL_FPROBE)) {
+ if (ctx->flags & (TPARG_FL_TEVENT | TPARG_FL_FPROBE)) {
/* eprobe and fprobe do not handle registers */
- trace_probe_log_err(offs, BAD_VAR);
+ trace_probe_log_err(ctx->offset, BAD_VAR);
break;
}
ret = regs_query_register_offset(arg + 1);
@@ -409,14 +449,14 @@ parse_probe_arg(char *arg, const struct fetch_type *type,
code->param = (unsigned int)ret;
ret = 0;
} else
- trace_probe_log_err(offs, BAD_REG_NAME);
+ trace_probe_log_err(ctx->offset, BAD_REG_NAME);
break;
case '@': /* memory, file-offset or symbol */
if (isdigit(arg[1])) {
ret = kstrtoul(arg + 1, 0, ¶m);
if (ret) {
- trace_probe_log_err(offs, BAD_MEM_ADDR);
+ trace_probe_log_err(ctx->offset, BAD_MEM_ADDR);
break;
}
/* load address */
@@ -424,13 +464,13 @@ parse_probe_arg(char *arg, const struct fetch_type *type,
code->immediate = param;
} else if (arg[1] == '+') {
/* kprobes don't support file offsets */
- if (flags & TPARG_FL_KERNEL) {
- trace_probe_log_err(offs, FILE_ON_KPROBE);
+ if (ctx->flags & TPARG_FL_KERNEL) {
+ trace_probe_log_err(ctx->offset, FILE_ON_KPROBE);
return -EINVAL;
}
ret = kstrtol(arg + 2, 0, &offset);
if (ret) {
- trace_probe_log_err(offs, BAD_FILE_OFFS);
+ trace_probe_log_err(ctx->offset, BAD_FILE_OFFS);
break;
}
@@ -438,8 +478,8 @@ parse_probe_arg(char *arg, const struct fetch_type *type,
code->immediate = (unsigned long)offset; // imm64?
} else {
/* uprobes don't support symbols */
- if (!(flags & TPARG_FL_KERNEL)) {
- trace_probe_log_err(offs, SYM_ON_UPROBE);
+ if (!(ctx->flags & TPARG_FL_KERNEL)) {
+ trace_probe_log_err(ctx->offset, SYM_ON_UPROBE);
return -EINVAL;
}
/* Preserve symbol for updating */
@@ -448,7 +488,7 @@ parse_probe_arg(char *arg, const struct fetch_type *type,
if (!code->data)
return -ENOMEM;
if (++code == end) {
- trace_probe_log_err(offs, TOO_MANY_OPS);
+ trace_probe_log_err(ctx->offset, TOO_MANY_OPS);
return -EINVAL;
}
code->op = FETCH_OP_IMM;
@@ -456,7 +496,7 @@ parse_probe_arg(char *arg, const struct fetch_type *type,
}
/* These are fetching from memory */
if (++code == end) {
- trace_probe_log_err(offs, TOO_MANY_OPS);
+ trace_probe_log_err(ctx->offset, TOO_MANY_OPS);
return -EINVAL;
}
*pcode = code;
@@ -475,36 +515,38 @@ parse_probe_arg(char *arg, const struct fetch_type *type,
arg++; /* Skip '+', because kstrtol() rejects it. */
tmp = strchr(arg, '(');
if (!tmp) {
- trace_probe_log_err(offs, DEREF_NEED_BRACE);
+ trace_probe_log_err(ctx->offset, DEREF_NEED_BRACE);
return -EINVAL;
}
*tmp = '\0';
ret = kstrtol(arg, 0, &offset);
if (ret) {
- trace_probe_log_err(offs, BAD_DEREF_OFFS);
+ trace_probe_log_err(ctx->offset, BAD_DEREF_OFFS);
break;
}
- offs += (tmp + 1 - arg) + (arg[0] != '-' ? 1 : 0);
+ ctx->offset += (tmp + 1 - arg) + (arg[0] != '-' ? 1 : 0);
arg = tmp + 1;
tmp = strrchr(arg, ')');
if (!tmp) {
- trace_probe_log_err(offs + strlen(arg),
+ trace_probe_log_err(ctx->offset + strlen(arg),
DEREF_OPEN_BRACE);
return -EINVAL;
} else {
- const struct fetch_type *t2 = find_fetch_type(NULL, flags);
+ const struct fetch_type *t2 = find_fetch_type(NULL, ctx->flags);
+ int cur_offs = ctx->offset;
*tmp = '\0';
- ret = parse_probe_arg(arg, t2, &code, end, flags, offs);
+ ret = parse_probe_arg(arg, t2, &code, end, ctx);
if (ret)
break;
+ ctx->offset = cur_offs;
if (code->op == FETCH_OP_COMM ||
code->op == FETCH_OP_DATA) {
- trace_probe_log_err(offs, COMM_CANT_DEREF);
+ trace_probe_log_err(ctx->offset, COMM_CANT_DEREF);
return -EINVAL;
}
if (++code == end) {
- trace_probe_log_err(offs, TOO_MANY_OPS);
+ trace_probe_log_err(ctx->offset, TOO_MANY_OPS);
return -EINVAL;
}
*pcode = code;
@@ -515,7 +557,7 @@ parse_probe_arg(char *arg, const struct fetch_type *type,
break;
case '\\': /* Immediate value */
if (arg[1] == '"') { /* Immediate string */
- ret = __parse_imm_string(arg + 2, &tmp, offs + 2);
+ ret = __parse_imm_string(arg + 2, &tmp, ctx->offset + 2);
if (ret)
break;
code->op = FETCH_OP_DATA;
@@ -523,7 +565,7 @@ parse_probe_arg(char *arg, const struct fetch_type *type,
} else {
ret = str_to_immediate(arg + 1, &code->immediate);
if (ret)
- trace_probe_log_err(offs + 1, BAD_IMM);
+ trace_probe_log_err(ctx->offset + 1, BAD_IMM);
else
code->op = FETCH_OP_IMM;
}
@@ -531,7 +573,7 @@ parse_probe_arg(char *arg, const struct fetch_type *type,
}
if (!ret && code->op == FETCH_OP_NOP) {
/* Parsed, but do not find fetch method */
- trace_probe_log_err(offs, BAD_FETCH_ARG);
+ trace_probe_log_err(ctx->offset, BAD_FETCH_ARG);
ret = -EINVAL;
}
return ret;
@@ -576,12 +618,13 @@ static int __parse_bitfield_probe_arg(const char *bf,
/* String length checking wrapper */
static int traceprobe_parse_probe_arg_body(const char *argv, ssize_t *size,
- struct probe_arg *parg, unsigned int flags, int offset)
+ struct probe_arg *parg,
+ struct traceprobe_parse_context *ctx)
{
struct fetch_insn *code, *scode, *tmp = NULL;
char *t, *t2, *t3;
- char *arg;
int ret, len;
+ char *arg;
arg = kstrdup(argv, GFP_KERNEL);
if (!arg)
@@ -590,10 +633,10 @@ static int traceprobe_parse_probe_arg_body(const char *argv, ssize_t *size,
ret = -EINVAL;
len = strlen(arg);
if (len > MAX_ARGSTR_LEN) {
- trace_probe_log_err(offset, ARG_TOO_LONG);
+ trace_probe_log_err(ctx->offset, ARG_TOO_LONG);
goto out;
} else if (len == 0) {
- trace_probe_log_err(offset, NO_ARG_BODY);
+ trace_probe_log_err(ctx->offset, NO_ARG_BODY);
goto out;
}
@@ -611,23 +654,24 @@ static int traceprobe_parse_probe_arg_body(const char *argv, ssize_t *size,
*t2++ = '\0';
t3 = strchr(t2, ']');
if (!t3) {
- offset += t2 + strlen(t2) - arg;
- trace_probe_log_err(offset,
+ int offs = t2 + strlen(t2) - arg;
+
+ trace_probe_log_err(ctx->offset + offs,
ARRAY_NO_CLOSE);
goto out;
} else if (t3[1] != '\0') {
- trace_probe_log_err(offset + t3 + 1 - arg,
+ trace_probe_log_err(ctx->offset + t3 + 1 - arg,
BAD_ARRAY_SUFFIX);
goto out;
}
*t3 = '\0';
if (kstrtouint(t2, 0, &parg->count) || !parg->count) {
- trace_probe_log_err(offset + t2 - arg,
+ trace_probe_log_err(ctx->offset + t2 - arg,
BAD_ARRAY_NUM);
goto out;
}
if (parg->count > MAX_ARRAY_LEN) {
- trace_probe_log_err(offset + t2 - arg,
+ trace_probe_log_err(ctx->offset + t2 - arg,
ARRAY_TOO_BIG);
goto out;
}
@@ -638,17 +682,17 @@ static int traceprobe_parse_probe_arg_body(const char *argv, ssize_t *size,
* Since $comm and immediate string can not be dereferenced,
* we can find those by strcmp. But ignore for eprobes.
*/
- if (!(flags & TPARG_FL_TEVENT) &&
+ if (!(ctx->flags & TPARG_FL_TEVENT) &&
(strcmp(arg, "$comm") == 0 || strcmp(arg, "$COMM") == 0 ||
strncmp(arg, "\\\"", 2) == 0)) {
/* The type of $comm must be "string", and not an array. */
if (parg->count || (t && strcmp(t, "string")))
goto out;
- parg->type = find_fetch_type("string", flags);
+ parg->type = find_fetch_type("string", ctx->flags);
} else
- parg->type = find_fetch_type(t, flags);
+ parg->type = find_fetch_type(t, ctx->flags);
if (!parg->type) {
- trace_probe_log_err(offset + (t ? (t - arg) : 0), BAD_TYPE);
+ trace_probe_log_err(ctx->offset + (t ? (t - arg) : 0), BAD_TYPE);
goto out;
}
parg->offset = *size;
@@ -670,7 +714,7 @@ static int traceprobe_parse_probe_arg_body(const char *argv, ssize_t *size,
code[FETCH_INSN_MAX - 1].op = FETCH_OP_END;
ret = parse_probe_arg(arg, parg->type, &code, &code[FETCH_INSN_MAX - 1],
- flags, offset);
+ ctx);
if (ret)
goto fail;
@@ -681,7 +725,7 @@ static int traceprobe_parse_probe_arg_body(const char *argv, ssize_t *size,
if (code->op != FETCH_OP_REG && code->op != FETCH_OP_STACK &&
code->op != FETCH_OP_RETVAL && code->op != FETCH_OP_ARG &&
code->op != FETCH_OP_DEREF && code->op != FETCH_OP_TP_ARG) {
- trace_probe_log_err(offset + (t ? (t - arg) : 0),
+ trace_probe_log_err(ctx->offset + (t ? (t - arg) : 0),
BAD_SYMSTRING);
goto fail;
}
@@ -689,7 +733,7 @@ static int traceprobe_parse_probe_arg_body(const char *argv, ssize_t *size,
if (code->op != FETCH_OP_DEREF && code->op != FETCH_OP_UDEREF &&
code->op != FETCH_OP_IMM && code->op != FETCH_OP_COMM &&
code->op != FETCH_OP_DATA && code->op != FETCH_OP_TP_ARG) {
- trace_probe_log_err(offset + (t ? (t - arg) : 0),
+ trace_probe_log_err(ctx->offset + (t ? (t - arg) : 0),
BAD_STRING);
goto fail;
}
@@ -708,7 +752,7 @@ static int traceprobe_parse_probe_arg_body(const char *argv, ssize_t *size,
*/
code++;
if (code->op != FETCH_OP_NOP) {
- trace_probe_log_err(offset, TOO_MANY_OPS);
+ trace_probe_log_err(ctx->offset, TOO_MANY_OPS);
goto fail;
}
}
@@ -731,7 +775,7 @@ static int traceprobe_parse_probe_arg_body(const char *argv, ssize_t *size,
} else {
code++;
if (code->op != FETCH_OP_NOP) {
- trace_probe_log_err(offset, TOO_MANY_OPS);
+ trace_probe_log_err(ctx->offset, TOO_MANY_OPS);
goto fail;
}
code->op = FETCH_OP_ST_RAW;
@@ -742,7 +786,7 @@ static int traceprobe_parse_probe_arg_body(const char *argv, ssize_t *size,
if (t != NULL) {
ret = __parse_bitfield_probe_arg(t, parg->type, &code);
if (ret) {
- trace_probe_log_err(offset + t - arg, BAD_BITFIELD);
+ trace_probe_log_err(ctx->offset + t - arg, BAD_BITFIELD);
goto fail;
}
}
@@ -752,13 +796,13 @@ static int traceprobe_parse_probe_arg_body(const char *argv, ssize_t *size,
if (scode->op != FETCH_OP_ST_MEM &&
scode->op != FETCH_OP_ST_STRING &&
scode->op != FETCH_OP_ST_USTRING) {
- trace_probe_log_err(offset + (t ? (t - arg) : 0),
+ trace_probe_log_err(ctx->offset + (t ? (t - arg) : 0),
BAD_STRING);
goto fail;
}
code++;
if (code->op != FETCH_OP_NOP) {
- trace_probe_log_err(offset, TOO_MANY_OPS);
+ trace_probe_log_err(ctx->offset, TOO_MANY_OPS);
goto fail;
}
code->op = FETCH_OP_LP_ARRAY;
@@ -807,7 +851,7 @@ static int traceprobe_conflict_field_name(const char *name,
}
int traceprobe_parse_probe_arg(struct trace_probe *tp, int i, const char *arg,
- unsigned int flags)
+ struct traceprobe_parse_context *ctx)
{
struct probe_arg *parg = &tp->args[i];
const char *body;
@@ -842,9 +886,9 @@ int traceprobe_parse_probe_arg(struct trace_probe *tp, int i, const char *arg,
trace_probe_log_err(0, USED_ARG_NAME);
return -EINVAL;
}
+ ctx->offset = body - arg;
/* Parse fetch argument */
- return traceprobe_parse_probe_arg_body(body, &tp->size, parg, flags,
- body - arg);
+ return traceprobe_parse_probe_arg_body(body, &tp->size, parg, ctx);
}
void traceprobe_free_probe_arg(struct probe_arg *arg)
diff --git a/kernel/trace/trace_probe.h b/kernel/trace/trace_probe.h
index 7a0703c00a7d..aa0991349c59 100644
--- a/kernel/trace/trace_probe.h
+++ b/kernel/trace/trace_probe.h
@@ -378,8 +378,15 @@ static inline bool tparg_is_function_entry(unsigned int flags)
return (flags & TPARG_FL_LOC_MASK) == (TPARG_FL_KERNEL | TPARG_FL_FENTRY);
}
+struct traceprobe_parse_context {
+ struct trace_event_call *event;
+ unsigned int flags;
+ int offset;
+};
+
extern int traceprobe_parse_probe_arg(struct trace_probe *tp, int i,
- const char *argv, unsigned int flags);
+ const char *argv,
+ struct traceprobe_parse_context *ctx);
extern int traceprobe_update_arg(struct probe_arg *arg);
extern void traceprobe_free_probe_arg(struct probe_arg *arg);
diff --git a/kernel/trace/trace_uprobe.c b/kernel/trace/trace_uprobe.c
index 8b92e34ff0c8..fa09b33ee731 100644
--- a/kernel/trace/trace_uprobe.c
+++ b/kernel/trace/trace_uprobe.c
@@ -686,10 +686,12 @@ static int __trace_uprobe_create(int argc, const char **argv)
/* parse arguments */
for (i = 0; i < argc && i < MAX_TRACE_ARGS; i++) {
+ struct traceprobe_parse_context ctx = {
+ .flags = (is_return ? TPARG_FL_RETURN : 0) | TPARG_FL_USER,
+ };
+
trace_probe_log_set_index(i + 2);
- ret = traceprobe_parse_probe_arg(&tu->tp, i, argv[i],
- (is_return ? TPARG_FL_RETURN : 0) |
- TPARG_FL_USER);
+ ret = traceprobe_parse_probe_arg(&tu->tp, i, argv[i], &ctx);
if (ret)
goto error;
}
^ permalink raw reply related [flat|nested] 17+ messages in thread* [PATCH v13 07/12] tracing/probes: Support function parameters if BTF is available
2023-05-26 4:17 [PATCH v13 00/12] tracing: Add fprobe/tracepoint events Masami Hiramatsu (Google)
` (5 preceding siblings ...)
2023-05-26 4:18 ` [PATCH v13 06/12] tracing/probes: Move event parameter fetching code to common parser Masami Hiramatsu (Google)
@ 2023-05-26 4:19 ` Masami Hiramatsu (Google)
2023-05-26 4:19 ` [PATCH v13 08/12] tracing/probes: Add $arg* meta argument for all function args Masami Hiramatsu (Google)
` (4 subsequent siblings)
11 siblings, 0 replies; 17+ messages in thread
From: Masami Hiramatsu (Google) @ 2023-05-26 4:19 UTC (permalink / raw)
To: linux-trace-kernel
Cc: linux-kernel, Steven Rostedt, mhiramat, Florent Revest,
Mark Rutland, Will Deacon, Mathieu Desnoyers, Martin KaFai Lau,
bpf, Bagas Sanjaya
From: Masami Hiramatsu (Google) <mhiramat@kernel.org>
Support function or tracepoint parameters by name if BTF support is enabled
and the event is for function entry (this feature can be used with kprobe-
events, fprobe-events and tracepoint probe events.)
Note that the BTF variable syntax does not require a prefix. If it starts
with an alphabetic character or an underscore ('_') without a prefix like
'$' and '%', it is considered as a BTF variable.
If you specify only the BTF variable name, the argument name will also
be the same name instead of 'arg*'.
# echo 'p vfs_read count pos' >> dynamic_events
# echo 'f vfs_write count pos' >> dynamic_events
# echo 't sched_overutilized_tp rd overutilized' >> dynamic_events
# cat dynamic_events
p:kprobes/p_vfs_read_0 vfs_read count=count pos=pos
f:fprobes/vfs_write__entry vfs_write count=count pos=pos
t:tracepoints/sched_overutilized_tp sched_overutilized_tp rd=rd overutilized=overutilized
Signed-off-by: Masami Hiramatsu (Google) <mhiramat@kernel.org>
Reviewed-by: Alan Maguire <alan.maguire@oracle.com>
Tested-by: Alan Maguire <alan.maguire@oracle.com>
---
Changes in v13:
- Fix to apply BTF arg name even if that has fetch_type.
Changes in v8:
- Update Kconfig help description.
Changes in v7:
- Add BTF_KIND_ENUM64 support
- Fix to check the ERR_PTR() type return value of bpf_get_btf_vmlinux().
Changes in v6:
- update patch description.
- use bpf_get_btf_vmlinux() instead of btf_parse_vmlinux().
---
kernel/trace/Kconfig | 12 ++
kernel/trace/trace.c | 4 +
kernel/trace/trace_fprobe.c | 53 ++++++-----
kernel/trace/trace_kprobe.c | 12 +-
kernel/trace/trace_probe.c | 211 +++++++++++++++++++++++++++++++++++++++++++
kernel/trace/trace_probe.h | 9 ++
6 files changed, 270 insertions(+), 31 deletions(-)
diff --git a/kernel/trace/Kconfig b/kernel/trace/Kconfig
index 8e10a9453c96..b3f90d602896 100644
--- a/kernel/trace/Kconfig
+++ b/kernel/trace/Kconfig
@@ -664,6 +664,18 @@ config FPROBE_EVENTS
and the kprobe events on function entry and exit will be
transparently converted to this fprobe events.
+config PROBE_EVENTS_BTF_ARGS
+ depends on HAVE_FUNCTION_ARG_ACCESS_API
+ depends on FPROBE_EVENTS || KPROBE_EVENTS
+ depends on DEBUG_INFO_BTF && BPF_SYSCALL
+ bool "Support BTF function arguments for probe events"
+ default y
+ help
+ The user can specify the arguments of the probe event using the names
+ of the arguments of the probed function, when the probe location is a
+ kernel function entry or a tracepoint.
+ This is available only if BTF (BPF Type Format) support is enabled.
+
config KPROBE_EVENTS
depends on KPROBES
depends on HAVE_REGS_AND_STACK_ACCESS_API
diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c
index abcc2ca615e8..7bd2741356ef 100644
--- a/kernel/trace/trace.c
+++ b/kernel/trace/trace.c
@@ -5670,7 +5670,11 @@ static const char readme_msg[] =
"\t args: <name>=fetcharg[:type]\n"
"\t fetcharg: (%<register>|$<efield>), @<address>, @<symbol>[+|-<offset>],\n"
#ifdef CONFIG_HAVE_FUNCTION_ARG_ACCESS_API
+#ifdef CONFIG_PROBE_EVENTS_BTF_ARGS
+ "\t $stack<index>, $stack, $retval, $comm, $arg<N>, <argname>\n"
+#else
"\t $stack<index>, $stack, $retval, $comm, $arg<N>,\n"
+#endif
#else
"\t $stack<index>, $stack, $retval, $comm,\n"
#endif
diff --git a/kernel/trace/trace_fprobe.c b/kernel/trace/trace_fprobe.c
index 7d144e4a3fb6..2dd884609321 100644
--- a/kernel/trace/trace_fprobe.c
+++ b/kernel/trace/trace_fprobe.c
@@ -366,6 +366,7 @@ static void free_trace_fprobe(struct trace_fprobe *tf)
static struct trace_fprobe *alloc_trace_fprobe(const char *group,
const char *event,
const char *symbol,
+ struct tracepoint *tpoint,
int maxactive,
int nargs, bool is_return)
{
@@ -385,6 +386,7 @@ static struct trace_fprobe *alloc_trace_fprobe(const char *group,
else
tf->fp.entry_handler = fentry_dispatcher;
+ tf->tpoint = tpoint;
tf->fp.nr_maxactive = maxactive;
ret = trace_probe_init(&tf->tp, event, group, false);
@@ -930,8 +932,12 @@ static int __trace_fprobe_create(int argc, const char *argv[])
int maxactive = 0;
char buf[MAX_EVENT_NAME_LEN];
char gbuf[MAX_EVENT_NAME_LEN];
- unsigned int flags = TPARG_FL_KERNEL | TPARG_FL_FPROBE;
+ char sbuf[KSYM_NAME_LEN];
bool is_tracepoint = false;
+ struct tracepoint *tpoint = NULL;
+ struct traceprobe_parse_context ctx = {
+ .flags = TPARG_FL_KERNEL | TPARG_FL_FPROBE,
+ };
if ((argv[0][0] != 'f' && argv[0][0] != 't') || argc < 2)
return -ECANCELED;
@@ -995,14 +1001,6 @@ static int __trace_fprobe_create(int argc, const char *argv[])
goto parse_error;
}
- if (is_return)
- flags |= TPARG_FL_RETURN;
- else
- flags |= TPARG_FL_FENTRY;
-
- if (is_tracepoint)
- flags |= TPARG_FL_TPOINT;
-
trace_probe_log_set_index(0);
if (event) {
ret = traceprobe_parse_event_name(&event, &group, gbuf,
@@ -1014,7 +1012,8 @@ static int __trace_fprobe_create(int argc, const char *argv[])
if (!event) {
/* Make a new event name */
if (is_tracepoint)
- strscpy(buf, symbol, MAX_EVENT_NAME_LEN);
+ snprintf(buf, MAX_EVENT_NAME_LEN, "%s%s",
+ isdigit(*symbol) ? "_" : "", symbol);
else
snprintf(buf, MAX_EVENT_NAME_LEN, "%s__%s", symbol,
is_return ? "exit" : "entry");
@@ -1022,8 +1021,27 @@ static int __trace_fprobe_create(int argc, const char *argv[])
event = buf;
}
+ if (is_return)
+ ctx.flags |= TPARG_FL_RETURN;
+ else
+ ctx.flags |= TPARG_FL_FENTRY;
+
+ if (is_tracepoint) {
+ ctx.flags |= TPARG_FL_TPOINT;
+ tpoint = find_tracepoint(symbol);
+ if (!tpoint) {
+ trace_probe_log_set_index(1);
+ trace_probe_log_err(0, NO_TRACEPOINT);
+ goto parse_error;
+ }
+ ctx.funcname = kallsyms_lookup(
+ (unsigned long)tpoint->probestub,
+ NULL, NULL, NULL, sbuf);
+ } else
+ ctx.funcname = symbol;
+
/* setup a probe */
- tf = alloc_trace_fprobe(group, event, symbol, maxactive,
+ tf = alloc_trace_fprobe(group, event, symbol, tpoint, maxactive,
argc - 2, is_return);
if (IS_ERR(tf)) {
ret = PTR_ERR(tf);
@@ -1032,24 +1050,15 @@ static int __trace_fprobe_create(int argc, const char *argv[])
goto out; /* We know tf is not allocated */
}
- if (is_tracepoint) {
- tf->tpoint = find_tracepoint(tf->symbol);
- if (!tf->tpoint) {
- trace_probe_log_set_index(1);
- trace_probe_log_err(0, NO_TRACEPOINT);
- goto parse_error;
- }
+ if (is_tracepoint)
tf->mod = __module_text_address(
(unsigned long)tf->tpoint->probestub);
- }
argc -= 2; argv += 2;
-
/* parse arguments */
for (i = 0; i < argc && i < MAX_TRACE_ARGS; i++) {
- struct traceprobe_parse_context ctx = { .flags = flags };
-
trace_probe_log_set_index(i + 2);
+ ctx.offset = 0;
ret = traceprobe_parse_probe_arg(&tf->tp, i, argv[i], &ctx);
if (ret)
goto error; /* This can be -ENOMEM */
diff --git a/kernel/trace/trace_kprobe.c b/kernel/trace/trace_kprobe.c
index 1a3497719ada..7cc32da3e8e8 100644
--- a/kernel/trace/trace_kprobe.c
+++ b/kernel/trace/trace_kprobe.c
@@ -742,7 +742,7 @@ static int __trace_kprobe_create(int argc, const char *argv[])
void *addr = NULL;
char buf[MAX_EVENT_NAME_LEN];
char gbuf[MAX_EVENT_NAME_LEN];
- unsigned int flags = TPARG_FL_KERNEL;
+ struct traceprobe_parse_context ctx = { .flags = TPARG_FL_KERNEL };
switch (argv[0][0]) {
case 'r':
@@ -823,10 +823,10 @@ static int __trace_kprobe_create(int argc, const char *argv[])
goto parse_error;
}
if (is_return)
- flags |= TPARG_FL_RETURN;
+ ctx.flags |= TPARG_FL_RETURN;
ret = kprobe_on_func_entry(NULL, symbol, offset);
if (ret == 0 && !is_return)
- flags |= TPARG_FL_FENTRY;
+ ctx.flags |= TPARG_FL_FENTRY;
/* Defer the ENOENT case until register kprobe */
if (ret == -EINVAL && is_return) {
trace_probe_log_err(0, BAD_RETPROBE);
@@ -856,7 +856,7 @@ static int __trace_kprobe_create(int argc, const char *argv[])
/* setup a probe */
tk = alloc_trace_kprobe(group, event, addr, symbol, offset, maxactive,
- argc - 2, is_return);
+ argc - 2, is_return);
if (IS_ERR(tk)) {
ret = PTR_ERR(tk);
/* This must return -ENOMEM, else there is a bug */
@@ -866,10 +866,10 @@ static int __trace_kprobe_create(int argc, const char *argv[])
argc -= 2; argv += 2;
/* parse arguments */
+ ctx.funcname = symbol;
for (i = 0; i < argc && i < MAX_TRACE_ARGS; i++) {
- struct traceprobe_parse_context ctx = { .flags = flags };
-
trace_probe_log_set_index(i + 2);
+ ctx.offset = 0;
ret = traceprobe_parse_probe_arg(&tk->tp, i, argv[i], &ctx);
if (ret)
goto error; /* This can be -ENOMEM */
diff --git a/kernel/trace/trace_probe.c b/kernel/trace/trace_probe.c
index 30eeac5e9b61..2a169ea2ada1 100644
--- a/kernel/trace/trace_probe.c
+++ b/kernel/trace/trace_probe.c
@@ -11,6 +11,8 @@
*/
#define pr_fmt(fmt) "trace_probe: " fmt
+#include <linux/bpf.h>
+
#include "trace_probe.h"
#undef C
@@ -300,6 +302,171 @@ static int parse_trace_event_arg(char *arg, struct fetch_insn *code,
return -ENOENT;
}
+#ifdef CONFIG_PROBE_EVENTS_BTF_ARGS
+
+static struct btf *traceprobe_get_btf(void)
+{
+ struct btf *btf = bpf_get_btf_vmlinux();
+
+ if (IS_ERR_OR_NULL(btf))
+ return NULL;
+
+ return btf;
+}
+
+static u32 btf_type_int(const struct btf_type *t)
+{
+ return *(u32 *)(t + 1);
+}
+
+static const char *type_from_btf_id(struct btf *btf, s32 id)
+{
+ const struct btf_type *t;
+ u32 intdata;
+ s32 tid;
+
+ /* TODO: const char * could be converted as a string */
+ t = btf_type_skip_modifiers(btf, id, &tid);
+
+ switch (BTF_INFO_KIND(t->info)) {
+ case BTF_KIND_ENUM:
+ /* enum is "int", so convert to "s32" */
+ return "s32";
+ case BTF_KIND_ENUM64:
+ return "s64";
+ case BTF_KIND_PTR:
+ /* pointer will be converted to "x??" */
+ if (IS_ENABLED(CONFIG_64BIT))
+ return "x64";
+ else
+ return "x32";
+ case BTF_KIND_INT:
+ intdata = btf_type_int(t);
+ if (BTF_INT_ENCODING(intdata) & BTF_INT_SIGNED) {
+ switch (BTF_INT_BITS(intdata)) {
+ case 8:
+ return "s8";
+ case 16:
+ return "s16";
+ case 32:
+ return "s32";
+ case 64:
+ return "s64";
+ }
+ } else { /* unsigned */
+ switch (BTF_INT_BITS(intdata)) {
+ case 8:
+ return "u8";
+ case 16:
+ return "u16";
+ case 32:
+ return "u32";
+ case 64:
+ return "u64";
+ }
+ }
+ }
+ /* TODO: support other types */
+
+ return NULL;
+}
+
+static const struct btf_param *find_btf_func_param(const char *funcname, s32 *nr)
+{
+ struct btf *btf = traceprobe_get_btf();
+ const struct btf_type *t;
+ s32 id;
+
+ if (!btf || !funcname || !nr)
+ return ERR_PTR(-EINVAL);
+
+ id = btf_find_by_name_kind(btf, funcname, BTF_KIND_FUNC);
+ if (id <= 0)
+ return ERR_PTR(-ENOENT);
+
+ /* Get BTF_KIND_FUNC type */
+ t = btf_type_by_id(btf, id);
+ if (!btf_type_is_func(t))
+ return ERR_PTR(-ENOENT);
+
+ /* The type of BTF_KIND_FUNC is BTF_KIND_FUNC_PROTO */
+ t = btf_type_by_id(btf, t->type);
+ if (!btf_type_is_func_proto(t))
+ return ERR_PTR(-ENOENT);
+
+ *nr = btf_type_vlen(t);
+
+ if (*nr)
+ return (const struct btf_param *)(t + 1);
+ else
+ return NULL;
+}
+
+static int parse_btf_arg(const char *varname, struct fetch_insn *code,
+ struct traceprobe_parse_context *ctx)
+{
+ struct btf *btf = traceprobe_get_btf();
+ const struct btf_param *params;
+ int i;
+
+ if (!btf) {
+ trace_probe_log_err(ctx->offset, NOSUP_BTFARG);
+ return -EOPNOTSUPP;
+ }
+
+ if (WARN_ON_ONCE(!ctx->funcname))
+ return -EINVAL;
+
+ if (!ctx->params) {
+ params = find_btf_func_param(ctx->funcname, &ctx->nr_params);
+ if (IS_ERR(params)) {
+ trace_probe_log_err(ctx->offset, NO_BTF_ENTRY);
+ return PTR_ERR(params);
+ }
+ ctx->params = params;
+ } else
+ params = ctx->params;
+
+ for (i = 0; i < ctx->nr_params; i++) {
+ const char *name = btf_name_by_offset(btf, params[i].name_off);
+
+ if (name && !strcmp(name, varname)) {
+ code->op = FETCH_OP_ARG;
+ code->param = i;
+ return 0;
+ }
+ }
+ trace_probe_log_err(ctx->offset, NO_BTFARG);
+ return -ENOENT;
+}
+
+static const struct fetch_type *parse_btf_arg_type(int arg_idx,
+ struct traceprobe_parse_context *ctx)
+{
+ struct btf *btf = traceprobe_get_btf();
+ const char *typestr = NULL;
+
+ if (btf && ctx->params)
+ typestr = type_from_btf_id(btf, ctx->params[arg_idx].type);
+
+ return find_fetch_type(typestr, ctx->flags);
+}
+#else
+static struct btf *traceprobe_get_btf(void)
+{
+ return NULL;
+}
+
+static int parse_btf_arg(const char *varname, struct fetch_insn *code,
+ struct traceprobe_parse_context *ctx)
+{
+ trace_probe_log_err(ctx->offset, NOSUP_BTFARG);
+ return -EOPNOTSUPP;
+}
+#define parse_btf_arg_type(idx, ctx) \
+ find_fetch_type(NULL, ctx->flags)
+#endif
+
#define PARAM_MAX_STACK (THREAD_SIZE / sizeof(unsigned long))
static int parse_probe_vars(char *arg, const struct fetch_type *t,
@@ -570,6 +737,15 @@ parse_probe_arg(char *arg, const struct fetch_type *type,
code->op = FETCH_OP_IMM;
}
break;
+ default:
+ if (isalpha(arg[0]) || arg[0] == '_') { /* BTF variable */
+ if (!tparg_is_function_entry(ctx->flags)) {
+ trace_probe_log_err(ctx->offset, NOSUP_BTFARG);
+ return -EINVAL;
+ }
+ ret = parse_btf_arg(arg, code, ctx);
+ break;
+ }
}
if (!ret && code->op == FETCH_OP_NOP) {
/* Parsed, but do not find fetch method */
@@ -718,6 +894,11 @@ static int traceprobe_parse_probe_arg_body(const char *argv, ssize_t *size,
if (ret)
goto fail;
+ /* Update storing type if BTF is available */
+ if (IS_ENABLED(CONFIG_PROBE_EVENTS_BTF_ARGS) &&
+ !t && code->op == FETCH_OP_ARG)
+ parg->type = parse_btf_arg_type(code->param, ctx);
+
ret = -EINVAL;
/* Store operation */
if (parg->type->is_string) {
@@ -850,6 +1031,33 @@ static int traceprobe_conflict_field_name(const char *name,
return 0;
}
+static char *generate_probe_arg_name(const char *arg, int idx)
+{
+ char *name = NULL;
+ const char *end;
+
+ /*
+ * If argument name is omitted, try arg as a name (BTF variable)
+ * or "argN".
+ */
+ if (IS_ENABLED(CONFIG_PROBE_EVENTS_BTF_ARGS)) {
+ end = strchr(arg, ':');
+ if (!end)
+ end = arg + strlen(arg);
+
+ name = kmemdup_nul(arg, end - arg, GFP_KERNEL);
+ if (!name || !is_good_name(name)) {
+ kfree(name);
+ name = NULL;
+ }
+ }
+
+ if (!name)
+ name = kasprintf(GFP_KERNEL, "arg%d", idx + 1);
+
+ return name;
+}
+
int traceprobe_parse_probe_arg(struct trace_probe *tp, int i, const char *arg,
struct traceprobe_parse_context *ctx)
{
@@ -871,8 +1079,7 @@ int traceprobe_parse_probe_arg(struct trace_probe *tp, int i, const char *arg,
parg->name = kmemdup_nul(arg, body - arg, GFP_KERNEL);
body++;
} else {
- /* If argument name is omitted, set "argN" */
- parg->name = kasprintf(GFP_KERNEL, "arg%d", i + 1);
+ parg->name = generate_probe_arg_name(arg, i);
body = arg;
}
if (!parg->name)
diff --git a/kernel/trace/trace_probe.h b/kernel/trace/trace_probe.h
index aa0991349c59..2c7e7e9e9ce7 100644
--- a/kernel/trace/trace_probe.h
+++ b/kernel/trace/trace_probe.h
@@ -23,6 +23,7 @@
#include <linux/limits.h>
#include <linux/uaccess.h>
#include <linux/bitops.h>
+#include <linux/btf.h>
#include <asm/bitsperlong.h>
#include "trace.h"
@@ -380,6 +381,9 @@ static inline bool tparg_is_function_entry(unsigned int flags)
struct traceprobe_parse_context {
struct trace_event_call *event;
+ const struct btf_param *params;
+ s32 nr_params;
+ const char *funcname;
unsigned int flags;
int offset;
};
@@ -478,7 +482,10 @@ extern int traceprobe_define_arg_fields(struct trace_event_call *event_call,
C(NO_EVENT_INFO, "This requires both group and event name to attach"),\
C(BAD_ATTACH_EVENT, "Attached event does not exist"),\
C(BAD_ATTACH_ARG, "Attached event does not have this field"),\
- C(NO_EP_FILTER, "No filter rule after 'if'"),
+ C(NO_EP_FILTER, "No filter rule after 'if'"), \
+ C(NOSUP_BTFARG, "BTF is not available or not supported"), \
+ C(NO_BTFARG, "This variable is not found at this probe point"),\
+ C(NO_BTF_ENTRY, "No BTF entry for this probe point"),
#undef C
#define C(a, b) TP_ERR_##a
^ permalink raw reply related [flat|nested] 17+ messages in thread* [PATCH v13 08/12] tracing/probes: Add $arg* meta argument for all function args
2023-05-26 4:17 [PATCH v13 00/12] tracing: Add fprobe/tracepoint events Masami Hiramatsu (Google)
` (6 preceding siblings ...)
2023-05-26 4:19 ` [PATCH v13 07/12] tracing/probes: Support function parameters if BTF is available Masami Hiramatsu (Google)
@ 2023-05-26 4:19 ` Masami Hiramatsu (Google)
2023-05-26 4:19 ` [PATCH v13 09/12] tracing/probes: Add BTF retval type support Masami Hiramatsu (Google)
` (3 subsequent siblings)
11 siblings, 0 replies; 17+ messages in thread
From: Masami Hiramatsu (Google) @ 2023-05-26 4:19 UTC (permalink / raw)
To: linux-trace-kernel
Cc: linux-kernel, Steven Rostedt, mhiramat, Florent Revest,
Mark Rutland, Will Deacon, Mathieu Desnoyers, Martin KaFai Lau,
bpf, Bagas Sanjaya
From: Masami Hiramatsu (Google) <mhiramat@kernel.org>
Add the '$arg*' meta fetch argument for function-entry probe events. This
will be expanded to the all arguments of the function and the tracepoint
using BTF function argument information.
e.g.
# echo 'p vfs_read $arg*' >> dynamic_events
# echo 'f vfs_write $arg*' >> dynamic_events
# echo 't sched_overutilized_tp $arg*' >> dynamic_events
# cat dynamic_events
p:kprobes/p_vfs_read_0 vfs_read file=file buf=buf count=count pos=pos
f:fprobes/vfs_write__entry vfs_write file=file buf=buf count=count pos=pos
t:tracepoints/sched_overutilized_tp sched_overutilized_tp rd=rd overutilized=overutilized
Also, single '$arg[0-9]*' will be converted to the BTF function argument.
NOTE: This seems like a wildcard, but a fake one at this moment. This
is just for telling user that this can be expanded to several arguments.
And it is not like other $-vars, you can not use this $arg* as a part of
fetch args, e.g. specifying name "foo=$arg*" and using it in dereferences
"+0($arg*)" will lead a parse error.
Signed-off-by: Masami Hiramatsu (Google) <mhiramat@kernel.org>
---
Changes in v13:
- Ignore BTF error if '$arg*' is not there.
- Fix to remove meaningless $argN format check.
- Fix kernel crash when only pass the $argN without '$arg*'.
- Fix to show $argN instead of $argN+1
Changes in v11:
- Change $args to $arg*.
- Convert single '$arg[0-9]*' to BTF function argument.
- Cleanup code.
Changes in v10:
- Change $$args to $args so that user can use $$ for current task's pid.
Changes in v6:
- update patch description.
---
kernel/trace/trace_fprobe.c | 21 +++++
kernel/trace/trace_kprobe.c | 23 +++++-
kernel/trace/trace_probe.c | 169 ++++++++++++++++++++++++++++++++++++++++++-
kernel/trace/trace_probe.h | 11 +++
4 files changed, 212 insertions(+), 12 deletions(-)
diff --git a/kernel/trace/trace_fprobe.c b/kernel/trace/trace_fprobe.c
index 2dd884609321..dfe2e546acdc 100644
--- a/kernel/trace/trace_fprobe.c
+++ b/kernel/trace/trace_fprobe.c
@@ -925,14 +925,16 @@ static int __trace_fprobe_create(int argc, const char *argv[])
* FETCHARG:TYPE : use TYPE instead of unsigned long.
*/
struct trace_fprobe *tf = NULL;
- int i, len, ret = 0;
+ int i, len, new_argc = 0, ret = 0;
bool is_return = false;
char *symbol = NULL, *tmp = NULL;
const char *event = NULL, *group = FPROBE_EVENT_SYSTEM;
+ const char **new_argv = NULL;
int maxactive = 0;
char buf[MAX_EVENT_NAME_LEN];
char gbuf[MAX_EVENT_NAME_LEN];
char sbuf[KSYM_NAME_LEN];
+ char abuf[MAX_BTF_ARGS_LEN];
bool is_tracepoint = false;
struct tracepoint *tpoint = NULL;
struct traceprobe_parse_context ctx = {
@@ -1040,9 +1042,22 @@ static int __trace_fprobe_create(int argc, const char *argv[])
} else
ctx.funcname = symbol;
+ argc -= 2; argv += 2;
+ new_argv = traceprobe_expand_meta_args(argc, argv, &new_argc,
+ abuf, MAX_BTF_ARGS_LEN, &ctx);
+ if (IS_ERR(new_argv)) {
+ ret = PTR_ERR(new_argv);
+ new_argv = NULL;
+ goto out;
+ }
+ if (new_argv) {
+ argc = new_argc;
+ argv = new_argv;
+ }
+
/* setup a probe */
tf = alloc_trace_fprobe(group, event, symbol, tpoint, maxactive,
- argc - 2, is_return);
+ argc, is_return);
if (IS_ERR(tf)) {
ret = PTR_ERR(tf);
/* This must return -ENOMEM, else there is a bug */
@@ -1054,7 +1069,6 @@ static int __trace_fprobe_create(int argc, const char *argv[])
tf->mod = __module_text_address(
(unsigned long)tf->tpoint->probestub);
- argc -= 2; argv += 2;
/* parse arguments */
for (i = 0; i < argc && i < MAX_TRACE_ARGS; i++) {
trace_probe_log_set_index(i + 2);
@@ -1083,6 +1097,7 @@ static int __trace_fprobe_create(int argc, const char *argv[])
out:
trace_probe_log_clear();
+ kfree(new_argv);
kfree(symbol);
return ret;
diff --git a/kernel/trace/trace_kprobe.c b/kernel/trace/trace_kprobe.c
index 7cc32da3e8e8..74adb82331dd 100644
--- a/kernel/trace/trace_kprobe.c
+++ b/kernel/trace/trace_kprobe.c
@@ -732,9 +732,10 @@ static int __trace_kprobe_create(int argc, const char *argv[])
* FETCHARG:TYPE : use TYPE instead of unsigned long.
*/
struct trace_kprobe *tk = NULL;
- int i, len, ret = 0;
+ int i, len, new_argc = 0, ret = 0;
bool is_return = false;
char *symbol = NULL, *tmp = NULL;
+ const char **new_argv = NULL;
const char *event = NULL, *group = KPROBE_EVENT_SYSTEM;
enum probe_print_type ptype;
int maxactive = 0;
@@ -742,6 +743,7 @@ static int __trace_kprobe_create(int argc, const char *argv[])
void *addr = NULL;
char buf[MAX_EVENT_NAME_LEN];
char gbuf[MAX_EVENT_NAME_LEN];
+ char abuf[MAX_BTF_ARGS_LEN];
struct traceprobe_parse_context ctx = { .flags = TPARG_FL_KERNEL };
switch (argv[0][0]) {
@@ -854,19 +856,31 @@ static int __trace_kprobe_create(int argc, const char *argv[])
event = buf;
}
+ argc -= 2; argv += 2;
+ ctx.funcname = symbol;
+ new_argv = traceprobe_expand_meta_args(argc, argv, &new_argc,
+ abuf, MAX_BTF_ARGS_LEN, &ctx);
+ if (IS_ERR(new_argv)) {
+ ret = PTR_ERR(new_argv);
+ new_argv = NULL;
+ goto out;
+ }
+ if (new_argv) {
+ argc = new_argc;
+ argv = new_argv;
+ }
+
/* setup a probe */
tk = alloc_trace_kprobe(group, event, addr, symbol, offset, maxactive,
- argc - 2, is_return);
+ argc, is_return);
if (IS_ERR(tk)) {
ret = PTR_ERR(tk);
/* This must return -ENOMEM, else there is a bug */
WARN_ON_ONCE(ret != -ENOMEM);
goto out; /* We know tk is not allocated */
}
- argc -= 2; argv += 2;
/* parse arguments */
- ctx.funcname = symbol;
for (i = 0; i < argc && i < MAX_TRACE_ARGS; i++) {
trace_probe_log_set_index(i + 2);
ctx.offset = 0;
@@ -894,6 +908,7 @@ static int __trace_kprobe_create(int argc, const char *argv[])
out:
trace_probe_log_clear();
+ kfree(new_argv);
kfree(symbol);
return ret;
diff --git a/kernel/trace/trace_probe.c b/kernel/trace/trace_probe.c
index 2a169ea2ada1..7318642aceb3 100644
--- a/kernel/trace/trace_probe.c
+++ b/kernel/trace/trace_probe.c
@@ -371,9 +371,11 @@ static const char *type_from_btf_id(struct btf *btf, s32 id)
return NULL;
}
-static const struct btf_param *find_btf_func_param(const char *funcname, s32 *nr)
+static const struct btf_param *find_btf_func_param(const char *funcname, s32 *nr,
+ bool tracepoint)
{
struct btf *btf = traceprobe_get_btf();
+ const struct btf_param *param;
const struct btf_type *t;
s32 id;
@@ -395,9 +397,16 @@ static const struct btf_param *find_btf_func_param(const char *funcname, s32 *nr
return ERR_PTR(-ENOENT);
*nr = btf_type_vlen(t);
+ param = (const struct btf_param *)(t + 1);
- if (*nr)
- return (const struct btf_param *)(t + 1);
+ /* Hide the first 'data' argument of tracepoint */
+ if (tracepoint) {
+ (*nr)--;
+ param++;
+ }
+
+ if (*nr > 0)
+ return param;
else
return NULL;
}
@@ -418,7 +427,8 @@ static int parse_btf_arg(const char *varname, struct fetch_insn *code,
return -EINVAL;
if (!ctx->params) {
- params = find_btf_func_param(ctx->funcname, &ctx->nr_params);
+ params = find_btf_func_param(ctx->funcname, &ctx->nr_params,
+ ctx->flags & TPARG_FL_TPOINT);
if (IS_ERR(params)) {
trace_probe_log_err(ctx->offset, NO_BTF_ENTRY);
return PTR_ERR(params);
@@ -451,12 +461,19 @@ static const struct fetch_type *parse_btf_arg_type(int arg_idx,
return find_fetch_type(typestr, ctx->flags);
}
+
#else
static struct btf *traceprobe_get_btf(void)
{
return NULL;
}
+static const struct btf_param *find_btf_func_param(const char *funcname, s32 *nr,
+ bool tracepoint)
+{
+ return ERR_PTR(-EOPNOTSUPP);
+}
+
static int parse_btf_arg(const char *varname, struct fetch_insn *code,
struct traceprobe_parse_context *ctx)
{
@@ -1114,6 +1131,150 @@ void traceprobe_free_probe_arg(struct probe_arg *arg)
kfree(arg->fmt);
}
+static int argv_has_var_arg(int argc, const char *argv[], int *args_idx,
+ struct traceprobe_parse_context *ctx)
+{
+ int i, found = 0;
+
+ for (i = 0; i < argc; i++)
+ if (str_has_prefix(argv[i], "$arg")) {
+ trace_probe_log_set_index(i + 2);
+
+ if (!tparg_is_function_entry(ctx->flags)) {
+ trace_probe_log_err(0, NOFENTRY_ARGS);
+ return -EINVAL;
+ }
+
+ if (isdigit(argv[i][4])) {
+ found = 1;
+ continue;
+ }
+
+ if (argv[i][4] != '*') {
+ trace_probe_log_err(0, BAD_VAR);
+ return -EINVAL;
+ }
+
+ if (*args_idx >= 0 && *args_idx < argc) {
+ trace_probe_log_err(0, DOUBLE_ARGS);
+ return -EINVAL;
+ }
+ found = 1;
+ *args_idx = i;
+ }
+
+ return found;
+}
+
+static int sprint_nth_btf_arg(int idx, const char *type,
+ char *buf, int bufsize,
+ struct traceprobe_parse_context *ctx)
+{
+ struct btf *btf = traceprobe_get_btf();
+ const char *name;
+ int ret;
+
+ if (idx >= ctx->nr_params) {
+ trace_probe_log_err(0, NO_BTFARG);
+ return -ENOENT;
+ }
+ name = btf_name_by_offset(btf, ctx->params[idx].name_off);
+ if (!name) {
+ trace_probe_log_err(0, NO_BTF_ENTRY);
+ return -ENOENT;
+ }
+ ret = snprintf(buf, bufsize, "%s%s", name, type);
+ if (ret >= bufsize) {
+ trace_probe_log_err(0, ARGS_2LONG);
+ return -E2BIG;
+ }
+ return ret;
+}
+
+/* Return new_argv which must be freed after use */
+const char **traceprobe_expand_meta_args(int argc, const char *argv[],
+ int *new_argc, char *buf, int bufsize,
+ struct traceprobe_parse_context *ctx)
+{
+ const struct btf_param *params = NULL;
+ int i, j, n, used, ret, args_idx = -1;
+ const char **new_argv = NULL;
+ int nr_params;
+
+ ret = argv_has_var_arg(argc, argv, &args_idx, ctx);
+ if (ret < 0)
+ return ERR_PTR(ret);
+
+ if (!ret) {
+ *new_argc = argc;
+ return NULL;
+ }
+
+ params = find_btf_func_param(ctx->funcname, &nr_params,
+ ctx->flags & TPARG_FL_TPOINT);
+ if (IS_ERR(params)) {
+ if (args_idx != -1) {
+ /* $arg* requires BTF info */
+ trace_probe_log_err(0, NOSUP_BTFARG);
+ return (const char **)params;
+ }
+ return 0;
+ }
+ ctx->params = params;
+ ctx->nr_params = nr_params;
+
+ if (args_idx >= 0)
+ *new_argc = argc + ctx->nr_params - 1;
+ else
+ *new_argc = argc;
+
+ new_argv = kcalloc(*new_argc, sizeof(char *), GFP_KERNEL);
+ if (!new_argv)
+ return ERR_PTR(-ENOMEM);
+
+ used = 0;
+ for (i = 0, j = 0; i < argc; i++) {
+ trace_probe_log_set_index(i + 2);
+ if (i == args_idx) {
+ for (n = 0; n < nr_params; n++) {
+ ret = sprint_nth_btf_arg(n, "", buf + used,
+ bufsize - used, ctx);
+ if (ret < 0)
+ goto error;
+
+ new_argv[j++] = buf + used;
+ used += ret + 1;
+ }
+ continue;
+ }
+
+ if (str_has_prefix(argv[i], "$arg")) {
+ char *type = NULL;
+
+ n = simple_strtoul(argv[i] + 4, &type, 10);
+ if (type && !(*type == ':' || *type == '\0')) {
+ trace_probe_log_err(0, BAD_VAR);
+ ret = -ENOENT;
+ goto error;
+ }
+ /* Note: $argN starts from $arg1 */
+ ret = sprint_nth_btf_arg(n - 1, type, buf + used,
+ bufsize - used, ctx);
+ if (ret < 0)
+ goto error;
+ new_argv[j++] = buf + used;
+ used += ret + 1;
+ } else
+ new_argv[j++] = argv[i];
+ }
+
+ return new_argv;
+
+error:
+ kfree(new_argv);
+ return ERR_PTR(ret);
+}
+
int traceprobe_update_arg(struct probe_arg *arg)
{
struct fetch_insn *code = arg->code;
diff --git a/kernel/trace/trace_probe.h b/kernel/trace/trace_probe.h
index 2c7e7e9e9ce7..c864e6dea10f 100644
--- a/kernel/trace/trace_probe.h
+++ b/kernel/trace/trace_probe.h
@@ -33,7 +33,9 @@
#define MAX_ARGSTR_LEN 63
#define MAX_ARRAY_LEN 64
#define MAX_ARG_NAME_LEN 32
+#define MAX_BTF_ARGS_LEN 128
#define MAX_STRING_SIZE PATH_MAX
+#define MAX_ARG_BUF_LEN (MAX_TRACE_ARGS * MAX_ARG_NAME_LEN)
/* Reserved field names */
#define FIELD_STRING_IP "__probe_ip"
@@ -391,6 +393,9 @@ struct traceprobe_parse_context {
extern int traceprobe_parse_probe_arg(struct trace_probe *tp, int i,
const char *argv,
struct traceprobe_parse_context *ctx);
+const char **traceprobe_expand_meta_args(int argc, const char *argv[],
+ int *new_argc, char *buf, int bufsize,
+ struct traceprobe_parse_context *ctx);
extern int traceprobe_update_arg(struct probe_arg *arg);
extern void traceprobe_free_probe_arg(struct probe_arg *arg);
@@ -485,7 +490,11 @@ extern int traceprobe_define_arg_fields(struct trace_event_call *event_call,
C(NO_EP_FILTER, "No filter rule after 'if'"), \
C(NOSUP_BTFARG, "BTF is not available or not supported"), \
C(NO_BTFARG, "This variable is not found at this probe point"),\
- C(NO_BTF_ENTRY, "No BTF entry for this probe point"),
+ C(NO_BTF_ENTRY, "No BTF entry for this probe point"), \
+ C(BAD_VAR_ARGS, "$arg* must be an independent parameter without name etc."),\
+ C(NOFENTRY_ARGS, "$arg* can be used only on function entry"), \
+ C(DOUBLE_ARGS, "$arg* can be used only once in the parameters"), \
+ C(ARGS_2LONG, "$arg* failed because the argument list is too long"),
#undef C
#define C(a, b) TP_ERR_##a
^ permalink raw reply related [flat|nested] 17+ messages in thread* [PATCH v13 09/12] tracing/probes: Add BTF retval type support
2023-05-26 4:17 [PATCH v13 00/12] tracing: Add fprobe/tracepoint events Masami Hiramatsu (Google)
` (7 preceding siblings ...)
2023-05-26 4:19 ` [PATCH v13 08/12] tracing/probes: Add $arg* meta argument for all function args Masami Hiramatsu (Google)
@ 2023-05-26 4:19 ` Masami Hiramatsu (Google)
2023-06-12 7:29 ` Donglin Peng
2023-05-26 4:19 ` [PATCH v13 10/12] selftests/ftrace: Add tracepoint probe test case Masami Hiramatsu (Google)
` (2 subsequent siblings)
11 siblings, 1 reply; 17+ messages in thread
From: Masami Hiramatsu (Google) @ 2023-05-26 4:19 UTC (permalink / raw)
To: linux-trace-kernel
Cc: linux-kernel, Steven Rostedt, mhiramat, Florent Revest,
Mark Rutland, Will Deacon, Mathieu Desnoyers, Martin KaFai Lau,
bpf, Bagas Sanjaya
From: Masami Hiramatsu (Google) <mhiramat@kernel.org>
Check the target function has non-void retval type and set the correct
fetch type if user doesn't specify it.
If the function returns void, $retval is rejected as below;
# echo 'f unregister_kprobes%return $retval' >> dynamic_events
sh: write error: No such file or directory
# cat error_log
[ 37.488397] trace_fprobe: error: This function returns 'void' type
Command: f unregister_kprobes%return $retval
^
Signed-off-by: Masami Hiramatsu (Google) <mhiramat@kernel.org>
---
Changes in v8:
- Fix wrong indentation.
Changes in v7:
- Introduce this as a new patch.
---
kernel/trace/trace_probe.c | 69 ++++++++++++++++++++++++++++++++++++++++----
kernel/trace/trace_probe.h | 1 +
2 files changed, 63 insertions(+), 7 deletions(-)
diff --git a/kernel/trace/trace_probe.c b/kernel/trace/trace_probe.c
index 7318642aceb3..dfe3e1823eec 100644
--- a/kernel/trace/trace_probe.c
+++ b/kernel/trace/trace_probe.c
@@ -371,15 +371,13 @@ static const char *type_from_btf_id(struct btf *btf, s32 id)
return NULL;
}
-static const struct btf_param *find_btf_func_param(const char *funcname, s32 *nr,
- bool tracepoint)
+static const struct btf_type *find_btf_func_proto(const char *funcname)
{
struct btf *btf = traceprobe_get_btf();
- const struct btf_param *param;
const struct btf_type *t;
s32 id;
- if (!btf || !funcname || !nr)
+ if (!btf || !funcname)
return ERR_PTR(-EINVAL);
id = btf_find_by_name_kind(btf, funcname, BTF_KIND_FUNC);
@@ -396,6 +394,22 @@ static const struct btf_param *find_btf_func_param(const char *funcname, s32 *nr
if (!btf_type_is_func_proto(t))
return ERR_PTR(-ENOENT);
+ return t;
+}
+
+static const struct btf_param *find_btf_func_param(const char *funcname, s32 *nr,
+ bool tracepoint)
+{
+ const struct btf_param *param;
+ const struct btf_type *t;
+
+ if (!funcname || !nr)
+ return ERR_PTR(-EINVAL);
+
+ t = find_btf_func_proto(funcname);
+ if (IS_ERR(t))
+ return (const struct btf_param *)t;
+
*nr = btf_type_vlen(t);
param = (const struct btf_param *)(t + 1);
@@ -462,6 +476,32 @@ static const struct fetch_type *parse_btf_arg_type(int arg_idx,
return find_fetch_type(typestr, ctx->flags);
}
+static const struct fetch_type *parse_btf_retval_type(
+ struct traceprobe_parse_context *ctx)
+{
+ struct btf *btf = traceprobe_get_btf();
+ const char *typestr = NULL;
+ const struct btf_type *t;
+
+ if (btf && ctx->funcname) {
+ t = find_btf_func_proto(ctx->funcname);
+ if (!IS_ERR(t))
+ typestr = type_from_btf_id(btf, t->type);
+ }
+
+ return find_fetch_type(typestr, ctx->flags);
+}
+
+static bool is_btf_retval_void(const char *funcname)
+{
+ const struct btf_type *t;
+
+ t = find_btf_func_proto(funcname);
+ if (IS_ERR(t))
+ return false;
+
+ return t->type == 0;
+}
#else
static struct btf *traceprobe_get_btf(void)
{
@@ -480,8 +520,15 @@ static int parse_btf_arg(const char *varname, struct fetch_insn *code,
trace_probe_log_err(ctx->offset, NOSUP_BTFARG);
return -EOPNOTSUPP;
}
+
#define parse_btf_arg_type(idx, ctx) \
find_fetch_type(NULL, ctx->flags)
+
+#define parse_btf_retval_type(ctx) \
+ find_fetch_type(NULL, ctx->flags)
+
+#define is_btf_retval_void(funcname) (false)
+
#endif
#define PARAM_MAX_STACK (THREAD_SIZE / sizeof(unsigned long))
@@ -512,6 +559,11 @@ static int parse_probe_vars(char *arg, const struct fetch_type *t,
if (strcmp(arg, "retval") == 0) {
if (ctx->flags & TPARG_FL_RETURN) {
+ if ((ctx->flags & TPARG_FL_KERNEL) &&
+ is_btf_retval_void(ctx->funcname)) {
+ err = TP_ERR_NO_RETVAL;
+ goto inval;
+ }
code->op = FETCH_OP_RETVAL;
return 0;
}
@@ -912,9 +964,12 @@ static int traceprobe_parse_probe_arg_body(const char *argv, ssize_t *size,
goto fail;
/* Update storing type if BTF is available */
- if (IS_ENABLED(CONFIG_PROBE_EVENTS_BTF_ARGS) &&
- !t && code->op == FETCH_OP_ARG)
- parg->type = parse_btf_arg_type(code->param, ctx);
+ if (IS_ENABLED(CONFIG_PROBE_EVENTS_BTF_ARGS) && !t) {
+ if (code->op == FETCH_OP_ARG)
+ parg->type = parse_btf_arg_type(code->param, ctx);
+ else if (code->op == FETCH_OP_RETVAL)
+ parg->type = parse_btf_retval_type(ctx);
+ }
ret = -EINVAL;
/* Store operation */
diff --git a/kernel/trace/trace_probe.h b/kernel/trace/trace_probe.h
index c864e6dea10f..eb43bea5c168 100644
--- a/kernel/trace/trace_probe.h
+++ b/kernel/trace/trace_probe.h
@@ -449,6 +449,7 @@ extern int traceprobe_define_arg_fields(struct trace_event_call *event_call,
C(BAD_EVENT_NAME, "Event name must follow the same rules as C identifiers"), \
C(EVENT_EXIST, "Given group/event name is already used by another event"), \
C(RETVAL_ON_PROBE, "$retval is not available on probe"), \
+ C(NO_RETVAL, "This function returns 'void' type"), \
C(BAD_STACK_NUM, "Invalid stack number"), \
C(BAD_ARG_NUM, "Invalid argument number"), \
C(BAD_VAR, "Invalid $-valiable specified"), \
^ permalink raw reply related [flat|nested] 17+ messages in thread* Re: [PATCH v13 09/12] tracing/probes: Add BTF retval type support
2023-05-26 4:19 ` [PATCH v13 09/12] tracing/probes: Add BTF retval type support Masami Hiramatsu (Google)
@ 2023-06-12 7:29 ` Donglin Peng
2023-07-07 14:12 ` Masami Hiramatsu
0 siblings, 1 reply; 17+ messages in thread
From: Donglin Peng @ 2023-06-12 7:29 UTC (permalink / raw)
To: Masami Hiramatsu (Google)
Cc: linux-kernel, Steven Rostedt, Florent Revest, Mark Rutland,
Will Deacon, Mathieu Desnoyers, Martin KaFai Lau, bpf,
Bagas Sanjaya, linux-trace-kernel, Ding Hui, huangcun
On 2023/5/26 12:19, Masami Hiramatsu (Google) wrote:
> From: Masami Hiramatsu (Google) <mhiramat@kernel.org>
>
> Check the target function has non-void retval type and set the correct
> fetch type if user doesn't specify it.
> If the function returns void, $retval is rejected as below;
>
> # echo 'f unregister_kprobes%return $retval' >> dynamic_events
> sh: write error: No such file or directory
> # cat error_log
> [ 37.488397] trace_fprobe: error: This function returns 'void' type
> Command: f unregister_kprobes%return $retval
> ^
>
> Signed-off-by: Masami Hiramatsu (Google) <mhiramat@kernel.org>
> ---
> Changes in v8:
> - Fix wrong indentation.
> Changes in v7:
> - Introduce this as a new patch.
> ---
> kernel/trace/trace_probe.c | 69 ++++++++++++++++++++++++++++++++++++++++----
> kernel/trace/trace_probe.h | 1 +
> 2 files changed, 63 insertions(+), 7 deletions(-)
>
> diff --git a/kernel/trace/trace_probe.c b/kernel/trace/trace_probe.c
> index 7318642aceb3..dfe3e1823eec 100644
> --- a/kernel/trace/trace_probe.c
> +++ b/kernel/trace/trace_probe.c
> @@ -371,15 +371,13 @@ static const char *type_from_btf_id(struct btf *btf, s32 id)
> return NULL;
> }
>
> -static const struct btf_param *find_btf_func_param(const char *funcname, s32 *nr,
> - bool tracepoint)
> +static const struct btf_type *find_btf_func_proto(const char *funcname)
> {
> struct btf *btf = traceprobe_get_btf();
> - const struct btf_param *param;
> const struct btf_type *t;
> s32 id;
>
> - if (!btf || !funcname || !nr)
> + if (!btf || !funcname)
> return ERR_PTR(-EINVAL);
>
> id = btf_find_by_name_kind(btf, funcname, BTF_KIND_FUNC);
> @@ -396,6 +394,22 @@ static const struct btf_param *find_btf_func_param(const char *funcname, s32 *nr
> if (!btf_type_is_func_proto(t))
> return ERR_PTR(-ENOENT);
>
> + return t;
> +}
> +
> +static const struct btf_param *find_btf_func_param(const char *funcname, s32 *nr,
> + bool tracepoint)
> +{
> + const struct btf_param *param;
> + const struct btf_type *t;
> +
> + if (!funcname || !nr)
> + return ERR_PTR(-EINVAL);
> +
> + t = find_btf_func_proto(funcname);
> + if (IS_ERR(t))
> + return (const struct btf_param *)t;
> +
> *nr = btf_type_vlen(t);
> param = (const struct btf_param *)(t + 1);
>
> @@ -462,6 +476,32 @@ static const struct fetch_type *parse_btf_arg_type(int arg_idx,
> return find_fetch_type(typestr, ctx->flags);
> }
>
> +static const struct fetch_type *parse_btf_retval_type(
> + struct traceprobe_parse_context *ctx)
> +{
Can we make this a common interface so that the funcgraph-retval can
also use it to get the return type?
-- Donglin Peng
> + struct btf *btf = traceprobe_get_btf();
> + const char *typestr = NULL;
> + const struct btf_type *t;
> +
> + if (btf && ctx->funcname) {
> + t = find_btf_func_proto(ctx->funcname);
> + if (!IS_ERR(t))
> + typestr = type_from_btf_id(btf, t->type);
> + }
> +
> + return find_fetch_type(typestr, ctx->flags);
> +}
> +
> +static bool is_btf_retval_void(const char *funcname)
> +{
> + const struct btf_type *t;
> +
> + t = find_btf_func_proto(funcname);
> + if (IS_ERR(t))
> + return false;
> +
> + return t->type == 0;
> +}
> #else
> static struct btf *traceprobe_get_btf(void)
> {
> @@ -480,8 +520,15 @@ static int parse_btf_arg(const char *varname, struct fetch_insn *code,
> trace_probe_log_err(ctx->offset, NOSUP_BTFARG);
> return -EOPNOTSUPP;
> }
> +
> #define parse_btf_arg_type(idx, ctx) \
> find_fetch_type(NULL, ctx->flags)
> +
> +#define parse_btf_retval_type(ctx) \
> + find_fetch_type(NULL, ctx->flags)
> +
> +#define is_btf_retval_void(funcname) (false)
> +
> #endif
>
> #define PARAM_MAX_STACK (THREAD_SIZE / sizeof(unsigned long))
> @@ -512,6 +559,11 @@ static int parse_probe_vars(char *arg, const struct fetch_type *t,
>
> if (strcmp(arg, "retval") == 0) {
> if (ctx->flags & TPARG_FL_RETURN) {
> + if ((ctx->flags & TPARG_FL_KERNEL) &&
> + is_btf_retval_void(ctx->funcname)) {
> + err = TP_ERR_NO_RETVAL;
> + goto inval;
> + }
> code->op = FETCH_OP_RETVAL;
> return 0;
> }
> @@ -912,9 +964,12 @@ static int traceprobe_parse_probe_arg_body(const char *argv, ssize_t *size,
> goto fail;
>
> /* Update storing type if BTF is available */
> - if (IS_ENABLED(CONFIG_PROBE_EVENTS_BTF_ARGS) &&
> - !t && code->op == FETCH_OP_ARG)
> - parg->type = parse_btf_arg_type(code->param, ctx);
> + if (IS_ENABLED(CONFIG_PROBE_EVENTS_BTF_ARGS) && !t) {
> + if (code->op == FETCH_OP_ARG)
> + parg->type = parse_btf_arg_type(code->param, ctx);
> + else if (code->op == FETCH_OP_RETVAL)
> + parg->type = parse_btf_retval_type(ctx);
> + }
>
> ret = -EINVAL;
> /* Store operation */
> diff --git a/kernel/trace/trace_probe.h b/kernel/trace/trace_probe.h
> index c864e6dea10f..eb43bea5c168 100644
> --- a/kernel/trace/trace_probe.h
> +++ b/kernel/trace/trace_probe.h
> @@ -449,6 +449,7 @@ extern int traceprobe_define_arg_fields(struct trace_event_call *event_call,
> C(BAD_EVENT_NAME, "Event name must follow the same rules as C identifiers"), \
> C(EVENT_EXIST, "Given group/event name is already used by another event"), \
> C(RETVAL_ON_PROBE, "$retval is not available on probe"), \
> + C(NO_RETVAL, "This function returns 'void' type"), \
> C(BAD_STACK_NUM, "Invalid stack number"), \
> C(BAD_ARG_NUM, "Invalid argument number"), \
> C(BAD_VAR, "Invalid $-valiable specified"), \
>
>
>
^ permalink raw reply [flat|nested] 17+ messages in thread* Re: [PATCH v13 09/12] tracing/probes: Add BTF retval type support
2023-06-12 7:29 ` Donglin Peng
@ 2023-07-07 14:12 ` Masami Hiramatsu
0 siblings, 0 replies; 17+ messages in thread
From: Masami Hiramatsu @ 2023-07-07 14:12 UTC (permalink / raw)
To: Donglin Peng
Cc: linux-kernel, Steven Rostedt, Florent Revest, Mark Rutland,
Will Deacon, Mathieu Desnoyers, Martin KaFai Lau, bpf,
Bagas Sanjaya, linux-trace-kernel, Ding Hui, huangcun
On Mon, 12 Jun 2023 15:29:00 +0800
Donglin Peng <pengdonglin@sangfor.com.cn> wrote:
> On 2023/5/26 12:19, Masami Hiramatsu (Google) wrote:
> > From: Masami Hiramatsu (Google) <mhiramat@kernel.org>
> >
> > Check the target function has non-void retval type and set the correct
> > fetch type if user doesn't specify it.
> > If the function returns void, $retval is rejected as below;
> >
> > # echo 'f unregister_kprobes%return $retval' >> dynamic_events
> > sh: write error: No such file or directory
> > # cat error_log
> > [ 37.488397] trace_fprobe: error: This function returns 'void' type
> > Command: f unregister_kprobes%return $retval
> > ^
> >
> > Signed-off-by: Masami Hiramatsu (Google) <mhiramat@kernel.org>
> > ---
> > Changes in v8:
> > - Fix wrong indentation.
> > Changes in v7:
> > - Introduce this as a new patch.
> > ---
> > kernel/trace/trace_probe.c | 69 ++++++++++++++++++++++++++++++++++++++++----
> > kernel/trace/trace_probe.h | 1 +
> > 2 files changed, 63 insertions(+), 7 deletions(-)
> >
> > diff --git a/kernel/trace/trace_probe.c b/kernel/trace/trace_probe.c
> > index 7318642aceb3..dfe3e1823eec 100644
> > --- a/kernel/trace/trace_probe.c
> > +++ b/kernel/trace/trace_probe.c
> > @@ -371,15 +371,13 @@ static const char *type_from_btf_id(struct btf *btf, s32 id)
> > return NULL;
> > }
> >
> > -static const struct btf_param *find_btf_func_param(const char *funcname, s32 *nr,
> > - bool tracepoint)
> > +static const struct btf_type *find_btf_func_proto(const char *funcname)
> > {
> > struct btf *btf = traceprobe_get_btf();
> > - const struct btf_param *param;
> > const struct btf_type *t;
> > s32 id;
> >
> > - if (!btf || !funcname || !nr)
> > + if (!btf || !funcname)
> > return ERR_PTR(-EINVAL);
> >
> > id = btf_find_by_name_kind(btf, funcname, BTF_KIND_FUNC);
> > @@ -396,6 +394,22 @@ static const struct btf_param *find_btf_func_param(const char *funcname, s32 *nr
> > if (!btf_type_is_func_proto(t))
> > return ERR_PTR(-ENOENT);
> >
> > + return t;
> > +}
> > +
> > +static const struct btf_param *find_btf_func_param(const char *funcname, s32 *nr,
> > + bool tracepoint)
> > +{
> > + const struct btf_param *param;
> > + const struct btf_type *t;
> > +
> > + if (!funcname || !nr)
> > + return ERR_PTR(-EINVAL);
> > +
> > + t = find_btf_func_proto(funcname);
> > + if (IS_ERR(t))
> > + return (const struct btf_param *)t;
> > +
> > *nr = btf_type_vlen(t);
> > param = (const struct btf_param *)(t + 1);
> >
> > @@ -462,6 +476,32 @@ static const struct fetch_type *parse_btf_arg_type(int arg_idx,
> > return find_fetch_type(typestr, ctx->flags);
> > }
> >
> > +static const struct fetch_type *parse_btf_retval_type(
> > + struct traceprobe_parse_context *ctx)
> > +{
>
> Can we make this a common interface so that the funcgraph-retval can
> also use it to get the return type?
It is possible to expose BTF part as an independent file so that
other ftrace tracers reuse it.
But you might need to store the result for each function somewhere
in the kernel. Would you have any idea?
Thank you,
>
> -- Donglin Peng
>
> > + struct btf *btf = traceprobe_get_btf();
> > + const char *typestr = NULL;
> > + const struct btf_type *t;
> > +
> > + if (btf && ctx->funcname) {
> > + t = find_btf_func_proto(ctx->funcname);
> > + if (!IS_ERR(t))
> > + typestr = type_from_btf_id(btf, t->type);
> > + }
> > +
> > + return find_fetch_type(typestr, ctx->flags);
> > +}
> > +
> > +static bool is_btf_retval_void(const char *funcname)
> > +{
> > + const struct btf_type *t;
> > +
> > + t = find_btf_func_proto(funcname);
> > + if (IS_ERR(t))
> > + return false;
> > +
> > + return t->type == 0;
> > +}
> > #else
> > static struct btf *traceprobe_get_btf(void)
> > {
> > @@ -480,8 +520,15 @@ static int parse_btf_arg(const char *varname, struct fetch_insn *code,
> > trace_probe_log_err(ctx->offset, NOSUP_BTFARG);
> > return -EOPNOTSUPP;
> > }
> > +
> > #define parse_btf_arg_type(idx, ctx) \
> > find_fetch_type(NULL, ctx->flags)
> > +
> > +#define parse_btf_retval_type(ctx) \
> > + find_fetch_type(NULL, ctx->flags)
> > +
> > +#define is_btf_retval_void(funcname) (false)
> > +
> > #endif
> >
> > #define PARAM_MAX_STACK (THREAD_SIZE / sizeof(unsigned long))
> > @@ -512,6 +559,11 @@ static int parse_probe_vars(char *arg, const struct fetch_type *t,
> >
> > if (strcmp(arg, "retval") == 0) {
> > if (ctx->flags & TPARG_FL_RETURN) {
> > + if ((ctx->flags & TPARG_FL_KERNEL) &&
> > + is_btf_retval_void(ctx->funcname)) {
> > + err = TP_ERR_NO_RETVAL;
> > + goto inval;
> > + }
> > code->op = FETCH_OP_RETVAL;
> > return 0;
> > }
> > @@ -912,9 +964,12 @@ static int traceprobe_parse_probe_arg_body(const char *argv, ssize_t *size,
> > goto fail;
> >
> > /* Update storing type if BTF is available */
> > - if (IS_ENABLED(CONFIG_PROBE_EVENTS_BTF_ARGS) &&
> > - !t && code->op == FETCH_OP_ARG)
> > - parg->type = parse_btf_arg_type(code->param, ctx);
> > + if (IS_ENABLED(CONFIG_PROBE_EVENTS_BTF_ARGS) && !t) {
> > + if (code->op == FETCH_OP_ARG)
> > + parg->type = parse_btf_arg_type(code->param, ctx);
> > + else if (code->op == FETCH_OP_RETVAL)
> > + parg->type = parse_btf_retval_type(ctx);
> > + }
> >
> > ret = -EINVAL;
> > /* Store operation */
> > diff --git a/kernel/trace/trace_probe.h b/kernel/trace/trace_probe.h
> > index c864e6dea10f..eb43bea5c168 100644
> > --- a/kernel/trace/trace_probe.h
> > +++ b/kernel/trace/trace_probe.h
> > @@ -449,6 +449,7 @@ extern int traceprobe_define_arg_fields(struct trace_event_call *event_call,
> > C(BAD_EVENT_NAME, "Event name must follow the same rules as C identifiers"), \
> > C(EVENT_EXIST, "Given group/event name is already used by another event"), \
> > C(RETVAL_ON_PROBE, "$retval is not available on probe"), \
> > + C(NO_RETVAL, "This function returns 'void' type"), \
> > C(BAD_STACK_NUM, "Invalid stack number"), \
> > C(BAD_ARG_NUM, "Invalid argument number"), \
> > C(BAD_VAR, "Invalid $-valiable specified"), \
> >
> >
> >
>
--
Masami Hiramatsu (Google) <mhiramat@kernel.org>
^ permalink raw reply [flat|nested] 17+ messages in thread
* [PATCH v13 10/12] selftests/ftrace: Add tracepoint probe test case
2023-05-26 4:17 [PATCH v13 00/12] tracing: Add fprobe/tracepoint events Masami Hiramatsu (Google)
` (8 preceding siblings ...)
2023-05-26 4:19 ` [PATCH v13 09/12] tracing/probes: Add BTF retval type support Masami Hiramatsu (Google)
@ 2023-05-26 4:19 ` Masami Hiramatsu (Google)
2023-05-26 4:19 ` [PATCH v13 11/12] selftests/ftrace: Add BTF arguments test cases Masami Hiramatsu (Google)
2023-05-26 4:19 ` [PATCH v13 12/12] Documentation: tracing/probes: Add fprobe event tracing document Masami Hiramatsu (Google)
11 siblings, 0 replies; 17+ messages in thread
From: Masami Hiramatsu (Google) @ 2023-05-26 4:19 UTC (permalink / raw)
To: linux-trace-kernel
Cc: linux-kernel, Steven Rostedt, mhiramat, Florent Revest,
Mark Rutland, Will Deacon, Mathieu Desnoyers, Martin KaFai Lau,
bpf, Bagas Sanjaya
From: Masami Hiramatsu (Google) <mhiramat@kernel.org>
Add test cases for tracepoint probe events.
Signed-off-by: Masami Hiramatsu (Google) <mhiramat@kernel.org>
---
Changes in v12:
- Fix to remove buggy space in requires line.
---
.../ftrace/test.d/dynevent/add_remove_fprobe.tc | 2
.../ftrace/test.d/dynevent/add_remove_tprobe.tc | 27 +++++++
.../ftrace/test.d/dynevent/fprobe_syntax_errors.tc | 2
.../ftrace/test.d/dynevent/tprobe_syntax_errors.tc | 82 ++++++++++++++++++++
4 files changed, 111 insertions(+), 2 deletions(-)
create mode 100644 tools/testing/selftests/ftrace/test.d/dynevent/add_remove_tprobe.tc
create mode 100644 tools/testing/selftests/ftrace/test.d/dynevent/tprobe_syntax_errors.tc
diff --git a/tools/testing/selftests/ftrace/test.d/dynevent/add_remove_fprobe.tc b/tools/testing/selftests/ftrace/test.d/dynevent/add_remove_fprobe.tc
index 53e0d5671687..dc25bcf4f9e2 100644
--- a/tools/testing/selftests/ftrace/test.d/dynevent/add_remove_fprobe.tc
+++ b/tools/testing/selftests/ftrace/test.d/dynevent/add_remove_fprobe.tc
@@ -1,7 +1,7 @@
#!/bin/sh
# SPDX-License-Identifier: GPL-2.0
# description: Generic dynamic event - add/remove fprobe events
-# requires: dynamic_events "f[:[<group>/][<event>]] <func-name>[%return] [<args>]": README
+# requires: dynamic_events "f[:[<group>/][<event>]] <func-name>[%return] [<args>]":README
echo 0 > events/enable
echo > dynamic_events
diff --git a/tools/testing/selftests/ftrace/test.d/dynevent/add_remove_tprobe.tc b/tools/testing/selftests/ftrace/test.d/dynevent/add_remove_tprobe.tc
new file mode 100644
index 000000000000..155792eaeee5
--- /dev/null
+++ b/tools/testing/selftests/ftrace/test.d/dynevent/add_remove_tprobe.tc
@@ -0,0 +1,27 @@
+#!/bin/sh
+# SPDX-License-Identifier: GPL-2.0
+# description: Generic dynamic event - add/remove tracepoint probe events
+# requires: dynamic_events "t[:[<group>/][<event>]] <tracepoint> [<args>]":README
+
+echo 0 > events/enable
+echo > dynamic_events
+
+TRACEPOINT1=kmem_cache_alloc
+TRACEPOINT2=kmem_cache_free
+
+echo "t:myevent1 $TRACEPOINT1" >> dynamic_events
+echo "t:myevent2 $TRACEPOINT2" >> dynamic_events
+
+grep -q myevent1 dynamic_events
+grep -q myevent2 dynamic_events
+test -d events/tracepoints/myevent1
+test -d events/tracepoints/myevent2
+
+echo "-:myevent2" >> dynamic_events
+
+grep -q myevent1 dynamic_events
+! grep -q myevent2 dynamic_events
+
+echo > dynamic_events
+
+clear_trace
diff --git a/tools/testing/selftests/ftrace/test.d/dynevent/fprobe_syntax_errors.tc b/tools/testing/selftests/ftrace/test.d/dynevent/fprobe_syntax_errors.tc
index 48e7a1a6aae3..4065a21efea1 100644
--- a/tools/testing/selftests/ftrace/test.d/dynevent/fprobe_syntax_errors.tc
+++ b/tools/testing/selftests/ftrace/test.d/dynevent/fprobe_syntax_errors.tc
@@ -1,7 +1,7 @@
#!/bin/sh
# SPDX-License-Identifier: GPL-2.0
# description: Fprobe event parser error log check
-# requires: dynamic_events "f[:[<group>/][<event>]] <func-name>[%return] [<args>]": README
+# requires: dynamic_events "f[:[<group>/][<event>]] <func-name>[%return] [<args>]":README
check_error() { # command-with-error-pos-by-^
ftrace_errlog_check 'trace_fprobe' "$1" 'dynamic_events'
diff --git a/tools/testing/selftests/ftrace/test.d/dynevent/tprobe_syntax_errors.tc b/tools/testing/selftests/ftrace/test.d/dynevent/tprobe_syntax_errors.tc
new file mode 100644
index 000000000000..da117b8f1d12
--- /dev/null
+++ b/tools/testing/selftests/ftrace/test.d/dynevent/tprobe_syntax_errors.tc
@@ -0,0 +1,82 @@
+#!/bin/sh
+# SPDX-License-Identifier: GPL-2.0
+# description: Tracepoint probe event parser error log check
+# requires: dynamic_events "t[:[<group>/][<event>]] <tracepoint> [<args>]":README
+
+check_error() { # command-with-error-pos-by-^
+ ftrace_errlog_check 'trace_fprobe' "$1" 'dynamic_events'
+}
+
+check_error 't^100 kfree' # BAD_MAXACT_TYPE
+
+check_error 't ^non_exist_tracepoint' # NO_TRACEPOINT
+check_error 't:^/bar kfree' # NO_GROUP_NAME
+check_error 't:^12345678901234567890123456789012345678901234567890123456789012345/bar kfree' # GROUP_TOO_LONG
+
+check_error 't:^foo.1/bar kfree' # BAD_GROUP_NAME
+check_error 't:^ kfree' # NO_EVENT_NAME
+check_error 't:foo/^12345678901234567890123456789012345678901234567890123456789012345 kfree' # EVENT_TOO_LONG
+check_error 't:foo/^bar.1 kfree' # BAD_EVENT_NAME
+
+check_error 't kfree ^$retval' # RETVAL_ON_PROBE
+check_error 't kfree ^$stack10000' # BAD_STACK_NUM
+
+check_error 't kfree ^$arg10000' # BAD_ARG_NUM
+
+check_error 't kfree ^$none_var' # BAD_VAR
+check_error 't kfree ^%rax' # BAD_VAR
+
+check_error 't kfree ^@12345678abcde' # BAD_MEM_ADDR
+check_error 't kfree ^@+10' # FILE_ON_KPROBE
+
+grep -q "imm-value" README && \
+check_error 't kfree arg1=\^x' # BAD_IMM
+grep -q "imm-string" README && \
+check_error 't kfree arg1=\"abcd^' # IMMSTR_NO_CLOSE
+
+check_error 't kfree ^+0@0)' # DEREF_NEED_BRACE
+check_error 't kfree ^+0ab1(@0)' # BAD_DEREF_OFFS
+check_error 't kfree +0(+0(@0^)' # DEREF_OPEN_BRACE
+
+if grep -A1 "fetcharg:" README | grep -q '\$comm' ; then
+check_error 't kfree +0(^$comm)' # COMM_CANT_DEREF
+fi
+
+check_error 't kfree ^&1' # BAD_FETCH_ARG
+
+
+# We've introduced this limitation with array support
+if grep -q ' <type>\\\[<array-size>\\\]' README; then
+check_error 't kfree +0(^+0(+0(+0(+0(+0(+0(+0(+0(+0(+0(+0(+0(+0(@0))))))))))))))' # TOO_MANY_OPS?
+check_error 't kfree +0(@11):u8[10^' # ARRAY_NO_CLOSE
+check_error 't kfree +0(@11):u8[10]^a' # BAD_ARRAY_SUFFIX
+check_error 't kfree +0(@11):u8[^10a]' # BAD_ARRAY_NUM
+check_error 't kfree +0(@11):u8[^256]' # ARRAY_TOO_BIG
+fi
+
+check_error 't kfree @11:^unknown_type' # BAD_TYPE
+check_error 't kfree $stack0:^string' # BAD_STRING
+check_error 't kfree @11:^b10@a/16' # BAD_BITFIELD
+
+check_error 't kfree ^arg123456789012345678901234567890=@11' # ARG_NAME_TOO_LOG
+check_error 't kfree ^=@11' # NO_ARG_NAME
+check_error 't kfree ^var.1=@11' # BAD_ARG_NAME
+check_error 't kfree var1=@11 ^var1=@12' # USED_ARG_NAME
+check_error 't kfree ^+1234567(+1234567(+1234567(+1234567(+1234567(+1234567(@1234))))))' # ARG_TOO_LONG
+check_error 't kfree arg1=^' # NO_ARG_BODY
+
+
+# multiprobe errors
+if grep -q "Create/append/" README && grep -q "imm-value" README; then
+echo "t:tracepoint/testevent kfree" > dynamic_events
+check_error '^f:tracepoint/testevent kfree' # DIFF_PROBE_TYPE
+
+# Explicitly use printf "%s" to not interpret \1
+printf "%s" "t:tracepoints/testevent kfree abcd=\\1" > dynamic_events
+check_error "t:tracepoints/testevent kfree ^bcd=\\1" # DIFF_ARG_TYPE
+check_error "t:tracepoints/testevent kfree ^abcd=\\1:u8" # DIFF_ARG_TYPE
+check_error "t:tracepoints/testevent kfree ^abcd=\\\"foo\"" # DIFF_ARG_TYPE
+check_error "^t:tracepoints/testevent kfree abcd=\\1" # SAME_PROBE
+fi
+
+exit 0
^ permalink raw reply related [flat|nested] 17+ messages in thread* [PATCH v13 11/12] selftests/ftrace: Add BTF arguments test cases
2023-05-26 4:17 [PATCH v13 00/12] tracing: Add fprobe/tracepoint events Masami Hiramatsu (Google)
` (9 preceding siblings ...)
2023-05-26 4:19 ` [PATCH v13 10/12] selftests/ftrace: Add tracepoint probe test case Masami Hiramatsu (Google)
@ 2023-05-26 4:19 ` Masami Hiramatsu (Google)
2023-05-26 4:19 ` [PATCH v13 12/12] Documentation: tracing/probes: Add fprobe event tracing document Masami Hiramatsu (Google)
11 siblings, 0 replies; 17+ messages in thread
From: Masami Hiramatsu (Google) @ 2023-05-26 4:19 UTC (permalink / raw)
To: linux-trace-kernel
Cc: linux-kernel, Steven Rostedt, mhiramat, Florent Revest,
Mark Rutland, Will Deacon, Mathieu Desnoyers, Martin KaFai Lau,
bpf, Bagas Sanjaya
From: Masami Hiramatsu (Google) <mhiramat@kernel.org>
Add test cases to check the BTF arguments correctly supported.
Signed-off-by: Masami Hiramatsu (Google) <mhiramat@kernel.org>
---
Changes in v13:
- Add a test case for only "$argN" argument case (no '$arg*')
Changes in v11:
- Change $args to $arg*
Changes in v10:
- Change $$args to $args
- Add new error patterns
- Fix non BTF case
Changes in v7:
- Add BTF void retval test case
---
.../ftrace/test.d/dynevent/add_remove_btfarg.tc | 58 ++++++++++++++++++++
.../ftrace/test.d/dynevent/fprobe_syntax_errors.tc | 14 +++++
.../ftrace/test.d/kprobe/kprobe_syntax_errors.tc | 14 +++++
3 files changed, 86 insertions(+)
create mode 100644 tools/testing/selftests/ftrace/test.d/dynevent/add_remove_btfarg.tc
diff --git a/tools/testing/selftests/ftrace/test.d/dynevent/add_remove_btfarg.tc b/tools/testing/selftests/ftrace/test.d/dynevent/add_remove_btfarg.tc
new file mode 100644
index 000000000000..b89de1771655
--- /dev/null
+++ b/tools/testing/selftests/ftrace/test.d/dynevent/add_remove_btfarg.tc
@@ -0,0 +1,58 @@
+#!/bin/sh
+# SPDX-License-Identifier: GPL-2.0
+# description: Generic dynamic event - add/remove probes with BTF arguments
+# requires: dynamic_events "<argname>":README
+
+KPROBES=
+FPROBES=
+
+if grep -qF "p[:[<group>/][<event>]] <place> [<args>]" README ; then
+ KPROBES=yes
+fi
+if grep -qF "f[:[<group>/][<event>]] <func-name>[%return] [<args>]" README ; then
+ FPROBES=yes
+fi
+
+if [ -z "$KPROBES" -a "$FPROBES" ] ; then
+ exit_unsupported
+fi
+
+echo 0 > events/enable
+echo > dynamic_events
+
+TP=kfree
+
+if [ "$FPROBES" ] ; then
+echo "f:fpevent $TP object" >> dynamic_events
+echo "t:tpevent $TP ptr" >> dynamic_events
+
+grep -q "fpevent.*object=object" dynamic_events
+grep -q "tpevent.*ptr=ptr" dynamic_events
+
+echo > dynamic_events
+
+echo "f:fpevent $TP "'$arg1' >> dynamic_events
+grep -q "fpevent.*object=object" dynamic_events
+echo > dynamic_events
+
+echo "f:fpevent $TP "'$arg*' >> dynamic_events
+echo "t:tpevent $TP "'$arg*' >> dynamic_events
+
+grep -q "fpevent.*object=object" dynamic_events
+grep -q "tpevent.*ptr=ptr" dynamic_events
+! grep -q "tpevent.*_data" dynamic_events
+fi
+
+echo > dynamic_events
+
+if [ "$KPROBES" ] ; then
+echo "p:kpevent $TP object" >> dynamic_events
+grep -q "kpevent.*object=object" dynamic_events
+
+echo > dynamic_events
+
+echo "p:kpevent $TP "'$arg*' >> dynamic_events
+grep -q "kpevent.*object=object" dynamic_events
+fi
+
+clear_trace
diff --git a/tools/testing/selftests/ftrace/test.d/dynevent/fprobe_syntax_errors.tc b/tools/testing/selftests/ftrace/test.d/dynevent/fprobe_syntax_errors.tc
index 4065a21efea1..812f5b3f6055 100644
--- a/tools/testing/selftests/ftrace/test.d/dynevent/fprobe_syntax_errors.tc
+++ b/tools/testing/selftests/ftrace/test.d/dynevent/fprobe_syntax_errors.tc
@@ -94,4 +94,18 @@ fi
# %return suffix errors
check_error 'f vfs_read^%hoge' # BAD_ADDR_SUFFIX
+# BTF arguments errors
+if grep -q "<argname>" README; then
+check_error 'f vfs_read args=^$arg*' # BAD_VAR_ARGS
+check_error 'f vfs_read +0(^$arg*)' # BAD_VAR_ARGS
+check_error 'f vfs_read $arg* ^$arg*' # DOUBLE_ARGS
+check_error 'f vfs_read%return ^$arg*' # NOFENTRY_ARGS
+check_error 'f vfs_read ^hoge' # NO_BTFARG
+check_error 'f kfree ^$arg10' # NO_BTFARG (exceed the number of parameters)
+check_error 'f kfree%return ^$retval' # NO_RETVAL
+else
+check_error 'f vfs_read ^$arg*' # NOSUP_BTFARG
+check_error 't kfree ^$arg*' # NOSUP_BTFARG
+fi
+
exit 0
diff --git a/tools/testing/selftests/ftrace/test.d/kprobe/kprobe_syntax_errors.tc b/tools/testing/selftests/ftrace/test.d/kprobe/kprobe_syntax_errors.tc
index 97c08867490a..65fbb26fd58c 100644
--- a/tools/testing/selftests/ftrace/test.d/kprobe/kprobe_syntax_errors.tc
+++ b/tools/testing/selftests/ftrace/test.d/kprobe/kprobe_syntax_errors.tc
@@ -103,4 +103,18 @@ check_error 'p vfs_read^%hoge' # BAD_ADDR_SUFFIX
check_error 'p ^vfs_read+10%return' # BAD_RETPROBE
fi
+# BTF arguments errors
+if grep -q "<argname>" README; then
+check_error 'p vfs_read args=^$arg*' # BAD_VAR_ARGS
+check_error 'p vfs_read +0(^$arg*)' # BAD_VAR_ARGS
+check_error 'p vfs_read $arg* ^$arg*' # DOUBLE_ARGS
+check_error 'r vfs_read ^$arg*' # NOFENTRY_ARGS
+check_error 'p vfs_read+8 ^$arg*' # NOFENTRY_ARGS
+check_error 'p vfs_read ^hoge' # NO_BTFARG
+check_error 'p kfree ^$arg10' # NO_BTFARG (exceed the number of parameters)
+check_error 'r kfree ^$retval' # NO_RETVAL
+else
+check_error 'p vfs_read ^$arg*' # NOSUP_BTFARG
+fi
+
exit 0
^ permalink raw reply related [flat|nested] 17+ messages in thread* [PATCH v13 12/12] Documentation: tracing/probes: Add fprobe event tracing document
2023-05-26 4:17 [PATCH v13 00/12] tracing: Add fprobe/tracepoint events Masami Hiramatsu (Google)
` (10 preceding siblings ...)
2023-05-26 4:19 ` [PATCH v13 11/12] selftests/ftrace: Add BTF arguments test cases Masami Hiramatsu (Google)
@ 2023-05-26 4:19 ` Masami Hiramatsu (Google)
11 siblings, 0 replies; 17+ messages in thread
From: Masami Hiramatsu (Google) @ 2023-05-26 4:19 UTC (permalink / raw)
To: linux-trace-kernel
Cc: linux-kernel, Steven Rostedt, mhiramat, Florent Revest,
Mark Rutland, Will Deacon, Mathieu Desnoyers, Martin KaFai Lau,
bpf, Bagas Sanjaya
From: Masami Hiramatsu (Google) <mhiramat@kernel.org>
Add a documentation about fprobe event tracing including
tracepoint probe event and BTF argument.
Signed-off-by: Masami Hiramatsu (Google) <mhiramat@kernel.org>
Reviewed-by: Bagas Sanjaya <bagasdotme@gmail.com>
---
Changes in v12:
- Use double backquotes for keywords and identifiers.
- Use the ref link instead of file.
Changes in v11:
- Fix $$args to $arg*.
- Update the document.
Changes in v10:
- Fix space before tab.
Changes in v7:
- Update about BTF auto type casting for $retval.
---
Documentation/trace/fprobetrace.rst | 188 +++++++++++++++++++++++++++++++++++
Documentation/trace/index.rst | 1
Documentation/trace/kprobetrace.rst | 2
3 files changed, 191 insertions(+)
create mode 100644 Documentation/trace/fprobetrace.rst
diff --git a/Documentation/trace/fprobetrace.rst b/Documentation/trace/fprobetrace.rst
new file mode 100644
index 000000000000..e949bc0cff05
--- /dev/null
+++ b/Documentation/trace/fprobetrace.rst
@@ -0,0 +1,188 @@
+.. SPDX-License-Identifier: GPL-2.0
+
+==========================
+Fprobe-based Event Tracing
+==========================
+
+.. Author: Masami Hiramatsu <mhiramat@kernel.org>
+
+Overview
+--------
+
+Fprobe event is similar to the kprobe event, but limited to probe on
+the function entry and exit only. It is good enough for many use cases
+which only traces some specific functions.
+
+This document also covers tracepoint probe events (tprobe) since this
+is also works only on the tracepoint entry. User can trace a part of
+tracepoint argument, or the tracepoint without trace-event, which is
+not exposed on tracefs.
+
+As same as other dynamic events, fprobe events and tracepoint probe
+events are defined via `dynamic_events` interface file on tracefs.
+
+Synopsis of fprobe-events
+-------------------------
+::
+
+ f[:[GRP1/][EVENT1]] SYM [FETCHARGS] : Probe on function entry
+ f[MAXACTIVE][:[GRP1/][EVENT1]] SYM%return [FETCHARGS] : Probe on function exit
+ t[:[GRP2/][EVENT2]] TRACEPOINT [FETCHARGS] : Probe on tracepoint
+
+ GRP1 : Group name for fprobe. If omitted, use "fprobes" for it.
+ GRP2 : Group name for tprobe. If omitted, use "tracepoints" for it.
+ EVENT1 : Event name for fprobe. If omitted, the event name is
+ "SYM__entry" or "SYM__exit".
+ EVENT2 : Event name for tprobe. If omitted, the event name is
+ the same as "TRACEPOINT", but if the "TRACEPOINT" starts
+ with a digit character, "_TRACEPOINT" is used.
+ MAXACTIVE : Maximum number of instances of the specified function that
+ can be probed simultaneously, or 0 for the default value
+ as defined in Documentation/trace/fprobes.rst
+
+ FETCHARGS : Arguments. Each probe can have up to 128 args.
+ ARG : Fetch "ARG" function argument using BTF (only for function
+ entry or tracepoint.) (\*1)
+ @ADDR : Fetch memory at ADDR (ADDR should be in kernel)
+ @SYM[+|-offs] : Fetch memory at SYM +|- offs (SYM should be a data symbol)
+ $stackN : Fetch Nth entry of stack (N >= 0)
+ $stack : Fetch stack address.
+ $argN : Fetch the Nth function argument. (N >= 1) (\*2)
+ $retval : Fetch return value.(\*3)
+ $comm : Fetch current task comm.
+ +|-[u]OFFS(FETCHARG) : Fetch memory at FETCHARG +|- OFFS address.(\*4)(\*5)
+ \IMM : Store an immediate value to the argument.
+ NAME=FETCHARG : Set NAME as the argument name of FETCHARG.
+ FETCHARG:TYPE : Set TYPE as the type of FETCHARG. Currently, basic types
+ (u8/u16/u32/u64/s8/s16/s32/s64), hexadecimal types
+ (x8/x16/x32/x64), "char", "string", "ustring", "symbol", "symstr"
+ and bitfield are supported.
+
+ (\*1) This is available only when BTF is enabled.
+ (\*2) only for the probe on function entry (offs == 0).
+ (\*3) only for return probe.
+ (\*4) this is useful for fetching a field of data structures.
+ (\*5) "u" means user-space dereference.
+
+For the details of TYPE, see :ref:`kprobetrace documentation <kprobetrace_types>`.
+
+BTF arguments
+-------------
+BTF (BPF Type Format) argument allows user to trace function and tracepoint
+parameters by its name instead of ``$argN``. This feature is available if the
+kernel is configured with CONFIG_BPF_SYSCALL and CONFIG_DEBUG_INFO_BTF.
+If user only specify the BTF argument, the event's argument name is also
+automatically set by the given name. ::
+
+ # echo 'f:myprobe vfs_read count pos' >> dynamic_events
+ # cat dynamic_events
+ f:fprobes/myprobe vfs_read count=count pos=pos
+
+It also chooses the fetch type from BTF information. For example, in the above
+example, the ``count`` is unsigned long, and the ``pos`` is a pointer. Thus, both
+are converted to 64bit unsigned long, but only ``pos`` has "%Lx" print-format as
+below ::
+
+ # cat events/fprobes/myprobe/format
+ name: myprobe
+ ID: 1313
+ 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 long __probe_ip; offset:8; size:8; signed:0;
+ field:u64 count; offset:16; size:8; signed:0;
+ field:u64 pos; offset:24; size:8; signed:0;
+
+ print fmt: "(%lx) count=%Lu pos=0x%Lx", REC->__probe_ip, REC->count, REC->pos
+
+If user unsures the name of arguments, ``$arg*`` will be helpful. The ``$arg*``
+is expanded to all function arguments of the function or the tracepoint. ::
+
+ # echo 'f:myprobe vfs_read $arg*' >> dynamic_events
+ # cat dynamic_events
+ f:fprobes/myprobe vfs_read file=file buf=buf count=count pos=pos
+
+BTF also affects the ``$retval``. If user doesn't set any type, the retval type is
+automatically picked from the BTF. If the function returns ``void``, ``$retval``
+is rejected.
+
+Usage examples
+--------------
+Here is an example to add fprobe events on ``vfs_read()`` function entry
+and exit, with BTF arguments.
+::
+
+ # echo 'f vfs_read $arg*' >> dynamic_events
+ # echo 'f vfs_read%return $retval' >> dynamic_events
+ # cat dynamic_events
+ f:fprobes/vfs_read__entry vfs_read file=file buf=buf count=count pos=pos
+ f:fprobes/vfs_read__exit vfs_read%return arg1=$retval
+ # echo 1 > events/fprobes/enable
+ # head -n 20 trace | tail
+ # TASK-PID CPU# ||||| TIMESTAMP FUNCTION
+ # | | | ||||| | |
+ sh-70 [000] ...1. 335.883195: vfs_read__entry: (vfs_read+0x4/0x340) file=0xffff888005cf9a80 buf=0x7ffef36c6879 count=1 pos=0xffffc900005aff08
+ sh-70 [000] ..... 335.883208: vfs_read__exit: (ksys_read+0x75/0x100 <- vfs_read) arg1=1
+ sh-70 [000] ...1. 335.883220: vfs_read__entry: (vfs_read+0x4/0x340) file=0xffff888005cf9a80 buf=0x7ffef36c6879 count=1 pos=0xffffc900005aff08
+ sh-70 [000] ..... 335.883224: vfs_read__exit: (ksys_read+0x75/0x100 <- vfs_read) arg1=1
+ sh-70 [000] ...1. 335.883232: vfs_read__entry: (vfs_read+0x4/0x340) file=0xffff888005cf9a80 buf=0x7ffef36c687a count=1 pos=0xffffc900005aff08
+ sh-70 [000] ..... 335.883237: vfs_read__exit: (ksys_read+0x75/0x100 <- vfs_read) arg1=1
+ sh-70 [000] ...1. 336.050329: vfs_read__entry: (vfs_read+0x4/0x340) file=0xffff888005cf9a80 buf=0x7ffef36c6879 count=1 pos=0xffffc900005aff08
+ sh-70 [000] ..... 336.050343: vfs_read__exit: (ksys_read+0x75/0x100 <- vfs_read) arg1=1
+
+You can see all function arguments and return values are recorded as signed int.
+
+Also, here is an example of tracepoint events on ``sched_switch`` tracepoint.
+To compare the result, this also enables the ``sched_switch`` traceevent too.
+::
+
+ # echo 't sched_switch $arg*' >> dynamic_events
+ # echo 1 > events/sched/sched_switch/enable
+ # echo 1 > events/tracepoints/sched_switch/enable
+ # echo > trace
+ # head -n 20 trace | tail
+ # TASK-PID CPU# ||||| TIMESTAMP FUNCTION
+ # | | | ||||| | |
+ sh-70 [000] d..2. 3912.083993: sched_switch: prev_comm=sh prev_pid=70 prev_prio=120 prev_state=S ==> next_comm=swapper/0 next_pid=0 next_prio=120
+ sh-70 [000] d..3. 3912.083995: sched_switch: (__probestub_sched_switch+0x4/0x10) preempt=0 prev=0xffff88800664e100 next=0xffffffff828229c0 prev_state=1
+ <idle>-0 [000] d..2. 3912.084183: sched_switch: prev_comm=swapper/0 prev_pid=0 prev_prio=120 prev_state=R ==> next_comm=rcu_preempt next_pid=16 next_prio=120
+ <idle>-0 [000] d..3. 3912.084184: sched_switch: (__probestub_sched_switch+0x4/0x10) preempt=0 prev=0xffffffff828229c0 next=0xffff888004208000 prev_state=0
+ rcu_preempt-16 [000] d..2. 3912.084196: sched_switch: prev_comm=rcu_preempt prev_pid=16 prev_prio=120 prev_state=I ==> next_comm=swapper/0 next_pid=0 next_prio=120
+ rcu_preempt-16 [000] d..3. 3912.084196: sched_switch: (__probestub_sched_switch+0x4/0x10) preempt=0 prev=0xffff888004208000 next=0xffffffff828229c0 prev_state=1026
+ <idle>-0 [000] d..2. 3912.085191: sched_switch: prev_comm=swapper/0 prev_pid=0 prev_prio=120 prev_state=R ==> next_comm=rcu_preempt next_pid=16 next_prio=120
+ <idle>-0 [000] d..3. 3912.085191: sched_switch: (__probestub_sched_switch+0x4/0x10) preempt=0 prev=0xffffffff828229c0 next=0xffff888004208000 prev_state=0
+
+As you can see, the ``sched_switch`` trace-event shows *cooked* parameters, on
+the other hand, the ``sched_switch`` tracepoint probe event shows *raw*
+parameters. This means you can access any field values in the task
+structure pointed by the ``prev`` and ``next`` arguments.
+
+For example, usually ``task_struct::start_time`` is not traced, but with this
+traceprobe event, you can trace it as below.
+::
+
+ # echo 't sched_switch comm=+1896(next):string start_time=+1728(next):u64' > dynamic_events
+ # head -n 20 trace | tail
+ # TASK-PID CPU# ||||| TIMESTAMP FUNCTION
+ # | | | ||||| | |
+ sh-70 [000] d..3. 5606.686577: sched_switch: (__probestub_sched_switch+0x4/0x10) comm="rcu_preempt" usage=1 start_time=245000000
+ rcu_preempt-16 [000] d..3. 5606.686602: sched_switch: (__probestub_sched_switch+0x4/0x10) comm="sh" usage=1 start_time=1596095526
+ sh-70 [000] d..3. 5606.686637: sched_switch: (__probestub_sched_switch+0x4/0x10) comm="swapper/0" usage=2 start_time=0
+ <idle>-0 [000] d..3. 5606.687190: sched_switch: (__probestub_sched_switch+0x4/0x10) comm="rcu_preempt" usage=1 start_time=245000000
+ rcu_preempt-16 [000] d..3. 5606.687202: sched_switch: (__probestub_sched_switch+0x4/0x10) comm="swapper/0" usage=2 start_time=0
+ <idle>-0 [000] d..3. 5606.690317: sched_switch: (__probestub_sched_switch+0x4/0x10) comm="kworker/0:1" usage=1 start_time=137000000
+ kworker/0:1-14 [000] d..3. 5606.690339: sched_switch: (__probestub_sched_switch+0x4/0x10) comm="swapper/0" usage=2 start_time=0
+ <idle>-0 [000] d..3. 5606.692368: sched_switch: (__probestub_sched_switch+0x4/0x10) comm="kworker/0:1" usage=1 start_time=137000000
+
+Currently, to find the offset of a specific field in the data structure,
+you need to build kernel with debuginfo and run `perf probe` command with
+`-D` option. e.g.
+::
+
+ # perf probe -D "__probestub_sched_switch next->comm:string next->start_time"
+ p:probe/__probestub_sched_switch __probestub_sched_switch+0 comm=+1896(%cx):string start_time=+1728(%cx):u64
+
+And replace the ``%cx`` with the ``next``.
diff --git a/Documentation/trace/index.rst b/Documentation/trace/index.rst
index ea25a9220f92..5092d6c13af5 100644
--- a/Documentation/trace/index.rst
+++ b/Documentation/trace/index.rst
@@ -13,6 +13,7 @@ Linux Tracing Technologies
kprobes
kprobetrace
uprobetracer
+ fprobetrace
tracepoints
events
events-kmem
diff --git a/Documentation/trace/kprobetrace.rst b/Documentation/trace/kprobetrace.rst
index 651f9ab53f3e..8a2dfee38145 100644
--- a/Documentation/trace/kprobetrace.rst
+++ b/Documentation/trace/kprobetrace.rst
@@ -66,6 +66,8 @@ Synopsis of kprobe_events
(\*3) this is useful for fetching a field of data structures.
(\*4) "u" means user-space dereference. See :ref:`user_mem_access`.
+.. _kprobetrace_types:
+
Types
-----
Several types are supported for fetchargs. Kprobe tracer will access memory
^ permalink raw reply related [flat|nested] 17+ messages in thread