diff mbox series

[v2,1/2] trace-cmd ftrace: print function retval in function_graph

Message ID 20240521213730.67993-2-jianfeng.w.wang@oracle.com (mailing list archive)
State Superseded
Headers show
Series trace-cmd ftrace: support function retval feature in function_graph | expand

Commit Message

Jianfeng Wang May 21, 2024, 9:37 p.m. UTC
The upstream Linux kernel has introduced a new feature, i.e.
funcgraph-retval for the function_graph tracer. (Commit ID:
a1be9ccc57f07d54278be34eed6bd679bc941c97). With this feature, the
function_graph tracer can record each function's return value along
with its execution time at the function_graph exit handler. This is
useful for debugging kernel issues, such as syscall errors.

This commit enhances 'trace-cmd report' by supporting the
function_graph retval feature if the underlying kernel supports it.
Once the feature is supported, 'trace-cmd report' will print each
function's return value in this format: '(ret=retval)" at the end
of every funcgraph leaf entry and exit event.

Example usage:
*  Record
> trace-cmd record -p function_graph -P <pid>

The recording process is not changed: trace-cmd can communicate
with the kernel to get the actual funcgraph_exit event format.
If the kernel supports ftrace-retval, trace-cmd will use the
updated event format that has "retval" field.

* Report
> trace-cmd report -O fgraph:depth -O fgraph:tailprint

If the kernel supports this feature, the output looks like this:

 <...>-1136215 [006] ..... 1724848.377495: funcgraph_entry:                   |        __cond_resched() { (3)
 <...>-1136215 [006] ..... 1724848.377495: funcgraph_entry:        0.200 us   |          rcu_all_qs(); (4) (ret=0x0)
 <...>-1136215 [006] ..... 1724848.377496: funcgraph_exit:         0.601 us   |        } /* __cond_resched */ (3) (ret=0x0)
 <...>-1136215 [006] ..... 1724848.377496: funcgraph_entry:        0.581 us   |        proc_reg_write(); (3) (ret=-5)
 <...>-1136215 [006] ..... 1724848.377497: funcgraph_exit:         5.059 us   |      } /* vfs_write */ (2) (ret=-5)
 <...>-1136215 [006] ..... 1724848.377497: funcgraph_exit:         6.332 us   |    } /* ksys_write */ (1) (ret=-5)
 <...>-1136215 [006] ..... 1724848.377497: funcgraph_exit:       + 18.414 us  |  } /* __x64_sys_write */ (0) (ret=-5)

Signed-off-by: Jianfeng Wang <jianfeng.w.wang@oracle.com>
---
 lib/trace-cmd/trace-ftrace.c | 37 ++++++++++++++++++++++++++++++++++++
 1 file changed, 37 insertions(+)

Comments

Steven Rostedt July 17, 2024, 6:48 p.m. UTC | #1
Sorry for the late reply, I'm finally able to look at userspace tooling.

First little nit, please start the subject with a capital letter:

  trace-cmd report: Print function retval in function_graph

Also note that it's for the report command.

On Tue, 21 May 2024 14:37:29 -0700
Jianfeng Wang <jianfeng.w.wang@oracle.com> wrote:

> The upstream Linux kernel has introduced a new feature, i.e.
> funcgraph-retval for the function_graph tracer. (Commit ID:
> a1be9ccc57f07d54278be34eed6bd679bc941c97). With this feature, the
> function_graph tracer can record each function's return value along
> with its execution time at the function_graph exit handler. This is
> useful for debugging kernel issues, such as syscall errors.
> 
> This commit enhances 'trace-cmd report' by supporting the
> function_graph retval feature if the underlying kernel supports it.
> Once the feature is supported, 'trace-cmd report' will print each
> function's return value in this format: '(ret=retval)" at the end
> of every funcgraph leaf entry and exit event.
> 
> Example usage:
> *  Record
> > trace-cmd record -p function_graph -P <pid>  
> 
> The recording process is not changed: trace-cmd can communicate
> with the kernel to get the actual funcgraph_exit event format.
> If the kernel supports ftrace-retval, trace-cmd will use the
> updated event format that has "retval" field.
> 
> * Report
> > trace-cmd report -O fgraph:depth -O fgraph:tailprint  
> 
> If the kernel supports this feature, the output looks like this:
> 
>  <...>-1136215 [006] ..... 1724848.377495: funcgraph_entry:                   |        __cond_resched() { (3)
>  <...>-1136215 [006] ..... 1724848.377495: funcgraph_entry:        0.200 us   |          rcu_all_qs(); (4) (ret=0x0)
>  <...>-1136215 [006] ..... 1724848.377496: funcgraph_exit:         0.601 us   |        } /* __cond_resched */ (3) (ret=0x0)
>  <...>-1136215 [006] ..... 1724848.377496: funcgraph_entry:        0.581 us   |        proc_reg_write(); (3) (ret=-5)
>  <...>-1136215 [006] ..... 1724848.377497: funcgraph_exit:         5.059 us   |      } /* vfs_write */ (2) (ret=-5)
>  <...>-1136215 [006] ..... 1724848.377497: funcgraph_exit:         6.332 us   |    } /* ksys_write */ (1) (ret=-5)
>  <...>-1136215 [006] ..... 1724848.377497: funcgraph_exit:       + 18.414 us  |  } /* __x64_sys_write */ (0) (ret=-5)
> 
> Signed-off-by: Jianfeng Wang <jianfeng.w.wang@oracle.com>
> ---
>  lib/trace-cmd/trace-ftrace.c | 37 ++++++++++++++++++++++++++++++++++++
>  1 file changed, 37 insertions(+)
> 
> diff --git a/lib/trace-cmd/trace-ftrace.c b/lib/trace-cmd/trace-ftrace.c
> index 89e46c3d..2b4e49a5 100644
> --- a/lib/trace-cmd/trace-ftrace.c
> +++ b/lib/trace-cmd/trace-ftrace.c
> @@ -10,6 +10,9 @@
>  
>  #include "trace-cmd-private.h"
>  
> +#define MAX_LINUX_ERRNO   4095
> +#define IS_LINUX_ERR_VALUE(x) ((unsigned long long)(void *)(x) >= (unsigned long long)-MAX_LINUX_ERRNO)
> +
>  struct tep_plugin_option trace_ftrace_options[] = {
>  	{
>  		.name = "tailprint",
> @@ -185,6 +188,8 @@ print_graph_entry_leaf(struct trace_seq *s,
>  	unsigned long long rettime, calltime;
>  	unsigned long long duration, depth;
>  	unsigned long long val;
> +	unsigned long long retval;
> +	bool fgraph_retval_supported = true;
>  	const char *func;
>  	int ret;
>  	int i;
> @@ -195,6 +200,13 @@ print_graph_entry_leaf(struct trace_seq *s,
>  	if (tep_get_field_val(s, finfo->fgraph_ret_event, "calltime", ret_rec, &calltime, 1))
>  		return trace_seq_putc(s, '!');
>  
> +	if (!tep_find_field(finfo->fgraph_ret_event, "retval"))
> +		fgraph_retval_supported = false;
> +	else {

This should follow Linux kernel syntax rules. That is, if one block of
a if statement has brackets, then the other should too.

Thanks,

-- Steve

> +		if (tep_get_field_val(s, finfo->fgraph_ret_event, "retval", ret_rec, &retval, 1))
> +			return trace_seq_putc(s, '!');
> +	}
> +
>  	duration = rettime - calltime;
diff mbox series

Patch

diff --git a/lib/trace-cmd/trace-ftrace.c b/lib/trace-cmd/trace-ftrace.c
index 89e46c3d..2b4e49a5 100644
--- a/lib/trace-cmd/trace-ftrace.c
+++ b/lib/trace-cmd/trace-ftrace.c
@@ -10,6 +10,9 @@ 
 
 #include "trace-cmd-private.h"
 
+#define MAX_LINUX_ERRNO   4095
+#define IS_LINUX_ERR_VALUE(x) ((unsigned long long)(void *)(x) >= (unsigned long long)-MAX_LINUX_ERRNO)
+
 struct tep_plugin_option trace_ftrace_options[] = {
 	{
 		.name = "tailprint",
@@ -185,6 +188,8 @@  print_graph_entry_leaf(struct trace_seq *s,
 	unsigned long long rettime, calltime;
 	unsigned long long duration, depth;
 	unsigned long long val;
+	unsigned long long retval;
+	bool fgraph_retval_supported = true;
 	const char *func;
 	int ret;
 	int i;
@@ -195,6 +200,13 @@  print_graph_entry_leaf(struct trace_seq *s,
 	if (tep_get_field_val(s, finfo->fgraph_ret_event, "calltime", ret_rec, &calltime, 1))
 		return trace_seq_putc(s, '!');
 
+	if (!tep_find_field(finfo->fgraph_ret_event, "retval"))
+		fgraph_retval_supported = false;
+	else {
+		if (tep_get_field_val(s, finfo->fgraph_ret_event, "retval", ret_rec, &retval, 1))
+			return trace_seq_putc(s, '!');
+	}
+
 	duration = rettime - calltime;
 
 	/* Overhead */
@@ -222,6 +234,14 @@  print_graph_entry_leaf(struct trace_seq *s,
 	if (ret && fgraph_depth->set)
 		ret = trace_seq_printf(s, " (%lld)", depth);
 
+	/* Return Value */
+	if (ret && fgraph_retval_supported) {
+		if (!IS_LINUX_ERR_VALUE(retval))
+			ret = trace_seq_printf(s, " (ret=0x%llx)", retval);
+		else
+			ret = trace_seq_printf(s, " (ret=%lld)", retval);
+	}
+
 	return ret;
 }
 
@@ -316,6 +336,8 @@  fgraph_ret_handler(struct trace_seq *s, struct tep_record *record,
 	unsigned long long duration, depth;
 	unsigned long long val;
 	const char *func;
+	unsigned long long retval;
+	bool fgraph_retval_supported = true;
 	int i;
 
 	ret_event_check(finfo, event->tep);
@@ -326,6 +348,13 @@  fgraph_ret_handler(struct trace_seq *s, struct tep_record *record,
 	if (tep_get_field_val(s, event, "calltime", record, &calltime, 1))
 		return trace_seq_putc(s, '!');
 
+	if (!tep_find_field(event, "retval"))
+		fgraph_retval_supported = false;
+	else {
+		if (tep_get_field_val(s, event, "retval", record, &retval, 1))
+			return trace_seq_putc(s, '!');
+	}
+
 	duration = rettime - calltime;
 
 	/* Overhead */
@@ -355,6 +384,14 @@  fgraph_ret_handler(struct trace_seq *s, struct tep_record *record,
 	if (fgraph_depth->set)
 		trace_seq_printf(s, " (%lld)", depth);
 
+	/* Return Value */
+	if (fgraph_retval_supported) {
+		if (!IS_LINUX_ERR_VALUE(retval))
+			trace_seq_printf(s, " (ret=0x%llx)", retval);
+		else
+			trace_seq_printf(s, " (ret=%lld)", retval);
+	}
+
 	return 0;
 }