Return EADDRNOTAVAIL when func matches several symbols during kprobe creation

RFC PATCH Superseded Francis Laniel 2 participants 6 messages Aug 24, 2023 Tracing

lore

Share & tools

lore

Francis Laniel

author

2023-08-24 10:38 UTC #0 lore raw

v2 cover Return EADDRNOTAVAIL when func matches several symbols during kprobe creation

Hi.

In the kernel source code, it exists different functions which share the same name but which have, of course, different addresses as they can be defined in different modules:
# Kernel was compiled with CONFIG_NTFS_FS and CONFIG_NTFS3_FS as built-in.

root@vm-amd64:~# grep ntfs_file_write_iter /proc/kallsyms
ffffffff814ce3c0 t __pfx_ntfs_file_write_iter
ffffffff814ce3d0 t ntfs_file_write_iter
ffffffff814fc8a0 t __pfx_ntfs_file_write_iter
ffffffff814fc8b0 t ntfs_file_write_iter
This can be source of troubles when you create a PMU kprobe for such a function,
as it will only install one for the first address (e.g. 0xffffffff814ce3d0 in
the above).
This could lead to some troubles were BPF based tools does not report any event
because the second function is not called:
root@vm-amd64:/mnt# mount | grep /mnt
/foo.img on /mnt type ntfs3 (rw,relatime,uid=0,gid=0,iocharset=utf8)
# ig is a tool which installs a PMU kprobe on ntfs_file_write_iter().
root@vm-amd64:/mnt# ig trace fsslower -m 0 -f ntfs3 --host &> /tmp/foo &
[1] 207
root@vm-amd64:/mnt# dd if=./foo of=./bar count=3
3+0 records in
3+0 records out
1536 bytes (1.5 kB, 1.5 KiB) copied, 0.00543323 s, 283 kB/s
root@vm-amd64:/mnt# fg
ig trace fsslower -m 0 -f ntfs3 --host &> /tmp/foo
^Croot@vm-amd64:/mnt# more /tmp/foo
RUNTIME.CONTAINERNAME          RUNTIME.CONTAIN… PID              COMM
  T      BYTES     OFFSET        LAT FILE
                                                214              dd
  R        512          0        766 foo
                                                214              dd
  R        512        512          9 foo
                                                214              dd
As you can see in the above, only read events are reported and no write because
the kprobe is installed for the old ntfs_file_write_iter() and not the ntfs3
one.
The same behavior occurs with sysfs kprobe:
root@vm-amd64:/# echo 'p:probe/ntfs_file_write_iter ntfs_file_write_iter' > /sys/kernel/tracing/kprobe_events
root@vm-amd64:/# cat /sys/kernel/tracing/kprobe_events
p:probe/ntfs_file_write_iter ntfs_file_write_iter
root@vm-amd64:/# mount | grep /mnt
/foo.img on /mnt type ntfs3 (rw,relatime,uid=0,gid=0,iocharset=utf8)
root@vm-amd64:/# perf record -e probe:ntfs_file_write_iter &
[1] 210
root@vm-amd64:/# cd /mnt/
root@vm-amd64:/mnt# dd if=./foo of=./bar count=3
3+0 records in
3+0 records out
1536 bytes (1.5 kB, 1.5 KiB) copied, 0.00234793 s, 654 kB/s
root@vm-amd64:/mnt# cd -
/
root@vm-amd64:/# fg
perf record -e probe:ntfs_file_write_iter
^C[ perf record: Woken up 1 times to write data ]
[ perf record: Captured and wrote 0.056 MB perf.data ]

root@vm-amd64:/# perf report
Error:
The perf.data data has no samples!
# To display the perf.data header info, please use --header/--header-only optio>
#

In this contribution, I modified the functions creating sysfs and PMU kprobes to test if the function name given as argument matches several symbols. In this case, these functions return EADDRNOTAVAIL to indicate the user to use addr and offs to remove this ambiguity.
So, when the above BPF tool is run, the following error message is printed:

root@vm-amd64:~# ig trace fsslower -m 0 -f ntfs3 --host &> /tmp/foo &
[1] 228
root@vm-amd64:~# more /tmp/foo
RUNTIME.CONTAINERNAME          RUNTIME.CONTAIN… PID              COMM
  T      BYTES     OFFSET        LAT FILE
Error: running gadget: running gadget: installing tracer: attaching kprobe: crea
ting perf_kprobe PMU (arch-specific fallback for "ntfs_file_write_iter"): token
ntfs_file_write_iter: opening perf event: cannot assign requested address
And the same with sysfs kprobe:
root@vm-amd64:/# echo 'p:probe/ntfs_file_write_iter ntfs_file_write_iter' > /sys/kernel/tracing/kprobe_events
-bash: echo: write error: Cannot assign requested address
Note that, this does not influence perf as it installs kprobes as offset on
_text:
root@vm-amd64:/# perf probe --add ntfs_file_write_iter
Added new events:
  probe:ntfs_file_write_iter (on ntfs_file_write_iter)
  probe:ntfs_file_write_iter (on ntfs_file_write_iter)
...
root@vm-amd64:/# cat /sys/kernel/tracing/kprobe_events
p:probe/ntfs_file_write_iter _text+5039088
p:probe/ntfs_file_write_iter _text+5228752

Note that, this contribution is the conclusion of a previous RFC which intended to install a PMU kprobe for all matching symbols [1, 2].

If you see any way to improve this contribution, particularly if you have an idea to add tests or documentation for this behavior, please share your feedback.

Changes since:
v1:
* Use EADDRNOTAVAIL instead of adding a new error code.
* Correct also this behavior for sysfs kprobe.

Francis Laniel (1):
tracing/kprobes: Return EADDRNOTAVAIL when func matches several
symbols

 kernel/trace/trace_kprobe.c | 42 +++++++++++++++++++++++++++++++++++++
 1 file changed, 42 insertions(+)

Best regards and thank you in advance.
---
[1]: https://lore.kernel.org/lkml/[email protected]/
[2]: https://lore.kernel.org/lkml/[email protected]/
--
2.34.1

Francis Laniel

author

2023-08-24 10:38 UTC #1 lore raw

RFC PATCH v2 1/1 tracing/kprobes: Return EADDRNOTAVAIL when func matches several symbols

Previously to this commit, if func matches several symbols, a kprobe, being either sysfs or PMU, would only be installed for the first matching address. This could lead to some misunderstanding when some BPF code was never called because it was attached to a function which was indeed not call, because the effectively called one has no kprobes.

So, this commit returns EADDRNOTAVAIL when func matches several symbols. This way, user needs to use addr to remove the ambiguity.

Signed-off-by Francis Laniel <[email protected]>

Link https://lore.kernel.org/lkml/[email protected]/ open here

---
 kernel/trace/trace_kprobe.c | 42 +++++++++++++++++++++++++++++++++++++
 1 file changed, 42 insertions(+)

kernel/trace/trace_kprobe.c +42 -0

index 23dba01831f7..0c8dd6ba650b 100644
@@ -705,6 +705,25 @@ static struct notifier_block trace_kprobe_module_nb = {
 	.priority = 1	/* Invoked after kprobe module callback */
 };
 
+static int count_symbols(void *data, unsigned long unused)
+{
+	unsigned int *count = data;
+
+	(*count)++;
+
+	return 0;
+}
+
+static unsigned int func_name_several_symbols(char *func_name)
+{
+	unsigned int count;
+
+	count = 0;
+	kallsyms_on_each_match_symbol(count_symbols, func_name, &count);
+
+	return count > 1;
+}
+
 static int __trace_kprobe_create(int argc, const char *argv[])
 {
 	/*

@@ -836,6 +855,18 @@ static int __trace_kprobe_create(int argc, const char *argv[])
 		}
 	}
 
+	/*
+	 * If user specifies KSYM, we check it does not correspond to several
+	 * symbols.
+	 * If this is the case, we return EADDRNOTAVAIL to indicate the user
+	 * he/she should use ADDR rather than KSYM to remove the ambiguity.
+	 */
+	if (symbol && func_name_several_symbols(symbol)) {
+		ret = -EADDRNOTAVAIL;
+
+		goto error;
+	}
+
 	trace_probe_log_set_index(0);
 	if (event) {
 		ret = traceprobe_parse_event_name(&event, &group, gbuf,

@@ -1699,6 +1730,7 @@ static int unregister_kprobe_event(struct trace_kprobe *tk)
 }
 
 #ifdef CONFIG_PERF_EVENTS
+
 /* create a trace_kprobe, but don't add it to global lists */
 struct trace_event_call *
 create_local_trace_kprobe(char *func, void *addr, unsigned long offs,

@@ -1709,6 +1741,16 @@ create_local_trace_kprobe(char *func, void *addr, unsigned long offs,
 	int ret;
 	char *event;
 
+	/*
+	 * If user specifies func, we check that function name does not
+	 * correspond to several symbols.
+	 * If this is the case, we return EADDRNOTAVAIL to indicate the user
+	 * he/she should use addr and offs rather than func to remove the
+	 * ambiguity.
+	 */
+	if (func && func_name_several_symbols(func))
+		return ERR_PTR(-EADDRNOTAVAIL);
+
 	/*
 	 * local trace_kprobes are not added to dyn_event, so they are never
 	 * searched in find_trace_kprobe(). Therefore, there is no concern of

Signature · 1 line

2.34.1

Masami Hiramatsu (Google)

maintainer

2023-08-24 13:03 UTC #2 lore raw

Re: [RFC PATCH v2 1/1] tracing/kprobes: Return EADDRNOTAVAIL when func matches several symbols

Quoted 8 lines · On Thu, 24 Aug 2023 12:37:34 +0200 Francis Laniel…

Previously to this commit, if func matches several symbols, a kprobe, being either sysfs or PMU, would only be installed for the first matching address. This could lead to some misunderstanding when some BPF code was never called because it was attached to a function which was indeed not call, because the effectively called one has no kprobes.
So, this commit returns EADDRNOTAVAIL when func matches several symbols. This way, user needs to use addr to remove the ambiguity.

Thanks for update the patch. I have some comments there.

Quoted 16 lines

Suggested-by Masami Hiramatsu <[email protected]>
Signed-off-by Francis Laniel <[email protected]>
Link https://lore.kernel.org/lkml/[email protected]/ open here
---
 kernel/trace/trace_kprobe.c | 42 +++++++++++++++++++++++++++++++++++++
 1 file changed, 42 insertions(+)
kernel/trace/trace_kprobe.c +0 -0
index 23dba01831f7..0c8dd6ba650b 100644
@@ -705,6 +705,25 @@ static struct notifier_block trace_kprobe_module_nb = {
 	.priority = 1	/* Invoked after kprobe module callback */
 };
 

kernel/trace/trace_kprobe.c static struct notifier_block trace_kprobe_module_nb = { · line 717

Quoted 10 lines

kernel/trace/trace_kprobe.c +10 -0

+static int count_symbols(void *data, unsigned long unused)
+{
+	unsigned int *count = data;
+
+	(*count)++;
+
+	return 0;
+}
+
+static unsigned int func_name_several_symbols(char *func_name)

If this returns boolean, please use 'bool' for return type. Also, I think 'func_name_is_unique()' is more natural.

Quoted 13 lines

kernel/trace/trace_kprobe.c +9 -0

+{
+	unsigned int count;
+
+	count = 0;
+	kallsyms_on_each_match_symbol(count_symbols, func_name, &count);
+
+	return count > 1;
+}
+
 static int __trace_kprobe_create(int argc, const char *argv[])
 {
 	/*

kernel/trace/trace_kprobe.c static int __trace_kprobe_create(int argc, const char *argv[]) · line 864

Quoted 10 lines

kernel/trace/trace_kprobe.c +7 -0

@@ -836,6 +855,18 @@ static int __trace_kprobe_create(int argc, const char *argv[])
 		}
 	}
 
+	/*
+	 * If user specifies KSYM, we check it does not correspond to several
+	 * symbols.
+	 * If this is the case, we return EADDRNOTAVAIL to indicate the user
+	 * he/she should use ADDR rather than KSYM to remove the ambiguity.
+	 */
+	if (symbol && func_name_several_symbols(symbol)) {

Then, here will be

	if (symbol && !func_name_is_unique(symbol)) {

Quoted 18 lines

kernel/trace/trace_kprobe.c +6 -0

+		ret = -EADDRNOTAVAIL;
+
+		goto error;
+	}
+
 	trace_probe_log_set_index(0);
 	if (event) {
 		ret = traceprobe_parse_event_name(&event, &group, gbuf,

@@ -1699,6 +1730,7 @@ static int unregister_kprobe_event(struct trace_kprobe *tk)
 }
 
 #ifdef CONFIG_PERF_EVENTS
+
 /* create a trace_kprobe, but don't add it to global lists */
 struct trace_event_call *
 create_local_trace_kprobe(char *func, void *addr, unsigned long offs,

@@ -1709,6 +1741,16 @@ create_local_trace_kprobe(char *func, void *addr, unsigned long offs,
 	int ret;

kernel/trace/trace_kprobe.c create_local_trace_kprobe(char *func, void *addr, unsigned long offs, · line 1751

Quoted 10 lines

kernel/trace/trace_kprobe.c +8 -0

 	char *event;
 
+	/*
+	 * If user specifies func, we check that function name does not
+	 * correspond to several symbols.
+	 * If this is the case, we return EADDRNOTAVAIL to indicate the user
+	 * he/she should use addr and offs rather than func to remove the
+	 * ambiguity.
+	 */
+	if (func && func_name_several_symbols(func))

Ditto.

Thanks!

Quoted 8 lines

kernel/trace/trace_kprobe.c +2 -0

+		return ERR_PTR(-EADDRNOTAVAIL);
+
 	/*
 	 * local trace_kprobes are not added to dyn_event, so they are never
 	 * searched in find_trace_kprobe(). Therefore, there is no concern of

Signature · 1 line

2.34.1

Signature · 1 line

Masami Hiramatsu (Google) <[email protected]>

Francis Laniel

author

2023-08-24 14:32 UTC #3 lore raw

Re: [RFC PATCH v2 1/1] tracing/kprobes: Return EADDRNOTAVAIL when func matches several symbols

Hi.

Quoted 34 lines · Le jeudi 24 août 2023, 15:02:27 CEST Masami Hiramatsu a écrit :

On Thu, 24 Aug 2023 12:37:34 +0200
Quoted 9 lines · Francis Laniel
Previously to this commit, if func matches several symbols, a kprobe, being
either sysfs or PMU, would only be installed for the first matching address. This could lead to some misunderstanding when some BPF code was never called because it was attached to a function which was indeed not call, because the effectively called one has no kprobes.
So, this commit returns EADDRNOTAVAIL when func matches several symbols. This way, user needs to use addr to remove the ambiguity.
Thanks for update the patch. I have some comments there.
Quoted 19 lines
Suggested-by Masami Hiramatsu <[email protected]>
Signed-off-by Francis Laniel <[email protected]>
Link:
https://lore.kernel.org/lkml/20230819101105.b0c104ae4494a7d1f2eea742@kern
el.org/ ---
 kernel/trace/trace_kprobe.c | 42 +++++++++++++++++++++++++++++++++++++
 1 file changed, 42 insertions(+)
kernel/trace/trace_kprobe.c +0 -0
index 23dba01831f7..0c8dd6ba650b 100644
@@ -705,6 +705,25 @@ static struct notifier_block trace_kprobe_module_nb =
{>
.priority = 1 /* Invoked after kprobe module callback */
};

Quoted 14 lines

kernel/trace/trace_kprobe.c static struct notifier_block trace_kprobe_module_nb = { · line 717
Quoted 10 lines
kernel/trace/trace_kprobe.c +10 -0
+static int count_symbols(void *data, unsigned long unused)
+{
+	unsigned int *count = data;
+
+	(*count)++;
+
+	return 0;
+}
+
+static unsigned int func_name_several_symbols(char *func_name)
If this returns boolean, please use 'bool' for return type. Also, I think 'func_name_is_unique()' is more natural.

This name sounds better but it means it will check count == 1. I am fine with it, but please see my below comment.

Quoted 31 lines

Quoted 18 lines

kernel/trace/trace_kprobe.c +9 -0

+{
+	unsigned int count;
+
+	count = 0;
+	kallsyms_on_each_match_symbol(count_symbols, func_name, &count);
+
+	return count > 1;
+}
+
 
 static int __trace_kprobe_create(int argc, const char *argv[])
 {
 
 	/*

@@ -836,6 +855,18 @@ static int __trace_kprobe_create(int argc, const char
 *argv[])> 
 		}

kernel/trace/trace_kprobe.c static int __trace_kprobe_create(int argc, const char · line 866

Quoted 10 lines

kernel/trace/trace_kprobe.c +7 -0

 
 	}
 
+	/*
+	 * If user specifies KSYM, we check it does not correspond to several
+	 * symbols.
+	 * If this is the case, we return EADDRNOTAVAIL to indicate the user
+	 * he/she should use ADDR rather than KSYM to remove the ambiguity.
+	 */
+	if (symbol && func_name_several_symbols(symbol)) {

Then, here will be

Quoted 2 lines

	if (symbol && !func_name_is_unique(symbol)) {

I am fine with the above, but it means if users gives an unknown symbol, we will return EADDRNOTAVAIL instead of currently returning ENOENT. Is it OK?

Quoted 41 lines

Quoted 26 lines

kernel/trace/trace_kprobe.c +5 -0

+		ret = -EADDRNOTAVAIL;
+
+		goto error;
+	}
+

kernel/trace/trace_kprobe.c +0 -0

 	trace_probe_log_set_index(0);
 	if (event) {
 
 		ret = traceprobe_parse_event_name(&event, &group, gbuf,

@@ -1699,6 +1730,7 @@ static int unregister_kprobe_event(struct

trace_kprobe *tk)>
}

#ifdef CONFIG_PERF_EVENTS

kernel/trace/trace_kprobe.c +1 -0

+
 
 /* create a trace_kprobe, but don't add it to global lists */
 struct trace_event_call *
 create_local_trace_kprobe(char *func, void *addr, unsigned long offs,

@@ -1709,6 +1741,16 @@ create_local_trace_kprobe(char *func, void *addr,

unsigned long offs,>

kernel/trace/trace_kprobe.c +0 -0

 	int ret;

kernel/trace/trace_kprobe.c create_local_trace_kprobe(char *func, void *addr, unsigned long offs, · line 1751

Quoted 10 lines

kernel/trace/trace_kprobe.c +8 -0

 	char *event;
 
+	/*
+	 * If user specifies func, we check that function name does not
+	 * correspond to several symbols.
+	 * If this is the case, we return EADDRNOTAVAIL to indicate the user
+	 * he/she should use addr and offs rather than func to remove the
+	 * ambiguity.
+	 */
+	if (func && func_name_several_symbols(func))

Ditto.

Thanks!

Quoted 7 lines

kernel/trace/trace_kprobe.c create_local_trace_kprobe(char *func, void *addr, unsigned long offs, · line 1756
Quoted 7 lines
kernel/trace/trace_kprobe.c +2 -0
+		return ERR_PTR(-EADDRNOTAVAIL);
+
 
 	/*
 
 	 * local trace_kprobes are not added to dyn_event, so they are never
 	 * searched in find_trace_kprobe(). Therefore, there is no concern of

Best regards.

Masami Hiramatsu (Google)

maintainer

2023-08-24 14:49 UTC #4 lore raw

Re: [RFC PATCH v2 1/1] tracing/kprobes: Return EADDRNOTAVAIL when func matches several symbols

Quoted 86 lines · On Thu, 24 Aug 2023 16:31:13 +0200 Francis Laniel…

Hi.
Quoted 34 lines · Le jeudi 24 août 2023, 15:02:27 CEST Masami Hiramatsu a écrit :
On Thu, 24 Aug 2023 12:37:34 +0200
Quoted 9 lines · Francis Laniel
Previously to this commit, if func matches several symbols, a kprobe, being
either sysfs or PMU, would only be installed for the first matching address. This could lead to some misunderstanding when some BPF code was never called because it was attached to a function which was indeed not call, because the effectively called one has no kprobes.
So, this commit returns EADDRNOTAVAIL when func matches several symbols. This way, user needs to use addr to remove the ambiguity.
Thanks for update the patch. I have some comments there.
Quoted 19 lines
Suggested-by Masami Hiramatsu <[email protected]>
Signed-off-by Francis Laniel <[email protected]>
Link:
https://lore.kernel.org/lkml/20230819101105.b0c104ae4494a7d1f2eea742@kern
el.org/ ---
 kernel/trace/trace_kprobe.c | 42 +++++++++++++++++++++++++++++++++++++
 1 file changed, 42 insertions(+)
kernel/trace/trace_kprobe.c +0 -0
index 23dba01831f7..0c8dd6ba650b 100644
@@ -705,6 +705,25 @@ static struct notifier_block trace_kprobe_module_nb =
{>
.priority = 1 /* Invoked after kprobe module callback */
};
Quoted 14 lines
kernel/trace/trace_kprobe.c static struct notifier_block trace_kprobe_module_nb = { · line 717
Quoted 10 lines
kernel/trace/trace_kprobe.c +10 -0
+static int count_symbols(void *data, unsigned long unused)
+{
+	unsigned int *count = data;
+
+	(*count)++;
+
+	return 0;
+}
+
+static unsigned int func_name_several_symbols(char *func_name)
If this returns boolean, please use 'bool' for return type. Also, I think 'func_name_is_unique()' is more natural.
This name sounds better but it means it will check count == 1. I am fine with it, but please see my below comment.
Quoted 31 lines
Quoted 18 lines
kernel/trace/trace_kprobe.c +9 -0
+{
+	unsigned int count;
+
+	count = 0;
+	kallsyms_on_each_match_symbol(count_symbols, func_name, &count);
+
+	return count > 1;
+}
+
 
 static int __trace_kprobe_create(int argc, const char *argv[])
 {
 
 	/*
@@ -836,6 +855,18 @@ static int __trace_kprobe_create(int argc, const char
 *argv[])> 
 		}
kernel/trace/trace_kprobe.c static int __trace_kprobe_create(int argc, const char · line 866
Quoted 10 lines
kernel/trace/trace_kprobe.c +7 -0
 
 	}
 
+	/*
+	 * If user specifies KSYM, we check it does not correspond to several
+	 * symbols.
+	 * If this is the case, we return EADDRNOTAVAIL to indicate the user
+	 * he/she should use ADDR rather than KSYM to remove the ambiguity.
+	 */
+	if (symbol && func_name_several_symbols(symbol)) {
Then, here will be

Quoted 6 lines

Quoted 2 lines
	if (symbol && !func_name_is_unique(symbol)) {
I am fine with the above, but it means if users gives an unknown symbol, we will return EADDRNOTAVAIL instead of currently returning ENOENT. Is it OK?

Ah, good catch! Hm, then what about 'int number_of_same_symbols()' ?

if (symbol) {
	num = number_of_same_symbols(symbol);
	if (num > 1)
		return -EADDRNOTAVAIL;
	else if (num == 0)
		return -ENOENT;
}

Thank you,

Quoted 53 lines

Quoted 42 lines
Quoted 27 lines
kernel/trace/trace_kprobe.c +5 -0
+		ret = -EADDRNOTAVAIL;
+
+		goto error;
+	}
+
 
kernel/trace/trace_kprobe.c +0 -0
 	trace_probe_log_set_index(0);
 	if (event) {
 
 		ret = traceprobe_parse_event_name(&event, &group, gbuf,
@@ -1699,6 +1730,7 @@ static int unregister_kprobe_event(struct
trace_kprobe *tk)>
}
#ifdef CONFIG_PERF_EVENTS
kernel/trace/trace_kprobe.c +1 -0
+
 
 /* create a trace_kprobe, but don't add it to global lists */
 struct trace_event_call *
 create_local_trace_kprobe(char *func, void *addr, unsigned long offs,
@@ -1709,6 +1741,16 @@ create_local_trace_kprobe(char *func, void *addr,
unsigned long offs,>
kernel/trace/trace_kprobe.c +0 -0
 	int ret;
kernel/trace/trace_kprobe.c create_local_trace_kprobe(char *func, void *addr, unsigned long offs, · line 1751
Quoted 10 lines
kernel/trace/trace_kprobe.c +8 -0
 	char *event;
 
+	/*
+	 * If user specifies func, we check that function name does not
+	 * correspond to several symbols.
+	 * If this is the case, we return EADDRNOTAVAIL to indicate the user
+	 * he/she should use addr and offs rather than func to remove the
+	 * ambiguity.
+	 */
+	if (func && func_name_several_symbols(func))
Ditto.
Thanks!
Quoted 7 lines
kernel/trace/trace_kprobe.c create_local_trace_kprobe(char *func, void *addr, unsigned long offs, · line 1756
Quoted 7 lines
kernel/trace/trace_kprobe.c +2 -0
+		return ERR_PTR(-EADDRNOTAVAIL);
+
 
 	/*
 
 	 * local trace_kprobes are not added to dyn_event, so they are never
 	 * searched in find_trace_kprobe(). Therefore, there is no concern of
Best regards.

Signature · 1 line

Masami Hiramatsu (Google) <[email protected]>

Francis Laniel

author

2023-08-24 16:10 UTC #5 lore raw

Re: [RFC PATCH v2 1/1] tracing/kprobes: Return EADDRNOTAVAIL when func matches several symbols

Quoted 108 lines · Le jeudi 24 août 2023, 16:47:21 CEST Masami Hiramatsu a écrit :

On Thu, 24 Aug 2023 16:31:13 +0200
Quoted 95 lines · Francis Laniel
Hi.
Quoted 41 lines · Le jeudi 24 août 2023, 15:02:27 CEST Masami Hiramatsu a écrit :
On Thu, 24 Aug 2023 12:37:34 +0200
Quoted 9 lines · Francis Laniel
Previously to this commit, if func matches several symbols, a kprobe, being
either sysfs or PMU, would only be installed for the first matching address. This could lead to some misunderstanding when some BPF code was
never called because it was attached to a function which was indeed not
call, because the effectively called one has no kprobes.
Quoted 3 lines
So, this commit returns EADDRNOTAVAIL when func matches several symbols.
This way, user needs to use addr to remove the ambiguity.
Thanks for update the patch. I have some comments there.
Quoted 23 lines
Suggested-by Masami Hiramatsu <[email protected]>
Signed-off-by Francis Laniel <[email protected]>
Link:
https://lore.kernel.org/lkml/20230819101105.b0c104ae4494a7d1f2eea742@k
ern
el.org/ ---
 kernel/trace/trace_kprobe.c | 42
 +++++++++++++++++++++++++++++++++++++
 1 file changed, 42 insertions(+)
kernel/trace/trace_kprobe.c +0 -0
index 23dba01831f7..0c8dd6ba650b 100644
@@ -705,6 +705,25 @@ static struct notifier_block
trace_kprobe_module_nb =
{>
.priority = 1 /* Invoked after kprobe module callback */
};
Quoted 13 lines
kernel/trace/trace_kprobe.c static struct notifier_block trace_kprobe_module_nb = { · line 717
Quoted 10 lines
kernel/trace/trace_kprobe.c +10 -0
+static int count_symbols(void *data, unsigned long unused)
+{
+	unsigned int *count = data;
+
+	(*count)++;
+
+	return 0;
+}
+
+static unsigned int func_name_several_symbols(char *func_name)
If this returns boolean, please use 'bool' for return type. Also, I think 'func_name_is_unique()' is more natural.
This name sounds better but it means it will check count == 1. I am fine with it, but please see my below comment.
Quoted 34 lines
Quoted 26 lines
kernel/trace/trace_kprobe.c +9 -0
+{
+	unsigned int count;
+
+	count = 0;
+	kallsyms_on_each_match_symbol(count_symbols, func_name, &count);
+
+	return count > 1;
+}
+
 
 static int __trace_kprobe_create(int argc, const char *argv[])
 {
 
 	/*
@@ -836,6 +855,18 @@ static int __trace_kprobe_create(int argc, const
char
*argv[])>
kernel/trace/trace_kprobe.c +2 -0
 		}
 
 	}
 
+	/*
+	 * If user specifies KSYM, we check it does not correspond to
several
kernel/trace/trace_kprobe.c static int __trace_kprobe_create(int argc, const char *argv[]) · line 864
Quoted 5 lines
kernel/trace/trace_kprobe.c +5 -0
+	 * symbols.
+	 * If this is the case, we return EADDRNOTAVAIL to indicate the user
+	 * he/she should use ADDR rather than KSYM to remove the ambiguity.
+	 */
+	if (symbol && func_name_several_symbols(symbol)) {
Then, here will be
Quoted 6 lines
Quoted 1 line
	if (symbol && !func_name_is_unique(symbol)) {
I am fine with the above, but it means if users gives an unknown symbol, we
will return EADDRNOTAVAIL instead of currently returning ENOENT. Is it OK?
Ah, good catch! Hm, then what about 'int number_of_same_symbols()' ?

Quoted 7 lines

if (symbol) {
	num = number_of_same_symbols(symbol);
	if (num > 1)
		return -EADDRNOTAVAIL;
	else if (num == 0)
		return -ENOENT;
}

Done in the v3 :D!

Quoted 56 lines

Thank you,
Quoted 54 lines
Quoted 44 lines
Quoted 29 lines
kernel/trace/trace_kprobe.c +5 -0
+		ret = -EADDRNOTAVAIL;
+
+		goto error;
+	}
+
 
 	trace_probe_log_set_index(0);
 	if (event) {
 
 		ret = traceprobe_parse_event_name(&event, &group, gbuf,
@@ -1699,6 +1730,7 @@ static int unregister_kprobe_event(struct
trace_kprobe *tk)>
}
#ifdef CONFIG_PERF_EVENTS
kernel/trace/trace_kprobe.c +1 -0
+
 
 /* create a trace_kprobe, but don't add it to global lists */
 struct trace_event_call *
 create_local_trace_kprobe(char *func, void *addr, unsigned long offs,
@@ -1709,6 +1741,16 @@ create_local_trace_kprobe(char *func, void
 *addr,
unsigned long offs,>
kernel/trace/trace_kprobe.c +0 -0
 	int ret;
kernel/trace/trace_kprobe.c create_local_trace_kprobe(char *func, void *addr, unsigned long offs, · line 1751
Quoted 10 lines
kernel/trace/trace_kprobe.c +8 -0
 	char *event;
 
+	/*
+	 * If user specifies func, we check that function name does not
+	 * correspond to several symbols.
+	 * If this is the case, we return EADDRNOTAVAIL to indicate the user
+	 * he/she should use addr and offs rather than func to remove the
+	 * ambiguity.
+	 */
+	if (func && func_name_several_symbols(func))
Ditto.
Thanks!
Quoted 8 lines
kernel/trace/trace_kprobe.c create_local_trace_kprobe(char *func, void *addr, unsigned long offs,
Quoted 8 lines
kernel/trace/trace_kprobe.c +2 -0
+		return ERR_PTR(-EADDRNOTAVAIL);
+
 
 	/*
 
 	 * local trace_kprobes are not added to dyn_event, so they are never
 	 * searched in find_trace_kprobe(). Therefore, there is no concern
 	 of
Best regards.