Skip to contentSkip to conversation

kprobes: Use module_name() macro

PATCH Applied Jarkko Sakkinen 4 participants 6 messages Aug 18, 2020 – Sep 25, 2020

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 kprobes: Use module_name() macro

It is advised to use module_name() macro instead of dereferencing mod->name directly. This makes sense for consistencys sake and also it prevents a hard dependency to CONFIG_MODULES.

Cc Andi Kleen <[email protected]>
Cc Ard Biesheuvel <[email protected]>
Cc Jessica Yu <[email protected]>
Cc Mark Rutland <[email protected]>,
Cc Masami Hiramatsu <[email protected]>
Cc Mike Rapoport <[email protected]>
Cc Peter Zijlstra <[email protected]>
Cc Russell King <[email protected]>
Cc Will Deacon <[email protected]>

---
I thought that to get things moving it would make sense to fix this low hanging fruit issue first. Similarly as Masami's fix kernel/kprobes.c this will make my patch set less rambling, and thus easier to follow.

 kernel/trace/trace_kprobe.c | 7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)
kernel/trace/trace_kprobe.c +4 -3
index aefb6065b508..19c00ee90945 100644
@@ -106,9 +106,10 @@ static nokprobe_inline bool trace_kprobe_has_gone(struct trace_kprobe *tk)
 static nokprobe_inline bool trace_kprobe_within_module(struct trace_kprobe *tk,
 						 struct module *mod)
 {
-	int len = strlen(mod->name);
+	int len = strlen(module_name(mod));
 	const char *name = trace_kprobe_symbol(tk);
-	return strncmp(mod->name, name, len) == 0 && name[len] == ':';
+
+	return strncmp(module_name(mod), name, len) == 0 && name[len] == ':';
 }
 
 static nokprobe_inline bool trace_kprobe_module_exist(struct trace_kprobe *tk)
@@ -688,7 +689,7 @@ static int trace_kprobe_module_callback(struct notifier_block *nb,
 			if (ret)
 				pr_warn("Failed to re-register probe %s on %s: %d\n",
 					trace_probe_name(&tk->tp),
-					mod->name, ret);
+					module_name(mod), ret);
 		}
 	}
 	mutex_unlock(&event_mutex);
Signature · 1 line
2.25.1
maintainer Acked-by on v1 1/1
#1 lore raw
Quoted 4 lines · On Tue, 18 Aug 2020 08:08:57 +0300 Jarkko Sakkinen…

It is advised to use module_name() macro instead of dereferencing mod->name directly. This makes sense for consistencys sake and also it prevents a hard dependency to CONFIG_MODULES.

Quoted 11 lines
Cc Andi Kleen <[email protected]>
Cc Ard Biesheuvel <[email protected]>
Cc Jessica Yu <[email protected]>
Cc Mark Rutland <[email protected]>,
Cc Masami Hiramatsu <[email protected]>
Cc Mike Rapoport <[email protected]>
Cc Peter Zijlstra <[email protected]>
Cc Russell King <[email protected]>
Cc Will Deacon <[email protected]>

OK, this looks good to me.

Thank you,

Quoted 36 lines

---
I thought that to get things moving it would make sense to fix this low hanging fruit issue first. Similarly as Masami's fix kernel/kprobes.c this will make my patch set less rambling, and thus easier to follow.

 kernel/trace/trace_kprobe.c | 7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)
kernel/trace/trace_kprobe.c +4 -3
index aefb6065b508..19c00ee90945 100644
@@ -106,9 +106,10 @@ static nokprobe_inline bool trace_kprobe_has_gone(struct trace_kprobe *tk)
 static nokprobe_inline bool trace_kprobe_within_module(struct trace_kprobe *tk,
 						 struct module *mod)
 {
-	int len = strlen(mod->name);
+	int len = strlen(module_name(mod));
 	const char *name = trace_kprobe_symbol(tk);
-	return strncmp(mod->name, name, len) == 0 && name[len] == ':';
+
+	return strncmp(module_name(mod), name, len) == 0 && name[len] == ':';
 }
 
 static nokprobe_inline bool trace_kprobe_module_exist(struct trace_kprobe *tk)
@@ -688,7 +689,7 @@ static int trace_kprobe_module_callback(struct notifier_block *nb,
 			if (ret)
 				pr_warn("Failed to re-register probe %s on %s: %d\n",
 					trace_probe_name(&tk->tp),
-					mod->name, ret);
+					module_name(mod), ret);
 		}
 	}
 	mutex_unlock(&event_mutex);
Signature · 1 line
2.25.1
Signature · 1 line
Masami Hiramatsu <[email protected]>
#2 lore raw
Quoted 19 lines · Masami Hiramatsu
Quoted 4 lines · On Tue, 18 Aug 2020 08:08:57 +0300 Jarkko Sakkinen…

It is advised to use module_name() macro instead of dereferencing mod->name directly. This makes sense for consistencys sake and also it prevents a hard dependency to CONFIG_MODULES.

Quoted 11 lines
Cc Andi Kleen <[email protected]>
Cc Ard Biesheuvel <[email protected]>
Cc Jessica Yu <[email protected]>
Cc Mark Rutland <[email protected]>,
Cc Masami Hiramatsu <[email protected]>
Cc Mike Rapoport <[email protected]>
Cc Peter Zijlstra <[email protected]>
Cc Russell King <[email protected]>
Cc Will Deacon <[email protected]>
Quoted 3 lines

OK, this looks good to me.

Great, thank you.

When this might get included to a PR, or at minimum land to linux-next?

Just thinking what to use as the baseline for the next version of my main series.

BR,
/Jarkko

maintainer
#3 lore raw
Quoted 8 lines · On Tue, 18 Aug 2020 19:33:56 +0300 Jarkko Sakkinen…
Quoted 1 line

Great, thank you.

When this might get included to a PR, or at minimum land to linux-next?

Just thinking what to use as the baseline for the next version of my main series.

I can apply this to my tree along with Masami's latest bootconfig patches. This will be for linux-next. I don't usually push to linux-next until around -rc3. Would that be too late?

-- Steve

#4 lore raw
Quoted 9 lines · Steven Rostedt
Quoted 6 lines · On Tue, 18 Aug 2020 19:33:56 +0300 Jarkko Sakkinen…
Quoted 1 line

Great, thank you.

When this might get included to a PR, or at minimum land to linux-next?

Quoted 8 lines
Quoted 2 lines

Just thinking what to use as the baseline for the next version of my main series.

I can apply this to my tree along with Masami's latest bootconfig patches. This will be for linux-next. I don't usually push to linux-next until around -rc3. Would that be too late?

-- Steve

Nope. I have piles of stuff to catch before getting to work with this (because coming back from vacation).

/Jarkko

#5 lore raw
In reply to Jarkko Sakkinen
The following commit has been merged into the perf/core branch of tip:

The following commit has been merged into the perf/core branch of tip:

Commit-ID:     e9ffc8c1b83931599663c21ba9082bcafa51d333
Gitweb:        https://git.kernel.org/tip/e9ffc8c1b83931599663c21ba9082bcafa51d333
Author:        Jarkko Sakkinen <[email protected]>
AuthorDate:    Mon, 21 Sep 2020 21:24:25 -04:00
Committer:     Peter Zijlstra <[email protected]>
CommitterDate: Thu, 24 Sep 2020 15:55:49 +02:00

kprobes: Use module_name() macro

It is advised to use module_name() macro instead of dereferencing mod->name directly. This makes sense for consistencys sake and also it prevents a hard dependency to CONFIG_MODULES.

---
 kernel/trace/trace_kprobe.c | 7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)
kernel/trace/trace_kprobe.c +4 -3
index aefb606..19c00ee 100644
@@ -106,9 +106,10 @@ static nokprobe_inline bool trace_kprobe_has_gone(struct trace_kprobe *tk)
 static nokprobe_inline bool trace_kprobe_within_module(struct trace_kprobe *tk,
 						 struct module *mod)
 {
-	int len = strlen(mod->name);
+	int len = strlen(module_name(mod));
 	const char *name = trace_kprobe_symbol(tk);
-	return strncmp(mod->name, name, len) == 0 && name[len] == ':';
+
+	return strncmp(module_name(mod), name, len) == 0 && name[len] == ':';
 }
 
 static nokprobe_inline bool trace_kprobe_module_exist(struct trace_kprobe *tk)
@@ -688,7 +689,7 @@ static int trace_kprobe_module_callback(struct notifier_block *nb,
 			if (ret)
 				pr_warn("Failed to re-register probe %s on %s: %d\n",
 					trace_probe_name(&tk->tp),
-					mod->name, ret);
+					module_name(mod), ret);
 		}
 	}
 	mutex_unlock(&event_mutex);