Skip to contentSkip to conversation

usermodehelper: kill the sub_info->path[0] check

PATCH In review Oleg Nesterov 3 participants 4 messages May 20, 2013 – May 22, 2013

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 usermodehelper: kill the sub_info->path[0] check

call_usermodehelper_exec() does nothing but returns success if path[0] == 0. The only user which needs this strange feature is request_module(), it can check modprobe_path[0] itself like other users do if they want to detect the "disabled by admin" case.

Kill it. Not only it looks strange, it can confuse other callers. And this allows us to revert 264b83c0 "usermodehelper: check subprocess_info->path != NULL", do_execve(NULL) is safe.

---
 kernel/kmod.c |   11 +++--------
 1 files changed, 3 insertions(+), 8 deletions(-)
kernel/kmod.c +3 -8
index 8241906..fb32636 100644
@@ -147,6 +147,9 @@ int __request_module(bool wait, const char *fmt, ...)
 	 */
 	WARN_ON_ONCE(wait && current_is_async());
 
+	if (!modprobe_path[0])
+		return 0;
+
 	va_start(args, fmt);
 	ret = vsnprintf(module_name, MODULE_NAME_LEN, fmt, args);
 	va_end(args);
@@ -569,14 +572,6 @@ int call_usermodehelper_exec(struct subprocess_info *sub_info, int wait)
 	int retval = 0;
 
 	helper_lock();
-	if (!sub_info->path) {
-		retval = -EINVAL;
-		goto out;
-	}
-
-	if (sub_info->path[0] == '\0')
-		goto out;
-
 	if (!khelper_wq || usermodehelper_disabled) {
 		retval = -EBUSY;
 		goto out;
Signature · 1 line
1.5.5.1
Acked-by on v1 1/1
#1 lore raw
Quoted 10 lines · Oleg Nesterov

call_usermodehelper_exec() does nothing but returns success if path[0] == 0. The only user which needs this strange feature is request_module(), it can check modprobe_path[0] itself like other users do if they want to detect the "disabled by admin" case.

Kill it. Not only it looks strange, it can confuse other callers. And this allows us to revert 264b83c0 "usermodehelper: check subprocess_info->path != NULL", do_execve(NULL) is safe.

Nice cleanup. It was definitely weird...

Thanks!
Rusty.

Quoted 35 lines
---
 kernel/kmod.c |   11 +++--------
 1 files changed, 3 insertions(+), 8 deletions(-)
kernel/kmod.c +3 -8
index 8241906..fb32636 100644
@@ -147,6 +147,9 @@ int __request_module(bool wait, const char *fmt, ...)
 	 */
 	WARN_ON_ONCE(wait && current_is_async());
 
+	if (!modprobe_path[0])
+		return 0;
+
 	va_start(args, fmt);
 	ret = vsnprintf(module_name, MODULE_NAME_LEN, fmt, args);
 	va_end(args);
@@ -569,14 +572,6 @@ int call_usermodehelper_exec(struct subprocess_info *sub_info, int wait)
 	int retval = 0;
 
 	helper_lock();
-	if (!sub_info->path) {
-		retval = -EINVAL;
-		goto out;
-	}
-
-	if (sub_info->path[0] == '\0')
-		goto out;
-
 	if (!khelper_wq || usermodehelper_disabled) {
 		retval = -EBUSY;
 		goto out;
Signature · 1 line
1.5.5.1
Acked-by on v1 1/1
#2 lore raw
In reply to Oleg Nesterov

Hi Oleg,

Quoted 10 lines · Oleg Nesterov

call_usermodehelper_exec() does nothing but returns success if path[0] == 0. The only user which needs this strange feature is request_module(), it can check modprobe_path[0] itself like other users do if they want to detect the "disabled by admin" case.

Kill it. Not only it looks strange, it can confuse other callers. And this allows us to revert 264b83c0 "usermodehelper: check subprocess_info->path != NULL", do_execve(NULL) is safe.

But... see below.

Quoted 9 lines
---
 kernel/kmod.c |   11 +++--------
 1 files changed, 3 insertions(+), 8 deletions(-)
kernel/kmod.c +0 -0
index 8241906..fb32636 100644

kernel/kmod.c int __request_module(bool wait, const char *fmt, ...) · line 152

Quoted 6 lines
kernel/kmod.c +3 -0
@@ -147,6 +147,9 @@ int __request_module(bool wait, const char *fmt, ...)
         */
        WARN_ON_ONCE(wait && current_is_async());
 
+       if (!modprobe_path[0])
+               return 0;
+

Any reason to not return -EINVAL here except for maintaining the previous behavior? Checking the callers reveals just a few of them actually check the return value and IMO this is no different than the binary not existing and failing later on exec.

Quoted 9 lines
kernel/kmod.c +0 -2
        va_start(args, fmt);
        ret = vsnprintf(module_name, MODULE_NAME_LEN, fmt, args);
        va_end(args);
@@ -569,14 +572,6 @@ int call_usermodehelper_exec(struct subprocess_info *sub_info, int wait)
        int retval = 0;
 
        helper_lock();
-       if (!sub_info->path) {
-               retval = -EINVAL;

kernel/kmod.c int call_usermodehelper_exec(struct subprocess_info *sub_info, int wait)

Quoted 10 lines
kernel/kmod.c +0 -7
-               goto out;
-       }
-
-       if (sub_info->path[0] == '\0')
-               goto out;
-
        if (!khelper_wq || usermodehelper_disabled) {
                retval = -EBUSY;
                goto out;
--

Lucas De Marchi

#3 lore raw

Hi Lucas,

Quoted 2 lines · Lucas De Marchi

Thanks.

Quoted 10 lines

kernel/kmod.c int __request_module(bool wait, const char *fmt, ...) · line 152

Quoted 7 lines
kernel/kmod.c +3 -0
@@ -147,6 +147,9 @@ int __request_module(bool wait, const char *fmt, ...)
         */
        WARN_ON_ONCE(wait && current_is_async());
 
+       if (!modprobe_path[0])
+               return 0;
+

Any reason to not return -EINVAL here except for maintaining the previous behavior?

But for what?

Keep the previous behaviour is important. And this matches, say, kobject_uevent_env().

Quoted 3 lines

Checking the callers reveals just a few of them
actually check the return value and IMO this is no different than the binary not existing and failing later on exec.

Yes, agreed. And perhaps request_module() is different. For example, search_binary_handler(). Perhaps we should change this, but imho this needs more patches/discussion.

This is like the previous commit 264b83c0 reverted by this patch, the change tries to be simple and conservative.

Oleg.