Skip to contentSkip to conversation

Fix IPMI watchdog set_param_str() using kstrdup

PATCH Sébastien Dugué 2 participants 3 messages Dec 20, 2006 IPMI

lore
Share & tools
lore
#0 lore raw

I've wondered for some time why my IPMI watchdog stopped working. I traced it
back to the following patch:

[PATCH] ipmi: strstrip conversion

Here is a patch to fix this issue.

Sébastien.

Fix IPMI watchdog set_param_str() using kstrdup

set_param_str() cannot use kstrdup() to duplicate the parameter. That's
fine when the driver is compiled as a module but it sure is not when built into the kernel as the kernel parameters are parsed before the kmalloc slabs are setup.

 ipmi_watchdog.c |   11 +++++------
 1 file changed, 5 insertions(+), 6 deletions(-)
drivers/char/ipmi/ipmi_watchdog.c +5 -6
Index: linux-2.6.20-rc1/drivers/char/ipmi/ipmi_watchdog.c
===================================================================
@@ -216,13 +216,13 @@ static int set_param_str(const char *val
 {
 	action_fn  fn = (action_fn) kp->arg;
 	int        rv = 0;
-	char       *dup, *s;
+	char       valcp[16];
+	char       *s;
 
-	dup = kstrdup(val, GFP_KERNEL);
-	if (!dup)
-		return -ENOMEM;
+	strncpy(valcp, val, 16);
+	valcp[15] = '\0';
 
-	s = strstrip(dup);
+	s = strstrip(valcp);
 
 	down_read(&register_sem);
 	rv = fn(s, NULL);
@@ -235,7 +235,6 @@ static int set_param_str(const char *val
 
  out_unlock:
 	up_read(&register_sem);
-	kfree(dup);
 	return rv;
 }
 
#1 lore raw
Quoted 4 lines · Sébastien Dugué

set_param_str() cannot use kstrdup() to duplicate the parameter. That's
fine when the driver is compiled as a module but it sure is not when built into the kernel as the kernel parameters are parsed before the kmalloc slabs are setup.

Aah. I wonder though, if we could defer parsing driver parameters until kmalloc caches are up...

#2 lore raw
Quoted 8 lines · On Wed, 20 Dec 2006 13:19:09 +0200 "Pekka Enberg" <[email protected]>…
Quoted 4 lines · Sébastien Dugué

set_param_str() cannot use kstrdup() to duplicate the parameter. That's
fine when the driver is compiled as a module but it sure is not when built into the kernel as the kernel parameters are parsed before the kmalloc slabs are setup.

Aah. I wonder though, if we could defer parsing driver parameters until kmalloc caches are up...

From a general point of view, why not. Except that for the
watchdog case, I want it to start as soon as possible for it to be able to reboot in case of an early crash.

Sébastien.