pccard_store_cis: fix wrong error handling

The test for the error from pcmcia_replace_cis() was incorrect, and
would always trigger (because if an error didn't happen, the "ret" value
would not be zero, it would be the passed-in count).

Reported and debugged by Fabrice Bellet <fabrice@bellet.info>

Rather than just fix the single broken test, make the code in question
use an understandable code-sequence instead, fixing the whole function
to be more readable.
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

drivers/pcmcia/socket_sysfs.c

@@ -298,7 +298,7 @@ static ssize_t pccard_store_cis(struct kobject *kobj, char *buf, loff_t off, siz
 {
 	struct pcmcia_socket *s = to_socket(container_of(kobj, struct class_device, kobj));
 	cisdump_t *cis;
-	ssize_t ret = count;
+	int error;
 
 	if (off)
 		return -EINVAL;
@@ -316,25 +316,22 @@ static ssize_t pccard_store_cis(struct kobject *kobj, char *buf, loff_t off, siz
 	cis->Length = count + 1;
 	memcpy(cis->Data, buf, count);
 
-	if (pcmcia_replace_cis(s, cis))
-		ret  = -EIO;
-
+	error = pcmcia_replace_cis(s, cis);
 	kfree(cis);
+	if (error)
+		return -EIO;
 
-	if (!ret) {
-		mutex_lock(&s->skt_mutex);
-		if ((s->callback) && (s->state & SOCKET_PRESENT) &&
-		    !(s->state & SOCKET_CARDBUS)) {
-			if (try_module_get(s->callback->owner)) {
-				s->callback->requery(s);
-				module_put(s->callback->owner);
-			}
+	mutex_lock(&s->skt_mutex);
+	if ((s->callback) && (s->state & SOCKET_PRESENT) &&
+	    !(s->state & SOCKET_CARDBUS)) {
+		if (try_module_get(s->callback->owner)) {
+			s->callback->requery(s);
+			module_put(s->callback->owner);
 		}
-		mutex_unlock(&s->skt_mutex);
 	}
+	mutex_unlock(&s->skt_mutex);
 
-
-	return (ret);
+	return count;
 }