[PATCH] kernel-2.6.9.rc4 lib/kobject.c

From: Andrew
Date: Sat Oct 16 2004 - 22:50:58 EST




There are two conditions where kset_hotplug can exit without the call to call_usermodehelper() after incrementing the sequence_number. This patch eliminates the first by getting the kobj_path before the sequence_number++. It also reduces the likelihood of the second by decrementing the sequence_number (if it can) if the call to kset->hotplug_ops->hotplug() fails.

Some user space programs (rightly or wrongly) are expecting there would be no "gaps" in hotplug event sequence numbers, and hang waiting for the "missing" events.


Signed-off-by: Andrew Duggan <cmkrnl@xxxxxxxxxxxxx>



--- lib/kobject.c.orig 2004-10-16 20:51:01.450973973 -0400
+++ lib/kobject.c 2004-10-16 21:08:19.961602269 -0400
@@ -177,6 +177,10 @@ static void kset_hotplug(const char *act
envp [i++] = scratch;
scratch += sprintf(scratch, "ACTION=%s", action) + 1;

+ kobj_path = kobject_get_path(kset, kobj, GFP_KERNEL);
+ if (!kobj_path)
+ goto exit;
+
spin_lock(&sequence_lock);
seq = sequence_num++;
spin_unlock(&sequence_lock);
@@ -184,10 +188,6 @@ static void kset_hotplug(const char *act
envp [i++] = scratch;
scratch += sprintf(scratch, "SEQNUM=%ld", seq) + 1;

- kobj_path = kobject_get_path(kset, kobj, GFP_KERNEL);
- if (!kobj_path)
- goto exit;
-
envp [i++] = scratch;
scratch += sprintf (scratch, "DEVPATH=%s", kobj_path) + 1;

@@ -199,6 +199,13 @@ static void kset_hotplug(const char *act
if (retval) {
pr_debug ("%s - hotplug() returned %d\n",
__FUNCTION__, retval);
+ /* decr sequence_num since no event will happen
+ but only if it is consistent */
+ spin_lock(&sequence_lock);
+ if (sequence_num == seq+1)
+ sequence_num--;
+ spin_unlock(&sequence_lock);
+
goto exit;
}
}



-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/