[PATCH] genetlink: Optimize genl_register_family()

[PATCH] genetlink: Optimize genl_register_family()

[Krishna Kumar]

From: Krishna Kumar <krkumar2@in.ibm.com>

genl_register_family() doesn't need to call genl_family_find_byid
when GENL_ID_GENERATE is passed during register.

Patch on net-next-2.6, compile and reboot testing only.

Signed-off-by: Krishna Kumar <krkumar2@in.ibm.com>
---
diff -ruNp org/net/netlink/genetlink.c new/net/netlink/genetlink.c
--- org/net/netlink/genetlink.c	2009-10-06 08:34:10.000000000 +0530
+++ new/net/netlink/genetlink.c	2009-10-12 13:57:38.000000000 +0530
@@ -374,11 +374,6 @@ int genl_register_family(struct genl_fam
 		goto errout_locked;
 	}
 
-	if (genl_family_find_byid(family->id)) {
-		err = -EEXIST;
-		goto errout_locked;
-	}
-
 	if (family->id == GENL_ID_GENERATE) {
 		u16 newid = genl_generate_id();
 
@@ -388,6 +383,9 @@ int genl_register_family(struct genl_fam
 		}
 
 		family->id = newid;
+	} else if (genl_family_find_byid(family->id)) {
+		err = -EEXIST;
+		goto errout_locked;
 	}
 
 	if (family->maxattr) {

[PATCH] genetlink: Optimize and one bug fix in genl_generate_id()

[Krishna Kumar]

From: Krishna Kumar <krkumar2@in.ibm.com>

1. GENL_MIN_ID is a valid id -> no need to start at
   GENL_MIN_ID + 1.
2. Avoid going through the ids two times: If we start at
   GENL_MIN_ID+1 (*or bigger*) and all ids are over!, the
   code iterates through the list twice (*or lesser*).
3. Simplify code - no need to start at idx=0 which gets
   reset to GENL_MIN_ID.

Patch on net-next-2.6. Reboot test shows that first id
passed to genl_register_family was 16, next two were
GENL_ID_GENERATE and genl_generate_id returned 17 & 18
(user level testing of same code shows expected values
across entire range of MIN/MAX).

Signed-off-by: Krishna Kumar <krkumar2@in.ibm.com>
---
diff -ruNp org/net/netlink/genetlink.c new/net/netlink/genetlink.c
--- org/net/netlink/genetlink.c	2009-10-12 13:57:38.000000000 +0530
+++ new/net/netlink/genetlink.c	2009-10-12 13:57:59.000000000 +0530
@@ -97,25 +97,17 @@ static struct genl_ops *genl_get_cmd(u8 
 */
 static inline u16 genl_generate_id(void)
 {
-	static u16 id_gen_idx;
-	int overflowed = 0;
+	static u16 id_gen_idx = GENL_MIN_ID;
+	int i;
 
-	do {
-		if (id_gen_idx == 0)
+	for (i = 0; i <= GENL_MAX_ID - GENL_MIN_ID; i++) {
+		if (!genl_family_find_byid(id_gen_idx))
+			return id_gen_idx;
+		if (++id_gen_idx > GENL_MAX_ID)
 			id_gen_idx = GENL_MIN_ID;
+	}
 
-		if (++id_gen_idx > GENL_MAX_ID) {
-			if (!overflowed) {
-				overflowed = 1;
-				id_gen_idx = 0;
-				continue;
-			} else
-				return 0;
-		}
-
-	} while (genl_family_find_byid(id_gen_idx));
-
-	return id_gen_idx;
+	return 0;
 }
 
 static struct genl_multicast_group notify_grp;

Re: [PATCH] genetlink: Optimize and one bug fix in genl_generate_id()

[David Miller]

From: Krishna Kumar <krkumar2@in.ibm.com>
Date: Thu, 15 Oct 2009 11:25:07 +0530

> From: Krishna Kumar <krkumar2@in.ibm.com>
> 
> 1. GENL_MIN_ID is a valid id -> no need to start at
>    GENL_MIN_ID + 1.
> 2. Avoid going through the ids two times: If we start at
>    GENL_MIN_ID+1 (*or bigger*) and all ids are over!, the
>    code iterates through the list twice (*or lesser*).
> 3. Simplify code - no need to start at idx=0 which gets
>    reset to GENL_MIN_ID.
> 
> Patch on net-next-2.6. Reboot test shows that first id
> passed to genl_register_family was 16, next two were
> GENL_ID_GENERATE and genl_generate_id returned 17 & 18
> (user level testing of same code shows expected values
> across entire range of MIN/MAX).
> 
> Signed-off-by: Krishna Kumar <krkumar2@in.ibm.com>

Applied, thanks.

Re: [PATCH] genetlink: Optimize genl_register_family()

[David Miller]

From: Krishna Kumar <krkumar2@in.ibm.com>
Date: Thu, 15 Oct 2009 11:24:53 +0530

> From: Krishna Kumar <krkumar2@in.ibm.com>
> 
> genl_register_family() doesn't need to call genl_family_find_byid
> when GENL_ID_GENERATE is passed during register.
> 
> Patch on net-next-2.6, compile and reboot testing only.
> 
> Signed-off-by: Krishna Kumar <krkumar2@in.ibm.com>

Applied.