Skip to contentSkip to conversation

Trivial patch against mempool

PATCH Balbir Singh 1 participant 1 message Feb 23, 2002

lore
Share & tools
lore
#0 lore raw

That is a good suggestion too, I have redone the patch so that the alloc_fn and free_fn are checked before calling kmalloc(). I think the BUG_ON is also a good solution

I have the following now

1.

mempool.c Fri Feb 22 17:26:02 2002 renamed from mempool.c.org Fri Feb 22 12:00:58 2002 +3 -0
@@ -34,6 +34,9 @@
       mempool_t *pool;
       int i;
 
+       BUG_ON(!alloc_fn);
+       BUG_ON(!free_fn);
+
       pool = kmalloc(sizeof(*pool), GFP_KERNEL);
       if (!pool)
               return NULL;

2.

mempool.c Fri Feb 22 17:38:52 2002 renamed from mempool.c.org Fri Feb 22 12:00:58 2002 +2 -0
@@ -34,6 +34,8 @@
       mempool_t *pool;
       int i;
 
+       BUG_ON(!(alloc_fn && free_fn));
+
       pool = kmalloc(sizeof(*pool), GFP_KERNEL);
       if (!pool)
               return NULL;

I think (1) is more readable, what do you say?
Balbir

Quoted 7 lines

From: Benjamin LaHaise <[email protected]>
To: Balbir Singh <[email protected]>
CC: [email protected]
Subject: Re: [PATCH] Trivial patch against mempool
Date: Fri, 22 Feb 2002 19:02:56 -0500

On Fri, Feb 22, 2002 at 12:28:14PM -0800, Balbir Singh wrote:

Quoted 8 lines
Quoted 3 lines

Check if the alloc_fn and free_fn are not NULL. The caller generally ensures that alloc_fn and free_fn are valid. It would not harm to check. This makes the checking in mempool_create() more complete.

Rather than leak memory in that case, why not just BUG_ON null function pointers so that people know what code is at fault?

-ben

_________________________________________________________________
MSN Photos is the easiest way to share and print your photos: http://photos.msn.com/support/worldwide.aspx