[ofa-general] [PATCH][TRIVIAL]mad.c: Need parens to kmalloc correct amount of memory

[Roland Dreier]

thanks a lot for sending this with the Signed-off-by line...

But actually, looking at the code it seems that if we're going to touch
this area anyway, we might as well use krealloc() too.  So I'll queue
the following to merge:

IB/mad: Use krealloc() to resize snoop table

Use krealloc() instead of kmalloc() followed by memcpy() when resizing
the MAD module's snoop table.

Also put parentheses around the new table size to avoid calculating
the wrong size to allocate, which fixes a bug pointed out by Haven
Hash <haven.hash at isilon.com>.

Signed-off-by: Roland Dreier <rolandd at cisco.com>
---
 drivers/infiniband/core/mad.c |   14 +++++---------
 1 files changed, 5 insertions(+), 9 deletions(-)

diff --git a/drivers/infiniband/core/mad.c b/drivers/infiniband/core/mad.c
index 49c45fe..5c54fc2 100644
--- a/drivers/infiniband/core/mad.c
+++ b/drivers/infiniband/core/mad.c
@@ -406,19 +406,15 @@ static int register_snoop_agent(struct ib_mad_qp_info *qp_info,
 
 	if (i == qp_info->snoop_table_size) {
 		/* Grow table. */
-		new_snoop_table = kmalloc(sizeof mad_snoop_priv *
-					  qp_info->snoop_table_size + 1,
-					  GFP_ATOMIC);
+		new_snoop_table = krealloc(qp_info->snoop_table,
+					   sizeof mad_snoop_priv *
+					   (qp_info->snoop_table_size + 1),
+					   GFP_ATOMIC);
 		if (!new_snoop_table) {
 			i = -ENOMEM;
 			goto out;
 		}
-		if (qp_info->snoop_table) {
-			memcpy(new_snoop_table, qp_info->snoop_table,
-			       sizeof mad_snoop_priv *
-			       qp_info->snoop_table_size);
-			kfree(qp_info->snoop_table);
-		}
+
 		qp_info->snoop_table = new_snoop_table;
 		qp_info->snoop_table_size++;
 	}
-- 
1.6.0.1