The arithmetic to zero pad the last 64-bit word in the push buffer is not
correct.

1. It should be pdata + length to get to the end.
2. 'pdata' is void pointer and passing it to PTR_ALIGN() will cast the
aligned pointer to void.  Pass 'end' which is u64 pointer to PTR_ALIGN()
instead so that the aligned pointer - 1 is the last 64-bit pointer to data.

Signed-off-by: Michael Chan <michael.c...@broadcom.com>
---
 drivers/net/ethernet/broadcom/bnxt/bnxt.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt.c 
b/drivers/net/ethernet/broadcom/bnxt/bnxt.c
index 8ab000d..82f1913 100644
--- a/drivers/net/ethernet/broadcom/bnxt/bnxt.c
+++ b/drivers/net/ethernet/broadcom/bnxt/bnxt.c
@@ -248,7 +248,8 @@ static netdev_tx_t bnxt_start_xmit(struct sk_buff *skb, 
struct net_device *dev)
                tx_push1->tx_bd_cfa_meta = cpu_to_le32(vlan_tag_flags);
                tx_push1->tx_bd_cfa_action = cpu_to_le32(cfa_action);
 
-               end = PTR_ALIGN(pdata + length + 1, 8) - 1;
+               end = pdata + length;
+               end = PTR_ALIGN(end, 8) - 1;
                *end = 0;
 
                skb_copy_from_linear_data(skb, pdata, len);
-- 
1.8.3.1

Reply via email to