bnx2x: Accessing un-mapped page

The allocated RX buffer size was 64 bytes bigger than the PCI mapped
size with no good reason. If the packet was actually using the buffer up
to its limit and if the last 64 bytes of the buffer crossed 4KB boundary
then an unmapped PCI page was accessed. The fix is to use only one
parameter for the buffer size - there is no need to differentiate
between the buffer size and the PCI mapping size since the extra 64
bytes can actually be used by the FW to align the Ethernet payload to
64 bytes.

Also updating the driver version and date

Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

drivers/net/bnx2x.h

@@ -151,6 +151,8 @@ struct sw_rx_page {
 #define PAGES_PER_SGE_SHIFT		0
 #define PAGES_PER_SGE			(1 << PAGES_PER_SGE_SHIFT)
 
+#define BCM_RX_ETH_PAYLOAD_ALIGN	64
+
 /* SGE ring related macros */
 #define NUM_RX_SGE_PAGES		2
 #define RX_SGE_CNT		(BCM_PAGE_SIZE / sizeof(struct eth_rx_sge))
@@ -750,8 +752,7 @@ struct bnx2x {
 
 	u32			rx_csum;
 	u32			rx_offset;
-	u32			rx_buf_use_size;	/* useable size */
-	u32			rx_buf_size;		/* with alignment */
+	u32			rx_buf_size;
 #define ETH_OVREHEAD			(ETH_HLEN + 8)	/* 8 for CRC + VLAN */
 #define ETH_MIN_PACKET_SIZE		60
 #define ETH_MAX_PACKET_SIZE		1500

drivers/net/bnx2x_main.c

@@ -59,8 +59,8 @@
 #include "bnx2x.h"
 #include "bnx2x_init.h"
 
-#define DRV_MODULE_VERSION	"1.45.20"
-#define DRV_MODULE_RELDATE	"2008/08/25"
+#define DRV_MODULE_VERSION	"1.45.21"
+#define DRV_MODULE_RELDATE	"2008/09/03"
 #define BNX2X_BC_VER		0x040200
 
 /* Time in jiffies before concluding the transmitter is hung */
@@ -1027,7 +1027,7 @@ static inline int bnx2x_alloc_rx_skb(struct bnx2x *bp,
 	if (unlikely(skb == NULL))
 		return -ENOMEM;
 
-	mapping = pci_map_single(bp->pdev, skb->data, bp->rx_buf_use_size,
+	mapping = pci_map_single(bp->pdev, skb->data, bp->rx_buf_size,
 				 PCI_DMA_FROMDEVICE);
 	if (unlikely(dma_mapping_error(&bp->pdev->dev, mapping))) {
 		dev_kfree_skb(skb);
@@ -1169,7 +1169,7 @@ static void bnx2x_tpa_start(struct bnx2x_fastpath *fp, u16 queue,
 	/* move empty skb from pool to prod and map it */
 	prod_rx_buf->skb = fp->tpa_pool[queue].skb;
 	mapping = pci_map_single(bp->pdev, fp->tpa_pool[queue].skb->data,
-				 bp->rx_buf_use_size, PCI_DMA_FROMDEVICE);
+				 bp->rx_buf_size, PCI_DMA_FROMDEVICE);
 	pci_unmap_addr_set(prod_rx_buf, mapping, mapping);
 
 	/* move partial skb from cons to pool (don't unmap yet) */
@@ -1276,7 +1276,7 @@ static void bnx2x_tpa_stop(struct bnx2x *bp, struct bnx2x_fastpath *fp,
 	   pool entry status to BNX2X_TPA_STOP even if new skb allocation
 	   fails. */
 	pci_unmap_single(bp->pdev, pci_unmap_addr(rx_buf, mapping),
-			 bp->rx_buf_use_size, PCI_DMA_FROMDEVICE);
+			 bp->rx_buf_size, PCI_DMA_FROMDEVICE);
 
 	if (likely(new_skb)) {
 		/* fix ip xsum and give it to the stack */
@@ -1520,7 +1520,7 @@ static int bnx2x_rx_int(struct bnx2x_fastpath *fp, int budget)
 			} else if (bnx2x_alloc_rx_skb(bp, fp, bd_prod) == 0) {
 				pci_unmap_single(bp->pdev,
 					pci_unmap_addr(rx_buf, mapping),
-						 bp->rx_buf_use_size,
+						 bp->rx_buf_size,
 						 PCI_DMA_FROMDEVICE);
 				skb_reserve(skb, pad);
 				skb_put(skb, len);
@@ -4229,7 +4229,7 @@ static inline void bnx2x_free_tpa_pool(struct bnx2x *bp,
 		if (fp->tpa_state[i] == BNX2X_TPA_START)
 			pci_unmap_single(bp->pdev,
 					 pci_unmap_addr(rx_buf, mapping),
-					 bp->rx_buf_use_size,
+					 bp->rx_buf_size,
 					 PCI_DMA_FROMDEVICE);
 
 		dev_kfree_skb(skb);
@@ -4245,15 +4245,14 @@ static void bnx2x_init_rx_rings(struct bnx2x *bp)
 	u16 ring_prod, cqe_ring_prod;
 	int i, j;
 
-	bp->rx_buf_use_size = bp->dev->mtu;
-	bp->rx_buf_use_size += bp->rx_offset + ETH_OVREHEAD;
-	bp->rx_buf_size = bp->rx_buf_use_size + 64;
+	bp->rx_buf_size = bp->dev->mtu;
+	bp->rx_buf_size += bp->rx_offset + ETH_OVREHEAD +
+		BCM_RX_ETH_PAYLOAD_ALIGN;
 
 	if (bp->flags & TPA_ENABLE_FLAG) {
 		DP(NETIF_MSG_IFUP,
-		   "rx_buf_use_size %d  rx_buf_size %d  effective_mtu %d\n",
-		   bp->rx_buf_use_size, bp->rx_buf_size,
-		   bp->dev->mtu + ETH_OVREHEAD);
+		   "rx_buf_size %d  effective_mtu %d\n",
+		   bp->rx_buf_size, bp->dev->mtu + ETH_OVREHEAD);
 
 		for_each_queue(bp, j) {
 			struct bnx2x_fastpath *fp = &bp->fp[j];
@@ -4462,9 +4461,10 @@ static void bnx2x_init_context(struct bnx2x *bp)
 		context->ustorm_st_context.common.status_block_id = sb_id;
 		context->ustorm_st_context.common.flags =
 			USTORM_ETH_ST_CONTEXT_CONFIG_ENABLE_MC_ALIGNMENT;
-		context->ustorm_st_context.common.mc_alignment_size = 64;
+		context->ustorm_st_context.common.mc_alignment_size =
+			BCM_RX_ETH_PAYLOAD_ALIGN;
 		context->ustorm_st_context.common.bd_buff_size =
-						bp->rx_buf_use_size;
+						bp->rx_buf_size;
 		context->ustorm_st_context.common.bd_page_base_hi =
 						U64_HI(fp->rx_desc_mapping);
 		context->ustorm_st_context.common.bd_page_base_lo =
@@ -4717,7 +4717,7 @@ static void bnx2x_init_internal_func(struct bnx2x *bp)
 	}
 
 	/* Init CQ ring mapping and aggregation size */
-	max_agg_size = min((u32)(bp->rx_buf_use_size +
+	max_agg_size = min((u32)(bp->rx_buf_size +
 				 8*BCM_PAGE_SIZE*PAGES_PER_SGE),
 			   (u32)0xffff);
 	for_each_queue(bp, i) {
@@ -5940,7 +5940,7 @@ static void bnx2x_free_rx_skbs(struct bnx2x *bp)
 
 			pci_unmap_single(bp->pdev,
 					 pci_unmap_addr(rx_buf, mapping),
-					 bp->rx_buf_use_size,
+					 bp->rx_buf_size,
 					 PCI_DMA_FROMDEVICE);
 
 			rx_buf->skb = NULL;