* [Patch net 1/4] net: hns: fix length and page_offset overflow when CONFIG_ARM64_64K_PAGES
2018-08-23 3:10 [Patch net 0/4] net: hns: bug fixes & optimization for HNS driver Huazhong Tan
@ 2018-08-23 3:10 ` Huazhong Tan
2018-08-23 3:10 ` [Patch net 2/4] net: hns: modify variable type in hns_nic_reuse_page Huazhong Tan
` (3 subsequent siblings)
4 siblings, 0 replies; 6+ messages in thread
From: Huazhong Tan @ 2018-08-23 3:10 UTC (permalink / raw)
To: davem; +Cc: netdev, linuxarm, salil.mehta, yisen.zhuang, lipeng321,
tanhuazhong
When enable the config item "CONFIG_ARM64_64K_PAGES", the size of PAGE_SIZE
is 65536(64K). But the type of length and page_offset are u16, they will
overflow. So change them to u32.
Fixes: 6fe6611ff275 ("net: add Hisilicon Network Subsystem hnae framework support")
Signed-off-by: Huazhong Tan <tanhuazhong@huawei.com>
Signed-off-by: Salil Mehta <salil.mehta@huawei.com>
---
drivers/net/ethernet/hisilicon/hns/hnae.h | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/drivers/net/ethernet/hisilicon/hns/hnae.h b/drivers/net/ethernet/hisilicon/hns/hnae.h
index fa5b30f..cad52bd 100644
--- a/drivers/net/ethernet/hisilicon/hns/hnae.h
+++ b/drivers/net/ethernet/hisilicon/hns/hnae.h
@@ -220,10 +220,10 @@ struct hnae_desc_cb {
/* priv data for the desc, e.g. skb when use with ip stack*/
void *priv;
- u16 page_offset;
- u16 reuse_flag;
+ u32 page_offset;
+ u32 length; /* length of the buffer */
- u16 length; /* length of the buffer */
+ u16 reuse_flag;
/* desc type, used by the ring user to mark the type of the priv data */
u16 type;
--
1.9.1
^ permalink raw reply related [flat|nested] 6+ messages in thread
* [Patch net 2/4] net: hns: modify variable type in hns_nic_reuse_page
2018-08-23 3:10 [Patch net 0/4] net: hns: bug fixes & optimization for HNS driver Huazhong Tan
2018-08-23 3:10 ` [Patch net 1/4] net: hns: fix length and page_offset overflow when CONFIG_ARM64_64K_PAGES Huazhong Tan
@ 2018-08-23 3:10 ` Huazhong Tan
2018-08-23 3:10 ` [Patch net 3/4] net: hns: fix skb->truesize underestimation Huazhong Tan
` (2 subsequent siblings)
4 siblings, 0 replies; 6+ messages in thread
From: Huazhong Tan @ 2018-08-23 3:10 UTC (permalink / raw)
To: davem; +Cc: netdev, linuxarm, salil.mehta, yisen.zhuang, lipeng321,
tanhuazhong
'truesize' is supposed to be u32, not int, so fix it.
Signed-off-by: Huazhong tan <tanhuazhong@huawei.com>
Signed-off-by: Salil Mehta <salil.mehta@huawei.com>
---
drivers/net/ethernet/hisilicon/hns/hns_enet.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/drivers/net/ethernet/hisilicon/hns/hns_enet.c b/drivers/net/ethernet/hisilicon/hns/hns_enet.c
index 9f2b552..c8c0b03 100644
--- a/drivers/net/ethernet/hisilicon/hns/hns_enet.c
+++ b/drivers/net/ethernet/hisilicon/hns/hns_enet.c
@@ -512,7 +512,8 @@ static void hns_nic_reuse_page(struct sk_buff *skb, int i,
struct hnae_desc_cb *desc_cb)
{
struct hnae_desc *desc;
- int truesize, size;
+ u32 truesize;
+ int size;
int last_offset;
bool twobufs;
--
1.9.1
^ permalink raw reply related [flat|nested] 6+ messages in thread
* [Patch net 3/4] net: hns: fix skb->truesize underestimation
2018-08-23 3:10 [Patch net 0/4] net: hns: bug fixes & optimization for HNS driver Huazhong Tan
2018-08-23 3:10 ` [Patch net 1/4] net: hns: fix length and page_offset overflow when CONFIG_ARM64_64K_PAGES Huazhong Tan
2018-08-23 3:10 ` [Patch net 2/4] net: hns: modify variable type in hns_nic_reuse_page Huazhong Tan
@ 2018-08-23 3:10 ` Huazhong Tan
2018-08-23 3:10 ` [Patch net 4/4] net: hns: use eth_get_headlen interface instead of hns_nic_get_headlen Huazhong Tan
2018-08-23 4:48 ` [Patch net 0/4] net: hns: bug fixes & optimization for HNS driver David Miller
4 siblings, 0 replies; 6+ messages in thread
From: Huazhong Tan @ 2018-08-23 3:10 UTC (permalink / raw)
To: davem; +Cc: netdev, linuxarm, salil.mehta, yisen.zhuang, lipeng321,
tanhuazhong
skb->truesize is not meant to be tracking amount of used bytes in a skb,
but amount of reserved/consumed bytes in memory.
For instance, if we use a single byte in last page fragment, we have to
account the full size of the fragment.
So skb_add_rx_frag needs to calculate the length of the entire buffer into
turesize.
Fixes: 9cbe9fd5214e ("net: hns: optimize XGE capability by reducing cpu usage")
Signed-off-by: Huazhong tan <tanhuazhong@huawei.com>
Signed-off-by: Salil Mehta <salil.mehta@huawei.com>
---
drivers/net/ethernet/hisilicon/hns/hns_enet.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/net/ethernet/hisilicon/hns/hns_enet.c b/drivers/net/ethernet/hisilicon/hns/hns_enet.c
index c8c0b03..71bd3bf 100644
--- a/drivers/net/ethernet/hisilicon/hns/hns_enet.c
+++ b/drivers/net/ethernet/hisilicon/hns/hns_enet.c
@@ -531,7 +531,7 @@ static void hns_nic_reuse_page(struct sk_buff *skb, int i,
}
skb_add_rx_frag(skb, i, desc_cb->priv, desc_cb->page_offset + pull_len,
- size - pull_len, truesize - pull_len);
+ size - pull_len, truesize);
/* avoid re-using remote pages,flag default unreuse */
if (unlikely(page_to_nid(desc_cb->priv) != numa_node_id()))
--
1.9.1
^ permalink raw reply related [flat|nested] 6+ messages in thread
* [Patch net 4/4] net: hns: use eth_get_headlen interface instead of hns_nic_get_headlen
2018-08-23 3:10 [Patch net 0/4] net: hns: bug fixes & optimization for HNS driver Huazhong Tan
` (2 preceding siblings ...)
2018-08-23 3:10 ` [Patch net 3/4] net: hns: fix skb->truesize underestimation Huazhong Tan
@ 2018-08-23 3:10 ` Huazhong Tan
2018-08-23 4:48 ` [Patch net 0/4] net: hns: bug fixes & optimization for HNS driver David Miller
4 siblings, 0 replies; 6+ messages in thread
From: Huazhong Tan @ 2018-08-23 3:10 UTC (permalink / raw)
To: davem; +Cc: netdev, linuxarm, salil.mehta, yisen.zhuang, lipeng321,
tanhuazhong
Update hns to drop the hns_nic_get_headlen function in favour of
eth_get_headlen, and hence also removes now redundant hns_nic_get_headlen.
Signed-off-by: Huazhong Tan <tanhuazhong@huawei.com>
Signed-off-by: Salil Mehta <salil.mehta@huawei.com>
---
drivers/net/ethernet/hisilicon/hns/hns_enet.c | 103 +-------------------------
1 file changed, 1 insertion(+), 102 deletions(-)
diff --git a/drivers/net/ethernet/hisilicon/hns/hns_enet.c b/drivers/net/ethernet/hisilicon/hns/hns_enet.c
index 71bd3bf..02a0ba2 100644
--- a/drivers/net/ethernet/hisilicon/hns/hns_enet.c
+++ b/drivers/net/ethernet/hisilicon/hns/hns_enet.c
@@ -406,107 +406,6 @@ netdev_tx_t hns_nic_net_xmit_hw(struct net_device *ndev,
return NETDEV_TX_BUSY;
}
-/**
- * hns_nic_get_headlen - determine size of header for RSC/LRO/GRO/FCOE
- * @data: pointer to the start of the headers
- * @max: total length of section to find headers in
- *
- * This function is meant to determine the length of headers that will
- * be recognized by hardware for LRO, GRO, and RSC offloads. The main
- * motivation of doing this is to only perform one pull for IPv4 TCP
- * packets so that we can do basic things like calculating the gso_size
- * based on the average data per packet.
- **/
-static unsigned int hns_nic_get_headlen(unsigned char *data, u32 flag,
- unsigned int max_size)
-{
- unsigned char *network;
- u8 hlen;
-
- /* this should never happen, but better safe than sorry */
- if (max_size < ETH_HLEN)
- return max_size;
-
- /* initialize network frame pointer */
- network = data;
-
- /* set first protocol and move network header forward */
- network += ETH_HLEN;
-
- /* handle any vlan tag if present */
- if (hnae_get_field(flag, HNS_RXD_VLAN_M, HNS_RXD_VLAN_S)
- == HNS_RX_FLAG_VLAN_PRESENT) {
- if ((typeof(max_size))(network - data) > (max_size - VLAN_HLEN))
- return max_size;
-
- network += VLAN_HLEN;
- }
-
- /* handle L3 protocols */
- if (hnae_get_field(flag, HNS_RXD_L3ID_M, HNS_RXD_L3ID_S)
- == HNS_RX_FLAG_L3ID_IPV4) {
- if ((typeof(max_size))(network - data) >
- (max_size - sizeof(struct iphdr)))
- return max_size;
-
- /* access ihl as a u8 to avoid unaligned access on ia64 */
- hlen = (network[0] & 0x0F) << 2;
-
- /* verify hlen meets minimum size requirements */
- if (hlen < sizeof(struct iphdr))
- return network - data;
-
- /* record next protocol if header is present */
- } else if (hnae_get_field(flag, HNS_RXD_L3ID_M, HNS_RXD_L3ID_S)
- == HNS_RX_FLAG_L3ID_IPV6) {
- if ((typeof(max_size))(network - data) >
- (max_size - sizeof(struct ipv6hdr)))
- return max_size;
-
- /* record next protocol */
- hlen = sizeof(struct ipv6hdr);
- } else {
- return network - data;
- }
-
- /* relocate pointer to start of L4 header */
- network += hlen;
-
- /* finally sort out TCP/UDP */
- if (hnae_get_field(flag, HNS_RXD_L4ID_M, HNS_RXD_L4ID_S)
- == HNS_RX_FLAG_L4ID_TCP) {
- if ((typeof(max_size))(network - data) >
- (max_size - sizeof(struct tcphdr)))
- return max_size;
-
- /* access doff as a u8 to avoid unaligned access on ia64 */
- hlen = (network[12] & 0xF0) >> 2;
-
- /* verify hlen meets minimum size requirements */
- if (hlen < sizeof(struct tcphdr))
- return network - data;
-
- network += hlen;
- } else if (hnae_get_field(flag, HNS_RXD_L4ID_M, HNS_RXD_L4ID_S)
- == HNS_RX_FLAG_L4ID_UDP) {
- if ((typeof(max_size))(network - data) >
- (max_size - sizeof(struct udphdr)))
- return max_size;
-
- network += sizeof(struct udphdr);
- }
-
- /* If everything has gone correctly network should be the
- * data section of the packet and will be the end of the header.
- * If not then it probably represents the end of the last recognized
- * header.
- */
- if ((typeof(max_size))(network - data) < max_size)
- return network - data;
- else
- return max_size;
-}
-
static void hns_nic_reuse_page(struct sk_buff *skb, int i,
struct hnae_ring *ring, int pull_len,
struct hnae_desc_cb *desc_cb)
@@ -696,7 +595,7 @@ static int hns_nic_poll_rx_skb(struct hns_nic_ring_data *ring_data,
} else {
ring->stats.seg_pkt_cnt++;
- pull_len = hns_nic_get_headlen(va, bnum_flag, HNS_RX_HEAD_SIZE);
+ pull_len = eth_get_headlen(va, HNS_RX_HEAD_SIZE);
memcpy(__skb_put(skb, pull_len), va,
ALIGN(pull_len, sizeof(long)));
--
1.9.1
^ permalink raw reply related [flat|nested] 6+ messages in thread
* Re: [Patch net 0/4] net: hns: bug fixes & optimization for HNS driver
2018-08-23 3:10 [Patch net 0/4] net: hns: bug fixes & optimization for HNS driver Huazhong Tan
` (3 preceding siblings ...)
2018-08-23 3:10 ` [Patch net 4/4] net: hns: use eth_get_headlen interface instead of hns_nic_get_headlen Huazhong Tan
@ 2018-08-23 4:48 ` David Miller
4 siblings, 0 replies; 6+ messages in thread
From: David Miller @ 2018-08-23 4:48 UTC (permalink / raw)
To: tanhuazhong; +Cc: netdev, linuxarm, salil.mehta, yisen.zhuang, lipeng321
From: Huazhong Tan <tanhuazhong@huawei.com>
Date: Thu, 23 Aug 2018 11:10:09 +0800
> This patchset presents some bug fixes found out when
> CONFIG_ARM64_64K_PAGES enable and an optimization for HNS driver.
Series applied, thank you.
^ permalink raw reply [flat|nested] 6+ messages in thread