| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Date: Thu, 16 Jun 2016 16:18:53 +0200
From: Ursula Braun <ubraun@...ux.vnet.ibm.com>
To: davem@...emloft.net
Cc: netdev@...r.kernel.org, linux-s390@...r.kernel.org,
schwidefsky@...ibm.com, heiko.carstens@...ibm.com,
utz.bacher@...ibm.com
Subject: [PATCH net-next 03/13] qeth: clean up condition when tso is used
From: Eugene Crosser <Eugene.Crosser@...ibm.com>
Make conditions under which TSO is activated more stringent.
Make calculation of SBALEs required for the skb more accurate.
Signed-off-by: Eugene Crosser <Eugene.Crosser@...ibm.com>
Signed-off-by: Ursula Braun <ubraun@...ux.vnet.ibm.com>
---
drivers/s390/net/qeth_core.h | 6 ++++++
drivers/s390/net/qeth_l3_main.c | 44 ++++++++++++++++++++++-------------------
2 files changed, 30 insertions(+), 20 deletions(-)
diff --git a/drivers/s390/net/qeth_core.h b/drivers/s390/net/qeth_core.h
index eb8f434..c3ab956 100644
--- a/drivers/s390/net/qeth_core.h
+++ b/drivers/s390/net/qeth_core.h
@@ -19,6 +19,7 @@
#include <linux/seq_file.h>
#include <linux/ethtool.h>
#include <linux/hashtable.h>
+#include <linux/ip.h>
#include <net/ipv6.h>
#include <net/if_inet6.h>
@@ -878,6 +879,11 @@ static inline int qeth_get_ip_version(struct sk_buff *skb)
}
}
+static inline int qeth_get_ip_protocol(struct sk_buff *skb)
+{
+ return ip_hdr(skb)->protocol;
+}
+
static inline void qeth_put_buffer_pool_entry(struct qeth_card *card,
struct qeth_buffer_pool_entry *entry)
{
diff --git a/drivers/s390/net/qeth_l3_main.c b/drivers/s390/net/qeth_l3_main.c
index 7c9968a..86b3bc2 100644
--- a/drivers/s390/net/qeth_l3_main.c
+++ b/drivers/s390/net/qeth_l3_main.c
@@ -2794,10 +2794,10 @@ static void qeth_tso_fill_header(struct qeth_card *card,
}
/**
- * qeth_get_elements_no_tso() - find number of SBALEs for skb data, inc. frags.
- * @card: qeth card structure, to check max. elems.
- * @skb: SKB address
- * @extra_elems: extra elems needed, to check against max.
+ * qeth_l3_get_elements_no_tso() - find number of SBALEs for skb data for tso
+ * @card: qeth card structure, to check max. elems.
+ * @skb: SKB address
+ * @extra_elems: extra elems needed, to check against max.
*
* Returns the number of pages, and thus QDIO buffer elements, needed to cover
* skb data, including linear part and fragments, but excluding TCP header.
@@ -2806,7 +2806,7 @@ static void qeth_tso_fill_header(struct qeth_card *card,
* Returns 0 if it does not.
* Note: extra_elems is not included in the returned result.
*/
-static int qeth_get_elements_no_tso(struct qeth_card *card,
+static int qeth_l3_get_elements_no_tso(struct qeth_card *card,
struct sk_buff *skb, int extra_elems)
{
addr_t tcpdptr = (addr_t)tcp_hdr(skb) + tcp_hdrlen(skb);
@@ -2841,7 +2841,7 @@ static int qeth_l3_hard_start_xmit(struct sk_buff *skb, struct net_device *dev)
qeth_get_priority_queue(card, skb, ipv, cast_type) :
card->qdio.default_out_queue];
int tx_bytes = skb->len;
- bool large_send;
+ bool use_tso;
int data_offset = -1;
int nr_frags;
@@ -2866,10 +2866,12 @@ static int qeth_l3_hard_start_xmit(struct sk_buff *skb, struct net_device *dev)
card->perf_stats.outbound_start_time = qeth_get_micros();
}
- large_send = skb_is_gso(skb);
+ /* Ignore segment size from skb_is_gso(), 1 page is always used. */
+ use_tso = skb_is_gso(skb) &&
+ (qeth_get_ip_protocol(skb) == IPPROTO_TCP) && (ipv == 4);
- if ((card->info.type == QETH_CARD_TYPE_IQD) && (!large_send) &&
- (skb_shinfo(skb)->nr_frags == 0)) {
+ if ((card->info.type == QETH_CARD_TYPE_IQD) &&
+ !skb_is_nonlinear(skb)) {
new_skb = skb;
if (new_skb->protocol == ETH_P_AF_IUCV)
data_offset = 0;
@@ -2913,16 +2915,16 @@ static int qeth_l3_hard_start_xmit(struct sk_buff *skb, struct net_device *dev)
/* fix hardware limitation: as long as we do not have sbal
* chaining we can not send long frag lists
*/
- if (large_send) {
- if (!qeth_get_elements_no_tso(card, new_skb, 1)) {
- if (skb_linearize(new_skb))
- goto tx_drop;
- if (card->options.performance_stats)
- card->perf_stats.tx_lin++;
- }
+ if ((card->info.type != QETH_CARD_TYPE_IQD) &&
+ ((use_tso && !qeth_l3_get_elements_no_tso(card, new_skb, 1)) ||
+ (!use_tso && !qeth_get_elements_no(card, new_skb, 0)))) {
+ if (skb_linearize(new_skb))
+ goto tx_drop;
+ if (card->options.performance_stats)
+ card->perf_stats.tx_lin++;
}
- if (large_send && (cast_type == RTN_UNSPEC)) {
+ if (use_tso) {
hdr = (struct qeth_hdr *)skb_push(new_skb,
sizeof(struct qeth_hdr_tso));
memset(hdr, 0, sizeof(struct qeth_hdr_tso));
@@ -2949,7 +2951,9 @@ static int qeth_l3_hard_start_xmit(struct sk_buff *skb, struct net_device *dev)
qeth_l3_hdr_csum(card, hdr, new_skb);
}
- elements = qeth_get_elements_no(card, new_skb, hdr_elements);
+ elements = use_tso ?
+ qeth_l3_get_elements_no_tso(card, new_skb, hdr_elements) :
+ qeth_get_elements_no(card, new_skb, hdr_elements);
if (!elements) {
if (data_offset >= 0)
kmem_cache_free(qeth_core_header_cache, hdr);
@@ -2959,7 +2963,7 @@ static int qeth_l3_hard_start_xmit(struct sk_buff *skb, struct net_device *dev)
if (card->info.type != QETH_CARD_TYPE_IQD) {
int len;
- if (large_send)
+ if (use_tso)
len = ((unsigned long)tcp_hdr(new_skb) +
tcp_hdrlen(new_skb)) -
(unsigned long)new_skb->data;
@@ -2980,7 +2984,7 @@ static int qeth_l3_hard_start_xmit(struct sk_buff *skb, struct net_device *dev)
dev_kfree_skb_any(skb);
if (card->options.performance_stats) {
nr_frags = skb_shinfo(new_skb)->nr_frags;
- if (large_send) {
+ if (use_tso) {
card->perf_stats.large_send_bytes += tx_bytes;
card->perf_stats.large_send_cnt++;
}
--
2.6.6
Powered by blists - more mailing lists