* [PATCH net-next 01/13] qeth: Include error message for "OS Mismatch"
2016-06-16 14:18 [PATCH net-next 00/13] s390: qeth patches Ursula Braun
@ 2016-06-16 14:18 ` Ursula Braun
2016-06-16 14:18 ` [PATCH net-next 02/13] qeth: refactor calculation of SBALE count Ursula Braun
` (12 subsequent siblings)
13 siblings, 0 replies; 15+ messages in thread
From: Ursula Braun @ 2016-06-16 14:18 UTC (permalink / raw)
To: davem; +Cc: netdev, linux-s390, schwidefsky, heiko.carstens, utz.bacher
From: Eugene Crosser <Eugene.Crosser@ru.ibm.com>
Having understood the semantics of BRIDGEPORT error code 0x0010,
we can introduce a meaningful error message.
Signed-off-by: Eugene Crosser <Eugene.Crosser@ru.ibm.com>
Signed-off-by: Ursula Braun <ubraun@linux.vnet.ibm.com>
---
drivers/s390/net/qeth_l2_main.c | 6 ++++++
1 file changed, 6 insertions(+)
diff --git a/drivers/s390/net/qeth_l2_main.c b/drivers/s390/net/qeth_l2_main.c
index 80b1979..faab0b6 100644
--- a/drivers/s390/net/qeth_l2_main.c
+++ b/drivers/s390/net/qeth_l2_main.c
@@ -1800,6 +1800,12 @@ static int qeth_bridgeport_makerc(struct qeth_card *card,
dev_err(&card->gdev->dev,
"The device is not configured as a Bridge Port\n");
break;
+ case 0x2B10:
+ case 0x0010: /* OS mismatch */
+ rc = -EPERM;
+ dev_err(&card->gdev->dev,
+ "A Bridge Port is already configured by a different operating system\n");
+ break;
case 0x2B14:
case 0x0014: /* Another device is Primary */
switch (setcmd) {
--
2.6.6
^ permalink raw reply related [flat|nested] 15+ messages in thread* [PATCH net-next 02/13] qeth: refactor calculation of SBALE count
2016-06-16 14:18 [PATCH net-next 00/13] s390: qeth patches Ursula Braun
2016-06-16 14:18 ` [PATCH net-next 01/13] qeth: Include error message for "OS Mismatch" Ursula Braun
@ 2016-06-16 14:18 ` Ursula Braun
2016-06-16 14:18 ` [PATCH net-next 03/13] qeth: clean up condition when tso is used Ursula Braun
` (11 subsequent siblings)
13 siblings, 0 replies; 15+ messages in thread
From: Ursula Braun @ 2016-06-16 14:18 UTC (permalink / raw)
To: davem; +Cc: netdev, linux-s390, schwidefsky, heiko.carstens, utz.bacher
From: Eugene Crosser <Eugene.Crosser@ru.ibm.com>
Rewrite the functions that calculate the required number of buffer
elements needed to represent SKB data, to make them hopefully more
comprehensible. Plus a few cleanups.
Signed-off-by: Eugene Crosser <Eugene.Crosser@ru.ibm.com>
Signed-off-by: Ursula Braun <ubraun@linux.vnet.ibm.com>
---
drivers/s390/net/qeth_core.h | 13 +++++++++
drivers/s390/net/qeth_core_main.c | 55 +++++++++++++++++++++++--------------
drivers/s390/net/qeth_l3_main.c | 58 +++++++++++++++++++++++++--------------
3 files changed, 85 insertions(+), 41 deletions(-)
diff --git a/drivers/s390/net/qeth_core.h b/drivers/s390/net/qeth_core.h
index ec2e014..eb8f434 100644
--- a/drivers/s390/net/qeth_core.h
+++ b/drivers/s390/net/qeth_core.h
@@ -844,6 +844,19 @@ struct qeth_trap_id {
/*some helper functions*/
#define QETH_CARD_IFNAME(card) (((card)->dev)? (card)->dev->name : "")
+/**
+ * qeth_get_elements_for_range() - find number of SBALEs to cover range.
+ * @start: Start of the address range.
+ * @end: Address after the end of the range.
+ *
+ * Returns the number of pages, and thus QDIO buffer elements, needed to cover
+ * the specified address range.
+ */
+static inline int qeth_get_elements_for_range(addr_t start, addr_t end)
+{
+ return PFN_UP(end - 1) - PFN_DOWN(start);
+}
+
static inline int qeth_get_micros(void)
{
return (int) (get_tod_clock() >> 12);
diff --git a/drivers/s390/net/qeth_core_main.c b/drivers/s390/net/qeth_core_main.c
index b7b7477..a91a31d 100644
--- a/drivers/s390/net/qeth_core_main.c
+++ b/drivers/s390/net/qeth_core_main.c
@@ -3810,41 +3810,54 @@ int qeth_get_priority_queue(struct qeth_card *card, struct sk_buff *skb,
}
EXPORT_SYMBOL_GPL(qeth_get_priority_queue);
+/**
+ * qeth_get_elements_for_frags() - find number of SBALEs for skb frags.
+ * @skb: SKB address
+ *
+ * Returns the number of pages, and thus QDIO buffer elements, needed to cover
+ * fragmented part of the SKB. Returns zero for linear SKB.
+ */
int qeth_get_elements_for_frags(struct sk_buff *skb)
{
- int cnt, length, e, elements = 0;
- struct skb_frag_struct *frag;
- char *data;
+ int cnt, elements = 0;
for (cnt = 0; cnt < skb_shinfo(skb)->nr_frags; cnt++) {
- frag = &skb_shinfo(skb)->frags[cnt];
- data = (char *)page_to_phys(skb_frag_page(frag)) +
- frag->page_offset;
- length = frag->size;
- e = PFN_UP((unsigned long)data + length - 1) -
- PFN_DOWN((unsigned long)data);
- elements += e;
+ struct skb_frag_struct *frag = &skb_shinfo(skb)->frags[cnt];
+
+ elements += qeth_get_elements_for_range(
+ (addr_t)skb_frag_address(frag),
+ (addr_t)skb_frag_address(frag) + skb_frag_size(frag));
}
return elements;
}
EXPORT_SYMBOL_GPL(qeth_get_elements_for_frags);
+/**
+ * qeth_get_elements_no() - 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.
+ *
+ * Returns the number of pages, and thus QDIO buffer elements, needed to cover
+ * skb data, including linear part and fragments. Checks if the result plus
+ * extra_elems fits under the limit for the card. Returns 0 if it does not.
+ * Note: extra_elems is not included in the returned result.
+ */
int qeth_get_elements_no(struct qeth_card *card,
- struct sk_buff *skb, int elems)
+ struct sk_buff *skb, int extra_elems)
{
- int dlen = skb->len - skb->data_len;
- int elements_needed = PFN_UP((unsigned long)skb->data + dlen - 1) -
- PFN_DOWN((unsigned long)skb->data);
-
- elements_needed += qeth_get_elements_for_frags(skb);
+ int elements = qeth_get_elements_for_range(
+ (addr_t)skb->data,
+ (addr_t)skb->data + skb_headlen(skb)) +
+ qeth_get_elements_for_frags(skb);
- if ((elements_needed + elems) > QETH_MAX_BUFFER_ELEMENTS(card)) {
+ if ((elements + extra_elems) > QETH_MAX_BUFFER_ELEMENTS(card)) {
QETH_DBF_MESSAGE(2, "Invalid size of IP packet "
"(Number=%d / Length=%d). Discarded.\n",
- (elements_needed+elems), skb->len);
+ elements + extra_elems, skb->len);
return 0;
}
- return elements_needed;
+ return elements;
}
EXPORT_SYMBOL_GPL(qeth_get_elements_no);
@@ -3859,7 +3872,7 @@ int qeth_hdr_chk_and_bounce(struct sk_buff *skb, struct qeth_hdr **hdr, int len)
rest = len - inpage;
if (rest > hroom)
return 1;
- memmove(skb->data - rest, skb->data, skb->len - skb->data_len);
+ memmove(skb->data - rest, skb->data, skb_headlen(skb));
skb->data -= rest;
skb->tail -= rest;
*hdr = (struct qeth_hdr *)skb->data;
@@ -3873,7 +3886,7 @@ static inline void __qeth_fill_buffer(struct sk_buff *skb,
struct qdio_buffer *buffer, int is_tso, int *next_element_to_fill,
int offset)
{
- int length = skb->len - skb->data_len;
+ int length = skb_headlen(skb);
int length_here;
int element;
char *data;
diff --git a/drivers/s390/net/qeth_l3_main.c b/drivers/s390/net/qeth_l3_main.c
index ac54433..7c9968a 100644
--- a/drivers/s390/net/qeth_l3_main.c
+++ b/drivers/s390/net/qeth_l3_main.c
@@ -2793,15 +2793,34 @@ static void qeth_tso_fill_header(struct qeth_card *card,
}
}
-static inline int qeth_l3_tso_elements(struct sk_buff *skb)
+/**
+ * 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.
+ *
+ * Returns the number of pages, and thus QDIO buffer elements, needed to cover
+ * skb data, including linear part and fragments, but excluding TCP header.
+ * (Exclusion of TCP header distinguishes it from qeth_get_elements_no().)
+ * Checks if the result plus extra_elems fits under the limit for the 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,
+ struct sk_buff *skb, int extra_elems)
{
- unsigned long tcpd = (unsigned long)tcp_hdr(skb) +
- tcp_hdr(skb)->doff * 4;
- int tcpd_len = skb_headlen(skb) - (tcpd - (unsigned long)skb->data);
- int elements = PFN_UP(tcpd + tcpd_len - 1) - PFN_DOWN(tcpd);
-
- elements += qeth_get_elements_for_frags(skb);
+ addr_t tcpdptr = (addr_t)tcp_hdr(skb) + tcp_hdrlen(skb);
+ int elements = qeth_get_elements_for_range(
+ tcpdptr,
+ (addr_t)skb->data + skb_headlen(skb)) +
+ qeth_get_elements_for_frags(skb);
+ if ((elements + extra_elems) > QETH_MAX_BUFFER_ELEMENTS(card)) {
+ QETH_DBF_MESSAGE(2,
+ "Invalid size of TSO IP packet (Number=%d / Length=%d). Discarded.\n",
+ elements + extra_elems, skb->len);
+ return 0;
+ }
return elements;
}
@@ -2810,8 +2829,8 @@ static int qeth_l3_hard_start_xmit(struct sk_buff *skb, struct net_device *dev)
int rc;
u16 *tag;
struct qeth_hdr *hdr = NULL;
- int elements_needed = 0;
- int elems;
+ int hdr_elements = 0;
+ int elements;
struct qeth_card *card = dev->ml_priv;
struct sk_buff *new_skb = NULL;
int ipv = qeth_get_ip_version(skb);
@@ -2859,7 +2878,7 @@ static int qeth_l3_hard_start_xmit(struct sk_buff *skb, struct net_device *dev)
hdr = kmem_cache_alloc(qeth_core_header_cache, GFP_ATOMIC);
if (!hdr)
goto tx_drop;
- elements_needed++;
+ hdr_elements++;
} else {
/* create a clone with writeable headroom */
new_skb = skb_realloc_headroom(skb, sizeof(struct qeth_hdr_tso)
@@ -2895,7 +2914,7 @@ static int qeth_l3_hard_start_xmit(struct sk_buff *skb, struct net_device *dev)
* chaining we can not send long frag lists
*/
if (large_send) {
- if (qeth_l3_tso_elements(new_skb) + 1 > 16) {
+ if (!qeth_get_elements_no_tso(card, new_skb, 1)) {
if (skb_linearize(new_skb))
goto tx_drop;
if (card->options.performance_stats)
@@ -2909,7 +2928,7 @@ static int qeth_l3_hard_start_xmit(struct sk_buff *skb, struct net_device *dev)
memset(hdr, 0, sizeof(struct qeth_hdr_tso));
qeth_l3_fill_header(card, hdr, new_skb, ipv, cast_type);
qeth_tso_fill_header(card, hdr, new_skb);
- elements_needed++;
+ hdr_elements++;
} else {
if (data_offset < 0) {
hdr = (struct qeth_hdr *)skb_push(new_skb,
@@ -2930,31 +2949,29 @@ static int qeth_l3_hard_start_xmit(struct sk_buff *skb, struct net_device *dev)
qeth_l3_hdr_csum(card, hdr, new_skb);
}
- elems = qeth_get_elements_no(card, new_skb, elements_needed);
- if (!elems) {
+ elements = qeth_get_elements_no(card, new_skb, hdr_elements);
+ if (!elements) {
if (data_offset >= 0)
kmem_cache_free(qeth_core_header_cache, hdr);
goto tx_drop;
}
- elements_needed += elems;
- nr_frags = skb_shinfo(new_skb)->nr_frags;
+ elements += hdr_elements;
if (card->info.type != QETH_CARD_TYPE_IQD) {
int len;
if (large_send)
len = ((unsigned long)tcp_hdr(new_skb) +
- tcp_hdr(new_skb)->doff * 4) -
+ tcp_hdrlen(new_skb)) -
(unsigned long)new_skb->data;
else
len = sizeof(struct qeth_hdr_layer3);
if (qeth_hdr_chk_and_bounce(new_skb, &hdr, len))
goto tx_drop;
- rc = qeth_do_send_packet(card, queue, new_skb, hdr,
- elements_needed);
+ rc = qeth_do_send_packet(card, queue, new_skb, hdr, elements);
} else
rc = qeth_do_send_packet_fast(card, queue, new_skb, hdr,
- elements_needed, data_offset, 0);
+ elements, data_offset, 0);
if (!rc) {
card->stats.tx_packets++;
@@ -2962,6 +2979,7 @@ static int qeth_l3_hard_start_xmit(struct sk_buff *skb, struct net_device *dev)
if (new_skb != skb)
dev_kfree_skb_any(skb);
if (card->options.performance_stats) {
+ nr_frags = skb_shinfo(new_skb)->nr_frags;
if (large_send) {
card->perf_stats.large_send_bytes += tx_bytes;
card->perf_stats.large_send_cnt++;
--
2.6.6
^ permalink raw reply related [flat|nested] 15+ messages in thread* [PATCH net-next 03/13] qeth: clean up condition when tso is used
2016-06-16 14:18 [PATCH net-next 00/13] s390: qeth patches Ursula Braun
2016-06-16 14:18 ` [PATCH net-next 01/13] qeth: Include error message for "OS Mismatch" Ursula Braun
2016-06-16 14:18 ` [PATCH net-next 02/13] qeth: refactor calculation of SBALE count Ursula Braun
@ 2016-06-16 14:18 ` Ursula Braun
2016-06-16 14:18 ` [PATCH net-next 04/13] qeth: fill netdevice->gso_* attributes accurately Ursula Braun
` (10 subsequent siblings)
13 siblings, 0 replies; 15+ messages in thread
From: Ursula Braun @ 2016-06-16 14:18 UTC (permalink / raw)
To: davem; +Cc: netdev, linux-s390, schwidefsky, heiko.carstens, utz.bacher
From: Eugene Crosser <Eugene.Crosser@ru.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@ru.ibm.com>
Signed-off-by: Ursula Braun <ubraun@linux.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
^ permalink raw reply related [flat|nested] 15+ messages in thread* [PATCH net-next 04/13] qeth: fill netdevice->gso_* attributes accurately
2016-06-16 14:18 [PATCH net-next 00/13] s390: qeth patches Ursula Braun
` (2 preceding siblings ...)
2016-06-16 14:18 ` [PATCH net-next 03/13] qeth: clean up condition when tso is used Ursula Braun
@ 2016-06-16 14:18 ` Ursula Braun
2016-06-16 14:18 ` [PATCH net-next 05/13] qeth: enable scatter/gather in layer 2 mode Ursula Braun
` (9 subsequent siblings)
13 siblings, 0 replies; 15+ messages in thread
From: Ursula Braun @ 2016-06-16 14:18 UTC (permalink / raw)
To: davem; +Cc: netdev, linux-s390, schwidefsky, heiko.carstens, utz.bacher
From: Eugene Crosser <Eugene.Crosser@ru.ibm.com>
Use QETH_MAX_BUFFER_ELEMENTS(card) instead of constant 16.
Also fill gso_max_segs and gso_min_segs.
Signed-off-by: Eugene Crosser <Eugene.Crosser@ru.ibm.com>
Signed-off-by: Ursula Braun <ubraun@linux.vnet.ibm.com>
---
drivers/s390/net/qeth_l3_main.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/drivers/s390/net/qeth_l3_main.c b/drivers/s390/net/qeth_l3_main.c
index 86b3bc2..76ddd18 100644
--- a/drivers/s390/net/qeth_l3_main.c
+++ b/drivers/s390/net/qeth_l3_main.c
@@ -3217,7 +3217,9 @@ static int qeth_l3_setup_netdev(struct qeth_card *card)
NETIF_F_HW_VLAN_CTAG_RX |
NETIF_F_HW_VLAN_CTAG_FILTER;
netif_keep_dst(card->dev);
- card->dev->gso_max_size = 15 * PAGE_SIZE;
+ card->dev->gso_max_size = (QETH_MAX_BUFFER_ELEMENTS(card) - 1) *
+ PAGE_SIZE;
+ card->dev->gso_max_segs = (QETH_MAX_BUFFER_ELEMENTS(card) - 1);
SET_NETDEV_DEV(card->dev, &card->gdev->dev);
netif_napi_add(card->dev, &card->napi, qeth_l3_poll, QETH_NAPI_WEIGHT);
--
2.6.6
^ permalink raw reply related [flat|nested] 15+ messages in thread* [PATCH net-next 05/13] qeth: enable scatter/gather in layer 2 mode
2016-06-16 14:18 [PATCH net-next 00/13] s390: qeth patches Ursula Braun
` (3 preceding siblings ...)
2016-06-16 14:18 ` [PATCH net-next 04/13] qeth: fill netdevice->gso_* attributes accurately Ursula Braun
@ 2016-06-16 14:18 ` Ursula Braun
2016-06-16 14:18 ` [PATCH net-next 06/13] qeth: enable scatter/gather by default Ursula Braun
` (8 subsequent siblings)
13 siblings, 0 replies; 15+ messages in thread
From: Ursula Braun @ 2016-06-16 14:18 UTC (permalink / raw)
To: davem; +Cc: netdev, linux-s390, schwidefsky, heiko.carstens, utz.bacher
From: Eugene Crosser <Eugene.Crosser@ru.ibm.com>
The patch enables NETIF_F_SG flag for OSA in layer 2 mode.
It also adds performance accounting for fragmented sends,
adds a conditional skb_linearize() attempt if the skb had
too many fragments for QDIO SBAL, and fills netdevice->gso_*
attributes.
Signed-off-by: Eugene Crosser <Eugene.Crosser@ru.ibm.com>
Signed-off-by: Ursula Braun <ubraun@linux.vnet.ibm.com>
Reviewed-by: Lakhvich Dmitriy <ldmitriy@ru.ibm.com>
Reviewed-by: Thomas Richter <tmricht@de.ibm.com>
---
drivers/s390/net/qeth_l2_main.c | 26 +++++++++++++++++++++++++-
1 file changed, 25 insertions(+), 1 deletion(-)
diff --git a/drivers/s390/net/qeth_l2_main.c b/drivers/s390/net/qeth_l2_main.c
index faab0b6..ec163e4 100644
--- a/drivers/s390/net/qeth_l2_main.c
+++ b/drivers/s390/net/qeth_l2_main.c
@@ -869,6 +869,7 @@ static int qeth_l2_hard_start_xmit(struct sk_buff *skb, struct net_device *dev)
int data_offset = -1;
int elements_needed = 0;
int hd_len = 0;
+ int nr_frags;
if (card->qdio.do_prio_queueing || (cast_type &&
card->info.is_multicast_different))
@@ -892,6 +893,17 @@ static int qeth_l2_hard_start_xmit(struct sk_buff *skb, struct net_device *dev)
}
netif_stop_queue(dev);
+ /* fix hardware limitation: as long as we do not have sbal
+ * chaining we can not send long frag lists
+ */
+ if ((card->info.type != QETH_CARD_TYPE_IQD) &&
+ !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 (card->info.type == QETH_CARD_TYPE_OSN)
hdr = (struct qeth_hdr *)skb->data;
else {
@@ -943,6 +955,14 @@ static int qeth_l2_hard_start_xmit(struct sk_buff *skb, struct net_device *dev)
if (!rc) {
card->stats.tx_packets++;
card->stats.tx_bytes += tx_bytes;
+ if (card->options.performance_stats) {
+ nr_frags = skb_shinfo(new_skb)->nr_frags;
+ if (nr_frags) {
+ card->perf_stats.sg_skbs_sent++;
+ /* nr_frags + skb->data */
+ card->perf_stats.sg_frags_sent += nr_frags + 1;
+ }
+ }
if (new_skb != skb)
dev_kfree_skb_any(skb);
rc = NETDEV_TX_OK;
@@ -1118,12 +1138,16 @@ static int qeth_l2_setup_netdev(struct qeth_card *card)
&qeth_l2_ethtool_ops : &qeth_l2_osn_ops;
card->dev->features |= NETIF_F_HW_VLAN_CTAG_FILTER;
if (card->info.type == QETH_CARD_TYPE_OSD && !card->info.guestlan) {
- card->dev->hw_features = NETIF_F_IP_CSUM | NETIF_F_RXCSUM;
+ card->dev->hw_features = NETIF_F_IP_CSUM | NETIF_F_RXCSUM |
+ NETIF_F_SG;
/* Turn on RX offloading per default */
card->dev->features |= NETIF_F_RXCSUM;
}
card->info.broadcast_capable = 1;
qeth_l2_request_initial_mac(card);
+ card->dev->gso_max_size = (QETH_MAX_BUFFER_ELEMENTS(card) - 1) *
+ PAGE_SIZE;
+ card->dev->gso_max_segs = (QETH_MAX_BUFFER_ELEMENTS(card) - 1);
SET_NETDEV_DEV(card->dev, &card->gdev->dev);
netif_napi_add(card->dev, &card->napi, qeth_l2_poll, QETH_NAPI_WEIGHT);
netif_carrier_off(card->dev);
--
2.6.6
^ permalink raw reply related [flat|nested] 15+ messages in thread* [PATCH net-next 06/13] qeth: enable scatter/gather by default
2016-06-16 14:18 [PATCH net-next 00/13] s390: qeth patches Ursula Braun
` (4 preceding siblings ...)
2016-06-16 14:18 ` [PATCH net-next 05/13] qeth: enable scatter/gather in layer 2 mode Ursula Braun
@ 2016-06-16 14:18 ` Ursula Braun
2016-06-16 14:18 ` [PATCH net-next 07/13] qeth: introduce linearization fail count to stats Ursula Braun
` (7 subsequent siblings)
13 siblings, 0 replies; 15+ messages in thread
From: Ursula Braun @ 2016-06-16 14:18 UTC (permalink / raw)
To: davem; +Cc: netdev, linux-s390, schwidefsky, heiko.carstens, utz.bacher
From: Eugene Crosser <Eugene.Crosser@ru.ibm.com>
Set scatter/gather ON by default on OSA, for both layer 2 and
layer 3 modes. We always use fragmentation over QDIO anyway,
so let the upper layers of the stack take advantage of that.
Signed-off-by: Eugene Crosser <Eugene.Crosser@ru.ibm.com>
Signed-off-by: Ursula Braun <ubraun@linux.vnet.ibm.com>
Reviewed-by: Lakhvich Dmitriy <ldmitriy@ru.ibm.com>
Reviewed-by: Thomas Richter <tmricht@de.ibm.com>
---
drivers/s390/net/qeth_l2_main.c | 4 ++--
drivers/s390/net/qeth_l3_main.c | 3 ++-
2 files changed, 4 insertions(+), 3 deletions(-)
diff --git a/drivers/s390/net/qeth_l2_main.c b/drivers/s390/net/qeth_l2_main.c
index ec163e4..e24627f 100644
--- a/drivers/s390/net/qeth_l2_main.c
+++ b/drivers/s390/net/qeth_l2_main.c
@@ -1140,8 +1140,8 @@ static int qeth_l2_setup_netdev(struct qeth_card *card)
if (card->info.type == QETH_CARD_TYPE_OSD && !card->info.guestlan) {
card->dev->hw_features = NETIF_F_IP_CSUM | NETIF_F_RXCSUM |
NETIF_F_SG;
- /* Turn on RX offloading per default */
- card->dev->features |= NETIF_F_RXCSUM;
+ /* Turn on RX offloading and SG per default */
+ card->dev->features |= NETIF_F_RXCSUM | NETIF_F_SG;
}
card->info.broadcast_capable = 1;
qeth_l2_request_initial_mac(card);
diff --git a/drivers/s390/net/qeth_l3_main.c b/drivers/s390/net/qeth_l3_main.c
index 76ddd18..0afdb14 100644
--- a/drivers/s390/net/qeth_l3_main.c
+++ b/drivers/s390/net/qeth_l3_main.c
@@ -3191,7 +3191,8 @@ static int qeth_l3_setup_netdev(struct qeth_card *card)
card->dev->hw_features = NETIF_F_SG |
NETIF_F_RXCSUM | NETIF_F_IP_CSUM |
NETIF_F_TSO;
- card->dev->features = NETIF_F_RXCSUM;
+ card->dev->features = NETIF_F_RXCSUM |
+ NETIF_F_SG;
}
}
} else if (card->info.type == QETH_CARD_TYPE_IQD) {
--
2.6.6
^ permalink raw reply related [flat|nested] 15+ messages in thread* [PATCH net-next 07/13] qeth: introduce linearization fail count to stats
2016-06-16 14:18 [PATCH net-next 00/13] s390: qeth patches Ursula Braun
` (5 preceding siblings ...)
2016-06-16 14:18 ` [PATCH net-next 06/13] qeth: enable scatter/gather by default Ursula Braun
@ 2016-06-16 14:18 ` Ursula Braun
2016-06-16 14:18 ` [PATCH net-next 08/13] qeth: optimize IP handling in rx_mode callback Ursula Braun
` (6 subsequent siblings)
13 siblings, 0 replies; 15+ messages in thread
From: Ursula Braun @ 2016-06-16 14:18 UTC (permalink / raw)
To: davem; +Cc: netdev, linux-s390, schwidefsky, heiko.carstens, utz.bacher
From: Eugene Crosser <Eugene.Crosser@ru.ibm.com>
When skb data touches too many pages, skb_linearize() is called
opportunistically in the hope that less pages will be required
for a big linear buffer than for multiple fragments. This patch
intoduces a separate counter in ethtool statistics structure
representing _failed_ linearization attempts.
Signed-off-by: Eugene Crosser <Eugene.Crosser@ru.ibm.com>
Signed-off-by: Ursula Braun <ubraun@linux.vnet.ibm.com>
Reviewed-by: Lakhvich Dmitriy <ldmitriy@ru.ibm.com>
Reviewed-by: Thomas Richter <tmricht@de.ibm.com>
---
drivers/s390/net/qeth_core.h | 1 +
drivers/s390/net/qeth_core_main.c | 6 ++++--
drivers/s390/net/qeth_l2_main.c | 12 +++++++++---
drivers/s390/net/qeth_l3_main.c | 12 +++++++++---
4 files changed, 23 insertions(+), 8 deletions(-)
diff --git a/drivers/s390/net/qeth_core.h b/drivers/s390/net/qeth_core.h
index c3ab956..b4f14e4 100644
--- a/drivers/s390/net/qeth_core.h
+++ b/drivers/s390/net/qeth_core.h
@@ -145,6 +145,7 @@ struct qeth_perf_stats {
unsigned int sg_alloc_page_rx;
unsigned int tx_csum;
unsigned int tx_lin;
+ unsigned int tx_linfail;
};
/* Routing stuff */
diff --git a/drivers/s390/net/qeth_core_main.c b/drivers/s390/net/qeth_core_main.c
index a91a31d..9806ee0 100644
--- a/drivers/s390/net/qeth_core_main.c
+++ b/drivers/s390/net/qeth_core_main.c
@@ -5801,6 +5801,7 @@ static struct {
{"tx do_QDIO count"},
{"tx csum"},
{"tx lin"},
+ {"tx linfail"},
{"cq handler count"},
{"cq handler time"}
};
@@ -5861,8 +5862,9 @@ void qeth_core_get_ethtool_stats(struct net_device *dev,
data[32] = card->perf_stats.outbound_do_qdio_cnt;
data[33] = card->perf_stats.tx_csum;
data[34] = card->perf_stats.tx_lin;
- data[35] = card->perf_stats.cq_cnt;
- data[36] = card->perf_stats.cq_time;
+ data[35] = card->perf_stats.tx_linfail;
+ data[36] = card->perf_stats.cq_cnt;
+ data[37] = card->perf_stats.cq_time;
}
EXPORT_SYMBOL_GPL(qeth_core_get_ethtool_stats);
diff --git a/drivers/s390/net/qeth_l2_main.c b/drivers/s390/net/qeth_l2_main.c
index e24627f..08b5fa9 100644
--- a/drivers/s390/net/qeth_l2_main.c
+++ b/drivers/s390/net/qeth_l2_main.c
@@ -898,10 +898,16 @@ static int qeth_l2_hard_start_xmit(struct sk_buff *skb, struct net_device *dev)
*/
if ((card->info.type != QETH_CARD_TYPE_IQD) &&
!qeth_get_elements_no(card, new_skb, 0)) {
- if (skb_linearize(new_skb))
+ int lin_rc = skb_linearize(new_skb);
+
+ if (card->options.performance_stats) {
+ if (lin_rc)
+ card->perf_stats.tx_linfail++;
+ else
+ card->perf_stats.tx_lin++;
+ }
+ if (lin_rc)
goto tx_drop;
- if (card->options.performance_stats)
- card->perf_stats.tx_lin++;
}
if (card->info.type == QETH_CARD_TYPE_OSN)
diff --git a/drivers/s390/net/qeth_l3_main.c b/drivers/s390/net/qeth_l3_main.c
index 0afdb14..51077fb 100644
--- a/drivers/s390/net/qeth_l3_main.c
+++ b/drivers/s390/net/qeth_l3_main.c
@@ -2918,10 +2918,16 @@ static int qeth_l3_hard_start_xmit(struct sk_buff *skb, struct net_device *dev)
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))
+ int lin_rc = skb_linearize(new_skb);
+
+ if (card->options.performance_stats) {
+ if (lin_rc)
+ card->perf_stats.tx_linfail++;
+ else
+ card->perf_stats.tx_lin++;
+ }
+ if (lin_rc)
goto tx_drop;
- if (card->options.performance_stats)
- card->perf_stats.tx_lin++;
}
if (use_tso) {
--
2.6.6
^ permalink raw reply related [flat|nested] 15+ messages in thread* [PATCH net-next 08/13] qeth: optimize IP handling in rx_mode callback
2016-06-16 14:18 [PATCH net-next 00/13] s390: qeth patches Ursula Braun
` (6 preceding siblings ...)
2016-06-16 14:18 ` [PATCH net-next 07/13] qeth: introduce linearization fail count to stats Ursula Braun
@ 2016-06-16 14:18 ` Ursula Braun
2016-06-16 14:18 ` [PATCH net-next 09/13] qeth layer 2 and layer 3 common feature handling Ursula Braun
` (5 subsequent siblings)
13 siblings, 0 replies; 15+ messages in thread
From: Ursula Braun @ 2016-06-16 14:18 UTC (permalink / raw)
To: davem; +Cc: netdev, linux-s390, schwidefsky, heiko.carstens, utz.bacher
From: Lakhvich Dmitriy <LDmitriy@ru.ibm.com>
In layer3 mode of the qeth driver, multicast IP addresses
from struct net_device and other type of IP addresses
from other sources require mapping to the OSA-card.
This patch simplifies the IP address mapping logic, and changes imple-
mentation of ndo_set_rx_mode callback and ip notifier events.
Addresses are stored in private hashtables instead of lists now.
It allows hardware registration/removal for new/deleted multicast
addresses only.
Signed-off-by: Lakhvich Dmitriy <ldmitriy@ru.ibm.com>
Signed-off-by: Ursula Braun <ubraun@linux.vnet.ibm.com>
Reviewed-by: Evgeny Cherkashin <Eugene.Crosser@ru.ibm.com>
Reviewed-by: Thomas Richter <tmricht@linux.vnet.ibm.com>
---
drivers/s390/net/qeth_core.h | 20 +-
drivers/s390/net/qeth_core_main.c | 10 -
drivers/s390/net/qeth_l2.h | 7 +
drivers/s390/net/qeth_l2_main.c | 12 +-
drivers/s390/net/qeth_l3.h | 31 +-
drivers/s390/net/qeth_l3_main.c | 782 +++++++++++++++++++-------------------
drivers/s390/net/qeth_l3_sys.c | 78 ++--
7 files changed, 482 insertions(+), 458 deletions(-)
diff --git a/drivers/s390/net/qeth_core.h b/drivers/s390/net/qeth_core.h
index b4f14e4..ab0a171 100644
--- a/drivers/s390/net/qeth_core.h
+++ b/drivers/s390/net/qeth_core.h
@@ -561,7 +561,6 @@ enum qeth_ip_types {
QETH_IP_TYPE_NORMAL,
QETH_IP_TYPE_VIPA,
QETH_IP_TYPE_RXIP,
- QETH_IP_TYPE_DEL_ALL_MC,
};
enum qeth_cmd_buffer_state {
@@ -742,17 +741,10 @@ struct qeth_vlan_vid {
unsigned short vid;
};
-enum qeth_mac_disposition {
- QETH_DISP_MAC_DELETE = 0,
- QETH_DISP_MAC_DO_NOTHING = 1,
- QETH_DISP_MAC_ADD = 2,
-};
-
-struct qeth_mac {
- u8 mac_addr[OSA_ADDR_LEN];
- u8 is_uc:1;
- u8 disp_flag:2;
- struct hlist_node hnode;
+enum qeth_addr_disposition {
+ QETH_DISP_ADDR_DELETE = 0,
+ QETH_DISP_ADDR_DO_NOTHING = 1,
+ QETH_DISP_ADDR_ADD = 2,
};
struct qeth_rx {
@@ -800,6 +792,8 @@ struct qeth_card {
unsigned long active_vlans[BITS_TO_LONGS(VLAN_N_VID)];
struct list_head vid_list;
DECLARE_HASHTABLE(mac_htable, 4);
+ DECLARE_HASHTABLE(ip_htable, 4);
+ DECLARE_HASHTABLE(ip_mc_htable, 4);
struct work_struct kernel_thread_starter;
spinlock_t thread_mask_lock;
unsigned long thread_start_mask;
@@ -807,8 +801,6 @@ struct qeth_card {
unsigned long thread_running_mask;
struct task_struct *recovery_task;
spinlock_t ip_lock;
- struct list_head ip_list;
- struct list_head *ip_tbd_list;
struct qeth_ipato ipato;
struct list_head cmd_waiter_list;
/* QDIO buffer handling */
diff --git a/drivers/s390/net/qeth_core_main.c b/drivers/s390/net/qeth_core_main.c
index 9806ee0..ede9ed8 100644
--- a/drivers/s390/net/qeth_core_main.c
+++ b/drivers/s390/net/qeth_core_main.c
@@ -1464,8 +1464,6 @@ static int qeth_setup_card(struct qeth_card *card)
card->thread_allowed_mask = 0;
card->thread_running_mask = 0;
INIT_WORK(&card->kernel_thread_starter, qeth_start_kernel_thread);
- INIT_LIST_HEAD(&card->ip_list);
- INIT_LIST_HEAD(card->ip_tbd_list);
INIT_LIST_HEAD(&card->cmd_waiter_list);
init_waitqueue_head(&card->wait_q);
/* initial options */
@@ -1500,11 +1498,6 @@ static struct qeth_card *qeth_alloc_card(void)
if (!card)
goto out;
QETH_DBF_HEX(SETUP, 2, &card, sizeof(void *));
- card->ip_tbd_list = kzalloc(sizeof(struct list_head), GFP_KERNEL);
- if (!card->ip_tbd_list) {
- QETH_DBF_TEXT(SETUP, 0, "iptbdnom");
- goto out_card;
- }
if (qeth_setup_channel(&card->read))
goto out_ip;
if (qeth_setup_channel(&card->write))
@@ -1517,8 +1510,6 @@ static struct qeth_card *qeth_alloc_card(void)
out_channel:
qeth_clean_channel(&card->read);
out_ip:
- kfree(card->ip_tbd_list);
-out_card:
kfree(card);
out:
return NULL;
@@ -4980,7 +4971,6 @@ static void qeth_core_free_card(struct qeth_card *card)
qeth_clean_channel(&card->write);
if (card->dev)
free_netdev(card->dev);
- kfree(card->ip_tbd_list);
qeth_free_qdio_buffers(card);
unregister_service_level(&card->qeth_service_level);
kfree(card);
diff --git a/drivers/s390/net/qeth_l2.h b/drivers/s390/net/qeth_l2.h
index 0767556..29d9fb3 100644
--- a/drivers/s390/net/qeth_l2.h
+++ b/drivers/s390/net/qeth_l2.h
@@ -12,4 +12,11 @@ int qeth_l2_create_device_attributes(struct device *);
void qeth_l2_remove_device_attributes(struct device *);
void qeth_l2_setup_bridgeport_attrs(struct qeth_card *card);
+struct qeth_mac {
+ u8 mac_addr[OSA_ADDR_LEN];
+ u8 is_uc:1;
+ u8 disp_flag:2;
+ struct hlist_node hnode;
+};
+
#endif /* __QETH_L2_H__ */
diff --git a/drivers/s390/net/qeth_l2_main.c b/drivers/s390/net/qeth_l2_main.c
index 08b5fa9..2a331d1 100644
--- a/drivers/s390/net/qeth_l2_main.c
+++ b/drivers/s390/net/qeth_l2_main.c
@@ -780,7 +780,7 @@ qeth_l2_add_mac(struct qeth_card *card, struct netdev_hw_addr *ha, u8 is_uc)
qeth_l2_mac_hash(ha->addr)) {
if (is_uc == mac->is_uc &&
!memcmp(ha->addr, mac->mac_addr, OSA_ADDR_LEN)) {
- mac->disp_flag = QETH_DISP_MAC_DO_NOTHING;
+ mac->disp_flag = QETH_DISP_ADDR_DO_NOTHING;
return;
}
}
@@ -792,7 +792,7 @@ qeth_l2_add_mac(struct qeth_card *card, struct netdev_hw_addr *ha, u8 is_uc)
memcpy(mac->mac_addr, ha->addr, OSA_ADDR_LEN);
mac->is_uc = is_uc;
- mac->disp_flag = QETH_DISP_MAC_ADD;
+ mac->disp_flag = QETH_DISP_ADDR_ADD;
hash_add(card->mac_htable, &mac->hnode,
qeth_l2_mac_hash(mac->mac_addr));
@@ -825,7 +825,7 @@ static void qeth_l2_set_rx_mode(struct net_device *dev)
qeth_l2_add_mac(card, ha, 1);
hash_for_each_safe(card->mac_htable, i, tmp, mac, hnode) {
- if (mac->disp_flag == QETH_DISP_MAC_DELETE) {
+ if (mac->disp_flag == QETH_DISP_ADDR_DELETE) {
if (!mac->is_uc)
rc = qeth_l2_send_delgroupmac(card,
mac->mac_addr);
@@ -837,15 +837,15 @@ static void qeth_l2_set_rx_mode(struct net_device *dev)
hash_del(&mac->hnode);
kfree(mac);
- } else if (mac->disp_flag == QETH_DISP_MAC_ADD) {
+ } else if (mac->disp_flag == QETH_DISP_ADDR_ADD) {
rc = qeth_l2_write_mac(card, mac);
if (rc) {
hash_del(&mac->hnode);
kfree(mac);
} else
- mac->disp_flag = QETH_DISP_MAC_DELETE;
+ mac->disp_flag = QETH_DISP_ADDR_DELETE;
} else
- mac->disp_flag = QETH_DISP_MAC_DELETE;
+ mac->disp_flag = QETH_DISP_ADDR_DELETE;
}
spin_unlock_bh(&card->mclock);
diff --git a/drivers/s390/net/qeth_l3.h b/drivers/s390/net/qeth_l3.h
index 551a4b4..26f7953 100644
--- a/drivers/s390/net/qeth_l3.h
+++ b/drivers/s390/net/qeth_l3.h
@@ -10,16 +10,23 @@
#define __QETH_L3_H__
#include "qeth_core.h"
+#include <linux/hashtable.h>
#define QETH_SNIFF_AVAIL 0x0008
struct qeth_ipaddr {
- struct list_head entry;
+ struct hlist_node hnode;
enum qeth_ip_types type;
enum qeth_ipa_setdelip_flags set_flags;
enum qeth_ipa_setdelip_flags del_flags;
- int is_multicast;
- int users;
+ u8 is_multicast:1;
+ u8 in_progress:1;
+ u8 disp_flag:2;
+
+ /* is changed only for normal ip addresses
+ * for non-normal addresses it always is 1
+ */
+ int ref_counter;
enum qeth_prot_versions proto;
unsigned char mac[OSA_ADDR_LEN];
union {
@@ -32,7 +39,24 @@ struct qeth_ipaddr {
unsigned int pfxlen;
} a6;
} u;
+
};
+static inline u64 qeth_l3_ipaddr_hash(struct qeth_ipaddr *addr)
+{
+ u64 ret = 0;
+ u8 *point;
+
+ if (addr->proto == QETH_PROT_IPV6) {
+ point = (u8 *) &addr->u.a6.addr;
+ ret = get_unaligned((u64 *)point) ^
+ get_unaligned((u64 *) (point + 8));
+ }
+ if (addr->proto == QETH_PROT_IPV4) {
+ point = (u8 *) &addr->u.a4.addr;
+ ret = get_unaligned((u32 *) point);
+ }
+ return ret;
+}
struct qeth_ipato_entry {
struct list_head entry;
@@ -60,6 +84,5 @@ int qeth_l3_is_addr_covered_by_ipato(struct qeth_card *, struct qeth_ipaddr *);
struct qeth_ipaddr *qeth_l3_get_addr_buffer(enum qeth_prot_versions);
int qeth_l3_add_ip(struct qeth_card *, struct qeth_ipaddr *);
int qeth_l3_delete_ip(struct qeth_card *, struct qeth_ipaddr *);
-void qeth_l3_set_ip_addr_list(struct qeth_card *);
#endif /* __QETH_L3_H__ */
diff --git a/drivers/s390/net/qeth_l3_main.c b/drivers/s390/net/qeth_l3_main.c
index 51077fb..6e7d06c 100644
--- a/drivers/s390/net/qeth_l3_main.c
+++ b/drivers/s390/net/qeth_l3_main.c
@@ -30,6 +30,7 @@
#include <net/ip6_fib.h>
#include <net/ip6_checksum.h>
#include <net/iucv/af_iucv.h>
+#include <linux/hashtable.h>
#include "qeth_l3.h"
@@ -57,7 +58,7 @@ static int qeth_l3_isxdigit(char *buf)
static void qeth_l3_ipaddr4_to_string(const __u8 *addr, char *buf)
{
- sprintf(buf, "%i.%i.%i.%i", addr[0], addr[1], addr[2], addr[3]);
+ sprintf(buf, "%pI4", addr);
}
static int qeth_l3_string_to_ipaddr4(const char *buf, __u8 *addr)
@@ -204,104 +205,129 @@ int qeth_l3_is_addr_covered_by_ipato(struct qeth_card *card,
return rc;
}
-/*
- * Add IP to be added to todo list. If there is already an "add todo"
- * in this list we just incremenent the reference count.
- * Returns 0 if we just incremented reference count.
- */
-static int __qeth_l3_insert_ip_todo(struct qeth_card *card,
- struct qeth_ipaddr *addr, int add)
+inline int
+qeth_l3_ipaddrs_is_equal(struct qeth_ipaddr *addr1, struct qeth_ipaddr *addr2)
{
- struct qeth_ipaddr *tmp, *t;
- int found = 0;
+ return addr1->proto == addr2->proto &&
+ !memcmp(&addr1->u, &addr2->u, sizeof(addr1->u)) &&
+ !memcmp(&addr1->mac, &addr2->mac, sizeof(addr1->mac));
+}
- if (card->options.sniffer)
- return 0;
- list_for_each_entry_safe(tmp, t, card->ip_tbd_list, entry) {
- if ((addr->type == QETH_IP_TYPE_DEL_ALL_MC) &&
- (tmp->type == QETH_IP_TYPE_DEL_ALL_MC))
- return 0;
- if ((tmp->proto == QETH_PROT_IPV4) &&
- (addr->proto == QETH_PROT_IPV4) &&
- (tmp->type == addr->type) &&
- (tmp->is_multicast == addr->is_multicast) &&
- (tmp->u.a4.addr == addr->u.a4.addr) &&
- (tmp->u.a4.mask == addr->u.a4.mask)) {
- found = 1;
- break;
- }
- if ((tmp->proto == QETH_PROT_IPV6) &&
- (addr->proto == QETH_PROT_IPV6) &&
- (tmp->type == addr->type) &&
- (tmp->is_multicast == addr->is_multicast) &&
- (tmp->u.a6.pfxlen == addr->u.a6.pfxlen) &&
- (memcmp(&tmp->u.a6.addr, &addr->u.a6.addr,
- sizeof(struct in6_addr)) == 0)) {
- found = 1;
- break;
- }
- }
- if (found) {
- if (addr->users != 0)
- tmp->users += addr->users;
- else
- tmp->users += add ? 1 : -1;
- if (tmp->users == 0) {
- list_del(&tmp->entry);
- kfree(tmp);
- }
- return 0;
+static struct qeth_ipaddr *
+qeth_l3_ip_from_hash(struct qeth_card *card, struct qeth_ipaddr *tmp_addr)
+{
+ struct qeth_ipaddr *addr;
+
+ if (tmp_addr->is_multicast) {
+ hash_for_each_possible(card->ip_mc_htable, addr,
+ hnode, qeth_l3_ipaddr_hash(tmp_addr))
+ if (qeth_l3_ipaddrs_is_equal(tmp_addr, addr))
+ return addr;
} else {
- if (addr->type == QETH_IP_TYPE_DEL_ALL_MC)
- list_add(&addr->entry, card->ip_tbd_list);
- else {
- if (addr->users == 0)
- addr->users += add ? 1 : -1;
- if (add && (addr->type == QETH_IP_TYPE_NORMAL) &&
- qeth_l3_is_addr_covered_by_ipato(card, addr)) {
- QETH_CARD_TEXT(card, 2, "tkovaddr");
- addr->set_flags |= QETH_IPA_SETIP_TAKEOVER_FLAG;
- }
- list_add_tail(&addr->entry, card->ip_tbd_list);
- }
- return 1;
+ hash_for_each_possible(card->ip_htable, addr,
+ hnode, qeth_l3_ipaddr_hash(tmp_addr))
+ if (qeth_l3_ipaddrs_is_equal(tmp_addr, addr))
+ return addr;
}
+
+ return NULL;
}
-int qeth_l3_delete_ip(struct qeth_card *card, struct qeth_ipaddr *addr)
+int qeth_l3_delete_ip(struct qeth_card *card, struct qeth_ipaddr *tmp_addr)
{
- unsigned long flags;
int rc = 0;
+ struct qeth_ipaddr *addr;
QETH_CARD_TEXT(card, 4, "delip");
- if (addr->proto == QETH_PROT_IPV4)
- QETH_CARD_HEX(card, 4, &addr->u.a4.addr, 4);
+ if (tmp_addr->proto == QETH_PROT_IPV4)
+ QETH_CARD_HEX(card, 4, &tmp_addr->u.a4.addr, 4);
else {
- QETH_CARD_HEX(card, 4, &addr->u.a6.addr, 8);
- QETH_CARD_HEX(card, 4, ((char *)&addr->u.a6.addr) + 8, 8);
+ QETH_CARD_HEX(card, 4, &tmp_addr->u.a6.addr, 8);
+ QETH_CARD_HEX(card, 4, ((char *)&tmp_addr->u.a6.addr) + 8, 8);
}
- spin_lock_irqsave(&card->ip_lock, flags);
- rc = __qeth_l3_insert_ip_todo(card, addr, 0);
- spin_unlock_irqrestore(&card->ip_lock, flags);
+
+ addr = qeth_l3_ip_from_hash(card, tmp_addr);
+ if (!addr)
+ return -ENOENT;
+
+ addr->ref_counter--;
+ if (addr->type == QETH_IP_TYPE_NORMAL && addr->ref_counter > 0)
+ return rc;
+ if (addr->in_progress)
+ return -EINPROGRESS;
+
+ rc = qeth_l3_deregister_addr_entry(card, addr);
+
+ hash_del(&addr->hnode);
+ kfree(addr);
+
return rc;
}
-int qeth_l3_add_ip(struct qeth_card *card, struct qeth_ipaddr *addr)
+int qeth_l3_add_ip(struct qeth_card *card, struct qeth_ipaddr *tmp_addr)
{
- unsigned long flags;
int rc = 0;
+ struct qeth_ipaddr *addr;
QETH_CARD_TEXT(card, 4, "addip");
- if (addr->proto == QETH_PROT_IPV4)
- QETH_CARD_HEX(card, 4, &addr->u.a4.addr, 4);
+
+ if (tmp_addr->proto == QETH_PROT_IPV4)
+ QETH_CARD_HEX(card, 4, &tmp_addr->u.a4.addr, 4);
else {
- QETH_CARD_HEX(card, 4, &addr->u.a6.addr, 8);
- QETH_CARD_HEX(card, 4, ((char *)&addr->u.a6.addr) + 8, 8);
+ QETH_CARD_HEX(card, 4, &tmp_addr->u.a6.addr, 8);
+ QETH_CARD_HEX(card, 4, ((char *)&tmp_addr->u.a6.addr) + 8, 8);
+ }
+
+ addr = qeth_l3_ip_from_hash(card, tmp_addr);
+ if (!addr) {
+ addr = qeth_l3_get_addr_buffer(tmp_addr->proto);
+ if (!addr)
+ return -ENOMEM;
+
+ memcpy(addr, tmp_addr, sizeof(struct qeth_ipaddr));
+ addr->ref_counter = 1;
+
+ if (addr->type == QETH_IP_TYPE_NORMAL &&
+ qeth_l3_is_addr_covered_by_ipato(card, addr)) {
+ QETH_CARD_TEXT(card, 2, "tkovaddr");
+ addr->set_flags |= QETH_IPA_SETIP_TAKEOVER_FLAG;
+ }
+ hash_add(card->ip_htable, &addr->hnode,
+ qeth_l3_ipaddr_hash(addr));
+
+ /* qeth_l3_register_addr_entry can go to sleep
+ * if we add a IPV4 addr. It is caused by the reason
+ * that SETIP ipa cmd starts ARP staff for IPV4 addr.
+ * Thus we should unlock spinlock, and make a protection
+ * using in_progress variable to indicate that there is
+ * an hardware operation with this IPV4 address
+ */
+ if (addr->proto == QETH_PROT_IPV4) {
+ addr->in_progress = 1;
+ spin_unlock_bh(&card->ip_lock);
+ rc = qeth_l3_register_addr_entry(card, addr);
+ spin_lock_bh(&card->ip_lock);
+ addr->in_progress = 0;
+ } else
+ rc = qeth_l3_register_addr_entry(card, addr);
+
+ if (!rc || (rc == IPA_RC_DUPLICATE_IP_ADDRESS) ||
+ (rc == IPA_RC_LAN_OFFLINE)) {
+ addr->disp_flag = QETH_DISP_ADDR_DO_NOTHING;
+ if (addr->ref_counter < 1) {
+ qeth_l3_delete_ip(card, addr);
+ kfree(addr);
+ }
+ } else {
+ hash_del(&addr->hnode);
+ kfree(addr);
+ }
+ } else {
+ if (addr->type == QETH_IP_TYPE_NORMAL)
+ addr->ref_counter++;
}
- spin_lock_irqsave(&card->ip_lock, flags);
- rc = __qeth_l3_insert_ip_todo(card, addr, 1);
- spin_unlock_irqrestore(&card->ip_lock, flags);
+
return rc;
}
@@ -312,229 +338,88 @@ struct qeth_ipaddr *qeth_l3_get_addr_buffer(
struct qeth_ipaddr *addr;
addr = kzalloc(sizeof(struct qeth_ipaddr), GFP_ATOMIC);
- if (addr == NULL) {
+ if (!addr)
return NULL;
- }
+
addr->type = QETH_IP_TYPE_NORMAL;
+ addr->disp_flag = QETH_DISP_ADDR_DO_NOTHING;
addr->proto = prot;
- return addr;
-}
-static void qeth_l3_delete_mc_addresses(struct qeth_card *card)
-{
- struct qeth_ipaddr *iptodo;
- unsigned long flags;
-
- QETH_CARD_TEXT(card, 4, "delmc");
- iptodo = qeth_l3_get_addr_buffer(QETH_PROT_IPV4);
- if (!iptodo) {
- QETH_CARD_TEXT(card, 2, "dmcnomem");
- return;
- }
- iptodo->type = QETH_IP_TYPE_DEL_ALL_MC;
- spin_lock_irqsave(&card->ip_lock, flags);
- if (!__qeth_l3_insert_ip_todo(card, iptodo, 0))
- kfree(iptodo);
- spin_unlock_irqrestore(&card->ip_lock, flags);
+ return addr;
}
-/*
- * Add/remove address to/from card's ip list, i.e. try to add or remove
- * reference to/from an IP address that is already registered on the card.
- * Returns:
- * 0 address was on card and its reference count has been adjusted,
- * but is still > 0, so nothing has to be done
- * also returns 0 if card was not on card and the todo was to delete
- * the address -> there is also nothing to be done
- * 1 address was not on card and the todo is to add it to the card's ip
- * list
- * -1 address was on card and its reference count has been decremented
- * to <= 0 by the todo -> address must be removed from card
- */
-static int __qeth_l3_ref_ip_on_card(struct qeth_card *card,
- struct qeth_ipaddr *todo, struct qeth_ipaddr **__addr)
+static void qeth_l3_clear_ip_htable(struct qeth_card *card, int recover)
{
struct qeth_ipaddr *addr;
- int found = 0;
-
- list_for_each_entry(addr, &card->ip_list, entry) {
- if ((addr->proto == QETH_PROT_IPV4) &&
- (todo->proto == QETH_PROT_IPV4) &&
- (addr->type == todo->type) &&
- (addr->u.a4.addr == todo->u.a4.addr) &&
- (addr->u.a4.mask == todo->u.a4.mask)) {
- found = 1;
- break;
- }
- if ((addr->proto == QETH_PROT_IPV6) &&
- (todo->proto == QETH_PROT_IPV6) &&
- (addr->type == todo->type) &&
- (addr->u.a6.pfxlen == todo->u.a6.pfxlen) &&
- (memcmp(&addr->u.a6.addr, &todo->u.a6.addr,
- sizeof(struct in6_addr)) == 0)) {
- found = 1;
- break;
- }
- }
- if (found) {
- addr->users += todo->users;
- if (addr->users <= 0) {
- *__addr = addr;
- return -1;
- } else {
- /* for VIPA and RXIP limit refcount to 1 */
- if (addr->type != QETH_IP_TYPE_NORMAL)
- addr->users = 1;
- return 0;
- }
- }
- if (todo->users > 0) {
- /* for VIPA and RXIP limit refcount to 1 */
- if (todo->type != QETH_IP_TYPE_NORMAL)
- todo->users = 1;
- return 1;
- } else
- return 0;
-}
-
-static void __qeth_l3_delete_all_mc(struct qeth_card *card,
- unsigned long *flags)
-{
- struct list_head fail_list;
- struct qeth_ipaddr *addr, *tmp;
- int rc;
-
- INIT_LIST_HEAD(&fail_list);
-again:
- list_for_each_entry_safe(addr, tmp, &card->ip_list, entry) {
- if (addr->is_multicast) {
- list_del(&addr->entry);
- spin_unlock_irqrestore(&card->ip_lock, *flags);
- rc = qeth_l3_deregister_addr_entry(card, addr);
- spin_lock_irqsave(&card->ip_lock, *flags);
- if (!rc || (rc == IPA_RC_MC_ADDR_NOT_FOUND))
- kfree(addr);
- else
- list_add_tail(&addr->entry, &fail_list);
- goto again;
- }
- }
- list_splice(&fail_list, &card->ip_list);
-}
-
-void qeth_l3_set_ip_addr_list(struct qeth_card *card)
-{
- struct list_head *tbd_list;
- struct qeth_ipaddr *todo, *addr;
- unsigned long flags;
- int rc;
+ struct hlist_node *tmp;
+ int i;
- QETH_CARD_TEXT(card, 2, "sdiplist");
- QETH_CARD_HEX(card, 2, &card, sizeof(void *));
+ QETH_CARD_TEXT(card, 4, "clearip");
- if (!qeth_card_hw_is_reachable(card) || card->options.sniffer)
+ if (recover && card->options.sniffer)
return;
- spin_lock_irqsave(&card->ip_lock, flags);
- tbd_list = card->ip_tbd_list;
- card->ip_tbd_list = kzalloc(sizeof(struct list_head), GFP_ATOMIC);
- if (!card->ip_tbd_list) {
- QETH_CARD_TEXT(card, 0, "silnomem");
- card->ip_tbd_list = tbd_list;
- spin_unlock_irqrestore(&card->ip_lock, flags);
- return;
- } else
- INIT_LIST_HEAD(card->ip_tbd_list);
-
- while (!list_empty(tbd_list)) {
- todo = list_entry(tbd_list->next, struct qeth_ipaddr, entry);
- list_del(&todo->entry);
- if (todo->type == QETH_IP_TYPE_DEL_ALL_MC) {
- __qeth_l3_delete_all_mc(card, &flags);
- kfree(todo);
+ spin_lock_bh(&card->ip_lock);
+
+ hash_for_each_safe(card->ip_htable, i, tmp, addr, hnode) {
+ if (!recover) {
+ hash_del(&addr->hnode);
+ kfree(addr);
continue;
}
- rc = __qeth_l3_ref_ip_on_card(card, todo, &addr);
- if (rc == 0) {
- /* nothing to be done; only adjusted refcount */
- kfree(todo);
- } else if (rc == 1) {
- /* new entry to be added to on-card list */
- spin_unlock_irqrestore(&card->ip_lock, flags);
- rc = qeth_l3_register_addr_entry(card, todo);
- spin_lock_irqsave(&card->ip_lock, flags);
- if (!rc || (rc == IPA_RC_LAN_OFFLINE))
- list_add_tail(&todo->entry, &card->ip_list);
- else
- kfree(todo);
- } else if (rc == -1) {
- /* on-card entry to be removed */
- list_del_init(&addr->entry);
- spin_unlock_irqrestore(&card->ip_lock, flags);
- rc = qeth_l3_deregister_addr_entry(card, addr);
- spin_lock_irqsave(&card->ip_lock, flags);
- if (!rc || (rc == IPA_RC_IP_ADDRESS_NOT_DEFINED))
- kfree(addr);
- else
- list_add_tail(&addr->entry, &card->ip_list);
- kfree(todo);
- }
+ addr->disp_flag = QETH_DISP_ADDR_ADD;
}
- spin_unlock_irqrestore(&card->ip_lock, flags);
- kfree(tbd_list);
-}
-static void qeth_l3_clear_ip_list(struct qeth_card *card, int recover)
-{
- struct qeth_ipaddr *addr, *tmp;
- unsigned long flags;
+ spin_unlock_bh(&card->ip_lock);
- QETH_CARD_TEXT(card, 4, "clearip");
- if (recover && card->options.sniffer)
- return;
- spin_lock_irqsave(&card->ip_lock, flags);
- /* clear todo list */
- list_for_each_entry_safe(addr, tmp, card->ip_tbd_list, entry) {
- list_del(&addr->entry);
+ spin_lock_bh(&card->mclock);
+
+ hash_for_each_safe(card->ip_mc_htable, i, tmp, addr, hnode) {
+ hash_del(&addr->hnode);
kfree(addr);
}
- while (!list_empty(&card->ip_list)) {
- addr = list_entry(card->ip_list.next,
- struct qeth_ipaddr, entry);
- list_del_init(&addr->entry);
- if (!recover || addr->is_multicast) {
- kfree(addr);
- continue;
- }
- list_add_tail(&addr->entry, card->ip_tbd_list);
- }
- spin_unlock_irqrestore(&card->ip_lock, flags);
-}
+ spin_unlock_bh(&card->mclock);
+
-static int qeth_l3_address_exists_in_list(struct list_head *list,
- struct qeth_ipaddr *addr, int same_type)
+}
+static void qeth_l3_recover_ip(struct qeth_card *card)
{
- struct qeth_ipaddr *tmp;
+ struct qeth_ipaddr *addr;
+ struct hlist_node *tmp;
+ int i;
+ int rc;
- list_for_each_entry(tmp, list, entry) {
- if ((tmp->proto == QETH_PROT_IPV4) &&
- (addr->proto == QETH_PROT_IPV4) &&
- ((same_type && (tmp->type == addr->type)) ||
- (!same_type && (tmp->type != addr->type))) &&
- (tmp->u.a4.addr == addr->u.a4.addr))
- return 1;
+ QETH_CARD_TEXT(card, 4, "recoverip");
+
+ spin_lock_bh(&card->ip_lock);
+
+ hash_for_each_safe(card->ip_htable, i, tmp, addr, hnode) {
+ if (addr->disp_flag == QETH_DISP_ADDR_ADD) {
+ if (addr->proto == QETH_PROT_IPV4) {
+ addr->in_progress = 1;
+ spin_unlock_bh(&card->ip_lock);
+ rc = qeth_l3_register_addr_entry(card, addr);
+ spin_lock_bh(&card->ip_lock);
+ addr->in_progress = 0;
+ } else
+ rc = qeth_l3_register_addr_entry(card, addr);
+
+ if (!rc) {
+ addr->disp_flag = QETH_DISP_ADDR_DO_NOTHING;
+ if (addr->ref_counter < 1) {
+ qeth_l3_delete_ip(card, addr);
+ kfree(addr);
+ }
+ } else {
+ hash_del(&addr->hnode);
+ kfree(addr);
+ }
+ }
+ }
- if ((tmp->proto == QETH_PROT_IPV6) &&
- (addr->proto == QETH_PROT_IPV6) &&
- ((same_type && (tmp->type == addr->type)) ||
- (!same_type && (tmp->type != addr->type))) &&
- (memcmp(&tmp->u.a6.addr, &addr->u.a6.addr,
- sizeof(struct in6_addr)) == 0))
- return 1;
+ spin_unlock_bh(&card->ip_lock);
- }
- return 0;
}
static int qeth_l3_send_setdelmc(struct qeth_card *card,
@@ -712,27 +597,28 @@ int qeth_l3_setrouting_v6(struct qeth_card *card)
*/
static void qeth_l3_clear_ipato_list(struct qeth_card *card)
{
-
struct qeth_ipato_entry *ipatoe, *tmp;
- unsigned long flags;
- spin_lock_irqsave(&card->ip_lock, flags);
+ spin_lock_bh(&card->ip_lock);
+
list_for_each_entry_safe(ipatoe, tmp, &card->ipato.entries, entry) {
list_del(&ipatoe->entry);
kfree(ipatoe);
}
- spin_unlock_irqrestore(&card->ip_lock, flags);
+
+ spin_unlock_bh(&card->ip_lock);
}
int qeth_l3_add_ipato_entry(struct qeth_card *card,
struct qeth_ipato_entry *new)
{
struct qeth_ipato_entry *ipatoe;
- unsigned long flags;
int rc = 0;
QETH_CARD_TEXT(card, 2, "addipato");
- spin_lock_irqsave(&card->ip_lock, flags);
+
+ spin_lock_bh(&card->ip_lock);
+
list_for_each_entry(ipatoe, &card->ipato.entries, entry) {
if (ipatoe->proto != new->proto)
continue;
@@ -743,10 +629,12 @@ int qeth_l3_add_ipato_entry(struct qeth_card *card,
break;
}
}
+
if (!rc)
list_add_tail(&new->entry, &card->ipato.entries);
- spin_unlock_irqrestore(&card->ip_lock, flags);
+ spin_unlock_bh(&card->ip_lock);
+
return rc;
}
@@ -754,10 +642,11 @@ void qeth_l3_del_ipato_entry(struct qeth_card *card,
enum qeth_prot_versions proto, u8 *addr, int mask_bits)
{
struct qeth_ipato_entry *ipatoe, *tmp;
- unsigned long flags;
QETH_CARD_TEXT(card, 2, "delipato");
- spin_lock_irqsave(&card->ip_lock, flags);
+
+ spin_lock_bh(&card->ip_lock);
+
list_for_each_entry_safe(ipatoe, tmp, &card->ipato.entries, entry) {
if (ipatoe->proto != proto)
continue;
@@ -768,7 +657,8 @@ void qeth_l3_del_ipato_entry(struct qeth_card *card,
kfree(ipatoe);
}
}
- spin_unlock_irqrestore(&card->ip_lock, flags);
+
+ spin_unlock_bh(&card->ip_lock);
}
/*
@@ -778,7 +668,6 @@ int qeth_l3_add_vipa(struct qeth_card *card, enum qeth_prot_versions proto,
const u8 *addr)
{
struct qeth_ipaddr *ipaddr;
- unsigned long flags;
int rc = 0;
ipaddr = qeth_l3_get_addr_buffer(proto);
@@ -797,18 +686,18 @@ int qeth_l3_add_vipa(struct qeth_card *card, enum qeth_prot_versions proto,
ipaddr->del_flags = QETH_IPA_DELIP_VIPA_FLAG;
} else
return -ENOMEM;
- spin_lock_irqsave(&card->ip_lock, flags);
- if (qeth_l3_address_exists_in_list(&card->ip_list, ipaddr, 0) ||
- qeth_l3_address_exists_in_list(card->ip_tbd_list, ipaddr, 0))
+
+ spin_lock_bh(&card->ip_lock);
+
+ if (!qeth_l3_ip_from_hash(card, ipaddr))
rc = -EEXIST;
- spin_unlock_irqrestore(&card->ip_lock, flags);
- if (rc) {
- kfree(ipaddr);
- return rc;
- }
- if (!qeth_l3_add_ip(card, ipaddr))
- kfree(ipaddr);
- qeth_l3_set_ip_addr_list(card);
+ else
+ qeth_l3_add_ip(card, ipaddr);
+
+ spin_unlock_bh(&card->ip_lock);
+
+ kfree(ipaddr);
+
return rc;
}
@@ -831,9 +720,12 @@ void qeth_l3_del_vipa(struct qeth_card *card, enum qeth_prot_versions proto,
ipaddr->type = QETH_IP_TYPE_VIPA;
} else
return;
- if (!qeth_l3_delete_ip(card, ipaddr))
- kfree(ipaddr);
- qeth_l3_set_ip_addr_list(card);
+
+ spin_lock_bh(&card->ip_lock);
+ qeth_l3_delete_ip(card, ipaddr);
+ spin_unlock_bh(&card->ip_lock);
+
+ kfree(ipaddr);
}
/*
@@ -843,7 +735,6 @@ int qeth_l3_add_rxip(struct qeth_card *card, enum qeth_prot_versions proto,
const u8 *addr)
{
struct qeth_ipaddr *ipaddr;
- unsigned long flags;
int rc = 0;
ipaddr = qeth_l3_get_addr_buffer(proto);
@@ -857,24 +748,25 @@ int qeth_l3_add_rxip(struct qeth_card *card, enum qeth_prot_versions proto,
memcpy(&ipaddr->u.a6.addr, addr, 16);
ipaddr->u.a6.pfxlen = 0;
}
+
ipaddr->type = QETH_IP_TYPE_RXIP;
ipaddr->set_flags = QETH_IPA_SETIP_TAKEOVER_FLAG;
ipaddr->del_flags = 0;
} else
return -ENOMEM;
- spin_lock_irqsave(&card->ip_lock, flags);
- if (qeth_l3_address_exists_in_list(&card->ip_list, ipaddr, 0) ||
- qeth_l3_address_exists_in_list(card->ip_tbd_list, ipaddr, 0))
+
+ spin_lock_bh(&card->ip_lock);
+
+ if (!qeth_l3_ip_from_hash(card, ipaddr))
rc = -EEXIST;
- spin_unlock_irqrestore(&card->ip_lock, flags);
- if (rc) {
- kfree(ipaddr);
- return rc;
- }
- if (!qeth_l3_add_ip(card, ipaddr))
- kfree(ipaddr);
- qeth_l3_set_ip_addr_list(card);
- return 0;
+ else
+ qeth_l3_add_ip(card, ipaddr);
+
+ spin_unlock_bh(&card->ip_lock);
+
+ kfree(ipaddr);
+
+ return rc;
}
void qeth_l3_del_rxip(struct qeth_card *card, enum qeth_prot_versions proto,
@@ -896,9 +788,12 @@ void qeth_l3_del_rxip(struct qeth_card *card, enum qeth_prot_versions proto,
ipaddr->type = QETH_IP_TYPE_RXIP;
} else
return;
- if (!qeth_l3_delete_ip(card, ipaddr))
- kfree(ipaddr);
- qeth_l3_set_ip_addr_list(card);
+
+ spin_lock_bh(&card->ip_lock);
+ qeth_l3_delete_ip(card, ipaddr);
+ spin_unlock_bh(&card->ip_lock);
+
+ kfree(ipaddr);
}
static int qeth_l3_register_addr_entry(struct qeth_card *card,
@@ -908,6 +803,7 @@ static int qeth_l3_register_addr_entry(struct qeth_card *card,
int rc = 0;
int cnt = 3;
+
if (addr->proto == QETH_PROT_IPV4) {
QETH_CARD_TEXT(card, 2, "setaddr4");
QETH_CARD_HEX(card, 3, &addr->u.a4.addr, sizeof(int));
@@ -1507,31 +1403,99 @@ qeth_diags_trace(struct qeth_card *card, enum qeth_diags_trace_cmds diags_cmd)
return qeth_send_ipa_cmd(card, iob, qeth_diags_trace_cb, NULL);
}
-static void qeth_l3_get_mac_for_ipm(__u32 ipm, char *mac,
- struct net_device *dev)
+static void qeth_l3_get_mac_for_ipm(__u32 ipm, char *mac)
{
ip_eth_mc_map(ipm, mac);
}
-static void qeth_l3_add_mc(struct qeth_card *card, struct in_device *in4_dev)
+static void qeth_l3_mark_all_mc_to_be_deleted(struct qeth_card *card)
+{
+ struct qeth_ipaddr *addr;
+ int i;
+
+ hash_for_each(card->ip_mc_htable, i, addr, hnode)
+ addr->disp_flag = QETH_DISP_ADDR_DELETE;
+
+}
+
+static void qeth_l3_add_all_new_mc(struct qeth_card *card)
+{
+ struct qeth_ipaddr *addr;
+ struct hlist_node *tmp;
+ int i;
+ int rc;
+
+ hash_for_each_safe(card->ip_mc_htable, i, tmp, addr, hnode) {
+ if (addr->disp_flag == QETH_DISP_ADDR_ADD) {
+ rc = qeth_l3_register_addr_entry(card, addr);
+ if (!rc || (rc == IPA_RC_LAN_OFFLINE))
+ addr->ref_counter = 1;
+ else {
+ hash_del(&addr->hnode);
+ kfree(addr);
+ }
+ }
+ }
+
+}
+
+static void qeth_l3_delete_nonused_mc(struct qeth_card *card)
+{
+ struct qeth_ipaddr *addr;
+ struct hlist_node *tmp;
+ int i;
+ int rc;
+
+ hash_for_each_safe(card->ip_mc_htable, i, tmp, addr, hnode) {
+ if (addr->disp_flag == QETH_DISP_ADDR_DELETE) {
+ rc = qeth_l3_deregister_addr_entry(card, addr);
+ if (!rc || (rc == IPA_RC_MC_ADDR_NOT_FOUND)) {
+ hash_del(&addr->hnode);
+ kfree(addr);
+ }
+ }
+ }
+
+}
+
+
+static void
+qeth_l3_add_mc_to_hash(struct qeth_card *card, struct in_device *in4_dev)
{
- struct qeth_ipaddr *ipm;
struct ip_mc_list *im4;
+ struct qeth_ipaddr *tmp, *ipm;
char buf[MAX_ADDR_LEN];
QETH_CARD_TEXT(card, 4, "addmc");
+
+ tmp = qeth_l3_get_addr_buffer(QETH_PROT_IPV4);
+ if (!tmp)
+ return;
+
for (im4 = rcu_dereference(in4_dev->mc_list); im4 != NULL;
im4 = rcu_dereference(im4->next_rcu)) {
- qeth_l3_get_mac_for_ipm(im4->multiaddr, buf, in4_dev->dev);
- ipm = qeth_l3_get_addr_buffer(QETH_PROT_IPV4);
- if (!ipm)
- continue;
- ipm->u.a4.addr = im4->multiaddr;
- memcpy(ipm->mac, buf, OSA_ADDR_LEN);
- ipm->is_multicast = 1;
- if (!qeth_l3_add_ip(card, ipm))
- kfree(ipm);
+ qeth_l3_get_mac_for_ipm(im4->multiaddr, buf);
+
+ tmp->u.a4.addr = im4->multiaddr;
+ memcpy(tmp->mac, buf, sizeof(tmp->mac));
+
+ ipm = qeth_l3_ip_from_hash(card, tmp);
+ if (ipm) {
+ ipm->disp_flag = QETH_DISP_ADDR_DO_NOTHING;
+ } else {
+ ipm = qeth_l3_get_addr_buffer(QETH_PROT_IPV4);
+ if (!ipm)
+ continue;
+ memcpy(ipm->mac, buf, sizeof(tmp->mac));
+ ipm->u.a4.addr = im4->multiaddr;
+ ipm->is_multicast = 1;
+ ipm->disp_flag = QETH_DISP_ADDR_ADD;
+ hash_add(card->ip_mc_htable,
+ &ipm->hnode, qeth_l3_ipaddr_hash(ipm));
+ }
}
+
+ kfree(tmp);
}
/* called with rcu_read_lock */
@@ -1541,6 +1505,7 @@ static void qeth_l3_add_vlan_mc(struct qeth_card *card)
u16 vid;
QETH_CARD_TEXT(card, 4, "addmcvl");
+
if (!qeth_is_supported(card, IPA_FULL_VLAN))
return;
@@ -1555,7 +1520,7 @@ static void qeth_l3_add_vlan_mc(struct qeth_card *card)
in_dev = __in_dev_get_rcu(netdev);
if (!in_dev)
continue;
- qeth_l3_add_mc(card, in_dev);
+ qeth_l3_add_mc_to_hash(card, in_dev);
}
}
@@ -1564,36 +1529,60 @@ static void qeth_l3_add_multicast_ipv4(struct qeth_card *card)
struct in_device *in4_dev;
QETH_CARD_TEXT(card, 4, "chkmcv4");
+
rcu_read_lock();
in4_dev = __in_dev_get_rcu(card->dev);
if (in4_dev == NULL)
goto unlock;
- qeth_l3_add_mc(card, in4_dev);
+ qeth_l3_add_mc_to_hash(card, in4_dev);
qeth_l3_add_vlan_mc(card);
unlock:
rcu_read_unlock();
}
#ifdef CONFIG_QETH_IPV6
-static void qeth_l3_add_mc6(struct qeth_card *card, struct inet6_dev *in6_dev)
+static void
+qeth_l3_add_mc6_to_hash(struct qeth_card *card, struct inet6_dev *in6_dev)
{
struct qeth_ipaddr *ipm;
struct ifmcaddr6 *im6;
+ struct qeth_ipaddr *tmp;
char buf[MAX_ADDR_LEN];
QETH_CARD_TEXT(card, 4, "addmc6");
+
+ tmp = qeth_l3_get_addr_buffer(QETH_PROT_IPV6);
+ if (!tmp)
+ return;
+
for (im6 = in6_dev->mc_list; im6 != NULL; im6 = im6->next) {
ndisc_mc_map(&im6->mca_addr, buf, in6_dev->dev, 0);
+
+ memcpy(tmp->mac, buf, sizeof(tmp->mac));
+ memcpy(&tmp->u.a6.addr, &im6->mca_addr.s6_addr,
+ sizeof(struct in6_addr));
+ tmp->is_multicast = 1;
+
+ ipm = qeth_l3_ip_from_hash(card, tmp);
+ if (ipm) {
+ ipm->disp_flag = QETH_DISP_ADDR_DO_NOTHING;
+ continue;
+ }
+
ipm = qeth_l3_get_addr_buffer(QETH_PROT_IPV6);
if (!ipm)
continue;
- ipm->is_multicast = 1;
+
memcpy(ipm->mac, buf, OSA_ADDR_LEN);
memcpy(&ipm->u.a6.addr, &im6->mca_addr.s6_addr,
sizeof(struct in6_addr));
- if (!qeth_l3_add_ip(card, ipm))
- kfree(ipm);
+ ipm->is_multicast = 1;
+ ipm->disp_flag = QETH_DISP_ADDR_ADD;
+ hash_add(card->ip_mc_htable,
+ &ipm->hnode, qeth_l3_ipaddr_hash(ipm));
+
}
+ kfree(tmp);
}
/* called with rcu_read_lock */
@@ -1603,6 +1592,7 @@ static void qeth_l3_add_vlan_mc6(struct qeth_card *card)
u16 vid;
QETH_CARD_TEXT(card, 4, "admc6vl");
+
if (!qeth_is_supported(card, IPA_FULL_VLAN))
return;
@@ -1618,7 +1608,7 @@ static void qeth_l3_add_vlan_mc6(struct qeth_card *card)
if (!in_dev)
continue;
read_lock_bh(&in_dev->lock);
- qeth_l3_add_mc6(card, in_dev);
+ qeth_l3_add_mc6_to_hash(card, in_dev);
read_unlock_bh(&in_dev->lock);
in6_dev_put(in_dev);
}
@@ -1629,14 +1619,16 @@ static void qeth_l3_add_multicast_ipv6(struct qeth_card *card)
struct inet6_dev *in6_dev;
QETH_CARD_TEXT(card, 4, "chkmcv6");
+
if (!qeth_is_supported(card, IPA_IPV6))
return ;
in6_dev = in6_dev_get(card->dev);
- if (in6_dev == NULL)
+ if (!in6_dev)
return;
+
rcu_read_lock();
read_lock_bh(&in6_dev->lock);
- qeth_l3_add_mc6(card, in6_dev);
+ qeth_l3_add_mc6_to_hash(card, in6_dev);
qeth_l3_add_vlan_mc6(card);
read_unlock_bh(&in6_dev->lock);
rcu_read_unlock();
@@ -1660,16 +1652,23 @@ static void qeth_l3_free_vlan_addresses4(struct qeth_card *card,
in_dev = in_dev_get(netdev);
if (!in_dev)
return;
+
+ addr = qeth_l3_get_addr_buffer(QETH_PROT_IPV4);
+ if (!addr)
+ return;
+
+ spin_lock_bh(&card->ip_lock);
+
for (ifa = in_dev->ifa_list; ifa; ifa = ifa->ifa_next) {
- addr = qeth_l3_get_addr_buffer(QETH_PROT_IPV4);
- if (addr) {
- addr->u.a4.addr = ifa->ifa_address;
- addr->u.a4.mask = ifa->ifa_mask;
- addr->type = QETH_IP_TYPE_NORMAL;
- if (!qeth_l3_delete_ip(card, addr))
- kfree(addr);
- }
+ addr->u.a4.addr = ifa->ifa_address;
+ addr->u.a4.mask = ifa->ifa_mask;
+ addr->type = QETH_IP_TYPE_NORMAL;
+ qeth_l3_delete_ip(card, addr);
}
+
+ spin_unlock_bh(&card->ip_lock);
+
+ kfree(addr);
in_dev_put(in_dev);
}
@@ -1687,20 +1686,28 @@ static void qeth_l3_free_vlan_addresses6(struct qeth_card *card,
netdev = __vlan_find_dev_deep_rcu(card->dev, htons(ETH_P_8021Q), vid);
if (!netdev)
return;
+
in6_dev = in6_dev_get(netdev);
if (!in6_dev)
return;
+
+ addr = qeth_l3_get_addr_buffer(QETH_PROT_IPV6);
+ if (!addr)
+ return;
+
+ spin_lock_bh(&card->ip_lock);
+
list_for_each_entry(ifa, &in6_dev->addr_list, if_list) {
- addr = qeth_l3_get_addr_buffer(QETH_PROT_IPV6);
- if (addr) {
- memcpy(&addr->u.a6.addr, &ifa->addr,
- sizeof(struct in6_addr));
- addr->u.a6.pfxlen = ifa->prefix_len;
- addr->type = QETH_IP_TYPE_NORMAL;
- if (!qeth_l3_delete_ip(card, addr))
- kfree(addr);
- }
+ memcpy(&addr->u.a6.addr, &ifa->addr,
+ sizeof(struct in6_addr));
+ addr->u.a6.pfxlen = ifa->prefix_len;
+ addr->type = QETH_IP_TYPE_NORMAL;
+ qeth_l3_delete_ip(card, addr);
}
+
+ spin_unlock_bh(&card->ip_lock);
+
+ kfree(addr);
in6_dev_put(in6_dev);
#endif /* CONFIG_QETH_IPV6 */
}
@@ -1727,18 +1734,16 @@ static int qeth_l3_vlan_rx_kill_vid(struct net_device *dev,
__be16 proto, u16 vid)
{
struct qeth_card *card = dev->ml_priv;
- unsigned long flags;
QETH_CARD_TEXT_(card, 4, "kid:%d", vid);
+
if (qeth_wait_for_threads(card, QETH_RECOVER_THREAD)) {
QETH_CARD_TEXT(card, 3, "kidREC");
return 0;
}
- spin_lock_irqsave(&card->vlanlock, flags);
/* unregister IP addresses of vlan device */
qeth_l3_free_vlan_addresses(card, vid);
clear_bit(vid, card->active_vlans);
- spin_unlock_irqrestore(&card->vlanlock, flags);
qeth_l3_set_multicast_list(card->dev);
return 0;
}
@@ -1994,8 +1999,8 @@ static int qeth_l3_verify_vlan_dev(struct net_device *dev,
static int qeth_l3_verify_dev(struct net_device *dev)
{
struct qeth_card *card;
- unsigned long flags;
int rc = 0;
+ unsigned long flags;
read_lock_irqsave(&qeth_core_card_list.rwlock, flags);
list_for_each_entry(card, &qeth_core_card_list.list, list) {
@@ -2051,7 +2056,7 @@ static void qeth_l3_stop_card(struct qeth_card *card, int recovery_mode)
card->state = CARD_STATE_SOFTSETUP;
}
if (card->state == CARD_STATE_SOFTSETUP) {
- qeth_l3_clear_ip_list(card, 1);
+ qeth_l3_clear_ip_htable(card, 1);
qeth_clear_ipacmd_list(card);
card->state = CARD_STATE_HARDSETUP;
}
@@ -2106,12 +2111,20 @@ static void qeth_l3_set_multicast_list(struct net_device *dev)
(card->state != CARD_STATE_UP))
return;
if (!card->options.sniffer) {
- qeth_l3_delete_mc_addresses(card);
+
+ spin_lock_bh(&card->mclock);
+
+ qeth_l3_mark_all_mc_to_be_deleted(card);
+
qeth_l3_add_multicast_ipv4(card);
#ifdef CONFIG_QETH_IPV6
qeth_l3_add_multicast_ipv6(card);
#endif
- qeth_l3_set_ip_addr_list(card);
+ qeth_l3_delete_nonused_mc(card);
+ qeth_l3_add_all_new_mc(card);
+
+ spin_unlock_bh(&card->mclock);
+
if (!qeth_adp_supported(card, IPA_SETADP_SET_PROMISC_MODE))
return;
}
@@ -3261,7 +3274,7 @@ static void qeth_l3_remove_device(struct ccwgroup_device *cgdev)
card->dev = NULL;
}
- qeth_l3_clear_ip_list(card, 0);
+ qeth_l3_clear_ip_htable(card, 0);
qeth_l3_clear_ipato_list(card);
return;
}
@@ -3346,7 +3359,7 @@ contin:
card->state = CARD_STATE_SOFTSETUP;
qeth_set_allowed_threads(card, 0xffffffff, 0);
- qeth_l3_set_ip_addr_list(card);
+ qeth_l3_recover_ip(card);
if (card->lan_online)
netif_carrier_on(card->dev);
else
@@ -3547,6 +3560,7 @@ EXPORT_SYMBOL_GPL(qeth_l3_discipline);
static int qeth_l3_ip_event(struct notifier_block *this,
unsigned long event, void *ptr)
{
+
struct in_ifaddr *ifa = (struct in_ifaddr *)ptr;
struct net_device *dev = (struct net_device *)ifa->ifa_dev->dev;
struct qeth_ipaddr *addr;
@@ -3561,27 +3575,27 @@ static int qeth_l3_ip_event(struct notifier_block *this,
QETH_CARD_TEXT(card, 3, "ipevent");
addr = qeth_l3_get_addr_buffer(QETH_PROT_IPV4);
- if (addr != NULL) {
+ if (addr) {
addr->u.a4.addr = ifa->ifa_address;
addr->u.a4.mask = ifa->ifa_mask;
addr->type = QETH_IP_TYPE_NORMAL;
} else
- goto out;
+ return NOTIFY_DONE;
switch (event) {
case NETDEV_UP:
- if (!qeth_l3_add_ip(card, addr))
- kfree(addr);
+ spin_lock_bh(&card->ip_lock);
+ qeth_l3_add_ip(card, addr);
+ spin_unlock_bh(&card->ip_lock);
break;
case NETDEV_DOWN:
- if (!qeth_l3_delete_ip(card, addr))
- kfree(addr);
- break;
- default:
+ spin_lock_bh(&card->ip_lock);
+ qeth_l3_delete_ip(card, addr);
+ spin_unlock_bh(&card->ip_lock);
break;
}
- qeth_l3_set_ip_addr_list(card);
-out:
+
+ kfree(addr);
return NOTIFY_DONE;
}
@@ -3610,27 +3624,27 @@ static int qeth_l3_ip6_event(struct notifier_block *this,
return NOTIFY_DONE;
addr = qeth_l3_get_addr_buffer(QETH_PROT_IPV6);
- if (addr != NULL) {
+ if (addr) {
memcpy(&addr->u.a6.addr, &ifa->addr, sizeof(struct in6_addr));
addr->u.a6.pfxlen = ifa->prefix_len;
addr->type = QETH_IP_TYPE_NORMAL;
} else
- goto out;
+ return NOTIFY_DONE;
switch (event) {
case NETDEV_UP:
- if (!qeth_l3_add_ip(card, addr))
- kfree(addr);
+ spin_lock_bh(&card->ip_lock);
+ qeth_l3_add_ip(card, addr);
+ spin_unlock_bh(&card->ip_lock);
break;
case NETDEV_DOWN:
- if (!qeth_l3_delete_ip(card, addr))
- kfree(addr);
- break;
- default:
+ spin_lock_bh(&card->ip_lock);
+ qeth_l3_delete_ip(card, addr);
+ spin_unlock_bh(&card->ip_lock);
break;
}
- qeth_l3_set_ip_addr_list(card);
-out:
+
+ kfree(addr);
return NOTIFY_DONE;
}
diff --git a/drivers/s390/net/qeth_l3_sys.c b/drivers/s390/net/qeth_l3_sys.c
index 386eb7b..65645b1 100644
--- a/drivers/s390/net/qeth_l3_sys.c
+++ b/drivers/s390/net/qeth_l3_sys.c
@@ -8,6 +8,7 @@
#include <linux/slab.h>
#include <asm/ebcdic.h>
+#include <linux/hashtable.h>
#include "qeth_l3.h"
#define QETH_DEVICE_ATTR(_id, _name, _mode, _show, _store) \
@@ -285,19 +286,19 @@ static ssize_t qeth_l3_dev_hsuid_store(struct device *dev,
if (card->options.hsuid[0]) {
/* delete old ip address */
addr = qeth_l3_get_addr_buffer(QETH_PROT_IPV6);
- if (addr != NULL) {
- addr->u.a6.addr.s6_addr32[0] = 0xfe800000;
- addr->u.a6.addr.s6_addr32[1] = 0x00000000;
- for (i = 8; i < 16; i++)
- addr->u.a6.addr.s6_addr[i] =
- card->options.hsuid[i - 8];
- addr->u.a6.pfxlen = 0;
- addr->type = QETH_IP_TYPE_NORMAL;
- } else
+ if (!addr)
return -ENOMEM;
- if (!qeth_l3_delete_ip(card, addr))
- kfree(addr);
- qeth_l3_set_ip_addr_list(card);
+
+ addr->u.a6.addr.s6_addr32[0] = 0xfe800000;
+ addr->u.a6.addr.s6_addr32[1] = 0x00000000;
+ for (i = 8; i < 16; i++)
+ addr->u.a6.addr.s6_addr[i] =
+ card->options.hsuid[i - 8];
+ addr->u.a6.pfxlen = 0;
+ addr->type = QETH_IP_TYPE_NORMAL;
+
+ qeth_l3_delete_ip(card, addr);
+ kfree(addr);
}
if (strlen(tmp) == 0) {
@@ -328,9 +329,8 @@ static ssize_t qeth_l3_dev_hsuid_store(struct device *dev,
addr->type = QETH_IP_TYPE_NORMAL;
} else
return -ENOMEM;
- if (!qeth_l3_add_ip(card, addr))
- kfree(addr);
- qeth_l3_set_ip_addr_list(card);
+ qeth_l3_add_ip(card, addr);
+ kfree(addr);
return count;
}
@@ -367,8 +367,8 @@ static ssize_t qeth_l3_dev_ipato_enable_store(struct device *dev,
struct device_attribute *attr, const char *buf, size_t count)
{
struct qeth_card *card = dev_get_drvdata(dev);
- struct qeth_ipaddr *tmpipa, *t;
- int rc = 0;
+ struct qeth_ipaddr *addr;
+ int i, rc = 0;
if (!card)
return -EINVAL;
@@ -384,21 +384,20 @@ static ssize_t qeth_l3_dev_ipato_enable_store(struct device *dev,
card->ipato.enabled = (card->ipato.enabled)? 0 : 1;
} else if (sysfs_streq(buf, "1")) {
card->ipato.enabled = 1;
- list_for_each_entry_safe(tmpipa, t, card->ip_tbd_list, entry) {
- if ((tmpipa->type == QETH_IP_TYPE_NORMAL) &&
- qeth_l3_is_addr_covered_by_ipato(card, tmpipa))
- tmpipa->set_flags |=
+ hash_for_each(card->ip_htable, i, addr, hnode) {
+ if ((addr->type == QETH_IP_TYPE_NORMAL) &&
+ qeth_l3_is_addr_covered_by_ipato(card, addr))
+ addr->set_flags |=
QETH_IPA_SETIP_TAKEOVER_FLAG;
- }
-
+ }
} else if (sysfs_streq(buf, "0")) {
card->ipato.enabled = 0;
- list_for_each_entry_safe(tmpipa, t, card->ip_tbd_list, entry) {
- if (tmpipa->set_flags &
- QETH_IPA_SETIP_TAKEOVER_FLAG)
- tmpipa->set_flags &=
- ~QETH_IPA_SETIP_TAKEOVER_FLAG;
- }
+ hash_for_each(card->ip_htable, i, addr, hnode) {
+ if (addr->set_flags &
+ QETH_IPA_SETIP_TAKEOVER_FLAG)
+ addr->set_flags &=
+ ~QETH_IPA_SETIP_TAKEOVER_FLAG;
+ }
} else
rc = -EINVAL;
out:
@@ -452,7 +451,6 @@ static ssize_t qeth_l3_dev_ipato_add_show(char *buf, struct qeth_card *card,
enum qeth_prot_versions proto)
{
struct qeth_ipato_entry *ipatoe;
- unsigned long flags;
char addr_str[40];
int entry_len; /* length of 1 entry string, differs between v4 and v6 */
int i = 0;
@@ -460,7 +458,7 @@ static ssize_t qeth_l3_dev_ipato_add_show(char *buf, struct qeth_card *card,
entry_len = (proto == QETH_PROT_IPV4)? 12 : 40;
/* add strlen for "/<mask>\n" */
entry_len += (proto == QETH_PROT_IPV4)? 5 : 6;
- spin_lock_irqsave(&card->ip_lock, flags);
+ spin_lock_bh(&card->ip_lock);
list_for_each_entry(ipatoe, &card->ipato.entries, entry) {
if (ipatoe->proto != proto)
continue;
@@ -473,7 +471,7 @@ static ssize_t qeth_l3_dev_ipato_add_show(char *buf, struct qeth_card *card,
i += snprintf(buf + i, PAGE_SIZE - i,
"%s/%i\n", addr_str, ipatoe->mask_bits);
}
- spin_unlock_irqrestore(&card->ip_lock, flags);
+ spin_unlock_bh(&card->ip_lock);
i += snprintf(buf + i, PAGE_SIZE - i, "\n");
return i;
@@ -689,15 +687,15 @@ static ssize_t qeth_l3_dev_vipa_add_show(char *buf, struct qeth_card *card,
enum qeth_prot_versions proto)
{
struct qeth_ipaddr *ipaddr;
+ struct hlist_node *tmp;
char addr_str[40];
int entry_len; /* length of 1 entry string, differs between v4 and v6 */
- unsigned long flags;
int i = 0;
entry_len = (proto == QETH_PROT_IPV4)? 12 : 40;
entry_len += 2; /* \n + terminator */
- spin_lock_irqsave(&card->ip_lock, flags);
- list_for_each_entry(ipaddr, &card->ip_list, entry) {
+ spin_lock_bh(&card->ip_lock);
+ hash_for_each_safe(card->ip_htable, i, tmp, ipaddr, hnode) {
if (ipaddr->proto != proto)
continue;
if (ipaddr->type != QETH_IP_TYPE_VIPA)
@@ -711,7 +709,7 @@ static ssize_t qeth_l3_dev_vipa_add_show(char *buf, struct qeth_card *card,
addr_str);
i += snprintf(buf + i, PAGE_SIZE - i, "%s\n", addr_str);
}
- spin_unlock_irqrestore(&card->ip_lock, flags);
+ spin_unlock_bh(&card->ip_lock);
i += snprintf(buf + i, PAGE_SIZE - i, "\n");
return i;
@@ -851,15 +849,15 @@ static ssize_t qeth_l3_dev_rxip_add_show(char *buf, struct qeth_card *card,
enum qeth_prot_versions proto)
{
struct qeth_ipaddr *ipaddr;
+ struct hlist_node *tmp;
char addr_str[40];
int entry_len; /* length of 1 entry string, differs between v4 and v6 */
- unsigned long flags;
int i = 0;
entry_len = (proto == QETH_PROT_IPV4)? 12 : 40;
entry_len += 2; /* \n + terminator */
- spin_lock_irqsave(&card->ip_lock, flags);
- list_for_each_entry(ipaddr, &card->ip_list, entry) {
+ spin_lock_bh(&card->ip_lock);
+ hash_for_each_safe(card->ip_htable, i, tmp, ipaddr, hnode) {
if (ipaddr->proto != proto)
continue;
if (ipaddr->type != QETH_IP_TYPE_RXIP)
@@ -873,7 +871,7 @@ static ssize_t qeth_l3_dev_rxip_add_show(char *buf, struct qeth_card *card,
addr_str);
i += snprintf(buf + i, PAGE_SIZE - i, "%s\n", addr_str);
}
- spin_unlock_irqrestore(&card->ip_lock, flags);
+ spin_unlock_bh(&card->ip_lock);
i += snprintf(buf + i, PAGE_SIZE - i, "\n");
return i;
--
2.6.6
^ permalink raw reply related [flat|nested] 15+ messages in thread* [PATCH net-next 09/13] qeth layer 2 and layer 3 common feature handling
2016-06-16 14:18 [PATCH net-next 00/13] s390: qeth patches Ursula Braun
` (7 preceding siblings ...)
2016-06-16 14:18 ` [PATCH net-next 08/13] qeth: optimize IP handling in rx_mode callback Ursula Braun
@ 2016-06-16 14:18 ` Ursula Braun
2016-06-16 14:19 ` [PATCH net-next 10/13] qeth: add network device features for VLAN devices Ursula Braun
` (4 subsequent siblings)
13 siblings, 0 replies; 15+ messages in thread
From: Ursula Braun @ 2016-06-16 14:18 UTC (permalink / raw)
To: davem; +Cc: netdev, linux-s390, schwidefsky, heiko.carstens, utz.bacher
From: Thomas Richter <tmricht@linux.vnet.ibm.com>
This patch introduces a common set of fix_features and set_features
functions for layer 2 and layer 3. The RX, TX and TSO offload
functionality on the OSA card is enabled using ethtool at user's
request and not at device initialization as done before.
For layer 3 the RX checksum offloading is disabled at device
initialization time.
Signed-off-by: Thomas Richter <tmricht@linux.vnet.ibm.com>
Signed-off-by: Ursula Braun <ubraun@linux.vnet.ibm.com>
---
drivers/s390/net/qeth_core.h | 5 +-
drivers/s390/net/qeth_core_main.c | 127 ++++++++++++++++++++++++++------------
drivers/s390/net/qeth_l2_main.c | 54 ++++------------
drivers/s390/net/qeth_l3_main.c | 123 +++---------------------------------
4 files changed, 111 insertions(+), 198 deletions(-)
diff --git a/drivers/s390/net/qeth_core.h b/drivers/s390/net/qeth_core.h
index ab0a171..bf40063 100644
--- a/drivers/s390/net/qeth_core.h
+++ b/drivers/s390/net/qeth_core.h
@@ -993,12 +993,13 @@ int qeth_send_setassparms(struct qeth_card *, struct qeth_cmd_buffer *, __u16,
int (*reply_cb)(struct qeth_card *,
struct qeth_reply *, unsigned long),
void *);
+int qeth_setassparms_cb(struct qeth_card *, struct qeth_reply *, unsigned long);
struct qeth_cmd_buffer *qeth_get_setassparms_cmd(struct qeth_card *,
enum qeth_ipa_funcs,
__u16, __u16,
enum qeth_prot_versions);
-int qeth_start_ipa_tx_checksum(struct qeth_card *);
-int qeth_set_rx_csum(struct qeth_card *, int);
+int qeth_set_features(struct net_device *, netdev_features_t);
+netdev_features_t qeth_fix_features(struct net_device *, netdev_features_t);
/* exports for OSN */
int qeth_osn_assist(struct net_device *, void *, int);
diff --git a/drivers/s390/net/qeth_core_main.c b/drivers/s390/net/qeth_core_main.c
index ede9ed8..19a6ee0 100644
--- a/drivers/s390/net/qeth_core_main.c
+++ b/drivers/s390/net/qeth_core_main.c
@@ -5268,8 +5268,8 @@ no_mem:
}
EXPORT_SYMBOL_GPL(qeth_core_get_next_skb);
-static int qeth_setassparms_cb(struct qeth_card *card,
- struct qeth_reply *reply, unsigned long data)
+int qeth_setassparms_cb(struct qeth_card *card,
+ struct qeth_reply *reply, unsigned long data)
{
struct qeth_ipa_cmd *cmd;
@@ -5297,6 +5297,7 @@ static int qeth_setassparms_cb(struct qeth_card *card,
return 0;
}
+EXPORT_SYMBOL_GPL(qeth_setassparms_cb);
struct qeth_cmd_buffer *qeth_get_setassparms_cmd(struct qeth_card *card,
enum qeth_ipa_funcs ipa_func,
@@ -6053,74 +6054,120 @@ int qeth_core_ethtool_get_settings(struct net_device *netdev,
}
EXPORT_SYMBOL_GPL(qeth_core_ethtool_get_settings);
-static int qeth_send_checksum_command(struct qeth_card *card)
+static int qeth_send_checksum_on(struct qeth_card *card, int cstype)
{
+ long rxtx_arg;
int rc;
- rc = qeth_send_simple_setassparms(card, IPA_INBOUND_CHECKSUM,
- IPA_CMD_ASS_START, 0);
+ rc = qeth_send_simple_setassparms(card, cstype, IPA_CMD_ASS_START, 0);
if (rc) {
- dev_warn(&card->gdev->dev, "Starting HW checksumming for %s "
- "failed, using SW checksumming\n",
- QETH_CARD_IFNAME(card));
+ dev_warn(&card->gdev->dev,
+ "Starting HW checksumming for %s failed, using SW checksumming\n",
+ QETH_CARD_IFNAME(card));
return rc;
}
- rc = qeth_send_simple_setassparms(card, IPA_INBOUND_CHECKSUM,
- IPA_CMD_ASS_ENABLE,
- card->info.csum_mask);
+ rxtx_arg = (cstype == IPA_OUTBOUND_CHECKSUM) ? card->info.tx_csum_mask
+ : card->info.csum_mask;
+ rc = qeth_send_simple_setassparms(card, cstype, IPA_CMD_ASS_ENABLE,
+ rxtx_arg);
if (rc) {
- dev_warn(&card->gdev->dev, "Enabling HW checksumming for %s "
- "failed, using SW checksumming\n",
- QETH_CARD_IFNAME(card));
+ dev_warn(&card->gdev->dev,
+ "Enabling HW checksumming for %s failed, using SW checksumming\n",
+ QETH_CARD_IFNAME(card));
return rc;
}
+
+ dev_info(&card->gdev->dev, "HW Checksumming (%sbound) enabled\n",
+ cstype == IPA_INBOUND_CHECKSUM ? "in" : "out");
return 0;
}
-int qeth_set_rx_csum(struct qeth_card *card, int on)
+static int qeth_set_ipa_csum(struct qeth_card *card, int on, int cstype)
{
int rc;
if (on) {
- rc = qeth_send_checksum_command(card);
+ rc = qeth_send_checksum_on(card, cstype);
if (rc)
return -EIO;
- dev_info(&card->gdev->dev,
- "HW Checksumming (inbound) enabled\n");
} else {
- rc = qeth_send_simple_setassparms(card,
- IPA_INBOUND_CHECKSUM, IPA_CMD_ASS_STOP, 0);
+ rc = qeth_send_simple_setassparms(card, cstype,
+ IPA_CMD_ASS_STOP, 0);
if (rc)
return -EIO;
}
return 0;
}
-EXPORT_SYMBOL_GPL(qeth_set_rx_csum);
-int qeth_start_ipa_tx_checksum(struct qeth_card *card)
+static int qeth_set_ipa_tso(struct qeth_card *card, int on)
{
- int rc = 0;
+ int rc;
- if (!qeth_is_supported(card, IPA_OUTBOUND_CHECKSUM))
- return rc;
- rc = qeth_send_simple_setassparms(card, IPA_OUTBOUND_CHECKSUM,
- IPA_CMD_ASS_START, 0);
- if (rc)
- goto err_out;
- rc = qeth_send_simple_setassparms(card, IPA_OUTBOUND_CHECKSUM,
- IPA_CMD_ASS_ENABLE,
- card->info.tx_csum_mask);
- if (rc)
- goto err_out;
+ QETH_CARD_TEXT(card, 3, "sttso");
- dev_info(&card->gdev->dev, "HW TX Checksumming enabled\n");
- return rc;
-err_out:
- dev_warn(&card->gdev->dev, "Enabling HW TX checksumming for %s "
- "failed, using SW TX checksumming\n", QETH_CARD_IFNAME(card));
+ if (on) {
+ rc = qeth_send_simple_setassparms(card, IPA_OUTBOUND_TSO,
+ IPA_CMD_ASS_START, 0);
+ if (rc) {
+ dev_warn(&card->gdev->dev,
+ "Starting outbound TCP segmentation offload for %s failed\n",
+ QETH_CARD_IFNAME(card));
+ return -EIO;
+ }
+ dev_info(&card->gdev->dev, "Outbound TSO enabled\n");
+ } else {
+ rc = qeth_send_simple_setassparms(card, IPA_OUTBOUND_TSO,
+ IPA_CMD_ASS_STOP, 0);
+ }
return rc;
}
-EXPORT_SYMBOL_GPL(qeth_start_ipa_tx_checksum);
+
+int qeth_set_features(struct net_device *dev, netdev_features_t features)
+{
+ struct qeth_card *card = dev->ml_priv;
+ netdev_features_t changed = card->dev->features ^ features;
+ int rc = 0;
+
+ QETH_DBF_TEXT(SETUP, 2, "setfeat");
+ QETH_DBF_HEX(SETUP, 2, &features, sizeof(features));
+
+ if (card->state == CARD_STATE_DOWN ||
+ card->state == CARD_STATE_RECOVER)
+ return 0;
+
+ if ((changed & NETIF_F_IP_CSUM))
+ rc = qeth_set_ipa_csum(card,
+ features & NETIF_F_IP_CSUM ? 1 : 0,
+ IPA_OUTBOUND_CHECKSUM);
+ if ((changed & NETIF_F_RXCSUM))
+ rc |= qeth_set_ipa_csum(card,
+ features & NETIF_F_RXCSUM ? 1 : 0,
+ IPA_INBOUND_CHECKSUM);
+ if ((changed & NETIF_F_TSO))
+ rc |= qeth_set_ipa_tso(card, features & NETIF_F_TSO ? 1 : 0);
+ return rc ? -EIO : 0;
+}
+EXPORT_SYMBOL_GPL(qeth_set_features);
+
+netdev_features_t qeth_fix_features(struct net_device *dev,
+ netdev_features_t features)
+{
+ struct qeth_card *card = dev->ml_priv;
+
+ QETH_DBF_TEXT(SETUP, 2, "fixfeat");
+ if (!qeth_is_supported(card, IPA_OUTBOUND_CHECKSUM))
+ features &= ~NETIF_F_IP_CSUM;
+ if (!qeth_is_supported(card, IPA_INBOUND_CHECKSUM))
+ features &= ~NETIF_F_RXCSUM;
+ if (!qeth_is_supported(card, IPA_OUTBOUND_TSO)) {
+ features &= ~NETIF_F_TSO;
+ dev_info(&card->gdev->dev, "Outbound TSO not supported on %s\n",
+ QETH_CARD_IFNAME(card));
+ }
+ QETH_DBF_HEX(SETUP, 2, &features, sizeof(features));
+ return features;
+}
+EXPORT_SYMBOL_GPL(qeth_fix_features);
static int __init qeth_core_init(void)
{
diff --git a/drivers/s390/net/qeth_l2_main.c b/drivers/s390/net/qeth_l2_main.c
index 2a331d1..928a4ad 100644
--- a/drivers/s390/net/qeth_l2_main.c
+++ b/drivers/s390/net/qeth_l2_main.c
@@ -404,38 +404,6 @@ static int qeth_l2_vlan_rx_kill_vid(struct net_device *dev,
return rc;
}
-static netdev_features_t qeth_l2_fix_features(struct net_device *dev,
- netdev_features_t features)
-{
- struct qeth_card *card = dev->ml_priv;
-
- QETH_DBF_TEXT(SETUP, 2, "fixfeat");
- if (!qeth_is_supported(card, IPA_OUTBOUND_CHECKSUM))
- features &= ~NETIF_F_IP_CSUM;
- if (!qeth_is_supported(card, IPA_INBOUND_CHECKSUM))
- features &= ~NETIF_F_RXCSUM;
- QETH_DBF_HEX(SETUP, 2, &features, sizeof(features));
- return features;
-}
-
-static int qeth_l2_set_features(struct net_device *dev,
- netdev_features_t features)
-{
- struct qeth_card *card = dev->ml_priv;
- netdev_features_t changed = dev->features ^ features;
-
- QETH_DBF_TEXT(SETUP, 2, "setfeat");
- QETH_DBF_HEX(SETUP, 2, &features, sizeof(features));
-
- if (card->state == CARD_STATE_DOWN ||
- card->state == CARD_STATE_RECOVER)
- return 0;
-
- if (!(changed & NETIF_F_RXCSUM))
- return 0;
- return qeth_set_rx_csum(card, features & NETIF_F_RXCSUM ? 1 : 0);
-}
-
static void qeth_l2_stop_card(struct qeth_card *card, int recovery_mode)
{
QETH_DBF_TEXT(SETUP , 2, "stopcard");
@@ -1112,8 +1080,8 @@ static const struct net_device_ops qeth_l2_netdev_ops = {
.ndo_vlan_rx_add_vid = qeth_l2_vlan_rx_add_vid,
.ndo_vlan_rx_kill_vid = qeth_l2_vlan_rx_kill_vid,
.ndo_tx_timeout = qeth_tx_timeout,
- .ndo_fix_features = qeth_l2_fix_features,
- .ndo_set_features = qeth_l2_set_features
+ .ndo_fix_features = qeth_fix_features,
+ .ndo_set_features = qeth_set_features
};
static int qeth_l2_setup_netdev(struct qeth_card *card)
@@ -1144,10 +1112,14 @@ static int qeth_l2_setup_netdev(struct qeth_card *card)
&qeth_l2_ethtool_ops : &qeth_l2_osn_ops;
card->dev->features |= NETIF_F_HW_VLAN_CTAG_FILTER;
if (card->info.type == QETH_CARD_TYPE_OSD && !card->info.guestlan) {
- card->dev->hw_features = NETIF_F_IP_CSUM | NETIF_F_RXCSUM |
- NETIF_F_SG;
- /* Turn on RX offloading and SG per default */
- card->dev->features |= NETIF_F_RXCSUM | NETIF_F_SG;
+ card->dev->hw_features = NETIF_F_SG;
+ /* OSA 3S and earlier has no RX/TX support */
+ if (qeth_is_supported(card, IPA_OUTBOUND_CHECKSUM))
+ card->dev->hw_features |= NETIF_F_IP_CSUM;
+ if (qeth_is_supported(card, IPA_INBOUND_CHECKSUM))
+ card->dev->hw_features |= NETIF_F_RXCSUM;
+ /* Turn on SG per default */
+ card->dev->features |= NETIF_F_SG;
}
card->info.broadcast_capable = 1;
qeth_l2_request_initial_mac(card);
@@ -1165,9 +1137,6 @@ static int qeth_l2_start_ipassists(struct qeth_card *card)
/* configure isolation level */
if (qeth_set_access_ctrl_online(card, 0))
return -ENODEV;
- if (qeth_is_supported(card, IPA_INBOUND_CHECKSUM))
- qeth_set_rx_csum(card, 1);
- qeth_start_ipa_tx_checksum(card);
return 0;
}
@@ -1236,7 +1205,8 @@ static int __qeth_l2_set_online(struct ccwgroup_device *gdev, int recovery_mode)
contin:
if ((card->info.type == QETH_CARD_TYPE_OSD) ||
(card->info.type == QETH_CARD_TYPE_OSX)) {
- if (qeth_l2_start_ipassists(card))
+ rc = qeth_l2_start_ipassists(card);
+ if (rc)
goto out_remove;
}
diff --git a/drivers/s390/net/qeth_l3_main.c b/drivers/s390/net/qeth_l3_main.c
index 6e7d06c..fc81776 100644
--- a/drivers/s390/net/qeth_l3_main.c
+++ b/drivers/s390/net/qeth_l3_main.c
@@ -909,36 +909,6 @@ static int qeth_l3_setadapter_parms(struct qeth_card *card)
return rc;
}
-static int qeth_l3_default_setassparms_cb(struct qeth_card *card,
- struct qeth_reply *reply, unsigned long data)
-{
- struct qeth_ipa_cmd *cmd;
-
- QETH_CARD_TEXT(card, 4, "defadpcb");
-
- cmd = (struct qeth_ipa_cmd *) data;
- if (cmd->hdr.return_code == 0) {
- cmd->hdr.return_code = cmd->data.setassparms.hdr.return_code;
- if (cmd->hdr.prot_version == QETH_PROT_IPV4)
- card->options.ipa4.enabled_funcs = cmd->hdr.ipa_enabled;
- if (cmd->hdr.prot_version == QETH_PROT_IPV6)
- card->options.ipa6.enabled_funcs = cmd->hdr.ipa_enabled;
- }
- if (cmd->data.setassparms.hdr.assist_no == IPA_INBOUND_CHECKSUM &&
- cmd->data.setassparms.hdr.command_code == IPA_CMD_ASS_START) {
- card->info.csum_mask = cmd->data.setassparms.data.flags_32bit;
- QETH_CARD_TEXT_(card, 3, "csum:%d", card->info.csum_mask);
- }
- if (cmd->data.setassparms.hdr.assist_no == IPA_OUTBOUND_CHECKSUM &&
- cmd->data.setassparms.hdr.command_code == IPA_CMD_ASS_START) {
- card->info.tx_csum_mask =
- cmd->data.setassparms.data.flags_32bit;
- QETH_CARD_TEXT_(card, 3, "tcsu:%d", card->info.tx_csum_mask);
- }
-
- return 0;
-}
-
#ifdef CONFIG_QETH_IPV6
static int qeth_l3_send_simple_setassparms_ipv6(struct qeth_card *card,
enum qeth_ipa_funcs ipa_func, __u16 cmd_code)
@@ -952,7 +922,7 @@ static int qeth_l3_send_simple_setassparms_ipv6(struct qeth_card *card,
if (!iob)
return -ENOMEM;
rc = qeth_send_setassparms(card, iob, 0, 0,
- qeth_l3_default_setassparms_cb, NULL);
+ qeth_setassparms_cb, NULL);
return rc;
}
#endif
@@ -1187,47 +1157,6 @@ out:
return rc;
}
-static void qeth_l3_start_ipa_checksum(struct qeth_card *card)
-{
- QETH_CARD_TEXT(card, 3, "strtcsum");
- if (qeth_is_supported(card, IPA_INBOUND_CHECKSUM)
- && (card->dev->features & NETIF_F_RXCSUM))
- qeth_set_rx_csum(card, 1);
-}
-
-static void qeth_l3_start_ipa_tx_checksum(struct qeth_card *card)
-{
- QETH_CARD_TEXT(card, 3, "strttxcs");
- qeth_start_ipa_tx_checksum(card);
-}
-
-static int qeth_l3_start_ipa_tso(struct qeth_card *card)
-{
- int rc;
-
- QETH_CARD_TEXT(card, 3, "sttso");
-
- if (!qeth_is_supported(card, IPA_OUTBOUND_TSO)) {
- dev_info(&card->gdev->dev,
- "Outbound TSO not supported on %s\n",
- QETH_CARD_IFNAME(card));
- rc = -EOPNOTSUPP;
- } else {
- rc = qeth_send_simple_setassparms(card, IPA_OUTBOUND_TSO,
- IPA_CMD_ASS_START, 0);
- if (rc)
- dev_warn(&card->gdev->dev, "Starting outbound TCP "
- "segmentation offload for %s failed\n",
- QETH_CARD_IFNAME(card));
- else
- dev_info(&card->gdev->dev,
- "Outbound TSO enabled\n");
- }
- if (rc)
- card->dev->features &= ~NETIF_F_TSO;
- return rc;
-}
-
static int qeth_l3_start_ipassists(struct qeth_card *card)
{
QETH_CARD_TEXT(card, 3, "strtipas");
@@ -1241,9 +1170,6 @@ static int qeth_l3_start_ipassists(struct qeth_card *card)
qeth_l3_start_ipa_multicast(card); /* go on*/
qeth_l3_start_ipa_ipv6(card); /* go on*/
qeth_l3_start_ipa_broadcast(card); /* go on*/
- qeth_l3_start_ipa_checksum(card); /* go on*/
- qeth_l3_start_ipa_tx_checksum(card);
- qeth_l3_start_ipa_tso(card); /* go on*/
return 0;
}
@@ -2440,7 +2366,7 @@ static int qeth_l3_arp_add_entry(struct qeth_card *card,
rc = qeth_send_setassparms(card, iob,
sizeof(struct qeth_arp_cache_entry),
(unsigned long) entry,
- qeth_l3_default_setassparms_cb, NULL);
+ qeth_setassparms_cb, NULL);
if (rc) {
tmp = rc;
qeth_l3_ipaddr4_to_string((u8 *)entry->ipaddr, buf);
@@ -2480,7 +2406,7 @@ static int qeth_l3_arp_remove_entry(struct qeth_card *card,
return -ENOMEM;
rc = qeth_send_setassparms(card, iob,
12, (unsigned long)buf,
- qeth_l3_default_setassparms_cb, NULL);
+ qeth_setassparms_cb, NULL);
if (rc) {
tmp = rc;
memset(buf, 0, 16);
@@ -2826,7 +2752,7 @@ static int qeth_l3_get_elements_no_tso(struct qeth_card *card,
int elements = qeth_get_elements_for_range(
tcpdptr,
(addr_t)skb->data + skb_headlen(skb)) +
- qeth_get_elements_for_frags(skb);
+ qeth_get_elements_for_frags(skb);
if ((elements + extra_elems) > QETH_MAX_BUFFER_ELEMENTS(card)) {
QETH_DBF_MESSAGE(2,
@@ -3089,36 +3015,6 @@ static int qeth_l3_stop(struct net_device *dev)
return 0;
}
-static netdev_features_t qeth_l3_fix_features(struct net_device *dev,
- netdev_features_t features)
-{
- struct qeth_card *card = dev->ml_priv;
-
- if (!qeth_is_supported(card, IPA_OUTBOUND_CHECKSUM))
- features &= ~NETIF_F_IP_CSUM;
- if (!qeth_is_supported(card, IPA_OUTBOUND_TSO))
- features &= ~NETIF_F_TSO;
- if (!qeth_is_supported(card, IPA_INBOUND_CHECKSUM))
- features &= ~NETIF_F_RXCSUM;
- return features;
-}
-
-static int qeth_l3_set_features(struct net_device *dev,
- netdev_features_t features)
-{
- struct qeth_card *card = dev->ml_priv;
- netdev_features_t changed = dev->features ^ features;
-
- if (!(changed & NETIF_F_RXCSUM))
- return 0;
-
- if (card->state == CARD_STATE_DOWN ||
- card->state == CARD_STATE_RECOVER)
- return 0;
-
- return qeth_set_rx_csum(card, features & NETIF_F_RXCSUM ? 1 : 0);
-}
-
static const struct ethtool_ops qeth_l3_ethtool_ops = {
.get_link = ethtool_op_get_link,
.get_strings = qeth_core_get_strings,
@@ -3161,8 +3057,8 @@ static const struct net_device_ops qeth_l3_netdev_ops = {
.ndo_set_rx_mode = qeth_l3_set_multicast_list,
.ndo_do_ioctl = qeth_l3_do_ioctl,
.ndo_change_mtu = qeth_change_mtu,
- .ndo_fix_features = qeth_l3_fix_features,
- .ndo_set_features = qeth_l3_set_features,
+ .ndo_fix_features = qeth_fix_features,
+ .ndo_set_features = qeth_set_features,
.ndo_vlan_rx_add_vid = qeth_l3_vlan_rx_add_vid,
.ndo_vlan_rx_kill_vid = qeth_l3_vlan_rx_kill_vid,
.ndo_tx_timeout = qeth_tx_timeout,
@@ -3177,8 +3073,8 @@ static const struct net_device_ops qeth_l3_osa_netdev_ops = {
.ndo_set_rx_mode = qeth_l3_set_multicast_list,
.ndo_do_ioctl = qeth_l3_do_ioctl,
.ndo_change_mtu = qeth_change_mtu,
- .ndo_fix_features = qeth_l3_fix_features,
- .ndo_set_features = qeth_l3_set_features,
+ .ndo_fix_features = qeth_fix_features,
+ .ndo_set_features = qeth_set_features,
.ndo_vlan_rx_add_vid = qeth_l3_vlan_rx_add_vid,
.ndo_vlan_rx_kill_vid = qeth_l3_vlan_rx_kill_vid,
.ndo_tx_timeout = qeth_tx_timeout,
@@ -3210,8 +3106,7 @@ static int qeth_l3_setup_netdev(struct qeth_card *card)
card->dev->hw_features = NETIF_F_SG |
NETIF_F_RXCSUM | NETIF_F_IP_CSUM |
NETIF_F_TSO;
- card->dev->features = NETIF_F_RXCSUM |
- NETIF_F_SG;
+ card->dev->features = NETIF_F_SG;
}
}
} else if (card->info.type == QETH_CARD_TYPE_IQD) {
--
2.6.6
^ permalink raw reply related [flat|nested] 15+ messages in thread* [PATCH net-next 10/13] qeth: add network device features for VLAN devices
2016-06-16 14:18 [PATCH net-next 00/13] s390: qeth patches Ursula Braun
` (8 preceding siblings ...)
2016-06-16 14:18 ` [PATCH net-next 09/13] qeth layer 2 and layer 3 common feature handling Ursula Braun
@ 2016-06-16 14:19 ` Ursula Braun
2016-06-16 14:19 ` [PATCH net-next 11/13] qeth: improve set_features error handling Ursula Braun
` (3 subsequent siblings)
13 siblings, 0 replies; 15+ messages in thread
From: Ursula Braun @ 2016-06-16 14:19 UTC (permalink / raw)
To: davem; +Cc: netdev, linux-s390, schwidefsky, heiko.carstens, utz.bacher
From: Hans Wippel <hwippel@linux.vnet.ibm.com>
Network device features indicate the capabilities of network devices (e.g.,
TX checksum offloading and TSO) and their configuration state. Additional
network device features (vlan_features) indicate for each network device,
which capabilities can be used on VLAN devices, that are configured on the
respective base network device.
In the current qeth implementation, network device features are only set
for the base network devices and not for the VLAN devices. Thus, features
like TX checksum offloading cannot be used on VLAN devices.
This patch adds network device features to vlan_features, so they can be
used by VLAN devices.
Signed-off-by: Hans Wippel <hwippel@linux.vnet.ibm.com>
Signed-off-by: Ursula Braun <ubraun@linux.vnet.ibm.com>
---
drivers/s390/net/qeth_l2_main.c | 9 +++++++--
drivers/s390/net/qeth_l3_main.c | 3 +++
2 files changed, 10 insertions(+), 2 deletions(-)
diff --git a/drivers/s390/net/qeth_l2_main.c b/drivers/s390/net/qeth_l2_main.c
index 928a4ad..9fd48de 100644
--- a/drivers/s390/net/qeth_l2_main.c
+++ b/drivers/s390/net/qeth_l2_main.c
@@ -1113,11 +1113,16 @@ static int qeth_l2_setup_netdev(struct qeth_card *card)
card->dev->features |= NETIF_F_HW_VLAN_CTAG_FILTER;
if (card->info.type == QETH_CARD_TYPE_OSD && !card->info.guestlan) {
card->dev->hw_features = NETIF_F_SG;
+ card->dev->vlan_features = NETIF_F_SG;
/* OSA 3S and earlier has no RX/TX support */
- if (qeth_is_supported(card, IPA_OUTBOUND_CHECKSUM))
+ if (qeth_is_supported(card, IPA_OUTBOUND_CHECKSUM)) {
card->dev->hw_features |= NETIF_F_IP_CSUM;
- if (qeth_is_supported(card, IPA_INBOUND_CHECKSUM))
+ card->dev->vlan_features |= NETIF_F_IP_CSUM;
+ }
+ if (qeth_is_supported(card, IPA_INBOUND_CHECKSUM)) {
card->dev->hw_features |= NETIF_F_RXCSUM;
+ card->dev->vlan_features |= NETIF_F_RXCSUM;
+ }
/* Turn on SG per default */
card->dev->features |= NETIF_F_SG;
}
diff --git a/drivers/s390/net/qeth_l3_main.c b/drivers/s390/net/qeth_l3_main.c
index fc81776..05f764c 100644
--- a/drivers/s390/net/qeth_l3_main.c
+++ b/drivers/s390/net/qeth_l3_main.c
@@ -3106,6 +3106,9 @@ static int qeth_l3_setup_netdev(struct qeth_card *card)
card->dev->hw_features = NETIF_F_SG |
NETIF_F_RXCSUM | NETIF_F_IP_CSUM |
NETIF_F_TSO;
+ card->dev->vlan_features = NETIF_F_SG |
+ NETIF_F_RXCSUM | NETIF_F_IP_CSUM |
+ NETIF_F_TSO;
card->dev->features = NETIF_F_SG;
}
}
--
2.6.6
^ permalink raw reply related [flat|nested] 15+ messages in thread* [PATCH net-next 11/13] qeth: improve set_features error handling
2016-06-16 14:18 [PATCH net-next 00/13] s390: qeth patches Ursula Braun
` (9 preceding siblings ...)
2016-06-16 14:19 ` [PATCH net-next 10/13] qeth: add network device features for VLAN devices Ursula Braun
@ 2016-06-16 14:19 ` Ursula Braun
2016-06-16 14:19 ` [PATCH net-next 12/13] qeth: omit outbound queue 3 for unicast packets in Priority Queuing on HiperSockets Ursula Braun
` (2 subsequent siblings)
13 siblings, 0 replies; 15+ messages in thread
From: Ursula Braun @ 2016-06-16 14:19 UTC (permalink / raw)
To: davem; +Cc: netdev, linux-s390, schwidefsky, heiko.carstens, utz.bacher
From: Hans Wippel <hwippel@linux.vnet.ibm.com>
The function set_features is called to configure network device features
on the hardware. If errors occur, the network device features should
reflect the changed hardware state and the function should return an
error in order to notify the user.
In case of an error, the current implementation does not necessarily
save the changed hardware state in the network device features before an
error is returned.
This patch improves error handling by saving features, that could be
changed, to the network device features before returning an error. If
the device is not running, an additional check in fix_features removes
features, that require hardware changes, before they are passed to
set_features. Thus, the corresponding check was removed in set_features.
Signed-off-by: Hans Wippel <hwippel@linux.vnet.ibm.com>
Signed-off-by: Ursula Braun <ubraun@linux.vnet.ibm.com>
---
drivers/s390/net/qeth_core_main.c | 38 +++++++++++++++++++++++++++-----------
1 file changed, 27 insertions(+), 11 deletions(-)
diff --git a/drivers/s390/net/qeth_core_main.c b/drivers/s390/net/qeth_core_main.c
index 19a6ee0..44d3200 100644
--- a/drivers/s390/net/qeth_core_main.c
+++ b/drivers/s390/net/qeth_core_main.c
@@ -6125,27 +6125,38 @@ static int qeth_set_ipa_tso(struct qeth_card *card, int on)
int qeth_set_features(struct net_device *dev, netdev_features_t features)
{
struct qeth_card *card = dev->ml_priv;
- netdev_features_t changed = card->dev->features ^ features;
+ netdev_features_t changed = dev->features ^ features;
int rc = 0;
QETH_DBF_TEXT(SETUP, 2, "setfeat");
QETH_DBF_HEX(SETUP, 2, &features, sizeof(features));
- if (card->state == CARD_STATE_DOWN ||
- card->state == CARD_STATE_RECOVER)
- return 0;
-
- if ((changed & NETIF_F_IP_CSUM))
+ if ((changed & NETIF_F_IP_CSUM)) {
rc = qeth_set_ipa_csum(card,
features & NETIF_F_IP_CSUM ? 1 : 0,
IPA_OUTBOUND_CHECKSUM);
- if ((changed & NETIF_F_RXCSUM))
- rc |= qeth_set_ipa_csum(card,
+ if (rc)
+ changed ^= NETIF_F_IP_CSUM;
+ }
+ if ((changed & NETIF_F_RXCSUM)) {
+ rc = qeth_set_ipa_csum(card,
features & NETIF_F_RXCSUM ? 1 : 0,
IPA_INBOUND_CHECKSUM);
- if ((changed & NETIF_F_TSO))
- rc |= qeth_set_ipa_tso(card, features & NETIF_F_TSO ? 1 : 0);
- return rc ? -EIO : 0;
+ if (rc)
+ changed ^= NETIF_F_RXCSUM;
+ }
+ if ((changed & NETIF_F_TSO)) {
+ rc = qeth_set_ipa_tso(card, features & NETIF_F_TSO ? 1 : 0);
+ if (rc)
+ changed ^= NETIF_F_TSO;
+ }
+
+ /* everything changed successfully? */
+ if ((dev->features ^ features) == changed)
+ return 0;
+ /* something went wrong. save changed features and return error */
+ dev->features ^= changed;
+ return -EIO;
}
EXPORT_SYMBOL_GPL(qeth_set_features);
@@ -6164,6 +6175,11 @@ netdev_features_t qeth_fix_features(struct net_device *dev,
dev_info(&card->gdev->dev, "Outbound TSO not supported on %s\n",
QETH_CARD_IFNAME(card));
}
+ /* if the card isn't up, remove features that require hw changes */
+ if (card->state == CARD_STATE_DOWN ||
+ card->state == CARD_STATE_RECOVER)
+ features = features & ~(NETIF_F_IP_CSUM | NETIF_F_RXCSUM |
+ NETIF_F_TSO);
QETH_DBF_HEX(SETUP, 2, &features, sizeof(features));
return features;
}
--
2.6.6
^ permalink raw reply related [flat|nested] 15+ messages in thread* [PATCH net-next 12/13] qeth: omit outbound queue 3 for unicast packets in Priority Queuing on HiperSockets
2016-06-16 14:18 [PATCH net-next 00/13] s390: qeth patches Ursula Braun
` (10 preceding siblings ...)
2016-06-16 14:19 ` [PATCH net-next 11/13] qeth: improve set_features error handling Ursula Braun
@ 2016-06-16 14:19 ` Ursula Braun
2016-06-16 14:19 ` [PATCH net-next 13/13] s390/qeth: fix indentation in qeth_l3_arp_query Ursula Braun
2016-06-17 5:16 ` [PATCH net-next 00/13] s390: qeth patches David Miller
13 siblings, 0 replies; 15+ messages in thread
From: Ursula Braun @ 2016-06-16 14:19 UTC (permalink / raw)
To: davem; +Cc: netdev, linux-s390, schwidefsky, heiko.carstens, utz.bacher
From: Hans Wippel <hwippel@linux.vnet.ibm.com>
On HiperSockets only outbound queues 0 to 2 are available for unicast
packets. Current Priority Queuing implementation in the qeth driver puts
outgoing packets in outbound queues 0 to 3.
This puts outgoing unicast packets into outbound queue 2 instead of
outbound queue 3 when using Priority Queuing on a HiperSocket.
Additionally, the default outbound queue cannot be set to outbound queue 3
on HiperSockets.
Signed-off-by: Hans Wippel <hwippel@linux.vnet.ibm.com>
Signed-off-by: Ursula Braun <ubraun@linux.vnet.ibm.com>
---
drivers/s390/net/qeth_core_main.c | 17 +++++++++++++----
drivers/s390/net/qeth_core_sys.c | 4 ++++
2 files changed, 17 insertions(+), 4 deletions(-)
diff --git a/drivers/s390/net/qeth_core_main.c b/drivers/s390/net/qeth_core_main.c
index 44d3200..7dba6c8 100644
--- a/drivers/s390/net/qeth_core_main.c
+++ b/drivers/s390/net/qeth_core_main.c
@@ -3748,6 +3748,14 @@ void qeth_qdio_output_handler(struct ccw_device *ccwdev,
}
EXPORT_SYMBOL_GPL(qeth_qdio_output_handler);
+/* We cannot use outbound queue 3 for unicast packets on HiperSockets */
+static inline int qeth_cut_iqd_prio(struct qeth_card *card, int queue_num)
+{
+ if ((card->info.type == QETH_CARD_TYPE_IQD) && (queue_num == 3))
+ return 2;
+ return queue_num;
+}
+
/**
* Note: Function assumes that we have 4 outbound queues.
*/
@@ -3775,9 +3783,9 @@ int qeth_get_priority_queue(struct qeth_card *card, struct sk_buff *skb,
return card->qdio.default_out_queue;
}
if (card->qdio.do_prio_queueing == QETH_PRIO_Q_ING_PREC)
- return ~tos >> 6 & 3;
+ return qeth_cut_iqd_prio(card, ~tos >> 6 & 3);
if (tos & IPTOS_MINCOST)
- return 3;
+ return qeth_cut_iqd_prio(card, 3);
if (tos & IPTOS_RELIABILITY)
return 2;
if (tos & IPTOS_THROUGHPUT)
@@ -3788,11 +3796,12 @@ int qeth_get_priority_queue(struct qeth_card *card, struct sk_buff *skb,
case QETH_PRIO_Q_ING_SKB:
if (skb->priority > 5)
return 0;
- return ~skb->priority >> 1 & 3;
+ return qeth_cut_iqd_prio(card, ~skb->priority >> 1 & 3);
case QETH_PRIO_Q_ING_VLAN:
tci = &((struct ethhdr *)skb->data)->h_proto;
if (*tci == ETH_P_8021Q)
- return ~*(tci + 1) >> (VLAN_PRIO_SHIFT + 1) & 3;
+ return qeth_cut_iqd_prio(card, ~*(tci + 1) >>
+ (VLAN_PRIO_SHIFT + 1) & 3);
break;
default:
break;
diff --git a/drivers/s390/net/qeth_core_sys.c b/drivers/s390/net/qeth_core_sys.c
index e6e5b96..75b29fd2 100644
--- a/drivers/s390/net/qeth_core_sys.c
+++ b/drivers/s390/net/qeth_core_sys.c
@@ -243,6 +243,10 @@ static ssize_t qeth_dev_prioqing_store(struct device *dev,
card->qdio.do_prio_queueing = QETH_NO_PRIO_QUEUEING;
card->qdio.default_out_queue = 2;
} else if (sysfs_streq(buf, "no_prio_queueing:3")) {
+ if (card->info.type == QETH_CARD_TYPE_IQD) {
+ rc = -EPERM;
+ goto out;
+ }
card->qdio.do_prio_queueing = QETH_NO_PRIO_QUEUEING;
card->qdio.default_out_queue = 3;
} else if (sysfs_streq(buf, "no_prio_queueing")) {
--
2.6.6
^ permalink raw reply related [flat|nested] 15+ messages in thread* [PATCH net-next 13/13] s390/qeth: fix indentation in qeth_l3_arp_query
2016-06-16 14:18 [PATCH net-next 00/13] s390: qeth patches Ursula Braun
` (11 preceding siblings ...)
2016-06-16 14:19 ` [PATCH net-next 12/13] qeth: omit outbound queue 3 for unicast packets in Priority Queuing on HiperSockets Ursula Braun
@ 2016-06-16 14:19 ` Ursula Braun
2016-06-17 5:16 ` [PATCH net-next 00/13] s390: qeth patches David Miller
13 siblings, 0 replies; 15+ messages in thread
From: Ursula Braun @ 2016-06-16 14:19 UTC (permalink / raw)
To: davem; +Cc: netdev, linux-s390, schwidefsky, heiko.carstens, utz.bacher
From: Sebastian Ott <sebott@linux.vnet.ibm.com>
gcc-6 warns about obviously wrong indentation:
drivers/s390/net/qeth_l3_main.c: In function 'qeth_l3_arp_query':
drivers/s390/net/qeth_l3_main.c:2315:3: warning: this 'if' clause does not
guard... [-Wmisleading-indentation]
if (copy_to_user(udata, qinfo.udata, 4))
^~
drivers/s390/net/qeth_l3_main.c:2317:4: note: ...this statement, but the
latter is misleadingly indented as if it is guarded by the 'if'
goto free_and_out;
^~~~
Although this particular case is harmless, fix the indentation to get rid
of that warning and improve readability.
Signed-off-by: Sebastian Ott <sebott@linux.vnet.ibm.com>
Signed-off-by: Ursula Braun <ubraun@linux.vnet.ibm.com>
---
drivers/s390/net/qeth_l3_main.c | 25 ++++++++++++-------------
1 file changed, 12 insertions(+), 13 deletions(-)
diff --git a/drivers/s390/net/qeth_l3_main.c b/drivers/s390/net/qeth_l3_main.c
index 05f764c..bcd324e 100644
--- a/drivers/s390/net/qeth_l3_main.c
+++ b/drivers/s390/net/qeth_l3_main.c
@@ -2314,22 +2314,21 @@ static int qeth_l3_arp_query(struct qeth_card *card, char __user *udata)
if (rc) {
if (copy_to_user(udata, qinfo.udata, 4))
rc = -EFAULT;
- goto free_and_out;
- } else {
+ goto free_and_out;
+ }
#ifdef CONFIG_QETH_IPV6
- if (qinfo.mask_bits & QETH_QARP_WITH_IPV6) {
- /* fails in case of GuestLAN QDIO mode */
- qeth_l3_query_arp_cache_info(card, QETH_PROT_IPV6,
- &qinfo);
- }
+ if (qinfo.mask_bits & QETH_QARP_WITH_IPV6) {
+ /* fails in case of GuestLAN QDIO mode */
+ qeth_l3_query_arp_cache_info(card, QETH_PROT_IPV6, &qinfo);
+ }
#endif
- if (copy_to_user(udata, qinfo.udata, qinfo.udata_len)) {
- QETH_CARD_TEXT(card, 4, "qactf");
- rc = -EFAULT;
- goto free_and_out;
- }
- QETH_CARD_TEXT(card, 4, "qacts");
+ if (copy_to_user(udata, qinfo.udata, qinfo.udata_len)) {
+ QETH_CARD_TEXT(card, 4, "qactf");
+ rc = -EFAULT;
+ goto free_and_out;
}
+ QETH_CARD_TEXT(card, 4, "qacts");
+
free_and_out:
kfree(qinfo.udata);
out:
--
2.6.6
^ permalink raw reply related [flat|nested] 15+ messages in thread* Re: [PATCH net-next 00/13] s390: qeth patches
2016-06-16 14:18 [PATCH net-next 00/13] s390: qeth patches Ursula Braun
` (12 preceding siblings ...)
2016-06-16 14:19 ` [PATCH net-next 13/13] s390/qeth: fix indentation in qeth_l3_arp_query Ursula Braun
@ 2016-06-17 5:16 ` David Miller
13 siblings, 0 replies; 15+ messages in thread
From: David Miller @ 2016-06-17 5:16 UTC (permalink / raw)
To: ubraun; +Cc: netdev, linux-s390, schwidefsky, heiko.carstens, utz.bacher
From: Ursula Braun <ubraun@linux.vnet.ibm.com>
Date: Thu, 16 Jun 2016 16:18:50 +0200
> here are patches for the s390 qeth driver for net-next:
>
> Patch 01 is a minor improvement for the bridgeport code in qeth.
>
> Patches 02-07 take care about scatter/gather handling in qeth.
>
> Patch 08 improves handling of multicast IP addresses in the qeth layer3
> discipline.
>
> Patches 09-11 improve netdev features related functions in qeth.
>
> Patch 12 implements an outbound queue restriction for HiperSockets.
>
> Patch 13 fixes a wrong indentation in qeth_l3_main.c causing a warning
> with gcc-6.
Series applied, thanks.
^ permalink raw reply [flat|nested] 15+ messages in thread