summaryrefslogtreecommitdiff
path: root/drivers/infiniband/hw
diff options
context:
space:
mode:
Diffstat (limited to 'drivers/infiniband/hw')
-rw-r--r--drivers/infiniband/hw/amso1100/c2_provider.c23
-rw-r--r--drivers/infiniband/hw/cxgb3/iwch_provider.c19
-rw-r--r--drivers/infiniband/hw/cxgb4/Kconfig6
-rw-r--r--drivers/infiniband/hw/cxgb4/cm.c242
-rw-r--r--drivers/infiniband/hw/cxgb4/cq.c55
-rw-r--r--drivers/infiniband/hw/cxgb4/device.c220
-rw-r--r--drivers/infiniband/hw/cxgb4/iw_cxgb4.h14
-rw-r--r--drivers/infiniband/hw/cxgb4/mem.c61
-rw-r--r--drivers/infiniband/hw/cxgb4/provider.c47
-rw-r--r--drivers/infiniband/hw/cxgb4/qp.c217
-rw-r--r--drivers/infiniband/hw/cxgb4/resource.c10
-rw-r--r--drivers/infiniband/hw/cxgb4/t4.h78
-rw-r--r--drivers/infiniband/hw/cxgb4/t4fw_ri_api.h14
-rw-r--r--drivers/infiniband/hw/cxgb4/user.h5
-rw-r--r--drivers/infiniband/hw/ehca/ehca_classes.h2
-rw-r--r--drivers/infiniband/hw/ehca/ehca_cq.c1
-rw-r--r--drivers/infiniband/hw/ehca/ehca_mrmw.c257
-rw-r--r--drivers/infiniband/hw/ipath/ipath_diag.c66
-rw-r--r--drivers/infiniband/hw/ipath/ipath_dma.c43
-rw-r--r--drivers/infiniband/hw/ipath/ipath_mr.c39
-rw-r--r--drivers/infiniband/hw/mlx4/alias_GUID.c2
-rw-r--r--drivers/infiniband/hw/mlx4/cm.c80
-rw-r--r--drivers/infiniband/hw/mlx4/cq.c42
-rw-r--r--drivers/infiniband/hw/mlx4/doorbell.c4
-rw-r--r--drivers/infiniband/hw/mlx4/mad.c122
-rw-r--r--drivers/infiniband/hw/mlx4/main.c46
-rw-r--r--drivers/infiniband/hw/mlx4/mcg.c5
-rw-r--r--drivers/infiniband/hw/mlx4/mlx4_ib.h24
-rw-r--r--drivers/infiniband/hw/mlx4/mr.c39
-rw-r--r--drivers/infiniband/hw/mlx4/qp.c321
-rw-r--r--drivers/infiniband/hw/mlx4/sysfs.c5
-rw-r--r--drivers/infiniband/hw/mlx5/cq.c62
-rw-r--r--drivers/infiniband/hw/mlx5/doorbell.c4
-rw-r--r--drivers/infiniband/hw/mlx5/main.c18
-rw-r--r--drivers/infiniband/hw/mlx5/mem.c80
-rw-r--r--drivers/infiniband/hw/mlx5/mlx5_ib.h14
-rw-r--r--drivers/infiniband/hw/mlx5/mr.c157
-rw-r--r--drivers/infiniband/hw/mlx5/qp.c552
-rw-r--r--drivers/infiniband/hw/mthca/mthca_main.c8
-rw-r--r--drivers/infiniband/hw/mthca/mthca_provider.c43
-rw-r--r--drivers/infiniband/hw/nes/nes_cm.c121
-rw-r--r--drivers/infiniband/hw/nes/nes_cm.h3
-rw-r--r--drivers/infiniband/hw/nes/nes_user.h5
-rw-r--r--drivers/infiniband/hw/nes/nes_verbs.c261
-rw-r--r--drivers/infiniband/hw/nes/nes_verbs.h1
-rw-r--r--drivers/infiniband/hw/ocrdma/Makefile2
-rw-r--r--drivers/infiniband/hw/ocrdma/ocrdma.h110
-rw-r--r--drivers/infiniband/hw/ocrdma/ocrdma_abi.h7
-rw-r--r--drivers/infiniband/hw/ocrdma/ocrdma_ah.c2
-rw-r--r--drivers/infiniband/hw/ocrdma/ocrdma_hw.c299
-rw-r--r--drivers/infiniband/hw/ocrdma/ocrdma_hw.h6
-rw-r--r--drivers/infiniband/hw/ocrdma/ocrdma_main.c81
-rw-r--r--drivers/infiniband/hw/ocrdma/ocrdma_sli.h261
-rw-r--r--drivers/infiniband/hw/ocrdma/ocrdma_stats.c623
-rw-r--r--drivers/infiniband/hw/ocrdma/ocrdma_stats.h54
-rw-r--r--drivers/infiniband/hw/ocrdma/ocrdma_verbs.c233
-rw-r--r--drivers/infiniband/hw/qib/qib.h12
-rw-r--r--drivers/infiniband/hw/qib/qib_diag.c52
-rw-r--r--drivers/infiniband/hw/qib/qib_dma.c21
-rw-r--r--drivers/infiniband/hw/qib/qib_file_ops.c5
-rw-r--r--drivers/infiniband/hw/qib/qib_fs.c1
-rw-r--r--drivers/infiniband/hw/qib/qib_iba6120.c11
-rw-r--r--drivers/infiniband/hw/qib/qib_iba7220.c12
-rw-r--r--drivers/infiniband/hw/qib/qib_iba7322.c37
-rw-r--r--drivers/infiniband/hw/qib/qib_init.c96
-rw-r--r--drivers/infiniband/hw/qib/qib_mad.c44
-rw-r--r--drivers/infiniband/hw/qib/qib_mr.c14
-rw-r--r--drivers/infiniband/hw/qib/qib_pcie.c55
-rw-r--r--drivers/infiniband/hw/qib/qib_rc.c2
-rw-r--r--drivers/infiniband/hw/qib/qib_ruc.c1
-rw-r--r--drivers/infiniband/hw/qib/qib_ud.c6
-rw-r--r--drivers/infiniband/hw/qib/qib_user_sdma.c136
-rw-r--r--drivers/infiniband/hw/qib/qib_verbs.c8
-rw-r--r--drivers/infiniband/hw/qib/qib_verbs.h16
-rw-r--r--drivers/infiniband/hw/usnic/usnic_uiom.c2
75 files changed, 4230 insertions, 1417 deletions
diff --git a/drivers/infiniband/hw/amso1100/c2_provider.c b/drivers/infiniband/hw/amso1100/c2_provider.c
index 07eb3a8067d8..8af33cf1fc4e 100644
--- a/drivers/infiniband/hw/amso1100/c2_provider.c
+++ b/drivers/infiniband/hw/amso1100/c2_provider.c
@@ -431,9 +431,9 @@ static struct ib_mr *c2_reg_user_mr(struct ib_pd *pd, u64 start, u64 length,
u64 *pages;
u64 kva = 0;
int shift, n, len;
- int i, j, k;
+ int i, k, entry;
int err = 0;
- struct ib_umem_chunk *chunk;
+ struct scatterlist *sg;
struct c2_pd *c2pd = to_c2pd(pd);
struct c2_mr *c2mr;
@@ -452,10 +452,7 @@ static struct ib_mr *c2_reg_user_mr(struct ib_pd *pd, u64 start, u64 length,
}
shift = ffs(c2mr->umem->page_size) - 1;
-
- n = 0;
- list_for_each_entry(chunk, &c2mr->umem->chunk_list, list)
- n += chunk->nents;
+ n = c2mr->umem->nmap;
pages = kmalloc(n * sizeof(u64), GFP_KERNEL);
if (!pages) {
@@ -464,14 +461,12 @@ static struct ib_mr *c2_reg_user_mr(struct ib_pd *pd, u64 start, u64 length,
}
i = 0;
- list_for_each_entry(chunk, &c2mr->umem->chunk_list, list) {
- for (j = 0; j < chunk->nmap; ++j) {
- len = sg_dma_len(&chunk->page_list[j]) >> shift;
- for (k = 0; k < len; ++k) {
- pages[i++] =
- sg_dma_address(&chunk->page_list[j]) +
- (c2mr->umem->page_size * k);
- }
+ for_each_sg(c2mr->umem->sg_head.sgl, sg, c2mr->umem->nmap, entry) {
+ len = sg_dma_len(sg) >> shift;
+ for (k = 0; k < len; ++k) {
+ pages[i++] =
+ sg_dma_address(sg) +
+ (c2mr->umem->page_size * k);
}
}
diff --git a/drivers/infiniband/hw/cxgb3/iwch_provider.c b/drivers/infiniband/hw/cxgb3/iwch_provider.c
index d2283837d451..811b24a539c0 100644
--- a/drivers/infiniband/hw/cxgb3/iwch_provider.c
+++ b/drivers/infiniband/hw/cxgb3/iwch_provider.c
@@ -618,14 +618,13 @@ static struct ib_mr *iwch_reg_user_mr(struct ib_pd *pd, u64 start, u64 length,
{
__be64 *pages;
int shift, n, len;
- int i, j, k;
+ int i, k, entry;
int err = 0;
- struct ib_umem_chunk *chunk;
struct iwch_dev *rhp;
struct iwch_pd *php;
struct iwch_mr *mhp;
struct iwch_reg_user_mr_resp uresp;
-
+ struct scatterlist *sg;
PDBG("%s ib_pd %p\n", __func__, pd);
php = to_iwch_pd(pd);
@@ -645,9 +644,7 @@ static struct ib_mr *iwch_reg_user_mr(struct ib_pd *pd, u64 start, u64 length,
shift = ffs(mhp->umem->page_size) - 1;
- n = 0;
- list_for_each_entry(chunk, &mhp->umem->chunk_list, list)
- n += chunk->nents;
+ n = mhp->umem->nmap;
err = iwch_alloc_pbl(mhp, n);
if (err)
@@ -661,12 +658,10 @@ static struct ib_mr *iwch_reg_user_mr(struct ib_pd *pd, u64 start, u64 length,
i = n = 0;
- list_for_each_entry(chunk, &mhp->umem->chunk_list, list)
- for (j = 0; j < chunk->nmap; ++j) {
- len = sg_dma_len(&chunk->page_list[j]) >> shift;
+ for_each_sg(mhp->umem->sg_head.sgl, sg, mhp->umem->nmap, entry) {
+ len = sg_dma_len(sg) >> shift;
for (k = 0; k < len; ++k) {
- pages[i++] = cpu_to_be64(sg_dma_address(
- &chunk->page_list[j]) +
+ pages[i++] = cpu_to_be64(sg_dma_address(sg) +
mhp->umem->page_size * k);
if (i == PAGE_SIZE / sizeof *pages) {
err = iwch_write_pbl(mhp, pages, i, n);
@@ -676,7 +671,7 @@ static struct ib_mr *iwch_reg_user_mr(struct ib_pd *pd, u64 start, u64 length,
i = 0;
}
}
- }
+ }
if (i)
err = iwch_write_pbl(mhp, pages, i, n);
diff --git a/drivers/infiniband/hw/cxgb4/Kconfig b/drivers/infiniband/hw/cxgb4/Kconfig
index d4e8983fba53..23f38cf2c5cd 100644
--- a/drivers/infiniband/hw/cxgb4/Kconfig
+++ b/drivers/infiniband/hw/cxgb4/Kconfig
@@ -1,10 +1,10 @@
config INFINIBAND_CXGB4
- tristate "Chelsio T4 RDMA Driver"
+ tristate "Chelsio T4/T5 RDMA Driver"
depends on CHELSIO_T4 && INET && (IPV6 || IPV6=n)
select GENERIC_ALLOCATOR
---help---
- This is an iWARP/RDMA driver for the Chelsio T4 1GbE and
- 10GbE adapters.
+ This is an iWARP/RDMA driver for the Chelsio T4 and T5
+ 1GbE, 10GbE adapters and T5 40GbE adapter.
For general information about Chelsio and our products, visit
our website at <http://www.chelsio.com>.
diff --git a/drivers/infiniband/hw/cxgb4/cm.c b/drivers/infiniband/hw/cxgb4/cm.c
index d286bdebe2ab..1f863a96a480 100644
--- a/drivers/infiniband/hw/cxgb4/cm.c
+++ b/drivers/infiniband/hw/cxgb4/cm.c
@@ -98,9 +98,9 @@ int c4iw_debug;
module_param(c4iw_debug, int, 0644);
MODULE_PARM_DESC(c4iw_debug, "Enable debug logging (default=0)");
-static int peer2peer;
+static int peer2peer = 1;
module_param(peer2peer, int, 0644);
-MODULE_PARM_DESC(peer2peer, "Support peer2peer ULPs (default=0)");
+MODULE_PARM_DESC(peer2peer, "Support peer2peer ULPs (default=1)");
static int p2p_type = FW_RI_INIT_P2PTYPE_READ_REQ;
module_param(p2p_type, int, 0644);
@@ -173,12 +173,15 @@ static void start_ep_timer(struct c4iw_ep *ep)
add_timer(&ep->timer);
}
-static void stop_ep_timer(struct c4iw_ep *ep)
+static int stop_ep_timer(struct c4iw_ep *ep)
{
PDBG("%s ep %p stopping\n", __func__, ep);
del_timer_sync(&ep->timer);
- if (!test_and_set_bit(TIMEOUT, &ep->com.flags))
+ if (!test_and_set_bit(TIMEOUT, &ep->com.flags)) {
c4iw_put_ep(&ep->com);
+ return 0;
+ }
+ return 1;
}
static int c4iw_l2t_send(struct c4iw_rdev *rdev, struct sk_buff *skb,
@@ -400,7 +403,8 @@ static struct dst_entry *find_route(struct c4iw_dev *dev, __be32 local_ip,
n = dst_neigh_lookup(&rt->dst, &peer_ip);
if (!n)
return NULL;
- if (!our_interface(dev, n->dev)) {
+ if (!our_interface(dev, n->dev) &&
+ !(n->dev->flags & IFF_LOOPBACK)) {
dst_release(&rt->dst);
return NULL;
}
@@ -583,6 +587,10 @@ static int send_connect(struct c4iw_ep *ep)
opt2 |= SACK_EN(1);
if (wscale && enable_tcp_window_scaling)
opt2 |= WND_SCALE_EN(1);
+ if (is_t5(ep->com.dev->rdev.lldi.adapter_type)) {
+ opt2 |= T5_OPT_2_VALID;
+ opt2 |= V_CONG_CNTRL(CONG_ALG_TAHOE);
+ }
t4_set_arp_err_handler(skb, NULL, act_open_req_arp_failure);
if (is_t4(ep->com.dev->rdev.lldi.adapter_type)) {
@@ -759,8 +767,9 @@ static void send_mpa_req(struct c4iw_ep *ep, struct sk_buff *skb,
ep->mpa_skb = skb;
c4iw_l2t_send(&ep->com.dev->rdev, skb, ep->l2t);
start_ep_timer(ep);
- state_set(&ep->com, MPA_REQ_SENT);
+ __state_set(&ep->com, MPA_REQ_SENT);
ep->mpa_attr.initiator = 1;
+ ep->snd_seq += mpalen;
return;
}
@@ -840,6 +849,7 @@ static int send_mpa_reject(struct c4iw_ep *ep, const void *pdata, u8 plen)
t4_set_arp_err_handler(skb, NULL, arp_failure_discard);
BUG_ON(ep->mpa_skb);
ep->mpa_skb = skb;
+ ep->snd_seq += mpalen;
return c4iw_l2t_send(&ep->com.dev->rdev, skb, ep->l2t);
}
@@ -923,7 +933,8 @@ static int send_mpa_reply(struct c4iw_ep *ep, const void *pdata, u8 plen)
skb_get(skb);
t4_set_arp_err_handler(skb, NULL, arp_failure_discard);
ep->mpa_skb = skb;
- state_set(&ep->com, MPA_REP_SENT);
+ __state_set(&ep->com, MPA_REP_SENT);
+ ep->snd_seq += mpalen;
return c4iw_l2t_send(&ep->com.dev->rdev, skb, ep->l2t);
}
@@ -940,6 +951,7 @@ static int act_establish(struct c4iw_dev *dev, struct sk_buff *skb)
PDBG("%s ep %p tid %u snd_isn %u rcv_isn %u\n", __func__, ep, tid,
be32_to_cpu(req->snd_isn), be32_to_cpu(req->rcv_isn));
+ mutex_lock(&ep->com.mutex);
dst_confirm(ep->dst);
/* setup the hwtid for this connection */
@@ -963,17 +975,18 @@ static int act_establish(struct c4iw_dev *dev, struct sk_buff *skb)
send_mpa_req(ep, skb, 1);
else
send_mpa_req(ep, skb, mpa_rev);
-
+ mutex_unlock(&ep->com.mutex);
return 0;
}
-static void close_complete_upcall(struct c4iw_ep *ep)
+static void close_complete_upcall(struct c4iw_ep *ep, int status)
{
struct iw_cm_event event;
PDBG("%s ep %p tid %u\n", __func__, ep, ep->hwtid);
memset(&event, 0, sizeof(event));
event.event = IW_CM_EVENT_CLOSE;
+ event.status = status;
if (ep->com.cm_id) {
PDBG("close complete delivered ep %p cm_id %p tid %u\n",
ep, ep->com.cm_id, ep->hwtid);
@@ -987,8 +1000,7 @@ static void close_complete_upcall(struct c4iw_ep *ep)
static int abort_connection(struct c4iw_ep *ep, struct sk_buff *skb, gfp_t gfp)
{
PDBG("%s ep %p tid %u\n", __func__, ep, ep->hwtid);
- close_complete_upcall(ep);
- state_set(&ep->com, ABORTING);
+ __state_set(&ep->com, ABORTING);
set_bit(ABORT_CONN, &ep->com.history);
return send_abort(ep, skb, gfp);
}
@@ -1066,9 +1078,10 @@ static void connect_reply_upcall(struct c4iw_ep *ep, int status)
}
}
-static void connect_request_upcall(struct c4iw_ep *ep)
+static int connect_request_upcall(struct c4iw_ep *ep)
{
struct iw_cm_event event;
+ int ret;
PDBG("%s ep %p tid %u\n", __func__, ep, ep->hwtid);
memset(&event, 0, sizeof(event));
@@ -1093,15 +1106,14 @@ static void connect_request_upcall(struct c4iw_ep *ep)
event.private_data_len = ep->plen;
event.private_data = ep->mpa_pkt + sizeof(struct mpa_message);
}
- if (state_read(&ep->parent_ep->com) != DEAD) {
- c4iw_get_ep(&ep->com);
- ep->parent_ep->com.cm_id->event_handler(
- ep->parent_ep->com.cm_id,
- &event);
- }
+ c4iw_get_ep(&ep->com);
+ ret = ep->parent_ep->com.cm_id->event_handler(ep->parent_ep->com.cm_id,
+ &event);
+ if (ret)
+ c4iw_put_ep(&ep->com);
set_bit(CONNREQ_UPCALL, &ep->com.history);
c4iw_put_ep(&ep->parent_ep->com);
- ep->parent_ep = NULL;
+ return ret;
}
static void established_upcall(struct c4iw_ep *ep)
@@ -1146,7 +1158,7 @@ static int update_rx_credits(struct c4iw_ep *ep, u32 credits)
return credits;
}
-static void process_mpa_reply(struct c4iw_ep *ep, struct sk_buff *skb)
+static int process_mpa_reply(struct c4iw_ep *ep, struct sk_buff *skb)
{
struct mpa_message *mpa;
struct mpa_v2_conn_params *mpa_v2_params;
@@ -1156,17 +1168,17 @@ static void process_mpa_reply(struct c4iw_ep *ep, struct sk_buff *skb)
struct c4iw_qp_attributes attrs;
enum c4iw_qp_attr_mask mask;
int err;
+ int disconnect = 0;
PDBG("%s ep %p tid %u\n", __func__, ep, ep->hwtid);
/*
- * Stop mpa timer. If it expired, then the state has
- * changed and we bail since ep_timeout already aborted
- * the connection.
+ * Stop mpa timer. If it expired, then
+ * we ignore the MPA reply. process_timeout()
+ * will abort the connection.
*/
- stop_ep_timer(ep);
- if (state_read(&ep->com) != MPA_REQ_SENT)
- return;
+ if (stop_ep_timer(ep))
+ return 0;
/*
* If we get more than the supported amount of private data
@@ -1188,7 +1200,7 @@ static void process_mpa_reply(struct c4iw_ep *ep, struct sk_buff *skb)
* if we don't even have the mpa message, then bail.
*/
if (ep->mpa_pkt_len < sizeof(*mpa))
- return;
+ return 0;
mpa = (struct mpa_message *) ep->mpa_pkt;
/* Validate MPA header. */
@@ -1228,7 +1240,7 @@ static void process_mpa_reply(struct c4iw_ep *ep, struct sk_buff *skb)
* We'll continue process when more data arrives.
*/
if (ep->mpa_pkt_len < (sizeof(*mpa) + plen))
- return;
+ return 0;
if (mpa->flags & MPA_REJECT) {
err = -ECONNREFUSED;
@@ -1240,7 +1252,7 @@ static void process_mpa_reply(struct c4iw_ep *ep, struct sk_buff *skb)
* start reply message including private data. And
* the MPA header is valid.
*/
- state_set(&ep->com, FPDU_MODE);
+ __state_set(&ep->com, FPDU_MODE);
ep->mpa_attr.crc_enabled = (mpa->flags & MPA_CRC) | crc_enabled ? 1 : 0;
ep->mpa_attr.recv_marker_enabled = markers_enabled;
ep->mpa_attr.xmit_marker_enabled = mpa->flags & MPA_MARKERS ? 1 : 0;
@@ -1330,9 +1342,11 @@ static void process_mpa_reply(struct c4iw_ep *ep, struct sk_buff *skb)
attrs.layer_etype = LAYER_MPA | DDP_LLP;
attrs.ecode = MPA_NOMATCH_RTR;
attrs.next_state = C4IW_QP_STATE_TERMINATE;
+ attrs.send_term = 1;
err = c4iw_modify_qp(ep->com.qp->rhp, ep->com.qp,
- C4IW_QP_ATTR_NEXT_STATE, &attrs, 0);
+ C4IW_QP_ATTR_NEXT_STATE, &attrs, 1);
err = -ENOMEM;
+ disconnect = 1;
goto out;
}
@@ -1348,18 +1362,20 @@ static void process_mpa_reply(struct c4iw_ep *ep, struct sk_buff *skb)
attrs.layer_etype = LAYER_MPA | DDP_LLP;
attrs.ecode = MPA_INSUFF_IRD;
attrs.next_state = C4IW_QP_STATE_TERMINATE;
+ attrs.send_term = 1;
err = c4iw_modify_qp(ep->com.qp->rhp, ep->com.qp,
- C4IW_QP_ATTR_NEXT_STATE, &attrs, 0);
+ C4IW_QP_ATTR_NEXT_STATE, &attrs, 1);
err = -ENOMEM;
+ disconnect = 1;
goto out;
}
goto out;
err:
- state_set(&ep->com, ABORTING);
+ __state_set(&ep->com, ABORTING);
send_abort(ep, skb, GFP_KERNEL);
out:
connect_reply_upcall(ep, err);
- return;
+ return disconnect;
}
static void process_mpa_request(struct c4iw_ep *ep, struct sk_buff *skb)
@@ -1370,15 +1386,12 @@ static void process_mpa_request(struct c4iw_ep *ep, struct sk_buff *skb)
PDBG("%s ep %p tid %u\n", __func__, ep, ep->hwtid);
- if (state_read(&ep->com) != MPA_REQ_WAIT)
- return;
-
/*
* If we get more than the supported amount of private data
* then we must fail this connection.
*/
if (ep->mpa_pkt_len + skb->len > sizeof(ep->mpa_pkt)) {
- stop_ep_timer(ep);
+ (void)stop_ep_timer(ep);
abort_connection(ep, skb, GFP_KERNEL);
return;
}
@@ -1400,7 +1413,6 @@ static void process_mpa_request(struct c4iw_ep *ep, struct sk_buff *skb)
return;
PDBG("%s enter (%s line %u)\n", __func__, __FILE__, __LINE__);
- stop_ep_timer(ep);
mpa = (struct mpa_message *) ep->mpa_pkt;
/*
@@ -1409,13 +1421,13 @@ static void process_mpa_request(struct c4iw_ep *ep, struct sk_buff *skb)
if (mpa->revision > mpa_rev) {
printk(KERN_ERR MOD "%s MPA version mismatch. Local = %d,"
" Received = %d\n", __func__, mpa_rev, mpa->revision);
- stop_ep_timer(ep);
+ (void)stop_ep_timer(ep);
abort_connection(ep, skb, GFP_KERNEL);
return;
}
if (memcmp(mpa->key, MPA_KEY_REQ, sizeof(mpa->key))) {
- stop_ep_timer(ep);
+ (void)stop_ep_timer(ep);
abort_connection(ep, skb, GFP_KERNEL);
return;
}
@@ -1426,7 +1438,7 @@ static void process_mpa_request(struct c4iw_ep *ep, struct sk_buff *skb)
* Fail if there's too much private data.
*/
if (plen > MPA_MAX_PRIVATE_DATA) {
- stop_ep_timer(ep);
+ (void)stop_ep_timer(ep);
abort_connection(ep, skb, GFP_KERNEL);
return;
}
@@ -1435,7 +1447,7 @@ static void process_mpa_request(struct c4iw_ep *ep, struct sk_buff *skb)
* If plen does not account for pkt size
*/
if (ep->mpa_pkt_len > (sizeof(*mpa) + plen)) {
- stop_ep_timer(ep);
+ (void)stop_ep_timer(ep);
abort_connection(ep, skb, GFP_KERNEL);
return;
}
@@ -1492,10 +1504,24 @@ static void process_mpa_request(struct c4iw_ep *ep, struct sk_buff *skb)
ep->mpa_attr.xmit_marker_enabled, ep->mpa_attr.version,
ep->mpa_attr.p2p_type);
- state_set(&ep->com, MPA_REQ_RCVD);
-
- /* drive upcall */
- connect_request_upcall(ep);
+ /*
+ * If the endpoint timer already expired, then we ignore
+ * the start request. process_timeout() will abort
+ * the connection.
+ */
+ if (!stop_ep_timer(ep)) {
+ __state_set(&ep->com, MPA_REQ_RCVD);
+
+ /* drive upcall */
+ mutex_lock(&ep->parent_ep->com.mutex);
+ if (ep->parent_ep->com.state != DEAD) {
+ if (connect_request_upcall(ep))
+ abort_connection(ep, skb, GFP_KERNEL);
+ } else {
+ abort_connection(ep, skb, GFP_KERNEL);
+ }
+ mutex_unlock(&ep->parent_ep->com.mutex);
+ }
return;
}
@@ -1507,19 +1533,23 @@ static int rx_data(struct c4iw_dev *dev, struct sk_buff *skb)
unsigned int tid = GET_TID(hdr);
struct tid_info *t = dev->rdev.lldi.tids;
__u8 status = hdr->status;
+ int disconnect = 0;
ep = lookup_tid(t, tid);
+ if (!ep)
+ return 0;
PDBG("%s ep %p tid %u dlen %u\n", __func__, ep, ep->hwtid, dlen);
skb_pull(skb, sizeof(*hdr));
skb_trim(skb, dlen);
+ mutex_lock(&ep->com.mutex);
/* update RX credits */
update_rx_credits(ep, dlen);
- switch (state_read(&ep->com)) {
+ switch (ep->com.state) {
case MPA_REQ_SENT:
ep->rcv_seq += dlen;
- process_mpa_reply(ep, skb);
+ disconnect = process_mpa_reply(ep, skb);
break;
case MPA_REQ_WAIT:
ep->rcv_seq += dlen;
@@ -1532,15 +1562,19 @@ static int rx_data(struct c4iw_dev *dev, struct sk_buff *skb)
pr_err("%s Unexpected streaming data." \
" qpid %u ep %p state %d tid %u status %d\n",
__func__, ep->com.qp->wq.sq.qid, ep,
- state_read(&ep->com), ep->hwtid, status);
+ ep->com.state, ep->hwtid, status);
attrs.next_state = C4IW_QP_STATE_TERMINATE;
c4iw_modify_qp(ep->com.qp->rhp, ep->com.qp,
- C4IW_QP_ATTR_NEXT_STATE, &attrs, 0);
+ C4IW_QP_ATTR_NEXT_STATE, &attrs, 1);
+ disconnect = 1;
break;
}
default:
break;
}
+ mutex_unlock(&ep->com.mutex);
+ if (disconnect)
+ c4iw_ep_disconnect(ep, 0, GFP_KERNEL);
return 0;
}
@@ -1647,6 +1681,15 @@ static inline int act_open_has_tid(int status)
status != CPL_ERR_ARP_MISS;
}
+/* Returns whether a CPL status conveys negative advice.
+ */
+static int is_neg_adv(unsigned int status)
+{
+ return status == CPL_ERR_RTX_NEG_ADVICE ||
+ status == CPL_ERR_PERSIST_NEG_ADVICE ||
+ status == CPL_ERR_KEEPALV_NEG_ADVICE;
+}
+
#define ACT_OPEN_RETRY_COUNT 2
static int import_ep(struct c4iw_ep *ep, int iptype, __u8 *peer_ip,
@@ -1835,7 +1878,7 @@ static int act_open_rpl(struct c4iw_dev *dev, struct sk_buff *skb)
PDBG("%s ep %p atid %u status %u errno %d\n", __func__, ep, atid,
status, status2errno(status));
- if (status == CPL_ERR_RTX_NEG_ADVICE) {
+ if (is_neg_adv(status)) {
printk(KERN_WARNING MOD "Connection problems for atid %u\n",
atid);
return 0;
@@ -1979,6 +2022,10 @@ static void accept_cr(struct c4iw_ep *ep, struct sk_buff *skb,
if (tcph->ece && tcph->cwr)
opt2 |= CCTRL_ECN(1);
}
+ if (is_t5(ep->com.dev->rdev.lldi.adapter_type)) {
+ opt2 |= T5_OPT_2_VALID;
+ opt2 |= V_CONG_CNTRL(CONG_ALG_TAHOE);
+ }
rpl = cplhdr(skb);
INIT_TP_WR(rpl, ep->hwtid);
@@ -2240,13 +2287,13 @@ static int peer_close(struct c4iw_dev *dev, struct sk_buff *skb)
disconnect = 0;
break;
case MORIBUND:
- stop_ep_timer(ep);
+ (void)stop_ep_timer(ep);
if (ep->com.cm_id && ep->com.qp) {
attrs.next_state = C4IW_QP_STATE_IDLE;
c4iw_modify_qp(ep->com.qp->rhp, ep->com.qp,
C4IW_QP_ATTR_NEXT_STATE, &attrs, 1);
}
- close_complete_upcall(ep);
+ close_complete_upcall(ep, 0);
__state_set(&ep->com, DEAD);
release = 1;
disconnect = 0;
@@ -2265,15 +2312,6 @@ static int peer_close(struct c4iw_dev *dev, struct sk_buff *skb)
return 0;
}
-/*
- * Returns whether an ABORT_REQ_RSS message is a negative advice.
- */
-static int is_neg_adv_abort(unsigned int status)
-{
- return status == CPL_ERR_RTX_NEG_ADVICE ||
- status == CPL_ERR_PERSIST_NEG_ADVICE;
-}
-
static int peer_abort(struct c4iw_dev *dev, struct sk_buff *skb)
{
struct cpl_abort_req_rss *req = cplhdr(skb);
@@ -2287,7 +2325,7 @@ static int peer_abort(struct c4iw_dev *dev, struct sk_buff *skb)
unsigned int tid = GET_TID(req);
ep = lookup_tid(t, tid);
- if (is_neg_adv_abort(req->status)) {
+ if (is_neg_adv(req->status)) {
PDBG("%s neg_adv_abort ep %p tid %u\n", __func__, ep,
ep->hwtid);
return 0;
@@ -2309,10 +2347,10 @@ static int peer_abort(struct c4iw_dev *dev, struct sk_buff *skb)
case CONNECTING:
break;
case MPA_REQ_WAIT:
- stop_ep_timer(ep);
+ (void)stop_ep_timer(ep);
break;
case MPA_REQ_SENT:
- stop_ep_timer(ep);
+ (void)stop_ep_timer(ep);
if (mpa_rev == 1 || (mpa_rev == 2 && ep->tried_with_mpa_v1))
connect_reply_upcall(ep, -ECONNRESET);
else {
@@ -2417,7 +2455,7 @@ static int close_con_rpl(struct c4iw_dev *dev, struct sk_buff *skb)
__state_set(&ep->com, MORIBUND);
break;
case MORIBUND:
- stop_ep_timer(ep);
+ (void)stop_ep_timer(ep);
if ((ep->com.cm_id) && (ep->com.qp)) {
attrs.next_state = C4IW_QP_STATE_IDLE;
c4iw_modify_qp(ep->com.qp->rhp,
@@ -2425,7 +2463,7 @@ static int close_con_rpl(struct c4iw_dev *dev, struct sk_buff *skb)
C4IW_QP_ATTR_NEXT_STATE,
&attrs, 1);
}
- close_complete_upcall(ep);
+ close_complete_upcall(ep, 0);
__state_set(&ep->com, DEAD);
release = 1;
break;
@@ -2500,22 +2538,28 @@ static int fw4_ack(struct c4iw_dev *dev, struct sk_buff *skb)
int c4iw_reject_cr(struct iw_cm_id *cm_id, const void *pdata, u8 pdata_len)
{
- int err;
+ int err = 0;
+ int disconnect = 0;
struct c4iw_ep *ep = to_ep(cm_id);
PDBG("%s ep %p tid %u\n", __func__, ep, ep->hwtid);
- if (state_read(&ep->com) == DEAD) {
+ mutex_lock(&ep->com.mutex);
+ if (ep->com.state == DEAD) {
+ mutex_unlock(&ep->com.mutex);
c4iw_put_ep(&ep->com);
return -ECONNRESET;
}
set_bit(ULP_REJECT, &ep->com.history);
- BUG_ON(state_read(&ep->com) != MPA_REQ_RCVD);
+ BUG_ON(ep->com.state != MPA_REQ_RCVD);
if (mpa_rev == 0)
abort_connection(ep, NULL, GFP_KERNEL);
else {
err = send_mpa_reject(ep, pdata, pdata_len);
- err = c4iw_ep_disconnect(ep, 0, GFP_KERNEL);
+ disconnect = 1;
}
+ mutex_unlock(&ep->com.mutex);
+ if (disconnect)
+ err = c4iw_ep_disconnect(ep, 0, GFP_KERNEL);
c4iw_put_ep(&ep->com);
return 0;
}
@@ -2530,12 +2574,14 @@ int c4iw_accept_cr(struct iw_cm_id *cm_id, struct iw_cm_conn_param *conn_param)
struct c4iw_qp *qp = get_qhp(h, conn_param->qpn);
PDBG("%s ep %p tid %u\n", __func__, ep, ep->hwtid);
- if (state_read(&ep->com) == DEAD) {
+
+ mutex_lock(&ep->com.mutex);
+ if (ep->com.state == DEAD) {
err = -ECONNRESET;
goto err;
}
- BUG_ON(state_read(&ep->com) != MPA_REQ_RCVD);
+ BUG_ON(ep->com.state != MPA_REQ_RCVD);
BUG_ON(!qp);
set_bit(ULP_ACCEPT, &ep->com.history);
@@ -2604,14 +2650,16 @@ int c4iw_accept_cr(struct iw_cm_id *cm_id, struct iw_cm_conn_param *conn_param)
if (err)
goto err1;
- state_set(&ep->com, FPDU_MODE);
+ __state_set(&ep->com, FPDU_MODE);
established_upcall(ep);
+ mutex_unlock(&ep->com.mutex);
c4iw_put_ep(&ep->com);
return 0;
err1:
ep->com.cm_id = NULL;
cm_id->rem_ref(cm_id);
err:
+ mutex_unlock(&ep->com.mutex);
c4iw_put_ep(&ep->com);
return err;
}
@@ -2980,7 +3028,7 @@ int c4iw_ep_disconnect(struct c4iw_ep *ep, int abrupt, gfp_t gfp)
rdev = &ep->com.dev->rdev;
if (c4iw_fatal_error(rdev)) {
fatal = 1;
- close_complete_upcall(ep);
+ close_complete_upcall(ep, -EIO);
ep->com.state = DEAD;
}
switch (ep->com.state) {
@@ -3002,7 +3050,7 @@ int c4iw_ep_disconnect(struct c4iw_ep *ep, int abrupt, gfp_t gfp)
if (!test_and_set_bit(CLOSE_SENT, &ep->com.flags)) {
close = 1;
if (abrupt) {
- stop_ep_timer(ep);
+ (void)stop_ep_timer(ep);
ep->com.state = ABORTING;
} else
ep->com.state = MORIBUND;
@@ -3022,7 +3070,7 @@ int c4iw_ep_disconnect(struct c4iw_ep *ep, int abrupt, gfp_t gfp)
if (close) {
if (abrupt) {
set_bit(EP_DISC_ABORT, &ep->com.history);
- close_complete_upcall(ep);
+ close_complete_upcall(ep, -ECONNRESET);
ret = send_abort(ep, NULL, gfp);
} else {
set_bit(EP_DISC_CLOSE, &ep->com.history);
@@ -3203,6 +3251,7 @@ static void send_fw_pass_open_req(struct c4iw_dev *dev, struct sk_buff *skb,
struct sk_buff *req_skb;
struct fw_ofld_connection_wr *req;
struct cpl_pass_accept_req *cpl = cplhdr(skb);
+ int ret;
req_skb = alloc_skb(sizeof(struct fw_ofld_connection_wr), GFP_KERNEL);
req = (struct fw_ofld_connection_wr *)__skb_put(req_skb, sizeof(*req));
@@ -3239,7 +3288,13 @@ static void send_fw_pass_open_req(struct c4iw_dev *dev, struct sk_buff *skb,
req->cookie = (unsigned long)skb;
set_wr_txq(req_skb, CPL_PRIORITY_CONTROL, port_id);
- cxgb4_ofld_send(dev->rdev.lldi.ports[0], req_skb);
+ ret = cxgb4_ofld_send(dev->rdev.lldi.ports[0], req_skb);
+ if (ret < 0) {
+ pr_err("%s - cxgb4_ofld_send error %d - dropping\n", __func__,
+ ret);
+ kfree_skb(skb);
+ kfree_skb(req_skb);
+ }
}
/*
@@ -3346,13 +3401,13 @@ static int rx_pkt(struct c4iw_dev *dev, struct sk_buff *skb)
pi = (struct port_info *)netdev_priv(pdev);
tx_chan = cxgb4_port_chan(pdev);
}
+ neigh_release(neigh);
if (!e) {
pr_err("%s - failed to allocate l2t entry!\n",
__func__);
goto free_dst;
}
- neigh_release(neigh);
step = dev->rdev.lldi.nrxq / dev->rdev.lldi.nchan;
rss_qid = dev->rdev.lldi.rxq_ids[pi->port_id * step];
window = (__force u16) htons((__force u16)tcph->window);
@@ -3427,15 +3482,26 @@ static void process_timeout(struct c4iw_ep *ep)
&attrs, 1);
}
__state_set(&ep->com, ABORTING);
+ close_complete_upcall(ep, -ETIMEDOUT);
+ break;
+ case ABORTING:
+ case DEAD:
+
+ /*
+ * These states are expected if the ep timed out at the same
+ * time as another thread was calling stop_ep_timer().
+ * So we silently do nothing for these states.
+ */
+ abort = 0;
break;
default:
WARN(1, "%s unexpected state ep %p tid %u state %u\n",
__func__, ep, ep->hwtid, ep->com.state);
abort = 0;
}
- mutex_unlock(&ep->com.mutex);
if (abort)
abort_connection(ep, NULL, GFP_KERNEL);
+ mutex_unlock(&ep->com.mutex);
c4iw_put_ep(&ep->com);
}
@@ -3449,6 +3515,8 @@ static void process_timedout_eps(void)
tmp = timeout_list.next;
list_del(tmp);
+ tmp->next = NULL;
+ tmp->prev = NULL;
spin_unlock_irq(&timeout_lock);
ep = list_entry(tmp, struct c4iw_ep, entry);
process_timeout(ep);
@@ -3465,6 +3533,7 @@ static void process_work(struct work_struct *work)
unsigned int opcode;
int ret;
+ process_timedout_eps();
while ((skb = skb_dequeue(&rxq))) {
rpl = cplhdr(skb);
dev = *((struct c4iw_dev **) (skb->cb + sizeof(void *)));
@@ -3474,8 +3543,8 @@ static void process_work(struct work_struct *work)
ret = work_handlers[opcode](dev, skb);
if (!ret)
kfree_skb(skb);
+ process_timedout_eps();
}
- process_timedout_eps();
}
static DECLARE_WORK(skb_work, process_work);
@@ -3487,8 +3556,13 @@ static void ep_timeout(unsigned long arg)
spin_lock(&timeout_lock);
if (!test_and_set_bit(TIMEOUT, &ep->com.flags)) {
- list_add_tail(&ep->entry, &timeout_list);
- kickit = 1;
+ /*
+ * Only insert if it is not already on the list.
+ */
+ if (!ep->entry.next) {
+ list_add_tail(&ep->entry, &timeout_list);
+ kickit = 1;
+ }
}
spin_unlock(&timeout_lock);
if (kickit)
@@ -3570,7 +3644,7 @@ static int peer_abort_intr(struct c4iw_dev *dev, struct sk_buff *skb)
kfree_skb(skb);
return 0;
}
- if (is_neg_adv_abort(req->status)) {
+ if (is_neg_adv(req->status)) {
PDBG("%s neg_adv_abort ep %p tid %u\n", __func__, ep,
ep->hwtid);
kfree_skb(skb);
diff --git a/drivers/infiniband/hw/cxgb4/cq.c b/drivers/infiniband/hw/cxgb4/cq.c
index 88de3aa9c5b0..cfaa56ada189 100644
--- a/drivers/infiniband/hw/cxgb4/cq.c
+++ b/drivers/infiniband/hw/cxgb4/cq.c
@@ -235,27 +235,21 @@ int c4iw_flush_sq(struct c4iw_qp *qhp)
struct t4_cq *cq = &chp->cq;
int idx;
struct t4_swsqe *swsqe;
- int error = (qhp->attr.state != C4IW_QP_STATE_CLOSING &&
- qhp->attr.state != C4IW_QP_STATE_IDLE);
if (wq->sq.flush_cidx == -1)
wq->sq.flush_cidx = wq->sq.cidx;
idx = wq->sq.flush_cidx;
BUG_ON(idx >= wq->sq.size);
while (idx != wq->sq.pidx) {
- if (error) {
- swsqe = &wq->sq.sw_sq[idx];
- BUG_ON(swsqe->flushed);
- swsqe->flushed = 1;
- insert_sq_cqe(wq, cq, swsqe);
- if (wq->sq.oldest_read == swsqe) {
- BUG_ON(swsqe->opcode != FW_RI_READ_REQ);
- advance_oldest_read(wq);
- }
- flushed++;
- } else {
- t4_sq_consume(wq);
+ swsqe = &wq->sq.sw_sq[idx];
+ BUG_ON(swsqe->flushed);
+ swsqe->flushed = 1;
+ insert_sq_cqe(wq, cq, swsqe);
+ if (wq->sq.oldest_read == swsqe) {
+ BUG_ON(swsqe->opcode != FW_RI_READ_REQ);
+ advance_oldest_read(wq);
}
+ flushed++;
if (++idx == wq->sq.size)
idx = 0;
}
@@ -365,8 +359,14 @@ void c4iw_flush_hw_cq(struct c4iw_cq *chp)
if (CQE_OPCODE(hw_cqe) == FW_RI_READ_RESP) {
- /*
- * drop peer2peer RTR reads.
+ /* If we have reached here because of async
+ * event or other error, and have egress error
+ * then drop
+ */
+ if (CQE_TYPE(hw_cqe) == 1)
+ goto next_cqe;
+
+ /* drop peer2peer RTR reads.
*/
if (CQE_WRID_STAG(hw_cqe) == 1)
goto next_cqe;
@@ -511,8 +511,18 @@ static int poll_cq(struct t4_wq *wq, struct t4_cq *cq, struct t4_cqe *cqe,
*/
if (RQ_TYPE(hw_cqe) && (CQE_OPCODE(hw_cqe) == FW_RI_READ_RESP)) {
- /*
- * If this is an unsolicited read response, then the read
+ /* If we have reached here because of async
+ * event or other error, and have egress error
+ * then drop
+ */
+ if (CQE_TYPE(hw_cqe) == 1) {
+ if (CQE_STATUS(hw_cqe))
+ t4_set_wq_in_error(wq);
+ ret = -EAGAIN;
+ goto skip_cqe;
+ }
+
+ /* If this is an unsolicited read response, then the read
* was generated by the kernel driver as part of peer-2-peer
* connection setup. So ignore the completion.
*/
@@ -603,7 +613,7 @@ proc_cqe:
*/
if (SQ_TYPE(hw_cqe)) {
int idx = CQE_WRID_SQ_IDX(hw_cqe);
- BUG_ON(idx > wq->sq.size);
+ BUG_ON(idx >= wq->sq.size);
/*
* Account for any unsignaled completions completed by
@@ -617,7 +627,7 @@ proc_cqe:
wq->sq.in_use -= wq->sq.size + idx - wq->sq.cidx;
else
wq->sq.in_use -= idx - wq->sq.cidx;
- BUG_ON(wq->sq.in_use < 0 && wq->sq.in_use < wq->sq.size);
+ BUG_ON(wq->sq.in_use <= 0 && wq->sq.in_use >= wq->sq.size);
wq->sq.cidx = (uint16_t)idx;
PDBG("%s completing sq idx %u\n", __func__, wq->sq.cidx);
@@ -662,7 +672,7 @@ skip_cqe:
static int c4iw_poll_cq_one(struct c4iw_cq *chp, struct ib_wc *wc)
{
struct c4iw_qp *qhp = NULL;
- struct t4_cqe cqe = {0, 0}, *rd_cqe;
+ struct t4_cqe uninitialized_var(cqe), *rd_cqe;
struct t4_wq *wq;
u32 credit = 0;
u8 cqe_flushed;
@@ -881,7 +891,7 @@ struct ib_cq *c4iw_create_cq(struct ib_device *ibdev, int entries,
/*
* Make actual HW queue 2x to avoid cdix_inc overflows.
*/
- hwentries = entries * 2;
+ hwentries = min(entries * 2, T4_MAX_IQ_SIZE);
/*
* Make HW queue at least 64 entries so GTS updates aren't too
@@ -930,6 +940,7 @@ struct ib_cq *c4iw_create_cq(struct ib_device *ibdev, int entries,
if (!mm2)
goto err4;
+ memset(&uresp, 0, sizeof(uresp));
uresp.qid_mask = rhp->rdev.cqmask;
uresp.cqid = chp->cq.cqid;
uresp.size = chp->cq.size;
diff --git a/drivers/infiniband/hw/cxgb4/device.c b/drivers/infiniband/hw/cxgb4/device.c
index 4a033853312e..f4fa50a609e2 100644
--- a/drivers/infiniband/hw/cxgb4/device.c
+++ b/drivers/infiniband/hw/cxgb4/device.c
@@ -64,6 +64,10 @@ struct uld_ctx {
static LIST_HEAD(uld_ctx_list);
static DEFINE_MUTEX(dev_mutex);
+#define DB_FC_RESUME_SIZE 64
+#define DB_FC_RESUME_DELAY 1
+#define DB_FC_DRAIN_THRESH 0
+
static struct dentry *c4iw_debugfs_root;
struct c4iw_debugfs_data {
@@ -282,7 +286,7 @@ static const struct file_operations stag_debugfs_fops = {
.llseek = default_llseek,
};
-static char *db_state_str[] = {"NORMAL", "FLOW_CONTROL", "RECOVERY"};
+static char *db_state_str[] = {"NORMAL", "FLOW_CONTROL", "RECOVERY", "STOPPED"};
static int stats_show(struct seq_file *seq, void *v)
{
@@ -311,9 +315,10 @@ static int stats_show(struct seq_file *seq, void *v)
seq_printf(seq, " DB FULL: %10llu\n", dev->rdev.stats.db_full);
seq_printf(seq, " DB EMPTY: %10llu\n", dev->rdev.stats.db_empty);
seq_printf(seq, " DB DROP: %10llu\n", dev->rdev.stats.db_drop);
- seq_printf(seq, " DB State: %s Transitions %llu\n",
+ seq_printf(seq, " DB State: %s Transitions %llu FC Interruptions %llu\n",
db_state_str[dev->db_state],
- dev->rdev.stats.db_state_transitions);
+ dev->rdev.stats.db_state_transitions,
+ dev->rdev.stats.db_fc_interruptions);
seq_printf(seq, "TCAM_FULL: %10llu\n", dev->rdev.stats.tcam_full);
seq_printf(seq, "ACT_OFLD_CONN_FAILS: %10llu\n",
dev->rdev.stats.act_ofld_conn_fails);
@@ -643,6 +648,12 @@ static int c4iw_rdev_open(struct c4iw_rdev *rdev)
printk(KERN_ERR MOD "error %d initializing ocqp pool\n", err);
goto err4;
}
+ rdev->status_page = (struct t4_dev_status_page *)
+ __get_free_page(GFP_KERNEL);
+ if (!rdev->status_page) {
+ pr_err(MOD "error allocating status page\n");
+ goto err4;
+ }
return 0;
err4:
c4iw_rqtpool_destroy(rdev);
@@ -656,6 +667,7 @@ err1:
static void c4iw_rdev_close(struct c4iw_rdev *rdev)
{
+ free_page((unsigned long)rdev->status_page);
c4iw_pblpool_destroy(rdev);
c4iw_rqtpool_destroy(rdev);
c4iw_destroy_resource(&rdev->resource);
@@ -670,7 +682,10 @@ static void c4iw_dealloc(struct uld_ctx *ctx)
idr_destroy(&ctx->dev->hwtid_idr);
idr_destroy(&ctx->dev->stid_idr);
idr_destroy(&ctx->dev->atid_idr);
- iounmap(ctx->dev->rdev.oc_mw_kva);
+ if (ctx->dev->rdev.bar2_kva)
+ iounmap(ctx->dev->rdev.bar2_kva);
+ if (ctx->dev->rdev.oc_mw_kva)
+ iounmap(ctx->dev->rdev.oc_mw_kva);
ib_dealloc_device(&ctx->dev->ibdev);
ctx->dev = NULL;
}
@@ -703,18 +718,6 @@ static struct c4iw_dev *c4iw_alloc(const struct cxgb4_lld_info *infop)
pr_info("%s: On-Chip Queues not supported on this device.\n",
pci_name(infop->pdev));
- if (!is_t4(infop->adapter_type)) {
- if (!allow_db_fc_on_t5) {
- db_fc_threshold = 100000;
- pr_info("DB Flow Control Disabled.\n");
- }
-
- if (!allow_db_coalescing_on_t5) {
- db_coalescing_threshold = -1;
- pr_info("DB Coalescing Disabled.\n");
- }
- }
-
devp = (struct c4iw_dev *)ib_alloc_device(sizeof(*devp));
if (!devp) {
printk(KERN_ERR MOD "Cannot allocate ib device\n");
@@ -722,11 +725,31 @@ static struct c4iw_dev *c4iw_alloc(const struct cxgb4_lld_info *infop)
}
devp->rdev.lldi = *infop;
- devp->rdev.oc_mw_pa = pci_resource_start(devp->rdev.lldi.pdev, 2) +
- (pci_resource_len(devp->rdev.lldi.pdev, 2) -
- roundup_pow_of_two(devp->rdev.lldi.vr->ocq.size));
- devp->rdev.oc_mw_kva = ioremap_wc(devp->rdev.oc_mw_pa,
- devp->rdev.lldi.vr->ocq.size);
+ /*
+ * For T5 devices, we map all of BAR2 with WC.
+ * For T4 devices with onchip qp mem, we map only that part
+ * of BAR2 with WC.
+ */
+ devp->rdev.bar2_pa = pci_resource_start(devp->rdev.lldi.pdev, 2);
+ if (is_t5(devp->rdev.lldi.adapter_type)) {
+ devp->rdev.bar2_kva = ioremap_wc(devp->rdev.bar2_pa,
+ pci_resource_len(devp->rdev.lldi.pdev, 2));
+ if (!devp->rdev.bar2_kva) {
+ pr_err(MOD "Unable to ioremap BAR2\n");
+ return ERR_PTR(-EINVAL);
+ }
+ } else if (ocqp_supported(infop)) {
+ devp->rdev.oc_mw_pa =
+ pci_resource_start(devp->rdev.lldi.pdev, 2) +
+ pci_resource_len(devp->rdev.lldi.pdev, 2) -
+ roundup_pow_of_two(devp->rdev.lldi.vr->ocq.size);
+ devp->rdev.oc_mw_kva = ioremap_wc(devp->rdev.oc_mw_pa,
+ devp->rdev.lldi.vr->ocq.size);
+ if (!devp->rdev.oc_mw_kva) {
+ pr_err(MOD "Unable to ioremap onchip mem\n");
+ return ERR_PTR(-EINVAL);
+ }
+ }
PDBG(KERN_INFO MOD "ocq memory: "
"hw_start 0x%x size %u mw_pa 0x%lx mw_kva %p\n",
@@ -749,6 +772,7 @@ static struct c4iw_dev *c4iw_alloc(const struct cxgb4_lld_info *infop)
spin_lock_init(&devp->lock);
mutex_init(&devp->rdev.stats.lock);
mutex_init(&devp->db_mutex);
+ INIT_LIST_HEAD(&devp->db_fc_list);
if (c4iw_debugfs_root) {
devp->debugfs_root = debugfs_create_dir(
@@ -897,11 +921,13 @@ static int c4iw_uld_rx_handler(void *handle, const __be64 *rsp,
}
opcode = *(u8 *)rsp;
- if (c4iw_handlers[opcode])
+ if (c4iw_handlers[opcode]) {
c4iw_handlers[opcode](dev, skb);
- else
+ } else {
pr_info("%s no handler opcode 0x%x...\n", __func__,
opcode);
+ kfree_skb(skb);
+ }
return 0;
nomem:
@@ -977,13 +1003,16 @@ static int disable_qp_db(int id, void *p, void *data)
static void stop_queues(struct uld_ctx *ctx)
{
- spin_lock_irq(&ctx->dev->lock);
- if (ctx->dev->db_state == NORMAL) {
- ctx->dev->rdev.stats.db_state_transitions++;
- ctx->dev->db_state = FLOW_CONTROL;
+ unsigned long flags;
+
+ spin_lock_irqsave(&ctx->dev->lock, flags);
+ ctx->dev->rdev.stats.db_state_transitions++;
+ ctx->dev->db_state = STOPPED;
+ if (ctx->dev->rdev.flags & T4_STATUS_PAGE_DISABLED)
idr_for_each(&ctx->dev->qpidr, disable_qp_db, NULL);
- }
- spin_unlock_irq(&ctx->dev->lock);
+ else
+ ctx->dev->rdev.status_page->db_off = 1;
+ spin_unlock_irqrestore(&ctx->dev->lock, flags);
}
static int enable_qp_db(int id, void *p, void *data)
@@ -994,15 +1023,72 @@ static int enable_qp_db(int id, void *p, void *data)
return 0;
}
+static void resume_rc_qp(struct c4iw_qp *qp)
+{
+ spin_lock(&qp->lock);
+ t4_ring_sq_db(&qp->wq, qp->wq.sq.wq_pidx_inc,
+ is_t5(qp->rhp->rdev.lldi.adapter_type), NULL);
+ qp->wq.sq.wq_pidx_inc = 0;
+ t4_ring_rq_db(&qp->wq, qp->wq.rq.wq_pidx_inc,
+ is_t5(qp->rhp->rdev.lldi.adapter_type), NULL);
+ qp->wq.rq.wq_pidx_inc = 0;
+ spin_unlock(&qp->lock);
+}
+
+static void resume_a_chunk(struct uld_ctx *ctx)
+{
+ int i;
+ struct c4iw_qp *qp;
+
+ for (i = 0; i < DB_FC_RESUME_SIZE; i++) {
+ qp = list_first_entry(&ctx->dev->db_fc_list, struct c4iw_qp,
+ db_fc_entry);
+ list_del_init(&qp->db_fc_entry);
+ resume_rc_qp(qp);
+ if (list_empty(&ctx->dev->db_fc_list))
+ break;
+ }
+}
+
static void resume_queues(struct uld_ctx *ctx)
{
spin_lock_irq(&ctx->dev->lock);
- if (ctx->dev->qpcnt <= db_fc_threshold &&
- ctx->dev->db_state == FLOW_CONTROL) {
- ctx->dev->db_state = NORMAL;
- ctx->dev->rdev.stats.db_state_transitions++;
- idr_for_each(&ctx->dev->qpidr, enable_qp_db, NULL);
+ if (ctx->dev->db_state != STOPPED)
+ goto out;
+ ctx->dev->db_state = FLOW_CONTROL;
+ while (1) {
+ if (list_empty(&ctx->dev->db_fc_list)) {
+ WARN_ON(ctx->dev->db_state != FLOW_CONTROL);
+ ctx->dev->db_state = NORMAL;
+ ctx->dev->rdev.stats.db_state_transitions++;
+ if (ctx->dev->rdev.flags & T4_STATUS_PAGE_DISABLED) {
+ idr_for_each(&ctx->dev->qpidr, enable_qp_db,
+ NULL);
+ } else {
+ ctx->dev->rdev.status_page->db_off = 0;
+ }
+ break;
+ } else {
+ if (cxgb4_dbfifo_count(ctx->dev->rdev.lldi.ports[0], 1)
+ < (ctx->dev->rdev.lldi.dbfifo_int_thresh <<
+ DB_FC_DRAIN_THRESH)) {
+ resume_a_chunk(ctx);
+ }
+ if (!list_empty(&ctx->dev->db_fc_list)) {
+ spin_unlock_irq(&ctx->dev->lock);
+ if (DB_FC_RESUME_DELAY) {
+ set_current_state(TASK_UNINTERRUPTIBLE);
+ schedule_timeout(DB_FC_RESUME_DELAY);
+ }
+ spin_lock_irq(&ctx->dev->lock);
+ if (ctx->dev->db_state != FLOW_CONTROL)
+ break;
+ }
+ }
}
+out:
+ if (ctx->dev->db_state != NORMAL)
+ ctx->dev->rdev.stats.db_fc_interruptions++;
spin_unlock_irq(&ctx->dev->lock);
}
@@ -1028,12 +1114,12 @@ static int count_qps(int id, void *p, void *data)
return 0;
}
-static void deref_qps(struct qp_list qp_list)
+static void deref_qps(struct qp_list *qp_list)
{
int idx;
- for (idx = 0; idx < qp_list.idx; idx++)
- c4iw_qp_rem_ref(&qp_list.qps[idx]->ibqp);
+ for (idx = 0; idx < qp_list->idx; idx++)
+ c4iw_qp_rem_ref(&qp_list->qps[idx]->ibqp);
}
static void recover_lost_dbs(struct uld_ctx *ctx, struct qp_list *qp_list)
@@ -1044,17 +1130,22 @@ static void recover_lost_dbs(struct uld_ctx *ctx, struct qp_list *qp_list)
for (idx = 0; idx < qp_list->idx; idx++) {
struct c4iw_qp *qp = qp_list->qps[idx];
+ spin_lock_irq(&qp->rhp->lock);
+ spin_lock(&qp->lock);
ret = cxgb4_sync_txq_pidx(qp->rhp->rdev.lldi.ports[0],
qp->wq.sq.qid,
t4_sq_host_wq_pidx(&qp->wq),
t4_sq_wq_size(&qp->wq));
if (ret) {
- printk(KERN_ERR MOD "%s: Fatal error - "
+ pr_err(KERN_ERR MOD "%s: Fatal error - "
"DB overflow recovery failed - "
"error syncing SQ qid %u\n",
pci_name(ctx->lldi.pdev), qp->wq.sq.qid);
+ spin_unlock(&qp->lock);
+ spin_unlock_irq(&qp->rhp->lock);
return;
}
+ qp->wq.sq.wq_pidx_inc = 0;
ret = cxgb4_sync_txq_pidx(qp->rhp->rdev.lldi.ports[0],
qp->wq.rq.qid,
@@ -1062,12 +1153,17 @@ static void recover_lost_dbs(struct uld_ctx *ctx, struct qp_list *qp_list)
t4_rq_wq_size(&qp->wq));
if (ret) {
- printk(KERN_ERR MOD "%s: Fatal error - "
+ pr_err(KERN_ERR MOD "%s: Fatal error - "
"DB overflow recovery failed - "
"error syncing RQ qid %u\n",
pci_name(ctx->lldi.pdev), qp->wq.rq.qid);
+ spin_unlock(&qp->lock);
+ spin_unlock_irq(&qp->rhp->lock);
return;
}
+ qp->wq.rq.wq_pidx_inc = 0;
+ spin_unlock(&qp->lock);
+ spin_unlock_irq(&qp->rhp->lock);
/* Wait for the dbfifo to drain */
while (cxgb4_dbfifo_count(qp->rhp->rdev.lldi.ports[0], 1) > 0) {
@@ -1083,36 +1179,22 @@ static void recover_queues(struct uld_ctx *ctx)
struct qp_list qp_list;
int ret;
- /* lock out kernel db ringers */
- mutex_lock(&ctx->dev->db_mutex);
-
- /* put all queues in to recovery mode */
- spin_lock_irq(&ctx->dev->lock);
- ctx->dev->db_state = RECOVERY;
- ctx->dev->rdev.stats.db_state_transitions++;
- idr_for_each(&ctx->dev->qpidr, disable_qp_db, NULL);
- spin_unlock_irq(&ctx->dev->lock);
-
/* slow everybody down */
set_current_state(TASK_UNINTERRUPTIBLE);
schedule_timeout(usecs_to_jiffies(1000));
- /* Wait for the dbfifo to completely drain. */
- while (cxgb4_dbfifo_count(ctx->dev->rdev.lldi.ports[0], 1) > 0) {
- set_current_state(TASK_UNINTERRUPTIBLE);
- schedule_timeout(usecs_to_jiffies(10));
- }
-
/* flush the SGE contexts */
ret = cxgb4_flush_eq_cache(ctx->dev->rdev.lldi.ports[0]);
if (ret) {
printk(KERN_ERR MOD "%s: Fatal error - DB overflow recovery failed\n",
pci_name(ctx->lldi.pdev));
- goto out;
+ return;
}
/* Count active queues so we can build a list of queues to recover */
spin_lock_irq(&ctx->dev->lock);
+ WARN_ON(ctx->dev->db_state != STOPPED);
+ ctx->dev->db_state = RECOVERY;
idr_for_each(&ctx->dev->qpidr, count_qps, &count);
qp_list.qps = kzalloc(count * sizeof *qp_list.qps, GFP_ATOMIC);
@@ -1120,7 +1202,7 @@ static void recover_queues(struct uld_ctx *ctx)
printk(KERN_ERR MOD "%s: Fatal error - DB overflow recovery failed\n",
pci_name(ctx->lldi.pdev));
spin_unlock_irq(&ctx->dev->lock);
- goto out;
+ return;
}
qp_list.idx = 0;
@@ -1133,29 +1215,13 @@ static void recover_queues(struct uld_ctx *ctx)
recover_lost_dbs(ctx, &qp_list);
/* we're almost done! deref the qps and clean up */
- deref_qps(qp_list);
+ deref_qps(&qp_list);
kfree(qp_list.qps);
- /* Wait for the dbfifo to completely drain again */
- while (cxgb4_dbfifo_count(ctx->dev->rdev.lldi.ports[0], 1) > 0) {
- set_current_state(TASK_UNINTERRUPTIBLE);
- schedule_timeout(usecs_to_jiffies(10));
- }
-
- /* resume the queues */
spin_lock_irq(&ctx->dev->lock);
- if (ctx->dev->qpcnt > db_fc_threshold)
- ctx->dev->db_state = FLOW_CONTROL;
- else {
- ctx->dev->db_state = NORMAL;
- idr_for_each(&ctx->dev->qpidr, enable_qp_db, NULL);
- }
- ctx->dev->rdev.stats.db_state_transitions++;
+ WARN_ON(ctx->dev->db_state != RECOVERY);
+ ctx->dev->db_state = STOPPED;
spin_unlock_irq(&ctx->dev->lock);
-
-out:
- /* start up kernel db ringers again */
- mutex_unlock(&ctx->dev->db_mutex);
}
static int c4iw_uld_control(void *handle, enum cxgb4_control control, ...)
@@ -1165,9 +1231,7 @@ static int c4iw_uld_control(void *handle, enum cxgb4_control control, ...)
switch (control) {
case CXGB4_CONTROL_DB_FULL:
stop_queues(ctx);
- mutex_lock(&ctx->dev->rdev.stats.lock);
ctx->dev->rdev.stats.db_full++;
- mutex_unlock(&ctx->dev->rdev.stats.lock);
break;
case CXGB4_CONTROL_DB_EMPTY:
resume_queues(ctx);
diff --git a/drivers/infiniband/hw/cxgb4/iw_cxgb4.h b/drivers/infiniband/hw/cxgb4/iw_cxgb4.h
index 23eaeabab93b..7474b490760a 100644
--- a/drivers/infiniband/hw/cxgb4/iw_cxgb4.h
+++ b/drivers/infiniband/hw/cxgb4/iw_cxgb4.h
@@ -109,6 +109,7 @@ struct c4iw_dev_ucontext {
enum c4iw_rdev_flags {
T4_FATAL_ERROR = (1<<0),
+ T4_STATUS_PAGE_DISABLED = (1<<1),
};
struct c4iw_stat {
@@ -130,6 +131,7 @@ struct c4iw_stats {
u64 db_empty;
u64 db_drop;
u64 db_state_transitions;
+ u64 db_fc_interruptions;
u64 tcam_full;
u64 act_ofld_conn_fails;
u64 pas_ofld_conn_fails;
@@ -147,9 +149,12 @@ struct c4iw_rdev {
struct gen_pool *ocqp_pool;
u32 flags;
struct cxgb4_lld_info lldi;
+ unsigned long bar2_pa;
+ void __iomem *bar2_kva;
unsigned long oc_mw_pa;
void __iomem *oc_mw_kva;
struct c4iw_stats stats;
+ struct t4_dev_status_page *status_page;
};
static inline int c4iw_fatal_error(struct c4iw_rdev *rdev)
@@ -211,7 +216,8 @@ static inline int c4iw_wait_for_reply(struct c4iw_rdev *rdev,
enum db_state {
NORMAL = 0,
FLOW_CONTROL = 1,
- RECOVERY = 2
+ RECOVERY = 2,
+ STOPPED = 3
};
struct c4iw_dev {
@@ -225,10 +231,10 @@ struct c4iw_dev {
struct mutex db_mutex;
struct dentry *debugfs_root;
enum db_state db_state;
- int qpcnt;
struct idr hwtid_idr;
struct idr atid_idr;
struct idr stid_idr;
+ struct list_head db_fc_list;
};
static inline struct c4iw_dev *to_c4iw_dev(struct ib_device *ibdev)
@@ -369,6 +375,7 @@ struct c4iw_fr_page_list {
DEFINE_DMA_UNMAP_ADDR(mapping);
dma_addr_t dma_addr;
struct c4iw_dev *dev;
+ int pll_len;
};
static inline struct c4iw_fr_page_list *to_c4iw_fr_page_list(
@@ -428,10 +435,12 @@ struct c4iw_qp_attributes {
u8 ecode;
u16 sq_db_inc;
u16 rq_db_inc;
+ u8 send_term;
};
struct c4iw_qp {
struct ib_qp ibqp;
+ struct list_head db_fc_entry;
struct c4iw_dev *rhp;
struct c4iw_ep *ep;
struct c4iw_qp_attributes attr;
@@ -441,6 +450,7 @@ struct c4iw_qp {
atomic_t refcnt;
wait_queue_head_t wait;
struct timer_list timer;
+ int sq_sig_all;
};
static inline struct c4iw_qp *to_c4iw_qp(struct ib_qp *ibqp)
diff --git a/drivers/infiniband/hw/cxgb4/mem.c b/drivers/infiniband/hw/cxgb4/mem.c
index 41b11951a30a..ec7a2988a703 100644
--- a/drivers/infiniband/hw/cxgb4/mem.c
+++ b/drivers/infiniband/hw/cxgb4/mem.c
@@ -37,9 +37,9 @@
#include "iw_cxgb4.h"
-int use_dsgl = 1;
+int use_dsgl = 0;
module_param(use_dsgl, int, 0644);
-MODULE_PARM_DESC(use_dsgl, "Use DSGL for PBL/FastReg (default=1)");
+MODULE_PARM_DESC(use_dsgl, "Use DSGL for PBL/FastReg (default=0)");
#define T4_ULPTX_MIN_IO 32
#define C4IW_MAX_INLINE_SIZE 96
@@ -259,8 +259,12 @@ static int write_tpt_entry(struct c4iw_rdev *rdev, u32 reset_tpt_entry,
if ((!reset_tpt_entry) && (*stag == T4_STAG_UNSET)) {
stag_idx = c4iw_get_resource(&rdev->resource.tpt_table);
- if (!stag_idx)
+ if (!stag_idx) {
+ mutex_lock(&rdev->stats.lock);
+ rdev->stats.stag.fail++;
+ mutex_unlock(&rdev->stats.lock);
return -ENOMEM;
+ }
mutex_lock(&rdev->stats.lock);
rdev->stats.stag.cur += 32;
if (rdev->stats.stag.cur > rdev->stats.stag.max)
@@ -678,9 +682,9 @@ struct ib_mr *c4iw_reg_user_mr(struct ib_pd *pd, u64 start, u64 length,
{
__be64 *pages;
int shift, n, len;
- int i, j, k;
+ int i, k, entry;
int err = 0;
- struct ib_umem_chunk *chunk;
+ struct scatterlist *sg;
struct c4iw_dev *rhp;
struct c4iw_pd *php;
struct c4iw_mr *mhp;
@@ -710,10 +714,7 @@ struct ib_mr *c4iw_reg_user_mr(struct ib_pd *pd, u64 start, u64 length,
shift = ffs(mhp->umem->page_size) - 1;
- n = 0;
- list_for_each_entry(chunk, &mhp->umem->chunk_list, list)
- n += chunk->nents;
-
+ n = mhp->umem->nmap;
err = alloc_pbl(mhp, n);
if (err)
goto err;
@@ -726,24 +727,22 @@ struct ib_mr *c4iw_reg_user_mr(struct ib_pd *pd, u64 start, u64 length,
i = n = 0;
- list_for_each_entry(chunk, &mhp->umem->chunk_list, list)
- for (j = 0; j < chunk->nmap; ++j) {
- len = sg_dma_len(&chunk->page_list[j]) >> shift;
- for (k = 0; k < len; ++k) {
- pages[i++] = cpu_to_be64(sg_dma_address(
- &chunk->page_list[j]) +
- mhp->umem->page_size * k);
- if (i == PAGE_SIZE / sizeof *pages) {
- err = write_pbl(&mhp->rhp->rdev,
- pages,
- mhp->attr.pbl_addr + (n << 3), i);
- if (err)
- goto pbl_done;
- n += i;
- i = 0;
- }
+ for_each_sg(mhp->umem->sg_head.sgl, sg, mhp->umem->nmap, entry) {
+ len = sg_dma_len(sg) >> shift;
+ for (k = 0; k < len; ++k) {
+ pages[i++] = cpu_to_be64(sg_dma_address(sg) +
+ mhp->umem->page_size * k);
+ if (i == PAGE_SIZE / sizeof *pages) {
+ err = write_pbl(&mhp->rhp->rdev,
+ pages,
+ mhp->attr.pbl_addr + (n << 3), i);
+ if (err)
+ goto pbl_done;
+ n += i;
+ i = 0;
}
}
+ }
if (i)
err = write_pbl(&mhp->rhp->rdev, pages,
@@ -903,7 +902,11 @@ struct ib_fast_reg_page_list *c4iw_alloc_fastreg_pbl(struct ib_device *device,
dma_unmap_addr_set(c4pl, mapping, dma_addr);
c4pl->dma_addr = dma_addr;
c4pl->dev = dev;
- c4pl->ibpl.max_page_list_len = pll_len;
+ c4pl->pll_len = pll_len;
+
+ PDBG("%s c4pl %p pll_len %u page_list %p dma_addr %pad\n",
+ __func__, c4pl, c4pl->pll_len, c4pl->ibpl.page_list,
+ &c4pl->dma_addr);
return &c4pl->ibpl;
}
@@ -912,8 +915,12 @@ void c4iw_free_fastreg_pbl(struct ib_fast_reg_page_list *ibpl)
{
struct c4iw_fr_page_list *c4pl = to_c4iw_fr_page_list(ibpl);
+ PDBG("%s c4pl %p pll_len %u page_list %p dma_addr %pad\n",
+ __func__, c4pl, c4pl->pll_len, c4pl->ibpl.page_list,
+ &c4pl->dma_addr);
+
dma_free_coherent(&c4pl->dev->rdev.lldi.pdev->dev,
- c4pl->ibpl.max_page_list_len,
+ c4pl->pll_len,
c4pl->ibpl.page_list, dma_unmap_addr(c4pl, mapping));
kfree(c4pl);
}
diff --git a/drivers/infiniband/hw/cxgb4/provider.c b/drivers/infiniband/hw/cxgb4/provider.c
index 7e94c9a656a1..a94a3e12c349 100644
--- a/drivers/infiniband/hw/cxgb4/provider.c
+++ b/drivers/infiniband/hw/cxgb4/provider.c
@@ -106,15 +106,56 @@ static struct ib_ucontext *c4iw_alloc_ucontext(struct ib_device *ibdev,
{
struct c4iw_ucontext *context;
struct c4iw_dev *rhp = to_c4iw_dev(ibdev);
+ static int warned;
+ struct c4iw_alloc_ucontext_resp uresp;
+ int ret = 0;
+ struct c4iw_mm_entry *mm = NULL;
PDBG("%s ibdev %p\n", __func__, ibdev);
context = kzalloc(sizeof(*context), GFP_KERNEL);
- if (!context)
- return ERR_PTR(-ENOMEM);
+ if (!context) {
+ ret = -ENOMEM;
+ goto err;
+ }
+
c4iw_init_dev_ucontext(&rhp->rdev, &context->uctx);
INIT_LIST_HEAD(&context->mmaps);
spin_lock_init(&context->mmap_lock);
+
+ if (udata->outlen < sizeof(uresp)) {
+ if (!warned++)
+ pr_err(MOD "Warning - downlevel libcxgb4 (non-fatal), device status page disabled.");
+ rhp->rdev.flags |= T4_STATUS_PAGE_DISABLED;
+ } else {
+ mm = kmalloc(sizeof(*mm), GFP_KERNEL);
+ if (!mm) {
+ ret = -ENOMEM;
+ goto err_free;
+ }
+
+ uresp.status_page_size = PAGE_SIZE;
+
+ spin_lock(&context->mmap_lock);
+ uresp.status_page_key = context->key;
+ context->key += PAGE_SIZE;
+ spin_unlock(&context->mmap_lock);
+
+ ret = ib_copy_to_udata(udata, &uresp, sizeof(uresp));
+ if (ret)
+ goto err_mm;
+
+ mm->key = uresp.status_page_key;
+ mm->addr = virt_to_phys(rhp->rdev.status_page);
+ mm->len = PAGE_SIZE;
+ insert_mmap(context, mm);
+ }
return &context->ibucontext;
+err_mm:
+ kfree(mm);
+err_free:
+ kfree(context);
+err:
+ return ERR_PTR(ret);
}
static int c4iw_mmap(struct ib_ucontext *context, struct vm_area_struct *vma)
@@ -287,7 +328,7 @@ static int c4iw_query_device(struct ib_device *ibdev,
props->max_mr = c4iw_num_stags(&dev->rdev);
props->max_pd = T4_MAX_NUM_PD;
props->local_ca_ack_delay = 0;
- props->max_fast_reg_page_list_len = T4_MAX_FR_DEPTH;
+ props->max_fast_reg_page_list_len = t4_max_fr_depth(use_dsgl);
return 0;
}
diff --git a/drivers/infiniband/hw/cxgb4/qp.c b/drivers/infiniband/hw/cxgb4/qp.c
index 582936708e6e..086f62f5dc9e 100644
--- a/drivers/infiniband/hw/cxgb4/qp.c
+++ b/drivers/infiniband/hw/cxgb4/qp.c
@@ -212,13 +212,23 @@ static int create_qp(struct c4iw_rdev *rdev, struct t4_wq *wq,
wq->db = rdev->lldi.db_reg;
wq->gts = rdev->lldi.gts_reg;
- if (user) {
- wq->sq.udb = (u64)pci_resource_start(rdev->lldi.pdev, 2) +
- (wq->sq.qid << rdev->qpshift);
- wq->sq.udb &= PAGE_MASK;
- wq->rq.udb = (u64)pci_resource_start(rdev->lldi.pdev, 2) +
- (wq->rq.qid << rdev->qpshift);
- wq->rq.udb &= PAGE_MASK;
+ if (user || is_t5(rdev->lldi.adapter_type)) {
+ u32 off;
+
+ off = (wq->sq.qid << rdev->qpshift) & PAGE_MASK;
+ if (user) {
+ wq->sq.udb = (u64 __iomem *)(rdev->bar2_pa + off);
+ } else {
+ off += 128 * (wq->sq.qid & rdev->qpmask) + 8;
+ wq->sq.udb = (u64 __iomem *)(rdev->bar2_kva + off);
+ }
+ off = (wq->rq.qid << rdev->qpshift) & PAGE_MASK;
+ if (user) {
+ wq->rq.udb = (u64 __iomem *)(rdev->bar2_pa + off);
+ } else {
+ off += 128 * (wq->rq.qid & rdev->qpmask) + 8;
+ wq->rq.udb = (u64 __iomem *)(rdev->bar2_kva + off);
+ }
}
wq->rdev = rdev;
wq->rq.msn = 1;
@@ -299,9 +309,10 @@ static int create_qp(struct c4iw_rdev *rdev, struct t4_wq *wq,
if (ret)
goto free_dma;
- PDBG("%s sqid 0x%x rqid 0x%x kdb 0x%p squdb 0x%llx rqudb 0x%llx\n",
+ PDBG("%s sqid 0x%x rqid 0x%x kdb 0x%p squdb 0x%lx rqudb 0x%lx\n",
__func__, wq->sq.qid, wq->rq.qid, wq->db,
- (unsigned long long)wq->sq.udb, (unsigned long long)wq->rq.udb);
+ (__force unsigned long) wq->sq.udb,
+ (__force unsigned long) wq->rq.udb);
return 0;
free_dma:
@@ -425,6 +436,8 @@ static int build_rdma_send(struct t4_sq *sq, union t4_wr *wqe,
default:
return -EINVAL;
}
+ wqe->send.r3 = 0;
+ wqe->send.r4 = 0;
plen = 0;
if (wr->num_sge) {
@@ -555,7 +568,8 @@ static int build_fastreg(struct t4_sq *sq, union t4_wr *wqe,
int pbllen = roundup(wr->wr.fast_reg.page_list_len * sizeof(u64), 32);
int rem;
- if (wr->wr.fast_reg.page_list_len > T4_MAX_FR_DEPTH)
+ if (wr->wr.fast_reg.page_list_len >
+ t4_max_fr_depth(use_dsgl))
return -EINVAL;
wqe->fr.qpbinde_to_dcacpu = 0;
@@ -638,6 +652,48 @@ void c4iw_qp_rem_ref(struct ib_qp *qp)
wake_up(&(to_c4iw_qp(qp)->wait));
}
+static void add_to_fc_list(struct list_head *head, struct list_head *entry)
+{
+ if (list_empty(entry))
+ list_add_tail(entry, head);
+}
+
+static int ring_kernel_sq_db(struct c4iw_qp *qhp, u16 inc)
+{
+ unsigned long flags;
+
+ spin_lock_irqsave(&qhp->rhp->lock, flags);
+ spin_lock(&qhp->lock);
+ if (qhp->rhp->db_state == NORMAL)
+ t4_ring_sq_db(&qhp->wq, inc,
+ is_t5(qhp->rhp->rdev.lldi.adapter_type), NULL);
+ else {
+ add_to_fc_list(&qhp->rhp->db_fc_list, &qhp->db_fc_entry);
+ qhp->wq.sq.wq_pidx_inc += inc;
+ }
+ spin_unlock(&qhp->lock);
+ spin_unlock_irqrestore(&qhp->rhp->lock, flags);
+ return 0;
+}
+
+static int ring_kernel_rq_db(struct c4iw_qp *qhp, u16 inc)
+{
+ unsigned long flags;
+
+ spin_lock_irqsave(&qhp->rhp->lock, flags);
+ spin_lock(&qhp->lock);
+ if (qhp->rhp->db_state == NORMAL)
+ t4_ring_rq_db(&qhp->wq, inc,
+ is_t5(qhp->rhp->rdev.lldi.adapter_type), NULL);
+ else {
+ add_to_fc_list(&qhp->rhp->db_fc_list, &qhp->db_fc_entry);
+ qhp->wq.rq.wq_pidx_inc += inc;
+ }
+ spin_unlock(&qhp->lock);
+ spin_unlock_irqrestore(&qhp->rhp->lock, flags);
+ return 0;
+}
+
int c4iw_post_send(struct ib_qp *ibqp, struct ib_send_wr *wr,
struct ib_send_wr **bad_wr)
{
@@ -646,7 +702,7 @@ int c4iw_post_send(struct ib_qp *ibqp, struct ib_send_wr *wr,
enum fw_wr_opcodes fw_opcode = 0;
enum fw_ri_wr_flags fw_flags;
struct c4iw_qp *qhp;
- union t4_wr *wqe;
+ union t4_wr *wqe = NULL;
u32 num_wrs;
struct t4_swsqe *swsqe;
unsigned long flag;
@@ -675,7 +731,7 @@ int c4iw_post_send(struct ib_qp *ibqp, struct ib_send_wr *wr,
fw_flags = 0;
if (wr->send_flags & IB_SEND_SOLICITED)
fw_flags |= FW_RI_SOLICITED_EVENT_FLAG;
- if (wr->send_flags & IB_SEND_SIGNALED)
+ if (wr->send_flags & IB_SEND_SIGNALED || qhp->sq_sig_all)
fw_flags |= FW_RI_COMPLETION_FLAG;
swsqe = &qhp->wq.sq.sw_sq[qhp->wq.sq.pidx];
switch (wr->opcode) {
@@ -736,7 +792,8 @@ int c4iw_post_send(struct ib_qp *ibqp, struct ib_send_wr *wr,
}
swsqe->idx = qhp->wq.sq.pidx;
swsqe->complete = 0;
- swsqe->signaled = (wr->send_flags & IB_SEND_SIGNALED);
+ swsqe->signaled = (wr->send_flags & IB_SEND_SIGNALED) ||
+ qhp->sq_sig_all;
swsqe->flushed = 0;
swsqe->wr_id = wr->wr_id;
@@ -750,9 +807,14 @@ int c4iw_post_send(struct ib_qp *ibqp, struct ib_send_wr *wr,
t4_sq_produce(&qhp->wq, len16);
idx += DIV_ROUND_UP(len16*16, T4_EQ_ENTRY_SIZE);
}
- if (t4_wq_db_enabled(&qhp->wq))
- t4_ring_sq_db(&qhp->wq, idx);
- spin_unlock_irqrestore(&qhp->lock, flag);
+ if (!qhp->rhp->rdev.status_page->db_off) {
+ t4_ring_sq_db(&qhp->wq, idx,
+ is_t5(qhp->rhp->rdev.lldi.adapter_type), wqe);
+ spin_unlock_irqrestore(&qhp->lock, flag);
+ } else {
+ spin_unlock_irqrestore(&qhp->lock, flag);
+ ring_kernel_sq_db(qhp, idx);
+ }
return err;
}
@@ -761,7 +823,7 @@ int c4iw_post_receive(struct ib_qp *ibqp, struct ib_recv_wr *wr,
{
int err = 0;
struct c4iw_qp *qhp;
- union t4_recv_wr *wqe;
+ union t4_recv_wr *wqe = NULL;
u32 num_wrs;
u8 len16 = 0;
unsigned long flag;
@@ -812,9 +874,14 @@ int c4iw_post_receive(struct ib_qp *ibqp, struct ib_recv_wr *wr,
wr = wr->next;
num_wrs--;
}
- if (t4_wq_db_enabled(&qhp->wq))
- t4_ring_rq_db(&qhp->wq, idx);
- spin_unlock_irqrestore(&qhp->lock, flag);
+ if (!qhp->rhp->rdev.status_page->db_off) {
+ t4_ring_rq_db(&qhp->wq, idx,
+ is_t5(qhp->rhp->rdev.lldi.adapter_type), wqe);
+ spin_unlock_irqrestore(&qhp->lock, flag);
+ } else {
+ spin_unlock_irqrestore(&qhp->lock, flag);
+ ring_kernel_rq_db(qhp, idx);
+ }
return err;
}
@@ -1200,35 +1267,6 @@ out:
return ret;
}
-/*
- * Called by the library when the qp has user dbs disabled due to
- * a DB_FULL condition. This function will single-thread all user
- * DB rings to avoid overflowing the hw db-fifo.
- */
-static int ring_kernel_db(struct c4iw_qp *qhp, u32 qid, u16 inc)
-{
- int delay = db_delay_usecs;
-
- mutex_lock(&qhp->rhp->db_mutex);
- do {
-
- /*
- * The interrupt threshold is dbfifo_int_thresh << 6. So
- * make sure we don't cross that and generate an interrupt.
- */
- if (cxgb4_dbfifo_count(qhp->rhp->rdev.lldi.ports[0], 1) <
- (qhp->rhp->rdev.lldi.dbfifo_int_thresh << 5)) {
- writel(QID(qid) | PIDX(inc), qhp->wq.db);
- break;
- }
- set_current_state(TASK_UNINTERRUPTIBLE);
- schedule_timeout(usecs_to_jiffies(delay));
- delay = min(delay << 1, 2000);
- } while (1);
- mutex_unlock(&qhp->rhp->db_mutex);
- return 0;
-}
-
int c4iw_modify_qp(struct c4iw_dev *rhp, struct c4iw_qp *qhp,
enum c4iw_qp_attr_mask mask,
struct c4iw_qp_attributes *attrs,
@@ -1278,11 +1316,11 @@ int c4iw_modify_qp(struct c4iw_dev *rhp, struct c4iw_qp *qhp,
}
if (mask & C4IW_QP_ATTR_SQ_DB) {
- ret = ring_kernel_db(qhp, qhp->wq.sq.qid, attrs->sq_db_inc);
+ ret = ring_kernel_sq_db(qhp, attrs->sq_db_inc);
goto out;
}
if (mask & C4IW_QP_ATTR_RQ_DB) {
- ret = ring_kernel_db(qhp, qhp->wq.rq.qid, attrs->rq_db_inc);
+ ret = ring_kernel_rq_db(qhp, attrs->rq_db_inc);
goto out;
}
@@ -1332,6 +1370,7 @@ int c4iw_modify_qp(struct c4iw_dev *rhp, struct c4iw_qp *qhp,
switch (attrs->next_state) {
case C4IW_QP_STATE_CLOSING:
BUG_ON(atomic_read(&qhp->ep->com.kref.refcount) < 2);
+ t4_set_wq_in_error(&qhp->wq);
set_state(qhp, C4IW_QP_STATE_CLOSING);
ep = qhp->ep;
if (!internal) {
@@ -1339,30 +1378,30 @@ int c4iw_modify_qp(struct c4iw_dev *rhp, struct c4iw_qp *qhp,
disconnect = 1;
c4iw_get_ep(&qhp->ep->com);
}
- t4_set_wq_in_error(&qhp->wq);
ret = rdma_fini(rhp, qhp, ep);
if (ret)
goto err;
break;
case C4IW_QP_STATE_TERMINATE:
+ t4_set_wq_in_error(&qhp->wq);
set_state(qhp, C4IW_QP_STATE_TERMINATE);
qhp->attr.layer_etype = attrs->layer_etype;
qhp->attr.ecode = attrs->ecode;
- t4_set_wq_in_error(&qhp->wq);
ep = qhp->ep;
- disconnect = 1;
- if (!internal)
+ if (!internal) {
+ c4iw_get_ep(&qhp->ep->com);
terminate = 1;
- else {
+ disconnect = 1;
+ } else {
+ terminate = qhp->attr.send_term;
ret = rdma_fini(rhp, qhp, ep);
if (ret)
goto err;
}
- c4iw_get_ep(&qhp->ep->com);
break;
case C4IW_QP_STATE_ERROR:
- set_state(qhp, C4IW_QP_STATE_ERROR);
t4_set_wq_in_error(&qhp->wq);
+ set_state(qhp, C4IW_QP_STATE_ERROR);
if (!internal) {
abort = 1;
disconnect = 1;
@@ -1465,14 +1504,6 @@ out:
return ret;
}
-static int enable_qp_db(int id, void *p, void *data)
-{
- struct c4iw_qp *qp = p;
-
- t4_enable_wq_db(&qp->wq);
- return 0;
-}
-
int c4iw_destroy_qp(struct ib_qp *ib_qp)
{
struct c4iw_dev *rhp;
@@ -1490,22 +1521,15 @@ int c4iw_destroy_qp(struct ib_qp *ib_qp)
c4iw_modify_qp(rhp, qhp, C4IW_QP_ATTR_NEXT_STATE, &attrs, 0);
wait_event(qhp->wait, !qhp->ep);
- spin_lock_irq(&rhp->lock);
- remove_handle_nolock(rhp, &rhp->qpidr, qhp->wq.sq.qid);
- rhp->qpcnt--;
- BUG_ON(rhp->qpcnt < 0);
- if (rhp->qpcnt <= db_fc_threshold && rhp->db_state == FLOW_CONTROL) {
- rhp->rdev.stats.db_state_transitions++;
- rhp->db_state = NORMAL;
- idr_for_each(&rhp->qpidr, enable_qp_db, NULL);
- }
- if (db_coalescing_threshold >= 0)
- if (rhp->qpcnt <= db_coalescing_threshold)
- cxgb4_enable_db_coalescing(rhp->rdev.lldi.ports[0]);
- spin_unlock_irq(&rhp->lock);
+ remove_handle(rhp, &rhp->qpidr, qhp->wq.sq.qid);
atomic_dec(&qhp->refcnt);
wait_event(qhp->wait, !atomic_read(&qhp->refcnt));
+ spin_lock_irq(&rhp->lock);
+ if (!list_empty(&qhp->db_fc_entry))
+ list_del_init(&qhp->db_fc_entry);
+ spin_unlock_irq(&rhp->lock);
+
ucontext = ib_qp->uobject ?
to_c4iw_ucontext(ib_qp->uobject->context) : NULL;
destroy_qp(&rhp->rdev, &qhp->wq,
@@ -1516,14 +1540,6 @@ int c4iw_destroy_qp(struct ib_qp *ib_qp)
return 0;
}
-static int disable_qp_db(int id, void *p, void *data)
-{
- struct c4iw_qp *qp = p;
-
- t4_disable_wq_db(&qp->wq);
- return 0;
-}
-
struct ib_qp *c4iw_create_qp(struct ib_pd *pd, struct ib_qp_init_attr *attrs,
struct ib_udata *udata)
{
@@ -1533,7 +1549,7 @@ struct ib_qp *c4iw_create_qp(struct ib_pd *pd, struct ib_qp_init_attr *attrs,
struct c4iw_cq *schp;
struct c4iw_cq *rchp;
struct c4iw_create_qp_resp uresp;
- int sqsize, rqsize;
+ unsigned int sqsize, rqsize;
struct c4iw_ucontext *ucontext;
int ret;
struct c4iw_mm_entry *mm1, *mm2, *mm3, *mm4, *mm5 = NULL;
@@ -1605,25 +1621,13 @@ struct ib_qp *c4iw_create_qp(struct ib_pd *pd, struct ib_qp_init_attr *attrs,
qhp->attr.enable_bind = 1;
qhp->attr.max_ord = 1;
qhp->attr.max_ird = 1;
+ qhp->sq_sig_all = attrs->sq_sig_type == IB_SIGNAL_ALL_WR;
spin_lock_init(&qhp->lock);
mutex_init(&qhp->mutex);
init_waitqueue_head(&qhp->wait);
atomic_set(&qhp->refcnt, 1);
- spin_lock_irq(&rhp->lock);
- if (rhp->db_state != NORMAL)
- t4_disable_wq_db(&qhp->wq);
- rhp->qpcnt++;
- if (rhp->qpcnt > db_fc_threshold && rhp->db_state == NORMAL) {
- rhp->rdev.stats.db_state_transitions++;
- rhp->db_state = FLOW_CONTROL;
- idr_for_each(&rhp->qpidr, disable_qp_db, NULL);
- }
- if (db_coalescing_threshold >= 0)
- if (rhp->qpcnt > db_coalescing_threshold)
- cxgb4_disable_db_coalescing(rhp->rdev.lldi.ports[0]);
- ret = insert_handle_nolock(rhp, &rhp->qpidr, qhp, qhp->wq.sq.qid);
- spin_unlock_irq(&rhp->lock);
+ ret = insert_handle(rhp, &rhp->qpidr, qhp, qhp->wq.sq.qid);
if (ret)
goto err2;
@@ -1692,11 +1696,11 @@ struct ib_qp *c4iw_create_qp(struct ib_pd *pd, struct ib_qp_init_attr *attrs,
mm2->len = PAGE_ALIGN(qhp->wq.rq.memsize);
insert_mmap(ucontext, mm2);
mm3->key = uresp.sq_db_gts_key;
- mm3->addr = qhp->wq.sq.udb;
+ mm3->addr = (__force unsigned long) qhp->wq.sq.udb;
mm3->len = PAGE_SIZE;
insert_mmap(ucontext, mm3);
mm4->key = uresp.rq_db_gts_key;
- mm4->addr = qhp->wq.rq.udb;
+ mm4->addr = (__force unsigned long) qhp->wq.rq.udb;
mm4->len = PAGE_SIZE;
insert_mmap(ucontext, mm4);
if (mm5) {
@@ -1709,6 +1713,7 @@ struct ib_qp *c4iw_create_qp(struct ib_pd *pd, struct ib_qp_init_attr *attrs,
}
qhp->ibqp.qp_num = qhp->wq.sq.qid;
init_timer(&(qhp->timer));
+ INIT_LIST_HEAD(&qhp->db_fc_entry);
PDBG("%s qhp %p sq_num_entries %d, rq_num_entries %d qpid 0x%0x\n",
__func__, qhp, qhp->attr.sq_num_entries, qhp->attr.rq_num_entries,
qhp->wq.sq.qid);
@@ -1772,11 +1777,15 @@ int c4iw_ib_modify_qp(struct ib_qp *ibqp, struct ib_qp_attr *attr,
/*
* Use SQ_PSN and RQ_PSN to pass in IDX_INC values for
* ringing the queue db when we're in DB_FULL mode.
+ * Only allow this on T4 devices.
*/
attrs.sq_db_inc = attr->sq_psn;
attrs.rq_db_inc = attr->rq_psn;
mask |= (attr_mask & IB_QP_SQ_PSN) ? C4IW_QP_ATTR_SQ_DB : 0;
mask |= (attr_mask & IB_QP_RQ_PSN) ? C4IW_QP_ATTR_RQ_DB : 0;
+ if (is_t5(to_c4iw_qp(ibqp)->rhp->rdev.lldi.adapter_type) &&
+ (mask & (C4IW_QP_ATTR_SQ_DB|C4IW_QP_ATTR_RQ_DB)))
+ return -EINVAL;
return c4iw_modify_qp(rhp, qhp, mask, &attrs, 0);
}
diff --git a/drivers/infiniband/hw/cxgb4/resource.c b/drivers/infiniband/hw/cxgb4/resource.c
index cdef4d7fb6d8..67df71a7012e 100644
--- a/drivers/infiniband/hw/cxgb4/resource.c
+++ b/drivers/infiniband/hw/cxgb4/resource.c
@@ -179,8 +179,12 @@ u32 c4iw_get_qpid(struct c4iw_rdev *rdev, struct c4iw_dev_ucontext *uctx)
kfree(entry);
} else {
qid = c4iw_get_resource(&rdev->resource.qid_table);
- if (!qid)
+ if (!qid) {
+ mutex_lock(&rdev->stats.lock);
+ rdev->stats.qid.fail++;
+ mutex_unlock(&rdev->stats.lock);
goto out;
+ }
mutex_lock(&rdev->stats.lock);
rdev->stats.qid.cur += rdev->qpmask + 1;
mutex_unlock(&rdev->stats.lock);
@@ -322,8 +326,8 @@ u32 c4iw_rqtpool_alloc(struct c4iw_rdev *rdev, int size)
unsigned long addr = gen_pool_alloc(rdev->rqt_pool, size << 6);
PDBG("%s addr 0x%x size %d\n", __func__, (u32)addr, size << 6);
if (!addr)
- printk_ratelimited(KERN_WARNING MOD "%s: Out of RQT memory\n",
- pci_name(rdev->lldi.pdev));
+ pr_warn_ratelimited(MOD "%s: Out of RQT memory\n",
+ pci_name(rdev->lldi.pdev));
mutex_lock(&rdev->stats.lock);
if (addr) {
rdev->stats.rqt.cur += roundup(size << 6, 1 << MIN_RQT_SHIFT);
diff --git a/drivers/infiniband/hw/cxgb4/t4.h b/drivers/infiniband/hw/cxgb4/t4.h
index e73ace739183..2178f3198410 100644
--- a/drivers/infiniband/hw/cxgb4/t4.h
+++ b/drivers/infiniband/hw/cxgb4/t4.h
@@ -84,7 +84,14 @@ struct t4_status_page {
sizeof(struct fw_ri_isgl)) / sizeof(struct fw_ri_sge))
#define T4_MAX_FR_IMMD ((T4_SQ_NUM_BYTES - sizeof(struct fw_ri_fr_nsmr_wr) - \
sizeof(struct fw_ri_immd)) & ~31UL)
-#define T4_MAX_FR_DEPTH (1024 / sizeof(u64))
+#define T4_MAX_FR_IMMD_DEPTH (T4_MAX_FR_IMMD / sizeof(u64))
+#define T4_MAX_FR_DSGL 1024
+#define T4_MAX_FR_DSGL_DEPTH (T4_MAX_FR_DSGL / sizeof(u64))
+
+static inline int t4_max_fr_depth(int use_dsgl)
+{
+ return use_dsgl ? T4_MAX_FR_DSGL_DEPTH : T4_MAX_FR_IMMD_DEPTH;
+}
#define T4_RQ_NUM_SLOTS 2
#define T4_RQ_NUM_BYTES (T4_EQ_ENTRY_SIZE * T4_RQ_NUM_SLOTS)
@@ -292,7 +299,7 @@ struct t4_sq {
unsigned long phys_addr;
struct t4_swsqe *sw_sq;
struct t4_swsqe *oldest_read;
- u64 udb;
+ u64 __iomem *udb;
size_t memsize;
u32 qid;
u16 in_use;
@@ -300,6 +307,7 @@ struct t4_sq {
u16 cidx;
u16 pidx;
u16 wq_pidx;
+ u16 wq_pidx_inc;
u16 flags;
short flush_cidx;
};
@@ -313,7 +321,7 @@ struct t4_rq {
dma_addr_t dma_addr;
DEFINE_DMA_UNMAP_ADDR(mapping);
struct t4_swrqe *sw_rq;
- u64 udb;
+ u64 __iomem *udb;
size_t memsize;
u32 qid;
u32 msn;
@@ -324,6 +332,7 @@ struct t4_rq {
u16 cidx;
u16 pidx;
u16 wq_pidx;
+ u16 wq_pidx_inc;
};
struct t4_wq {
@@ -433,15 +442,67 @@ static inline u16 t4_sq_wq_size(struct t4_wq *wq)
return wq->sq.size * T4_SQ_NUM_SLOTS;
}
-static inline void t4_ring_sq_db(struct t4_wq *wq, u16 inc)
+/* This function copies 64 byte coalesced work request to memory
+ * mapped BAR2 space. For coalesced WRs, the SGE fetches data
+ * from the FIFO instead of from Host.
+ */
+static inline void pio_copy(u64 __iomem *dst, u64 *src)
+{
+ int count = 8;
+
+ while (count) {
+ writeq(*src, dst);
+ src++;
+ dst++;
+ count--;
+ }
+}
+
+static inline void t4_ring_sq_db(struct t4_wq *wq, u16 inc, u8 t5,
+ union t4_wr *wqe)
{
+
+ /* Flush host queue memory writes. */
wmb();
+ if (t5) {
+ if (inc == 1 && wqe) {
+ PDBG("%s: WC wq->sq.pidx = %d\n",
+ __func__, wq->sq.pidx);
+ pio_copy(wq->sq.udb + 7, (void *)wqe);
+ } else {
+ PDBG("%s: DB wq->sq.pidx = %d\n",
+ __func__, wq->sq.pidx);
+ writel(PIDX_T5(inc), wq->sq.udb);
+ }
+
+ /* Flush user doorbell area writes. */
+ wmb();
+ return;
+ }
writel(QID(wq->sq.qid) | PIDX(inc), wq->db);
}
-static inline void t4_ring_rq_db(struct t4_wq *wq, u16 inc)
+static inline void t4_ring_rq_db(struct t4_wq *wq, u16 inc, u8 t5,
+ union t4_recv_wr *wqe)
{
+
+ /* Flush host queue memory writes. */
wmb();
+ if (t5) {
+ if (inc == 1 && wqe) {
+ PDBG("%s: WC wq->rq.pidx = %d\n",
+ __func__, wq->rq.pidx);
+ pio_copy(wq->rq.udb + 7, (void *)wqe);
+ } else {
+ PDBG("%s: DB wq->rq.pidx = %d\n",
+ __func__, wq->rq.pidx);
+ writel(PIDX_T5(inc), wq->rq.udb);
+ }
+
+ /* Flush user doorbell area writes. */
+ wmb();
+ return;
+ }
writel(QID(wq->rq.qid) | PIDX(inc), wq->db);
}
@@ -566,6 +627,9 @@ static inline int t4_next_hw_cqe(struct t4_cq *cq, struct t4_cqe **cqe)
printk(KERN_ERR MOD "cq overflow cqid %u\n", cq->cqid);
BUG_ON(1);
} else if (t4_valid_cqe(cq, &cq->queue[cq->cidx])) {
+
+ /* Ensure CQE is flushed to memory */
+ rmb();
*cqe = &cq->queue[cq->cidx];
ret = 0;
} else
@@ -609,3 +673,7 @@ static inline void t4_set_cq_in_error(struct t4_cq *cq)
((struct t4_status_page *)&cq->queue[cq->size])->qp_err = 1;
}
#endif
+
+struct t4_dev_status_page {
+ u8 db_off;
+};
diff --git a/drivers/infiniband/hw/cxgb4/t4fw_ri_api.h b/drivers/infiniband/hw/cxgb4/t4fw_ri_api.h
index dc193c292671..6121ca08fe58 100644
--- a/drivers/infiniband/hw/cxgb4/t4fw_ri_api.h
+++ b/drivers/infiniband/hw/cxgb4/t4fw_ri_api.h
@@ -836,4 +836,18 @@ struct ulptx_idata {
#define V_RX_DACK_CHANGE(x) ((x) << S_RX_DACK_CHANGE)
#define F_RX_DACK_CHANGE V_RX_DACK_CHANGE(1U)
+enum { /* TCP congestion control algorithms */
+ CONG_ALG_RENO,
+ CONG_ALG_TAHOE,
+ CONG_ALG_NEWRENO,
+ CONG_ALG_HIGHSPEED
+};
+
+#define S_CONG_CNTRL 14
+#define M_CONG_CNTRL 0x3
+#define V_CONG_CNTRL(x) ((x) << S_CONG_CNTRL)
+#define G_CONG_CNTRL(x) (((x) >> S_CONG_CNTRL) & M_CONG_CNTRL)
+
+#define T5_OPT_2_VALID (1 << 31)
+
#endif /* _T4FW_RI_API_H_ */
diff --git a/drivers/infiniband/hw/cxgb4/user.h b/drivers/infiniband/hw/cxgb4/user.h
index 32b754c35ab7..11ccd276e5d9 100644
--- a/drivers/infiniband/hw/cxgb4/user.h
+++ b/drivers/infiniband/hw/cxgb4/user.h
@@ -70,4 +70,9 @@ struct c4iw_create_qp_resp {
__u32 qid_mask;
__u32 flags;
};
+
+struct c4iw_alloc_ucontext_resp {
+ __u64 status_page_key;
+ __u32 status_page_size;
+};
#endif
diff --git a/drivers/infiniband/hw/ehca/ehca_classes.h b/drivers/infiniband/hw/ehca/ehca_classes.h
index f08f6eaf3fa8..bd45e0f3923f 100644
--- a/drivers/infiniband/hw/ehca/ehca_classes.h
+++ b/drivers/infiniband/hw/ehca/ehca_classes.h
@@ -322,7 +322,7 @@ struct ehca_mr_pginfo {
} phy;
struct { /* type EHCA_MR_PGI_USER section */
struct ib_umem *region;
- struct ib_umem_chunk *next_chunk;
+ struct scatterlist *next_sg;
u64 next_nmap;
} usr;
struct { /* type EHCA_MR_PGI_FMR section */
diff --git a/drivers/infiniband/hw/ehca/ehca_cq.c b/drivers/infiniband/hw/ehca/ehca_cq.c
index 212150c25ea0..8cc837537768 100644
--- a/drivers/infiniband/hw/ehca/ehca_cq.c
+++ b/drivers/infiniband/hw/ehca/ehca_cq.c
@@ -283,6 +283,7 @@ struct ib_cq *ehca_create_cq(struct ib_device *device, int cqe, int comp_vector,
(my_cq->galpas.user.fw_handle & (PAGE_SIZE - 1));
if (ib_copy_to_udata(udata, &resp, sizeof(resp))) {
ehca_err(device, "Copy to udata failed.");
+ cq = ERR_PTR(-EFAULT);
goto create_cq_exit4;
}
}
diff --git a/drivers/infiniband/hw/ehca/ehca_mrmw.c b/drivers/infiniband/hw/ehca/ehca_mrmw.c
index bcfb0c183620..3488e8c9fcb4 100644
--- a/drivers/infiniband/hw/ehca/ehca_mrmw.c
+++ b/drivers/infiniband/hw/ehca/ehca_mrmw.c
@@ -400,10 +400,7 @@ reg_user_mr_fallback:
pginfo.num_hwpages = num_hwpages;
pginfo.u.usr.region = e_mr->umem;
pginfo.next_hwpage = e_mr->umem->offset / hwpage_size;
- pginfo.u.usr.next_chunk = list_prepare_entry(pginfo.u.usr.next_chunk,
- (&e_mr->umem->chunk_list),
- list);
-
+ pginfo.u.usr.next_sg = pginfo.u.usr.region->sg_head.sgl;
ret = ehca_reg_mr(shca, e_mr, (u64 *)virt, length, mr_access_flags,
e_pd, &pginfo, &e_mr->ib.ib_mr.lkey,
&e_mr->ib.ib_mr.rkey, EHCA_REG_MR);
@@ -1858,61 +1855,39 @@ static int ehca_set_pagebuf_user1(struct ehca_mr_pginfo *pginfo,
u64 *kpage)
{
int ret = 0;
- struct ib_umem_chunk *prev_chunk;
- struct ib_umem_chunk *chunk;
u64 pgaddr;
- u32 i = 0;
u32 j = 0;
int hwpages_per_kpage = PAGE_SIZE / pginfo->hwpage_size;
-
- /* loop over desired chunk entries */
- chunk = pginfo->u.usr.next_chunk;
- prev_chunk = pginfo->u.usr.next_chunk;
- list_for_each_entry_continue(
- chunk, (&(pginfo->u.usr.region->chunk_list)), list) {
- for (i = pginfo->u.usr.next_nmap; i < chunk->nmap; ) {
- pgaddr = page_to_pfn(sg_page(&chunk->page_list[i]))
- << PAGE_SHIFT ;
- *kpage = pgaddr + (pginfo->next_hwpage *
- pginfo->hwpage_size);
- if ( !(*kpage) ) {
- ehca_gen_err("pgaddr=%llx "
- "chunk->page_list[i]=%llx "
- "i=%x next_hwpage=%llx",
- pgaddr, (u64)sg_dma_address(
- &chunk->page_list[i]),
- i, pginfo->next_hwpage);
- return -EFAULT;
- }
- (pginfo->hwpage_cnt)++;
- (pginfo->next_hwpage)++;
- kpage++;
- if (pginfo->next_hwpage % hwpages_per_kpage == 0) {
- (pginfo->kpage_cnt)++;
- (pginfo->u.usr.next_nmap)++;
- pginfo->next_hwpage = 0;
- i++;
- }
- j++;
- if (j >= number) break;
+ struct scatterlist **sg = &pginfo->u.usr.next_sg;
+
+ while (*sg != NULL) {
+ pgaddr = page_to_pfn(sg_page(*sg))
+ << PAGE_SHIFT;
+ *kpage = pgaddr + (pginfo->next_hwpage *
+ pginfo->hwpage_size);
+ if (!(*kpage)) {
+ ehca_gen_err("pgaddr=%llx "
+ "sg_dma_address=%llx "
+ "entry=%llx next_hwpage=%llx",
+ pgaddr, (u64)sg_dma_address(*sg),
+ pginfo->u.usr.next_nmap,
+ pginfo->next_hwpage);
+ return -EFAULT;
}
- if ((pginfo->u.usr.next_nmap >= chunk->nmap) &&
- (j >= number)) {
- pginfo->u.usr.next_nmap = 0;
- prev_chunk = chunk;
- break;
- } else if (pginfo->u.usr.next_nmap >= chunk->nmap) {
- pginfo->u.usr.next_nmap = 0;
- prev_chunk = chunk;
- } else if (j >= number)
+ (pginfo->hwpage_cnt)++;
+ (pginfo->next_hwpage)++;
+ kpage++;
+ if (pginfo->next_hwpage % hwpages_per_kpage == 0) {
+ (pginfo->kpage_cnt)++;
+ (pginfo->u.usr.next_nmap)++;
+ pginfo->next_hwpage = 0;
+ *sg = sg_next(*sg);
+ }
+ j++;
+ if (j >= number)
break;
- else
- prev_chunk = chunk;
}
- pginfo->u.usr.next_chunk =
- list_prepare_entry(prev_chunk,
- (&(pginfo->u.usr.region->chunk_list)),
- list);
+
return ret;
}
@@ -1920,20 +1895,19 @@ static int ehca_set_pagebuf_user1(struct ehca_mr_pginfo *pginfo,
* check given pages for contiguous layout
* last page addr is returned in prev_pgaddr for further check
*/
-static int ehca_check_kpages_per_ate(struct scatterlist *page_list,
- int start_idx, int end_idx,
+static int ehca_check_kpages_per_ate(struct scatterlist **sg,
+ int num_pages,
u64 *prev_pgaddr)
{
- int t;
- for (t = start_idx; t <= end_idx; t++) {
- u64 pgaddr = page_to_pfn(sg_page(&page_list[t])) << PAGE_SHIFT;
+ for (; *sg && num_pages > 0; *sg = sg_next(*sg), num_pages--) {
+ u64 pgaddr = page_to_pfn(sg_page(*sg)) << PAGE_SHIFT;
if (ehca_debug_level >= 3)
ehca_gen_dbg("chunk_page=%llx value=%016llx", pgaddr,
*(u64 *)__va(pgaddr));
if (pgaddr - PAGE_SIZE != *prev_pgaddr) {
ehca_gen_err("uncontiguous page found pgaddr=%llx "
- "prev_pgaddr=%llx page_list_i=%x",
- pgaddr, *prev_pgaddr, t);
+ "prev_pgaddr=%llx entries_left_in_hwpage=%x",
+ pgaddr, *prev_pgaddr, num_pages);
return -EINVAL;
}
*prev_pgaddr = pgaddr;
@@ -1947,111 +1921,80 @@ static int ehca_set_pagebuf_user2(struct ehca_mr_pginfo *pginfo,
u64 *kpage)
{
int ret = 0;
- struct ib_umem_chunk *prev_chunk;
- struct ib_umem_chunk *chunk;
u64 pgaddr, prev_pgaddr;
- u32 i = 0;
u32 j = 0;
int kpages_per_hwpage = pginfo->hwpage_size / PAGE_SIZE;
int nr_kpages = kpages_per_hwpage;
+ struct scatterlist **sg = &pginfo->u.usr.next_sg;
+
+ while (*sg != NULL) {
- /* loop over desired chunk entries */
- chunk = pginfo->u.usr.next_chunk;
- prev_chunk = pginfo->u.usr.next_chunk;
- list_for_each_entry_continue(
- chunk, (&(pginfo->u.usr.region->chunk_list)), list) {
- for (i = pginfo->u.usr.next_nmap; i < chunk->nmap; ) {
- if (nr_kpages == kpages_per_hwpage) {
- pgaddr = ( page_to_pfn(sg_page(&chunk->page_list[i]))
- << PAGE_SHIFT );
- *kpage = pgaddr;
- if ( !(*kpage) ) {
- ehca_gen_err("pgaddr=%llx i=%x",
- pgaddr, i);
+ if (nr_kpages == kpages_per_hwpage) {
+ pgaddr = (page_to_pfn(sg_page(*sg))
+ << PAGE_SHIFT);
+ *kpage = pgaddr;
+ if (!(*kpage)) {
+ ehca_gen_err("pgaddr=%llx entry=%llx",
+ pgaddr, pginfo->u.usr.next_nmap);
+ ret = -EFAULT;
+ return ret;
+ }
+ /*
+ * The first page in a hwpage must be aligned;
+ * the first MR page is exempt from this rule.
+ */
+ if (pgaddr & (pginfo->hwpage_size - 1)) {
+ if (pginfo->hwpage_cnt) {
+ ehca_gen_err(
+ "invalid alignment "
+ "pgaddr=%llx entry=%llx "
+ "mr_pgsize=%llx",
+ pgaddr, pginfo->u.usr.next_nmap,
+ pginfo->hwpage_size);
ret = -EFAULT;
return ret;
}
- /*
- * The first page in a hwpage must be aligned;
- * the first MR page is exempt from this rule.
- */
- if (pgaddr & (pginfo->hwpage_size - 1)) {
- if (pginfo->hwpage_cnt) {
- ehca_gen_err(
- "invalid alignment "
- "pgaddr=%llx i=%x "
- "mr_pgsize=%llx",
- pgaddr, i,
- pginfo->hwpage_size);
- ret = -EFAULT;
- return ret;
- }
- /* first MR page */
- pginfo->kpage_cnt =
- (pgaddr &
- (pginfo->hwpage_size - 1)) >>
- PAGE_SHIFT;
- nr_kpages -= pginfo->kpage_cnt;
- *kpage = pgaddr &
- ~(pginfo->hwpage_size - 1);
- }
- if (ehca_debug_level >= 3) {
- u64 val = *(u64 *)__va(pgaddr);
- ehca_gen_dbg("kpage=%llx chunk_page=%llx "
- "value=%016llx",
- *kpage, pgaddr, val);
- }
- prev_pgaddr = pgaddr;
- i++;
- pginfo->kpage_cnt++;
- pginfo->u.usr.next_nmap++;
- nr_kpages--;
- if (!nr_kpages)
- goto next_kpage;
- continue;
+ /* first MR page */
+ pginfo->kpage_cnt =
+ (pgaddr &
+ (pginfo->hwpage_size - 1)) >>
+ PAGE_SHIFT;
+ nr_kpages -= pginfo->kpage_cnt;
+ *kpage = pgaddr &
+ ~(pginfo->hwpage_size - 1);
}
- if (i + nr_kpages > chunk->nmap) {
- ret = ehca_check_kpages_per_ate(
- chunk->page_list, i,
- chunk->nmap - 1, &prev_pgaddr);
- if (ret) return ret;
- pginfo->kpage_cnt += chunk->nmap - i;
- pginfo->u.usr.next_nmap += chunk->nmap - i;
- nr_kpages -= chunk->nmap - i;
- break;
+ if (ehca_debug_level >= 3) {
+ u64 val = *(u64 *)__va(pgaddr);
+ ehca_gen_dbg("kpage=%llx page=%llx "
+ "value=%016llx",
+ *kpage, pgaddr, val);
}
+ prev_pgaddr = pgaddr;
+ *sg = sg_next(*sg);
+ pginfo->kpage_cnt++;
+ pginfo->u.usr.next_nmap++;
+ nr_kpages--;
+ if (!nr_kpages)
+ goto next_kpage;
+ continue;
+ }
+
+ ret = ehca_check_kpages_per_ate(sg, nr_kpages,
+ &prev_pgaddr);
+ if (ret)
+ return ret;
+ pginfo->kpage_cnt += nr_kpages;
+ pginfo->u.usr.next_nmap += nr_kpages;
- ret = ehca_check_kpages_per_ate(chunk->page_list, i,
- i + nr_kpages - 1,
- &prev_pgaddr);
- if (ret) return ret;
- i += nr_kpages;
- pginfo->kpage_cnt += nr_kpages;
- pginfo->u.usr.next_nmap += nr_kpages;
next_kpage:
- nr_kpages = kpages_per_hwpage;
- (pginfo->hwpage_cnt)++;
- kpage++;
- j++;
- if (j >= number) break;
- }
- if ((pginfo->u.usr.next_nmap >= chunk->nmap) &&
- (j >= number)) {
- pginfo->u.usr.next_nmap = 0;
- prev_chunk = chunk;
- break;
- } else if (pginfo->u.usr.next_nmap >= chunk->nmap) {
- pginfo->u.usr.next_nmap = 0;
- prev_chunk = chunk;
- } else if (j >= number)
+ nr_kpages = kpages_per_hwpage;
+ (pginfo->hwpage_cnt)++;
+ kpage++;
+ j++;
+ if (j >= number)
break;
- else
- prev_chunk = chunk;
}
- pginfo->u.usr.next_chunk =
- list_prepare_entry(prev_chunk,
- (&(pginfo->u.usr.region->chunk_list)),
- list);
+
return ret;
}
@@ -2591,16 +2534,6 @@ static void ehca_dma_unmap_sg(struct ib_device *dev, struct scatterlist *sg,
/* This is only a stub; nothing to be done here */
}
-static u64 ehca_dma_address(struct ib_device *dev, struct scatterlist *sg)
-{
- return sg->dma_address;
-}
-
-static unsigned int ehca_dma_len(struct ib_device *dev, struct scatterlist *sg)
-{
- return sg->length;
-}
-
static void ehca_dma_sync_single_for_cpu(struct ib_device *dev, u64 addr,
size_t size,
enum dma_data_direction dir)
@@ -2653,8 +2586,6 @@ struct ib_dma_mapping_ops ehca_dma_mapping_ops = {
.unmap_page = ehca_dma_unmap_page,
.map_sg = ehca_dma_map_sg,
.unmap_sg = ehca_dma_unmap_sg,
- .dma_address = ehca_dma_address,
- .dma_len = ehca_dma_len,
.sync_single_for_cpu = ehca_dma_sync_single_for_cpu,
.sync_single_for_device = ehca_dma_sync_single_for_device,
.alloc_coherent = ehca_dma_alloc_coherent,
diff --git a/drivers/infiniband/hw/ipath/ipath_diag.c b/drivers/infiniband/hw/ipath/ipath_diag.c
index 714293b78518..e2f9a51f4a38 100644
--- a/drivers/infiniband/hw/ipath/ipath_diag.c
+++ b/drivers/infiniband/hw/ipath/ipath_diag.c
@@ -326,7 +326,7 @@ static ssize_t ipath_diagpkt_write(struct file *fp,
size_t count, loff_t *off)
{
u32 __iomem *piobuf;
- u32 plen, clen, pbufn;
+ u32 plen, pbufn, maxlen_reserve;
struct ipath_diag_pkt odp;
struct ipath_diag_xpkt dp;
u32 *tmpbuf = NULL;
@@ -335,51 +335,29 @@ static ssize_t ipath_diagpkt_write(struct file *fp,
u64 val;
u32 l_state, lt_state; /* LinkState, LinkTrainingState */
- if (count < sizeof(odp)) {
- ret = -EINVAL;
- goto bail;
- }
if (count == sizeof(dp)) {
if (copy_from_user(&dp, data, sizeof(dp))) {
ret = -EFAULT;
goto bail;
}
- } else if (copy_from_user(&odp, data, sizeof(odp))) {
- ret = -EFAULT;
+ } else if (count == sizeof(odp)) {
+ if (copy_from_user(&odp, data, sizeof(odp))) {
+ ret = -EFAULT;
+ goto bail;
+ }
+ } else {
+ ret = -EINVAL;
goto bail;
}
- /*
- * Due to padding/alignment issues (lessened with new struct)
- * the old and new structs are the same length. We need to
- * disambiguate them, which we can do because odp.len has never
- * been less than the total of LRH+BTH+DETH so far, while
- * dp.unit (same offset) unit is unlikely to get that high.
- * Similarly, dp.data, the pointer to user at the same offset
- * as odp.unit, is almost certainly at least one (512byte)page
- * "above" NULL. The if-block below can be omitted if compatibility
- * between a new driver and older diagnostic code is unimportant.
- * compatibility the other direction (new diags, old driver) is
- * handled in the diagnostic code, with a warning.
- */
- if (dp.unit >= 20 && dp.data < 512) {
- /* very probable version mismatch. Fix it up */
- memcpy(&odp, &dp, sizeof(odp));
- /* We got a legacy dp, copy elements to dp */
- dp.unit = odp.unit;
- dp.data = odp.data;
- dp.len = odp.len;
- dp.pbc_wd = 0; /* Indicate we need to compute PBC wd */
- }
-
/* send count must be an exact number of dwords */
if (dp.len & 3) {
ret = -EINVAL;
goto bail;
}
- clen = dp.len >> 2;
+ plen = dp.len >> 2;
dd = ipath_lookup(dp.unit);
if (!dd || !(dd->ipath_flags & IPATH_PRESENT) ||
@@ -422,16 +400,22 @@ static ssize_t ipath_diagpkt_write(struct file *fp,
goto bail;
}
- /* need total length before first word written */
- /* +1 word is for the qword padding */
- plen = sizeof(u32) + dp.len;
-
- if ((plen + 4) > dd->ipath_ibmaxlen) {
+ /*
+ * need total length before first word written, plus 2 Dwords. One Dword
+ * is for padding so we get the full user data when not aligned on
+ * a word boundary. The other Dword is to make sure we have room for the
+ * ICRC which gets tacked on later.
+ */
+ maxlen_reserve = 2 * sizeof(u32);
+ if (dp.len > dd->ipath_ibmaxlen - maxlen_reserve) {
ipath_dbg("Pkt len 0x%x > ibmaxlen %x\n",
- plen - 4, dd->ipath_ibmaxlen);
+ dp.len, dd->ipath_ibmaxlen);
ret = -EINVAL;
- goto bail; /* before writing pbc */
+ goto bail;
}
+
+ plen = sizeof(u32) + dp.len;
+
tmpbuf = vmalloc(plen);
if (!tmpbuf) {
dev_info(&dd->pcidev->dev, "Unable to allocate tmp buffer, "
@@ -473,11 +457,11 @@ static ssize_t ipath_diagpkt_write(struct file *fp,
*/
if (dd->ipath_flags & IPATH_PIO_FLUSH_WC) {
ipath_flush_wc();
- __iowrite32_copy(piobuf + 2, tmpbuf, clen - 1);
+ __iowrite32_copy(piobuf + 2, tmpbuf, plen - 1);
ipath_flush_wc();
- __raw_writel(tmpbuf[clen - 1], piobuf + clen + 1);
+ __raw_writel(tmpbuf[plen - 1], piobuf + plen + 1);
} else
- __iowrite32_copy(piobuf + 2, tmpbuf, clen);
+ __iowrite32_copy(piobuf + 2, tmpbuf, plen);
ipath_flush_wc();
diff --git a/drivers/infiniband/hw/ipath/ipath_dma.c b/drivers/infiniband/hw/ipath/ipath_dma.c
index 644c2c74e054..123a8c053539 100644
--- a/drivers/infiniband/hw/ipath/ipath_dma.c
+++ b/drivers/infiniband/hw/ipath/ipath_dma.c
@@ -115,6 +115,10 @@ static int ipath_map_sg(struct ib_device *dev, struct scatterlist *sgl,
ret = 0;
break;
}
+ sg->dma_address = addr + sg->offset;
+#ifdef CONFIG_NEED_SG_DMA_LENGTH
+ sg->dma_length = sg->length;
+#endif
}
return ret;
}
@@ -126,21 +130,6 @@ static void ipath_unmap_sg(struct ib_device *dev,
BUG_ON(!valid_dma_direction(direction));
}
-static u64 ipath_sg_dma_address(struct ib_device *dev, struct scatterlist *sg)
-{
- u64 addr = (u64) page_address(sg_page(sg));
-
- if (addr)
- addr += sg->offset;
- return addr;
-}
-
-static unsigned int ipath_sg_dma_len(struct ib_device *dev,
- struct scatterlist *sg)
-{
- return sg->length;
-}
-
static void ipath_sync_single_for_cpu(struct ib_device *dev,
u64 addr,
size_t size,
@@ -176,17 +165,15 @@ static void ipath_dma_free_coherent(struct ib_device *dev, size_t size,
}
struct ib_dma_mapping_ops ipath_dma_mapping_ops = {
- ipath_mapping_error,
- ipath_dma_map_single,
- ipath_dma_unmap_single,
- ipath_dma_map_page,
- ipath_dma_unmap_page,
- ipath_map_sg,
- ipath_unmap_sg,
- ipath_sg_dma_address,
- ipath_sg_dma_len,
- ipath_sync_single_for_cpu,
- ipath_sync_single_for_device,
- ipath_dma_alloc_coherent,
- ipath_dma_free_coherent
+ .mapping_error = ipath_mapping_error,
+ .map_single = ipath_dma_map_single,
+ .unmap_single = ipath_dma_unmap_single,
+ .map_page = ipath_dma_map_page,
+ .unmap_page = ipath_dma_unmap_page,
+ .map_sg = ipath_map_sg,
+ .unmap_sg = ipath_unmap_sg,
+ .sync_single_for_cpu = ipath_sync_single_for_cpu,
+ .sync_single_for_device = ipath_sync_single_for_device,
+ .alloc_coherent = ipath_dma_alloc_coherent,
+ .free_coherent = ipath_dma_free_coherent
};
diff --git a/drivers/infiniband/hw/ipath/ipath_mr.c b/drivers/infiniband/hw/ipath/ipath_mr.c
index e346d3890a0e..5e61e9bff697 100644
--- a/drivers/infiniband/hw/ipath/ipath_mr.c
+++ b/drivers/infiniband/hw/ipath/ipath_mr.c
@@ -188,8 +188,8 @@ struct ib_mr *ipath_reg_user_mr(struct ib_pd *pd, u64 start, u64 length,
{
struct ipath_mr *mr;
struct ib_umem *umem;
- struct ib_umem_chunk *chunk;
- int n, m, i;
+ int n, m, entry;
+ struct scatterlist *sg;
struct ib_mr *ret;
if (length == 0) {
@@ -202,10 +202,7 @@ struct ib_mr *ipath_reg_user_mr(struct ib_pd *pd, u64 start, u64 length,
if (IS_ERR(umem))
return (void *) umem;
- n = 0;
- list_for_each_entry(chunk, &umem->chunk_list, list)
- n += chunk->nents;
-
+ n = umem->nmap;
mr = alloc_mr(n, &to_idev(pd->device)->lk_table);
if (!mr) {
ret = ERR_PTR(-ENOMEM);
@@ -224,22 +221,20 @@ struct ib_mr *ipath_reg_user_mr(struct ib_pd *pd, u64 start, u64 length,
m = 0;
n = 0;
- list_for_each_entry(chunk, &umem->chunk_list, list) {
- for (i = 0; i < chunk->nents; i++) {
- void *vaddr;
-
- vaddr = page_address(sg_page(&chunk->page_list[i]));
- if (!vaddr) {
- ret = ERR_PTR(-EINVAL);
- goto bail;
- }
- mr->mr.map[m]->segs[n].vaddr = vaddr;
- mr->mr.map[m]->segs[n].length = umem->page_size;
- n++;
- if (n == IPATH_SEGSZ) {
- m++;
- n = 0;
- }
+ for_each_sg(umem->sg_head.sgl, sg, umem->nmap, entry) {
+ void *vaddr;
+
+ vaddr = page_address(sg_page(sg));
+ if (!vaddr) {
+ ret = ERR_PTR(-EINVAL);
+ goto bail;
+ }
+ mr->mr.map[m]->segs[n].vaddr = vaddr;
+ mr->mr.map[m]->segs[n].length = umem->page_size;
+ n++;
+ if (n == IPATH_SEGSZ) {
+ m++;
+ n = 0;
}
}
ret = &mr->ibmr;
diff --git a/drivers/infiniband/hw/mlx4/alias_GUID.c b/drivers/infiniband/hw/mlx4/alias_GUID.c
index 2f215b93db6b..0eb141c41416 100644
--- a/drivers/infiniband/hw/mlx4/alias_GUID.c
+++ b/drivers/infiniband/hw/mlx4/alias_GUID.c
@@ -154,7 +154,7 @@ void mlx4_ib_notify_slaves_on_guid_change(struct mlx4_ib_dev *dev,
continue;
slave_id = (block_num * NUM_ALIAS_GUID_IN_REC) + i ;
- if (slave_id >= dev->dev->num_slaves)
+ if (slave_id >= dev->dev->num_vfs + 1)
return;
tmp_cur_ag = *(__be64 *)&p_data[i * GUID_REC_SIZE];
form_cache_ag = get_cached_alias_guid(dev, port_num,
diff --git a/drivers/infiniband/hw/mlx4/cm.c b/drivers/infiniband/hw/mlx4/cm.c
index d1f5f1dd77b0..56a593e0ae5d 100644
--- a/drivers/infiniband/hw/mlx4/cm.c
+++ b/drivers/infiniband/hw/mlx4/cm.c
@@ -61,6 +61,11 @@ struct cm_generic_msg {
__be32 remote_comm_id;
};
+struct cm_sidr_generic_msg {
+ struct ib_mad_hdr hdr;
+ __be32 request_id;
+};
+
struct cm_req_msg {
unsigned char unused[0x60];
union ib_gid primary_path_sgid;
@@ -69,28 +74,62 @@ struct cm_req_msg {
static void set_local_comm_id(struct ib_mad *mad, u32 cm_id)
{
- struct cm_generic_msg *msg = (struct cm_generic_msg *)mad;
- msg->local_comm_id = cpu_to_be32(cm_id);
+ if (mad->mad_hdr.attr_id == CM_SIDR_REQ_ATTR_ID) {
+ struct cm_sidr_generic_msg *msg =
+ (struct cm_sidr_generic_msg *)mad;
+ msg->request_id = cpu_to_be32(cm_id);
+ } else if (mad->mad_hdr.attr_id == CM_SIDR_REP_ATTR_ID) {
+ pr_err("trying to set local_comm_id in SIDR_REP\n");
+ return;
+ } else {
+ struct cm_generic_msg *msg = (struct cm_generic_msg *)mad;
+ msg->local_comm_id = cpu_to_be32(cm_id);
+ }
}
static u32 get_local_comm_id(struct ib_mad *mad)
{
- struct cm_generic_msg *msg = (struct cm_generic_msg *)mad;
-
- return be32_to_cpu(msg->local_comm_id);
+ if (mad->mad_hdr.attr_id == CM_SIDR_REQ_ATTR_ID) {
+ struct cm_sidr_generic_msg *msg =
+ (struct cm_sidr_generic_msg *)mad;
+ return be32_to_cpu(msg->request_id);
+ } else if (mad->mad_hdr.attr_id == CM_SIDR_REP_ATTR_ID) {
+ pr_err("trying to set local_comm_id in SIDR_REP\n");
+ return -1;
+ } else {
+ struct cm_generic_msg *msg = (struct cm_generic_msg *)mad;
+ return be32_to_cpu(msg->local_comm_id);
+ }
}
static void set_remote_comm_id(struct ib_mad *mad, u32 cm_id)
{
- struct cm_generic_msg *msg = (struct cm_generic_msg *)mad;
- msg->remote_comm_id = cpu_to_be32(cm_id);
+ if (mad->mad_hdr.attr_id == CM_SIDR_REP_ATTR_ID) {
+ struct cm_sidr_generic_msg *msg =
+ (struct cm_sidr_generic_msg *)mad;
+ msg->request_id = cpu_to_be32(cm_id);
+ } else if (mad->mad_hdr.attr_id == CM_SIDR_REQ_ATTR_ID) {
+ pr_err("trying to set remote_comm_id in SIDR_REQ\n");
+ return;
+ } else {
+ struct cm_generic_msg *msg = (struct cm_generic_msg *)mad;
+ msg->remote_comm_id = cpu_to_be32(cm_id);
+ }
}
static u32 get_remote_comm_id(struct ib_mad *mad)
{
- struct cm_generic_msg *msg = (struct cm_generic_msg *)mad;
-
- return be32_to_cpu(msg->remote_comm_id);
+ if (mad->mad_hdr.attr_id == CM_SIDR_REP_ATTR_ID) {
+ struct cm_sidr_generic_msg *msg =
+ (struct cm_sidr_generic_msg *)mad;
+ return be32_to_cpu(msg->request_id);
+ } else if (mad->mad_hdr.attr_id == CM_SIDR_REQ_ATTR_ID) {
+ pr_err("trying to set remote_comm_id in SIDR_REQ\n");
+ return -1;
+ } else {
+ struct cm_generic_msg *msg = (struct cm_generic_msg *)mad;
+ return be32_to_cpu(msg->remote_comm_id);
+ }
}
static union ib_gid gid_from_req_msg(struct ib_device *ibdev, struct ib_mad *mad)
@@ -282,19 +321,21 @@ int mlx4_ib_multiplex_cm_handler(struct ib_device *ibdev, int port, int slave_id
u32 sl_cm_id;
int pv_cm_id = -1;
- sl_cm_id = get_local_comm_id(mad);
-
if (mad->mad_hdr.attr_id == CM_REQ_ATTR_ID ||
- mad->mad_hdr.attr_id == CM_REP_ATTR_ID) {
+ mad->mad_hdr.attr_id == CM_REP_ATTR_ID ||
+ mad->mad_hdr.attr_id == CM_SIDR_REQ_ATTR_ID) {
+ sl_cm_id = get_local_comm_id(mad);
id = id_map_alloc(ibdev, slave_id, sl_cm_id);
if (IS_ERR(id)) {
mlx4_ib_warn(ibdev, "%s: id{slave: %d, sl_cm_id: 0x%x} Failed to id_map_alloc\n",
__func__, slave_id, sl_cm_id);
return PTR_ERR(id);
}
- } else if (mad->mad_hdr.attr_id == CM_REJ_ATTR_ID) {
+ } else if (mad->mad_hdr.attr_id == CM_REJ_ATTR_ID ||
+ mad->mad_hdr.attr_id == CM_SIDR_REP_ATTR_ID) {
return 0;
} else {
+ sl_cm_id = get_local_comm_id(mad);
id = id_map_get(ibdev, &pv_cm_id, slave_id, sl_cm_id);
}
@@ -315,14 +356,18 @@ int mlx4_ib_multiplex_cm_handler(struct ib_device *ibdev, int port, int slave_id
}
int mlx4_ib_demux_cm_handler(struct ib_device *ibdev, int port, int *slave,
- struct ib_mad *mad)
+ struct ib_mad *mad)
{
u32 pv_cm_id;
struct id_map_entry *id;
- if (mad->mad_hdr.attr_id == CM_REQ_ATTR_ID) {
+ if (mad->mad_hdr.attr_id == CM_REQ_ATTR_ID ||
+ mad->mad_hdr.attr_id == CM_SIDR_REQ_ATTR_ID) {
union ib_gid gid;
+ if (!slave)
+ return 0;
+
gid = gid_from_req_msg(ibdev, mad);
*slave = mlx4_ib_find_real_gid(ibdev, port, gid.global.interface_id);
if (*slave < 0) {
@@ -341,7 +386,8 @@ int mlx4_ib_demux_cm_handler(struct ib_device *ibdev, int port, int *slave,
return -ENOENT;
}
- *slave = id->slave_id;
+ if (slave)
+ *slave = id->slave_id;
set_remote_comm_id(mad, id->sl_cm_id);
if (mad->mad_hdr.attr_id == CM_DREQ_ATTR_ID)
diff --git a/drivers/infiniband/hw/mlx4/cq.c b/drivers/infiniband/hw/mlx4/cq.c
index cc40f08ca8f1..5f640814cc81 100644
--- a/drivers/infiniband/hw/mlx4/cq.c
+++ b/drivers/infiniband/hw/mlx4/cq.c
@@ -564,7 +564,7 @@ static int mlx4_ib_ipoib_csum_ok(__be16 status, __be16 checksum)
}
static int use_tunnel_data(struct mlx4_ib_qp *qp, struct mlx4_ib_cq *cq, struct ib_wc *wc,
- unsigned tail, struct mlx4_cqe *cqe)
+ unsigned tail, struct mlx4_cqe *cqe, int is_eth)
{
struct mlx4_ib_proxy_sqp_hdr *hdr;
@@ -574,12 +574,20 @@ static int use_tunnel_data(struct mlx4_ib_qp *qp, struct mlx4_ib_cq *cq, struct
DMA_FROM_DEVICE);
hdr = (struct mlx4_ib_proxy_sqp_hdr *) (qp->sqp_proxy_rcv[tail].addr);
wc->pkey_index = be16_to_cpu(hdr->tun.pkey_index);
- wc->slid = be16_to_cpu(hdr->tun.slid_mac_47_32);
- wc->sl = (u8) (be16_to_cpu(hdr->tun.sl_vid) >> 12);
wc->src_qp = be32_to_cpu(hdr->tun.flags_src_qp) & 0xFFFFFF;
wc->wc_flags |= (hdr->tun.g_ml_path & 0x80) ? (IB_WC_GRH) : 0;
wc->dlid_path_bits = 0;
+ if (is_eth) {
+ wc->vlan_id = be16_to_cpu(hdr->tun.sl_vid);
+ memcpy(&(wc->smac[0]), (char *)&hdr->tun.mac_31_0, 4);
+ memcpy(&(wc->smac[4]), (char *)&hdr->tun.slid_mac_47_32, 2);
+ wc->wc_flags |= (IB_WC_WITH_VLAN | IB_WC_WITH_SMAC);
+ } else {
+ wc->slid = be16_to_cpu(hdr->tun.slid_mac_47_32);
+ wc->sl = (u8) (be16_to_cpu(hdr->tun.sl_vid) >> 12);
+ }
+
return 0;
}
@@ -594,6 +602,7 @@ static int mlx4_ib_poll_one(struct mlx4_ib_cq *cq,
struct mlx4_srq *msrq = NULL;
int is_send;
int is_error;
+ int is_eth;
u32 g_mlpath_rqpn;
u16 wqe_ctr;
unsigned tail = 0;
@@ -778,11 +787,15 @@ repoll:
break;
}
+ is_eth = (rdma_port_get_link_layer(wc->qp->device,
+ (*cur_qp)->port) ==
+ IB_LINK_LAYER_ETHERNET);
if (mlx4_is_mfunc(to_mdev(cq->ibcq.device)->dev)) {
if ((*cur_qp)->mlx4_ib_qp_type &
(MLX4_IB_QPT_PROXY_SMI_OWNER |
MLX4_IB_QPT_PROXY_SMI | MLX4_IB_QPT_PROXY_GSI))
- return use_tunnel_data(*cur_qp, cq, wc, tail, cqe);
+ return use_tunnel_data(*cur_qp, cq, wc, tail,
+ cqe, is_eth);
}
wc->slid = be16_to_cpu(cqe->rlid);
@@ -793,20 +806,21 @@ repoll:
wc->pkey_index = be32_to_cpu(cqe->immed_rss_invalid) & 0x7f;
wc->wc_flags |= mlx4_ib_ipoib_csum_ok(cqe->status,
cqe->checksum) ? IB_WC_IP_CSUM_OK : 0;
- if (rdma_port_get_link_layer(wc->qp->device,
- (*cur_qp)->port) == IB_LINK_LAYER_ETHERNET)
+ if (is_eth) {
wc->sl = be16_to_cpu(cqe->sl_vid) >> 13;
- else
- wc->sl = be16_to_cpu(cqe->sl_vid) >> 12;
- if (be32_to_cpu(cqe->vlan_my_qpn) & MLX4_CQE_VLAN_PRESENT_MASK) {
- wc->vlan_id = be16_to_cpu(cqe->sl_vid) &
- MLX4_CQE_VID_MASK;
+ if (be32_to_cpu(cqe->vlan_my_qpn) &
+ MLX4_CQE_VLAN_PRESENT_MASK) {
+ wc->vlan_id = be16_to_cpu(cqe->sl_vid) &
+ MLX4_CQE_VID_MASK;
+ } else {
+ wc->vlan_id = 0xffff;
+ }
+ memcpy(wc->smac, cqe->smac, ETH_ALEN);
+ wc->wc_flags |= (IB_WC_WITH_VLAN | IB_WC_WITH_SMAC);
} else {
+ wc->sl = be16_to_cpu(cqe->sl_vid) >> 12;
wc->vlan_id = 0xffff;
}
- wc->wc_flags |= IB_WC_WITH_VLAN;
- memcpy(wc->smac, cqe->smac, ETH_ALEN);
- wc->wc_flags |= IB_WC_WITH_SMAC;
}
return 0;
diff --git a/drivers/infiniband/hw/mlx4/doorbell.c b/drivers/infiniband/hw/mlx4/doorbell.c
index 8aee4233b388..c51740986367 100644
--- a/drivers/infiniband/hw/mlx4/doorbell.c
+++ b/drivers/infiniband/hw/mlx4/doorbell.c
@@ -45,7 +45,6 @@ int mlx4_ib_db_map_user(struct mlx4_ib_ucontext *context, unsigned long virt,
struct mlx4_db *db)
{
struct mlx4_ib_user_db_page *page;
- struct ib_umem_chunk *chunk;
int err = 0;
mutex_lock(&context->db_page_mutex);
@@ -73,8 +72,7 @@ int mlx4_ib_db_map_user(struct mlx4_ib_ucontext *context, unsigned long virt,
list_add(&page->list, &context->db_page_list);
found:
- chunk = list_entry(page->umem->chunk_list.next, struct ib_umem_chunk, list);
- db->dma = sg_dma_address(chunk->page_list) + (virt & ~PAGE_MASK);
+ db->dma = sg_dma_address(page->umem->sg_head.sgl) + (virt & ~PAGE_MASK);
db->u.user_page = page;
++page->refcnt;
diff --git a/drivers/infiniband/hw/mlx4/mad.c b/drivers/infiniband/hw/mlx4/mad.c
index f2a3f48107e7..fd36ec672632 100644
--- a/drivers/infiniband/hw/mlx4/mad.c
+++ b/drivers/infiniband/hw/mlx4/mad.c
@@ -467,6 +467,7 @@ int mlx4_ib_send_to_slave(struct mlx4_ib_dev *dev, int slave, u8 port,
int ret = 0;
u16 tun_pkey_ix;
u16 cached_pkey;
+ u8 is_eth = dev->dev->caps.port_type[port] == MLX4_PORT_TYPE_ETH;
if (dest_qpt > IB_QPT_GSI)
return -EINVAL;
@@ -509,6 +510,10 @@ int mlx4_ib_send_to_slave(struct mlx4_ib_dev *dev, int slave, u8 port,
* The driver will set the force loopback bit in post_send */
memset(&attr, 0, sizeof attr);
attr.port_num = port;
+ if (is_eth) {
+ memcpy(&attr.grh.dgid.raw[0], &grh->dgid.raw[0], 16);
+ attr.ah_flags = IB_AH_GRH;
+ }
ah = ib_create_ah(tun_ctx->pd, &attr);
if (IS_ERR(ah))
return -ENOMEM;
@@ -540,11 +545,36 @@ int mlx4_ib_send_to_slave(struct mlx4_ib_dev *dev, int slave, u8 port,
/* adjust tunnel data */
tun_mad->hdr.pkey_index = cpu_to_be16(tun_pkey_ix);
- tun_mad->hdr.sl_vid = cpu_to_be16(((u16)(wc->sl)) << 12);
- tun_mad->hdr.slid_mac_47_32 = cpu_to_be16(wc->slid);
tun_mad->hdr.flags_src_qp = cpu_to_be32(wc->src_qp & 0xFFFFFF);
tun_mad->hdr.g_ml_path = (grh && (wc->wc_flags & IB_WC_GRH)) ? 0x80 : 0;
+ if (is_eth) {
+ u16 vlan = 0;
+ if (mlx4_get_slave_default_vlan(dev->dev, port, slave, &vlan,
+ NULL)) {
+ /* VST mode */
+ if (vlan != wc->vlan_id)
+ /* Packet vlan is not the VST-assigned vlan.
+ * Drop the packet.
+ */
+ goto out;
+ else
+ /* Remove the vlan tag before forwarding
+ * the packet to the VF.
+ */
+ vlan = 0xffff;
+ } else {
+ vlan = wc->vlan_id;
+ }
+
+ tun_mad->hdr.sl_vid = cpu_to_be16(vlan);
+ memcpy((char *)&tun_mad->hdr.mac_31_0, &(wc->smac[0]), 4);
+ memcpy((char *)&tun_mad->hdr.slid_mac_47_32, &(wc->smac[4]), 2);
+ } else {
+ tun_mad->hdr.sl_vid = cpu_to_be16(((u16)(wc->sl)) << 12);
+ tun_mad->hdr.slid_mac_47_32 = cpu_to_be16(wc->slid);
+ }
+
ib_dma_sync_single_for_device(&dev->ib_dev,
tun_qp->tx_ring[tun_tx_ix].buf.map,
sizeof (struct mlx4_rcv_tunnel_mad),
@@ -580,6 +610,41 @@ static int mlx4_ib_demux_mad(struct ib_device *ibdev, u8 port,
int err;
int slave;
u8 *slave_id;
+ int is_eth = 0;
+
+ if (rdma_port_get_link_layer(ibdev, port) == IB_LINK_LAYER_INFINIBAND)
+ is_eth = 0;
+ else
+ is_eth = 1;
+
+ if (is_eth) {
+ if (!(wc->wc_flags & IB_WC_GRH)) {
+ mlx4_ib_warn(ibdev, "RoCE grh not present.\n");
+ return -EINVAL;
+ }
+ if (mad->mad_hdr.mgmt_class != IB_MGMT_CLASS_CM) {
+ mlx4_ib_warn(ibdev, "RoCE mgmt class is not CM\n");
+ return -EINVAL;
+ }
+ if (mlx4_get_slave_from_roce_gid(dev->dev, port, grh->dgid.raw, &slave)) {
+ mlx4_ib_warn(ibdev, "failed matching grh\n");
+ return -ENOENT;
+ }
+ if (slave >= dev->dev->caps.sqp_demux) {
+ mlx4_ib_warn(ibdev, "slave id: %d is bigger than allowed:%d\n",
+ slave, dev->dev->caps.sqp_demux);
+ return -ENOENT;
+ }
+
+ if (mlx4_ib_demux_cm_handler(ibdev, port, NULL, mad))
+ return 0;
+
+ err = mlx4_ib_send_to_slave(dev, slave, port, wc->qp->qp_type, wc, grh, mad);
+ if (err)
+ pr_debug("failed sending to slave %d via tunnel qp (%d)\n",
+ slave, err);
+ return 0;
+ }
/* Initially assume that this mad is for us */
slave = mlx4_master_func_num(dev->dev);
@@ -1076,8 +1141,9 @@ static int is_proxy_qp0(struct mlx4_ib_dev *dev, int qpn, int slave)
int mlx4_ib_send_to_wire(struct mlx4_ib_dev *dev, int slave, u8 port,
- enum ib_qp_type dest_qpt, u16 pkey_index, u32 remote_qpn,
- u32 qkey, struct ib_ah_attr *attr, struct ib_mad *mad)
+ enum ib_qp_type dest_qpt, u16 pkey_index,
+ u32 remote_qpn, u32 qkey, struct ib_ah_attr *attr,
+ u8 *s_mac, struct ib_mad *mad)
{
struct ib_sge list;
struct ib_send_wr wr, *bad_wr;
@@ -1166,6 +1232,9 @@ int mlx4_ib_send_to_wire(struct mlx4_ib_dev *dev, int slave, u8 port,
wr.num_sge = 1;
wr.opcode = IB_WR_SEND;
wr.send_flags = IB_SEND_SIGNALED;
+ if (s_mac)
+ memcpy(to_mah(ah)->av.eth.s_mac, s_mac, 6);
+
ret = ib_post_send(send_qp, &wr, &bad_wr);
out:
@@ -1174,6 +1243,22 @@ out:
return ret;
}
+static int get_slave_base_gid_ix(struct mlx4_ib_dev *dev, int slave, int port)
+{
+ if (rdma_port_get_link_layer(&dev->ib_dev, port) == IB_LINK_LAYER_INFINIBAND)
+ return slave;
+ return mlx4_get_base_gid_ix(dev->dev, slave, port);
+}
+
+static void fill_in_real_sgid_index(struct mlx4_ib_dev *dev, int slave, int port,
+ struct ib_ah_attr *ah_attr)
+{
+ if (rdma_port_get_link_layer(&dev->ib_dev, port) == IB_LINK_LAYER_INFINIBAND)
+ ah_attr->grh.sgid_index = slave;
+ else
+ ah_attr->grh.sgid_index += get_slave_base_gid_ix(dev, slave, port);
+}
+
static void mlx4_ib_multiplex_mad(struct mlx4_ib_demux_pv_ctx *ctx, struct ib_wc *wc)
{
struct mlx4_ib_dev *dev = to_mdev(ctx->ib_dev);
@@ -1184,6 +1269,7 @@ static void mlx4_ib_multiplex_mad(struct mlx4_ib_demux_pv_ctx *ctx, struct ib_wc
struct ib_ah_attr ah_attr;
u8 *slave_id;
int slave;
+ int port;
/* Get slave that sent this packet */
if (wc->src_qp < dev->dev->phys_caps.base_proxy_sqpn ||
@@ -1260,12 +1346,18 @@ static void mlx4_ib_multiplex_mad(struct mlx4_ib_demux_pv_ctx *ctx, struct ib_wc
memcpy(&ah.av, &tunnel->hdr.av, sizeof (struct mlx4_av));
ah.ibah.device = ctx->ib_dev;
mlx4_ib_query_ah(&ah.ibah, &ah_attr);
- if ((ah_attr.ah_flags & IB_AH_GRH) &&
- (ah_attr.grh.sgid_index != slave)) {
- mlx4_ib_warn(ctx->ib_dev, "slave:%d accessed invalid sgid_index:%d\n",
- slave, ah_attr.grh.sgid_index);
+ if (ah_attr.ah_flags & IB_AH_GRH)
+ fill_in_real_sgid_index(dev, slave, ctx->port, &ah_attr);
+
+ port = mlx4_slave_convert_port(dev->dev, slave, ah_attr.port_num);
+ if (port < 0)
return;
- }
+ ah_attr.port_num = port;
+ memcpy(ah_attr.dmac, tunnel->hdr.mac, 6);
+ ah_attr.vlan_id = be16_to_cpu(tunnel->hdr.vlan);
+ /* if slave have default vlan use it */
+ mlx4_get_slave_default_vlan(dev->dev, ctx->port, slave,
+ &ah_attr.vlan_id, &ah_attr.sl);
mlx4_ib_send_to_wire(dev, slave, ctx->port,
is_proxy_qp0(dev, wc->src_qp, slave) ?
@@ -1273,7 +1365,7 @@ static void mlx4_ib_multiplex_mad(struct mlx4_ib_demux_pv_ctx *ctx, struct ib_wc
be16_to_cpu(tunnel->hdr.pkey_index),
be32_to_cpu(tunnel->hdr.remote_qpn),
be32_to_cpu(tunnel->hdr.qkey),
- &ah_attr, &tunnel->mad);
+ &ah_attr, wc->smac, &tunnel->mad);
}
static int mlx4_ib_alloc_pv_bufs(struct mlx4_ib_demux_pv_ctx *ctx,
@@ -1850,7 +1942,15 @@ static int mlx4_ib_alloc_demux_ctx(struct mlx4_ib_dev *dev,
ctx->port = port;
ctx->ib_dev = &dev->ib_dev;
- for (i = 0; i < dev->dev->caps.sqp_demux; i++) {
+ for (i = 0;
+ i < min(dev->dev->caps.sqp_demux, (u16)(dev->dev->num_vfs + 1));
+ i++) {
+ struct mlx4_active_ports actv_ports =
+ mlx4_get_active_ports(dev->dev, i);
+
+ if (!test_bit(port - 1, actv_ports.ports))
+ continue;
+
ret = alloc_pv_object(dev, i, port, &ctx->tun[i]);
if (ret) {
ret = -ENOMEM;
diff --git a/drivers/infiniband/hw/mlx4/main.c b/drivers/infiniband/hw/mlx4/main.c
index e81c5547e647..1b6dbe156a37 100644
--- a/drivers/infiniband/hw/mlx4/main.c
+++ b/drivers/infiniband/hw/mlx4/main.c
@@ -53,8 +53,8 @@
#include "user.h"
#define DRV_NAME MLX4_IB_DRV_NAME
-#define DRV_VERSION "1.0"
-#define DRV_RELDATE "April 4, 2008"
+#define DRV_VERSION "2.2-1"
+#define DRV_RELDATE "Feb 2014"
#define MLX4_IB_FLOW_MAX_PRIO 0xFFF
#define MLX4_IB_FLOW_QPN_MASK 0xFFFFFF
@@ -1546,7 +1546,7 @@ static int mlx4_ib_addr_event(int event, struct net_device *event_netdev,
iboe = &ibdev->iboe;
spin_lock(&iboe->lock);
- for (port = 1; port <= MLX4_MAX_PORTS; ++port)
+ for (port = 1; port <= ibdev->dev->caps.num_ports; ++port)
if ((netif_is_bond_master(real_dev) &&
(real_dev == iboe->masters[port - 1])) ||
(!netif_is_bond_master(real_dev) &&
@@ -1569,14 +1569,14 @@ static u8 mlx4_ib_get_dev_port(struct net_device *dev,
iboe = &ibdev->iboe;
- for (port = 1; port <= MLX4_MAX_PORTS; ++port)
+ for (port = 1; port <= ibdev->dev->caps.num_ports; ++port)
if ((netif_is_bond_master(real_dev) &&
(real_dev == iboe->masters[port - 1])) ||
(!netif_is_bond_master(real_dev) &&
(real_dev == iboe->netdevs[port - 1])))
break;
- if ((port == 0) || (port > MLX4_MAX_PORTS))
+ if ((port == 0) || (port > ibdev->dev->caps.num_ports))
return 0;
else
return port;
@@ -1626,7 +1626,7 @@ static void mlx4_ib_get_dev_addr(struct net_device *dev,
union ib_gid gid;
- if ((port == 0) || (port > MLX4_MAX_PORTS))
+ if ((port == 0) || (port > ibdev->dev->caps.num_ports))
return;
/* IPv4 gids */
@@ -1803,7 +1803,7 @@ static void init_pkeys(struct mlx4_ib_dev *ibdev)
static void mlx4_ib_alloc_eqs(struct mlx4_dev *dev, struct mlx4_ib_dev *ibdev)
{
- char name[32];
+ char name[80];
int eq_per_port = 0;
int added_eqs = 0;
int total_eqs = 0;
@@ -1833,8 +1833,8 @@ static void mlx4_ib_alloc_eqs(struct mlx4_dev *dev, struct mlx4_ib_dev *ibdev)
eq = 0;
mlx4_foreach_port(i, dev, MLX4_PORT_TYPE_IB) {
for (j = 0; j < eq_per_port; j++) {
- sprintf(name, "mlx4-ib-%d-%d@%s",
- i, j, dev->pdev->bus->name);
+ snprintf(name, sizeof(name), "mlx4-ib-%d-%d@%s",
+ i, j, dev->pdev->bus->name);
/* Set IRQ for specific name (per ring) */
if (mlx4_assign_eq(dev, name, NULL,
&ibdev->eq_table[eq])) {
@@ -1888,14 +1888,6 @@ static void *mlx4_ib_add(struct mlx4_dev *dev)
pr_info_once("%s", mlx4_ib_version);
- mlx4_foreach_non_ib_transport_port(i, dev)
- num_ports++;
-
- if (mlx4_is_mfunc(dev) && num_ports) {
- dev_err(&dev->pdev->dev, "RoCE is not supported over SRIOV as yet\n");
- return NULL;
- }
-
num_ports = 0;
mlx4_foreach_ib_transport_port(i, dev)
num_ports++;
@@ -2056,8 +2048,9 @@ static void *mlx4_ib_add(struct mlx4_dev *dev)
err = mlx4_counter_alloc(ibdev->dev, &ibdev->counters[i]);
if (err)
ibdev->counters[i] = -1;
- } else
- ibdev->counters[i] = -1;
+ } else {
+ ibdev->counters[i] = -1;
+ }
}
mlx4_foreach_port(i, dev, MLX4_PORT_TYPE_IB)
@@ -2331,17 +2324,24 @@ static void do_slave_init(struct mlx4_ib_dev *ibdev, int slave, int do_init)
struct mlx4_dev *dev = ibdev->dev;
int i;
unsigned long flags;
+ struct mlx4_active_ports actv_ports;
+ unsigned int ports;
+ unsigned int first_port;
if (!mlx4_is_master(dev))
return;
- dm = kcalloc(dev->caps.num_ports, sizeof *dm, GFP_ATOMIC);
+ actv_ports = mlx4_get_active_ports(dev, slave);
+ ports = bitmap_weight(actv_ports.ports, dev->caps.num_ports);
+ first_port = find_first_bit(actv_ports.ports, dev->caps.num_ports);
+
+ dm = kcalloc(ports, sizeof(*dm), GFP_ATOMIC);
if (!dm) {
pr_err("failed to allocate memory for tunneling qp update\n");
goto out;
}
- for (i = 0; i < dev->caps.num_ports; i++) {
+ for (i = 0; i < ports; i++) {
dm[i] = kmalloc(sizeof (struct mlx4_ib_demux_work), GFP_ATOMIC);
if (!dm[i]) {
pr_err("failed to allocate memory for tunneling qp update work struct\n");
@@ -2353,9 +2353,9 @@ static void do_slave_init(struct mlx4_ib_dev *ibdev, int slave, int do_init)
}
}
/* initialize or tear down tunnel QPs for the slave */
- for (i = 0; i < dev->caps.num_ports; i++) {
+ for (i = 0; i < ports; i++) {
INIT_WORK(&dm[i]->work, mlx4_ib_tunnels_update_work);
- dm[i]->port = i + 1;
+ dm[i]->port = first_port + i + 1;
dm[i]->slave = slave;
dm[i]->do_init = do_init;
dm[i]->dev = ibdev;
diff --git a/drivers/infiniband/hw/mlx4/mcg.c b/drivers/infiniband/hw/mlx4/mcg.c
index 25b2cdff00f8..ed327e6c8fdc 100644
--- a/drivers/infiniband/hw/mlx4/mcg.c
+++ b/drivers/infiniband/hw/mlx4/mcg.c
@@ -215,8 +215,9 @@ static int send_mad_to_wire(struct mlx4_ib_demux_ctx *ctx, struct ib_mad *mad)
}
mlx4_ib_query_ah(dev->sm_ah[ctx->port - 1], &ah_attr);
spin_unlock(&dev->sm_lock);
- return mlx4_ib_send_to_wire(dev, mlx4_master_func_num(dev->dev), ctx->port,
- IB_QPT_GSI, 0, 1, IB_QP1_QKEY, &ah_attr, mad);
+ return mlx4_ib_send_to_wire(dev, mlx4_master_func_num(dev->dev),
+ ctx->port, IB_QPT_GSI, 0, 1, IB_QP1_QKEY,
+ &ah_attr, NULL, mad);
}
static int send_mad_to_slave(int slave, struct mlx4_ib_demux_ctx *ctx,
diff --git a/drivers/infiniband/hw/mlx4/mlx4_ib.h b/drivers/infiniband/hw/mlx4/mlx4_ib.h
index a230683af940..f589522fddfd 100644
--- a/drivers/infiniband/hw/mlx4/mlx4_ib.h
+++ b/drivers/infiniband/hw/mlx4/mlx4_ib.h
@@ -241,6 +241,22 @@ struct mlx4_ib_proxy_sqp_hdr {
struct mlx4_rcv_tunnel_hdr tun;
} __packed;
+struct mlx4_roce_smac_vlan_info {
+ u64 smac;
+ int smac_index;
+ int smac_port;
+ u64 candidate_smac;
+ int candidate_smac_index;
+ int candidate_smac_port;
+ u16 vid;
+ int vlan_index;
+ int vlan_port;
+ u16 candidate_vid;
+ int candidate_vlan_index;
+ int candidate_vlan_port;
+ int update_vid;
+};
+
struct mlx4_ib_qp {
struct ib_qp ibqp;
struct mlx4_qp mqp;
@@ -273,8 +289,9 @@ struct mlx4_ib_qp {
struct list_head gid_list;
struct list_head steering_rules;
struct mlx4_ib_buf *sqp_proxy_rcv;
+ struct mlx4_roce_smac_vlan_info pri;
+ struct mlx4_roce_smac_vlan_info alt;
u64 reg_id;
-
};
struct mlx4_ib_srq {
@@ -720,9 +737,12 @@ void mlx4_ib_tunnels_update_work(struct work_struct *work);
int mlx4_ib_send_to_slave(struct mlx4_ib_dev *dev, int slave, u8 port,
enum ib_qp_type qpt, struct ib_wc *wc,
struct ib_grh *grh, struct ib_mad *mad);
+
int mlx4_ib_send_to_wire(struct mlx4_ib_dev *dev, int slave, u8 port,
enum ib_qp_type dest_qpt, u16 pkey_index, u32 remote_qpn,
- u32 qkey, struct ib_ah_attr *attr, struct ib_mad *mad);
+ u32 qkey, struct ib_ah_attr *attr, u8 *s_mac,
+ struct ib_mad *mad);
+
__be64 mlx4_ib_get_new_demux_tid(struct mlx4_ib_demux_ctx *ctx);
int mlx4_ib_demux_cm_handler(struct ib_device *ibdev, int port, int *slave,
diff --git a/drivers/infiniband/hw/mlx4/mr.c b/drivers/infiniband/hw/mlx4/mr.c
index e471f089ff00..cb2a8727f3fb 100644
--- a/drivers/infiniband/hw/mlx4/mr.c
+++ b/drivers/infiniband/hw/mlx4/mr.c
@@ -90,11 +90,11 @@ int mlx4_ib_umem_write_mtt(struct mlx4_ib_dev *dev, struct mlx4_mtt *mtt,
struct ib_umem *umem)
{
u64 *pages;
- struct ib_umem_chunk *chunk;
- int i, j, k;
+ int i, k, entry;
int n;
int len;
int err = 0;
+ struct scatterlist *sg;
pages = (u64 *) __get_free_page(GFP_KERNEL);
if (!pages)
@@ -102,26 +102,25 @@ int mlx4_ib_umem_write_mtt(struct mlx4_ib_dev *dev, struct mlx4_mtt *mtt,
i = n = 0;
- list_for_each_entry(chunk, &umem->chunk_list, list)
- for (j = 0; j < chunk->nmap; ++j) {
- len = sg_dma_len(&chunk->page_list[j]) >> mtt->page_shift;
- for (k = 0; k < len; ++k) {
- pages[i++] = sg_dma_address(&chunk->page_list[j]) +
- umem->page_size * k;
- /*
- * Be friendly to mlx4_write_mtt() and
- * pass it chunks of appropriate size.
- */
- if (i == PAGE_SIZE / sizeof (u64)) {
- err = mlx4_write_mtt(dev->dev, mtt, n,
- i, pages);
- if (err)
- goto out;
- n += i;
- i = 0;
- }
+ for_each_sg(umem->sg_head.sgl, sg, umem->nmap, entry) {
+ len = sg_dma_len(sg) >> mtt->page_shift;
+ for (k = 0; k < len; ++k) {
+ pages[i++] = sg_dma_address(sg) +
+ umem->page_size * k;
+ /*
+ * Be friendly to mlx4_write_mtt() and
+ * pass it chunks of appropriate size.
+ */
+ if (i == PAGE_SIZE / sizeof (u64)) {
+ err = mlx4_write_mtt(dev->dev, mtt, n,
+ i, pages);
+ if (err)
+ goto out;
+ n += i;
+ i = 0;
}
}
+ }
if (i)
err = mlx4_write_mtt(dev->dev, mtt, n, i, pages);
diff --git a/drivers/infiniband/hw/mlx4/qp.c b/drivers/infiniband/hw/mlx4/qp.c
index d8f4d1fe8494..41308af4163c 100644
--- a/drivers/infiniband/hw/mlx4/qp.c
+++ b/drivers/infiniband/hw/mlx4/qp.c
@@ -662,10 +662,14 @@ static int create_qp_common(struct mlx4_ib_dev *dev, struct ib_pd *pd,
if (!sqp)
return -ENOMEM;
qp = &sqp->qp;
+ qp->pri.vid = 0xFFFF;
+ qp->alt.vid = 0xFFFF;
} else {
qp = kzalloc(sizeof (struct mlx4_ib_qp), GFP_KERNEL);
if (!qp)
return -ENOMEM;
+ qp->pri.vid = 0xFFFF;
+ qp->alt.vid = 0xFFFF;
}
} else
qp = *caller_qp;
@@ -940,11 +944,32 @@ static void destroy_qp_common(struct mlx4_ib_dev *dev, struct mlx4_ib_qp *qp,
{
struct mlx4_ib_cq *send_cq, *recv_cq;
- if (qp->state != IB_QPS_RESET)
+ if (qp->state != IB_QPS_RESET) {
if (mlx4_qp_modify(dev->dev, NULL, to_mlx4_state(qp->state),
MLX4_QP_STATE_RST, NULL, 0, 0, &qp->mqp))
pr_warn("modify QP %06x to RESET failed.\n",
qp->mqp.qpn);
+ if (qp->pri.smac) {
+ mlx4_unregister_mac(dev->dev, qp->pri.smac_port, qp->pri.smac);
+ qp->pri.smac = 0;
+ }
+ if (qp->alt.smac) {
+ mlx4_unregister_mac(dev->dev, qp->alt.smac_port, qp->alt.smac);
+ qp->alt.smac = 0;
+ }
+ if (qp->pri.vid < 0x1000) {
+ mlx4_unregister_vlan(dev->dev, qp->pri.vlan_port, qp->pri.vid);
+ qp->pri.vid = 0xFFFF;
+ qp->pri.candidate_vid = 0xFFFF;
+ qp->pri.update_vid = 0;
+ }
+ if (qp->alt.vid < 0x1000) {
+ mlx4_unregister_vlan(dev->dev, qp->alt.vlan_port, qp->alt.vid);
+ qp->alt.vid = 0xFFFF;
+ qp->alt.candidate_vid = 0xFFFF;
+ qp->alt.update_vid = 0;
+ }
+ }
get_cqs(qp, &send_cq, &recv_cq);
@@ -1057,6 +1082,8 @@ struct ib_qp *mlx4_ib_create_qp(struct ib_pd *pd,
qp = kzalloc(sizeof *qp, GFP_KERNEL);
if (!qp)
return ERR_PTR(-ENOMEM);
+ qp->pri.vid = 0xFFFF;
+ qp->alt.vid = 0xFFFF;
/* fall through */
case IB_QPT_UD:
{
@@ -1188,12 +1215,13 @@ static void mlx4_set_sched(struct mlx4_qp_path *path, u8 port)
static int _mlx4_set_path(struct mlx4_ib_dev *dev, const struct ib_ah_attr *ah,
u64 smac, u16 vlan_tag, struct mlx4_qp_path *path,
- u8 port)
+ struct mlx4_roce_smac_vlan_info *smac_info, u8 port)
{
int is_eth = rdma_port_get_link_layer(&dev->ib_dev, port) ==
IB_LINK_LAYER_ETHERNET;
int vidx;
int smac_index;
+ int err;
path->grh_mylmc = ah->src_path_bits & 0x7f;
@@ -1223,61 +1251,103 @@ static int _mlx4_set_path(struct mlx4_ib_dev *dev, const struct ib_ah_attr *ah,
}
if (is_eth) {
- path->sched_queue = MLX4_IB_DEFAULT_SCHED_QUEUE |
- ((port - 1) << 6) | ((ah->sl & 7) << 3);
-
if (!(ah->ah_flags & IB_AH_GRH))
return -1;
- memcpy(path->dmac, ah->dmac, ETH_ALEN);
- path->ackto = MLX4_IB_LINK_TYPE_ETH;
- /* find the index into MAC table for IBoE */
- if (!is_zero_ether_addr((const u8 *)&smac)) {
- if (mlx4_find_cached_mac(dev->dev, port, smac,
- &smac_index))
- return -ENOENT;
- } else {
- smac_index = 0;
- }
-
- path->grh_mylmc &= 0x80 | smac_index;
+ path->sched_queue = MLX4_IB_DEFAULT_SCHED_QUEUE |
+ ((port - 1) << 6) | ((ah->sl & 7) << 3);
path->feup |= MLX4_FEUP_FORCE_ETH_UP;
if (vlan_tag < 0x1000) {
- if (mlx4_find_cached_vlan(dev->dev, port, vlan_tag, &vidx))
- return -ENOENT;
-
- path->vlan_index = vidx;
- path->fl = 1 << 6;
+ if (smac_info->vid < 0x1000) {
+ /* both valid vlan ids */
+ if (smac_info->vid != vlan_tag) {
+ /* different VIDs. unreg old and reg new */
+ err = mlx4_register_vlan(dev->dev, port, vlan_tag, &vidx);
+ if (err)
+ return err;
+ smac_info->candidate_vid = vlan_tag;
+ smac_info->candidate_vlan_index = vidx;
+ smac_info->candidate_vlan_port = port;
+ smac_info->update_vid = 1;
+ path->vlan_index = vidx;
+ } else {
+ path->vlan_index = smac_info->vlan_index;
+ }
+ } else {
+ /* no current vlan tag in qp */
+ err = mlx4_register_vlan(dev->dev, port, vlan_tag, &vidx);
+ if (err)
+ return err;
+ smac_info->candidate_vid = vlan_tag;
+ smac_info->candidate_vlan_index = vidx;
+ smac_info->candidate_vlan_port = port;
+ smac_info->update_vid = 1;
+ path->vlan_index = vidx;
+ }
path->feup |= MLX4_FVL_FORCE_ETH_VLAN;
+ path->fl = 1 << 6;
+ } else {
+ /* have current vlan tag. unregister it at modify-qp success */
+ if (smac_info->vid < 0x1000) {
+ smac_info->candidate_vid = 0xFFFF;
+ smac_info->update_vid = 1;
+ }
}
- } else
+
+ /* get smac_index for RoCE use.
+ * If no smac was yet assigned, register one.
+ * If one was already assigned, but the new mac differs,
+ * unregister the old one and register the new one.
+ */
+ if (!smac_info->smac || smac_info->smac != smac) {
+ /* register candidate now, unreg if needed, after success */
+ smac_index = mlx4_register_mac(dev->dev, port, smac);
+ if (smac_index >= 0) {
+ smac_info->candidate_smac_index = smac_index;
+ smac_info->candidate_smac = smac;
+ smac_info->candidate_smac_port = port;
+ } else {
+ return -EINVAL;
+ }
+ } else {
+ smac_index = smac_info->smac_index;
+ }
+
+ memcpy(path->dmac, ah->dmac, 6);
+ path->ackto = MLX4_IB_LINK_TYPE_ETH;
+ /* put MAC table smac index for IBoE */
+ path->grh_mylmc = (u8) (smac_index) | 0x80;
+ } else {
path->sched_queue = MLX4_IB_DEFAULT_SCHED_QUEUE |
((port - 1) << 6) | ((ah->sl & 0xf) << 2);
+ }
return 0;
}
static int mlx4_set_path(struct mlx4_ib_dev *dev, const struct ib_qp_attr *qp,
enum ib_qp_attr_mask qp_attr_mask,
+ struct mlx4_ib_qp *mqp,
struct mlx4_qp_path *path, u8 port)
{
return _mlx4_set_path(dev, &qp->ah_attr,
mlx4_mac_to_u64((u8 *)qp->smac),
(qp_attr_mask & IB_QP_VID) ? qp->vlan_id : 0xffff,
- path, port);
+ path, &mqp->pri, port);
}
static int mlx4_set_alt_path(struct mlx4_ib_dev *dev,
const struct ib_qp_attr *qp,
enum ib_qp_attr_mask qp_attr_mask,
+ struct mlx4_ib_qp *mqp,
struct mlx4_qp_path *path, u8 port)
{
return _mlx4_set_path(dev, &qp->alt_ah_attr,
mlx4_mac_to_u64((u8 *)qp->alt_smac),
(qp_attr_mask & IB_QP_ALT_VID) ?
qp->alt_vlan_id : 0xffff,
- path, port);
+ path, &mqp->alt, port);
}
static void update_mcg_macs(struct mlx4_ib_dev *dev, struct mlx4_ib_qp *qp)
@@ -1292,6 +1362,37 @@ static void update_mcg_macs(struct mlx4_ib_dev *dev, struct mlx4_ib_qp *qp)
}
}
+static int handle_eth_ud_smac_index(struct mlx4_ib_dev *dev, struct mlx4_ib_qp *qp, u8 *smac,
+ struct mlx4_qp_context *context)
+{
+ struct net_device *ndev;
+ u64 u64_mac;
+ int smac_index;
+
+
+ ndev = dev->iboe.netdevs[qp->port - 1];
+ if (ndev) {
+ smac = ndev->dev_addr;
+ u64_mac = mlx4_mac_to_u64(smac);
+ } else {
+ u64_mac = dev->dev->caps.def_mac[qp->port];
+ }
+
+ context->pri_path.sched_queue = MLX4_IB_DEFAULT_SCHED_QUEUE | ((qp->port - 1) << 6);
+ if (!qp->pri.smac) {
+ smac_index = mlx4_register_mac(dev->dev, qp->port, u64_mac);
+ if (smac_index >= 0) {
+ qp->pri.candidate_smac_index = smac_index;
+ qp->pri.candidate_smac = u64_mac;
+ qp->pri.candidate_smac_port = qp->port;
+ context->pri_path.grh_mylmc = 0x80 | (u8) smac_index;
+ } else {
+ return -ENOENT;
+ }
+ }
+ return 0;
+}
+
static int __mlx4_ib_modify_qp(struct ib_qp *ibqp,
const struct ib_qp_attr *attr, int attr_mask,
enum ib_qp_state cur_state, enum ib_qp_state new_state)
@@ -1403,7 +1504,7 @@ static int __mlx4_ib_modify_qp(struct ib_qp *ibqp,
}
if (attr_mask & IB_QP_AV) {
- if (mlx4_set_path(dev, attr, attr_mask, &context->pri_path,
+ if (mlx4_set_path(dev, attr, attr_mask, qp, &context->pri_path,
attr_mask & IB_QP_PORT ?
attr->port_num : qp->port))
goto out;
@@ -1426,7 +1527,8 @@ static int __mlx4_ib_modify_qp(struct ib_qp *ibqp,
dev->dev->caps.pkey_table_len[attr->alt_port_num])
goto out;
- if (mlx4_set_alt_path(dev, attr, attr_mask, &context->alt_path,
+ if (mlx4_set_alt_path(dev, attr, attr_mask, qp,
+ &context->alt_path,
attr->alt_port_num))
goto out;
@@ -1532,6 +1634,20 @@ static int __mlx4_ib_modify_qp(struct ib_qp *ibqp,
context->pri_path.fl = 0x80;
context->pri_path.sched_queue |= MLX4_IB_DEFAULT_SCHED_QUEUE;
}
+ if (rdma_port_get_link_layer(&dev->ib_dev, qp->port) ==
+ IB_LINK_LAYER_ETHERNET) {
+ if (qp->mlx4_ib_qp_type == MLX4_IB_QPT_TUN_GSI ||
+ qp->mlx4_ib_qp_type == MLX4_IB_QPT_GSI)
+ context->pri_path.feup = 1 << 7; /* don't fsm */
+ /* handle smac_index */
+ if (qp->mlx4_ib_qp_type == MLX4_IB_QPT_UD ||
+ qp->mlx4_ib_qp_type == MLX4_IB_QPT_PROXY_GSI ||
+ qp->mlx4_ib_qp_type == MLX4_IB_QPT_TUN_GSI) {
+ err = handle_eth_ud_smac_index(dev, qp, (u8 *)attr->smac, context);
+ if (err)
+ return -EINVAL;
+ }
+ }
}
if (qp->ibqp.qp_type == IB_QPT_RAW_PACKET)
@@ -1619,28 +1735,113 @@ static int __mlx4_ib_modify_qp(struct ib_qp *ibqp,
* If we moved a kernel QP to RESET, clean up all old CQ
* entries and reinitialize the QP.
*/
- if (new_state == IB_QPS_RESET && !ibqp->uobject) {
- mlx4_ib_cq_clean(recv_cq, qp->mqp.qpn,
- ibqp->srq ? to_msrq(ibqp->srq): NULL);
- if (send_cq != recv_cq)
- mlx4_ib_cq_clean(send_cq, qp->mqp.qpn, NULL);
+ if (new_state == IB_QPS_RESET) {
+ if (!ibqp->uobject) {
+ mlx4_ib_cq_clean(recv_cq, qp->mqp.qpn,
+ ibqp->srq ? to_msrq(ibqp->srq) : NULL);
+ if (send_cq != recv_cq)
+ mlx4_ib_cq_clean(send_cq, qp->mqp.qpn, NULL);
+
+ qp->rq.head = 0;
+ qp->rq.tail = 0;
+ qp->sq.head = 0;
+ qp->sq.tail = 0;
+ qp->sq_next_wqe = 0;
+ if (qp->rq.wqe_cnt)
+ *qp->db.db = 0;
- qp->rq.head = 0;
- qp->rq.tail = 0;
- qp->sq.head = 0;
- qp->sq.tail = 0;
- qp->sq_next_wqe = 0;
- if (qp->rq.wqe_cnt)
- *qp->db.db = 0;
+ if (qp->flags & MLX4_IB_QP_NETIF)
+ mlx4_ib_steer_qp_reg(dev, qp, 0);
+ }
+ if (qp->pri.smac) {
+ mlx4_unregister_mac(dev->dev, qp->pri.smac_port, qp->pri.smac);
+ qp->pri.smac = 0;
+ }
+ if (qp->alt.smac) {
+ mlx4_unregister_mac(dev->dev, qp->alt.smac_port, qp->alt.smac);
+ qp->alt.smac = 0;
+ }
+ if (qp->pri.vid < 0x1000) {
+ mlx4_unregister_vlan(dev->dev, qp->pri.vlan_port, qp->pri.vid);
+ qp->pri.vid = 0xFFFF;
+ qp->pri.candidate_vid = 0xFFFF;
+ qp->pri.update_vid = 0;
+ }
- if (qp->flags & MLX4_IB_QP_NETIF)
- mlx4_ib_steer_qp_reg(dev, qp, 0);
+ if (qp->alt.vid < 0x1000) {
+ mlx4_unregister_vlan(dev->dev, qp->alt.vlan_port, qp->alt.vid);
+ qp->alt.vid = 0xFFFF;
+ qp->alt.candidate_vid = 0xFFFF;
+ qp->alt.update_vid = 0;
+ }
}
-
out:
if (err && steer_qp)
mlx4_ib_steer_qp_reg(dev, qp, 0);
kfree(context);
+ if (qp->pri.candidate_smac) {
+ if (err) {
+ mlx4_unregister_mac(dev->dev, qp->pri.candidate_smac_port, qp->pri.candidate_smac);
+ } else {
+ if (qp->pri.smac)
+ mlx4_unregister_mac(dev->dev, qp->pri.smac_port, qp->pri.smac);
+ qp->pri.smac = qp->pri.candidate_smac;
+ qp->pri.smac_index = qp->pri.candidate_smac_index;
+ qp->pri.smac_port = qp->pri.candidate_smac_port;
+ }
+ qp->pri.candidate_smac = 0;
+ qp->pri.candidate_smac_index = 0;
+ qp->pri.candidate_smac_port = 0;
+ }
+ if (qp->alt.candidate_smac) {
+ if (err) {
+ mlx4_unregister_mac(dev->dev, qp->alt.candidate_smac_port, qp->alt.candidate_smac);
+ } else {
+ if (qp->alt.smac)
+ mlx4_unregister_mac(dev->dev, qp->alt.smac_port, qp->alt.smac);
+ qp->alt.smac = qp->alt.candidate_smac;
+ qp->alt.smac_index = qp->alt.candidate_smac_index;
+ qp->alt.smac_port = qp->alt.candidate_smac_port;
+ }
+ qp->alt.candidate_smac = 0;
+ qp->alt.candidate_smac_index = 0;
+ qp->alt.candidate_smac_port = 0;
+ }
+
+ if (qp->pri.update_vid) {
+ if (err) {
+ if (qp->pri.candidate_vid < 0x1000)
+ mlx4_unregister_vlan(dev->dev, qp->pri.candidate_vlan_port,
+ qp->pri.candidate_vid);
+ } else {
+ if (qp->pri.vid < 0x1000)
+ mlx4_unregister_vlan(dev->dev, qp->pri.vlan_port,
+ qp->pri.vid);
+ qp->pri.vid = qp->pri.candidate_vid;
+ qp->pri.vlan_port = qp->pri.candidate_vlan_port;
+ qp->pri.vlan_index = qp->pri.candidate_vlan_index;
+ }
+ qp->pri.candidate_vid = 0xFFFF;
+ qp->pri.update_vid = 0;
+ }
+
+ if (qp->alt.update_vid) {
+ if (err) {
+ if (qp->alt.candidate_vid < 0x1000)
+ mlx4_unregister_vlan(dev->dev, qp->alt.candidate_vlan_port,
+ qp->alt.candidate_vid);
+ } else {
+ if (qp->alt.vid < 0x1000)
+ mlx4_unregister_vlan(dev->dev, qp->alt.vlan_port,
+ qp->alt.vid);
+ qp->alt.vid = qp->alt.candidate_vid;
+ qp->alt.vlan_port = qp->alt.candidate_vlan_port;
+ qp->alt.vlan_index = qp->alt.candidate_vlan_index;
+ }
+ qp->alt.candidate_vid = 0xFFFF;
+ qp->alt.update_vid = 0;
+ }
+
return err;
}
@@ -1842,9 +2043,9 @@ static int build_mlx_header(struct mlx4_ib_sqp *sqp, struct ib_send_wr *wr,
{
struct ib_device *ib_dev = sqp->qp.ibqp.device;
struct mlx4_wqe_mlx_seg *mlx = wqe;
+ struct mlx4_wqe_ctrl_seg *ctrl = wqe;
struct mlx4_wqe_inline_seg *inl = wqe + sizeof *mlx;
struct mlx4_ib_ah *ah = to_mah(wr->wr.ud.ah);
- struct net_device *ndev;
union ib_gid sgid;
u16 pkey;
int send_size;
@@ -1868,12 +2069,11 @@ static int build_mlx_header(struct mlx4_ib_sqp *sqp, struct ib_send_wr *wr,
/* When multi-function is enabled, the ib_core gid
* indexes don't necessarily match the hw ones, so
* we must use our own cache */
- sgid.global.subnet_prefix =
- to_mdev(ib_dev)->sriov.demux[sqp->qp.port - 1].
- subnet_prefix;
- sgid.global.interface_id =
- to_mdev(ib_dev)->sriov.demux[sqp->qp.port - 1].
- guid_cache[ah->av.ib.gid_index];
+ err = mlx4_get_roce_gid_from_slave(to_mdev(ib_dev)->dev,
+ be32_to_cpu(ah->av.ib.port_pd) >> 24,
+ ah->av.ib.gid_index, &sgid.raw[0]);
+ if (err)
+ return err;
} else {
err = ib_get_cached_gid(ib_dev,
be32_to_cpu(ah->av.ib.port_pd) >> 24,
@@ -1882,7 +2082,7 @@ static int build_mlx_header(struct mlx4_ib_sqp *sqp, struct ib_send_wr *wr,
return err;
}
- if (ah->av.eth.vlan != 0xffff) {
+ if (ah->av.eth.vlan != cpu_to_be16(0xffff)) {
vlan = be16_to_cpu(ah->av.eth.vlan) & 0x0fff;
is_vlan = 1;
}
@@ -1902,6 +2102,9 @@ static int build_mlx_header(struct mlx4_ib_sqp *sqp, struct ib_send_wr *wr,
sqp->ud_header.grh.flow_label =
ah->av.ib.sl_tclass_flowlabel & cpu_to_be32(0xfffff);
sqp->ud_header.grh.hop_limit = ah->av.ib.hop_limit;
+ if (is_eth)
+ memcpy(sqp->ud_header.grh.source_gid.raw, sgid.raw, 16);
+ else {
if (mlx4_is_mfunc(to_mdev(ib_dev)->dev)) {
/* When multi-function is enabled, the ib_core gid
* indexes don't necessarily match the hw ones, so
@@ -1917,6 +2120,7 @@ static int build_mlx_header(struct mlx4_ib_sqp *sqp, struct ib_send_wr *wr,
be32_to_cpu(ah->av.ib.port_pd) >> 24,
ah->av.ib.gid_index,
&sqp->ud_header.grh.source_gid);
+ }
memcpy(sqp->ud_header.grh.destination_gid.raw,
ah->av.ib.dgid, 16);
}
@@ -1949,16 +2153,23 @@ static int build_mlx_header(struct mlx4_ib_sqp *sqp, struct ib_send_wr *wr,
if (is_eth) {
u8 *smac;
+ struct in6_addr in6;
+
u16 pcp = (be32_to_cpu(ah->av.ib.sl_tclass_flowlabel) >> 29) << 13;
mlx->sched_prio = cpu_to_be16(pcp);
memcpy(sqp->ud_header.eth.dmac_h, ah->av.eth.mac, 6);
/* FIXME: cache smac value? */
- ndev = to_mdev(sqp->qp.ibqp.device)->iboe.netdevs[sqp->qp.port - 1];
- if (!ndev)
- return -ENODEV;
- smac = ndev->dev_addr;
+ memcpy(&ctrl->srcrb_flags16[0], ah->av.eth.mac, 2);
+ memcpy(&ctrl->imm, ah->av.eth.mac + 2, 4);
+ memcpy(&in6, sgid.raw, sizeof(in6));
+
+ if (!mlx4_is_mfunc(to_mdev(ib_dev)->dev))
+ smac = to_mdev(sqp->qp.ibqp.device)->
+ iboe.netdevs[sqp->qp.port - 1]->dev_addr;
+ else /* use the src mac of the tunnel */
+ smac = ah->av.eth.s_mac;
memcpy(sqp->ud_header.eth.smac_h, smac, 6);
if (!memcmp(sqp->ud_header.eth.smac_h, sqp->ud_header.eth.dmac_h, 6))
mlx->flags |= cpu_to_be32(MLX4_WQE_CTRL_FORCE_LOOPBACK);
@@ -2190,6 +2401,8 @@ static void build_tunnel_header(struct ib_send_wr *wr, void *wqe, unsigned *mlx_
hdr.remote_qpn = cpu_to_be32(wr->wr.ud.remote_qpn);
hdr.pkey_index = cpu_to_be16(wr->wr.ud.pkey_index);
hdr.qkey = cpu_to_be32(wr->wr.ud.remote_qkey);
+ memcpy(hdr.mac, ah->av.eth.mac, 6);
+ hdr.vlan = ah->av.eth.vlan;
spc = MLX4_INLINE_ALIGN -
((unsigned long) (inl + 1) & (MLX4_INLINE_ALIGN - 1));
diff --git a/drivers/infiniband/hw/mlx4/sysfs.c b/drivers/infiniband/hw/mlx4/sysfs.c
index db2ea31df832..5a38e43eca65 100644
--- a/drivers/infiniband/hw/mlx4/sysfs.c
+++ b/drivers/infiniband/hw/mlx4/sysfs.c
@@ -627,6 +627,7 @@ static int register_one_pkey_tree(struct mlx4_ib_dev *dev, int slave)
int port;
struct kobject *p, *t;
struct mlx4_port *mport;
+ struct mlx4_active_ports actv_ports;
get_name(dev, name, slave, sizeof name);
@@ -649,7 +650,11 @@ static int register_one_pkey_tree(struct mlx4_ib_dev *dev, int slave)
goto err_ports;
}
+ actv_ports = mlx4_get_active_ports(dev->dev, slave);
+
for (port = 1; port <= dev->dev->caps.num_ports; ++port) {
+ if (!test_bit(port - 1, actv_ports.ports))
+ continue;
err = add_port(dev, port, slave);
if (err)
goto err_add;
diff --git a/drivers/infiniband/hw/mlx5/cq.c b/drivers/infiniband/hw/mlx5/cq.c
index b1705ce6eb88..62bb6b49dc1d 100644
--- a/drivers/infiniband/hw/mlx5/cq.c
+++ b/drivers/infiniband/hw/mlx5/cq.c
@@ -366,6 +366,38 @@ static void free_cq_buf(struct mlx5_ib_dev *dev, struct mlx5_ib_cq_buf *buf)
mlx5_buf_free(&dev->mdev, &buf->buf);
}
+static void get_sig_err_item(struct mlx5_sig_err_cqe *cqe,
+ struct ib_sig_err *item)
+{
+ u16 syndrome = be16_to_cpu(cqe->syndrome);
+
+#define GUARD_ERR (1 << 13)
+#define APPTAG_ERR (1 << 12)
+#define REFTAG_ERR (1 << 11)
+
+ if (syndrome & GUARD_ERR) {
+ item->err_type = IB_SIG_BAD_GUARD;
+ item->expected = be32_to_cpu(cqe->expected_trans_sig) >> 16;
+ item->actual = be32_to_cpu(cqe->actual_trans_sig) >> 16;
+ } else
+ if (syndrome & REFTAG_ERR) {
+ item->err_type = IB_SIG_BAD_REFTAG;
+ item->expected = be32_to_cpu(cqe->expected_reftag);
+ item->actual = be32_to_cpu(cqe->actual_reftag);
+ } else
+ if (syndrome & APPTAG_ERR) {
+ item->err_type = IB_SIG_BAD_APPTAG;
+ item->expected = be32_to_cpu(cqe->expected_trans_sig) & 0xffff;
+ item->actual = be32_to_cpu(cqe->actual_trans_sig) & 0xffff;
+ } else {
+ pr_err("Got signature completion error with bad syndrome %04x\n",
+ syndrome);
+ }
+
+ item->sig_err_offset = be64_to_cpu(cqe->err_offset);
+ item->key = be32_to_cpu(cqe->mkey);
+}
+
static int mlx5_poll_one(struct mlx5_ib_cq *cq,
struct mlx5_ib_qp **cur_qp,
struct ib_wc *wc)
@@ -375,6 +407,9 @@ static int mlx5_poll_one(struct mlx5_ib_cq *cq,
struct mlx5_cqe64 *cqe64;
struct mlx5_core_qp *mqp;
struct mlx5_ib_wq *wq;
+ struct mlx5_sig_err_cqe *sig_err_cqe;
+ struct mlx5_core_mr *mmr;
+ struct mlx5_ib_mr *mr;
uint8_t opcode;
uint32_t qpn;
u16 wqe_ctr;
@@ -475,6 +510,33 @@ repoll:
}
}
break;
+ case MLX5_CQE_SIG_ERR:
+ sig_err_cqe = (struct mlx5_sig_err_cqe *)cqe64;
+
+ read_lock(&dev->mdev.priv.mr_table.lock);
+ mmr = __mlx5_mr_lookup(&dev->mdev,
+ mlx5_base_mkey(be32_to_cpu(sig_err_cqe->mkey)));
+ if (unlikely(!mmr)) {
+ read_unlock(&dev->mdev.priv.mr_table.lock);
+ mlx5_ib_warn(dev, "CQE@CQ %06x for unknown MR %6x\n",
+ cq->mcq.cqn, be32_to_cpu(sig_err_cqe->mkey));
+ return -EINVAL;
+ }
+
+ mr = to_mibmr(mmr);
+ get_sig_err_item(sig_err_cqe, &mr->sig->err_item);
+ mr->sig->sig_err_exists = true;
+ mr->sig->sigerr_count++;
+
+ mlx5_ib_warn(dev, "CQN: 0x%x Got SIGERR on key: 0x%x err_type %x err_offset %llx expected %x actual %x\n",
+ cq->mcq.cqn, mr->sig->err_item.key,
+ mr->sig->err_item.err_type,
+ mr->sig->err_item.sig_err_offset,
+ mr->sig->err_item.expected,
+ mr->sig->err_item.actual);
+
+ read_unlock(&dev->mdev.priv.mr_table.lock);
+ goto repoll;
}
return 0;
diff --git a/drivers/infiniband/hw/mlx5/doorbell.c b/drivers/infiniband/hw/mlx5/doorbell.c
index 256a23344f28..ece028fc47d6 100644
--- a/drivers/infiniband/hw/mlx5/doorbell.c
+++ b/drivers/infiniband/hw/mlx5/doorbell.c
@@ -47,7 +47,6 @@ int mlx5_ib_db_map_user(struct mlx5_ib_ucontext *context, unsigned long virt,
struct mlx5_db *db)
{
struct mlx5_ib_user_db_page *page;
- struct ib_umem_chunk *chunk;
int err = 0;
mutex_lock(&context->db_page_mutex);
@@ -75,8 +74,7 @@ int mlx5_ib_db_map_user(struct mlx5_ib_ucontext *context, unsigned long virt,
list_add(&page->list, &context->db_page_list);
found:
- chunk = list_entry(page->umem->chunk_list.next, struct ib_umem_chunk, list);
- db->dma = sg_dma_address(chunk->page_list) + (virt & ~PAGE_MASK);
+ db->dma = sg_dma_address(page->umem->sg_head.sgl) + (virt & ~PAGE_MASK);
db->u.user_page = page;
++page->refcnt;
diff --git a/drivers/infiniband/hw/mlx5/main.c b/drivers/infiniband/hw/mlx5/main.c
index aa03e732b6a8..364d4b6937f5 100644
--- a/drivers/infiniband/hw/mlx5/main.c
+++ b/drivers/infiniband/hw/mlx5/main.c
@@ -46,8 +46,8 @@
#include "mlx5_ib.h"
#define DRIVER_NAME "mlx5_ib"
-#define DRIVER_VERSION "1.0"
-#define DRIVER_RELDATE "June 2013"
+#define DRIVER_VERSION "2.2-1"
+#define DRIVER_RELDATE "Feb 2014"
MODULE_AUTHOR("Eli Cohen <eli@mellanox.com>");
MODULE_DESCRIPTION("Mellanox Connect-IB HCA IB driver");
@@ -273,6 +273,17 @@ static int mlx5_ib_query_device(struct ib_device *ibdev,
if (flags & MLX5_DEV_CAP_FLAG_XRC)
props->device_cap_flags |= IB_DEVICE_XRC;
props->device_cap_flags |= IB_DEVICE_MEM_MGT_EXTENSIONS;
+ if (flags & MLX5_DEV_CAP_FLAG_SIG_HAND_OVER) {
+ props->device_cap_flags |= IB_DEVICE_SIGNATURE_HANDOVER;
+ /* At this stage no support for signature handover */
+ props->sig_prot_cap = IB_PROT_T10DIF_TYPE_1 |
+ IB_PROT_T10DIF_TYPE_2 |
+ IB_PROT_T10DIF_TYPE_3;
+ props->sig_guard_cap = IB_GUARD_T10DIF_CRC |
+ IB_GUARD_T10DIF_CSUM;
+ }
+ if (flags & MLX5_DEV_CAP_FLAG_BLOCK_MCAST)
+ props->device_cap_flags |= IB_DEVICE_BLOCK_MULTICAST_LOOPBACK;
props->vendor_id = be32_to_cpup((__be32 *)(out_mad->data + 36)) &
0xffffff;
@@ -1423,12 +1434,15 @@ static int init_one(struct pci_dev *pdev,
dev->ib_dev.get_dma_mr = mlx5_ib_get_dma_mr;
dev->ib_dev.reg_user_mr = mlx5_ib_reg_user_mr;
dev->ib_dev.dereg_mr = mlx5_ib_dereg_mr;
+ dev->ib_dev.destroy_mr = mlx5_ib_destroy_mr;
dev->ib_dev.attach_mcast = mlx5_ib_mcg_attach;
dev->ib_dev.detach_mcast = mlx5_ib_mcg_detach;
dev->ib_dev.process_mad = mlx5_ib_process_mad;
+ dev->ib_dev.create_mr = mlx5_ib_create_mr;
dev->ib_dev.alloc_fast_reg_mr = mlx5_ib_alloc_fast_reg_mr;
dev->ib_dev.alloc_fast_reg_page_list = mlx5_ib_alloc_fast_reg_page_list;
dev->ib_dev.free_fast_reg_page_list = mlx5_ib_free_fast_reg_page_list;
+ dev->ib_dev.check_mr_status = mlx5_ib_check_mr_status;
if (mdev->caps.flags & MLX5_DEV_CAP_FLAG_XRC) {
dev->ib_dev.alloc_xrcd = mlx5_ib_alloc_xrcd;
diff --git a/drivers/infiniband/hw/mlx5/mem.c b/drivers/infiniband/hw/mlx5/mem.c
index 3a5322870b96..8499aec94db6 100644
--- a/drivers/infiniband/hw/mlx5/mem.c
+++ b/drivers/infiniband/hw/mlx5/mem.c
@@ -44,16 +44,17 @@
void mlx5_ib_cont_pages(struct ib_umem *umem, u64 addr, int *count, int *shift,
int *ncont, int *order)
{
- struct ib_umem_chunk *chunk;
unsigned long tmp;
unsigned long m;
- int i, j, k;
+ int i, k;
u64 base = 0;
int p = 0;
int skip;
int mask;
u64 len;
u64 pfn;
+ struct scatterlist *sg;
+ int entry;
addr = addr >> PAGE_SHIFT;
tmp = (unsigned long)addr;
@@ -61,32 +62,31 @@ void mlx5_ib_cont_pages(struct ib_umem *umem, u64 addr, int *count, int *shift,
skip = 1 << m;
mask = skip - 1;
i = 0;
- list_for_each_entry(chunk, &umem->chunk_list, list)
- for (j = 0; j < chunk->nmap; j++) {
- len = sg_dma_len(&chunk->page_list[j]) >> PAGE_SHIFT;
- pfn = sg_dma_address(&chunk->page_list[j]) >> PAGE_SHIFT;
- for (k = 0; k < len; k++) {
- if (!(i & mask)) {
- tmp = (unsigned long)pfn;
- m = min(m, find_first_bit(&tmp, sizeof(tmp)));
+ for_each_sg(umem->sg_head.sgl, sg, umem->nmap, entry) {
+ len = sg_dma_len(sg) >> PAGE_SHIFT;
+ pfn = sg_dma_address(sg) >> PAGE_SHIFT;
+ for (k = 0; k < len; k++) {
+ if (!(i & mask)) {
+ tmp = (unsigned long)pfn;
+ m = min(m, find_first_bit(&tmp, sizeof(tmp)));
+ skip = 1 << m;
+ mask = skip - 1;
+ base = pfn;
+ p = 0;
+ } else {
+ if (base + p != pfn) {
+ tmp = (unsigned long)p;
+ m = find_first_bit(&tmp, sizeof(tmp));
skip = 1 << m;
mask = skip - 1;
base = pfn;
p = 0;
- } else {
- if (base + p != pfn) {
- tmp = (unsigned long)p;
- m = find_first_bit(&tmp, sizeof(tmp));
- skip = 1 << m;
- mask = skip - 1;
- base = pfn;
- p = 0;
- }
}
- p++;
- i++;
}
+ p++;
+ i++;
}
+ }
if (i) {
m = min_t(unsigned long, ilog2(roundup_pow_of_two(i)), m);
@@ -112,32 +112,32 @@ void mlx5_ib_populate_pas(struct mlx5_ib_dev *dev, struct ib_umem *umem,
{
int shift = page_shift - PAGE_SHIFT;
int mask = (1 << shift) - 1;
- struct ib_umem_chunk *chunk;
- int i, j, k;
+ int i, k;
u64 cur = 0;
u64 base;
int len;
+ struct scatterlist *sg;
+ int entry;
i = 0;
- list_for_each_entry(chunk, &umem->chunk_list, list)
- for (j = 0; j < chunk->nmap; j++) {
- len = sg_dma_len(&chunk->page_list[j]) >> PAGE_SHIFT;
- base = sg_dma_address(&chunk->page_list[j]);
- for (k = 0; k < len; k++) {
- if (!(i & mask)) {
- cur = base + (k << PAGE_SHIFT);
- if (umr)
- cur |= 3;
+ for_each_sg(umem->sg_head.sgl, sg, umem->nmap, entry) {
+ len = sg_dma_len(sg) >> PAGE_SHIFT;
+ base = sg_dma_address(sg);
+ for (k = 0; k < len; k++) {
+ if (!(i & mask)) {
+ cur = base + (k << PAGE_SHIFT);
+ if (umr)
+ cur |= 3;
- pas[i >> shift] = cpu_to_be64(cur);
- mlx5_ib_dbg(dev, "pas[%d] 0x%llx\n",
- i >> shift, be64_to_cpu(pas[i >> shift]));
- } else
- mlx5_ib_dbg(dev, "=====> 0x%llx\n",
- base + (k << PAGE_SHIFT));
- i++;
- }
+ pas[i >> shift] = cpu_to_be64(cur);
+ mlx5_ib_dbg(dev, "pas[%d] 0x%llx\n",
+ i >> shift, be64_to_cpu(pas[i >> shift]));
+ } else
+ mlx5_ib_dbg(dev, "=====> 0x%llx\n",
+ base + (k << PAGE_SHIFT));
+ i++;
}
+ }
}
int mlx5_ib_get_buf_offset(u64 addr, int page_shift, u32 *offset)
diff --git a/drivers/infiniband/hw/mlx5/mlx5_ib.h b/drivers/infiniband/hw/mlx5/mlx5_ib.h
index 389e31965773..50541586e0a6 100644
--- a/drivers/infiniband/hw/mlx5/mlx5_ib.h
+++ b/drivers/infiniband/hw/mlx5/mlx5_ib.h
@@ -189,6 +189,9 @@ struct mlx5_ib_qp {
int create_type;
u32 pa_lkey;
+
+ /* Store signature errors */
+ bool signature_en;
};
struct mlx5_ib_cq_buf {
@@ -265,6 +268,7 @@ struct mlx5_ib_mr {
enum ib_wc_status status;
struct mlx5_ib_dev *dev;
struct mlx5_create_mkey_mbox_out out;
+ struct mlx5_core_sig_ctx *sig;
};
struct mlx5_ib_fast_reg_page_list {
@@ -396,6 +400,11 @@ static inline struct mlx5_ib_qp *to_mibqp(struct mlx5_core_qp *mqp)
return container_of(mqp, struct mlx5_ib_qp, mqp);
}
+static inline struct mlx5_ib_mr *to_mibmr(struct mlx5_core_mr *mmr)
+{
+ return container_of(mmr, struct mlx5_ib_mr, mmr);
+}
+
static inline struct mlx5_ib_pd *to_mpd(struct ib_pd *ibpd)
{
return container_of(ibpd, struct mlx5_ib_pd, ibpd);
@@ -495,6 +504,9 @@ struct ib_mr *mlx5_ib_reg_user_mr(struct ib_pd *pd, u64 start, u64 length,
u64 virt_addr, int access_flags,
struct ib_udata *udata);
int mlx5_ib_dereg_mr(struct ib_mr *ibmr);
+int mlx5_ib_destroy_mr(struct ib_mr *ibmr);
+struct ib_mr *mlx5_ib_create_mr(struct ib_pd *pd,
+ struct ib_mr_init_attr *mr_init_attr);
struct ib_mr *mlx5_ib_alloc_fast_reg_mr(struct ib_pd *pd,
int max_page_list_len);
struct ib_fast_reg_page_list *mlx5_ib_alloc_fast_reg_page_list(struct ib_device *ibdev,
@@ -530,6 +542,8 @@ int mlx5_mr_cache_init(struct mlx5_ib_dev *dev);
int mlx5_mr_cache_cleanup(struct mlx5_ib_dev *dev);
int mlx5_mr_ib_cont_pages(struct ib_umem *umem, u64 addr, int *count, int *shift);
void mlx5_umr_cq_handler(struct ib_cq *cq, void *cq_context);
+int mlx5_ib_check_mr_status(struct ib_mr *ibmr, u32 check_mask,
+ struct ib_mr_status *mr_status);
static inline void init_query_mad(struct ib_smp *mad)
{
diff --git a/drivers/infiniband/hw/mlx5/mr.c b/drivers/infiniband/hw/mlx5/mr.c
index 7c95ca1f0c25..81392b26d078 100644
--- a/drivers/infiniband/hw/mlx5/mr.c
+++ b/drivers/infiniband/hw/mlx5/mr.c
@@ -992,6 +992,122 @@ int mlx5_ib_dereg_mr(struct ib_mr *ibmr)
return 0;
}
+struct ib_mr *mlx5_ib_create_mr(struct ib_pd *pd,
+ struct ib_mr_init_attr *mr_init_attr)
+{
+ struct mlx5_ib_dev *dev = to_mdev(pd->device);
+ struct mlx5_create_mkey_mbox_in *in;
+ struct mlx5_ib_mr *mr;
+ int access_mode, err;
+ int ndescs = roundup(mr_init_attr->max_reg_descriptors, 4);
+
+ mr = kzalloc(sizeof(*mr), GFP_KERNEL);
+ if (!mr)
+ return ERR_PTR(-ENOMEM);
+
+ in = kzalloc(sizeof(*in), GFP_KERNEL);
+ if (!in) {
+ err = -ENOMEM;
+ goto err_free;
+ }
+
+ in->seg.status = 1 << 6; /* free */
+ in->seg.xlt_oct_size = cpu_to_be32(ndescs);
+ in->seg.qpn_mkey7_0 = cpu_to_be32(0xffffff << 8);
+ in->seg.flags_pd = cpu_to_be32(to_mpd(pd)->pdn);
+ access_mode = MLX5_ACCESS_MODE_MTT;
+
+ if (mr_init_attr->flags & IB_MR_SIGNATURE_EN) {
+ u32 psv_index[2];
+
+ in->seg.flags_pd = cpu_to_be32(be32_to_cpu(in->seg.flags_pd) |
+ MLX5_MKEY_BSF_EN);
+ in->seg.bsfs_octo_size = cpu_to_be32(MLX5_MKEY_BSF_OCTO_SIZE);
+ mr->sig = kzalloc(sizeof(*mr->sig), GFP_KERNEL);
+ if (!mr->sig) {
+ err = -ENOMEM;
+ goto err_free_in;
+ }
+
+ /* create mem & wire PSVs */
+ err = mlx5_core_create_psv(&dev->mdev, to_mpd(pd)->pdn,
+ 2, psv_index);
+ if (err)
+ goto err_free_sig;
+
+ access_mode = MLX5_ACCESS_MODE_KLM;
+ mr->sig->psv_memory.psv_idx = psv_index[0];
+ mr->sig->psv_wire.psv_idx = psv_index[1];
+
+ mr->sig->sig_status_checked = true;
+ mr->sig->sig_err_exists = false;
+ /* Next UMR, Arm SIGERR */
+ ++mr->sig->sigerr_count;
+ }
+
+ in->seg.flags = MLX5_PERM_UMR_EN | access_mode;
+ err = mlx5_core_create_mkey(&dev->mdev, &mr->mmr, in, sizeof(*in),
+ NULL, NULL, NULL);
+ if (err)
+ goto err_destroy_psv;
+
+ mr->ibmr.lkey = mr->mmr.key;
+ mr->ibmr.rkey = mr->mmr.key;
+ mr->umem = NULL;
+ kfree(in);
+
+ return &mr->ibmr;
+
+err_destroy_psv:
+ if (mr->sig) {
+ if (mlx5_core_destroy_psv(&dev->mdev,
+ mr->sig->psv_memory.psv_idx))
+ mlx5_ib_warn(dev, "failed to destroy mem psv %d\n",
+ mr->sig->psv_memory.psv_idx);
+ if (mlx5_core_destroy_psv(&dev->mdev,
+ mr->sig->psv_wire.psv_idx))
+ mlx5_ib_warn(dev, "failed to destroy wire psv %d\n",
+ mr->sig->psv_wire.psv_idx);
+ }
+err_free_sig:
+ kfree(mr->sig);
+err_free_in:
+ kfree(in);
+err_free:
+ kfree(mr);
+ return ERR_PTR(err);
+}
+
+int mlx5_ib_destroy_mr(struct ib_mr *ibmr)
+{
+ struct mlx5_ib_dev *dev = to_mdev(ibmr->device);
+ struct mlx5_ib_mr *mr = to_mmr(ibmr);
+ int err;
+
+ if (mr->sig) {
+ if (mlx5_core_destroy_psv(&dev->mdev,
+ mr->sig->psv_memory.psv_idx))
+ mlx5_ib_warn(dev, "failed to destroy mem psv %d\n",
+ mr->sig->psv_memory.psv_idx);
+ if (mlx5_core_destroy_psv(&dev->mdev,
+ mr->sig->psv_wire.psv_idx))
+ mlx5_ib_warn(dev, "failed to destroy wire psv %d\n",
+ mr->sig->psv_wire.psv_idx);
+ kfree(mr->sig);
+ }
+
+ err = mlx5_core_destroy_mkey(&dev->mdev, &mr->mmr);
+ if (err) {
+ mlx5_ib_warn(dev, "failed to destroy mkey 0x%x (%d)\n",
+ mr->mmr.key, err);
+ return err;
+ }
+
+ kfree(mr);
+
+ return err;
+}
+
struct ib_mr *mlx5_ib_alloc_fast_reg_mr(struct ib_pd *pd,
int max_page_list_len)
{
@@ -1077,3 +1193,44 @@ void mlx5_ib_free_fast_reg_page_list(struct ib_fast_reg_page_list *page_list)
kfree(mfrpl->ibfrpl.page_list);
kfree(mfrpl);
}
+
+int mlx5_ib_check_mr_status(struct ib_mr *ibmr, u32 check_mask,
+ struct ib_mr_status *mr_status)
+{
+ struct mlx5_ib_mr *mmr = to_mmr(ibmr);
+ int ret = 0;
+
+ if (check_mask & ~IB_MR_CHECK_SIG_STATUS) {
+ pr_err("Invalid status check mask\n");
+ ret = -EINVAL;
+ goto done;
+ }
+
+ mr_status->fail_status = 0;
+ if (check_mask & IB_MR_CHECK_SIG_STATUS) {
+ if (!mmr->sig) {
+ ret = -EINVAL;
+ pr_err("signature status check requested on a non-signature enabled MR\n");
+ goto done;
+ }
+
+ mmr->sig->sig_status_checked = true;
+ if (!mmr->sig->sig_err_exists)
+ goto done;
+
+ if (ibmr->lkey == mmr->sig->err_item.key)
+ memcpy(&mr_status->sig_err, &mmr->sig->err_item,
+ sizeof(mr_status->sig_err));
+ else {
+ mr_status->sig_err.err_type = IB_SIG_BAD_GUARD;
+ mr_status->sig_err.sig_err_offset = 0;
+ mr_status->sig_err.key = mmr->sig->err_item.key;
+ }
+
+ mmr->sig->sig_err_exists = false;
+ mr_status->fail_status |= IB_MR_CHECK_SIG_STATUS;
+ }
+
+done:
+ return ret;
+}
diff --git a/drivers/infiniband/hw/mlx5/qp.c b/drivers/infiniband/hw/mlx5/qp.c
index 7dfe8a1c84cf..dc930ed21eca 100644
--- a/drivers/infiniband/hw/mlx5/qp.c
+++ b/drivers/infiniband/hw/mlx5/qp.c
@@ -256,8 +256,11 @@ static int calc_send_wqe(struct ib_qp_init_attr *attr)
}
size += attr->cap.max_send_sge * sizeof(struct mlx5_wqe_data_seg);
-
- return ALIGN(max_t(int, inl_size, size), MLX5_SEND_WQE_BB);
+ if (attr->create_flags & IB_QP_CREATE_SIGNATURE_EN &&
+ ALIGN(max_t(int, inl_size, size), MLX5_SEND_WQE_BB) < MLX5_SIG_WQE_SIZE)
+ return MLX5_SIG_WQE_SIZE;
+ else
+ return ALIGN(max_t(int, inl_size, size), MLX5_SEND_WQE_BB);
}
static int calc_sq_size(struct mlx5_ib_dev *dev, struct ib_qp_init_attr *attr,
@@ -284,6 +287,9 @@ static int calc_sq_size(struct mlx5_ib_dev *dev, struct ib_qp_init_attr *attr,
sizeof(struct mlx5_wqe_inline_seg);
attr->cap.max_inline_data = qp->max_inline_data;
+ if (attr->create_flags & IB_QP_CREATE_SIGNATURE_EN)
+ qp->signature_en = true;
+
wq_size = roundup_pow_of_two(attr->cap.max_send_wr * wqe_size);
qp->sq.wqe_cnt = wq_size / MLX5_SEND_WQE_BB;
if (qp->sq.wqe_cnt > dev->mdev.caps.max_wqes) {
@@ -665,7 +671,7 @@ static int create_kernel_qp(struct mlx5_ib_dev *dev,
int err;
uuari = &dev->mdev.priv.uuari;
- if (init_attr->create_flags)
+ if (init_attr->create_flags & ~IB_QP_CREATE_SIGNATURE_EN)
return -EINVAL;
if (init_attr->qp_type == MLX5_IB_QPT_REG_UMR)
@@ -801,6 +807,15 @@ static int create_qp_common(struct mlx5_ib_dev *dev, struct ib_pd *pd,
spin_lock_init(&qp->sq.lock);
spin_lock_init(&qp->rq.lock);
+ if (init_attr->create_flags & IB_QP_CREATE_BLOCK_MULTICAST_LOOPBACK) {
+ if (!(dev->mdev.caps.flags & MLX5_DEV_CAP_FLAG_BLOCK_MCAST)) {
+ mlx5_ib_dbg(dev, "block multicast loopback isn't supported\n");
+ return -EINVAL;
+ } else {
+ qp->flags |= MLX5_IB_QP_BLOCK_MULTICAST_LOOPBACK;
+ }
+ }
+
if (init_attr->sq_sig_type == IB_SIGNAL_ALL_WR)
qp->sq_signal_bits = MLX5_WQE_CTRL_CQ_UPDATE;
@@ -872,6 +887,9 @@ static int create_qp_common(struct mlx5_ib_dev *dev, struct ib_pd *pd,
if (qp->wq_sig)
in->ctx.flags_pd |= cpu_to_be32(MLX5_QP_ENABLE_SIG);
+ if (qp->flags & MLX5_IB_QP_BLOCK_MULTICAST_LOOPBACK)
+ in->ctx.flags_pd |= cpu_to_be32(MLX5_QP_BLOCK_MCAST);
+
if (qp->scat_cqe && is_connected(init_attr->qp_type)) {
int rcqe_sz;
int scqe_sz;
@@ -1771,6 +1789,27 @@ static __be64 frwr_mkey_mask(void)
return cpu_to_be64(result);
}
+static __be64 sig_mkey_mask(void)
+{
+ u64 result;
+
+ result = MLX5_MKEY_MASK_LEN |
+ MLX5_MKEY_MASK_PAGE_SIZE |
+ MLX5_MKEY_MASK_START_ADDR |
+ MLX5_MKEY_MASK_EN_SIGERR |
+ MLX5_MKEY_MASK_EN_RINVAL |
+ MLX5_MKEY_MASK_KEY |
+ MLX5_MKEY_MASK_LR |
+ MLX5_MKEY_MASK_LW |
+ MLX5_MKEY_MASK_RR |
+ MLX5_MKEY_MASK_RW |
+ MLX5_MKEY_MASK_SMALL_FENCE |
+ MLX5_MKEY_MASK_FREE |
+ MLX5_MKEY_MASK_BSF_EN;
+
+ return cpu_to_be64(result);
+}
+
static void set_frwr_umr_segment(struct mlx5_wqe_umr_ctrl_seg *umr,
struct ib_send_wr *wr, int li)
{
@@ -1826,7 +1865,7 @@ static u8 get_umr_flags(int acc)
(acc & IB_ACCESS_REMOTE_WRITE ? MLX5_PERM_REMOTE_WRITE : 0) |
(acc & IB_ACCESS_REMOTE_READ ? MLX5_PERM_REMOTE_READ : 0) |
(acc & IB_ACCESS_LOCAL_WRITE ? MLX5_PERM_LOCAL_WRITE : 0) |
- MLX5_PERM_LOCAL_READ | MLX5_PERM_UMR_EN | MLX5_ACCESS_MODE_MTT;
+ MLX5_PERM_LOCAL_READ | MLX5_PERM_UMR_EN;
}
static void set_mkey_segment(struct mlx5_mkey_seg *seg, struct ib_send_wr *wr,
@@ -1838,7 +1877,8 @@ static void set_mkey_segment(struct mlx5_mkey_seg *seg, struct ib_send_wr *wr,
return;
}
- seg->flags = get_umr_flags(wr->wr.fast_reg.access_flags);
+ seg->flags = get_umr_flags(wr->wr.fast_reg.access_flags) |
+ MLX5_ACCESS_MODE_MTT;
*writ = seg->flags & (MLX5_PERM_LOCAL_WRITE | IB_ACCESS_REMOTE_WRITE);
seg->qpn_mkey7_0 = cpu_to_be32((wr->wr.fast_reg.rkey & 0xff) | 0xffffff00);
seg->flags_pd = cpu_to_be32(MLX5_MKEY_REMOTE_INVAL);
@@ -1954,6 +1994,342 @@ static int set_data_inl_seg(struct mlx5_ib_qp *qp, struct ib_send_wr *wr,
return 0;
}
+static u16 prot_field_size(enum ib_signature_type type)
+{
+ switch (type) {
+ case IB_SIG_TYPE_T10_DIF:
+ return MLX5_DIF_SIZE;
+ default:
+ return 0;
+ }
+}
+
+static u8 bs_selector(int block_size)
+{
+ switch (block_size) {
+ case 512: return 0x1;
+ case 520: return 0x2;
+ case 4096: return 0x3;
+ case 4160: return 0x4;
+ case 1073741824: return 0x5;
+ default: return 0;
+ }
+}
+
+static int format_selector(struct ib_sig_attrs *attr,
+ struct ib_sig_domain *domain,
+ int *selector)
+{
+
+#define FORMAT_DIF_NONE 0
+#define FORMAT_DIF_CRC_INC 8
+#define FORMAT_DIF_CRC_NO_INC 12
+#define FORMAT_DIF_CSUM_INC 13
+#define FORMAT_DIF_CSUM_NO_INC 14
+
+ switch (domain->sig.dif.type) {
+ case IB_T10DIF_NONE:
+ /* No DIF */
+ *selector = FORMAT_DIF_NONE;
+ break;
+ case IB_T10DIF_TYPE1: /* Fall through */
+ case IB_T10DIF_TYPE2:
+ switch (domain->sig.dif.bg_type) {
+ case IB_T10DIF_CRC:
+ *selector = FORMAT_DIF_CRC_INC;
+ break;
+ case IB_T10DIF_CSUM:
+ *selector = FORMAT_DIF_CSUM_INC;
+ break;
+ default:
+ return 1;
+ }
+ break;
+ case IB_T10DIF_TYPE3:
+ switch (domain->sig.dif.bg_type) {
+ case IB_T10DIF_CRC:
+ *selector = domain->sig.dif.type3_inc_reftag ?
+ FORMAT_DIF_CRC_INC :
+ FORMAT_DIF_CRC_NO_INC;
+ break;
+ case IB_T10DIF_CSUM:
+ *selector = domain->sig.dif.type3_inc_reftag ?
+ FORMAT_DIF_CSUM_INC :
+ FORMAT_DIF_CSUM_NO_INC;
+ break;
+ default:
+ return 1;
+ }
+ break;
+ default:
+ return 1;
+ }
+
+ return 0;
+}
+
+static int mlx5_set_bsf(struct ib_mr *sig_mr,
+ struct ib_sig_attrs *sig_attrs,
+ struct mlx5_bsf *bsf, u32 data_size)
+{
+ struct mlx5_core_sig_ctx *msig = to_mmr(sig_mr)->sig;
+ struct mlx5_bsf_basic *basic = &bsf->basic;
+ struct ib_sig_domain *mem = &sig_attrs->mem;
+ struct ib_sig_domain *wire = &sig_attrs->wire;
+ int ret, selector;
+
+ switch (sig_attrs->mem.sig_type) {
+ case IB_SIG_TYPE_T10_DIF:
+ if (sig_attrs->wire.sig_type != IB_SIG_TYPE_T10_DIF)
+ return -EINVAL;
+
+ /* Input domain check byte mask */
+ basic->check_byte_mask = sig_attrs->check_mask;
+ if (mem->sig.dif.pi_interval == wire->sig.dif.pi_interval &&
+ mem->sig.dif.type == wire->sig.dif.type) {
+ /* Same block structure */
+ basic->bsf_size_sbs = 1 << 4;
+ if (mem->sig.dif.bg_type == wire->sig.dif.bg_type)
+ basic->wire.copy_byte_mask = 0xff;
+ else
+ basic->wire.copy_byte_mask = 0x3f;
+ } else
+ basic->wire.bs_selector = bs_selector(wire->sig.dif.pi_interval);
+
+ basic->mem.bs_selector = bs_selector(mem->sig.dif.pi_interval);
+ basic->raw_data_size = cpu_to_be32(data_size);
+
+ ret = format_selector(sig_attrs, mem, &selector);
+ if (ret)
+ return -EINVAL;
+ basic->m_bfs_psv = cpu_to_be32(selector << 24 |
+ msig->psv_memory.psv_idx);
+
+ ret = format_selector(sig_attrs, wire, &selector);
+ if (ret)
+ return -EINVAL;
+ basic->w_bfs_psv = cpu_to_be32(selector << 24 |
+ msig->psv_wire.psv_idx);
+ break;
+
+ default:
+ return -EINVAL;
+ }
+
+ return 0;
+}
+
+static int set_sig_data_segment(struct ib_send_wr *wr, struct mlx5_ib_qp *qp,
+ void **seg, int *size)
+{
+ struct ib_sig_attrs *sig_attrs = wr->wr.sig_handover.sig_attrs;
+ struct ib_mr *sig_mr = wr->wr.sig_handover.sig_mr;
+ struct mlx5_bsf *bsf;
+ u32 data_len = wr->sg_list->length;
+ u32 data_key = wr->sg_list->lkey;
+ u64 data_va = wr->sg_list->addr;
+ int ret;
+ int wqe_size;
+
+ if (!wr->wr.sig_handover.prot) {
+ /**
+ * Source domain doesn't contain signature information
+ * So need construct:
+ * ------------------
+ * | data_klm |
+ * ------------------
+ * | BSF |
+ * ------------------
+ **/
+ struct mlx5_klm *data_klm = *seg;
+
+ data_klm->bcount = cpu_to_be32(data_len);
+ data_klm->key = cpu_to_be32(data_key);
+ data_klm->va = cpu_to_be64(data_va);
+ wqe_size = ALIGN(sizeof(*data_klm), 64);
+ } else {
+ /**
+ * Source domain contains signature information
+ * So need construct a strided block format:
+ * ---------------------------
+ * | stride_block_ctrl |
+ * ---------------------------
+ * | data_klm |
+ * ---------------------------
+ * | prot_klm |
+ * ---------------------------
+ * | BSF |
+ * ---------------------------
+ **/
+ struct mlx5_stride_block_ctrl_seg *sblock_ctrl;
+ struct mlx5_stride_block_entry *data_sentry;
+ struct mlx5_stride_block_entry *prot_sentry;
+ u32 prot_key = wr->wr.sig_handover.prot->lkey;
+ u64 prot_va = wr->wr.sig_handover.prot->addr;
+ u16 block_size = sig_attrs->mem.sig.dif.pi_interval;
+ int prot_size;
+
+ sblock_ctrl = *seg;
+ data_sentry = (void *)sblock_ctrl + sizeof(*sblock_ctrl);
+ prot_sentry = (void *)data_sentry + sizeof(*data_sentry);
+
+ prot_size = prot_field_size(sig_attrs->mem.sig_type);
+ if (!prot_size) {
+ pr_err("Bad block size given: %u\n", block_size);
+ return -EINVAL;
+ }
+ sblock_ctrl->bcount_per_cycle = cpu_to_be32(block_size +
+ prot_size);
+ sblock_ctrl->op = cpu_to_be32(MLX5_STRIDE_BLOCK_OP);
+ sblock_ctrl->repeat_count = cpu_to_be32(data_len / block_size);
+ sblock_ctrl->num_entries = cpu_to_be16(2);
+
+ data_sentry->bcount = cpu_to_be16(block_size);
+ data_sentry->key = cpu_to_be32(data_key);
+ data_sentry->va = cpu_to_be64(data_va);
+ prot_sentry->bcount = cpu_to_be16(prot_size);
+ prot_sentry->key = cpu_to_be32(prot_key);
+
+ if (prot_key == data_key && prot_va == data_va) {
+ /**
+ * The data and protection are interleaved
+ * in a single memory region
+ **/
+ prot_sentry->va = cpu_to_be64(data_va + block_size);
+ prot_sentry->stride = cpu_to_be16(block_size + prot_size);
+ data_sentry->stride = prot_sentry->stride;
+ } else {
+ /* The data and protection are two different buffers */
+ prot_sentry->va = cpu_to_be64(prot_va);
+ data_sentry->stride = cpu_to_be16(block_size);
+ prot_sentry->stride = cpu_to_be16(prot_size);
+ }
+ wqe_size = ALIGN(sizeof(*sblock_ctrl) + sizeof(*data_sentry) +
+ sizeof(*prot_sentry), 64);
+ }
+
+ *seg += wqe_size;
+ *size += wqe_size / 16;
+ if (unlikely((*seg == qp->sq.qend)))
+ *seg = mlx5_get_send_wqe(qp, 0);
+
+ bsf = *seg;
+ ret = mlx5_set_bsf(sig_mr, sig_attrs, bsf, data_len);
+ if (ret)
+ return -EINVAL;
+
+ *seg += sizeof(*bsf);
+ *size += sizeof(*bsf) / 16;
+ if (unlikely((*seg == qp->sq.qend)))
+ *seg = mlx5_get_send_wqe(qp, 0);
+
+ return 0;
+}
+
+static void set_sig_mkey_segment(struct mlx5_mkey_seg *seg,
+ struct ib_send_wr *wr, u32 nelements,
+ u32 length, u32 pdn)
+{
+ struct ib_mr *sig_mr = wr->wr.sig_handover.sig_mr;
+ u32 sig_key = sig_mr->rkey;
+ u8 sigerr = to_mmr(sig_mr)->sig->sigerr_count & 1;
+
+ memset(seg, 0, sizeof(*seg));
+
+ seg->flags = get_umr_flags(wr->wr.sig_handover.access_flags) |
+ MLX5_ACCESS_MODE_KLM;
+ seg->qpn_mkey7_0 = cpu_to_be32((sig_key & 0xff) | 0xffffff00);
+ seg->flags_pd = cpu_to_be32(MLX5_MKEY_REMOTE_INVAL | sigerr << 26 |
+ MLX5_MKEY_BSF_EN | pdn);
+ seg->len = cpu_to_be64(length);
+ seg->xlt_oct_size = cpu_to_be32(be16_to_cpu(get_klm_octo(nelements)));
+ seg->bsfs_octo_size = cpu_to_be32(MLX5_MKEY_BSF_OCTO_SIZE);
+}
+
+static void set_sig_umr_segment(struct mlx5_wqe_umr_ctrl_seg *umr,
+ struct ib_send_wr *wr, u32 nelements)
+{
+ memset(umr, 0, sizeof(*umr));
+
+ umr->flags = MLX5_FLAGS_INLINE | MLX5_FLAGS_CHECK_FREE;
+ umr->klm_octowords = get_klm_octo(nelements);
+ umr->bsf_octowords = cpu_to_be16(MLX5_MKEY_BSF_OCTO_SIZE);
+ umr->mkey_mask = sig_mkey_mask();
+}
+
+
+static int set_sig_umr_wr(struct ib_send_wr *wr, struct mlx5_ib_qp *qp,
+ void **seg, int *size)
+{
+ struct mlx5_ib_mr *sig_mr = to_mmr(wr->wr.sig_handover.sig_mr);
+ u32 pdn = get_pd(qp)->pdn;
+ u32 klm_oct_size;
+ int region_len, ret;
+
+ if (unlikely(wr->num_sge != 1) ||
+ unlikely(wr->wr.sig_handover.access_flags &
+ IB_ACCESS_REMOTE_ATOMIC) ||
+ unlikely(!sig_mr->sig) || unlikely(!qp->signature_en) ||
+ unlikely(!sig_mr->sig->sig_status_checked))
+ return -EINVAL;
+
+ /* length of the protected region, data + protection */
+ region_len = wr->sg_list->length;
+ if (wr->wr.sig_handover.prot)
+ region_len += wr->wr.sig_handover.prot->length;
+
+ /**
+ * KLM octoword size - if protection was provided
+ * then we use strided block format (3 octowords),
+ * else we use single KLM (1 octoword)
+ **/
+ klm_oct_size = wr->wr.sig_handover.prot ? 3 : 1;
+
+ set_sig_umr_segment(*seg, wr, klm_oct_size);
+ *seg += sizeof(struct mlx5_wqe_umr_ctrl_seg);
+ *size += sizeof(struct mlx5_wqe_umr_ctrl_seg) / 16;
+ if (unlikely((*seg == qp->sq.qend)))
+ *seg = mlx5_get_send_wqe(qp, 0);
+
+ set_sig_mkey_segment(*seg, wr, klm_oct_size, region_len, pdn);
+ *seg += sizeof(struct mlx5_mkey_seg);
+ *size += sizeof(struct mlx5_mkey_seg) / 16;
+ if (unlikely((*seg == qp->sq.qend)))
+ *seg = mlx5_get_send_wqe(qp, 0);
+
+ ret = set_sig_data_segment(wr, qp, seg, size);
+ if (ret)
+ return ret;
+
+ sig_mr->sig->sig_status_checked = false;
+ return 0;
+}
+
+static int set_psv_wr(struct ib_sig_domain *domain,
+ u32 psv_idx, void **seg, int *size)
+{
+ struct mlx5_seg_set_psv *psv_seg = *seg;
+
+ memset(psv_seg, 0, sizeof(*psv_seg));
+ psv_seg->psv_num = cpu_to_be32(psv_idx);
+ switch (domain->sig_type) {
+ case IB_SIG_TYPE_T10_DIF:
+ psv_seg->transient_sig = cpu_to_be32(domain->sig.dif.bg << 16 |
+ domain->sig.dif.app_tag);
+ psv_seg->ref_tag = cpu_to_be32(domain->sig.dif.ref_tag);
+
+ *seg += sizeof(*psv_seg);
+ *size += sizeof(*psv_seg) / 16;
+ break;
+
+ default:
+ pr_err("Bad signature type given.\n");
+ return 1;
+ }
+
+ return 0;
+}
+
static int set_frwr_li_wr(void **seg, struct ib_send_wr *wr, int *size,
struct mlx5_core_dev *mdev, struct mlx5_ib_pd *pd, struct mlx5_ib_qp *qp)
{
@@ -2041,6 +2417,59 @@ static u8 get_fence(u8 fence, struct ib_send_wr *wr)
}
}
+static int begin_wqe(struct mlx5_ib_qp *qp, void **seg,
+ struct mlx5_wqe_ctrl_seg **ctrl,
+ struct ib_send_wr *wr, int *idx,
+ int *size, int nreq)
+{
+ int err = 0;
+
+ if (unlikely(mlx5_wq_overflow(&qp->sq, nreq, qp->ibqp.send_cq))) {
+ err = -ENOMEM;
+ return err;
+ }
+
+ *idx = qp->sq.cur_post & (qp->sq.wqe_cnt - 1);
+ *seg = mlx5_get_send_wqe(qp, *idx);
+ *ctrl = *seg;
+ *(uint32_t *)(*seg + 8) = 0;
+ (*ctrl)->imm = send_ieth(wr);
+ (*ctrl)->fm_ce_se = qp->sq_signal_bits |
+ (wr->send_flags & IB_SEND_SIGNALED ?
+ MLX5_WQE_CTRL_CQ_UPDATE : 0) |
+ (wr->send_flags & IB_SEND_SOLICITED ?
+ MLX5_WQE_CTRL_SOLICITED : 0);
+
+ *seg += sizeof(**ctrl);
+ *size = sizeof(**ctrl) / 16;
+
+ return err;
+}
+
+static void finish_wqe(struct mlx5_ib_qp *qp,
+ struct mlx5_wqe_ctrl_seg *ctrl,
+ u8 size, unsigned idx, u64 wr_id,
+ int nreq, u8 fence, u8 next_fence,
+ u32 mlx5_opcode)
+{
+ u8 opmod = 0;
+
+ ctrl->opmod_idx_opcode = cpu_to_be32(((u32)(qp->sq.cur_post) << 8) |
+ mlx5_opcode | ((u32)opmod << 24));
+ ctrl->qpn_ds = cpu_to_be32(size | (qp->mqp.qpn << 8));
+ ctrl->fm_ce_se |= fence;
+ qp->fm_cache = next_fence;
+ if (unlikely(qp->wq_sig))
+ ctrl->signature = wq_sig(ctrl);
+
+ qp->sq.wrid[idx] = wr_id;
+ qp->sq.w_list[idx].opcode = mlx5_opcode;
+ qp->sq.wqe_head[idx] = qp->sq.head + nreq;
+ qp->sq.cur_post += DIV_ROUND_UP(size * 16, MLX5_SEND_WQE_BB);
+ qp->sq.w_list[idx].next = qp->sq.cur_post;
+}
+
+
int mlx5_ib_post_send(struct ib_qp *ibqp, struct ib_send_wr *wr,
struct ib_send_wr **bad_wr)
{
@@ -2048,13 +2477,13 @@ int mlx5_ib_post_send(struct ib_qp *ibqp, struct ib_send_wr *wr,
struct mlx5_ib_dev *dev = to_mdev(ibqp->device);
struct mlx5_core_dev *mdev = &dev->mdev;
struct mlx5_ib_qp *qp = to_mqp(ibqp);
+ struct mlx5_ib_mr *mr;
struct mlx5_wqe_data_seg *dpseg;
struct mlx5_wqe_xrc_seg *xrc;
struct mlx5_bf *bf = qp->bf;
int uninitialized_var(size);
void *qend = qp->sq.qend;
unsigned long flags;
- u32 mlx5_opcode;
unsigned idx;
int err = 0;
int inl = 0;
@@ -2063,7 +2492,6 @@ int mlx5_ib_post_send(struct ib_qp *ibqp, struct ib_send_wr *wr,
int nreq;
int i;
u8 next_fence = 0;
- u8 opmod = 0;
u8 fence;
spin_lock_irqsave(&qp->sq.lock, flags);
@@ -2076,36 +2504,23 @@ int mlx5_ib_post_send(struct ib_qp *ibqp, struct ib_send_wr *wr,
goto out;
}
- if (unlikely(mlx5_wq_overflow(&qp->sq, nreq, qp->ibqp.send_cq))) {
+ fence = qp->fm_cache;
+ num_sge = wr->num_sge;
+ if (unlikely(num_sge > qp->sq.max_gs)) {
mlx5_ib_warn(dev, "\n");
err = -ENOMEM;
*bad_wr = wr;
goto out;
}
- fence = qp->fm_cache;
- num_sge = wr->num_sge;
- if (unlikely(num_sge > qp->sq.max_gs)) {
+ err = begin_wqe(qp, &seg, &ctrl, wr, &idx, &size, nreq);
+ if (err) {
mlx5_ib_warn(dev, "\n");
err = -ENOMEM;
*bad_wr = wr;
goto out;
}
- idx = qp->sq.cur_post & (qp->sq.wqe_cnt - 1);
- seg = mlx5_get_send_wqe(qp, idx);
- ctrl = seg;
- *(uint32_t *)(seg + 8) = 0;
- ctrl->imm = send_ieth(wr);
- ctrl->fm_ce_se = qp->sq_signal_bits |
- (wr->send_flags & IB_SEND_SIGNALED ?
- MLX5_WQE_CTRL_CQ_UPDATE : 0) |
- (wr->send_flags & IB_SEND_SOLICITED ?
- MLX5_WQE_CTRL_SOLICITED : 0);
-
- seg += sizeof(*ctrl);
- size = sizeof(*ctrl) / 16;
-
switch (ibqp->qp_type) {
case IB_QPT_XRC_INI:
xrc = seg;
@@ -2158,6 +2573,73 @@ int mlx5_ib_post_send(struct ib_qp *ibqp, struct ib_send_wr *wr,
num_sge = 0;
break;
+ case IB_WR_REG_SIG_MR:
+ qp->sq.wr_data[idx] = IB_WR_REG_SIG_MR;
+ mr = to_mmr(wr->wr.sig_handover.sig_mr);
+
+ ctrl->imm = cpu_to_be32(mr->ibmr.rkey);
+ err = set_sig_umr_wr(wr, qp, &seg, &size);
+ if (err) {
+ mlx5_ib_warn(dev, "\n");
+ *bad_wr = wr;
+ goto out;
+ }
+
+ finish_wqe(qp, ctrl, size, idx, wr->wr_id,
+ nreq, get_fence(fence, wr),
+ next_fence, MLX5_OPCODE_UMR);
+ /*
+ * SET_PSV WQEs are not signaled and solicited
+ * on error
+ */
+ wr->send_flags &= ~IB_SEND_SIGNALED;
+ wr->send_flags |= IB_SEND_SOLICITED;
+ err = begin_wqe(qp, &seg, &ctrl, wr,
+ &idx, &size, nreq);
+ if (err) {
+ mlx5_ib_warn(dev, "\n");
+ err = -ENOMEM;
+ *bad_wr = wr;
+ goto out;
+ }
+
+ err = set_psv_wr(&wr->wr.sig_handover.sig_attrs->mem,
+ mr->sig->psv_memory.psv_idx, &seg,
+ &size);
+ if (err) {
+ mlx5_ib_warn(dev, "\n");
+ *bad_wr = wr;
+ goto out;
+ }
+
+ finish_wqe(qp, ctrl, size, idx, wr->wr_id,
+ nreq, get_fence(fence, wr),
+ next_fence, MLX5_OPCODE_SET_PSV);
+ err = begin_wqe(qp, &seg, &ctrl, wr,
+ &idx, &size, nreq);
+ if (err) {
+ mlx5_ib_warn(dev, "\n");
+ err = -ENOMEM;
+ *bad_wr = wr;
+ goto out;
+ }
+
+ next_fence = MLX5_FENCE_MODE_INITIATOR_SMALL;
+ err = set_psv_wr(&wr->wr.sig_handover.sig_attrs->wire,
+ mr->sig->psv_wire.psv_idx, &seg,
+ &size);
+ if (err) {
+ mlx5_ib_warn(dev, "\n");
+ *bad_wr = wr;
+ goto out;
+ }
+
+ finish_wqe(qp, ctrl, size, idx, wr->wr_id,
+ nreq, get_fence(fence, wr),
+ next_fence, MLX5_OPCODE_SET_PSV);
+ num_sge = 0;
+ goto skip_psv;
+
default:
break;
}
@@ -2238,22 +2720,10 @@ int mlx5_ib_post_send(struct ib_qp *ibqp, struct ib_send_wr *wr,
}
}
- mlx5_opcode = mlx5_ib_opcode[wr->opcode];
- ctrl->opmod_idx_opcode = cpu_to_be32(((u32)(qp->sq.cur_post) << 8) |
- mlx5_opcode |
- ((u32)opmod << 24));
- ctrl->qpn_ds = cpu_to_be32(size | (qp->mqp.qpn << 8));
- ctrl->fm_ce_se |= get_fence(fence, wr);
- qp->fm_cache = next_fence;
- if (unlikely(qp->wq_sig))
- ctrl->signature = wq_sig(ctrl);
-
- qp->sq.wrid[idx] = wr->wr_id;
- qp->sq.w_list[idx].opcode = mlx5_opcode;
- qp->sq.wqe_head[idx] = qp->sq.head + nreq;
- qp->sq.cur_post += DIV_ROUND_UP(size * 16, MLX5_SEND_WQE_BB);
- qp->sq.w_list[idx].next = qp->sq.cur_post;
-
+ finish_wqe(qp, ctrl, size, idx, wr->wr_id, nreq,
+ get_fence(fence, wr), next_fence,
+ mlx5_ib_opcode[wr->opcode]);
+skip_psv:
if (0)
dump_wqe(qp, idx, size);
}
diff --git a/drivers/infiniband/hw/mthca/mthca_main.c b/drivers/infiniband/hw/mthca/mthca_main.c
index 87897b95666d..ded76c101dde 100644
--- a/drivers/infiniband/hw/mthca/mthca_main.c
+++ b/drivers/infiniband/hw/mthca/mthca_main.c
@@ -858,13 +858,9 @@ static int mthca_enable_msi_x(struct mthca_dev *mdev)
entries[1].entry = 1;
entries[2].entry = 2;
- err = pci_enable_msix(mdev->pdev, entries, ARRAY_SIZE(entries));
- if (err) {
- if (err > 0)
- mthca_info(mdev, "Only %d MSI-X vectors available, "
- "not using MSI-X\n", err);
+ err = pci_enable_msix_exact(mdev->pdev, entries, ARRAY_SIZE(entries));
+ if (err)
return err;
- }
mdev->eq_table.eq[MTHCA_EQ_COMP ].msi_x_vector = entries[0].vector;
mdev->eq_table.eq[MTHCA_EQ_ASYNC].msi_x_vector = entries[1].vector;
diff --git a/drivers/infiniband/hw/mthca/mthca_provider.c b/drivers/infiniband/hw/mthca/mthca_provider.c
index 5b71d43bd89c..415f8e1a54db 100644
--- a/drivers/infiniband/hw/mthca/mthca_provider.c
+++ b/drivers/infiniband/hw/mthca/mthca_provider.c
@@ -695,6 +695,7 @@ static struct ib_cq *mthca_create_cq(struct ib_device *ibdev, int entries,
if (context && ib_copy_to_udata(udata, &cq->cqn, sizeof (__u32))) {
mthca_free_cq(to_mdev(ibdev), cq);
+ err = -EFAULT;
goto err_free;
}
@@ -976,12 +977,12 @@ static struct ib_mr *mthca_reg_user_mr(struct ib_pd *pd, u64 start, u64 length,
u64 virt, int acc, struct ib_udata *udata)
{
struct mthca_dev *dev = to_mdev(pd->device);
- struct ib_umem_chunk *chunk;
+ struct scatterlist *sg;
struct mthca_mr *mr;
struct mthca_reg_mr ucmd;
u64 *pages;
int shift, n, len;
- int i, j, k;
+ int i, k, entry;
int err = 0;
int write_mtt_size;
@@ -1009,10 +1010,7 @@ static struct ib_mr *mthca_reg_user_mr(struct ib_pd *pd, u64 start, u64 length,
}
shift = ffs(mr->umem->page_size) - 1;
-
- n = 0;
- list_for_each_entry(chunk, &mr->umem->chunk_list, list)
- n += chunk->nents;
+ n = mr->umem->nmap;
mr->mtt = mthca_alloc_mtt(dev, n);
if (IS_ERR(mr->mtt)) {
@@ -1030,25 +1028,24 @@ static struct ib_mr *mthca_reg_user_mr(struct ib_pd *pd, u64 start, u64 length,
write_mtt_size = min(mthca_write_mtt_size(dev), (int) (PAGE_SIZE / sizeof *pages));
- list_for_each_entry(chunk, &mr->umem->chunk_list, list)
- for (j = 0; j < chunk->nmap; ++j) {
- len = sg_dma_len(&chunk->page_list[j]) >> shift;
- for (k = 0; k < len; ++k) {
- pages[i++] = sg_dma_address(&chunk->page_list[j]) +
- mr->umem->page_size * k;
- /*
- * Be friendly to write_mtt and pass it chunks
- * of appropriate size.
- */
- if (i == write_mtt_size) {
- err = mthca_write_mtt(dev, mr->mtt, n, pages, i);
- if (err)
- goto mtt_done;
- n += i;
- i = 0;
- }
+ for_each_sg(mr->umem->sg_head.sgl, sg, mr->umem->nmap, entry) {
+ len = sg_dma_len(sg) >> shift;
+ for (k = 0; k < len; ++k) {
+ pages[i++] = sg_dma_address(sg) +
+ mr->umem->page_size * k;
+ /*
+ * Be friendly to write_mtt and pass it chunks
+ * of appropriate size.
+ */
+ if (i == write_mtt_size) {
+ err = mthca_write_mtt(dev, mr->mtt, n, pages, i);
+ if (err)
+ goto mtt_done;
+ n += i;
+ i = 0;
}
}
+ }
if (i)
err = mthca_write_mtt(dev, mr->mtt, n, pages, i);
diff --git a/drivers/infiniband/hw/nes/nes_cm.c b/drivers/infiniband/hw/nes/nes_cm.c
index 9c9f2f57e960..dfa9df484505 100644
--- a/drivers/infiniband/hw/nes/nes_cm.c
+++ b/drivers/infiniband/hw/nes/nes_cm.c
@@ -128,6 +128,7 @@ static void build_mpa_v1(struct nes_cm_node *, void *, u8);
static void build_rdma0_msg(struct nes_cm_node *, struct nes_qp **);
static void print_core(struct nes_cm_core *core);
+static void record_ird_ord(struct nes_cm_node *, u16, u16);
/* External CM API Interface */
/* instance of function pointers for client API */
@@ -317,7 +318,6 @@ static int parse_mpa(struct nes_cm_node *cm_node, u8 *buffer, u32 *type,
}
}
-
if (priv_data_len + mpa_hdr_len != len) {
nes_debug(NES_DBG_CM, "The received ietf buffer was not right"
" complete (%x + %x != %x)\n",
@@ -356,25 +356,57 @@ static int parse_mpa(struct nes_cm_node *cm_node, u8 *buffer, u32 *type,
/* send reset */
return -EINVAL;
}
+ if (ird_size == IETF_NO_IRD_ORD || ord_size == IETF_NO_IRD_ORD)
+ cm_node->mpav2_ird_ord = IETF_NO_IRD_ORD;
- if (cm_node->state != NES_CM_STATE_MPAREQ_SENT) {
+ if (cm_node->mpav2_ird_ord != IETF_NO_IRD_ORD) {
/* responder */
- if (cm_node->ord_size > ird_size)
- cm_node->ord_size = ird_size;
- } else {
- /* initiator */
- if (cm_node->ord_size > ird_size)
- cm_node->ord_size = ird_size;
-
- if (cm_node->ird_size < ord_size) {
- /* no resources available */
- /* send terminate message */
- return -EINVAL;
+ if (cm_node->state != NES_CM_STATE_MPAREQ_SENT) {
+ /* we are still negotiating */
+ if (ord_size > NES_MAX_IRD) {
+ cm_node->ird_size = NES_MAX_IRD;
+ } else {
+ cm_node->ird_size = ord_size;
+ if (ord_size == 0 &&
+ (rtr_ctrl_ord & IETF_RDMA0_READ)) {
+ cm_node->ird_size = 1;
+ nes_debug(NES_DBG_CM,
+ "%s: Remote peer doesn't support RDMA0_READ (ord=%u)\n",
+ __func__, ord_size);
+ }
+ }
+ if (ird_size > NES_MAX_ORD)
+ cm_node->ord_size = NES_MAX_ORD;
+ else
+ cm_node->ord_size = ird_size;
+ } else { /* initiator */
+ if (ord_size > NES_MAX_IRD) {
+ nes_debug(NES_DBG_CM,
+ "%s: Unable to support the requested (ord =%u)\n",
+ __func__, ord_size);
+ return -EINVAL;
+ }
+ cm_node->ird_size = ord_size;
+
+ if (ird_size > NES_MAX_ORD) {
+ cm_node->ord_size = NES_MAX_ORD;
+ } else {
+ if (ird_size == 0 &&
+ (rtr_ctrl_ord & IETF_RDMA0_READ)) {
+ nes_debug(NES_DBG_CM,
+ "%s: Remote peer doesn't support RDMA0_READ (ird=%u)\n",
+ __func__, ird_size);
+ return -EINVAL;
+ } else {
+ cm_node->ord_size = ird_size;
+ }
+ }
}
}
if (rtr_ctrl_ord & IETF_RDMA0_READ) {
cm_node->send_rdma0_op = SEND_RDMA_READ_ZERO;
+
} else if (rtr_ctrl_ord & IETF_RDMA0_WRITE) {
cm_node->send_rdma0_op = SEND_RDMA_WRITE_ZERO;
} else { /* Not supported RDMA0 operation */
@@ -514,6 +546,19 @@ static void print_core(struct nes_cm_core *core)
nes_debug(NES_DBG_CM, "-------------- end core ---------------\n");
}
+static void record_ird_ord(struct nes_cm_node *cm_node,
+ u16 conn_ird, u16 conn_ord)
+{
+ if (conn_ird > NES_MAX_IRD)
+ conn_ird = NES_MAX_IRD;
+
+ if (conn_ord > NES_MAX_ORD)
+ conn_ord = NES_MAX_ORD;
+
+ cm_node->ird_size = conn_ird;
+ cm_node->ord_size = conn_ord;
+}
+
/**
* cm_build_mpa_frame - build a MPA V1 frame or MPA V2 frame
*/
@@ -557,11 +602,13 @@ static void build_mpa_v2(struct nes_cm_node *cm_node,
mpa_frame->priv_data_len += htons(IETF_RTR_MSG_SIZE);
/* initialize RTR msg */
- ctrl_ird = (cm_node->ird_size > IETF_NO_IRD_ORD) ?
- IETF_NO_IRD_ORD : cm_node->ird_size;
- ctrl_ord = (cm_node->ord_size > IETF_NO_IRD_ORD) ?
- IETF_NO_IRD_ORD : cm_node->ord_size;
-
+ if (cm_node->mpav2_ird_ord == IETF_NO_IRD_ORD) {
+ ctrl_ird = IETF_NO_IRD_ORD;
+ ctrl_ord = IETF_NO_IRD_ORD;
+ } else {
+ ctrl_ird = cm_node->ird_size & IETF_NO_IRD_ORD;
+ ctrl_ord = cm_node->ord_size & IETF_NO_IRD_ORD;
+ }
ctrl_ird |= IETF_PEER_TO_PEER;
ctrl_ird |= IETF_FLPDU_ZERO_LEN;
@@ -610,7 +657,7 @@ static void build_rdma0_msg(struct nes_cm_node *cm_node, struct nes_qp **nesqp_a
struct nes_qp *nesqp = *nesqp_addr;
struct nes_hw_qp_wqe *wqe = &nesqp->hwqp.sq_vbase[0];
- u64temp = (unsigned long)nesqp;
+ u64temp = (unsigned long)nesqp->nesuqp_addr;
u64temp |= NES_SW_CONTEXT_ALIGN >> 1;
set_wqe_64bit_value(wqe->wqe_words, NES_IWARP_SQ_WQE_COMP_CTX_LOW_IDX, u64temp);
@@ -1409,8 +1456,9 @@ static struct nes_cm_node *make_cm_node(struct nes_cm_core *cm_core,
cm_node->mpa_frame_rev = mpa_version;
cm_node->send_rdma0_op = SEND_RDMA_READ_ZERO;
- cm_node->ird_size = IETF_NO_IRD_ORD;
- cm_node->ord_size = IETF_NO_IRD_ORD;
+ cm_node->mpav2_ird_ord = 0;
+ cm_node->ird_size = 0;
+ cm_node->ord_size = 0;
nes_debug(NES_DBG_CM, "Make node addresses : loc = %pI4:%x, rem = %pI4:%x\n",
&cm_node->loc_addr, cm_node->loc_port,
@@ -3027,11 +3075,11 @@ int nes_accept(struct iw_cm_id *cm_id, struct iw_cm_conn_param *conn_param)
rem_ref_cm_node(cm_node->cm_core, cm_node);
return -ECONNRESET;
}
-
/* associate the node with the QP */
nesqp->cm_node = (void *)cm_node;
cm_node->nesqp = nesqp;
+
nes_debug(NES_DBG_CM, "QP%u, cm_node=%p, jiffies = %lu listener = %p\n",
nesqp->hwqp.qp_id, cm_node, jiffies, cm_node->listener);
atomic_inc(&cm_accepts);
@@ -3054,6 +3102,11 @@ int nes_accept(struct iw_cm_id *cm_id, struct iw_cm_conn_param *conn_param)
if (cm_node->mpa_frame_rev == IETF_MPA_V1)
mpa_frame_offset = 4;
+ if (cm_node->mpa_frame_rev == IETF_MPA_V1 ||
+ cm_node->mpav2_ird_ord == IETF_NO_IRD_ORD) {
+ record_ird_ord(cm_node, (u16)conn_param->ird, (u16)conn_param->ord);
+ }
+
memcpy(mpa_v2_frame->priv_data, conn_param->private_data,
conn_param->private_data_len);
@@ -3117,7 +3170,6 @@ int nes_accept(struct iw_cm_id *cm_id, struct iw_cm_conn_param *conn_param)
}
nesqp->skip_lsmm = 1;
-
/* Cache the cm_id in the qp */
nesqp->cm_id = cm_id;
cm_node->cm_id = cm_id;
@@ -3154,7 +3206,7 @@ int nes_accept(struct iw_cm_id *cm_id, struct iw_cm_conn_param *conn_param)
nesqp->nesqp_context->ird_ord_sizes |= cpu_to_le32(
((u32)1 << NES_QPCONTEXT_ORDIRD_IWARP_MODE_SHIFT));
nesqp->nesqp_context->ird_ord_sizes |=
- cpu_to_le32((u32)conn_param->ord);
+ cpu_to_le32((u32)cm_node->ord_size);
memset(&nes_quad, 0, sizeof(nes_quad));
nes_quad.DstIpAdrIndex =
@@ -3194,6 +3246,9 @@ int nes_accept(struct iw_cm_id *cm_id, struct iw_cm_conn_param *conn_param)
cm_event.remote_addr = cm_id->remote_addr;
cm_event.private_data = NULL;
cm_event.private_data_len = 0;
+ cm_event.ird = cm_node->ird_size;
+ cm_event.ord = cm_node->ord_size;
+
ret = cm_id->event_handler(cm_id, &cm_event);
attr.qp_state = IB_QPS_RTS;
nes_modify_qp(&nesqp->ibqp, &attr, IB_QP_STATE, NULL);
@@ -3290,14 +3345,8 @@ int nes_connect(struct iw_cm_id *cm_id, struct iw_cm_conn_param *conn_param)
/* cache the cm_id in the qp */
nesqp->cm_id = cm_id;
-
cm_id->provider_data = nesqp;
-
nesqp->private_data_len = conn_param->private_data_len;
- nesqp->nesqp_context->ird_ord_sizes |= cpu_to_le32((u32)conn_param->ord);
- /* space for rdma0 read msg */
- if (conn_param->ord == 0)
- nesqp->nesqp_context->ird_ord_sizes |= cpu_to_le32(1);
nes_debug(NES_DBG_CM, "requested ord = 0x%08X.\n", (u32)conn_param->ord);
nes_debug(NES_DBG_CM, "mpa private data len =%u\n",
@@ -3334,6 +3383,11 @@ int nes_connect(struct iw_cm_id *cm_id, struct iw_cm_conn_param *conn_param)
return -ENOMEM;
}
+ record_ird_ord(cm_node, (u16)conn_param->ird, (u16)conn_param->ord);
+ if (cm_node->send_rdma0_op == SEND_RDMA_READ_ZERO &&
+ cm_node->ord_size == 0)
+ cm_node->ord_size = 1;
+
cm_node->apbvt_set = apbvt_set;
nesqp->cm_node = cm_node;
cm_node->nesqp = nesqp;
@@ -3530,6 +3584,8 @@ static void cm_event_connected(struct nes_cm_event *event)
nesqp->nesqp_context->ird_ord_sizes |=
cpu_to_le32((u32)1 <<
NES_QPCONTEXT_ORDIRD_IWARP_MODE_SHIFT);
+ nesqp->nesqp_context->ird_ord_sizes |=
+ cpu_to_le32((u32)cm_node->ord_size);
/* Adjust tail for not having a LSMM */
/*nesqp->hwqp.sq_tail = 1;*/
@@ -3742,8 +3798,13 @@ static void cm_event_mpa_req(struct nes_cm_event *event)
cm_event_raddr->sin_addr.s_addr = htonl(event->cm_info.rem_addr);
cm_event.private_data = cm_node->mpa_frame_buf;
cm_event.private_data_len = (u8)cm_node->mpa_frame_size;
+ if (cm_node->mpa_frame_rev == IETF_MPA_V1) {
+ cm_event.ird = NES_MAX_IRD;
+ cm_event.ord = NES_MAX_ORD;
+ } else {
cm_event.ird = cm_node->ird_size;
cm_event.ord = cm_node->ord_size;
+ }
ret = cm_id->event_handler(cm_id, &cm_event);
if (ret)
diff --git a/drivers/infiniband/hw/nes/nes_cm.h b/drivers/infiniband/hw/nes/nes_cm.h
index 4646e6666087..522c99cd07c4 100644
--- a/drivers/infiniband/hw/nes/nes_cm.h
+++ b/drivers/infiniband/hw/nes/nes_cm.h
@@ -58,6 +58,8 @@
#define IETF_RDMA0_WRITE 0x8000
#define IETF_RDMA0_READ 0x4000
#define IETF_NO_IRD_ORD 0x3FFF
+#define NES_MAX_IRD 0x40
+#define NES_MAX_ORD 0x7F
enum ietf_mpa_flags {
IETF_MPA_FLAGS_MARKERS = 0x80, /* receive Markers */
@@ -333,6 +335,7 @@ struct nes_cm_node {
enum mpa_frame_version mpa_frame_rev;
u16 ird_size;
u16 ord_size;
+ u16 mpav2_ird_ord;
u16 mpa_frame_size;
struct iw_cm_id *cm_id;
diff --git a/drivers/infiniband/hw/nes/nes_user.h b/drivers/infiniband/hw/nes/nes_user.h
index 4926de744488..529c421bb15c 100644
--- a/drivers/infiniband/hw/nes/nes_user.h
+++ b/drivers/infiniband/hw/nes/nes_user.h
@@ -39,8 +39,8 @@
#include <linux/types.h>
-#define NES_ABI_USERSPACE_VER 1
-#define NES_ABI_KERNEL_VER 1
+#define NES_ABI_USERSPACE_VER 2
+#define NES_ABI_KERNEL_VER 2
/*
* Make sure that all structs defined in this file remain laid out so
@@ -78,6 +78,7 @@ struct nes_create_cq_req {
struct nes_create_qp_req {
__u64 user_wqe_buffers;
+ __u64 user_qp_buffer;
};
enum iwnes_memreg_type {
diff --git a/drivers/infiniband/hw/nes/nes_verbs.c b/drivers/infiniband/hw/nes/nes_verbs.c
index 8308e3634767..218dd3574285 100644
--- a/drivers/infiniband/hw/nes/nes_verbs.c
+++ b/drivers/infiniband/hw/nes/nes_verbs.c
@@ -1186,11 +1186,13 @@ static struct ib_qp *nes_create_qp(struct ib_pd *ibpd,
nes_free_resource(nesadapter, nesadapter->allocated_qps, qp_num);
kfree(nesqp->allocated_buffer);
nes_debug(NES_DBG_QP, "ib_copy_from_udata() Failed \n");
- return NULL;
+ return ERR_PTR(-EFAULT);
}
if (req.user_wqe_buffers) {
virt_wqs = 1;
}
+ if (req.user_qp_buffer)
+ nesqp->nesuqp_addr = req.user_qp_buffer;
if ((ibpd->uobject) && (ibpd->uobject->context)) {
nesqp->user_mode = 1;
nes_ucontext = to_nesucontext(ibpd->uobject->context);
@@ -2307,7 +2309,7 @@ static struct ib_mr *nes_reg_user_mr(struct ib_pd *pd, u64 start, u64 length,
struct nes_device *nesdev = nesvnic->nesdev;
struct nes_adapter *nesadapter = nesdev->nesadapter;
struct ib_mr *ibmr = ERR_PTR(-EINVAL);
- struct ib_umem_chunk *chunk;
+ struct scatterlist *sg;
struct nes_ucontext *nes_ucontext;
struct nes_pbl *nespbl;
struct nes_mr *nesmr;
@@ -2315,7 +2317,7 @@ static struct ib_mr *nes_reg_user_mr(struct ib_pd *pd, u64 start, u64 length,
struct nes_mem_reg_req req;
struct nes_vpbl vpbl;
struct nes_root_vpbl root_vpbl;
- int nmap_index, page_index;
+ int entry, page_index;
int page_count = 0;
int err, pbl_depth = 0;
int chunk_pages;
@@ -2330,6 +2332,7 @@ static struct ib_mr *nes_reg_user_mr(struct ib_pd *pd, u64 start, u64 length,
u16 pbl_count;
u8 single_page = 1;
u8 stag_key;
+ int first_page = 1;
region = ib_umem_get(pd->uobject->context, start, length, acc, 0);
if (IS_ERR(region)) {
@@ -2380,128 +2383,125 @@ static struct ib_mr *nes_reg_user_mr(struct ib_pd *pd, u64 start, u64 length,
}
nesmr->region = region;
- list_for_each_entry(chunk, &region->chunk_list, list) {
- nes_debug(NES_DBG_MR, "Chunk: nents = %u, nmap = %u .\n",
- chunk->nents, chunk->nmap);
- for (nmap_index = 0; nmap_index < chunk->nmap; ++nmap_index) {
- if (sg_dma_address(&chunk->page_list[nmap_index]) & ~PAGE_MASK) {
- ib_umem_release(region);
- nes_free_resource(nesadapter, nesadapter->allocated_mrs, stag_index);
- nes_debug(NES_DBG_MR, "Unaligned Memory Buffer: 0x%x\n",
- (unsigned int) sg_dma_address(&chunk->page_list[nmap_index]));
- ibmr = ERR_PTR(-EINVAL);
- kfree(nesmr);
- goto reg_user_mr_err;
- }
+ for_each_sg(region->sg_head.sgl, sg, region->nmap, entry) {
+ if (sg_dma_address(sg) & ~PAGE_MASK) {
+ ib_umem_release(region);
+ nes_free_resource(nesadapter, nesadapter->allocated_mrs, stag_index);
+ nes_debug(NES_DBG_MR, "Unaligned Memory Buffer: 0x%x\n",
+ (unsigned int) sg_dma_address(sg));
+ ibmr = ERR_PTR(-EINVAL);
+ kfree(nesmr);
+ goto reg_user_mr_err;
+ }
- if (!sg_dma_len(&chunk->page_list[nmap_index])) {
- ib_umem_release(region);
- nes_free_resource(nesadapter, nesadapter->allocated_mrs,
- stag_index);
- nes_debug(NES_DBG_MR, "Invalid Buffer Size\n");
- ibmr = ERR_PTR(-EINVAL);
- kfree(nesmr);
- goto reg_user_mr_err;
- }
+ if (!sg_dma_len(sg)) {
+ ib_umem_release(region);
+ nes_free_resource(nesadapter, nesadapter->allocated_mrs,
+ stag_index);
+ nes_debug(NES_DBG_MR, "Invalid Buffer Size\n");
+ ibmr = ERR_PTR(-EINVAL);
+ kfree(nesmr);
+ goto reg_user_mr_err;
+ }
- region_length += sg_dma_len(&chunk->page_list[nmap_index]);
- chunk_pages = sg_dma_len(&chunk->page_list[nmap_index]) >> 12;
- region_length -= skip_pages << 12;
- for (page_index=skip_pages; page_index < chunk_pages; page_index++) {
- skip_pages = 0;
- if ((page_count!=0)&&(page_count<<12)-(region->offset&(4096-1))>=region->length)
- goto enough_pages;
- if ((page_count&0x01FF) == 0) {
- if (page_count >= 1024 * 512) {
+ region_length += sg_dma_len(sg);
+ chunk_pages = sg_dma_len(sg) >> 12;
+ region_length -= skip_pages << 12;
+ for (page_index = skip_pages; page_index < chunk_pages; page_index++) {
+ skip_pages = 0;
+ if ((page_count != 0) && (page_count<<12)-(region->offset&(4096-1)) >= region->length)
+ goto enough_pages;
+ if ((page_count&0x01FF) == 0) {
+ if (page_count >= 1024 * 512) {
+ ib_umem_release(region);
+ nes_free_resource(nesadapter,
+ nesadapter->allocated_mrs, stag_index);
+ kfree(nesmr);
+ ibmr = ERR_PTR(-E2BIG);
+ goto reg_user_mr_err;
+ }
+ if (root_pbl_index == 1) {
+ root_vpbl.pbl_vbase = pci_alloc_consistent(nesdev->pcidev,
+ 8192, &root_vpbl.pbl_pbase);
+ nes_debug(NES_DBG_MR, "Allocating root PBL, va = %p, pa = 0x%08X\n",
+ root_vpbl.pbl_vbase, (unsigned int)root_vpbl.pbl_pbase);
+ if (!root_vpbl.pbl_vbase) {
ib_umem_release(region);
- nes_free_resource(nesadapter,
- nesadapter->allocated_mrs, stag_index);
+ pci_free_consistent(nesdev->pcidev, 4096, vpbl.pbl_vbase,
+ vpbl.pbl_pbase);
+ nes_free_resource(nesadapter, nesadapter->allocated_mrs,
+ stag_index);
kfree(nesmr);
- ibmr = ERR_PTR(-E2BIG);
+ ibmr = ERR_PTR(-ENOMEM);
goto reg_user_mr_err;
}
- if (root_pbl_index == 1) {
- root_vpbl.pbl_vbase = pci_alloc_consistent(nesdev->pcidev,
- 8192, &root_vpbl.pbl_pbase);
- nes_debug(NES_DBG_MR, "Allocating root PBL, va = %p, pa = 0x%08X\n",
- root_vpbl.pbl_vbase, (unsigned int)root_vpbl.pbl_pbase);
- if (!root_vpbl.pbl_vbase) {
- ib_umem_release(region);
- pci_free_consistent(nesdev->pcidev, 4096, vpbl.pbl_vbase,
- vpbl.pbl_pbase);
- nes_free_resource(nesadapter, nesadapter->allocated_mrs,
- stag_index);
- kfree(nesmr);
- ibmr = ERR_PTR(-ENOMEM);
- goto reg_user_mr_err;
- }
- root_vpbl.leaf_vpbl = kzalloc(sizeof(*root_vpbl.leaf_vpbl)*1024,
- GFP_KERNEL);
- if (!root_vpbl.leaf_vpbl) {
- ib_umem_release(region);
- pci_free_consistent(nesdev->pcidev, 8192, root_vpbl.pbl_vbase,
- root_vpbl.pbl_pbase);
- pci_free_consistent(nesdev->pcidev, 4096, vpbl.pbl_vbase,
- vpbl.pbl_pbase);
- nes_free_resource(nesadapter, nesadapter->allocated_mrs,
- stag_index);
- kfree(nesmr);
- ibmr = ERR_PTR(-ENOMEM);
- goto reg_user_mr_err;
- }
- root_vpbl.pbl_vbase[0].pa_low =
- cpu_to_le32((u32)vpbl.pbl_pbase);
- root_vpbl.pbl_vbase[0].pa_high =
- cpu_to_le32((u32)((((u64)vpbl.pbl_pbase) >> 32)));
- root_vpbl.leaf_vpbl[0] = vpbl;
- }
- vpbl.pbl_vbase = pci_alloc_consistent(nesdev->pcidev, 4096,
- &vpbl.pbl_pbase);
- nes_debug(NES_DBG_MR, "Allocating leaf PBL, va = %p, pa = 0x%08X\n",
- vpbl.pbl_vbase, (unsigned int)vpbl.pbl_pbase);
- if (!vpbl.pbl_vbase) {
+ root_vpbl.leaf_vpbl = kzalloc(sizeof(*root_vpbl.leaf_vpbl)*1024,
+ GFP_KERNEL);
+ if (!root_vpbl.leaf_vpbl) {
ib_umem_release(region);
- nes_free_resource(nesadapter, nesadapter->allocated_mrs, stag_index);
- ibmr = ERR_PTR(-ENOMEM);
+ pci_free_consistent(nesdev->pcidev, 8192, root_vpbl.pbl_vbase,
+ root_vpbl.pbl_pbase);
+ pci_free_consistent(nesdev->pcidev, 4096, vpbl.pbl_vbase,
+ vpbl.pbl_pbase);
+ nes_free_resource(nesadapter, nesadapter->allocated_mrs,
+ stag_index);
kfree(nesmr);
+ ibmr = ERR_PTR(-ENOMEM);
goto reg_user_mr_err;
}
- if (1 <= root_pbl_index) {
- root_vpbl.pbl_vbase[root_pbl_index].pa_low =
- cpu_to_le32((u32)vpbl.pbl_pbase);
- root_vpbl.pbl_vbase[root_pbl_index].pa_high =
- cpu_to_le32((u32)((((u64)vpbl.pbl_pbase)>>32)));
- root_vpbl.leaf_vpbl[root_pbl_index] = vpbl;
- }
- root_pbl_index++;
- cur_pbl_index = 0;
+ root_vpbl.pbl_vbase[0].pa_low =
+ cpu_to_le32((u32)vpbl.pbl_pbase);
+ root_vpbl.pbl_vbase[0].pa_high =
+ cpu_to_le32((u32)((((u64)vpbl.pbl_pbase) >> 32)));
+ root_vpbl.leaf_vpbl[0] = vpbl;
}
- if (single_page) {
- if (page_count != 0) {
- if ((last_dma_addr+4096) !=
- (sg_dma_address(&chunk->page_list[nmap_index])+
- (page_index*4096)))
- single_page = 0;
- last_dma_addr = sg_dma_address(&chunk->page_list[nmap_index])+
- (page_index*4096);
- } else {
- first_dma_addr = sg_dma_address(&chunk->page_list[nmap_index])+
- (page_index*4096);
- last_dma_addr = first_dma_addr;
- }
+ vpbl.pbl_vbase = pci_alloc_consistent(nesdev->pcidev, 4096,
+ &vpbl.pbl_pbase);
+ nes_debug(NES_DBG_MR, "Allocating leaf PBL, va = %p, pa = 0x%08X\n",
+ vpbl.pbl_vbase, (unsigned int)vpbl.pbl_pbase);
+ if (!vpbl.pbl_vbase) {
+ ib_umem_release(region);
+ nes_free_resource(nesadapter, nesadapter->allocated_mrs, stag_index);
+ ibmr = ERR_PTR(-ENOMEM);
+ kfree(nesmr);
+ goto reg_user_mr_err;
+ }
+ if (1 <= root_pbl_index) {
+ root_vpbl.pbl_vbase[root_pbl_index].pa_low =
+ cpu_to_le32((u32)vpbl.pbl_pbase);
+ root_vpbl.pbl_vbase[root_pbl_index].pa_high =
+ cpu_to_le32((u32)((((u64)vpbl.pbl_pbase)>>32)));
+ root_vpbl.leaf_vpbl[root_pbl_index] = vpbl;
+ }
+ root_pbl_index++;
+ cur_pbl_index = 0;
+ }
+ if (single_page) {
+ if (page_count != 0) {
+ if ((last_dma_addr+4096) !=
+ (sg_dma_address(sg)+
+ (page_index*4096)))
+ single_page = 0;
+ last_dma_addr = sg_dma_address(sg)+
+ (page_index*4096);
+ } else {
+ first_dma_addr = sg_dma_address(sg)+
+ (page_index*4096);
+ last_dma_addr = first_dma_addr;
}
-
- vpbl.pbl_vbase[cur_pbl_index].pa_low =
- cpu_to_le32((u32)(sg_dma_address(&chunk->page_list[nmap_index])+
- (page_index*4096)));
- vpbl.pbl_vbase[cur_pbl_index].pa_high =
- cpu_to_le32((u32)((((u64)(sg_dma_address(&chunk->page_list[nmap_index])+
- (page_index*4096))) >> 32)));
- cur_pbl_index++;
- page_count++;
}
+
+ vpbl.pbl_vbase[cur_pbl_index].pa_low =
+ cpu_to_le32((u32)(sg_dma_address(sg)+
+ (page_index*4096)));
+ vpbl.pbl_vbase[cur_pbl_index].pa_high =
+ cpu_to_le32((u32)((((u64)(sg_dma_address(sg)+
+ (page_index*4096))) >> 32)));
+ cur_pbl_index++;
+ page_count++;
}
}
+
enough_pages:
nes_debug(NES_DBG_MR, "calculating stag, stag_index=0x%08x, driver_key=0x%08x,"
" stag_key=0x%08x\n",
@@ -2613,25 +2613,28 @@ static struct ib_mr *nes_reg_user_mr(struct ib_pd *pd, u64 start, u64 length,
nespbl->pbl_size, (unsigned long) nespbl->pbl_pbase,
(void *) nespbl->pbl_vbase, nespbl->user_base);
- list_for_each_entry(chunk, &region->chunk_list, list) {
- for (nmap_index = 0; nmap_index < chunk->nmap; ++nmap_index) {
- chunk_pages = sg_dma_len(&chunk->page_list[nmap_index]) >> 12;
- chunk_pages += (sg_dma_len(&chunk->page_list[nmap_index]) & (4096-1)) ? 1 : 0;
- nespbl->page = sg_page(&chunk->page_list[0]);
- for (page_index=0; page_index<chunk_pages; page_index++) {
- ((__le32 *)pbl)[0] = cpu_to_le32((u32)
- (sg_dma_address(&chunk->page_list[nmap_index])+
- (page_index*4096)));
- ((__le32 *)pbl)[1] = cpu_to_le32(((u64)
- (sg_dma_address(&chunk->page_list[nmap_index])+
- (page_index*4096)))>>32);
- nes_debug(NES_DBG_MR, "pbl=%p, *pbl=0x%016llx, 0x%08x%08x\n", pbl,
- (unsigned long long)*pbl,
- le32_to_cpu(((__le32 *)pbl)[1]), le32_to_cpu(((__le32 *)pbl)[0]));
- pbl++;
- }
+ for_each_sg(region->sg_head.sgl, sg, region->nmap, entry) {
+ chunk_pages = sg_dma_len(sg) >> 12;
+ chunk_pages += (sg_dma_len(sg) & (4096-1)) ? 1 : 0;
+ if (first_page) {
+ nespbl->page = sg_page(sg);
+ first_page = 0;
+ }
+
+ for (page_index = 0; page_index < chunk_pages; page_index++) {
+ ((__le32 *)pbl)[0] = cpu_to_le32((u32)
+ (sg_dma_address(sg)+
+ (page_index*4096)));
+ ((__le32 *)pbl)[1] = cpu_to_le32(((u64)
+ (sg_dma_address(sg)+
+ (page_index*4096)))>>32);
+ nes_debug(NES_DBG_MR, "pbl=%p, *pbl=0x%016llx, 0x%08x%08x\n", pbl,
+ (unsigned long long)*pbl,
+ le32_to_cpu(((__le32 *)pbl)[1]), le32_to_cpu(((__le32 *)pbl)[0]));
+ pbl++;
}
}
+
if (req.reg_type == IWNES_MEMREG_TYPE_QP) {
list_add_tail(&nespbl->list, &nes_ucontext->qp_reg_mem_list);
} else {
@@ -3134,9 +3137,7 @@ int nes_modify_qp(struct ib_qp *ibqp, struct ib_qp_attr *attr,
" original_last_aeq = 0x%04X. last_aeq = 0x%04X.\n",
nesqp->hwqp.qp_id, atomic_read(&nesqp->refcount),
original_last_aeq, nesqp->last_aeq);
- if ((!ret) ||
- ((original_last_aeq != NES_AEQE_AEID_RDMAP_ROE_BAD_LLP_CLOSE) &&
- (ret))) {
+ if (!ret || original_last_aeq != NES_AEQE_AEID_RDMAP_ROE_BAD_LLP_CLOSE) {
if (dont_wait) {
if (nesqp->cm_id && nesqp->hw_tcp_state != 0) {
nes_debug(NES_DBG_MOD_QP, "QP%u Queuing fake disconnect for QP refcount (%d),"
diff --git a/drivers/infiniband/hw/nes/nes_verbs.h b/drivers/infiniband/hw/nes/nes_verbs.h
index 0eff7c44d76b..309b31c31ae1 100644
--- a/drivers/infiniband/hw/nes/nes_verbs.h
+++ b/drivers/infiniband/hw/nes/nes_verbs.h
@@ -184,5 +184,6 @@ struct nes_qp {
u8 pau_busy;
u8 pau_pending;
u8 pau_state;
+ __u64 nesuqp_addr;
};
#endif /* NES_VERBS_H */
diff --git a/drivers/infiniband/hw/ocrdma/Makefile b/drivers/infiniband/hw/ocrdma/Makefile
index 06a5bed12e43..d1bfd4f4cdde 100644
--- a/drivers/infiniband/hw/ocrdma/Makefile
+++ b/drivers/infiniband/hw/ocrdma/Makefile
@@ -2,4 +2,4 @@ ccflags-y := -Idrivers/net/ethernet/emulex/benet
obj-$(CONFIG_INFINIBAND_OCRDMA) += ocrdma.o
-ocrdma-y := ocrdma_main.o ocrdma_verbs.o ocrdma_hw.o ocrdma_ah.o
+ocrdma-y := ocrdma_main.o ocrdma_verbs.o ocrdma_hw.o ocrdma_ah.o ocrdma_stats.o
diff --git a/drivers/infiniband/hw/ocrdma/ocrdma.h b/drivers/infiniband/hw/ocrdma/ocrdma.h
index 7c001b97b23f..19011dbb930f 100644
--- a/drivers/infiniband/hw/ocrdma/ocrdma.h
+++ b/drivers/infiniband/hw/ocrdma/ocrdma.h
@@ -35,17 +35,27 @@
#include <rdma/ib_verbs.h>
#include <rdma/ib_user_verbs.h>
+#include <rdma/ib_addr.h>
#include <be_roce.h>
#include "ocrdma_sli.h"
-#define OCRDMA_ROCE_DEV_VERSION "1.0.0"
+#define OCRDMA_ROCE_DRV_VERSION "10.2.145.0u"
+
+#define OCRDMA_ROCE_DRV_DESC "Emulex OneConnect RoCE Driver"
#define OCRDMA_NODE_DESC "Emulex OneConnect RoCE HCA"
+#define OC_NAME_SH OCRDMA_NODE_DESC "(Skyhawk)"
+#define OC_NAME_UNKNOWN OCRDMA_NODE_DESC "(Unknown)"
+
+#define OC_SKH_DEVICE_PF 0x720
+#define OC_SKH_DEVICE_VF 0x728
#define OCRDMA_MAX_AH 512
#define OCRDMA_UVERBS(CMD_NAME) (1ull << IB_USER_VERBS_CMD_##CMD_NAME)
+#define convert_to_64bit(lo, hi) ((u64)hi << 32 | (u64)lo)
+
struct ocrdma_dev_attr {
u8 fw_ver[32];
u32 vendor_id;
@@ -65,6 +75,7 @@ struct ocrdma_dev_attr {
int max_mr;
u64 max_mr_size;
u32 max_num_mr_pbl;
+ int max_mw;
int max_fmr;
int max_map_per_fmr;
int max_pages_per_frmr;
@@ -83,6 +94,12 @@ struct ocrdma_dev_attr {
u8 num_ird_pages;
};
+struct ocrdma_dma_mem {
+ void *va;
+ dma_addr_t pa;
+ u32 size;
+};
+
struct ocrdma_pbl {
void *va;
dma_addr_t pa;
@@ -148,6 +165,26 @@ struct ocrdma_mr {
struct ocrdma_hw_mr hwmr;
};
+struct ocrdma_stats {
+ u8 type;
+ struct ocrdma_dev *dev;
+};
+
+struct stats_mem {
+ struct ocrdma_mqe mqe;
+ void *va;
+ dma_addr_t pa;
+ u32 size;
+ char *debugfs_mem;
+};
+
+struct phy_info {
+ u16 auto_speeds_supported;
+ u16 fixed_speeds_supported;
+ u16 phy_type;
+ u16 interface_type;
+};
+
struct ocrdma_dev {
struct ib_device ibdev;
struct ocrdma_dev_attr attr;
@@ -191,12 +228,30 @@ struct ocrdma_dev {
struct mqe_ctx mqe_ctx;
struct be_dev_info nic_info;
+ struct phy_info phy;
+ char model_number[32];
+ u32 hba_port_num;
struct list_head entry;
struct rcu_head rcu;
int id;
- struct ocrdma_mr *stag_arr[OCRDMA_MAX_STAG];
+ u64 stag_arr[OCRDMA_MAX_STAG];
u16 pvid;
+ u32 asic_id;
+
+ ulong last_stats_time;
+ struct mutex stats_lock; /* provide synch for debugfs operations */
+ struct stats_mem stats_mem;
+ struct ocrdma_stats rsrc_stats;
+ struct ocrdma_stats rx_stats;
+ struct ocrdma_stats wqe_stats;
+ struct ocrdma_stats tx_stats;
+ struct ocrdma_stats db_err_stats;
+ struct ocrdma_stats tx_qp_err_stats;
+ struct ocrdma_stats rx_qp_err_stats;
+ struct ocrdma_stats tx_dbg_stats;
+ struct ocrdma_stats rx_dbg_stats;
+ struct dentry *dir;
};
struct ocrdma_cq {
@@ -209,8 +264,8 @@ struct ocrdma_cq {
*/
u32 max_hw_cqe;
bool phase_change;
- bool armed, solicited;
- bool arm_needed;
+ bool deferred_arm, deferred_sol;
+ bool first_arm;
spinlock_t cq_lock ____cacheline_aligned; /* provide synchronization
* to cq polling
@@ -223,6 +278,7 @@ struct ocrdma_cq {
struct ocrdma_ucontext *ucontext;
dma_addr_t pa;
u32 len;
+ u32 cqe_cnt;
/* head of all qp's sq and rq for which cqes need to be flushed
* by the software.
@@ -232,7 +288,6 @@ struct ocrdma_cq {
struct ocrdma_pd {
struct ib_pd ibpd;
- struct ocrdma_dev *dev;
struct ocrdma_ucontext *uctx;
u32 id;
int num_dpp_qp;
@@ -317,10 +372,8 @@ struct ocrdma_qp {
bool dpp_enabled;
u8 *ird_q_va;
bool signaled;
- u16 db_cache;
};
-
struct ocrdma_ucontext {
struct ib_ucontext ibucontext;
@@ -385,13 +438,6 @@ static inline struct ocrdma_srq *get_ocrdma_srq(struct ib_srq *ibsrq)
return container_of(ibsrq, struct ocrdma_srq, ibsrq);
}
-
-static inline int ocrdma_get_num_posted_shift(struct ocrdma_qp *qp)
-{
- return ((qp->dev->nic_info.dev_family == OCRDMA_GEN2_FAMILY &&
- qp->id < 128) ? 24 : 16);
-}
-
static inline int is_cqe_valid(struct ocrdma_cq *cq, struct ocrdma_cqe *cqe)
{
int cqe_valid;
@@ -436,4 +482,40 @@ static inline int ocrdma_resolve_dmac(struct ocrdma_dev *dev,
return 0;
}
+static inline char *hca_name(struct ocrdma_dev *dev)
+{
+ switch (dev->nic_info.pdev->device) {
+ case OC_SKH_DEVICE_PF:
+ case OC_SKH_DEVICE_VF:
+ return OC_NAME_SH;
+ default:
+ return OC_NAME_UNKNOWN;
+ }
+}
+
+static inline int ocrdma_get_eq_table_index(struct ocrdma_dev *dev,
+ int eqid)
+{
+ int indx;
+
+ for (indx = 0; indx < dev->eq_cnt; indx++) {
+ if (dev->eq_tbl[indx].q.id == eqid)
+ return indx;
+ }
+
+ return -EINVAL;
+}
+
+static inline u8 ocrdma_get_asic_type(struct ocrdma_dev *dev)
+{
+ if (dev->nic_info.dev_family == 0xF && !dev->asic_id) {
+ pci_read_config_dword(
+ dev->nic_info.pdev,
+ OCRDMA_SLI_ASIC_ID_OFFSET, &dev->asic_id);
+ }
+
+ return (dev->asic_id & OCRDMA_SLI_ASIC_GEN_NUM_MASK) >>
+ OCRDMA_SLI_ASIC_GEN_NUM_SHIFT;
+}
+
#endif
diff --git a/drivers/infiniband/hw/ocrdma/ocrdma_abi.h b/drivers/infiniband/hw/ocrdma/ocrdma_abi.h
index fbac8eb44036..1554cca5712a 100644
--- a/drivers/infiniband/hw/ocrdma/ocrdma_abi.h
+++ b/drivers/infiniband/hw/ocrdma/ocrdma_abi.h
@@ -28,7 +28,8 @@
#ifndef __OCRDMA_ABI_H__
#define __OCRDMA_ABI_H__
-#define OCRDMA_ABI_VERSION 1
+#define OCRDMA_ABI_VERSION 2
+#define OCRDMA_BE_ROCE_ABI_VERSION 1
/* user kernel communication data structures. */
struct ocrdma_alloc_ucontext_resp {
@@ -107,9 +108,7 @@ struct ocrdma_create_qp_uresp {
u32 db_sq_offset;
u32 db_rq_offset;
u32 db_shift;
- u64 rsvd1;
- u64 rsvd2;
- u64 rsvd3;
+ u64 rsvd[11];
} __packed;
struct ocrdma_create_srq_uresp {
diff --git a/drivers/infiniband/hw/ocrdma/ocrdma_ah.c b/drivers/infiniband/hw/ocrdma/ocrdma_ah.c
index 34071143006e..d4cc01f10c01 100644
--- a/drivers/infiniband/hw/ocrdma/ocrdma_ah.c
+++ b/drivers/infiniband/hw/ocrdma/ocrdma_ah.c
@@ -100,7 +100,7 @@ struct ib_ah *ocrdma_create_ah(struct ib_pd *ibpd, struct ib_ah_attr *attr)
if (!(attr->ah_flags & IB_AH_GRH))
return ERR_PTR(-EINVAL);
- ah = kzalloc(sizeof *ah, GFP_ATOMIC);
+ ah = kzalloc(sizeof(*ah), GFP_ATOMIC);
if (!ah)
return ERR_PTR(-ENOMEM);
diff --git a/drivers/infiniband/hw/ocrdma/ocrdma_hw.c b/drivers/infiniband/hw/ocrdma/ocrdma_hw.c
index 1664d648cbfc..3bbf2010a821 100644
--- a/drivers/infiniband/hw/ocrdma/ocrdma_hw.c
+++ b/drivers/infiniband/hw/ocrdma/ocrdma_hw.c
@@ -32,7 +32,6 @@
#include <rdma/ib_verbs.h>
#include <rdma/ib_user_verbs.h>
-#include <rdma/ib_addr.h>
#include "ocrdma.h"
#include "ocrdma_hw.h"
@@ -243,6 +242,23 @@ static int ocrdma_get_mbx_errno(u32 status)
return err_num;
}
+char *port_speed_string(struct ocrdma_dev *dev)
+{
+ char *str = "";
+ u16 speeds_supported;
+
+ speeds_supported = dev->phy.fixed_speeds_supported |
+ dev->phy.auto_speeds_supported;
+ if (speeds_supported & OCRDMA_PHY_SPEED_40GBPS)
+ str = "40Gbps ";
+ else if (speeds_supported & OCRDMA_PHY_SPEED_10GBPS)
+ str = "10Gbps ";
+ else if (speeds_supported & OCRDMA_PHY_SPEED_1GBPS)
+ str = "1Gbps ";
+
+ return str;
+}
+
static int ocrdma_get_mbx_cqe_errno(u16 cqe_status)
{
int err_num = -EINVAL;
@@ -332,6 +348,11 @@ static void *ocrdma_init_emb_mqe(u8 opcode, u32 cmd_len)
return mqe;
}
+static void *ocrdma_alloc_mqe(void)
+{
+ return kzalloc(sizeof(struct ocrdma_mqe), GFP_KERNEL);
+}
+
static void ocrdma_free_q(struct ocrdma_dev *dev, struct ocrdma_queue_info *q)
{
dma_free_coherent(&dev->nic_info.pdev->dev, q->size, q->va, q->dma);
@@ -364,8 +385,8 @@ static void ocrdma_build_q_pages(struct ocrdma_pa *q_pa, int cnt,
}
}
-static int ocrdma_mbx_delete_q(struct ocrdma_dev *dev, struct ocrdma_queue_info *q,
- int queue_type)
+static int ocrdma_mbx_delete_q(struct ocrdma_dev *dev,
+ struct ocrdma_queue_info *q, int queue_type)
{
u8 opcode = 0;
int status;
@@ -444,7 +465,7 @@ mbx_err:
return status;
}
-static int ocrdma_get_irq(struct ocrdma_dev *dev, struct ocrdma_eq *eq)
+int ocrdma_get_irq(struct ocrdma_dev *dev, struct ocrdma_eq *eq)
{
int irq;
@@ -574,6 +595,7 @@ static int ocrdma_create_mq(struct ocrdma_dev *dev)
if (status)
goto alloc_err;
+ dev->eq_tbl[0].cq_cnt++;
status = ocrdma_mbx_mq_cq_create(dev, &dev->mq.cq, &dev->eq_tbl[0].q);
if (status)
goto mbx_cq_free;
@@ -639,7 +661,7 @@ static void ocrdma_dispatch_ibevent(struct ocrdma_dev *dev,
{
struct ocrdma_qp *qp = NULL;
struct ocrdma_cq *cq = NULL;
- struct ib_event ib_evt;
+ struct ib_event ib_evt = { 0 };
int cq_event = 0;
int qp_event = 1;
int srq_event = 0;
@@ -664,6 +686,8 @@ static void ocrdma_dispatch_ibevent(struct ocrdma_dev *dev,
case OCRDMA_CQ_OVERRUN_ERROR:
ib_evt.element.cq = &cq->ibcq;
ib_evt.event = IB_EVENT_CQ_ERR;
+ cq_event = 1;
+ qp_event = 0;
break;
case OCRDMA_CQ_QPCAT_ERROR:
ib_evt.element.qp = &qp->ibqp;
@@ -725,6 +749,7 @@ static void ocrdma_dispatch_ibevent(struct ocrdma_dev *dev,
qp->srq->ibsrq.
srq_context);
} else if (dev_event) {
+ pr_err("%s: Fatal event received\n", dev->ibdev.name);
ib_dispatch_event(&ib_evt);
}
@@ -752,7 +777,6 @@ static void ocrdma_process_grp5_aync(struct ocrdma_dev *dev,
}
}
-
static void ocrdma_process_acqe(struct ocrdma_dev *dev, void *ae_cqe)
{
/* async CQE processing */
@@ -799,8 +823,6 @@ static int ocrdma_mq_cq_handler(struct ocrdma_dev *dev, u16 cq_id)
ocrdma_process_acqe(dev, cqe);
else if (cqe->valid_ae_cmpl_cons & OCRDMA_MCQE_CMPL_MASK)
ocrdma_process_mcqe(dev, cqe);
- else
- pr_err("%s() cqe->compl is not set.\n", __func__);
memset(cqe, 0, sizeof(struct ocrdma_mcqe));
ocrdma_mcq_inc_tail(dev);
}
@@ -858,16 +880,8 @@ static void ocrdma_qp_cq_handler(struct ocrdma_dev *dev, u16 cq_idx)
BUG();
cq = dev->cq_tbl[cq_idx];
- if (cq == NULL) {
- pr_err("%s%d invalid id=0x%x\n", __func__, dev->id, cq_idx);
+ if (cq == NULL)
return;
- }
- spin_lock_irqsave(&cq->cq_lock, flags);
- cq->armed = false;
- cq->solicited = false;
- spin_unlock_irqrestore(&cq->cq_lock, flags);
-
- ocrdma_ring_cq_db(dev, cq->id, false, false, 0);
if (cq->ibcq.comp_handler) {
spin_lock_irqsave(&cq->comp_handler_lock, flags);
@@ -892,27 +906,35 @@ static irqreturn_t ocrdma_irq_handler(int irq, void *handle)
struct ocrdma_dev *dev = eq->dev;
struct ocrdma_eqe eqe;
struct ocrdma_eqe *ptr;
- u16 eqe_popped = 0;
u16 cq_id;
- while (1) {
+ int budget = eq->cq_cnt;
+
+ do {
ptr = ocrdma_get_eqe(eq);
eqe = *ptr;
ocrdma_le32_to_cpu(&eqe, sizeof(eqe));
if ((eqe.id_valid & OCRDMA_EQE_VALID_MASK) == 0)
break;
- eqe_popped += 1;
+
ptr->id_valid = 0;
+ /* ring eq doorbell as soon as its consumed. */
+ ocrdma_ring_eq_db(dev, eq->q.id, false, true, 1);
/* check whether its CQE or not. */
if ((eqe.id_valid & OCRDMA_EQE_FOR_CQE_MASK) == 0) {
cq_id = eqe.id_valid >> OCRDMA_EQE_RESOURCE_ID_SHIFT;
ocrdma_cq_handler(dev, cq_id);
}
ocrdma_eq_inc_tail(eq);
- }
- ocrdma_ring_eq_db(dev, eq->q.id, true, true, eqe_popped);
- /* Ring EQ doorbell with num_popped to 0 to enable interrupts again. */
- if (dev->nic_info.intr_mode == BE_INTERRUPT_MODE_INTX)
- ocrdma_ring_eq_db(dev, eq->q.id, true, true, 0);
+
+ /* There can be a stale EQE after the last bound CQ is
+ * destroyed. EQE valid and budget == 0 implies this.
+ */
+ if (budget)
+ budget--;
+
+ } while (budget);
+
+ ocrdma_ring_eq_db(dev, eq->q.id, true, true, 0);
return IRQ_HANDLED;
}
@@ -949,7 +971,8 @@ static int ocrdma_mbx_cmd(struct ocrdma_dev *dev, struct ocrdma_mqe *mqe)
{
int status = 0;
u16 cqe_status, ext_status;
- struct ocrdma_mqe *rsp;
+ struct ocrdma_mqe *rsp_mqe;
+ struct ocrdma_mbx_rsp *rsp = NULL;
mutex_lock(&dev->mqe_ctx.lock);
ocrdma_post_mqe(dev, mqe);
@@ -958,23 +981,61 @@ static int ocrdma_mbx_cmd(struct ocrdma_dev *dev, struct ocrdma_mqe *mqe)
goto mbx_err;
cqe_status = dev->mqe_ctx.cqe_status;
ext_status = dev->mqe_ctx.ext_status;
- rsp = ocrdma_get_mqe_rsp(dev);
- ocrdma_copy_le32_to_cpu(mqe, rsp, (sizeof(*mqe)));
+ rsp_mqe = ocrdma_get_mqe_rsp(dev);
+ ocrdma_copy_le32_to_cpu(mqe, rsp_mqe, (sizeof(*mqe)));
+ if ((mqe->hdr.spcl_sge_cnt_emb & OCRDMA_MQE_HDR_EMB_MASK) >>
+ OCRDMA_MQE_HDR_EMB_SHIFT)
+ rsp = &mqe->u.rsp;
+
if (cqe_status || ext_status) {
- pr_err("%s() opcode=0x%x, cqe_status=0x%x, ext_status=0x%x\n",
- __func__,
- (rsp->u.rsp.subsys_op & OCRDMA_MBX_RSP_OPCODE_MASK) >>
- OCRDMA_MBX_RSP_OPCODE_SHIFT, cqe_status, ext_status);
+ pr_err("%s() cqe_status=0x%x, ext_status=0x%x,",
+ __func__, cqe_status, ext_status);
+ if (rsp) {
+ /* This is for embedded cmds. */
+ pr_err("opcode=0x%x, subsystem=0x%x\n",
+ (rsp->subsys_op & OCRDMA_MBX_RSP_OPCODE_MASK) >>
+ OCRDMA_MBX_RSP_OPCODE_SHIFT,
+ (rsp->subsys_op & OCRDMA_MBX_RSP_SUBSYS_MASK) >>
+ OCRDMA_MBX_RSP_SUBSYS_SHIFT);
+ }
status = ocrdma_get_mbx_cqe_errno(cqe_status);
goto mbx_err;
}
- if (mqe->u.rsp.status & OCRDMA_MBX_RSP_STATUS_MASK)
+ /* For non embedded, rsp errors are handled in ocrdma_nonemb_mbx_cmd */
+ if (rsp && (mqe->u.rsp.status & OCRDMA_MBX_RSP_STATUS_MASK))
status = ocrdma_get_mbx_errno(mqe->u.rsp.status);
mbx_err:
mutex_unlock(&dev->mqe_ctx.lock);
return status;
}
+static int ocrdma_nonemb_mbx_cmd(struct ocrdma_dev *dev, struct ocrdma_mqe *mqe,
+ void *payload_va)
+{
+ int status = 0;
+ struct ocrdma_mbx_rsp *rsp = payload_va;
+
+ if ((mqe->hdr.spcl_sge_cnt_emb & OCRDMA_MQE_HDR_EMB_MASK) >>
+ OCRDMA_MQE_HDR_EMB_SHIFT)
+ BUG();
+
+ status = ocrdma_mbx_cmd(dev, mqe);
+ if (!status)
+ /* For non embedded, only CQE failures are handled in
+ * ocrdma_mbx_cmd. We need to check for RSP errors.
+ */
+ if (rsp->status & OCRDMA_MBX_RSP_STATUS_MASK)
+ status = ocrdma_get_mbx_errno(rsp->status);
+
+ if (status)
+ pr_err("opcode=0x%x, subsystem=0x%x\n",
+ (rsp->subsys_op & OCRDMA_MBX_RSP_OPCODE_MASK) >>
+ OCRDMA_MBX_RSP_OPCODE_SHIFT,
+ (rsp->subsys_op & OCRDMA_MBX_RSP_SUBSYS_MASK) >>
+ OCRDMA_MBX_RSP_SUBSYS_SHIFT);
+ return status;
+}
+
static void ocrdma_get_attr(struct ocrdma_dev *dev,
struct ocrdma_dev_attr *attr,
struct ocrdma_mbx_query_config *rsp)
@@ -985,6 +1046,9 @@ static void ocrdma_get_attr(struct ocrdma_dev *dev,
attr->max_qp =
(rsp->qp_srq_cq_ird_ord & OCRDMA_MBX_QUERY_CFG_MAX_QP_MASK) >>
OCRDMA_MBX_QUERY_CFG_MAX_QP_SHIFT;
+ attr->max_srq =
+ (rsp->max_srq_rpir_qps & OCRDMA_MBX_QUERY_CFG_MAX_SRQ_MASK) >>
+ OCRDMA_MBX_QUERY_CFG_MAX_SRQ_OFFSET;
attr->max_send_sge = ((rsp->max_write_send_sge &
OCRDMA_MBX_QUERY_CFG_MAX_SEND_SGE_MASK) >>
OCRDMA_MBX_QUERY_CFG_MAX_SEND_SGE_SHIFT);
@@ -1000,9 +1064,6 @@ static void ocrdma_get_attr(struct ocrdma_dev *dev,
attr->max_ord_per_qp = (rsp->max_ird_ord_per_qp &
OCRDMA_MBX_QUERY_CFG_MAX_ORD_PER_QP_MASK) >>
OCRDMA_MBX_QUERY_CFG_MAX_ORD_PER_QP_SHIFT;
- attr->max_srq =
- (rsp->max_srq_rpir_qps & OCRDMA_MBX_QUERY_CFG_MAX_SRQ_MASK) >>
- OCRDMA_MBX_QUERY_CFG_MAX_SRQ_OFFSET;
attr->max_ird_per_qp = (rsp->max_ird_ord_per_qp &
OCRDMA_MBX_QUERY_CFG_MAX_IRD_PER_QP_MASK) >>
OCRDMA_MBX_QUERY_CFG_MAX_IRD_PER_QP_SHIFT;
@@ -1015,6 +1076,7 @@ static void ocrdma_get_attr(struct ocrdma_dev *dev,
attr->local_ca_ack_delay = (rsp->max_pd_ca_ack_delay &
OCRDMA_MBX_QUERY_CFG_CA_ACK_DELAY_MASK) >>
OCRDMA_MBX_QUERY_CFG_CA_ACK_DELAY_SHIFT;
+ attr->max_mw = rsp->max_mw;
attr->max_mr = rsp->max_mr;
attr->max_mr_size = ~0ull;
attr->max_fmr = 0;
@@ -1036,7 +1098,7 @@ static void ocrdma_get_attr(struct ocrdma_dev *dev,
attr->max_inline_data =
attr->wqe_size - (sizeof(struct ocrdma_hdr_wqe) +
sizeof(struct ocrdma_sge));
- if (dev->nic_info.dev_family == OCRDMA_GEN2_FAMILY) {
+ if (ocrdma_get_asic_type(dev) == OCRDMA_ASIC_GEN_SKH_R) {
attr->ird = 1;
attr->ird_page_size = OCRDMA_MIN_Q_PAGE_SIZE;
attr->num_ird_pages = MAX_OCRDMA_IRD_PAGES;
@@ -1110,6 +1172,96 @@ mbx_err:
return status;
}
+int ocrdma_mbx_rdma_stats(struct ocrdma_dev *dev, bool reset)
+{
+ struct ocrdma_rdma_stats_req *req = dev->stats_mem.va;
+ struct ocrdma_mqe *mqe = &dev->stats_mem.mqe;
+ struct ocrdma_rdma_stats_resp *old_stats = NULL;
+ int status;
+
+ old_stats = kzalloc(sizeof(*old_stats), GFP_KERNEL);
+ if (old_stats == NULL)
+ return -ENOMEM;
+
+ memset(mqe, 0, sizeof(*mqe));
+ mqe->hdr.pyld_len = dev->stats_mem.size;
+ mqe->hdr.spcl_sge_cnt_emb |=
+ (1 << OCRDMA_MQE_HDR_SGE_CNT_SHIFT) &
+ OCRDMA_MQE_HDR_SGE_CNT_MASK;
+ mqe->u.nonemb_req.sge[0].pa_lo = (u32) (dev->stats_mem.pa & 0xffffffff);
+ mqe->u.nonemb_req.sge[0].pa_hi = (u32) upper_32_bits(dev->stats_mem.pa);
+ mqe->u.nonemb_req.sge[0].len = dev->stats_mem.size;
+
+ /* Cache the old stats */
+ memcpy(old_stats, req, sizeof(struct ocrdma_rdma_stats_resp));
+ memset(req, 0, dev->stats_mem.size);
+
+ ocrdma_init_mch((struct ocrdma_mbx_hdr *)req,
+ OCRDMA_CMD_GET_RDMA_STATS,
+ OCRDMA_SUBSYS_ROCE,
+ dev->stats_mem.size);
+ if (reset)
+ req->reset_stats = reset;
+
+ status = ocrdma_nonemb_mbx_cmd(dev, mqe, dev->stats_mem.va);
+ if (status)
+ /* Copy from cache, if mbox fails */
+ memcpy(req, old_stats, sizeof(struct ocrdma_rdma_stats_resp));
+ else
+ ocrdma_le32_to_cpu(req, dev->stats_mem.size);
+
+ kfree(old_stats);
+ return status;
+}
+
+static int ocrdma_mbx_get_ctrl_attribs(struct ocrdma_dev *dev)
+{
+ int status = -ENOMEM;
+ struct ocrdma_dma_mem dma;
+ struct ocrdma_mqe *mqe;
+ struct ocrdma_get_ctrl_attribs_rsp *ctrl_attr_rsp;
+ struct mgmt_hba_attribs *hba_attribs;
+
+ mqe = ocrdma_alloc_mqe();
+ if (!mqe)
+ return status;
+ memset(mqe, 0, sizeof(*mqe));
+
+ dma.size = sizeof(struct ocrdma_get_ctrl_attribs_rsp);
+ dma.va = dma_alloc_coherent(&dev->nic_info.pdev->dev,
+ dma.size, &dma.pa, GFP_KERNEL);
+ if (!dma.va)
+ goto free_mqe;
+
+ mqe->hdr.pyld_len = dma.size;
+ mqe->hdr.spcl_sge_cnt_emb |=
+ (1 << OCRDMA_MQE_HDR_SGE_CNT_SHIFT) &
+ OCRDMA_MQE_HDR_SGE_CNT_MASK;
+ mqe->u.nonemb_req.sge[0].pa_lo = (u32) (dma.pa & 0xffffffff);
+ mqe->u.nonemb_req.sge[0].pa_hi = (u32) upper_32_bits(dma.pa);
+ mqe->u.nonemb_req.sge[0].len = dma.size;
+
+ memset(dma.va, 0, dma.size);
+ ocrdma_init_mch((struct ocrdma_mbx_hdr *)dma.va,
+ OCRDMA_CMD_GET_CTRL_ATTRIBUTES,
+ OCRDMA_SUBSYS_COMMON,
+ dma.size);
+
+ status = ocrdma_nonemb_mbx_cmd(dev, mqe, dma.va);
+ if (!status) {
+ ctrl_attr_rsp = (struct ocrdma_get_ctrl_attribs_rsp *)dma.va;
+ hba_attribs = &ctrl_attr_rsp->ctrl_attribs.hba_attribs;
+
+ dev->hba_port_num = hba_attribs->phy_port;
+ strncpy(dev->model_number,
+ hba_attribs->controller_model_number, 31);
+ }
+ dma_free_coherent(&dev->nic_info.pdev->dev, dma.size, dma.va, dma.pa);
+free_mqe:
+ kfree(mqe);
+ return status;
+}
+
static int ocrdma_mbx_query_dev(struct ocrdma_dev *dev)
{
int status = -ENOMEM;
@@ -1157,6 +1309,35 @@ mbx_err:
return status;
}
+static int ocrdma_mbx_get_phy_info(struct ocrdma_dev *dev)
+{
+ int status = -ENOMEM;
+ struct ocrdma_mqe *cmd;
+ struct ocrdma_get_phy_info_rsp *rsp;
+
+ cmd = ocrdma_init_emb_mqe(OCRDMA_CMD_PHY_DETAILS, sizeof(*cmd));
+ if (!cmd)
+ return status;
+
+ ocrdma_init_mch((struct ocrdma_mbx_hdr *)&cmd->u.cmd[0],
+ OCRDMA_CMD_PHY_DETAILS, OCRDMA_SUBSYS_COMMON,
+ sizeof(*cmd));
+
+ status = ocrdma_mbx_cmd(dev, (struct ocrdma_mqe *)cmd);
+ if (status)
+ goto mbx_err;
+
+ rsp = (struct ocrdma_get_phy_info_rsp *)cmd;
+ dev->phy.phy_type = le16_to_cpu(rsp->phy_type);
+ dev->phy.auto_speeds_supported =
+ le16_to_cpu(rsp->auto_speeds_supported);
+ dev->phy.fixed_speeds_supported =
+ le16_to_cpu(rsp->fixed_speeds_supported);
+mbx_err:
+ kfree(cmd);
+ return status;
+}
+
int ocrdma_mbx_alloc_pd(struct ocrdma_dev *dev, struct ocrdma_pd *pd)
{
int status = -ENOMEM;
@@ -1226,7 +1407,7 @@ static int ocrdma_build_q_conf(u32 *num_entries, int entry_size,
static int ocrdma_mbx_create_ah_tbl(struct ocrdma_dev *dev)
{
- int i ;
+ int i;
int status = 0;
int max_ah;
struct ocrdma_create_ah_tbl *cmd;
@@ -1357,12 +1538,10 @@ static void ocrdma_unbind_eq(struct ocrdma_dev *dev, u16 eq_id)
int i;
mutex_lock(&dev->dev_lock);
- for (i = 0; i < dev->eq_cnt; i++) {
- if (dev->eq_tbl[i].q.id != eq_id)
- continue;
- dev->eq_tbl[i].cq_cnt -= 1;
- break;
- }
+ i = ocrdma_get_eq_table_index(dev, eq_id);
+ if (i == -EINVAL)
+ BUG();
+ dev->eq_tbl[i].cq_cnt -= 1;
mutex_unlock(&dev->dev_lock);
}
@@ -1380,7 +1559,7 @@ int ocrdma_mbx_create_cq(struct ocrdma_dev *dev, struct ocrdma_cq *cq,
__func__, dev->id, dev->attr.max_cqe, entries);
return -EINVAL;
}
- if (dpp_cq && (dev->nic_info.dev_family != OCRDMA_GEN2_FAMILY))
+ if (dpp_cq && (ocrdma_get_asic_type(dev) != OCRDMA_ASIC_GEN_SKH_R))
return -EINVAL;
if (dpp_cq) {
@@ -1417,6 +1596,7 @@ int ocrdma_mbx_create_cq(struct ocrdma_dev *dev, struct ocrdma_cq *cq,
cq->eqn = ocrdma_bind_eq(dev);
cmd->cmd.req.rsvd_version = OCRDMA_CREATE_CQ_VER3;
cqe_count = cq->len / cqe_size;
+ cq->cqe_cnt = cqe_count;
if (cqe_count > 1024) {
/* Set cnt to 3 to indicate more than 1024 cq entries */
cmd->cmd.ev_cnt_flags |= (0x3 << OCRDMA_CREATE_CQ_CNT_SHIFT);
@@ -1439,7 +1619,7 @@ int ocrdma_mbx_create_cq(struct ocrdma_dev *dev, struct ocrdma_cq *cq,
}
/* shared eq between all the consumer cqs. */
cmd->cmd.eqn = cq->eqn;
- if (dev->nic_info.dev_family == OCRDMA_GEN2_FAMILY) {
+ if (ocrdma_get_asic_type(dev) == OCRDMA_ASIC_GEN_SKH_R) {
if (dpp_cq)
cmd->cmd.pgsz_pgcnt |= OCRDMA_CREATE_CQ_DPP <<
OCRDMA_CREATE_CQ_TYPE_SHIFT;
@@ -1484,12 +1664,9 @@ int ocrdma_mbx_destroy_cq(struct ocrdma_dev *dev, struct ocrdma_cq *cq)
(cq->id << OCRDMA_DESTROY_CQ_QID_SHIFT) &
OCRDMA_DESTROY_CQ_QID_MASK;
- ocrdma_unbind_eq(dev, cq->eqn);
status = ocrdma_mbx_cmd(dev, (struct ocrdma_mqe *)cmd);
- if (status)
- goto mbx_err;
+ ocrdma_unbind_eq(dev, cq->eqn);
dma_free_coherent(&dev->nic_info.pdev->dev, cq->len, cq->va, cq->pa);
-mbx_err:
kfree(cmd);
return status;
}
@@ -2029,8 +2206,7 @@ int ocrdma_mbx_create_qp(struct ocrdma_qp *qp, struct ib_qp_init_attr *attrs,
OCRDMA_CREATE_QP_REQ_RQ_CQID_MASK;
qp->rq_cq = cq;
- if (pd->dpp_enabled && attrs->cap.max_inline_data && pd->num_dpp_qp &&
- (attrs->cap.max_inline_data <= dev->attr.max_inline_data)) {
+ if (pd->dpp_enabled && pd->num_dpp_qp) {
ocrdma_set_create_qp_dpp_cmd(cmd, pd, qp, enable_dpp_cq,
dpp_cq_id);
}
@@ -2099,7 +2275,7 @@ static int ocrdma_set_av_params(struct ocrdma_qp *qp,
memcpy(&cmd->params.dgid[0], &ah_attr->grh.dgid.raw[0],
sizeof(cmd->params.dgid));
status = ocrdma_query_gid(&qp->dev->ibdev, 1,
- ah_attr->grh.sgid_index, &sgid);
+ ah_attr->grh.sgid_index, &sgid);
if (status)
return status;
@@ -2127,8 +2303,7 @@ static int ocrdma_set_av_params(struct ocrdma_qp *qp,
static int ocrdma_set_qp_params(struct ocrdma_qp *qp,
struct ocrdma_modify_qp *cmd,
- struct ib_qp_attr *attrs, int attr_mask,
- enum ib_qp_state old_qps)
+ struct ib_qp_attr *attrs, int attr_mask)
{
int status = 0;
@@ -2233,8 +2408,7 @@ pmtu_err:
}
int ocrdma_mbx_modify_qp(struct ocrdma_dev *dev, struct ocrdma_qp *qp,
- struct ib_qp_attr *attrs, int attr_mask,
- enum ib_qp_state old_qps)
+ struct ib_qp_attr *attrs, int attr_mask)
{
int status = -ENOMEM;
struct ocrdma_modify_qp *cmd;
@@ -2257,7 +2431,7 @@ int ocrdma_mbx_modify_qp(struct ocrdma_dev *dev, struct ocrdma_qp *qp,
OCRDMA_QP_PARAMS_STATE_MASK;
}
- status = ocrdma_set_qp_params(qp, cmd, attrs, attr_mask, old_qps);
+ status = ocrdma_set_qp_params(qp, cmd, attrs, attr_mask);
if (status)
goto mbx_err;
status = ocrdma_mbx_cmd(dev, (struct ocrdma_mqe *)cmd);
@@ -2488,7 +2662,7 @@ static int ocrdma_create_eqs(struct ocrdma_dev *dev)
for (i = 0; i < num_eq; i++) {
status = ocrdma_create_eq(dev, &dev->eq_tbl[i],
- OCRDMA_EQ_LEN);
+ OCRDMA_EQ_LEN);
if (status) {
status = -EINVAL;
break;
@@ -2533,6 +2707,13 @@ int ocrdma_init_hw(struct ocrdma_dev *dev)
status = ocrdma_mbx_create_ah_tbl(dev);
if (status)
goto conf_err;
+ status = ocrdma_mbx_get_phy_info(dev);
+ if (status)
+ goto conf_err;
+ status = ocrdma_mbx_get_ctrl_attribs(dev);
+ if (status)
+ goto conf_err;
+
return 0;
conf_err:
diff --git a/drivers/infiniband/hw/ocrdma/ocrdma_hw.h b/drivers/infiniband/hw/ocrdma/ocrdma_hw.h
index 82fe332ae6c6..e513f7293142 100644
--- a/drivers/infiniband/hw/ocrdma/ocrdma_hw.h
+++ b/drivers/infiniband/hw/ocrdma/ocrdma_hw.h
@@ -112,8 +112,7 @@ int ocrdma_mbx_create_qp(struct ocrdma_qp *, struct ib_qp_init_attr *attrs,
u8 enable_dpp_cq, u16 dpp_cq_id, u16 *dpp_offset,
u16 *dpp_credit_lmt);
int ocrdma_mbx_modify_qp(struct ocrdma_dev *, struct ocrdma_qp *,
- struct ib_qp_attr *attrs, int attr_mask,
- enum ib_qp_state old_qps);
+ struct ib_qp_attr *attrs, int attr_mask);
int ocrdma_mbx_query_qp(struct ocrdma_dev *, struct ocrdma_qp *,
struct ocrdma_qp_params *param);
int ocrdma_mbx_destroy_qp(struct ocrdma_dev *, struct ocrdma_qp *);
@@ -132,5 +131,8 @@ int ocrdma_qp_state_change(struct ocrdma_qp *, enum ib_qp_state new_state,
bool ocrdma_is_qp_in_sq_flushlist(struct ocrdma_cq *, struct ocrdma_qp *);
bool ocrdma_is_qp_in_rq_flushlist(struct ocrdma_cq *, struct ocrdma_qp *);
void ocrdma_flush_qp(struct ocrdma_qp *);
+int ocrdma_get_irq(struct ocrdma_dev *dev, struct ocrdma_eq *eq);
+int ocrdma_mbx_rdma_stats(struct ocrdma_dev *, bool reset);
+char *port_speed_string(struct ocrdma_dev *dev);
#endif /* __OCRDMA_HW_H__ */
diff --git a/drivers/infiniband/hw/ocrdma/ocrdma_main.c b/drivers/infiniband/hw/ocrdma/ocrdma_main.c
index 1a8a945efa60..7c504e079744 100644
--- a/drivers/infiniband/hw/ocrdma/ocrdma_main.c
+++ b/drivers/infiniband/hw/ocrdma/ocrdma_main.c
@@ -39,10 +39,11 @@
#include "ocrdma_ah.h"
#include "be_roce.h"
#include "ocrdma_hw.h"
+#include "ocrdma_stats.h"
#include "ocrdma_abi.h"
-MODULE_VERSION(OCRDMA_ROCE_DEV_VERSION);
-MODULE_DESCRIPTION("Emulex RoCE HCA Driver");
+MODULE_VERSION(OCRDMA_ROCE_DRV_VERSION);
+MODULE_DESCRIPTION(OCRDMA_ROCE_DRV_DESC " " OCRDMA_ROCE_DRV_VERSION);
MODULE_AUTHOR("Emulex Corporation");
MODULE_LICENSE("GPL");
@@ -286,7 +287,7 @@ static int ocrdma_register_device(struct ocrdma_dev *dev)
dev->ibdev.process_mad = ocrdma_process_mad;
- if (dev->nic_info.dev_family == OCRDMA_GEN2_FAMILY) {
+ if (ocrdma_get_asic_type(dev) == OCRDMA_ASIC_GEN_SKH_R) {
dev->ibdev.uverbs_cmd_mask |=
OCRDMA_UVERBS(CREATE_SRQ) |
OCRDMA_UVERBS(MODIFY_SRQ) |
@@ -338,9 +339,42 @@ static void ocrdma_free_resources(struct ocrdma_dev *dev)
kfree(dev->sgid_tbl);
}
+/* OCRDMA sysfs interface */
+static ssize_t show_rev(struct device *device, struct device_attribute *attr,
+ char *buf)
+{
+ struct ocrdma_dev *dev = dev_get_drvdata(device);
+
+ return scnprintf(buf, PAGE_SIZE, "0x%x\n", dev->nic_info.pdev->vendor);
+}
+
+static ssize_t show_fw_ver(struct device *device, struct device_attribute *attr,
+ char *buf)
+{
+ struct ocrdma_dev *dev = dev_get_drvdata(device);
+
+ return scnprintf(buf, PAGE_SIZE, "%s", &dev->attr.fw_ver[0]);
+}
+
+static DEVICE_ATTR(hw_rev, S_IRUGO, show_rev, NULL);
+static DEVICE_ATTR(fw_ver, S_IRUGO, show_fw_ver, NULL);
+
+static struct device_attribute *ocrdma_attributes[] = {
+ &dev_attr_hw_rev,
+ &dev_attr_fw_ver
+};
+
+static void ocrdma_remove_sysfiles(struct ocrdma_dev *dev)
+{
+ int i;
+
+ for (i = 0; i < ARRAY_SIZE(ocrdma_attributes); i++)
+ device_remove_file(&dev->ibdev.dev, ocrdma_attributes[i]);
+}
+
static struct ocrdma_dev *ocrdma_add(struct be_dev_info *dev_info)
{
- int status = 0;
+ int status = 0, i;
struct ocrdma_dev *dev;
dev = (struct ocrdma_dev *)ib_alloc_device(sizeof(struct ocrdma_dev));
@@ -369,11 +403,25 @@ static struct ocrdma_dev *ocrdma_add(struct be_dev_info *dev_info)
if (status)
goto alloc_err;
+ for (i = 0; i < ARRAY_SIZE(ocrdma_attributes); i++)
+ if (device_create_file(&dev->ibdev.dev, ocrdma_attributes[i]))
+ goto sysfs_err;
spin_lock(&ocrdma_devlist_lock);
list_add_tail_rcu(&dev->entry, &ocrdma_dev_list);
spin_unlock(&ocrdma_devlist_lock);
+ /* Init stats */
+ ocrdma_add_port_stats(dev);
+
+ pr_info("%s %s: %s \"%s\" port %d\n",
+ dev_name(&dev->nic_info.pdev->dev), hca_name(dev),
+ port_speed_string(dev), dev->model_number,
+ dev->hba_port_num);
+ pr_info("%s ocrdma%d driver loaded successfully\n",
+ dev_name(&dev->nic_info.pdev->dev), dev->id);
return dev;
+sysfs_err:
+ ocrdma_remove_sysfiles(dev);
alloc_err:
ocrdma_free_resources(dev);
ocrdma_cleanup_hw(dev);
@@ -400,6 +448,9 @@ static void ocrdma_remove(struct ocrdma_dev *dev)
/* first unregister with stack to stop all the active traffic
* of the registered clients.
*/
+ ocrdma_rem_port_stats(dev);
+ ocrdma_remove_sysfiles(dev);
+
ib_unregister_device(&dev->ibdev);
spin_lock(&ocrdma_devlist_lock);
@@ -437,7 +488,7 @@ static int ocrdma_close(struct ocrdma_dev *dev)
cur_qp = dev->qp_tbl;
for (i = 0; i < OCRDMA_MAX_QP; i++) {
qp = cur_qp[i];
- if (qp) {
+ if (qp && qp->ibqp.qp_type != IB_QPT_GSI) {
/* change the QP state to ERROR */
_ocrdma_modify_qp(&qp->ibqp, &attrs, attr_mask);
@@ -478,6 +529,7 @@ static struct ocrdma_driver ocrdma_drv = {
.add = ocrdma_add,
.remove = ocrdma_remove,
.state_change_handler = ocrdma_event_handler,
+ .be_abi_version = OCRDMA_BE_ROCE_ABI_VERSION,
};
static void ocrdma_unregister_inet6addr_notifier(void)
@@ -487,10 +539,17 @@ static void ocrdma_unregister_inet6addr_notifier(void)
#endif
}
+static void ocrdma_unregister_inetaddr_notifier(void)
+{
+ unregister_inetaddr_notifier(&ocrdma_inetaddr_notifier);
+}
+
static int __init ocrdma_init_module(void)
{
int status;
+ ocrdma_init_debugfs();
+
status = register_inetaddr_notifier(&ocrdma_inetaddr_notifier);
if (status)
return status;
@@ -498,13 +557,19 @@ static int __init ocrdma_init_module(void)
#if IS_ENABLED(CONFIG_IPV6)
status = register_inet6addr_notifier(&ocrdma_inet6addr_notifier);
if (status)
- return status;
+ goto err_notifier6;
#endif
status = be_roce_register_driver(&ocrdma_drv);
if (status)
- ocrdma_unregister_inet6addr_notifier();
+ goto err_be_reg;
+ return 0;
+
+err_be_reg:
+ ocrdma_unregister_inet6addr_notifier();
+err_notifier6:
+ ocrdma_unregister_inetaddr_notifier();
return status;
}
@@ -512,6 +577,8 @@ static void __exit ocrdma_exit_module(void)
{
be_roce_unregister_driver(&ocrdma_drv);
ocrdma_unregister_inet6addr_notifier();
+ ocrdma_unregister_inetaddr_notifier();
+ ocrdma_rem_debugfs();
}
module_init(ocrdma_init_module);
diff --git a/drivers/infiniband/hw/ocrdma/ocrdma_sli.h b/drivers/infiniband/hw/ocrdma/ocrdma_sli.h
index 60d5ac23ea80..96c9ee602ba4 100644
--- a/drivers/infiniband/hw/ocrdma/ocrdma_sli.h
+++ b/drivers/infiniband/hw/ocrdma/ocrdma_sli.h
@@ -30,8 +30,16 @@
#define Bit(_b) (1 << (_b))
-#define OCRDMA_GEN1_FAMILY 0xB
-#define OCRDMA_GEN2_FAMILY 0x0F
+enum {
+ OCRDMA_ASIC_GEN_SKH_R = 0x04,
+ OCRDMA_ASIC_GEN_LANCER = 0x0B
+};
+
+enum {
+ OCRDMA_ASIC_REV_A0 = 0x00,
+ OCRDMA_ASIC_REV_B0 = 0x10,
+ OCRDMA_ASIC_REV_C0 = 0x20
+};
#define OCRDMA_SUBSYS_ROCE 10
enum {
@@ -64,6 +72,7 @@ enum {
OCRDMA_CMD_ATTACH_MCAST,
OCRDMA_CMD_DETACH_MCAST,
+ OCRDMA_CMD_GET_RDMA_STATS,
OCRDMA_CMD_MAX
};
@@ -74,12 +83,14 @@ enum {
OCRDMA_CMD_CREATE_CQ = 12,
OCRDMA_CMD_CREATE_EQ = 13,
OCRDMA_CMD_CREATE_MQ = 21,
+ OCRDMA_CMD_GET_CTRL_ATTRIBUTES = 32,
OCRDMA_CMD_GET_FW_VER = 35,
OCRDMA_CMD_DELETE_MQ = 53,
OCRDMA_CMD_DELETE_CQ = 54,
OCRDMA_CMD_DELETE_EQ = 55,
OCRDMA_CMD_GET_FW_CONFIG = 58,
- OCRDMA_CMD_CREATE_MQ_EXT = 90
+ OCRDMA_CMD_CREATE_MQ_EXT = 90,
+ OCRDMA_CMD_PHY_DETAILS = 102
};
enum {
@@ -103,7 +114,10 @@ enum {
OCRDMA_DB_GEN2_SRQ_OFFSET = OCRDMA_DB_GEN2_RQ_OFFSET,
OCRDMA_DB_CQ_OFFSET = 0x120,
OCRDMA_DB_EQ_OFFSET = OCRDMA_DB_CQ_OFFSET,
- OCRDMA_DB_MQ_OFFSET = 0x140
+ OCRDMA_DB_MQ_OFFSET = 0x140,
+
+ OCRDMA_DB_SQ_SHIFT = 16,
+ OCRDMA_DB_RQ_SHIFT = 24
};
#define OCRDMA_DB_CQ_RING_ID_MASK 0x3FF /* bits 0 - 9 */
@@ -138,6 +152,10 @@ enum {
#define OCRDMA_MIN_Q_PAGE_SIZE (4096)
#define OCRDMA_MAX_Q_PAGES (8)
+#define OCRDMA_SLI_ASIC_ID_OFFSET 0x9C
+#define OCRDMA_SLI_ASIC_REV_MASK 0x000000FF
+#define OCRDMA_SLI_ASIC_GEN_NUM_MASK 0x0000FF00
+#define OCRDMA_SLI_ASIC_GEN_NUM_SHIFT 0x08
/*
# 0: 4K Bytes
# 1: 8K Bytes
@@ -562,6 +580,30 @@ enum {
OCRDMA_FN_MODE_RDMA = 0x4
};
+struct ocrdma_get_phy_info_rsp {
+ struct ocrdma_mqe_hdr hdr;
+ struct ocrdma_mbx_rsp rsp;
+
+ u16 phy_type;
+ u16 interface_type;
+ u32 misc_params;
+ u16 ext_phy_details;
+ u16 rsvd;
+ u16 auto_speeds_supported;
+ u16 fixed_speeds_supported;
+ u32 future_use[2];
+};
+
+enum {
+ OCRDMA_PHY_SPEED_ZERO = 0x0,
+ OCRDMA_PHY_SPEED_10MBPS = 0x1,
+ OCRDMA_PHY_SPEED_100MBPS = 0x2,
+ OCRDMA_PHY_SPEED_1GBPS = 0x4,
+ OCRDMA_PHY_SPEED_10GBPS = 0x8,
+ OCRDMA_PHY_SPEED_40GBPS = 0x20
+};
+
+
struct ocrdma_get_link_speed_rsp {
struct ocrdma_mqe_hdr hdr;
struct ocrdma_mbx_rsp rsp;
@@ -590,7 +632,7 @@ enum {
enum {
OCRDMA_CREATE_CQ_VER2 = 2,
- OCRDMA_CREATE_CQ_VER3 = 3,
+ OCRDMA_CREATE_CQ_VER3 = 3,
OCRDMA_CREATE_CQ_PAGE_CNT_MASK = 0xFFFF,
OCRDMA_CREATE_CQ_PAGE_SIZE_SHIFT = 16,
@@ -1050,6 +1092,7 @@ enum {
OCRDMA_MODIFY_QP_RSP_MAX_ORD_MASK = 0xFFFF <<
OCRDMA_MODIFY_QP_RSP_MAX_ORD_SHIFT
};
+
struct ocrdma_modify_qp_rsp {
struct ocrdma_mqe_hdr hdr;
struct ocrdma_mbx_rsp rsp;
@@ -1062,8 +1105,8 @@ struct ocrdma_query_qp {
struct ocrdma_mqe_hdr hdr;
struct ocrdma_mbx_hdr req;
-#define OCRDMA_QUERY_UP_QP_ID_SHIFT 0
-#define OCRDMA_QUERY_UP_QP_ID_MASK 0xFFFFFF
+#define OCRDMA_QUERY_UP_QP_ID_SHIFT 0
+#define OCRDMA_QUERY_UP_QP_ID_MASK 0xFFFFFF
u32 qp_id;
};
@@ -1703,4 +1746,208 @@ struct ocrdma_av {
u32 valid;
} __packed;
+struct ocrdma_rsrc_stats {
+ u32 dpp_pds;
+ u32 non_dpp_pds;
+ u32 rc_dpp_qps;
+ u32 uc_dpp_qps;
+ u32 ud_dpp_qps;
+ u32 rc_non_dpp_qps;
+ u32 rsvd;
+ u32 uc_non_dpp_qps;
+ u32 ud_non_dpp_qps;
+ u32 rsvd1;
+ u32 srqs;
+ u32 rbqs;
+ u32 r64K_nsmr;
+ u32 r64K_to_2M_nsmr;
+ u32 r2M_to_44M_nsmr;
+ u32 r44M_to_1G_nsmr;
+ u32 r1G_to_4G_nsmr;
+ u32 nsmr_count_4G_to_32G;
+ u32 r32G_to_64G_nsmr;
+ u32 r64G_to_128G_nsmr;
+ u32 r128G_to_higher_nsmr;
+ u32 embedded_nsmr;
+ u32 frmr;
+ u32 prefetch_qps;
+ u32 ondemand_qps;
+ u32 phy_mr;
+ u32 mw;
+ u32 rsvd2[7];
+};
+
+struct ocrdma_db_err_stats {
+ u32 sq_doorbell_errors;
+ u32 cq_doorbell_errors;
+ u32 rq_srq_doorbell_errors;
+ u32 cq_overflow_errors;
+ u32 rsvd[4];
+};
+
+struct ocrdma_wqe_stats {
+ u32 large_send_rc_wqes_lo;
+ u32 large_send_rc_wqes_hi;
+ u32 large_write_rc_wqes_lo;
+ u32 large_write_rc_wqes_hi;
+ u32 rsvd[4];
+ u32 read_wqes_lo;
+ u32 read_wqes_hi;
+ u32 frmr_wqes_lo;
+ u32 frmr_wqes_hi;
+ u32 mw_bind_wqes_lo;
+ u32 mw_bind_wqes_hi;
+ u32 invalidate_wqes_lo;
+ u32 invalidate_wqes_hi;
+ u32 rsvd1[2];
+ u32 dpp_wqe_drops;
+ u32 rsvd2[5];
+};
+
+struct ocrdma_tx_stats {
+ u32 send_pkts_lo;
+ u32 send_pkts_hi;
+ u32 write_pkts_lo;
+ u32 write_pkts_hi;
+ u32 read_pkts_lo;
+ u32 read_pkts_hi;
+ u32 read_rsp_pkts_lo;
+ u32 read_rsp_pkts_hi;
+ u32 ack_pkts_lo;
+ u32 ack_pkts_hi;
+ u32 send_bytes_lo;
+ u32 send_bytes_hi;
+ u32 write_bytes_lo;
+ u32 write_bytes_hi;
+ u32 read_req_bytes_lo;
+ u32 read_req_bytes_hi;
+ u32 read_rsp_bytes_lo;
+ u32 read_rsp_bytes_hi;
+ u32 ack_timeouts;
+ u32 rsvd[5];
+};
+
+
+struct ocrdma_tx_qp_err_stats {
+ u32 local_length_errors;
+ u32 local_protection_errors;
+ u32 local_qp_operation_errors;
+ u32 retry_count_exceeded_errors;
+ u32 rnr_retry_count_exceeded_errors;
+ u32 rsvd[3];
+};
+
+struct ocrdma_rx_stats {
+ u32 roce_frame_bytes_lo;
+ u32 roce_frame_bytes_hi;
+ u32 roce_frame_icrc_drops;
+ u32 roce_frame_payload_len_drops;
+ u32 ud_drops;
+ u32 qp1_drops;
+ u32 psn_error_request_packets;
+ u32 psn_error_resp_packets;
+ u32 rnr_nak_timeouts;
+ u32 rnr_nak_receives;
+ u32 roce_frame_rxmt_drops;
+ u32 nak_count_psn_sequence_errors;
+ u32 rc_drop_count_lookup_errors;
+ u32 rq_rnr_naks;
+ u32 srq_rnr_naks;
+ u32 roce_frames_lo;
+ u32 roce_frames_hi;
+ u32 rsvd;
+};
+
+struct ocrdma_rx_qp_err_stats {
+ u32 nak_invalid_requst_errors;
+ u32 nak_remote_operation_errors;
+ u32 nak_count_remote_access_errors;
+ u32 local_length_errors;
+ u32 local_protection_errors;
+ u32 local_qp_operation_errors;
+ u32 rsvd[2];
+};
+
+struct ocrdma_tx_dbg_stats {
+ u32 data[100];
+};
+
+struct ocrdma_rx_dbg_stats {
+ u32 data[200];
+};
+
+struct ocrdma_rdma_stats_req {
+ struct ocrdma_mbx_hdr hdr;
+ u8 reset_stats;
+ u8 rsvd[3];
+} __packed;
+
+struct ocrdma_rdma_stats_resp {
+ struct ocrdma_mbx_hdr hdr;
+ struct ocrdma_rsrc_stats act_rsrc_stats;
+ struct ocrdma_rsrc_stats th_rsrc_stats;
+ struct ocrdma_db_err_stats db_err_stats;
+ struct ocrdma_wqe_stats wqe_stats;
+ struct ocrdma_tx_stats tx_stats;
+ struct ocrdma_tx_qp_err_stats tx_qp_err_stats;
+ struct ocrdma_rx_stats rx_stats;
+ struct ocrdma_rx_qp_err_stats rx_qp_err_stats;
+ struct ocrdma_tx_dbg_stats tx_dbg_stats;
+ struct ocrdma_rx_dbg_stats rx_dbg_stats;
+} __packed;
+
+
+struct mgmt_hba_attribs {
+ u8 flashrom_version_string[32];
+ u8 manufacturer_name[32];
+ u32 supported_modes;
+ u32 rsvd0[3];
+ u8 ncsi_ver_string[12];
+ u32 default_extended_timeout;
+ u8 controller_model_number[32];
+ u8 controller_description[64];
+ u8 controller_serial_number[32];
+ u8 ip_version_string[32];
+ u8 firmware_version_string[32];
+ u8 bios_version_string[32];
+ u8 redboot_version_string[32];
+ u8 driver_version_string[32];
+ u8 fw_on_flash_version_string[32];
+ u32 functionalities_supported;
+ u16 max_cdblength;
+ u8 asic_revision;
+ u8 generational_guid[16];
+ u8 hba_port_count;
+ u16 default_link_down_timeout;
+ u8 iscsi_ver_min_max;
+ u8 multifunction_device;
+ u8 cache_valid;
+ u8 hba_status;
+ u8 max_domains_supported;
+ u8 phy_port;
+ u32 firmware_post_status;
+ u32 hba_mtu[8];
+ u32 rsvd1[4];
+};
+
+struct mgmt_controller_attrib {
+ struct mgmt_hba_attribs hba_attribs;
+ u16 pci_vendor_id;
+ u16 pci_device_id;
+ u16 pci_sub_vendor_id;
+ u16 pci_sub_system_id;
+ u8 pci_bus_number;
+ u8 pci_device_number;
+ u8 pci_function_number;
+ u8 interface_type;
+ u64 unique_identifier;
+ u32 rsvd0[5];
+};
+
+struct ocrdma_get_ctrl_attribs_rsp {
+ struct ocrdma_mbx_hdr hdr;
+ struct mgmt_controller_attrib ctrl_attribs;
+};
+
+
#endif /* __OCRDMA_SLI_H__ */
diff --git a/drivers/infiniband/hw/ocrdma/ocrdma_stats.c b/drivers/infiniband/hw/ocrdma/ocrdma_stats.c
new file mode 100644
index 000000000000..6c54106f5e64
--- /dev/null
+++ b/drivers/infiniband/hw/ocrdma/ocrdma_stats.c
@@ -0,0 +1,623 @@
+/*******************************************************************
+ * This file is part of the Emulex RoCE Device Driver for *
+ * RoCE (RDMA over Converged Ethernet) adapters. *
+ * Copyright (C) 2008-2014 Emulex. All rights reserved. *
+ * EMULEX and SLI are trademarks of Emulex. *
+ * www.emulex.com *
+ * *
+ * This program is free software; you can redistribute it and/or *
+ * modify it under the terms of version 2 of the GNU General *
+ * Public License as published by the Free Software Foundation. *
+ * This program is distributed in the hope that it will be useful. *
+ * ALL EXPRESS OR IMPLIED CONDITIONS, REPRESENTATIONS AND *
+ * WARRANTIES, INCLUDING ANY IMPLIED WARRANTY OF MERCHANTABILITY, *
+ * FITNESS FOR A PARTICULAR PURPOSE, OR NON-INFRINGEMENT, ARE *
+ * DISCLAIMED, EXCEPT TO THE EXTENT THAT SUCH DISCLAIMERS ARE HELD *
+ * TO BE LEGALLY INVALID. See the GNU General Public License for *
+ * more details, a copy of which can be found in the file COPYING *
+ * included with this package. *
+ *
+ * Contact Information:
+ * linux-drivers@emulex.com
+ *
+ * Emulex
+ * 3333 Susan Street
+ * Costa Mesa, CA 92626
+ *******************************************************************/
+
+#include <rdma/ib_addr.h>
+#include "ocrdma_stats.h"
+
+static struct dentry *ocrdma_dbgfs_dir;
+
+static int ocrdma_add_stat(char *start, char *pcur,
+ char *name, u64 count)
+{
+ char buff[128] = {0};
+ int cpy_len = 0;
+
+ snprintf(buff, 128, "%s: %llu\n", name, count);
+ cpy_len = strlen(buff);
+
+ if (pcur + cpy_len > start + OCRDMA_MAX_DBGFS_MEM) {
+ pr_err("%s: No space in stats buff\n", __func__);
+ return 0;
+ }
+
+ memcpy(pcur, buff, cpy_len);
+ return cpy_len;
+}
+
+static bool ocrdma_alloc_stats_mem(struct ocrdma_dev *dev)
+{
+ struct stats_mem *mem = &dev->stats_mem;
+
+ /* Alloc mbox command mem*/
+ mem->size = max_t(u32, sizeof(struct ocrdma_rdma_stats_req),
+ sizeof(struct ocrdma_rdma_stats_resp));
+
+ mem->va = dma_alloc_coherent(&dev->nic_info.pdev->dev, mem->size,
+ &mem->pa, GFP_KERNEL);
+ if (!mem->va) {
+ pr_err("%s: stats mbox allocation failed\n", __func__);
+ return false;
+ }
+
+ memset(mem->va, 0, mem->size);
+
+ /* Alloc debugfs mem */
+ mem->debugfs_mem = kzalloc(OCRDMA_MAX_DBGFS_MEM, GFP_KERNEL);
+ if (!mem->debugfs_mem) {
+ pr_err("%s: stats debugfs mem allocation failed\n", __func__);
+ return false;
+ }
+
+ return true;
+}
+
+static void ocrdma_release_stats_mem(struct ocrdma_dev *dev)
+{
+ struct stats_mem *mem = &dev->stats_mem;
+
+ if (mem->va)
+ dma_free_coherent(&dev->nic_info.pdev->dev, mem->size,
+ mem->va, mem->pa);
+ kfree(mem->debugfs_mem);
+}
+
+static char *ocrdma_resource_stats(struct ocrdma_dev *dev)
+{
+ char *stats = dev->stats_mem.debugfs_mem, *pcur;
+ struct ocrdma_rdma_stats_resp *rdma_stats =
+ (struct ocrdma_rdma_stats_resp *)dev->stats_mem.va;
+ struct ocrdma_rsrc_stats *rsrc_stats = &rdma_stats->act_rsrc_stats;
+
+ memset(stats, 0, (OCRDMA_MAX_DBGFS_MEM));
+
+ pcur = stats;
+ pcur += ocrdma_add_stat(stats, pcur, "active_dpp_pds",
+ (u64)rsrc_stats->dpp_pds);
+ pcur += ocrdma_add_stat(stats, pcur, "active_non_dpp_pds",
+ (u64)rsrc_stats->non_dpp_pds);
+ pcur += ocrdma_add_stat(stats, pcur, "active_rc_dpp_qps",
+ (u64)rsrc_stats->rc_dpp_qps);
+ pcur += ocrdma_add_stat(stats, pcur, "active_uc_dpp_qps",
+ (u64)rsrc_stats->uc_dpp_qps);
+ pcur += ocrdma_add_stat(stats, pcur, "active_ud_dpp_qps",
+ (u64)rsrc_stats->ud_dpp_qps);
+ pcur += ocrdma_add_stat(stats, pcur, "active_rc_non_dpp_qps",
+ (u64)rsrc_stats->rc_non_dpp_qps);
+ pcur += ocrdma_add_stat(stats, pcur, "active_uc_non_dpp_qps",
+ (u64)rsrc_stats->uc_non_dpp_qps);
+ pcur += ocrdma_add_stat(stats, pcur, "active_ud_non_dpp_qps",
+ (u64)rsrc_stats->ud_non_dpp_qps);
+ pcur += ocrdma_add_stat(stats, pcur, "active_srqs",
+ (u64)rsrc_stats->srqs);
+ pcur += ocrdma_add_stat(stats, pcur, "active_rbqs",
+ (u64)rsrc_stats->rbqs);
+ pcur += ocrdma_add_stat(stats, pcur, "active_64K_nsmr",
+ (u64)rsrc_stats->r64K_nsmr);
+ pcur += ocrdma_add_stat(stats, pcur, "active_64K_to_2M_nsmr",
+ (u64)rsrc_stats->r64K_to_2M_nsmr);
+ pcur += ocrdma_add_stat(stats, pcur, "active_2M_to_44M_nsmr",
+ (u64)rsrc_stats->r2M_to_44M_nsmr);
+ pcur += ocrdma_add_stat(stats, pcur, "active_44M_to_1G_nsmr",
+ (u64)rsrc_stats->r44M_to_1G_nsmr);
+ pcur += ocrdma_add_stat(stats, pcur, "active_1G_to_4G_nsmr",
+ (u64)rsrc_stats->r1G_to_4G_nsmr);
+ pcur += ocrdma_add_stat(stats, pcur, "active_nsmr_count_4G_to_32G",
+ (u64)rsrc_stats->nsmr_count_4G_to_32G);
+ pcur += ocrdma_add_stat(stats, pcur, "active_32G_to_64G_nsmr",
+ (u64)rsrc_stats->r32G_to_64G_nsmr);
+ pcur += ocrdma_add_stat(stats, pcur, "active_64G_to_128G_nsmr",
+ (u64)rsrc_stats->r64G_to_128G_nsmr);
+ pcur += ocrdma_add_stat(stats, pcur, "active_128G_to_higher_nsmr",
+ (u64)rsrc_stats->r128G_to_higher_nsmr);
+ pcur += ocrdma_add_stat(stats, pcur, "active_embedded_nsmr",
+ (u64)rsrc_stats->embedded_nsmr);
+ pcur += ocrdma_add_stat(stats, pcur, "active_frmr",
+ (u64)rsrc_stats->frmr);
+ pcur += ocrdma_add_stat(stats, pcur, "active_prefetch_qps",
+ (u64)rsrc_stats->prefetch_qps);
+ pcur += ocrdma_add_stat(stats, pcur, "active_ondemand_qps",
+ (u64)rsrc_stats->ondemand_qps);
+ pcur += ocrdma_add_stat(stats, pcur, "active_phy_mr",
+ (u64)rsrc_stats->phy_mr);
+ pcur += ocrdma_add_stat(stats, pcur, "active_mw",
+ (u64)rsrc_stats->mw);
+
+ /* Print the threshold stats */
+ rsrc_stats = &rdma_stats->th_rsrc_stats;
+
+ pcur += ocrdma_add_stat(stats, pcur, "threshold_dpp_pds",
+ (u64)rsrc_stats->dpp_pds);
+ pcur += ocrdma_add_stat(stats, pcur, "threshold_non_dpp_pds",
+ (u64)rsrc_stats->non_dpp_pds);
+ pcur += ocrdma_add_stat(stats, pcur, "threshold_rc_dpp_qps",
+ (u64)rsrc_stats->rc_dpp_qps);
+ pcur += ocrdma_add_stat(stats, pcur, "threshold_uc_dpp_qps",
+ (u64)rsrc_stats->uc_dpp_qps);
+ pcur += ocrdma_add_stat(stats, pcur, "threshold_ud_dpp_qps",
+ (u64)rsrc_stats->ud_dpp_qps);
+ pcur += ocrdma_add_stat(stats, pcur, "threshold_rc_non_dpp_qps",
+ (u64)rsrc_stats->rc_non_dpp_qps);
+ pcur += ocrdma_add_stat(stats, pcur, "threshold_uc_non_dpp_qps",
+ (u64)rsrc_stats->uc_non_dpp_qps);
+ pcur += ocrdma_add_stat(stats, pcur, "threshold_ud_non_dpp_qps",
+ (u64)rsrc_stats->ud_non_dpp_qps);
+ pcur += ocrdma_add_stat(stats, pcur, "threshold_srqs",
+ (u64)rsrc_stats->srqs);
+ pcur += ocrdma_add_stat(stats, pcur, "threshold_rbqs",
+ (u64)rsrc_stats->rbqs);
+ pcur += ocrdma_add_stat(stats, pcur, "threshold_64K_nsmr",
+ (u64)rsrc_stats->r64K_nsmr);
+ pcur += ocrdma_add_stat(stats, pcur, "threshold_64K_to_2M_nsmr",
+ (u64)rsrc_stats->r64K_to_2M_nsmr);
+ pcur += ocrdma_add_stat(stats, pcur, "threshold_2M_to_44M_nsmr",
+ (u64)rsrc_stats->r2M_to_44M_nsmr);
+ pcur += ocrdma_add_stat(stats, pcur, "threshold_44M_to_1G_nsmr",
+ (u64)rsrc_stats->r44M_to_1G_nsmr);
+ pcur += ocrdma_add_stat(stats, pcur, "threshold_1G_to_4G_nsmr",
+ (u64)rsrc_stats->r1G_to_4G_nsmr);
+ pcur += ocrdma_add_stat(stats, pcur, "threshold_nsmr_count_4G_to_32G",
+ (u64)rsrc_stats->nsmr_count_4G_to_32G);
+ pcur += ocrdma_add_stat(stats, pcur, "threshold_32G_to_64G_nsmr",
+ (u64)rsrc_stats->r32G_to_64G_nsmr);
+ pcur += ocrdma_add_stat(stats, pcur, "threshold_64G_to_128G_nsmr",
+ (u64)rsrc_stats->r64G_to_128G_nsmr);
+ pcur += ocrdma_add_stat(stats, pcur, "threshold_128G_to_higher_nsmr",
+ (u64)rsrc_stats->r128G_to_higher_nsmr);
+ pcur += ocrdma_add_stat(stats, pcur, "threshold_embedded_nsmr",
+ (u64)rsrc_stats->embedded_nsmr);
+ pcur += ocrdma_add_stat(stats, pcur, "threshold_frmr",
+ (u64)rsrc_stats->frmr);
+ pcur += ocrdma_add_stat(stats, pcur, "threshold_prefetch_qps",
+ (u64)rsrc_stats->prefetch_qps);
+ pcur += ocrdma_add_stat(stats, pcur, "threshold_ondemand_qps",
+ (u64)rsrc_stats->ondemand_qps);
+ pcur += ocrdma_add_stat(stats, pcur, "threshold_phy_mr",
+ (u64)rsrc_stats->phy_mr);
+ pcur += ocrdma_add_stat(stats, pcur, "threshold_mw",
+ (u64)rsrc_stats->mw);
+ return stats;
+}
+
+static char *ocrdma_rx_stats(struct ocrdma_dev *dev)
+{
+ char *stats = dev->stats_mem.debugfs_mem, *pcur;
+ struct ocrdma_rdma_stats_resp *rdma_stats =
+ (struct ocrdma_rdma_stats_resp *)dev->stats_mem.va;
+ struct ocrdma_rx_stats *rx_stats = &rdma_stats->rx_stats;
+
+ memset(stats, 0, (OCRDMA_MAX_DBGFS_MEM));
+
+ pcur = stats;
+ pcur += ocrdma_add_stat
+ (stats, pcur, "roce_frame_bytes",
+ convert_to_64bit(rx_stats->roce_frame_bytes_lo,
+ rx_stats->roce_frame_bytes_hi));
+ pcur += ocrdma_add_stat(stats, pcur, "roce_frame_icrc_drops",
+ (u64)rx_stats->roce_frame_icrc_drops);
+ pcur += ocrdma_add_stat(stats, pcur, "roce_frame_payload_len_drops",
+ (u64)rx_stats->roce_frame_payload_len_drops);
+ pcur += ocrdma_add_stat(stats, pcur, "ud_drops",
+ (u64)rx_stats->ud_drops);
+ pcur += ocrdma_add_stat(stats, pcur, "qp1_drops",
+ (u64)rx_stats->qp1_drops);
+ pcur += ocrdma_add_stat(stats, pcur, "psn_error_request_packets",
+ (u64)rx_stats->psn_error_request_packets);
+ pcur += ocrdma_add_stat(stats, pcur, "psn_error_resp_packets",
+ (u64)rx_stats->psn_error_resp_packets);
+ pcur += ocrdma_add_stat(stats, pcur, "rnr_nak_timeouts",
+ (u64)rx_stats->rnr_nak_timeouts);
+ pcur += ocrdma_add_stat(stats, pcur, "rnr_nak_receives",
+ (u64)rx_stats->rnr_nak_receives);
+ pcur += ocrdma_add_stat(stats, pcur, "roce_frame_rxmt_drops",
+ (u64)rx_stats->roce_frame_rxmt_drops);
+ pcur += ocrdma_add_stat(stats, pcur, "nak_count_psn_sequence_errors",
+ (u64)rx_stats->nak_count_psn_sequence_errors);
+ pcur += ocrdma_add_stat(stats, pcur, "rc_drop_count_lookup_errors",
+ (u64)rx_stats->rc_drop_count_lookup_errors);
+ pcur += ocrdma_add_stat(stats, pcur, "rq_rnr_naks",
+ (u64)rx_stats->rq_rnr_naks);
+ pcur += ocrdma_add_stat(stats, pcur, "srq_rnr_naks",
+ (u64)rx_stats->srq_rnr_naks);
+ pcur += ocrdma_add_stat(stats, pcur, "roce_frames",
+ convert_to_64bit(rx_stats->roce_frames_lo,
+ rx_stats->roce_frames_hi));
+
+ return stats;
+}
+
+static char *ocrdma_tx_stats(struct ocrdma_dev *dev)
+{
+ char *stats = dev->stats_mem.debugfs_mem, *pcur;
+ struct ocrdma_rdma_stats_resp *rdma_stats =
+ (struct ocrdma_rdma_stats_resp *)dev->stats_mem.va;
+ struct ocrdma_tx_stats *tx_stats = &rdma_stats->tx_stats;
+
+ memset(stats, 0, (OCRDMA_MAX_DBGFS_MEM));
+
+ pcur = stats;
+ pcur += ocrdma_add_stat(stats, pcur, "send_pkts",
+ convert_to_64bit(tx_stats->send_pkts_lo,
+ tx_stats->send_pkts_hi));
+ pcur += ocrdma_add_stat(stats, pcur, "write_pkts",
+ convert_to_64bit(tx_stats->write_pkts_lo,
+ tx_stats->write_pkts_hi));
+ pcur += ocrdma_add_stat(stats, pcur, "read_pkts",
+ convert_to_64bit(tx_stats->read_pkts_lo,
+ tx_stats->read_pkts_hi));
+ pcur += ocrdma_add_stat(stats, pcur, "read_rsp_pkts",
+ convert_to_64bit(tx_stats->read_rsp_pkts_lo,
+ tx_stats->read_rsp_pkts_hi));
+ pcur += ocrdma_add_stat(stats, pcur, "ack_pkts",
+ convert_to_64bit(tx_stats->ack_pkts_lo,
+ tx_stats->ack_pkts_hi));
+ pcur += ocrdma_add_stat(stats, pcur, "send_bytes",
+ convert_to_64bit(tx_stats->send_bytes_lo,
+ tx_stats->send_bytes_hi));
+ pcur += ocrdma_add_stat(stats, pcur, "write_bytes",
+ convert_to_64bit(tx_stats->write_bytes_lo,
+ tx_stats->write_bytes_hi));
+ pcur += ocrdma_add_stat(stats, pcur, "read_req_bytes",
+ convert_to_64bit(tx_stats->read_req_bytes_lo,
+ tx_stats->read_req_bytes_hi));
+ pcur += ocrdma_add_stat(stats, pcur, "read_rsp_bytes",
+ convert_to_64bit(tx_stats->read_rsp_bytes_lo,
+ tx_stats->read_rsp_bytes_hi));
+ pcur += ocrdma_add_stat(stats, pcur, "ack_timeouts",
+ (u64)tx_stats->ack_timeouts);
+
+ return stats;
+}
+
+static char *ocrdma_wqe_stats(struct ocrdma_dev *dev)
+{
+ char *stats = dev->stats_mem.debugfs_mem, *pcur;
+ struct ocrdma_rdma_stats_resp *rdma_stats =
+ (struct ocrdma_rdma_stats_resp *)dev->stats_mem.va;
+ struct ocrdma_wqe_stats *wqe_stats = &rdma_stats->wqe_stats;
+
+ memset(stats, 0, (OCRDMA_MAX_DBGFS_MEM));
+
+ pcur = stats;
+ pcur += ocrdma_add_stat(stats, pcur, "large_send_rc_wqes",
+ convert_to_64bit(wqe_stats->large_send_rc_wqes_lo,
+ wqe_stats->large_send_rc_wqes_hi));
+ pcur += ocrdma_add_stat(stats, pcur, "large_write_rc_wqes",
+ convert_to_64bit(wqe_stats->large_write_rc_wqes_lo,
+ wqe_stats->large_write_rc_wqes_hi));
+ pcur += ocrdma_add_stat(stats, pcur, "read_wqes",
+ convert_to_64bit(wqe_stats->read_wqes_lo,
+ wqe_stats->read_wqes_hi));
+ pcur += ocrdma_add_stat(stats, pcur, "frmr_wqes",
+ convert_to_64bit(wqe_stats->frmr_wqes_lo,
+ wqe_stats->frmr_wqes_hi));
+ pcur += ocrdma_add_stat(stats, pcur, "mw_bind_wqes",
+ convert_to_64bit(wqe_stats->mw_bind_wqes_lo,
+ wqe_stats->mw_bind_wqes_hi));
+ pcur += ocrdma_add_stat(stats, pcur, "invalidate_wqes",
+ convert_to_64bit(wqe_stats->invalidate_wqes_lo,
+ wqe_stats->invalidate_wqes_hi));
+ pcur += ocrdma_add_stat(stats, pcur, "dpp_wqe_drops",
+ (u64)wqe_stats->dpp_wqe_drops);
+ return stats;
+}
+
+static char *ocrdma_db_errstats(struct ocrdma_dev *dev)
+{
+ char *stats = dev->stats_mem.debugfs_mem, *pcur;
+ struct ocrdma_rdma_stats_resp *rdma_stats =
+ (struct ocrdma_rdma_stats_resp *)dev->stats_mem.va;
+ struct ocrdma_db_err_stats *db_err_stats = &rdma_stats->db_err_stats;
+
+ memset(stats, 0, (OCRDMA_MAX_DBGFS_MEM));
+
+ pcur = stats;
+ pcur += ocrdma_add_stat(stats, pcur, "sq_doorbell_errors",
+ (u64)db_err_stats->sq_doorbell_errors);
+ pcur += ocrdma_add_stat(stats, pcur, "cq_doorbell_errors",
+ (u64)db_err_stats->cq_doorbell_errors);
+ pcur += ocrdma_add_stat(stats, pcur, "rq_srq_doorbell_errors",
+ (u64)db_err_stats->rq_srq_doorbell_errors);
+ pcur += ocrdma_add_stat(stats, pcur, "cq_overflow_errors",
+ (u64)db_err_stats->cq_overflow_errors);
+ return stats;
+}
+
+static char *ocrdma_rxqp_errstats(struct ocrdma_dev *dev)
+{
+ char *stats = dev->stats_mem.debugfs_mem, *pcur;
+ struct ocrdma_rdma_stats_resp *rdma_stats =
+ (struct ocrdma_rdma_stats_resp *)dev->stats_mem.va;
+ struct ocrdma_rx_qp_err_stats *rx_qp_err_stats =
+ &rdma_stats->rx_qp_err_stats;
+
+ memset(stats, 0, (OCRDMA_MAX_DBGFS_MEM));
+
+ pcur = stats;
+ pcur += ocrdma_add_stat(stats, pcur, "nak_invalid_requst_errors",
+ (u64)rx_qp_err_stats->nak_invalid_requst_errors);
+ pcur += ocrdma_add_stat(stats, pcur, "nak_remote_operation_errors",
+ (u64)rx_qp_err_stats->nak_remote_operation_errors);
+ pcur += ocrdma_add_stat(stats, pcur, "nak_count_remote_access_errors",
+ (u64)rx_qp_err_stats->nak_count_remote_access_errors);
+ pcur += ocrdma_add_stat(stats, pcur, "local_length_errors",
+ (u64)rx_qp_err_stats->local_length_errors);
+ pcur += ocrdma_add_stat(stats, pcur, "local_protection_errors",
+ (u64)rx_qp_err_stats->local_protection_errors);
+ pcur += ocrdma_add_stat(stats, pcur, "local_qp_operation_errors",
+ (u64)rx_qp_err_stats->local_qp_operation_errors);
+ return stats;
+}
+
+static char *ocrdma_txqp_errstats(struct ocrdma_dev *dev)
+{
+ char *stats = dev->stats_mem.debugfs_mem, *pcur;
+ struct ocrdma_rdma_stats_resp *rdma_stats =
+ (struct ocrdma_rdma_stats_resp *)dev->stats_mem.va;
+ struct ocrdma_tx_qp_err_stats *tx_qp_err_stats =
+ &rdma_stats->tx_qp_err_stats;
+
+ memset(stats, 0, (OCRDMA_MAX_DBGFS_MEM));
+
+ pcur = stats;
+ pcur += ocrdma_add_stat(stats, pcur, "local_length_errors",
+ (u64)tx_qp_err_stats->local_length_errors);
+ pcur += ocrdma_add_stat(stats, pcur, "local_protection_errors",
+ (u64)tx_qp_err_stats->local_protection_errors);
+ pcur += ocrdma_add_stat(stats, pcur, "local_qp_operation_errors",
+ (u64)tx_qp_err_stats->local_qp_operation_errors);
+ pcur += ocrdma_add_stat(stats, pcur, "retry_count_exceeded_errors",
+ (u64)tx_qp_err_stats->retry_count_exceeded_errors);
+ pcur += ocrdma_add_stat(stats, pcur, "rnr_retry_count_exceeded_errors",
+ (u64)tx_qp_err_stats->rnr_retry_count_exceeded_errors);
+ return stats;
+}
+
+static char *ocrdma_tx_dbg_stats(struct ocrdma_dev *dev)
+{
+ int i;
+ char *pstats = dev->stats_mem.debugfs_mem;
+ struct ocrdma_rdma_stats_resp *rdma_stats =
+ (struct ocrdma_rdma_stats_resp *)dev->stats_mem.va;
+ struct ocrdma_tx_dbg_stats *tx_dbg_stats =
+ &rdma_stats->tx_dbg_stats;
+
+ memset(pstats, 0, (OCRDMA_MAX_DBGFS_MEM));
+
+ for (i = 0; i < 100; i++)
+ pstats += snprintf(pstats, 80, "DW[%d] = 0x%x\n", i,
+ tx_dbg_stats->data[i]);
+
+ return dev->stats_mem.debugfs_mem;
+}
+
+static char *ocrdma_rx_dbg_stats(struct ocrdma_dev *dev)
+{
+ int i;
+ char *pstats = dev->stats_mem.debugfs_mem;
+ struct ocrdma_rdma_stats_resp *rdma_stats =
+ (struct ocrdma_rdma_stats_resp *)dev->stats_mem.va;
+ struct ocrdma_rx_dbg_stats *rx_dbg_stats =
+ &rdma_stats->rx_dbg_stats;
+
+ memset(pstats, 0, (OCRDMA_MAX_DBGFS_MEM));
+
+ for (i = 0; i < 200; i++)
+ pstats += snprintf(pstats, 80, "DW[%d] = 0x%x\n", i,
+ rx_dbg_stats->data[i]);
+
+ return dev->stats_mem.debugfs_mem;
+}
+
+static void ocrdma_update_stats(struct ocrdma_dev *dev)
+{
+ ulong now = jiffies, secs;
+ int status = 0;
+
+ secs = jiffies_to_msecs(now - dev->last_stats_time) / 1000U;
+ if (secs) {
+ /* update */
+ status = ocrdma_mbx_rdma_stats(dev, false);
+ if (status)
+ pr_err("%s: stats mbox failed with status = %d\n",
+ __func__, status);
+ dev->last_stats_time = jiffies;
+ }
+}
+
+static ssize_t ocrdma_dbgfs_ops_read(struct file *filp, char __user *buffer,
+ size_t usr_buf_len, loff_t *ppos)
+{
+ struct ocrdma_stats *pstats = filp->private_data;
+ struct ocrdma_dev *dev = pstats->dev;
+ ssize_t status = 0;
+ char *data = NULL;
+
+ /* No partial reads */
+ if (*ppos != 0)
+ return 0;
+
+ mutex_lock(&dev->stats_lock);
+
+ ocrdma_update_stats(dev);
+
+ switch (pstats->type) {
+ case OCRDMA_RSRC_STATS:
+ data = ocrdma_resource_stats(dev);
+ break;
+ case OCRDMA_RXSTATS:
+ data = ocrdma_rx_stats(dev);
+ break;
+ case OCRDMA_WQESTATS:
+ data = ocrdma_wqe_stats(dev);
+ break;
+ case OCRDMA_TXSTATS:
+ data = ocrdma_tx_stats(dev);
+ break;
+ case OCRDMA_DB_ERRSTATS:
+ data = ocrdma_db_errstats(dev);
+ break;
+ case OCRDMA_RXQP_ERRSTATS:
+ data = ocrdma_rxqp_errstats(dev);
+ break;
+ case OCRDMA_TXQP_ERRSTATS:
+ data = ocrdma_txqp_errstats(dev);
+ break;
+ case OCRDMA_TX_DBG_STATS:
+ data = ocrdma_tx_dbg_stats(dev);
+ break;
+ case OCRDMA_RX_DBG_STATS:
+ data = ocrdma_rx_dbg_stats(dev);
+ break;
+
+ default:
+ status = -EFAULT;
+ goto exit;
+ }
+
+ if (usr_buf_len < strlen(data)) {
+ status = -ENOSPC;
+ goto exit;
+ }
+
+ status = simple_read_from_buffer(buffer, usr_buf_len, ppos, data,
+ strlen(data));
+exit:
+ mutex_unlock(&dev->stats_lock);
+ return status;
+}
+
+static int ocrdma_debugfs_open(struct inode *inode, struct file *file)
+{
+ if (inode->i_private)
+ file->private_data = inode->i_private;
+ return 0;
+}
+
+static const struct file_operations ocrdma_dbg_ops = {
+ .owner = THIS_MODULE,
+ .open = ocrdma_debugfs_open,
+ .read = ocrdma_dbgfs_ops_read,
+};
+
+void ocrdma_add_port_stats(struct ocrdma_dev *dev)
+{
+ if (!ocrdma_dbgfs_dir)
+ return;
+
+ /* Create post stats base dir */
+ dev->dir = debugfs_create_dir(dev->ibdev.name, ocrdma_dbgfs_dir);
+ if (!dev->dir)
+ goto err;
+
+ dev->rsrc_stats.type = OCRDMA_RSRC_STATS;
+ dev->rsrc_stats.dev = dev;
+ if (!debugfs_create_file("resource_stats", S_IRUSR, dev->dir,
+ &dev->rsrc_stats, &ocrdma_dbg_ops))
+ goto err;
+
+ dev->rx_stats.type = OCRDMA_RXSTATS;
+ dev->rx_stats.dev = dev;
+ if (!debugfs_create_file("rx_stats", S_IRUSR, dev->dir,
+ &dev->rx_stats, &ocrdma_dbg_ops))
+ goto err;
+
+ dev->wqe_stats.type = OCRDMA_WQESTATS;
+ dev->wqe_stats.dev = dev;
+ if (!debugfs_create_file("wqe_stats", S_IRUSR, dev->dir,
+ &dev->wqe_stats, &ocrdma_dbg_ops))
+ goto err;
+
+ dev->tx_stats.type = OCRDMA_TXSTATS;
+ dev->tx_stats.dev = dev;
+ if (!debugfs_create_file("tx_stats", S_IRUSR, dev->dir,
+ &dev->tx_stats, &ocrdma_dbg_ops))
+ goto err;
+
+ dev->db_err_stats.type = OCRDMA_DB_ERRSTATS;
+ dev->db_err_stats.dev = dev;
+ if (!debugfs_create_file("db_err_stats", S_IRUSR, dev->dir,
+ &dev->db_err_stats, &ocrdma_dbg_ops))
+ goto err;
+
+
+ dev->tx_qp_err_stats.type = OCRDMA_TXQP_ERRSTATS;
+ dev->tx_qp_err_stats.dev = dev;
+ if (!debugfs_create_file("tx_qp_err_stats", S_IRUSR, dev->dir,
+ &dev->tx_qp_err_stats, &ocrdma_dbg_ops))
+ goto err;
+
+ dev->rx_qp_err_stats.type = OCRDMA_RXQP_ERRSTATS;
+ dev->rx_qp_err_stats.dev = dev;
+ if (!debugfs_create_file("rx_qp_err_stats", S_IRUSR, dev->dir,
+ &dev->rx_qp_err_stats, &ocrdma_dbg_ops))
+ goto err;
+
+
+ dev->tx_dbg_stats.type = OCRDMA_TX_DBG_STATS;
+ dev->tx_dbg_stats.dev = dev;
+ if (!debugfs_create_file("tx_dbg_stats", S_IRUSR, dev->dir,
+ &dev->tx_dbg_stats, &ocrdma_dbg_ops))
+ goto err;
+
+ dev->rx_dbg_stats.type = OCRDMA_RX_DBG_STATS;
+ dev->rx_dbg_stats.dev = dev;
+ if (!debugfs_create_file("rx_dbg_stats", S_IRUSR, dev->dir,
+ &dev->rx_dbg_stats, &ocrdma_dbg_ops))
+ goto err;
+
+ /* Now create dma_mem for stats mbx command */
+ if (!ocrdma_alloc_stats_mem(dev))
+ goto err;
+
+ mutex_init(&dev->stats_lock);
+
+ return;
+err:
+ ocrdma_release_stats_mem(dev);
+ debugfs_remove_recursive(dev->dir);
+ dev->dir = NULL;
+}
+
+void ocrdma_rem_port_stats(struct ocrdma_dev *dev)
+{
+ if (!dev->dir)
+ return;
+ mutex_destroy(&dev->stats_lock);
+ ocrdma_release_stats_mem(dev);
+ debugfs_remove(dev->dir);
+}
+
+void ocrdma_init_debugfs(void)
+{
+ /* Create base dir in debugfs root dir */
+ ocrdma_dbgfs_dir = debugfs_create_dir("ocrdma", NULL);
+}
+
+void ocrdma_rem_debugfs(void)
+{
+ debugfs_remove_recursive(ocrdma_dbgfs_dir);
+}
diff --git a/drivers/infiniband/hw/ocrdma/ocrdma_stats.h b/drivers/infiniband/hw/ocrdma/ocrdma_stats.h
new file mode 100644
index 000000000000..5f5e20c46d7c
--- /dev/null
+++ b/drivers/infiniband/hw/ocrdma/ocrdma_stats.h
@@ -0,0 +1,54 @@
+/*******************************************************************
+ * This file is part of the Emulex RoCE Device Driver for *
+ * RoCE (RDMA over Converged Ethernet) adapters. *
+ * Copyright (C) 2008-2014 Emulex. All rights reserved. *
+ * EMULEX and SLI are trademarks of Emulex. *
+ * www.emulex.com *
+ * *
+ * This program is free software; you can redistribute it and/or *
+ * modify it under the terms of version 2 of the GNU General *
+ * Public License as published by the Free Software Foundation. *
+ * This program is distributed in the hope that it will be useful. *
+ * ALL EXPRESS OR IMPLIED CONDITIONS, REPRESENTATIONS AND *
+ * WARRANTIES, INCLUDING ANY IMPLIED WARRANTY OF MERCHANTABILITY, *
+ * FITNESS FOR A PARTICULAR PURPOSE, OR NON-INFRINGEMENT, ARE *
+ * DISCLAIMED, EXCEPT TO THE EXTENT THAT SUCH DISCLAIMERS ARE HELD *
+ * TO BE LEGALLY INVALID. See the GNU General Public License for *
+ * more details, a copy of which can be found in the file COPYING *
+ * included with this package. *
+ *
+ * Contact Information:
+ * linux-drivers@emulex.com
+ *
+ * Emulex
+ * 3333 Susan Street
+ * Costa Mesa, CA 92626
+ *******************************************************************/
+
+#ifndef __OCRDMA_STATS_H__
+#define __OCRDMA_STATS_H__
+
+#include <linux/debugfs.h>
+#include "ocrdma.h"
+#include "ocrdma_hw.h"
+
+#define OCRDMA_MAX_DBGFS_MEM 4096
+
+enum OCRDMA_STATS_TYPE {
+ OCRDMA_RSRC_STATS,
+ OCRDMA_RXSTATS,
+ OCRDMA_WQESTATS,
+ OCRDMA_TXSTATS,
+ OCRDMA_DB_ERRSTATS,
+ OCRDMA_RXQP_ERRSTATS,
+ OCRDMA_TXQP_ERRSTATS,
+ OCRDMA_TX_DBG_STATS,
+ OCRDMA_RX_DBG_STATS
+};
+
+void ocrdma_rem_debugfs(void);
+void ocrdma_init_debugfs(void);
+void ocrdma_rem_port_stats(struct ocrdma_dev *dev);
+void ocrdma_add_port_stats(struct ocrdma_dev *dev);
+
+#endif /* __OCRDMA_STATS_H__ */
diff --git a/drivers/infiniband/hw/ocrdma/ocrdma_verbs.c b/drivers/infiniband/hw/ocrdma/ocrdma_verbs.c
index e0cc201be41a..edf6211d84b8 100644
--- a/drivers/infiniband/hw/ocrdma/ocrdma_verbs.c
+++ b/drivers/infiniband/hw/ocrdma/ocrdma_verbs.c
@@ -53,7 +53,7 @@ int ocrdma_query_gid(struct ib_device *ibdev, u8 port,
dev = get_ocrdma_dev(ibdev);
memset(sgid, 0, sizeof(*sgid));
- if (index >= OCRDMA_MAX_SGID)
+ if (index > OCRDMA_MAX_SGID)
return -EINVAL;
memcpy(sgid, &dev->sgid_tbl[index], sizeof(*sgid));
@@ -89,7 +89,7 @@ int ocrdma_query_device(struct ib_device *ibdev, struct ib_device_attr *attr)
attr->max_cq = dev->attr.max_cq;
attr->max_cqe = dev->attr.max_cqe;
attr->max_mr = dev->attr.max_mr;
- attr->max_mw = 0;
+ attr->max_mw = dev->attr.max_mw;
attr->max_pd = dev->attr.max_pd;
attr->atomic_cap = 0;
attr->max_fmr = 0;
@@ -144,7 +144,6 @@ static inline void get_link_speed_and_width(struct ocrdma_dev *dev,
}
}
-
int ocrdma_query_port(struct ib_device *ibdev,
u8 port, struct ib_port_attr *props)
{
@@ -267,7 +266,7 @@ static struct ocrdma_pd *_ocrdma_alloc_pd(struct ocrdma_dev *dev,
if (udata && uctx) {
pd->dpp_enabled =
- dev->nic_info.dev_family == OCRDMA_GEN2_FAMILY;
+ ocrdma_get_asic_type(dev) == OCRDMA_ASIC_GEN_SKH_R;
pd->num_dpp_qp =
pd->dpp_enabled ? OCRDMA_PD_MAX_DPP_ENABLED_QP : 0;
}
@@ -726,10 +725,10 @@ static void build_user_pbes(struct ocrdma_dev *dev, struct ocrdma_mr *mr,
u32 num_pbes)
{
struct ocrdma_pbe *pbe;
- struct ib_umem_chunk *chunk;
+ struct scatterlist *sg;
struct ocrdma_pbl *pbl_tbl = mr->hwmr.pbl_table;
struct ib_umem *umem = mr->umem;
- int i, shift, pg_cnt, pages, pbe_cnt, total_num_pbes = 0;
+ int shift, pg_cnt, pages, pbe_cnt, entry, total_num_pbes = 0;
if (!mr->hwmr.num_pbes)
return;
@@ -739,39 +738,37 @@ static void build_user_pbes(struct ocrdma_dev *dev, struct ocrdma_mr *mr,
shift = ilog2(umem->page_size);
- list_for_each_entry(chunk, &umem->chunk_list, list) {
- /* get all the dma regions from the chunk. */
- for (i = 0; i < chunk->nmap; i++) {
- pages = sg_dma_len(&chunk->page_list[i]) >> shift;
- for (pg_cnt = 0; pg_cnt < pages; pg_cnt++) {
- /* store the page address in pbe */
- pbe->pa_lo =
- cpu_to_le32(sg_dma_address
- (&chunk->page_list[i]) +
- (umem->page_size * pg_cnt));
- pbe->pa_hi =
- cpu_to_le32(upper_32_bits
- ((sg_dma_address
- (&chunk->page_list[i]) +
- umem->page_size * pg_cnt)));
- pbe_cnt += 1;
- total_num_pbes += 1;
- pbe++;
-
- /* if done building pbes, issue the mbx cmd. */
- if (total_num_pbes == num_pbes)
- return;
-
- /* if the given pbl is full storing the pbes,
- * move to next pbl.
- */
- if (pbe_cnt ==
- (mr->hwmr.pbl_size / sizeof(u64))) {
- pbl_tbl++;
- pbe = (struct ocrdma_pbe *)pbl_tbl->va;
- pbe_cnt = 0;
- }
+ for_each_sg(umem->sg_head.sgl, sg, umem->nmap, entry) {
+ pages = sg_dma_len(sg) >> shift;
+ for (pg_cnt = 0; pg_cnt < pages; pg_cnt++) {
+ /* store the page address in pbe */
+ pbe->pa_lo =
+ cpu_to_le32(sg_dma_address
+ (sg) +
+ (umem->page_size * pg_cnt));
+ pbe->pa_hi =
+ cpu_to_le32(upper_32_bits
+ ((sg_dma_address
+ (sg) +
+ umem->page_size * pg_cnt)));
+ pbe_cnt += 1;
+ total_num_pbes += 1;
+ pbe++;
+
+ /* if done building pbes, issue the mbx cmd. */
+ if (total_num_pbes == num_pbes)
+ return;
+
+ /* if the given pbl is full storing the pbes,
+ * move to next pbl.
+ */
+ if (pbe_cnt ==
+ (mr->hwmr.pbl_size / sizeof(u64))) {
+ pbl_tbl++;
+ pbe = (struct ocrdma_pbe *)pbl_tbl->va;
+ pbe_cnt = 0;
}
+
}
}
}
@@ -840,8 +837,7 @@ int ocrdma_dereg_mr(struct ib_mr *ib_mr)
status = ocrdma_mbx_dealloc_lkey(dev, mr->hwmr.fr_mr, mr->hwmr.lkey);
- if (mr->hwmr.fr_mr == 0)
- ocrdma_free_mr_pbl_tbl(dev, &mr->hwmr);
+ ocrdma_free_mr_pbl_tbl(dev, &mr->hwmr);
/* it could be user registered memory. */
if (mr->umem)
@@ -910,6 +906,7 @@ struct ib_cq *ocrdma_create_cq(struct ib_device *ibdev, int entries, int vector,
spin_lock_init(&cq->comp_handler_lock);
INIT_LIST_HEAD(&cq->sq_head);
INIT_LIST_HEAD(&cq->rq_head);
+ cq->first_arm = true;
if (ib_ctx) {
uctx = get_ocrdma_ucontext(ib_ctx);
@@ -927,9 +924,7 @@ struct ib_cq *ocrdma_create_cq(struct ib_device *ibdev, int entries, int vector,
goto ctx_err;
}
cq->phase = OCRDMA_CQE_VALID;
- cq->arm_needed = true;
dev->cq_tbl[cq->id] = cq;
-
return &cq->ibcq;
ctx_err:
@@ -952,15 +947,52 @@ int ocrdma_resize_cq(struct ib_cq *ibcq, int new_cnt,
return status;
}
+static void ocrdma_flush_cq(struct ocrdma_cq *cq)
+{
+ int cqe_cnt;
+ int valid_count = 0;
+ unsigned long flags;
+
+ struct ocrdma_dev *dev = get_ocrdma_dev(cq->ibcq.device);
+ struct ocrdma_cqe *cqe = NULL;
+
+ cqe = cq->va;
+ cqe_cnt = cq->cqe_cnt;
+
+ /* Last irq might have scheduled a polling thread
+ * sync-up with it before hard flushing.
+ */
+ spin_lock_irqsave(&cq->cq_lock, flags);
+ while (cqe_cnt) {
+ if (is_cqe_valid(cq, cqe))
+ valid_count++;
+ cqe++;
+ cqe_cnt--;
+ }
+ ocrdma_ring_cq_db(dev, cq->id, false, false, valid_count);
+ spin_unlock_irqrestore(&cq->cq_lock, flags);
+}
+
int ocrdma_destroy_cq(struct ib_cq *ibcq)
{
int status;
struct ocrdma_cq *cq = get_ocrdma_cq(ibcq);
+ struct ocrdma_eq *eq = NULL;
struct ocrdma_dev *dev = get_ocrdma_dev(ibcq->device);
int pdid = 0;
+ u32 irq, indx;
- status = ocrdma_mbx_destroy_cq(dev, cq);
+ dev->cq_tbl[cq->id] = NULL;
+ indx = ocrdma_get_eq_table_index(dev, cq->eqn);
+ if (indx == -EINVAL)
+ BUG();
+ eq = &dev->eq_tbl[indx];
+ irq = ocrdma_get_irq(dev, eq);
+ synchronize_irq(irq);
+ ocrdma_flush_cq(cq);
+
+ status = ocrdma_mbx_destroy_cq(dev, cq);
if (cq->ucontext) {
pdid = cq->ucontext->cntxt_pd->id;
ocrdma_del_mmap(cq->ucontext, (u64) cq->pa,
@@ -969,7 +1001,6 @@ int ocrdma_destroy_cq(struct ib_cq *ibcq)
ocrdma_get_db_addr(dev, pdid),
dev->nic_info.db_page_size);
}
- dev->cq_tbl[cq->id] = NULL;
kfree(cq);
return status;
@@ -1092,15 +1123,9 @@ static int ocrdma_copy_qp_uresp(struct ocrdma_qp *qp,
}
uresp.db_page_addr = usr_db;
uresp.db_page_size = dev->nic_info.db_page_size;
- if (dev->nic_info.dev_family == OCRDMA_GEN2_FAMILY) {
- uresp.db_sq_offset = OCRDMA_DB_GEN2_SQ_OFFSET;
- uresp.db_rq_offset = OCRDMA_DB_GEN2_RQ_OFFSET;
- uresp.db_shift = 24;
- } else {
- uresp.db_sq_offset = OCRDMA_DB_SQ_OFFSET;
- uresp.db_rq_offset = OCRDMA_DB_RQ_OFFSET;
- uresp.db_shift = 16;
- }
+ uresp.db_sq_offset = OCRDMA_DB_GEN2_SQ_OFFSET;
+ uresp.db_rq_offset = OCRDMA_DB_GEN2_RQ_OFFSET;
+ uresp.db_shift = OCRDMA_DB_RQ_SHIFT;
if (qp->dpp_enabled) {
uresp.dpp_credit = dpp_credit_lmt;
@@ -1132,7 +1157,7 @@ err:
static void ocrdma_set_qp_db(struct ocrdma_dev *dev, struct ocrdma_qp *qp,
struct ocrdma_pd *pd)
{
- if (dev->nic_info.dev_family == OCRDMA_GEN2_FAMILY) {
+ if (ocrdma_get_asic_type(dev) == OCRDMA_ASIC_GEN_SKH_R) {
qp->sq_db = dev->nic_info.db +
(pd->id * dev->nic_info.db_page_size) +
OCRDMA_DB_GEN2_SQ_OFFSET;
@@ -1182,7 +1207,6 @@ static void ocrdma_set_qp_init_params(struct ocrdma_qp *qp,
qp->signaled = (attrs->sq_sig_type == IB_SIGNAL_ALL_WR) ? true : false;
}
-
static void ocrdma_store_gsi_qp_cq(struct ocrdma_dev *dev,
struct ib_qp_init_attr *attrs)
{
@@ -1268,17 +1292,6 @@ gen_err:
return ERR_PTR(status);
}
-
-static void ocrdma_flush_rq_db(struct ocrdma_qp *qp)
-{
- if (qp->db_cache) {
- u32 val = qp->rq.dbid | (qp->db_cache <<
- ocrdma_get_num_posted_shift(qp));
- iowrite32(val, qp->rq_db);
- qp->db_cache = 0;
- }
-}
-
int _ocrdma_modify_qp(struct ib_qp *ibqp, struct ib_qp_attr *attr,
int attr_mask)
{
@@ -1296,9 +1309,7 @@ int _ocrdma_modify_qp(struct ib_qp *ibqp, struct ib_qp_attr *attr,
*/
if (status < 0)
return status;
- status = ocrdma_mbx_modify_qp(dev, qp, attr, attr_mask, old_qps);
- if (!status && attr_mask & IB_QP_STATE && attr->qp_state == IB_QPS_RTR)
- ocrdma_flush_rq_db(qp);
+ status = ocrdma_mbx_modify_qp(dev, qp, attr, attr_mask);
return status;
}
@@ -1510,7 +1521,7 @@ static void ocrdma_discard_cqes(struct ocrdma_qp *qp, struct ocrdma_cq *cq)
int discard_cnt = 0;
u32 cur_getp, stop_getp;
struct ocrdma_cqe *cqe;
- u32 qpn = 0;
+ u32 qpn = 0, wqe_idx = 0;
spin_lock_irqsave(&cq->cq_lock, cq_flags);
@@ -1539,24 +1550,29 @@ static void ocrdma_discard_cqes(struct ocrdma_qp *qp, struct ocrdma_cq *cq)
if (qpn == 0 || qpn != qp->id)
goto skip_cqe;
- /* mark cqe discarded so that it is not picked up later
- * in the poll_cq().
- */
- discard_cnt += 1;
- cqe->cmn.qpn = 0;
if (is_cqe_for_sq(cqe)) {
ocrdma_hwq_inc_tail(&qp->sq);
} else {
if (qp->srq) {
+ wqe_idx = (le32_to_cpu(cqe->rq.buftag_qpn) >>
+ OCRDMA_CQE_BUFTAG_SHIFT) &
+ qp->srq->rq.max_wqe_idx;
+ if (wqe_idx < 1)
+ BUG();
spin_lock_irqsave(&qp->srq->q_lock, flags);
ocrdma_hwq_inc_tail(&qp->srq->rq);
- ocrdma_srq_toggle_bit(qp->srq, cur_getp);
+ ocrdma_srq_toggle_bit(qp->srq, wqe_idx - 1);
spin_unlock_irqrestore(&qp->srq->q_lock, flags);
} else {
ocrdma_hwq_inc_tail(&qp->rq);
}
}
+ /* mark cqe discarded so that it is not picked up later
+ * in the poll_cq().
+ */
+ discard_cnt += 1;
+ cqe->cmn.qpn = 0;
skip_cqe:
cur_getp = (cur_getp + 1) % cq->max_hw_cqe;
} while (cur_getp != stop_getp);
@@ -1659,7 +1675,7 @@ static int ocrdma_copy_srq_uresp(struct ocrdma_dev *dev, struct ocrdma_srq *srq,
(srq->pd->id * dev->nic_info.db_page_size);
uresp.db_page_size = dev->nic_info.db_page_size;
uresp.num_rqe_allocated = srq->rq.max_cnt;
- if (dev->nic_info.dev_family == OCRDMA_GEN2_FAMILY) {
+ if (ocrdma_get_asic_type(dev) == OCRDMA_ASIC_GEN_SKH_R) {
uresp.db_rq_offset = OCRDMA_DB_GEN2_RQ_OFFSET;
uresp.db_shift = 24;
} else {
@@ -2009,15 +2025,15 @@ static int ocrdma_build_fr(struct ocrdma_qp *qp, struct ocrdma_hdr_wqe *hdr,
fast_reg->num_sges = wr->wr.fast_reg.page_list_len;
fast_reg->size_sge =
get_encoded_page_size(1 << wr->wr.fast_reg.page_shift);
- mr = (struct ocrdma_mr *) (unsigned long) qp->dev->stag_arr[(hdr->lkey >> 8) &
- (OCRDMA_MAX_STAG - 1)];
+ mr = (struct ocrdma_mr *) (unsigned long)
+ qp->dev->stag_arr[(hdr->lkey >> 8) & (OCRDMA_MAX_STAG - 1)];
build_frmr_pbes(wr, mr->hwmr.pbl_table, &mr->hwmr);
return 0;
}
static void ocrdma_ring_sq_db(struct ocrdma_qp *qp)
{
- u32 val = qp->sq.dbid | (1 << 16);
+ u32 val = qp->sq.dbid | (1 << OCRDMA_DB_SQ_SHIFT);
iowrite32(val, qp->sq_db);
}
@@ -2122,12 +2138,9 @@ int ocrdma_post_send(struct ib_qp *ibqp, struct ib_send_wr *wr,
static void ocrdma_ring_rq_db(struct ocrdma_qp *qp)
{
- u32 val = qp->rq.dbid | (1 << ocrdma_get_num_posted_shift(qp));
+ u32 val = qp->rq.dbid | (1 << OCRDMA_DB_RQ_SHIFT);
- if (qp->state != OCRDMA_QPS_INIT)
- iowrite32(val, qp->rq_db);
- else
- qp->db_cache++;
+ iowrite32(val, qp->rq_db);
}
static void ocrdma_build_rqe(struct ocrdma_hdr_wqe *rqe, struct ib_recv_wr *wr,
@@ -2213,7 +2226,7 @@ static int ocrdma_srq_get_idx(struct ocrdma_srq *srq)
if (row == srq->bit_fields_len)
BUG();
- return indx;
+ return indx + 1; /* Use from index 1 */
}
static void ocrdma_ring_srq_db(struct ocrdma_srq *srq)
@@ -2550,10 +2563,13 @@ static void ocrdma_update_free_srq_cqe(struct ib_wc *ibwc,
srq = get_ocrdma_srq(qp->ibqp.srq);
wqe_idx = (le32_to_cpu(cqe->rq.buftag_qpn) >>
- OCRDMA_CQE_BUFTAG_SHIFT) & srq->rq.max_wqe_idx;
+ OCRDMA_CQE_BUFTAG_SHIFT) & srq->rq.max_wqe_idx;
+ if (wqe_idx < 1)
+ BUG();
+
ibwc->wr_id = srq->rqe_wr_id_tbl[wqe_idx];
spin_lock_irqsave(&srq->q_lock, flags);
- ocrdma_srq_toggle_bit(srq, wqe_idx);
+ ocrdma_srq_toggle_bit(srq, wqe_idx - 1);
spin_unlock_irqrestore(&srq->q_lock, flags);
ocrdma_hwq_inc_tail(&srq->rq);
}
@@ -2705,10 +2721,18 @@ expand_cqe:
}
stop_cqe:
cq->getp = cur_getp;
- if (polled_hw_cqes || expand || stop) {
- ocrdma_ring_cq_db(dev, cq->id, cq->armed, cq->solicited,
+ if (cq->deferred_arm) {
+ ocrdma_ring_cq_db(dev, cq->id, true, cq->deferred_sol,
+ polled_hw_cqes);
+ cq->deferred_arm = false;
+ cq->deferred_sol = false;
+ } else {
+ /* We need to pop the CQE. No need to arm */
+ ocrdma_ring_cq_db(dev, cq->id, false, cq->deferred_sol,
polled_hw_cqes);
+ cq->deferred_sol = false;
}
+
return i;
}
@@ -2780,30 +2804,28 @@ int ocrdma_arm_cq(struct ib_cq *ibcq, enum ib_cq_notify_flags cq_flags)
struct ocrdma_cq *cq = get_ocrdma_cq(ibcq);
struct ocrdma_dev *dev = get_ocrdma_dev(ibcq->device);
u16 cq_id;
- u16 cur_getp;
- struct ocrdma_cqe *cqe;
unsigned long flags;
+ bool arm_needed = false, sol_needed = false;
cq_id = cq->id;
spin_lock_irqsave(&cq->cq_lock, flags);
if (cq_flags & IB_CQ_NEXT_COMP || cq_flags & IB_CQ_SOLICITED)
- cq->armed = true;
+ arm_needed = true;
if (cq_flags & IB_CQ_SOLICITED)
- cq->solicited = true;
-
- cur_getp = cq->getp;
- cqe = cq->va + cur_getp;
+ sol_needed = true;
- /* check whether any valid cqe exist or not, if not then safe to
- * arm. If cqe is not yet consumed, then let it get consumed and then
- * we arm it to avoid false interrupts.
- */
- if (!is_cqe_valid(cq, cqe) || cq->arm_needed) {
- cq->arm_needed = false;
- ocrdma_ring_cq_db(dev, cq_id, cq->armed, cq->solicited, 0);
+ if (cq->first_arm) {
+ ocrdma_ring_cq_db(dev, cq_id, arm_needed, sol_needed, 0);
+ cq->first_arm = false;
+ goto skip_defer;
}
+ cq->deferred_arm = true;
+
+skip_defer:
+ cq->deferred_sol = sol_needed;
spin_unlock_irqrestore(&cq->cq_lock, flags);
+
return 0;
}
@@ -2838,7 +2860,8 @@ struct ib_mr *ocrdma_alloc_frmr(struct ib_pd *ibpd, int max_page_list_len)
goto mbx_err;
mr->ibmr.rkey = mr->hwmr.lkey;
mr->ibmr.lkey = mr->hwmr.lkey;
- dev->stag_arr[(mr->hwmr.lkey >> 8) & (OCRDMA_MAX_STAG - 1)] = mr;
+ dev->stag_arr[(mr->hwmr.lkey >> 8) & (OCRDMA_MAX_STAG - 1)] =
+ (unsigned long) mr;
return &mr->ibmr;
mbx_err:
ocrdma_free_mr_pbl_tbl(dev, &mr->hwmr);
diff --git a/drivers/infiniband/hw/qib/qib.h b/drivers/infiniband/hw/qib/qib.h
index 1946101419a3..c00ae093b6f8 100644
--- a/drivers/infiniband/hw/qib/qib.h
+++ b/drivers/infiniband/hw/qib/qib.h
@@ -868,8 +868,10 @@ struct qib_devdata {
/* last buffer for user use */
u32 lastctxt_piobuf;
- /* saturating counter of (non-port-specific) device interrupts */
- u32 int_counter;
+ /* reset value */
+ u64 z_int_counter;
+ /* percpu intcounter */
+ u64 __percpu *int_counter;
/* pio bufs allocated per ctxt */
u32 pbufsctxt;
@@ -1184,7 +1186,7 @@ int qib_setup_eagerbufs(struct qib_ctxtdata *);
void qib_set_ctxtcnt(struct qib_devdata *);
int qib_create_ctxts(struct qib_devdata *dd);
struct qib_ctxtdata *qib_create_ctxtdata(struct qib_pportdata *, u32, int);
-void qib_init_pportdata(struct qib_pportdata *, struct qib_devdata *, u8, u8);
+int qib_init_pportdata(struct qib_pportdata *, struct qib_devdata *, u8, u8);
void qib_free_ctxtdata(struct qib_devdata *, struct qib_ctxtdata *);
u32 qib_kreceive(struct qib_ctxtdata *, u32 *, u32 *);
@@ -1449,6 +1451,10 @@ void qib_nomsi(struct qib_devdata *);
void qib_nomsix(struct qib_devdata *);
void qib_pcie_getcmd(struct qib_devdata *, u16 *, u8 *, u8 *);
void qib_pcie_reenable(struct qib_devdata *, u16, u8, u8);
+/* interrupts for device */
+u64 qib_int_counter(struct qib_devdata *);
+/* interrupt for all devices */
+u64 qib_sps_ints(void);
/*
* dma_addr wrappers - all 0's invalid for hw
diff --git a/drivers/infiniband/hw/qib/qib_diag.c b/drivers/infiniband/hw/qib/qib_diag.c
index 1686fd4bda87..5dfda4c5cc9c 100644
--- a/drivers/infiniband/hw/qib/qib_diag.c
+++ b/drivers/infiniband/hw/qib/qib_diag.c
@@ -546,7 +546,7 @@ static ssize_t qib_diagpkt_write(struct file *fp,
size_t count, loff_t *off)
{
u32 __iomem *piobuf;
- u32 plen, clen, pbufn;
+ u32 plen, pbufn, maxlen_reserve;
struct qib_diag_xpkt dp;
u32 *tmpbuf = NULL;
struct qib_devdata *dd;
@@ -590,15 +590,20 @@ static ssize_t qib_diagpkt_write(struct file *fp,
}
ppd = &dd->pport[dp.port - 1];
- /* need total length before first word written */
- /* +1 word is for the qword padding */
- plen = sizeof(u32) + dp.len;
- clen = dp.len >> 2;
-
- if ((plen + 4) > ppd->ibmaxlen) {
+ /*
+ * need total length before first word written, plus 2 Dwords. One Dword
+ * is for padding so we get the full user data when not aligned on
+ * a word boundary. The other Dword is to make sure we have room for the
+ * ICRC which gets tacked on later.
+ */
+ maxlen_reserve = 2 * sizeof(u32);
+ if (dp.len > ppd->ibmaxlen - maxlen_reserve) {
ret = -EINVAL;
- goto bail; /* before writing pbc */
+ goto bail;
}
+
+ plen = sizeof(u32) + dp.len;
+
tmpbuf = vmalloc(plen);
if (!tmpbuf) {
qib_devinfo(dd->pcidev,
@@ -638,11 +643,11 @@ static ssize_t qib_diagpkt_write(struct file *fp,
*/
if (dd->flags & QIB_PIO_FLUSH_WC) {
qib_flush_wc();
- qib_pio_copy(piobuf + 2, tmpbuf, clen - 1);
+ qib_pio_copy(piobuf + 2, tmpbuf, plen - 1);
qib_flush_wc();
- __raw_writel(tmpbuf[clen - 1], piobuf + clen + 1);
+ __raw_writel(tmpbuf[plen - 1], piobuf + plen + 1);
} else
- qib_pio_copy(piobuf + 2, tmpbuf, clen);
+ qib_pio_copy(piobuf + 2, tmpbuf, plen);
if (dd->flags & QIB_USE_SPCL_TRIG) {
u32 spcl_off = (pbufn >= dd->piobcnt2k) ? 2047 : 1023;
@@ -689,28 +694,23 @@ int qib_register_observer(struct qib_devdata *dd,
const struct diag_observer *op)
{
struct diag_observer_list_elt *olp;
- int ret = -EINVAL;
+ unsigned long flags;
if (!dd || !op)
- goto bail;
- ret = -ENOMEM;
+ return -EINVAL;
olp = vmalloc(sizeof *olp);
if (!olp) {
pr_err("vmalloc for observer failed\n");
- goto bail;
+ return -ENOMEM;
}
- if (olp) {
- unsigned long flags;
- spin_lock_irqsave(&dd->qib_diag_trans_lock, flags);
- olp->op = op;
- olp->next = dd->diag_observer_list;
- dd->diag_observer_list = olp;
- spin_unlock_irqrestore(&dd->qib_diag_trans_lock, flags);
- ret = 0;
- }
-bail:
- return ret;
+ spin_lock_irqsave(&dd->qib_diag_trans_lock, flags);
+ olp->op = op;
+ olp->next = dd->diag_observer_list;
+ dd->diag_observer_list = olp;
+ spin_unlock_irqrestore(&dd->qib_diag_trans_lock, flags);
+
+ return 0;
}
/* Remove all registered observers when device is closed */
diff --git a/drivers/infiniband/hw/qib/qib_dma.c b/drivers/infiniband/hw/qib/qib_dma.c
index 2920bb39a65b..59fe092b4b0f 100644
--- a/drivers/infiniband/hw/qib/qib_dma.c
+++ b/drivers/infiniband/hw/qib/qib_dma.c
@@ -108,6 +108,10 @@ static int qib_map_sg(struct ib_device *dev, struct scatterlist *sgl,
ret = 0;
break;
}
+ sg->dma_address = addr + sg->offset;
+#ifdef CONFIG_NEED_SG_DMA_LENGTH
+ sg->dma_length = sg->length;
+#endif
}
return ret;
}
@@ -119,21 +123,6 @@ static void qib_unmap_sg(struct ib_device *dev,
BUG_ON(!valid_dma_direction(direction));
}
-static u64 qib_sg_dma_address(struct ib_device *dev, struct scatterlist *sg)
-{
- u64 addr = (u64) page_address(sg_page(sg));
-
- if (addr)
- addr += sg->offset;
- return addr;
-}
-
-static unsigned int qib_sg_dma_len(struct ib_device *dev,
- struct scatterlist *sg)
-{
- return sg->length;
-}
-
static void qib_sync_single_for_cpu(struct ib_device *dev, u64 addr,
size_t size, enum dma_data_direction dir)
{
@@ -173,8 +162,6 @@ struct ib_dma_mapping_ops qib_dma_mapping_ops = {
.unmap_page = qib_dma_unmap_page,
.map_sg = qib_map_sg,
.unmap_sg = qib_unmap_sg,
- .dma_address = qib_sg_dma_address,
- .dma_len = qib_sg_dma_len,
.sync_single_for_cpu = qib_sync_single_for_cpu,
.sync_single_for_device = qib_sync_single_for_device,
.alloc_coherent = qib_dma_alloc_coherent,
diff --git a/drivers/infiniband/hw/qib/qib_file_ops.c b/drivers/infiniband/hw/qib/qib_file_ops.c
index 275f247f9fca..b15e34eeef68 100644
--- a/drivers/infiniband/hw/qib/qib_file_ops.c
+++ b/drivers/infiniband/hw/qib/qib_file_ops.c
@@ -1459,7 +1459,7 @@ static int get_a_ctxt(struct file *fp, const struct qib_user_info *uinfo,
cused++;
else
cfree++;
- if (pusable && cfree && cused < inuse) {
+ if (cfree && cused < inuse) {
udd = dd;
inuse = cused;
}
@@ -1578,7 +1578,7 @@ static int do_qib_user_sdma_queue_create(struct file *fp)
struct qib_ctxtdata *rcd = fd->rcd;
struct qib_devdata *dd = rcd->dd;
- if (dd->flags & QIB_HAS_SEND_DMA)
+ if (dd->flags & QIB_HAS_SEND_DMA) {
fd->pq = qib_user_sdma_queue_create(&dd->pcidev->dev,
dd->unit,
@@ -1586,6 +1586,7 @@ static int do_qib_user_sdma_queue_create(struct file *fp)
fd->subctxt);
if (!fd->pq)
return -ENOMEM;
+ }
return 0;
}
diff --git a/drivers/infiniband/hw/qib/qib_fs.c b/drivers/infiniband/hw/qib/qib_fs.c
index c61e2a92b3c1..cab610ccd50e 100644
--- a/drivers/infiniband/hw/qib/qib_fs.c
+++ b/drivers/infiniband/hw/qib/qib_fs.c
@@ -105,6 +105,7 @@ static int create_file(const char *name, umode_t mode,
static ssize_t driver_stats_read(struct file *file, char __user *buf,
size_t count, loff_t *ppos)
{
+ qib_stats.sps_ints = qib_sps_ints();
return simple_read_from_buffer(buf, count, ppos, &qib_stats,
sizeof qib_stats);
}
diff --git a/drivers/infiniband/hw/qib/qib_iba6120.c b/drivers/infiniband/hw/qib/qib_iba6120.c
index 84e593d6007b..d68266ac7619 100644
--- a/drivers/infiniband/hw/qib/qib_iba6120.c
+++ b/drivers/infiniband/hw/qib/qib_iba6120.c
@@ -1634,9 +1634,7 @@ static irqreturn_t qib_6120intr(int irq, void *data)
goto bail;
}
- qib_stats.sps_ints++;
- if (dd->int_counter != (u32) -1)
- dd->int_counter++;
+ this_cpu_inc(*dd->int_counter);
if (unlikely(istat & (~QLOGIC_IB_I_BITSEXTANT |
QLOGIC_IB_I_GPIO | QLOGIC_IB_I_ERROR)))
@@ -1808,7 +1806,8 @@ static int qib_6120_setup_reset(struct qib_devdata *dd)
* isn't set.
*/
dd->flags &= ~(QIB_INITTED | QIB_PRESENT);
- dd->int_counter = 0; /* so we check interrupts work again */
+ /* so we check interrupts work again */
+ dd->z_int_counter = qib_int_counter(dd);
val = dd->control | QLOGIC_IB_C_RESET;
writeq(val, &dd->kregbase[kr_control]);
mb(); /* prevent compiler re-ordering around actual reset */
@@ -3266,7 +3265,9 @@ static int init_6120_variables(struct qib_devdata *dd)
dd->eep_st_masks[2].errs_to_log = ERR_MASK(ResetNegated);
- qib_init_pportdata(ppd, dd, 0, 1);
+ ret = qib_init_pportdata(ppd, dd, 0, 1);
+ if (ret)
+ goto bail;
ppd->link_width_supported = IB_WIDTH_1X | IB_WIDTH_4X;
ppd->link_speed_supported = QIB_IB_SDR;
ppd->link_width_enabled = IB_WIDTH_4X;
diff --git a/drivers/infiniband/hw/qib/qib_iba7220.c b/drivers/infiniband/hw/qib/qib_iba7220.c
index 454c2e7668fe..7dec89fdc124 100644
--- a/drivers/infiniband/hw/qib/qib_iba7220.c
+++ b/drivers/infiniband/hw/qib/qib_iba7220.c
@@ -1962,10 +1962,7 @@ static irqreturn_t qib_7220intr(int irq, void *data)
goto bail;
}
- qib_stats.sps_ints++;
- if (dd->int_counter != (u32) -1)
- dd->int_counter++;
-
+ this_cpu_inc(*dd->int_counter);
if (unlikely(istat & (~QLOGIC_IB_I_BITSEXTANT |
QLOGIC_IB_I_GPIO | QLOGIC_IB_I_ERROR)))
unlikely_7220_intr(dd, istat);
@@ -2120,7 +2117,8 @@ static int qib_setup_7220_reset(struct qib_devdata *dd)
* isn't set.
*/
dd->flags &= ~(QIB_INITTED | QIB_PRESENT);
- dd->int_counter = 0; /* so we check interrupts work again */
+ /* so we check interrupts work again */
+ dd->z_int_counter = qib_int_counter(dd);
val = dd->control | QLOGIC_IB_C_RESET;
writeq(val, &dd->kregbase[kr_control]);
mb(); /* prevent compiler reordering around actual reset */
@@ -4061,7 +4059,9 @@ static int qib_init_7220_variables(struct qib_devdata *dd)
init_waitqueue_head(&cpspec->autoneg_wait);
INIT_DELAYED_WORK(&cpspec->autoneg_work, autoneg_7220_work);
- qib_init_pportdata(ppd, dd, 0, 1);
+ ret = qib_init_pportdata(ppd, dd, 0, 1);
+ if (ret)
+ goto bail;
ppd->link_width_supported = IB_WIDTH_1X | IB_WIDTH_4X;
ppd->link_speed_supported = QIB_IB_SDR | QIB_IB_DDR;
diff --git a/drivers/infiniband/hw/qib/qib_iba7322.c b/drivers/infiniband/hw/qib/qib_iba7322.c
index d1bd21319d7d..a7eb32517a04 100644
--- a/drivers/infiniband/hw/qib/qib_iba7322.c
+++ b/drivers/infiniband/hw/qib/qib_iba7322.c
@@ -3115,9 +3115,7 @@ static irqreturn_t qib_7322intr(int irq, void *data)
goto bail;
}
- qib_stats.sps_ints++;
- if (dd->int_counter != (u32) -1)
- dd->int_counter++;
+ this_cpu_inc(*dd->int_counter);
/* handle "errors" of various kinds first, device ahead of port */
if (unlikely(istat & (~QIB_I_BITSEXTANT | QIB_I_GPIO |
@@ -3186,9 +3184,7 @@ static irqreturn_t qib_7322pintr(int irq, void *data)
*/
return IRQ_HANDLED;
- qib_stats.sps_ints++;
- if (dd->int_counter != (u32) -1)
- dd->int_counter++;
+ this_cpu_inc(*dd->int_counter);
/* Clear the interrupt bit we expect to be set. */
qib_write_kreg(dd, kr_intclear, ((1ULL << QIB_I_RCVAVAIL_LSB) |
@@ -3215,9 +3211,7 @@ static irqreturn_t qib_7322bufavail(int irq, void *data)
*/
return IRQ_HANDLED;
- qib_stats.sps_ints++;
- if (dd->int_counter != (u32) -1)
- dd->int_counter++;
+ this_cpu_inc(*dd->int_counter);
/* Clear the interrupt bit we expect to be set. */
qib_write_kreg(dd, kr_intclear, QIB_I_SPIOBUFAVAIL);
@@ -3248,9 +3242,7 @@ static irqreturn_t sdma_intr(int irq, void *data)
*/
return IRQ_HANDLED;
- qib_stats.sps_ints++;
- if (dd->int_counter != (u32) -1)
- dd->int_counter++;
+ this_cpu_inc(*dd->int_counter);
/* Clear the interrupt bit we expect to be set. */
qib_write_kreg(dd, kr_intclear, ppd->hw_pidx ?
@@ -3277,9 +3269,7 @@ static irqreturn_t sdma_idle_intr(int irq, void *data)
*/
return IRQ_HANDLED;
- qib_stats.sps_ints++;
- if (dd->int_counter != (u32) -1)
- dd->int_counter++;
+ this_cpu_inc(*dd->int_counter);
/* Clear the interrupt bit we expect to be set. */
qib_write_kreg(dd, kr_intclear, ppd->hw_pidx ?
@@ -3306,9 +3296,7 @@ static irqreturn_t sdma_progress_intr(int irq, void *data)
*/
return IRQ_HANDLED;
- qib_stats.sps_ints++;
- if (dd->int_counter != (u32) -1)
- dd->int_counter++;
+ this_cpu_inc(*dd->int_counter);
/* Clear the interrupt bit we expect to be set. */
qib_write_kreg(dd, kr_intclear, ppd->hw_pidx ?
@@ -3336,9 +3324,7 @@ static irqreturn_t sdma_cleanup_intr(int irq, void *data)
*/
return IRQ_HANDLED;
- qib_stats.sps_ints++;
- if (dd->int_counter != (u32) -1)
- dd->int_counter++;
+ this_cpu_inc(*dd->int_counter);
/* Clear the interrupt bit we expect to be set. */
qib_write_kreg(dd, kr_intclear, ppd->hw_pidx ?
@@ -3723,7 +3709,8 @@ static int qib_do_7322_reset(struct qib_devdata *dd)
dd->pport->cpspec->ibsymdelta = 0;
dd->pport->cpspec->iblnkerrdelta = 0;
dd->pport->cpspec->ibmalfdelta = 0;
- dd->int_counter = 0; /* so we check interrupts work again */
+ /* so we check interrupts work again */
+ dd->z_int_counter = qib_int_counter(dd);
/*
* Keep chip from being accessed until we are ready. Use
@@ -6557,7 +6544,11 @@ static int qib_init_7322_variables(struct qib_devdata *dd)
}
dd->num_pports++;
- qib_init_pportdata(ppd, dd, pidx, dd->num_pports);
+ ret = qib_init_pportdata(ppd, dd, pidx, dd->num_pports);
+ if (ret) {
+ dd->num_pports--;
+ goto bail;
+ }
ppd->link_width_supported = IB_WIDTH_1X | IB_WIDTH_4X;
ppd->link_width_enabled = IB_WIDTH_4X;
diff --git a/drivers/infiniband/hw/qib/qib_init.c b/drivers/infiniband/hw/qib/qib_init.c
index 24e802f4ea2f..5b7aeb224a30 100644
--- a/drivers/infiniband/hw/qib/qib_init.c
+++ b/drivers/infiniband/hw/qib/qib_init.c
@@ -130,7 +130,6 @@ void qib_set_ctxtcnt(struct qib_devdata *dd)
int qib_create_ctxts(struct qib_devdata *dd)
{
unsigned i;
- int ret;
int local_node_id = pcibus_to_node(dd->pcidev->bus);
if (local_node_id < 0)
@@ -145,8 +144,7 @@ int qib_create_ctxts(struct qib_devdata *dd)
if (!dd->rcd) {
qib_dev_err(dd,
"Unable to allocate ctxtdata array, failing\n");
- ret = -ENOMEM;
- goto done;
+ return -ENOMEM;
}
/* create (one or more) kctxt */
@@ -163,15 +161,14 @@ int qib_create_ctxts(struct qib_devdata *dd)
if (!rcd) {
qib_dev_err(dd,
"Unable to allocate ctxtdata for Kernel ctxt, failing\n");
- ret = -ENOMEM;
- goto done;
+ kfree(dd->rcd);
+ dd->rcd = NULL;
+ return -ENOMEM;
}
rcd->pkeys[0] = QIB_DEFAULT_P_KEY;
rcd->seq_cnt = 1;
}
- ret = 0;
-done:
- return ret;
+ return 0;
}
/*
@@ -233,7 +230,7 @@ struct qib_ctxtdata *qib_create_ctxtdata(struct qib_pportdata *ppd, u32 ctxt,
/*
* Common code for initializing the physical port structure.
*/
-void qib_init_pportdata(struct qib_pportdata *ppd, struct qib_devdata *dd,
+int qib_init_pportdata(struct qib_pportdata *ppd, struct qib_devdata *dd,
u8 hw_pidx, u8 port)
{
int size;
@@ -243,6 +240,7 @@ void qib_init_pportdata(struct qib_pportdata *ppd, struct qib_devdata *dd,
spin_lock_init(&ppd->sdma_lock);
spin_lock_init(&ppd->lflags_lock);
+ spin_lock_init(&ppd->cc_shadow_lock);
init_waitqueue_head(&ppd->state_wait);
init_timer(&ppd->symerr_clear_timer);
@@ -250,8 +248,10 @@ void qib_init_pportdata(struct qib_pportdata *ppd, struct qib_devdata *dd,
ppd->symerr_clear_timer.data = (unsigned long)ppd;
ppd->qib_wq = NULL;
-
- spin_lock_init(&ppd->cc_shadow_lock);
+ ppd->ibport_data.pmastats =
+ alloc_percpu(struct qib_pma_counters);
+ if (!ppd->ibport_data.pmastats)
+ return -ENOMEM;
if (qib_cc_table_size < IB_CCT_MIN_ENTRIES)
goto bail;
@@ -299,7 +299,7 @@ void qib_init_pportdata(struct qib_pportdata *ppd, struct qib_devdata *dd,
goto bail_3;
}
- return;
+ return 0;
bail_3:
kfree(ppd->ccti_entries_shadow);
@@ -313,7 +313,7 @@ bail_1:
bail:
/* User is intentionally disabling the congestion control agent */
if (!qib_cc_table_size)
- return;
+ return 0;
if (qib_cc_table_size < IB_CCT_MIN_ENTRIES) {
qib_cc_table_size = 0;
@@ -324,7 +324,7 @@ bail:
qib_dev_err(dd, "Congestion Control Agent disabled for port %d\n",
port);
- return;
+ return 0;
}
static int init_pioavailregs(struct qib_devdata *dd)
@@ -525,6 +525,7 @@ static void enable_chip(struct qib_devdata *dd)
static void verify_interrupt(unsigned long opaque)
{
struct qib_devdata *dd = (struct qib_devdata *) opaque;
+ u64 int_counter;
if (!dd)
return; /* being torn down */
@@ -533,7 +534,8 @@ static void verify_interrupt(unsigned long opaque)
* If we don't have a lid or any interrupts, let the user know and
* don't bother checking again.
*/
- if (dd->int_counter == 0) {
+ int_counter = qib_int_counter(dd) - dd->z_int_counter;
+ if (int_counter == 0) {
if (!dd->f_intr_fallback(dd))
dev_err(&dd->pcidev->dev,
"No interrupts detected, not usable.\n");
@@ -633,6 +635,12 @@ wq_error:
return -ENOMEM;
}
+static void qib_free_pportdata(struct qib_pportdata *ppd)
+{
+ free_percpu(ppd->ibport_data.pmastats);
+ ppd->ibport_data.pmastats = NULL;
+}
+
/**
* qib_init - do the actual initialization sequence on the chip
* @dd: the qlogic_ib device
@@ -920,6 +928,7 @@ static void qib_shutdown_device(struct qib_devdata *dd)
destroy_workqueue(ppd->qib_wq);
ppd->qib_wq = NULL;
}
+ qib_free_pportdata(ppd);
}
qib_update_eeprom_log(dd);
@@ -1079,9 +1088,34 @@ void qib_free_devdata(struct qib_devdata *dd)
#ifdef CONFIG_DEBUG_FS
qib_dbg_ibdev_exit(&dd->verbs_dev);
#endif
+ free_percpu(dd->int_counter);
ib_dealloc_device(&dd->verbs_dev.ibdev);
}
+u64 qib_int_counter(struct qib_devdata *dd)
+{
+ int cpu;
+ u64 int_counter = 0;
+
+ for_each_possible_cpu(cpu)
+ int_counter += *per_cpu_ptr(dd->int_counter, cpu);
+ return int_counter;
+}
+
+u64 qib_sps_ints(void)
+{
+ unsigned long flags;
+ struct qib_devdata *dd;
+ u64 sps_ints = 0;
+
+ spin_lock_irqsave(&qib_devs_lock, flags);
+ list_for_each_entry(dd, &qib_dev_list, list) {
+ sps_ints += qib_int_counter(dd);
+ }
+ spin_unlock_irqrestore(&qib_devs_lock, flags);
+ return sps_ints;
+}
+
/*
* Allocate our primary per-unit data structure. Must be done via verbs
* allocator, because the verbs cleanup process both does cleanup and
@@ -1097,14 +1131,10 @@ struct qib_devdata *qib_alloc_devdata(struct pci_dev *pdev, size_t extra)
int ret;
dd = (struct qib_devdata *) ib_alloc_device(sizeof(*dd) + extra);
- if (!dd) {
- dd = ERR_PTR(-ENOMEM);
- goto bail;
- }
+ if (!dd)
+ return ERR_PTR(-ENOMEM);
-#ifdef CONFIG_DEBUG_FS
- qib_dbg_ibdev_init(&dd->verbs_dev);
-#endif
+ INIT_LIST_HEAD(&dd->list);
idr_preload(GFP_KERNEL);
spin_lock_irqsave(&qib_devs_lock, flags);
@@ -1121,11 +1151,13 @@ struct qib_devdata *qib_alloc_devdata(struct pci_dev *pdev, size_t extra)
if (ret < 0) {
qib_early_err(&pdev->dev,
"Could not allocate unit ID: error %d\n", -ret);
-#ifdef CONFIG_DEBUG_FS
- qib_dbg_ibdev_exit(&dd->verbs_dev);
-#endif
- ib_dealloc_device(&dd->verbs_dev.ibdev);
- dd = ERR_PTR(ret);
+ goto bail;
+ }
+ dd->int_counter = alloc_percpu(u64);
+ if (!dd->int_counter) {
+ ret = -ENOMEM;
+ qib_early_err(&pdev->dev,
+ "Could not allocate per-cpu int_counter\n");
goto bail;
}
@@ -1139,9 +1171,15 @@ struct qib_devdata *qib_alloc_devdata(struct pci_dev *pdev, size_t extra)
qib_early_err(&pdev->dev,
"Could not alloc cpulist info, cpu affinity might be wrong\n");
}
-
-bail:
+#ifdef CONFIG_DEBUG_FS
+ qib_dbg_ibdev_init(&dd->verbs_dev);
+#endif
return dd;
+bail:
+ if (!list_empty(&dd->list))
+ list_del_init(&dd->list);
+ ib_dealloc_device(&dd->verbs_dev.ibdev);
+ return ERR_PTR(ret);;
}
/*
diff --git a/drivers/infiniband/hw/qib/qib_mad.c b/drivers/infiniband/hw/qib/qib_mad.c
index ccb119143d20..edad991d60ed 100644
--- a/drivers/infiniband/hw/qib/qib_mad.c
+++ b/drivers/infiniband/hw/qib/qib_mad.c
@@ -1634,6 +1634,23 @@ static int pma_get_portcounters_cong(struct ib_pma_mad *pmp,
return reply((struct ib_smp *)pmp);
}
+static void qib_snapshot_pmacounters(
+ struct qib_ibport *ibp,
+ struct qib_pma_counters *pmacounters)
+{
+ struct qib_pma_counters *p;
+ int cpu;
+
+ memset(pmacounters, 0, sizeof(*pmacounters));
+ for_each_possible_cpu(cpu) {
+ p = per_cpu_ptr(ibp->pmastats, cpu);
+ pmacounters->n_unicast_xmit += p->n_unicast_xmit;
+ pmacounters->n_unicast_rcv += p->n_unicast_rcv;
+ pmacounters->n_multicast_xmit += p->n_multicast_xmit;
+ pmacounters->n_multicast_rcv += p->n_multicast_rcv;
+ }
+}
+
static int pma_get_portcounters_ext(struct ib_pma_mad *pmp,
struct ib_device *ibdev, u8 port)
{
@@ -1642,6 +1659,7 @@ static int pma_get_portcounters_ext(struct ib_pma_mad *pmp,
struct qib_ibport *ibp = to_iport(ibdev, port);
struct qib_pportdata *ppd = ppd_from_ibp(ibp);
u64 swords, rwords, spkts, rpkts, xwait;
+ struct qib_pma_counters pma;
u8 port_select = p->port_select;
memset(pmp->data, 0, sizeof(pmp->data));
@@ -1664,10 +1682,17 @@ static int pma_get_portcounters_ext(struct ib_pma_mad *pmp,
p->port_rcv_data = cpu_to_be64(rwords);
p->port_xmit_packets = cpu_to_be64(spkts);
p->port_rcv_packets = cpu_to_be64(rpkts);
- p->port_unicast_xmit_packets = cpu_to_be64(ibp->n_unicast_xmit);
- p->port_unicast_rcv_packets = cpu_to_be64(ibp->n_unicast_rcv);
- p->port_multicast_xmit_packets = cpu_to_be64(ibp->n_multicast_xmit);
- p->port_multicast_rcv_packets = cpu_to_be64(ibp->n_multicast_rcv);
+
+ qib_snapshot_pmacounters(ibp, &pma);
+
+ p->port_unicast_xmit_packets = cpu_to_be64(pma.n_unicast_xmit
+ - ibp->z_unicast_xmit);
+ p->port_unicast_rcv_packets = cpu_to_be64(pma.n_unicast_rcv
+ - ibp->z_unicast_rcv);
+ p->port_multicast_xmit_packets = cpu_to_be64(pma.n_multicast_xmit
+ - ibp->z_multicast_xmit);
+ p->port_multicast_rcv_packets = cpu_to_be64(pma.n_multicast_rcv
+ - ibp->z_multicast_rcv);
bail:
return reply((struct ib_smp *) pmp);
@@ -1795,6 +1820,7 @@ static int pma_set_portcounters_ext(struct ib_pma_mad *pmp,
struct qib_ibport *ibp = to_iport(ibdev, port);
struct qib_pportdata *ppd = ppd_from_ibp(ibp);
u64 swords, rwords, spkts, rpkts, xwait;
+ struct qib_pma_counters pma;
qib_snapshot_counters(ppd, &swords, &rwords, &spkts, &rpkts, &xwait);
@@ -1810,17 +1836,19 @@ static int pma_set_portcounters_ext(struct ib_pma_mad *pmp,
if (p->counter_select & IB_PMA_SELX_PORT_RCV_PACKETS)
ibp->z_port_rcv_packets = rpkts;
+ qib_snapshot_pmacounters(ibp, &pma);
+
if (p->counter_select & IB_PMA_SELX_PORT_UNI_XMIT_PACKETS)
- ibp->n_unicast_xmit = 0;
+ ibp->z_unicast_xmit = pma.n_unicast_xmit;
if (p->counter_select & IB_PMA_SELX_PORT_UNI_RCV_PACKETS)
- ibp->n_unicast_rcv = 0;
+ ibp->z_unicast_rcv = pma.n_unicast_rcv;
if (p->counter_select & IB_PMA_SELX_PORT_MULTI_XMIT_PACKETS)
- ibp->n_multicast_xmit = 0;
+ ibp->z_multicast_xmit = pma.n_multicast_xmit;
if (p->counter_select & IB_PMA_SELX_PORT_MULTI_RCV_PACKETS)
- ibp->n_multicast_rcv = 0;
+ ibp->z_multicast_rcv = pma.n_multicast_rcv;
return pma_get_portcounters_ext(pmp, ibdev, port);
}
diff --git a/drivers/infiniband/hw/qib/qib_mr.c b/drivers/infiniband/hw/qib/qib_mr.c
index e6687ded8210..9bbb55347cc1 100644
--- a/drivers/infiniband/hw/qib/qib_mr.c
+++ b/drivers/infiniband/hw/qib/qib_mr.c
@@ -232,8 +232,8 @@ struct ib_mr *qib_reg_user_mr(struct ib_pd *pd, u64 start, u64 length,
{
struct qib_mr *mr;
struct ib_umem *umem;
- struct ib_umem_chunk *chunk;
- int n, m, i;
+ struct scatterlist *sg;
+ int n, m, entry;
struct ib_mr *ret;
if (length == 0) {
@@ -246,9 +246,7 @@ struct ib_mr *qib_reg_user_mr(struct ib_pd *pd, u64 start, u64 length,
if (IS_ERR(umem))
return (void *) umem;
- n = 0;
- list_for_each_entry(chunk, &umem->chunk_list, list)
- n += chunk->nents;
+ n = umem->nmap;
mr = alloc_mr(n, pd);
if (IS_ERR(mr)) {
@@ -268,11 +266,10 @@ struct ib_mr *qib_reg_user_mr(struct ib_pd *pd, u64 start, u64 length,
mr->mr.page_shift = ilog2(umem->page_size);
m = 0;
n = 0;
- list_for_each_entry(chunk, &umem->chunk_list, list) {
- for (i = 0; i < chunk->nents; i++) {
+ for_each_sg(umem->sg_head.sgl, sg, umem->nmap, entry) {
void *vaddr;
- vaddr = page_address(sg_page(&chunk->page_list[i]));
+ vaddr = page_address(sg_page(sg));
if (!vaddr) {
ret = ERR_PTR(-EINVAL);
goto bail;
@@ -284,7 +281,6 @@ struct ib_mr *qib_reg_user_mr(struct ib_pd *pd, u64 start, u64 length,
m++;
n = 0;
}
- }
}
ret = &mr->ibmr;
diff --git a/drivers/infiniband/hw/qib/qib_pcie.c b/drivers/infiniband/hw/qib/qib_pcie.c
index c8d9c4ab142b..61a0046efb76 100644
--- a/drivers/infiniband/hw/qib/qib_pcie.c
+++ b/drivers/infiniband/hw/qib/qib_pcie.c
@@ -197,46 +197,47 @@ static void qib_msix_setup(struct qib_devdata *dd, int pos, u32 *msixcnt,
struct qib_msix_entry *qib_msix_entry)
{
int ret;
- u32 tabsize = 0;
- u16 msix_flags;
+ int nvec = *msixcnt;
struct msix_entry *msix_entry;
int i;
+ ret = pci_msix_vec_count(dd->pcidev);
+ if (ret < 0)
+ goto do_intx;
+
+ nvec = min(nvec, ret);
+
/* We can't pass qib_msix_entry array to qib_msix_setup
* so use a dummy msix_entry array and copy the allocated
* irq back to the qib_msix_entry array. */
- msix_entry = kmalloc(*msixcnt * sizeof(*msix_entry), GFP_KERNEL);
- if (!msix_entry) {
- ret = -ENOMEM;
+ msix_entry = kmalloc(nvec * sizeof(*msix_entry), GFP_KERNEL);
+ if (!msix_entry)
goto do_intx;
- }
- for (i = 0; i < *msixcnt; i++)
+
+ for (i = 0; i < nvec; i++)
msix_entry[i] = qib_msix_entry[i].msix;
- pci_read_config_word(dd->pcidev, pos + PCI_MSIX_FLAGS, &msix_flags);
- tabsize = 1 + (msix_flags & PCI_MSIX_FLAGS_QSIZE);
- if (tabsize > *msixcnt)
- tabsize = *msixcnt;
- ret = pci_enable_msix(dd->pcidev, msix_entry, tabsize);
- if (ret > 0) {
- tabsize = ret;
- ret = pci_enable_msix(dd->pcidev, msix_entry, tabsize);
- }
-do_intx:
- if (ret) {
- qib_dev_err(dd,
- "pci_enable_msix %d vectors failed: %d, falling back to INTx\n",
- tabsize, ret);
- tabsize = 0;
- }
- for (i = 0; i < tabsize; i++)
+ ret = pci_enable_msix_range(dd->pcidev, msix_entry, 1, nvec);
+ if (ret < 0)
+ goto free_msix_entry;
+ else
+ nvec = ret;
+
+ for (i = 0; i < nvec; i++)
qib_msix_entry[i].msix = msix_entry[i];
+
kfree(msix_entry);
- *msixcnt = tabsize;
+ *msixcnt = nvec;
+ return;
- if (ret)
- qib_enable_intx(dd->pcidev);
+free_msix_entry:
+ kfree(msix_entry);
+do_intx:
+ qib_dev_err(dd, "pci_enable_msix_range %d vectors failed: %d, "
+ "falling back to INTx\n", nvec, ret);
+ *msixcnt = 0;
+ qib_enable_intx(dd->pcidev);
}
/**
diff --git a/drivers/infiniband/hw/qib/qib_rc.c b/drivers/infiniband/hw/qib/qib_rc.c
index 3ab341320ead..2f2501890c4e 100644
--- a/drivers/infiniband/hw/qib/qib_rc.c
+++ b/drivers/infiniband/hw/qib/qib_rc.c
@@ -752,7 +752,7 @@ void qib_send_rc_ack(struct qib_qp *qp)
qib_flush_wc();
qib_sendbuf_done(dd, pbufn);
- ibp->n_unicast_xmit++;
+ this_cpu_inc(ibp->pmastats->n_unicast_xmit);
goto done;
queue_ack:
diff --git a/drivers/infiniband/hw/qib/qib_ruc.c b/drivers/infiniband/hw/qib/qib_ruc.c
index 357b6cfcd46c..4c07a8b34ffe 100644
--- a/drivers/infiniband/hw/qib/qib_ruc.c
+++ b/drivers/infiniband/hw/qib/qib_ruc.c
@@ -703,6 +703,7 @@ void qib_make_ruc_header(struct qib_qp *qp, struct qib_other_headers *ohdr,
ohdr->bth[0] = cpu_to_be32(bth0);
ohdr->bth[1] = cpu_to_be32(qp->remote_qpn);
ohdr->bth[2] = cpu_to_be32(bth2);
+ this_cpu_inc(ibp->pmastats->n_unicast_xmit);
}
/**
diff --git a/drivers/infiniband/hw/qib/qib_ud.c b/drivers/infiniband/hw/qib/qib_ud.c
index 3ad651c3356c..aaf7039f8ed2 100644
--- a/drivers/infiniband/hw/qib/qib_ud.c
+++ b/drivers/infiniband/hw/qib/qib_ud.c
@@ -280,11 +280,11 @@ int qib_make_ud_req(struct qib_qp *qp)
ah_attr = &to_iah(wqe->wr.wr.ud.ah)->attr;
if (ah_attr->dlid >= QIB_MULTICAST_LID_BASE) {
if (ah_attr->dlid != QIB_PERMISSIVE_LID)
- ibp->n_multicast_xmit++;
+ this_cpu_inc(ibp->pmastats->n_multicast_xmit);
else
- ibp->n_unicast_xmit++;
+ this_cpu_inc(ibp->pmastats->n_unicast_xmit);
} else {
- ibp->n_unicast_xmit++;
+ this_cpu_inc(ibp->pmastats->n_unicast_xmit);
lid = ah_attr->dlid & ~((1 << ppd->lmc) - 1);
if (unlikely(lid == ppd->lid)) {
/*
diff --git a/drivers/infiniband/hw/qib/qib_user_sdma.c b/drivers/infiniband/hw/qib/qib_user_sdma.c
index 165aee2ca8a0..d2806cae234c 100644
--- a/drivers/infiniband/hw/qib/qib_user_sdma.c
+++ b/drivers/infiniband/hw/qib/qib_user_sdma.c
@@ -52,6 +52,17 @@
/* attempt to drain the queue for 5secs */
#define QIB_USER_SDMA_DRAIN_TIMEOUT 500
+/*
+ * track how many times a process open this driver.
+ */
+static struct rb_root qib_user_sdma_rb_root = RB_ROOT;
+
+struct qib_user_sdma_rb_node {
+ struct rb_node node;
+ int refcount;
+ pid_t pid;
+};
+
struct qib_user_sdma_pkt {
struct list_head list; /* list element */
@@ -120,15 +131,60 @@ struct qib_user_sdma_queue {
/* dma page table */
struct rb_root dma_pages_root;
+ struct qib_user_sdma_rb_node *sdma_rb_node;
+
/* protect everything above... */
struct mutex lock;
};
+static struct qib_user_sdma_rb_node *
+qib_user_sdma_rb_search(struct rb_root *root, pid_t pid)
+{
+ struct qib_user_sdma_rb_node *sdma_rb_node;
+ struct rb_node *node = root->rb_node;
+
+ while (node) {
+ sdma_rb_node = container_of(node,
+ struct qib_user_sdma_rb_node, node);
+ if (pid < sdma_rb_node->pid)
+ node = node->rb_left;
+ else if (pid > sdma_rb_node->pid)
+ node = node->rb_right;
+ else
+ return sdma_rb_node;
+ }
+ return NULL;
+}
+
+static int
+qib_user_sdma_rb_insert(struct rb_root *root, struct qib_user_sdma_rb_node *new)
+{
+ struct rb_node **node = &(root->rb_node);
+ struct rb_node *parent = NULL;
+ struct qib_user_sdma_rb_node *got;
+
+ while (*node) {
+ got = container_of(*node, struct qib_user_sdma_rb_node, node);
+ parent = *node;
+ if (new->pid < got->pid)
+ node = &((*node)->rb_left);
+ else if (new->pid > got->pid)
+ node = &((*node)->rb_right);
+ else
+ return 0;
+ }
+
+ rb_link_node(&new->node, parent, node);
+ rb_insert_color(&new->node, root);
+ return 1;
+}
+
struct qib_user_sdma_queue *
qib_user_sdma_queue_create(struct device *dev, int unit, int ctxt, int sctxt)
{
struct qib_user_sdma_queue *pq =
kmalloc(sizeof(struct qib_user_sdma_queue), GFP_KERNEL);
+ struct qib_user_sdma_rb_node *sdma_rb_node;
if (!pq)
goto done;
@@ -138,6 +194,7 @@ qib_user_sdma_queue_create(struct device *dev, int unit, int ctxt, int sctxt)
pq->num_pending = 0;
pq->num_sending = 0;
pq->added = 0;
+ pq->sdma_rb_node = NULL;
INIT_LIST_HEAD(&pq->sent);
spin_lock_init(&pq->sent_lock);
@@ -163,8 +220,30 @@ qib_user_sdma_queue_create(struct device *dev, int unit, int ctxt, int sctxt)
pq->dma_pages_root = RB_ROOT;
+ sdma_rb_node = qib_user_sdma_rb_search(&qib_user_sdma_rb_root,
+ current->pid);
+ if (sdma_rb_node) {
+ sdma_rb_node->refcount++;
+ } else {
+ int ret;
+ sdma_rb_node = kmalloc(sizeof(
+ struct qib_user_sdma_rb_node), GFP_KERNEL);
+ if (!sdma_rb_node)
+ goto err_rb;
+
+ sdma_rb_node->refcount = 1;
+ sdma_rb_node->pid = current->pid;
+
+ ret = qib_user_sdma_rb_insert(&qib_user_sdma_rb_root,
+ sdma_rb_node);
+ BUG_ON(ret == 0);
+ }
+ pq->sdma_rb_node = sdma_rb_node;
+
goto done;
+err_rb:
+ dma_pool_destroy(pq->header_cache);
err_slab:
kmem_cache_destroy(pq->pkt_slab);
err_kfree:
@@ -1020,8 +1099,13 @@ void qib_user_sdma_queue_destroy(struct qib_user_sdma_queue *pq)
if (!pq)
return;
- kmem_cache_destroy(pq->pkt_slab);
+ pq->sdma_rb_node->refcount--;
+ if (pq->sdma_rb_node->refcount == 0) {
+ rb_erase(&pq->sdma_rb_node->node, &qib_user_sdma_rb_root);
+ kfree(pq->sdma_rb_node);
+ }
dma_pool_destroy(pq->header_cache);
+ kmem_cache_destroy(pq->pkt_slab);
kfree(pq);
}
@@ -1241,26 +1325,52 @@ static int qib_user_sdma_push_pkts(struct qib_pportdata *ppd,
struct qib_user_sdma_queue *pq,
struct list_head *pktlist, int count)
{
- int ret = 0;
unsigned long flags;
if (unlikely(!(ppd->lflags & QIBL_LINKACTIVE)))
return -ECOMM;
- spin_lock_irqsave(&ppd->sdma_lock, flags);
-
- if (unlikely(!__qib_sdma_running(ppd))) {
- ret = -ECOMM;
- goto unlock;
+ /* non-blocking mode */
+ if (pq->sdma_rb_node->refcount > 1) {
+ spin_lock_irqsave(&ppd->sdma_lock, flags);
+ if (unlikely(!__qib_sdma_running(ppd))) {
+ spin_unlock_irqrestore(&ppd->sdma_lock, flags);
+ return -ECOMM;
+ }
+ pq->num_pending += count;
+ list_splice_tail_init(pktlist, &ppd->sdma_userpending);
+ qib_user_sdma_send_desc(ppd, &ppd->sdma_userpending);
+ spin_unlock_irqrestore(&ppd->sdma_lock, flags);
+ return 0;
}
+ /* In this case, descriptors from this process are not
+ * linked to ppd pending queue, interrupt handler
+ * won't update this process, it is OK to directly
+ * modify without sdma lock.
+ */
+
+
pq->num_pending += count;
- list_splice_tail_init(pktlist, &ppd->sdma_userpending);
- qib_user_sdma_send_desc(ppd, &ppd->sdma_userpending);
+ /*
+ * Blocking mode for single rail process, we must
+ * release/regain sdma_lock to give other process
+ * chance to make progress. This is important for
+ * performance.
+ */
+ do {
+ spin_lock_irqsave(&ppd->sdma_lock, flags);
+ if (unlikely(!__qib_sdma_running(ppd))) {
+ spin_unlock_irqrestore(&ppd->sdma_lock, flags);
+ return -ECOMM;
+ }
+ qib_user_sdma_send_desc(ppd, pktlist);
+ if (!list_empty(pktlist))
+ qib_sdma_make_progress(ppd);
+ spin_unlock_irqrestore(&ppd->sdma_lock, flags);
+ } while (!list_empty(pktlist));
-unlock:
- spin_unlock_irqrestore(&ppd->sdma_lock, flags);
- return ret;
+ return 0;
}
int qib_user_sdma_writev(struct qib_ctxtdata *rcd,
@@ -1290,7 +1400,7 @@ int qib_user_sdma_writev(struct qib_ctxtdata *rcd,
qib_user_sdma_queue_clean(ppd, pq);
while (dim) {
- int mxp = 8;
+ int mxp = 1;
int ndesc = 0;
ret = qib_user_sdma_queue_pkts(dd, ppd, pq,
diff --git a/drivers/infiniband/hw/qib/qib_verbs.c b/drivers/infiniband/hw/qib/qib_verbs.c
index 092b0bb1bb78..9bcfbd842980 100644
--- a/drivers/infiniband/hw/qib/qib_verbs.c
+++ b/drivers/infiniband/hw/qib/qib_verbs.c
@@ -662,7 +662,7 @@ void qib_ib_rcv(struct qib_ctxtdata *rcd, void *rhdr, void *data, u32 tlen)
mcast = qib_mcast_find(ibp, &hdr->u.l.grh.dgid);
if (mcast == NULL)
goto drop;
- ibp->n_multicast_rcv++;
+ this_cpu_inc(ibp->pmastats->n_multicast_rcv);
list_for_each_entry_rcu(p, &mcast->qp_list, list)
qib_qp_rcv(rcd, hdr, 1, data, tlen, p->qp);
/*
@@ -678,8 +678,8 @@ void qib_ib_rcv(struct qib_ctxtdata *rcd, void *rhdr, void *data, u32 tlen)
&rcd->lookaside_qp->refcount))
wake_up(
&rcd->lookaside_qp->wait);
- rcd->lookaside_qp = NULL;
- }
+ rcd->lookaside_qp = NULL;
+ }
}
if (!rcd->lookaside_qp) {
qp = qib_lookup_qpn(ibp, qp_num);
@@ -689,7 +689,7 @@ void qib_ib_rcv(struct qib_ctxtdata *rcd, void *rhdr, void *data, u32 tlen)
rcd->lookaside_qpn = qp_num;
} else
qp = rcd->lookaside_qp;
- ibp->n_unicast_rcv++;
+ this_cpu_inc(ibp->pmastats->n_unicast_rcv);
qib_qp_rcv(rcd, hdr, lnh == QIB_LRH_GRH, data, tlen, qp);
}
return;
diff --git a/drivers/infiniband/hw/qib/qib_verbs.h b/drivers/infiniband/hw/qib/qib_verbs.h
index a01c7d2cf541..bfc8948fdd35 100644
--- a/drivers/infiniband/hw/qib/qib_verbs.h
+++ b/drivers/infiniband/hw/qib/qib_verbs.h
@@ -664,6 +664,13 @@ struct qib_opcode_stats_perctx {
struct qib_opcode_stats stats[128];
};
+struct qib_pma_counters {
+ u64 n_unicast_xmit; /* total unicast packets sent */
+ u64 n_unicast_rcv; /* total unicast packets received */
+ u64 n_multicast_xmit; /* total multicast packets sent */
+ u64 n_multicast_rcv; /* total multicast packets received */
+};
+
struct qib_ibport {
struct qib_qp __rcu *qp0;
struct qib_qp __rcu *qp1;
@@ -680,10 +687,11 @@ struct qib_ibport {
__be64 mkey;
__be64 guids[QIB_GUIDS_PER_PORT - 1]; /* writable GUIDs */
u64 tid; /* TID for traps */
- u64 n_unicast_xmit; /* total unicast packets sent */
- u64 n_unicast_rcv; /* total unicast packets received */
- u64 n_multicast_xmit; /* total multicast packets sent */
- u64 n_multicast_rcv; /* total multicast packets received */
+ struct qib_pma_counters __percpu *pmastats;
+ u64 z_unicast_xmit; /* starting count for PMA */
+ u64 z_unicast_rcv; /* starting count for PMA */
+ u64 z_multicast_xmit; /* starting count for PMA */
+ u64 z_multicast_rcv; /* starting count for PMA */
u64 z_symbol_error_counter; /* starting count for PMA */
u64 z_link_error_recovery_counter; /* starting count for PMA */
u64 z_link_downed_counter; /* starting count for PMA */
diff --git a/drivers/infiniband/hw/usnic/usnic_uiom.c b/drivers/infiniband/hw/usnic/usnic_uiom.c
index 16755cdab2c0..801a1d6937e4 100644
--- a/drivers/infiniband/hw/usnic/usnic_uiom.c
+++ b/drivers/infiniband/hw/usnic/usnic_uiom.c
@@ -286,7 +286,7 @@ iter_chunk:
err = iommu_map(pd->domain, va_start, pa_start,
size, flags);
if (err) {
- usnic_err("Failed to map va 0x%lx pa 0x%pa size 0x%zx with err %d\n",
+ usnic_err("Failed to map va 0x%lx pa %pa size 0x%zx with err %d\n",
va_start, &pa_start, size, err);
goto err_out;
}