* [PATCH v1 08/14] xprtrdma: Squelch "max send, max recv" messages at connect time
From: Chuck Lever @ 2016-11-09 19:05 UTC (permalink / raw)
To: linux-rdma-u79uwXL29TY76Z2rM5mHXA,
linux-nfs-u79uwXL29TY76Z2rM5mHXA
In-Reply-To: <20161109184735.15007.96507.stgit-FYjufvaPoItvLzlybtyyYzGyq/o6K9yX@public.gmane.org>
Clean up: This message was intended to be a dprintk, as it is on the
server-side.
Fixes: 87cfb9a0c85c ('xprtrdma: Client-side support for ...')
Signed-off-by: Chuck Lever <chuck.lever-QHcLZuEGTsvQT0dZR+AlfA@public.gmane.org>
---
net/sunrpc/xprtrdma/verbs.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/net/sunrpc/xprtrdma/verbs.c b/net/sunrpc/xprtrdma/verbs.c
index cbb1885..9f66fd8 100644
--- a/net/sunrpc/xprtrdma/verbs.c
+++ b/net/sunrpc/xprtrdma/verbs.c
@@ -223,8 +223,8 @@
cdata->inline_rsize = rsize;
if (wsize < cdata->inline_wsize)
cdata->inline_wsize = wsize;
- pr_info("rpcrdma: max send %u, max recv %u\n",
- cdata->inline_wsize, cdata->inline_rsize);
+ dprintk("RPC: %s: max send %u, max recv %u\n",
+ __func__, cdata->inline_wsize, cdata->inline_rsize);
rpcrdma_set_max_header_sizes(r_xprt);
}
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply related
* [PATCH v1 09/14] xprtrdma: Shorten QP access error message
From: Chuck Lever @ 2016-11-09 19:06 UTC (permalink / raw)
To: linux-rdma-u79uwXL29TY76Z2rM5mHXA,
linux-nfs-u79uwXL29TY76Z2rM5mHXA
In-Reply-To: <20161109184735.15007.96507.stgit-FYjufvaPoItvLzlybtyyYzGyq/o6K9yX@public.gmane.org>
Clean up: The convention for this type of warning message is not to
show the function name or "RPC: ".
Signed-off-by: Chuck Lever <chuck.lever-QHcLZuEGTsvQT0dZR+AlfA@public.gmane.org>
---
net/sunrpc/xprtrdma/verbs.c | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/net/sunrpc/xprtrdma/verbs.c b/net/sunrpc/xprtrdma/verbs.c
index 9f66fd8..11d0774 100644
--- a/net/sunrpc/xprtrdma/verbs.c
+++ b/net/sunrpc/xprtrdma/verbs.c
@@ -103,9 +103,9 @@
{
struct rpcrdma_ep *ep = context;
- pr_err("RPC: %s: %s on device %s ep %p\n",
- __func__, ib_event_msg(event->event),
- event->device->name, context);
+ pr_err("rpcrdma: %s on device %s ep %p\n",
+ ib_event_msg(event->event), event->device->name, context);
+
if (ep->rep_connected == 1) {
ep->rep_connected = -EIO;
rpcrdma_conn_func(ep);
--
To unsubscribe from this list: send the line "unsubscribe linux-nfs" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply related
* [PATCH v1 10/14] xprtrdma: Update dprintk in rpcrdma_count_chunks
From: Chuck Lever @ 2016-11-09 19:06 UTC (permalink / raw)
To: linux-rdma-u79uwXL29TY76Z2rM5mHXA,
linux-nfs-u79uwXL29TY76Z2rM5mHXA
In-Reply-To: <20161109184735.15007.96507.stgit-FYjufvaPoItvLzlybtyyYzGyq/o6K9yX@public.gmane.org>
Clean up: offset and handle should be zero-filled, just like in the
chunk encoders.
Signed-off-by: Chuck Lever <chuck.lever-QHcLZuEGTsvQT0dZR+AlfA@public.gmane.org>
---
net/sunrpc/xprtrdma/rpc_rdma.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/net/sunrpc/xprtrdma/rpc_rdma.c b/net/sunrpc/xprtrdma/rpc_rdma.c
index d987c2d..01f5cbc 100644
--- a/net/sunrpc/xprtrdma/rpc_rdma.c
+++ b/net/sunrpc/xprtrdma/rpc_rdma.c
@@ -786,7 +786,7 @@ static bool rpcrdma_results_inline(struct rpcrdma_xprt *r_xprt,
ifdebug(FACILITY) {
u64 off;
xdr_decode_hyper((__be32 *)&seg->rs_offset, &off);
- dprintk("RPC: %s: chunk %d@0x%llx:0x%x\n",
+ dprintk("RPC: %s: chunk %d@0x%016llx:0x%08x\n",
__func__,
be32_to_cpu(seg->rs_length),
(unsigned long long)off,
--
To unsubscribe from this list: send the line "unsubscribe linux-nfs" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply related
* [PATCH v1 11/14] xprtrdma: Relocate connection helper functions
From: Chuck Lever @ 2016-11-09 19:06 UTC (permalink / raw)
To: linux-rdma-u79uwXL29TY76Z2rM5mHXA,
linux-nfs-u79uwXL29TY76Z2rM5mHXA
In-Reply-To: <20161109184735.15007.96507.stgit-FYjufvaPoItvLzlybtyyYzGyq/o6K9yX@public.gmane.org>
Clean up: Disentangle connection helpers from RPC-over-RDMA reply
decoding functions.
Signed-off-by: Chuck Lever <chuck.lever-QHcLZuEGTsvQT0dZR+AlfA@public.gmane.org>
---
net/sunrpc/xprtrdma/rpc_rdma.c | 34 ----------------------------------
net/sunrpc/xprtrdma/transport.c | 28 ++++++++++++++++++++++++++++
net/sunrpc/xprtrdma/xprt_rdma.h | 10 +++-------
3 files changed, 31 insertions(+), 41 deletions(-)
diff --git a/net/sunrpc/xprtrdma/rpc_rdma.c b/net/sunrpc/xprtrdma/rpc_rdma.c
index 01f5cbc..c52e0f2 100644
--- a/net/sunrpc/xprtrdma/rpc_rdma.c
+++ b/net/sunrpc/xprtrdma/rpc_rdma.c
@@ -906,28 +906,6 @@ static bool rpcrdma_results_inline(struct rpcrdma_xprt *r_xprt,
return fixup_copy_count;
}
-void
-rpcrdma_connect_worker(struct work_struct *work)
-{
- struct rpcrdma_ep *ep =
- container_of(work, struct rpcrdma_ep, rep_connect_worker.work);
- struct rpcrdma_xprt *r_xprt =
- container_of(ep, struct rpcrdma_xprt, rx_ep);
- struct rpc_xprt *xprt = &r_xprt->rx_xprt;
-
- spin_lock_bh(&xprt->transport_lock);
- if (++xprt->connect_cookie == 0) /* maintain a reserved value */
- ++xprt->connect_cookie;
- if (ep->rep_connected > 0) {
- if (!xprt_test_and_set_connected(xprt))
- xprt_wake_pending_tasks(xprt, 0);
- } else {
- if (xprt_test_and_clear_connected(xprt))
- xprt_wake_pending_tasks(xprt, -ENOTCONN);
- }
- spin_unlock_bh(&xprt->transport_lock);
-}
-
#if defined(CONFIG_SUNRPC_BACKCHANNEL)
/* By convention, backchannel calls arrive via rdma_msg type
* messages, and never populate the chunk lists. This makes
@@ -959,18 +937,6 @@ static bool rpcrdma_results_inline(struct rpcrdma_xprt *r_xprt,
}
#endif /* CONFIG_SUNRPC_BACKCHANNEL */
-/*
- * This function is called when an async event is posted to
- * the connection which changes the connection state. All it
- * does at this point is mark the connection up/down, the rpc
- * timers do the rest.
- */
-void
-rpcrdma_conn_func(struct rpcrdma_ep *ep)
-{
- schedule_delayed_work(&ep->rep_connect_worker, 0);
-}
-
/* Process received RPC/RDMA messages.
*
* Errors must result in the RPC task either being awakened, or
diff --git a/net/sunrpc/xprtrdma/transport.c b/net/sunrpc/xprtrdma/transport.c
index 545d3fc..534c178 100644
--- a/net/sunrpc/xprtrdma/transport.c
+++ b/net/sunrpc/xprtrdma/transport.c
@@ -219,6 +219,34 @@
}
}
+void
+rpcrdma_conn_func(struct rpcrdma_ep *ep)
+{
+ schedule_delayed_work(&ep->rep_connect_worker, 0);
+}
+
+void
+rpcrdma_connect_worker(struct work_struct *work)
+{
+ struct rpcrdma_ep *ep =
+ container_of(work, struct rpcrdma_ep, rep_connect_worker.work);
+ struct rpcrdma_xprt *r_xprt =
+ container_of(ep, struct rpcrdma_xprt, rx_ep);
+ struct rpc_xprt *xprt = &r_xprt->rx_xprt;
+
+ spin_lock_bh(&xprt->transport_lock);
+ if (++xprt->connect_cookie == 0) /* maintain a reserved value */
+ ++xprt->connect_cookie;
+ if (ep->rep_connected > 0) {
+ if (!xprt_test_and_set_connected(xprt))
+ xprt_wake_pending_tasks(xprt, 0);
+ } else {
+ if (xprt_test_and_clear_connected(xprt))
+ xprt_wake_pending_tasks(xprt, -ENOTCONN);
+ }
+ spin_unlock_bh(&xprt->transport_lock);
+}
+
static void
xprt_rdma_connect_worker(struct work_struct *work)
{
diff --git a/net/sunrpc/xprtrdma/xprt_rdma.h b/net/sunrpc/xprtrdma/xprt_rdma.h
index b8424fa..e35efd4 100644
--- a/net/sunrpc/xprtrdma/xprt_rdma.h
+++ b/net/sunrpc/xprtrdma/xprt_rdma.h
@@ -490,6 +490,7 @@ int rpcrdma_ep_create(struct rpcrdma_ep *, struct rpcrdma_ia *,
struct rpcrdma_create_data_internal *);
void rpcrdma_ep_destroy(struct rpcrdma_ep *, struct rpcrdma_ia *);
int rpcrdma_ep_connect(struct rpcrdma_ep *, struct rpcrdma_ia *);
+void rpcrdma_conn_func(struct rpcrdma_ep *ep);
void rpcrdma_ep_disconnect(struct rpcrdma_ep *, struct rpcrdma_ia *);
int rpcrdma_ep_post(struct rpcrdma_ia *, struct rpcrdma_ep *,
@@ -549,13 +550,6 @@ struct rpcrdma_regbuf *rpcrdma_alloc_regbuf(size_t, enum dma_data_direction,
}
/*
- * RPC/RDMA connection management calls - xprtrdma/rpc_rdma.c
- */
-void rpcrdma_connect_worker(struct work_struct *);
-void rpcrdma_conn_func(struct rpcrdma_ep *);
-void rpcrdma_reply_handler(struct work_struct *);
-
-/*
* RPC/RDMA protocol calls - xprtrdma/rpc_rdma.c
*/
@@ -572,12 +566,14 @@ bool rpcrdma_prepare_send_sges(struct rpcrdma_ia *, struct rpcrdma_req *,
void rpcrdma_unmap_sges(struct rpcrdma_ia *, struct rpcrdma_req *);
int rpcrdma_marshal_req(struct rpc_rqst *);
void rpcrdma_set_max_header_sizes(struct rpcrdma_xprt *);
+void rpcrdma_reply_handler(struct work_struct *work);
/* RPC/RDMA module init - xprtrdma/transport.c
*/
extern unsigned int xprt_rdma_max_inline_read;
void xprt_rdma_format_addresses(struct rpc_xprt *xprt, struct sockaddr *sap);
void xprt_rdma_free_addresses(struct rpc_xprt *xprt);
+void rpcrdma_connect_worker(struct work_struct *work);
void xprt_rdma_print_stats(struct rpc_xprt *xprt, struct seq_file *seq);
int xprt_rdma_init(void);
void xprt_rdma_cleanup(void);
--
To unsubscribe from this list: send the line "unsubscribe linux-nfs" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply related
* [PATCH v1 12/14] xprtrdma: Simplify synopsis of rpcrdma_ep_connect()
From: Chuck Lever @ 2016-11-09 19:06 UTC (permalink / raw)
To: linux-rdma-u79uwXL29TY76Z2rM5mHXA,
linux-nfs-u79uwXL29TY76Z2rM5mHXA
In-Reply-To: <20161109184735.15007.96507.stgit-FYjufvaPoItvLzlybtyyYzGyq/o6K9yX@public.gmane.org>
Clean up: Make the rpcrdma_xprt visible in the whole function.
The caller has it already, so eliminate the container_of noise, and
just pass it in explicitly.
Signed-off-by: Chuck Lever <chuck.lever-QHcLZuEGTsvQT0dZR+AlfA@public.gmane.org>
---
net/sunrpc/xprtrdma/transport.c | 2 +-
net/sunrpc/xprtrdma/verbs.c | 17 +++++++----------
net/sunrpc/xprtrdma/xprt_rdma.h | 2 +-
3 files changed, 9 insertions(+), 12 deletions(-)
diff --git a/net/sunrpc/xprtrdma/transport.c b/net/sunrpc/xprtrdma/transport.c
index 534c178..349903b 100644
--- a/net/sunrpc/xprtrdma/transport.c
+++ b/net/sunrpc/xprtrdma/transport.c
@@ -259,7 +259,7 @@
dprintk("RPC: %s: %sconnect\n", __func__,
r_xprt->rx_ep.rep_connected != 0 ? "re" : "");
- rc = rpcrdma_ep_connect(&r_xprt->rx_ep, &r_xprt->rx_ia);
+ rc = rpcrdma_ep_connect(r_xprt);
if (rc)
xprt_wake_pending_tasks(xprt, rc);
diff --git a/net/sunrpc/xprtrdma/verbs.c b/net/sunrpc/xprtrdma/verbs.c
index 11d0774..901c3ab 100644
--- a/net/sunrpc/xprtrdma/verbs.c
+++ b/net/sunrpc/xprtrdma/verbs.c
@@ -634,26 +634,25 @@ static void rpcrdma_destroy_id(struct rdma_cm_id *id)
ib_free_cq(ep->rep_attr.send_cq);
}
-/*
- * Connect unconnected endpoint.
- */
int
-rpcrdma_ep_connect(struct rpcrdma_ep *ep, struct rpcrdma_ia *ia)
+rpcrdma_ep_connect(struct rpcrdma_xprt *r_xprt)
{
+ struct rpcrdma_ep *ep = &r_xprt->rx_ep;
+ struct rpcrdma_ia *ia = &r_xprt->rx_ia;
struct rdma_cm_id *id, *old;
int rc = 0;
int retry_count = 0;
if (ep->rep_connected != 0) {
- struct rpcrdma_xprt *xprt;
+ struct sockaddr *sap;
+
retry:
dprintk("RPC: %s: reconnecting...\n", __func__);
rpcrdma_ep_disconnect(ep, ia);
- xprt = container_of(ia, struct rpcrdma_xprt, rx_ia);
- id = rpcrdma_create_id(xprt, ia,
- (struct sockaddr *)&xprt->rx_data.addr);
+ sap = (struct sockaddr *)&r_xprt->rx_data.addr;
+ id = rpcrdma_create_id(r_xprt, ia, sap);
if (IS_ERR(id)) {
rc = -EHOSTUNREACH;
goto out;
@@ -735,12 +734,10 @@ static void rpcrdma_destroy_id(struct rdma_cm_id *id)
}
rc = ep->rep_connected;
} else {
- struct rpcrdma_xprt *r_xprt;
unsigned int extras;
dprintk("RPC: %s: connected\n", __func__);
- r_xprt = container_of(ia, struct rpcrdma_xprt, rx_ia);
extras = r_xprt->rx_buf.rb_bc_srv_max_requests;
if (extras) {
diff --git a/net/sunrpc/xprtrdma/xprt_rdma.h b/net/sunrpc/xprtrdma/xprt_rdma.h
index e35efd4..74760d9 100644
--- a/net/sunrpc/xprtrdma/xprt_rdma.h
+++ b/net/sunrpc/xprtrdma/xprt_rdma.h
@@ -489,7 +489,7 @@ struct rpcrdma_xprt {
int rpcrdma_ep_create(struct rpcrdma_ep *, struct rpcrdma_ia *,
struct rpcrdma_create_data_internal *);
void rpcrdma_ep_destroy(struct rpcrdma_ep *, struct rpcrdma_ia *);
-int rpcrdma_ep_connect(struct rpcrdma_ep *, struct rpcrdma_ia *);
+int rpcrdma_ep_connect(struct rpcrdma_xprt *r_xprt);
void rpcrdma_conn_func(struct rpcrdma_ep *ep);
void rpcrdma_ep_disconnect(struct rpcrdma_ep *, struct rpcrdma_ia *);
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply related
* [PATCH v1 13/14] xprtrdma: Refactor FRMR invalidation
From: Chuck Lever @ 2016-11-09 19:06 UTC (permalink / raw)
To: linux-rdma-u79uwXL29TY76Z2rM5mHXA,
linux-nfs-u79uwXL29TY76Z2rM5mHXA
In-Reply-To: <20161109184735.15007.96507.stgit-FYjufvaPoItvLzlybtyyYzGyq/o6K9yX@public.gmane.org>
Clean up: After some recent updates, clarifications can be made to
the FRMR invalidation logic.
- Both the remote and local invalidation case mark the frmr INVALID,
so make that a common path.
- Manage the WR list more "tastefully" by replacing the conditional
that discriminates between the list head and ->next pointers.
- Use mw->mw_handle in all cases, since that has the same value as
f->fr_mr->rkey, and is already in cache.
Signed-off-by: Chuck Lever <chuck.lever-QHcLZuEGTsvQT0dZR+AlfA@public.gmane.org>
---
net/sunrpc/xprtrdma/frwr_ops.c | 57 +++++++++++++++-------------------------
1 file changed, 21 insertions(+), 36 deletions(-)
diff --git a/net/sunrpc/xprtrdma/frwr_ops.c b/net/sunrpc/xprtrdma/frwr_ops.c
index e99bf61..900dc40 100644
--- a/net/sunrpc/xprtrdma/frwr_ops.c
+++ b/net/sunrpc/xprtrdma/frwr_ops.c
@@ -457,26 +457,6 @@
return -ENOTCONN;
}
-static struct ib_send_wr *
-__frwr_prepare_linv_wr(struct rpcrdma_mw *mw)
-{
- struct rpcrdma_frmr *f = &mw->frmr;
- struct ib_send_wr *invalidate_wr;
-
- dprintk("RPC: %s: invalidating frmr %p\n", __func__, f);
-
- f->fr_state = FRMR_IS_INVALID;
- invalidate_wr = &f->fr_invwr;
-
- memset(invalidate_wr, 0, sizeof(*invalidate_wr));
- f->fr_cqe.done = frwr_wc_localinv;
- invalidate_wr->wr_cqe = &f->fr_cqe;
- invalidate_wr->opcode = IB_WR_LOCAL_INV;
- invalidate_wr->ex.invalidate_rkey = f->fr_mr->rkey;
-
- return invalidate_wr;
-}
-
/* Invalidate all memory regions that were registered for "req".
*
* Sleeps until it is safe for the host CPU to access the
@@ -487,7 +467,7 @@
static void
frwr_op_unmap_sync(struct rpcrdma_xprt *r_xprt, struct rpcrdma_req *req)
{
- struct ib_send_wr *invalidate_wrs, *pos, *prev, *bad_wr;
+ struct ib_send_wr *first, **prev, *last, *bad_wr;
struct rpcrdma_rep *rep = req->rl_reply;
struct rpcrdma_ia *ia = &r_xprt->rx_ia;
struct rpcrdma_mw *mw, *tmp;
@@ -503,23 +483,28 @@
*/
f = NULL;
count = 0;
- invalidate_wrs = pos = prev = NULL;
+ prev = &first;
list_for_each_entry(mw, &req->rl_registered, mw_list) {
+ mw->frmr.fr_state = FRMR_IS_INVALID;
+
if ((rep->rr_wc_flags & IB_WC_WITH_INVALIDATE) &&
- (mw->mw_handle == rep->rr_inv_rkey)) {
- mw->frmr.fr_state = FRMR_IS_INVALID;
+ (mw->mw_handle == rep->rr_inv_rkey))
continue;
- }
- pos = __frwr_prepare_linv_wr(mw);
+ f = &mw->frmr;
+ dprintk("RPC: %s: invalidating frmr %p\n",
+ __func__, f);
+
+ f->fr_cqe.done = frwr_wc_localinv;
+ last = &f->fr_invwr;
+ memset(last, 0, sizeof(*last));
+ last->wr_cqe = &f->fr_cqe;
+ last->opcode = IB_WR_LOCAL_INV;
+ last->ex.invalidate_rkey = mw->mw_handle;
count++;
- if (!invalidate_wrs)
- invalidate_wrs = pos;
- else
- prev->next = pos;
- prev = pos;
- f = &mw->frmr;
+ *prev = last;
+ prev = &last->next;
}
if (!f)
goto unmap;
@@ -528,7 +513,7 @@
* last WR in the chain completes, all WRs in the chain
* are complete.
*/
- f->fr_invwr.send_flags = IB_SEND_SIGNALED;
+ last->send_flags = IB_SEND_SIGNALED;
f->fr_cqe.done = frwr_wc_localinv_wake;
reinit_completion(&f->fr_linv_done);
@@ -543,7 +528,7 @@
* unless ri_id->qp is a valid pointer.
*/
r_xprt->rx_stats.local_inv_needed++;
- rc = ib_post_send(ia->ri_id->qp, invalidate_wrs, &bad_wr);
+ rc = ib_post_send(ia->ri_id->qp, first, &bad_wr);
if (rc)
goto reset_mrs;
@@ -554,7 +539,7 @@
*/
unmap:
list_for_each_entry_safe(mw, tmp, &req->rl_registered, mw_list) {
- dprintk("RPC: %s: unmapping frmr %p\n",
+ dprintk("RPC: %s: DMA unmapping frmr %p\n",
__func__, &mw->frmr);
list_del_init(&mw->mw_list);
ib_dma_unmap_sg(ia->ri_device,
@@ -572,7 +557,7 @@
*/
list_for_each_entry(mw, &req->rl_registered, mw_list) {
f = &mw->frmr;
- if (mw->frmr.fr_mr->rkey == bad_wr->ex.invalidate_rkey) {
+ if (mw->mw_handle == bad_wr->ex.invalidate_rkey) {
__frwr_reset_mr(ia, mw);
bad_wr = bad_wr->next;
}
--
To unsubscribe from this list: send the line "unsubscribe linux-nfs" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply related
* [PATCH v1 14/14] xprtrdma: Update documenting comment
From: Chuck Lever @ 2016-11-09 19:06 UTC (permalink / raw)
To: linux-rdma-u79uwXL29TY76Z2rM5mHXA,
linux-nfs-u79uwXL29TY76Z2rM5mHXA
In-Reply-To: <20161109184735.15007.96507.stgit-FYjufvaPoItvLzlybtyyYzGyq/o6K9yX@public.gmane.org>
Clean up: If reset fails, FRMRs are no longer abandoned, rather
they are released immediately. Update the comment to reflect this.
Fixes: 2ffc871a574d ('xprtrdma: Release orphaned MRs immediately')
Signed-off-by: Chuck Lever <chuck.lever-QHcLZuEGTsvQT0dZR+AlfA@public.gmane.org>
---
net/sunrpc/xprtrdma/frwr_ops.c | 4 ----
1 file changed, 4 deletions(-)
diff --git a/net/sunrpc/xprtrdma/frwr_ops.c b/net/sunrpc/xprtrdma/frwr_ops.c
index 900dc40..47bed53 100644
--- a/net/sunrpc/xprtrdma/frwr_ops.c
+++ b/net/sunrpc/xprtrdma/frwr_ops.c
@@ -171,10 +171,6 @@
}
/* Reset of a single FRMR. Generate a fresh rkey by replacing the MR.
- *
- * There's no recovery if this fails. The FRMR is abandoned, but
- * remains in rb_all. It will be cleaned up when the transport is
- * destroyed.
*/
static void
frwr_op_recover_mr(struct rpcrdma_mw *mw)
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply related
* [PATCH net 0/2] qed: Fix RoCE infrastructure
From: Yuval Mintz @ 2016-11-09 20:48 UTC (permalink / raw)
To: davem-fT/PcQaiUtIeIZ0/mPfg9Q, netdev-u79uwXL29TY76Z2rM5mHXA
Cc: linux-rdma-u79uwXL29TY76Z2rM5mHXA,
dledford-H+wXaHxf7aLQT0dZR+AlfA,
Ram.Amrani-YGCgFSpz5w/QT0dZR+AlfA, Yuval Mintz
This series fixes 2 basic issues with RoCE support,
one handles a missing configuration in the initial infrastructure
support while the other is a regression introduced by one of the
initial fix submissions.
Dave,
I've built this series against 'net' [specifically on top of
cdb26d3387f0 ("net: bgmac: fix reversed checks for clock control flag")],
but I'm not entirely convinced whether that's the right place for them -
On one-hand, they're RoCE-specific fixes, but they're also infrastructure
fixes that are wholly contained inside qed.
Would like to hear your [and Doug's] preference on whether we should
target such patches to net or linux-rdma.
Thanks,
Yuval
Ram Amrani (2):
qed: configure ll2 RoCE v1/v2 flavor correctly
qed: correct rdma params configuration
drivers/net/ethernet/qlogic/qed/qed_hsi.h | 3 ---
drivers/net/ethernet/qlogic/qed/qed_ll2.c | 1 +
drivers/net/ethernet/qlogic/qed/qed_main.c | 17 ++++++++---------
3 files changed, 9 insertions(+), 12 deletions(-)
--
1.9.3
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply
* [PATCH net 1/2] qed: configure ll2 RoCE v1/v2 flavor correctly
From: Yuval Mintz @ 2016-11-09 20:48 UTC (permalink / raw)
To: davem-fT/PcQaiUtIeIZ0/mPfg9Q, netdev-u79uwXL29TY76Z2rM5mHXA
Cc: linux-rdma-u79uwXL29TY76Z2rM5mHXA,
dledford-H+wXaHxf7aLQT0dZR+AlfA,
Ram.Amrani-YGCgFSpz5w/QT0dZR+AlfA, Yuval Mintz
In-Reply-To: <1478724524-16940-1-git-send-email-Yuval.Mintz-YGCgFSpz5w/QT0dZR+AlfA@public.gmane.org>
From: Ram Amrani <Ram.Amrani-YGCgFSpz5w/QT0dZR+AlfA@public.gmane.org>
Currently RoCE v2 won't operate with RDMA CM due to missing setting of
the roce-flavour in the ll2 configuration.
This patch properly sets the flavour, and deletes incorrect HSI
that doesn't [yet] exist.
Fixes: abd49676c707 ("qed: Add RoCE ll2 & GSI support")
Signed-off-by: Ram Amrani <Ram.Amrani-YGCgFSpz5w/QT0dZR+AlfA@public.gmane.org>
Signed-off-by: Yuval Mintz <Yuval.Mintz-YGCgFSpz5w/QT0dZR+AlfA@public.gmane.org>
---
drivers/net/ethernet/qlogic/qed/qed_hsi.h | 3 ---
drivers/net/ethernet/qlogic/qed/qed_ll2.c | 1 +
2 files changed, 1 insertion(+), 3 deletions(-)
diff --git a/drivers/net/ethernet/qlogic/qed/qed_hsi.h b/drivers/net/ethernet/qlogic/qed/qed_hsi.h
index 72eee29..2777d5b 100644
--- a/drivers/net/ethernet/qlogic/qed/qed_hsi.h
+++ b/drivers/net/ethernet/qlogic/qed/qed_hsi.h
@@ -727,9 +727,6 @@ struct core_tx_bd_flags {
#define CORE_TX_BD_FLAGS_L4_PROTOCOL_SHIFT 6
#define CORE_TX_BD_FLAGS_L4_PSEUDO_CSUM_MODE_MASK 0x1
#define CORE_TX_BD_FLAGS_L4_PSEUDO_CSUM_MODE_SHIFT 7
-#define CORE_TX_BD_FLAGS_ROCE_FLAV_MASK 0x1
-#define CORE_TX_BD_FLAGS_ROCE_FLAV_SHIFT 12
-
};
struct core_tx_bd {
diff --git a/drivers/net/ethernet/qlogic/qed/qed_ll2.c b/drivers/net/ethernet/qlogic/qed/qed_ll2.c
index 63e1a1b..f95385c 100644
--- a/drivers/net/ethernet/qlogic/qed/qed_ll2.c
+++ b/drivers/net/ethernet/qlogic/qed/qed_ll2.c
@@ -1119,6 +1119,7 @@ static void qed_ll2_prepare_tx_packet_set_bd(struct qed_hwfn *p_hwfn,
start_bd->bd_flags.as_bitfield |= CORE_TX_BD_FLAGS_START_BD_MASK <<
CORE_TX_BD_FLAGS_START_BD_SHIFT;
SET_FIELD(start_bd->bitfield0, CORE_TX_BD_NBDS, num_of_bds);
+ SET_FIELD(start_bd->bitfield0, CORE_TX_BD_ROCE_FLAV, type);
DMA_REGPAIR_LE(start_bd->addr, first_frag);
start_bd->nbytes = cpu_to_le16(first_frag_len);
--
1.9.3
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply related
* [PATCH net 2/2] qed: Correct rdma params configuration
From: Yuval Mintz @ 2016-11-09 20:48 UTC (permalink / raw)
To: davem-fT/PcQaiUtIeIZ0/mPfg9Q, netdev-u79uwXL29TY76Z2rM5mHXA
Cc: linux-rdma-u79uwXL29TY76Z2rM5mHXA,
dledford-H+wXaHxf7aLQT0dZR+AlfA,
Ram.Amrani-YGCgFSpz5w/QT0dZR+AlfA, Yuval Mintz
In-Reply-To: <1478724524-16940-1-git-send-email-Yuval.Mintz-YGCgFSpz5w/QT0dZR+AlfA@public.gmane.org>
From: Ram Amrani <Ram.Amrani-YGCgFSpz5w/QT0dZR+AlfA@public.gmane.org>
Previous fix has broken RoCE support as the rdma_pf_params are now
being set into the parameters only after the params are alrady assigned
into the hw-function.
Fixes: 0189efb8f4f8 ("qed*: Fix Kconfig dependencies with INFINIBAND_QEDR")
Signed-off-by: Ram Amrani <Ram.Amrani-YGCgFSpz5w/QT0dZR+AlfA@public.gmane.org>
Signed-off-by: Yuval Mintz <Yuval.Mintz-YGCgFSpz5w/QT0dZR+AlfA@public.gmane.org>
---
drivers/net/ethernet/qlogic/qed/qed_main.c | 17 ++++++++---------
1 file changed, 8 insertions(+), 9 deletions(-)
diff --git a/drivers/net/ethernet/qlogic/qed/qed_main.c b/drivers/net/ethernet/qlogic/qed/qed_main.c
index c418360..333c744 100644
--- a/drivers/net/ethernet/qlogic/qed/qed_main.c
+++ b/drivers/net/ethernet/qlogic/qed/qed_main.c
@@ -839,20 +839,19 @@ static void qed_update_pf_params(struct qed_dev *cdev,
{
int i;
+ if (IS_ENABLED(CONFIG_QED_RDMA)) {
+ params->rdma_pf_params.num_qps = QED_ROCE_QPS;
+ params->rdma_pf_params.min_dpis = QED_ROCE_DPIS;
+ /* divide by 3 the MRs to avoid MF ILT overflow */
+ params->rdma_pf_params.num_mrs = RDMA_MAX_TIDS;
+ params->rdma_pf_params.gl_pi = QED_ROCE_PROTOCOL_INDEX;
+ }
+
for (i = 0; i < cdev->num_hwfns; i++) {
struct qed_hwfn *p_hwfn = &cdev->hwfns[i];
p_hwfn->pf_params = *params;
}
-
- if (!IS_ENABLED(CONFIG_QED_RDMA))
- return;
-
- params->rdma_pf_params.num_qps = QED_ROCE_QPS;
- params->rdma_pf_params.min_dpis = QED_ROCE_DPIS;
- /* divide by 3 the MRs to avoid MF ILT overflow */
- params->rdma_pf_params.num_mrs = RDMA_MAX_TIDS;
- params->rdma_pf_params.gl_pi = QED_ROCE_PROTOCOL_INDEX;
}
static int qed_slowpath_start(struct qed_dev *cdev,
--
1.9.3
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply related
* Re: [PATCH rdma-next 2/4] IB/core: Support rate limit for packet pacing
From: Bodong Wang @ 2016-11-09 21:00 UTC (permalink / raw)
To: Hefty, Sean, Leon Romanovsky,
dledford-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org
Cc: linux-rdma-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
In-Reply-To: <1828884A29C6694DAF28B7E6B8A82373AB0A7F70-P5GAC/sN6hkd3b2yrw5b5LfspsVTdybXVpNB7YpNyf8@public.gmane.org>
On 11/9/2016 11:27 AM, Hefty, Sean wrote:
>> enum ib_qp_state {
>> @@ -1151,6 +1152,7 @@ struct ib_qp_attr {
>> u8 rnr_retry;
>> u8 alt_port_num;
>> u8 alt_timeout;
>> + u32 rate_limit;
>> };
> We already have ib_qp_attr::ib_ah_attr::static_rate, and that accounts for both the primary and alternate paths. We should not add a conflicting rate_limit field. Either use static_rate as defined by the spec, or replace/update it, with corresponding changes to how it is used in conjunction with SM data.
They are different features. Static rate has a limitation on how many
different speeds we could get, while packet pacing(rate limit) allows
customer to set any number between the min and max range.
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply
* Re: wireshark's RPC-over-RDMA dissector
From: Chuck Lever @ 2016-11-09 22:42 UTC (permalink / raw)
To: Parav Pandit; +Cc: Linux NFS Mailing List, List Linux RDMA Mailing
In-Reply-To: <CAG53R5WQRCD32iEefGoaRqggAT1hyv_FSpCAN3c3Z7sVspzx1Q-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
> On Nov 9, 2016, at 11:41 AM, Parav Pandit <pandit.parav-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote:
>
> Hi Chuck,
>
> Just FYI.
> I have committed few fixes in wireshark trunk for RoCE and IB
> dissectors which will in general benefit other ULPs as well (primary
> for statefulness of ULPs) last month.
>
> I have few patches pending in my sandbox in area of RoCE and for other
> ULP that I will push in coming days, likely next week.
> I am currently actively testing them and have made steady progress so far.
>
> I will try to find sometime to review them next week.
>
> Regards,
> Parav Pandit
Stateful dissection might be helpful for associating RPC-over-RDMA
headers with RDMA Reads and Writes on the same QP. Thanks Parav,
I'll watch for your work.
> On Wed, Nov 9, 2016 at 9:35 PM, Chuck Lever <chuck.lever-QHcLZuEGTsvQT0dZR+AlfA@public.gmane.org> wrote:
>> Hi-
>>
>> Thanks to Yan Berman, for a couple of years now we've had a basic
>> RPC-over-RDMA dissector in wireshark that can be used with ibdump
>> captures. There have been some bugs noted, but no-one has had the
>> cycles to dig in and address.
>>
>> Recently Tom Haynes helped me set up my own wireshark build so I
>> could help address some of the known issues.
>>
>> http://git.linux-nfs.org/?p=cel/wireshark.git;a=shortlog;h=refs/heads/rpc-rdma-fixes
>>
>> Posting here for review before I take the next steps to push these
>> to the wireshark community. Constructive critique and other
>> suggestions are welcome.
>>
>> The fixes so far focus on dissecting transport headers correctly.
>> There continue to be significant open issues with the dissector:
>> • There does not appear to be any support for dissecting
>> RPC-over-RDMA on iWARP or RoCE
>> • The NFS dissector does not handle portions of the XDR
>> stream that were transmitted via RDMA Read/Write
>> • RPC messages conveyed via RDMA_NOMSG are not recognized
>> or dissected
>> • There is no association between RDMA Reads and Writes
>> and the RPC-over-RDMA message they go with
>> • A CREQ / CREP pair are needed to identify which QP
>> numbers are used for RPC-over-RDMA traffic
>> • With TCP, the dissector fully outdents the RPC and NFS
>> dissection results; but with RDMA, the dissector places
>> the results in the tree under the Infiniband header
>> • Not enough error detection in the dissector
>>
>> --
>> Chuck Lever
--
Chuck Lever
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply
* [PATCH] i40iw: Use vector when creating CQs
From: Henry Orosco @ 2016-11-10 3:24 UTC (permalink / raw)
To: dledford-H+wXaHxf7aLQT0dZR+AlfA
Cc: linux-rdma-u79uwXL29TY76Z2rM5mHXA,
e1000-rdma-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f, Henry Orosco
Assign each CEQ vector to a different CPU when possible, then
when creating a CQ, use the vector for the CEQ id. This
allows completion work to be distributed over multiple cores.
Signed-off-by: Mustafa Ismail <mustafa.ismail-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
Signed-off-by: Henry Orosco <henry.orosco-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
---
drivers/infiniband/hw/i40iw/i40iw_main.c | 8 +++++++-
drivers/infiniband/hw/i40iw/i40iw_verbs.c | 5 +++--
2 files changed, 10 insertions(+), 3 deletions(-)
diff --git a/drivers/infiniband/hw/i40iw/i40iw_main.c b/drivers/infiniband/hw/i40iw/i40iw_main.c
index 078d784..154a9d4 100644
--- a/drivers/infiniband/hw/i40iw/i40iw_main.c
+++ b/drivers/infiniband/hw/i40iw/i40iw_main.c
@@ -270,6 +270,7 @@ static void i40iw_disable_irq(struct i40iw_sc_dev *dev,
i40iw_wr32(dev->hw, I40E_PFINT_DYN_CTLN(msix_vec->idx - 1), 0);
else
i40iw_wr32(dev->hw, I40E_VFINT_DYN_CTLN1(msix_vec->idx - 1), 0);
+ irq_set_affinity_hint(msix_vec->irq, NULL);
free_irq(msix_vec->irq, dev_id);
}
@@ -688,6 +689,7 @@ static enum i40iw_status_code i40iw_configure_ceq_vector(struct i40iw_device *iw
struct i40iw_msix_vector *msix_vec)
{
enum i40iw_status_code status;
+ cpumask_t mask;
if (iwdev->msix_shared && !ceq_id) {
tasklet_init(&iwdev->dpc_tasklet, i40iw_dpc, (unsigned long)iwdev);
@@ -697,12 +699,15 @@ static enum i40iw_status_code i40iw_configure_ceq_vector(struct i40iw_device *iw
status = request_irq(msix_vec->irq, i40iw_ceq_handler, 0, "CEQ", iwceq);
}
+ cpumask_clear(&mask);
+ cpumask_set_cpu(msix_vec->cpu_affinity, &mask);
+ irq_set_affinity_hint(msix_vec->irq, &mask);
+
if (status) {
i40iw_pr_err("ceq irq config fail\n");
return I40IW_ERR_CONFIG;
}
msix_vec->ceq_id = ceq_id;
- msix_vec->cpu_affinity = 0;
return 0;
}
@@ -1384,6 +1389,7 @@ static enum i40iw_status_code i40iw_save_msix_info(struct i40iw_device *iwdev,
for (i = 0, ceq_idx = 0; i < iwdev->msix_count; i++, iw_qvinfo++) {
iwdev->iw_msixtbl[i].idx = ldev->msix_entries[i].entry;
iwdev->iw_msixtbl[i].irq = ldev->msix_entries[i].vector;
+ iwdev->iw_msixtbl[i].cpu_affinity = ceq_idx;
if (i == 0) {
iw_qvinfo->aeq_idx = 0;
if (iwdev->msix_shared)
diff --git a/drivers/infiniband/hw/i40iw/i40iw_verbs.c b/drivers/infiniband/hw/i40iw/i40iw_verbs.c
index 6329c97..2dbbf67 100644
--- a/drivers/infiniband/hw/i40iw/i40iw_verbs.c
+++ b/drivers/infiniband/hw/i40iw/i40iw_verbs.c
@@ -1137,7 +1137,8 @@ static struct ib_cq *i40iw_create_cq(struct ib_device *ibdev,
ukinfo->cq_id = cq_num;
iwcq->ibcq.cqe = info.cq_uk_init_info.cq_size;
info.ceqe_mask = 0;
- info.ceq_id = 0;
+ if (attr->comp_vector < iwdev->ceqs_count)
+ info.ceq_id = attr->comp_vector;
info.ceq_id_valid = true;
info.ceqe_mask = 1;
info.type = I40IW_CQ_TYPE_IWARP;
@@ -2621,7 +2622,7 @@ static struct i40iw_ib_device *i40iw_init_rdma_device(struct i40iw_device *iwdev
(1ull << IB_USER_VERBS_CMD_POST_RECV) |
(1ull << IB_USER_VERBS_CMD_POST_SEND);
iwibdev->ibdev.phys_port_cnt = 1;
- iwibdev->ibdev.num_comp_vectors = 1;
+ iwibdev->ibdev.num_comp_vectors = iwdev->ceqs_count;
iwibdev->ibdev.dma_device = &pcidev->dev;
iwibdev->ibdev.dev.parent = &pcidev->dev;
iwibdev->ibdev.query_port = i40iw_query_port;
--
1.8.3.1
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply related
* [PATCH] i40iw: Convert page_size to encoded value
From: Henry Orosco @ 2016-11-10 3:25 UTC (permalink / raw)
To: dledford-H+wXaHxf7aLQT0dZR+AlfA
Cc: linux-rdma-u79uwXL29TY76Z2rM5mHXA,
e1000-rdma-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f, Henry Orosco
Passed in page_size was used as encoded value for writing
the WQE and passed in value was usually 4096. This was
working out since bit 0 was 0 and implies 4KB pages,
but would not work for other page sizes.
Signed-off-by: Mustafa Ismail <mustafa.ismail-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
Signed-off-by: Henry Orosco <henry.orosco-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
---
drivers/infiniband/hw/i40iw/i40iw_ctrl.c | 12 +++++++++---
drivers/infiniband/hw/i40iw/i40iw_type.h | 5 +++++
2 files changed, 14 insertions(+), 3 deletions(-)
diff --git a/drivers/infiniband/hw/i40iw/i40iw_ctrl.c b/drivers/infiniband/hw/i40iw/i40iw_ctrl.c
index cd71444..bdb4421 100644
--- a/drivers/infiniband/hw/i40iw/i40iw_ctrl.c
+++ b/drivers/infiniband/hw/i40iw/i40iw_ctrl.c
@@ -2613,7 +2613,9 @@ static enum i40iw_status_code i40iw_sc_alloc_stag(
u64 *wqe;
struct i40iw_sc_cqp *cqp;
u64 header;
+ enum i40iw_page_size page_size;
+ page_size = (info->page_size == 0x200000) ? I40IW_PAGE_SIZE_2M : I40IW_PAGE_SIZE_4K;
cqp = dev->cqp;
wqe = i40iw_sc_cqp_get_next_send_wqe(cqp, scratch);
if (!wqe)
@@ -2633,7 +2635,7 @@ static enum i40iw_status_code i40iw_sc_alloc_stag(
LS_64(1, I40IW_CQPSQ_STAG_MR) |
LS_64(info->access_rights, I40IW_CQPSQ_STAG_ARIGHTS) |
LS_64(info->chunk_size, I40IW_CQPSQ_STAG_LPBLSIZE) |
- LS_64(info->page_size, I40IW_CQPSQ_STAG_HPAGESIZE) |
+ LS_64(page_size, I40IW_CQPSQ_STAG_HPAGESIZE) |
LS_64(info->remote_access, I40IW_CQPSQ_STAG_REMACCENABLED) |
LS_64(info->use_hmc_fcn_index, I40IW_CQPSQ_STAG_USEHMCFNIDX) |
LS_64(info->use_pf_rid, I40IW_CQPSQ_STAG_USEPFRID) |
@@ -2669,7 +2671,9 @@ static enum i40iw_status_code i40iw_sc_mr_reg_non_shared(
u32 pble_obj_cnt;
bool remote_access;
u8 addr_type;
+ enum i40iw_page_size page_size;
+ page_size = (info->page_size == 0x200000) ? I40IW_PAGE_SIZE_2M : I40IW_PAGE_SIZE_4K;
if (info->access_rights & (I40IW_ACCESS_FLAGS_REMOTEREAD_ONLY |
I40IW_ACCESS_FLAGS_REMOTEWRITE_ONLY))
remote_access = true;
@@ -2712,7 +2716,7 @@ static enum i40iw_status_code i40iw_sc_mr_reg_non_shared(
header = LS_64(I40IW_CQP_OP_REG_MR, I40IW_CQPSQ_OPCODE) |
LS_64(1, I40IW_CQPSQ_STAG_MR) |
LS_64(info->chunk_size, I40IW_CQPSQ_STAG_LPBLSIZE) |
- LS_64(info->page_size, I40IW_CQPSQ_STAG_HPAGESIZE) |
+ LS_64(page_size, I40IW_CQPSQ_STAG_HPAGESIZE) |
LS_64(info->access_rights, I40IW_CQPSQ_STAG_ARIGHTS) |
LS_64(remote_access, I40IW_CQPSQ_STAG_REMACCENABLED) |
LS_64(addr_type, I40IW_CQPSQ_STAG_VABASEDTO) |
@@ -2927,7 +2931,9 @@ enum i40iw_status_code i40iw_sc_mr_fast_register(
u64 temp, header;
u64 *wqe;
u32 wqe_idx;
+ enum i40iw_page_size page_size;
+ page_size = (info->page_size == 0x200000) ? I40IW_PAGE_SIZE_2M : I40IW_PAGE_SIZE_4K;
wqe = i40iw_qp_get_next_send_wqe(&qp->qp_uk, &wqe_idx, I40IW_QP_WQE_MIN_SIZE,
0, info->wr_id);
if (!wqe)
@@ -2954,7 +2960,7 @@ enum i40iw_status_code i40iw_sc_mr_fast_register(
LS_64(info->stag_idx, I40IWQPSQ_STAGINDEX) |
LS_64(I40IWQP_OP_FAST_REGISTER, I40IWQPSQ_OPCODE) |
LS_64(info->chunk_size, I40IWQPSQ_LPBLSIZE) |
- LS_64(info->page_size, I40IWQPSQ_HPAGESIZE) |
+ LS_64(page_size, I40IWQPSQ_HPAGESIZE) |
LS_64(info->access_rights, I40IWQPSQ_STAGRIGHTS) |
LS_64(info->addr_type, I40IWQPSQ_VABASEDTO) |
LS_64(info->read_fence, I40IWQPSQ_READFENCE) |
diff --git a/drivers/infiniband/hw/i40iw/i40iw_type.h b/drivers/infiniband/hw/i40iw/i40iw_type.h
index f60f0e2..1c58ba6 100644
--- a/drivers/infiniband/hw/i40iw/i40iw_type.h
+++ b/drivers/infiniband/hw/i40iw/i40iw_type.h
@@ -74,6 +74,11 @@ struct i40iw_cq_shadow_area {
struct i40iw_priv_cq_ops;
struct i40iw_hmc_ops;
+enum i40iw_page_size {
+ I40IW_PAGE_SIZE_4K = 0,
+ I40IW_PAGE_SIZE_2M
+};
+
enum i40iw_resource_indicator_type {
I40IW_RSRC_INDICATOR_TYPE_ADAPTER = 0,
I40IW_RSRC_INDICATOR_TYPE_CQ,
--
1.8.3.1
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply related
* [PATCH] i40iw: Correct values for max_recv_sge, max_send_sge
From: Henry Orosco @ 2016-11-10 3:26 UTC (permalink / raw)
To: dledford-H+wXaHxf7aLQT0dZR+AlfA
Cc: linux-rdma-u79uwXL29TY76Z2rM5mHXA,
e1000-rdma-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f, Henry Orosco
When creating QPs, ensure init_attr->cap.max_recv_sge
is clipped to MAX_FRAG_COUNT.
Expose MAX_FRAG_COUNT for max_recv_sge and max_send_sge in
i40iw_query_qp().
Signed-off-by: Shiraz Saleem <shiraz.saleem-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
Signed-off-by: Henry Orosco <henry.orosco-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
---
drivers/infiniband/hw/i40iw/i40iw_verbs.c | 6 +++++-
1 file changed, 5 insertions(+), 1 deletion(-)
diff --git a/drivers/infiniband/hw/i40iw/i40iw_verbs.c b/drivers/infiniband/hw/i40iw/i40iw_verbs.c
index 6329c97..d06fe593 100644
--- a/drivers/infiniband/hw/i40iw/i40iw_verbs.c
+++ b/drivers/infiniband/hw/i40iw/i40iw_verbs.c
@@ -610,6 +610,9 @@ static struct ib_qp *i40iw_create_qp(struct ib_pd *ibpd,
if (init_attr->cap.max_send_sge > I40IW_MAX_WQ_FRAGMENT_COUNT)
init_attr->cap.max_send_sge = I40IW_MAX_WQ_FRAGMENT_COUNT;
+ if (init_attr->cap.max_recv_sge > I40IW_MAX_WQ_FRAGMENT_COUNT)
+ init_attr->cap.max_recv_sge = I40IW_MAX_WQ_FRAGMENT_COUNT;
+
memset(&init_info, 0, sizeof(init_info));
sq_size = init_attr->cap.max_send_wr;
@@ -815,8 +818,9 @@ static int i40iw_query_qp(struct ib_qp *ibqp,
attr->qp_access_flags = 0;
attr->cap.max_send_wr = qp->qp_uk.sq_size;
attr->cap.max_recv_wr = qp->qp_uk.rq_size;
- attr->cap.max_recv_sge = 1;
attr->cap.max_inline_data = I40IW_MAX_INLINE_DATA_SIZE;
+ attr->cap.max_send_sge = I40IW_MAX_WQ_FRAGMENT_COUNT;
+ attr->cap.max_recv_sge = I40IW_MAX_WQ_FRAGMENT_COUNT;
init_attr->event_handler = iwqp->ibqp.event_handler;
init_attr->qp_context = iwqp->ibqp.qp_context;
init_attr->send_cq = iwqp->ibqp.send_cq;
--
1.8.3.1
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply related
* [PATCH] i40iw: Fix for LAN handler removal
From: Henry Orosco @ 2016-11-10 3:27 UTC (permalink / raw)
To: dledford-H+wXaHxf7aLQT0dZR+AlfA
Cc: linux-rdma-u79uwXL29TY76Z2rM5mHXA,
e1000-rdma-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f, Henry Orosco
If i40iw_open() fails for any reason, the LAN handler
is not being removed. Modify i40iw_deinit_device()
to always remove the handler.
Signed-off-by: Mustafa Ismail <mustafa.ismail-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
Signed-off-by: Henry Orosco <henry.orosco-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
---
drivers/infiniband/hw/i40iw/i40iw_main.c | 10 ++++------
1 file changed, 4 insertions(+), 6 deletions(-)
diff --git a/drivers/infiniband/hw/i40iw/i40iw_main.c b/drivers/infiniband/hw/i40iw/i40iw_main.c
index 078d784..79808cc 100644
--- a/drivers/infiniband/hw/i40iw/i40iw_main.c
+++ b/drivers/infiniband/hw/i40iw/i40iw_main.c
@@ -1404,12 +1404,11 @@ static enum i40iw_status_code i40iw_save_msix_info(struct i40iw_device *iwdev,
* i40iw_deinit_device - clean up the device resources
* @iwdev: iwarp device
* @reset: true if called before reset
- * @del_hdl: true if delete hdl entry
*
* Destroy the ib device interface, remove the mac ip entry and ipv4/ipv6 addresses,
* destroy the device queues and free the pble and the hmc objects
*/
-static void i40iw_deinit_device(struct i40iw_device *iwdev, bool reset, bool del_hdl)
+static void i40iw_deinit_device(struct i40iw_device *iwdev, bool reset)
{
struct i40e_info *ldev = iwdev->ldev;
@@ -1472,8 +1471,7 @@ static void i40iw_deinit_device(struct i40iw_device *iwdev, bool reset, bool del
break;
}
- if (del_hdl)
- i40iw_del_handler(i40iw_find_i40e_handler(ldev));
+ i40iw_del_handler(i40iw_find_i40e_handler(ldev));
kfree(iwdev->hdl);
}
@@ -1635,7 +1633,7 @@ static int i40iw_open(struct i40e_info *ldev, struct i40e_client *client)
} while (0);
i40iw_pr_err("status = %d last completion = %d\n", status, iwdev->init_state);
- i40iw_deinit_device(iwdev, false, false);
+ i40iw_deinit_device(iwdev, false);
return -ERESTART;
}
@@ -1680,7 +1678,7 @@ static void i40iw_close(struct i40e_info *ldev, struct i40e_client *client, bool
iwdev = &hdl->device;
destroy_workqueue(iwdev->virtchnl_wq);
- i40iw_deinit_device(iwdev, reset, true);
+ i40iw_deinit_device(iwdev, reset);
}
/**
--
1.8.3.1
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply related
* [PATCH] i40iw: Optimize inline data copy
From: Henry Orosco @ 2016-11-10 3:28 UTC (permalink / raw)
To: dledford-H+wXaHxf7aLQT0dZR+AlfA
Cc: linux-rdma-u79uwXL29TY76Z2rM5mHXA,
e1000-rdma-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f, Henry Orosco
Use memcpy for inline data copy in sends
and writes instead of byte by byte copy.
Signed-off-by: Mustafa Ismail <mustafa.ismail-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
Signed-off-by: Henry Orosco <henry.orosco-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
---
drivers/infiniband/hw/i40iw/i40iw_uk.c | 24 ++++++++++--------------
1 file changed, 10 insertions(+), 14 deletions(-)
diff --git a/drivers/infiniband/hw/i40iw/i40iw_uk.c b/drivers/infiniband/hw/i40iw/i40iw_uk.c
index 4d28c3c..47cb2e0 100644
--- a/drivers/infiniband/hw/i40iw/i40iw_uk.c
+++ b/drivers/infiniband/hw/i40iw/i40iw_uk.c
@@ -430,7 +430,7 @@ static enum i40iw_status_code i40iw_inline_rdma_write(struct i40iw_qp_uk *qp,
struct i40iw_inline_rdma_write *op_info;
u64 *push;
u64 header = 0;
- u32 i, wqe_idx;
+ u32 wqe_idx;
enum i40iw_status_code ret_code;
bool read_fence = false;
u8 wqe_size;
@@ -465,14 +465,12 @@ static enum i40iw_status_code i40iw_inline_rdma_write(struct i40iw_qp_uk *qp,
src = (u8 *)(op_info->data);
if (op_info->len <= 16) {
- for (i = 0; i < op_info->len; i++, src++, dest++)
- *dest = *src;
+ memcpy(dest, src, op_info->len);
} else {
- for (i = 0; i < 16; i++, src++, dest++)
- *dest = *src;
+ memcpy(dest, src, 16);
+ src += 16;
dest = (u8 *)wqe + 32;
- for (; i < op_info->len; i++, src++, dest++)
- *dest = *src;
+ memcpy(dest, src, op_info->len - 16);
}
wmb(); /* make sure WQE is populated before valid bit is set */
@@ -507,7 +505,7 @@ static enum i40iw_status_code i40iw_inline_send(struct i40iw_qp_uk *qp,
u8 *dest, *src;
struct i40iw_post_inline_send *op_info;
u64 header;
- u32 wqe_idx, i;
+ u32 wqe_idx;
enum i40iw_status_code ret_code;
bool read_fence = false;
u8 wqe_size;
@@ -540,14 +538,12 @@ static enum i40iw_status_code i40iw_inline_send(struct i40iw_qp_uk *qp,
src = (u8 *)(op_info->data);
if (op_info->len <= 16) {
- for (i = 0; i < op_info->len; i++, src++, dest++)
- *dest = *src;
+ memcpy(dest, src, op_info->len);
} else {
- for (i = 0; i < 16; i++, src++, dest++)
- *dest = *src;
+ memcpy(dest, src, 16);
+ src += 16;
dest = (u8 *)wqe + 32;
- for (; i < op_info->len; i++, src++, dest++)
- *dest = *src;
+ memcpy(dest, src, op_info->len - 16);
}
wmb(); /* make sure WQE is populated before valid bit is set */
--
1.8.3.1
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply related
* [PATCH] i40iw: Query device accounts for internal rsrc
From: Henry Orosco @ 2016-11-10 3:30 UTC (permalink / raw)
To: dledford-H+wXaHxf7aLQT0dZR+AlfA
Cc: linux-rdma-u79uwXL29TY76Z2rM5mHXA,
e1000-rdma-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f, Henry Orosco
Some resources are consumed internally and not available to the user.
After hw is initialized, figure out how many resources are consumed
and subtract those numbers from the initial max device capability in
i40iw_query_device().
Signed-off-by: Henry Orosco <henry.orosco-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
---
drivers/infiniband/hw/i40iw/i40iw.h | 4 ++++
drivers/infiniband/hw/i40iw/i40iw_hw.c | 2 +-
drivers/infiniband/hw/i40iw/i40iw_main.c | 15 +++++++++++++++
drivers/infiniband/hw/i40iw/i40iw_user.h | 3 ++-
drivers/infiniband/hw/i40iw/i40iw_verbs.c | 8 ++++----
5 files changed, 26 insertions(+), 6 deletions(-)
diff --git a/drivers/infiniband/hw/i40iw/i40iw.h b/drivers/infiniband/hw/i40iw/i40iw.h
index 8ec09e4..22af66a 100644
--- a/drivers/infiniband/hw/i40iw/i40iw.h
+++ b/drivers/infiniband/hw/i40iw/i40iw.h
@@ -297,6 +297,10 @@ struct i40iw_device {
u32 mr_stagmask;
u32 mpa_version;
bool dcb;
+ u32 used_pds;
+ u32 used_cqs;
+ u32 used_mrs;
+ u32 used_qps;
};
struct i40iw_ib_device {
diff --git a/drivers/infiniband/hw/i40iw/i40iw_hw.c b/drivers/infiniband/hw/i40iw/i40iw_hw.c
index 0c92a40..886b8c8 100644
--- a/drivers/infiniband/hw/i40iw/i40iw_hw.c
+++ b/drivers/infiniband/hw/i40iw/i40iw_hw.c
@@ -62,7 +62,7 @@ u32 i40iw_initialize_hw_resources(struct i40iw_device *iwdev)
max_mr = iwdev->sc_dev.hmc_info->hmc_obj[I40IW_HMC_IW_MR].cnt;
arp_table_size = iwdev->sc_dev.hmc_info->hmc_obj[I40IW_HMC_IW_ARP].cnt;
iwdev->max_cqe = 0xFFFFF;
- num_pds = max_qp * 4;
+ num_pds = I40IW_MAX_PDS;
resources_size = sizeof(struct i40iw_arp_entry) * arp_table_size;
resources_size += sizeof(unsigned long) * BITS_TO_LONGS(max_qp);
resources_size += sizeof(unsigned long) * BITS_TO_LONGS(max_mr);
diff --git a/drivers/infiniband/hw/i40iw/i40iw_main.c b/drivers/infiniband/hw/i40iw/i40iw_main.c
index 078d784..ba45b16 100644
--- a/drivers/infiniband/hw/i40iw/i40iw_main.c
+++ b/drivers/infiniband/hw/i40iw/i40iw_main.c
@@ -1540,6 +1540,20 @@ static enum i40iw_status_code i40iw_setup_init_state(struct i40iw_handler *hdl,
}
/**
+ * i40iw_get_used_rsrc - determine resources used internally
+ * @iwdev: iwarp device
+ *
+ * Called after internal allocations
+ */
+static void i40iw_get_used_rsrc(struct i40iw_device *iwdev)
+{
+ iwdev->used_pds = find_next_zero_bit(iwdev->allocated_pds, iwdev->max_pd, 0);
+ iwdev->used_qps = find_next_zero_bit(iwdev->allocated_qps, iwdev->max_qp, 0);
+ iwdev->used_cqs = find_next_zero_bit(iwdev->allocated_cqs, iwdev->max_cq, 0);
+ iwdev->used_mrs = find_next_zero_bit(iwdev->allocated_mrs, iwdev->max_mr, 0);
+}
+
+/**
* i40iw_open - client interface operation open for iwarp/uda device
* @ldev: lan device information
* @client: iwarp client information, provided during registration
@@ -1611,6 +1625,7 @@ static int i40iw_open(struct i40e_info *ldev, struct i40e_client *client)
status = i40iw_initialize_hw_resources(iwdev);
if (status)
break;
+ i40iw_get_used_rsrc(iwdev);
dev->ccq_ops->ccq_arm(dev->ccq);
status = i40iw_hmc_init_pble(&iwdev->sc_dev, iwdev->pble_rsrc);
if (status)
diff --git a/drivers/infiniband/hw/i40iw/i40iw_user.h b/drivers/infiniband/hw/i40iw/i40iw_user.h
index e65c2baa..66263fc 100644
--- a/drivers/infiniband/hw/i40iw/i40iw_user.h
+++ b/drivers/infiniband/hw/i40iw/i40iw_user.h
@@ -76,7 +76,8 @@ enum i40iw_device_capabilities_const {
I40IW_MAX_ORD_SIZE = 127,
I40IW_MAX_WQ_ENTRIES = 2048,
I40IW_Q2_BUFFER_SIZE = (248 + 100),
- I40IW_QP_CTX_SIZE = 248
+ I40IW_QP_CTX_SIZE = 248,
+ I40IW_MAX_PDS = 32768
};
#define i40iw_handle void *
diff --git a/drivers/infiniband/hw/i40iw/i40iw_verbs.c b/drivers/infiniband/hw/i40iw/i40iw_verbs.c
index 6329c97..bfea715 100644
--- a/drivers/infiniband/hw/i40iw/i40iw_verbs.c
+++ b/drivers/infiniband/hw/i40iw/i40iw_verbs.c
@@ -67,13 +67,13 @@ static int i40iw_query_device(struct ib_device *ibdev,
props->vendor_part_id = iwdev->ldev->pcidev->device;
props->hw_ver = (u32)iwdev->sc_dev.hw_rev;
props->max_mr_size = I40IW_MAX_OUTBOUND_MESSAGE_SIZE;
- props->max_qp = iwdev->max_qp;
+ props->max_qp = iwdev->max_qp - iwdev->used_qps;
props->max_qp_wr = (I40IW_MAX_WQ_ENTRIES >> 2) - 1;
props->max_sge = I40IW_MAX_WQ_FRAGMENT_COUNT;
- props->max_cq = iwdev->max_cq;
+ props->max_cq = iwdev->max_cq - iwdev->used_cqs;
props->max_cqe = iwdev->max_cqe;
- props->max_mr = iwdev->max_mr;
- props->max_pd = iwdev->max_pd;
+ props->max_mr = iwdev->max_mr - iwdev->used_mrs;
+ props->max_pd = iwdev->max_pd - iwdev->used_pds;
props->max_sge_rd = I40IW_MAX_SGE_RD;
props->max_qp_rd_atom = I40IW_MAX_IRD_SIZE;
props->max_qp_init_rd_atom = props->max_qp_rd_atom;
--
1.8.3.1
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply related
* [PATCH] i40iw: Remove checks for more than 48 bytes inline data
From: Henry Orosco @ 2016-11-10 3:32 UTC (permalink / raw)
To: dledford-H+wXaHxf7aLQT0dZR+AlfA
Cc: linux-rdma-u79uwXL29TY76Z2rM5mHXA,
e1000-rdma-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f, Henry Orosco
Remove dead code, which isn't executed because we
return error if the data size is greater than 48 bytes.
Inline data size greater than 48 bytes isn't supported
and the maximum WQE size is 64 bytes.
Signed-off-by: Tatyana Nikolova <tatyana.e.nikolova-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
Signed-off-by: Henry Orosco <henry.orosco-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
---
drivers/infiniband/hw/i40iw/i40iw_uk.c | 6 +-----
1 file changed, 1 insertion(+), 5 deletions(-)
diff --git a/drivers/infiniband/hw/i40iw/i40iw_uk.c b/drivers/infiniband/hw/i40iw/i40iw_uk.c
index 4d28c3c..338d164 100644
--- a/drivers/infiniband/hw/i40iw/i40iw_uk.c
+++ b/drivers/infiniband/hw/i40iw/i40iw_uk.c
@@ -1190,12 +1190,8 @@ enum i40iw_status_code i40iw_inline_data_size_to_wqesize(u32 data_size,
if (data_size <= 16)
*wqe_size = I40IW_QP_WQE_MIN_SIZE;
- else if (data_size <= 48)
- *wqe_size = 64;
- else if (data_size <= 80)
- *wqe_size = 96;
else
- *wqe_size = 128;
+ *wqe_size = 64;
return 0;
}
--
1.8.3.1
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply related
* [PATCH] i40iw: Remove NULL check for cm_node->iwdev
From: Henry Orosco @ 2016-11-10 3:32 UTC (permalink / raw)
To: dledford-H+wXaHxf7aLQT0dZR+AlfA
Cc: linux-rdma-u79uwXL29TY76Z2rM5mHXA,
e1000-rdma-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f, Henry Orosco
It is not necessary to check cm_node->iwdev in
i40iw_rem_ref_cm_node() as it can never be NULL after
a successful call out of i40iw_make_cm_node().
Signed-off-by: Chien Tin Tung <chien.tin.tung-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
Signed-off-by: Henry Orosco <henry.orosco-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
---
drivers/infiniband/hw/i40iw/i40iw_cm.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/infiniband/hw/i40iw/i40iw_cm.c b/drivers/infiniband/hw/i40iw/i40iw_cm.c
index c5e3d7a..71f2452 100644
--- a/drivers/infiniband/hw/i40iw/i40iw_cm.c
+++ b/drivers/infiniband/hw/i40iw/i40iw_cm.c
@@ -2224,7 +2224,7 @@ static void i40iw_rem_ref_cm_node(struct i40iw_cm_node *cm_node)
i40iw_dec_refcnt_listen(cm_core, cm_node->listener, 0, true);
} else {
if (!i40iw_listen_port_in_use(cm_core, cm_node->loc_port) &&
- cm_node->apbvt_set && cm_node->iwdev) {
+ cm_node->apbvt_set) {
i40iw_manage_apbvt(cm_node->iwdev,
cm_node->loc_port,
I40IW_MANAGE_APBVT_DEL);
--
1.8.3.1
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply related
* [PATCH] i40iw: Use actual page size
From: Henry Orosco @ 2016-11-10 3:33 UTC (permalink / raw)
To: dledford-H+wXaHxf7aLQT0dZR+AlfA
Cc: linux-rdma-u79uwXL29TY76Z2rM5mHXA,
e1000-rdma-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f, Henry Orosco
In i40iw_post_send, use the actual page size instead of
encoded page size. This is to be consistent with the
rest of the file.
Signed-off-by: Mustafa Ismail <mustafa.ismail-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
Signed-off-by: Henry Orosco <henry.orosco-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
---
drivers/infiniband/hw/i40iw/i40iw_verbs.c | 5 +----
1 file changed, 1 insertion(+), 4 deletions(-)
diff --git a/drivers/infiniband/hw/i40iw/i40iw_verbs.c b/drivers/infiniband/hw/i40iw/i40iw_verbs.c
index 6329c97..8f138b8 100644
--- a/drivers/infiniband/hw/i40iw/i40iw_verbs.c
+++ b/drivers/infiniband/hw/i40iw/i40iw_verbs.c
@@ -2142,7 +2142,6 @@ static int i40iw_post_send(struct ib_qp *ibqp,
case IB_WR_REG_MR:
{
struct i40iw_mr *iwmr = to_iwmr(reg_wr(ib_wr)->mr);
- int page_shift = ilog2(reg_wr(ib_wr)->mr->page_size);
int flags = reg_wr(ib_wr)->access;
struct i40iw_pble_alloc *palloc = &iwmr->iwpbl.pble_alloc;
struct i40iw_sc_dev *dev = &iwqp->iwdev->sc_dev;
@@ -2153,6 +2152,7 @@ static int i40iw_post_send(struct ib_qp *ibqp,
info.access_rights |= i40iw_get_user_access(flags);
info.stag_key = reg_wr(ib_wr)->key & 0xff;
info.stag_idx = reg_wr(ib_wr)->key >> 8;
+ info.page_size = reg_wr(ib_wr)->mr->page_size;
info.wr_id = ib_wr->wr_id;
info.addr_type = I40IW_ADDR_TYPE_VA_BASED;
@@ -2166,9 +2166,6 @@ static int i40iw_post_send(struct ib_qp *ibqp,
if (iwmr->npages > I40IW_MIN_PAGES_PER_FMR)
info.chunk_size = 1;
- if (page_shift == 21)
- info.page_size = 1; /* 2M page */
-
ret = dev->iw_priv_qp_ops->iw_mr_fast_register(&iwqp->sc_qp, &info, true);
if (ret)
err = -ENOMEM;
--
1.8.3.1
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply related
* [PATCH] i40iw: Use runtime check for IS_ENABLED(CONFIG_IPV6)
From: Henry Orosco @ 2016-11-10 3:34 UTC (permalink / raw)
To: dledford-H+wXaHxf7aLQT0dZR+AlfA
Cc: linux-rdma-u79uwXL29TY76Z2rM5mHXA,
e1000-rdma-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f, Henry Orosco
To be consistent, use the runtime check instead of
conditional compile.
Signed-off-by: Mustafa Ismail <mustafa.ismail-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
Signed-off-by: Henry Orosco <henry.orosco-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
---
drivers/infiniband/hw/i40iw/i40iw_cm.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/drivers/infiniband/hw/i40iw/i40iw_cm.c b/drivers/infiniband/hw/i40iw/i40iw_cm.c
index c5e3d7a..aaca1d0 100644
--- a/drivers/infiniband/hw/i40iw/i40iw_cm.c
+++ b/drivers/infiniband/hw/i40iw/i40iw_cm.c
@@ -1579,9 +1579,10 @@ static enum i40iw_status_code i40iw_del_multiple_qhash(
static struct net_device *i40iw_netdev_vlan_ipv6(u32 *addr, u16 *vlan_id, u8 *mac)
{
struct net_device *ip_dev = NULL;
-#if IS_ENABLED(CONFIG_IPV6)
struct in6_addr laddr6;
+ if (!IS_ENABLED(CONFIG_IPV6))
+ return NULL;
i40iw_copy_ip_htonl(laddr6.in6_u.u6_addr32, addr);
if (vlan_id)
*vlan_id = I40IW_NO_VLAN;
@@ -1598,7 +1599,6 @@ static struct net_device *i40iw_netdev_vlan_ipv6(u32 *addr, u16 *vlan_id, u8 *ma
}
}
rcu_read_unlock();
-#endif
return ip_dev;
}
--
1.8.3.1
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply related
* [PATCH 1/2] i40iw: Remove variable flush_code and check to set qp->sq_flush
From: Henry Orosco @ 2016-11-10 3:42 UTC (permalink / raw)
To: dledford-H+wXaHxf7aLQT0dZR+AlfA
Cc: linux-rdma-u79uwXL29TY76Z2rM5mHXA,
e1000-rdma-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f, Henry Orosco
The flush_code variable in i40iw_bld_terminate_hdr() is obsolete and
the check to set qp->sq_flush is unreachable. Currently flush code is
populated in setup_term_hdr() and both SQ and RQ are flushed always
as part of the tear down flow.
Signed-off-by: Shiraz Saleem <shiraz.saleem-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
Signed-off-by: Henry Orosco <henry.orosco-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
---
drivers/infiniband/hw/i40iw/i40iw_ctrl.c | 4 ----
1 file changed, 4 deletions(-)
diff --git a/drivers/infiniband/hw/i40iw/i40iw_ctrl.c b/drivers/infiniband/hw/i40iw/i40iw_ctrl.c
index cd71444..57a4236 100644
--- a/drivers/infiniband/hw/i40iw/i40iw_ctrl.c
+++ b/drivers/infiniband/hw/i40iw/i40iw_ctrl.c
@@ -4045,7 +4045,6 @@ static int i40iw_bld_terminate_hdr(struct i40iw_sc_qp *qp,
u16 ddp_seg_len;
int copy_len = 0;
u8 is_tagged = 0;
- enum i40iw_flush_opcode flush_code = FLUSH_INVALID;
u32 opcode;
struct i40iw_terminate_hdr *termhdr;
@@ -4218,9 +4217,6 @@ static int i40iw_bld_terminate_hdr(struct i40iw_sc_qp *qp,
if (copy_len)
memcpy(termhdr + 1, pkt, copy_len);
- if (flush_code && !info->in_rdrsp_wr)
- qp->sq_flush = (info->sq) ? true : false;
-
return sizeof(struct i40iw_terminate_hdr) + copy_len;
}
--
1.8.3.1
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply related
* [PATCH] i40iw: Remove check on return from device_init_pestat()
From: Henry Orosco @ 2016-11-10 3:42 UTC (permalink / raw)
To: dledford-H+wXaHxf7aLQT0dZR+AlfA
Cc: linux-rdma-u79uwXL29TY76Z2rM5mHXA,
e1000-rdma-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f, Henry Orosco
Remove unnecessary check for return code from
device_init_pestat() and change func to void.
Signed-off-by: Shiraz Saleem <shiraz.saleem-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
Signed-off-by: Henry Orosco <henry.orosco-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
---
drivers/infiniband/hw/i40iw/i40iw_ctrl.c | 10 ++--------
drivers/infiniband/hw/i40iw/i40iw_p.h | 2 +-
drivers/infiniband/hw/i40iw/i40iw_virtchnl.c | 6 +-----
3 files changed, 4 insertions(+), 14 deletions(-)
diff --git a/drivers/infiniband/hw/i40iw/i40iw_ctrl.c b/drivers/infiniband/hw/i40iw/i40iw_ctrl.c
index 57a4236..9a5f4b8 100644
--- a/drivers/infiniband/hw/i40iw/i40iw_ctrl.c
+++ b/drivers/infiniband/hw/i40iw/i40iw_ctrl.c
@@ -4709,10 +4709,9 @@ static void i40iw_hw_stat_refresh_all(struct i40iw_dev_pestat *devstat)
* i40iw_device_init_pestat - Initialize the pestat structure
* @dev: pestat struct
*/
-enum i40iw_status_code i40iw_device_init_pestat(struct i40iw_dev_pestat *devstat)
+void i40iw_device_init_pestat(struct i40iw_dev_pestat *devstat)
{
devstat->ops = iw_device_pestat_ops;
- return 0;
}
/**
@@ -4736,12 +4735,7 @@ enum i40iw_status_code i40iw_device_init(struct i40iw_sc_dev *dev,
dev->debug_mask = info->debug_mask;
- ret_code = i40iw_device_init_pestat(&dev->dev_pestat);
- if (ret_code) {
- i40iw_debug(dev, I40IW_DEBUG_DEV,
- "%s: i40iw_device_init_pestat failed\n", __func__);
- return ret_code;
- }
+ i40iw_device_init_pestat(&dev->dev_pestat);
dev->hmc_fn_id = info->hmc_fn_id;
dev->qs_handle = info->qs_handle;
dev->exception_lan_queue = info->exception_lan_queue;
diff --git a/drivers/infiniband/hw/i40iw/i40iw_p.h b/drivers/infiniband/hw/i40iw/i40iw_p.h
index a0b8ca1..8a59c74 100644
--- a/drivers/infiniband/hw/i40iw/i40iw_p.h
+++ b/drivers/infiniband/hw/i40iw/i40iw_p.h
@@ -47,7 +47,7 @@ void i40iw_debug_buf(struct i40iw_sc_dev *dev, enum i40iw_debug_flag mask,
enum i40iw_status_code i40iw_device_init(struct i40iw_sc_dev *dev,
struct i40iw_device_init_info *info);
-enum i40iw_status_code i40iw_device_init_pestat(struct i40iw_dev_pestat *);
+void i40iw_device_init_pestat(struct i40iw_dev_pestat *devstat);
void i40iw_sc_cqp_post_sq(struct i40iw_sc_cqp *cqp);
diff --git a/drivers/infiniband/hw/i40iw/i40iw_virtchnl.c b/drivers/infiniband/hw/i40iw/i40iw_virtchnl.c
index 3041003..dbd39c4 100644
--- a/drivers/infiniband/hw/i40iw/i40iw_virtchnl.c
+++ b/drivers/infiniband/hw/i40iw/i40iw_virtchnl.c
@@ -496,11 +496,7 @@ enum i40iw_status_code i40iw_vchnl_recv_pf(struct i40iw_sc_dev *dev,
i40iw_debug(dev, I40IW_DEBUG_VIRT,
"VF%u error CQP HMC Function operation.\n",
vf_id);
- ret_code = i40iw_device_init_pestat(&vf_dev->dev_pestat);
- if (ret_code)
- i40iw_debug(dev, I40IW_DEBUG_VIRT,
- "VF%u - i40iw_device_init_pestat failed\n",
- vf_id);
+ i40iw_device_init_pestat(&vf_dev->dev_pestat);
vf_dev->dev_pestat.ops.iw_hw_stat_init(&vf_dev->dev_pestat,
(u8)vf_dev->pmf_index,
dev->hw, false);
--
1.8.3.1
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply related
* RE: [PATCH 1/2] i40iw: Remove variable flush_code and check to set qp->sq_flush
From: Orosco, Henry @ 2016-11-10 4:17 UTC (permalink / raw)
To: dledford-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org
Cc: linux-rdma-u79uwXL29TY76Z2rM5mHXA@public.gmane.org,
e1000-rdma-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org
In-Reply-To: <20161110034207.18016-1-henry.orosco-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
Please disregard this one. Accidently sent this one as part of a series.
-----Original Message-----
From: Orosco, Henry
Sent: Wednesday, November 09, 2016 9:42 PM
To: dledford-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org
Cc: linux-rdma-u79uwXL29TY76Z2rM5mHXA@public.gmane.org; e1000-rdma-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org; Orosco, Henry <henry.orosco-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
Subject: [PATCH 1/2] i40iw: Remove variable flush_code and check to set qp->sq_flush
The flush_code variable in i40iw_bld_terminate_hdr() is obsolete and the check to set qp->sq_flush is unreachable. Currently flush code is populated in setup_term_hdr() and both SQ and RQ are flushed always as part of the tear down flow.
Signed-off-by: Shiraz Saleem <shiraz.saleem-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
Signed-off-by: Henry Orosco <henry.orosco-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
---
drivers/infiniband/hw/i40iw/i40iw_ctrl.c | 4 ----
1 file changed, 4 deletions(-)
diff --git a/drivers/infiniband/hw/i40iw/i40iw_ctrl.c b/drivers/infiniband/hw/i40iw/i40iw_ctrl.c
index cd71444..57a4236 100644
--- a/drivers/infiniband/hw/i40iw/i40iw_ctrl.c
+++ b/drivers/infiniband/hw/i40iw/i40iw_ctrl.c
@@ -4045,7 +4045,6 @@ static int i40iw_bld_terminate_hdr(struct i40iw_sc_qp *qp,
u16 ddp_seg_len;
int copy_len = 0;
u8 is_tagged = 0;
- enum i40iw_flush_opcode flush_code = FLUSH_INVALID;
u32 opcode;
struct i40iw_terminate_hdr *termhdr;
@@ -4218,9 +4217,6 @@ static int i40iw_bld_terminate_hdr(struct i40iw_sc_qp *qp,
if (copy_len)
memcpy(termhdr + 1, pkt, copy_len);
- if (flush_code && !info->in_rdrsp_wr)
- qp->sq_flush = (info->sq) ? true : false;
-
return sizeof(struct i40iw_terminate_hdr) + copy_len; }
--
1.8.3.1
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply related
page: next (older) | prev (newer) | latest
- recent:[subjects (threaded)|topics (new)|topics (active)]
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox