summaryrefslogtreecommitdiff
path: root/net
diff options
context:
space:
mode:
Diffstat (limited to 'net')
-rw-r--r--net/ceph/messenger.c40
-rw-r--r--net/ceph/messenger_v1.c6
-rw-r--r--net/ceph/messenger_v2.c2
-rw-r--r--net/sunrpc/clnt.c61
-rw-r--r--net/sunrpc/sched.c51
-rw-r--r--net/sunrpc/xprt.c9
-rw-r--r--net/sunrpc/xprtmultipath.c4
-rw-r--r--net/sunrpc/xprtrdma/backchannel.c2
-rw-r--r--net/sunrpc/xprtrdma/frwr_ops.c20
-rw-r--r--net/sunrpc/xprtrdma/svc_rdma_backchannel.c4
-rw-r--r--net/sunrpc/xprtrdma/transport.c3
-rw-r--r--net/sunrpc/xprtrdma/verbs.c52
-rw-r--r--net/sunrpc/xprtrdma/xprt_rdma.h10
-rw-r--r--net/sunrpc/xprtsock.c5
14 files changed, 150 insertions, 119 deletions
diff --git a/net/ceph/messenger.c b/net/ceph/messenger.c
index d3bb656308b4..dfa237fbd5a3 100644
--- a/net/ceph/messenger.c
+++ b/net/ceph/messenger.c
@@ -728,7 +728,6 @@ static void ceph_msg_data_bio_cursor_init(struct ceph_msg_data_cursor *cursor,
it->iter.bi_size = cursor->resid;
BUG_ON(cursor->resid < bio_iter_len(it->bio, it->iter));
- cursor->last_piece = cursor->resid == bio_iter_len(it->bio, it->iter);
}
static struct page *ceph_msg_data_bio_next(struct ceph_msg_data_cursor *cursor,
@@ -754,10 +753,8 @@ static bool ceph_msg_data_bio_advance(struct ceph_msg_data_cursor *cursor,
cursor->resid -= bytes;
bio_advance_iter(it->bio, &it->iter, bytes);
- if (!cursor->resid) {
- BUG_ON(!cursor->last_piece);
+ if (!cursor->resid)
return false; /* no more data */
- }
if (!bytes || (it->iter.bi_size && it->iter.bi_bvec_done &&
page == bio_iter_page(it->bio, it->iter)))
@@ -770,9 +767,7 @@ static bool ceph_msg_data_bio_advance(struct ceph_msg_data_cursor *cursor,
it->iter.bi_size = cursor->resid;
}
- BUG_ON(cursor->last_piece);
BUG_ON(cursor->resid < bio_iter_len(it->bio, it->iter));
- cursor->last_piece = cursor->resid == bio_iter_len(it->bio, it->iter);
return true;
}
#endif /* CONFIG_BLOCK */
@@ -788,8 +783,6 @@ static void ceph_msg_data_bvecs_cursor_init(struct ceph_msg_data_cursor *cursor,
cursor->bvec_iter.bi_size = cursor->resid;
BUG_ON(cursor->resid < bvec_iter_len(bvecs, cursor->bvec_iter));
- cursor->last_piece =
- cursor->resid == bvec_iter_len(bvecs, cursor->bvec_iter);
}
static struct page *ceph_msg_data_bvecs_next(struct ceph_msg_data_cursor *cursor,
@@ -815,19 +808,14 @@ static bool ceph_msg_data_bvecs_advance(struct ceph_msg_data_cursor *cursor,
cursor->resid -= bytes;
bvec_iter_advance(bvecs, &cursor->bvec_iter, bytes);
- if (!cursor->resid) {
- BUG_ON(!cursor->last_piece);
+ if (!cursor->resid)
return false; /* no more data */
- }
if (!bytes || (cursor->bvec_iter.bi_bvec_done &&
page == bvec_iter_page(bvecs, cursor->bvec_iter)))
return false; /* more bytes to process in this segment */
- BUG_ON(cursor->last_piece);
BUG_ON(cursor->resid < bvec_iter_len(bvecs, cursor->bvec_iter));
- cursor->last_piece =
- cursor->resid == bvec_iter_len(bvecs, cursor->bvec_iter);
return true;
}
@@ -853,7 +841,6 @@ static void ceph_msg_data_pages_cursor_init(struct ceph_msg_data_cursor *cursor,
BUG_ON(page_count > (int)USHRT_MAX);
cursor->page_count = (unsigned short)page_count;
BUG_ON(length > SIZE_MAX - cursor->page_offset);
- cursor->last_piece = cursor->page_offset + cursor->resid <= PAGE_SIZE;
}
static struct page *
@@ -868,11 +855,7 @@ ceph_msg_data_pages_next(struct ceph_msg_data_cursor *cursor,
BUG_ON(cursor->page_offset >= PAGE_SIZE);
*page_offset = cursor->page_offset;
- if (cursor->last_piece)
- *length = cursor->resid;
- else
- *length = PAGE_SIZE - *page_offset;
-
+ *length = min_t(size_t, cursor->resid, PAGE_SIZE - *page_offset);
return data->pages[cursor->page_index];
}
@@ -897,8 +880,6 @@ static bool ceph_msg_data_pages_advance(struct ceph_msg_data_cursor *cursor,
BUG_ON(cursor->page_index >= cursor->page_count);
cursor->page_index++;
- cursor->last_piece = cursor->resid <= PAGE_SIZE;
-
return true;
}
@@ -928,7 +909,6 @@ ceph_msg_data_pagelist_cursor_init(struct ceph_msg_data_cursor *cursor,
cursor->resid = min(length, pagelist->length);
cursor->page = page;
cursor->offset = 0;
- cursor->last_piece = cursor->resid <= PAGE_SIZE;
}
static struct page *
@@ -948,11 +928,7 @@ ceph_msg_data_pagelist_next(struct ceph_msg_data_cursor *cursor,
/* offset of first page in pagelist is always 0 */
*page_offset = cursor->offset & ~PAGE_MASK;
- if (cursor->last_piece)
- *length = cursor->resid;
- else
- *length = PAGE_SIZE - *page_offset;
-
+ *length = min_t(size_t, cursor->resid, PAGE_SIZE - *page_offset);
return cursor->page;
}
@@ -985,8 +961,6 @@ static bool ceph_msg_data_pagelist_advance(struct ceph_msg_data_cursor *cursor,
BUG_ON(list_is_last(&cursor->page->lru, &pagelist->head));
cursor->page = list_next_entry(cursor->page, lru);
- cursor->last_piece = cursor->resid <= PAGE_SIZE;
-
return true;
}
@@ -1044,8 +1018,7 @@ void ceph_msg_data_cursor_init(struct ceph_msg_data_cursor *cursor,
* Indicate whether this is the last piece in this data item.
*/
struct page *ceph_msg_data_next(struct ceph_msg_data_cursor *cursor,
- size_t *page_offset, size_t *length,
- bool *last_piece)
+ size_t *page_offset, size_t *length)
{
struct page *page;
@@ -1074,8 +1047,6 @@ struct page *ceph_msg_data_next(struct ceph_msg_data_cursor *cursor,
BUG_ON(*page_offset + *length > PAGE_SIZE);
BUG_ON(!*length);
BUG_ON(*length > cursor->resid);
- if (last_piece)
- *last_piece = cursor->last_piece;
return page;
}
@@ -1112,7 +1083,6 @@ void ceph_msg_data_advance(struct ceph_msg_data_cursor *cursor, size_t bytes)
cursor->total_resid -= bytes;
if (!cursor->resid && cursor->total_resid) {
- WARN_ON(!cursor->last_piece);
cursor->data++;
__ceph_msg_data_cursor_init(cursor);
new_piece = true;
diff --git a/net/ceph/messenger_v1.c b/net/ceph/messenger_v1.c
index 6b014eca3a13..3ddbde87e4d6 100644
--- a/net/ceph/messenger_v1.c
+++ b/net/ceph/messenger_v1.c
@@ -495,7 +495,7 @@ static int write_partial_message_data(struct ceph_connection *con)
continue;
}
- page = ceph_msg_data_next(cursor, &page_offset, &length, NULL);
+ page = ceph_msg_data_next(cursor, &page_offset, &length);
if (length == cursor->total_resid)
more = MSG_MORE;
ret = ceph_tcp_sendpage(con->sock, page, page_offset, length,
@@ -1008,7 +1008,7 @@ static int read_partial_msg_data(struct ceph_connection *con)
continue;
}
- page = ceph_msg_data_next(cursor, &page_offset, &length, NULL);
+ page = ceph_msg_data_next(cursor, &page_offset, &length);
ret = ceph_tcp_recvpage(con->sock, page, page_offset, length);
if (ret <= 0) {
if (do_datacrc)
@@ -1050,7 +1050,7 @@ static int read_partial_msg_data_bounce(struct ceph_connection *con)
continue;
}
- page = ceph_msg_data_next(cursor, &off, &len, NULL);
+ page = ceph_msg_data_next(cursor, &off, &len);
ret = ceph_tcp_recvpage(con->sock, con->bounce_page, 0, len);
if (ret <= 0) {
con->in_data_crc = crc;
diff --git a/net/ceph/messenger_v2.c b/net/ceph/messenger_v2.c
index c6e5bfc717d5..cc8ff81a50b7 100644
--- a/net/ceph/messenger_v2.c
+++ b/net/ceph/messenger_v2.c
@@ -862,7 +862,7 @@ static void get_bvec_at(struct ceph_msg_data_cursor *cursor,
ceph_msg_data_advance(cursor, 0);
/* get a piece of data, cursor isn't advanced */
- page = ceph_msg_data_next(cursor, &off, &len, NULL);
+ page = ceph_msg_data_next(cursor, &off, &len);
bv->bv_page = page;
bv->bv_offset = off;
diff --git a/net/sunrpc/clnt.c b/net/sunrpc/clnt.c
index c284efa3d1ef..993acf38af87 100644
--- a/net/sunrpc/clnt.c
+++ b/net/sunrpc/clnt.c
@@ -345,7 +345,7 @@ static int rpc_alloc_clid(struct rpc_clnt *clnt)
{
int clid;
- clid = ida_simple_get(&rpc_clids, 0, 0, GFP_KERNEL);
+ clid = ida_alloc(&rpc_clids, GFP_KERNEL);
if (clid < 0)
return clid;
clnt->cl_clid = clid;
@@ -354,7 +354,7 @@ static int rpc_alloc_clid(struct rpc_clnt *clnt)
static void rpc_free_clid(struct rpc_clnt *clnt)
{
- ida_simple_remove(&rpc_clids, clnt->cl_clid);
+ ida_free(&rpc_clids, clnt->cl_clid);
}
static struct rpc_clnt * rpc_new_client(const struct rpc_create_args *args,
@@ -873,6 +873,57 @@ void rpc_killall_tasks(struct rpc_clnt *clnt)
}
EXPORT_SYMBOL_GPL(rpc_killall_tasks);
+/**
+ * rpc_cancel_tasks - try to cancel a set of RPC tasks
+ * @clnt: Pointer to RPC client
+ * @error: RPC task error value to set
+ * @fnmatch: Pointer to selector function
+ * @data: User data
+ *
+ * Uses @fnmatch to define a set of RPC tasks that are to be cancelled.
+ * The argument @error must be a negative error value.
+ */
+unsigned long rpc_cancel_tasks(struct rpc_clnt *clnt, int error,
+ bool (*fnmatch)(const struct rpc_task *,
+ const void *),
+ const void *data)
+{
+ struct rpc_task *task;
+ unsigned long count = 0;
+
+ if (list_empty(&clnt->cl_tasks))
+ return 0;
+ /*
+ * Spin lock all_tasks to prevent changes...
+ */
+ spin_lock(&clnt->cl_lock);
+ list_for_each_entry(task, &clnt->cl_tasks, tk_task) {
+ if (!RPC_IS_ACTIVATED(task))
+ continue;
+ if (!fnmatch(task, data))
+ continue;
+ rpc_task_try_cancel(task, error);
+ count++;
+ }
+ spin_unlock(&clnt->cl_lock);
+ return count;
+}
+EXPORT_SYMBOL_GPL(rpc_cancel_tasks);
+
+static int rpc_clnt_disconnect_xprt(struct rpc_clnt *clnt,
+ struct rpc_xprt *xprt, void *dummy)
+{
+ if (xprt_connected(xprt))
+ xprt_force_disconnect(xprt);
+ return 0;
+}
+
+void rpc_clnt_disconnect(struct rpc_clnt *clnt)
+{
+ rpc_clnt_iterate_for_each_xprt(clnt, rpc_clnt_disconnect_xprt, NULL);
+}
+EXPORT_SYMBOL_GPL(rpc_clnt_disconnect);
+
/*
* Properly shut down an RPC client, terminating all outstanding
* requests.
@@ -1642,7 +1693,7 @@ static void
__rpc_call_rpcerror(struct rpc_task *task, int tk_status, int rpc_status)
{
trace_rpc_call_rpcerror(task, tk_status, rpc_status);
- task->tk_rpc_status = rpc_status;
+ rpc_task_set_rpc_status(task, rpc_status);
rpc_exit(task, tk_status);
}
@@ -2435,10 +2486,8 @@ rpc_check_timeout(struct rpc_task *task)
{
struct rpc_clnt *clnt = task->tk_client;
- if (RPC_SIGNALLED(task)) {
- rpc_call_rpcerror(task, -ERESTARTSYS);
+ if (RPC_SIGNALLED(task))
return;
- }
if (xprt_adjust_timeout(task->tk_rqstp) == 0)
return;
diff --git a/net/sunrpc/sched.c b/net/sunrpc/sched.c
index 46cbf151a50b..be587a308e05 100644
--- a/net/sunrpc/sched.c
+++ b/net/sunrpc/sched.c
@@ -65,6 +65,13 @@ gfp_t rpc_task_gfp_mask(void)
}
EXPORT_SYMBOL_GPL(rpc_task_gfp_mask);
+bool rpc_task_set_rpc_status(struct rpc_task *task, int rpc_status)
+{
+ if (cmpxchg(&task->tk_rpc_status, 0, rpc_status) == 0)
+ return true;
+ return false;
+}
+
unsigned long
rpc_task_timeout(const struct rpc_task *task)
{
@@ -853,12 +860,25 @@ void rpc_signal_task(struct rpc_task *task)
if (!RPC_IS_ACTIVATED(task))
return;
+ if (!rpc_task_set_rpc_status(task, -ERESTARTSYS))
+ return;
trace_rpc_task_signalled(task, task->tk_action);
set_bit(RPC_TASK_SIGNALLED, &task->tk_runstate);
smp_mb__after_atomic();
queue = READ_ONCE(task->tk_waitqueue);
if (queue)
- rpc_wake_up_queued_task_set_status(queue, task, -ERESTARTSYS);
+ rpc_wake_up_queued_task(queue, task);
+}
+
+void rpc_task_try_cancel(struct rpc_task *task, int error)
+{
+ struct rpc_wait_queue *queue;
+
+ if (!rpc_task_set_rpc_status(task, error))
+ return;
+ queue = READ_ONCE(task->tk_waitqueue);
+ if (queue)
+ rpc_wake_up_queued_task(queue, task);
}
void rpc_exit(struct rpc_task *task, int status)
@@ -905,10 +925,16 @@ static void __rpc_execute(struct rpc_task *task)
* Perform the next FSM step or a pending callback.
*
* tk_action may be NULL if the task has been killed.
- * In particular, note that rpc_killall_tasks may
- * do this at any time, so beware when dereferencing.
*/
do_action = task->tk_action;
+ /* Tasks with an RPC error status should exit */
+ if (do_action != rpc_exit_task &&
+ (status = READ_ONCE(task->tk_rpc_status)) != 0) {
+ task->tk_status = status;
+ if (do_action != NULL)
+ do_action = rpc_exit_task;
+ }
+ /* Callbacks override all actions */
if (task->tk_callback) {
do_action = task->tk_callback;
task->tk_callback = NULL;
@@ -931,14 +957,6 @@ static void __rpc_execute(struct rpc_task *task)
}
/*
- * Signalled tasks should exit rather than sleep.
- */
- if (RPC_SIGNALLED(task)) {
- task->tk_rpc_status = -ERESTARTSYS;
- rpc_exit(task, -ERESTARTSYS);
- }
-
- /*
* The queue->lock protects against races with
* rpc_make_runnable().
*
@@ -953,6 +971,12 @@ static void __rpc_execute(struct rpc_task *task)
spin_unlock(&queue->lock);
continue;
}
+ /* Wake up any task that has an exit status */
+ if (READ_ONCE(task->tk_rpc_status) != 0) {
+ rpc_wake_up_task_queue_locked(queue, task);
+ spin_unlock(&queue->lock);
+ continue;
+ }
rpc_clear_running(task);
spin_unlock(&queue->lock);
if (task_is_async)
@@ -970,10 +994,7 @@ static void __rpc_execute(struct rpc_task *task)
* clean up after sleeping on some queue, we don't
* break the loop here, but go around once more.
*/
- trace_rpc_task_signalled(task, task->tk_action);
- set_bit(RPC_TASK_SIGNALLED, &task->tk_runstate);
- task->tk_rpc_status = -ERESTARTSYS;
- rpc_exit(task, -ERESTARTSYS);
+ rpc_signal_task(task);
}
trace_rpc_task_sync_wake(task, task->tk_action);
}
diff --git a/net/sunrpc/xprt.c b/net/sunrpc/xprt.c
index f8fae7815649..71dc26373444 100644
--- a/net/sunrpc/xprt.c
+++ b/net/sunrpc/xprt.c
@@ -1788,7 +1788,7 @@ static int xprt_alloc_id(struct rpc_xprt *xprt)
{
int id;
- id = ida_simple_get(&rpc_xprt_ids, 0, 0, GFP_KERNEL);
+ id = ida_alloc(&rpc_xprt_ids, GFP_KERNEL);
if (id < 0)
return id;
@@ -1798,7 +1798,7 @@ static int xprt_alloc_id(struct rpc_xprt *xprt)
static void xprt_free_id(struct rpc_xprt *xprt)
{
- ida_simple_remove(&rpc_xprt_ids, xprt->id);
+ ida_free(&rpc_xprt_ids, xprt->id);
}
struct rpc_xprt *xprt_alloc(struct net *net, size_t size,
@@ -1822,10 +1822,7 @@ struct rpc_xprt *xprt_alloc(struct net *net, size_t size,
goto out_free;
list_add(&req->rq_list, &xprt->free);
}
- if (max_alloc > num_prealloc)
- xprt->max_reqs = max_alloc;
- else
- xprt->max_reqs = num_prealloc;
+ xprt->max_reqs = max_t(unsigned int, max_alloc, num_prealloc);
xprt->min_reqs = num_prealloc;
xprt->num_reqs = num_prealloc;
diff --git a/net/sunrpc/xprtmultipath.c b/net/sunrpc/xprtmultipath.c
index 685db598acbe..701250b305db 100644
--- a/net/sunrpc/xprtmultipath.c
+++ b/net/sunrpc/xprtmultipath.c
@@ -103,7 +103,7 @@ static int xprt_switch_alloc_id(struct rpc_xprt_switch *xps, gfp_t gfp_flags)
{
int id;
- id = ida_simple_get(&rpc_xprtswitch_ids, 0, 0, gfp_flags);
+ id = ida_alloc(&rpc_xprtswitch_ids, gfp_flags);
if (id < 0)
return id;
@@ -113,7 +113,7 @@ static int xprt_switch_alloc_id(struct rpc_xprt_switch *xps, gfp_t gfp_flags)
static void xprt_switch_free_id(struct rpc_xprt_switch *xps)
{
- ida_simple_remove(&rpc_xprtswitch_ids, xps->xps_id);
+ ida_free(&rpc_xprtswitch_ids, xps->xps_id);
}
/**
diff --git a/net/sunrpc/xprtrdma/backchannel.c b/net/sunrpc/xprtrdma/backchannel.c
index faba7136dd9a..e4d84a13c566 100644
--- a/net/sunrpc/xprtrdma/backchannel.c
+++ b/net/sunrpc/xprtrdma/backchannel.c
@@ -189,7 +189,7 @@ create_req:
return NULL;
size = min_t(size_t, r_xprt->rx_ep->re_inline_recv, PAGE_SIZE);
- req = rpcrdma_req_create(r_xprt, size, GFP_KERNEL);
+ req = rpcrdma_req_create(r_xprt, size);
if (!req)
return NULL;
if (rpcrdma_req_setup(r_xprt, req)) {
diff --git a/net/sunrpc/xprtrdma/frwr_ops.c b/net/sunrpc/xprtrdma/frwr_ops.c
index de0bdb6b729f..ffbf99894970 100644
--- a/net/sunrpc/xprtrdma/frwr_ops.c
+++ b/net/sunrpc/xprtrdma/frwr_ops.c
@@ -124,16 +124,16 @@ int frwr_mr_init(struct rpcrdma_xprt *r_xprt, struct rpcrdma_mr *mr)
unsigned int depth = ep->re_max_fr_depth;
struct scatterlist *sg;
struct ib_mr *frmr;
- int rc;
+
+ sg = kcalloc_node(depth, sizeof(*sg), XPRTRDMA_GFP_FLAGS,
+ ibdev_to_node(ep->re_id->device));
+ if (!sg)
+ return -ENOMEM;
frmr = ib_alloc_mr(ep->re_pd, ep->re_mrtype, depth);
if (IS_ERR(frmr))
goto out_mr_err;
- sg = kmalloc_array(depth, sizeof(*sg), GFP_KERNEL);
- if (!sg)
- goto out_list_err;
-
mr->mr_xprt = r_xprt;
mr->mr_ibmr = frmr;
mr->mr_device = NULL;
@@ -146,13 +146,9 @@ int frwr_mr_init(struct rpcrdma_xprt *r_xprt, struct rpcrdma_mr *mr)
return 0;
out_mr_err:
- rc = PTR_ERR(frmr);
- trace_xprtrdma_frwr_alloc(mr, rc);
- return rc;
-
-out_list_err:
- ib_dereg_mr(frmr);
- return -ENOMEM;
+ kfree(sg);
+ trace_xprtrdma_frwr_alloc(mr, PTR_ERR(frmr));
+ return PTR_ERR(frmr);
}
/**
diff --git a/net/sunrpc/xprtrdma/svc_rdma_backchannel.c b/net/sunrpc/xprtrdma/svc_rdma_backchannel.c
index 85c8cdda98b1..aa2227a7e552 100644
--- a/net/sunrpc/xprtrdma/svc_rdma_backchannel.c
+++ b/net/sunrpc/xprtrdma/svc_rdma_backchannel.c
@@ -119,12 +119,12 @@ xprt_rdma_bc_allocate(struct rpc_task *task)
return -EINVAL;
}
- page = alloc_page(RPCRDMA_DEF_GFP);
+ page = alloc_page(GFP_NOIO | __GFP_NOWARN);
if (!page)
return -ENOMEM;
rqst->rq_buffer = page_address(page);
- rqst->rq_rbuffer = kmalloc(rqst->rq_rcvsize, RPCRDMA_DEF_GFP);
+ rqst->rq_rbuffer = kmalloc(rqst->rq_rcvsize, GFP_NOIO | __GFP_NOWARN);
if (!rqst->rq_rbuffer) {
put_page(page);
return -ENOMEM;
diff --git a/net/sunrpc/xprtrdma/transport.c b/net/sunrpc/xprtrdma/transport.c
index bcb37b51adf6..10bb2b929c6d 100644
--- a/net/sunrpc/xprtrdma/transport.c
+++ b/net/sunrpc/xprtrdma/transport.c
@@ -494,8 +494,7 @@ xprt_rdma_connect(struct rpc_xprt *xprt, struct rpc_task *task)
xprt_reconnect_backoff(xprt, RPCRDMA_INIT_REEST_TO);
}
trace_xprtrdma_op_connect(r_xprt, delay);
- queue_delayed_work(xprtiod_workqueue, &r_xprt->rx_connect_worker,
- delay);
+ queue_delayed_work(system_long_wq, &r_xprt->rx_connect_worker, delay);
}
/**
diff --git a/net/sunrpc/xprtrdma/verbs.c b/net/sunrpc/xprtrdma/verbs.c
index 2fbe9aaeec34..44b87e4274b4 100644
--- a/net/sunrpc/xprtrdma/verbs.c
+++ b/net/sunrpc/xprtrdma/verbs.c
@@ -76,8 +76,7 @@ static void rpcrdma_mrs_destroy(struct rpcrdma_xprt *r_xprt);
static void rpcrdma_ep_get(struct rpcrdma_ep *ep);
static int rpcrdma_ep_put(struct rpcrdma_ep *ep);
static struct rpcrdma_regbuf *
-rpcrdma_regbuf_alloc(size_t size, enum dma_data_direction direction,
- gfp_t flags);
+rpcrdma_regbuf_alloc(size_t size, enum dma_data_direction direction);
static void rpcrdma_regbuf_dma_unmap(struct rpcrdma_regbuf *rb);
static void rpcrdma_regbuf_free(struct rpcrdma_regbuf *rb);
@@ -373,7 +372,7 @@ static int rpcrdma_ep_create(struct rpcrdma_xprt *r_xprt)
struct rpcrdma_ep *ep;
int rc;
- ep = kzalloc(sizeof(*ep), GFP_KERNEL);
+ ep = kzalloc(sizeof(*ep), XPRTRDMA_GFP_FLAGS);
if (!ep)
return -ENOTCONN;
ep->re_xprt = &r_xprt->rx_xprt;
@@ -606,7 +605,7 @@ static struct rpcrdma_sendctx *rpcrdma_sendctx_create(struct rpcrdma_ep *ep)
struct rpcrdma_sendctx *sc;
sc = kzalloc(struct_size(sc, sc_sges, ep->re_attr.cap.max_send_sge),
- GFP_KERNEL);
+ XPRTRDMA_GFP_FLAGS);
if (!sc)
return NULL;
@@ -629,7 +628,7 @@ static int rpcrdma_sendctxs_create(struct rpcrdma_xprt *r_xprt)
* Sends are posted.
*/
i = r_xprt->rx_ep->re_max_requests + RPCRDMA_MAX_BC_REQUESTS;
- buf->rb_sc_ctxs = kcalloc(i, sizeof(sc), GFP_KERNEL);
+ buf->rb_sc_ctxs = kcalloc(i, sizeof(sc), XPRTRDMA_GFP_FLAGS);
if (!buf->rb_sc_ctxs)
return -ENOMEM;
@@ -740,13 +739,16 @@ rpcrdma_mrs_create(struct rpcrdma_xprt *r_xprt)
{
struct rpcrdma_buffer *buf = &r_xprt->rx_buf;
struct rpcrdma_ep *ep = r_xprt->rx_ep;
+ struct ib_device *device = ep->re_id->device;
unsigned int count;
+ /* Try to allocate enough to perform one full-sized I/O */
for (count = 0; count < ep->re_max_rdma_segs; count++) {
struct rpcrdma_mr *mr;
int rc;
- mr = kzalloc(sizeof(*mr), GFP_KERNEL);
+ mr = kzalloc_node(sizeof(*mr), XPRTRDMA_GFP_FLAGS,
+ ibdev_to_node(device));
if (!mr)
break;
@@ -791,38 +793,33 @@ void rpcrdma_mrs_refresh(struct rpcrdma_xprt *r_xprt)
/* If there is no underlying connection, it's no use
* to wake the refresh worker.
*/
- if (ep->re_connect_status == 1) {
- /* The work is scheduled on a WQ_MEM_RECLAIM
- * workqueue in order to prevent MR allocation
- * from recursing into NFS during direct reclaim.
- */
- queue_work(xprtiod_workqueue, &buf->rb_refresh_worker);
- }
+ if (ep->re_connect_status != 1)
+ return;
+ queue_work(system_highpri_wq, &buf->rb_refresh_worker);
}
/**
* rpcrdma_req_create - Allocate an rpcrdma_req object
* @r_xprt: controlling r_xprt
* @size: initial size, in bytes, of send and receive buffers
- * @flags: GFP flags passed to memory allocators
*
* Returns an allocated and fully initialized rpcrdma_req or NULL.
*/
-struct rpcrdma_req *rpcrdma_req_create(struct rpcrdma_xprt *r_xprt, size_t size,
- gfp_t flags)
+struct rpcrdma_req *rpcrdma_req_create(struct rpcrdma_xprt *r_xprt,
+ size_t size)
{
struct rpcrdma_buffer *buffer = &r_xprt->rx_buf;
struct rpcrdma_req *req;
- req = kzalloc(sizeof(*req), flags);
+ req = kzalloc(sizeof(*req), XPRTRDMA_GFP_FLAGS);
if (req == NULL)
goto out1;
- req->rl_sendbuf = rpcrdma_regbuf_alloc(size, DMA_TO_DEVICE, flags);
+ req->rl_sendbuf = rpcrdma_regbuf_alloc(size, DMA_TO_DEVICE);
if (!req->rl_sendbuf)
goto out2;
- req->rl_recvbuf = rpcrdma_regbuf_alloc(size, DMA_NONE, flags);
+ req->rl_recvbuf = rpcrdma_regbuf_alloc(size, DMA_NONE);
if (!req->rl_recvbuf)
goto out3;
@@ -858,7 +855,7 @@ int rpcrdma_req_setup(struct rpcrdma_xprt *r_xprt, struct rpcrdma_req *req)
r_xprt->rx_ep->re_max_rdma_segs * rpcrdma_readchunk_maxsz;
maxhdrsize *= sizeof(__be32);
rb = rpcrdma_regbuf_alloc(__roundup_pow_of_two(maxhdrsize),
- DMA_TO_DEVICE, GFP_KERNEL);
+ DMA_TO_DEVICE);
if (!rb)
goto out;
@@ -929,12 +926,12 @@ struct rpcrdma_rep *rpcrdma_rep_create(struct rpcrdma_xprt *r_xprt,
struct rpcrdma_buffer *buf = &r_xprt->rx_buf;
struct rpcrdma_rep *rep;
- rep = kzalloc(sizeof(*rep), GFP_KERNEL);
+ rep = kzalloc(sizeof(*rep), XPRTRDMA_GFP_FLAGS);
if (rep == NULL)
goto out;
rep->rr_rdmabuf = rpcrdma_regbuf_alloc(r_xprt->rx_ep->re_inline_recv,
- DMA_FROM_DEVICE, GFP_KERNEL);
+ DMA_FROM_DEVICE);
if (!rep->rr_rdmabuf)
goto out_free;
@@ -1064,8 +1061,8 @@ int rpcrdma_buffer_create(struct rpcrdma_xprt *r_xprt)
for (i = 0; i < r_xprt->rx_xprt.max_reqs; i++) {
struct rpcrdma_req *req;
- req = rpcrdma_req_create(r_xprt, RPCRDMA_V1_DEF_INLINE_SIZE * 2,
- GFP_KERNEL);
+ req = rpcrdma_req_create(r_xprt,
+ RPCRDMA_V1_DEF_INLINE_SIZE * 2);
if (!req)
goto out;
list_add(&req->rl_list, &buf->rb_send_bufs);
@@ -1235,15 +1232,14 @@ void rpcrdma_buffer_put(struct rpcrdma_buffer *buffers, struct rpcrdma_req *req)
* or Replies they may be registered externally via frwr_map.
*/
static struct rpcrdma_regbuf *
-rpcrdma_regbuf_alloc(size_t size, enum dma_data_direction direction,
- gfp_t flags)
+rpcrdma_regbuf_alloc(size_t size, enum dma_data_direction direction)
{
struct rpcrdma_regbuf *rb;
- rb = kmalloc(sizeof(*rb), flags);
+ rb = kmalloc(sizeof(*rb), XPRTRDMA_GFP_FLAGS);
if (!rb)
return NULL;
- rb->rg_data = kmalloc(size, flags);
+ rb->rg_data = kmalloc(size, XPRTRDMA_GFP_FLAGS);
if (!rb->rg_data) {
kfree(rb);
return NULL;
diff --git a/net/sunrpc/xprtrdma/xprt_rdma.h b/net/sunrpc/xprtrdma/xprt_rdma.h
index c79f92eeda76..5e5ff6784ef5 100644
--- a/net/sunrpc/xprtrdma/xprt_rdma.h
+++ b/net/sunrpc/xprtrdma/xprt_rdma.h
@@ -149,7 +149,11 @@ static inline void *rdmab_data(const struct rpcrdma_regbuf *rb)
return rb->rg_data;
}
-#define RPCRDMA_DEF_GFP (GFP_NOIO | __GFP_NOWARN)
+/* Do not use emergency memory reserves, and fail quickly if memory
+ * cannot be allocated easily. These flags may be used wherever there
+ * is robust logic to handle a failure to allocate.
+ */
+#define XPRTRDMA_GFP_FLAGS (__GFP_NOMEMALLOC | __GFP_NORETRY | __GFP_NOWARN)
/* To ensure a transport can always make forward progress,
* the number of RDMA segments allowed in header chunk lists
@@ -467,8 +471,8 @@ void rpcrdma_post_recvs(struct rpcrdma_xprt *r_xprt, int needed, bool temp);
/*
* Buffer calls - xprtrdma/verbs.c
*/
-struct rpcrdma_req *rpcrdma_req_create(struct rpcrdma_xprt *r_xprt, size_t size,
- gfp_t flags);
+struct rpcrdma_req *rpcrdma_req_create(struct rpcrdma_xprt *r_xprt,
+ size_t size);
int rpcrdma_req_setup(struct rpcrdma_xprt *r_xprt, struct rpcrdma_req *req);
void rpcrdma_req_destroy(struct rpcrdma_req *req);
int rpcrdma_buffer_create(struct rpcrdma_xprt *);
diff --git a/net/sunrpc/xprtsock.c b/net/sunrpc/xprtsock.c
index e976007f4fd0..f34d5427b66c 100644
--- a/net/sunrpc/xprtsock.c
+++ b/net/sunrpc/xprtsock.c
@@ -261,7 +261,7 @@ static void xs_format_common_peer_addresses(struct rpc_xprt *xprt)
switch (sap->sa_family) {
case AF_LOCAL:
sun = xs_addr_un(xprt);
- strlcpy(buf, sun->sun_path, sizeof(buf));
+ strscpy(buf, sun->sun_path, sizeof(buf));
xprt->address_strings[RPC_DISPLAY_ADDR] =
kstrdup(buf, GFP_KERNEL);
break;
@@ -1978,8 +1978,7 @@ static void xs_local_connect(struct rpc_xprt *xprt, struct rpc_task *task)
* we'll need to figure out how to pass a namespace to
* connect.
*/
- task->tk_rpc_status = -ENOTCONN;
- rpc_exit(task, -ENOTCONN);
+ rpc_task_set_rpc_status(task, -ENOTCONN);
goto out_wake;
}
ret = xs_local_setup_socket(transport);