@@ -1687,12 +1687,15 @@ static bool xprt_throttle_congested(struct rpc_xprt *xprt, struct rpc_task *task
static struct rpc_rqst *xprt_dynamic_alloc_slot(struct rpc_xprt *xprt)
{
struct rpc_rqst *req = ERR_PTR(-EAGAIN);
+ gfp_t gfp_mask = GFP_NOFS;
if (xprt->num_reqs >= xprt->max_reqs)
goto out;
++xprt->num_reqs;
spin_unlock(&xprt->reserve_lock);
- req = kzalloc(sizeof(struct rpc_rqst), GFP_NOFS);
+ if (current->flags & PF_WQ_WORKER)
+ gfp_mask |= __GFP_NORETRY | __GFP_NOWARN;
+ req = kzalloc(sizeof(struct rpc_rqst), gfp_mask);
spin_lock(&xprt->reserve_lock);
if (req != NULL)
goto out;
@@ -521,7 +521,7 @@ xprt_rdma_alloc_slot(struct rpc_xprt *xprt, struct rpc_task *task)
return;
out_sleep:
- task->tk_status = -EAGAIN;
+ task->tk_status = -ENOMEM;
xprt_add_backlog(xprt, task);
}
When memory is short, new worker threads cannot be created and we depend on the minimum one rpciod thread to be able to handle everything. So it must not block waiting for memory. xprt_dynamic_alloc_slot can block indefinitely. This can tie up all workqueue threads and NFS can deadlock. So when called from a workqueue, set __GFP_NORETRY. The rdma alloc_slot already does not block. However it sets the error to -EAGAIN suggesting this will trigger a sleep. It does not. As we can see in call_reserveresult(), only -ENOMEM causes a sleep. -EAGAIN causes immediate retry. Signed-off-by: NeilBrown <neilb@suse.de> --- net/sunrpc/xprt.c | 5 ++++- net/sunrpc/xprtrdma/transport.c | 2 +- 2 files changed, 5 insertions(+), 2 deletions(-)