NVMe: nvme_submit_async_admin_req() must use atomic rq allocation

We are called for async event notification issues, and the
nvmeq lock is already held. If we fail the request allocation,
we'll just retry next time.

Reported-by: Julia Lawall <julia.lawall@lip6.fr>
Signed-off-by: Jens Axboe <axboe@fb.com>

diff --git a/drivers/block/nvme-core.c b/drivers/block/nvme-core.c
index bbac17f29fe74ef9b162da6fa2eac82503dd89e0..fb4b205317c68c369a8a8dd94f63e7314c6714b8 100644 (file)
--- a/drivers/block/nvme-core.c
+++ b/drivers/block/nvme-core.c
@@ -825,7 +825,7 @@ static int nvme_submit_async_admin_req(struct nvme_dev *dev)
        struct nvme_cmd_info *cmd_info;
        struct request *req;
 
-       req = blk_mq_alloc_request(dev->admin_q, WRITE, GFP_KERNEL, false);
+       req = blk_mq_alloc_request(dev->admin_q, WRITE, GFP_ATOMIC, false);
        if (IS_ERR(req))
                return PTR_ERR(req);