Commit 959401aa authored by Jens Axboe's avatar Jens Axboe

Merge branch 'nvmf-4.9-rc' of git://git.infradead.org/nvme-fabrics into for-linus

Sagi writes:

These are the relevant fixes for rc6
- fix possible crash in nvmet-rdma cm_handler from Bart
- fix possible memory leak in nvmet-rdma for connection failures
- fix possible use-after-free conditions in nvmet-rdma
- fix possible IO errors during reconnect stage from Christoph
- fix possible memory leak in nvme-rdma during IO queues connect
  failures from Steve
parents e76d21c4 14c862db
......@@ -83,6 +83,7 @@ enum nvme_rdma_queue_flags {
NVME_RDMA_Q_CONNECTED = (1 << 0),
NVME_RDMA_IB_QUEUE_ALLOCATED = (1 << 1),
NVME_RDMA_Q_DELETING = (1 << 2),
NVME_RDMA_Q_LIVE = (1 << 3),
};
struct nvme_rdma_queue {
......@@ -624,10 +625,18 @@ static int nvme_rdma_connect_io_queues(struct nvme_rdma_ctrl *ctrl)
for (i = 1; i < ctrl->queue_count; i++) {
ret = nvmf_connect_io_queue(&ctrl->ctrl, i);
if (ret)
break;
if (ret) {
dev_info(ctrl->ctrl.device,
"failed to connect i/o queue: %d\n", ret);
goto out_free_queues;
}
set_bit(NVME_RDMA_Q_LIVE, &ctrl->queues[i].flags);
}
return 0;
out_free_queues:
nvme_rdma_free_io_queues(ctrl);
return ret;
}
......@@ -712,6 +721,8 @@ static void nvme_rdma_reconnect_ctrl_work(struct work_struct *work)
if (ret)
goto stop_admin_q;
set_bit(NVME_RDMA_Q_LIVE, &ctrl->queues[0].flags);
ret = nvme_enable_ctrl(&ctrl->ctrl, ctrl->cap);
if (ret)
goto stop_admin_q;
......@@ -761,8 +772,10 @@ static void nvme_rdma_error_recovery_work(struct work_struct *work)
nvme_stop_keep_alive(&ctrl->ctrl);
for (i = 0; i < ctrl->queue_count; i++)
for (i = 0; i < ctrl->queue_count; i++) {
clear_bit(NVME_RDMA_Q_CONNECTED, &ctrl->queues[i].flags);
clear_bit(NVME_RDMA_Q_LIVE, &ctrl->queues[i].flags);
}
if (ctrl->queue_count > 1)
nvme_stop_queues(&ctrl->ctrl);
......@@ -1378,6 +1391,24 @@ nvme_rdma_timeout(struct request *rq, bool reserved)
return BLK_EH_HANDLED;
}
/*
* We cannot accept any other command until the Connect command has completed.
*/
static inline bool nvme_rdma_queue_is_ready(struct nvme_rdma_queue *queue,
struct request *rq)
{
if (unlikely(!test_bit(NVME_RDMA_Q_LIVE, &queue->flags))) {
struct nvme_command *cmd = (struct nvme_command *)rq->cmd;
if (rq->cmd_type != REQ_TYPE_DRV_PRIV ||
cmd->common.opcode != nvme_fabrics_command ||
cmd->fabrics.fctype != nvme_fabrics_type_connect)
return false;
}
return true;
}
static int nvme_rdma_queue_rq(struct blk_mq_hw_ctx *hctx,
const struct blk_mq_queue_data *bd)
{
......@@ -1394,6 +1425,9 @@ static int nvme_rdma_queue_rq(struct blk_mq_hw_ctx *hctx,
WARN_ON_ONCE(rq->tag < 0);
if (!nvme_rdma_queue_is_ready(queue, rq))
return BLK_MQ_RQ_QUEUE_BUSY;
dev = queue->device->dev;
ib_dma_sync_single_for_cpu(dev, sqe->dma,
sizeof(struct nvme_command), DMA_TO_DEVICE);
......@@ -1544,6 +1578,8 @@ static int nvme_rdma_configure_admin_queue(struct nvme_rdma_ctrl *ctrl)
if (error)
goto out_cleanup_queue;
set_bit(NVME_RDMA_Q_LIVE, &ctrl->queues[0].flags);
error = nvmf_reg_read64(&ctrl->ctrl, NVME_REG_CAP, &ctrl->cap);
if (error) {
dev_err(ctrl->ctrl.device,
......
......@@ -838,9 +838,13 @@ static void nvmet_fatal_error_handler(struct work_struct *work)
void nvmet_ctrl_fatal_error(struct nvmet_ctrl *ctrl)
{
mutex_lock(&ctrl->lock);
if (!(ctrl->csts & NVME_CSTS_CFS)) {
ctrl->csts |= NVME_CSTS_CFS;
INIT_WORK(&ctrl->fatal_err_work, nvmet_fatal_error_handler);
schedule_work(&ctrl->fatal_err_work);
}
mutex_unlock(&ctrl->lock);
}
EXPORT_SYMBOL_GPL(nvmet_ctrl_fatal_error);
......
......@@ -951,6 +951,7 @@ static int nvmet_rdma_create_queue_ib(struct nvmet_rdma_queue *queue)
static void nvmet_rdma_destroy_queue_ib(struct nvmet_rdma_queue *queue)
{
ib_drain_qp(queue->cm_id->qp);
rdma_destroy_qp(queue->cm_id);
ib_free_cq(queue->cq);
}
......@@ -1066,6 +1067,7 @@ nvmet_rdma_alloc_queue(struct nvmet_rdma_device *ndev,
spin_lock_init(&queue->rsp_wr_wait_lock);
INIT_LIST_HEAD(&queue->free_rsps);
spin_lock_init(&queue->rsps_lock);
INIT_LIST_HEAD(&queue->queue_list);
queue->idx = ida_simple_get(&nvmet_rdma_queue_ida, 0, 0, GFP_KERNEL);
if (queue->idx < 0) {
......@@ -1244,7 +1246,6 @@ static void __nvmet_rdma_queue_disconnect(struct nvmet_rdma_queue *queue)
if (disconnect) {
rdma_disconnect(queue->cm_id);
ib_drain_qp(queue->cm_id->qp);
schedule_work(&queue->release_work);
}
}
......@@ -1269,7 +1270,12 @@ static void nvmet_rdma_queue_connect_fail(struct rdma_cm_id *cm_id,
{
WARN_ON_ONCE(queue->state != NVMET_RDMA_Q_CONNECTING);
pr_err("failed to connect queue\n");
mutex_lock(&nvmet_rdma_queue_mutex);
if (!list_empty(&queue->queue_list))
list_del_init(&queue->queue_list);
mutex_unlock(&nvmet_rdma_queue_mutex);
pr_err("failed to connect queue %d\n", queue->idx);
schedule_work(&queue->release_work);
}
......@@ -1352,6 +1358,12 @@ static int nvmet_rdma_cm_handler(struct rdma_cm_id *cm_id,
case RDMA_CM_EVENT_ADDR_CHANGE:
case RDMA_CM_EVENT_DISCONNECTED:
case RDMA_CM_EVENT_TIMEWAIT_EXIT:
/*
* We might end up here when we already freed the qp
* which means queue release sequence is in progress,
* so don't get in the way...
*/
if (queue)
nvmet_rdma_queue_disconnect(queue);
break;
case RDMA_CM_EVENT_DEVICE_REMOVAL:
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment