Merge tag 'block-6.14-20250313' of git://git.kernel.dk/linux

Pull block fixes from Jens Axboe:

 - NVMe pull request via Keith:
     - Concurrent pci error and hotplug handling fix (Keith)
     - Endpoint function fixes (Damien)

 - Fix for a regression introduced in this cycle with error checking for
   batched request completions (Shin'ichiro)

* tag 'block-6.14-20250313' of git://git.kernel.dk/linux:
  block: change blk_mq_add_to_batch() third argument type to bool
  nvme: move error logging from nvme_end_req() to __nvme_end_req()
  nvmet: pci-epf: Do not add an IRQ vector if not needed
  nvmet: pci-epf: Set NVMET_PCI_EPF_Q_LIVE when a queue is fully created
  nvme-pci: fix stuck reset on concurrent DPC and HP

Author: torvalds

drivers/block/null_blk/main.c

@@ -1549,8 +1549,8 @@ static int null_poll(struct blk_mq_hw_ctx *hctx, struct io_comp_batch *iob)
 		cmd = blk_mq_rq_to_pdu(req);
 		cmd->error = null_process_cmd(cmd, req_op(req), blk_rq_pos(req),
 						blk_rq_sectors(req));
-		if (!blk_mq_add_to_batch(req, iob, (__force int) cmd->error,
-					blk_mq_end_request_batch))
+		if (!blk_mq_add_to_batch(req, iob, cmd->error != BLK_STS_OK,
+					 blk_mq_end_request_batch))
 			blk_mq_end_request(req, cmd->error);
 		nr++;
 	}

drivers/block/virtio_blk.c

@@ -1207,11 +1207,12 @@ static int virtblk_poll(struct blk_mq_hw_ctx *hctx, struct io_comp_batch *iob)
 
 	while ((vbr = virtqueue_get_buf(vq->vq, &len)) != NULL) {
 		struct request *req = blk_mq_rq_from_pdu(vbr);
+		u8 status = virtblk_vbr_status(vbr);
 
 		found++;
 		if (!blk_mq_complete_request_remote(req) &&
-		    !blk_mq_add_to_batch(req, iob, virtblk_vbr_status(vbr),
-						virtblk_complete_batch))
+		    !blk_mq_add_to_batch(req, iob, status != VIRTIO_BLK_S_OK,
+					 virtblk_complete_batch))
 			virtblk_request_done(req);
 	}
 

drivers/nvme/host/apple.c

@@ -599,7 +599,8 @@ static inline void apple_nvme_handle_cqe(struct apple_nvme_queue *q,
 	}
 
 	if (!nvme_try_complete_req(req, cqe->status, cqe->result) &&
-	    !blk_mq_add_to_batch(req, iob, nvme_req(req)->status,
+	    !blk_mq_add_to_batch(req, iob,
+				 nvme_req(req)->status != NVME_SC_SUCCESS,
 				 apple_nvme_complete_batch))
 		apple_nvme_complete_rq(req);
 }

drivers/nvme/host/core.c

@@ -431,6 +431,12 @@ static inline void nvme_end_req_zoned(struct request *req)
 
 static inline void __nvme_end_req(struct request *req)
 {
+	if (unlikely(nvme_req(req)->status && !(req->rq_flags & RQF_QUIET))) {
+		if (blk_rq_is_passthrough(req))
+			nvme_log_err_passthru(req);
+		else
+			nvme_log_error(req);
+	}
 	nvme_end_req_zoned(req);
 	nvme_trace_bio_complete(req);
 	if (req->cmd_flags & REQ_NVME_MPATH)
@@ -441,12 +447,6 @@ void nvme_end_req(struct request *req)
 {
 	blk_status_t status = nvme_error_status(nvme_req(req)->status);
 
-	if (unlikely(nvme_req(req)->status && !(req->rq_flags & RQF_QUIET))) {
-		if (blk_rq_is_passthrough(req))
-			nvme_log_err_passthru(req);
-		else
-			nvme_log_error(req);
-	}
 	__nvme_end_req(req);
 	blk_mq_end_request(req, status);
 }

drivers/nvme/host/pci.c

@@ -1130,8 +1130,9 @@ static inline void nvme_handle_cqe(struct nvme_queue *nvmeq,
 
 	trace_nvme_sq(req, cqe->sq_head, nvmeq->sq_tail);
 	if (!nvme_try_complete_req(req, cqe->status, cqe->result) &&
-	    !blk_mq_add_to_batch(req, iob, nvme_req(req)->status,
-					nvme_pci_complete_batch))
+	    !blk_mq_add_to_batch(req, iob,
+				 nvme_req(req)->status != NVME_SC_SUCCESS,
+				 nvme_pci_complete_batch))
 		nvme_pci_complete_rq(req);
 }
 
@@ -1411,17 +1412,28 @@ static enum blk_eh_timer_return nvme_timeout(struct request *req)
 	struct nvme_dev *dev = nvmeq->dev;
 	struct request *abort_req;
 	struct nvme_command cmd = { };
+	struct pci_dev *pdev = to_pci_dev(dev->dev);
 	u32 csts = readl(dev->bar + NVME_REG_CSTS);
 	u8 opcode;
 
+	/*
+	 * Shutdown the device immediately if we see it is disconnected. This
+	 * unblocks PCIe error handling if the nvme driver is waiting in
+	 * error_resume for a device that has been removed. We can't unbind the
+	 * driver while the driver's error callback is waiting to complete, so
+	 * we're relying on a timeout to break that deadlock if a removal
+	 * occurs while reset work is running.
+	 */
+	if (pci_dev_is_disconnected(pdev))
+		nvme_change_ctrl_state(&dev->ctrl, NVME_CTRL_DELETING);
 	if (nvme_state_terminal(&dev->ctrl))
 		goto disable;
 
 	/* If PCI error recovery process is happening, we cannot reset or
 	 * the recovery mechanism will surely fail.
 	 */
 	mb();
-	if (pci_channel_offline(to_pci_dev(dev->dev)))
+	if (pci_channel_offline(pdev))
 		return BLK_EH_RESET_TIMER;
 
 	/*

drivers/nvme/target/pci-epf.c

@@ -1265,15 +1265,12 @@ static u16 nvmet_pci_epf_create_cq(struct nvmet_ctrl *tctrl,
 	struct nvmet_pci_epf_queue *cq = &ctrl->cq[cqid];
 	u16 status;
 
-	if (test_and_set_bit(NVMET_PCI_EPF_Q_LIVE, &cq->flags))
+	if (test_bit(NVMET_PCI_EPF_Q_LIVE, &cq->flags))
 		return NVME_SC_QID_INVALID | NVME_STATUS_DNR;
 
 	if (!(flags & NVME_QUEUE_PHYS_CONTIG))
 		return NVME_SC_INVALID_QUEUE | NVME_STATUS_DNR;
 
-	if (flags & NVME_CQ_IRQ_ENABLED)
-		set_bit(NVMET_PCI_EPF_Q_IRQ_ENABLED, &cq->flags);
-
 	cq->pci_addr = pci_addr;
 	cq->qid = cqid;
 	cq->depth = qsize + 1;
@@ -1290,24 +1287,27 @@ static u16 nvmet_pci_epf_create_cq(struct nvmet_ctrl *tctrl,
 		cq->qes = ctrl->io_cqes;
 	cq->pci_size = cq->qes * cq->depth;
 
-	cq->iv = nvmet_pci_epf_add_irq_vector(ctrl, vector);
-	if (!cq->iv) {
-		status = NVME_SC_INTERNAL | NVME_STATUS_DNR;
-		goto err;
+	if (flags & NVME_CQ_IRQ_ENABLED) {
+		cq->iv = nvmet_pci_epf_add_irq_vector(ctrl, vector);
+		if (!cq->iv)
+			return NVME_SC_INTERNAL | NVME_STATUS_DNR;
+		set_bit(NVMET_PCI_EPF_Q_IRQ_ENABLED, &cq->flags);
 	}
 
 	status = nvmet_cq_create(tctrl, &cq->nvme_cq, cqid, cq->depth);
 	if (status != NVME_SC_SUCCESS)
 		goto err;
 
+	set_bit(NVMET_PCI_EPF_Q_LIVE, &cq->flags);
+
 	dev_dbg(ctrl->dev, "CQ[%u]: %u entries of %zu B, IRQ vector %u\n",
 		cqid, qsize, cq->qes, cq->vector);
 
 	return NVME_SC_SUCCESS;
 
 err:
-	clear_bit(NVMET_PCI_EPF_Q_IRQ_ENABLED, &cq->flags);
-	clear_bit(NVMET_PCI_EPF_Q_LIVE, &cq->flags);
+	if (test_and_clear_bit(NVMET_PCI_EPF_Q_IRQ_ENABLED, &cq->flags))
+		nvmet_pci_epf_remove_irq_vector(ctrl, cq->vector);
 	return status;
 }
 
@@ -1333,7 +1333,7 @@ static u16 nvmet_pci_epf_create_sq(struct nvmet_ctrl *tctrl,
 	struct nvmet_pci_epf_queue *sq = &ctrl->sq[sqid];
 	u16 status;
 
-	if (test_and_set_bit(NVMET_PCI_EPF_Q_LIVE, &sq->flags))
+	if (test_bit(NVMET_PCI_EPF_Q_LIVE, &sq->flags))
 		return NVME_SC_QID_INVALID | NVME_STATUS_DNR;
 
 	if (!(flags & NVME_QUEUE_PHYS_CONTIG))
@@ -1355,7 +1355,7 @@ static u16 nvmet_pci_epf_create_sq(struct nvmet_ctrl *tctrl,
 
 	status = nvmet_sq_create(tctrl, &sq->nvme_sq, sqid, sq->depth);
 	if (status != NVME_SC_SUCCESS)
-		goto out_clear_bit;
+		return status;
 
 	sq->iod_wq = alloc_workqueue("sq%d_wq", WQ_UNBOUND,
 				min_t(int, sq->depth, WQ_MAX_ACTIVE), sqid);
@@ -1365,15 +1365,15 @@ static u16 nvmet_pci_epf_create_sq(struct nvmet_ctrl *tctrl,
 		goto out_destroy_sq;
 	}
 
+	set_bit(NVMET_PCI_EPF_Q_LIVE, &sq->flags);
+
 	dev_dbg(ctrl->dev, "SQ[%u]: %u entries of %zu B\n",
 		sqid, qsize, sq->qes);
 
 	return NVME_SC_SUCCESS;
 
 out_destroy_sq:
 	nvmet_sq_destroy(&sq->nvme_sq);
-out_clear_bit:
-	clear_bit(NVMET_PCI_EPF_Q_LIVE, &sq->flags);
 	return status;
 }
 

include/linux/blk-mq.h

@@ -852,20 +852,28 @@ static inline bool blk_mq_is_reserved_rq(struct request *rq)
 	return rq->rq_flags & RQF_RESV;
 }
 
-/*
+/**
+ * blk_mq_add_to_batch() - add a request to the completion batch
+ * @req: The request to add to batch
+ * @iob: The batch to add the request
+ * @is_error: Specify true if the request failed with an error
+ * @complete: The completaion handler for the request
+ *
  * Batched completions only work when there is no I/O error and no special
  * ->end_io handler.
+ *
+ * Return: true when the request was added to the batch, otherwise false
  */
 static inline bool blk_mq_add_to_batch(struct request *req,
-				       struct io_comp_batch *iob, int ioerror,
+				       struct io_comp_batch *iob, bool is_error,
 				       void (*complete)(struct io_comp_batch *))
 {
 	/*
 	 * Check various conditions that exclude batch processing:
 	 * 1) No batch container
 	 * 2) Has scheduler data attached
 	 * 3) Not a passthrough request and end_io set
-	 * 4) Not a passthrough request and an ioerror
+	 * 4) Not a passthrough request and failed with an error
 	 */
 	if (!iob)
 		return false;
@@ -874,7 +882,7 @@ static inline bool blk_mq_add_to_batch(struct request *req,
 	if (!blk_rq_is_passthrough(req)) {
 		if (req->end_io)
 			return false;
-		if (ioerror < 0)
+		if (is_error)
 			return false;
 	}