[PATCH 2/4] nvme-fc: unmap cmd_iu DMA on rsp_iu mapping failure in init_request

[PATCH 2/4] nvme-fc: unmap cmd_iu DMA on rsp_iu mapping failure in init_request

[Guixin Liu]

__nvme_fc_init_request() maps cmd_iu and then rsp_iu for DMA. If the
rsp_iu mapping fails, the original code only recorded the error and fell
through: it left the already-mapped cmd_iu unmapped and still marked the
op as FCPOP_STATE_IDLE before returning. Since blk-mq does not call
.exit_request() when .init_request() fails, the cmd_iu mapping is leaked
for every op whose rsp_iu mapping fails.

Jump to an error path on rsp_iu mapping failure that unmaps cmd_iu and
returns the error without marking the op idle, so it stays in the
FCPOP_STATE_UNINIT state set by the initial memset().

Fixes: e399441de911 ("nvme-fabrics: Add host support for FC transport")
Signed-off-by: Guixin Liu <kanie@linux.alibaba.com>
---
 drivers/nvme/host/fc.c | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/drivers/nvme/host/fc.c b/drivers/nvme/host/fc.c
index 04363b9c4489..40f9da2833ff 100644
--- a/drivers/nvme/host/fc.c
+++ b/drivers/nvme/host/fc.c
@@ -2100,9 +2100,15 @@ __nvme_fc_init_request(struct nvme_fc_ctrl *ctrl,
 		dev_err(ctrl->dev,
 			"FCP Op failed - rspiu dma mapping failed.\n");
 		ret = -EFAULT;
+		goto out_unmap;
 	}
 
 	atomic_set(&op->state, FCPOP_STATE_IDLE);
+	return 0;
+
+out_unmap:
+	fc_dma_unmap_single(ctrl->lport->dev, op->fcp_req.cmddma,
+			sizeof(op->cmd_iu), DMA_TO_DEVICE);
 out_on_error:
 	return ret;
 }
-- 
2.43.7

Re: [PATCH 2/4] nvme-fc: unmap cmd_iu DMA on rsp_iu mapping failure in init_request

[Christoph Hellwig]

Looks good:

Reviewed-by: Christoph Hellwig <hch@lst.de>