scsi: storvsc: Fix swiotlb bounce buffer leak in confidential VM

storvsc_queuecommand() maps the scatter/gather list using scsi_dma_map(),
which in a confidential VM allocates swiotlb bounce buffers. If the I/O
submission fails in storvsc_do_io(), the I/O is typically retried by higher
level code, but the bounce buffer memory is never freed.  The mostly like
cause of I/O submission failure is a full VMBus channel ring buffer, which
is not uncommon under high I/O loads.  Eventually enough bounce buffer
memory leaks that the confidential VM can't do any I/O. The same problem
can arise in a non-confidential VM with kernel boot parameter
swiotlb=force.

Fix this by doing scsi_dma_unmap() in the case of an I/O submission
error, which frees the bounce buffer memory.

Fixes: 743b237c3a7b ("scsi: storvsc: Add Isolation VM support for storvsc driver")
Signed-off-by: Michael Kelley <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Tested-by: Dexuan Cui <[email protected]>
Reviewed-by: Dexuan Cui <[email protected]>
Reviewed-by: Tianyu Lan <[email protected]>
Signed-off-by: Martin K. Petersen <[email protected]>

1 files changed, 3 insertions, 0 deletions

diff --git a/drivers/scsi/storvsc_drv.c b/drivers/scsi/storvsc_drv.c
index a84194d82347..0252ee2dd844 100644
--- a/drivers/scsi/storvsc_drv.c
+++ b/drivers/scsi/storvsc_drv.c
@@ -1824,6 +1824,9 @@ static int storvsc_queuecommand(struct Scsi_Host *host, struct scsi_cmnd *scmnd)
 	ret = storvsc_do_io(dev, cmd_request, get_cpu());
 	put_cpu();
 
+	if (ret)
+		scsi_dma_unmap(scmnd);
+
 	if (ret == -EAGAIN) {
 		/* no more space */
 		ret = SCSI_MLQUEUE_DEVICE_BUSY;