[PATCHv2] nvme: sanitize metadata bounce buffer for reads

[Keith Busch]

From: Keith Busch <kbusch at kernel.org>

User can request more metadata bytes than the device will write. Ensure
kernel buffer is initialized so we are not leaking unsanitized memory on
the completion's copy-out.

Fixes: 0b7f1f26f95a51a ("nvme: use the block layer for userspace passthrough metadata")
Signed-off-by: Keith Busch <kbusch at kernel.org>
---
v1->v2:

  correctly split data direction handling (axboe)

 drivers/nvme/host/ioctl.c | 10 +++++++---
 1 file changed, 7 insertions(+), 3 deletions(-)

diff --git a/drivers/nvme/host/ioctl.c b/drivers/nvme/host/ioctl.c
index 788b36e7915ab..eb2ef3e149614 100644
--- a/drivers/nvme/host/ioctl.c
+++ b/drivers/nvme/host/ioctl.c
@@ -36,9 +36,13 @@ static void *nvme_add_user_metadata(struct request *req, void __user *ubuf,
 	if (!buf)
 		goto out;
 
-	ret = -EFAULT;
-	if ((req_op(req) == REQ_OP_DRV_OUT) && copy_from_user(buf, ubuf, len))
-		goto out_free_meta;
+	if (req_op(req) == REQ_OP_DRV_OUT) {
+		ret = -EFAULT;
+		if (copy_from_user(buf, ubuf, len))
+			goto out_free_meta;
+	} else {
+		memset(buf, 0, len);
+	}
 
 	bip = bio_integrity_alloc(bio, GFP_KERNEL, 1);
 	if (IS_ERR(bip)) {
-- 
2.34.1