[PATCH AUTOSEL 4.19 056/237] nvmet: avoid integer overflow in the discard code

From: Sasha Levin
Date: Sat Nov 16 2019 - 11:25:18 EST


From: Bart Van Assche <bvanassche@xxxxxxx>

[ Upstream commit 8eacd1bd21d6913ec27e6120e9a8733352e191d3 ]

Although I'm not sure whether it is a good idea to support large discard
commands, I think integer overflow for discard ranges larger than 4 GB
should be avoided. This patch avoids that smatch reports the following:

drivers/nvme/target/io-cmd-file.c:249:1 nvmet_file_execute_discard() warn: should '((range.nlb)) << req->ns->blksize_shift' be a 64 bit type?

Fixes: d5eff33ee6f8 ("nvmet: add simple file backed ns support")
Signed-off-by: Bart Van Assche <bvanassche@xxxxxxx>
Reviewed-by: Chaitanya Kulkarni <chaitanya.kulkarni@xxxxxxx>
Signed-off-by: Christoph Hellwig <hch@xxxxxx>
Signed-off-by: Sasha Levin <sashal@xxxxxxxxxx>
---
drivers/nvme/target/io-cmd-file.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/drivers/nvme/target/io-cmd-file.c b/drivers/nvme/target/io-cmd-file.c
index 81a9dc5290a87..39d972e2595f0 100644
--- a/drivers/nvme/target/io-cmd-file.c
+++ b/drivers/nvme/target/io-cmd-file.c
@@ -246,7 +246,8 @@ static void nvmet_file_execute_discard(struct nvmet_req *req)
break;

offset = le64_to_cpu(range.slba) << req->ns->blksize_shift;
- len = le32_to_cpu(range.nlb) << req->ns->blksize_shift;
+ len = le32_to_cpu(range.nlb);
+ len <<= req->ns->blksize_shift;
if (offset + len > req->ns->size) {
ret = NVME_SC_LBA_RANGE | NVME_SC_DNR;
break;
--
2.20.1