[ Upstream commit 7c2fd76048e95dd267055b5f5e0a48e6e7c81fd9 ]
On an NVMe namespace that does not support metadata, it is possible to send an IO command with metadata through io-passthru. This allows issues like [1] to trigger in the completion code path. nvme_map_user_request() doesn't check if the namespace supports metadata before sending it forward. It also allows admin commands with metadata to be processed as it ignores metadata when bdev == NULL and may report success.
Reject an IO command with metadata when the NVMe namespace doesn't support it and reject an admin command if it has metadata.
[1] https://lore.kernel.org/all/mb61pcylvnym8.fsf@amazon.com/
Suggested-by: Christoph Hellwig hch@lst.de Reviewed-by: Christoph Hellwig hch@lst.de Reviewed-by: Sagi Grimberg sagi@grimberg.me Reviewed-by: Anuj Gupta anuj20.g@samsung.com Signed-off-by: Keith Busch kbusch@kernel.org [ Minor changes to make it work on 6.1 ] Signed-off-by: Puranjay Mohan pjy@amazon.com --- drivers/nvme/host/ioctl.c | 8 +++++++- 1 file changed, 7 insertions(+), 1 deletion(-)
diff --git a/drivers/nvme/host/ioctl.c b/drivers/nvme/host/ioctl.c index b3e322e4ade38..a02873792890e 100644 --- a/drivers/nvme/host/ioctl.c +++ b/drivers/nvme/host/ioctl.c @@ -3,6 +3,7 @@ * Copyright (c) 2011-2014, Intel Corporation. * Copyright (c) 2017-2021 Christoph Hellwig. */ +#include <linux/blk-integrity.h> #include <linux/ptrace.h> /* for force_successful_syscall_return */ #include <linux/nvme_ioctl.h> #include <linux/io_uring.h> @@ -95,10 +96,15 @@ static int nvme_map_user_request(struct request *req, u64 ubuffer, struct request_queue *q = req->q; struct nvme_ns *ns = q->queuedata; struct block_device *bdev = ns ? ns->disk->part0 : NULL; + bool supports_metadata = bdev && blk_get_integrity(bdev->bd_disk); + bool has_metadata = meta_buffer && meta_len; struct bio *bio = NULL; void *meta = NULL; int ret;
+ if (has_metadata && !supports_metadata) + return -EINVAL; + if (ioucmd && (ioucmd->flags & IORING_URING_CMD_FIXED)) { struct iov_iter iter;
@@ -122,7 +128,7 @@ static int nvme_map_user_request(struct request *req, u64 ubuffer, if (bdev) bio_set_dev(bio, bdev);
- if (bdev && meta_buffer && meta_len) { + if (has_metadata) { meta = nvme_add_user_metadata(req, meta_buffer, meta_len, meta_seed); if (IS_ERR(meta)) {
On Wed, Oct 16, 2024 at 09:07:39AM +0000, Puranjay Mohan wrote:
[ Upstream commit 7c2fd76048e95dd267055b5f5e0a48e6e7c81fd9 ]
On an NVMe namespace that does not support metadata, it is possible to send an IO command with metadata through io-passthru. This allows issues like [1] to trigger in the completion code path. nvme_map_user_request() doesn't check if the namespace supports metadata before sending it forward. It also allows admin commands with metadata to be processed as it ignores metadata when bdev == NULL and may report success.
Reject an IO command with metadata when the NVMe namespace doesn't support it and reject an admin command if it has metadata.
[1] https://lore.kernel.org/all/mb61pcylvnym8.fsf@amazon.com/
Suggested-by: Christoph Hellwig hch@lst.de Reviewed-by: Christoph Hellwig hch@lst.de Reviewed-by: Sagi Grimberg sagi@grimberg.me Reviewed-by: Anuj Gupta anuj20.g@samsung.com Signed-off-by: Keith Busch kbusch@kernel.org [ Minor changes to make it work on 6.1 ]
What about 6.6? And 5.15? We can't take patches for only some branches, otherwise when you upgrade you would have regressions.
Please send ALL needed patches and then we will be glad to queue them up. I've dropped the two submissions you sent for now.
thanks,
greg k-h
linux-stable-mirror@lists.linaro.org