From: Shida Zhang <zhangshida@kylinos.cn>

Replace repetitive bio chaining patterns with bio_chain_and_submit.
Note that while the parameter order (prev vs new) differs from the
original code, the chaining order does not affect bio chain
functionality.

Signed-off-by: Shida Zhang <zhangshida@kylinos.cn>
---
 drivers/nvme/target/io-cmd-bdev.c | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)

diff --git a/drivers/nvme/target/io-cmd-bdev.c b/drivers/nvme/target/io-cmd-bdev.c
index 8d246b8ca60..4af45659bd2 100644
--- a/drivers/nvme/target/io-cmd-bdev.c
+++ b/drivers/nvme/target/io-cmd-bdev.c
@@ -312,8 +312,7 @@ static void nvmet_bdev_execute_rw(struct nvmet_req *req)
 					opf, GFP_KERNEL);
 			bio->bi_iter.bi_sector = sector;
 
-			bio_chain(bio, prev);
-			submit_bio(prev);
+			bio_chain_and_submit(prev, bio);
 		}
 
 		sector += sg->length >> 9;
-- 
2.34.1

Acked-by: Sagi Grimberg <sagi@grimberg.me>

Sagi Grimberg <sagi@grimberg.me> 于2025年12月1日周一 07:03写道：
>
> Acked-by: Sagi Grimberg <sagi@grimberg.me>

Hello,

I already dropped this patch in v3:
https://lore.kernel.org/all/20251129090122.2457896-1-zhangshida@kylinos.cn/
The reason is that the order of operations is critical. In the original code::
----------------
...
bio->bi_end_io = nvmet_bio_done;

for_each_sg(req->sg, sg, req->sg_cnt, i) {
...
          struct bio *prev = bio;
....
          bio_chain(bio, prev);
          submit_bio(prev);
}
----------------

the oldest bio (i.e., prev) retains the real bi_end_io function:

bio -> bio -> ... -> prev
However, using bio_chain_and_submit(prev, bio) would create the reverse chain:

prev -> prev -> ... -> bio

where the newest bio would hold the real bi_end_io function, which does not
match the required behavior in this context.

Thanks,
Shida

zhangshida <starzhangzsd@gmail.com> 于2025年11月28日周五 16:33写道：
>
> From: Shida Zhang <zhangshida@kylinos.cn>
>
> Replace repetitive bio chaining patterns with bio_chain_and_submit.
> Note that while the parameter order (prev vs new) differs from the
> original code, the chaining order does not affect bio chain
> functionality.
>
> Signed-off-by: Shida Zhang <zhangshida@kylinos.cn>
> ---
>  drivers/nvme/target/io-cmd-bdev.c | 3 +--
>  1 file changed, 1 insertion(+), 2 deletions(-)
>
> diff --git a/drivers/nvme/target/io-cmd-bdev.c b/drivers/nvme/target/io-cmd-bdev.c
> index 8d246b8ca60..4af45659bd2 100644
> --- a/drivers/nvme/target/io-cmd-bdev.c
> +++ b/drivers/nvme/target/io-cmd-bdev.c
> @@ -312,8 +312,7 @@ static void nvmet_bdev_execute_rw(struct nvmet_req *req)
>                                         opf, GFP_KERNEL);
>                         bio->bi_iter.bi_sector = sector;
>
> -                       bio_chain(bio, prev);
> -                       submit_bio(prev);
> +                       bio_chain_and_submit(prev, bio);
>                 }
>

My apologies.  I think the order really matters here, will drop this patch.

Thanks,
shida

>                 sector += sg->length >> 9;
> --
> 2.34.1
>