[PATCH] io_uring: return void from io_put_kbuf_comp()

[PATCH] io_uring: return void from io_put_kbuf_comp()

[Ming Lei]

The two callers don't handle the return value of io_put_kbuf_comp(), so
change its return type into void.

Signed-off-by: Ming Lei <ming.lei@redhat.com>
---
 io_uring/kbuf.h | 12 +++---------
 1 file changed, 3 insertions(+), 9 deletions(-)

diff --git a/io_uring/kbuf.h b/io_uring/kbuf.h
index 1c7b654ee726..86931fa655ad 100644
--- a/io_uring/kbuf.h
+++ b/io_uring/kbuf.h
@@ -119,18 +119,12 @@ static inline void __io_put_kbuf_list(struct io_kiocb *req,
 	}
 }
 
-static inline unsigned int io_put_kbuf_comp(struct io_kiocb *req)
+static inline void io_put_kbuf_comp(struct io_kiocb *req)
 {
-	unsigned int ret;
-
 	lockdep_assert_held(&req->ctx->completion_lock);
 
-	if (!(req->flags & (REQ_F_BUFFER_SELECTED|REQ_F_BUFFER_RING)))
-		return 0;
-
-	ret = IORING_CQE_F_BUFFER | (req->buf_index << IORING_CQE_BUFFER_SHIFT);
-	__io_put_kbuf_list(req, &req->ctx->io_buffers_comp);
-	return ret;
+	if (req->flags & (REQ_F_BUFFER_SELECTED|REQ_F_BUFFER_RING))
+		__io_put_kbuf_list(req, &req->ctx->io_buffers_comp);
 }
 
 static inline unsigned int io_put_kbuf(struct io_kiocb *req,
-- 
2.41.0

Re: [PATCH] io_uring: return void from io_put_kbuf_comp()

[Jens Axboe]

On 3/29/24 9:50 AM, Ming Lei wrote:
> The two callers don't handle the return value of io_put_kbuf_comp(), so
> change its return type into void.

We might want to consider changing the name of it too, it's a bit
different in that it's just recyling/dropping this kbuf rather than
posting a completion on behalf of it.

Maybe io_kbuf_drop() would be better. Would distuingish it from the
normal use cases of "drop this kbuf and return the cflags representation
of it, as I'm posting a completionw ith it".

> Signed-off-by: Ming Lei <ming.lei@redhat.com>
> ---
>  io_uring/kbuf.h | 12 +++---------
>  1 file changed, 3 insertions(+), 9 deletions(-)
> 
> diff --git a/io_uring/kbuf.h b/io_uring/kbuf.h
> index 1c7b654ee726..86931fa655ad 100644
> --- a/io_uring/kbuf.h
> +++ b/io_uring/kbuf.h
> @@ -119,18 +119,12 @@ static inline void __io_put_kbuf_list(struct io_kiocb *req,
>  	}
>  }
>  
> -static inline unsigned int io_put_kbuf_comp(struct io_kiocb *req)
> +static inline void io_put_kbuf_comp(struct io_kiocb *req)
>  {
> -	unsigned int ret;
> -
>  	lockdep_assert_held(&req->ctx->completion_lock);
>  
> -	if (!(req->flags & (REQ_F_BUFFER_SELECTED|REQ_F_BUFFER_RING)))
> -		return 0;
> -
> -	ret = IORING_CQE_F_BUFFER | (req->buf_index << IORING_CQE_BUFFER_SHIFT);
> -	__io_put_kbuf_list(req, &req->ctx->io_buffers_comp);
> -	return ret;
> +	if (req->flags & (REQ_F_BUFFER_SELECTED|REQ_F_BUFFER_RING))
> +		__io_put_kbuf_list(req, &req->ctx->io_buffers_comp);
>  }

If you post a v2 with the above suggestion, let's please just keep the
flags checking how it is. It's consistent with what we do elsewhere.

-- 
Jens Axboe

Re: [PATCH] io_uring: return void from io_put_kbuf_comp()

[Ming Lei]

On Fri, Mar 29, 2024 at 10:07:13AM -0600, Jens Axboe wrote:
> On 3/29/24 9:50 AM, Ming Lei wrote:
> > The two callers don't handle the return value of io_put_kbuf_comp(), so
> > change its return type into void.
> 
> We might want to consider changing the name of it too, it's a bit
> different in that it's just recyling/dropping this kbuf rather than
> posting a completion on behalf of it.
> 
> Maybe io_kbuf_drop() would be better. Would distuingish it from the
> normal use cases of "drop this kbuf and return the cflags representation
> of it, as I'm posting a completionw ith it".
> 
> > Signed-off-by: Ming Lei <ming.lei@redhat.com>
> > ---
> >  io_uring/kbuf.h | 12 +++---------
> >  1 file changed, 3 insertions(+), 9 deletions(-)
> > 
> > diff --git a/io_uring/kbuf.h b/io_uring/kbuf.h
> > index 1c7b654ee726..86931fa655ad 100644
> > --- a/io_uring/kbuf.h
> > +++ b/io_uring/kbuf.h
> > @@ -119,18 +119,12 @@ static inline void __io_put_kbuf_list(struct io_kiocb *req,
> >  	}
> >  }
> >  
> > -static inline unsigned int io_put_kbuf_comp(struct io_kiocb *req)
> > +static inline void io_put_kbuf_comp(struct io_kiocb *req)
> >  {
> > -	unsigned int ret;
> > -
> >  	lockdep_assert_held(&req->ctx->completion_lock);
> >  
> > -	if (!(req->flags & (REQ_F_BUFFER_SELECTED|REQ_F_BUFFER_RING)))
> > -		return 0;
> > -
> > -	ret = IORING_CQE_F_BUFFER | (req->buf_index << IORING_CQE_BUFFER_SHIFT);
> > -	__io_put_kbuf_list(req, &req->ctx->io_buffers_comp);
> > -	return ret;
> > +	if (req->flags & (REQ_F_BUFFER_SELECTED|REQ_F_BUFFER_RING))
> > +		__io_put_kbuf_list(req, &req->ctx->io_buffers_comp);
> >  }
> 
> If you post a v2 with the above suggestion, let's please just keep the
> flags checking how it is. It's consistent with what we do elsewhere.

OK, I can rename it as io_kbuf_drop() in v2 and keep the original check style.

Also the following patch removes one io_put_kbuf_comp(), I will post v2
when that patch gets feedback or merged.

https://lore.kernel.org/io-uring/20240329154712.1936153-1-ming.lei@redhat.com/


thanks,
Ming