[PATCH] io_uring/kbuf: fix memory leak if io_buffer_add_list fails

[PATCH] io_uring/kbuf: fix memory leak if io_buffer_add_list fails

[Jens Axboe]

io_register_pbuf_ring() ignores the return value of io_buffer_add_list(),
which can fail if xa_store() returns an error (e.g., -ENOMEM). When this
happens, the function returns 0 (success) to the caller, but the
io_buffer_list structure is neither added to the xarray nor freed.

In practice this requires failure injection to hit, hence not a real
issue. But it should get fixed up none the less.

Fixes: ef62de3c4ad5 ("io_uring/kbuf: use region api for pbuf rings")
Signed-off-by: Jens Axboe <axboe@kernel.dk>

---

diff --git a/io_uring/kbuf.c b/io_uring/kbuf.c
index 796d131107dd..67d4fe576473 100644
--- a/io_uring/kbuf.c
+++ b/io_uring/kbuf.c
@@ -669,8 +669,9 @@ int io_register_pbuf_ring(struct io_ring_ctx *ctx, void __user *arg)
 	bl->buf_ring = br;
 	if (reg.flags & IOU_PBUF_RING_INC)
 		bl->flags |= IOBL_INC;
-	io_buffer_add_list(ctx, bl, reg.bgid);
-	return 0;
+	ret = io_buffer_add_list(ctx, bl, reg.bgid);
+	if (!ret)
+		return 0;
 fail:
 	io_free_region(ctx->user, &bl->region);
 	kfree(bl);

-- 
Jens Axboe

Re: [PATCH] io_uring/kbuf: fix memory leak if io_buffer_add_list fails

[Pavel Begunkov]

On 2/5/26 15:43, Jens Axboe wrote:
> io_register_pbuf_ring() ignores the return value of io_buffer_add_list(),
> which can fail if xa_store() returns an error (e.g., -ENOMEM). When this
> happens, the function returns 0 (success) to the caller, but the
> io_buffer_list structure is neither added to the xarray nor freed.
> 
> In practice this requires failure injection to hit, hence not a real
> issue. But it should get fixed up none the less.
> 
> Fixes: ef62de3c4ad5 ("io_uring/kbuf: use region api for pbuf rings")

Looks like that patch just moved the call, and the tag should be
more like:

c7fb19428d67d ("io_uring: add support for ring mapped supplied buffers")

> Signed-off-by: Jens Axboe <axboe@kernel.dk>
> 
> ---
> 
> diff --git a/io_uring/kbuf.c b/io_uring/kbuf.c
> index 796d131107dd..67d4fe576473 100644
> --- a/io_uring/kbuf.c
> +++ b/io_uring/kbuf.c
> @@ -669,8 +669,9 @@ int io_register_pbuf_ring(struct io_ring_ctx *ctx, void __user *arg)
>   	bl->buf_ring = br;
>   	if (reg.flags & IOU_PBUF_RING_INC)
>   		bl->flags |= IOBL_INC;
> -	io_buffer_add_list(ctx, bl, reg.bgid);
> -	return 0;
> +	ret = io_buffer_add_list(ctx, bl, reg.bgid);
> +	if (!ret)
> +		return 0;
>   fail:
>   	io_free_region(ctx->user, &bl->region);
>   	kfree(bl);
> 

-- 
Pavel Begunkov

Re: [PATCH] io_uring/kbuf: fix memory leak if io_buffer_add_list fails

[Jens Axboe]

On 2/5/26 11:11 AM, Pavel Begunkov wrote:
> On 2/5/26 15:43, Jens Axboe wrote:
>> io_register_pbuf_ring() ignores the return value of io_buffer_add_list(),
>> which can fail if xa_store() returns an error (e.g., -ENOMEM). When this
>> happens, the function returns 0 (success) to the caller, but the
>> io_buffer_list structure is neither added to the xarray nor freed.
>>
>> In practice this requires failure injection to hit, hence not a real
>> issue. But it should get fixed up none the less.
>>
>> Fixes: ef62de3c4ad5 ("io_uring/kbuf: use region api for pbuf rings")
> 
> Looks like that patch just moved the call, and the tag should be
> more like:
> 
> c7fb19428d67d ("io_uring: add support for ring mapped supplied buffers")

Yep agree, I'll update the tag.

-- 
Jens Axboe