[PATCH v2] block/rnbd: Remove a useless mutex

[PATCH v2] block/rnbd: Remove a useless mutex

[Christophe JAILLET]

According to lib/idr.c,
   The IDA handles its own locking.  It is safe to call any of the IDA
   functions without synchronisation in your code.

so the 'ida_lock' mutex can just be removed.
It is here only to protect some ida_simple_get()/ida_simple_remove() calls.

While at it, switch to ida_alloc_XXX()/ida_free() instead to
ida_simple_get()/ida_simple_remove().
The latter is deprecated and more verbose.

Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
---
 drivers/block/rnbd/rnbd-clt.c | 11 +++--------
 1 file changed, 3 insertions(+), 8 deletions(-)

diff --git a/drivers/block/rnbd/rnbd-clt.c b/drivers/block/rnbd/rnbd-clt.c
index 9a880d559ab8..1f63f308eb39 100644
--- a/drivers/block/rnbd/rnbd-clt.c
+++ b/drivers/block/rnbd/rnbd-clt.c
@@ -23,7 +23,6 @@ MODULE_LICENSE("GPL");
 
 static int rnbd_client_major;
 static DEFINE_IDA(index_ida);
-static DEFINE_MUTEX(ida_lock);
 static DEFINE_MUTEX(sess_lock);
 static LIST_HEAD(sess_list);
 
@@ -55,9 +54,7 @@ static void rnbd_clt_put_dev(struct rnbd_clt_dev *dev)
 	if (!refcount_dec_and_test(&dev->refcount))
 		return;
 
-	mutex_lock(&ida_lock);
-	ida_simple_remove(&index_ida, dev->clt_device_id);
-	mutex_unlock(&ida_lock);
+	ida_free(&index_ida, dev->clt_device_id);
 	kfree(dev->hw_queues);
 	kfree(dev->pathname);
 	rnbd_clt_put_sess(dev->sess);
@@ -1460,10 +1457,8 @@ static struct rnbd_clt_dev *init_dev(struct rnbd_clt_session *sess,
 		goto out_alloc;
 	}
 
-	mutex_lock(&ida_lock);
-	ret = ida_simple_get(&index_ida, 0, 1 << (MINORBITS - RNBD_PART_BITS),
-			     GFP_KERNEL);
-	mutex_unlock(&ida_lock);
+	ret = ida_alloc_max(&index_ida, 1 << (MINORBITS - RNBD_PART_BITS),
+			    GFP_KERNEL);
 	if (ret < 0) {
 		pr_err("Failed to initialize device '%s' from session %s, allocating idr failed, err: %d\n",
 		       pathname, sess->sessname, ret);
-- 
2.32.0

Re: [PATCH v2] block/rnbd: Remove a useless mutex

[Jinpu Wang]

On Mon, Feb 7, 2022 at 9:48 PM Christophe JAILLET
<christophe.jaillet@wanadoo.fr> wrote:
>
> According to lib/idr.c,
>    The IDA handles its own locking.  It is safe to call any of the IDA
>    functions without synchronisation in your code.
>
> so the 'ida_lock' mutex can just be removed.
> It is here only to protect some ida_simple_get()/ida_simple_remove() calls.
>
> While at it, switch to ida_alloc_XXX()/ida_free() instead to
> ida_simple_get()/ida_simple_remove().
> The latter is deprecated and more verbose.
>
> Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Acked-by: Jack Wang <jinpu.wang@ionos.com>
Thanks!
> ---
>  drivers/block/rnbd/rnbd-clt.c | 11 +++--------
>  1 file changed, 3 insertions(+), 8 deletions(-)
>
> diff --git a/drivers/block/rnbd/rnbd-clt.c b/drivers/block/rnbd/rnbd-clt.c
> index 9a880d559ab8..1f63f308eb39 100644
> --- a/drivers/block/rnbd/rnbd-clt.c
> +++ b/drivers/block/rnbd/rnbd-clt.c
> @@ -23,7 +23,6 @@ MODULE_LICENSE("GPL");
>
>  static int rnbd_client_major;
>  static DEFINE_IDA(index_ida);
> -static DEFINE_MUTEX(ida_lock);
>  static DEFINE_MUTEX(sess_lock);
>  static LIST_HEAD(sess_list);
>
> @@ -55,9 +54,7 @@ static void rnbd_clt_put_dev(struct rnbd_clt_dev *dev)
>         if (!refcount_dec_and_test(&dev->refcount))
>                 return;
>
> -       mutex_lock(&ida_lock);
> -       ida_simple_remove(&index_ida, dev->clt_device_id);
> -       mutex_unlock(&ida_lock);
> +       ida_free(&index_ida, dev->clt_device_id);
>         kfree(dev->hw_queues);
>         kfree(dev->pathname);
>         rnbd_clt_put_sess(dev->sess);
> @@ -1460,10 +1457,8 @@ static struct rnbd_clt_dev *init_dev(struct rnbd_clt_session *sess,
>                 goto out_alloc;
>         }
>
> -       mutex_lock(&ida_lock);
> -       ret = ida_simple_get(&index_ida, 0, 1 << (MINORBITS - RNBD_PART_BITS),
> -                            GFP_KERNEL);
> -       mutex_unlock(&ida_lock);
> +       ret = ida_alloc_max(&index_ida, 1 << (MINORBITS - RNBD_PART_BITS),
> +                           GFP_KERNEL);
>         if (ret < 0) {
>                 pr_err("Failed to initialize device '%s' from session %s, allocating idr failed, err: %d\n",
>                        pathname, sess->sessname, ret);
> --
> 2.32.0
>

Re: [PATCH v2] block/rnbd: Remove a useless mutex

[Jens Axboe]

On Mon, 7 Feb 2022 21:48:19 +0100, Christophe JAILLET wrote:
> According to lib/idr.c,
>    The IDA handles its own locking.  It is safe to call any of the IDA
>    functions without synchronisation in your code.
> 
> so the 'ida_lock' mutex can just be removed.
> It is here only to protect some ida_simple_get()/ida_simple_remove() calls.
> 
> [...]

Applied, thanks!

[1/1] block/rnbd: Remove a useless mutex
      commit: a41cdf83a63122e5c5fb8c8550b812b25a332577

Best regards,
-- 
Jens Axboe