2019-06-13 19:54:46

by Christophe JAILLET

Subject: [PATCH net-next] hinic: Use devm_kasprintf instead of hard coding it

'devm_kasprintf' is less verbose than:
snprintf(NULL, 0, ...);
devm_kzalloc(...);
sprintf
so use it instead.

Signed-off-by: Christophe JAILLET <[email protected]>
---
drivers/net/ethernet/huawei/hinic/hinic_rx.c | 8 +++-----
1 file changed, 3 insertions(+), 5 deletions(-)

diff --git a/drivers/net/ethernet/huawei/hinic/hinic_rx.c b/drivers/net/ethernet/huawei/hinic/hinic_rx.c
index 9b4082557ad5..95b09fd110d3 100644
--- a/drivers/net/ethernet/huawei/hinic/hinic_rx.c
+++ b/drivers/net/ethernet/huawei/hinic/hinic_rx.c
@@ -493,7 +493,7 @@ int hinic_init_rxq(struct hinic_rxq *rxq, struct hinic_rq *rq,
struct net_device *netdev)
{
struct hinic_qp *qp = container_of(rq, struct hinic_qp, rq);
- int err, pkts, irqname_len;
+ int err, pkts;

rxq->netdev = netdev;
rxq->rq = rq;
@@ -502,13 +502,11 @@ int hinic_init_rxq(struct hinic_rxq *rxq, struct hinic_rq *rq,

rxq_stats_init(rxq);

- irqname_len = snprintf(NULL, 0, "hinic_rxq%d", qp->q_id) + 1;
- rxq->irq_name = devm_kzalloc(&netdev->dev, irqname_len, GFP_KERNEL);
+ rxq->irq_name = devm_kasprintf(&netdev->dev, GFP_KERNEL,
+ "hinic_rxq%d", qp->q_id);
if (!rxq->irq_name)
return -ENOMEM;

- sprintf(rxq->irq_name, "hinic_rxq%d", qp->q_id);
-
pkts = rx_alloc_pkts(rxq);
if (!pkts) {
err = -ENOMEM;
--
2.20.1

2019-06-14 12:12:08

by Christophe JAILLET

Subject: Re: [PATCH net-next] hinic: Use devm_kasprintf instead of hard coding it

Hi,

I got a:

<[email protected]>: host huawei.com[103.218.216.136] said: 550
5.1.1 Error: invalid recipients is found from 80.12.242.127

However, MAINTAINERS has:
HUAWEI ETHERNET DRIVER
M: Aviad Krawczyk <[email protected]>
L: [email protected]
S: Supported
F: Documentation/networking/hinic.txt
F: drivers/net/ethernet/huawei/hinic/

I don't know how this should be fixed (neither if it should be...), so if s.o. knows, please do.

Best regards,
Christophe Jaillet

Le 13/06/2019 à 21:54, Christophe JAILLET a écrit :
> 'devm_kasprintf' is less verbose than:
> snprintf(NULL, 0, ...);
> devm_kzalloc(...);
> sprintf
> so use it instead.
>
> Signed-off-by: Christophe JAILLET <[email protected]>
> ---
> drivers/net/ethernet/huawei/hinic/hinic_rx.c | 8 +++-----
> 1 file changed, 3 insertions(+), 5 deletions(-)
>
> diff --git a/drivers/net/ethernet/huawei/hinic/hinic_rx.c b/drivers/net/ethernet/huawei/hinic/hinic_rx.c
> index 9b4082557ad5..95b09fd110d3 100644
> --- a/drivers/net/ethernet/huawei/hinic/hinic_rx.c
> +++ b/drivers/net/ethernet/huawei/hinic/hinic_rx.c
> @@ -493,7 +493,7 @@ int hinic_init_rxq(struct hinic_rxq *rxq, struct hinic_rq *rq,
> struct net_device *netdev)
> {
> struct hinic_qp *qp = container_of(rq, struct hinic_qp, rq);
> - int err, pkts, irqname_len;
> + int err, pkts;
>
> rxq->netdev = netdev;
> rxq->rq = rq;
> @@ -502,13 +502,11 @@ int hinic_init_rxq(struct hinic_rxq *rxq, struct hinic_rq *rq,
>
> rxq_stats_init(rxq);
>
> - irqname_len = snprintf(NULL, 0, "hinic_rxq%d", qp->q_id) + 1;
> - rxq->irq_name = devm_kzalloc(&netdev->dev, irqname_len, GFP_KERNEL);
> + rxq->irq_name = devm_kasprintf(&netdev->dev, GFP_KERNEL,
> + "hinic_rxq%d", qp->q_id);
> if (!rxq->irq_name)
> return -ENOMEM;
>
> - sprintf(rxq->irq_name, "hinic_rxq%d", qp->q_id);
> -
> pkts = rx_alloc_pkts(rxq);
> if (!pkts) {
> err = -ENOMEM;

2019-06-14 12:23:17

by Julia Lawall

Subject: Re: [PATCH net-next] hinic: Use devm_kasprintf instead of hard coding it

On Fri, 14 Jun 2019, Christophe JAILLET wrote:

> Hi,
>
> I got a:
>
> <[email protected]>: host huawei.com[103.218.216.136] said: 550
> 5.1.1 Error: invalid recipients is found from 80.12.242.127
>
> However, MAINTAINERS has:
> HUAWEI ETHERNET DRIVER
> M: Aviad Krawczyk <[email protected]>
> L: [email protected]
> S: Supported
> F: Documentation/networking/hinic.txt
> F: drivers/net/ethernet/huawei/hinic/
>
> I don't know how this should be fixed (neither if it should be...), so if s.o.
> knows, please do.

Maybe this person would know, since he is also from Huawei and has signed
off on a patch by Aviad Krawczyk:

cde66f24c3bf42123647c5233447c5790d92557f
Signed-off-by: Zhao Chen <[email protected]>

julia

>
> Best regards,
> Christophe Jaillet
>
> Le 13/06/2019 à 21:54, Christophe JAILLET a écrit :
> > 'devm_kasprintf' is less verbose than:
> > snprintf(NULL, 0, ...);
> > devm_kzalloc(...);
> > sprintf
> > so use it instead.
> >
> > Signed-off-by: Christophe JAILLET <[email protected]>
> > ---
> > drivers/net/ethernet/huawei/hinic/hinic_rx.c | 8 +++-----
> > 1 file changed, 3 insertions(+), 5 deletions(-)
> >
> > diff --git a/drivers/net/ethernet/huawei/hinic/hinic_rx.c
> > b/drivers/net/ethernet/huawei/hinic/hinic_rx.c
> > index 9b4082557ad5..95b09fd110d3 100644
> > --- a/drivers/net/ethernet/huawei/hinic/hinic_rx.c
> > +++ b/drivers/net/ethernet/huawei/hinic/hinic_rx.c
> > @@ -493,7 +493,7 @@ int hinic_init_rxq(struct hinic_rxq *rxq, struct
> > hinic_rq *rq,
> > struct net_device *netdev)
> > {
> > struct hinic_qp *qp = container_of(rq, struct hinic_qp, rq);
> > - int err, pkts, irqname_len;
> > + int err, pkts;
> > rxq->netdev = netdev;
> > rxq->rq = rq;
> > @@ -502,13 +502,11 @@ int hinic_init_rxq(struct hinic_rxq *rxq, struct
> > hinic_rq *rq,
> > rxq_stats_init(rxq);
> > - irqname_len = snprintf(NULL, 0, "hinic_rxq%d", qp->q_id) + 1;
> > - rxq->irq_name = devm_kzalloc(&netdev->dev, irqname_len, GFP_KERNEL);
> > + rxq->irq_name = devm_kasprintf(&netdev->dev, GFP_KERNEL,
> > + "hinic_rxq%d", qp->q_id);
> > if (!rxq->irq_name)
> > return -ENOMEM;
> > - sprintf(rxq->irq_name, "hinic_rxq%d", qp->q_id);
> > -
> > pkts = rx_alloc_pkts(rxq);
> > if (!pkts) {
> > err = -ENOMEM;
>
>
>

2019-06-15 20:39:14

by David Miller

Subject: Re: [PATCH net-next] hinic: Use devm_kasprintf instead of hard coding it

From: Christophe JAILLET <[email protected]>
Date: Thu, 13 Jun 2019 21:54:12 +0200

> 'devm_kasprintf' is less verbose than:
> snprintf(NULL, 0, ...);
> devm_kzalloc(...);
> sprintf
> so use it instead.
>
> Signed-off-by: Christophe JAILLET <[email protected]>

Applied.