[dpdk-dev] i40evf: fix return value if command fails
Commit Message
Previously, if message is sent successfully, but no response is
received, function "i40evf_execute_vf_cmd" will return without error.
The root cause is value "err" is overwritten. This patch fixes it.
Fixes: ae19955e7c86 ("i40evf: support reporting PF reset")
Signed-off-by: Jingjing Wu <jingjing.wu@intel.com>
---
drivers/net/i40e/i40e_ethdev_vf.c | 8 +++-----
1 file changed, 3 insertions(+), 5 deletions(-)
Comments
On Thu, May 05, 2016 at 02:22:54PM +0800, Jingjing Wu wrote:
> Previously, if message is sent successfully, but no response is
> received, function "i40evf_execute_vf_cmd" will return without error.
> The root cause is value "err" is overwritten. This patch fixes it.
>
> Fixes: ae19955e7c86 ("i40evf: support reporting PF reset")
> Signed-off-by: Jingjing Wu <jingjing.wu@intel.com>
> ---
> drivers/net/i40e/i40e_ethdev_vf.c | 8 +++-----
> 1 file changed, 3 insertions(+), 5 deletions(-)
>
> diff --git a/drivers/net/i40e/i40e_ethdev_vf.c b/drivers/net/i40e/i40e_ethdev_vf.c
> index 2bce69b..9380019 100644
> --- a/drivers/net/i40e/i40e_ethdev_vf.c
> +++ b/drivers/net/i40e/i40e_ethdev_vf.c
> @@ -327,8 +327,7 @@ i40evf_execute_vf_cmd(struct rte_eth_dev *dev, struct vf_cmd_info *args)
> struct i40e_vf *vf = I40EVF_DEV_PRIVATE_TO_VF(dev->data->dev_private);
> struct i40evf_arq_msg_info info;
> enum i40evf_aq_result ret;
> - int err = -1;
> - int i = 0;
> + int err, i = 0;
>
> if (_atomic_set_cmd(vf, args->ops))
> return -1;
> @@ -346,6 +345,7 @@ i40evf_execute_vf_cmd(struct rte_eth_dev *dev, struct vf_cmd_info *args)
> return err;
> }
>
> + err = -1;
> switch (args->ops) {
> case I40E_VIRTCHNL_OP_RESET_VF:
> /*no need to process in this function */
Hi Jingjing,
this patch changes this return value in this case (I40E_VIRTCHNL_OP_RESET_VF).
Is this intentional - if so it probably needs to be documented in the commit
message. If not, the "err = -1" should be put at the start of the do { } while
loop for the GET_VF_RESOURCES command.
Perhaps to avoid future errors, each case in the switch should explicitly set
it's own default error value?
/Bruce
> -----Original Message-----
> From: Richardson, Bruce
> Sent: Friday, May 06, 2016 11:14 PM
> To: Wu, Jingjing
> Cc: Zhang, Helin; dev@dpdk.org
> Subject: Re: [dpdk-dev] [PATCH] i40evf: fix return value if command fails
>
> On Thu, May 05, 2016 at 02:22:54PM +0800, Jingjing Wu wrote:
> > Previously, if message is sent successfully, but no response is
> > received, function "i40evf_execute_vf_cmd" will return without error.
> > The root cause is value "err" is overwritten. This patch fixes it.
> >
> > Fixes: ae19955e7c86 ("i40evf: support reporting PF reset")
> > Signed-off-by: Jingjing Wu <jingjing.wu@intel.com>
> > ---
> > drivers/net/i40e/i40e_ethdev_vf.c | 8 +++-----
> > 1 file changed, 3 insertions(+), 5 deletions(-)
> >
> > diff --git a/drivers/net/i40e/i40e_ethdev_vf.c
> > b/drivers/net/i40e/i40e_ethdev_vf.c
> > index 2bce69b..9380019 100644
> > --- a/drivers/net/i40e/i40e_ethdev_vf.c
> > +++ b/drivers/net/i40e/i40e_ethdev_vf.c
> > @@ -327,8 +327,7 @@ i40evf_execute_vf_cmd(struct rte_eth_dev *dev,
> struct vf_cmd_info *args)
> > struct i40e_vf *vf = I40EVF_DEV_PRIVATE_TO_VF(dev->data-
> >dev_private);
> > struct i40evf_arq_msg_info info;
> > enum i40evf_aq_result ret;
> > - int err = -1;
> > - int i = 0;
> > + int err, i = 0;
> >
> > if (_atomic_set_cmd(vf, args->ops))
> > return -1;
> > @@ -346,6 +345,7 @@ i40evf_execute_vf_cmd(struct rte_eth_dev *dev,
> struct vf_cmd_info *args)
> > return err;
> > }
> >
> > + err = -1;
> > switch (args->ops) {
> > case I40E_VIRTCHNL_OP_RESET_VF:
> > /*no need to process in this function */
>
> Hi Jingjing,
>
> this patch changes this return value in this case
> (I40E_VIRTCHNL_OP_RESET_VF).
> Is this intentional - if so it probably needs to be documented in the commit
> message. If not, the "err = -1" should be put at the start of the do { } while
> loop for the GET_VF_RESOURCES command.
>
> Perhaps to avoid future errors, each case in the switch should explicitly set
> it's own default error value?
>
Yes, you reminded me the case I40E_VIRTCHNL_OP_RESET_VF. Let me rework this.
Thanks
Jingjing
> /Bruce
@@ -327,8 +327,7 @@ i40evf_execute_vf_cmd(struct rte_eth_dev *dev, struct vf_cmd_info *args)
struct i40e_vf *vf = I40EVF_DEV_PRIVATE_TO_VF(dev->data->dev_private);
struct i40evf_arq_msg_info info;
enum i40evf_aq_result ret;
- int err = -1;
- int i = 0;
+ int err, i = 0;
if (_atomic_set_cmd(vf, args->ops))
return -1;
@@ -346,6 +345,7 @@ i40evf_execute_vf_cmd(struct rte_eth_dev *dev, struct vf_cmd_info *args)
return err;
}
+ err = -1;
switch (args->ops) {
case I40E_VIRTCHNL_OP_RESET_VF:
/*no need to process in this function */
@@ -358,10 +358,8 @@ i40evf_execute_vf_cmd(struct rte_eth_dev *dev, struct vf_cmd_info *args)
if (ret == I40EVF_MSG_CMD) {
err = 0;
break;
- } else if (ret == I40EVF_MSG_ERR) {
- err = -1;
+ } else if (ret == I40EVF_MSG_ERR)
break;
- }
rte_delay_ms(ASQ_DELAY_MS);
/* If don't read msg or read sys event, continue */
} while (i++ < MAX_TRY_TIMES);