[PATCH] net/memif: allow stopping and closing device

[Ferruh Yigit]

On 11/24/2021 2:01 PM, yoursunny wrote:

Hi Junxiao, comment moved down, please don't top post,
it makes following the discussion very hard in the archive.

> Yours, Junxiao
> 
> On Wed, Nov 24, 2021, 06:02 Ferruh Yigit <ferruh.yigit at intel.com <mailto:ferruh.yigit at intel.com>> wrote:
> 
>     On 11/18/2021 5:33 PM, Junxiao Shi wrote:
>      > Bugzilla ID: 888
>      > Fixes: febc855b358e ("ethdev: forbid closing started device")
>      >
>      > Signed-off-by: Junxiao Shi <git at mail1.yoursunny.com <mailto:git at mail1.yoursunny.com>>
> 
>     Thanks Junxiao, +1 to this fix, cc'ed memif maintainer Jakub.
> 
>      > ---
>      >   drivers/net/memif/rte_eth_memif.c | 11 ++++++++---
>      >   1 file changed, 8 insertions(+), 3 deletions(-)
>      >
>      > diff --git a/drivers/net/memif/rte_eth_memif.c b/drivers/net/memif/rte_eth_memif.c
>      > index 43d7378329..e3d523af57 100644
>      > --- a/drivers/net/memif/rte_eth_memif.c
>      > +++ b/drivers/net/memif/rte_eth_memif.c
>      > @@ -1260,6 +1260,13 @@ memif_dev_start(struct rte_eth_dev *dev)
>      >       return ret;
>      >   }
>      >
>      > +static int
>      > +memif_dev_stop(struct rte_eth_dev *dev)
>      > +{
>      > +     memif_disconnect(dev);
> 
>     Is the 'memif_dev_stop()' safe to be called multiple times?
>     If 'memif_dev_close()' calls 'memif_dev_stop()' (see below), it is possible
>     to call 'memif_dev_stop()' multiple times, so it should be protected.
> 
>      > +     return 0;
>      > +}
>      > +
>      >   static int
>      >   memif_dev_close(struct rte_eth_dev *dev)
>      >   {
>      > @@ -1268,7 +1275,6 @@ memif_dev_close(struct rte_eth_dev *dev)
>      >
>      >       if (rte_eal_process_type() == RTE_PROC_PRIMARY) {
>      >               memif_msg_enq_disconnect(pmd->cc, "Device closed", 0);
>      > -             memif_disconnect(dev);
>      >
>      >               for (i = 0; i < dev->data->nb_rx_queues; i++)
>      >                       (*dev->dev_ops->rx_queue_release)(dev, i);
>      > @@ -1276,8 +1282,6 @@ memif_dev_close(struct rte_eth_dev *dev)
>      >                       (*dev->dev_ops->tx_queue_release)(dev, i);
>      >
>      >               memif_socket_remove_device(dev);
>      > -     } else {
>      > -             memif_disconnect(dev);
> 
>     Should we add 'memif_dev_stop()' within the close function?
>     Otherwise we are relying on user to stop, but at least in remove path
>     ('rte_pmd_memif_remove()') that may not be the case.
> 

> Hi Ferruh
> 
> You have to rely on user to call stop before calling close/remove.
> This is mandated in ethdev library, as implemented in:
> febc855b358e ("ethdev: forbid closing started device")
> 


Yes it is enforced,
and 'rte_pmd_memif_remove()' calls 'rte_eth_dev_close()' instead of
'memif_dev_close()', so agree that it won't mean much to add stop()
within close().

I will proceed with the patch without waiting review from Jakup
to have it in the release.

>      >       }
>      >
>      >       rte_free(dev->process_private);
>      > @@ -1515,6 +1519,7 @@ memif_rx_queue_intr_disable(struct rte_eth_dev *dev, uint16_t qid __rte_unused)
>      >
>      >   static const struct eth_dev_ops ops = {
>      >       .dev_start = memif_dev_start,
>      > +     .dev_stop = memif_dev_stop,
>      >       .dev_close = memif_dev_close,
>      >       .dev_infos_get = memif_dev_info,
>      >       .dev_configure = memif_dev_configure,
>      >
>