Re: [dpdk-stable] [dpdk-dev] [PATCH v2] app/testpmd: fix pmd_test_exit function for vdevs

[Thomas Monjalon <thomas@monjalon.net>]

21/05/2018 18:44, Ferruh Yigit:
> On 5/21/2018 5:40 PM, Thomas Monjalon wrote:
> > 21/05/2018 18:37, Ferruh Yigit:
> >> On 5/19/2018 3:19 PM, Thomas Monjalon wrote:
> >>> 18/05/2018 18:29, Ferruh Yigit:
> >>>> On 5/18/2018 4:55 PM, Matan Azrad wrote:
> >>>>> Hi all
> >>>>>
> >>>>> While this patch also applied I don't understand it.
> >>>>> Is it mandatory for each PMD to free all its resources in dev_close()?
> >>>>> Or it should be done by the rte_device remove function?
> >>>>>
> >>>>> If the resource cleanup should be done by the remove function I think it
> >>>>> should be called for all the devices (pci, vdev, etc).
> >>>>>
> >>>>> Is there an exit function for EAL to clean rte_eal_init()? If no, looks like we need it...
> >>>>
> >>>> Hi Matan,
> >>>>
> >>>> I believe there is a gap in resource cleanup.
> >>>> dev_close() it not for resource cleanup, it should be in PMD remove() functions,
> >>>> and PMDs have it. The problem is remove path is not called in application exit.
> >>>>
> >>>> As far as I know there is no simple API to clean the resources, having it may
> >>>> help application to do the cleanup.
> >>>>
> >>>> I have seen the rte_eal_cleanup() API by Harry, that can be extended to cover
> >>>> PMD resource cleanup if there is enough motivation for it.
> >>>
> >>> Yes, EAL resources should be removed by the function rte_eal_cleanup().
> >>> And ethdev ports must be removed by rte_eth_dev_close().
> >>
> >> There is probe() and remove() functions.
> >> There is dev_close devops, called by rte_eth_dev_close(), but there is no open()
> >> or equivalent.
> >>
> >> For example an ethdev allocated its private data during probe(), if
> >> rte_eth_dev_close() free it, how can a new ethdev can be allocated?
> > 
> > I don't understand the question.
> > You say closing a port and opening a new one.
> > So we allocate private data in the new probe.
> 
> the question is why resources allocated in probe() but freed in close() instead
> of remove()?

Because close() is the opposite of probe().

The function remove() does not exist in ethdev.
However it exists in EAL layer for rte_device.
Reminder: one rte_device can be common to several ethdev ports,
or other class of ports.

> Or should we have something like rte_eth_dev_open() ?

I don't think it is needed.