[dpdk-dev] doc: announce driver name changes

Message ID 1468083394-94907-1-git-send-email-pablo.de.lara.guarch@intel.com (mailing list archive)
State Accepted, archived
Headers

Commit Message

De Lara Guarch, Pablo July 9, 2016, 4:56 p.m. UTC
  Driver names for all the supported devices in DPDK do not have
a naming convention. Some are using a prefix, some are not
and some have long names. Driver names are used when creating
virtual devices, so it is useful to have consistency in the names.

Signed-off-by: Pablo de Lara <pablo.de.lara.guarch@intel.com>
---
 doc/guides/rel_notes/deprecation.rst | 5 +++++
 1 file changed, 5 insertions(+)
  

Comments

Christian Ehrhardt July 11, 2016, 5:08 a.m. UTC | #1
Hi,
I'm all in for consistent naming and this "only" is the deprecation notice
ahead of time which is great.
Looking ahead I wanted to ask for opinions if more than just me would
consider it useful to keep aliases on the old names when the rename happens.

That way any sort of old config would continue to work.
Of course I could do that in the scope of the scripts that consume the
names and apply the driver loads.
But then that is true for everybody out there who has some kind of
config/tooling around the old names - so I thought it is worth to ask for
opinions.

Kind Regards,
Christian



Christian Ehrhardt
Software Engineer, Ubuntu Server
Canonical Ltd

On Sat, Jul 9, 2016 at 6:56 PM, Pablo de Lara <
pablo.de.lara.guarch@intel.com> wrote:

> Driver names for all the supported devices in DPDK do not have
> a naming convention. Some are using a prefix, some are not
> and some have long names. Driver names are used when creating
> virtual devices, so it is useful to have consistency in the names.
>
> Signed-off-by: Pablo de Lara <pablo.de.lara.guarch@intel.com>
> ---
>  doc/guides/rel_notes/deprecation.rst | 5 +++++
>  1 file changed, 5 insertions(+)
>
> diff --git a/doc/guides/rel_notes/deprecation.rst
> b/doc/guides/rel_notes/deprecation.rst
> index f502f86..37d65c8 100644
> --- a/doc/guides/rel_notes/deprecation.rst
> +++ b/doc/guides/rel_notes/deprecation.rst
> @@ -41,3 +41,8 @@ Deprecation Notices
>  * The mempool functions for single/multi producer/consumer are deprecated
> and
>    will be removed in 16.11.
>    It is replaced by rte_mempool_generic_get/put functions.
> +
> +* Driver names are quite inconsistent among each others and they will be
> +  renamed to something more consistent (net_ prefix for net drivers and
> +  crypto_ for crypto drivers) in 16.11. Some of these driver names are
> used
> +  publicly, to create virtual devices, so a deprecation notice is
> necessary.
> --
> 2.7.4
>
>
  
De Lara Guarch, Pablo July 22, 2016, 12:37 p.m. UTC | #2
Hi,

> -----Original Message-----
> From: De Lara Guarch, Pablo
> Sent: Saturday, July 09, 2016 5:57 PM
> To: dev@dpdk.org
> Cc: Mcnamara, John; De Lara Guarch, Pablo
> Subject: [PATCH] doc: announce driver name changes
> 
> Driver names for all the supported devices in DPDK do not have
> a naming convention. Some are using a prefix, some are not
> and some have long names. Driver names are used when creating
> virtual devices, so it is useful to have consistency in the names.
> 
> Signed-off-by: Pablo de Lara <pablo.de.lara.guarch@intel.com>
> ---
>  doc/guides/rel_notes/deprecation.rst | 5 +++++
>  1 file changed, 5 insertions(+)
> 
> diff --git a/doc/guides/rel_notes/deprecation.rst
> b/doc/guides/rel_notes/deprecation.rst
> index f502f86..37d65c8 100644
> --- a/doc/guides/rel_notes/deprecation.rst
> +++ b/doc/guides/rel_notes/deprecation.rst
> @@ -41,3 +41,8 @@ Deprecation Notices
>  * The mempool functions for single/multi producer/consumer are
> deprecated and
>    will be removed in 16.11.
>    It is replaced by rte_mempool_generic_get/put functions.
> +
> +* Driver names are quite inconsistent among each others and they will be
> +  renamed to something more consistent (net_ prefix for net drivers and
> +  crypto_ for crypto drivers) in 16.11. Some of these driver names are used
> +  publicly, to create virtual devices, so a deprecation notice is necessary.
> --
> 2.7.4

Any more comments on this (apart from Christian Ehrhardt's)?
  
Adrien Mazarguil July 22, 2016, 12:54 p.m. UTC | #3
Hi Pablo,

On Fri, Jul 22, 2016 at 12:37:22PM +0000, De Lara Guarch, Pablo wrote:
> Hi,
> 
> > -----Original Message-----
> > From: De Lara Guarch, Pablo
> > Sent: Saturday, July 09, 2016 5:57 PM
> > To: dev@dpdk.org
> > Cc: Mcnamara, John; De Lara Guarch, Pablo
> > Subject: [PATCH] doc: announce driver name changes
> > 
> > Driver names for all the supported devices in DPDK do not have
> > a naming convention. Some are using a prefix, some are not
> > and some have long names. Driver names are used when creating
> > virtual devices, so it is useful to have consistency in the names.
> > 
> > Signed-off-by: Pablo de Lara <pablo.de.lara.guarch@intel.com>
> > ---
> >  doc/guides/rel_notes/deprecation.rst | 5 +++++
> >  1 file changed, 5 insertions(+)
> > 
> > diff --git a/doc/guides/rel_notes/deprecation.rst
> > b/doc/guides/rel_notes/deprecation.rst
> > index f502f86..37d65c8 100644
> > --- a/doc/guides/rel_notes/deprecation.rst
> > +++ b/doc/guides/rel_notes/deprecation.rst
> > @@ -41,3 +41,8 @@ Deprecation Notices
> >  * The mempool functions for single/multi producer/consumer are
> > deprecated and
> >    will be removed in 16.11.
> >    It is replaced by rte_mempool_generic_get/put functions.
> > +
> > +* Driver names are quite inconsistent among each others and they will be
> > +  renamed to something more consistent (net_ prefix for net drivers and
> > +  crypto_ for crypto drivers) in 16.11. Some of these driver names are used
> > +  publicly, to create virtual devices, so a deprecation notice is necessary.
> > --
> > 2.7.4
> 
> Any more comments on this (apart from Christian Ehrhardt's)?

Yes, since you're suggesting to prefix driver names, shall "librte_pmd_mlx5"
really become "net_librte_pmd_mlx5" or shortened to "net_mlx5" instead?

What about using a '/' separator instead of '_'?

Will this impact directories as well ("net/mlx5" -> "net/net_mlx5")?
  
Ferruh Yigit July 22, 2016, 1:18 p.m. UTC | #4
On 7/22/2016 1:54 PM, Adrien Mazarguil wrote:
> Hi Pablo,
> 
> On Fri, Jul 22, 2016 at 12:37:22PM +0000, De Lara Guarch, Pablo wrote:
>> Hi,
>>
>>> -----Original Message-----
>>> From: De Lara Guarch, Pablo
>>> Sent: Saturday, July 09, 2016 5:57 PM
>>> To: dev@dpdk.org
>>> Cc: Mcnamara, John; De Lara Guarch, Pablo
>>> Subject: [PATCH] doc: announce driver name changes
>>>
>>> Driver names for all the supported devices in DPDK do not have
>>> a naming convention. Some are using a prefix, some are not
>>> and some have long names. Driver names are used when creating
>>> virtual devices, so it is useful to have consistency in the names.
>>>
>>> Signed-off-by: Pablo de Lara <pablo.de.lara.guarch@intel.com>
>>> ---
>>>  doc/guides/rel_notes/deprecation.rst | 5 +++++
>>>  1 file changed, 5 insertions(+)
>>>
>>> diff --git a/doc/guides/rel_notes/deprecation.rst
>>> b/doc/guides/rel_notes/deprecation.rst
>>> index f502f86..37d65c8 100644
>>> --- a/doc/guides/rel_notes/deprecation.rst
>>> +++ b/doc/guides/rel_notes/deprecation.rst
>>> @@ -41,3 +41,8 @@ Deprecation Notices
>>>  * The mempool functions for single/multi producer/consumer are
>>> deprecated and
>>>    will be removed in 16.11.
>>>    It is replaced by rte_mempool_generic_get/put functions.
>>> +
>>> +* Driver names are quite inconsistent among each others and they will be
>>> +  renamed to something more consistent (net_ prefix for net drivers and
>>> +  crypto_ for crypto drivers) in 16.11. Some of these driver names are used
>>> +  publicly, to create virtual devices, so a deprecation notice is necessary.
>>> --
>>> 2.7.4
>>
>> Any more comments on this (apart from Christian Ehrhardt's)?
> 
> Yes, since you're suggesting to prefix driver names, shall "librte_pmd_mlx5"
> really become "net_librte_pmd_mlx5" or shortened to "net_mlx5" instead?
> 
> What about using a '/' separator instead of '_'?
> 
> Will this impact directories as well ("net/mlx5" -> "net/net_mlx5")?
> 

For physical net devices, driver name is same as folder name (mlnx5,
ixgbe ...)

For virtual net devices, driver name is folder name with "eth_" prefix
(eth_pcap, eth_ring)

Driver names for net devices looks consistent already, I don't know
about crypto devices but if crypto driver names are inconsistent what do
you think renaming crypto drivers only?

Thanks,
ferruh
  
De Lara Guarch, Pablo July 22, 2016, 2:15 p.m. UTC | #5
> -----Original Message-----
> From: Yigit, Ferruh
> Sent: Friday, July 22, 2016 2:19 PM
> To: De Lara Guarch, Pablo; dev@dpdk.org; Mcnamara, John
> Subject: Re: [dpdk-dev] [PATCH] doc: announce driver name changes
> 
> On 7/22/2016 1:54 PM, Adrien Mazarguil wrote:
> > Hi Pablo,
> >
> > On Fri, Jul 22, 2016 at 12:37:22PM +0000, De Lara Guarch, Pablo wrote:
> >> Hi,
> >>
> >>> -----Original Message-----
> >>> From: De Lara Guarch, Pablo
> >>> Sent: Saturday, July 09, 2016 5:57 PM
> >>> To: dev@dpdk.org
> >>> Cc: Mcnamara, John; De Lara Guarch, Pablo
> >>> Subject: [PATCH] doc: announce driver name changes
> >>>
> >>> Driver names for all the supported devices in DPDK do not have
> >>> a naming convention. Some are using a prefix, some are not
> >>> and some have long names. Driver names are used when creating
> >>> virtual devices, so it is useful to have consistency in the names.
> >>>
> >>> Signed-off-by: Pablo de Lara <pablo.de.lara.guarch@intel.com>
> >>> ---
> >>>  doc/guides/rel_notes/deprecation.rst | 5 +++++
> >>>  1 file changed, 5 insertions(+)
> >>>
> >>> diff --git a/doc/guides/rel_notes/deprecation.rst
> >>> b/doc/guides/rel_notes/deprecation.rst
> >>> index f502f86..37d65c8 100644
> >>> --- a/doc/guides/rel_notes/deprecation.rst
> >>> +++ b/doc/guides/rel_notes/deprecation.rst
> >>> @@ -41,3 +41,8 @@ Deprecation Notices
> >>>  * The mempool functions for single/multi producer/consumer are
> >>> deprecated and
> >>>    will be removed in 16.11.
> >>>    It is replaced by rte_mempool_generic_get/put functions.
> >>> +
> >>> +* Driver names are quite inconsistent among each others and they will
> be
> >>> +  renamed to something more consistent (net_ prefix for net drivers and
> >>> +  crypto_ for crypto drivers) in 16.11. Some of these driver names are
> used
> >>> +  publicly, to create virtual devices, so a deprecation notice is necessary.
> >>> --
> >>> 2.7.4
> >>
> >> Any more comments on this (apart from Christian Ehrhardt's)?
> >
> > Yes, since you're suggesting to prefix driver names, shall "librte_pmd_mlx5"
> > really become "net_librte_pmd_mlx5" or shortened to "net_mlx5" instead?
> >
> > What about using a '/' separator instead of '_'?
> >
> > Will this impact directories as well ("net/mlx5" -> "net/net_mlx5")?
> >

We will leave these untouched, although I don't think renaming the directories was necessary.

> 
> For physical net devices, driver name is same as folder name (mlnx5,
> ixgbe ...)
> 
> For virtual net devices, driver name is folder name with "eth_" prefix
> (eth_pcap, eth_ring)
> 
> Driver names for net devices looks consistent already, I don't know
> about crypto devices but if crypto driver names are inconsistent what do
> you think renaming crypto drivers only?

Sure, as long as virtual Ethernet devices are consistent, I think it is ok.
My main intention here was to have consistent (and short) driver names,
to call rte_eal_vdev_init (or --vdev in command line).

Thanks,
Pablo

> 
> Thanks,
> ferruh
  
Thomas Monjalon July 22, 2016, 2:20 p.m. UTC | #6
2016-07-22 14:15, De Lara Guarch, Pablo:
> From: Yigit, Ferruh
> > For physical net devices, driver name is same as folder name (mlnx5,
> > ixgbe ...)
> > 
> > For virtual net devices, driver name is folder name with "eth_" prefix
> > (eth_pcap, eth_ring)
> > 
> > Driver names for net devices looks consistent already, I don't know
> > about crypto devices but if crypto driver names are inconsistent what do
> > you think renaming crypto drivers only?
> 
> Sure, as long as virtual Ethernet devices are consistent, I think it is ok.
> My main intention here was to have consistent (and short) driver names,
> to call rte_eal_vdev_init (or --vdev in command line).

It is important to rename every drivers with net and crypto prefixes.
The long term goal is to have a consistent command line specification
to pass parameters to any device/driver (virtual or physical).
  
Adrien Mazarguil July 22, 2016, 3:13 p.m. UTC | #7
On Fri, Jul 22, 2016 at 02:15:39PM +0000, De Lara Guarch, Pablo wrote:
> 
> 
> > -----Original Message-----
> > From: Yigit, Ferruh
> > Sent: Friday, July 22, 2016 2:19 PM
> > To: De Lara Guarch, Pablo; dev@dpdk.org; Mcnamara, John
> > Subject: Re: [dpdk-dev] [PATCH] doc: announce driver name changes
> > 
> > On 7/22/2016 1:54 PM, Adrien Mazarguil wrote:
> > > Hi Pablo,
> > >
> > > On Fri, Jul 22, 2016 at 12:37:22PM +0000, De Lara Guarch, Pablo wrote:
> > >> Hi,
> > >>
> > >>> -----Original Message-----
> > >>> From: De Lara Guarch, Pablo
> > >>> Sent: Saturday, July 09, 2016 5:57 PM
> > >>> To: dev@dpdk.org
> > >>> Cc: Mcnamara, John; De Lara Guarch, Pablo
> > >>> Subject: [PATCH] doc: announce driver name changes
> > >>>
> > >>> Driver names for all the supported devices in DPDK do not have
> > >>> a naming convention. Some are using a prefix, some are not
> > >>> and some have long names. Driver names are used when creating
> > >>> virtual devices, so it is useful to have consistency in the names.
> > >>>
> > >>> Signed-off-by: Pablo de Lara <pablo.de.lara.guarch@intel.com>
> > >>> ---
> > >>>  doc/guides/rel_notes/deprecation.rst | 5 +++++
> > >>>  1 file changed, 5 insertions(+)
> > >>>
> > >>> diff --git a/doc/guides/rel_notes/deprecation.rst
> > >>> b/doc/guides/rel_notes/deprecation.rst
> > >>> index f502f86..37d65c8 100644
> > >>> --- a/doc/guides/rel_notes/deprecation.rst
> > >>> +++ b/doc/guides/rel_notes/deprecation.rst
> > >>> @@ -41,3 +41,8 @@ Deprecation Notices
> > >>>  * The mempool functions for single/multi producer/consumer are
> > >>> deprecated and
> > >>>    will be removed in 16.11.
> > >>>    It is replaced by rte_mempool_generic_get/put functions.
> > >>> +
> > >>> +* Driver names are quite inconsistent among each others and they will
> > be
> > >>> +  renamed to something more consistent (net_ prefix for net drivers and
> > >>> +  crypto_ for crypto drivers) in 16.11. Some of these driver names are
> > used
> > >>> +  publicly, to create virtual devices, so a deprecation notice is necessary.
> > >>> --
> > >>> 2.7.4
> > >>
> > >> Any more comments on this (apart from Christian Ehrhardt's)?
> > >
> > > Yes, since you're suggesting to prefix driver names, shall "librte_pmd_mlx5"
> > > really become "net_librte_pmd_mlx5" or shortened to "net_mlx5" instead?
> > >
> > > What about using a '/' separator instead of '_'?
> > >
> > > Will this impact directories as well ("net/mlx5" -> "net/net_mlx5")?
> > >
> 
> We will leave these untouched, although I don't think renaming the directories was necessary.

My feeling as well, the depreciation notice wasn't clear about the extent of
name changes.

> > For physical net devices, driver name is same as folder name (mlnx5,
> > ixgbe ...)
> > 
> > For virtual net devices, driver name is folder name with "eth_" prefix
> > (eth_pcap, eth_ring)
> > 
> > Driver names for net devices looks consistent already, I don't know
> > about crypto devices but if crypto driver names are inconsistent what do
> > you think renaming crypto drivers only?
> 
> Sure, as long as virtual Ethernet devices are consistent, I think it is ok.
> My main intention here was to have consistent (and short) driver names,
> to call rte_eal_vdev_init (or --vdev in command line).

So what about using "net/" instead of "net_" to share names with commit
prefixes and folders?
  
Thomas Monjalon July 27, 2016, 8:30 a.m. UTC | #8
2016-07-22 17:13, Adrien Mazarguil:
> > > >> From: De Lara Guarch, Pablo
> > > >>> Driver names for all the supported devices in DPDK do not have
> > > >>> a naming convention. Some are using a prefix, some are not
> > > >>> and some have long names. Driver names are used when creating
> > > >>> virtual devices, so it is useful to have consistency in the names.
> > > >>>
> > > >>> Signed-off-by: Pablo de Lara <pablo.de.lara.guarch@intel.com>
> 
> So what about using "net/" instead of "net_" to share names with commit
> prefixes and folders?

+1 for net/ in names

Anyway,
Acked-by: Thomas Monjalon <thomas.monjalon@6wind.com>
  
Maxime Coquelin July 28, 2016, 5:22 p.m. UTC | #9
On 07/09/2016 06:56 PM, Pablo de Lara wrote:
> Driver names for all the supported devices in DPDK do not have
> a naming convention. Some are using a prefix, some are not
> and some have long names. Driver names are used when creating
> virtual devices, so it is useful to have consistency in the names.
>
> Signed-off-by: Pablo de Lara <pablo.de.lara.guarch@intel.com>
> ---
>  doc/guides/rel_notes/deprecation.rst | 5 +++++
>  1 file changed, 5 insertions(+)
>
> diff --git a/doc/guides/rel_notes/deprecation.rst b/doc/guides/rel_notes/deprecation.rst
> index f502f86..37d65c8 100644
> --- a/doc/guides/rel_notes/deprecation.rst
> +++ b/doc/guides/rel_notes/deprecation.rst
> @@ -41,3 +41,8 @@ Deprecation Notices
>  * The mempool functions for single/multi producer/consumer are deprecated and
>    will be removed in 16.11.
>    It is replaced by rte_mempool_generic_get/put functions.
> +
> +* Driver names are quite inconsistent among each others and they will be
> +  renamed to something more consistent (net_ prefix for net drivers and
> +  crypto_ for crypto drivers) in 16.11. Some of these driver names are used
> +  publicly, to create virtual devices, so a deprecation notice is necessary.

Consistency is better.
Out of curiosity, will a deprecation message be displayed at least or 
virtual interfaces if probe is tried with deprecated name?

Acked-by: Maxime Coquelin <maxime.coquelin@redhat.com>

Thanks,
Maxime
  
Jerin Jacob July 28, 2016, 5:47 p.m. UTC | #10
On Sat, Jul 09, 2016 at 05:56:34PM +0100, Pablo de Lara wrote:
> Driver names for all the supported devices in DPDK do not have
> a naming convention. Some are using a prefix, some are not
> and some have long names. Driver names are used when creating
> virtual devices, so it is useful to have consistency in the names.
> 
> Signed-off-by: Pablo de Lara <pablo.de.lara.guarch@intel.com>

Acked-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>

> ---
>  doc/guides/rel_notes/deprecation.rst | 5 +++++
>  1 file changed, 5 insertions(+)
> 
> diff --git a/doc/guides/rel_notes/deprecation.rst b/doc/guides/rel_notes/deprecation.rst
> index f502f86..37d65c8 100644
> --- a/doc/guides/rel_notes/deprecation.rst
> +++ b/doc/guides/rel_notes/deprecation.rst
> @@ -41,3 +41,8 @@ Deprecation Notices
>  * The mempool functions for single/multi producer/consumer are deprecated and
>    will be removed in 16.11.
>    It is replaced by rte_mempool_generic_get/put functions.
> +
> +* Driver names are quite inconsistent among each others and they will be
> +  renamed to something more consistent (net_ prefix for net drivers and
> +  crypto_ for crypto drivers) in 16.11. Some of these driver names are used
> +  publicly, to create virtual devices, so a deprecation notice is necessary.
> -- 
> 2.7.4
>
  
Thomas Monjalon July 28, 2016, 5:54 p.m. UTC | #11
> > Driver names for all the supported devices in DPDK do not have
> > a naming convention. Some are using a prefix, some are not
> > and some have long names. Driver names are used when creating
> > virtual devices, so it is useful to have consistency in the names.
> > 
> > Signed-off-by: Pablo de Lara <pablo.de.lara.guarch@intel.com>
> 
> Acked-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
Acked-by: Thomas Monjalon <thomas.monjalon@6wind.com>
Acked-by: Maxime Coquelin <maxime.coquelin@redhat.com>

Applied, thanks
  

Patch

diff --git a/doc/guides/rel_notes/deprecation.rst b/doc/guides/rel_notes/deprecation.rst
index f502f86..37d65c8 100644
--- a/doc/guides/rel_notes/deprecation.rst
+++ b/doc/guides/rel_notes/deprecation.rst
@@ -41,3 +41,8 @@  Deprecation Notices
 * The mempool functions for single/multi producer/consumer are deprecated and
   will be removed in 16.11.
   It is replaced by rte_mempool_generic_get/put functions.
+
+* Driver names are quite inconsistent among each others and they will be
+  renamed to something more consistent (net_ prefix for net drivers and
+  crypto_ for crypto drivers) in 16.11. Some of these driver names are used
+  publicly, to create virtual devices, so a deprecation notice is necessary.