[PATCH net-next 00/15] net: sparx5: prepare for lan969x switch driver

Daniel Machon posted 15 patches 1 month, 3 weeks ago
There is a newer version of this series
drivers/net/ethernet/microchip/sparx5/Makefile     |    2 +-
.../ethernet/microchip/sparx5/sparx5_calendar.c    |   41 +-
drivers/net/ethernet/microchip/sparx5/sparx5_dcb.c |    5 +-
.../net/ethernet/microchip/sparx5/sparx5_ethtool.c |   33 +-
.../net/ethernet/microchip/sparx5/sparx5_fdma.c    |    6 +-
.../ethernet/microchip/sparx5/sparx5_mactable.c    |    6 +-
.../net/ethernet/microchip/sparx5/sparx5_main.c    |  198 +-
.../net/ethernet/microchip/sparx5/sparx5_main.h    |  111 +-
.../ethernet/microchip/sparx5/sparx5_main_regs.h   | 4128 +++++++++++---------
.../net/ethernet/microchip/sparx5/sparx5_netdev.c  |    8 +-
.../net/ethernet/microchip/sparx5/sparx5_packet.c  |    4 +-
.../net/ethernet/microchip/sparx5/sparx5_pgid.c    |   24 +-
.../net/ethernet/microchip/sparx5/sparx5_police.c  |    3 +-
.../net/ethernet/microchip/sparx5/sparx5_port.c    |   71 +-
.../net/ethernet/microchip/sparx5/sparx5_port.h    |   23 +-
.../net/ethernet/microchip/sparx5/sparx5_psfp.c    |   45 +-
drivers/net/ethernet/microchip/sparx5/sparx5_ptp.c |    6 +-
drivers/net/ethernet/microchip/sparx5/sparx5_qos.c |    8 +-
drivers/net/ethernet/microchip/sparx5/sparx5_qos.h |    4 +-
.../net/ethernet/microchip/sparx5/sparx5_regs.c    |  219 ++
.../net/ethernet/microchip/sparx5/sparx5_regs.h    |  244 ++
.../net/ethernet/microchip/sparx5/sparx5_sdlb.c    |   15 +-
.../ethernet/microchip/sparx5/sparx5_switchdev.c   |   53 +-
drivers/net/ethernet/microchip/sparx5/sparx5_tc.c  |    8 +-
.../net/ethernet/microchip/sparx5/sparx5_vlan.c    |   44 +-
25 files changed, 3191 insertions(+), 2118 deletions(-)
[PATCH net-next 00/15] net: sparx5: prepare for lan969x switch driver
Posted by Daniel Machon 1 month, 3 weeks ago
== Description:

This series is the first of a multi-part series, that prepares and adds
support for the new lan969x switch driver.

The upstreaming efforts is split into multiple series (might change a
bit as we go along):

    1) Prepare the Sparx5 driver for lan969x (this series)
    2) Add support lan969x (same basic features as Sparx5 provides +
       RGMII, excl.  FDMA and VCAP)
    3) Add support for lan969x FDMA
    4) Add support for lan969x VCAP

== Lan969x in short:

The lan969x Ethernet switch family [1] provides a rich set of
switching features and port configurations (up to 30 ports) from 10Mbps
to 10Gbps, with support for RGMII, SGMII, QSGMII, USGMII, and USXGMII,
ideal for industrial & process automation infrastructure applications,
transport, grid automation, power substation automation, and ring &
intra-ring topologies. The LAN969x family is hardware and software
compatible and scalable supporting 46Gbps to 102Gbps switch bandwidths.

== Preparing Sparx5 for lan969x:

The lan969x switch chip reuses many of the IP's of the Sparx5 switch
chip, therefore it has been decided to add support through the existing
Sparx5 driver, in order to avoid a bunch of duplicate code. However, in
order to reuse the Sparx5 switch driver, we have to introduce some
mechanisms to handle the chip differences that are there.  These
mechanisms are:

    - Platform match data to contain all the differences that needs to
      be handled (constants, ops etc.)

    - Register macro indirection layer so that we can reuse the existing
      register macros.

    - Function for branching out on platform type where required.

In some places we ops out functions and in other places we branch on the
chip type. Exactly when we choose one over the other, is an estimate in
each case.

After this series is applied, the Sparx5 driver will be prepared for
lan969x and still function exactly as before.

== Patch breakdown:

Patch #1     adds private match data

Patch #2     adds register macro indirection layer

Patch #3-#5  does some preparation work

Patch #6-#8  adds chip constants and updates the code to use them

Patch #9-#14 adds and uses ops for handling functions differently on the
             two platforms.

Patch #15    adds and uses a macro for branching out on the chip type

[1] https://www.microchip.com/en-us/product/lan9698

To: David S. Miller <davem@davemloft.net>
To: Eric Dumazet <edumazet@google.com>
To: Jakub Kicinski <kuba@kernel.org>
To: Paolo Abeni <pabeni@redhat.com>
To: Lars Povlsen <lars.povlsen@microchip.com>
To: Steen Hegelund <Steen.Hegelund@microchip.com>
To: horatiu.vultur@microchip.com
To: jensemil.schulzostergaard@microchip.com
To: UNGLinuxDriver@microchip.com
To: Richard Cochran <richardcochran@gmail.com>
To: horms@kernel.org
To: justinstitt@google.com
To: gal@nvidia.com
To: aakash.r.menon@gmail.com
To: jacob.e.keller@intel.com
Cc: netdev@vger.kernel.org
Cc: linux-arm-kernel@lists.infradead.org
Cc: linux-kernel@vger.kernel.org

Signed-off-by: Daniel Machon <daniel.machon@microchip.com>
---
Daniel Machon (15):
      net: sparx5: add support for private match data
      net: sparx5: add indirection layer to register macros
      net: sparx5: rename *spx5 to *sparx5 in a few places
      net: sparx5: modify SPX5_PORTS_ALL macro
      net: sparx5: add *sparx5 argument to a few functions
      net: sparx5: add constants to match data
      net: sparx5: use SPX5_CONST for constants which already have a symbol
      net: sparx5: use SPX5_CONST for constants which do not have a symbol
      net: sparx5: add ops to match data
      net: sparx5: ops out chip port to device index/bit functions
      net: sparx5: ops out functions for getting certain array values
      net: sparx5: ops out function for setting the port mux
      net: sparx5: ops out PTP IRQ handler
      net: sparx5: ops out function for DSM calendar calculation
      net: sparx5: add is_sparx5 macro and use it throughout

 drivers/net/ethernet/microchip/sparx5/Makefile     |    2 +-
 .../ethernet/microchip/sparx5/sparx5_calendar.c    |   41 +-
 drivers/net/ethernet/microchip/sparx5/sparx5_dcb.c |    5 +-
 .../net/ethernet/microchip/sparx5/sparx5_ethtool.c |   33 +-
 .../net/ethernet/microchip/sparx5/sparx5_fdma.c    |    6 +-
 .../ethernet/microchip/sparx5/sparx5_mactable.c    |    6 +-
 .../net/ethernet/microchip/sparx5/sparx5_main.c    |  198 +-
 .../net/ethernet/microchip/sparx5/sparx5_main.h    |  111 +-
 .../ethernet/microchip/sparx5/sparx5_main_regs.h   | 4128 +++++++++++---------
 .../net/ethernet/microchip/sparx5/sparx5_netdev.c  |    8 +-
 .../net/ethernet/microchip/sparx5/sparx5_packet.c  |    4 +-
 .../net/ethernet/microchip/sparx5/sparx5_pgid.c    |   24 +-
 .../net/ethernet/microchip/sparx5/sparx5_police.c  |    3 +-
 .../net/ethernet/microchip/sparx5/sparx5_port.c    |   71 +-
 .../net/ethernet/microchip/sparx5/sparx5_port.h    |   23 +-
 .../net/ethernet/microchip/sparx5/sparx5_psfp.c    |   45 +-
 drivers/net/ethernet/microchip/sparx5/sparx5_ptp.c |    6 +-
 drivers/net/ethernet/microchip/sparx5/sparx5_qos.c |    8 +-
 drivers/net/ethernet/microchip/sparx5/sparx5_qos.h |    4 +-
 .../net/ethernet/microchip/sparx5/sparx5_regs.c    |  219 ++
 .../net/ethernet/microchip/sparx5/sparx5_regs.h    |  244 ++
 .../net/ethernet/microchip/sparx5/sparx5_sdlb.c    |   15 +-
 .../ethernet/microchip/sparx5/sparx5_switchdev.c   |   53 +-
 drivers/net/ethernet/microchip/sparx5/sparx5_tc.c  |    8 +-
 .../net/ethernet/microchip/sparx5/sparx5_vlan.c    |   44 +-
 25 files changed, 3191 insertions(+), 2118 deletions(-)
---
base-commit: 3a39d672e7f48b8d6b91a09afa4b55352773b4b5
change-id: 20240927-b4-sparx5-lan969x-switch-driver-dfcd5277fa70

Best regards,
-- 
Daniel Machon <daniel.machon@microchip.com>
Re: [PATCH net-next 00/15] net: sparx5: prepare for lan969x switch driver
Posted by Jacob Keller 1 month, 3 weeks ago

On 10/1/2024 6:50 AM, Daniel Machon wrote:
> == Description:
> 
> This series is the first of a multi-part series, that prepares and adds
> support for the new lan969x switch driver.
> 
> The upstreaming efforts is split into multiple series (might change a
> bit as we go along):
> 
>     1) Prepare the Sparx5 driver for lan969x (this series)
>     2) Add support lan969x (same basic features as Sparx5 provides +
>        RGMII, excl.  FDMA and VCAP)
>     3) Add support for lan969x FDMA
>     4) Add support for lan969x VCAP
> 
> == Lan969x in short:
> 
> The lan969x Ethernet switch family [1] provides a rich set of
> switching features and port configurations (up to 30 ports) from 10Mbps
> to 10Gbps, with support for RGMII, SGMII, QSGMII, USGMII, and USXGMII,
> ideal for industrial & process automation infrastructure applications,
> transport, grid automation, power substation automation, and ring &
> intra-ring topologies. The LAN969x family is hardware and software
> compatible and scalable supporting 46Gbps to 102Gbps switch bandwidths.
> 
> == Preparing Sparx5 for lan969x:
> 
> The lan969x switch chip reuses many of the IP's of the Sparx5 switch
> chip, therefore it has been decided to add support through the existing
> Sparx5 driver, in order to avoid a bunch of duplicate code. However, in
> order to reuse the Sparx5 switch driver, we have to introduce some
> mechanisms to handle the chip differences that are there.  These
> mechanisms are:
> 
>     - Platform match data to contain all the differences that needs to
>       be handled (constants, ops etc.)
> 
>     - Register macro indirection layer so that we can reuse the existing
>       register macros.
> 
>     - Function for branching out on platform type where required.
> 
> In some places we ops out functions and in other places we branch on the
> chip type. Exactly when we choose one over the other, is an estimate in
> each case.
> 
> After this series is applied, the Sparx5 driver will be prepared for
> lan969x and still function exactly as before.
> 
> == Patch breakdown:
> 
> Patch #1     adds private match data
> 
> Patch #2     adds register macro indirection layer
> 
> Patch #3-#5  does some preparation work
> 
> Patch #6-#8  adds chip constants and updates the code to use them
> 
> Patch #9-#14 adds and uses ops for handling functions differently on the
>              two platforms.
> 
> Patch #15    adds and uses a macro for branching out on the chip type
> 
> [1] https://www.microchip.com/en-us/product/lan9698
> 

The series seems ok to me. I'm not personally a fan of the implicit
local variables used by macros. I do not know how common that is, or
what others on the list feel about this.

For everything else:

Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
Re: [PATCH net-next 00/15] net: sparx5: prepare for lan969x switch driver
Posted by Daniel Machon 1 month, 3 weeks ago
> > == Description:
> >
> > This series is the first of a multi-part series, that prepares and adds
> > support for the new lan969x switch driver.
> >
> > The upstreaming efforts is split into multiple series (might change a
> > bit as we go along):
> >
> >     1) Prepare the Sparx5 driver for lan969x (this series)
> >     2) Add support lan969x (same basic features as Sparx5 provides +
> >        RGMII, excl.  FDMA and VCAP)
> >     3) Add support for lan969x FDMA
> >     4) Add support for lan969x VCAP
> >
> > == Lan969x in short:
> >
> > The lan969x Ethernet switch family [1] provides a rich set of
> > switching features and port configurations (up to 30 ports) from 10Mbps
> > to 10Gbps, with support for RGMII, SGMII, QSGMII, USGMII, and USXGMII,
> > ideal for industrial & process automation infrastructure applications,
> > transport, grid automation, power substation automation, and ring &
> > intra-ring topologies. The LAN969x family is hardware and software
> > compatible and scalable supporting 46Gbps to 102Gbps switch bandwidths.
> >
> > == Preparing Sparx5 for lan969x:
> >
> > The lan969x switch chip reuses many of the IP's of the Sparx5 switch
> > chip, therefore it has been decided to add support through the existing
> > Sparx5 driver, in order to avoid a bunch of duplicate code. However, in
> > order to reuse the Sparx5 switch driver, we have to introduce some
> > mechanisms to handle the chip differences that are there.  These
> > mechanisms are:
> >
> >     - Platform match data to contain all the differences that needs to
> >       be handled (constants, ops etc.)
> >
> >     - Register macro indirection layer so that we can reuse the existing
> >       register macros.
> >
> >     - Function for branching out on platform type where required.
> >
> > In some places we ops out functions and in other places we branch on the
> > chip type. Exactly when we choose one over the other, is an estimate in
> > each case.
> >
> > After this series is applied, the Sparx5 driver will be prepared for
> > lan969x and still function exactly as before.
> >
> > == Patch breakdown:
> >
> > Patch #1     adds private match data
> >
> > Patch #2     adds register macro indirection layer
> >
> > Patch #3-#5  does some preparation work
> >
> > Patch #6-#8  adds chip constants and updates the code to use them
> >
> > Patch #9-#14 adds and uses ops for handling functions differently on the
> >              two platforms.
> >
> > Patch #15    adds and uses a macro for branching out on the chip type
> >
> > [1] https://www.microchip.com/en-us/product/lan9698
> >
> 
> The series seems ok to me. I'm not personally a fan of the implicit
> local variables used by macros. I do not know how common that is, or
> what others on the list feel about this.
> 
> For everything else:
> 
> Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>

Hi Jakob,

First off, thank you for your reviews - I really appreciate it.

Let me address your "variable scope" conerns:

I had the feeling that this could pontentially be somewhat contentious.

Basically, this comes down to making the least invasive changes to the
existing driver code. With this approach:

    For the SPX5_CONST macro this means shorter lines, and less 80 char
    wrapping.

    For the "*regs" variable this means not having to pass in the sparx5
    pointer to *all* register macros, which requires changes *all* over
    the code.

I thought the solution with an in-scope implicit variable was less
invasive (and maybe even more readable?). Just my opinion, given the
alternative.

Obviously I did a bit of research on this upfront, and I can point to
*many* places where drivers do the exact same (not justifying the use,
just pointing that out). Here is an intel driver that does the same [1]
(look at the *hw variable)

I am willing to come up with something different, if you really think
this is a no-go. Let me hear your thoughts. I think this applies to your
comments on #2, #3 and #6 as well.

/Daniel

[1] https://elixir.bootlin.com/linux/v6.12-rc1/source/drivers/net/ethernet/intel/igb/igb_main.c#L4746
Re: [PATCH net-next 00/15] net: sparx5: prepare for lan969x switch driver
Posted by Jacob Keller 1 month, 3 weeks ago

On 10/2/2024 12:47 AM, Daniel Machon wrote:
> 
> Hi Jakob,
> 
> First off, thank you for your reviews - I really appreciate it.
> 
> Let me address your "variable scope" conerns:
> 
> I had the feeling that this could pontentially be somewhat contentious.
> 
> Basically, this comes down to making the least invasive changes to the
> existing driver code. With this approach:
> 
>     For the SPX5_CONST macro this means shorter lines, and less 80 char
>     wrapping.
> 
>     For the "*regs" variable this means not having to pass in the sparx5
>     pointer to *all* register macros, which requires changes *all* over
>     the code.
> 
> I thought the solution with an in-scope implicit variable was less
> invasive (and maybe even more readable?). Just my opinion, given the
> alternative.
>

Obviously there is style preference here, and someone working
day-in/day-out on the code is likely to know which macros have which
variable dependencies. As an external reviewer, I would not know that,
so I would find it surprising when looking at some code which passes a
parameter which is never directly used.

> Obviously I did a bit of research on this upfront, and I can point to
> *many* places where drivers do the exact same (not justifying the use,
> just pointing that out). Here is an intel driver that does the same [1]
> (look at the *hw variable)

Yea, I'm sure a lot of the old Intel drivers have bad examples :D I've
spent a career trying to improve this.

> 
> I am willing to come up with something different, if you really think
> this is a no-go. Let me hear your thoughts. I think this applies to your
> comments on #2, #3 and #6 as well.
> 

It seems that Jakub Kicinski wants the entire macro removed, and his
opinion as maintainer matters more than mine.

Personally, I'm not opposed to a macro itself especially if the direct
access starts to get very long. However, I think the parameter being
accessed should be, well, a parameter of the macro.

> /Daniel
> 
> [1] https://elixir.bootlin.com/linux/v6.12-rc1/source/drivers/net/ethernet/intel/igb/igb_main.c#L4746
> 
> 

As an example of *why* I don't like this practice: It took me a while to
realize the hw variable was implicit to wr32. And I worked on this driver.

Thanks,
Jake