[U-Boot,11/11] powerpc, 8xx: Add a directory for soon to come boards from CSSI
diff mbox

Message ID d4588f31989ba31ebad611161a61c9f1771de2d5.1497591423.git.christophe.leroy@c-s.fr
State Changes Requested
Delegated to: Tom Rini
Headers show

Commit Message

Christophe Leroy June 21, 2017, 3:09 p.m. UTC
CS Systemes d'Information (CSSI) manufactures two boards, named MCR3000
and CMPC885 which are respectively based on MPC866 and MPC885.
This patch is the skeleton to allow soon inclusion of support for
those two boards.

Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
---
 The source code for those two boards needs consequent cleanup
before being suitable for inclusion, however we need to stop
8xx deletion hemoragia as soon as possible.

 board/cssi/MAINTAINERS | 5 +++++
 1 file changed, 5 insertions(+)
 create mode 100644 board/cssi/MAINTAINERS

Comments

Tom Rini June 21, 2017, 3:22 p.m. UTC | #1
On Wed, Jun 21, 2017 at 05:09:09PM +0200, Christophe Leroy wrote:

> CS Systemes d'Information (CSSI) manufactures two boards, named MCR3000
> and CMPC885 which are respectively based on MPC866 and MPC885.
> This patch is the skeleton to allow soon inclusion of support for
> those two boards.
> 
> Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
> ---
>  The source code for those two boards needs consequent cleanup
> before being suitable for inclusion, however we need to stop
> 8xx deletion hemoragia as soon as possible.

With what I'm going to send out as a v2 of the rest of the series, I'm
OK with not including an empty directory for now and bringing back
mpc8xx, thanks again!
Heiko Schocher June 21, 2017, 3:51 p.m. UTC | #2
Hello Christophe,

Am 21.06.2017 um 17:22 schrieb Tom Rini:
> On Wed, Jun 21, 2017 at 05:09:09PM +0200, Christophe Leroy wrote:
>
>> CS Systemes d'Information (CSSI) manufactures two boards, named MCR3000
>> and CMPC885 which are respectively based on MPC866 and MPC885.
>> This patch is the skeleton to allow soon inclusion of support for
>> those two boards.
>>
>> Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
>> ---
>>   The source code for those two boards needs consequent cleanup
>> before being suitable for inclusion, however we need to stop
>> 8xx deletion hemoragia as soon as possible.
>
> With what I'm going to send out as a v2 of the rest of the series, I'm
> OK with not including an empty directory for now and bringing back
> mpc8xx, thanks again!

Hmm.. do we really want back the old files?

Just a fast look in the patchstatistic:

  arch/powerpc/cpu/mpc8xx/spi.c          |  533 ++++++++++++

No, spi drivers are in drivers/spi nowadays

  arch/powerpc/cpu/mpc8xx/scc.c          |  472 +++++++++++
  arch/powerpc/cpu/mpc8xx/fec.c          |  933 +++++++++++++++++++++
  arch/powerpc/cpu/mpc8xx/fec.h          |   12 +

-> driver/net

  arch/powerpc/cpu/mpc8xx/serial.c       |  676 ++++++++++++++++

-> drivers/serial

Only fast look into:
arch/powerpc/cpu/mpc8xx/spi.c

         cp->cp_pbpar |=  0x0000000E;    /* set  bits    */

No use of ioaccesors, no defines for magic values ...:-(

DM support? Cleanup for Kconfig support?

One reason of the remove was, that the mpc8xx files wasn;t in a good
shape, and nobody volunteered for cleaning it up ...

Are there plans to clean this up ?

If no: NACK from me, sorry.

Why not posting a patchset, which introduce mpc8xx support only?

bye,
Heiko
Tom Rini June 21, 2017, 4:43 p.m. UTC | #3
On Wed, Jun 21, 2017 at 05:51:35PM +0200, Heiko Schocher wrote:
> Hello Christophe,
> 
> Am 21.06.2017 um 17:22 schrieb Tom Rini:
> >On Wed, Jun 21, 2017 at 05:09:09PM +0200, Christophe Leroy wrote:
> >
> >>CS Systemes d'Information (CSSI) manufactures two boards, named MCR3000
> >>and CMPC885 which are respectively based on MPC866 and MPC885.
> >>This patch is the skeleton to allow soon inclusion of support for
> >>those two boards.
> >>
> >>Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
> >>---
> >>  The source code for those two boards needs consequent cleanup
> >>before being suitable for inclusion, however we need to stop
> >>8xx deletion hemoragia as soon as possible.
> >
> >With what I'm going to send out as a v2 of the rest of the series, I'm
> >OK with not including an empty directory for now and bringing back
> >mpc8xx, thanks again!
> 
> Hmm.. do we really want back the old files?

Yes, we do.  Revert, and then improve.  We will not let mpc8xx fall
behind, but we need to bring things back to start with.

> Just a fast look in the patchstatistic:
> 
>  arch/powerpc/cpu/mpc8xx/spi.c          |  533 ++++++++++++
> 
> No, spi drivers are in drivers/spi nowadays
> 
>  arch/powerpc/cpu/mpc8xx/scc.c          |  472 +++++++++++
>  arch/powerpc/cpu/mpc8xx/fec.c          |  933 +++++++++++++++++++++
>  arch/powerpc/cpu/mpc8xx/fec.h          |   12 +
> 
> -> driver/net
> 
>  arch/powerpc/cpu/mpc8xx/serial.c       |  676 ++++++++++++++++
> 
> -> drivers/serial
> 
> Only fast look into:
> arch/powerpc/cpu/mpc8xx/spi.c
> 
>         cp->cp_pbpar |=  0x0000000E;    /* set  bits    */
> 
> No use of ioaccesors, no defines for magic values ...:-(
> 
> DM support? Cleanup for Kconfig support?
> 
> One reason of the remove was, that the mpc8xx files wasn;t in a good
> shape, and nobody volunteered for cleaning it up ...
> 
> Are there plans to clean this up ?
> 
> If no: NACK from me, sorry.

Yes, agreed, DM conversion, Kconfig work, and otherwise cleaning things
up to be as clean as newer architecture work is, is required to keep
mpc8xx in.

> Why not posting a patchset, which introduce mpc8xx support only?

This, fwiw, is what I sent out as v2, but needs to be v3'd.
Christophe Leroy June 21, 2017, 4:47 p.m. UTC | #4
Le 21/06/2017 à 17:51, Heiko Schocher a écrit :
> Hello Christophe,
> 
> Am 21.06.2017 um 17:22 schrieb Tom Rini:
>> On Wed, Jun 21, 2017 at 05:09:09PM +0200, Christophe Leroy wrote:
>>
>>> CS Systemes d'Information (CSSI) manufactures two boards, named MCR3000
>>> and CMPC885 which are respectively based on MPC866 and MPC885.
>>> This patch is the skeleton to allow soon inclusion of support for
>>> those two boards.
>>>
>>> Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
>>> ---
>>>   The source code for those two boards needs consequent cleanup
>>> before being suitable for inclusion, however we need to stop
>>> 8xx deletion hemoragia as soon as possible.
>>
>> With what I'm going to send out as a v2 of the rest of the series, I'm
>> OK with not including an empty directory for now and bringing back
>> mpc8xx, thanks again!
> 
> Hmm.. do we really want back the old files?

Yes, I don't like the idea of a big-bang. Lets clean it up step by step, 
as done for other targets.

> 
> Just a fast look in the patchstatistic:
> 
>   arch/powerpc/cpu/mpc8xx/spi.c          |  533 ++++++++++++
> 
> No, spi drivers are in drivers/spi nowadays
> 
>   arch/powerpc/cpu/mpc8xx/scc.c          |  472 +++++++++++
>   arch/powerpc/cpu/mpc8xx/fec.c          |  933 +++++++++++++++++++++
>   arch/powerpc/cpu/mpc8xx/fec.h          |   12 +
> 
> -> driver/net
> 
>   arch/powerpc/cpu/mpc8xx/serial.c       |  676 ++++++++++++++++
> 
> -> drivers/serial
> 
> Only fast look into:
> arch/powerpc/cpu/mpc8xx/spi.c
> 
>          cp->cp_pbpar |=  0x0000000E;    /* set  bits    */
> 
> No use of ioaccesors, no defines for magic values ...:-(
> 
> DM support? Cleanup for Kconfig support?
> 
> One reason of the remove was, that the mpc8xx files wasn;t in a good
> shape, and nobody volunteered for cleaning it up ...

I do volunteer for cleaning it up .... but definitly not for re-starting 
from scratch.

> 
> Are there plans to clean this up ?

I will do it, but I'd like to start from an existing situation, not 
start from scratch.

> 
> If no: NACK from me, sorry.
> 
> Why not posting a patchset, which introduce mpc8xx support only?

We have a basis that is maybe old, but works. For me it is a better 
starting point.

Christophe
Heiko Schocher June 22, 2017, 3:43 a.m. UTC | #5
Hello Christophe,

Am 21.06.2017 um 18:47 schrieb Christophe LEROY:
>
>
> Le 21/06/2017 à 17:51, Heiko Schocher a écrit :
>> Hello Christophe,
>>
>> Am 21.06.2017 um 17:22 schrieb Tom Rini:
>>> On Wed, Jun 21, 2017 at 05:09:09PM +0200, Christophe Leroy wrote:
>>>
>>>> CS Systemes d'Information (CSSI) manufactures two boards, named MCR3000
>>>> and CMPC885 which are respectively based on MPC866 and MPC885.
>>>> This patch is the skeleton to allow soon inclusion of support for
>>>> those two boards.
>>>>
>>>> Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
>>>> ---
>>>>   The source code for those two boards needs consequent cleanup
>>>> before being suitable for inclusion, however we need to stop
>>>> 8xx deletion hemoragia as soon as possible.
>>>
>>> With what I'm going to send out as a v2 of the rest of the series, I'm
>>> OK with not including an empty directory for now and bringing back
>>> mpc8xx, thanks again!
>>
>> Hmm.. do we really want back the old files?
>
> Yes, I don't like the idea of a big-bang. Lets clean it up step by step, as done for other targets.
>
>>
>> Just a fast look in the patchstatistic:
>>
>>   arch/powerpc/cpu/mpc8xx/spi.c          |  533 ++++++++++++
>>
>> No, spi drivers are in drivers/spi nowadays
>>
>>   arch/powerpc/cpu/mpc8xx/scc.c          |  472 +++++++++++
>>   arch/powerpc/cpu/mpc8xx/fec.c          |  933 +++++++++++++++++++++
>>   arch/powerpc/cpu/mpc8xx/fec.h          |   12 +
>>
>> -> driver/net
>>
>>   arch/powerpc/cpu/mpc8xx/serial.c       |  676 ++++++++++++++++
>>
>> -> drivers/serial
>>
>> Only fast look into:
>> arch/powerpc/cpu/mpc8xx/spi.c
>>
>>          cp->cp_pbpar |=  0x0000000E;    /* set  bits    */
>>
>> No use of ioaccesors, no defines for magic values ...:-(
>>
>> DM support? Cleanup for Kconfig support?
>>
>> One reason of the remove was, that the mpc8xx files wasn;t in a good
>> shape, and nobody volunteered for cleaning it up ...
>
> I do volunteer for cleaning it up .... but definitly not for re-starting from scratch.

Ok, thanks!

>> Are there plans to clean this up ?
>
> I will do it, but I'd like to start from an existing situation, not start from scratch.

Hmm... starting here from scratch is may better here, but thats just
just my opinion ...

>> If no: NACK from me, sorry.
>>
>> Why not posting a patchset, which introduce mpc8xx support only?
>
> We have a basis that is maybe old, but works. For me it is a better starting point.

Understand ... but a simple patch which introduces back the mpc8xx files
only, with some basic changes, like I mentioned would be cleaner.

And don;t forget to add you as mpc8xx maintainer ;-)

bye,
Heiko
Wolfgang Denk June 22, 2017, 7:20 a.m. UTC | #6
Dear Tom,

In message <20170621152200.GJ27196@bill-the-cat> you wrote:
> 
> With what I'm going to send out as a v2 of the rest of the series, I'm
> OK with not including an empty directory for now and bringing back
> mpc8xx, thanks again!

Sorry, but I disagree here.

This revert series is a mess.  It does not simply re-add 8xx
support, it reverts everything and then re-removes everything except
8xx.  This is totally unnecessary and ugly.  Please do not accept
these patches.


Also, as Heiko already mentioned, if we re-add 8xx support, this
should be done in a way that matches today's requirements. The code
needs cleanup everywhere: directory layout, Kconfig, DM support, I/O
accessors, etc.

NAK!




Best regards,

Wolfgang Denk
Christophe Leroy June 22, 2017, 8:39 a.m. UTC | #7
Le 22/06/2017 à 09:20, Wolfgang Denk a écrit :
> Dear Tom,
> 
> In message <20170621152200.GJ27196@bill-the-cat> you wrote:
>>
>> With what I'm going to send out as a v2 of the rest of the series, I'm
>> OK with not including an empty directory for now and bringing back
>> mpc8xx, thanks again!
> 
> Sorry, but I disagree here.
> 
> This revert series is a mess.  It does not simply re-add 8xx
> support, it reverts everything and then re-removes everything except
> 8xx.  This is totally unnecessary and ugly.  Please do not accept
> these patches.
> 
> 
> Also, as Heiko already mentioned, if we re-add 8xx support, this
> should be done in a way that matches today's requirements. The code
> needs cleanup everywhere: directory layout, Kconfig, DM support, I/O
> accessors, etc.

Yes I have accepted this task, but please don't make it more complicated 
than it is already. I need to start from an existing situation and go 
step by step just like it has been done with other targets.

Christophe


> 
> NAK!
> 
> 
> 
> 
> Best regards,
> 
> Wolfgang Denk
>

Patch
diff mbox

diff --git a/board/cssi/MAINTAINERS b/board/cssi/MAINTAINERS
new file mode 100644
index 0000000000..2c959de383
--- /dev/null
+++ b/board/cssi/MAINTAINERS
@@ -0,0 +1,5 @@ 
+BOARDS from CS Systemes d'Information
+M:	Christophe Leroy <christophe.leroy@c-s.fr>
+S:	Maintained
+F:	board/cssi/
+