Re: [PATCH v3 0/11] Add R8A7743/SK-RZG1M board support

[Geert Uytterhoeven]

Hi Simon,

On Thu, Oct 20, 2016 at 11:24 AM, Simon Horman  wrote:
> On Wed, Oct 12, 2016 at 11:29:56AM +0300, Sergei Shtylyov wrote:
>> On 10/12/2016 11:09 AM, Simon Horman wrote:
>> >>>  Here's the set of 11 patches against Simon Horman's 'renesas.git' repo's
>> >>>'renesas-devel-20161003-v4.8' tag. I'm adding the device tree support for
>> >>>the R8A7743-based SK-RZG1M board. The SoC is close to R8A7791 and the 
>> >>>board
>> >>>seems identical to the R8A7791/Porter board. The device tree patches 
>> >>>depend on
>> >>>the R8A7743 CPG/MSSR driver series just posted in order to compile and 
>> >>>work.
>> >>
>> >>   Forgot to mention that this version causes a regression with the sh_eth
>> >>driver (well, actually with phylib): since IRQC now gets a deferred 
>> >>probing,
>> >>PHY IRQ doesn't work anymore -- phylib falls back to polling.
>> >
>> >Is there a resolution to that problem?
>>
>>Geert has posted his IRQC driver patch recently. Not sure if it was
>> intended for merging but it solves the issue.
>
> Ok. I think it is ok to merge support for a new board even if there are
> problems as its not regressing (can't break something that previously
> didn't exist imho). But of course it would be best if things did work.
>
> Geert, do you have any comments on the relevance of your IRQC driver patch?

That patch is relevant, until the MDIO framework is fixed.
It will start to matter more when migrating more (R-Car Gen2) SoCs to the new
CPG/MSSR driver.

It would be good to give that patch some testing on more R-Car Gen2 boards
with NFS root (difficult for me with remote access and network latencies).

After that I can submit it to the irqchip maintainers, as I don't believe MDIO
will be fixed soonish...

Thanks for testing ;-)

Gr{oetje,eeting}s,

Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- ge...@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
-- Linus Torvalds

Re: [PATCH v3 0/11] Add R8A7743/SK-RZG1M board support

[Simon Horman]

On Wed, Oct 12, 2016 at 11:29:56AM +0300, Sergei Shtylyov wrote:
> On 10/12/2016 11:09 AM, Simon Horman wrote:
> 
> >>>  Here's the set of 11 patches against Simon Horman's 'renesas.git' repo's
> >>>'renesas-devel-20161003-v4.8' tag. I'm adding the device tree support for
> >>>the R8A7743-based SK-RZG1M board. The SoC is close to R8A7791 and the board
> >>>seems identical to the R8A7791/Porter board. The device tree patches 
> >>>depend on
> >>>the R8A7743 CPG/MSSR driver series just posted in order to compile and 
> >>>work.
> >>
> >>   Forgot to mention that this version causes a regression with the sh_eth
> >>driver (well, actually with phylib): since IRQC now gets a deferred probing,
> >>PHY IRQ doesn't work anymore -- phylib falls back to polling.
> >
> >Is there a resolution to that problem?
> 
>Geert has posted his IRQC driver patch recently. Not sure if it was
> intended for merging but it solves the issue.

Ok. I think it is ok to merge support for a new board even if there are
problems as its not regressing (can't break something that previously
didn't exist imho). But of course it would be best if things did work.

Geert, do you have any comments on the relevance of your IRQC driver patch?

Re: [PATCH v3 0/11] Add R8A7743/SK-RZG1M board support

[Sergei Shtylyov]

On 10/12/2016 11:09 AM, Simon Horman wrote:


  Here's the set of 11 patches against Simon Horman's 'renesas.git' repo's
'renesas-devel-20161003-v4.8' tag. I'm adding the device tree support for
the R8A7743-based SK-RZG1M board. The SoC is close to R8A7791 and the board
seems identical to the R8A7791/Porter board. The device tree patches depend on
the R8A7743 CPG/MSSR driver series just posted in order to compile and work.


   Forgot to mention that this version causes a regression with the sh_eth
driver (well, actually with phylib): since IRQC now gets a deferred probing,
PHY IRQ doesn't work anymore -- phylib falls back to polling.


Is there a resolution to that problem?


   Geert has posted his IRQC driver patch recently. Not sure if it was 
intended for merging but it solves the issue.


MBR, Sergei

Re: [PATCH v3 0/11] Add R8A7743/SK-RZG1M board support

[Sergei Shtylyov]

On 10/06/2016 12:23 AM, Sergei Shtylyov wrote:


   Here's the set of 11 patches against Simon Horman's 'renesas.git' repo's
'renesas-devel-20161003-v4.8' tag. I'm adding the device tree support for
the R8A7743-based SK-RZG1M board. The SoC is close to R8A7791 and the board
seems identical to the R8A7791/Porter board. The device tree patches depend on
the R8A7743 CPG/MSSR driver series just posted in order to compile and work.


   Forgot to mention that this version causes a regression with the sh_eth 
driver (well, actually with phylib): since IRQC now gets a deferred probing, 
PHY IRQ doesn't work anymore -- phylib falls back to polling.


MBR, Sergei