On Mon, Jun 04, 2018 at 05:22:52PM +0300, Sergei Shtylyov wrote:
> On 06/04/2018 01:33 PM, Simon Horman wrote:
>
> >> Specify EtherAVB PHY IRQ in the Condor board's device tree, now that
> >> we have the GPIO support (previously phylib had to resort to polling).
> >>
> >> Based on the original (and large) patch by Vladimir Barinov.
> >>
> >> Signed-off-by: Vladimir Barinov <vladimir.barinov@xxxxxxxxxxxxxxxxxx>
> >> Signed-off-by: Sergei Shtylyov <sergei.shtylyov@xxxxxxxxxxxxxxxxxx>
> >>
> >> ---
> >> arch/arm64/boot/dts/renesas/r8a77980-condor.dts | 2 ++
> >> 1 file changed, 2 insertions(+)
> >>
> >> Index: renesas/arch/arm64/boot/dts/renesas/r8a77980-condor.dts
> >> ===================================================================
> >> --- renesas.orig/arch/arm64/boot/dts/renesas/r8a77980-condor.dts
> >> +++ renesas/arch/arm64/boot/dts/renesas/r8a77980-condor.dts
> >> @@ -59,6 +59,8 @@
> >> phy0: ethernet-phy@0 {
> >> rxc-skew-ps = <1500>;
> >> reg = <0>;
> >> + interrupt-parent = <&gpio1>;
> >> + interrupts = <17 IRQ_TYPE_LEVEL_LOW>;
> >
> > I don't see this documented. Perhaps I'm missing something obvious.
>
> Have you looked into the V3H PFC section for where in the GPSRs AVB_PHY_INT
> is mapped?
Thanks, I see that now.
> The Condor schematics doesn't explicitly list the GPIO for AVB_PHY_INT
> because that signal is meant to be routed thru the MAC. Unfortunately, the
> sh_eth/ravb drivers don't support the PHY interrupt (the phylib function,
> phy_mac_interrupt() reporting the PHY interrupts routed thru MAC is clearly
> inadequate as it wants the link state as an argument), so we have to resort
> to the GPIO interrupts...
Understood.
> > Or you have some extra information or newer documentation?
>
> No.
>
> > Also, given Olof Johansson's recent comments in ("Re: [GIT PULL] Renesas
> > ARM64 Based SoC DT Updates for v4.18") please consider squashing this patch
> > and the following one.
>
> Hm... note that the different Ether cores are involved in these 2 PHY IRQ
> patches. If that's OK, I can merge the patches...
Tough call. Functionally these are both ethernet even though they are
different IP cores. So I think I prefer a squash.
I have applied 1/3 of this series and will push shortly.
