Re: [PATCH] ARM: mach-imx6q: add ksz9131rn_phy_fixup

[Oleksij Rempel <o.rempel@pengutronix.de>]

[-- Attachment #1: Type: text/plain, Size: 5914 bytes --]

Hi Philippe,

On Fri, Mar 06, 2020 at 09:55:06AM +0000, Philippe Schenker wrote:
> On Thu, 2020-03-05 at 15:38 +0100, Oleksij Rempel wrote:
> > Hi Philippe,
> > 
> > On Thu, Mar 05, 2020 at 02:49:28PM +0100, Philippe Schenker wrote:
> > > The MAC of the i.MX6 SoC is compliant with RGMII v1.3. The KSZ9131
> > > PHY
> > > is like KSZ9031 adhering to RGMII v2.0 specification. This means the
> > > MAC should provide a delay to the TXC line. Because the i.MX6 MAC
> > > does
> > > not provide this delay this has to be done in the PHY.
> > > 
> > > This patch adds by default ~1.6ns delay to the TXC line. This should
> > > be good for all boards that have the RGMII signals routed with the
> > > same length.
> > > 
> > > The KSZ9131 has relatively high tolerances on skew registers from
> > > MMD 2.4 to MMD 2.8. Therefore the new DLL-based delay of 2ns is used
> > > and then as little as possibly subtracted from that so we get more
> > > accurate delay. This is actually needed because the i.MX6 SoC has
> > > an asynchron skew on TXC from -100ps to 900ps, to get all RGMII
> > > values within spec.
> 
> Hi Oleksij! Thanks for your comments and review.
> > 
> > This configuration has nothing to do in mach-imx/* It belongs to the
> > board devicetree. Please see DT binding documentation for needed
> > properties:
> > Documentation/devicetree/bindings/net/micrel-ksz90x1.txt
> 
> I know that nowadays this stuff only belongs in the devicetree. I fully
> agree with you. I am also aware of the devicetree bindings.
> > 
> > All of this mach-imx fixups are evil and should be removed or disabled
> > by Kconfig
> > option. Since they will run on all i.MX based boards even if this PHY
> > are
> > connected to some switch and not connected to the FEC directly.
> > And.. If driver didn't made this configuration all this changes will
> > be lost on
> > suspend/resume cycle or on PHY reset.
> 
> I am also aware of this behaviour.

... ò_ô ...

> But the i.MX6 is a SoC used in
> embedded applications and I guess no one comes and plugs in a PCIe MAC
> card in an embedded device.

... hm ...

> But yes you're right you never know.

well, it is not theoretical discussion. This devices do exist.. With
this patch you will break other existing systems.

> Because the i.MX6 is an embedded processor I still think we should place
> that fixup in mach-imx. There is already a fixup for the predecessor
> KSZ9031 in that code. The KSZ9131 is pin-compatible with KSZ9031 and
> also software compatible, just not with the skew settings.

This fixups will be removed or disabled with Kconfig option:
https://lore.kernel.org/patchwork/patch/1164172/

> I really dislike reinventing the weel here for an old SoC.

Well, you are doing it not for a SoC (old or new), you are doing it for
PHY. PHY fixes belong to PHY driver.

> Philippe
> > 
> > Regards,
> > Oleksij
> > 
> > > Signed-off-by: Philippe Schenker <philippe.schenker@toradex.com>
> > > 
> > > ---
> > > 
> > >  arch/arm/mach-imx/mach-imx6q.c | 37
> > > ++++++++++++++++++++++++++++++++++
> > >  1 file changed, 37 insertions(+)
> > > 
> > > diff --git a/arch/arm/mach-imx/mach-imx6q.c b/arch/arm/mach-
> > > imx/mach-imx6q.c
> > > index edd26e0ffeec..8ae5f2fa33e2 100644
> > > --- a/arch/arm/mach-imx/mach-imx6q.c
> > > +++ b/arch/arm/mach-imx/mach-imx6q.c
> > > @@ -61,6 +61,14 @@ static void mmd_write_reg(struct phy_device *dev,
> > > int device, int reg, int val)
> > >  	phy_write(dev, 0x0e, val);
> > >  }
> > >  
> > > +static int mmd_read_reg(struct phy_device *dev, int device, int
> > > reg)
> > > +{
> > > +	phy_write(dev, 0x0d, device);
> > > +	phy_write(dev, 0x0e, reg);
> > > +	phy_write(dev, 0x0d, (1 << 14) | device);
> > > +	return phy_read(dev, 0x0e);
> > > +}
> > > +
> > >  static int ksz9031rn_phy_fixup(struct phy_device *dev)
> > >  {
> > >  	/*
> > > @@ -74,6 +82,33 @@ static int ksz9031rn_phy_fixup(struct phy_device
> > > *dev)
> > >  	return 0;
> > >  }
> > >  
> > > +#define KSZ9131_RXTXDLL_BYPASS	12
> > > +
> > > +static int ksz9131rn_phy_fixup(struct phy_device *dev)
> > > +{
> > > +	int tmp;
> > > +
> > > +	tmp = mmd_read_reg(dev, 2, 0x4c);
> > > +	/* disable rxdll bypass (enable 2ns skew delay on RXC) */
> > > +	tmp &= ~(1 << KSZ9131_RXTXDLL_BYPASS);
> > > +	mmd_write_reg(dev, 2, 0x4c, tmp);
> > > +
> > > +	tmp = mmd_read_reg(dev, 2, 0x4d);
> > > +	/* disable txdll bypass (enable 2ns skew delay on TXC) */
> > > +	tmp &= ~(1 << KSZ9131_RXTXDLL_BYPASS);
> > > +	mmd_write_reg(dev, 2, 0x4d, tmp);
> > > +
> > > +	/*
> > > +	 * Subtract ~0.6ns from txdll = ~1.4ns delay.
> > > +	 * leave RXC path untouched
> > > +	 */
> > > +	mmd_write_reg(dev, 2, 4, 0x007d);
> > > +	mmd_write_reg(dev, 2, 6, 0xdddd);
> > > +	mmd_write_reg(dev, 2, 8, 0x0007);
> > > +
> > > +	return 0;
> > > +}
> > > +
> > >  /*
> > >   * fixup for PLX PEX8909 bridge to configure GPIO1-7 as output High
> > >   * as they are used for slots1-7 PERST#
> > > @@ -167,6 +202,8 @@ static void __init imx6q_enet_phy_init(void)
> > >  				ksz9021rn_phy_fixup);
> > >  		phy_register_fixup_for_uid(PHY_ID_KSZ9031,
> > > MICREL_PHY_ID_MASK,
> > >  				ksz9031rn_phy_fixup);
> > > +		phy_register_fixup_for_uid(PHY_ID_KSZ9131,
> > > MICREL_PHY_ID_MASK,
> > > +				ksz9131rn_phy_fixup);
> > >  		phy_register_fixup_for_uid(PHY_ID_AR8031, 0xffffffef,
> > >  				ar8031_phy_fixup);
> > >  		phy_register_fixup_for_uid(PHY_ID_AR8035, 0xffffffef,
> > > -- 
> > > 2.25.1
> > > 
> > > 
> > > 

-- 
Pengutronix e.K.                           |                             |
Steuerwalder Str. 21                       | http://www.pengutronix.de/  |
31137 Hildesheim, Germany                  | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]