> Date: Tue, 27 Feb 2018 18:47:42 +1100
> From: Jonathan Gray <j...@jsg.id.au>
>
> On Mon, Feb 26, 2018 at 10:08:12PM +0100, Mark Kettenis wrote:
> > There is a new rgepy(4) variant (rev. 6, RTL8211F). Unfortunately
> > they moved the status register around on this new chip and also
> > twiddled the bits. Here is a diff that handles that. Somewhat
> > inspired by FreeBSD but I didn't add some of the workarounds tat
> > disable EEE and power save mode that are found in their rgephy(4). So
> > far it seems I don't really need those for my Rock64 board.
> >
> > ok?
> >
> >
> > Index: dev/mii/rgephy.c
> > ===================================================================
> > RCS file: /cvs/src/sys/dev/mii/rgephy.c,v
> > retrieving revision 1.39
> > diff -u -p -r1.39 rgephy.c
> > --- dev/mii/rgephy.c 11 Mar 2017 13:40:46 -0000 1.39
> > +++ dev/mii/rgephy.c 26 Feb 2018 20:59:48 -0000
> > @@ -255,6 +255,11 @@ setit:
> > sc->mii_ticks = 0;
> > break;
> > }
> > + } else if (sc->mii_rev == RGEPHY_8211F) {
> > + reg = PHY_READ(sc, RGEPHY_F_SR);
> > + if (reg & RGEPHY_F_SR_LINK) {
> > + sc->mii_ticks = 0;
> > + }
> > } else {
> > reg = PHY_READ(sc, RGEPHY_SR);
> > if (reg & RGEPHY_SR_LINK) {
> > @@ -309,6 +314,10 @@ rgephy_status(struct mii_softc *sc)
> > bmsr = PHY_READ(sc, RL_GMEDIASTAT);
> > if (bmsr & RL_GMEDIASTAT_LINK)
> > mii->mii_media_status |= IFM_ACTIVE;
> > + } else if (sc->mii_rev == RGEPHY_8211F) {
> > + bmsr = PHY_READ(sc, RGEPHY_F_SR);
> > + if (bmsr & RGEPHY_F_SR_LINK)
> > + mii->mii_media_status |= IFM_ACTIVE;
> > } else {
> > bmsr = PHY_READ(sc, RGEPHY_SR);
> > if (bmsr & RGEPHY_SR_LINK)
> > @@ -340,6 +349,20 @@ rgephy_status(struct mii_softc *sc)
> > mii->mii_media_active |= IFM_10_T;
> >
> > if (bmsr & RL_GMEDIASTAT_FDX)
> > + mii->mii_media_active |= mii_phy_flowstatus(sc) |
> > + IFM_FDX;
> > + else
> > + mii->mii_media_active |= IFM_HDX;
> > + } else if (sc->mii_rev == RGEPHY_8211F) {
> > + bmsr = PHY_READ(sc, RGEPHY_F_SR);
> > + if (RGEPHY_F_SR_SPEED(bmsr) == RGEPHY_F_SR_SPEED_1000MBPS)
> > + mii->mii_media_active |= IFM_1000_T;
> > + else if (RGEPHY_F_SR_SPEED(bmsr) == RGEPHY_F_SR_SPEED_100MBPS)
> > + mii->mii_media_active |= IFM_100_TX;
> > + else if (RGEPHY_F_SR_SPEED(bmsr) == RGEPHY_F_SR_SPEED_10MBPS)
> > + mii->mii_media_active |= IFM_10_T;
> > +
> > + if (bmsr & RGEPHY_F_SR_FDX)
> > mii->mii_media_active |= mii_phy_flowstatus(sc) |
> > IFM_FDX;
> > else
> > Index: dev/mii/rgephyreg.h
> > ===================================================================
> > RCS file: /cvs/src/sys/dev/mii/rgephyreg.h,v
> > retrieving revision 1.8
> > diff -u -p -r1.8 rgephyreg.h
> > --- dev/mii/rgephyreg.h 19 Jul 2015 06:35:18 -0000 1.8
> > +++ dev/mii/rgephyreg.h 26 Feb 2018 20:59:48 -0000
> > @@ -36,6 +36,10 @@
> > #ifndef _DEV_MII_RGEPHYREG_H_
> > #define _DEV_MII_RGEPHYREG_H_
> >
> > +#define RGEPHY_8211B 2
> > +#define RGEPHY_8211C 3
> > +#define RGEPHY_8211F 6
> > +
> > /*
> > * Realtek 8169S/8110S gigE PHY registers
> > */
> > @@ -52,6 +56,7 @@
> > #define RGEPHY_CR_ALDPS 0x0004 /* RTL8251 only */
> > #define RGEPHY_CR_JABBER_DIS 0x0001
> >
> > +/* RTL8211B(L)/RTL8211C(L) */
> > #define RGEPHY_SR 0x11 /* PHY Specific Status */
> > #define RGEPHY_SR_SPEED_1000MBPS 0x8000
> > #define RGEPHY_SR_SPEED_100MBPS 0x4000
> > @@ -65,6 +70,16 @@
> > #define RGEPHY_SR_ALDPS 0x0008 /* RTL8211C(L) only */
> > #define RGEPHY_SR_JABBER 0x0001 /* Jabber */
> > #define RGEPHY_SR_SPEED(X) ((X) & RGEPHY_SR_SPEED_MASK)
> > +
> > +/* RTL8211F */
> > +#define RGEPHY_F_SR 0x1A /* PHY Specific Status */
> > +#define RGEPHY_F_SR_SPEED_1000MBPS 0x0020
> > +#define RGEPHY_F_SR_SPEED_100MBPS 0x0010
> > +#define RGEPHY_F_SR_SPEED_10MBPS 0x0000
> > +#define RGEPHY_F_SR_SPEED_MASK 0x0030
>
> > +#define RGEPHY_F_SR_FDX 0x0004
> > +#define RGEPHY_F_SR_LINK 0x0002
>
> Are these purposefully different from FreeBSD? They have
>
> #define RGEPHY_F_SSR_FDX 0x0008 /* full duplex */
> #define RGEPHY_F_SSR_LINK 0x0004 /* link up */
> #define RGEPHY_F_SSR_MDI 0x0002 /* MDI/MDIX */
> #define RGEPHY_F_SSR_JABBER 0x0001 /* Jabber */
No. My values are wrong. Well spotted.
I'll test with the fixed values tonight. Ok with that change
(assuming it works)?
