allwinner question: New CLR4(sc->phy_ctrl, OTG_PHY_CFG, OTG_PHY_ROUTE_OTG) for phyno !=0 vs. CLR4(sc, 0, OTG_PHY_CFG, OTG_PHY_ROUTE_OTG) before. . .
Mark Millard
markmi at dsl-only.net
Mon Sep 18 20:42:38 UTC 2017
On 2017-Sep-18, at 1:30 PM, Mark Millard <markmi at dsl-only.net> wrote:
> It probably just my ignorance of the code's intent
> but for A64 it used to be that phyno ==1 had code
> that did CLR4 for phyno==0 (hard coded):
>
> if (sc->phy_type == AWUSBPHY_TYPE_A64) {
> CLR4(sc, phyno, PMU_UNK_H3, PMU_UNK_H3_CLR);
>
> /* EHCI0 and OTG share a PHY */
> if (phyno == 0)
> SET4(sc, 0, OTG_PHY_CFG, OTG_PHY_ROUTE_OTG);
> else if (phyno == 1)
> CLR4(sc, 0, OTG_PHY_CFG, OTG_PHY_ROUTE_OTG);
> }
>
> So: that last CLR4 manipulated phyno==0 as far as I can tell,
> no matter what the passed-in phyno was.
>
> In the new code there seems to be no hook for phyno==1
> to manipulate phyno==0 similarly:
>
> if (sc->phy_conf->phy0_route == true) {
> if (phyno == 0)
> SET4(sc->phy_ctrl, OTG_PHY_CFG, OTG_PHY_ROUTE_OTG);
> else
> CLR4(sc->phy_ctrl, OTG_PHY_CFG, OTG_PHY_ROUTE_OTG);
> }
>
> That CLR4 seems to be manipulating phyno==1 instead and
> seems to have no means of doing otherwise.
>
> Was the old code wrong?
May be I asked the reverse of the right question:
that first CLR 4 in the old code varied by phyno but
now always uses phy_ctrl:
if (sc->phy_conf->pmu_unk1 == true)
CLR4(sc->phy_ctrl, PMU_UNK_H3, PMU_UNK_H3_CLR);
Overall one part or the other seems to be a mismatch with the
old code for A64.
===
Mark Millard
markmi at dsl-only.net
More information about the freebsd-arm
mailing list