On 06/11/2012 09:30 PM, Zumeng Chen wrote:

2012/6/11 Jon Hunter <jon-hunter at ti.com <mailto:jon-hunter@ti.com>>


    On 06/11/2012 09:00 AM, Zumeng Chen wrote:
    > A typo fix for this cosmetic change and mute a failed message from
    > a unnecessary setting of some parent clk for usbhs_omap on OMAP3EVM.
    >
    > Signed-off-by: Zumeng Chen <zumeng.chen@gmail.com
    <mailto:zumeng.chen@gmail.com>>
    > ---
    >  drivers/mfd/omap-usb-host.c |    4 +++-
    >  1 files changed, 3 insertions(+), 1 deletions(-)
    >
    > diff --git a/drivers/mfd/omap-usb-host.c b/drivers/mfd/omap-usb-host.c
    > index 7e96bb2..9aaaf3c 100644
    > --- a/drivers/mfd/omap-usb-host.c
    > +++ b/drivers/mfd/omap-usb-host.c
    > @@ -698,8 +698,9 @@ static int __devinit usbhs_omap_probe(struct
    platform_device *pdev)
    >               goto err_usbtll_p2_fck;
    >       }
    >
    > +#ifndef CONFIG_MACH_OMAP3EVM
    > +     /* for OMAP3 , the clk set parent fails */
    >       if (is_ehci_phy_mode(pdata->port_mode[0])) {
    > -             /* for OMAP3 , the clk set paretn fails */
    >               ret = clk_set_parent(omap->utmi_p1_fck,
    >                                       omap->xclk60mhsp1_ck);
    >               if (ret != 0)

    This begs the question, why is port_mode[0] set to ehci phy mode if this
    is failing? Something does not seem right here but this does not look
    like the right fix.

Actually, for omap3530evm, there is only port-mode[1] is valid, so
port_mode[0]
will be skipped by "if". And it will report "xclk60mhsp2_ck set parent
failed"

But why I bracket both of them, because, both xclk60mhsp2_ck and
xclk60mhsp1_ck
have dummy_ck clk for omap3530evm. So we can skip them directly.

How about something like this ...

diff --git a/arch/arm/plat-omap/clock.c b/arch/arm/plat-omap/clock.c
index 62ec5c4..c1335f1 100644
--- a/arch/arm/plat-omap/clock.c
+++ b/arch/arm/plat-omap/clock.c

@@ -149,6 +149,10 @@ int clk_set_parent(struct clk *clk, struct clk *parent)
        if (!arch_clock || !arch_clock->clk_set_parent)
                return ret;

+       /* If the clock is a dummy clock just return */
+       if (!strcmp(clk->name, dummy_ck.name))
+               return 0;
+
        spin_lock_irqsave(&clockfw_lock, flags);
        if (clk->usecount == 0) {
                ret = arch_clock->clk_set_parent(clk, parent);

Can you give this a try?

Cheers
Jon