[PATCH] arm: omap4: panda: remove usb_nop_xceiv_register

[PATCH] arm: omap4: panda: remove usb_nop_xceiv_register

[tom.leiming]

From: Ming Lei <tom.leiming@gmail.com>

Panda uses twl6030 otg phy instead of internal phy,
so removes usb_nop_xceiv_register to make musb working.

Cc: Felipe Balbi <balbi@ti.com>
Signed-off-by: Ming Lei <tom.leiming@gmail.com>
---
 arch/arm/mach-omap2/board-omap4panda.c |    2 --
 1 files changed, 0 insertions(+), 2 deletions(-)

diff --git a/arch/arm/mach-omap2/board-omap4panda.c b/arch/arm/mach-omap2/board-omap4panda.c
index 1ecd0a6..802ae20 100644
--- a/arch/arm/mach-omap2/board-omap4panda.c
+++ b/arch/arm/mach-omap2/board-omap4panda.c
@@ -374,8 +374,6 @@ static void __init omap4_panda_init(void)
 	platform_add_devices(panda_devices, ARRAY_SIZE(panda_devices));
 	omap_serial_init();
 	omap4_twl6030_hsmmc_init(mmc);
-	/* OMAP4 Panda uses internal transceiver so register nop transceiver */
-	usb_nop_xceiv_register();
 	omap4_ehci_init();
 	/* FIXME: allow multi-omap to boot until musb is updated for omap4 */
 	if (!cpu_is_omap44xx())
-- 
1.7.3

Re: [PATCH] arm: omap4: panda: remove usb_nop_xceiv_register

[Kevin Hilman]

tom.leiming@gmail.com writes:

> From: Ming Lei <tom.leiming@gmail.com>
>
> Panda uses twl6030 otg phy instead of internal phy,
> so removes usb_nop_xceiv_register to make musb working.
>
> Cc: Felipe Balbi <balbi@ti.com>
> Signed-off-by: Ming Lei <tom.leiming@gmail.com>

Please Cc the linux-arm-kernel list for OMAP kernel patches targetted
for upstream.

Also, would be nice to get some Reviewed-by and/or Tested-by tags from
Panda users on this one as well.

Kevin

> ---
>  arch/arm/mach-omap2/board-omap4panda.c |    2 --
>  1 files changed, 0 insertions(+), 2 deletions(-)
>
> diff --git a/arch/arm/mach-omap2/board-omap4panda.c b/arch/arm/mach-omap2/board-omap4panda.c
> index 1ecd0a6..802ae20 100644
> --- a/arch/arm/mach-omap2/board-omap4panda.c
> +++ b/arch/arm/mach-omap2/board-omap4panda.c
> @@ -374,8 +374,6 @@ static void __init omap4_panda_init(void)
>  	platform_add_devices(panda_devices, ARRAY_SIZE(panda_devices));
>  	omap_serial_init();
>  	omap4_twl6030_hsmmc_init(mmc);
> -	/* OMAP4 Panda uses internal transceiver so register nop transceiver */
> -	usb_nop_xceiv_register();
>  	omap4_ehci_init();
>  	/* FIXME: allow multi-omap to boot until musb is updated for omap4 */
>  	if (!cpu_is_omap44xx())

Re: [PATCH] arm: omap4: panda: remove usb_nop_xceiv_register

[Felipe Balbi]

On Tue, Jan 04, 2011 at 04:29:04PM -0800, Kevin Hilman wrote:
> tom.leiming@gmail.com writes:
> 
> > From: Ming Lei <tom.leiming@gmail.com>
> >
> > Panda uses twl6030 otg phy instead of internal phy,

This is wrong, it uses both. TWL6030 handles VBUS and ID pin
and the internal PHY handles Data Lines. I know it's a bit
crazy but this kind of design is becoming more and more usual
on different boards.

What truly needs to happen is that we allow for multiple PHYs
to be added to otg_transceiver (BTW, that's a mis-naming as
transceiver aren't only for OTG use-cases).

Anyway, change this phrasing.

> > so removes usb_nop_xceiv_register to make musb working.
> >
> > Cc: Felipe Balbi <balbi@ti.com>
> > Signed-off-by: Ming Lei <tom.leiming@gmail.com>

Considering you change the commit log above:

Reviewed-by: Felipe Balbi <balbi@ti.com>

This is ok, but can only go after 2.6.38 merge window. There are patches
to add support for internal PHY. Ideally we would actually register both
PHYs and it would work. They would have different capabilities, though.

-- 
balbi

Re: [PATCH] arm: omap4: panda: remove usb_nop_xceiv_register

[Ming Lei]

Hi,

2011/1/5 Felipe Balbi <balbi@ti.com>:
> On Tue, Jan 04, 2011 at 04:29:04PM -0800, Kevin Hilman wrote:
>> tom.leiming@gmail.com writes:
>>
>> > From: Ming Lei <tom.leiming@gmail.com>
>> >
>> > Panda uses twl6030 otg phy instead of internal phy,
>
> This is wrong, it uses both. TWL6030 handles VBUS and ID pin
> and the internal PHY handles Data Lines. I know it's a bit
> crazy but this kind of design is becoming more and more usual
> on different boards.

I see now after your explanation.

>
> What truly needs to happen is that we allow for multiple PHYs
> to be added to otg_transceiver (BTW, that's a mis-naming as
> transceiver aren't only for OTG use-cases).
>
> Anyway, change this phrasing.
>
>> > so removes usb_nop_xceiv_register to make musb working.
>> >
>> > Cc: Felipe Balbi <balbi@ti.com>
>> > Signed-off-by: Ming Lei <tom.leiming@gmail.com>
>
> Considering you change the commit log above:
>
> Reviewed-by: Felipe Balbi <balbi@ti.com>
>
> This is ok, but can only go after 2.6.38 merge window. There are patches
> to add support for internal PHY. Ideally we would actually register both
> PHYs and it would work. They would have different capabilities, though.

The current otg only supports one global transceiver, so we have to remove
the nop on Panda.

thanks,
-- 
Lei Ming

Re: [PATCH] arm: omap4: panda: remove usb_nop_xceiv_register

[Felipe Balbi]

Hi,

On Wed, Jan 05, 2011 at 09:40:19PM +0800, Ming Lei wrote:
> > Reviewed-by: Felipe Balbi <balbi@ti.com>
> >
> > This is ok, but can only go after 2.6.38 merge window. There are patches
> > to add support for internal PHY. Ideally we would actually register both
> > PHYs and it would work. They would have different capabilities, though.
> 
> The current otg only supports one global transceiver, so we have to remove
> the nop on Panda.

Yes, of course. I just pointed out your commit log is wrong and that
ideally we would fix this limitation. Ideally we would we have some sort
of framework for transceivers where we could add several of them.

-- 
balbi