[PATCH] ARM: shmobile: mackerel: fixup usb module order

[PATCH] ARM: shmobile: mackerel: fixup usb module order

[Kuninori Morimoto]

renesas_usbhs driver can play role as both Host and Gadget.
In case of Gadget, it requires not only renesas_usbhs
but also usb gadget module (like g_ether).
So, renesas_usbhs driver calls usb_add_gadget_udc() on probe time.

Because of this behavior,
Host port plays also Gadget role if kernel has both Host/Gadget support.

In mackerel case, from 0ada2da51800a4914887a9bcf22d563be80e50be
(ARM: mach-shmobile: mackerel: use renesas_usbhs instead of r8a66597_hcd)
usb0 plays Gadget role, and usb1 plays Host role,
and current mackerel board probes as usb1 -> usb0.
Thus, 1st installed usb gadget module (like g_ether) will be
assigned to usb1 (= usb Host port), and 2nd module to usb0 (= usb Gadget port).
It is very confusable for user.
This patch fixup usb modes probing order as usb0 -> usb1.

Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
---
 arch/arm/mach-shmobile/board-mackerel.c |    3 ++-
 1 files changed, 2 insertions(+), 1 deletions(-)

diff --git a/arch/arm/mach-shmobile/board-mackerel.c b/arch/arm/mach-shmobile/board-mackerel.c
index 7b53cda..725bff5 100644
--- a/arch/arm/mach-shmobile/board-mackerel.c
+++ b/arch/arm/mach-shmobile/board-mackerel.c
@@ -679,6 +679,7 @@ static struct platform_device usbhs0_device = {
  *  - J30 "open"
  *  - modify usbhs1_get_id() USBHS_HOST -> USBHS_GADGET
  *  - add .get_vbus = usbhs_get_vbus in usbhs1_private
+ *  - check usbhs0_device(pio)/usbhs1_device(irq) order in mackerel_devices.
  */
 #define IRQ8 evt2irq(0x0300)
 #define USB_PHY_MODE		(1 << 4)
@@ -1262,8 +1263,8 @@ static struct platform_device *mackerel_devices[] __initdata = {
 	&nor_flash_device,
 	&smc911x_device,
 	&lcdc_device,
-	&usbhs1_device,
 	&usbhs0_device,
+	&usbhs1_device,
 	&leds_device,
 	&fsi_device,
 	&fsi_ak4643_device,
-- 
1.7.5.4

Re: [PATCH] ARM: shmobile: mackerel: fixup usb module order

[Simon Horman]

On Sun, Aug 05, 2012 at 10:47:00PM -0700, Kuninori Morimoto wrote:
> renesas_usbhs driver can play role as both Host and Gadget.
> In case of Gadget, it requires not only renesas_usbhs
> but also usb gadget module (like g_ether).
> So, renesas_usbhs driver calls usb_add_gadget_udc() on probe time.
> 
> Because of this behavior,
> Host port plays also Gadget role if kernel has both Host/Gadget support.
> 
> In mackerel case, from 0ada2da51800a4914887a9bcf22d563be80e50be
> (ARM: mach-shmobile: mackerel: use renesas_usbhs instead of r8a66597_hcd)
> usb0 plays Gadget role, and usb1 plays Host role,
> and current mackerel board probes as usb1 -> usb0.
> Thus, 1st installed usb gadget module (like g_ether) will be
> assigned to usb1 (= usb Host port), and 2nd module to usb0 (= usb Gadget port).
> It is very confusable for user.
> This patch fixup usb modes probing order as usb0 -> usb1.
> 
> Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>

Reviewed-by: Simon Horman <horms@verge.net.au>

Re: [PATCH] ARM: shmobile: mackerel: fixup usb module order

[Simon Horman]

On Mon, Aug 06, 2012 at 05:19:12PM +0900, Simon Horman wrote:
> On Sun, Aug 05, 2012 at 10:47:00PM -0700, Kuninori Morimoto wrote:
> > renesas_usbhs driver can play role as both Host and Gadget.
> > In case of Gadget, it requires not only renesas_usbhs
> > but also usb gadget module (like g_ether).
> > So, renesas_usbhs driver calls usb_add_gadget_udc() on probe time.
> > 
> > Because of this behavior,
> > Host port plays also Gadget role if kernel has both Host/Gadget support.
> > 
> > In mackerel case, from 0ada2da51800a4914887a9bcf22d563be80e50be
> > (ARM: mach-shmobile: mackerel: use renesas_usbhs instead of r8a66597_hcd)
> > usb0 plays Gadget role, and usb1 plays Host role,
> > and current mackerel board probes as usb1 -> usb0.
> > Thus, 1st installed usb gadget module (like g_ether) will be
> > assigned to usb1 (= usb Host port), and 2nd module to usb0 (= usb Gadget port).
> > It is very confusable for user.
> > This patch fixup usb modes probing order as usb0 -> usb1.
> > 
> > Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
> 
> Reviewed-by: Simon Horman <horms@verge.net.au>

I have pushed this to the mackerel branch of my renesas tree,
git://git.kernel.org/pub/scm/linux/kernel/git/horms/renesas.git

I will endeavour to get it included in 3.7.

Re: [PATCH] ARM: shmobile: mackerel: fixup usb module order

[Simon Horman]

On Mon, Aug 27, 2012 at 05:41:13PM +0900, Simon Horman wrote:
> On Mon, Aug 06, 2012 at 05:19:12PM +0900, Simon Horman wrote:
> > On Sun, Aug 05, 2012 at 10:47:00PM -0700, Kuninori Morimoto wrote:
> > > renesas_usbhs driver can play role as both Host and Gadget.
> > > In case of Gadget, it requires not only renesas_usbhs
> > > but also usb gadget module (like g_ether).
> > > So, renesas_usbhs driver calls usb_add_gadget_udc() on probe time.
> > > 
> > > Because of this behavior,
> > > Host port plays also Gadget role if kernel has both Host/Gadget support.
> > > 
> > > In mackerel case, from 0ada2da51800a4914887a9bcf22d563be80e50be
> > > (ARM: mach-shmobile: mackerel: use renesas_usbhs instead of r8a66597_hcd)
> > > usb0 plays Gadget role, and usb1 plays Host role,
> > > and current mackerel board probes as usb1 -> usb0.
> > > Thus, 1st installed usb gadget module (like g_ether) will be
> > > assigned to usb1 (= usb Host port), and 2nd module to usb0 (= usb Gadget port).
> > > It is very confusable for user.
> > > This patch fixup usb modes probing order as usb0 -> usb1.
> > > 
> > > Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
> > 
> > Reviewed-by: Simon Horman <horms@verge.net.au>
> 
> I have pushed this to the mackerel branch of my renesas tree,
> git://git.kernel.org/pub/scm/linux/kernel/git/horms/renesas.git
> 
> I will endeavour to get it included in 3.7.

On further consideration I have moved this patch to the
fixes branch (and removed the now empty mackerel branch).

I will endeavour to get it included in 3.6.