[PATCH 1/2] omap4: 4430sdp: drop ehci support

[PATCH 1/2] omap4: 4430sdp: drop ehci support

[Anand Gadiyar]

Most revisions of the OMAP4 Blaze/SDP platform do not have
the EHCI signals routed by default. The pads are routed
for the alternate HSI functionality instead, and explicit
board modifications are needed to route the signals to
the USB PHY on the board.

Also, turning on the PHY connected to the EHCI port causes
a board reboot during bootup due to an unintended short
on the rails - this affects many initial revisions of the
board, and needs a minor board mod to fix (or as a
workaround, one should not attempt to power on the
USB PHY).

Given that these boards need explicit board mods to even
get EHCI working (separate from the accidental short above),
we should not attempt to enable EHCI by default.

So drop the EHCI support from the board files for the
Blaze/SDP platforms.

Signed-off-by: Anand Gadiyar <gadiyar@ti.com>
Cc: Keshava Munegowda <keshava_mgowda@ti.com>
Cc: Tony Lindgren <tony@atomide.com>
---

 arch/arm/mach-omap2/board-4430sdp.c |   20 --------------------
 1 file changed, 20 deletions(-)

Index: linux-omap-usb/arch/arm/mach-omap2/board-4430sdp.c
===================================================================
--- linux-omap-usb.orig/arch/arm/mach-omap2/board-4430sdp.c
+++ linux-omap-usb/arch/arm/mach-omap2/board-4430sdp.c
@@ -44,7 +44,6 @@
 #define ETH_KS8851_IRQ			34
 #define ETH_KS8851_POWER_ON		48
 #define ETH_KS8851_QUART		138
-#define OMAP4SDP_MDM_PWR_EN_GPIO	157
 #define OMAP4_SFH7741_SENSOR_OUTPUT_GPIO	184
 #define OMAP4_SFH7741_ENABLE_GPIO		188
 
@@ -251,16 +250,6 @@ static void __init omap_4430sdp_init_irq
 	gic_init_irq();
 }
 
-static const struct ehci_hcd_omap_platform_data ehci_pdata __initconst = {
-	.port_mode[0]	= EHCI_HCD_OMAP_MODE_PHY,
-	.port_mode[1]	= EHCI_HCD_OMAP_MODE_UNKNOWN,
-	.port_mode[2]	= EHCI_HCD_OMAP_MODE_UNKNOWN,
-	.phy_reset	= false,
-	.reset_gpio_port[0]  = -EINVAL,
-	.reset_gpio_port[1]  = -EINVAL,
-	.reset_gpio_port[2]  = -EINVAL,
-};
-
 static struct omap_musb_board_data musb_board_data = {
 	.interface_type		= MUSB_INTERFACE_UTMI,
 	.mode			= MUSB_OTG,
@@ -576,14 +564,6 @@ static void __init omap_4430sdp_init(voi
 	omap_serial_init();
 	omap4_twl6030_hsmmc_init(mmc);
 
-	/* Power on the ULPI PHY */
-	status = gpio_request(OMAP4SDP_MDM_PWR_EN_GPIO, "USBB1 PHY VMDM_3V3");
-	if (status)
-		pr_err("%s: Could not get USBB1 PHY GPIO\n", __func__);
-	else
-		gpio_direction_output(OMAP4SDP_MDM_PWR_EN_GPIO, 1);
-
-	usb_ehci_init(&ehci_pdata);
 	usb_musb_init(&musb_board_data);
 
 	status = omap_ethernet_init();

[PATCH 2/2] omap4: 4430sdp: mux GPIO_157 to safe mode

[Anand Gadiyar]

Commit 7d4ca85a5 (omap4: Fix ULPI PHY init for ES1.0 SDP)
changed the mux mode of the pad for GPIO_157 to HSI_MODE
by accident - the original intent was to mux it properly
to solve a board reboot problem.

The board reboot is caused by a bug on many revs of the
board that causes an accidental short circuilt when this
pad is driven high. To ensure this never happens (even
with misbehaving bootloaders), mux the pad to safe mode instead.

Signed-off-by: Anand Gadiyar <gadiyar@ti.com>
Cc: Tony Lindgren <tony@atomide.com>
---
 arch/arm/mach-omap2/board-4430sdp.c |   10 +++++++++-
 1 file changed, 9 insertions(+), 1 deletion(-)

Index: linux-omap-usb/arch/arm/mach-omap2/board-4430sdp.c
===================================================================
--- linux-omap-usb.orig/arch/arm/mach-omap2/board-4430sdp.c
+++ linux-omap-usb/arch/arm/mach-omap2/board-4430sdp.c
@@ -543,7 +543,15 @@ static void __init omap_sfh7741prox_init
 
 #ifdef CONFIG_OMAP_MUX
 static struct omap_board_mux board_mux[] __initdata = {
-	OMAP4_MUX(USBB2_ULPITLL_CLK, OMAP_MUX_MODE4 | OMAP_PIN_OUTPUT),
+	/*
+	 * USBB2_ULPITLL_CLK/GPIO_157 controls the circuitry
+	 * that turns on the PHY connected to the EHCI port.
+	 * several revisions of the Blaze are affected by
+	 * an accidental short-circuit that causes a reboot
+	 * if this pad is driven high. So, explicitly configure
+	 * the pad in safe mode to ensure this never happens
+	 */
+	OMAP4_MUX(USBB2_ULPITLL_CLK, OMAP_MUX_MODE7),
 	{ .reg_offset = OMAP_MUX_TERMINATOR },
 };
 #else

[PATCH 1/2] omap4: 4430sdp: drop ehci support

[Felipe Balbi]

On Wed, Feb 16, 2011 at 04:47:19PM +0530, Anand Gadiyar wrote:
> Most revisions of the OMAP4 Blaze/SDP platform do not have
> the EHCI signals routed by default. The pads are routed
> for the alternate HSI functionality instead, and explicit
> board modifications are needed to route the signals to
> the USB PHY on the board.
> 
> Also, turning on the PHY connected to the EHCI port causes
> a board reboot during bootup due to an unintended short
> on the rails - this affects many initial revisions of the
> board, and needs a minor board mod to fix (or as a
> workaround, one should not attempt to power on the
> USB PHY).
> 
> Given that these boards need explicit board mods to even
> get EHCI working (separate from the accidental short above),
> we should not attempt to enable EHCI by default.
> 
> So drop the EHCI support from the board files for the
> Blaze/SDP platforms.
> 
> Signed-off-by: Anand Gadiyar <gadiyar@ti.com>
> Cc: Keshava Munegowda <keshava_mgowda@ti.com>
> Cc: Tony Lindgren <tony@atomide.com>

Tony, if you want to queue this one directly, here's my:

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

if you wish, I can send you a pull request as well. I already have
another patch for you anyway.

-- 
balbi

[PATCH 1/2] omap4: 4430sdp: drop ehci support

[Tony Lindgren]

* Felipe Balbi <balbi@ti.com> [110216 03:25]:
> On Wed, Feb 16, 2011 at 04:47:19PM +0530, Anand Gadiyar wrote:
> > Most revisions of the OMAP4 Blaze/SDP platform do not have
> > the EHCI signals routed by default. The pads are routed
> > for the alternate HSI functionality instead, and explicit
> > board modifications are needed to route the signals to
> > the USB PHY on the board.
> > 
> > Also, turning on the PHY connected to the EHCI port causes
> > a board reboot during bootup due to an unintended short
> > on the rails - this affects many initial revisions of the
> > board, and needs a minor board mod to fix (or as a
> > workaround, one should not attempt to power on the
> > USB PHY).
> > 
> > Given that these boards need explicit board mods to even
> > get EHCI working (separate from the accidental short above),
> > we should not attempt to enable EHCI by default.
> > 
> > So drop the EHCI support from the board files for the
> > Blaze/SDP platforms.
> > 
> > Signed-off-by: Anand Gadiyar <gadiyar@ti.com>
> > Cc: Keshava Munegowda <keshava_mgowda@ti.com>
> > Cc: Tony Lindgren <tony@atomide.com>
> 
> Tony, if you want to queue this one directly, here's my:
> 
> Acked-by: Felipe Balbi <balbi@ti.com>
> 
> if you wish, I can send you a pull request as well. I already have
> another patch for you anyway.

OK if you have more I'd rather pull.

Tony

[PATCH 1/2] omap4: 4430sdp: drop ehci support

[Felipe Balbi]

On Wed, Feb 16, 2011 at 05:42:10PM -0800, Tony Lindgren wrote:
> * Felipe Balbi <balbi@ti.com> [110216 03:25]:
> > On Wed, Feb 16, 2011 at 04:47:19PM +0530, Anand Gadiyar wrote:
> > > Most revisions of the OMAP4 Blaze/SDP platform do not have
> > > the EHCI signals routed by default. The pads are routed
> > > for the alternate HSI functionality instead, and explicit
> > > board modifications are needed to route the signals to
> > > the USB PHY on the board.
> > > 
> > > Also, turning on the PHY connected to the EHCI port causes
> > > a board reboot during bootup due to an unintended short
> > > on the rails - this affects many initial revisions of the
> > > board, and needs a minor board mod to fix (or as a
> > > workaround, one should not attempt to power on the
> > > USB PHY).
> > > 
> > > Given that these boards need explicit board mods to even
> > > get EHCI working (separate from the accidental short above),
> > > we should not attempt to enable EHCI by default.
> > > 
> > > So drop the EHCI support from the board files for the
> > > Blaze/SDP platforms.
> > > 
> > > Signed-off-by: Anand Gadiyar <gadiyar@ti.com>
> > > Cc: Keshava Munegowda <keshava_mgowda@ti.com>
> > > Cc: Tony Lindgren <tony@atomide.com>
> > 
> > Tony, if you want to queue this one directly, here's my:
> > 
> > Acked-by: Felipe Balbi <balbi@ti.com>
> > 
> > if you wish, I can send you a pull request as well. I already have
> > another patch for you anyway.
> 
> OK if you have more I'd rather pull.

ok. Taken.

-- 
balbi

[PATCH 1/2] omap4: 4430sdp: drop ehci support

[Felipe Balbi]

On Wed, Feb 16, 2011 at 04:47:19PM +0530, Anand Gadiyar wrote:
> Most revisions of the OMAP4 Blaze/SDP platform do not have
> the EHCI signals routed by default. The pads are routed
> for the alternate HSI functionality instead, and explicit
> board modifications are needed to route the signals to
> the USB PHY on the board.
> 
> Also, turning on the PHY connected to the EHCI port causes
> a board reboot during bootup due to an unintended short
> on the rails - this affects many initial revisions of the
> board, and needs a minor board mod to fix (or as a
> workaround, one should not attempt to power on the
> USB PHY).
> 
> Given that these boards need explicit board mods to even
> get EHCI working (separate from the accidental short above),
> we should not attempt to enable EHCI by default.
> 
> So drop the EHCI support from the board files for the
> Blaze/SDP platforms.
> 
> Signed-off-by: Anand Gadiyar <gadiyar@ti.com>
> Cc: Keshava Munegowda <keshava_mgowda@ti.com>
> Cc: Tony Lindgren <tony@atomide.com>

applied, thanks.

-- 
balbi

[PATCH 1/2] omap4: 4430sdp: drop ehci support

[Russell King - ARM Linux]

On Wed, Feb 16, 2011 at 04:47:19PM +0530, Anand Gadiyar wrote:
> Most revisions of the OMAP4 Blaze/SDP platform do not have
> the EHCI signals routed by default. The pads are routed
> for the alternate HSI functionality instead, and explicit
> board modifications are needed to route the signals to
> the USB PHY on the board.
> 
> Also, turning on the PHY connected to the EHCI port causes
> a board reboot during bootup due to an unintended short
> on the rails - this affects many initial revisions of the
> board, and needs a minor board mod to fix (or as a
> workaround, one should not attempt to power on the
> USB PHY).

It's probably a good idea to put some comment in the code to that
effect - people don't generally look at the history when adding new
features.  It's probably also a good idea to claim the GPIO for the
PHY power control and ensure that it's set to 'off' mode - that code
can then be a placeholder for the comment.

[PATCH 1/2] omap4: 4430sdp: drop ehci support

[Anand Gadiyar]

Russell King - ARM Linux wrote:
> On Wed, Feb 16, 2011 at 04:47:19PM +0530, Anand Gadiyar wrote:
> > Most revisions of the OMAP4 Blaze/SDP platform do not have
> > the EHCI signals routed by default. The pads are routed
> > for the alternate HSI functionality instead, and explicit
> > board modifications are needed to route the signals to
> > the USB PHY on the board.
> >
> > Also, turning on the PHY connected to the EHCI port causes
> > a board reboot during bootup due to an unintended short
> > on the rails - this affects many initial revisions of the
> > board, and needs a minor board mod to fix (or as a
> > workaround, one should not attempt to power on the
> > USB PHY).
>
> It's probably a good idea to put some comment in the code to that
> effect - people don't generally look at the history when adding new
> features.

I did part of this in patch 2/2.
I've muxed the pad in "safe mode" which turns off the output drivers
for the pad - so it can no longer drive a 1.

I've also placed a big warning above that line explaining why.

> It's probably also a good idea to claim the GPIO for the
> PHY power control and ensure that it's set to 'off' mode - that code
> can then be a placeholder for the comment.

This is a good idea too. Tony, if you concur, I'll go
respin this series.

- Anand