[PATCH v2 1/3] phy_sun4i_usb: set_mode: Allow using set_mode to force end the current session

[PATCH v2 1/3] phy_sun4i_usb: set_mode: Allow using set_mode to force end the current session

[Hans de Goede]

The sunxi musb has a bug where sometimes it will generate a babble
error on device disconnect instead of a disconnect irq. When this
happens the musb-controller switches from host mode to device mode
(it clears MUSB_DEVCTL_SESSION and sets MUSB_DEVCTL_BDEVICE) and
gets stuck in this state.

Clearing this requires reporting Vbus low for 200 or more ms, but
on some devices Vbus is simply always high (host-only mode, no Vbus
control).

This commit modifies sun4i_usb_phy_set_mode so that it will force
end the current session when called with the current mode, before this
commit calling set_mode with the current mode was a nop since id_det
would stay the same resulting in the detect_work not doing anything.

This allows the sunxi-musb glue to use sun4i_usb_phy_set_mode to force
end the current session without changing the mode, to fixup the stuck
state after a babble error.

Signed-off-by: Hans de Goede <hdegoede@redhat.com>
---
Changes in v2:
-New patch in v2 of this series replacing the
 "phy-sun4i-usb: Add sun4i_usb_phy_force_session_end() function"
 from v1
---
 drivers/phy/phy-sun4i-usb.c | 15 +++++++++++----
 1 file changed, 11 insertions(+), 4 deletions(-)

diff --git a/drivers/phy/phy-sun4i-usb.c b/drivers/phy/phy-sun4i-usb.c
index 43c0d98..d76538c 100644
--- a/drivers/phy/phy-sun4i-usb.c
+++ b/drivers/phy/phy-sun4i-usb.c
@@ -437,25 +437,32 @@ static int sun4i_usb_phy_set_mode(struct phy *_phy, enum phy_mode mode)
 {
 	struct sun4i_usb_phy *phy = phy_get_drvdata(_phy);
 	struct sun4i_usb_phy_data *data = to_sun4i_usb_phy_data(phy);
+	enum usb_dr_mode new_mode;
 
 	if (phy->index != 0)
 		return -EINVAL;
 
 	switch (mode) {
 	case PHY_MODE_USB_HOST:
-		data->dr_mode = USB_DR_MODE_HOST;
+		new_mode = USB_DR_MODE_HOST;
 		break;
 	case PHY_MODE_USB_DEVICE:
-		data->dr_mode = USB_DR_MODE_PERIPHERAL;
+		new_mode = USB_DR_MODE_PERIPHERAL;
 		break;
 	case PHY_MODE_USB_OTG:
-		data->dr_mode = USB_DR_MODE_OTG;
+		new_mode = USB_DR_MODE_OTG;
 		break;
 	default:
 		return -EINVAL;
 	}
 
-	dev_info(&_phy->dev, "Changing dr_mode to %d\n", (int)data->dr_mode);
+	if (new_mode != data->dr_mode) {
+		dev_info(&_phy->dev, "Changing dr_mode to %d\n",
+			 (int)data->dr_mode);
+		data->dr_mode = new_mode;
+	}
+
+	data->id_det = -1; /* Force reprocessing of id */
 	data->force_session_end = true;
 	queue_delayed_work(system_wq, &data->detect, 0);
 
-- 
2.9.3

[PATCH v2 2/3] musb: sunxi: Remove custom babble handling

[Hans de Goede]

The musb-core now a days always treats babble errors in host mode
as disconnects, so there is no need for the sunxi specific handling
of this anymore.

Signed-off-by: Hans de Goede <hdegoede@redhat.com>
---
Changes in v2:
-This is a new patch in v2 of this patch series
---
 drivers/usb/musb/sunxi.c | 10 ----------
 1 file changed, 10 deletions(-)

diff --git a/drivers/usb/musb/sunxi.c b/drivers/usb/musb/sunxi.c
index 1408245..82eba92 100644
--- a/drivers/usb/musb/sunxi.c
+++ b/drivers/usb/musb/sunxi.c
@@ -186,16 +186,6 @@ static irqreturn_t sunxi_musb_interrupt(int irq, void *__hci)
 	if (musb->int_usb)
 		writeb(musb->int_usb, musb->mregs + SUNXI_MUSB_INTRUSB);
 
-	/*
-	 * sunxi musb often signals babble on low / full speed device
-	 * disconnect, without ever raising MUSB_INTR_DISCONNECT, since
-	 * normally babble never happens treat it as disconnect.
-	 */
-	if ((musb->int_usb & MUSB_INTR_BABBLE) && is_host_active(musb)) {
-		musb->int_usb &= ~MUSB_INTR_BABBLE;
-		musb->int_usb |= MUSB_INTR_DISCONNECT;
-	}
-
 	if ((musb->int_usb & MUSB_INTR_RESET) && !is_host_active(musb)) {
 		/* ep0 FADDR must be 0 when (re)entering peripheral mode */
 		musb_ep_select(musb->mregs, 0);
-- 
2.9.3

[PATCH v2 3/3] musb: sunxi: Force session end on babble errors in host-mode

[Hans de Goede]

The sunxi musb has a bug where sometimes it will generate a babble
error on device disconnect instead of a disconnect irq. When this
happens the musb-controller switches from host mode to device mode
(it clears MUSB_DEVCTL_SESSION and sets MUSB_DEVCTL_BDEVICE) and
gets stuck in this state.

Clearing this requires reporting Vbus low for 200 or more ms, but
on some devices Vbus is simply always high (host-only mode, no Vbus
control).

This commit adds a sunxi_musb_recover() callback which makes
sunxi_musb_work call phy_set_mode with the current mode, which
will force end the current session.

This fixes the musb controller getting stuck in this state on systems
without Vbus control; and also fixes the need to unplug the usb-b ->
usb-a cable to get out of this state on systems with Vbus control.

Signed-off-by: Hans de Goede <hdegoede@redhat.com>
---
Changes in v2:
-Use musb_platform_recover callback instead of using DYI code in the
 interrupt handler
-Call phy_set_mode with the current mode instead of adding a new custom
 sunxi phy callback
---
 drivers/usb/musb/sunxi.c | 15 +++++++++++++++
 1 file changed, 15 insertions(+)

diff --git a/drivers/usb/musb/sunxi.c b/drivers/usb/musb/sunxi.c
index 82eba92..d0be0ea 100644
--- a/drivers/usb/musb/sunxi.c
+++ b/drivers/usb/musb/sunxi.c
@@ -380,6 +380,20 @@ static int sunxi_musb_set_mode(struct musb *musb, u8 mode)
 	return 0;
 }
 
+static int sunxi_musb_recover(struct musb *musb)
+{
+	struct sunxi_glue *glue = dev_get_drvdata(musb->controller->parent);
+
+	/*
+	 * Schedule a phy_set_mode with the current glue->phy_mode value,
+	 * this will force end the current session.
+	 */
+	set_bit(SUNXI_MUSB_FL_PHY_MODE_PEND, &glue->flags);
+	schedule_work(&glue->work);
+
+	return 0;
+}
+
 /*
  * sunxi musb register layout
  * 0x00 - 0x17	fifo regs, 1 long per fifo
@@ -608,6 +622,7 @@ static const struct musb_platform_ops sunxi_musb_ops = {
 	.dma_init	= sunxi_musb_dma_controller_create,
 	.dma_exit	= sunxi_musb_dma_controller_destroy,
 	.set_mode	= sunxi_musb_set_mode,
+	.recover	= sunxi_musb_recover,
 	.set_vbus	= sunxi_musb_set_vbus,
 	.pre_root_reset_end = sunxi_musb_pre_root_reset_end,
 	.post_root_reset_end = sunxi_musb_post_root_reset_end,
-- 
2.9.3

[PATCH v2 2/3] musb: sunxi: Remove custom babble handling

[Bin Liu]

Hi,

On Thu, Sep 22, 2016 at 02:19:00PM +0300, Hans de Goede wrote:
> The musb-core now a days always treats babble errors in host mode

I don't think this statement is accurate. You might want to change it to
"The musb core already handles babble interrupt" or something else.

Regards,
-Bin.

> as disconnects, so there is no need for the sunxi specific handling
> of this anymore.
> 
> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
> ---
> Changes in v2:
> -This is a new patch in v2 of this patch series
> ---
>  drivers/usb/musb/sunxi.c | 10 ----------
>  1 file changed, 10 deletions(-)
> 
> diff --git a/drivers/usb/musb/sunxi.c b/drivers/usb/musb/sunxi.c
> index 1408245..82eba92 100644
> --- a/drivers/usb/musb/sunxi.c
> +++ b/drivers/usb/musb/sunxi.c
> @@ -186,16 +186,6 @@ static irqreturn_t sunxi_musb_interrupt(int irq, void *__hci)
>  	if (musb->int_usb)
>  		writeb(musb->int_usb, musb->mregs + SUNXI_MUSB_INTRUSB);
>  
> -	/*
> -	 * sunxi musb often signals babble on low / full speed device
> -	 * disconnect, without ever raising MUSB_INTR_DISCONNECT, since
> -	 * normally babble never happens treat it as disconnect.
> -	 */
> -	if ((musb->int_usb & MUSB_INTR_BABBLE) && is_host_active(musb)) {
> -		musb->int_usb &= ~MUSB_INTR_BABBLE;
> -		musb->int_usb |= MUSB_INTR_DISCONNECT;
> -	}
> -
>  	if ((musb->int_usb & MUSB_INTR_RESET) && !is_host_active(musb)) {
>  		/* ep0 FADDR must be 0 when (re)entering peripheral mode */
>  		musb_ep_select(musb->mregs, 0);
> -- 
> 2.9.3
> 

[PATCH v2 2/3] musb: sunxi: Remove custom babble handling

[Hans de Goede]

Hi,

On 09/22/2016 04:54 PM, Bin Liu wrote:
> Hi,
>
> On Thu, Sep 22, 2016 at 02:19:00PM +0300, Hans de Goede wrote:
>> The musb-core now a days always treats babble errors in host mode
>
> I don't think this statement is accurate. You might want to change it to
> "The musb core already handles babble interrupt" or something else.

It is accurate if you look in the history at drivers/usb/musb
commits around 15-03-10 you will see 2 relevant commits:

https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/drivers/usb/musb?id=b4dc38fd45b63e3da2bc98db5d283a15a637a2fa

"usb: musb: core: simplify musb_recover_work()"

This commits introduces calling musb_root_disconnect(musb)
on babble errors, that was not happening before which is why
I added the custom babble error handling. to the sunxi glue.

And:

"usb: musb: core: always try to recover from babble"

Where the title says it all.

Take these together and I believe that my commit msg:

"The musb-core now a days always treats babble errors in host mode
as disconnects, so there is no need for the sunxi specific handling
of this anymore."

Is quite accurate.

Regards,

Hans


>>
>> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
>> ---
>> Changes in v2:
>> -This is a new patch in v2 of this patch series
>> ---
>>  drivers/usb/musb/sunxi.c | 10 ----------
>>  1 file changed, 10 deletions(-)
>>
>> diff --git a/drivers/usb/musb/sunxi.c b/drivers/usb/musb/sunxi.c
>> index 1408245..82eba92 100644
>> --- a/drivers/usb/musb/sunxi.c
>> +++ b/drivers/usb/musb/sunxi.c
>> @@ -186,16 +186,6 @@ static irqreturn_t sunxi_musb_interrupt(int irq, void *__hci)
>>  	if (musb->int_usb)
>>  		writeb(musb->int_usb, musb->mregs + SUNXI_MUSB_INTRUSB);
>>
>> -	/*
>> -	 * sunxi musb often signals babble on low / full speed device
>> -	 * disconnect, without ever raising MUSB_INTR_DISCONNECT, since
>> -	 * normally babble never happens treat it as disconnect.
>> -	 */
>> -	if ((musb->int_usb & MUSB_INTR_BABBLE) && is_host_active(musb)) {
>> -		musb->int_usb &= ~MUSB_INTR_BABBLE;
>> -		musb->int_usb |= MUSB_INTR_DISCONNECT;
>> -	}
>> -
>>  	if ((musb->int_usb & MUSB_INTR_RESET) && !is_host_active(musb)) {
>>  		/* ep0 FADDR must be 0 when (re)entering peripheral mode */
>>  		musb_ep_select(musb->mregs, 0);
>> --
>> 2.9.3
>>

[PATCH v2 2/3] musb: sunxi: Remove custom babble handling

[Bin Liu]

Hi,

On Thu, Sep 22, 2016 at 05:03:39PM +0300, Hans de Goede wrote:
> Hi,
> 
> On 09/22/2016 04:54 PM, Bin Liu wrote:
> >Hi,
> >
> >On Thu, Sep 22, 2016 at 02:19:00PM +0300, Hans de Goede wrote:
> >>The musb-core now a days always treats babble errors in host mode
> >
> >I don't think this statement is accurate. You might want to change it to
> >"The musb core already handles babble interrupt" or something else.
> 
> It is accurate if you look in the history at drivers/usb/musb
> commits around 15-03-10 you will see 2 relevant commits:
> 
> https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/drivers/usb/musb?id=b4dc38fd45b63e3da2bc98db5d283a15a637a2fa
> 
> "usb: musb: core: simplify musb_recover_work()"
> 
> This commits introduces calling musb_root_disconnect(musb)
> on babble errors, that was not happening before which is why

That is true, but calling musb_root_disconnect() is just one step of the
recovery in musb core, not all.

The statement of "treats babble errors in host mode as disconnects"
implies all the babble handling is just disconnect, which is not
accurate.

BTY, "babble errors in host mode" is also redundant. babble implies
host mode.

Regards,
-Bin.
> I added the custom babble error handling. to the sunxi glue.
> 
> And:
> 
> "usb: musb: core: always try to recover from babble"
> 
> Where the title says it all.
> 
> Take these together and I believe that my commit msg:
> 
> "The musb-core now a days always treats babble errors in host mode
> as disconnects, so there is no need for the sunxi specific handling
> of this anymore."
> 
> Is quite accurate.
> 
> Regards,
> 
> Hans
> 
> 
> >>
> >>Signed-off-by: Hans de Goede <hdegoede@redhat.com>
> >>---
> >>Changes in v2:
> >>-This is a new patch in v2 of this patch series
> >>---
> >> drivers/usb/musb/sunxi.c | 10 ----------
> >> 1 file changed, 10 deletions(-)
> >>
> >>diff --git a/drivers/usb/musb/sunxi.c b/drivers/usb/musb/sunxi.c
> >>index 1408245..82eba92 100644
> >>--- a/drivers/usb/musb/sunxi.c
> >>+++ b/drivers/usb/musb/sunxi.c
> >>@@ -186,16 +186,6 @@ static irqreturn_t sunxi_musb_interrupt(int irq, void *__hci)
> >> 	if (musb->int_usb)
> >> 		writeb(musb->int_usb, musb->mregs + SUNXI_MUSB_INTRUSB);
> >>
> >>-	/*
> >>-	 * sunxi musb often signals babble on low / full speed device
> >>-	 * disconnect, without ever raising MUSB_INTR_DISCONNECT, since
> >>-	 * normally babble never happens treat it as disconnect.
> >>-	 */
> >>-	if ((musb->int_usb & MUSB_INTR_BABBLE) && is_host_active(musb)) {
> >>-		musb->int_usb &= ~MUSB_INTR_BABBLE;
> >>-		musb->int_usb |= MUSB_INTR_DISCONNECT;
> >>-	}
> >>-
> >> 	if ((musb->int_usb & MUSB_INTR_RESET) && !is_host_active(musb)) {
> >> 		/* ep0 FADDR must be 0 when (re)entering peripheral mode */
> >> 		musb_ep_select(musb->mregs, 0);
> >>--
> >>2.9.3
> >>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-usb" in
> the body of a message to majordomo at vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

[PATCH v2 2/3] musb: sunxi: Remove custom babble handling

[Hans de Goede]

Hi,

On 09/22/2016 05:30 PM, Bin Liu wrote:
> Hi,
>
> On Thu, Sep 22, 2016 at 05:03:39PM +0300, Hans de Goede wrote:
>> Hi,
>>
>> On 09/22/2016 04:54 PM, Bin Liu wrote:
>>> Hi,
>>>
>>> On Thu, Sep 22, 2016 at 02:19:00PM +0300, Hans de Goede wrote:
>>>> The musb-core now a days always treats babble errors in host mode
>>>
>>> I don't think this statement is accurate. You might want to change it to
>>> "The musb core already handles babble interrupt" or something else.
>>
>> It is accurate if you look in the history at drivers/usb/musb
>> commits around 15-03-10 you will see 2 relevant commits:
>>
>> https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/drivers/usb/musb?id=b4dc38fd45b63e3da2bc98db5d283a15a637a2fa
>>
>> "usb: musb: core: simplify musb_recover_work()"
>>
>> This commits introduces calling musb_root_disconnect(musb)
>> on babble errors, that was not happening before which is why
>
> That is true, but calling musb_root_disconnect() is just one step of the
> recovery in musb core, not all.
>
> The statement of "treats babble errors in host mode as disconnects"
> implies all the babble handling is just disconnect, which is not
> accurate.

Ok, I'll send out a v3 with an improved commit msg.

> BTY, "babble errors in host mode" is also redundant. babble implies
> host mode.

Regards,

Hans




>
> Regards,
> -Bin.
>> I added the custom babble error handling. to the sunxi glue.
>>
>> And:
>>
>> "usb: musb: core: always try to recover from babble"
>>
>> Where the title says it all.
>>
>> Take these together and I believe that my commit msg:
>>
>> "The musb-core now a days always treats babble errors in host mode
>> as disconnects, so there is no need for the sunxi specific handling
>> of this anymore."
>>
>> Is quite accurate.
>>
>> Regards,
>>
>> Hans
>>
>>
>>>>
>>>> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
>>>> ---
>>>> Changes in v2:
>>>> -This is a new patch in v2 of this patch series
>>>> ---
>>>> drivers/usb/musb/sunxi.c | 10 ----------
>>>> 1 file changed, 10 deletions(-)
>>>>
>>>> diff --git a/drivers/usb/musb/sunxi.c b/drivers/usb/musb/sunxi.c
>>>> index 1408245..82eba92 100644
>>>> --- a/drivers/usb/musb/sunxi.c
>>>> +++ b/drivers/usb/musb/sunxi.c
>>>> @@ -186,16 +186,6 @@ static irqreturn_t sunxi_musb_interrupt(int irq, void *__hci)
>>>> 	if (musb->int_usb)
>>>> 		writeb(musb->int_usb, musb->mregs + SUNXI_MUSB_INTRUSB);
>>>>
>>>> -	/*
>>>> -	 * sunxi musb often signals babble on low / full speed device
>>>> -	 * disconnect, without ever raising MUSB_INTR_DISCONNECT, since
>>>> -	 * normally babble never happens treat it as disconnect.
>>>> -	 */
>>>> -	if ((musb->int_usb & MUSB_INTR_BABBLE) && is_host_active(musb)) {
>>>> -		musb->int_usb &= ~MUSB_INTR_BABBLE;
>>>> -		musb->int_usb |= MUSB_INTR_DISCONNECT;
>>>> -	}
>>>> -
>>>> 	if ((musb->int_usb & MUSB_INTR_RESET) && !is_host_active(musb)) {
>>>> 		/* ep0 FADDR must be 0 when (re)entering peripheral mode */
>>>> 		musb_ep_select(musb->mregs, 0);
>>>> --
>>>> 2.9.3
>>>>
>> --
>> To unsubscribe from this list: send the line "unsubscribe linux-usb" in
>> the body of a message to majordomo at vger.kernel.org
>> More majordomo info at  http://vger.kernel.org/majordomo-info.html

[PATCH v2 2/3] musb: sunxi: Remove custom babble handling

[Bin Liu]

On Fri, Sep 23, 2016 at 04:42:26PM +0300, Hans de Goede wrote:
> Hi,
> 
> On 09/22/2016 05:30 PM, Bin Liu wrote:
> >Hi,
> >
> >On Thu, Sep 22, 2016 at 05:03:39PM +0300, Hans de Goede wrote:
> >>Hi,
> >>
> >>On 09/22/2016 04:54 PM, Bin Liu wrote:
> >>>Hi,
> >>>
> >>>On Thu, Sep 22, 2016 at 02:19:00PM +0300, Hans de Goede wrote:
> >>>>The musb-core now a days always treats babble errors in host mode
> >>>
> >>>I don't think this statement is accurate. You might want to change it to
> >>>"The musb core already handles babble interrupt" or something else.
> >>
> >>It is accurate if you look in the history at drivers/usb/musb
> >>commits around 15-03-10 you will see 2 relevant commits:
> >>
> >>https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/drivers/usb/musb?id=b4dc38fd45b63e3da2bc98db5d283a15a637a2fa
> >>
> >>"usb: musb: core: simplify musb_recover_work()"
> >>
> >>This commits introduces calling musb_root_disconnect(musb)
> >>on babble errors, that was not happening before which is why
> >
> >That is true, but calling musb_root_disconnect() is just one step of the
> >recovery in musb core, not all.
> >
> >The statement of "treats babble errors in host mode as disconnects"
> >implies all the babble handling is just disconnect, which is not
> >accurate.
> 
> Ok, I'll send out a v3 with an improved commit msg.

Thanks.

If Kishon gives his Acked-by, I will take all the 3 patches. Or
I can just take the 2 musb patches into my tree.

Regards,
-Bin.