diff mbox

[U-Boot,2/4] usb: add 'bcm_udc_otg' support

Message ID 1421793730-8179-3-git-send-email-srae@broadcom.com
State Rejected
Delegated to: Łukasz Majewski
Headers show

Commit Message

Steve Rae Jan. 20, 2015, 10:42 p.m. UTC
Implement the UDC support for the USB OTG interface.

Signed-off-by: Steve Rae <srae@broadcom.com>
---
This commit is not checkpatch clean - however, there seems to be
no way to remove this warning:
    "warning: drivers/usb/gadget/bcm_udc_otg.c,97:
        Adding new packed members is to be done with care"

 drivers/usb/gadget/bcm_udc_otg.c | 1193 ++++++++++++++++++++++++++++++++++++++
 drivers/usb/gadget/bcm_udc_otg.h |   19 +
 2 files changed, 1212 insertions(+)
 create mode 100644 drivers/usb/gadget/bcm_udc_otg.c
 create mode 100644 drivers/usb/gadget/bcm_udc_otg.h

Comments

Marek Vasut Jan. 22, 2015, 7:05 a.m. UTC | #1
On Tuesday, January 20, 2015 at 11:42:08 PM, Steve Rae wrote:
> Implement the UDC support for the USB OTG interface.
> 
> Signed-off-by: Steve Rae <srae@broadcom.com>
> ---

General question -- this bcm controller you're adding here isn't by
any chance a DWC2 controller, or is it ? There's already a driver
for DWC2 in drivers/usb/gadget/s3c_udc_otg.c . This driver should really
be properly renamed though ;-/

If this is not DWC2, do you know what controller this is please ?

[...]

> +#define FASTBOOT_INTERFACE_CLASS	0xff
> +#define FASTBOOT_INTERFACE_SUB_CLASS	0x42
> +#define FASTBOOT_INTERFACE_PROTOCOL	0x03
> +
> +#define wfld_set(addr, fld_val, fld_mask) \
> +		(writel(((readl(addr) & ~(fld_mask)) | (fld_val)), (addr)))
> +#define wfld_clear(addr, fld_mask) \
> +		(writel((readl(addr) & ~(fld_mask)), (addr)))

The same functionality is implemented by clrsetbits32() and friends.

> +#define DEVICE_STRING_LANGUAGE_ID	0x0409 /* English (United States) */
> +
> +/*
> + * In high speed mode rx packets are 512
> + * In full speed mode rx packets are 64
> + */
> +#define RX_ENDPOINT_MAXIMUM_PACKET_SIZE		(0x0200)
> +#define TX_ENDPOINT_MAXIMUM_PACKET_SIZE		(0x0040)

No need for the brackets here.

> +#ifndef CONFIG_USB_BOARDNAME
> +#define CONFIG_USB_BOARDNAME "Board"
> +#endif
> +
> +#ifndef CONFIG_USB_CONFIGURATION
> +#define CONFIG_USB_CONFIGURATION "Fastboot"
> +#endif

What is all this stuff doing in generic USB UDC driver please ?
Or is this not a generic UDC driver ?

[...]

> +static void usb_turn_off_vdp(void)
> +{
> +	/* Check if it is standard host port (SHP) */
> +	if (readl(HSOTG_CTRL_BASE_ADDR + HSOTG_CTRL_STATUS_OFFSET) &
> +			HSOTG_CTRL_STATUS_SHP_MASK) {
> +		udelay(60000);	/* 50 ms + 20 % */

mdelay(60), this should be fixed all over the place please.

> +		/*
> +		 * force turn off VDP, enable sw_ovwr_set to take over the
> +		 * bc11 switches directly
> +		 */
> +		wfld_set(HSOTG_CTRL_BASE_ADDR + HSOTG_CTRL_CFG_OFFSET,
> +			 BC11_CFG_VDP_OFF,
> +			 HSOTG_CTRL_CFG_OVWR_KEY_MASK |
> +			 HSOTG_CTRL_CFG_SW_OVWR_EN_MASK |
> +			 HSOTG_CTRL_CFG_OVWR_SET_M0_MASK |
> +			 HSOTG_CTRL_CFG_OVWR_SET_P0_MASK);
> +
> +		udelay(160);	/* Allow time for switches to disengage */
> +	} else {
> +		udelay(120000);	/* 100 ms + 20 % */
> +	}
> +}

[...]
Steve Rae Jan. 23, 2015, 11:48 p.m. UTC | #2
On 15-01-21 11:05 PM, Marek Vasut wrote:
> On Tuesday, January 20, 2015 at 11:42:08 PM, Steve Rae wrote:
>> Implement the UDC support for the USB OTG interface.
>>
>> Signed-off-by: Steve Rae <srae@broadcom.com>
>> ---
>
> General question -- this bcm controller you're adding here isn't by
> any chance a DWC2 controller, or is it ? There's already a driver
> for DWC2 in drivers/usb/gadget/s3c_udc_otg.c . This driver should really
> be properly renamed though ;-/
>
> If this is not DWC2, do you know what controller this is please ?

yes -- it is a DWC2....

So, I have had a quick look at the s3c_udc_otg*.c code....

First observation is that there is a completely different philosophy in 
the implementation. For example, the the interrupt handler routine(s), 
Broadcom is using #defines (sysmap.h which BTW are autogenerated by our 
Device Team) whereas "S3C" is using a "s3c_usbotg_reg" structure for the 
'address' and #defines for the 'data/masks'. I'm not suggesting that one 
is better than the other, they are just different.

So, how do we proceed?
- is the ultimate goal to get to a proper gadget driver for DWC2? (I 
don't really know enough about this yet, so I apologize for these 
questions....)
- is the "S3C" code a proper 'gadget' driver and/or is it a better 
starting point to get to a gadget driver?
- I don't have enough time right now to really investigate the existing 
"S3C" and implement it on our board(s); they are significantly different 
and it looks like it will take a lot of effort - is there someone (Denx 
or community) that could assist me?
- Could we continue to review this patchset and accept it; then 
establish a team to provide a "DWC2 gadget" that could be used going 
forward?
    Note: this patchset is relatively small:
    6938ae5 usb: fastboot: implement fastboot
    M       drivers/usb/gadget/Makefile
    A       drivers/usb/gadget/bcm_usb_gadget.c
    M       include/configs/bcm28155_ap.h
    b85657d usb: update 'sysmap.h'
    M       arch/arm/include/asm/arch-bcm281xx/sysmap.h
    d590d41 usb: add 'bcm_udc_otg' support
    A       drivers/usb/gadget/bcm_udc_otg.c
    A       drivers/usb/gadget/bcm_udc_otg.h
    5f1d857 usb: gadget: fastboot: add CONFIG_FASTBOOT_NO_GADGET support
    M       drivers/usb/gadget/f_fastboot.c

Thanks in advance, Steve

>
> [...]
[... snip ...]
Marek Vasut Jan. 24, 2015, 11:47 a.m. UTC | #3
On Saturday, January 24, 2015 at 12:48:15 AM, Steve Rae wrote:
> On 15-01-21 11:05 PM, Marek Vasut wrote:
> > On Tuesday, January 20, 2015 at 11:42:08 PM, Steve Rae wrote:
> >> Implement the UDC support for the USB OTG interface.
> >> 
> >> Signed-off-by: Steve Rae <srae@broadcom.com>
> >> ---
> > 
> > General question -- this bcm controller you're adding here isn't by
> > any chance a DWC2 controller, or is it ? There's already a driver
> > for DWC2 in drivers/usb/gadget/s3c_udc_otg.c . This driver should really
> > be properly renamed though ;-/
> > 
> > If this is not DWC2, do you know what controller this is please ?
> 
> yes -- it is a DWC2....
> 
> So, I have had a quick look at the s3c_udc_otg*.c code....
> 
> First observation is that there is a completely different philosophy in
> the implementation. For example, the the interrupt handler routine(s),
> Broadcom is using #defines (sysmap.h which BTW are autogenerated by our
> Device Team) whereas "S3C" is using a "s3c_usbotg_reg" structure for the
> 'address' and #defines for the 'data/masks'. I'm not suggesting that one
> is better than the other, they are just different.

Hi!

The later is accepted by U-Boot as the preferred way, so the s3c is doing
things right.

> So, how do we proceed?

At this point, the Altera SoCFPGA and Samsung Exynos both use this s3c
driver to operate the DWC2 core in gadget mode. It should be trivial to
add support for the boardcom SoC, just follow the Altera example.

I would like to avoid situation where we have two drivers for a single
IP block.

> - is the ultimate goal to get to a proper gadget driver for DWC2? (I
> don't really know enough about this yet, so I apologize for these
> questions....)

Yes, but we already have one to my knowledge.

> - is the "S3C" code a proper 'gadget' driver and/or is it a better
> starting point to get to a gadget driver?

Yes, and it's thoroughly tested already.

> - I don't have enough time right now to really investigate the existing
> "S3C" and implement it on our board(s); they are significantly different
> and it looks like it will take a lot of effort - is there someone (Denx
> or community) that could assist me?

Sure, just start asking the questions.

> - Could we continue to review this patchset and accept it; then
> establish a team to provide a "DWC2 gadget" that could be used going
> forward?

No, unless there's a convincing technical argument that the currently
mainline DWC2 gadget driver (the s3c one) can absolutelly not be used
for the broadcom SoC, I want to avoid having two drivers for the same
IP core in mainline, sorry.

[...]

Sorry, at this point I have to push back a little.

Best regards,
Marek Vasut
Łukasz Majewski Jan. 26, 2015, 8:38 a.m. UTC | #4
Hi Marek,

> On Saturday, January 24, 2015 at 12:48:15 AM, Steve Rae wrote:
> > On 15-01-21 11:05 PM, Marek Vasut wrote:
> > > On Tuesday, January 20, 2015 at 11:42:08 PM, Steve Rae wrote:
> > >> Implement the UDC support for the USB OTG interface.
> > >> 
> > >> Signed-off-by: Steve Rae <srae@broadcom.com>
> > >> ---
> > > 
> > > General question -- this bcm controller you're adding here isn't
> > > by any chance a DWC2 controller, or is it ? There's already a
> > > driver for DWC2 in drivers/usb/gadget/s3c_udc_otg.c . This driver
> > > should really be properly renamed though ;-/
> > > 
> > > If this is not DWC2, do you know what controller this is please ?
> > 
> > yes -- it is a DWC2....
> > 
> > So, I have had a quick look at the s3c_udc_otg*.c code....
> > 
> > First observation is that there is a completely different
> > philosophy in the implementation. For example, the the interrupt
> > handler routine(s), Broadcom is using #defines (sysmap.h which BTW
> > are autogenerated by our Device Team) whereas "S3C" is using a
> > "s3c_usbotg_reg" structure for the 'address' and #defines for the
> > 'data/masks'. I'm not suggesting that one is better than the other,
> > they are just different.
> 
> Hi!
> 
> The later is accepted by U-Boot as the preferred way, so the s3c is
> doing things right.
> 
> > So, how do we proceed?
> 
> At this point, the Altera SoCFPGA and Samsung Exynos both use this s3c
> driver to operate the DWC2 core in gadget mode. It should be trivial
> to add support for the boardcom SoC, just follow the Altera example.
> 
> I would like to avoid situation where we have two drivers for a single
> IP block.
> 
> > - is the ultimate goal to get to a proper gadget driver for DWC2? (I
> > don't really know enough about this yet, so I apologize for these
> > questions....)
> 
> Yes, but we already have one to my knowledge.
> 
> > - is the "S3C" code a proper 'gadget' driver and/or is it a better
> > starting point to get to a gadget driver?
> 
> Yes, and it's thoroughly tested already.
> 
> > - I don't have enough time right now to really investigate the
> > existing "S3C" and implement it on our board(s); they are
> > significantly different and it looks like it will take a lot of
> > effort - is there someone (Denx or community) that could assist me?
> 
> Sure, just start asking the questions.
> 
> > - Could we continue to review this patchset and accept it; then
> > establish a team to provide a "DWC2 gadget" that could be used going
> > forward?
> 
> No, unless there's a convincing technical argument that the currently
> mainline DWC2 gadget driver (the s3c one) can absolutelly not be used
> for the broadcom SoC, I want to avoid having two drivers for the same
> IP core in mainline, sorry.

Maybe it is a highest time to think about renaming s3c_* to dwc2_* and
avoid further confusion.

> 
> [...]
> 
> Sorry, at this point I have to push back a little.
> 
> Best regards,
> Marek Vasut
Marek Vasut Jan. 26, 2015, 12:12 p.m. UTC | #5
On Monday, January 26, 2015 at 09:38:28 AM, Lukasz Majewski wrote:
> Hi Marek,

Hi!

[...]

> > No, unless there's a convincing technical argument that the currently
> > mainline DWC2 gadget driver (the s3c one) can absolutelly not be used
> > for the broadcom SoC, I want to avoid having two drivers for the same
> > IP core in mainline, sorry.
> 
> Maybe it is a highest time to think about renaming s3c_* to dwc2_* and
> avoid further confusion.

The time was ripe for a while now ;-) Do you want to do it or shall I send
a patch ?

Best regards,
Marek Vasut
Łukasz Majewski Jan. 26, 2015, 12:31 p.m. UTC | #6
Hi Marek,

> On Monday, January 26, 2015 at 09:38:28 AM, Lukasz Majewski wrote:
> > Hi Marek,
> 
> Hi!
> 
> [...]
> 
> > > No, unless there's a convincing technical argument that the
> > > currently mainline DWC2 gadget driver (the s3c one) can
> > > absolutelly not be used for the broadcom SoC, I want to avoid
> > > having two drivers for the same IP core in mainline, sorry.
> > 
> > Maybe it is a highest time to think about renaming s3c_* to dwc2_*
> > and avoid further confusion.
> 
> The time was ripe for a while now ;-) Do you want to do it or shall I
> send a patch ?

I do know that I will test it :-), so feel free to send the patch.

> 
> Best regards,
> Marek Vasut
Steve Rae Jan. 26, 2015, 5:44 p.m. UTC | #7
On 15-01-24 03:47 AM, Marek Vasut wrote:
> On Saturday, January 24, 2015 at 12:48:15 AM, Steve Rae wrote:
>> On 15-01-21 11:05 PM, Marek Vasut wrote:
>>> On Tuesday, January 20, 2015 at 11:42:08 PM, Steve Rae wrote:
>>>> Implement the UDC support for the USB OTG interface.
>>>>
>>>> Signed-off-by: Steve Rae <srae@broadcom.com>
>>>> ---
>>>
>>> General question -- this bcm controller you're adding here isn't by
>>> any chance a DWC2 controller, or is it ? There's already a driver
>>> for DWC2 in drivers/usb/gadget/s3c_udc_otg.c . This driver should really
>>> be properly renamed though ;-/
>>>
>>> If this is not DWC2, do you know what controller this is please ?
>>
>> yes -- it is a DWC2....
>>
>> So, I have had a quick look at the s3c_udc_otg*.c code....
>>
>> First observation is that there is a completely different philosophy in
>> the implementation. For example, the the interrupt handler routine(s),
>> Broadcom is using #defines (sysmap.h which BTW are autogenerated by our
>> Device Team) whereas "S3C" is using a "s3c_usbotg_reg" structure for the
>> 'address' and #defines for the 'data/masks'. I'm not suggesting that one
>> is better than the other, they are just different.
>
> Hi!
>
> The later is accepted by U-Boot as the preferred way, so the s3c is doing
> things right.
>
>> So, how do we proceed?
>
> At this point, the Altera SoCFPGA and Samsung Exynos both use this s3c
> driver to operate the DWC2 core in gadget mode. It should be trivial to
> add support for the boardcom SoC, just follow the Altera example.
>
> I would like to avoid situation where we have two drivers for a single
> IP block.
>
>> - is the ultimate goal to get to a proper gadget driver for DWC2? (I
>> don't really know enough about this yet, so I apologize for these
>> questions....)
>
> Yes, but we already have one to my knowledge.
>
>> - is the "S3C" code a proper 'gadget' driver and/or is it a better
>> starting point to get to a gadget driver?
>
> Yes, and it's thoroughly tested already.
>
>> - I don't have enough time right now to really investigate the existing
>> "S3C" and implement it on our board(s); they are significantly different
>> and it looks like it will take a lot of effort - is there someone (Denx
>> or community) that could assist me?
>
> Sure, just start asking the questions.
>
>> - Could we continue to review this patchset and accept it; then
>> establish a team to provide a "DWC2 gadget" that could be used going
>> forward?
>
> No, unless there's a convincing technical argument that the currently
> mainline DWC2 gadget driver (the s3c one) can absolutelly not be used
> for the broadcom SoC, I want to avoid having two drivers for the same
> IP core in mainline, sorry.
>
> [...]
>
> Sorry, at this point I have to push back a little.
OK -- Thanks for this information....
Steve
>
> Best regards,
> Marek Vasut
>
Marek Vasut Jan. 27, 2015, 11 p.m. UTC | #8
On Monday, January 26, 2015 at 06:44:14 PM, Steve Rae wrote:
[...]
> > No, unless there's a convincing technical argument that the currently
> > mainline DWC2 gadget driver (the s3c one) can absolutelly not be used
> > for the broadcom SoC, I want to avoid having two drivers for the same
> > IP core in mainline, sorry.
> > 
> > [...]
> > 
> > Sorry, at this point I have to push back a little.
> 
> OK -- Thanks for this information....

Hi!

do you plan to rework the series on top of the dwc2 (s3c) driver
by any chance please ?

Best regards,
Marek Vasut
Steve Rae Jan. 28, 2015, 7:04 p.m. UTC | #9
On 15-01-27 03:00 PM, Marek Vasut wrote:
> On Monday, January 26, 2015 at 06:44:14 PM, Steve Rae wrote:
> [...]
>>> No, unless there's a convincing technical argument that the currently
>>> mainline DWC2 gadget driver (the s3c one) can absolutelly not be used
>>> for the broadcom SoC, I want to avoid having two drivers for the same
>>> IP core in mainline, sorry.
>>>
>>> [...]
>>>
>>> Sorry, at this point I have to push back a little.
>>
>> OK -- Thanks for this information....
>
> Hi!
>
> do you plan to rework the series on top of the dwc2 (s3c) driver
> by any chance please ?
>
> Best regards,
> Marek Vasut
>

I have it on my TODO list, but I am not going to be able to work on it 
(for weeks; if not months...) Sorry.
But if someone else would rework the code, I can probably find some time 
to test it....
Thanks, Steve
Marek Vasut Jan. 30, 2015, 10:16 p.m. UTC | #10
On Monday, January 26, 2015 at 01:31:39 PM, Lukasz Majewski wrote:
> Hi Marek,
> 
> > On Monday, January 26, 2015 at 09:38:28 AM, Lukasz Majewski wrote:
> > > Hi Marek,
> > 
> > Hi!
> > 
> > [...]
> > 
> > > > No, unless there's a convincing technical argument that the
> > > > currently mainline DWC2 gadget driver (the s3c one) can
> > > > absolutelly not be used for the broadcom SoC, I want to avoid
> > > > having two drivers for the same IP core in mainline, sorry.
> > > 
> > > Maybe it is a highest time to think about renaming s3c_* to dwc2_*
> > > and avoid further confusion.
> > 
> > The time was ripe for a while now ;-) Do you want to do it or shall I
> > send a patch ?
> 
> I do know that I will test it :-), so feel free to send the patch.

Hi!

I'm dead busy until end of next week. Do you mind cooking such patch please?

Best regards,
Marek Vasut
Marek Vasut Jan. 30, 2015, 10:17 p.m. UTC | #11
On Wednesday, January 28, 2015 at 08:04:45 PM, Steve Rae wrote:
> On 15-01-27 03:00 PM, Marek Vasut wrote:
> > On Monday, January 26, 2015 at 06:44:14 PM, Steve Rae wrote:
> > [...]
> > 
> >>> No, unless there's a convincing technical argument that the currently
> >>> mainline DWC2 gadget driver (the s3c one) can absolutelly not be used
> >>> for the broadcom SoC, I want to avoid having two drivers for the same
> >>> IP core in mainline, sorry.
> >>> 
> >>> [...]
> >>> 
> >>> Sorry, at this point I have to push back a little.
> >> 
> >> OK -- Thanks for this information....
> > 
> > Hi!
> > 
> > do you plan to rework the series on top of the dwc2 (s3c) driver
> > by any chance please ?
> > 
> > Best regards,
> > Marek Vasut
> 
> I have it on my TODO list, but I am not going to be able to work on it
> (for weeks; if not months...) Sorry.
> But if someone else would rework the code, I can probably find some time
> to test it....

Aw, that's a bit sad. Porting the Altera stuff was a matter of a couple
of hours, are you sure you won't give it a spin please ? :)

Best regards,
Marek Vasut
Lukasz Majewski Jan. 31, 2015, 1:26 p.m. UTC | #12
On Fri, 30 Jan 2015 23:16:57 +0100
Marek Vasut <marex@denx.de> wrote:

> On Monday, January 26, 2015 at 01:31:39 PM, Lukasz Majewski wrote:
> > Hi Marek,
> > 
> > > On Monday, January 26, 2015 at 09:38:28 AM, Lukasz Majewski wrote:
> > > > Hi Marek,
> > > 
> > > Hi!
> > > 
> > > [...]
> > > 
> > > > > No, unless there's a convincing technical argument that the
> > > > > currently mainline DWC2 gadget driver (the s3c one) can
> > > > > absolutelly not be used for the broadcom SoC, I want to avoid
> > > > > having two drivers for the same IP core in mainline, sorry.
> > > > 
> > > > Maybe it is a highest time to think about renaming s3c_* to
> > > > dwc2_* and avoid further confusion.
> > > 
> > > The time was ripe for a while now ;-) Do you want to do it or
> > > shall I send a patch ?
> > 
> > I do know that I will test it :-), so feel free to send the patch.
> 
> Hi!
> 
> I'm dead busy until end of next week. Do you mind cooking such patch
> please?

I will try to find some time slack to prepare patch.

> 
> Best regards,
> Marek Vasut
> _______________________________________________
> U-Boot mailing list
> U-Boot@lists.denx.de
> http://lists.denx.de/mailman/listinfo/u-boot
Marek Vasut Jan. 31, 2015, 2:15 p.m. UTC | #13
On Saturday, January 31, 2015 at 02:26:16 PM, Lukasz Majewski wrote:
> On Fri, 30 Jan 2015 23:16:57 +0100
> 
> Marek Vasut <marex@denx.de> wrote:
> > On Monday, January 26, 2015 at 01:31:39 PM, Lukasz Majewski wrote:
> > > Hi Marek,
> > > 
> > > > On Monday, January 26, 2015 at 09:38:28 AM, Lukasz Majewski wrote:
> > > > > Hi Marek,
> > > > 
> > > > Hi!
> > > > 
> > > > [...]
> > > > 
> > > > > > No, unless there's a convincing technical argument that the
> > > > > > currently mainline DWC2 gadget driver (the s3c one) can
> > > > > > absolutelly not be used for the broadcom SoC, I want to avoid
> > > > > > having two drivers for the same IP core in mainline, sorry.
> > > > > 
> > > > > Maybe it is a highest time to think about renaming s3c_* to
> > > > > dwc2_* and avoid further confusion.
> > > > 
> > > > The time was ripe for a while now ;-) Do you want to do it or
> > > > shall I send a patch ?
> > > 
> > > I do know that I will test it :-), so feel free to send the patch.
> > 
> > Hi!
> > 
> > I'm dead busy until end of next week. Do you mind cooking such patch
> > please?
> 
> I will try to find some time slack to prepare patch.

Thanks! Shall I pick it then ?

Best regards,
Marek Vasut
Lukasz Majewski Feb. 1, 2015, 1:15 p.m. UTC | #14
On Sat, 31 Jan 2015 15:15:24 +0100
Marek Vasut <marex@denx.de> wrote:

> On Saturday, January 31, 2015 at 02:26:16 PM, Lukasz Majewski wrote:
> > On Fri, 30 Jan 2015 23:16:57 +0100
> > 
> > Marek Vasut <marex@denx.de> wrote:
> > > On Monday, January 26, 2015 at 01:31:39 PM, Lukasz Majewski wrote:
> > > > Hi Marek,
> > > > 
> > > > > On Monday, January 26, 2015 at 09:38:28 AM, Lukasz Majewski
> > > > > wrote:
> > > > > > Hi Marek,
> > > > > 
> > > > > Hi!
> > > > > 
> > > > > [...]
> > > > > 
> > > > > > > No, unless there's a convincing technical argument that
> > > > > > > the currently mainline DWC2 gadget driver (the s3c one)
> > > > > > > can absolutelly not be used for the broadcom SoC, I want
> > > > > > > to avoid having two drivers for the same IP core in
> > > > > > > mainline, sorry.
> > > > > > 
> > > > > > Maybe it is a highest time to think about renaming s3c_* to
> > > > > > dwc2_* and avoid further confusion.
> > > > > 
> > > > > The time was ripe for a while now ;-) Do you want to do it or
> > > > > shall I send a patch ?
> > > > 
> > > > I do know that I will test it :-), so feel free to send the
> > > > patch.
> > > 
> > > Hi!
> > > 
> > > I'm dead busy until end of next week. Do you mind cooking such
> > > patch please?
> > 
> > I will try to find some time slack to prepare patch.
> 
> Thanks! Shall I pick it then ?

For technical (testing) point of view it would be better for me to
place it -dfu tree, and then send a PR to you.


Best regards,
LUkasz Majewski

> 
> Best regards,
> Marek Vasut
Marek Vasut Feb. 1, 2015, 4:21 p.m. UTC | #15
On Sunday, February 01, 2015 at 02:15:45 PM, Lukasz Majewski wrote:
> On Sat, 31 Jan 2015 15:15:24 +0100
> 
> Marek Vasut <marex@denx.de> wrote:
> > On Saturday, January 31, 2015 at 02:26:16 PM, Lukasz Majewski wrote:
> > > On Fri, 30 Jan 2015 23:16:57 +0100
> > > 
> > > Marek Vasut <marex@denx.de> wrote:
> > > > On Monday, January 26, 2015 at 01:31:39 PM, Lukasz Majewski wrote:
> > > > > Hi Marek,
> > > > > 
> > > > > > On Monday, January 26, 2015 at 09:38:28 AM, Lukasz Majewski
> > > > > > 
> > > > > > wrote:
> > > > > > > Hi Marek,
> > > > > > 
> > > > > > Hi!
> > > > > > 
> > > > > > [...]
> > > > > > 
> > > > > > > > No, unless there's a convincing technical argument that
> > > > > > > > the currently mainline DWC2 gadget driver (the s3c one)
> > > > > > > > can absolutelly not be used for the broadcom SoC, I want
> > > > > > > > to avoid having two drivers for the same IP core in
> > > > > > > > mainline, sorry.
> > > > > > > 
> > > > > > > Maybe it is a highest time to think about renaming s3c_* to
> > > > > > > dwc2_* and avoid further confusion.
> > > > > > 
> > > > > > The time was ripe for a while now ;-) Do you want to do it or
> > > > > > shall I send a patch ?
> > > > > 
> > > > > I do know that I will test it :-), so feel free to send the
> > > > > patch.
> > > > 
> > > > Hi!
> > > > 
> > > > I'm dead busy until end of next week. Do you mind cooking such
> > > > patch please?
> > > 
> > > I will try to find some time slack to prepare patch.
> > 
> > Thanks! Shall I pick it then ?
> 
> For technical (testing) point of view it would be better for me to
> place it -dfu tree, and then send a PR to you.

Roger :)

Best regards,
Marek Vasut
diff mbox

Patch

diff --git a/drivers/usb/gadget/bcm_udc_otg.c b/drivers/usb/gadget/bcm_udc_otg.c
new file mode 100644
index 0000000..706e003
--- /dev/null
+++ b/drivers/usb/gadget/bcm_udc_otg.c
@@ -0,0 +1,1193 @@ 
+/*
+ * Copyright 2015 Broadcom Corporation.
+ *
+ * SPDX-License-Identifier:	GPL-2.0+
+ */
+
+#include <config.h>
+#include <asm/io.h>
+#include <asm/arch/sysmap.h>
+#include <asm/kona-common/clk.h>
+#include "bcm_udc_otg.h"
+
+#ifdef CONFIG_CMD_SERIALNO
+#include "serial_no.h"
+#endif
+
+#include <usbdevice.h>
+
+#define DEVICE_STRING_LANGUAGE_ID_INDEX		0
+#define DEVICE_STRING_PRODUCT_INDEX		1
+#define DEVICE_STRING_SERIAL_NUMBER_INDEX	2
+#define DEVICE_STRING_CONFIG_INDEX		3
+#define DEVICE_STRING_INTERFACE_INDEX		4
+#define DEVICE_STRING_MANUFACTURER_INDEX	5
+#define DEVICE_STRING_MAX_INDEX		DEVICE_STRING_MANUFACTURER_INDEX
+
+#define FASTBOOT_INTERFACE_CLASS	0xff
+#define FASTBOOT_INTERFACE_SUB_CLASS	0x42
+#define FASTBOOT_INTERFACE_PROTOCOL	0x03
+
+#define wfld_set(addr, fld_val, fld_mask) \
+		(writel(((readl(addr) & ~(fld_mask)) | (fld_val)), (addr)))
+#define wfld_clear(addr, fld_mask) \
+		(writel((readl(addr) & ~(fld_mask)), (addr)))
+
+#define DEVICE_STRING_LANGUAGE_ID	0x0409 /* English (United States) */
+
+/*
+ * In high speed mode rx packets are 512
+ * In full speed mode rx packets are 64
+ */
+#define RX_ENDPOINT_MAXIMUM_PACKET_SIZE		(0x0200)
+#define TX_ENDPOINT_MAXIMUM_PACKET_SIZE		(0x0040)
+
+#ifndef CONFIG_USB_BOARDNAME
+#define CONFIG_USB_BOARDNAME "Board"
+#endif
+
+#ifndef CONFIG_USB_CONFIGURATION
+#define CONFIG_USB_CONFIGURATION "Fastboot"
+#endif
+
+#ifndef CONFIG_USB_INTERFACE
+#define CONFIG_USB_INTERFACE "Fastboot"
+#endif
+
+#ifndef CONFIG_USB_SERIALNO
+#define CONFIG_USB_SERIALNO "1234567890"
+#endif
+
+#ifndef CONFIG_USBID_ADDR
+#error	CONFIG_USBID_ADDR must be defined!
+#endif
+
+static char *usb_device_strings[DEVICE_STRING_MANUFACTURER_INDEX + 1] = {
+	/* Language =		*/	"",
+	/* Product =		*/	CONFIG_USB_BOARDNAME,
+	/* Serial =		*/	CONFIG_USB_SERIALNO,
+	/* Config =		*/	CONFIG_USB_CONFIGURATION,
+	/* Interface =		*/	CONFIG_USB_INTERFACE,
+	/* Manufacturer =	*/	CONFIG_G_DNL_MANUFACTURER,
+};
+
+static struct usb_device_descriptor dfu_dev_descriptor
+				__aligned(4) = {
+	.bLength			= 0x12,
+	.bDescriptorType		= USB_DT_DEVICE,
+	.bcdUSB				= 0x100,
+	.bDeviceClass			= 0xff,
+	.bDeviceSubClass		= 0xff,
+	.bDeviceProtocol		= 0xff,
+	.bMaxPacketSize0		= 64,	/* depends on enum speed */
+	.idVendor			= CONFIG_G_DNL_VENDOR_NUM,
+	.idProduct			= CONFIG_G_DNL_PRODUCT_NUM,
+	.bcdDevice			= 0x0001,
+	.iManufacturer			= DEVICE_STRING_MANUFACTURER_INDEX,
+	.iProduct			= DEVICE_STRING_PRODUCT_INDEX,
+	.iSerialNumber			= DEVICE_STRING_SERIAL_NUMBER_INDEX,
+	.bNumConfigurations		= 0x1,
+};
+
+struct full_configuration_descriptor {
+	struct usb_configuration_descriptor c;
+	struct usb_interface_descriptor i;
+	struct usb_endpoint_descriptor e1;
+	struct usb_endpoint_descriptor e2;
+} __packed;
+
+static struct full_configuration_descriptor full_desc
+				__aligned(4) = {
+.c = {
+	.bLength		= 0x9,
+	.bDescriptorType	= USB_DT_CONFIG,
+	.wTotalLength		= sizeof(struct usb_configuration_descriptor) +
+				  sizeof(struct usb_interface_descriptor) +
+				  sizeof(struct usb_endpoint_descriptor) +
+				  sizeof(struct usb_endpoint_descriptor),
+	.bNumInterfaces		= 1,
+	.bConfigurationValue	= 1,
+	.iConfiguration		= DEVICE_STRING_CONFIG_INDEX,
+	.bmAttributes		= BMATTRIBUTE_RESERVED |
+				  BMATTRIBUTE_SELF_POWERED,
+	.bMaxPower		= 0,
+},
+.i = {
+	.bLength		= 0x9,
+	.bDescriptorType	= USB_DT_INTERFACE,
+	.bInterfaceNumber	= 0x0,
+	.bAlternateSetting	= 0x0,	/* CheckMe */
+	.bNumEndpoints		= 0x2,
+	.bInterfaceClass	= FASTBOOT_INTERFACE_CLASS,
+	.bInterfaceSubClass	= FASTBOOT_INTERFACE_SUB_CLASS,
+	.bInterfaceProtocol	= FASTBOOT_INTERFACE_PROTOCOL,
+	.iInterface		= DEVICE_STRING_INTERFACE_INDEX,
+},
+.e1 = {
+	.bLength		= sizeof(struct usb_endpoint_descriptor),
+	.bDescriptorType	= USB_DT_ENDPOINT,
+	.bEndpointAddress	= USB_DIR_IN | 1,
+	.bmAttributes		= USB_ENDPOINT_XFER_BULK,
+	.wMaxPacketSize		= TX_ENDPOINT_MAXIMUM_PACKET_SIZE,
+	.bInterval		= 0x00,
+},
+.e2 = {
+	.bLength		= sizeof(struct usb_endpoint_descriptor),
+	.bDescriptorType	= USB_DT_ENDPOINT,
+	.bEndpointAddress	= USB_DIR_OUT | 1,
+	.bmAttributes		= USB_ENDPOINT_XFER_BULK,
+	.wMaxPacketSize		= RX_ENDPOINT_MAXIMUM_PACKET_SIZE,
+	.bInterval		= 0x00,
+},
+};
+
+static rx_type rx_callback;
+
+#define EP0_SETUP_BUF_SIZE		16	/* hold 2 setup packets */
+#define EP0_IN_BUF_SIZE			128
+#define EP0_OUT_BUF_SIZE		16
+
+#define RX_FIFO_SIZE			512		/* 512 x 4 bytes */
+#define TX_FIFO_SIZE			256		/* 256 x 4 bytes */
+#define TX_FIFO_OFFSET			RX_FIFO_SIZE
+#define TX_FIFO1_SIZE			256		/* 256 x 4 bytes */
+#define TX_FIFO1_OFFSET			(TX_FIFO_SIZE + TX_FIFO_OFFSET)
+
+#define TURN_AROUND_TIME		9
+
+#define EP_MISMATCH_CNT			1
+#define PERIODIC_FRM_INTERVAL		1
+#define BC11_CFG_VDP_OFF		0x55570000
+
+#define EP1_IN_BUF_SIZE			512
+#define EP1_OUT_BUF_SIZE		512
+
+#define HSOTG_CTRL_STATUS_OFFSET HSOTG_CTRL_BC_STATUS_OFFSET
+#define HSOTG_CTRL_STATUS_SHP_MASK HSOTG_CTRL_BC_STATUS_SDP_MASK
+#define HSOTG_CTRL_CFG_OFFSET HSOTG_CTRL_BC_CFG_OFFSET
+#define HSOTG_CTRL_CFG_OVWR_KEY_MASK HSOTG_CTRL_BC_CFG_BC_OVWR_KEY_MASK
+#define HSOTG_CTRL_CFG_SW_OVWR_EN_MASK HSOTG_CTRL_BC_CFG_SW_OVWR_EN_MASK
+#define HSOTG_CTRL_CFG_OVWR_SET_M0_MASK HSOTG_CTRL_BC_CFG_BC_OVWR_SET_M0_MASK
+#define HSOTG_CTRL_CFG_OVWR_SET_P0_MASK HSOTG_CTRL_BC_CFG_BC_OVWR_SET_P0_MASK
+
+/* Align buffers to 32 bytes so cache invalidate/flush routines work. */
+static uint32_t ep0_setup_buf[EP0_SETUP_BUF_SIZE]	__aligned(32);
+static uint32_t ep0_out_buf[EP0_OUT_BUF_SIZE]		__aligned(32);
+static uint32_t ep0_in_buf[EP0_IN_BUF_SIZE]		__aligned(32);
+
+static uint8_t ep1_out_buf[2][EP1_OUT_BUF_SIZE]		__aligned(32);
+static uint8_t ep1_in_buf[EP1_IN_BUF_SIZE]		__aligned(32);
+
+static int ep1_out_buf_sel;
+static uint32_t usb_speed;
+
+/**
+ * @brief: usb_soft_reset - Soft Reset USB Core.
+ * @param[in]:	none
+ * @param[out]:	none
+ * @return: 0
+ */
+static int usb_soft_reset(void)
+{
+	uint32_t val;
+
+	/* Add hclk soft reset after interface setting */
+	writel(HSOTG_RSTCTL_CSFTRST_MASK,
+	       HSOTG_BASE_ADDR + HSOTG_RSTCTL_OFFSET);
+
+	udelay(1000);
+
+	/* Poll until Reset complete and AHB idle */
+	do {
+		val = readl(HSOTG_BASE_ADDR + HSOTG_RSTCTL_OFFSET);
+	} while ((val & HSOTG_RSTCTL_CSFTRST_MASK) ||
+		 (!(val & HSOTG_RSTCTL_AHBIDLE_MASK)));
+
+	return 0;
+}
+
+/**
+ * @brief: setup_device_fifo - Configure USB FIFO.
+ * @param[in]:	none
+ * @param[out]:	none
+ * @return: none
+ */
+static void setup_device_fifo(void)
+{
+	/* Receive FIFO size register;
+512*4bytes, number is in terms of 32bits */
+	writel((RX_FIFO_SIZE << HSOTG_RXFSIZ_RXFDEP_SHIFT),
+	       HSOTG_BASE_ADDR + HSOTG_RXFSIZ_OFFSET);
+
+	/*
+	 * Receive FIFO size register
+	 *  256*4bytes -- Tx FIFO depth
+	 *  512*4bytes -- TX RAM start address
+	 *  (number is in terms of 32bits)
+	 */
+	writel((TX_FIFO_SIZE << HSOTG_NPTXFSIZ_NPTXFDEP_SHIFT) |
+	       (TX_FIFO_OFFSET << HSOTG_NPTXFSIZ_NPTXFSTADDR_SHIFT),
+	       HSOTG_BASE_ADDR + HSOTG_NPTXFSIZ_OFFSET);
+
+	writel((TX_FIFO1_SIZE << HSOTG_DIEPTXF1_INEPNTXFDEP_SHIFT) |
+	       (TX_FIFO1_OFFSET << HSOTG_DIEPTXF1_INEPNTXFSTADDR_SHIFT),
+	       HSOTG_BASE_ADDR + HSOTG_DIEPTXF1_OFFSET);
+}
+
+/**
+ * @brief: dfu_setup_device_mode - Configure USB device mode.
+ * @param[in]:	none
+ * @param[out]:	none
+ * @return: 0 or ERROR ID
+ */
+static int dfu_setup_device_mode(void)
+{
+	setup_device_fifo();
+
+	writel(HSOTG_AHBCFG_NPTXFEMPLVL_MASK |
+				/* int indicates TXFIFO empty */
+	       HSOTG_AHBCFG_DMAEN_MASK |
+				/* DMA mode */
+	       HSOTG_AHBCFG_GLBLINTRMSK_MASK,
+				/* Unmask the interrupt assertion */
+	       HSOTG_BASE_ADDR + HSOTG_AHBCFG_OFFSET);
+
+	writel(TURN_AROUND_TIME << HSOTG_CFG_USBTRDTIM_SHIFT,
+	       HSOTG_BASE_ADDR + HSOTG_CFG_OFFSET);
+
+	/* UTMI+ 8bit interface */
+	wfld_clear(HSOTG_BASE_ADDR + HSOTG_CFG_OFFSET,
+		   HSOTG_CFG_ULPI_UTMI_SEL_MASK);
+
+	writel((EP_MISMATCH_CNT << HSOTG_DCFG_EPMISCNT_SHIFT) |
+				/* IN Mismatch Cnt */
+	       (PERIODIC_FRM_INTERVAL << HSOTG_DCFG_PERFRINT_SHIFT),
+				/* Frame Interval */
+	       HSOTG_BASE_ADDR + HSOTG_DCFG_OFFSET);
+
+	/* check if OTG is in device mode */
+	if (readl(HSOTG_BASE_ADDR + HSOTG_INTSTS_OFFSET) &
+		  HSOTG_INTSTS_CURMOD_MASK) {
+		error("Not in device mode");
+		return 1;
+	}
+
+	return 0;
+}
+
+/**
+ * @brief: usb_clr_interrupt - Clear USB interrupt.
+ * @param[in]:	none
+ * @param[out]:	none
+ * @return: 0
+ */
+static int usb_clr_interrupt(void)
+{
+	/* clear all interrupts */
+	writel(0xffffffff, HSOTG_BASE_ADDR + HSOTG_INTSTS_OFFSET);
+	return 0;
+}
+
+/**
+ * @brief: usb_phy_connect - Connect USB Phy to host.
+ * @param[in]:	none
+ * @param[out]:	none
+ * @return: 0
+ */
+static int usb_phy_connect(void)
+{
+	/* set Phy to driving mode */
+	wfld_clear(HSOTG_CTRL_BASE_ADDR + HSOTG_CTRL_PHY_P1CTL_OFFSET,
+		   HSOTG_CTRL_PHY_P1CTL_NON_DRIVING_MASK);
+
+	udelay(100);
+
+	/* Clear Soft Disconnect */
+	wfld_clear(HSOTG_BASE_ADDR + HSOTG_DCTL_OFFSET,
+		   HSOTG_DCTL_SFTDISCON_MASK);
+
+	/* software reset Phy, active low */
+	wfld_clear(HSOTG_CTRL_BASE_ADDR + HSOTG_CTRL_PHY_P1CTL_OFFSET,
+		   HSOTG_CTRL_PHY_P1CTL_SOFT_RESET_MASK);
+
+	udelay(10000);
+
+	/* */
+	wfld_set(HSOTG_CTRL_BASE_ADDR + HSOTG_CTRL_PHY_P1CTL_OFFSET,
+		 HSOTG_CTRL_PHY_P1CTL_SOFT_RESET_MASK,
+		 HSOTG_CTRL_PHY_P1CTL_SOFT_RESET_MASK);
+
+	return 0;
+}
+
+/**
+ * @brief: usb_phy_disconnect - disconnect USB Phy from Host.
+ * @param[in]:	none
+ * @param[out]:	none
+ * @return: 0
+ */
+static int usb_phy_disconnect(void)
+{
+	/* Soft Disconnect */
+	wfld_set(HSOTG_BASE_ADDR + HSOTG_DCTL_OFFSET,
+		 HSOTG_DCTL_SFTDISCON_MASK,
+		 HSOTG_DCTL_SFTDISCON_MASK);
+
+	/* set Phy to non-driving (reset) mode */
+	wfld_set(HSOTG_CTRL_BASE_ADDR + HSOTG_CTRL_PHY_P1CTL_OFFSET,
+		 HSOTG_CTRL_PHY_P1CTL_NON_DRIVING_MASK,
+		 HSOTG_CTRL_PHY_P1CTL_NON_DRIVING_MASK);
+	return 0;
+}
+
+/**
+ * @brief: usb_wait_for_vbus - wait for vbus turning 5.0v by polling STAT2.
+ * @param[in]:	none
+ * @param[out]:	none
+ * @return: none
+ */
+static void usb_wait_for_vbus(void)
+{
+	uint32_t val;
+
+	/*
+	 * If there is no PMU, then the VBUS signal from the connector will
+	 * not necessarily be connected to STAT2. We can get around this by
+	 * telling the usb core to proceed irregardless by triggering the
+	 * core with internal STAT1/STAT2 signals in software.
+	 */
+	val = readl(HSOTG_CTRL_BASE_ADDR + HSOTG_CTRL_USBOTGCONTROL_OFFSET);
+	val |= (HSOTG_CTRL_USBOTGCONTROL_REG_OTGSTAT2_MASK |
+		HSOTG_CTRL_USBOTGCONTROL_REG_OTGSTAT1_MASK |
+		HSOTG_CTRL_USBOTGCONTROL_OTGSTAT_CTRL_MASK);
+	writel(val, HSOTG_CTRL_BASE_ADDR + HSOTG_CTRL_USBOTGCONTROL_OFFSET);
+}
+
+/**
+ * @brief: usb_turn_off_vdp - disable vdp.
+ * @param[in]:	none
+ * @param[out]:	none
+ * @return: none
+ */
+static void usb_turn_off_vdp(void)
+{
+	/* Check if it is standard host port (SHP) */
+	if (readl(HSOTG_CTRL_BASE_ADDR + HSOTG_CTRL_STATUS_OFFSET) &
+			HSOTG_CTRL_STATUS_SHP_MASK) {
+		udelay(60000);	/* 50 ms + 20 % */
+
+		/*
+		 * force turn off VDP, enable sw_ovwr_set to take over the
+		 * bc11 switches directly
+		 */
+		wfld_set(HSOTG_CTRL_BASE_ADDR + HSOTG_CTRL_CFG_OFFSET,
+			 BC11_CFG_VDP_OFF,
+			 HSOTG_CTRL_CFG_OVWR_KEY_MASK |
+			 HSOTG_CTRL_CFG_SW_OVWR_EN_MASK |
+			 HSOTG_CTRL_CFG_OVWR_SET_M0_MASK |
+			 HSOTG_CTRL_CFG_OVWR_SET_P0_MASK);
+
+		udelay(160);	/* Allow time for switches to disengage */
+	} else {
+		udelay(120000);	/* 100 ms + 20 % */
+	}
+}
+
+/**
+ * @brief: usb_ep0_setup_prime - Prepare receiving EP0 Setup packet.
+ * @param[in]:	none
+ * @param[out]:	none
+ * @return: 0
+ */
+static int usb_ep0_setup_prime(void)
+{
+	uint32_t mask;
+
+	invalidate_dcache_range((uint32_t)ep0_setup_buf,
+				((uint32_t)ep0_setup_buf +
+				 sizeof(ep0_setup_buf)));
+
+	/* Device OUT Endpoint 0 DMA Address Register */
+	writel((uint32_t)ep0_setup_buf,
+	       HSOTG_BASE_ADDR + HSOTG_DOEPDMA0_OFFSET);
+
+	/* Device OUT Endpoint 0 Transfer Size Register */
+	writel((0x3 << HSOTG_DOEPTSIZ0_SUPCNT_SHIFT) |	/* 3 SUPCnt */
+	       (0x3 << HSOTG_DOEPTSIZ0_PKTCNT_SHIFT) |	/* 3 PktCnt */
+	       (0x8 << HSOTG_DOEPTSIZ0_XFERSIZE_SHIFT),	/* XferSize 8 bytes */
+	       HSOTG_BASE_ADDR + HSOTG_DOEPTSIZ0_OFFSET);
+
+	/* Device OUT Endpoint 0 Control Register */
+	mask =	HSOTG_DOEPCTL0_EPENA_MASK |		/* Endpoint Enable */
+		HSOTG_DOEPCTL0_CNAK_MASK;		/* clear NAK bit */
+	wfld_set(HSOTG_BASE_ADDR + HSOTG_DOEPCTL0_OFFSET, mask, mask);
+
+	return 0;
+}
+
+/**
+ * @brief: usb_ep0_recv_data - Receive packet on EP0-OUT endpoint.
+ * @param[in]:	uint8_t *data_buf - Data packet buffer
+ * @param[in]:	uint32_t cnt - Number of bytes to receive
+ * @param[out]:	none
+ * @return: 0 or ERROR ID
+ */
+static int usb_ep0_recv_data(uint8_t *data_buf, uint32_t cnt)
+{
+	uint32_t mask;
+
+	if (cnt > sizeof(ep0_out_buf)) {
+		printf("EP0_OUT overflow\n");
+		return 1;
+	}
+
+	invalidate_dcache_range((uint32_t)ep0_out_buf,
+				(uint32_t)ep0_out_buf + sizeof(ep0_out_buf));
+
+	/* Device OUT Endpoint 0 DMA Address Register */
+	writel((uint32_t)ep0_out_buf,
+	       HSOTG_BASE_ADDR + HSOTG_DOEPDMA0_OFFSET);
+
+	/* Device OUT Endpoint 0 Transfer Size Register */
+	writel((1 << HSOTG_DOEPTSIZ0_PKTCNT_SHIFT) | cnt,
+	       HSOTG_BASE_ADDR + HSOTG_DOEPTSIZ0_OFFSET);
+
+	/* Device OUT Endpoint 0 Control Register */
+	mask =	HSOTG_DOEPCTL0_EPENA_MASK |		/* Endpoint Enable */
+		HSOTG_DOEPCTL0_CNAK_MASK;		/* clear NAK bit */
+	wfld_set(HSOTG_BASE_ADDR + HSOTG_DOEPCTL0_OFFSET, mask, mask);
+
+	if (cnt > 0)
+		memcpy(data_buf, ep0_out_buf, cnt);
+
+	return 0;
+}
+
+/**
+ * @brief: usb_ep0_recv_zlp - Receive zero length packet on EP0-OUT endpoint.
+ * @param[in]:	none
+ * @param[out]:	none
+ * @return: 0
+ */
+static int usb_ep0_recv_zlp(void)
+{
+	usb_ep0_recv_data(NULL, 0);
+	return 0;
+}
+
+/**
+ * @brief: usb_ep0_send_data - Send packet on EP0-IN endpoint.
+ * @param[in]:	uint8_t *data_buf - Data packet to send
+ * @param[in]:	uint32_t cnt - Number of bytes to send
+ * @param[out]:	none
+ * @return: 0 or ERROR ID
+ */
+static int usb_ep0_send_data(uint8_t *data_buf, uint32_t cnt)
+{
+	if (cnt > sizeof(ep0_in_buf)) {
+		printf("EP0_IN overflow\n");
+		return 1;
+	} else if (cnt > 0) {
+		/* use local DMA buffer */
+		memcpy(ep0_in_buf, data_buf, cnt);
+	}
+
+	flush_dcache_range((uint32_t)ep0_in_buf,
+			   (uint32_t)ep0_in_buf + sizeof(ep0_in_buf));
+
+	/* Device IN Endpoint 0 DMA Address Register */
+	writel((uint32_t)ep0_in_buf,
+	       HSOTG_BASE_ADDR + HSOTG_DIEPDMA0_OFFSET);
+
+	/* Device IN Endpoint 0 Transfer Size Register */
+	writel((1 << HSOTG_DIEPTSIZ0_PKTCNT_SHIFT) |	/* 1 pakcnt */
+	       (cnt << HSOTG_DIEPTSIZ0_XFERSIZE_SHIFT),	/* XferSize cnt */
+	       HSOTG_BASE_ADDR + HSOTG_DIEPTSIZ0_OFFSET);
+
+	/* Device IN Endpoint 0 Control Register */
+	writel(HSOTG_DIEPCTL0_EPENA_MASK |		/* Endpoint Enable */
+	       HSOTG_DIEPCTL0_CNAK_MASK |		/* clear NAK BIT */
+	       HSOTG_DIEPCTL0_USBACTEP_MASK |		/* USB ACTIVE EP */
+	       (0 << HSOTG_DIEPCTL0_NEXTEP_SHIFT),	/* next EP */
+	       HSOTG_BASE_ADDR + HSOTG_DIEPCTL0_OFFSET);
+
+	return 0;
+}
+
+/**
+ * @brief: usb_ep0_send_zlp - Send zero length packet on EP0-IN endpoint.
+ * @param[in]:	none
+ * @param[out]:	none
+ * @return: 0
+ */
+static int usb_ep0_send_zlp(void)
+{
+	usb_ep0_send_data(NULL, 0);
+	return 0;
+}
+
+/**
+ * @brief: usb_ep0_send_stall - Send STALL condition on EP0-IN endpoint.
+ * @param[in]:	none
+ * @param[out]:	none
+ * @return: 0
+ */
+static int usb_ep0_send_stall(void)
+{
+	/* Device Control IN Endpoint 0 Control Register */
+	wfld_set(HSOTG_BASE_ADDR + HSOTG_DIEPCTL0_OFFSET,
+		 HSOTG_DIEPCTL0_STALL_MASK, HSOTG_DIEPCTL0_STALL_MASK);
+	usb_ep0_setup_prime();
+	return 0;
+}
+
+/**
+ * @brief: usb_ep1_out_prime - Prepare receiving EP1-OUT packet.
+ * @param[in]:	none
+ * @param[out]:	none
+ * @return: 0
+ */
+static int usb_ep1_out_prime(void)
+{
+	/* toggle which buffer is used */
+	ep1_out_buf_sel = ep1_out_buf_sel ? 0 : 1;
+
+	invalidate_dcache_range((uint32_t)ep1_out_buf[ep1_out_buf_sel],
+				(uint32_t)ep1_out_buf[ep1_out_buf_sel] +
+				sizeof(ep1_out_buf[ep1_out_buf_sel]));
+
+	/* Device OUT Endpoint 1 DMA Address Register */
+	writel((uint32_t)ep1_out_buf[ep1_out_buf_sel],
+	       HSOTG_BASE_ADDR + HSOTG_DOEPDMA1_OFFSET);
+
+	/* Device OUT Endpoint 1 Transfer Size Register */
+	writel((0x1 << HSOTG_DOEPTSIZ1_PKTCNT_SHIFT) |
+				/* 1 PktCnt */
+	       (EP1_OUT_BUF_SIZE << HSOTG_DOEPTSIZ1_XFERSIZE_SHIFT),
+				/* XferSize */
+	       HSOTG_BASE_ADDR + HSOTG_DOEPTSIZ1_OFFSET);
+
+	/* Device OUT Endpoint 1 Control Register */
+	writel((1 << HSOTG_DOEPCTL1_EPENA_SHIFT) |	/* Endpoint Enable */
+	       (2 << HSOTG_DOEPCTL1_EPTYPE_SHIFT) |	/* Bulk Endpoint */
+	       (1 << HSOTG_DOEPCTL1_CNAK_SHIFT) |	/* clear NAK bit */
+	       (1 << HSOTG_DOEPCTL1_USBACTEP_SHIFT) |	/* USB ACTIVE EP */
+	       (RX_ENDPOINT_MAXIMUM_PACKET_SIZE << HSOTG_DOEPCTL1_MPS_SHIFT),
+	       HSOTG_BASE_ADDR + HSOTG_DOEPCTL1_OFFSET);
+
+	/* */
+	writel(0x00030003, HSOTG_BASE_ADDR + HSOTG_DAINTMSK_OFFSET);
+
+	return 0;
+}
+
+/**
+ * @brief: usb_ep1_in_prime - Prepare for sending EP1-IN packet.
+ * @param[in]:	none
+ * @param[out]:	none
+ * @return: 0
+ */
+static int usb_ep1_in_prime(void)
+{
+	/* Device Control IN Endpoint 1 Control Register */
+	writel((2 << HSOTG_DIEPCTL1_EPTYPE_SHIFT) |	/* Bulk Endpoint */
+	       (1 << HSOTG_DIEPCTL1_CNAK_SHIFT) |	/* clear NAK BIT */
+	       (1 << HSOTG_DIEPCTL1_USBACTEP_SHIFT) |	/* USB ACTIVE EP */
+	       (1 << HSOTG_DIEPCTL1_NEXTEP_SHIFT) |	/* next EP */
+	       (1 << HSOTG_DIEPCTL1_TXFNUM_SHIFT) |	/* use FIFO1 for ep1 */
+	       (TX_ENDPOINT_MAXIMUM_PACKET_SIZE << HSOTG_DIEPCTL1_MPS_SHIFT),
+	       HSOTG_BASE_ADDR + HSOTG_DIEPCTL1_OFFSET);
+
+	return 0;
+}
+
+/**
+ * @brief: usb_send_data_ep1_in - Send packet on EP1-IN endpoint.
+ * @param[in]:	uint8_t *data_buf - Data packet to send
+ * @param[in]:	uint32_t cnt - Number of bytes to send
+ * @param[out]:	none
+ * @return: 0 or ERROR ID
+ */
+static int usb_send_data_ep1_in(uint8_t *data_buf, uint32_t cnt)
+{
+	if (cnt > sizeof(ep1_in_buf)) {
+		printf("EP1_IN overflow\n");
+		return 1;
+	} else if (cnt > 0) {
+		/* use local DMA buffer */
+		memcpy(ep1_in_buf, data_buf, cnt);
+	}
+
+	flush_dcache_range((uint32_t)ep1_in_buf,
+			   (uint32_t)ep1_in_buf + sizeof(ep1_in_buf));
+
+	/* Device IN Endpoint 1 DMA Address Register */
+	writel((uint32_t)ep1_in_buf,
+	       HSOTG_BASE_ADDR + HSOTG_DIEPDMA1_OFFSET);
+
+	/* Device IN Endpoint 1 Transfer Size Register */
+	writel((1 << HSOTG_DIEPTSIZ1_PKTCNT_SHIFT) |	/* 1 pakcnt */
+	       (cnt << HSOTG_DIEPTSIZ1_XFERSIZE_SHIFT),	/* XferSize cnt */
+	       HSOTG_BASE_ADDR + HSOTG_DIEPTSIZ1_OFFSET);
+
+	/* Device IN Endpoint 1 Control Register */
+	writel((1 << HSOTG_DIEPCTL1_EPENA_SHIFT) |	/* Enable EP */
+	       (2 << HSOTG_DIEPCTL1_EPTYPE_SHIFT) |	/* Bulk Endpoint */
+	       (1 << HSOTG_DIEPCTL1_CNAK_SHIFT) |	/* clear NAK BIT */
+	       (1 << HSOTG_DIEPCTL1_USBACTEP_SHIFT) |	/* USB ACTIVE EP */
+	       (1 << HSOTG_DIEPCTL1_NEXTEP_SHIFT) |	/* next EP */
+	       (1 << HSOTG_DIEPCTL1_TXFNUM_SHIFT) |	/* use FIFO1 for ep1 */
+	       (TX_ENDPOINT_MAXIMUM_PACKET_SIZE << HSOTG_DIEPCTL1_MPS_SHIFT),
+	       HSOTG_BASE_ADDR + HSOTG_DIEPCTL1_OFFSET);
+
+	return 0;
+}
+
+static void process_usb_req_get_descriptor(uint16_t desc_type,
+		uint16_t desc_index, uint16_t wLength)
+{
+	switch (desc_type) {
+	case USB_DT_DEVICE:
+		/* Reply Device Descriptor */
+		debug("DEVICE DESC\n");
+		usb_ep0_send_data((uint8_t *)&dfu_dev_descriptor,
+				  min(sizeof(dfu_dev_descriptor),
+				      (size_t)wLength));
+		usb_ep0_recv_zlp();
+		break;
+	case USB_DT_CONFIG:
+		/* Reply Configuration Descriptor */
+		debug("CONFIG DESC\n");
+		usb_ep0_send_data((uint8_t *)&full_desc,
+				  min(sizeof(full_desc),
+				      (size_t)wLength));
+		usb_ep0_recv_zlp();
+		break;
+	case USB_DT_STRING:
+		debug("STRING DESC\n");
+		if (desc_index == DEVICE_STRING_LANGUAGE_ID_INDEX) {
+			uint8_t temp[100];
+			temp[0] = 4;
+			temp[1] = USB_DT_STRING;
+			temp[2] = DEVICE_STRING_LANGUAGE_ID & 0xFF;
+			temp[3] = DEVICE_STRING_LANGUAGE_ID >> 8;
+
+			usb_ep0_send_data(temp, 4);
+			usb_ep0_recv_zlp();
+		} else if (desc_index <= DEVICE_STRING_MAX_INDEX) {
+			int index;
+			int sl = strlen(&usb_device_strings[desc_index][0]);
+			uint8_t temp[100];
+			temp[0] = 2 + (2*sl);
+			temp[1] = USB_DT_STRING;
+
+			for (index = 0; index < sl; index++) {
+				int i = 2 + (2*index);
+				temp[i] = usb_device_strings[desc_index][index];
+				temp[i + 1] = 0;
+			}
+
+			usb_ep0_send_data(temp, 2 + (2*sl));
+			usb_ep0_recv_zlp();
+		} else if (desc_index == 238) {
+			debug("string index [%d] is ignored\n", desc_index);
+			usb_ep0_send_stall();
+		} else {
+			error("bad string index [%d]", desc_index);
+			usb_ep0_send_stall();
+		}
+		break;
+	default:
+		error("bad descriptor request");
+		usb_ep0_send_stall();
+		break;
+	}
+}
+
+static int process_usb_set_addr(uint32_t dev_addr)
+{
+	/* Device Configuration Register */
+	writel((EP_MISMATCH_CNT << HSOTG_DCFG_EPMISCNT_SHIFT) |
+				/* IN Mismatch Cnt */
+	       (PERIODIC_FRM_INTERVAL << HSOTG_DCFG_PERFRINT_SHIFT) |
+				/* Frame Interval */
+	       (dev_addr << HSOTG_DCFG_DEVADDR_SHIFT),
+				/* Device Address */
+	       HSOTG_BASE_ADDR + HSOTG_DCFG_OFFSET);
+
+	return 0;
+}
+
+/**
+ * @brief: dfu_ep0_handler - Decode EP0 setup commands and process CH9 requests.
+ * @param[in]:	uint8_t *setup_buf - EP0 packet buffer.
+ * @param[out]:	none
+ * @return: none
+ */
+static void dfu_ep0_handler(u8 *setup_buf)
+{
+	uint16_t wLength;
+	uint16_t desc_type;
+	uint16_t desc_index;
+	struct usb_device_request ctrl;
+
+	memcpy(&ctrl, setup_buf, sizeof(ctrl));
+
+	wLength = ctrl.wLength;
+
+	if (ctrl.bmRequestType & USB_TYPE_VENDOR) {
+		debug("USB_TYPE_VENDOR\n");
+		if (ctrl.bmRequestType & 0x80) {
+			/* VENDOR_REQ_GET_USBID */
+			if (ctrl.bRequest == 0x3) {
+				u8 temp[2];
+				memcpy(temp, (void *)CONFIG_USBID_ADDR, 2);
+				usb_ep0_send_data(temp, 2);
+				usb_ep0_recv_zlp();
+			} else {
+				u8 temp = 0;
+				usb_ep0_send_data(&temp, 1);
+				usb_ep0_recv_zlp();
+			}
+		} else {
+			usb_ep0_setup_prime();
+			usb_ep0_send_zlp();
+		}
+	} else if (ctrl.bmRequestType & USB_TYPE_CLASS) {
+		debug("USB_TYPE_CLASS\n");
+		usb_ep0_send_stall();
+	} else {
+		/* CH9 command */
+		switch (ctrl.bRequest) {
+		case USB_REQ_SET_ADDRESS:
+			debug("SET_ADDR = %d\n", ctrl.wValue);
+			process_usb_set_addr(ctrl.wValue);
+			usb_ep0_setup_prime();
+			usb_ep0_send_zlp();
+			return;
+
+		case USB_REQ_CLEAR_FEATURE:
+		case USB_REQ_SET_FEATURE:
+			debug("USB_SET/CLR_FEATURE\n");
+			if (ctrl.wValue == USB_TEST_MODE) {
+				unsigned int testctl = ctrl.wIndex >> 8;
+
+				usb_ep0_setup_prime();
+				usb_ep0_send_zlp();
+
+				/*
+				 * Wait till after status phase to set test mode
+				 */
+				wfld_set(HSOTG_BASE_ADDR + HSOTG_DCTL_OFFSET,
+					 testctl, HSOTG_DCTL_TSTCTL_MASK);
+			} else {
+				usb_ep0_setup_prime();
+				usb_ep0_send_zlp();
+			}
+			return;
+
+		case USB_REQ_GET_CONFIGURATION:
+			debug("GET_CONFIG\n");
+			usb_ep0_setup_prime();
+			usb_ep0_send_zlp();
+			break;
+
+		case USB_REQ_SET_CONFIGURATION:
+			debug("SET_CONFIG\n");
+			usb_ep0_setup_prime();
+			usb_ep1_out_prime();
+			usb_ep1_in_prime();
+			usb_ep0_send_zlp();
+			break;
+
+		case USB_REQ_GET_DESCRIPTOR:
+			debug("GET_DESC -> ");
+			desc_type = ctrl.wValue >> 8;
+			desc_index = ctrl.wValue & 0xff;
+			process_usb_req_get_descriptor(desc_type,
+						       desc_index,
+						       wLength);
+			break;
+
+		case USB_REQ_SET_INTERFACE:
+		case USB_REQ_GET_INTERFACE:
+			debug("USB_SET/GET_INTERFACE\n");
+			usb_ep0_setup_prime();
+			usb_ep0_send_zlp();
+			break;
+
+		case USB_REQ_GET_STATUS:
+		case USB_REQ_SET_DESCRIPTOR:
+		case USB_REQ_SYNCH_FRAME:
+			debug("USB misc CH9 cmd: ignore\n");
+			usb_ep0_setup_prime();
+			usb_ep0_send_zlp();
+			break;
+
+		default:
+			error("Unknown EP0 cmd");
+			usb_ep0_send_stall();
+			break;
+		}
+	}
+}
+
+/**
+ * @brief: usb_handle_ep0_in_int - Process USB EP0-IN interrupt.
+ * @param[in]:	none
+ * @param[out]:	none
+ * @return: 0
+ */
+static int usb_handle_ep0_in_int(void)
+{
+	uint32_t diep0int;
+
+	/* clear Device IN Endpoint 0 Interrupt Register */
+	diep0int = readl(HSOTG_BASE_ADDR + HSOTG_DIEPINT0_OFFSET);
+	writel(diep0int, HSOTG_BASE_ADDR + HSOTG_DIEPINT0_OFFSET);
+
+	return 0;
+}
+
+/**
+ * @brief: usb_handle_ep1_in_int - Process USB EP1-IN interrupt.
+ * @param[in]:	none
+ * @param[out]:	none
+ * @return: 0
+ */
+static int usb_handle_ep1_in_int(void)
+{
+	uint32_t diep1int;
+
+	/* clear Device IN Endpoint 1 Interrupt Register */
+	diep1int = readl(HSOTG_BASE_ADDR + HSOTG_DIEPINT1_OFFSET);
+	writel(diep1int, HSOTG_BASE_ADDR + HSOTG_DIEPINT1_OFFSET);
+
+	return 0;
+}
+
+/**
+ * @brief: usb_handle_ep0_out_int - Process USB EP0-OUT interrupt.
+ * @param[in]:	none
+ * @param[out]:	none
+ * @return: 0 or ERROR ID
+ */
+static int usb_handle_ep0_out_int(void)
+{
+	uint32_t doep0int;
+
+	/* clear Device OUT Endpoint 0 Interrupt Register */
+	doep0int = readl(HSOTG_BASE_ADDR + HSOTG_DOEPINT0_OFFSET);
+	writel(doep0int, HSOTG_BASE_ADDR + HSOTG_DOEPINT0_OFFSET);
+
+	if (doep0int & HSOTG_DOEPINT0_TIMEOUT_MASK) {
+		/* Timeout Condition */
+		debug("timeout\n");
+
+		/* Reset EP0 buffer */
+		dfu_ep0_handler((uint8_t *)ep0_setup_buf);
+	} else if (doep0int & HSOTG_DOEPINT0_XFERCOMPL_MASK) {
+		/* Transfer Completed Interrupt */
+
+		/* re-arm EP0-Setup */
+		usb_ep0_setup_prime();
+	} else {
+		debug(" ... not timeout or transfer complete?\n");
+		return 1;
+	}
+	return 0;
+}
+
+/**
+ * @brief: usb_handle_ep1_out_int - Process USB EP1-OUT interrupt.
+ * @param[in]:	none
+ * @param[out]:	none
+ * @return: 0 or ERROR ID
+ */
+static int usb_handle_ep1_out_int(void)
+{
+	uint32_t doepint1;
+
+	/* clear Device OUT Endpoint 1 Interrupt Register */
+	doepint1 = readl(HSOTG_BASE_ADDR + HSOTG_DOEPINT1_OFFSET);
+	writel(doepint1, HSOTG_BASE_ADDR + HSOTG_DOEPINT1_OFFSET);
+
+	if (doepint1 & HSOTG_DOEPINT1_TIMEOUT_MASK) {
+		/* Timeout Condition */
+		debug("timeout\n");
+	} else if (doepint1 & HSOTG_DOEPINT1_XFERCOMPL_MASK) {
+		/* Transfer Completed Interrupt */
+
+		/* Get pointer to current buffer */
+		uint8_t *buffer = ep1_out_buf[ep1_out_buf_sel];
+
+		/* Compute size of transfer */
+		int transfer_size = EP1_OUT_BUF_SIZE -
+			(readl(HSOTG_BASE_ADDR + HSOTG_DOEPTSIZ1_OFFSET) &
+			       HSOTG_DOEPTSIZ1_XFERSIZE_MASK);
+
+		/* Stuff a NULL in to help terminate command strings */
+		if (transfer_size != EP1_OUT_BUF_SIZE)
+			buffer[transfer_size] = 0;
+
+		/*
+		 * Before processing received data, re-arm EP1-OUT
+		 * (toggles active buffer)
+		 */
+		usb_ep1_out_prime();
+
+		/* Now, process data */
+		if (rx_callback)
+			rx_callback(buffer, transfer_size);
+
+	} else {
+		debug(" ... not timeout or transfer complete?\n");
+		return 1;
+	}
+	return 0;
+}
+
+/**
+ * @brief: usb_chirp_enum - Perform USB CHIRP.
+ * @param[in]:	none
+ * @param[out]:	none
+ * @return: 0
+ */
+static inline int usb_chirp_enum(void)
+{
+	/*
+	 * Enumerated Speed:
+	 *	00: High speed (PHY clock is running at 30 or 60 MHz)
+	 *	01: Full speed (PHY clock is running at 30 or 60 MHz)
+	 *	10: Low speed  (PHY clock is running at 6 MHz)
+	 *	11: Full speed (PHY clock is running at 48 MHz)
+	 */
+	usb_speed = readl(HSOTG_BASE_ADDR + HSOTG_DSTS_OFFSET);
+	usb_speed = ((usb_speed & HSOTG_DSTS_ENUMSPD_MASK) >>
+		     HSOTG_DSTS_ENUMSPD_SHIFT);
+
+	printf("USB speed: %s\n", (usb_speed == 0) ? "High" : "Full/Low");
+
+	return 0;
+}
+
+/**
+ * @brief: usb_power_up - Power-up USB.
+ * @param[in]:	none
+ * @param[out]:	none
+ * @return: none
+ */
+static inline void usb_power_up(void)
+{
+	/* clear "stop PHY clock" */
+	writel((0 << HSOTG_PCGCR_STOPPCLK_SHIFT),
+	       HSOTG_BASE_ADDR + HSOTG_PCGCR_OFFSET);
+	printf("Enable USB PHY clock!\n");
+}
+
+/*
+ * UDC interface
+ */
+
+int usb_init_otg(rx_type rx)
+{
+	uint32_t mask;
+
+	rx_callback = rx;
+
+	clk_usb_otg_enable((void *)HSOTG_BASE_ADDR);
+
+#ifdef CONFIG_CMD_SERIALNO
+	if (get_serial_no())
+		usb_device_strings[DEVICE_STRING_SERIAL_NUMBER_INDEX] =
+				(char *)get_serial_no();
+#endif
+
+	mask =	HSOTG_CTRL_USBOTGCONTROL_UTMIOTG_IDDIG_SW_MASK |
+			/* SW sets device mode bit 26 */
+		HSOTG_CTRL_USBOTGCONTROL_USB_ON_MASK |
+			/* turn on OTG core */
+		HSOTG_CTRL_USBOTGCONTROL_USB_ON_IS_HCLK_EN_MASK |
+			/* use usb_on as source of AHB clock enable */
+		HSOTG_CTRL_USBOTGCONTROL_USB_HCLK_EN_DIRECT_MASK;
+			/* explicit source of AHB clock enable */
+
+	wfld_set(HSOTG_CTRL_BASE_ADDR + HSOTG_CTRL_USBOTGCONTROL_OFFSET,
+		 mask, mask);
+
+	udelay(1000);
+	usb_soft_reset();
+	udelay(1000);
+
+	/* Initialize OTG device core */
+	if (dfu_setup_device_mode()) {
+		usb_phy_disconnect();
+		return 1;
+	}
+
+	/* Device All Endpoints Interrupt Mask Register */
+	writel(HSOTG_INTMSK_WKUPINTMSK_MASK |
+			/* Resume/Remote Wakeup Detected */
+	       HSOTG_INTMSK_DISCONNINTMSK_MASK |
+			/* Disconnect Detected */
+	       HSOTG_INTMSK_OEPINTMSK_MASK |	/* OUT Endpoints */
+	       HSOTG_INTMSK_INEPINTMSK_MASK |	/* IN Endpoints */
+	       HSOTG_INTMSK_ENUMDONEMSK_MASK |	/* Enumeration Done */
+	       HSOTG_INTMSK_USBRSTMSK_MASK |	/* USB Reset */
+	       HSOTG_INTMSK_USBSUSPMSK_MASK |	/* USB Suspend */
+	       HSOTG_INTMSK_ERLYSUSPMSK_MASK |	/* Early Suspend */
+	       HSOTG_INTMSK_OTGINTMSK_MASK,	/* OTG Interrupt */
+	       HSOTG_BASE_ADDR + HSOTG_INTMSK_OFFSET);
+
+	/* Device IN Endpoint Common Interrupt Mask Register */
+	writel(HSOTG_DIEPMSK_TIMEOUTMSK_MASK |	/* Timeout Condition */
+	       HSOTG_DIEPMSK_AHBERRMSK_MASK |	/* AHB Error */
+	       HSOTG_DIEPMSK_EPDISBLDMSK_MASK |	/* Endpoint Disabled */
+	       HSOTG_DIEPMSK_XFERCOMPLMSK_MASK,	/* Transfer Completed */
+	       HSOTG_BASE_ADDR + HSOTG_DIEPMSK_OFFSET);
+
+	/* Device OUT Endpoint Common Interrupt Mask Register */
+	writel(HSOTG_DOEPMSK_SETUPMSK_MASK |	/* SETUP Phase Done */
+	       HSOTG_DOEPMSK_AHBERRMSK_MASK |	/* AHB Error */
+	       HSOTG_DOEPMSK_EPDISBLDMSK_MASK |	/* Endpoint Disabled */
+	       HSOTG_DOEPMSK_XFERCOMPLMSK_MASK,	/* Transfer Completed */
+	       HSOTG_BASE_ADDR + HSOTG_DOEPMSK_OFFSET);
+
+	memset(ep0_setup_buf, 0, sizeof(ep0_setup_buf));
+
+	usb_ep0_setup_prime();
+	usb_clr_interrupt();
+
+	usb_phy_connect();
+
+	/* dataline or Vbus Pulsing */
+	mask =	HSOTG_CFG_SRPCAP_MASK |		/* SRP-Capable */
+		HSOTG_CFG_HNPCAP_MASK;		/* HNP-Capable */
+	wfld_set(HSOTG_BASE_ADDR + HSOTG_CFG_OFFSET, mask, mask);
+
+	mask =	HSOTG_OTGCTL_SESREQ_MASK;
+	wfld_set(HSOTG_BASE_ADDR + HSOTG_OTGCTL_OFFSET, mask, mask);
+
+	usb_speed = 0;
+
+	return 0;
+}
+
+void usb_handle_ints_otg(void)
+{
+	uint32_t regval;
+	uint32_t intsts;
+
+	intsts = readl(HSOTG_BASE_ADDR + HSOTG_INTSTS_OFFSET);
+
+	if (intsts) {
+		/* IN Endpoints Interrupt */
+		if (intsts & HSOTG_INTSTS_IEPINT_MASK) {
+			uint32_t daint = readl(HSOTG_BASE_ADDR +
+					       HSOTG_DAINT_OFFSET);
+
+#ifdef CONFIG_DEBUG_IEPINT
+			debug("daint = 0x%08X - I\n", daint);
+#endif
+
+			if (daint & (1 << (0 + HSOTG_DAINTMSK_INEPMSK_SHIFT)))
+				usb_handle_ep0_in_int();
+			if (daint & (1 << (1 + HSOTG_DAINTMSK_INEPMSK_SHIFT)))
+				usb_handle_ep1_in_int();
+		}
+
+		/* OUT Endpoints Interrupt */
+		if (intsts & HSOTG_INTSTS_OEPINT_MASK) {
+			uint32_t daint = readl(HSOTG_BASE_ADDR +
+					       HSOTG_DAINT_OFFSET);
+
+#ifdef CONFIG_DEBUG_OEPINT
+			debug("daint = 0x%08X - O\n", daint);
+#endif
+
+			if (daint & (1 << (0 + HSOTG_DAINTMSK_OUTEPMSK_SHIFT)))
+				usb_handle_ep0_out_int();
+			if (daint & (1 << (1 + HSOTG_DAINTMSK_OUTEPMSK_SHIFT)))
+				usb_handle_ep1_out_int();
+		}
+
+		/* OTG Interrupt */
+		if (intsts & HSOTG_INTSTS_OTGINT_MASK) {
+			debug("G\n");
+			/* clear OTG Interrupt Register */
+			regval = readl(HSOTG_BASE_ADDR + HSOTG_OTGINT_OFFSET);
+			writel(regval, HSOTG_BASE_ADDR + HSOTG_OTGINT_OFFSET);
+			debug("OTG: 0x%08X\n", regval);
+
+			/* Session Request Success Status Change */
+			if (regval & HSOTG_OTGINT_SESREQSUCSTSCHNG_MASK)
+				debug("r\n");
+
+			/* Session End Detected */
+			if (regval & HSOTG_OTGINT_SESENDDET_MASK) {
+				debug("e\n");
+				udelay(2500);
+				wfld_set(HSOTG_BASE_ADDR + HSOTG_OTGCTL_OFFSET,
+					 HSOTG_OTGCTL_SESREQ_MASK,
+					 HSOTG_OTGCTL_SESREQ_MASK);
+			}
+
+			/* Debounce Done */
+			if (regval & HSOTG_OTGINT_DBNCEDONE_MASK)
+				debug("d\n");
+
+			/* A-Device Timeout Change */
+			if (regval & HSOTG_OTGINT_ADEVTOUTCHG_MASK)
+				debug("a\n");
+		}
+
+		/* Enumeration Done */
+		if (intsts & HSOTG_INTSTS_ENUMDONE_MASK) {
+			debug("E\n");
+			usb_chirp_enum();
+			writel(HSOTG_INTSTS_ENUMDONE_MASK,
+			       HSOTG_BASE_ADDR + HSOTG_INTSTS_OFFSET);
+		}
+
+		/* Early Suspend */
+		if (intsts & HSOTG_INTSTS_ERLYSUSP_MASK) {
+			debug("eS\n");
+			writel(HSOTG_INTSTS_ERLYSUSP_MASK,
+			       HSOTG_BASE_ADDR + HSOTG_INTSTS_OFFSET);
+		}
+
+		/* USB Suspend */
+		if (intsts & HSOTG_INTSTS_USBSUSP_MASK) {
+			debug("Su\n");
+			writel(HSOTG_INTSTS_USBSUSP_MASK,
+			       HSOTG_BASE_ADDR + HSOTG_INTSTS_OFFSET);
+		}
+
+		/* Wake Up */
+		if (intsts & HSOTG_INTSTS_WKUPINT_MASK) {
+			usb_power_up();
+			debug("Rm\n");
+			writel(HSOTG_INTSTS_WKUPINT_MASK,
+			       HSOTG_BASE_ADDR + HSOTG_INTSTS_OFFSET);
+		}
+	}
+}
+
+int usb_send_bulk_otg(uint8_t *buffer, uint32_t len)
+{
+	return usb_send_data_ep1_in(buffer, len);
+}
+
+void usb_shutdown_otg(void)
+{
+	/* */
+	writel(0x04008C4C, HSOTG_CTRL_BASE_ADDR +
+			   HSOTG_CTRL_USBOTGCONTROL_OFFSET);
+
+	/* wait for clock to settle down before checking vbus */
+	udelay(32);
+
+	usb_wait_for_vbus();
+
+	usb_turn_off_vdp();
+}
diff --git a/drivers/usb/gadget/bcm_udc_otg.h b/drivers/usb/gadget/bcm_udc_otg.h
new file mode 100644
index 0000000..d5b1805
--- /dev/null
+++ b/drivers/usb/gadget/bcm_udc_otg.h
@@ -0,0 +1,19 @@ 
+/*
+ * Copyright 2015 Broadcom Corporation.
+ *
+ * SPDX-License-Identifier:	GPL-2.0+
+ */
+
+#ifndef __BCM_UDC_OTG_H
+#define __BCM_UDC_OTG_H
+
+#include <linux/types.h>
+
+typedef int (*rx_type)(const unsigned char *buffer, unsigned int buffer_size);
+
+int usb_init_otg(rx_type rx);
+void usb_handle_ints_otg(void);
+int usb_send_bulk_otg(uint8_t *buffer, uint32_t len);
+void usb_shutdown_otg(void);
+
+#endif