[U-Boot] [PATCH v3 00/13] usb: xhci: Add interrupt transfer support and full speed device support

[Bin Meng]

Hi Marek,

On Fri, Jul 21, 2017 at 8:36 PM, Marek Vasut <marex at denx.de> wrote:
> On 07/21/2017 02:34 PM, Bin Meng wrote:
>> Hi Marek,
>>
>> On Fri, Jul 21, 2017 at 6:06 PM, Marek Vasut <marex at denx.de> wrote:
>>> On 07/21/2017 12:04 PM, Bin Meng wrote:
>>>> Hi Marek,
>>>>
>>>> On Fri, Jul 21, 2017 at 4:56 PM, Marek Vasut <marex at denx.de> wrote:
>>>>> On 07/21/2017 10:54 AM, Bin Meng wrote:
>>>>>> Hi Marek,
>>>>>>
>>>>>> On Thu, Jul 20, 2017 at 8:20 PM, Marek Vasut <marex at denx.de> wrote:
>>>>>>> On 07/20/2017 02:16 PM, Bin Meng wrote:
>>>>>>>> Hi Marek,
>>>>>>>>
>>>>>>>> On Thu, Jul 20, 2017 at 7:12 AM, Bin Meng <bmeng.cn at gmail.com> wrote:
>>>>>>>>> This series is the final series of the xHCI driver update.
>>>>>>>>>
>>>>>>>>> This adds the missing interrupt transfer support to xHCI driver, so
>>>>>>>>> that devices like USB keyboard that uses interrupt transfer when
>>>>>>>>> CONFIG_SYS_USB_EVENT_POLL is defined can work.
>>>>>>>>>
>>>>>>>>> This also adds full speed device support. Unlike low/high/super speed
>>>>>>>>> devices, full speed device may report its endpoint 0 max packet size
>>>>>>>>> as 8/16/32/64. xHCI driver guesses 64 for the first attempt to get the
>>>>>>>>> initial 8 bytes device descriptor, and later adjusts the context for
>>>>>>>>> endpoint 0.
>>>>>>>>>
>>>>>>>>> This series also made several updates to xHCI driver to conform with
>>>>>>>>> the spec.
>>>>>>>>>
>>>>>>>>> Previous two series:
>>>>>>>>> [1]: usb: xhci: Fix USB xHCI support on Intel platform
>>>>>>>>> https://lists.denx.de/pipermail/u-boot/2017-June/296166.html
>>>>>>>>> [2]: usb: hub: Support USB 3.0 hubs
>>>>>>>>> https://lists.denx.de/pipermail/u-boot/2017-June/296284.html
>>>>>>>>>
>>>>>>>>> This series is available at u-boot-x86/xhci-working3 for testing.
>>>>>>>>>
>>>>>>>>> Changes in v3:
>>>>>>>>> - rebase on u-boot-usb/master
>>>>>>>>> - add changes to poplar.h
>>>>>>>>>
>>>>>>>>> Bin Meng (13):
>>>>>>>>>   usb: xhci: Add interrupt transfer support
>>>>>>>>>   usb: configs: Clean up CONFIG_SYS_USB_EVENT_POLL_(xxx) in board
>>>>>>>>>     configs
>>>>>>>>>   usb: kbd: kconfig: Set a default polling mechanism for USB keyboard
>>>>>>>>>   usb: Only get 64 bytes device descriptor for full speed devices
>>>>>>>>>   usb: Read device descriptor after device is addressed for xHCI
>>>>>>>>>   usb: xhci: Fix max packet size for full speed device endpoint 0
>>>>>>>>>   usb: hub: Clear port reset before usb_hub_port_connect_change()
>>>>>>>>>   usb: hub: Clear BH reset status change for a 3.0 hub
>>>>>>>>>   usb: Handle audio extension endpoint descriptor in usb_parse_config()
>>>>>>>>>   usb: xhci: Honor endpoint's interval
>>>>>>>>>   usb: xhci: Program max burst size for endpoint
>>>>>>>>>   usb: xhci: Set 'Error Count' to 0 for isoch endpoints
>>>>>>>>>   usb: xhci: Set 'Average TRB Length' to 8 for control endpoints
>>>>>>>>>
>>>>>>>>
>>>>>>>> This is the last xHCI series I have. Can you please take a look and apply?
>>>>>>>
>>>>>>> I'd like to see some RB/TB from Stefan Roese first if possible.
>>>>>>>
>>>>>>
>>>>>> OK, if that's the case, can you send a pull request to Tom to get usb
>>>>>> in? So that I can apply Stefan's x86 patch series on top of this. (I
>>>>>> cannot send PR for x86 before the usb changes)
>>>>>
>>>>> I actually cannot because the USB is currently broken, sorry, I'm
>>>>> waiting for a fix before I do that.
>>>>>
>>>>> Stefan, can you look at this series please?
>>>>
>>>> What "broken" issues are you seeing?
>>>
>>> I think Layerscape stopped building. Look at email with this subject
>>> "Compile failure encountered on latest (2017-07-20)
>>> git://git.denx.de/u-boot-usb.git branch master for target
>>> ls1021atwr_nor_defconfig"
>>
>> I see Patrice is working on a fix for the build failure.
>
> Correct, it's almost done. Sorry for the delay.

FYI, since it was reported by Ran this series has some issues on NXP
LS1021ATWR, I will need allocate some time to figure out why.

BTW: when will the u-boot-usb PR be sent? Not for any push, but
Stefan's x86 patches are in the x86 repo which depends on contents on
u-boot-usb and lots of Kconfig was changed in u-boot/master. The
longer time our patches are staying off stream, the more chances we
get rebase conflicts.

Regards,
Bin