Re: [BISSECTED]: BUG: usb: dwc3: usb ports not working anymore on odroid-XU4

[Richard Genoud]

2017-01-23 12:45 GMT+01:00 Felipe Balbi <balbi@xxxxxxxxxx>:
>
> Hi,
>
> Richard Genoud <richard.genoud@xxxxxxxxx> writes:
>> On 19/01/2017 09:03, Felipe Balbi wrote:
>>>
>>> Hi,
>>>
>>> Richard Genoud <richard.genoud@xxxxxxxxx> writes:
>>>> Hi,
>>>> Since commit c499ff71ff2a2 ("usb: dwc3: core: re-factor init and exit paths")
>>>> (merged in 4.8), the usb ports on odroid-XU4 don't work anymore.
>>>>
>>>> [ Actually, it's commit 2164a476205ccc ("usb: dwc3: set SUSPHY bit for all cores"), cf below ]
>>>>
>>>> Inserting an usb key (USB2.0) on the USB3.0 port result in:
>>>> [   64.488264] xhci-hcd xhci-hcd.2.auto: Port resume took longer than 20000 msec, port status = 0xc400fe3
>>>> [   74.568156] xhci-hcd xhci-hcd.2.auto: xHCI host not responding to stop endpoint command.
>>>> [   74.574806] xhci-hcd xhci-hcd.2.auto: Assuming host is dying, halting host.
>>>> [   74.601970] xhci-hcd xhci-hcd.2.auto: HC died; cleaning up
>>>> [   74.606276] usb 3-1: USB disconnect, device number 2
>>>> [   74.613565] usb 4-1: USB disconnect, device number 2
>>>> [   74.621208] usb usb3-port1: couldn't allocate usb_device
>>>> NB: it's not related to USB2.0 devices, I get the same result with an USB3.0 device (SATA to USB3 for instance).
>>>> NB2: it doesn't happen on an odriod-XU3 board, that doesn't have the realtek RTL8153 chip.
>>>>
>>>> I instrumented what was read/written in the registers before and after this patch, and I found that the culprit is:
>>>>     if (dwc->revision > DWC3_REVISION_194A)
>>>>             reg |= DWC3_GUSB3PIPECTL_SUSPHY;
>>>> Before commit c499ff71ff2a2 ("usb: dwc3: core: re-factor init and exit paths"), dwc3_phy_setup() was
>>>> done early in dwc3_probe() and thus, dwc->revision wasn't set yet (==0)
>>>> After this commit, dwc3_phy_setup() is done in dwc3_core_init(), after setting dwc->revision (it's done in dwc3_core_is_valid()).
>>>>
>>>> (The dwc3->revision on odroid-XU4 is 5533200a.)
>>>>
>>>> If I comment out the 2 lines:
>>>>     if (dwc->revision > DWC3_REVISION_194A)
>>>>             reg |= DWC3_GUSB3PIPECTL_SUSPHY;
>>>> The usb key is recognized right away:
>>>> [   38.008158] usb 3-1.2: new high-speed USB device number 3 using xhci-hcd
>>>> [   38.138924] usb 3-1.2: New USB device found, idVendor=abcd, idProduct=1234
>>>> [   38.144399] usb 3-1.2: New USB device strings: Mfr=1, Product=2, SerialNumber=3
>>>>
>>>> I took a look at the history behind those 2 lines, and they've been introduced by:
>>>> commit 2164a476205ccc ("usb: dwc3: set SUSPHY bit for all cores") in 3.19.
>>>> In 3.19, the dwc->revision dwc3_phy_setup() was set in dwc3_core_init().
>>>> And, booting a 3.19 kernel on XU4 gives the same error (whereas booting a 3.18 kernel is ok)
>>>>
>>>> [ So, I should have started this email with:
>>>> Since commit 2164a476205ccc ("usb: dwc3: set SUSPHY bit for all cores"),
>>>> the usb ports on odroid-XU4 don't work anymore. ]
>>>>
>>>> The funny thing is that commit 45bb7de213d8("usb: dwc3: setup phys earlier") (merged in 4.2)
>>>> moved dwc3_phy_setup() in dwc3_probe(), way before dwc->revision was set, acting like a revert
>>>> on commit 2164a476205c("usb: dwc3: set SUSPHY bit for all cores")
>>>> That's why on 4.2, inserting an USB key on odroid-XU4 worked again,
>>>>
>>>> So, to make a resume:
>>>> xxx-> 3.18: usb ok
>>>> 3.19->4.1: usb error (due to commit 2164a476205ccc ("usb: dwc3: set SUSPHY bit for all cores"))
>>>> 4.2->4.7: usb ok (due to commit 45bb7de213d8("usb: dwc3: setup phys earlier"))
>>>> 4.8->now: usb error (due to commit c499ff71ff2a2 ("usb: dwc3: core: re-factor init and exit paths"))
>>>>
>>>>
>>>> any idea on this ?
>>>
>>> Thanks for bisecting. This could be caused by a similar case as
>>> described at [1]. Can you check if adding quirk
>>> "snps,dis_u3_susphy_quirk" helps ?
>>>
>>> [1] https://marc.info/?i=2B3535C5ECE8B5419E3ECBE300772909021B3FD37E@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
>>>
>> Thanks for your quick answer !
>>
>> yes, adding this patch:
>>
>> diff --git a/arch/arm/boot/dts/exynos5422-odroidxu4.dts b/arch/arm/boot/dts/exynos5422-odroidxu4.dts
>> index 2faf88627a48..c76e76bc9ade 100644
>> --- a/arch/arm/boot/dts/exynos5422-odroidxu4.dts
>> +++ b/arch/arm/boot/dts/exynos5422-odroidxu4.dts
>> @@ -43,6 +43,9 @@
>>       status = "okay";
>>  };
>>
>> +&usbdrd_dwc3_0 {
>> +     snps,dis_u3_susphy_quirk;
>> +};
>>  &usbdrd_dwc3_1 {
>>       dr_mode = "host";
>>  };
>>
>> removes the error "Port resume took longer than 20000 msec, port
>> status = 0xc400fe3" and the USB devices are detected.
>
> okay great. Temporarily, you can probably push a patch doing that to
> your DTS. At least you're gonna have things working until we understand
> what's going on with DWC3.
Ok.

Thanks !

Richard.