[edk2-devel] [PATCH] MdeModulePkg/UsbBusDxe: some USB PenDisk fails enumeration.

Feng Libo lbfeng at zd-tech.com.cn
Thu Aug 27 01:07:00 UTC 2020


Hello, 


How about the progress of the Patch Review?


Thanks


--

Best Regards


Feng Libo
ZD Technology (Beijing) Co., Ltd

发件人:"丰立波" <lbfeng at zd-tech.com.cn>
发送日期:2020-08-13 10:49:33
收件人:"Jiang, Guomin" <guomin.jiang at intel.com>
抄送人:"devel at edk2.groups.io" <devel at edk2.groups.io>,"jeremy.linton at arm.com" <jeremy.linton at arm.com>,"Wu, Hao A" <hao.a.wu at intel.com>,"Ni, Ray" <ray.ni at intel.com>,"张超" <czhang at zd-tech.com.cn>
主题:Re:RE: [edk2-devel] [PATCH] MdeModulePkg/UsbBusDxe: some USB PenDisk fails enumeration.

Hello, Mr. Jiang,


Most USB PenDisk work fine with the original enumeration sequence. A few can't.
 
In Microsoft Post, they explained the cause :
In the early days of USB some USB devices would become confused by a second request for the Device Descriptor if they did not return the complete Device Descriptor for the first request.  To allow these devices to enumerate successfully it was necessary to reset the port between the first and second requests for the Device Descriptor.


In our experience, only three Pendisk with Innostor USB controller chip (VID=0x1F75, PID=0x917, USB3.1) fail the enumeration. We can only observed the Pendisk not responding the second Device Descriptor Request for a full Descriptor. The first Device Descriptor Request could impact some Pendisk or as Microsoft said "confuse".  


So, We add a second port reset after the MaxPacketSize0 request. and this second reset can clear the address that is already assigned. Then, we move the MaxPacketSize request before the address assignation.


That is all patch.


Furthermore, the USB Spec states the MaxPacketSize Request could be read through the default Pipe, the address 0. So I think it implies the MaxPacketSize Request should be prior to the address assignation, just as the enumeration sequence in the Microsoft Post.


Thanks


--

Best Regards


Feng Libo
ZD Technology (Beijing) Co., Ltd

发件人:"Jiang, Guomin" <guomin.jiang at intel.com>
发送日期:2020-08-12 16:03:42
收件人:"devel at edk2.groups.io" <devel at edk2.groups.io>,"Jiang, Guomin" <guomin.jiang at intel.com>,"lbfeng at zd-tech.com.cn" <lbfeng at zd-tech.com.cn>
抄送人:"jeremy.linton at arm.com" <jeremy.linton at arm.com>,"Wu, Hao A" <hao.a.wu at intel.com>,"Ni, Ray" <ray.ni at intel.com>
主题:RE: [edk2-devel] [PATCH] MdeModulePkg/UsbBusDxe: some USB PenDisk fails enumeration.


Hi Libo,

 

I review the USB Spec 2.0 and have some confusion when check the spec as below:

From the USB 2.0 spec:

 

Section 5.5.3 Control Transfer Packet Size Constraints

...

In order to determine the maximum packet size for the Default Control Pipe, the USB System Software

reads the device descriptor. The host will read the first eight bytes of the device descriptor. The device

always responds with at least these initial bytes in a single packet. After the host reads the initial part of the

device descriptor, it is guaranteed to have read this default pipe’s wMaxPacketSize field (byte 7 of the

device descriptor). It will then allow the correct size for all subsequent transactions. For all other control

endpoints, the maximum data payload size is known after configuration so that the USB System Software

can ensure that no data payload will be sent to the endpoint that is larger than the supported size.

...

 

Also,

Section 9.1.2 Bus Enumeration

...

3. Now that the host knows the port to which the new device has been attached, the host then waits for at

least 100 ms to allow completion of an insertion process and for power at the device to become stable.

The host then issues a port enable and reset command to that port. Refer to Section 7.1.7.5 for

sequence of events and timings of connection through device reset.

4. The hub performs the required reset processing for that port (see Section 11.5.1.5). When the reset

signal is released, the port has been enabled. The USB device is now in the Default state and can draw

no more than 100 mA from VBUS. All of its registers and state have been reset and it answers to the

default address.

5. The host assigns a unique address to the USB device, moving the device to the Address state.

6. Before the USB device receives a unique address, its Default Control Pipe is still accessible via the

default address. The host reads the device descriptor to determine what actual maximum data payload

size this USB device’s default pipe can use.

7. The host reads the configuration information from the device by reading each configuration zero to

n-1, where n is the number of configurations. This process may take several milliseconds to complete.

...

 

It seem that the original behavior follow the spec, but I don’t know why the device will not response and must reset it.

 

I notice that  you obtain the patch from https://techcommunity.microsoft.com/t5/microsoft-usb-blog/how-does-usb-stack-enumerate-a-device/ba-p/270685#:~:text=%20How%20does%20USB%20stack%20enumerate%20a%20device%3F,a%20request%20for%20the%20USB%20Device...%20More%20.

 

Do you know the device behavior from the device side when the issue happened?

 

Thanks.

Guomin

From: devel at edk2.groups.io <devel at edk2.groups.io> On Behalf Of Guomin Jiang
Sent: Tuesday, August 11, 2020 7:17 PM
To: devel at edk2.groups.io; lbfeng at zd-tech.com.cn
Cc: jeremy.linton at arm.com; Wu, Hao A <hao.a.wu at intel.com>; Ni, Ray <ray.ni at intel.com>
Subject: Re: [edk2-devel] [PATCH] MdeModulePkg/UsbBusDxe: some USB PenDisk fails enumeration.

 

+Hao, Ray,

 

Hi Libo, thanks for your explanation.

 

So I think the patch is improvement for current logic.

 

Hi Hao and Ray,

 

Can you give some comments for the change.

 

Hi Jeremy,

 

It may be helpful for the ASSERT issue https://edk2.groups.io/g/devel/message/62651,can you try it?

 

Best Regards

Guomin

From:devel at edk2.groups.io <devel at edk2.groups.io> On Behalf Of Feng Libo
Sent: Tuesday, August 11, 2020 5:50 PM
To: Jiang, Guomin <guomin.jiang at intel.com>
Cc:devel at edk2.groups.io; jeremy.linton at arm.com
Subject: Re: [edk2-devel] [PATCH] MdeModulePkg/UsbBusDxe: some USB PenDisk fails enumeration.

 

Hello, Mr. Jiang,

 

Thank for the review.

 

The original enumeration steps in the function of UsbEnumerateNewDev of file UsbEnumer.c: 1 reset the port, 2 set the usb device address, 3 get the Max Packet Size, 4 get the full device descriptor. However, when plugging a USB PenDisk with Innostor USB

controller chip (VID=0x1F75, PID=0x917, USB3.1), the fourth step always fails, trace as below:

 

========

XhcCheckUrbResult: TRANSACTION_ERROR! Completecode = 4 XhcControlTransfer: error - Device Error, transfer - 40 UsbGetOneConfig: failed to get full descript Device Error UsbBuildDescTable: failed to get configure (index 0) UsbEnumerateNewDev: failed to build descriptor table - Device Error

=======

 

The host controller need to get the full device descriptor, but this moment, the Pendisk device doesn't response any more. Then timeout. and UsbEnumerateNewDev complains : failed to build descriptor.

 

We have three Pendisks from different manufacturers, all with Innostor USB controller chip. they all can't be enumerated all. And we observed the problem on both Huawei KunPeng(华为鲲鹏)and Loognson(龙芯)platforms.

 

The three Pendisks always fail the USB enumeration. Other USB 2.0 and USB 3.0 on hand can work well.

 

With the patch, the three pendisks and other pendisks can all work well.

 

THanks

 

--

Best Regards

 

Feng Libo

ZD Technology (Beijing) Co., Ltd


发件人:"Jiang, Guomin" <guomin.jiang at intel.com>
发送日期:2020-08-11 08:21:10
收件人:"devel at edk2.groups.io" <devel at edk2.groups.io>,"Jiang, Guomin" <guomin.jiang at intel.com>,"lbfeng at zd-tech.com.cn" <lbfeng at zd-tech.com.cn>
抄送人:"jeremy.linton at arm.com" <jeremy.linton at arm.com>
主题:RE: [edk2-devel] [PATCH] MdeModulePkg/UsbBusDxe: some USB PenDisk fails enumeration.




+Jeremy,

 

I review the patch and think it is reasonable, but I want to know some more detail information

Can you provide the detail debug log about USB?
The symptom always can be seen or have fail rate?

 

Best Regards

Guomin

From:devel at edk2.groups.io <devel at edk2.groups.io> On Behalf Of Guomin Jiang
Sent: Thursday, August 6, 2020 12:29 PM
To:devel at edk2.groups.io; lbfeng at zd-tech.com.cn
Subject: Re: [edk2-devel] [PATCH] MdeModulePkg/UsbBusDxe: some USB PenDisk fails enumeration.

 

I will review it by next weekend(8/14).

 

Thanks.

From:devel at edk2.groups.io <devel at edk2.groups.io> On Behalf Of Feng Libo
Sent: Thursday, August 6, 2020 9:25 AM
To: Feng Libo <lbfeng at zd-tech.com.cn>; devel at edk2.groups.io
Subject: Re: [edk2-devel] [PATCH] MdeModulePkg/UsbBusDxe: some USB PenDisk fails enumeration.

 

Hello, 

could anyone review this PATCH?

We encountered the USB enumeration problem and the patch is based on the Microsoft post as below.

https://techcommunity.microsoft.com/t5/microsoft-usb-blog/how-does-usb-stack-enumerate-a-device/ba-p/270685#:~:text=%20How%20does%20USB%20stack%20enumerate%20a%20device%3F,a%20request%20for%20the%20USB%20Device...%20More%20

Thanks

Best Regards

Feng Libo


-=-=-=-=-=-=-=-=-=-=-=-
Groups.io Links: You receive all messages sent to this group.

View/Reply Online (#64666): https://edk2.groups.io/g/devel/message/64666
Mute This Topic: https://groups.io/mt/75608816/1813853
Group Owner: devel+owner at edk2.groups.io
Unsubscribe: https://edk2.groups.io/g/devel/unsub  [edk2-devel-archive at redhat.com]
-=-=-=-=-=-=-=-=-=-=-=-

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listman.redhat.com/archives/edk2-devel-archive/attachments/20200827/05f4417e/attachment.htm>


More information about the edk2-devel-archive mailing list