124 replies to this topic

[darren rose]

@all
How to provide USB drivers in that case?

Idea1: Use default .script zz_drivers, but add a USB30_x86.7z files USB30_amd64.7z
Can *.7z files added to a winbuilder project server?

Idea2: Or include USB3 drivers to a script, extract files to WorkbenchCommonDrivers_x86USB30
Run zz_drivers to include drivers to PE.
Idea2: Or include USB3 drivers to a script and run dism.exe from the script.

Any other idea?

think the most obvious solution is as medevil suggested - ie. adding these amd usb 3.0 drivers to the already existing usb3 support script - then we will have support for AMD usb 3.0, NEC and reneas

[MedEvil]

Yes but just for tests. I don't think, it's a good idea to add all 3 to a build, since at least in theory, they all are for the same hardware ids.

The goal should be, to find one set of USB 3.0 drivers, that work for everyone, which then get their own script.

[darren rose]

fair point - but can't see why it wouldn't work if we can't find a universal solution - I have got the usb 3 script in my build and added the AMD usb 3 drivers using zzz_drivers - and the build works fine on both the Dell Vostro laptops with AMD usb 3.0 and on my desktop which has renesas usb 3.0

Personally don't think there will be a generic set of drivers that will work for AMD, NEC, Renesas and Intel usb 3.0 controllers

[cdob]

my desktop which has renesas usb 3.0

Which controller version do you use?
Can you list HardwareIDs?

Personally don't think there will be a generic set of drivers that will work for AMD, NEC, Renesas and Intel usb 3.0 controllers

AMD, Renesas2, Renesas3 and Intel usb 3.0 drivers
http://cdob.reboot.p...es/USB30_x86.7z

[darren rose]

Renesas Electronics USB 3.0 Host Controller

PCIVEN_1033&DEV_0194&SUBSYS_84131043&REV_03
PCIVEN_1033&DEV_0194&SUBSYS_84131043
PCIVEN_1033&DEV_0194&CC_0C0330
PCIVEN_1033&DEV_0194&CC_0C03

Is that all you need? let me know if you need anything else

[MedEvil]

I would like to know, if you throw all 3 in, if different drivers get chosen for different hardware, or if it defaults always to the same.

btw. Since Windows 8 has only 1 set of drivers for all USB 3 controllers, it should be possible to get away with just one set.

[darren rose]

I would like to know, if you throw all 3 in, if different drivers get chosen for different hardware, or if it defaults always to the same.

well my PE has the usb3 script and the drivers for the AMD usb 3.0 hub and it uses different drivers depending on whether I boot it on my desktop with reneas or the laptop with AMD USB

[darren rose]

btw. Since Windows 8 has only 1 set of drivers for all USB 3 controllers, it should be possible to get away with just one set.

fair point - but are we sure that windows 8 uses these same "1 set of" drivers if I installed it on a vostro laptop with AMD USB 3.0?

[MedEvil]

I have a crazy idea for a new driver test.
Stay tuned!

[Wonko the Sane]

I have a crazy idea for a new driver test.
Stay tuned!

Wonko approves of crazy ideas!


Wonko

[cdob]

PCI\VEN_1033&DEV_0194&CC_0C03

This seems to be a µPD720200A controller
http://www.renesas.e..._host/index.jsp
http://www.renesas.e...0200a/index.jsp
http://documentation...9ej0100_usb.pdf

The AMD HardwareID and Renesas HardwareID is listed in package at post #54.

Which driver revision do you use?


Does anybody uses a µPD720200?
http://www.renesas.e...20200/index.jsp
Which HardwareIDs refers to this controller?

I have a crazy idea for a new driver test.

AMD licensed USB3 controller from Renesas.
However a AMD driver won't work at Renesas controller and reverse.

Crystal ball:
there is a generic USB3 driver in Windows 7 SP2.

[darren rose]

driver version is - amdhub30.sys file version 1.0.0.60 product version 2.1.16.0 and same for amdxhc.sys

[MedEvil]

However a AMD driver won't work at Renesas controller and reverse.

Are you sure about that?
Does Window actually accept the other driver package as valid drivers for the hardware, but it still doesn't work?

Cause my crazy idea was, to create a own universal driver, by creating a custom inf file.

If there is real incompatibility on the hardware side between different USB controllers, then it can't be done.

btw. My crystal ball says, M$ will not release a free SP for Win7, when they can sell Win8 instead!
Remember: No DirectX 10 for XP!

[cdob]

Are you sure about that?

I got a yellow quesiton mark in device manager in the past.
AMD Huson 75 and Renesas 2 drivers: yellow quesiton mark
NEC USB3 controller and AMD drivers: yellow quesiton mark
Sorry, no version details in memory.

Does Window actually accept the other driver package as valid drivers for the hardware, but it still doesn't work?

Windows trust *.inf files. PNP install the driver. And driver is launched at boot.
However a driver *.sys check hardware itself at initialisation.
It the driver dosn't validate the hardware, the driver reports a error.
If you patch the binary file driver...

Cause my crazy idea was, to create a own universal driver, by creating a custom inf file.

I was crazy in the past and I am crazy still: new game with Intel USB 3 drivers?

[MedEvil]

I thought hoped, that the problem would purely lie within the inf file, which each manufacturer narrows down to just work with their hardware.

If that idea is a bust as you say, i see no way to do it. I don't know enough about USB 3 to even try to fix something up in the binaries.

At least we can skip the other compatibility test, i had envisioned previously.

But there's one thing, i don't understand. When the inf are written in a way, that each driver is accepted by Windows for each hardware, how can darren have all 3 drivers in his build at the same time and have Windows not always pick the same driver as best match?

[darren rose]

just out of curiosity I installed Windows 8 on the AMD Dell Vostro to see what it detects the USB as

It displays it as AMD USB 3.0 eXtensible Host Controller (0096) & USB Root Hub (xHCI) and does indeed use the same universal windows 8 driver as used for other usb3 ports

[MedEvil]

That's how it should be with USB drivers. One size fits all.

I just don't get, why the tested drivers only work with their specific hardware.
The controllers should all show, the driver, the exact same interface.




Unless.... they keyed the binaries to specific hardware ids.

[ChrisR]

I just returned from a week's holiday
Thank you for this nice thread and for all tests

Idea2: Or include USB3 drivers to a script, extract files to WorkbenchCommonDrivers_x86USB30
Idea2: Or include USB3 drivers to a script and run dism.exe from the script.

think the most obvious solution is as medevil suggested - ie. adding these amd usb 3.0 drivers to the already existing usb3 support script - then we will have support for AMD usb 3.0, NEC and reneas

I tend to agree also with this second Idea

to test, here RemovableDeviceUSB3.script modified with cdob USB30_x86.7z and USB30_x64.7z just built
with inside: AMD 1.0.0.0072 - Intel 1.0.4.225 - Renesas 2.1.28.0 - Renesas 3.0.16.0

RemovableDeviceUSB3.script

edit: i don't have any USB 3 controller for test
sorry for the interruption, continue

[darren rose]

@ChrisR

Just downloaded and now doing a build so I can test it on the Dell Vostro with AMD 3.0

[MedEvil]

btw. Just for making things more clear, we should not talk about NEC anything anymore. NEC is now Renesas.
So it makes little no sense to see in the USB 3.0 script, Renesas and NEC drivers listed.

[MedEvil]

@darren
When possible, could you try to test each of the drivers, with each of the two controllers you have?

We need to find some common ground here.

[darren rose]

@darren
When possible, could you try to test each of the drivers, with each of the two controllers you have?

We need to find some common ground here.

yes no problem at all

I had already tried the original usb3 script and it didn't work with my AMD but did work with my Reneas
The slimmed down AMD drivers above work fine with the AMD but did NOT work with Reneas
Now trying the updated usb3 script ......

UPDATE:

The updated usb3 script by chrisr above works fine with the AMD usb3.0 - will try it later on reneas

[MedEvil]

There are 2 Reneas drivers, 1 AMD and 1 Intel in the script.

We would like to find a single one, that works with all hardware. Since you already ruled out 2 by previous tests, only the second Renesas and the Intel one remain.

[ChrisR]

To assist, perhaps, testing, in the meaning of a single driver
here the same script with 4 CheckBox to choose the drivers to install : AMD, Intel, Renesas2 and/or Renesas3 for testing.
http://www.mediafire...al84b04dtw5bjq1
I'm not sure that it brings something more

There are 2 Reneas drivers, 1 AMD and 1 Intel in the script.

We would like to find a single one, that works with all hardware. Since you already ruled out 2 by previous tests, only the second Renesas and the Intel one remain.

ideal, is there a chance ?
Renesas drivers not seem to work with AMD controller and vice versa. Probably the same for Intel.

#
in this script, only the pe2/3 part was modified
and I do not know how to write txtsetup.sif for PE1 and AMD, Intel and Renesas3 and drivers

Here for Renesas2 ported from the cdob's BartPE plugin http://www.911cd.net...showtopic=23637

ExtractFile,%ScriptFile%,Renesas_%sourcearch%,nusb3co.dll,%target_sys%

ExtractFile,%ScriptFile%,Renesas_%sourcearch%,nusb3hub.sys,%target_sys%\drivers

ExtractFile,%ScriptFile%,Renesas_%sourcearch%,nusb3xhc.sys,%target_sys%\drivers



//Host Controller

IniWrite,%target_win%\txtsetup.sif,SourceDisksFiles,nusb3xhc.sys,"1,,,,,,4_,4,1,,,1,4"

IniWrite,%target_win%\txtsetup.sif,BusExtenders.Load,nusb3xhc,nusb3xhc.sys

IniWrite,%target_win%\txtsetup.sif,HardwareIdsDatabase,PCI\VEN_1033&DEV_0194,nusb3xhc

//Hub

IniWrite,%target_win%\txtsetup.sif,SourceDisksFiles,nusb3hub.sys,"1,,,,,,4_,4,1,,,1,4"

IniWrite,%target_win%\txtsetup.sif,BusExtenders.Load,nusb3hub,nusb3hub.sys

IniWrite,%target_win%\txtsetup.sif,HardwareIdsDatabase,NUSB3\ROOT_HUB30,nusb3hub

IniWrite,%target_win%\txtsetup.sif,HardwareIdsDatabase,NUSB3\CLASS_09,nusb3hub

[MedEvil]

ideal, is there a chance ?

Like discussed before, USB 3.0 is a standard, manufacturers can't just do what they want. If they don't stick to the standard, they're legally not allowed to call their controller a USB 3.0 controller, that simple.

Secondly, by darrens test with Win8, it is proven that a single driver can be used for at least Renesas and AMD controllers.

So yes it can be done and it should work out of the box. We just have to figure out, why it doesn't. My guess goes toward, that the manufacturers have crippled their drivers intentionally, to not work with devices from other manufacturers.
Wouldn't be the first time i see this nonsense.