Problems with VIA USB 2.0 in XP SP2 and especially repair installs

[clasys]

I am a newcomer to this forum, but not a newbie. Hopefully I can add some value to the general discussion.

I have some specific issues and a problem. Hopefully others have been there and done that so I can the problem solved. At least I can make others aware of the problem's existence.

The problem has been around since before SP2, but since SP2, XP has become broken with respect to the EHCI driver. While USUALLY fixable, I am stuck in a situation where I don't know what/how to do a fix.

First, the problem:

Despite the fact that your system appears to totally work [or at least System Device Manager APPEARS to be OK], there is actually a problem:

If you look in the Driver File Details, you will see only a stack of five files. This indicates a problem!

The problem is that there is missing a SIXTH file named:

\WINDOWS\system32\Drivers\vulfnth.sys

The lack of this file makes the USB 2.0 work flaky depending on what you hook up to the machine. Sometimes you would swear you have a marginal hardware error. Fixing the driver makes the entire problem go away.

Back before SP2, the then-prevailing VIA drivers [usually on CDs from the mobo manufacturer] solved the problem; they were usually known as "XP SP1" drivers, etc.

However, when SP2 came out, I checked in great detail what was provided. In essence it was two distinct collections, both wrong:

If you installed or slipstreamed SP2 you got one broken collection. When the RTM release came out, it was specifically different, but equally broken.

In essence, the old problem reappeared, namely that the vulfnth.sys file was yet again among the missing, etc.

Fortunately, more recent releases from VIA have fixed this, so if you install same, you are in great shape!

I have read some posts here that some people believe they have solved their driver problem by getting XP/SP2 to install the built-in driver. I would respectfully dispute this in light of the above. Can anyone confirm that they were able to get XP SP2 to actually provide a driver stack that includes the sixth file? If so, I would be really interested in an example of it setup correctly. And if not, i.e., they only have the five files, the VIA package [270 or perhaps 260] should fix it. Towards that end, if I have helped anyone get their USB 2.0drivers into better shape, I am glad to help.

Now to my specific problem:

I had a system based on 8235 [six USB ports] on a KT400-based mobo, and all of the above applied, totally working, etc. The problem is that the system got dinged so badly that it would no longer boot. A repair install was attempted, with a lot of problems solved I won't bring up here; suffice to say I got all of the REST of the problems solved.

But, yet again, that pesky problem is back! However, I cannot get the 270 package to solve it. In essence, it is ignoring my attempt to install [or uninstall] the messed-up state apparently caused by installing over a system that had it right, but partially forced back to the SP2 level. [Repair installs wipe out all drivers and system hotfixes, etc.]

So here I am with the same problem haunting me, but can this be fixed, or am I doomed to having to totally reinstall for the want of this nail?

Any help appreciated; will try anything short of a re-install [which wouldn't have this problem, etc.]

cjl [veteran of over 100 VIA-based boards]

[tcsenter]

Originally Posted by clasys

In essence, the old problem reappeared, namely that the vulfnth.sys file was yet again among the missing, etc.

This file is not provided by Microsoft in-box on any version of Windows or Service Packs. It is only provided by VIA or by hardware vendors on driver CDs or the vendor website.

So it is not going to magically appear on any Windows CD, slipstreamed or not, since Microsoft doesn't supply this driver at all. Someone would need to custom prepare a Windows install CD to include the VIA drivers. There are methods to accomplish this:

How to Add OEM Plug and Play Drivers to Windows XP

How to add OEM Plug and Play drivers to Windows installations

Svrops.com - Windows XP Unattended Install

nLite - Deployment Tool for the bootable Unattended Windows installation

I have read some posts here that some people believe they have solved their driver problem by getting XP/SP2 to install the built-in driver. I would respectfully dispute this in light of the above. Can anyone confirm that they were able to get XP SP2 to actually provide a driver stack that includes the sixth file?

I can confirm the "in-box" USB drivers supplied with SP2 and SP3 are fully functioning with numerous different USB devices on dozens of VIA-based computers I have either built or serviced in the past four years, whether it is through the integrated USB controllers on VT8235 and VT8237 Southbridges, or the discrete VT6212 and VT6214 chips used on USB2.0 add-in PCI cards.

That is not to suggest I have never encountered glitches or problems with USB devices on VIA controllers, because I have. Most of which I encountered prior to Service Pack 2 for XP, particularly with the VT6202 PCI chip or older VIA motherboards released prior to 2004. In fact, I have for a few years now recommended avoiding PCI cards with VT6202 chip.

Although the VT6212 and VT6214 are much improved, and I personally have used both chips with no problems, I recommend NEC USB2.0 chips over VIA, if only because so many others report problems with the VIA chips.

That said, the Microsoft drivers supplied with SP2 and later are correct for VIA USB2.0 controllers. The drivers are not 'incomplete' if vulfnth.sys is not installed. This file is not supposed to be installed unless there is a particular reason to do so, such as an actual problem with a USB device.

I haven't installed the VIA USB drivers on any system in a few years now. Even when I have encountered USB problems in the past, installing the VIA drivers usually did not solve anything.

[clasys]

Originally Posted by tcsenter

This file is not provided by Microsoft in-box on any version of Windows or Service Packs. It is only provided by VIA or by hardware vendors on driver CDs or the vendor website.

So it is not going to magically appear on any Windows CD, slipstreamed or not, since Microsoft doesn't supply this driver at all. Someone would need to custom prepare a Windows install CD to include the VIA drivers. There are methods to accomplish this:

How to Add OEM Plug and Play Drivers to Windows XP

How to add OEM Plug and Play drivers to Windows installations

Svrops.com - Windows XP Unattended Install

nLite - Deployment Tool for the bootable Unattended Windows installation


I can confirm the "in-box" USB drivers supplied with SP2 and SP3 are fully functioning with numerous different USB devices on dozens of VIA-based computers I have either built or serviced in the past four years, whether it is through the integrated USB controllers on VT8235 and VT8237 Southbridges, or the discrete VT6212 and VT6214 chips used on USB2.0 add-in PCI cards.

That is not to suggest I have never encountered glitches or problems with USB devices on VIA controllers, because I have. Most of which I encountered prior to Service Pack 2 for XP, particularly with the VT6202 PCI chip or older VIA motherboards released prior to 2004. In fact, I have for a few years now recommended avoiding PCI cards with VT6202 chip.

Although the VT6212 and VT6214 are much improved, and I personally have used both chips with no problems, I recommend NEC USB2.0 chips over VIA, if only because so many others report problems with the VIA chips.

That said, the Microsoft drivers supplied with SP2 and later are correct for VIA USB2.0 controllers. The drivers are not 'incomplete' if vulfnth.sys is not installed. This file is not supposed to be installed unless there is a particular reason to do so, such as an actual problem with a USB device.

I haven't installed the VIA USB drivers on any system in a few years now. Even when I have encountered USB problems in the past, installing the VIA drivers usually did not solve anything.

I appreciate your response but must beg to differ with much of it:

1) It is a fact that the USB driver supplement is actually available here on viaarena and literally has been for years, despite the earlier problems we agree where MS demanding that only CD versions distributed with motherboards were [at that time] the only way to go.

2) You are specifically incorrect. I define "broken" as the absence of the filter file. The specific facts are that a) A broken version of the VIA driver is in Service Pack 2; go check for yourself, b) the RTM SP2 XP release also contains the driver, albeit broken in a specifically different way with regard to driver dates. Again, go confirm for yourself. Either way, I can install using only SP2 XP install media and get the driver installed directly without using any VIA-supplied download files or CDs. The board is a TYAN S2495ANRS that I am using for these purposes. It was supplied with a now-obsolete CD that only would work post-SP1, and not at all on SP2.

3) Whenever any of these drivers are installed, vulfnth.sys is specifically missing. By installing the VIA files available here, including the current 270 release after the fact, the vulfnth.sys file is added on. Clearly it is not part of any release from MS. But for SP2 purposes, the 260 or 270 release [perhaps some earlier formerly available here; I never tried those earlier versions] do upgrade the driver to include the filter file.

4) Whenever the filter file is not present, the board predictably exhibits behavior suggesting the hardware is flakey. Whenever the file is restored, the hardware works perfectly fine. Clearly, the file is mandatory for this class of hardware to work. If you are saying that for some newer implementations it MIGHT be unnecessary, I cannot speak to that, but clearly there is no advantage to NOT installing the file, if you can [if you can elect that; in my repair install I have no choice but to not get it]. Thus, I can only surmise that certain users are setting themselves up for problems because they provably do not have the file; we agree that without having the 270 or similar VIA files, the file is definitely not present.

5) When I repair installed this system, I now have lost the ability to put back the file which I have demonstrated is vital to this board. Anyone in the same situation therefore will inevitably be in trouble should they elect to do a repair install. For myself, I seem to now be forced into a total reinstall from scratch unless I can figure out how to get the 270package to override the situation caused by the contamination effect of an otherwise workable repair install, etc.

Thus, my twofold interest is in spreading the clearly true nature of the general install problem, suggesting people are free to upgrade the driver with available [from here] downloads, and doing so would at worst do no harm, and more likely will upgrade the system's reliability depending on exact configuration; I don't see a scenario suggesting not to try.

The other half is to seek help getting past my problem where now my system has become predictably unstable unless I can solve the problem the repair install has caused. Thus, my seeking help in making the 270 package installation more robust so that it installs correctly as it did when the system was initially installed before many, many hours of subsequent apps, etc. installs, which is unfortunately a truly large amount of work merely to overcome the quirkiness of a driver install, etc.

I simply cannot relate to a system where adding in the file doesn't fix anything yet was broken; my experience is the total opposite where the file must be present to prevent demonstrable instability. I use a variety of USB devices ranging from cameras to USB CF etc. drives to USB 2.0-based hard disks, etc. Perhaps you never got the right combination to experience the horror, etc. that then was totally solved by upgrading the driver.

This also applies to VIA-chip-based add-in cards BTW.

I totally agree with you that NEC chip-based cards totally work and are no hassle. The built-in drivers in XP since at least going back to the original SP1 just work fine. In fact, the Adaptec PCMCIA version of the NEC chipset is provided with an optional driver; the built-in XP SP2 driver works better on this card!

I will check out your links to see if anything there is applicable to adding in the 270-version of the driver over and above the already-present no-file version that absolutely IS in SP2, which in turn is apparently consistent with the posts of others here, etc.

I'll report back if anything works, or I can invent some other way of getting around the repair install problem, etc. In any case, thanks for your input.

cjl

[tcsenter]

Originally Posted by clasys

1) It is a fact that the USB driver supplement is actually available here on viaarena and literally has been for years

I'm aware the USB driver is available for download from VIA. I have multiple posts dating back to 2005 in which I've provided the download link for this driver.

2) You are specifically incorrect. I define "broken" as the absence of the filter file. The specific facts are that a) A broken version of the VIA driver is in Service Pack 2; go check for yourself, b) the RTM SP2 XP release also contains the driver, albeit broken in a specifically different way with regard to driver dates. Again, go confirm for yourself.

A driver which is not supposed to be included by Microsoft in any version of Windows XP or Service Packs cannot be "broken" merely by its absence from those products. By definition, the absence of something that is supposed to be absent is the antithesis of "broken". While you are certainly free to invent your own definition that is not used by anyone else on earth, it has the predictable and obvious consequence of making effective communication extremely difficult (i.e. broken).

Again, the VIA USB filter driver (vulfnth.sys) is not and has never been provided by Microsoft in any form or manner, nor is it supposed to be. The purpose of this lower filter driver on Windows XP was to address specific issues unique to VIA's hardware that Microsoft-provided drivers were not able to provide for at the time.

Through a combination of improvements in VIA's USB firmware (BIOS) code and Microsoft's USB driver stack, VIA's filter driver is no longer required, except for older hardware that never received the updated firmware or BIOS code. This is the reason VIA has not updated this driver since 2005 - its no longer relevant (except for older hardware as mentioned).

4) Whenever the filter file is not present, the board predictably exhibits behavior suggesting the hardware is flakey. Whenever the file is restored, the hardware works perfectly fine. Clearly, the file is mandatory for this class of hardware to work.

The VIA filter driver may be required for this particular Tyan motherboard due to BIOS or firmware flaws that never got resolved before Tyan phased-out support, but it is not mandatory for "this class of hardware" to function.

I simply cannot relate to a system where adding in the file doesn't fix anything yet was broken; my experience is the total opposite where the file must be present to prevent demonstrable instability. I use a variety of USB devices ranging from cameras to USB CF etc. drives to USB 2.0-based hard disks, etc. Perhaps you never got the right combination to experience the horror, etc. that then was totally solved by upgrading the driver.

See my post in this topic from 2005:

USB on an MSI MS-6738 + Win2k

To that dated list of tested and fully working USB devices (and VIA Southbridges), you may add:

Canon MP530 AIO printer
Nikon Coolpix Digital Camera
Several HP printer models
Several USB2.0 flash memory drives
Several USB2.0 external 2.5" HDD enclosures with Alcor Micro chips
Several USB2.0 WLAN adapters with Zydas, Ralink, and Broadcom WLAN chips

VIA VT8237R
VIA VT8237A
VIA VT8251

[clasys]

To tscenter:

In the interest of brevity, I am not quoting word-for-word what you posted, I will attempt to paraphrase and minimize the points made as necessary, etc.

To reiterate the problem:

I have a board that has the problem I stated, and nothing can be done to "handwave" away the problem as I stated it before. I'm glad for you that you have a totally different board that doesn't exhibit my specifically and repeatedly observed problems.

I never made any claims about Microsoft providing the relevant filter file, however, since clearly the contents first of the released XP Service Pack 2, and then later, the somewhat different replacement VIA innards vis-a-vis this problem clearly indicate that VIA and Microsoft were attempting to include driver support for XP once raised to the SP2 install level. That the two attempts differ from each other despite merely the short time frame between release of a stand-alone SP and the RTM release embedding it into the install CD, tells me that the work performed back then left something to be desired at the minimum.

Relative to my problem, neither of them work. Before SP2, the drivers provided on the CD associated with the motherboard worked fine for post-SP1 systems; the subsequent VIA downloads brought the same support for post-SP2 systems, as long as the system was never repair-installed.

Few people actually do repair installs; fewer still are prepared to notice nuanced problems caused by the repair install; I am for whatever dubious worth implied, one of them, since I made heavy use of this motherboard having generally good performance and reliability in all other aspects driver-related, and when this problem of the filter file was fixed, no problems whatsoever, etc.

However, the driver, not worked on since 2005, is in fact defective. I defined my term "broken" and you can object to it all you want. Regardless, the driver as provided by VIA is not capable of providing what it perfectly is capable of doing on an initial installation on a repair installation. And in my specific case with this particular motherboard, that is essentially a fatal problem. It is perfectly understandable that VIA never tested it through a repair-install cycle, etc.

Regarding some of your other points, I fail to see how your experience with a board that used different VIA chips from mine negates my statement about "this class of motherboard" since AFAIK you should have the same stuff on it to qualify. It seems you have better, not the same, etc. and perhaps that explains why you need not care about this issue; unfortunately, I don't have that "luxury" etc. Part of my original post was to pass that warning on to others who might potentially have the same problem.

Additionally, XP takes nothing from the BIOS for USB other than gross identification purposes which are primarily self-discovered by XP's plug-and-play software. About the only thing relevent is the exact quantity of root hubs defined as a function of how many interfaces are enabled. That's not a BIOS rev level issue, rather just enabling or not the hardware for XP to self-discover, etc.

In any case, I leave it to VIA as the final authority. When 260 or 270 is installed on an initial installation on this board, the filter file is always available. However, once a repair install mangles the driver to no longer have it, this same VIA code is incapable of restoring the original driver stack. Clearly, this is a bug since you cannot get identical results before and after a repair install. [At least one of the situations is broken; you can voice your own opinion about which one is which.]

I guess we can just chalk this up to the fact that we have separate experiences, and neither of us should be commenting on the "validity" of the other, etc.

Thus, I am still seeking a remedy, to cause the before and the after to agree with each other. The only other alternative for this otherwise serviceable board is to start the install all over again. What I actually do to make a complete installation is a rather laborious process I would sincerely like to avoid given that it would appear to otherwise have been successful, etc.

cjl