Hopefully, this Usenet post will reach people on the web. God knows I
have no interest in using Reddit again.

If you have an ASUS Zephyrus G14, or any laptop using a Goodix
fingerprint reader under Windows 11, you might have lost functionality
over the past year. The fingerprint sensor comes and goes as if the
hardware was defective.

It turns out that Windows 11 is forcing an update of the hardware
through Windows Update. In my case, it replaced the 3.0.21.260 driver
with a 3.0.21.290 driver. While the .260 driver works fine, the newer
one is causing the sensor itself to crash, requiring a _hardware_ reset
to get it going again. While downgrading to Windows 10 would work since
..290 seems to have been designed with that operating system in mind, not
everyone is too keen on using an older version of Windows.

To remedy the situation without too much trouble, in the GA401Qx models
at least:

1) Disable the fingerprint sensor in BIOS
2) Log into Windows using your pin
3) Go to Device Manager, click View, select Show Hidden Devices
4) In Universal Series Bus controllers, remove every instance of "Device
descriptor failed"
5) In Biometric devices, choose the Goodix fingerprint sensor,
Properties, Driver, Uninstall Device (select to try to remove the driver)
6) Download the older driver for the sensor. In my case,
<https://shorturl.at/oNU01>
7) Extract the driver to a specific folder of your choice (don't install it)
8) Restart the computer, go into BIOS, re-enable the fingerprint sensor
9) Whether Windows installs the driver automatically at bootup or
doesn't, go back into Device Manager and select your fingerprint sensor,
right-click then Properties.
10) Select Driver then Update Driver.
11) Select Browse my Computer for Drivers and enter the folder where you
extracted the driver.
12) Restart.

While not a complicated process, it is much more complicated than it
needs to be simply because Windows *insists* on using the poorly-written
..290 driver. On a fresh install, the possibility of rolling back is
simply not available, so you have to force the operating system to use
the previous version.

--
TG: @RabidPedagog

On 2023-12-08 14:48, RabidPedagog wrote:

....

> 6) Download the older driver for the sensor. In my case,
> <https://shorturl.at/oNU01>

Please do not use short urls here. Use the full URL.

<https://dlcdnets.asus.com/pub/ASUS/GamingNB/DriverforWin10/Fingerprint/FINGERPRINTWBFUSB_ROG_GOODIX_Z_V3.0.21.260_20076.EXE?model=GA401QM>

--
Cheers,
Carlos E.R.

"Carlos E. R." <robin_listas@es.invalid> wrote:

> RabidPedagog wrote:
>
>> ...
>> 6) Download the older driver for the sensor. In my case,
>> <https://shorturl.at/oNU01>
>> ...
>
> Please do not use short urls here. Use the full URL.
> <https://dlcdnets.asus.com/pub/ASUS/GamingNB/DriverforWin10/Fingerprint/FINGERPRINTWBFUSB_ROG_GOODIX_Z_V3.0.21.260_20076.EXE?model=GA401QM>

Or show both so the reader can choose depending on whether or not their
choice of NNTP client will retain or slice apart (insert newlines in) a
long URL, like:

6) Download the older driver for the sensor. In my case,
Long URL: https://dlcdnets.asus.com/pub/ASUS/GamingNB/DriverforWin10/Fingerprint/FINGERPRINTWBFUSB_ROG_GOODIX_Z_V3.0.21.260_20076.EXE?model=GA401QM
Short URL: https://shorturl.at/oNU01

Or, just show the long URL, and let readers handle whatever screwups
their client performs on long URLs, or on any lines physically longer
than the width of the viewing window (the client should *not* wrap long
physical lines, but instead provide for horizontal scrolling).

Short URLs are subject to link rot. If the shortening service goes
away, all the short URLs become useless. Short URLs may be removed if
the service gets reports of abuse or malicious use.

https://www.godaddy.com/whois/results.aspx?domain=shorturl.at

shorturl.at is registered with GoDaddy who shows the domain was first
registered back in 2016. Usenet is a lot older than that, and despite
the rumors Usenet has not yet died, and will probably last a lot longer.
Meanwhile redirection services come and go.

There's also the privacy concern about a shortening service tracking the
use of their redirections, like who used them (IP tracking) and to where
did they go. I'm not familiar with the privacy laws of Brazil where is
the domain's registrant. They have a privacy policy at:

https://www.shorturl.at/privacy-policy.php

But that requires you trust they perform as stated.

RabidPedagog <rabid@pedag.og> wrote:

> Hopefully, this Usenet post will reach people on the web. God knows I
> have no interest in using Reddit again.

This is not the Web. That uses HTTP[S]. This is Usenet which uses
NNTP. The Web is a bunch of sites. Usenet is a forum or bulletin board
type of communication.

https://en.wikipedia.org/wiki/World_Wide_Web
https://en.wikipedia.org/wiki/Usenet

Perhaps instead of "web" you meant to say "Internet". Web and Usenet
are subordinate under Internet.

> If you have an ASUS Zephyrus G14, or any laptop using a Goodix
> fingerprint reader under Windows 11, you might have lost functionality
> over the past year. The fingerprint sensor comes and goes as if the
> hardware was defective.
>
> It turns out that Windows 11 is forcing an update of the hardware
> through Windows Update. In my case, it replaced the 3.0.21.260 driver
> with a 3.0.21.290 driver. While the .260 driver works fine, the newer
> one is causing the sensor itself to crash, requiring a _hardware_ reset
> to get it going again. While downgrading to Windows 10 would work since
> .290 seems to have been designed with that operating system in mind, not
> everyone is too keen on using an older version of Windows.
>
> To remedy the situation ...

I don't see your manipulations will prevent WU from later attempting to
update the driver again. You manually force selection of a particular
driver and version of it, but how does that stop WU from "updating" it
again sometime later?

> While not a complicated process, it is much more complicated than it
> needs to be simply because Windows *insists* on using the
> poorly-written .290 driver. On a fresh install, the possibility of
> rolling back is simply not available, so you have to force the
> operating system to use the previous version.

Configure Windows to *not* perform hardware updates. Microsoft is too
often incorrect when performing brain surgery on kernel-mode interfaces
to hardware (aka drivers). Just because a new driver is available does
not mean it applies to your hardware nor provide any new functionality.
If the hardware driver is working, and doesn't lack any support for your
hardware, don't change it.

I've use WinAero Tweaker to disable hardware updates. Go to Behavior ->
Disable hardware updates. However, you can do the same with a registry
edit. WinAero has their own help article:

https://winaero.com/how-to-turn-off-driver-updates-in-windows-update-in-windows-10/

The instructions are for Windows 10. It mentions using itself along
with using the group policy editor (gpedit.msc) the latter of which is
not available in Home editions of Windows. However, all policies are
registry entries, so they show how to edit the registry to effect the
policy to disable hardware updates. Unfortunately Microsoft has the
habit of ignoring or abandoning old registry settings, so it's possible
Windows 11 does not obey this registry setting.

https://www.minitool.com/news/enable-disable-automatic-driver-updates-windows-11.html

That mentions a different policy aka registry edit to disable hardware
updates in Windows 11. Microsoft likes to keep a moving target on how
they implement update control.

If possible, only you should be retrieving hardware updates to apply
them after investigating if they apply to your particular hardware
(instead of, for example, a family of products for a specific version of
which yours is not one). For example, different versions of a driver
may support different hardware chips, so you have to select the driver
version that matches your hardware, not a family of hardware. If a new
driver does not apply to your specific hardware, you don't suffer the
bugs mentioned as fixed in the new driver version, or adds features that
you don't need, then stick with the old driver that is working.

Don't fix what ain't broke. A newer version does not mandate it is
necessary or even desirable. New code might fix old bugs, but new code
also introduces new bugs that have you waiting until yet another version
to fix. Another problem is the time to get a corrected driver into
Microsoft catalog. A vendor releases a driver, it takes months before
it gets into the MS catalog, the vendor finds a defect in the new
driver, but it takes months to get the bad one out of MS catalog to
replace with a new driver. During that time, users are getting the
defective driver that is way too slow to get updates in the MS catalog.

Disable hardware updates, and YOU do them yourself. You are usurping
the role of sysadmin on your computer, so do the job yourself. Leaving
it to Microsoft ensures you'll get incorrect or wrong drivers. New
doesn't help at all in a working scenario.

On 2023-12-08 11:29 a.m., VanguardLH wrote:
> RabidPedagog <rabid@pedag.og> wrote:
>
>> Hopefully, this Usenet post will reach people on the web. God knows I
>> have no interest in using Reddit again.
>
> This is not the Web. That uses HTTP[S]. This is Usenet which uses
> NNTP. The Web is a bunch of sites. Usenet is a forum or bulletin board
> type of communication.
>
> https://en.wikipedia.org/wiki/World_Wide_Web
> https://en.wikipedia.org/wiki/Usenet
>
> Perhaps instead of "web" you meant to say "Internet". Web and Usenet
> are subordinate under Internet.

Except that Google makes Usenet articles accessible through a search on
their engine. I thought that was common knowledge.

>> If you have an ASUS Zephyrus G14, or any laptop using a Goodix
>> fingerprint reader under Windows 11, you might have lost functionality
>> over the past year. The fingerprint sensor comes and goes as if the
>> hardware was defective.
>>
>> It turns out that Windows 11 is forcing an update of the hardware
>> through Windows Update. In my case, it replaced the 3.0.21.260 driver
>> with a 3.0.21.290 driver. While the .260 driver works fine, the newer
>> one is causing the sensor itself to crash, requiring a _hardware_ reset
>> to get it going again. While downgrading to Windows 10 would work since
>> .290 seems to have been designed with that operating system in mind, not
>> everyone is too keen on using an older version of Windows.
>>
>> To remedy the situation ...
>
> I don't see your manipulations will prevent WU from later attempting to
> update the driver again. You manually force selection of a particular
> driver and version of it, but how does that stop WU from "updating" it
> again sometime later?

It might, but the old one is still accessible this way through a
rollback. As far as I know, Windows will update a driver but respect a
user's decision to roll back after the fact and never try again.

< snip good advice >

--
TG: @RabidPedagog

On 2023-12-08 17:29, VanguardLH wrote:
> "Carlos E. R." <robin_listas@es.invalid> wrote:
>
>> RabidPedagog wrote:
>>
>>> ...
>>> 6) Download the older driver for the sensor. In my case,
>>> <https://shorturl.at/oNU01 " rel="nofollow" target="_blank">https://shorturl.at/oNU01>
>>> ...
>>
>> Please do not use short urls here. Use the full URL.
>> <https://dlcdnets.asus.com/pub/ASUS/GamingNB/DriverforWin10/Fingerprint/FINGERPRINTWBFUSB_ROG_GOODIX_Z_V3.0.21.260_20076.EXE?model=GA401QM " rel="nofollow" target="_blank">https://dlcdnets.asus.com/pub/ASUS/GamingNB/DriverforWin10/Fingerprint/FINGERPRINTWBFUSB_ROG_GOODIX_Z_V3.0.21.260_20076.EXE?model=GA401QM>
>
> Or show both so the reader can choose depending on whether or not their
> choice of NNTP client will retain or slice apart (insert newlines in) a
> long URL, like:
>
> 6) Download the older driver for the sensor. In my case,
> Long URL: https://dlcdnets.asus.com/pub/ASUS/GamingNB/DriverforWin10/Fingerprint/FINGERPRINTWBFUSB_ROG_GOODIX_Z_V3.0.21.260_20076.EXE?model=GA401QM
> Short URL: https://shorturl.at/oNU01
>
> Or, just show the long URL, and let readers handle whatever screwups
> their client performs on long URLs, or on any lines physically longer
> than the width of the viewing window (the client should *not* wrap long
> physical lines, but instead provide for horizontal scrolling).
>
> Short URLs are subject to link rot. If the shortening service goes
> away, all the short URLs become useless. Short URLs may be removed if
> the service gets reports of abuse or malicious use.
Besides the short url service disappearing or removing the link, the
short link is devoid of information. There is no way to know, before
hand, what it may point to. It could even be a malevolent site.

Just looking at the full link we can see it is a link to an asus site so
we know what kind of think it downloads, without clicking on it. Even if
the short link dies, or the destination is removed, we can see in the
future what it was.

And in this case, the download starts instantly, it is difficult to see
the link.

--
Cheers,
Carlos E.R.

On 2023-12-08 17:51, RabidPedagog wrote:
> On 2023-12-08 11:29 a.m., VanguardLH wrote:
>> RabidPedagog <rabid@pedag.og> wrote:
>>
>>> Hopefully, this Usenet post will reach people on the web. God knows I
>>> have no interest in using Reddit again.
>>
>> This is not the Web.  That uses HTTP[S].  This is Usenet which uses
>> NNTP.  The Web is a bunch of sites.  Usenet is a forum or bulletin board
>> type of communication.
>>
>> https://en.wikipedia.org/wiki/World_Wide_Web
>> https://en.wikipedia.org/wiki/Usenet
>>
>> Perhaps instead of "web" you meant to say "Internet".  Web and Usenet
>> are subordinate under Internet.
>
> Except that Google makes Usenet articles accessible through a search on
> their engine. I thought that was common knowledge.

A search engine that apparently is broken.

And that archive is under attack from spammers, while its administrators
do nothing. It could die any day.

--
Cheers,
Carlos E.R.

On 2023-12-08 12:26 p.m., Carlos E. R. wrote:
> On 2023-12-08 17:51, RabidPedagog wrote:
>> On 2023-12-08 11:29 a.m., VanguardLH wrote:
>>> RabidPedagog <rabid@pedag.og> wrote:
>>>
>>>> Hopefully, this Usenet post will reach people on the web. God knows I
>>>> have no interest in using Reddit again.
>>>
>>> This is not the Web.  That uses HTTP[S].  This is Usenet which uses
>>> NNTP.  The Web is a bunch of sites.  Usenet is a forum or bulletin board
>>> type of communication.
>>>
>>> https://en.wikipedia.org/wiki/World_Wide_Web
>>> https://en.wikipedia.org/wiki/Usenet
>>>
>>> Perhaps instead of "web" you meant to say "Internet".  Web and Usenet
>>> are subordinate under Internet.
>>
>> Except that Google makes Usenet articles accessible through a search
>> on their engine. I thought that was common knowledge.
>
> A search engine that apparently is broken.
>
> And that archive is under attack from spammers, while its administrators
> do nothing. It could die any day.

Either way, if the post helps even a single person, it'll be worth it.

--
TG: @RabidPedagog

On 2023-12-08 18:46, RabidPedagog wrote:
> On 2023-12-08 12:26 p.m., Carlos E. R. wrote:

> Either way, if the post helps even a single person, it'll be worth it.
>

I have saved the post to my personal archive.

--
Cheers,
Carlos E.R.

RabidPedagog <rabid@pedag.og> wrote:

> On 2023-12-08 11:29 a.m., VanguardLH wrote:
>> RabidPedagog <rabid@pedag.og> wrote:
>>
>>> Hopefully, this Usenet post will reach people on the web. God knows I
>>> have no interest in using Reddit again.
>>
>> This is not the Web. That uses HTTP[S]. This is Usenet which uses
>> NNTP. The Web is a bunch of sites. Usenet is a forum or bulletin board
>> type of communication.
>>
>> https://en.wikipedia.org/wiki/World_Wide_Web
>> https://en.wikipedia.org/wiki/Usenet
>>
>> Perhaps instead of "web" you meant to say "Internet". Web and Usenet
>> are subordinate under Internet.
>
> Except that Google makes Usenet articles accessible through a search on
> their engine. I thought that was common knowledge.

If you're talking about Google Groups, that is Google's web-based forum.
Google acquired the DejaNews archive back around 2002. Google operates
a peering NNTP server, so articles in the Usenet *forums* (accessed via
HTTPS, so the Web) get gatewayed to Usenet. You cannot post to Google
Groups using NNTP, only via HTTP. When using Google Groups, you are
using the Web. There are other forum web sites that implement an
HTTP-to-NNTP gateway, or mailing lists that implement an email-to-NNTP
gateway. These gateways connect Web or email to Usenet.

>>> If you have an ASUS Zephyrus G14, or any laptop using a Goodix
>>> fingerprint reader under Windows 11, you might have lost functionality
>>> over the past year. The fingerprint sensor comes and goes as if the
>>> hardware was defective.
>>>
>>> It turns out that Windows 11 is forcing an update of the hardware
>>> through Windows Update. In my case, it replaced the 3.0.21.260 driver
>>> with a 3.0.21.290 driver. While the .260 driver works fine, the newer
>>> one is causing the sensor itself to crash, requiring a _hardware_ reset
>>> to get it going again. While downgrading to Windows 10 would work since
>>> .290 seems to have been designed with that operating system in mind, not
>>> everyone is too keen on using an older version of Windows.
>>>
>>> To remedy the situation ...
>>
>> I don't see your manipulations will prevent WU from later attempting to
>> update the driver again. You manually force selection of a particular
>> driver and version of it, but how does that stop WU from "updating" it
>> again sometime later?
>
> It might, but the old one is still accessible this way through a
> rollback. As far as I know, Windows will update a driver but respect a
> user's decision to roll back after the fact and never try again.

I'd like some confirmation that rollbacks on drivers remain permanent.
I'll hunt around. I've seen online instructions on how to rollback to a
prior driver version, and even with instructions on how to disable
hardware updates, but I've not seen where rolling back a driver makes
that decision permanent. That's why I mention doing backups, so you can
restore to a state that was working before. Alternatively, disable
hardware updates, reenable when you want to check, but YOU choose which,
if any, to apply, and then re-disable hardware updates. Easier is just
for you to go check the vendor's web site for new drivers, and YOU
decide if you want them, and apply them.

Windows only keeps one copy of the prior driver version. You cannot
walk back through multiple versions to find which was a working version.
Else, you have to point to the installer to get back to that version.

You can use DISM to keep an archive of the current drivers to save for
later reuse as older versions when a hardware update occurs.

dism /online/export-driver /destination:<savepath>

<savepath> could be, for example, C:\Drivers.

Even if rollback were permanent, this won't work if there is no prior
driver to rollback to, like for a fresh install or you wiped the prior
config from the registry. Rollbacks only work for hardware managed by
Device Manager. You don't have to rely on rollbacks, sfc, system
restore points, or other Windows supplied recovery methods if you
maintain your own scheduled image backups that let you restore the setup
to the exact state it was at the time of the backup. Backups works.
Windows-supplied recovery methods are iffy.