[Bug 42424] Multiple programs using Driver Store installer "dpinst.exe" fail to install drivers (NtQueryInformationToken missing support for 'TokenUIAccess') (Renesas Flash Programmer V3.02.01, Hantek 6000B)

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=42424

Zebediah Figura <z.figura12 at gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |z.figura12 at gmail.com
            Summary|Renesas Flash Programmer    |Multiple programs using
                   |V3.02.01 installer fails to |Driver Store installer
                   |install USB drivers         |"dpinst.exe" fail to
                   |(NtQueryInformationToken    |install drivers
                   |missing support for         |(NtQueryInformationToken
                   |'TokenUIAccess')            |missing support for
                   |                            |'TokenUIAccess') (Renesas
                   |                            |Flash Programmer V3.02.01,
                   |                            |Hantek 6000B)

--- Comment #3 from Zebediah Figura <z.figura12 at gmail.com> ---
Another one, from wine-devel:

<https://www.winehq.org/pipermail/wine-devel/2022-January/205250.html>

These programs are both using Microsoft's Driver Store redistributable
installer dpinst.exe (or dpinst64.exe).

Unlike Anastasius at the time this bug was filed, I think it's possible and
worthwhile to get some USB device drivers to work. Some Windows drivers can't
work due to requiring kernel features (the big one is probably going to be
arbitrary user space memory access via non-buffered ioctls), but others should
be able to without difficulty. I've already introduced wineusb.sys, which as of
the time of writing has enough functionality for at least one Windows driver to
work.

Anyway, stubbing TokenUIAccess gets the installer further, but then it breaks
later due to missing Driver Store functions in setupapi. These functions are
undocumented and pretty much nobody has reverse-engineered them either. Given
that dpinst.exe is a Microsoft component itself, and actually has at least some
documentation to it (mostly in the form of its own help message printed when
run with the /? switch) I think implementing a dpinst.exe replacement would be
a more feasible approach.

-- 
Do not reply to this email, post in Bugzilla using the
above URL to reply.
You are receiving this mail because:
You are watching all bug changes.