[Avamander]

A few of the drivers they install (or want to install) are also on Microsoft's vulnerable actively exploited driver blacklist. So that's fun, they have no intention of fixing it because they do not support "third party software". I'm also pretty sure their installer doesn't work without unencrypted HTTP traffic being let through. Plus they keep offering bloatware as "updates" to you.

On top of it all, the software they offer is slow and buggy on brand-new hardware.

But most of those issues also exist with AMD's or Gigabyte's drivers, most hardware vendors seem trashy like that. Like, if you install Samsung Magician (for their SSDs) then that even asks you if you're in the EEA (because of the privacy laws I suspect), it's absolutely crazy.

Microsoft should make it *significantly* harder to ship drivers outside of Windows Update and they should forbid any telemetry/analytics without consent.

I find Linux's hardware support model significantly nicer, although some rarer things do not work OOB, there's none of this bullshit.

[matheusmoreira]

Hardware manufacturers consistently ship out the worst softwares in existence. It's just a cost center to them. They've already sold the thing, it doesn't matter anymore.

My laptop has a fan and keyboard LED application that requires kernel access and takes over a minute to display a window on screen. Not to mention being Windows only.

Words can barely describe just how aggravating that thing was. One of the best things I've ever done is reverse engineer that piece of crap and create a Linux free software replacement. Mine works instantly, I just feed it a configuration file. I intend to do this for every piece of hardware I buy from now on.

[Avamander]

I really wish someone made such software for ASUS and Gigabyte both, without dangerous kernel drivers.

In that sense fwupd has been an amazing development, as there's now a chance that you can update the firmware of your hardware on Linux and don't have to boot Windows.

[matheusmoreira]

Actually having hardware lying around to reverse engineer is the limiting factor for me. I suppose I could give it a shot if people who own the devices sent me the required data. I'd need their help with testing.

USB stuff was really nice to work with. Wireshark made it really easy to intercept the control commands. For example, to configure my keyboard's RGB LEDs I need to send 0xCC01llrrggbb7f over the USB control channel; the ll identifies the LED and rrggbb sets the color. Given this sort of data it's a simple matter to make a program to send it.

Reverse engineering ACPI stuff seems to be more involved. I wasn't able to intercept communications on the Windows side. On Linux I managed to dump DSDT tables and decompile WMI methods but that just gave me stub code. If there's anything in there it must be somehow hidden. I'm hoping someone more experienced will provide some pointers in this thread.

[Hilift]

I thought it was revealing that the driver blacklist (sipolicy.p7b) was not updated for several years until Will Dormann happened to notice it in 2022.

https://www.csoonline.com/article/573965/how-to-update-your-...

[userbinator]

> Microsoft should make it significantly harder to ship drivers outside of Windows Update

No. No no no no no no no NO! That just centralises even more control to MS.

What we really need is for more people to develop open-source Windows drivers for existing hardware, or encourage the use of Linux.

[Avamander]

That will never work on Windows, neither do I want some rando's drivers on my machine.