$1900 Bug Bounty to Fix the Lenovo Legion Pro 7 16iax10h's Speakers on Linux

There is a common problem with Realtek ALC3306 on Linux (Kernel Bug 213159). This affects many Lenovo laptop models. For example, my fairly old Legion S7 15IMH5 laptop also does not work.

I'm not willing to pay $1000 for a fix (it's easier for me to buy a new laptop that will work with Linux), but $100 is probably okay. :)

If there's some Lenovo EU rep with enough budget for their department, this person should really be contracted to fix the audio from their laptops one by one. This seemed to take 2 weeks with a dedicated person and remote access but debugging with access to the device could be much faster!

I heard from a frontier coder at deepmind that Gemini, whilst not great at novel frontier coding, is actually pretty good at debugging. Wonder if someone is going to automate a bug bounty pipeline with AI tools.

This has been an old problem with Legion laptops. All this will be available free of cost to everyone! Mad respect for people who are pledging their own money and the person who fixed it.

Also, Lenovo Legion Pro 7* are not cheap (not that this would have been justified for cheap laptops).

Shame on Lenovo/<big company> who should have fixed this years ago.

I would stay away from anything that's using Nvidia, especially a laptop.

Funnily enough the sound doesn't work properly on Windows 10 on this laptop. In exactly same manner as on Linux. Did open source win again?

Wow... one solution is of course to deactivate these. Which is what I did for my Legion 5 Gen 10. Speakers don't seem to be much in use these days anyway.

Still, I didn't expect this amount of custom configuration for my new laptop. Most importantly Bluetooth sound and getting Nvidia driver support. For Bluetooth I ended up writing my own tiny daemon. While driver support exists there seems to be a race condition somewhere between Pipewire, systemd and the bluetooth drivers. And for Nvidia drivers I ended up using the CUDA driver repository which is curiously only available for Debian 12.

If only people would do these bounties for ITE sensor hubs.

Many of these firmware blobs are non redistributable, if you require this blob, you should grab it and back it up. I know a lot of firwmare can't be legally distributable in linux-firmware.

Lenovo may not be as friendly as IBM to its opensource.

Can someone explain exactly what's happening here? https://github.com/nadimkobeissi/16iax10h-linux-sound-saga/i...

It seems like there's a lot of personal information being asked for / thrown around... including a debit/credit card number?

Is there no better way to handle the bounty payment?

Interestingly the Awinc aw88399 smart amplifier chip at the core of this issue allegedly got supporting in 6.7, in 2023. https://www.phoronix.com/news/Linux-6.7-Sound

I have a couple old-ish Samsung Galaxy Book x86 tablets that have a similar issue, that I have never quite goaded myself into trying to reverse engineer. I'd love some better material on trying to reverse engineer windows drivers: presumably maybe running windows in qemu with some kind of intercepting pass through?

IF $1900 is the bounty, it means it doesn't hurt enough.

I remember going for the highest paying bounty in the Ethereum VM several years ago (I think it was ~$400 DAI/SAI). I did it because I wanted to force myself to learn the internals and to see for myself if the bounty system works. I think I spent a few weeks debugging and ended up splitting the bounty.

As long as the user-facing issues are disconnected from the technical issues, it's going to be hard to get the true value.

Maybe related:

- https://gist.github.com/felipelalli/6179aac72735fd35ea3a9854... - https://github.com/NixOS/nixos-hardware/issues/1039

Intel HDA was supposed to be a better standard than the AC'97 it was meant to replace. IMHO the blame lies solely on the codec makers for not working with the default settings (they can add additional functionality, but the base audio I/O should work with a generic HDA driver.)