[gentoo-user] Google Chrome now requires wayland and jack audio?

[Grant Edwards]

It looks like www-client/google-chrome just added wayland and jack
audio to the dependancies. So now I have to have Pulse _and_ Jack?

--
Grant

[Julien Roy]

One of the side effects of using proprietary software : you can't
control with which flags it gets built.
With chromium-bin, there is a wayland USE flag, but nothing for jack.

--
Julien

[Mark Knecht]

Is that truly a Chrome requirement, like the company Google wrote the ebuild, or is 

this something a Gentoo dev did for some reason?

I'm curious as the USB disconnect problem seems somehow to be related

to using Chrome on the host machine for sites that do a lot of audio, like

YouTube. A clean boot of the host machine, followed by a clean boot of the VM

and I've run for at least an hour with no disconnection problems. I can use 

Chrome for email, messaging and reading newspapers with no problem, but 

I run YouTube and twice I've had USB problems in the VM.

[Grant Edwards]

On 2022-07-15, Julien Roy <jul...@jroy.ca> wrote:

> One of the side effects of using proprietary software : you can't
> control with which flags it gets built.

Yep. I didn't used to have the chrome binary package installed, but
there are a couple things that I've never gotten to work in Chromium
(e.g. Webex).

> With chromium-bin, there is a wayland USE flag, but nothing for
> jack.

I looked into that more, and I had misread the emerge output. It
wasn't google-chrome that depended on jack, and now I can't figure out
why it was installed. I did

# emerge -C virtual/jack media-sound/jack-audio-connection-kit
# emerge -auvND world

It didn't get reinstalled. And then a subsequenct

# emerge --depclean --ask

removed another half-dozen audio-related packagets (zita-* and
realtime-*, whatever they are). I'm sure the next time I try to use
audio on that machine it won't work.

I used to think that someday Linux sound support would get
straightened out, but it just keeps getting worse...

--
Grant

[Grant Edwards]

On 2022-07-15, Mark Knecht <markk...@gmail.com> wrote:
> On Fri, Jul 15, 2022 at 12:28 PM Grant Edwards <grant.b...@gmail.com>
> wrote:
>>
>> It looks like www-client/google-chrome just added wayland and jack
>> audio to the dependancies. So now I have to have Pulse _and_ Jack?

> Is that truly a Chrome requirement, like the company Google wrote
> the ebuild, or is this something a Gentoo dev did for some reason?

Google doesn't provide an ebuild. The ebuild is written maintained by
the kind volunteers of the Chromium in Gentoo Project. For the binary
distribution from Google, those devs have no control over what
libraries the Chrome executables are built to use. All they can do is
try to figure out which libraries Chrome needs, and reflect that in
the ebuild so that after the binary from Google gets installed, it
works.

That said, there was no jack audio requirement for Chrome. I misread
the emerge output. The two new requirements that google-chrome was
pulling in were

dev-libs/wayland
dev-util/wayland-scanner

You don't have to be running Wayland, but you now need the above
wayland pieces.

There isn't actually a pulse audio requirement in the google-chrome
ebuild either, but if I don't have pulse installed, some audio stuff
in Chrome doesn't work. In web apps like Google Voice

* I can select my headset mic as audio in, but it won't work.

* I can't select headset as audio out.

Installing pulse audio fixed those problems.

> I'm curious as the USB disconnect problem seems somehow to be
> related to using Chrome on the host machine for sites that do a lot
> of audio, like YouTube. A clean boot of the host machine, followed
> by a clean boot of the VM and I've run for at least an hour with no
> disconnection problems. I can use Chrome for email, messaging and
> reading newspapers with no problem, but I run YouTube and twice I've
> had USB problems in the VM.

Yep, it sounds like doing audio via Chrome is disrupting the the USB
audio device that's in-use by the VM. Are there Linux audio drivers
for that hardware that you could uninstall to keep Chrome from seeing
it?

--
Grant

[Jigme Datse]

Pipewire will allow you to handle both in a pretty seamless way.
Though it does take a bit of configuration to get it working.

[Mark Knecht]

On Fri, Jul 15, 2022 at 1:56 PM Grant Edwards <grant.b...@gmail.com> wrote:
>
> On 2022-07-15, Mark Knecht <markk...@gmail.com> wrote:
> > On Fri, Jul 15, 2022 at 12:28 PM Grant Edwards <grant.b...@gmail.com>
> > wrote:

<SNIP>

> > I'm curious as the USB disconnect problem seems somehow to be
> > related to using Chrome on the host machine for sites that do a lot
> > of audio, like YouTube. A clean boot of the host machine, followed
> > by a clean boot of the VM and I've run for at least an hour with no
> > disconnection problems. I can use Chrome for email, messaging and
> > reading newspapers with no problem, but I run YouTube and twice I've
> > had USB problems in the VM.
>
>
> Yep, it sounds like doing audio via Chrome is disrupting the the USB
> audio device that's in-use by the VM. Are there Linux audio drivers
> for that hardware that you could uninstall to keep Chrome from seeing
> it?

There is no support in Linux for this hardware. From the computer's 

POV it's just an external USB device, partially an audio device, and 

partially just controlled over USB. I've told pulseaudio and KDE in 

general not to use it but I continue to see problems. I have no idea

what functionality the USB control port is providing.

I think the next step is to actually blacklist the device by its

USB device ID ala something like this:

https://www.projectgus.com/2014/09/blacklisting-a-single-usb-device-from-linux/

and see what happens.

This whole thing isn't overly critical to me. The device itself is 

stand alone in operation. It's only attached to a computer to do editing

which actually can be done on the device's GUI without a computer, or

I can hook it to a Windows laptop, or even this machine if booted into 

Windows. I was just wanting to be in Linux but open a VM to allow

me to edit more easily, which I actually can do but I have to hit the 

reconnect button in software or pull the USB cable, both of which

work but are hacks.

[Daniel Frey]

How is the device for audio set up on the host? It may actually be a
USB-audio device it provides to the VM itself, and that would explain
the issue you are having.

Dan

[Mark Knecht]

Hi Dan,

   It is not set up at all for audio on the host. I don't intend to use it in Linux

at all at this time. Line6 doesn't support Linux, they provide no Linux 

software and my use on the computer at all, at least initially, is simply

to use the DSP edit control software to create & tweak patches. That will

either be by rebooting this machine into Windows, or solving the 

disconnection problems in the Windows VM. However I have found in the

Line6 forums that some people see this disconnection problem even 

running in Windows natively so possibly I've just been lucky and now

seen it in the 2 or so hours I forced myself to use Windows to update

the DSP firmware and do an initial checkout of the machine.

   That said the device is a USB audio interface providing 8 USB input

and 8 USB output channels from the DSP + a stereo dry input over 

USB. (And possibly more stuff over USB but at this time I haven't 

investigated that.) I track, in both Linux and Windows, using Harrison

Mixbus32C (the for pay version of Ardour) and I will certainly be 

investigating that at some future date but for now I simply intend to

take the audio output from the DSP to a normal input preamp interface

for recording and not use USB at all

   From the standpoint of the Linux host my preference is that KDE 

and most Linux apps don't even know the device is present.

Hope this helps,

Mark