Linux 5.17 Features From New AMD P-State To Xilinx Drivers, Lots Of New Hardware

[Mez']

Yea but with telemetry, people can't claim that a feature that's only used by them and 100 other people worldwide is actually super vital and used by most people.

True.
Gnome devs must fear telemetry like hell. When they discover how much people rely on extensions, how millions have installed some extension and how nobody ever uses the NIMBY Gnome apps, such as Web, Boxes or Maps, they´re gonna cry in their bunker.

[Myownfriend]

Gnome devs must fear telemetry like hell.

If Gnome added telemetry then the "Gnome is trying to take over the world" conspiracy theorist would flood forums like mad.

When they discover how much people rely on extensions, how millions have installed some extension and how nobody ever uses the NIMBY Gnome apps, such as Web, Boxes or Maps, they´re gonna cry in their bunker.

I actually use Boxes lol Overall the extension info would be more useful than whether or not people use their apps. You can just uninstall their apps if you don't use them.

[tuxd3v]

humm,
Anything that put your private data at risk, is a very bad thing, and if introduced, it was with a very well established intention..
Now lets wait and see, what kind of telemetry is being used, because the name itself even tough is scary, doesn't specify..

[Quackdoc]

humm,
Anything that put your private data at risk, is a very bad thing, and if introduced, it was with a very well established intention..
Now lets wait and see, what kind of telemetry is being used, because the name itself even tough is scary, doesn't specify..

you don't have to wait. go read the code if you don't trust the kernel devs who this entire time they have been at it, have been steadfast in their principals against risking the user. you can see exactly what the code does.

[Alex/AT]

Well, having some telemetry does not make it much worse after the moment huge eBPF backdoor had been introduced.

[Myownfriend]

you can see exactly what the code does.

Feels like the most paranoid people in the OSS community constantly forget that that means the source code freely available for them to see.

[kescherPh]

I feel like some people that have replied after my earlier reply still don't understand what this so-called "telemetry" is for. It is not meant for sending data back to a manufacturer or dev team, but to administrators of machines. Think, detailed sensor data and other stuff the firmware collects alredy, being available to user space so it can be collected and made available to e.g. a monitoring daemon an admin uses.

Please actually read the patches. Or the Phoronix articles, for that matter.

[Alex/AT]

I feel like some people that have replied after my earlier reply still don't understand what this so-called "telemetry" is for

It's a way to read firmware-specific blob with the intent of sending it somewhere (assuming it's called telemetry and not monitoring).

[tuxd3v]

Well, having some telemetry does not make it much worse after the moment huge eBPF backdoor had been introduced.

I believe you can disable unprivileged use with:

sysctl kernel.unprivileged_bpf_disabled=1

Or write it to '/etc/sysctl.conf'( to be permanent.. ), and

sysctl -w -p /etc/sysctl.conf

You are talking about vulnerabilities like this, I believe, well yes a big big problem..

[tuxd3v]

you don't have to wait. go read the code if you don't trust the kernel devs who this entire time they have been at it, have been steadfast in their principals against risking the user. you can see exactly what the code does.

Sometimes even reading the code is not enough..
Like it was already previously proved its possible to make commits that seems to do nothing in the kernel, while in the long run, someone will take advantage of those "ingenuous commits", and exploit them.

No, I am not blaming the Core kernel developers, just this types of things are very sensitive..