Firefox 83 vs. Chrome 87 On Intel Tiger Lake + AMD Renoir Under Linux

[duby229]

Well, that patch breaks nVidia drivers when trying to shim gpl only symbols, so is at least evidence of a possible wrong doing on nVidia side. Now for an actual proof you need access to nVidia source code to prove materially that shim is doing something in violation of the license, hence the kernel have 2 choices:

1.) Actually use that evidence to bring nvidia to trial and request an code audit (i can guess that won't be pretty and will be expensive AF)

2.) keep hiding GPL_only symbols until nVidia have no other way that find an alternative course

The obvious alternative is for nVidia to license their modifications in a GPL compatible license so that they can contribute them upstream. The GPLv2 does allow for binary kernel modules as long as they can stand on their own. nVidia can fix the situation by contributing upstream whatever they need to so that their modules can stand on their own. They won't of course.

[birdie]

Every single time time nVidia gets caught using internal interfaces in their proprietary binary, that -IS- proof of them modifying the kernel and hiding their modifications. They've been caught red handed dozens of times.

"Using" is not equal to "modifying" for f's sake. One of us is not using the English language correctly or you make such leaps of logic it's just pure insanity like I said earlier. Again, the choice to compile and use NVIDIA drivers is on you, and NVIDIA themselves are not technically infringing on GPL as they do not distribute anything derived from or linking to GPL code. They just don't.

Let's just say you're on the side of Linux kernel developers who decide which interfaces are GPL only and which are not at whim and it's your/their grief. I get it. Have fun, just stop misuing the English language, don't bring irrelevant links and throw vapid accusations, thank you very much.

Oh, and make sure you've called for making all kernel APIs GPL only if you're so concerned about NVIDIA's "dirty" tactics. Let's just royally f-k everyone who cannot/doesn't want to streamline their kernel drivers under the God bestowed license and of course hundreds of thousands people depending on such drivers.

Be consistent!

[jrch2k8]

"Using" is not equal to "modifying" for f's sake. One of us is not using the English language correctly or you make such leaps of logic it's just pure insanity like I said earlier. Again, the choice to compile and use NVIDIA drivers is on you, and NVIDIA themselves are not technically infringing on GPL as they do not distribute anything derived from or linking to GPL code. They just don't.

Let's just say you're on the side of Linux kernel developers who decide which interfaces are GPL only and which are not at whim and it's your/their grief. I get it. Have fun, just stop misuing the English language, don't bring irrelevant links and throw vapid accusations, thank you very much.

Oh, and make sure you've called for making all kernel APIs GPL only if you're so concerned about NVIDIA's "dirty" tactics. Let's just royally f-k everyone who cannot/doesn't want to streamline their kernel drivers under the God bestowed license and of course hundreds of thousands people depending on such drivers.

Be consistent!

It is word play on both sides since you both are half right.

1.) Yes, nVidia is USING GPL-only interfaces, the problem lies that unlike regular API you see on user space, this symbols(API) don't provide functionality but implementation access(maybe is wrong call it API? but it is what it is) to something inside the kernel, so in a sense nVidia could be overriding some internal kernel functionality of something through that interface, so technically could be a MODIFICATION hence the code should follow the GPL license.

2.) Kernel's problem is not about screw nVidia but a real security concern since the BLOB could implement all kind of crazy crap from back doors to key loggers through the kernel directly and no one can audit that and nVidia has been very adamant in their refuse through the years which make kernel devs even more wary of the BLOB

3.) There is any proof nVidia is using this symbols in some dark way? nope, since no one have found a way through the lawyer wall at nVidia but in the same sense there is no proof they aren't.

4.) But the drivers are user installed !!!, Well it depends on the legislation, since you are forced to install a driver to use a piece of hardware and said driver compile this module by default without any alternative to provide said hardware functionality without this compiled code, again depending the legislation it may not be considered "user installed" and be free of liability. (there are other countries outside of the US, just saying)

5.) Well, Linux has been GPL(1991) since before nVidia(1993) existed as a company. So nVidia is royally screwing their user not the other way around since the devs are enforcing what nVidia has know from day 0 was there.

6.) If the kernel simply hide all GPL symbols making impossible for nVidia to continue with the current format, they won't loose anything since people will still buy their products while some 0.1% will whine it won't work on Linux and going back to Windows and the customers they actually care about will be easily solved through some arrangement through Red Hat/Oracle(HPC/AI markets) and remain completely unaffected.

The best case scenario could be(assuming 6 applies in the future) is Valve jump in the wagon and get GPU docs from nVidia and bring nouveau up to speed, that way users get a foss driver, kernel devs are happy and we let all fortune 500 slug it out with each other since probably nVidia will allow private reviews for Red Hat/Oracle if needed or any other big fat wallet customer.

[qarium]

You on the other hand can go enjoy incomplete open source AMD drivers which are not bug-free (in fact there are dozens of critical bugs, I welcome you to inspect: FreeDesktop Bugzilla and kernel bugzilla) and which do not support all the features found in their proprietary drivers. So much for fucking freedom and open source. You're living a lie. Good luck with that.

I use Fedora33 on a Threadripper 1920X with an Vega64... and i can not see any bug ????

what kind of bug do you mean ? "dozens of critical bugs" well would this not hit my daily useage ?

yes i know vega64 works good because it is old right now 5700xT works well to because it is old.

the only problem we have right now is the RX 6800(XT) because it is very new.

[bridgman]

the only problem we have right now is the RX 6800(XT) because it is very new.

Have you seen any bugs with the 6800/XT yet other than ROCm support for Navi in general ? I'm sure they must be some bugs but haven't seen any reported yet.

There is an issue with AMDVLK and Proton games but that seems to be related to a regression in Proton 5.13-2 and affects all GPUs using a driver in /opt, not just the 6800(XT).

[qarium]

Have you seen any bugs with the 6800/XT yet other than ROCm support for Navi in general ? I'm sure they must be some bugs but haven't seen any reported yet.
There is an issue with AMDVLK and Proton games but that seems to be related to a regression in Proton 5.13-2 and affects all GPUs using a driver in /opt, not just the 6800(XT).

i was more pointing out that most stable linux distro releases do not have 6800 support yet...
stable debian? even debian testing?...
the only distro who has 6800 support is Fedora33 but not on the orginal installer medium.
this means you can use 6800 if you use fedora33 with the newest updates with any other linux distro you will not have so much luck.

i do not have a 6800 yet because: there is none if i search on geizhals.de there is none.
only mindfactory.de have some somethimes but it is out of stock and totally overpriced.
this means it needs 1-2 month to find some in stock on low prices.

this means in 2-3 month i buy a 6800xt but then i am sure all bugs are already fixed and i can not help with bug reports.

[Citan]

At this rate, I expect to see Firefox catch up to Chrome's performance in about 300 years.

You mean you expect Chrome performance to catch up to Firefox before end 2030 right?
Because, let me tell you, Chrome's superior speed is just a facade.

Sure, if you compulse a website heavy with HTML structure and/or JS, and you have less than 5 tabs, and you have no extensions... Chrome will indeed be (possibly much) faster.

Who still nowadays, however, has such a tightened use?
Apart from people who are still uncomfortable around the numeric tools (ok, admitedly not a small part of all population, probably XD), everyone now hoards between 10 and several dozens of tabs at any given time, because on top of the 6-7 websites one always visits every day, work/personal hobbies/curiosity make you look through several searches, each leading itself to many tabs.

There are also the extensions: last time I checked (a few months ago), Chrome didn't have yet any proper way to handle nested, vertical tabs, which are a huge productivity improvement (at least for 90% of the few dozen people I demonstrated it to), and following on that Firefox has many extensions that provide customization to ergonomy, bookmark management or the like...
It still works flawless under pressure.

Concret example: my Chrome starts getting awry after 40+ tabs, Firefox holds up to 600 without any trouble. Unless you try to reload all of them at the same time XD.
On my everyday usage, I constently turn around 150 tabs of topics I have to research and analyse, or dozens of documentation page that have a decent chance to be useful in the day.
Only Firefox and its extensions allow me an efficient way to "industrialize" information research and grouping without slowing down computer / hogging resources.
And while I admit I'm very peculiar in my use of browsers, most people I know have an average number of opened tabs of ~15-20 any given day, with a mix of music player, news website, webmail (the "sticky ones" and otherwise "regular" site (the "information browing").
I'd really like to get automated comparison reproducing this kind of actual real-life use.

[birdie]

I use Fedora33 on a Threadripper 1920X with an Vega64... and i can not see any bug ????

what kind of bug do you mean ? "dozens of critical bugs" well would this not hit my daily useage ?

yes i know vega64 works good because it is old right now 5700xT works well to because it is old.

the only problem we have right now is the RX 6800(XT) because it is very new.

Anecdotal evidence/a sample of one. Yeah, it's always "worked" like that. And people who are having issues are hallucinating according to you - you happen to own 150 different GPUs and 15000 different hardware configurations, right.

[Citan]

Yeap, all true and like you i run everything stock and my RAM are factory 2666.

But as i said someone from mozilla is mailing me and it seems my crashes are even nastier cuz the crash are not even registering.

The weird thing is whatever it is, only affects Firefox and Firefox only but well eventually we should find out what the mystery is, for now Brave does the work perfectly and finally migrated most of my data to it, so my workflow is back to normal so later on i can go harder at debugging FF.

Following the whole "subthread", I have to say, congrats to you for sticking up with that mindset.
After so many tries and investigations most people would have dropped the ball and moved away.

I really hope you and Mozilla's interlocutor can find the cause and resolve it without it taking too much of your respective time.
And thank you both for staying in that fight like a bulldog never lets its prey go.

[qarium]

Anecdotal evidence/a sample of one. Yeah, it's always "worked" like that. And people who are having issues are hallucinating according to you - you happen to own 150 different GPUs and 15000 different hardware configurations, right.

Well you maybe think you are "smart" but in the end the one who as a working system is "smart"

I use Firefox 83 64bit with fedora33 on AMD-TR-1920X+ vega64 with wayland and gnome3.

you try to be smart with 150 different GPUs and 15000 different hardware configurations...
but hey why not just try some sane common configuration like i have ?

"And people who are having issues are hallucinating according to you"

wrong accoding to me they do not hallucinating but in fact you are the one who weaponize it.
you do not look for a solution with AMD hardware because on linux Nvidia IS NOT a solution.
you are only here to defend a very very very evil company: Nvidia...

and everyone here can see this. everyone can see you do not search for a solution!!

so i give you a good advice: admit defeat of your Nvidia nightmare dream and come back if you are ready to work on the solution to defeat evil companies like Nvidia.

before this point happens your writing in this forum is complete pointless.