Ubuntu 22.04 LTS Changes Default For NVIDIA Driver Back To Using X.Org Rather Than Wayland

[xhustler]

Being a "customer" is an entitlement class.

Unless you have $800 dollar PCI-E paperweights........ ?

A video card isn't just a piece of hardware. It needs a driver to work. We are Linux users, and we expect the card to work when we plug it in. And it isn't a blind expectation that we made up because it feels good. Nvidia promises working Linux support. Their problem is that the Linux world has left X.org behind and all we have left now is Wayland. X.org has been declared abandonware by all of its former developers. Those distros that have X.org, yes, it's still fully functional but it's unmaintained abandonware.

There are no x.org developers left for X11.

Then use their product the way the manufacturer intended - with X11.

At the moment, expecting Wayland to work flawlessly with Nvidia hardware is akin to you driving your expensive luxury sedan out in the desert and then complaining that it's not giving you the mileage you expect.

[Developer12]

Then use their product the way the manufacturer intended - with X11.

At the moment, expecting Wayland to work flawlessly with Nvidia hardware is akin to you driving your expensive luxury sedan out in the desert and then complaining that it's not giving you the mileage you expect.

BigOilCorp would really prefer you drive your car on gasoline and not think about global warming so much.

[Slartifartblast]

Being a "customer" is an entitlement class.

Unless you have $800 dollar PCI-E paperweights........ ?

A video card isn't just a piece of hardware. It needs a driver to work. We are Linux users, and we expect the card to work when we plug it in. And it isn't a blind expectation that we made up because it feels good. Nvidia promises working Linux support. Their problem is that the Linux world has left X.org behind and all we have left now is Wayland. X.org has been declared abandonware by all of its former developers. Those distros that have X.org, yes, it's still fully functional but it's unmaintained abandonware.

There are no x.org developers left for X11.

Really, funny how my X11 on Leap keeps getting updated ?

xorg-x11-server - X

Fri 18 Mar 2022 12:00:00 GMT
[email protected]
- U_glamor-Make-pixmap-exportable-from-gbm_bo_from_pixma.patch
* avoid consequently failing page flip (boo#1197269)
Sun 13 Mar 2022 12:00:00 GMT
[email protected]
- u_sync-pci-ids-with-Mesa-20.2.4.patch
* sync pci ids with Mesa 20.2.4 (related to boo#1197046)
Fri 11 Mar 2022 12:00:00 GMT
[email protected]
- U_0002-DRI2-Add-another-Coffeelake-PCI-ID.patch
U_0003-dri2-Sync-i965_pci_ids.h-from-mesa.patch
U_0004-dri2-Set-fallback-driver-names-for-Intel-and-AMD-chi.patch
U_0005-dri2-Sync-i965_pci_ids.h-from-mesa-iris_pci_ids.h.patch
* sync GL driver PCI IDs with Mesa (boo#1197045)
Fri 04 Mar 2022 12:00:00 GMT
[email protected]
- U_xfree86-Fix-NULL-pointer-dereference-crash.patch
* Fix a regression in
u_xfree86-Change-displays-array-to-pointers-array-to-f.patch
(boo#1196577)
* Credits go to Simon Lees ([email protected]) for finding the fix!
- renamed u_xfree86-Change-displays-array-to-pointers-array-to-f.patch
to U_xfree86-Change-displays-array-to-pointers-array-to-f.patch
since it's a backport from an upstream patch
Mon 07 Feb 2022 12:00:00 GMT
[email protected]
- u_xfree86-Change-displays-array-to-pointers-array-to-f.patch
Fix segmentation fault during terminal switches with multiple attached
displays (bsc#1188970)
Tue 14 Dec 2021 12:00:00 GMT
[email protected]
- U_xfixes-Fix-out-of-bounds-access-in-ProcXFixesCreateP.patch
* CVE-2021-4009/ZDI-CAN-14950 (bsc#1190487)
The handler for the CreatePointerBarrier request of the XFixes
extension does not properly validate the request length leading
to out of bounds memory write.
- U_Xext-Fix-out-of-bounds-access-in-SProcScreenSaverSus.patch
* CVE-2021-4010/ZDI-CAN-14951 (bsc#1190488)
The handler for the Suspend request of the Screen Saver extension
does not properly validate the request length leading to out of
bounds memory write.
- U_record-Fix-out-of-bounds-access-in-SwapCreateRegiste.patch
* CVE-2021-4011/ZDI-CAN-14952 (bsc#1190489)
The handlers for the RecordCreateContext and RecordRegisterClients
requests of the Record extension do not properly validate the request
length leading to out of bounds memory write.

Methinks you are parroting this

https://www.phoronix.com/scan.php?page=news_item&px=XServer-Abandonware

Only to have eaten humble pie here....

https://www.phoronix.com/scan.php?page=news_item&px=X.Org-Server-21.1

Reports of my death are greatly exaggerated.

Mark Twain

[wertigon]

Then use their product the way the manufacturer intended - with X11.

At the moment, expecting Wayland to work flawlessly with Nvidia hardware is akin to you driving your expensive luxury sedan out in the desert and then complaining that it's not giving you the mileage you expect.

Nope. Nvidia is desperately trying to compete with a petrol car in a country where everyone wants to buy EVs. You know, like GM tries to do in Norway.

[skeevy420]

Nope. Nvidia is desperately trying to compete with a petrol car in a country where everyone wants to buy EVs. You know, like GM tries to do in Norway.

I don't give a shit one way or the other about NVIDIA and Wayland, but buying an EV is rather moot if the world doesn't shift from fossil fuel power plants. You don't lower any pollution levels since the coal plants have to burn more coal so you can drive your EV. Currently, all buying an EV does is move your pollution from the car's exhaust to the coal plant's exhaust.

FWIW, I blame a lot of it on America for sabotaging green and nuclear power as well as not creating regulations like mandatory solar on new developments and mandatory wind turbines on whirly birds...the metal spinny things on roofs...that's free wind and thermal energy I see spinning on the top of every house.

There could also be whirly bird style turbines installed on highways. Anytime big trucks drive by they can spin the whirly birds with their wind wake.

[ssokolow]

I don't give a shit one way or the other about NVIDIA and Wayland, but buying an EV is rather moot if the world doesn't shift from fossil fuel power plants. You don't lower any pollution levels since the coal plants have to burn more coal so you can drive your EV. Currently, all buying an EV does is move your pollution from the car's exhaust to the coal plant's exhaust.

Not necessarily. I don't know about gasoline vs. coal but, when you're comparing apples to apples, big power plants are generally more energy efficient because it's easier to design more fuel-efficiency into a big, stationary system than a small, mobile one.

(Aside from natural gas furnaces and water heaters, which are very high-efficiency because the "waste" heat contributes to the goal if you pipe it to the premises and burn it on-site... thus why it's taken so long for heat pumps to become competitive for heating in North America despite having the ability to have greater than 100% efficiency when calculated as "how many watts of heat do I get out per watt of electricity I put in?")

[piotrj3]

I don't think there's a 'fix' possible in their current framework. I just don't think there's a way for NVIDIA to have their cake and eat it too with this. They're always going to be weird and out-of-date with this strategy. The only good solution I see for them is to have at least one viable option working on the tried-and-true path that AMD and Intel are on.

Funny thing is that all they'd need to do is put out an open-source Vulkan MESA driver with no implied support. Zink will get users OpenGL on it.

Out-of-date is design here of Wayland. And that design makes Intel/AMD engineers very often scratch their head to implement Vulkan properly and Nvidia who doesn't have issue implementing Vulkan, scratches their head how to best implement all implicit sync nuances without totally ruining their performance.

At this point, Metal, Vulkan, DX12, WDDM, Android all moved to explicit sync. Android developers even called it "best decision they ever made", while everyone has issues with Wayland and buffer management around it. And thing is you can implement implicit sync over explicit sync API without issues, but you cannot implement explicit sync over implicit sync correctly.

[JMB9]

Mah! Maybe in some use cases, but on a normal notebook with AMD graphics to me the latest version of Plasma works flawlessly, never crashed, and does not lack any functionality, indeed there are features that are not available on Xorg I am. For example, the gestures of Gnome and those arriving in Plasma.
Security ... a script is enough to compromise Xorg privacy and security, arguing that this is not a problem now that the problem is known is absurd.

I am happy that you have a good experience with Wayland ... but there were still some functionality missing - and after revoking Wayland as default with Nvidia and SDL2
revoking Wayland by default, I don't think anybody can really say that Wayland is stable and has all X.org functionality as a subset.
There are still users complaining that X.org is still use as otherwise more development will go to Wayland ... and more bugs will show up.
But it is the other way round: on professional systems a change in a component is done when everything is better and proven so.
When KDE devs will say that the experience with KDE is better under Wayland than under X.org I will start thinking about a change - not before.
So we will see what is in KDE neon 'user edition' - and after my experience this will work - as those devs do care and really work with their system.

As I have hardened servers I know that X.org is a target - but I was also responsible for workstation and I never saw any problems with stability or security in real life.
And I am using X.org since 1994 on my Linux installations.
A script is something I (or an admin) puts on the system - so there is no security problem. You don't run things from Internet resources you do not know, right?
Otherwise there is no security at all. Root must know what she/he does - otherwise hell breaks lose. That's a professional system. And it works as proven by reality!
From my point of view X.org as extremely insecure which eats your date and put every personal bit you have on social media is just witch hunting.
If the share of Wayland would be well above 50%, we may see if this not so new kid on the block keeps its promises ... maybe, but we will see ...
And still using XWayland ... is this code bullet proof?

[birdie]

Try again. That's among gamers, a *tiny* percentage of linux users.

The real breakdown?
intel: 67% (iGPUs for the win)
nvidia: 18%
amd: 15%

I don't freaking care about built-in GPUs. If we are talking about real GPUs NVIDIA is king by a huge wide margin, period.

[skeevy420]

Not necessarily. I don't know about gasoline vs. coal but, when you're comparing apples to apples, big power plants are generally more energy efficient because it's easier to design more fuel-efficiency into a big, stationary system than a small, mobile one.

(Aside from natural gas furnaces and water heaters, which are very high-efficiency because the "waste" heat contributes to the goal if you pipe it to the premises and burn it on-site... thus why it's taken so long for heat pumps to become competitive for heating in North America despite having the ability to have greater than 100% efficiency when calculated as "how many watts of heat do I get out per watt of electricity I put in?")

On the environmental side, the polluting properties of coal—starting with mining and lasting long after burning—and the large amounts of energy required to liquefy it mean that liquid coal produces more than twice the global warming emissions as regular gasoline and almost double those of ordinary diesel.

Pounds of CO₂ emitted per million British thermal units (Btu) of energy for various fuels

Coal (anthracite)	228.60
Coal (bituminous)	205.40
Coal (lignite)	216.24
Coal (subbituminous)	214.13
Diesel fuel and heating oil	163.45
Gasoline (without ethanol)	155.77
Propane	138.63
Natural gas	116.65

Basically, using gasoline and internal combustion engines is more environmental friendly than EVs and coal plants. And since it's basically established science that gasoline and internal combustion engines are bad for the environment, replacing them with EVs powered by coal that's 50-200% more dirty no matter how you fudge the numbers is clearly not the correct solution.

Using green vehicles without creating green energy first is like putting a fresh coat of paint over black mold spores. You didn't fix anything; you just hid the problem from view.