Intel's X.Org Driver Still Being Developed, No Sign Of 3.0 Released

[phoronix]

Phoronix: Intel's X.Org Driver Still Being Developed, No Sign Of 3.0 Released

While it has been three years since the last stable xf86-video-intel DDX driver release, almost two years since the last development release, and distributions/users beginning to get fed up by this DDX driver release management, Intel remains mum on the manner...

http://www.phoronix.com/scan.php?page=news_item&px=Intel-DDX-August-2016

[chuckula]

It's kind of ironic that the major improvements to OpenGL compatibility that we've seen from Intel mean that the generic GLAMOR interfaces can replace this driver in many scenarios.

[starshipeleven]

It's kind of ironic that the major improvements to OpenGL compatibility that we've seen from Intel mean that the generic GLAMOR interfaces can replace this driver in many scenarios.

Does the modesetting driver need OpenGL 4.x to do 2D acceleration? I think not.
imho that's unrelated, and probably also bad planning on Intel's part. If they focused on migrating to modesetting instead of leaving a lone guy pulling the whole driver, it would have been much better.

[andre30correia]

looks like the new intel chips will not have a 2d chip, like the amd ones, but they should put some stable snapshots sometimes like theu do with beignet

[milkylainen]

There is a reason for the driver not to be called a stable release. Because nothing about xf86-video-intel has been stable.
Intel still does not give a serious impression when talking drivers for their graphics hardware even though they had 20+ years trying to get things right.
I have lost hope in all Intel GFX driver development a long time ago. Still sticking to the only reasonable alternative from my point of view.
A non hybrid hardware running proprietary Nvidia drivers (on x86). This solution has given me the least amount of headache over the years.
From time to time I go back to trying to run an Intel GFX solution only to discover it is totally broken one way or another.
Last time was a hybrid Intel/Nvidia solution on Broadwell. Boy was that a bunch of crap.

[smitty3268]

This is all about backroom politics at intel.

Some of their devs think SNA is the way to go and that they need to get every bit of performance out of the hardware. Others think it's a giant mess of bugs and going to be unmaintainable, and they've managed to block the official release of 3.0 for years. Unfortunately, they haven't really provided a great alternative. Hopefully modesetting is finally in good enough shape.

[DanL]

It's kind of ironic that the major improvements to OpenGL compatibility that we've seen from Intel mean that the generic GLAMOR interfaces can replace this driver in many scenarios.

I think it has more to do with the fact that modern GPU's are moving away from dedicated 2D acceleration hardware and that Intel lacks the dev manpower to make dedicated 2D accel a priority

[axfelix]

SNA and Ivy/Haswell has actually worked really well for me with DRI3 for a long time now -- fewer multi-monitor heat/throttling issues than Nvidia proprietary, even -- but I think that's probably one of the only hardware combinations it behaves well under, and if it's never going to be the default, then it's probably not worth supporting.

[andre30correia]

There is a reason for the driver not to be called a stable release. Because nothing about xf86-video-intel has been stable.
Intel still does not give a serious impression when talking drivers for their graphics hardware even though they had 20+ years trying to get things right.
I have lost hope in all Intel GFX driver development a long time ago. Still sticking to the only reasonable alternative from my point of view.
A non hybrid hardware running proprietary Nvidia drivers (on x86). This solution has given me the least amount of headache over the years.
From time to time I go back to trying to run an Intel GFX solution only to discover it is totally broken one way or another.
Last time was a hybrid Intel/Nvidia solution on Broadwell. Boy was that a bunch of crap.

i have a optimus laptop for years and i never had such problems, i use bumblebee and nvidia prime (because vmware workstation only) the intel drivers working fine skylake here, the last bug i check was some flicker in chromium, it was a kernel driver bug who was resolve in one week

[microcode]

There is a reason for the driver not to be called a stable release. Because nothing about xf86-video-intel has been stable.
Intel still does not give a serious impression when talking drivers for their graphics hardware even though they had 20+ years trying to get things right.
I have lost hope in all Intel GFX driver development a long time ago. Still sticking to the only reasonable alternative from my point of view.
A non hybrid hardware running proprietary Nvidia drivers (on x86). This solution has given me the least amount of headache over the years.
From time to time I go back to trying to run an Intel GFX solution only to discover it is totally broken one way or another.
Last time was a hybrid Intel/Nvidia solution on Broadwell. Boy was that a bunch of crap.

I have had absolutely zero trouble with SNA and xf86-video-intel for the last five years on SNB, IVB, and HSW. I don't have a BDW machine running X11, so maybe it works worse there. The SKL stability issues with it seem to have nothing to do with the X.Org driver, and everything to do with the kernel and firmware.