Linux 6.9-rc7 Released: The Kernel Is Looking Good

[phoronix]

Phoronix: Linux 6.9-rc7 Released: The Kernel Is Looking Good

Linus Torvalds just released Linux 6.9-rc7 and it's looking good like the Linux 6.9 stable kernel will hopefully be out next Sunday...

https://www.phoronix.com/news/Linux-6.9-rc7-Released

[xpris]

I hope they can finally fix bug where many heavy games crash at launch or just after launch if there is no resize bar/above 4g decoding support or it is disabled. Bug introduced in kernel 6.9 rc5 up to rc7 and even backported to stable 6.8.9 and LTS branch with 6.6.30. Upstream know about this bug but so far without any resolution. There is bug entry in freedesktop drm and few posts on Linux reddit about this bug and multiple on valve GitHub where people reporting same things.
Example: https://github.com/ValveSoftware/csg...ux/issues/3728
Example2: https://github.com/ValveSoftware/Dota-2/issues/2694
BUG report: https://gitlab.freedesktop.org/drm/amd/-/issues/3343

[luno]

are gamescope changes for NVK in this release ?

[ojsl1]

I hope they can finally fix bug where many heavy games crash at launch or just after launch if there is no resize bar/above 4g decoding support or it is disabled. Bug introduced in kernel 6.9 rc5 up to rc7 and even backported to stable 6.8.9 and LTS branch with 6.6.30. Upstream know about this bug but so far without any resolution. There is bug entry in freedesktop drm and few posts on Linux reddit about this bug and multiple on valve GitHub where people reporting same things.
Example: https://github.com/ValveSoftware/csg...ux/issues/3728
Example2: https://github.com/ValveSoftware/Dota-2/issues/2694
BUG report: https://gitlab.freedesktop.org/drm/amd/-/issues/3343

Oh so that's what it was! I almost lost my mind combing through wine traces for clues on that without any results, thanks for the links

[pWe00Iri3e7Z9lHOX2Qx]

Oh so that's what it was! I almost lost my mind combing through wine traces for clues on that without any results, thanks for the links

Gotta love it when shit gets backported to the "stable" LTS kernels and breaks stuff...

[scottishduck]

are gamescope changes for NVK in this release ?

The kernel side of nvk is just nouveau modesetting. You’re most likely going to be waiting for Mesa 24.2

[ffs_]

I hope they can finally fix bug where many heavy games crash at launch or just after launch if there is no resize bar/above 4g decoding support or it is disabled. Bug introduced in kernel 6.9 rc5 up to rc7 and even backported to stable 6.8.9 and LTS branch with 6.6.30. Upstream know about this bug but so far without any resolution. There is bug entry in freedesktop drm and few posts on Linux reddit about this bug and multiple on valve GitHub where people reporting same things.
Example: https://github.com/ValveSoftware/csg...ux/issues/3728
Example2: https://github.com/ValveSoftware/Dota-2/issues/2694
BUG report: https://gitlab.freedesktop.org/drm/amd/-/issues/3343

It seems only AMD GPUs affected. CS2 runs fine at 6.9-rc6 with NV GPU.

[darkbasic]

Gotta love it when shit gets backported to the "stable" LTS kernels and breaks stuff...

The backport itself isn't the problem. The problem is that it doesn't get IMMEDIATELY reverted once reports hit bug trackers.

[EphemeralEft]

The backport itself isn't the problem. The problem is that it doesn't get IMMEDIATELY reverted once reports hit bug trackers.

Even if major issues were always fixed the next day, I still think too many are being added to LTS versions. I know it’s impossible to test all the possible combinations of hardware and software, but there has to be a better way.

Moving kernel-space features to user-space would probably help a lot. For example, why does Linux need its own TLS implementation or SMB client?

[kylew77]

Moving kernel-space features to user-space would probably help a lot. For example, why does Linux need its own TLS implementation or SMB client?

Agree and disagree with you. Agree because it would make things faster to debug, but disagree because the reason that kind of stuff is in kernel is because of speed reasons. There is a performance hit when transitioning from user space to kernel space and it has just gotten worse and worse on AMD64 due to all the cpu bugs. One could argue that modern chips are fast enough and just incur that penalty but I don't know how big a hit it would be. 1 or 2% isn't that bad, but if moving it to userspace would make things say 20 to 50% slower people will be mad.