New Features On The Horizon For The Linux Kernel
To recap some of the active kernel development patches that have been floating around and could make Linux 3.19 or another near-future kernel release include:
- KDBUS! Upstream developers felt KDBUS would be ready this year and they're darn close. Since the end of October, KDBUS has been under review on the kernel mailing list. We could see KDBUS for Linux 3.19 otherwise Linux 3.20 would be quite reasonable to assume.
- Kernel live patching... There's kGraft, Kpatch, and then the other approach that tries to take the best of both worlds. These technologies are about applying live kernel patches to a running system for maintenance/security purposes without having to reboot the system for environments where minimal downtime is a must.
- While it's not coming in Linux 3.19, there's an Intel patch to significantly boost the HD Graphics performance. This patch might first appear in the SteamOS kernel as there's some bad interaction issues for now with video acceleration, but we'll see soon enough.
- Nouveau perf counter work that's currently slated for Linux 3.20 to build a NVPerfKit-like solution for Linux.
- In the Nouveau pull request for Linux 3.19, Ben Skeggs of Red Hat indicated he's working on a couple of bigger projects for this open-source NVIDIA Linux driver. We don't know what all that entails, but hopefully re-clocking improvements. There's also GeForce GTX 900 acceleration support to come once NVIDIA releases the new signed firmware/microcode files -- one could hope by Linux 3.20.
- The new AMDGPU kernel driver to be used both by Radeon Gallium3D and the closed-source Catalyst user-space blobs still has yet to see the light of day. Patches have yet to be made public and it won't be mainlined then until at least Linux 3.20. The new AMDGPU kernel driver is crucial to AMD's new unified Linux driver strategy.
- Linux kernel optimizations for SSHDs at long last.
- TPM 2.0 support for the kernel if you care about trusted computing.
- Continued changes that make Clang incredibly close to building the mainline Linux kernel. With each kernel release, the patch-set needed by LLVMLinux to build a working kernel is lower and soon enough we'll hopefully be able to use Clang to compile a vanilla kernel as an alternative to GCC.
What else are you hoping to see materialize within the mainline Linux kernel in 2015? Let us know in the forums. Maybe in 2015 we'll see fresh starts out of Tux3 and Reiser 4? It's also important to remember that the Linux 4.0 kernel is probably coming in 2015.