Linus Torvalds Still Rejects KVM Tool From Kernel
Linus Torvalds has reaffirmed that at this point he doesn't intend to pull KVMTool into the mainline Linux kernel.
KVMTool is the lightweight QEMU-free native KVM tool. KVMTool has been developed by several open-source developers for nearly two years.
This tool for toying with KVM virtualization supports many features like SMP, various virtio features, QCOW2 read-write support, etc. The developers behind the tool have wanted to merge it into the mainline Linux kernel as a separate tool/utility (similar to Intel's turbostat and various other small user-space utilities living within the kernel tree) since this tool can stress the Kernel-based Virtual Machine components without QEMU.
KVMTool has been living within the "linux-next" tree, but hasn't been accepted by Linus Torvalds into his mainline tree, including for the latest Linux 3.7 merge window.
Torvalds has now explained his reasoning for not accepting this pull: "I have yet to see a compelling argument for merging it. It's tons of code, it doesnt match the original "small simple" model, and I think it would be better off as a separate project."
So now if KVMTool is to survive it will likely just need to be maintained as its own separate project.
KVMTool is the lightweight QEMU-free native KVM tool. KVMTool has been developed by several open-source developers for nearly two years.
This tool for toying with KVM virtualization supports many features like SMP, various virtio features, QCOW2 read-write support, etc. The developers behind the tool have wanted to merge it into the mainline Linux kernel as a separate tool/utility (similar to Intel's turbostat and various other small user-space utilities living within the kernel tree) since this tool can stress the Kernel-based Virtual Machine components without QEMU.
KVMTool has been living within the "linux-next" tree, but hasn't been accepted by Linus Torvalds into his mainline tree, including for the latest Linux 3.7 merge window.
Torvalds has now explained his reasoning for not accepting this pull: "I have yet to see a compelling argument for merging it. It's tons of code, it doesnt match the original "small simple" model, and I think it would be better off as a separate project."
So now if KVMTool is to survive it will likely just need to be maintained as its own separate project.
9 Comments