Thanks for the confirmation KY. And sorry for the HTML email, fixed. Best regards, Alex Ionescu On Tue, Sep 19, 2023 at 1:10 PM KY Srinivasan <kys@xxxxxxxxxxxxx> wrote: > > Linux support for Hyper-V has been completely upstreamed and our goal is to upstream all new development we do in this space. We welcome contributions to extend/improve what we currently have and will be happy to review contributions. > > > > Regards, > > > > K. Y > > > > Sent from Outlook > > From: Alex Ionescu <aionescu@xxxxxxxxx> > Sent: Tuesday, September 19, 2023 9:59 AM > To: linux-hyperv@xxxxxxxxxxxxxxx > Cc: KY Srinivasan <kys@xxxxxxxxxxxxx>; Dexuan Cui <decui@xxxxxxxxxxxxx>; Haiyang Zhang <haiyangz@xxxxxxxxxxxxx>; Michael Kelley (LINUX) <mikelley@xxxxxxxxxxxxx>; Saurabh Sengar <ssengar@xxxxxxxxxxxxxxxxxxx> > Subject: [EXTERNAL] Question regarding patches/discussions > > > > You don't often get email from aionescu@xxxxxxxxx. Learn why this is important > > Dear linux-hyperv Maintainers, > > > > I understand that the main goal of this submodule/project is to provide a first-class top-tier experience to delight both developers and users when running Linux hypervisor on Windows systems by Hyper-V (both in traditional VM scenarios, as well as in Docker and WSL2 scenarios), as well as to provide a number of strategic capabilities for Azure-related workloads, both internal to Microsoft and external (including VTL2/HCL, TDX/SNP Isolation, Linux-as-Root-Partition, etc) -- some of which are not expected to be used outside of specific Azure scenarios within Microsoft. > > > > That being said, there are some potential additional ideas and projects that could help improve the experience, stability, security, etc., of Linux-under-HyperV, which are not in Microsoft's current or future plans. Are you able and willing to review patches (not necessarily promise to approve) for linux-hyperv as other linux kernel maintainers normally would, or do you consider this a "closed" project for Microsoft purposes only? > > > > For example, as part of a project at UC Berkeley, I would have some small extensibility patches that I would like to try upstreaming, and some small refactoring (for example, new definitions in the GDK headers were not made available in the TLFS headers -- even stable ones, or some new hypercall support code was added, but only to the mshv driver/module, and not in the kernel itself, etc...). As you've pushed more new code into the new mshv module for VTL2/launching VMs, this prevents (or requires code duplication) lighting up some of those features for other kernel-level functionality that I'd like to leverage for hardening/etc. I would also potentially like to investigate access to hypercalls/hyperv earlier than today (closer to Windows' model at UEFI boot). > > > > But before going too deep into those discussions/proposals, I wanted to see if such room exists to begin with (and of course assuming it doesn't conflict/harm Microsoft's goals/plans). > > > > Thank you! > > > > > > Best regards, > Alex Ionescu
