From: Hamza Mahfooz <[email protected]> Sent: Friday, November 13, 2026 1:59 PM > > I was doing a deep dive into the host spec and noticed some gaps. So, > I went ahead and hacked together the patches in this series for items > that seemed worthwhile to address.
Are you (or someone on the Microsoft side) testing these changes in a Generation 1 VM, as well as on back revs of the Hyper-V host? Testing should go back to Windows Server 2016/Windows 10 for both Gen1 and Gen2. Same questions apply to your earlier cursor plane support. I just did basic smoke testing of the cursor plane support on both Gen1 and Gen2 VMs using vmconnect.exe on an up-to-date Windows 11 host -- no obvious problems. But I don't have access to older hosts like WS2016/Win10 to do the back rev testing. There are multiple frame buffer differences in Gen1 and Gen2 VMs, but most (or all?) are in the lowest layer in how the frame buffer is allocated and mapped. Hopefully these higher-level changes you are making are agnostic of the VM generation, but thoroughness would dictate testing both. I'm less sure about what might have changed in current Hyper-V versions and WS2016 from 10 years ago. My general impression is that very little, if anything, has changed in the Hyper-V frame buffer support, but again it would be better to not get a bad surprise after the code has gone in. Michael > > Hamza Mahfooz (2): > drm/hyperv: add modes for all resolutions offered by the host > drm/hyperv: batch dirty rectangles > > drivers/gpu/drm/hyperv/hyperv_drm.h | 15 ++++ > drivers/gpu/drm/hyperv/hyperv_drm_drv.c | 4 ++ > drivers/gpu/drm/hyperv/hyperv_drm_modeset.c | 32 ++++++++- > drivers/gpu/drm/hyperv/hyperv_drm_proto.c | 79 ++++++++++++++++----- > 4 files changed, 110 insertions(+), 20 deletions(-) > > -- > 2.56.0
