From: Saurabh Singh Sengar <ssengar@xxxxxxxxxxxxxxxxxxx> Sent: Tuesday, November 5, 2024 9:15 AM > > On Sun, Nov 03, 2024 at 05:17:19AM +0000, Michael Kelley wrote: > > I was running the linux-next-20241101 kernel in a Hyper-V guest VM today, > > and the Hyper-V guest HID (mouse) driver is failing to probe during boot. > > I bisected the problem to commit 8b7fd6a15f8c: > > > > HID: bpf: move HID-BPF report descriptor fixup earlier > > [snip] > > > > I'll do some additional debugging to try to narrow down what's going > > wrong. I'm experienced with Linux guests on Hyper-V, but don't have > > any existing knowledge of the hid-hyperv.c driver or how it interacts with > > the hid-core.c code. Nonetheless, I should be able to peel the onion a > > bit more. > > > > Benjamin -- if you have thoughts on what to look for, I would appreciate > > any pointers that could save me some time. It's certainly possible that > > there's a latent bug in how the Hyper-V HID driver interacts with the > > HID core, so I'll look at both sides of the interaction. > > > > Thanks, > > > > Michael Kelley > > Vitaly patch seems to fix this issue: > https://lore.kernel.org/linux-hyperv/20241105171141.GA13863@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx/ > Indeed it does! So commit 8b7fd6a15f8c is good, and the root cause is that probing in the Hyper-V mouse HID driver is structured incorrectly. Somehow the driver has worked the past 12 years, until commit 8b7fd6a15f8c came along. But the fix should be in the Hyper-V mouse driver. I looked to see if any other HID drivers have the same problem as the Hyper-V mouse HID driver, and I don't think there are any. So this is truly a one-off issue. This thread can be considered closed. We'll provide the resolution via the above patch from Vitaly Kuznetsov. Michael Kelley
