On Jun 27, 2022 / 23:18, Christoph Hellwig wrote: > On Tue, Jun 28, 2022 at 08:12:51AM +0200, Greg Kroah-Hartman wrote: > > Ah, so it's a fake PCI device, or is it a real one? > > Whatever the blktests configuration points to. But even if it is > "fake" that fake would come from the hypervisor. FYI, the WARN was observed with real PCI NVMe device also. I observed the WARN with v5.19-rc1 and still observe with v5.19-rc5. I'm not sure which version introduced the WARN. I once reported this failure, and shared with linux-pci list. https://lore.kernel.org/linux-block/20220614010907.bvbrgbz7nnvpnw5w@shindev/ -- Shin'ichiro Kawasaki
