problems with virtio and microvm in ACPI mode

[Mark Harmstone]

Hi all,

I've encountered a really weird bug when trying to run the fstest 
generic/476 in a loop on a microvm VM. The block device is a
virtio-blk-device, formatted as btrfs.

With acpi=on, it eventually grinds to a halt, and prints a hang message. 
I tracked the problem down to be within wait_dev_supers, i.e. it's 
apparently issuing a write for the superblock that never returns. The 
strange thing is that with acpi=off everything works as it should.

acpi=off adds this to the cmdline: virtio_mmio.device=512@0xfeb00e00:12

acpi=on adds this to the DSDT:

Device (VR07)
{
     Name (_HID, "LNRO0005")  // _HID: Hardware ID
     Name (_UID, 0x07)  // _UID: Unique ID
     Name (_CCA, One)  // _CCA: Cache Coherency Attribute
     Name (_CRS, ResourceTemplate ()  // _CRS: Current Resource Settings
     {
         Memory32Fixed (ReadWrite,
             0xFEB00E00,         // Address Base
             0x00000200,         // Address Length
             )
         Interrupt (ResourceConsumer, Level, ActiveHigh, Exclusive, ,, )
         {
             0x0000000C,
         }
     })
}

...and to my untrained eye it's not obvious if there's something wrong 
there.

Does this ring a bell to anyone? Is the bug in Qemu, the kernel, or am I
inadvertently doing something stupid? I'm happy to help someone 
reproduce this.

Thanks

Mark