commit d05ae948cc887054495977855b0859d0d4ab2613
Author: Nir Soffer <nsoffer@redhat.com>
Date:   Fri Jun 28 23:20:58 2024 +0300

    Consider discard option when writing zeros
    
    When opening an image with discard=off, we punch hole in the image when
    writing zeroes, making the image sparse. This breaks users that want to
    ensure that writes cannot fail with ENOSPACE by using fully allocated
    images[1].
    
    bdrv_co_pwrite_zeroes() correctly disables BDRV_REQ_MAY_UNMAP if we
    opened the child without discard=unmap or discard=on. But we don't go
    through this function when accessing the top node. Move the check down
    to bdrv_co_do_pwrite_zeroes() which seems to be used in all code paths.
    
    This change implements the documented behavior, punching holes only when
    opening the image with discard=on or discard=unmap. This may not be the
    best default but can improve it later.
    
    The test depends on a file system supporting discard, deallocating the
    entire file when punching hole with the length of the entire file.
    Tested with xfs, ext4, and tmpfs.
    
    [1] https://lists.nongnu.org/archive/html/qemu-discuss/2024-06/msg00003.html
    
    Signed-off-by: Nir Soffer <nsoffer@redhat.com>
    Message-id: 20240628202058.1964986-3-nsoffer@redhat.com
    Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>

-blockdev '{"driver":"file","filename":"/var/lib/libvirt/images/raw.img","node-name":"libvirt-1-storage","read-only":false}'
-device '{"driver":"scsi-hd","bus":"scsi0.0","channel":0,"scsi-id":0,"lun":0,"device_id":"drive-scsi0-0-0-0","drive":"libvirt-1-storage","id":"scsi0-0-0-0"}'

{"execute":"block-export-add","arguments":{"type":"nbd","id":"libvirt-nbd-libvirt-1-storage","node-name":"libvirt-1-storage","writable":true,"name":"drive-scsi0-0-0-0"},"id":"libvirt-446"}

-blockdev '{"driver":"file","filename":"/var/lib/libvirt/images/raw.img","node-name":"libvirt-1-storage","read-only":false}'
-device '{"driver":"scsi-hd","bus":"scsi0.0","channel":0,"scsi-id":0,"lun":0,"device_id":"drive-scsi0-0-0-0","drive":"libvirt-1-storage","id":"scsi0-0-0-0"}'

{"execute":"block-export-add","arguments":{"type":"nbd","id":"libvirt-nbd-libvirt-1-storage","node-name":"libvirt-1-storage","writable":true,"name":"drive-scsi0-0-0-0"},"id":"libvirt-463"}

{
  "execute": "blockdev-add",
  "arguments": {
    "driver": "nbd",
    "server": {
      "type": "inet",
      "host": "XXX1",
      "port": "49153"
    },
    "export": "drive-scsi0-0-0-0",
    "node-name": "migration-sda-storage",
    "read-only": false,
    "discard": "unmap",
    "detect-zeroes": "unmap"
  },
  "id": "libvirt-455"
}

{"execute":"blockdev-mirror","arguments":{"job-id":"drive-scsi0-0-0-0","device":"libvirt-1-storage","target":"migration-sda-storage","speed":9223372036853727232,"sync":"full","auto-finalize":true,"auto-dismiss":false},"id":"libvirt-456"}

{
  "execute": "blockdev-add",
  "arguments": {
    "driver": "nbd",
    "server": {
      "type": "inet",
      "host": "XXX2",
      "port": "49153"
    },
    "export": "drive-scsi0-0-0-0",
    "node-name": "migration-sda-storage",
    "read-only": false,
    "discard": "unmap",
    "detect-zeroes": "unmap"
  },
  "id": "libvirt-472"
}

{"execute":"blockdev-mirror","arguments":{"job-id":"drive-scsi0-0-0-0","device":"libvirt-1-storage","target":"migration-sda-storage","speed":9223372036853727232,"sync":"full","auto-finalize":true,"auto-dismiss":false},"id":"libvirt-473"}

-blockdev '{"driver":"file","filename":"/var/lib/libvirt/images/raw.img","node-name":"libvirt-1-storage","read-only":false}' 
-device '{"driver":"scsi-hd","bus":"scsi0.0","channel":0,"scsi-id":0,"lun":0,"device_id":"drive-scsi0-0-0-0","drive":"libvirt-1-storage","id":"scsi0-0-0-0"}'

Test steps the same with RHEL 9.6.

Final check raw image info on dst host:
 # qemu-img info /var/lib/libvirt/images/raw.img
image: /var/lib/libvirt/images/raw.img
file format: raw
virtual size: 1 GiB (1073741824 bytes)
disk size: 0 B
Child node '/file':
...
 

Test steps:

Create raw image on src and dst host:
# qemu-img create -f raw /var/lib/libvirt/images/raw.img 1G

Guest xml without "discard='unmap'":
<disk type='file' device='disk'>
   <driver name='qemu' type='raw'/>
   <source file='/var/lib/libvirt/images/raw.img'/>
   <target dev='sda' bus='scsi'/>
   <address type='drive' controller='0' bus='0' target='0' unit='0'/>
</disk>

Zero out the sda block device in guest.
# dd if=/dev/zero of=/dev/sda bs=512 count=2097152

Check raw image info on src host.
# qemu-img info /var/lib/libvirt/images/raw.img
image: /var/lib/libvirt/images/raw.img
file format: raw
virtual size: 1 GiB (1073741824 bytes)
disk size: 1 GiB
...

Migrate VM.
# virsh migrate vm1 qemu+tcp://XXX/system --verbose --live --p2p --copy-storage-all --migrate-disks-detect-zeroes sda

Check raw image info on dst host:
# qemu-img info /var/lib/libvirt/images/raw.img
image: /var/lib/libvirt/images/raw.img
file format: raw
virtual size: 1 GiB (1073741824 bytes)
disk size: 1 GiB
...