fstrim dont work after live migrate
Description of problem:
We have use lvm thin pool and after live migration non-shared storage fstrim cannot free data usage Data% without reboot, after reboot fstim work fine

```
  LV      VG Attr       LSize   Pool Origin Data%  Meta%  Move Log Cpy%Sync Convert
  p639937 vm Vwi-aotz--  30.00g pool        99.35

virsh qemu-agent-command p639937 '{"execute":"guest-fstrim"}'
{"return":{"paths":[{"minimum":0,"path":"/","trimmed":0}]}}

virsh shutdown p639937
Domain 'p639937' is being shutdown

virsh start p639937
Domain 'p639937' started

virsh qemu-agent-command p639937 '{"execute":"guest-fstrim"}'
{"return":{"paths":[{"minimum":0,"path":"/","trimmed":29178654720}]}}

lvs|grep p639937
  p639937 vm Vwi-aotz--  30.00g pool        9.58
```

On source host before migration:
```
  LV      VG Attr       LSize  Pool Origin Data%  Meta%  Move Log Cpy%Sync Convert
  p639937 vm Vwi-a-tz-- 30.00g pool        9.48
```

migration script 
```
SSH_OPTS='-o StrictHostKeyChecking=no -o PasswordAuthentication=no '
MIGR_OPTS="--live --copy-storage-all --verbose --persistent --undefinesource"
ssh $SSH_OPTS $HOST -t "[ -b /dev/vm/$ACCT ] || /usr/sbin/lvcreate -V${SIZE}G -T vm/pool -n$ACCT" || f_print_err "Error: creation lvm"
virsh migrate $MIGR_OPTS $ACCT qemu+ssh://$SERV/system  tcp://local.$SERV/ || f_print_err "Error on step: virsh migrate"
echo "Waiting for trim start..."
sleep 10
ssh $SSH_OPTS $HOST -t "/usr/bin/virsh qemu-agent-command $ACCT --timeout 60 '{\"execute\":\"guest-fstrim\"}' >/dev/null 2>&1"
```

Disc config:
```
    <disk type='block' device='disk'>
      <driver name='qemu' type='raw' cache='none' io='threads' discard='unmap'/>
      <source dev='/dev/vm/p639937'/>
      <backingStore/>
      <target dev='sda' bus='scsi'/>
      <iotune>
        <write_bytes_sec>104857600</write_bytes_sec>
        <write_bytes_sec_max>524288000</write_bytes_sec_max>
        <write_bytes_sec_max_length>120</write_bytes_sec_max_length>
      </iotune>
      <address type='drive' controller='0' bus='0' target='0' unit='0'/>
    </disk>
```

Sometimes trimming working after migration, bit this is very rare.
We have try rescanning disc, drop caches on vm after migration, but didnt help.

Inside vm's ext4 fs and almalinux 8/ubuntu 20+/debian 10-11