X-Git-Url: https://git.proxmox.com/?p=qemu-server.git;a=blobdiff_plain;f=PVE%2FQemuServer.pm;h=035d8b7084145d50283b34cb992bdde68a8b7350;hp=deb58eddd46a82d605d0c1cb9b5edecc87cdac5f;hb=b4496b9ed5994cb624806db4edf5dbc24576bd9e;hpb=9444c6e42c2b3c79f1e506511623b05603755589 diff --git a/PVE/QemuServer.pm b/PVE/QemuServer.pm index deb58ed..035d8b7 100644 --- a/PVE/QemuServer.pm +++ b/PVE/QemuServer.pm @@ -34,12 +34,22 @@ use PVE::QemuServer::PCI qw(print_pci_addr print_pcie_addr); use PVE::QemuServer::Memory; use PVE::QemuServer::USB qw(parse_usb_device); use PVE::QemuServer::Cloudinit; +use PVE::Systemd; use Time::HiRes qw(gettimeofday); use File::Copy qw(copy); use URI::Escape; -my $OVMF_CODE = '/usr/share/kvm/OVMF_CODE-pure-efi.fd'; -my $OVMF_VARS = '/usr/share/kvm/OVMF_VARS-pure-efi.fd'; +my $EDK2_FW_BASE = '/usr/share/pve-edk2-firmware/'; +my $OVMF = { + x86_64 => [ + "$EDK2_FW_BASE/OVMF_CODE.fd", + "$EDK2_FW_BASE/OVMF_VARS.fd" + ], + aarch64 => [ + "$EDK2_FW_BASE/AAVMF_CODE.fd", + "$EDK2_FW_BASE/AAVMF_VARS.fd" + ], +}; my $qemu_snap_storage = {rbd => 1, sheepdog => 1}; @@ -78,6 +88,14 @@ PVE::JSONSchema::register_standard_option('pve-qm-image-format', { optional => 1, }); +PVE::JSONSchema::register_standard_option('pve-qemu-machine', { + description => "Specifies the Qemu machine type.", + type => 'string', + pattern => '(pc|pc(-i440fx)?-\d+\.\d+(\.pxe)?|q35|pc-q35-\d+\.\d+(\.pxe)?|virt(?:-\d+\.\d+)?)', + maxLength => 40, + optional => 1, +}); + #no warnings 'redefine'; sub cgroups_write { @@ -153,7 +171,7 @@ my $cpu_vendor_list = { max => 'default', }; -my $cpu_flag = qr/[+-](pcid|spec-ctrl)/; +my $cpu_flag = qr/[+-](pcid|spec-ctrl|ibpb|ssbd|virt-ssbd|amd-ssbd|amd-no-ssb|pdpe1gb)/; my $cpu_fmt = { cputype => { @@ -172,7 +190,7 @@ my $cpu_fmt = { flags => { description => "List of additional CPU flags separated by ';'." . " Use '+FLAG' to enable, '-FLAG' to disable a flag." - . " Currently supported flags: 'pcid', 'spec-ctrl'.", + . " Currently supported flags: 'pcid', 'spec-ctrl', 'ibpb', 'ssbd', 'virt-ssbd', 'amd-ssbd', 'amd-no-ssb', 'pdpe1gb'.", format_description => '+FLAG[;-FLAG...]', type => 'string', pattern => qr/$cpu_flag(;$cpu_flag)*/, @@ -198,6 +216,39 @@ my $watchdog_fmt = { }; PVE::JSONSchema::register_format('pve-qm-watchdog', $watchdog_fmt); +my $agent_fmt = { + enabled => { + description => "Enable/disable Qemu GuestAgent.", + type => 'boolean', + default => 0, + default_key => 1, + }, + fstrim_cloned_disks => { + description => "Run fstrim after cloning/moving a disk.", + type => 'boolean', + optional => 1, + default => 0 + }, +}; + +my $vga_fmt = { + type => { + description => "Select the VGA type.", + type => 'string', + default => 'std', + optional => 1, + default_key => 1, + enum => [qw(cirrus qxl qxl2 qxl3 qxl4 serial0 serial1 serial2 serial3 std virtio vmware)], + }, + memory => { + description => "Sets the VGA memory (in MiB). Has no effect with serial display.", + type => 'integer', + optional => 1, + minimum => 4, + maximum => 512, + }, +}; + my $confdesc = { onboot => { optional => 1, @@ -263,7 +314,7 @@ my $confdesc = { shares => { optional => 1, type => 'integer', - description => "Amount of memory shares for auto-ballooning. The larger the number is, the more memory this VM gets. Number is relative to weights of all other running VMs. Using zero disables auto-ballooning", + description => "Amount of memory shares for auto-ballooning. The larger the number is, the more memory this VM gets. Number is relative to weights of all other running VMs. Using zero disables auto-ballooning. Auto-ballooning is done by pvestatd.", minimum => 0, maximum => 50000, default => 1000, @@ -271,7 +322,7 @@ my $confdesc = { keyboard => { optional => 1, type => 'string', - description => "Keybord layout for vnc server. Default is read from the '/etc/pve/datacenter.conf' configuration file.". + description => "Keybord layout for vnc server. Default is read from the '/etc/pve/datacenter.cfg' configuration file.". "It should not be necessary to set it.", enum => PVE::Tools::kvmkeymaplist(), default => undef, @@ -378,9 +429,9 @@ EODESC }, agent => { optional => 1, - type => 'boolean', - description => "Enable/disable Qemu GuestAgent.", - default => 0, + description => "Enable/disable Qemu GuestAgent and its properties.", + type => 'string', + format => $agent_fmt, }, kvm => { optional => 1, @@ -406,17 +457,16 @@ EODESC }, vga => { optional => 1, - type => 'string', - description => "Select the VGA type.", - verbose_description => "Select the VGA type. If you want to use high resolution" . - " modes (>= 1280x1024x16) then you should use the options " . - "'std' or 'vmware'. Default is 'std' for win8/win7/w2k8, and " . - "'cirrus' for other OS types. The 'qxl' option enables the SPICE " . - "display sever. For win* OS you can select how many independent " . - "displays you want, Linux guests can add displays them self. " . - "You can also run without any graphic card, using a serial device" . - " as terminal.", - enum => [qw(std cirrus vmware qxl serial0 serial1 serial2 serial3 qxl2 qxl3 qxl4)], + type => 'string', format => $vga_fmt, + description => "Configure the VGA hardware.", + verbose_description => "Configure the VGA Hardware. If you want to use ". + "high resolution modes (>= 1280x1024x16) you may need to increase " . + "the vga memory option. Since QEMU 2.9 the default VGA display type " . + "is 'std' for all OS types besides some Windows versions (XP and " . + "older) which use 'cirrus'. The 'qxl' option enables the SPICE " . + "display server. For win* OS you can select how many independent " . + "displays you want, Linux guests can add displays them self.\n". + "You can also run without any graphic card, using a serial device as terminal.", }, watchdog => { optional => 1, @@ -511,12 +561,15 @@ EODESCR description => "Default storage for VM state volumes/files.", optional => 1, }), - machine => { - description => "Specific the Qemu machine type.", - type => 'string', - pattern => '(pc|pc(-i440fx)?-\d+\.\d+(\.pxe)?|q35|pc-q35-\d+\.\d+(\.pxe)?)', - maxLength => 40, + runningmachine => get_standard_option('pve-qemu-machine', { + description => "Specifies the Qemu machine type of the running vm. This is used internally for snapshots.", + }), + machine => get_standard_option('pve-qemu-machine'), + arch => { + description => "Virtual processor architecture. Defaults to the host.", optional => 1, + type => 'string', + enum => [qw(x86_64 aarch64)], }, smbios1 => { description => "Specify SMBIOS type 1 fields.", @@ -537,6 +590,24 @@ EODESCR description => "Select BIOS implementation.", default => 'seabios', }, + vmgenid => { + type => 'string', + pattern => '(?:[a-fA-F0-9]{8}(?:-[a-fA-F0-9]{4}){3}-[a-fA-F0-9]{12}|[01])', + format_description => 'UUID', + description => "Set VM Generation ID. Use '1' to autogenerate on create or update, pass '0' to disable explicitly.", + verbose_description => "The VM generation ID (vmgenid) device exposes a". + " 128-bit integer value identifier to the guest OS. This allows to". + " notify the guest operating system when the virtual machine is". + " executed with a different configuration (e.g. snapshot execution". + " or creation from a template). The guest operating system notices". + " the change, and is then able to react as appropriate by marking". + " its copies of distributed databases as dirty, re-initializing its". + " random number generator, etc.\n". + "Note that auto-creation only works when done throug API/CLI create". + " or update methods, but not when manually editing the config file.", + default => "1 (autogenerated)", + optional => 1, + }, }; my $confdesc_cloudinit = { @@ -601,7 +672,7 @@ my $MAX_VIRTIO_DISKS = 16; my $MAX_SATA_DISKS = 6; my $MAX_USB_DEVICES = 5; my $MAX_NETS = 32; -my $MAX_UNUSED_DISKS = 8; +my $MAX_UNUSED_DISKS = 256; my $MAX_HOSTPCI_DEVICES = 4; my $MAX_SERIAL_PORTS = 4; my $MAX_PARALLEL_PORTS = 3; @@ -960,6 +1031,14 @@ my %scsiblock_fmt = ( }, ); +my %ssd_fmt = ( + ssd => { + type => 'boolean', + description => "Whether to expose this drive as an SSD, rather than a rotational hard disk.", + optional => 1, + }, +); + my $add_throttle_desc = sub { my ($key, $type, $what, $unit, $longunit, $minimum) = @_; my $d = { @@ -1007,6 +1086,7 @@ $drivedesc_base{'iops_wr_length'} = { alias => 'iops_wr_max_length' }; my $ide_fmt = { %drivedesc_base, %model_fmt, + %ssd_fmt, }; PVE::JSONSchema::register_format("pve-qm-ide", $ide_fmt); @@ -1022,6 +1102,7 @@ my $scsi_fmt = { %iothread_fmt, %queues_fmt, %scsiblock_fmt, + %ssd_fmt, }; my $scsidesc = { optional => 1, @@ -1032,6 +1113,7 @@ PVE::JSONSchema::register_standard_option("pve-qm-scsi", $scsidesc); my $sata_fmt = { %drivedesc_base, + %ssd_fmt, }; my $satadesc = { optional => 1, @@ -1057,6 +1139,7 @@ my $alldrive_fmt = { %model_fmt, %queues_fmt, %scsiblock_fmt, + %ssd_fmt, }; my $efidisk_fmt = { @@ -1262,19 +1345,15 @@ for (my $i = 0; $i < $MAX_UNUSED_DISKS; $i++) { my $kvm_api_version = 0; sub kvm_version { - return $kvm_api_version if $kvm_api_version; - my $fh = IO::File->new("ioctl(KVM_GET_API_VERSION(), 0)) { - $kvm_api_version = $v; - } + # 0xae00 => KVM_GET_API_VERSION + $kvm_api_version = ioctl($fh, 0xae00, 0); - $fh->close(); - - return $kvm_api_version; + return $kvm_api_version; } my $kvm_user_version; @@ -1611,24 +1690,37 @@ sub machine_type_is_q35 { } sub print_tabletdevice_full { - my ($conf) = @_; + my ($conf, $arch) = @_; my $q35 = machine_type_is_q35($conf); # we use uhci for old VMs because tablet driver was buggy in older qemu - my $usbbus = $q35 ? "ehci" : "uhci"; + my $usbbus; + if (machine_type_is_q35($conf) || $arch eq 'aarch64') { + $usbbus = 'ehci'; + } else { + $usbbus = 'uhci'; + } return "usb-tablet,id=tablet,bus=$usbbus.0,port=1"; } +sub print_keyboarddevice_full { + my ($conf, $arch, $machine) = @_; + + return undef if $arch ne 'aarch64'; + + return "usb-kbd,id=keyboard,bus=ehci.0,port=2"; +} + sub print_drivedevice_full { - my ($storecfg, $conf, $vmid, $drive, $bridges) = @_; + my ($storecfg, $conf, $vmid, $drive, $bridges, $arch, $machine_type) = @_; my $device = ''; my $maxdev = 0; if ($drive->{interface} eq 'virtio') { - my $pciaddr = print_pci_addr("$drive->{interface}$drive->{index}", $bridges); + my $pciaddr = print_pci_addr("$drive->{interface}$drive->{index}", $bridges, $arch, $machine_type); $device = "virtio-blk-pci,drive=drive-$drive->{interface}$drive->{index},id=$drive->{interface}$drive->{index}$pciaddr"; $device .= ",iothread=iothread-$drive->{interface}$drive->{index}" if $drive->{iothread}; } elsif ($drive->{interface} eq 'scsi') { @@ -1664,21 +1756,33 @@ sub print_drivedevice_full { $device = "scsi-$devicetype,bus=$controller_prefix$controller.0,channel=0,scsi-id=0,lun=$drive->{index},drive=drive-$drive->{interface}$drive->{index},id=$drive->{interface}$drive->{index}"; } - } elsif ($drive->{interface} eq 'ide'){ - $maxdev = 2; + if ($drive->{ssd} && ($devicetype eq 'block' || $devicetype eq 'hd')) { + $device .= ",rotation_rate=1"; + } + + } elsif ($drive->{interface} eq 'ide' || $drive->{interface} eq 'sata') { + my $maxdev = ($drive->{interface} eq 'sata') ? $MAX_SATA_DISKS : 2; my $controller = int($drive->{index} / $maxdev); my $unit = $drive->{index} % $maxdev; my $devicetype = ($drive->{media} && $drive->{media} eq 'cdrom') ? "cd" : "hd"; - $device = "ide-$devicetype,bus=ide.$controller,unit=$unit,drive=drive-$drive->{interface}$drive->{index},id=$drive->{interface}$drive->{index}"; - if ($devicetype eq 'hd' && (my $model = $drive->{model})) { - $model = URI::Escape::uri_unescape($model); - $device .= ",model=$model"; + $device = "ide-$devicetype"; + if ($drive->{interface} eq 'ide') { + $device .= ",bus=ide.$controller,unit=$unit"; + } else { + $device .= ",bus=ahci$controller.$unit"; + } + $device .= ",drive=drive-$drive->{interface}$drive->{index},id=$drive->{interface}$drive->{index}"; + + if ($devicetype eq 'hd') { + if (my $model = $drive->{model}) { + $model = URI::Escape::uri_unescape($model); + $device .= ",model=$model"; + } + if ($drive->{ssd}) { + $device .= ",rotation_rate=1"; + } } - } elsif ($drive->{interface} eq 'sata'){ - my $controller = int($drive->{index} / $MAX_SATA_DISKS); - my $unit = $drive->{index} % $MAX_SATA_DISKS; - $device = "ide-drive,bus=ahci$controller.$unit,drive=drive-$drive->{interface}$drive->{index},id=$drive->{interface}$drive->{index}"; } elsif ($drive->{interface} eq 'usb') { die "implement me"; # -device ide-drive,bus=ide.1,unit=0,drive=drive-ide0-1-0,id=ide0-1-0 @@ -1688,6 +1792,12 @@ sub print_drivedevice_full { $device .= ",bootindex=$drive->{bootindex}" if $drive->{bootindex}; + if (my $serial = $drive->{serial}) { + $serial = URI::Escape::uri_unescape($serial); + $device .= ",serial=$serial"; + } + + return $device; } @@ -1760,11 +1870,6 @@ sub print_drive_full { } } - if (my $serial = $drive->{serial}) { - $serial = URI::Escape::uri_unescape($serial); - $opts .= ",serial=$serial"; - } - $opts .= ",format=$format" if $format && !$drive->{format}; my $cache_direct = 0; @@ -1804,7 +1909,7 @@ sub print_drive_full { } sub print_netdevice_full { - my ($vmid, $conf, $net, $netid, $bridges, $use_old_bios_files) = @_; + my ($vmid, $conf, $net, $netid, $bridges, $use_old_bios_files, $arch, $machine_type) = @_; my $bootorder = $conf->{boot} || $confdesc->{boot}->{default}; @@ -1813,7 +1918,7 @@ sub print_netdevice_full { $device = 'virtio-net-pci'; }; - my $pciaddr = print_pci_addr("$netid", $bridges); + my $pciaddr = print_pci_addr("$netid", $bridges, $arch, $machine_type); my $tmpstr = "$device,mac=$net->{macaddr},netdev=$netid$pciaddr,id=$netid"; if ($net->{queues} && $net->{queues} > 1 && $net->{model} eq 'virtio'){ #Consider we have N queues, the number of vectors needed is 2*N + 2 (plus one config interrupt and control vq) @@ -1842,7 +1947,7 @@ sub print_netdevice_full { } sub print_netdev_full { - my ($vmid, $conf, $net, $netid, $hotplug) = @_; + my ($vmid, $conf, $arch, $net, $netid, $hotplug) = @_; my $i = ''; if ($netid =~ m/^net(\d+)$/) { @@ -1858,7 +1963,9 @@ sub print_netdev_full { if length($ifname) >= 16; my $vhostparam = ''; - $vhostparam = ',vhost=on' if $kernel_has_vhost_net && $net->{model} eq 'virtio'; + if (is_native($arch)) { + $vhostparam = ',vhost=on' if $kernel_has_vhost_net && $net->{model} eq 'virtio'; + } my $vmname = $conf->{name} || "vm$vmid"; @@ -1896,6 +2003,58 @@ sub print_cpu_device { return "$cpu-x86_64-cpu,id=cpu$id,socket-id=$current_socket,core-id=$current_core,thread-id=0"; } +my $vga_map = { + 'cirrus' => 'cirrus-vga', + 'std' => 'VGA', + 'vmware' => 'vmware-svga', + 'virtio' => 'virtio-vga', +}; + +sub print_vga_device { + my ($conf, $vga, $arch, $machine, $id, $qxlnum, $bridges) = @_; + + my $type = $vga_map->{$vga->{type}}; + if ($type eq 'virtio-vga' && $arch eq 'aarch64') { + $type = 'virtio-gpu'; + } + my $vgamem_mb = $vga->{memory}; + if ($qxlnum) { + $type = $id ? 'qxl' : 'qxl-vga'; + } + die "no devicetype for $vga->{type}\n" if !$type; + + my $memory = ""; + if ($vgamem_mb) { + if ($vga->{type} eq 'virtio') { + my $bytes = PVE::Tools::convert_size($vgamem_mb, "mb" => "b"); + $memory = ",max_hostmem=$bytes"; + } elsif ($qxlnum) { + # from https://www.spice-space.org/multiple-monitors.html + $memory = ",vgamem_mb=$vga->{memory}"; + my $ram = $vgamem_mb * 4; + my $vram = $vgamem_mb * 2; + $memory .= ",ram_size_mb=$ram,vram_size_mb=$vram"; + } else { + $memory = ",vgamem_mb=$vga->{memory}"; + } + } elsif ($qxlnum && $id) { + $memory = ",ram_size=67108864,vram_size=33554432"; + } + + my $q35 = machine_type_is_q35($conf); + my $vgaid = "vga" . ($id // ''); + my $pciaddr; + + if ($q35 && $vgaid eq 'vga') { + # the first display uses pcie.0 bus on q35 machines + $pciaddr = print_pcie_addr($vgaid, $bridges, $arch, $machine); + } else { + $pciaddr = print_pci_addr($vgaid, $bridges, $arch, $machine); + } + + return "$type,id=${vgaid}${memory}${pciaddr}"; +} + sub drive_is_cloudinit { my ($drive) = @_; return $drive->{file} =~ m@[:/]vm-\d+-cloudinit(?:\.$QEMU_FORMAT_RE)?$@; @@ -2208,6 +2367,28 @@ sub parse_watchdog { return $res; } +sub parse_guest_agent { + my ($value) = @_; + + return {} if !defined($value->{agent}); + + my $res = eval { PVE::JSONSchema::parse_property_string($agent_fmt, $value->{agent}) }; + warn $@ if $@; + + # if the agent is disabled ignore the other potentially set properties + return {} if !$res->{enabled}; + return $res; +} + +sub parse_vga { + my ($value) = @_; + + return {} if !$value; + my $res = eval { PVE::JSONSchema::parse_property_string($vga_fmt, $value) }; + warn $@ if $@; + return $res; +} + PVE::JSONSchema::register_format('pve-qm-usb-device', \&verify_usb_device); sub verify_usb_device { my ($value, $noerr) = @_; @@ -2224,7 +2405,7 @@ sub json_config_properties { my $prop = shift; foreach my $opt (keys %$confdesc) { - next if $opt eq 'parent' || $opt eq 'snaptime' || $opt eq 'vmstate'; + next if $opt eq 'parent' || $opt eq 'snaptime' || $opt eq 'vmstate' || $opt eq 'runningmachine'; $prop->{$opt} = $confdesc->{$opt}; } @@ -2675,7 +2856,7 @@ sub check_cmdline { my @param = split(/\0/, $line); my $cmd = $param[0]; - return if !$cmd || ($cmd !~ m|kvm$| && $cmd !~ m|qemu-system-x86_64$|); + return if !$cmd || ($cmd !~ m|kvm$| && $cmd !~ m@(?:^|/)qemu-system-[^/]+$@); for (my $i = 0; $i < scalar (@param); $i++) { my $p = $param[$i]; @@ -2761,6 +2942,53 @@ sub disksize { return $drive->{size}; } +our $vmstatus_return_properties = { + vmid => get_standard_option('pve-vmid'), + status => { + description => "Qemu process status.", + type => 'string', + enum => ['stopped', 'running'], + }, + maxmem => { + description => "Maximum memory in bytes.", + type => 'integer', + optional => 1, + renderer => 'bytes', + }, + maxdisk => { + description => "Root disk size in bytes.", + type => 'integer', + optional => 1, + renderer => 'bytes', + }, + name => { + description => "VM name.", + type => 'string', + optional => 1, + }, + qmpstatus => { + description => "Qemu QMP agent status.", + type => 'string', + optional => 1, + }, + pid => { + description => "PID of running qemu process.", + type => 'integer', + optional => 1, + }, + uptime => { + description => "Uptime.", + type => 'integer', + optional => 1, + renderer => 'duration', + }, + cpus => { + description => "Maximum usable CPUs.", + type => 'number', + optional => 1, + }, +}; + my $last_proc_pid_stat; # get VM status information @@ -2786,7 +3014,7 @@ sub vmstatus { my $cfspath = PVE::QemuConfig->cfs_config_path($vmid); my $conf = PVE::Cluster::cfs_read_file($cfspath) || {}; - my $d = {}; + my $d = { vmid => $vmid }; $d->{pid} = $list->{$vmid}->{pid}; # fixme: better status? @@ -3042,11 +3270,113 @@ sub conf_has_serial { sub vga_conf_has_spice { my ($vga) = @_; - return 0 if !$vga || $vga !~ m/^qxl([234])?$/; + my $vgaconf = parse_vga($vga); + my $vgatype = $vgaconf->{type}; + return 0 if !$vgatype || $vgatype !~ m/^qxl([234])?$/; return $1 || 1; } +my $host_arch; # FIXME: fix PVE::Tools::get_host_arch +sub get_host_arch() { + $host_arch = (POSIX::uname())[4] if !$host_arch; + return $host_arch; +} + +sub is_native($) { + my ($arch) = @_; + return get_host_arch() eq $arch; +} + +my $default_machines = { + x86_64 => 'pc', + aarch64 => 'virt', +}; + +sub get_basic_machine_info { + my ($conf, $forcemachine) = @_; + + my $arch = $conf->{arch} // get_host_arch(); + my $machine = $forcemachine || $conf->{machine} || $default_machines->{$arch}; + return ($arch, $machine); +} + +sub get_ovmf_files($) { + my ($arch) = @_; + + my $ovmf = $OVMF->{$arch} + or die "no OVMF images known for architecture '$arch'\n"; + + return @$ovmf; +} + +my $Arch2Qemu = { + aarch64 => '/usr/bin/qemu-system-aarch64', + x86_64 => '/usr/bin/qemu-system-x86_64', +}; +sub get_command_for_arch($) { + my ($arch) = @_; + return '/usr/bin/kvm' if is_native($arch); + + my $cmd = $Arch2Qemu->{$arch} + or die "don't know how to emulate architecture '$arch'\n"; + return $cmd; +} + +sub get_cpu_options { + my ($conf, $arch, $kvm, $machine_type, $kvm_off, $kvmver, $winversion, $gpu_passthrough) = @_; + + my $cpuFlags = []; + my $ostype = $conf->{ostype}; + + my $cpu = $kvm ? "kvm64" : "qemu64"; + if ($arch eq 'aarch64') { + $cpu = 'cortex-a57'; + } + if (my $cputype = $conf->{cpu}) { + my $cpuconf = PVE::JSONSchema::parse_property_string($cpu_fmt, $cputype) + or die "Cannot parse cpu description: $cputype\n"; + $cpu = $cpuconf->{cputype}; + $kvm_off = 1 if $cpuconf->{hidden}; + + if (defined(my $flags = $cpuconf->{flags})) { + push @$cpuFlags, split(";", $flags); + } + } + + push @$cpuFlags , '+lahf_lm' if $cpu eq 'kvm64' && $arch eq 'x86_64'; + + push @$cpuFlags , '-x2apic' + if $conf->{ostype} && $conf->{ostype} eq 'solaris'; + + push @$cpuFlags, '+sep' if $cpu eq 'kvm64' || $cpu eq 'kvm32'; + + push @$cpuFlags, '-rdtscp' if $cpu =~ m/^Opteron/; + + if (qemu_machine_feature_enabled ($machine_type, $kvmver, 2, 3) && $arch eq 'x86_64') { + + push @$cpuFlags , '+kvm_pv_unhalt' if $kvm; + push @$cpuFlags , '+kvm_pv_eoi' if $kvm; + } + + add_hyperv_enlightenments($cpuFlags, $winversion, $machine_type, $kvmver, $conf->{bios}, $gpu_passthrough) if $kvm; + + push @$cpuFlags, 'enforce' if $cpu ne 'host' && $kvm && $arch eq 'x86_64'; + + push @$cpuFlags, 'kvm=off' if $kvm_off; + + if (my $cpu_vendor = $cpu_vendor_list->{$cpu}) { + push @$cpuFlags, "vendor=${cpu_vendor}" + if $cpu_vendor ne 'default'; + } elsif ($arch ne 'aarch64') { + die "internal error"; # should not happen + } + + $cpu .= "," . join(',', @$cpuFlags) if scalar(@$cpuFlags); + + return ('-cpu', $cpu); +} + sub config_to_command { my ($storecfg, $vmid, $conf, $defaults, $forcemachine) = @_; @@ -3054,7 +3384,6 @@ sub config_to_command { my $globalFlags = []; my $machineFlags = []; my $rtcFlags = []; - my $cpuFlags = []; my $devices = []; my $pciaddr = ''; my $bridges = {}; @@ -3062,9 +3391,15 @@ sub config_to_command { my $vernum = 0; # unknown my $ostype = $conf->{ostype}; my $winversion = windows_version($ostype); - my $kvm = $conf->{kvm} // 1; + my $kvm = $conf->{kvm}; - die "KVM virtualisation configured, but not available. Either disable in VM configuration or enable in BIOS.\n" if (!$cpuinfo->{hvm} && $kvm); + my ($arch, $machine_type) = get_basic_machine_info($conf, $forcemachine); + $kvm //= 1 if is_native($arch); + + if ($kvm) { + die "KVM virtualisation configured, but not available. Either disable in VM configuration or enable in BIOS.\n" + if !defined kvm_version(); + } if ($kvmver =~ m/^(\d+)\.(\d+)$/) { $vernum = $1*1000000+$2*1000; @@ -3078,14 +3413,13 @@ sub config_to_command { my $q35 = machine_type_is_q35($conf); my $hotplug_features = parse_hotplug_features(defined($conf->{hotplug}) ? $conf->{hotplug} : '1'); - my $machine_type = $forcemachine || $conf->{machine}; my $use_old_bios_files = undef; ($use_old_bios_files, $machine_type) = qemu_use_old_bios_files($machine_type); my $cpuunits = defined($conf->{cpuunits}) ? $conf->{cpuunits} : $defaults->{cpuunits}; - push @$cmd, '/usr/bin/kvm'; + push @$cmd, get_command_for_arch($arch); push @$cmd, '-id', $vmid; @@ -3099,6 +3433,10 @@ sub config_to_command { push @$cmd, '-chardev', "socket,id=qmp,path=$qmpsocket,server,nowait"; push @$cmd, '-mon', "chardev=qmp,mode=control"; + if (qemu_machine_feature_enabled($machine_type, $kvmver, 2, 12)) { + push @$cmd, '-chardev', "socket,id=qmp-event,path=/var/run/qmeventd.sock,reconnect=5"; + push @$cmd, '-mon', "chardev=qmp-event,mode=control"; + } push @$cmd, '-pidfile' , pidfile_name($vmid); @@ -3108,8 +3446,13 @@ sub config_to_command { push @$cmd, '-smbios', "type=1,$conf->{smbios1}"; } + if ($conf->{vmgenid}) { + push @$devices, '-device', 'vmgenid,guid='.$conf->{vmgenid}; + } + + my ($ovmf_code, $ovmf_vars) = get_ovmf_files($arch); if ($conf->{bios} && $conf->{bios} eq 'ovmf') { - die "uefi base image not found\n" if ! -f $OVMF_CODE; + die "uefi base image not found\n" if ! -f $ovmf_code; my $path; my $format; @@ -3131,28 +3474,30 @@ sub config_to_command { } else { warn "no efidisk configured! Using temporary efivars disk.\n"; $path = "/tmp/$vmid-ovmf.fd"; - PVE::Tools::file_copy($OVMF_VARS, $path, -s $OVMF_VARS); + PVE::Tools::file_copy($ovmf_vars, $path, -s $ovmf_vars); $format = 'raw'; } - push @$cmd, '-drive', "if=pflash,unit=0,format=raw,readonly,file=$OVMF_CODE"; + push @$cmd, '-drive', "if=pflash,unit=0,format=raw,readonly,file=$ovmf_code"; push @$cmd, '-drive', "if=pflash,unit=1,format=$format,id=drive-efidisk0,file=$path"; } # add usb controllers - my @usbcontrollers = PVE::QemuServer::USB::get_usb_controllers($conf, $bridges, $q35, $usbdesc->{format}, $MAX_USB_DEVICES); + my @usbcontrollers = PVE::QemuServer::USB::get_usb_controllers($conf, $bridges, $arch, $machine_type, $usbdesc->{format}, $MAX_USB_DEVICES); push @$devices, @usbcontrollers if @usbcontrollers; - my $vga = $conf->{vga}; + my $vga = parse_vga($conf->{vga}); - my $qxlnum = vga_conf_has_spice($vga); - $vga = 'qxl' if $qxlnum; + my $qxlnum = vga_conf_has_spice($conf->{vga}); + $vga->{type} = 'qxl' if $qxlnum; - if (!$vga) { - if (qemu_machine_feature_enabled($machine_type, $kvmver, 2, 9)) { - $vga = (!$winversion || $winversion >= 6) ? 'std' : 'cirrus'; + if (!$vga->{type}) { + if ($arch eq 'aarch64') { + $vga->{type} = 'virtio'; + } elsif (qemu_machine_feature_enabled($machine_type, $kvmver, 2, 9)) { + $vga->{type} = (!$winversion || $winversion >= 6) ? 'std' : 'cirrus'; } else { - $vga = ($winversion >= 6) ? 'std' : 'cirrus'; + $vga->{type} = ($winversion >= 6) ? 'std' : 'cirrus'; } } @@ -3163,10 +3508,14 @@ sub config_to_command { } else { $tablet = $defaults->{tablet}; $tablet = 0 if $qxlnum; # disable for spice because it is not needed - $tablet = 0 if $vga =~ m/^serial\d+$/; # disable if we use serial terminal (no vga card) + $tablet = 0 if $vga->{type} =~ m/^serial\d+$/; # disable if we use serial terminal (no vga card) } - push @$devices, '-device', print_tabletdevice_full($conf) if $tablet; + if ($tablet) { + push @$devices, '-device', print_tabletdevice_full($conf, $arch) if $tablet; + my $kbd = print_keyboarddevice_full($conf, $arch); + push @$devices, '-device', $kbd if defined($kbd); + } my $kvm_off = 0; my $gpu_passthrough; @@ -3181,7 +3530,7 @@ sub config_to_command { die "q35 machine model is not enabled" if !$q35; $pciaddr = print_pcie_addr("hostpci$i"); }else{ - $pciaddr = print_pci_addr("hostpci$i", $bridges); + $pciaddr = print_pci_addr("hostpci$i", $bridges, $arch, $machine_type); } my $rombar = defined($d->{rombar}) && !$d->{rombar} ? ',rombar=0' : ''; @@ -3191,7 +3540,7 @@ sub config_to_command { if ($d->{'x-vga'}) { $xvga = ',x-vga=on'; $kvm_off = 1; - $vga = 'none'; + $vga->{type} = 'none'; $gpu_passthrough = 1; if ($conf->{bios} && $conf->{bios} eq 'ovmf') { @@ -3230,7 +3579,14 @@ sub config_to_command { if ($path eq 'socket') { my $socket = "/var/run/qemu-server/${vmid}.serial$i"; push @$devices, '-chardev', "socket,id=serial$i,path=$socket,server,nowait"; - push @$devices, '-device', "isa-serial,chardev=serial$i"; + # On aarch64, serial0 is the UART device. Qemu only allows + # connecting UART devices via the '-serial' command line, as + # the device has a fixed slot on the hardware... + if ($arch eq 'aarch64' && $i == 0) { + push @$devices, '-serial', "chardev:serial$i"; + } else { + push @$devices, '-device', "isa-serial,chardev=serial$i"; + } } else { die "no such serial device\n" if ! -c $path; push @$devices, '-chardev', "tty,id=serial$i,path=$path"; @@ -3294,12 +3650,12 @@ sub config_to_command { push @$cmd, '-no-reboot' if defined($conf->{reboot}) && $conf->{reboot} == 0; - push @$cmd, '-vga', $vga if $vga && $vga !~ m/^serial\d+$/; # for kvm 77 and later - - if ($vga && $vga !~ m/^serial\d+$/ && $vga ne 'none'){ + if ($vga->{type} && $vga->{type} !~ m/^serial\d+$/ && $vga->{type} ne 'none'){ + push @$devices, '-device', print_vga_device($conf, $vga, $arch, $machine_type, undef, $qxlnum, $bridges); my $socket = vnc_socket($vmid); push @$cmd, '-vnc', "unix:$socket,x509,password"; } else { + push @$cmd, '-vga', 'none' if $vga->{type} eq 'none'; push @$cmd, '-nographic'; } @@ -3338,48 +3694,7 @@ sub config_to_command { push @$rtcFlags, 'base=localtime'; } - my $cpu = $kvm ? "kvm64" : "qemu64"; - if (my $cputype = $conf->{cpu}) { - my $cpuconf = PVE::JSONSchema::parse_property_string($cpu_fmt, $cputype) - or die "Cannot parse cpu description: $cputype\n"; - $cpu = $cpuconf->{cputype}; - $kvm_off = 1 if $cpuconf->{hidden}; - - if (defined(my $flags = $cpuconf->{flags})) { - push @$cpuFlags, split(";", $flags); - } - } - - push @$cpuFlags , '+lahf_lm' if $cpu eq 'kvm64'; - - push @$cpuFlags , '-x2apic' - if $conf->{ostype} && $conf->{ostype} eq 'solaris'; - - push @$cpuFlags, '+sep' if $cpu eq 'kvm64' || $cpu eq 'kvm32'; - - push @$cpuFlags, '-rdtscp' if $cpu =~ m/^Opteron/; - - if (qemu_machine_feature_enabled ($machine_type, $kvmver, 2, 3)) { - - push @$cpuFlags , '+kvm_pv_unhalt' if $kvm; - push @$cpuFlags , '+kvm_pv_eoi' if $kvm; - } - - add_hyperv_enlightenments($cpuFlags, $winversion, $machine_type, $kvmver, $conf->{bios}, $gpu_passthrough) if $kvm; - - push @$cpuFlags, 'enforce' if $cpu ne 'host' && $kvm; - - push @$cpuFlags, 'kvm=off' if $kvm_off; - - my $cpu_vendor = $cpu_vendor_list->{$cpu} || - die "internal error"; # should not happen - - push @$cpuFlags, "vendor=${cpu_vendor}" - if $cpu_vendor ne 'default'; - - $cpu .= "," . join(',', @$cpuFlags) if scalar(@$cpuFlags); - - push @$cmd, '-cpu', $cpu; + push @$cmd, get_cpu_options($conf, $arch, $kvm, $machine_type, $kvm_off, $kvmver, $winversion, $gpu_passthrough); PVE::QemuServer::Memory::config($conf, $vmid, $sockets, $cores, $defaults, $hotplug_features, $cmd); @@ -3392,9 +3707,9 @@ sub config_to_command { #push @$cmd, '-soundhw', 'es1370'; #push @$cmd, '-soundhw', $soundhw if $soundhw; - if($conf->{agent}) { + if (parse_guest_agent($conf)->{enabled}) { my $qgasocket = qmp_socket($vmid, 1); - my $pciaddr = print_pci_addr("qga0", $bridges); + my $pciaddr = print_pci_addr("qga0", $bridges, $arch, $machine_type); push @$devices, '-chardev', "socket,path=$qgasocket,server,nowait,id=qga0"; push @$devices, '-device', "virtio-serial,id=qga0$pciaddr"; push @$devices, '-device', 'virtserialport,chardev=qga0,name=org.qemu.guest_agent.0'; @@ -3406,17 +3721,21 @@ sub config_to_command { if ($qxlnum > 1) { if ($winversion){ for(my $i = 1; $i < $qxlnum; $i++){ - my $pciaddr = print_pci_addr("vga$i", $bridges); - push @$cmd, '-device', "qxl,id=vga$i,ram_size=67108864,vram_size=33554432$pciaddr"; + push @$devices, '-device', print_vga_device($conf, $vga, $arch, $machine_type, $i, $qxlnum, $bridges); } } else { # assume other OS works like Linux - push @$cmd, '-global', 'qxl-vga.ram_size=134217728'; - push @$cmd, '-global', 'qxl-vga.vram_size=67108864'; + my ($ram, $vram) = ("134217728", "67108864"); + if ($vga->{memory}) { + $ram = PVE::Tools::convert_size($qxlnum*4*$vga->{memory}, 'mb' => 'b'); + $vram = PVE::Tools::convert_size($qxlnum*2*$vga->{memory}, 'mb' => 'b'); + } + push @$cmd, '-global', "qxl-vga.ram_size=$ram"; + push @$cmd, '-global', "qxl-vga.vram_size=$vram"; } } - my $pciaddr = print_pci_addr("spice", $bridges); + my $pciaddr = print_pci_addr("spice", $bridges, $arch, $machine_type); my $nodename = PVE::INotify::nodename(); my $pfamily = PVE::Tools::get_host_address_family($nodename); @@ -3434,13 +3753,13 @@ sub config_to_command { # enable balloon by default, unless explicitly disabled if (!defined($conf->{balloon}) || $conf->{balloon}) { - $pciaddr = print_pci_addr("balloon0", $bridges); + $pciaddr = print_pci_addr("balloon0", $bridges, $arch, $machine_type); push @$devices, '-device', "virtio-balloon-pci,id=balloon0$pciaddr"; } if ($conf->{watchdog}) { my $wdopts = parse_watchdog($conf->{watchdog}); - $pciaddr = print_pci_addr("watchdog", $bridges); + $pciaddr = print_pci_addr("watchdog", $bridges, $arch, $machine_type); my $watchdog = $wdopts->{model} || 'i6300esb'; push @$devices, '-device', "$watchdog$pciaddr"; push @$devices, '-watchdog-action', $wdopts->{action} if $wdopts->{action}; @@ -3488,7 +3807,7 @@ sub config_to_command { my ($maxdev, $controller, $controller_prefix) = scsihw_infos($conf, $drive); - $pciaddr = print_pci_addr("$controller_prefix$controller", $bridges); + $pciaddr = print_pci_addr("$controller_prefix$controller", $bridges, $arch, $machine_type); my $scsihw_type = $scsihw =~ m/^virtio-scsi-single/ ? "virtio-scsi-pci" : $scsihw; my $iothread = ''; @@ -3510,14 +3829,14 @@ sub config_to_command { if ($drive->{interface} eq 'sata') { my $controller = int($drive->{index} / $MAX_SATA_DISKS); - $pciaddr = print_pci_addr("ahci$controller", $bridges); + $pciaddr = print_pci_addr("ahci$controller", $bridges, $arch, $machine_type); push @$devices, '-device', "ahci,id=ahci$controller,multifunction=on$pciaddr" if !$ahcicontroller->{$controller}; $ahcicontroller->{$controller}=1; } my $drive_cmd = print_drive_full($storecfg, $vmid, $drive); push @$devices, '-drive',$drive_cmd; - push @$devices, '-device', print_drivedevice_full($storecfg, $conf, $vmid, $drive, $bridges); + push @$devices, '-device', print_drivedevice_full($storecfg, $conf, $vmid, $drive, $bridges, $arch, $machine_type); }); for (my $i = 0; $i < $MAX_NETS; $i++) { @@ -3532,10 +3851,10 @@ sub config_to_command { $bootindex_hash->{n} += 1; } - my $netdevfull = print_netdev_full($vmid,$conf,$d,"net$i"); + my $netdevfull = print_netdev_full($vmid, $conf, $arch, $d, "net$i"); push @$devices, '-netdev', $netdevfull; - my $netdevicefull = print_netdevice_full($vmid, $conf, $d, "net$i", $bridges, $use_old_bios_files); + my $netdevicefull = print_netdevice_full($vmid, $conf, $d, "net$i", $bridges, $use_old_bios_files, $arch, $machine_type); push @$devices, '-device', $netdevicefull; } @@ -3549,7 +3868,7 @@ sub config_to_command { $bridges->{3} = 1 if $scsihw =~ m/^virtio-scsi-single/; while (my ($k, $v) = each %$bridges) { - $pciaddr = print_pci_addr("pci.$k"); + $pciaddr = print_pci_addr("pci.$k", undef, $arch, $machine_type); unshift @$devices, '-device', "pci-bridge,id=pci.$k,chassis_nr=$k$pciaddr" if $k > 0; } } @@ -3585,9 +3904,10 @@ sub spice_port { } sub qmp_socket { - my ($vmid, $qga) = @_; + my ($vmid, $qga, $name) = @_; my $sockettype = $qga ? 'qga' : 'qmp'; - return "${var_run_tmpdir}/$vmid.$sockettype"; + my $ext = $name ? '-'.$name : ''; + return "${var_run_tmpdir}/$vmid$ext.$sockettype"; } sub pidfile_name { @@ -3599,21 +3919,22 @@ sub vm_devices_list { my ($vmid) = @_; my $res = vm_mon_cmd($vmid, 'query-pci'); + my $devices_to_check = []; my $devices = {}; foreach my $pcibus (@$res) { - foreach my $device (@{$pcibus->{devices}}) { - next if !$device->{'qdev_id'}; - if ($device->{'pci_bridge'}) { - $devices->{$device->{'qdev_id'}} = 1; - foreach my $bridge_device (@{$device->{'pci_bridge'}->{devices}}) { - next if !$bridge_device->{'qdev_id'}; - $devices->{$bridge_device->{'qdev_id'}} = 1; - $devices->{$device->{'qdev_id'}}++; - } - } else { - $devices->{$device->{'qdev_id'}} = 1; - } + push @$devices_to_check, @{$pcibus->{devices}}, + } + + while (@$devices_to_check) { + my $to_check = []; + for my $d (@$devices_to_check) { + $devices->{$d->{'qdev_id'}} = 1 if $d->{'qdev_id'}; + next if !$d->{'pci_bridge'}; + + $devices->{$d->{'qdev_id'}} += scalar(@{$d->{'pci_bridge'}->{devices}}); + push @$to_check, @{$d->{'pci_bridge'}->{devices}}; } + $devices_to_check = $to_check; } my $resblock = vm_mon_cmd($vmid, 'query-block'); @@ -3645,18 +3966,22 @@ sub vm_devices_list { } sub vm_deviceplug { - my ($storecfg, $conf, $vmid, $deviceid, $device) = @_; + my ($storecfg, $conf, $vmid, $deviceid, $device, $arch, $machine_type) = @_; my $q35 = machine_type_is_q35($conf); my $devices_list = vm_devices_list($vmid); return 1 if defined($devices_list->{$deviceid}); - qemu_add_pci_bridge($storecfg, $conf, $vmid, $deviceid); # add PCI bridge if we need it for the device + qemu_add_pci_bridge($storecfg, $conf, $vmid, $deviceid, $arch, $machine_type); # add PCI bridge if we need it for the device if ($deviceid eq 'tablet') { - qemu_deviceadd($vmid, print_tabletdevice_full($conf)); + qemu_deviceadd($vmid, print_tabletdevice_full($conf, $arch)); + + } elsif ($deviceid eq 'keyboard') { + + qemu_deviceadd($vmid, print_keyboarddevice_full($conf, $arch)); } elsif ($deviceid =~ m/^usb(\d+)$/) { @@ -3671,7 +3996,7 @@ sub vm_deviceplug { qemu_iothread_add($vmid, $deviceid, $device); qemu_driveadd($storecfg, $vmid, $device); - my $devicefull = print_drivedevice_full($storecfg, $conf, $vmid, $device); + my $devicefull = print_drivedevice_full($storecfg, $conf, $vmid, $device, $arch, $machine_type); qemu_deviceadd($vmid, $devicefull); eval { qemu_deviceaddverify($vmid, $deviceid); }; @@ -3685,7 +4010,7 @@ sub vm_deviceplug { my $scsihw = defined($conf->{scsihw}) ? $conf->{scsihw} : "lsi"; - my $pciaddr = print_pci_addr($deviceid); + my $pciaddr = print_pci_addr($deviceid, undef, $arch, $machine_type); my $scsihw_type = $scsihw eq 'virtio-scsi-single' ? "virtio-scsi-pci" : $scsihw; my $devicefull = "$scsihw_type,id=$deviceid$pciaddr"; @@ -3704,10 +4029,10 @@ sub vm_deviceplug { } elsif ($deviceid =~ m/^(scsi)(\d+)$/) { - qemu_findorcreatescsihw($storecfg,$conf, $vmid, $device); + qemu_findorcreatescsihw($storecfg,$conf, $vmid, $device, $arch, $machine_type); qemu_driveadd($storecfg, $vmid, $device); - my $devicefull = print_drivedevice_full($storecfg, $conf, $vmid, $device); + my $devicefull = print_drivedevice_full($storecfg, $conf, $vmid, $device, $arch, $machine_type); eval { qemu_deviceadd($vmid, $devicefull); }; if (my $err = $@) { eval { qemu_drivedel($vmid, $deviceid); }; @@ -3717,13 +4042,13 @@ sub vm_deviceplug { } elsif ($deviceid =~ m/^(net)(\d+)$/) { - return undef if !qemu_netdevadd($vmid, $conf, $device, $deviceid); + return undef if !qemu_netdevadd($vmid, $conf, $arch, $device, $deviceid); my $machine_type = PVE::QemuServer::qemu_machine_pxe($vmid, $conf); my $use_old_bios_files = undef; ($use_old_bios_files, $machine_type) = qemu_use_old_bios_files($machine_type); - my $netdevicefull = print_netdevice_full($vmid, $conf, $device, $deviceid, undef, $use_old_bios_files); + my $netdevicefull = print_netdevice_full($vmid, $conf, $device, $deviceid, undef, $use_old_bios_files, $arch, $machine_type); qemu_deviceadd($vmid, $netdevicefull); eval { qemu_deviceaddverify($vmid, $deviceid); }; if (my $err = $@) { @@ -3735,7 +4060,7 @@ sub vm_deviceplug { } elsif (!$q35 && $deviceid =~ m/^(pci\.)(\d+)$/) { my $bridgeid = $2; - my $pciaddr = print_pci_addr($deviceid); + my $pciaddr = print_pci_addr($deviceid, undef, $arch, $machine_type); my $devicefull = "pci-bridge,id=pci.$bridgeid,chassis_nr=$bridgeid$pciaddr"; qemu_deviceadd($vmid, $devicefull); @@ -3757,7 +4082,7 @@ sub vm_deviceunplug { die "can't unplug bootdisk" if $conf->{bootdisk} && $conf->{bootdisk} eq $deviceid; - if ($deviceid eq 'tablet') { + if ($deviceid eq 'tablet' || $deviceid eq 'keyboard') { qemu_devicedel($vmid, $deviceid); @@ -3908,7 +4233,7 @@ sub qemu_devicedelverify { } sub qemu_findorcreatescsihw { - my ($storecfg, $conf, $vmid, $device) = @_; + my ($storecfg, $conf, $vmid, $device, $arch, $machine_type) = @_; my ($maxdev, $controller, $controller_prefix) = scsihw_infos($conf, $device); @@ -3916,7 +4241,7 @@ sub qemu_findorcreatescsihw { my $devices_list = vm_devices_list($vmid); if(!defined($devices_list->{$scsihwid})) { - vm_deviceplug($storecfg, $conf, $vmid, $scsihwid, $device); + vm_deviceplug($storecfg, $conf, $vmid, $scsihwid, $device, $arch, $machine_type); } return 1; @@ -3952,13 +4277,13 @@ sub qemu_deletescsihw { } sub qemu_add_pci_bridge { - my ($storecfg, $conf, $vmid, $device) = @_; + my ($storecfg, $conf, $vmid, $device, $arch, $machine_type) = @_; my $bridges = {}; my $bridgeid; - print_pci_addr($device, $bridges); + print_pci_addr($device, $bridges, $arch, $machine_type); while (my ($k, $v) = each %$bridges) { $bridgeid = $k; @@ -3969,7 +4294,7 @@ sub qemu_add_pci_bridge { my $devices_list = vm_devices_list($vmid); if (!defined($devices_list->{$bridge})) { - vm_deviceplug($storecfg, $conf, $vmid, $bridge); + vm_deviceplug($storecfg, $conf, $vmid, $bridge, $arch, $machine_type); } return 1; @@ -3983,9 +4308,9 @@ sub qemu_set_link_status { } sub qemu_netdevadd { - my ($vmid, $conf, $device, $deviceid) = @_; + my ($vmid, $conf, $arch, $device, $deviceid) = @_; - my $netdev = print_netdev_full($vmid, $conf, $device, $deviceid, 1); + my $netdev = print_netdev_full($vmid, $conf, $arch, $device, $deviceid, 1); my %options = split(/[=,]/, $netdev); vm_mon_cmd($vmid, "netdev_add", %options); @@ -3999,7 +4324,7 @@ sub qemu_netdevdel { } sub qemu_usb_hotplug { - my ($storecfg, $conf, $vmid, $deviceid, $device) = @_; + my ($storecfg, $conf, $vmid, $deviceid, $device, $arch, $machine_type) = @_; return if !$device; @@ -4012,7 +4337,7 @@ sub qemu_usb_hotplug { my $devicelist = vm_devices_list($vmid); if (!$devicelist->{xhci}) { - my $pciaddr = print_pci_addr("xhci"); + my $pciaddr = print_pci_addr("xhci", undef, $arch, $machine_type); qemu_deviceadd($vmid, "nec-usb-xhci,id=xhci$pciaddr"); } } @@ -4020,7 +4345,7 @@ sub qemu_usb_hotplug { $d->{usb3} = $device->{usb3}; # add the new one - vm_deviceplug($storecfg, $conf, $vmid, $deviceid, $d); + vm_deviceplug($storecfg, $conf, $vmid, $deviceid, $d, $arch, $machine_type); } sub qemu_cpu_hotplug { @@ -4182,7 +4507,7 @@ sub qemu_volume_snapshot { my $running = check_running($vmid); if ($running && do_snapshots_with_qemu($storecfg, $volid)){ - vm_mon_cmd($vmid, "snapshot-drive", device => $deviceid, name => $snap); + vm_mon_cmd($vmid, 'blockdev-snapshot-internal-sync', device => $deviceid, name => $snap); } else { PVE::Storage::volume_snapshot($storecfg, $volid, $snap); } @@ -4193,8 +4518,18 @@ sub qemu_volume_snapshot_delete { my $running = check_running($vmid); + if($running) { + + $running = undef; + my $conf = PVE::QemuConfig->load_config($vmid); + foreach_drive($conf, sub { + my ($ds, $drive) = @_; + $running = 1 if $drive->{file} eq $volid; + }); + } + if ($running && do_snapshots_with_qemu($storecfg, $volid)){ - vm_mon_cmd($vmid, "delete-drive-snapshot", device => $deviceid, name => $snap); + vm_mon_cmd($vmid, 'blockdev-snapshot-delete-internal-sync', device => $deviceid, name => $snap); } else { PVE::Storage::volume_snapshot_delete($storecfg, $volid, $snap, $running); } @@ -4244,6 +4579,7 @@ sub vmconfig_hotplug_pending { my ($vmid, $conf, $storecfg, $selection, $errors) = @_; my $defaults = load_defaults(); + my ($arch, $machine_type) = get_basic_machine_info($conf, undef); # commit values which do not have any impact on running VM first # Note: those option cannot raise errors, we we do not care about @@ -4279,9 +4615,12 @@ sub vmconfig_hotplug_pending { } elsif ($opt eq 'tablet') { die "skip\n" if !$hotplug_features->{usb}; if ($defaults->{tablet}) { - vm_deviceplug($storecfg, $conf, $vmid, $opt); + vm_deviceplug($storecfg, $conf, $vmid, 'tablet', $arch, $machine_type); + vm_deviceplug($storecfg, $conf, $vmid, 'keyboard', $arch, $machine_type) + if $arch eq 'aarch64'; } else { - vm_deviceunplug($vmid, $conf, $opt); + vm_deviceunplug($vmid, $conf, 'tablet'); + vm_deviceunplug($vmid, $conf, 'keyboard') if $arch eq 'aarch64'; } } elsif ($opt =~ m/^usb\d+/) { die "skip\n"; @@ -4294,7 +4633,10 @@ sub vmconfig_hotplug_pending { qemu_cpu_hotplug($vmid, $conf, undef); } elsif ($opt eq 'balloon') { # enable balloon device is not hotpluggable - die "skip\n" if !defined($conf->{balloon}) || $conf->{balloon}; + die "skip\n" if defined($conf->{balloon}) && $conf->{balloon} == 0; + # here we reset the ballooning value to memory + my $balloon = $conf->{memory} || $defaults->{memory}; + vm_mon_cmd($vmid, "balloon", value => $balloon*1024*1024); } elsif ($fast_plug_option->{$opt}) { # do nothing } elsif ($opt =~ m/^net(\d+)$/) { @@ -4351,9 +4693,12 @@ sub vmconfig_hotplug_pending { } elsif ($opt eq 'tablet') { die "skip\n" if !$hotplug_features->{usb}; if ($value == 1) { - vm_deviceplug($storecfg, $conf, $vmid, $opt); + vm_deviceplug($storecfg, $conf, $vmid, 'tablet', $arch, $machine_type); + vm_deviceplug($storecfg, $conf, $vmid, 'keyboard', $arch, $machine_type) + if $arch eq 'aarch64'; } elsif ($value == 0) { - vm_deviceunplug($vmid, $conf, $opt); + vm_deviceunplug($vmid, $conf, 'tablet'); + vm_deviceunplug($vmid, $conf, 'keyboard') if $arch eq 'aarch64'; } } elsif ($opt =~ m/^usb\d+$/) { die "skip\n"; @@ -4362,7 +4707,7 @@ sub vmconfig_hotplug_pending { die "skip\n" if !$hotplug_features->{usb} || $value =~ m/spice/i; my $d = eval { PVE::JSONSchema::parse_property_string($usbdesc->{format}, $value) }; die "skip\n" if !$d; - qemu_usb_hotplug($storecfg, $conf, $vmid, $opt, $d); + qemu_usb_hotplug($storecfg, $conf, $vmid, $opt, $d, $arch, $machine_type); } elsif ($opt eq 'vcpus') { die "skip\n" if !$hotplug_features->{cpu}; qemu_cpu_hotplug($vmid, $conf, $value); @@ -4380,7 +4725,7 @@ sub vmconfig_hotplug_pending { } elsif ($opt =~ m/^net(\d+)$/) { # some changes can be done without hotplug vmconfig_update_net($storecfg, $conf, $hotplug_features->{network}, - $vmid, $opt, $value); + $vmid, $opt, $value, $arch, $machine_type); } elsif (is_valid_drivename($opt)) { # some changes can be done without hotplug my $drive = parse_drive($opt, $value); @@ -4388,7 +4733,7 @@ sub vmconfig_hotplug_pending { &$apply_pending_cloudinit($opt, $value); } vmconfig_update_disk($storecfg, $conf, $hotplug_features->{disk}, - $vmid, $opt, $value, 1); + $vmid, $opt, $value, 1, $arch, $machine_type); } elsif ($opt =~ m/^memory$/) { #dimms die "skip\n" if !$hotplug_features->{memory}; $value = PVE::QemuServer::Memory::qemu_memory_hotplug($vmid, $conf, $defaults, $opt, $value); @@ -4517,7 +4862,7 @@ my $safe_string_ne = sub { }; sub vmconfig_update_net { - my ($storecfg, $conf, $hotplug, $vmid, $opt, $value) = @_; + my ($storecfg, $conf, $hotplug, $vmid, $opt, $value, $arch, $machine_type) = @_; my $newnet = parse_net($value); @@ -4558,14 +4903,14 @@ sub vmconfig_update_net { } if ($hotplug) { - vm_deviceplug($storecfg, $conf, $vmid, $opt, $newnet); + vm_deviceplug($storecfg, $conf, $vmid, $opt, $newnet, $arch, $machine_type); } else { die "skip\n"; } } sub vmconfig_update_disk { - my ($storecfg, $conf, $hotplug, $vmid, $opt, $value, $force) = @_; + my ($storecfg, $conf, $hotplug, $vmid, $opt, $value, $force, $arch, $machine_type) = @_; # fixme: do we need force? @@ -4666,7 +5011,7 @@ sub vmconfig_update_disk { die "skip\n" if !$hotplug || $opt =~ m/(ide|sata)(\d+)/; # hotplug new disks PVE::Storage::activate_volumes($storecfg, [$drive->{file}]) if $drive->{file} !~ m|^/dev/.+|; - vm_deviceplug($storecfg, $conf, $vmid, $opt, $drive); + vm_deviceplug($storecfg, $conf, $vmid, $opt, $drive, $arch, $machine_type); } sub vm_start { @@ -4838,6 +5183,13 @@ sub vm_start { } $properties{timeout} = 10 if $statefile; # setting up the scope shoul be quick + my $run_qemu = sub { + PVE::Tools::run_fork sub { + PVE::Systemd::enter_systemd_scope($vmid, "Proxmox VE VM $vmid", %properties); + run_command($cmd, %run_params); + }; + }; + if ($conf->{hugepages}) { my $code = sub { @@ -4847,11 +5199,7 @@ sub vm_start { PVE::QemuServer::Memory::hugepages_mount(); PVE::QemuServer::Memory::hugepages_allocate($hugepages_topology, $hugepages_host_topology); - eval { - PVE::Tools::enter_systemd_scope($vmid, "Proxmox VE VM $vmid", %properties); - run_command($cmd, %run_params); - }; - + eval { $run_qemu->() }; if (my $err = $@) { PVE::QemuServer::Memory::hugepages_reset($hugepages_host_topology); die $err; @@ -4862,10 +5210,7 @@ sub vm_start { eval { PVE::QemuServer::Memory::hugepages_update_locked($code); }; } else { - eval { - PVE::Tools::enter_systemd_scope($vmid, "Proxmox VE VM $vmid", %properties); - run_command($cmd, %run_params); - }; + eval { $run_qemu->() }; } if (my $err = $@) { @@ -4916,10 +5261,8 @@ sub vm_start { } } else { - if (!$statefile && (!defined($conf->{balloon}) || $conf->{balloon})) { - vm_mon_cmd_nocheck($vmid, "balloon", value => $conf->{balloon}*1024*1024) - if $conf->{balloon}; - } + vm_mon_cmd_nocheck($vmid, "balloon", value => $conf->{balloon}*1024*1024) + if !$statefile && $conf->{balloon}; foreach my $opt (keys %$conf) { next if $opt !~ m/^net\d+$/; @@ -5090,7 +5433,7 @@ sub vm_stop { eval { if ($shutdown) { - if (defined($conf) && $conf->{agent}) { + if (defined($conf) && parse_guest_agent($conf)->{enabled}) { vm_qmp_command($vmid, { execute => "guest-shutdown" }, $nocheck); } else { vm_qmp_command($vmid, { execute => "system_powerdown" }, $nocheck); @@ -5166,6 +5509,13 @@ sub vm_resume { PVE::QemuConfig->lock_config($vmid, sub { + my $res = vm_mon_cmd($vmid, 'query-status'); + my $resume_cmd = 'cont'; + + if ($res->{status} && $res->{status} eq 'suspended') { + $resume_cmd = 'system_wakeup'; + } + if (!$nocheck) { my $conf = PVE::QemuConfig->load_config($vmid); @@ -5173,10 +5523,10 @@ sub vm_resume { PVE::QemuConfig->check_lock($conf) if !($skiplock || PVE::QemuConfig->has_lock($conf, 'backup')); - vm_mon_cmd($vmid, "cont"); + vm_mon_cmd($vmid, $resume_cmd); } else { - vm_mon_cmd_nocheck($vmid, "cont"); + vm_mon_cmd_nocheck($vmid, $resume_cmd); } }); } @@ -5452,6 +5802,13 @@ sub restore_update_config_line { } else { print $outfd $line; } + } elsif (($line =~ m/^vmgenid: (.*)/)) { + my $vmgenid = $1; + if ($vmgenid ne '0') { + # always generate a new vmgenid if there was a valid one setup + $vmgenid = generate_uuid(); + } + print $outfd "vmgenid: $vmgenid\n"; } elsif (($line =~ m/^(smbios1: )(.*)/) && $unique) { my ($uuid, $uuid_str); UUID::generate($uuid); @@ -5526,6 +5883,7 @@ sub update_disksize { my ($vmid, $conf, $volid_hash) = @_; my $changes; + my $prefix = "VM $vmid:"; # used and unused disks my $referenced = {}; @@ -5557,6 +5915,7 @@ sub update_disksize { if ($new ne $conf->{$opt}) { $changes = 1; $conf->{$opt} = $new; + print "$prefix update disk '$opt' information.\n"; } } } @@ -5567,6 +5926,7 @@ sub update_disksize { my $volid = $conf->{$opt}; my $path = $volid_hash->{$volid}->{path} if $volid_hash->{$volid}; if ($referenced->{$volid} || ($path && $referencedpath->{$path})) { + print "$prefix remove entry '$opt', its volume '$volid' is in use.\n"; $changes = 1; delete $conf->{$opt}; } @@ -5582,7 +5942,8 @@ sub update_disksize { next if !$path; # just to be sure next if $referencedpath->{$path}; $changes = 1; - PVE::QemuConfig->add_unused_volume($conf, $volid); + my $key = PVE::QemuConfig->add_unused_volume($conf, $volid); + print "$prefix add unreferenced volume '$volid' as '$key' to config.\n"; $referencedpath->{$path} = 1; # avoid to add more than once (aliases) } @@ -5590,10 +5951,17 @@ sub update_disksize { } sub rescan { - my ($vmid, $nolock) = @_; + my ($vmid, $nolock, $dryrun) = @_; my $cfg = PVE::Storage::config(); + # FIXME: Remove once our RBD plugin can handle CT and VM on a single storage + # see: https://pve.proxmox.com/pipermail/pve-devel/2018-July/032900.html + foreach my $stor (keys %{$cfg->{ids}}) { + delete($cfg->{ids}->{$stor}) if ! $cfg->{ids}->{$stor}->{content}->{images}; + } + + print "rescan volumes...\n"; my $volid_hash = scan_volids($cfg, $vmid); my $updatefn = sub { @@ -5611,7 +5979,7 @@ sub rescan { my $changes = update_disksize($vmid, $conf, $vm_volids); - PVE::QemuConfig->write_config($vmid, $conf) if $changes; + PVE::QemuConfig->write_config($vmid, $conf) if $changes && !$dryrun; }; if (defined($vmid)) { @@ -6091,11 +6459,11 @@ sub do_snapshots_with_qemu { } sub qga_check_running { - my ($vmid) = @_; + my ($vmid, $nowarn) = @_; eval { vm_mon_cmd($vmid, "guest-ping", timeout => 3); }; if ($@) { - warn "Qemu Guest Agent is not running - $@"; + warn "Qemu Guest Agent is not running - $@" if !$nowarn; return 0; } return 1; @@ -6144,7 +6512,9 @@ sub qemu_img_convert { my $cmd = []; push @$cmd, '/usr/bin/qemu-img', 'convert', '-p', '-n'; - push @$cmd, '-s', $snapname if($snapname && $src_format eq "qcow2"); + push @$cmd, '-l', "snapshot.name=$snapname" if($snapname && $src_format eq "qcow2"); + push @$cmd, '-t', 'none' if $dst_scfg->{type} eq 'zfspool'; + push @$cmd, '-T', 'none' if $src_scfg->{type} eq 'zfspool'; push @$cmd, '-f', $src_format, '-O', $dst_format, $src_path; if ($is_zero_initialized) { push @$cmd, "zeroinit:$dst_path"; @@ -6439,7 +6809,7 @@ sub qemu_machine_feature_enabled { my $current_major; my $current_minor; - if ($machine && $machine =~ m/^(pc(-i440fx|-q35)?-(\d+)\.(\d+))/) { + if ($machine && $machine =~ m/^((?:pc(-i440fx|-q35)?|virt)-(\d+)\.(\d+))/) { $current_major = $3; $current_minor = $4; @@ -6450,9 +6820,9 @@ sub qemu_machine_feature_enabled { $current_minor = $2; } - return 1 if $current_major >= $version_major && $current_minor >= $version_minor; - - + return 1 if $current_major > $version_major || + ($current_major == $version_major && + $current_minor >= $version_minor); } sub qemu_machine_pxe { @@ -6460,13 +6830,8 @@ sub qemu_machine_pxe { $machine = PVE::QemuServer::get_current_qemu_machine($vmid) if !$machine; - foreach my $opt (keys %$conf) { - next if $opt !~ m/^net(\d+)$/; - my $net = PVE::QemuServer::parse_net($conf->{$opt}); - next if !$net; - my $romfile = PVE::QemuServer::vm_mon_cmd_nocheck($vmid, 'qom-get', path => $opt, property => 'romfile'); - return $machine.".pxe" if $romfile =~ m/pxe/; - last; + if ($conf->{machine} && $conf->{machine} =~ m/\.pxe$/) { + $machine .= '.pxe'; } return $machine; @@ -6494,18 +6859,19 @@ sub qemu_use_old_bios_files { return ($use_old_bios_files, $machine_type); } -sub create_efidisk { - my ($storecfg, $storeid, $vmid, $fmt) = @_; +sub create_efidisk($$$$$) { + my ($storecfg, $storeid, $vmid, $fmt, $arch) = @_; - die "EFI vars default image not found\n" if ! -f $OVMF_VARS; + my (undef, $ovmf_vars) = get_ovmf_files($arch); + die "EFI vars default image not found\n" if ! -f $ovmf_vars; - my $vars_size = PVE::Tools::convert_size(-s $OVMF_VARS, 'b' => 'kb'); + my $vars_size = PVE::Tools::convert_size(-s $ovmf_vars, 'b' => 'kb'); my $volid = PVE::Storage::vdisk_alloc($storecfg, $storeid, $vmid, $fmt, undef, $vars_size); PVE::Storage::activate_volumes($storecfg, [$volid]); my $path = PVE::Storage::path($storecfg, $volid); eval { - run_command(['/usr/bin/qemu-img', 'convert', '-n', '-f', 'raw', '-O', $fmt, $OVMF_VARS, $path]); + run_command(['/usr/bin/qemu-img', 'convert', '-n', '-f', 'raw', '-O', $fmt, $ovmf_vars, $path]); }; die "Copying EFI vars image failed: $@" if $@; @@ -6587,6 +6953,11 @@ sub add_hyperv_enlightenments { if ($winversion >= 7) { push @$cpuFlags , 'hv_relaxed'; + + if (qemu_machine_feature_enabled ($machine_type, $kvmver, 2, 12)) { + push @$cpuFlags , 'hv_synic'; + push @$cpuFlags , 'hv_stimer'; + } } } @@ -6641,11 +7012,21 @@ sub resolve_first_disk { return $firstdisk; } -sub generate_smbios1_uuid { +sub generate_uuid { my ($uuid, $uuid_str); UUID::generate($uuid); UUID::unparse($uuid, $uuid_str); - return "uuid=$uuid_str"; + return $uuid_str; +} + +sub generate_smbios1_uuid { + return "uuid=".generate_uuid(); +} + +sub nbd_stop { + my ($vmid) = @_; + + vm_mon_cmd($vmid, 'nbd-server-stop'); } # bash completion helper