use strict;
use warnings;
+
use POSIX;
use IO::Handle;
use IO::Select;
use Storable qw(dclone);
use PVE::Exception qw(raise raise_param_exc);
use PVE::Storage;
-use PVE::Tools qw(run_command lock_file lock_file_full file_read_firstline dir_glob_foreach);
+use PVE::Tools qw(run_command lock_file lock_file_full file_read_firstline dir_glob_foreach $IPV6RE);
use PVE::JSONSchema qw(get_standard_option);
use PVE::Cluster qw(cfs_register_file cfs_read_file cfs_write_file cfs_lock_file);
use PVE::INotify;
use PVE::QemuConfig;
use PVE::QMPClient;
use PVE::RPCEnvironment;
+use PVE::GuestHelpers;
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::SysFSTools;
+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 $OVMF_IMG = '/usr/share/kvm/OVMF-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};
my $cpuinfo = PVE::ProcFSTools::read_cpuinfo();
+my $QEMU_FORMAT_RE = qr/raw|cow|qcow|qcow2|qed|vmdk|cloop/;
+
# Note about locking: we use flock on the config file protect
# against concurent actions.
# Aditionaly, we have a 'lock' setting in the config file. This
optional => 1,
});
-PVE::JSONSchema::register_standard_option('pve-snapshot-name', {
- description => "The name of the snapshot.",
- type => 'string', format => 'pve-configid',
- maxLength => 40,
-});
-
PVE::JSONSchema::register_standard_option('pve-qm-image-format', {
type => 'string',
enum => [qw(raw cow qcow qed qcow2 vmdk cloop)],
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 {
my $lock_dir = "/var/lock/qemu-server";
mkdir $lock_dir;
-my $pcisysfs = "/sys/bus/pci";
-
my $cpu_vendor_list = {
# Intel CPUs
486 => 'GenuineIntel',
coreduo => 'GenuineIntel',
core2duo => 'GenuineIntel',
Conroe => 'GenuineIntel',
- Penryn => 'GenuineIntel',
+ Penryn => 'GenuineIntel',
Nehalem => 'GenuineIntel',
+ 'Nehalem-IBRS' => 'GenuineIntel',
Westmere => 'GenuineIntel',
+ 'Westmere-IBRS' => 'GenuineIntel',
SandyBridge => 'GenuineIntel',
+ 'SandyBridge-IBRS' => 'GenuineIntel',
IvyBridge => 'GenuineIntel',
+ 'IvyBridge-IBRS' => 'GenuineIntel',
Haswell => 'GenuineIntel',
+ 'Haswell-IBRS' => 'GenuineIntel',
'Haswell-noTSX' => 'GenuineIntel',
+ 'Haswell-noTSX-IBRS' => 'GenuineIntel',
Broadwell => 'GenuineIntel',
+ 'Broadwell-IBRS' => 'GenuineIntel',
'Broadwell-noTSX' => 'GenuineIntel',
+ 'Broadwell-noTSX-IBRS' => 'GenuineIntel',
'Skylake-Client' => 'GenuineIntel',
-
+ 'Skylake-Client-IBRS' => 'GenuineIntel',
+ 'Skylake-Server' => 'GenuineIntel',
+ 'Skylake-Server-IBRS' => 'GenuineIntel',
+
# AMD CPUs
athlon => 'AuthenticAMD',
phenom => 'AuthenticAMD',
Opteron_G3 => 'AuthenticAMD',
Opteron_G4 => 'AuthenticAMD',
Opteron_G5 => 'AuthenticAMD',
+ EPYC => 'AuthenticAMD',
+ 'EPYC-IBPB' => 'AuthenticAMD',
# generic types, use vendor from host node
host => 'default',
kvm64 => 'default',
qemu32 => 'default',
qemu64 => 'default',
+ max => 'default',
};
+my $cpu_flag = qr/[+-](pcid|spec-ctrl|ibpb|ssbd|virt-ssbd|amd-ssbd|amd-no-ssb|pdpe1gb)/;
+
my $cpu_fmt = {
cputype => {
description => "Emulated CPU type.",
optional => 1,
default => 0
},
+ 'hv-vendor-id' => {
+ type => 'string',
+ pattern => qr/[a-zA-Z0-9]{1,12}/,
+ format_description => 'vendor-id',
+ description => 'The Hyper-V vendor ID. Some drivers or programs inside Windows guests need a specific ID.',
+ optional => 1,
+ },
+ flags => {
+ description => "List of additional CPU flags separated by ';'."
+ . " Use '+FLAG' to enable, '-FLAG' to disable a flag."
+ . " Currently supported flags: 'pcid', 'spec-ctrl', 'ibpb', 'ssbd', 'virt-ssbd', 'amd-ssbd', 'amd-no-ssb', 'pdpe1gb'.",
+ format_description => '+FLAG[;-FLAG...]',
+ type => 'string',
+ pattern => qr/$cpu_flag(;$cpu_flag)*/,
+ optional => 1,
+ },
};
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 none 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 $ivshmem_fmt = {
+ size => {
+ type => 'integer',
+ minimum => 1,
+ description => "The size of the file in MB.",
+ },
+ name => {
+ type => 'string',
+ pattern => '[a-zA-Z0-9\-]+',
+ optional => 1,
+ format_description => 'string',
+ description => "The name of the file. Will be prefixed with 'pve-shm-'. Default is the VMID. Will be deleted when the VM is stopped.",
+ },
+};
+
my $confdesc = {
onboot => {
optional => 1,
optional => 1,
type => 'string',
description => "Lock/unlock the VM.",
- enum => [qw(migrate backup snapshot rollback)],
+ enum => [qw(backup clone create migrate rollback snapshot snapshot-delete suspending suspended)],
},
cpulimit => {
optional => 1,
optional => 1,
type => 'integer',
description => "CPU weight for a VM.",
- verbose_description => "CPU weight for a VM. Argument is used in the kernel fair scheduler. The larger the number is, the more CPU time this VM gets. Number is relative to weights of all the other running VMs.\n\nNOTE: You can disable fair-scheduler configuration by setting this to 0.",
- minimum => 0,
- maximum => 500000,
+ verbose_description => "CPU weight for a VM. Argument is used in the kernel fair scheduler. The larger the number is, the more CPU time this VM gets. Number is relative to weights of all the other running VMs.",
+ minimum => 2,
+ maximum => 262144,
default => 1024,
},
memory => {
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,
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 => 'en-us',
+ default => undef,
},
name => {
optional => 1,
w2k8;; Microsoft Windows 2008
wvista;; Microsoft Windows Vista
win7;; Microsoft Windows 7
-win8;; Microsoft Windows 8/2012
+win8;; Microsoft Windows 8/2012/2012r2
+win10;; Microsoft Windows 10/2016
l24;; Linux 2.4 Kernel
l26;; Linux 2.6/3.X Kernel
solaris;; Solaris/OpenSolaris/OpenIndiania kernel
},
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,
},
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,
type => 'string', format => 'pve-volume-id',
description => "Reference to a volume which stores the VM state. This is used internally for snapshots.",
},
- machine => {
- description => "Specific the Qemu machine type.",
- type => 'string',
- pattern => '(pc|pc(-i440fx)?-\d+\.\d+(\.pxe)?|q35|pc-q35-\d+\.\d+(\.pxe)?)',
- maxLength => 40,
+ vmstatestorage => get_standard_option('pve-storage-id', {
+ description => "Default storage for VM state volumes/files.",
+ optional => 1,
+ }),
+ 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.",
maxLength => 256,
optional => 1,
},
- replica => {
- optional => 1,
- description => "Storage replica for local storage.",
- type => 'boolean',
- default => 0,
- },
- replica_rate_limit => {
- optional => 1,
- description => "Storage replica rate limit in KBytes/s.",
- type => 'integer',
- minimum => 1,
- },
- replica_interval => {
- optional => 1,
- description => "Storage replica sync interval",
- type => 'integer',
- minimum => 1,
- maximum => 1440,
- default => 15,
- },
- replica_target => {
- optional => 1,
- description => "Storage replica taget node.",
- type => 'string',
- },
protection => {
optional => 1,
type => 'boolean',
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,
+ },
+ hookscript => {
+ type => 'string',
+ format => 'pve-volume-id',
+ optional => 1,
+ description => "Script that will be executed during various steps in the vms lifetime.",
+ },
+ ivshmem => {
+ type => 'string',
+ format => $ivshmem_fmt,
+ description => "Inter-VM shared memory. Useful for direct communication between VMs, or to the host.",
+ optional => 1,
+ }
+};
+
+my $cicustom_fmt = {
+ meta => {
+ type => 'string',
+ optional => 1,
+ description => 'Specify a custom file containing all meta data passed to the VM via cloud-init. This is provider specific meaning configdrive2 and nocloud differ.',
+ format => 'pve-volume-id',
+ format_description => 'volume',
+ },
+ network => {
+ type => 'string',
+ optional => 1,
+ description => 'Specify a custom file containing all network data passed to the VM via cloud-init.',
+ format => 'pve-volume-id',
+ format_description => 'volume',
+ },
+ user => {
+ type => 'string',
+ optional => 1,
+ description => 'Specify a custom file containing all user data passed to the VM via cloud-init.',
+ format => 'pve-volume-id',
+ format_description => 'volume',
+ },
+};
+PVE::JSONSchema::register_format('pve-qm-cicustom', $cicustom_fmt);
+
+my $confdesc_cloudinit = {
+ citype => {
+ optional => 1,
+ type => 'string',
+ description => 'Specifies the cloud-init configuration format. The default depends on the configured operating system type (`ostype`. We use the `nocloud` format for Linux, and `configdrive2` for windows.',
+ enum => ['configdrive2', 'nocloud'],
+ },
+ ciuser => {
+ optional => 1,
+ type => 'string',
+ description => "cloud-init: User name to change ssh keys and password for instead of the image's configured default user.",
+ },
+ cipassword => {
+ optional => 1,
+ type => 'string',
+ description => 'cloud-init: Password to assign the user. Using this is generally not recommended. Use ssh keys instead. Also note that older cloud-init versions do not support hashed passwords.',
+ },
+ cicustom => {
+ optional => 1,
+ type => 'string',
+ description => 'cloud-init: Specify custom files to replace the automatically generated ones at start.',
+ format => 'pve-qm-cicustom',
+ },
+ searchdomain => {
+ optional => 1,
+ type => 'string',
+ description => "cloud-init: Sets DNS search domains for a container. Create will automatically use the setting from the host if neither searchdomain nor nameserver are set.",
+ },
+ nameserver => {
+ optional => 1,
+ type => 'string', format => 'address-list',
+ description => "cloud-init: Sets DNS server IP address for a container. Create will automatically use the setting from the host if neither searchdomain nor nameserver are set.",
+ },
+ sshkeys => {
+ optional => 1,
+ type => 'string',
+ format => 'urlencoded',
+ description => "cloud-init: Setup public SSH keys (one key per line, OpenSSH format).",
+ },
};
# what about other qemu settings ?
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;
__EOD__
my $net_fmt = {
- macaddr => {
- type => 'string',
- pattern => qr/[0-9a-f]{2}(?::[0-9a-f]{2}){5}/i,
+ macaddr => get_standard_option('mac-addr', {
description => "MAC address. That address must be unique withing your network. This is automatically generated if not specified.",
- format_description => "XX:XX:XX:XX:XX:XX",
- optional => 1,
- },
+ }),
model => {
type => 'string',
description => "Network Card Model. The 'virtio' model provides the best performance with very low CPU overhead. If your guest does not support this driver, it is usually best to use 'e1000'.",
PVE::JSONSchema::register_standard_option("pve-qm-net", $netdesc);
+my $ipconfig_fmt = {
+ ip => {
+ type => 'string',
+ format => 'pve-ipv4-config',
+ format_description => 'IPv4Format/CIDR',
+ description => 'IPv4 address in CIDR format.',
+ optional => 1,
+ default => 'dhcp',
+ },
+ gw => {
+ type => 'string',
+ format => 'ipv4',
+ format_description => 'GatewayIPv4',
+ description => 'Default gateway for IPv4 traffic.',
+ optional => 1,
+ requires => 'ip',
+ },
+ ip6 => {
+ type => 'string',
+ format => 'pve-ipv6-config',
+ format_description => 'IPv6Format/CIDR',
+ description => 'IPv6 address in CIDR format.',
+ optional => 1,
+ default => 'dhcp',
+ },
+ gw6 => {
+ type => 'string',
+ format => 'ipv6',
+ format_description => 'GatewayIPv6',
+ description => 'Default gateway for IPv6 traffic.',
+ optional => 1,
+ requires => 'ip6',
+ },
+};
+PVE::JSONSchema::register_format('pve-qm-ipconfig', $ipconfig_fmt);
+my $ipconfigdesc = {
+ optional => 1,
+ type => 'string', format => 'pve-qm-ipconfig',
+ description => <<'EODESCR',
+cloud-init: Specify IP addresses and gateways for the corresponding interface.
+
+IP addresses use CIDR notation, gateways are optional but need an IP of the same type specified.
+
+The special string 'dhcp' can be used for IP addresses to use DHCP, in which case no explicit gateway should be provided.
+For IPv6 the special string 'auto' can be used to use stateless autoconfiguration.
+
+If cloud-init is enabled and neither an IPv4 nor an IPv6 address is specified, it defaults to using dhcp on IPv4.
+EODESCR
+};
+PVE::JSONSchema::register_standard_option("pve-qm-ipconfig", $netdesc);
+
for (my $i = 0; $i < $MAX_NETS; $i++) {
$confdesc->{"net$i"} = $netdesc;
+ $confdesc_cloudinit->{"ipconfig$i"} = $ipconfigdesc;
+}
+
+foreach my $key (keys %$confdesc_cloudinit) {
+ $confdesc->{$key} = $confdesc_cloudinit->{$key};
}
PVE::JSONSchema::register_format('pve-volume-id-or-qm-path', \&verify_volume_id_or_qm_path);
},
snapshot => {
type => 'boolean',
- description => "Whether the drive should be included when making snapshots.",
+ description => "Controls qemu's snapshot mode feature."
+ . " If activated, changes made to the disk are temporary and will"
+ . " be discarded when the VM is shutdown.",
optional => 1,
},
cache => {
description => "Whether the drive should be included when making backups.",
optional => 1,
},
- replica => {
+ replicate => {
type => 'boolean',
- description => 'Will include this drive to an storage replical job.',
+ description => 'Whether the drive should considered for replication jobs.',
optional => 1,
default => 1,
},
maxLength => 20*3, # *3 since it's %xx url enoded
description => "The drive's reported serial number, url-encoded, up to 20 bytes long.",
optional => 1,
+ },
+ shared => {
+ type => 'boolean',
+ description => 'Mark this locally-managed volume as available on all nodes',
+ verbose_description => "Mark this locally-managed volume as available on all nodes.\n\nWARNING: This option does not share the volume automatically, it assumes it is shared already!",
+ optional => 1,
+ default => 0,
}
);
},
);
+my %ssd_fmt = (
+ ssd => {
+ type => 'boolean',
+ description => "Whether to expose this drive as an SSD, rather than a rotational hard disk.",
+ optional => 1,
+ },
+);
+
+my %wwn_fmt = (
+ wwn => {
+ type => 'string',
+ pattern => qr/^(0x)[0-9a-fA-F]{16}/,
+ format_description => 'wwn',
+ description => "The drive's worldwide name, encoded as 16 bytes hex string, prefixed by '0x'.",
+ optional => 1,
+ },
+);
+
my $add_throttle_desc = sub {
my ($key, $type, $what, $unit, $longunit, $minimum) = @_;
my $d = {
$add_throttle_desc->('iops_wr_max', 'integer', 'unthrottled write I/O pool', 'iops', 'operations per second');
# burst lengths
-$add_throttle_desc->('bps_max_length', 'integer', 'length of I/O bursts', 'seconds', 'seconds', 1);
-$add_throttle_desc->('bps_rd_length', 'integer', 'length of read I/O bursts', 'seconds', 'seconds', 1);
-$add_throttle_desc->('bps_wr_length', 'integer', 'length of write I/O bursts', 'seconds', 'seconds', 1);
-$add_throttle_desc->('iops_max_length', 'integer', 'length of I/O bursts', 'seconds', 'seconds', 1);
-$add_throttle_desc->('iops_rd_length', 'integer', 'length of read I/O bursts', 'seconds', 'seconds', 1);
-$add_throttle_desc->('iops_wr_length', 'integer', 'length of write I/O bursts', 'seconds', 'seconds', 1);
+$add_throttle_desc->('bps_max_length', 'integer', 'length of I/O bursts', 'seconds', 'seconds', 1);
+$add_throttle_desc->('bps_rd_max_length', 'integer', 'length of read I/O bursts', 'seconds', 'seconds', 1);
+$add_throttle_desc->('bps_wr_max_length', 'integer', 'length of write I/O bursts', 'seconds', 'seconds', 1);
+$add_throttle_desc->('iops_max_length', 'integer', 'length of I/O bursts', 'seconds', 'seconds', 1);
+$add_throttle_desc->('iops_rd_max_length', 'integer', 'length of read I/O bursts', 'seconds', 'seconds', 1);
+$add_throttle_desc->('iops_wr_max_length', 'integer', 'length of write I/O bursts', 'seconds', 'seconds', 1);
+
+# legacy support
+$drivedesc_base{'bps_rd_length'} = { alias => 'bps_rd_max_length' };
+$drivedesc_base{'bps_wr_length'} = { alias => 'bps_wr_max_length' };
+$drivedesc_base{'iops_rd_length'} = { alias => 'iops_rd_max_length' };
+$drivedesc_base{'iops_wr_length'} = { alias => 'iops_wr_max_length' };
my $ide_fmt = {
%drivedesc_base,
%model_fmt,
+ %ssd_fmt,
+ %wwn_fmt,
};
PVE::JSONSchema::register_format("pve-qm-ide", $ide_fmt);
%iothread_fmt,
%queues_fmt,
%scsiblock_fmt,
+ %ssd_fmt,
+ %wwn_fmt,
};
my $scsidesc = {
optional => 1,
my $sata_fmt = {
%drivedesc_base,
+ %ssd_fmt,
+ %wwn_fmt,
};
my $satadesc = {
optional => 1,
%model_fmt,
%queues_fmt,
%scsiblock_fmt,
+ %ssd_fmt,
+ %wwn_fmt,
};
my $efidisk_fmt = {
};
PVE::JSONSchema::register_standard_option("pve-qm-usb", $usbdesc);
-# NOTE: the match-groups of this regex are used in parse_hostpci
-my $PCIRE = qr/([a-f0-9]{2}:[a-f0-9]{2})(?:\.([a-f0-9]))?/;
+my $PCIRE = qr/[a-f0-9]{2}:[a-f0-9]{2}(?:\.[a-f0-9])?/;
my $hostpci_fmt = {
host => {
default_key => 1,
pattern => qr/$PCIRE(;$PCIRE)*/,
format_description => 'HOSTPCIID[;HOSTPCIID2...]',
description => <<EODESCR,
-Host PCI device pass through. The PCI ID of a host's PCI device or a list
+Host PCI device pass through. The PCI ID of a host's PCI device or a list
of PCI virtual functions of the host. HOSTPCIID syntax is:
'bus:dev.func' (hexadecimal numbers)
optional => 1,
default => 0,
},
+ 'mdev' => {
+ type => 'string',
+ format_description => 'string',
+ pattern => '[^/\.:]+',
+ optional => 1,
+ description => <<EODESCR
+The type of mediated device to use.
+An instance of this type will be created on startup of the VM and
+will be cleaned up when the VM stops.
+EODESCR
+ }
};
PVE::JSONSchema::register_format('pve-qm-hostpci', $hostpci_fmt);
verbose_description => <<EODESCR,
Map host PCI devices into guest.
-NOTE: This option allows direct access to host hardware. So it is no longer
+NOTE: This option allows direct access to host hardware. So it is no longer
possible to migrate such machines - use with special care.
CAUTION: Experimental! User reported problems with this option.
my $kvm_api_version = 0;
sub kvm_version {
-
return $kvm_api_version if $kvm_api_version;
- my $fh = IO::File->new("</dev/kvm") ||
- return 0;
-
- if (my $v = $fh->ioctl(KVM_GET_API_VERSION(), 0)) {
- $kvm_api_version = $v;
- }
+ open my $fh, '<', '/dev/kvm'
+ or return undef;
- $fh->close();
+ # 0xae00 => KVM_GET_API_VERSION
+ $kvm_api_version = ioctl($fh, 0xae00, 0);
- return $kvm_api_version;
+ return $kvm_api_version;
}
my $kvm_user_version;
}
-my $kernel_has_vhost_net = -c '/dev/vhost-net';
+sub kernel_has_vhost_net {
+ return -c '/dev/vhost-net';
+}
sub valid_drive_names {
# order is important - used to autoselect boot disk
sub filename_to_volume_id {
my ($vmid, $file, $media) = @_;
- if (!($file eq 'none' || $file eq 'cdrom' ||
+ if (!($file eq 'none' || $file eq 'cdrom' ||
$file =~ m|^/dev/.+| || $file =~ m/^([^:]+):(.+)$/)) {
return undef if $file =~ m|/|;
# can't use the schema's 'requires' because of the mbps* => bps* "transforming aliases"
for my $requirement (
+ [mbps_max => 'mbps'],
+ [mbps_rd_max => 'mbps_rd'],
+ [mbps_wr_max => 'mbps_wr'],
+ [miops_max => 'miops'],
+ [miops_rd_max => 'miops_rd'],
+ [miops_wr_max => 'miops_wr'],
[bps_max_length => 'mbps_max'],
[bps_rd_max_length => 'mbps_rd_max'],
[bps_wr_max_length => 'mbps_wr_max'],
}
sub print_tabletdevice_full {
- my ($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') {
$device = "scsi-$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";
+ }
+ $device .= ",wwn=$drive->{wwn}" if $drive->{wwn};
+
+ } elsif ($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}";
+ $device .= ",wwn=$drive->{wwn}" if $drive->{wwn};
} elsif ($drive->{interface} eq 'usb') {
die "implement me";
# -device ide-drive,bus=ide.1,unit=0,drive=drive-ide0-1-0,id=ide0-1-0
$device .= ",bootindex=$drive->{bootindex}" if $drive->{bootindex};
+ if (my $serial = $drive->{serial}) {
+ $serial = URI::Escape::uri_unescape($serial);
+ $device .= ",serial=$serial";
+ }
+
+
return $device;
}
my $path;
my $volid = $drive->{file};
my $format;
-
+
if (drive_is_cdrom($drive)) {
$path = get_iso_path($storecfg, $vmid, $volid);
} else {
}
my $opts = '';
- my @qemu_drive_options = qw(heads secs cyls trans media format cache snapshot rerror werror aio discard iops iops_rd iops_wr iops_max iops_rd_max iops_wr_max);
+ my @qemu_drive_options = qw(heads secs cyls trans media format cache rerror werror aio discard);
foreach my $o (@qemu_drive_options) {
- $opts .= ",$o=$drive->{$o}" if $drive->{$o};
- }
- if (my $serial = $drive->{serial}) {
- $serial = URI::Escape::uri_unescape($serial);
- $opts .= ",serial=$serial";
+ $opts .= ",$o=$drive->{$o}" if defined($drive->{$o});
}
- $opts .= ",format=$format" if $format && !$drive->{format};
+ # snapshot only accepts on|off
+ if (defined($drive->{snapshot})) {
+ my $v = $drive->{snapshot} ? 'on' : 'off';
+ $opts .= ",snapshot=$v";
+ }
- foreach my $o (qw(bps bps_rd bps_wr)) {
- my $v = $drive->{"m$o"};
- $opts .= ",$o=" . int($v*1024*1024) if $v;
+ foreach my $type (['', '-total'], [_rd => '-read'], [_wr => '-write']) {
+ my ($dir, $qmpname) = @$type;
+ if (my $v = $drive->{"mbps$dir"}) {
+ $opts .= ",throttling.bps$qmpname=".int($v*1024*1024);
+ }
+ if (my $v = $drive->{"mbps${dir}_max"}) {
+ $opts .= ",throttling.bps$qmpname-max=".int($v*1024*1024);
+ }
+ if (my $v = $drive->{"bps${dir}_max_length"}) {
+ $opts .= ",throttling.bps$qmpname-max-length=$v";
+ }
+ if (my $v = $drive->{"iops${dir}"}) {
+ $opts .= ",throttling.iops$qmpname=$v";
+ }
+ if (my $v = $drive->{"iops${dir}_max"}) {
+ $opts .= ",throttling.iops$qmpname-max=$v";
+ }
+ if (my $v = $drive->{"iops${dir}_max_length"}) {
+ $opts .= ",throttling.iops$qmpname-max-length=$v";
+ }
}
+ $opts .= ",format=$format" if $format && !$drive->{format};
+
my $cache_direct = 0;
if (my $cache = $drive->{cache}) {
}
sub print_netdevice_full {
- my ($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};
$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 && $net->{model} eq 'virtio'){
#Consider we have N queues, the number of vectors needed is 2*N + 2 (plus one config interrupt and control vq)
}
sub print_netdev_full {
- my ($vmid, $conf, $net, $netid, $hotplug) = @_;
+ my ($vmid, $conf, $arch, $net, $netid, $hotplug) = @_;
my $i = '';
if ($netid =~ m/^net(\d+)$/) {
if length($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";
sub print_cpu_device {
my ($conf, $id) = @_;
- my $nokvm = defined($conf->{kvm}) && $conf->{kvm} == 0 ? 1 : 0;
- my $cpu = $nokvm ? "qemu64" : "kvm64";
+ my $kvm = $conf->{kvm} // 1;
+ 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};
}
- my $sockets = 1;
- $sockets = $conf->{sockets} if $conf->{sockets};
my $cores = $conf->{cores} || 1;
my $current_core = ($id - 1) % $cores;
- my $current_socket = int(($id - $current_core)/$cores);
+ my $current_socket = int(($id - 1 - $current_core)/$cores);
return "$cpu-x86_64-cpu,id=cpu$id,socket-id=$current_socket,core-id=$current_core,thread-id=0";
}
-sub drive_is_cdrom {
+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 ($arch eq 'aarch64' && defined($type) && $type eq 'virtio-vga') {
+ $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)?$@;
+}
+
+sub drive_is_cdrom {
+ my ($drive, $exclude_cloudinit) = @_;
+
+ return 0 if $exclude_cloudinit && drive_is_cloudinit($drive);
return $drive && $drive->{media} && ($drive->{media} eq 'cdrom');
my @idlist = split(/;/, $res->{host});
delete $res->{host};
foreach my $id (@idlist) {
- if ($id =~ /^$PCIRE$/) {
- if (defined($2)) {
- push @{$res->{pciid}}, { id => $1, function => $2 };
- } else {
- my $pcidevices = lspci($1);
- $res->{pciid} = $pcidevices->{$1};
- }
- } else {
- # should have been caught by parse_property_string already
- die "failed to parse PCI id: $id\n";
+ if ($id =~ m/\./) { # full id 00:00.1
+ push @{$res->{pciid}}, {
+ id => $id,
+ };
+ } else { # partial id 00:00
+ $res->{pciid} = PVE::SysFSTools::lspci($id);
}
}
return $res;
return $res;
}
+# ipconfigX ip=cidr,gw=ip,ip6=cidr,gw6=ip
+sub parse_ipconfig {
+ my ($data) = @_;
+
+ my $res = eval { PVE::JSONSchema::parse_property_string($ipconfig_fmt, $data) };
+ if ($@) {
+ warn $@;
+ return undef;
+ }
+
+ if ($res->{gw} && !$res->{ip}) {
+ warn 'gateway specified without specifying an IP address';
+ return undef;
+ }
+ if ($res->{gw6} && !$res->{ip6}) {
+ warn 'IPv6 gateway specified without specifying an IPv6 address';
+ return undef;
+ }
+ if ($res->{gw} && $res->{ip} eq 'dhcp') {
+ warn 'gateway specified together with DHCP';
+ return undef;
+ }
+ if ($res->{gw6} && $res->{ip6} !~ /^$IPV6RE/) {
+ # gw6 + auto/dhcp
+ warn "IPv6 gateway specified together with $res->{ip6} address";
+ return undef;
+ }
+
+ if (!$res->{ip} && !$res->{ip6}) {
+ return { ip => 'dhcp', ip6 => 'dhcp' };
+ }
+
+ return $res;
+}
+
sub print_net {
my $net = shift;
sub vmconfig_register_unused_drive {
my ($storecfg, $vmid, $conf, $drive) = @_;
- if (!drive_is_cdrom($drive)) {
+ if (drive_is_cloudinit($drive)) {
+ eval { PVE::Storage::vdisk_free($storecfg, $drive->{file}) };
+ warn $@ if $@;
+ } elsif (!drive_is_cdrom($drive)) {
my $volid = $drive->{file};
if (vm_is_volid_owner($storecfg, $vmid, $volid)) {
PVE::QemuConfig->add_unused_volume($conf, $volid, $vmid);
return $res;
}
-PVE::JSONSchema::register_format('pve-qm-usb-device', \&verify_usb_device);
-sub verify_usb_device {
- my ($value, $noerr) = @_;
+sub parse_guest_agent {
+ my ($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) = @_;
return $value if parse_usb_device($value);
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};
}
return $prop;
}
+# return copy of $confdesc_cloudinit to generate documentation
+sub cloudinit_config_properties {
+
+ return dclone($confdesc_cloudinit);
+}
+
sub check_type {
my ($key, $value) = @_;
}
}
-sub check_iommu_support{
- #fixme : need to check IOMMU support
- #http://www.linux-kvm.org/page/How_to_assign_devices_with_VT-d_in_KVM
-
- my $iommu=1;
- return $iommu;
-
-}
-
sub touch_config {
my ($vmid) = @_;
PVE::QemuConfig->check_lock($conf) if !$skiplock;
+ if ($conf->{template}) {
+ # check if any base image is still used by a linked clone
+ foreach_drive($conf, sub {
+ my ($ds, $drive) = @_;
+
+ return if drive_is_cdrom($drive);
+
+ my $volid = $drive->{file};
+
+ return if !$volid || $volid =~ m|^/|;
+
+ die "base volume '$volid' is still in use by linked cloned\n"
+ if PVE::Storage::volume_is_base_and_used($storecfg, $volid);
+
+ });
+ }
+
# only remove disks owned by this VM
foreach_drive($conf, sub {
my ($ds, $drive) = @_;
- return if drive_is_cdrom($drive);
+ return if drive_is_cdrom($drive, 1);
my $volid = $drive->{file};
} else {
warn "vm $vmid - propertry 'delete' is only allowed in [PENDING]\n";
}
- } elsif ($line =~ m/^([a-z][a-z_]*\d*):\s*(\S+)\s*$/) {
+ } elsif ($line =~ m/^([a-z][a-z_]*\d*):\s*(.+?)\s*$/) {
my $key = $1;
my $value = $2;
eval { $value = check_type($key, $value); };
}
}
- my $conf = PVE::Cluster::cfs_read_file('datacenter.cfg');
- $res->{keyboard} = $conf->{keyboard} if $conf->{keyboard};
-
return $res;
}
sub check_local_resources {
my ($conf, $noerr) = @_;
- my $loc_res = 0;
+ my @loc_res = ();
+
+ push @loc_res, "hostusb" if $conf->{hostusb}; # old syntax
+ push @loc_res, "hostpci" if $conf->{hostpci}; # old syntax
- $loc_res = 1 if $conf->{hostusb}; # old syntax
- $loc_res = 1 if $conf->{hostpci}; # old syntax
+ push @loc_res, "ivshmem" if $conf->{ivshmem};
foreach my $k (keys %$conf) {
next if $k =~ m/^usb/ && ($conf->{$k} eq 'spice');
# sockets are safe: they will recreated be on the target side post-migrate
next if $k =~ m/^serial/ && ($conf->{$k} eq 'socket');
- $loc_res = 1 if $k =~ m/^(usb|hostpci|serial|parallel)\d+$/;
+ push @loc_res, $k if $k =~ m/^(usb|hostpci|serial|parallel)\d+$/;
}
- die "VM uses local resources\n" if $loc_res && !$noerr;
+ die "VM uses local resources\n" if scalar @loc_res && !$noerr;
- return $loc_res;
+ return \@loc_res;
}
# check if used storages are available on all nodes (use by migrate)
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];
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,
+ },
+ lock => {
+ description => "The current config lock, if any.",
+ type => 'string',
+ optional => 1,
+ }
+};
+
my $last_proc_pid_stat;
# get VM status information
my $storecfg = PVE::Storage::config();
my $list = vzlist();
+ my $defaults = load_defaults();
+
my ($uptime) = PVE::ProcFSTools::read_proc_uptime(1);
my $cpucount = $cpuinfo->{cpus} || 1;
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?
$d->{maxdisk} = 0;
}
- $d->{cpus} = ($conf->{sockets} || 1) * ($conf->{cores} || 1);
+ $d->{cpus} = ($conf->{sockets} || $defaults->{sockets})
+ * ($conf->{cores} || $defaults->{cores});
$d->{cpus} = $cpucount if $d->{cpus} > $cpucount;
$d->{cpus} = $conf->{vcpus} if $conf->{vcpus};
$d->{name} = $conf->{name} || "VM $vmid";
- $d->{maxmem} = $conf->{memory} ? $conf->{memory}*(1024*1024) : 0;
+ $d->{maxmem} = $conf->{memory} ? $conf->{memory}*(1024*1024)
+ : $defaults->{memory}*(1024*1024);
if ($conf->{balloon}) {
$d->{balloon_min} = $conf->{balloon}*(1024*1024);
- $d->{shares} = defined($conf->{shares}) ? $conf->{shares} : 1000;
+ $d->{shares} = defined($conf->{shares}) ? $conf->{shares}
+ : $defaults->{shares};
}
$d->{uptime} = 0;
$d->{template} = PVE::QemuConfig->is_template($conf);
+ $d->{serial} = 1 if conf_has_serial($conf);
+ $d->{lock} = $conf->{lock} if $conf->{lock};
+
$res->{$vmid} = $d;
}
my $volhash = {};
my $test_volid = sub {
- my ($volid, $is_cdrom) = @_;
+ my ($volid, $is_cdrom, $replicate, $shared, $snapname) = @_;
return if !$volid;
- $volhash->{$volid} = $is_cdrom || 0;
+ $volhash->{$volid}->{cdrom} //= 1;
+ $volhash->{$volid}->{cdrom} = 0 if !$is_cdrom;
+
+ $volhash->{$volid}->{replicate} //= 0;
+ $volhash->{$volid}->{replicate} = 1 if $replicate;
+
+ $volhash->{$volid}->{shared} //= 0;
+ $volhash->{$volid}->{shared} = 1 if $shared;
+
+ $volhash->{$volid}->{referenced_in_config} //= 0;
+ $volhash->{$volid}->{referenced_in_config} = 1 if !defined($snapname);
+
+ $volhash->{$volid}->{referenced_in_snapshot}->{$snapname} = 1
+ if defined($snapname);
};
foreach_drive($conf, sub {
my ($ds, $drive) = @_;
- &$test_volid($drive->{file}, drive_is_cdrom($drive));
+ $test_volid->($drive->{file}, drive_is_cdrom($drive), $drive->{replicate} // 1, $drive->{shared}, undef);
});
foreach my $snapname (keys %{$conf->{snapshots}}) {
my $snap = $conf->{snapshots}->{$snapname};
- &$test_volid($snap->{vmstate}, 0);
+ $test_volid->($snap->{vmstate}, 0, 1, $snapname);
foreach_drive($snap, sub {
my ($ds, $drive) = @_;
- &$test_volid($drive->{file}, drive_is_cdrom($drive));
+ $test_volid->($drive->{file}, drive_is_cdrom($drive), $drive->{replicate} // 1, $drive->{shared}, $snapname);
});
}
}
}
+sub conf_has_serial {
+ my ($conf) = @_;
+
+ for (my $i = 0; $i < $MAX_SERIAL_PORTS; $i++) {
+ if ($conf->{"serial$i"}) {
+ return 1;
+ }
+ }
+
+ return 0;
+}
+
sub vga_conf_has_spice {
my ($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';
+ }
+ my $hv_vendor_id;
+ 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};
+ $hv_vendor_id = $cpuconf->{'hv-vendor-id'};
+
+ 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, $hv_vendor_id) 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) = @_;
my $globalFlags = [];
my $machineFlags = [];
my $rtcFlags = [];
- my $cpuFlags = [];
my $devices = [];
my $pciaddr = '';
my $bridges = {};
my $vernum = 0; # unknown
my $ostype = $conf->{ostype};
my $winversion = windows_version($ostype);
+ my $kvm = $conf->{kvm};
+
+ my ($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;
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;
+ my $vmname = $conf->{name} || "vm$vmid";
+
+ push @$cmd, '-name', $vmname;
+
my $use_virtio = 0;
my $qmpsocket = qmp_socket($vmid);
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($vmid);
push @$cmd, '-smbios', "type=1,$conf->{smbios1}";
}
- if ($conf->{bios} && $conf->{bios} eq 'ovmf') {
- my $ovmfbase;
-
- # prefer the OVMF_CODE variant
- if (-f $OVMF_CODE) {
- $ovmfbase = $OVMF_CODE;
- } elsif (-f $OVMF_IMG) {
- $ovmfbase = $OVMF_IMG;
- }
+ if ($conf->{vmgenid}) {
+ push @$devices, '-device', 'vmgenid,guid='.$conf->{vmgenid};
+ }
- die "no uefi base img found\n" if !$ovmfbase;
- push @$cmd, '-drive', "if=pflash,unit=0,format=raw,readonly,file=$ovmfbase";
+ my ($ovmf_code, $ovmf_vars) = get_ovmf_files($arch);
+ if ($conf->{bios} && $conf->{bios} eq 'ovmf') {
+ die "uefi base image not found\n" if ! -f $ovmf_code;
- if (defined($conf->{efidisk0}) && ($ovmfbase eq $OVMF_CODE)) {
- my $d = PVE::JSONSchema::parse_property_string($efidisk_fmt, $conf->{efidisk0});
- my $format = $d->{format} // 'raw';
- my $path;
+ my $path;
+ my $format;
+ if (my $efidisk = $conf->{efidisk0}) {
+ my $d = PVE::JSONSchema::parse_property_string($efidisk_fmt, $efidisk);
my ($storeid, $volname) = PVE::Storage::parse_volume_id($d->{file}, 1);
+ $format = $d->{format};
if ($storeid) {
$path = PVE::Storage::path($storecfg, $d->{file});
- my $scfg = PVE::Storage::storage_config($storecfg, $storeid);
- $format = qemu_img_format($scfg, $volname);
+ if (!defined($format)) {
+ my $scfg = PVE::Storage::storage_config($storecfg, $storeid);
+ $format = qemu_img_format($scfg, $volname);
+ }
} else {
$path = $d->{file};
- $format = "raw";
+ die "efidisk format must be specified\n"
+ if !defined($format);
}
- push @$cmd, '-drive', "if=pflash,unit=1,id=drive-efidisk0,format=$format,file=$path";
- } elsif ($ovmfbase eq $OVMF_CODE) {
- warn "using uefi without permanent efivars disk\n";
- my $ovmfvar_dst = "/tmp/$vmid-ovmf.fd";
- PVE::Tools::file_copy($OVMF_VARS, $ovmfvar_dst, 256*1024);
- push @$cmd, '-drive', "if=pflash,unit=1,format=raw,file=$ovmfvar_dst";
} else {
- # if the base img is not OVMF_CODE, we do not have to bother
- # to create/use a vars image, since it will not be used anyway
- # this can only happen if someone manually deletes the OVMF_CODE image
- # or has an old pve-qemu-kvm version installed.
- # both should not happen, but we ignore it here
+ warn "no efidisk configured! Using temporary efivars disk.\n";
+ $path = "/tmp/$vmid-ovmf.fd";
+ 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=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) {
- $vga = $winversion >= 6 ? 'std' : 'cirrus';
+ if (!$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 {
+ $vga->{type} = ($winversion >= 6) ? 'std' : 'cirrus';
+ }
}
# enable absolute mouse coordinates (needed by vnc)
} 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 $gpu_passthrough;
next if !$d;
my $pcie = $d->{pcie};
- if($pcie){
+ if ($pcie) {
die "q35 machine model is not enabled" if !$q35;
- $pciaddr = print_pcie_addr("hostpci$i");
- }else{
- $pciaddr = print_pci_addr("hostpci$i", $bridges);
+ # win7 wants to have the pcie devices directly on the pcie bus
+ # instead of in the root port
+ if ($winversion == 7) {
+ $pciaddr = print_pcie_addr("hostpci${i}bus0");
+ } else {
+ $pciaddr = print_pcie_addr("hostpci$i");
+ }
+ } else {
+ $pciaddr = print_pci_addr("hostpci$i", $bridges, $arch, $machine_type);
}
my $rombar = defined($d->{rombar}) && !$d->{rombar} ? ',rombar=0' : '';
if ($d->{'x-vga'}) {
$xvga = ',x-vga=on';
$kvm_off = 1;
- $vga = 'none';
+ $vga->{type} = 'none' if !defined($conf->{vga});
$gpu_passthrough = 1;
if ($conf->{bios} && $conf->{bios} eq 'ovmf') {
}
my $pcidevices = $d->{pciid};
my $multifunction = 1 if @$pcidevices > 1;
+ my $sysfspath;
+ if ($d->{mdev} && scalar(@$pcidevices) == 1) {
+ my $id = $pcidevices->[0]->{id};
+ my $uuid = PVE::SysFSTools::generate_mdev_uuid($vmid, $i);
+ $sysfspath = "/sys/bus/pci/devices/0000:$id/$uuid";
+ } elsif ($d->{mdev}) {
+ warn "ignoring mediated device with multifunction device\n";
+ }
my $j=0;
foreach my $pcidevice (@$pcidevices) {
$id .= ".$j" if $multifunction;
my $addr = $pciaddr;
$addr .= ".$j" if $multifunction;
- my $devicestr = "vfio-pci,host=$pcidevice->{id}.$pcidevice->{function},id=$id$addr";
+ my $devicestr = "vfio-pci";
+ if ($sysfspath) {
+ $devicestr .= ",sysfsdev=$sysfspath";
+ } else {
+ $devicestr .= ",host=$pcidevice->{id}";
+ }
+ $devicestr .= ",id=$id$addr";
if($j == 0){
$devicestr .= "$rombar$xvga";
if ($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";
}
}
- my $vmname = $conf->{name} || "vm$vmid";
-
- push @$cmd, '-name', $vmname;
my $sockets = 1;
$sockets = $conf->{smp} if $conf->{smp}; # old style - no longer iused
push @$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';
}
# time drift fix
my $tdf = defined($conf->{tdf}) ? $conf->{tdf} : $defaults->{tdf};
- my $nokvm = defined($conf->{kvm}) && $conf->{kvm} == 0 ? 1 : 0;
my $useLocaltime = $conf->{localtime};
if ($winversion >= 5) { # windows
push @$rtcFlags, 'driftfix=slew' if $tdf;
- if ($nokvm) {
+ if (!$kvm) {
push @$machineFlags, 'accel=tcg';
- } else {
- die "No accelerator found!\n" if !$cpuinfo->{hvm};
}
if ($machine_type) {
push @$rtcFlags, 'base=localtime';
}
- my $cpu = $nokvm ? "qemu64" : "kvm64";
- 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 $cpuconf->{hidden};
- }
-
- push @$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 !$nokvm;
- push @$cpuFlags , '+kvm_pv_eoi' if !$nokvm;
- }
-
- add_hyperv_enlighments($cpuFlags, $winversion, $machine_type, $kvmver, $nokvm, $conf->{bios}, $gpu_passthrough);
-
- push @$cpuFlags, 'enforce' if $cpu ne 'host' && !$nokvm;
-
- 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, $cmd);
-
+
push @$cmd, '-S' if $conf->{freeze};
- # set keyboard layout
- my $kb = $conf->{keyboard} || $defaults->{keyboard};
- push @$cmd, '-k', $kb if $kb;
+ push @$cmd, '-k', $conf->{keyboard} if defined($conf->{keyboard});
# enable sound
#my $soundhw = $conf->{soundhw} || $defaults->{soundhw};
#push @$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';
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);
- $spice_port = PVE::Tools::next_spice_port($pfamily);
+ my @nodeaddrs = PVE::Tools::getaddrinfo_all('localhost', family => $pfamily);
+ die "failed to get an ip address of type $pfamily for 'localhost'\n" if !@nodeaddrs;
+ my $localhost = PVE::Network::addr_to_ip($nodeaddrs[0]->{addr});
+ $spice_port = PVE::Tools::next_spice_port($pfamily, $localhost);
- push @$devices, '-spice', "tls-port=${spice_port},addr=localhost,tls-ciphers=HIGH,seamless-migration=on";
+ push @$devices, '-spice', "tls-port=${spice_port},addr=$localhost,tls-ciphers=HIGH,seamless-migration=on";
push @$devices, '-device', "virtio-serial,id=spice$pciaddr";
push @$devices, '-chardev', "spicevmc,id=vdagent,name=vdagent";
# enable balloon by default, unless explicitly disabled
if (!defined($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 $wdopts->{action};
push @$vollist, $drive->{file};
}
+ # ignore efidisk here, already added in bios/fw handling code above
+ return if $drive->{interface} eq 'efidisk';
+
$use_virtio = 1 if $ds =~ m/^virtio/;
if (drive_is_cdrom ($drive)) {
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 = '';
my $queues = '';
if($conf->{scsihw} && $conf->{scsihw} eq "virtio-scsi-single" && $drive->{queues}){
$queues = ",num_queues=$drive->{queues}";
- }
+ }
push @$devices, '-device', "$scsihw_type,id=$controller_prefix$controller$pciaddr$iothread$queues" if !$scsicontroller->{$controller};
$scsicontroller->{$controller}=1;
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;
}
- if ($drive->{interface} eq 'efidisk') {
- # this will be added somewhere else
- return;
- }
-
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++) {
$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', $netdevicefull;
}
+ if ($conf->{ivshmem}) {
+ my $ivshmem = PVE::JSONSchema::parse_property_string($ivshmem_fmt, $conf->{ivshmem});
+
+ my $bus;
+ if ($q35) {
+ $bus = print_pcie_addr("ivshmem");
+ } else {
+ $bus = print_pci_addr("ivshmem", $bridges, $arch, $machine_type);
+ }
+
+ my $ivshmem_name = $ivshmem->{name} // $vmid;
+ my $path = '/dev/shm/pve-shm-' . $ivshmem_name;
+
+ push @$devices, '-device', "ivshmem-plain,memdev=ivshmem$bus,";
+ push @$devices, '-object', "memory-backend-file,id=ivshmem,share=on,mem-path=$path,size=$ivshmem->{size}M";
+ }
+
if (!$q35) {
# add pci bridges
if (qemu_machine_feature_enabled ($machine_type, $kvmver, 2, 3)) {
$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 $k > 0;
}
}
- # add custom args
- if ($conf->{args}) {
- my $aa = PVE::Tools::split_args($conf->{args});
- push @$cmd, @$aa;
- }
-
push @$cmd, @$devices;
push @$cmd, '-rtc', join(',', @$rtcFlags)
if scalar(@$rtcFlags);
push @$cmd, '-global', join(',', @$globalFlags)
if scalar(@$globalFlags);
+ if (my $vmstate = $conf->{vmstate}) {
+ my $statepath = PVE::Storage::path($storecfg, $vmstate);
+ PVE::Storage::activate_volumes($storecfg, [$vmstate]);
+ push @$cmd, '-loadstate', $statepath;
+ }
+
+ # add custom args
+ if ($conf->{args}) {
+ my $aa = PVE::Tools::split_args($conf->{args});
+ push @$cmd, @$aa;
+ }
+
return wantarray ? ($cmd, $vollist, $spice_port) : $cmd;
}
}
sub qmp_socket {
- my ($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 {
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');
}
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+)$/) {
qemu_iothread_add($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); };
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";
} 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, $deviceid); };
} elsif ($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 $machine_type = PVE::QemuServer::qemu_machine_pxe($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, $use_old_bios_files);
- qemu_deviceadd($vmid, $netdevicefull);
- eval { qemu_deviceaddverify($vmid, $deviceid); };
+ 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, $deviceid);
+ qemu_set_link_status($vmid, $deviceid, !$device->{link_down});
+ };
if (my $err = $@) {
eval { qemu_netdevdel($vmid, $deviceid); };
warn $@ if $@;
die $err;
- }
+ }
} elsif (!$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);
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, $deviceid);
} elsif ($deviceid =~ m/^(scsi)(\d+)$/) {
- #qemu 2.3 segfault on drive_del with virtioscsi + iothread
- my $device = parse_drive($deviceid, $conf->{$deviceid});
- die "virtioscsi with iothread is not hot-unplugglable currently" if $device->{iothread};
-
qemu_devicedel($vmid, $deviceid);
qemu_drivedel($vmid, $deviceid);
qemu_deletescsihw($conf, $vmid, $deviceid);
sub qemu_iothread_del {
my($conf, $vmid, $deviceid) = @_;
- my $device = parse_drive($deviceid, $conf->{$deviceid});
+ my $confid = $deviceid;
+ if ($deviceid =~ m/^(?:virtioscsi|scsihw)(\d+)$/) {
+ $confid = 'scsi' . $1;
+ }
+ my $device = parse_drive($confid, $conf->{$confid});
if ($device->{iothread}) {
my $iothreads = vm_iothreads_list($vmid);
qemu_objectdel($vmid, "iothread-$deviceid") if $iothreads->{"iothread-$deviceid"};
}
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);
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, $machine_type);
}
return 1;
}
sub qemu_add_pci_bridge {
- my ($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;
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, $machine_type);
}
return 1;
}
sub qemu_netdevadd {
- my ($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);
}
sub qemu_usb_hotplug {
- my ($storecfg, $conf, $vmid, $deviceid, $device) = @_;
+ my ($storecfg, $conf, $vmid, $deviceid, $device, $arch, $machine_type) = @_;
return if !$device;
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");
}
}
$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 {
}
}
- die "monitor read timeout\n" if !scalar(@ready);
-
- return $res;
-}
-
-# old code, only used to shutdown old VM after update
-sub vm_monitor_command {
- my ($vmid, $cmdstr, $nocheck) = @_;
-
- my $res;
-
- eval {
- die "VM $vmid not running\n" if !check_running($vmid, $nocheck);
-
- my $sname = "${var_run_tmpdir}/$vmid.mon";
-
- my $sock = IO::Socket::UNIX->new( Peer => $sname ) ||
- die "unable to connect to VM $vmid socket - $!\n";
-
- my $timeout = 3;
-
- # hack: migrate sometime blocks the monitor (when migrate_downtime
- # is set)
- if ($cmdstr =~ m/^(info\s+migrate|migrate\s)/) {
- $timeout = 60*60; # 1 hour
- }
-
- # read banner;
- my $data = __read_avail($sock, $timeout);
-
- if ($data !~ m/^QEMU\s+(\S+)\s+monitor\s/) {
- die "got unexpected qemu monitor banner\n";
- }
-
- my $sel = new IO::Select;
- $sel->add($sock);
-
- if (!scalar(my @ready = $sel->can_write($timeout))) {
- die "monitor write error - timeout";
- }
-
- my $fullcmd = "$cmdstr\r";
-
- # syslog('info', "VM $vmid monitor command: $cmdstr");
-
- my $b;
- if (!($b = $sock->syswrite($fullcmd)) || ($b != length($fullcmd))) {
- die "monitor write error - $!";
- }
-
- return if ($cmdstr eq 'q') || ($cmdstr eq 'quit');
-
- $timeout = 20;
-
- if ($cmdstr =~ m/^(info\s+migrate|migrate\s)/) {
- $timeout = 60*60; # 1 hour
- } elsif ($cmdstr =~ m/^(eject|change)/) {
- $timeout = 60; # note: cdrom mount command is slow
- }
- if ($res = __read_avail($sock, $timeout)) {
-
- my @lines = split("\r?\n", $res);
-
- shift @lines if $lines[0] !~ m/^unknown command/; # skip echo
-
- $res = join("\n", @lines);
- $res .= "\n";
- }
- };
-
- my $err = $@;
-
- if ($err) {
- syslog("err", "VM $vmid monitor command failed - $err");
- die $err;
- }
-
+ die "monitor read timeout\n" if !scalar(@ready);
+
return $res;
}
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);
}
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, $running);
}
'startup' => 1,
'description' => 1,
'protection' => 1,
+ 'vmstatestorage' => 1,
+ 'hookscript' => 1,
};
# hotplug changes in [PENDING]
my ($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
} 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";
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+)$/) {
}
}
+ my $apply_pending_cloudinit;
+ $apply_pending_cloudinit = sub {
+ my ($key, $value) = @_;
+ $apply_pending_cloudinit = sub {}; # once is enough
+
+ my @cloudinit_opts = keys %$confdesc_cloudinit;
+ foreach my $opt (keys %{$conf->{pending}}) {
+ next if !grep { $_ eq $opt } @cloudinit_opts;
+ $conf->{$opt} = delete $conf->{pending}->{$opt};
+ }
+
+ my $new_conf = { %$conf };
+ $new_conf->{$key} = $value;
+ PVE::QemuServer::Cloudinit::generate_cloudinitconfig($new_conf, $vmid);
+ };
+
foreach my $opt (keys %{$conf->{pending}}) {
next if $selection && !$selection->{$opt};
my $value = $conf->{pending}->{$opt};
} 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+$/) {
die "skip\n";
die "skip\n" if !$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, $value);
} elsif ($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, $value);
+ if (drive_is_cloudinit($drive)) {
+ &$apply_pending_cloudinit($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, $value);
};
sub vmconfig_update_net {
- my ($storecfg, $conf, $hotplug, $vmid, $opt, $value) = @_;
+ my ($storecfg, $conf, $hotplug, $vmid, $opt, $value, $arch, $machine_type) = @_;
my $newnet = parse_net($value);
}
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, $machine_type) = @_;
# fixme: do we need force?
if ($drive->{file} eq 'none') {
vm_mon_cmd($vmid, "eject",force => JSON::true,device => "drive-$opt");
+ if (drive_is_cloudinit($old_drive)) {
+ vmconfig_register_unused_drive($storecfg, $vmid, $conf, $old_drive);
+ }
} else {
my $path = get_iso_path($storecfg, $vmid, $drive->{file});
vm_mon_cmd($vmid, "eject", force => JSON::true,device => "drive-$opt"); # force eject if locked
die "skip\n" if !$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, $machine_type);
}
sub vm_start {
die "you can't start a vm if it's a template\n" if PVE::QemuConfig->is_template($conf);
- PVE::QemuConfig->check_lock($conf) if !$skiplock;
+ my $is_suspended = PVE::QemuConfig->has_lock($conf, 'suspended');
+
+ PVE::QemuConfig->check_lock($conf)
+ if !($skiplock || $is_suspended);
die "VM $vmid already running\n" if check_running($vmid, undef, $migratedfrom);
$conf = PVE::QemuConfig->load_config($vmid); # update/reload
}
+ PVE::QemuServer::Cloudinit::generate_cloudinitconfig($conf, $vmid);
+
my $defaults = load_defaults();
# set environment variable useful inside network script
}
}
+ PVE::GuestHelpers::exec_hookscript($conf, $vmid, 'pre-start', 1);
+
+ if ($is_suspended) {
+ # enforce machine type on suspended vm to ensure HW compatibility
+ $forcemachine = $conf->{runningmachine};
+ print "Resuming suspended VM\n";
+ }
+
my ($cmd, $vollist, $spice_port) = config_to_command($storecfg, $vmid, $conf, $defaults, $forcemachine);
my $migrate_port = 0;
next if !$d;
my $pcidevices = $d->{pciid};
foreach my $pcidevice (@$pcidevices) {
- my $pciid = $pcidevice->{id}.".".$pcidevice->{function};
+ my $pciid = $pcidevice->{id};
- my $info = pci_device_info("0000:$pciid");
- die "IOMMU not present\n" if !check_iommu_support();
+ my $info = PVE::SysFSTools::pci_device_info("0000:$pciid");
+ die "IOMMU not present\n" if !PVE::SysFSTools::check_iommu_support();
die "no pci device info for device '$pciid'\n" if !$info;
- die "can't unbind/bind pci group to vfio '$pciid'\n" if !pci_dev_group_bind_to_vfio($pciid);
- die "can't reset pci device '$pciid'\n" if $info->{has_fl_reset} and !pci_dev_reset($info);
+
+ if ($d->{mdev}) {
+ my $uuid = PVE::SysFSTools::generate_mdev_uuid($vmid, $i);
+ PVE::SysFSTools::pci_create_mdev_device($pciid, $uuid, $d->{mdev});
+ } else {
+ die "can't unbind/bind pci group to vfio '$pciid'\n"
+ if !PVE::SysFSTools::pci_dev_group_bind_to_vfio($pciid);
+ die "can't reset pci device '$pciid'\n"
+ if $info->{has_fl_reset} and !PVE::SysFSTools::pci_dev_reset($info);
+ }
}
}
PVE::Storage::activate_volumes($storecfg, $vollist);
- if (!check_running($vmid, 1) && -d "/sys/fs/cgroup/systemd/qemu.slice/$vmid.scope") {
- my $cmd = [];
- push @$cmd, '/bin/systemctl', 'stop', "$vmid.scope";
- eval { run_command($cmd); };
+ if (-d "/sys/fs/cgroup/systemd/qemu.slice/$vmid.scope") {
+ eval {
+ run_command(['/bin/systemctl', 'stop', "$vmid.scope"],
+ outfunc => sub {}, errfunc => sub {});
+ };
}
my $cpuunits = defined($conf->{cpuunits}) ? $conf->{cpuunits}
: $defaults->{cpuunits};
- my %run_params = (timeout => $statefile ? undef : 30, umask => 0077);
+ my $start_timeout = ($conf->{hugepages} || $is_suspended) ? 300 : 30;
+ my %run_params = (timeout => $statefile ? undef : $start_timeout, umask => 0077);
my %properties = (
Slice => 'qemu.slice',
}
$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 $code = sub {
PVE::QemuServer::Memory::hugepages_mount();
PVE::QemuServer::Memory::hugepages_allocate($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 $err;
eval { PVE::QemuServer::Memory::hugepages_update_locked($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 = $@) {
}
} else {
- if (!$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 $opt !~ m/^net\d+$/;
property => "guest-stats-polling-interval",
value => 2) if (!defined($conf->{balloon}) || $conf->{balloon});
+ if ($is_suspended && (my $vmstate = $conf->{vmstate})) {
+ print "Resumed VM, removing state\n";
+ delete $conf->@{qw(lock vmstate runningmachine)};
+ PVE::Storage::deactivate_volumes($storecfg, [$vmstate]);
+ PVE::Storage::vdisk_free($storecfg, $vmstate);
+ PVE::QemuConfig->write_config($vmid, $conf);
+ }
+
+ PVE::GuestHelpers::exec_hookscript($conf, $vmid, 'post-start');
});
}
my $qmpclient = PVE::QMPClient->new();
$res = $qmpclient->cmd($vmid, $cmd, $timeout);
- } elsif (-e "${var_run_tmpdir}/$vmid.mon") {
- die "can't execute complex command on old monitor - stop/start your vm to fix the problem\n"
- if scalar(%{$cmd->{arguments}});
- vm_monitor_command($vmid, $cmd->{execute}, $nocheck);
} else {
die "unable to open monitor socket\n";
}
}
sub vm_commandline {
- my ($storecfg, $vmid) = @_;
+ my ($storecfg, $vmid, $snapname) = @_;
my $conf = PVE::QemuConfig->load_config($vmid);
+ if ($snapname) {
+ my $snapshot = $conf->{snapshots}->{$snapname};
+ die "snapshot '$snapname' does not exist\n" if !defined($snapshot);
+
+ $snapshot->{digest} = $conf->{digest}; # keep file digest for API
+
+ $conf = $snapshot;
+ }
+
my $defaults = load_defaults();
my $cmd = config_to_command($storecfg, $vmid, $conf, $defaults);
my $vollist = [];
foreach_volid($conf, sub {
- my ($volid, $is_cdrom) = @_;
+ my ($volid, $attr) = @_;
return if $volid =~ m|^/|;
unlink "/var/run/qemu-server/${vmid}.$ext";
}
+ if ($conf->{ivshmem}) {
+ my $ivshmem = PVE::JSONSchema::parse_property_string($ivshmem_fmt, $conf->{ivshmem});
+ # just delete it for now, VMs which have this already open do not
+ # are affected, but new VMs will get a separated one. If this
+ # becomes an issue we either add some sort of ref-counting or just
+ # add a "don't delete on stop" flag to the ivshmem format.
+ unlink '/dev/shm/pve-shm-' . ($ivshmem->{name} // $vmid);
+ }
+
+ foreach my $key (keys %$conf) {
+ next if $key !~ m/^hostpci(\d+)$/;
+ my $hostpciindex = $1;
+ my $d = parse_hostpci($conf->{$key});
+ my $uuid = PVE::SysFSTools::generate_mdev_uuid($vmid, $hostpciindex);
+
+ foreach my $pci (@{$d->{pciid}}) {
+ my $pciid = $pci->{id};
+ PVE::SysFSTools::pci_cleanup_mdev_device($pciid, $uuid);
+ }
+ }
+
vmconfig_apply_pending($vmid, $conf, $storecfg) if $apply_pending_changes;
};
warn $@ if $@; # avoid errors - just warn
my $opts = PVE::JSONSchema::pve_parse_startup_order($conf->{startup});
$timeout = $opts->{down} if $opts->{down};
}
+ PVE::GuestHelpers::exec_hookscript($conf, $vmid, 'pre-stop');
}
$timeout = 60 if !defined($timeout);
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" }, $nocheck);
}
sub vm_suspend {
- my ($vmid, $skiplock) = @_;
+ my ($vmid, $skiplock, $includestate, $statestorage) = @_;
+
+ my $conf;
+ my $path;
+ my $storecfg;
+ my $vmstate;
PVE::QemuConfig->lock_config($vmid, sub {
- my $conf = PVE::QemuConfig->load_config($vmid);
+ $conf = PVE::QemuConfig->load_config($vmid);
+ my $is_backing_up = PVE::QemuConfig->has_lock($conf, 'backup');
PVE::QemuConfig->check_lock($conf)
- if !($skiplock || PVE::QemuConfig->has_lock($conf, 'backup'));
+ if !($skiplock || $is_backing_up);
- vm_mon_cmd($vmid, "stop");
+ die "cannot suspend to disk during backup\n"
+ if $is_backing_up && $includestate;
+
+ if ($includestate) {
+ $conf->{lock} = 'suspending';
+ my $date = strftime("%Y-%m-%d", localtime(time()));
+ $storecfg = PVE::Storage::config();
+ $vmstate = PVE::QemuConfig->__snapshot_save_vmstate($vmid, $conf, "suspend-$date", $storecfg, $statestorage, 1);
+ $path = PVE::Storage::path($storecfg, $vmstate);
+ PVE::QemuConfig->write_config($vmid, $conf);
+ } else {
+ vm_mon_cmd($vmid, "stop");
+ }
});
+
+ if ($includestate) {
+ # save vm state
+ PVE::Storage::activate_volumes($storecfg, [$vmstate]);
+
+ eval {
+ vm_mon_cmd($vmid, "savevm-start", statefile => $path);
+ for(;;) {
+ my $state = vm_mon_cmd_nocheck($vmid, "query-savevm");
+ if (!$state->{status}) {
+ die "savevm not active\n";
+ } elsif ($state->{status} eq 'active') {
+ sleep(1);
+ next;
+ } elsif ($state->{status} eq 'completed') {
+ print "State saved, quitting\n";
+ last;
+ } elsif ($state->{status} eq 'failed' && $state->{error}) {
+ die "query-savevm failed with error '$state->{error}'\n"
+ } else {
+ die "query-savevm returned status '$state->{status}'\n";
+ }
+ }
+ };
+ my $err = $@;
+
+ PVE::QemuConfig->lock_config($vmid, sub {
+ $conf = PVE::QemuConfig->load_config($vmid);
+ if ($err) {
+ # cleanup, but leave suspending lock, to indicate something went wrong
+ eval {
+ vm_mon_cmd($vmid, "savevm-end");
+ PVE::Storage::deactivate_volumes($storecfg, [$vmstate]);
+ PVE::Storage::vdisk_free($storecfg, $vmstate);
+ delete $conf->@{qw(vmstate runningmachine)};
+ PVE::QemuConfig->write_config($vmid, $conf);
+ };
+ warn $@ if $@;
+ die $err;
+ }
+
+ die "lock changed unexpectedly\n"
+ if !PVE::QemuConfig->has_lock($conf, 'suspending');
+
+ vm_qmp_command($vmid, { execute => "quit" });
+ $conf->{lock} = 'suspended';
+ PVE::QemuConfig->write_config($vmid, $conf);
+ });
+ }
}
sub vm_resume {
PVE::QemuConfig->lock_config($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);
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, $resume_cmd);
}
});
}
});
}
-# pci helpers
-
-sub file_write {
- my ($filename, $buf) = @_;
-
- my $fh = IO::File->new($filename, "w");
- return undef if !$fh;
-
- my $res = print $fh $buf;
-
- $fh->close();
-
- return $res;
-}
-
-sub pci_device_info {
- my ($name) = @_;
-
- my $res;
-
- return undef if $name !~ m/^([a-f0-9]{4}):([a-f0-9]{2}):([a-f0-9]{2})\.([a-f0-9])$/;
- my ($domain, $bus, $slot, $func) = ($1, $2, $3, $4);
-
- my $irq = file_read_firstline("$pcisysfs/devices/$name/irq");
- return undef if !defined($irq) || $irq !~ m/^\d+$/;
-
- my $vendor = file_read_firstline("$pcisysfs/devices/$name/vendor");
- return undef if !defined($vendor) || $vendor !~ s/^0x//;
-
- my $product = file_read_firstline("$pcisysfs/devices/$name/device");
- return undef if !defined($product) || $product !~ s/^0x//;
-
- $res = {
- name => $name,
- vendor => $vendor,
- product => $product,
- domain => $domain,
- bus => $bus,
- slot => $slot,
- func => $func,
- irq => $irq,
- has_fl_reset => -f "$pcisysfs/devices/$name/reset" || 0,
- };
-
- return $res;
-}
-
-sub pci_dev_reset {
- my ($dev) = @_;
-
- my $name = $dev->{name};
-
- my $fn = "$pcisysfs/devices/$name/reset";
-
- return file_write($fn, "1");
-}
-
-sub pci_dev_bind_to_vfio {
- my ($dev) = @_;
-
- my $name = $dev->{name};
-
- my $vfio_basedir = "$pcisysfs/drivers/vfio-pci";
-
- if (!-d $vfio_basedir) {
- system("/sbin/modprobe vfio-pci >/dev/null 2>/dev/null");
- }
- die "Cannot find vfio-pci module!\n" if !-d $vfio_basedir;
-
- my $testdir = "$vfio_basedir/$name";
- return 1 if -d $testdir;
-
- my $data = "$dev->{vendor} $dev->{product}";
- return undef if !file_write("$vfio_basedir/new_id", $data);
-
- my $fn = "$pcisysfs/devices/$name/driver/unbind";
- if (!file_write($fn, $name)) {
- return undef if -f $fn;
- }
-
- $fn = "$vfio_basedir/bind";
- if (! -d $testdir) {
- return undef if !file_write($fn, $name);
- }
-
- return -d $testdir;
-}
-
-sub pci_dev_group_bind_to_vfio {
- my ($pciid) = @_;
-
- my $vfio_basedir = "$pcisysfs/drivers/vfio-pci";
-
- if (!-d $vfio_basedir) {
- system("/sbin/modprobe vfio-pci >/dev/null 2>/dev/null");
- }
- die "Cannot find vfio-pci module!\n" if !-d $vfio_basedir;
-
- # get IOMMU group devices
- opendir(my $D, "$pcisysfs/devices/0000:$pciid/iommu_group/devices/") || die "Cannot open iommu_group: $!\n";
- my @devs = grep /^0000:/, readdir($D);
- closedir($D);
-
- foreach my $pciid (@devs) {
- $pciid =~ m/^([:\.\da-f]+)$/ or die "PCI ID $pciid not valid!\n";
-
- # pci bridges, switches or root ports are not supported
- # they have a pci_bus subdirectory so skip them
- next if (-e "$pcisysfs/devices/$pciid/pci_bus");
-
- my $info = pci_device_info($1);
- pci_dev_bind_to_vfio($info) || die "Cannot bind $pciid to vfio\n";
- }
-
- return 1;
-}
-
# vzdump restore implementaion
sub tar_archive_read_firstfile {
return if $line =~ m/^lock:/;
return if $line =~ m/^unused\d+:/;
return if $line =~ m/^parent:/;
- return if $line =~ m/^template:/; # restored VM is never a template
my $dc = PVE::Cluster::cfs_read_file('datacenter.cfg');
if (($line =~ m/^(vlan(\d+)):\s*(\S+)\s*$/)) {
} 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);
+ UUID::unparse($uuid, $uuid_str);
+ my $smbios1 = parse_smbios1($2);
+ $smbios1->{uuid} = $uuid_str;
+ print $outfd $1.print_smbios1($smbios1)."\n";
} else {
print $outfd $line;
}
my ($vmid, $conf, $volid_hash) = @_;
my $changes;
+ my $prefix = "VM $vmid:";
- my $used = {};
+ # used and unused disks
+ my $referenced = {};
# Note: it is allowed to define multiple storages with same path (alias), so
# we need to check both 'volid' and real 'path' (two different volid can point
# to the same path).
- my $usedpath = {};
+ my $referencedpath = {};
# update size info
foreach my $opt (keys %$conf) {
my $volid = $drive->{file};
next if !$volid;
- $used->{$volid} = 1;
+ $referenced->{$volid} = 1;
if ($volid_hash->{$volid} &&
(my $path = $volid_hash->{$volid}->{path})) {
- $usedpath->{$path} = 1;
+ $referencedpath->{$path} = 1;
}
next if drive_is_cdrom($drive);
if ($new ne $conf->{$opt}) {
$changes = 1;
$conf->{$opt} = $new;
+ print "$prefix update disk '$opt' information.\n";
}
}
}
next if $opt !~ m/^unused\d+$/;
my $volid = $conf->{$opt};
my $path = $volid_hash->{$volid}->{path} if $volid_hash->{$volid};
- if ($used->{$volid} || ($path && $usedpath->{$path})) {
+ if ($referenced->{$volid} || ($path && $referencedpath->{$path})) {
+ print "$prefix remove entry '$opt', its volume '$volid' is in use.\n";
$changes = 1;
delete $conf->{$opt};
}
+
+ $referenced->{$volid} = 1;
+ $referencedpath->{$path} = 1 if $path;
}
foreach my $volid (sort keys %$volid_hash) {
next if $volid =~ m/vm-$vmid-state-/;
- next if $used->{$volid};
+ next if $referenced->{$volid};
my $path = $volid_hash->{$volid}->{path};
next if !$path; # just to be sure
- next if $usedpath->{$path};
+ next if $referencedpath->{$path};
$changes = 1;
- PVE::QemuConfig->add_unused_volume($conf, $volid);
- $usedpath->{$path} = 1; # avoid to add more than once (aliases)
+ 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)
}
return $changes;
}
sub rescan {
- my ($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 {
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($vmid)) {
sub restore_vma_archive {
my ($archive, $vmid, $user, $opts, $comp) = @_;
- my $input = $archive eq '-' ? "<&STDIN" : undef;
my $readfrom = $archive;
- my $uncomp = '';
- if ($comp) {
+ my $cfg = PVE::Storage::config();
+ my $commands = [];
+ my $bwlimit = $opts->{bwlimit};
+
+ my $dbg_cmdstring = '';
+ my $add_pipe = sub {
+ my ($cmd) = @_;
+ push @$commands, $cmd;
+ $dbg_cmdstring .= ' | ' if length($dbg_cmdstring);
+ $dbg_cmdstring .= PVE::Tools::cmd2string($cmd);
$readfrom = '-';
- my $qarchive = PVE::Tools::shellquote($archive);
+ };
+
+ my $input = undef;
+ if ($archive eq '-') {
+ $input = '<&STDIN';
+ } else {
+ # If we use a backup from a PVE defined storage we also consider that
+ # storage's rate limit:
+ my (undef, $volid) = PVE::Storage::path_to_volume_id($cfg, $archive);
+ if (defined($volid)) {
+ my ($sid, undef) = PVE::Storage::parse_volume_id($volid);
+ my $readlimit = PVE::Storage::get_bandwidth_limit('restore', [$sid], $bwlimit);
+ if ($readlimit) {
+ print STDERR "applying read rate limit: $readlimit\n";
+ my $cstream = ['cstream', '-t', $readlimit*1024, '--', $readfrom];
+ $add_pipe->($cstream);
+ }
+ }
+ }
+
+ if ($comp) {
+ my $cmd;
if ($comp eq 'gzip') {
- $uncomp = "zcat $qarchive|";
+ $cmd = ['zcat', $readfrom];
} elsif ($comp eq 'lzop') {
- $uncomp = "lzop -d -c $qarchive|";
+ $cmd = ['lzop', '-d', '-c', $readfrom];
} else {
die "unknown compression method '$comp'\n";
}
-
+ $add_pipe->($cmd);
}
my $tmpdir = "/var/tmp/vzdumptmp$$";
rmtree $tmpdir;
# disable interrupts (always do cleanups)
- local $SIG{INT} = $SIG{TERM} = $SIG{QUIT} = $SIG{HUP} = sub {
- warn "got interrupt - ignored\n";
- };
+ local $SIG{INT} =
+ local $SIG{TERM} =
+ local $SIG{QUIT} =
+ local $SIG{HUP} = sub { warn "got interrupt - ignored\n"; };
my $mapfifo = "/var/tmp/vzdumptmp$$.fifo";
POSIX::mkfifo($mapfifo, 0600);
open($fifofh, '>', $mapfifo) || die $!;
};
- my $cmd = "${uncomp}vma extract -v -r $mapfifo $readfrom $tmpdir";
+ $add_pipe->(['vma', 'extract', '-v', '-r', $mapfifo, $readfrom, $tmpdir]);
my $oldtimeout;
my $timeout = 5;
my $cfs_path = PVE::QemuConfig->cfs_config_path($vmid);
my $oldconf = PVE::Cluster::cfs_read_file($cfs_path);
+ my %storage_limits;
+
my $print_devmap = sub {
my $virtdev_hash = {};
$rpcenv->check($user, "/storage/$storeid", ['Datastore.AllocateSpace']);
}
+ $storage_limits{$storeid} = $bwlimit;
+
$virtdev_hash->{$virtdev} = $devinfo->{$devname};
}
}
+ foreach my $key (keys %storage_limits) {
+ my $limit = PVE::Storage::get_bandwidth_limit('restore', [$key], $bwlimit);
+ next if !$limit;
+ print STDERR "rate limit for storage $key: $limit KiB/s\n";
+ $storage_limits{$key} = $limit * 1024;
+ }
+
foreach my $devname (keys %$devinfo) {
die "found no device mapping information for device '$devname'\n"
if !$devinfo->{$devname}->{virtdev};
}
- my $cfg = PVE::Storage::config();
-
# create empty/temp config
if ($oldconf) {
PVE::Tools::file_set_contents($conffile, "memory: 128\n");
foreach my $virtdev (sort keys %$virtdev_hash) {
my $d = $virtdev_hash->{$virtdev};
my $alloc_size = int(($d->{size} + 1024 - 1)/1024);
- my $scfg = PVE::Storage::storage_config($cfg, $d->{storeid});
+ my $storeid = $d->{storeid};
+ my $scfg = PVE::Storage::storage_config($cfg, $storeid);
+
+ my $map_opts = '';
+ if (my $limit = $storage_limits{$storeid}) {
+ $map_opts .= "throttling.bps=$limit:throttling.group=$storeid:";
+ }
# test if requested format is supported
- my ($defFormat, $validFormats) = PVE::Storage::storage_default_format($cfg, $d->{storeid});
+ my ($defFormat, $validFormats) = PVE::Storage::storage_default_format($cfg, $storeid);
my $supported = grep { $_ eq $d->{format} } @$validFormats;
$d->{format} = $defFormat if !$supported;
- my $volid = PVE::Storage::vdisk_alloc($cfg, $d->{storeid}, $vmid,
+ my $volid = PVE::Storage::vdisk_alloc($cfg, $storeid, $vmid,
$d->{format}, undef, $alloc_size);
print STDERR "new volume ID is '$volid'\n";
$d->{volid} = $volid;
$write_zeros = 0;
}
- print $fifofh "format=$d->{format}:${write_zeros}:$d->{devname}=$path\n";
+ print $fifofh "${map_opts}format=$d->{format}:${write_zeros}:$d->{devname}=$path\n";
print "map '$d->{devname}' to '$path' (write zeros = ${write_zeros})\n";
$map->{$virtdev} = $volid;
eval {
# enable interrupts
- local $SIG{INT} = $SIG{TERM} = $SIG{QUIT} = $SIG{HUP} = $SIG{PIPE} = sub {
- die "interrupted by signal\n";
- };
+ local $SIG{INT} =
+ local $SIG{TERM} =
+ local $SIG{QUIT} =
+ local $SIG{HUP} =
+ local $SIG{PIPE} = sub { die "interrupted by signal\n"; };
local $SIG{ALRM} = sub { die "got timeout\n"; };
$oldtimeout = alarm($timeout);
}
};
- print "restore vma archive: $cmd\n";
- run_command($cmd, input => $input, outfunc => $parser, afterfork => $openfifo);
+ print "restore vma archive: $dbg_cmdstring\n";
+ run_command($commands, input => $input, outfunc => $parser, afterfork => $openfifo);
};
my $err = $@;
push @$vollist, $volid if $volid;
}
- my $cfg = PVE::Storage::config();
PVE::Storage::deactivate_volumes($cfg, $vollist);
unlink $mapfifo;
my $tmpfn = "$conffile.$$.tmp";
# disable interrupts (always do cleanups)
- local $SIG{INT} = $SIG{TERM} = $SIG{QUIT} = $SIG{HUP} = sub {
- print STDERR "got interrupt - ignored\n";
- };
+ local $SIG{INT} =
+ local $SIG{TERM} =
+ local $SIG{QUIT} =
+ local $SIG{HUP} = sub { print STDERR "got interrupt - ignored\n"; };
eval {
# enable interrupts
- local $SIG{INT} = $SIG{TERM} = $SIG{QUIT} = $SIG{HUP} = $SIG{PIPE} = sub {
- die "interrupted by signal\n";
- };
+ local $SIG{INT} =
+ local $SIG{TERM} =
+ local $SIG{QUIT} =
+ local $SIG{HUP} =
+ local $SIG{PIPE} = sub { die "interrupted by signal\n"; };
if ($archive eq '-') {
print "extracting archive from STDIN\n";
warn $@ if $@;
};
-sub foreach_writable_storage {
+sub foreach_storage_used_by_vm {
my ($conf, $func) = @_;
my $sidhash = {};
- foreach my $ds (keys %$conf) {
- next if !is_valid_drivename($ds);
-
- my $drive = parse_drive($ds, $conf->{$ds});
- next if !$drive;
- next if drive_is_cdrom($drive);
+ foreach_drive($conf, sub {
+ my ($ds, $drive) = @_;
+ return if drive_is_cdrom($drive);
my $volid = $drive->{file};
my ($sid, $volname) = PVE::Storage::parse_volume_id($volid, 1);
$sidhash->{$sid} = $sid if $sid;
- }
+ });
foreach my $sid (sort keys %$sidhash) {
&$func($sid);
my $storage_name = PVE::Storage::parse_volume_id($volid);
- if ($qemu_snap_storage->{$storecfg->{ids}->{$storage_name}->{type}}
+ if ($qemu_snap_storage->{$storecfg->{ids}->{$storage_name}->{type}}
&& !$storecfg->{ids}->{$storage_name}->{krbd}){
return 1;
}
}
sub qga_check_running {
- my ($vmid) = @_;
+ my ($vmid, $nowarn) = @_;
eval { vm_mon_cmd($vmid, "guest-ping", timeout => 3); };
if ($@) {
- warn "Qemu Guest Agent are not running - $@";
+ warn "Qemu Guest Agent is not running - $@" if !$nowarn;
return 0;
}
return 1;
});
}
+sub convert_iscsi_path {
+ my ($path) = @_;
+
+ if ($path =~ m|^iscsi://([^/]+)/([^/]+)/(.+)$|) {
+ my $portal = $1;
+ my $target = $2;
+ my $lun = $3;
+
+ my $initiator_name = get_initiator_name();
+
+ return "file.driver=iscsi,file.transport=tcp,file.initiator-name=$initiator_name,".
+ "file.portal=$portal,file.target=$target,file.lun=$lun,driver=raw";
+ }
+
+ die "cannot convert iscsi path '$path', unkown format\n";
+}
+
sub qemu_img_convert {
my ($src_volid, $dst_volid, $size, $snapname, $is_zero_initialized) = @_;
my $src_path = PVE::Storage::path($storecfg, $src_volid, $snapname);
my $dst_path = PVE::Storage::path($storecfg, $dst_volid);
+ my $src_is_iscsi = ($src_path =~ m|^iscsi://|);
+ my $dst_is_iscsi = ($dst_path =~ m|^iscsi://|);
+
my $cmd = [];
push @$cmd, '/usr/bin/qemu-img', 'convert', '-p', '-n';
- push @$cmd, '-s', $snapname if($snapname && $src_format eq "qcow2");
- push @$cmd, '-f', $src_format, '-O', $dst_format, $src_path;
- if ($is_zero_initialized) {
+ 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 $src_scfg->{type} eq 'zfspool';
+
+ if ($src_is_iscsi) {
+ push @$cmd, '--image-opts';
+ $src_path = convert_iscsi_path($src_path);
+ } else {
+ push @$cmd, '-f', $src_format;
+ }
+
+ if ($dst_is_iscsi) {
+ push @$cmd, '--target-image-opts';
+ $dst_path = convert_iscsi_path($dst_path);
+ } else {
+ push @$cmd, '-O', $dst_format;
+ }
+
+ push @$cmd, $src_path;
+
+ if (!$dst_is_iscsi && $is_zero_initialized) {
push @$cmd, "zeroinit:$dst_path";
} else {
push @$cmd, $dst_path;
sub qemu_img_format {
my ($scfg, $volname) = @_;
- if ($scfg->{path} && $volname =~ m/\.(raw|cow|qcow|qcow2|qed|vmdk|cloop)$/) {
+ if ($scfg->{path} && $volname =~ m/\.($QEMU_FORMAT_RE)$/) {
return $1;
} else {
return "raw";
}
sub qemu_drive_mirror {
- my ($vmid, $drive, $dst_volid, $vmiddst, $is_zero_initialized, $jobs, $skipcomplete, $qga) = @_;
+ my ($vmid, $drive, $dst_volid, $vmiddst, $is_zero_initialized, $jobs, $skipcomplete, $qga, $bwlimit) = @_;
$jobs = {} if !$jobs;
my $format;
$jobs->{"drive-$drive"} = {};
- if ($dst_volid =~ /^nbd:(localhost|[\d\.]+|\[[\d\.:a-fA-F]+\]):(\d+):exportname=(\S+)/) {
- my $server = $1;
- my $port = $2;
- my $exportname = $3;
-
+ if ($dst_volid =~ /^nbd:/) {
+ $qemu_target = $dst_volid;
$format = "nbd";
- my $unixsocket = "/run/qemu-server/$vmid.mirror-drive-$drive";
- $qemu_target = "nbd+unix:///$exportname?socket=$unixsocket";
- my $cmd = ['socat', '-T30', "UNIX-LISTEN:$unixsocket,fork", "TCP:$server:$2,connect-timeout=5"];
-
- my $pid = fork();
- if (!defined($pid)) {
- die "forking socat tunnel failed\n";
- } elsif ($pid == 0) {
- exec(@$cmd);
- warn "exec failed: $!\n";
- POSIX::_exit(-1);
- }
- $jobs->{"drive-$drive"}->{pid} = $pid;
-
- my $timeout = 0;
- while (!-S $unixsocket) {
- die "nbd connection helper timed out\n"
- if $timeout++ > 5;
- sleep 1;
- }
} else {
my $storecfg = PVE::Storage::config();
my ($dst_storeid, $dst_volname) = PVE::Storage::parse_volume_id($dst_volid);
my $opts = { timeout => 10, device => "drive-$drive", mode => "existing", sync => "full", target => $qemu_target };
$opts->{format} = $format if $format;
- print "drive mirror is starting for drive-$drive\n";
-
- eval { vm_mon_cmd($vmid, "drive-mirror", %$opts); }; #if a job already run for this device,it's throw an error
+ if (defined($bwlimit)) {
+ $opts->{speed} = $bwlimit * 1024;
+ print "drive mirror is starting for drive-$drive with bandwidth limit: ${bwlimit} KB/s\n";
+ } else {
+ print "drive mirror is starting for drive-$drive\n";
+ }
+ # if a job already runs for this device we get an error, catch it for cleanup
+ eval { vm_mon_cmd($vmid, "drive-mirror", %$opts); };
if (my $err = $@) {
eval { PVE::QemuServer::qemu_blockjobs_cancel($vmid, $jobs) };
- die "mirroring error: $err";
+ warn "$@\n" if $@;
+ die "mirroring error: $err\n";
}
qemu_drive_mirror_monitor ($vmid, $vmiddst, $jobs, $skipcomplete, $qga);
last if $skipcomplete; #do the complete later
if ($vmiddst && $vmiddst != $vmid) {
- if ($qga) {
+ my $agent_running = $qga && qga_check_running($vmid);
+ if ($agent_running) {
print "freeze filesystem\n";
eval { PVE::QemuServer::vm_mon_cmd($vmid, "guest-fsfreeze-freeze"); };
} else {
# if we clone a disk for a new target vm, we don't switch the disk
PVE::QemuServer::qemu_blockjobs_cancel($vmid, $jobs);
- if ($qga) {
+ if ($agent_running) {
print "unfreeze filesystem\n";
eval { PVE::QemuServer::vm_mon_cmd($vmid, "guest-fsfreeze-thaw"); };
} else {
}else {
print "$job: Completed successfully.\n";
$jobs->{$job}->{complete} = 1;
- eval { qemu_blockjobs_finish_tunnel($vmid, $job, $jobs->{$job}->{pid}) } ;
}
}
}
if (defined($jobs->{$job}->{cancel}) && !defined($running_jobs->{$job})) {
print "$job: Done.\n";
- eval { qemu_blockjobs_finish_tunnel($vmid, $job, $jobs->{$job}->{pid}) } ;
delete $jobs->{$job};
}
}
}
}
-sub qemu_blockjobs_finish_tunnel {
- my ($vmid, $job, $cpid) = @_;
-
- return if !$cpid;
-
- for (my $i = 1; $i < 20; $i++) {
- my $waitpid = waitpid($cpid, WNOHANG);
- last if (defined($waitpid) && ($waitpid == $cpid));
-
- if ($i == 10) {
- kill(15, $cpid);
- } elsif ($i >= 15) {
- kill(9, $cpid);
- }
- sleep (1);
- }
- unlink "/run/qemu-server/$vmid.mirror-$job";
-}
-
sub clone_disk {
my ($storecfg, $vmid, $running, $drivename, $drive, $snapname,
- $newvmid, $storage, $format, $full, $newvollist, $jobs, $skipcomplete, $qga) = @_;
+ $newvmid, $storage, $format, $full, $newvollist, $jobs, $skipcomplete, $qga, $bwlimit) = @_;
my $newvolid;
my ($storeid, $volname) = PVE::Storage::parse_volume_id($drive->{file});
$storeid = $storage if $storage;
- my ($defFormat, $validFormats) = PVE::Storage::storage_default_format($storecfg, $storeid);
- if (!$format) {
- my $scfg = PVE::Storage::storage_config($storecfg, $storeid);
- $format = qemu_img_format($scfg, $volname);
- }
-
- # test if requested format is supported - else use default
- my $supported = grep { $_ eq $format } @$validFormats;
- $format = $defFormat if !$supported;
-
+ my $dst_format = resolve_dst_disk_format($storecfg, $storeid, $volname, $format);
my ($size) = PVE::Storage::volume_size_info($storecfg, $drive->{file}, 3);
print "create full clone of drive $drivename ($drive->{file})\n";
- $newvolid = PVE::Storage::vdisk_alloc($storecfg, $storeid, $newvmid, $format, undef, ($size/1024));
+ my $name = undef;
+ if (drive_is_cloudinit($drive)) {
+ $name = "vm-$newvmid-cloudinit";
+ $snapname = undef;
+ # cloudinit only supports raw and qcow2 atm:
+ if ($dst_format eq 'qcow2') {
+ $name .= '.qcow2';
+ } elsif ($dst_format ne 'raw') {
+ die "clone: unhandled format for cloudinit image\n";
+ }
+ }
+ $newvolid = PVE::Storage::vdisk_alloc($storecfg, $storeid, $newvmid, $dst_format, $name, ($size/1024));
push @$newvollist, $newvolid;
PVE::Storage::activate_volumes($storecfg, [$newvolid]);
my $sparseinit = PVE::Storage::volume_has_feature($storecfg, 'sparseinit', $newvolid);
if (!$running || $snapname) {
+ # TODO: handle bwlimits
qemu_img_convert($drive->{file}, $newvolid, $size, $snapname, $sparseinit);
} else {
if $drive->{iothread};
}
- qemu_drive_mirror($vmid, $drivename, $newvolid, $newvmid, $sparseinit, $jobs, $skipcomplete, $qga);
+ qemu_drive_mirror($vmid, $drivename, $newvolid, $newvmid, $sparseinit, $jobs, $skipcomplete, $qga, $bwlimit);
}
}
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;
$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 {
$machine = PVE::QemuServer::get_current_qemu_machine($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 $machine;
return ($use_old_bios_files, $machine_type);
}
-sub lspci {
+sub create_efidisk($$$$$) {
+ my ($storecfg, $storeid, $vmid, $fmt, $arch) = @_;
- my $devices = {};
+ my (undef, $ovmf_vars) = get_ovmf_files($arch);
+ die "EFI vars default image not found\n" if ! -f $ovmf_vars;
- dir_glob_foreach("$pcisysfs/devices", '[a-f0-9]{4}:([a-f0-9]{2}:[a-f0-9]{2})\.([0-9])', sub {
- my (undef, $id, $function) = @_;
- my $res = { id => $id, function => $function};
- push @{$devices->{$id}}, $res;
- });
+ 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]);
- return $devices;
+ my $path = PVE::Storage::path($storecfg, $volid);
+ eval {
+ run_command(['/usr/bin/qemu-img', 'convert', '-n', '-f', 'raw', '-O', $fmt, $ovmf_vars, $path]);
+ };
+ die "Copying EFI vars image failed: $@" if $@;
+
+ return ($volid, $vars_size);
}
sub vm_iothreads_list {
return ($maxdev, $controller, $controller_prefix);
}
-sub add_hyperv_enlighments {
- my ($cpuFlags, $winversion, $machine_type, $kvmver, $nokvm, $bios, $gpu_passthrough) = @_;
+sub add_hyperv_enlightenments {
+ my ($cpuFlags, $winversion, $machine_type, $kvmver, $bios, $gpu_passthrough, $hv_vendor_id) = @_;
- return if $nokvm;
return if $winversion < 6;
return if $bios && $bios eq 'ovmf' && $winversion < 8;
- push @$cpuFlags , 'hv_vendor_id=proxmox' if $gpu_passthrough;
+ if ($gpu_passthrough || defined($hv_vendor_id)) {
+ $hv_vendor_id //= 'proxmox';
+ push @$cpuFlags , "hv_vendor_id=$hv_vendor_id";
+ }
if (qemu_machine_feature_enabled ($machine_type, $kvmver, 2, 3)) {
push @$cpuFlags , 'hv_spinlocks=0x1fff';
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';
+ }
}
}
return $winversion;
}
+sub resolve_dst_disk_format {
+ my ($storecfg, $storeid, $src_volname, $format) = @_;
+ my ($defFormat, $validFormats) = PVE::Storage::storage_default_format($storecfg, $storeid);
+
+ if (!$format) {
+ # if no target format is specified, use the source disk format as hint
+ if ($src_volname) {
+ my $scfg = PVE::Storage::storage_config($storecfg, $storeid);
+ $format = qemu_img_format($scfg, $src_volname);
+ } else {
+ return $defFormat;
+ }
+ }
+
+ # test if requested format is supported - else use default
+ my $supported = grep { $_ eq $format } @$validFormats;
+ $format = $defFormat if !$supported;
+ return $format;
+}
+
+sub resolve_first_disk {
+ my $conf = shift;
+ my @disks = PVE::QemuServer::valid_drive_names();
+ my $firstdisk;
+ foreach my $ds (reverse @disks) {
+ next if !$conf->{$ds};
+ my $disk = PVE::QemuServer::parse_drive($ds, $conf->{$ds});
+ next if PVE::QemuServer::drive_is_cdrom($disk);
+ $firstdisk = $ds;
+ }
+ return $firstdisk;
+}
+
+sub generate_uuid {
+ my ($uuid, $uuid_str);
+ UUID::generate($uuid);
+ UUID::unparse($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
sub complete_backup_archives {
return $res;
}
-sub nbd_stop {
- my ($vmid) = @_;
-
- vm_mon_cmd($vmid, 'nbd-server-stop');
-}
-
1;