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);
+use PVE::Tools qw(run_command lock_file lock_file_full file_read_firstline dir_glob_foreach);
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;
my $pcisysfs = "/sys/bus/pci";
my $confdesc = {
+ iothread => {
+ optional => 1,
+ type => 'boolean',
+ description => "Enable iothread dataplane.",
+ default => 0,
+ },
onboot => {
optional => 1,
type => 'boolean',
minimum => 1,
default => 1,
},
+ numa => {
+ optional => 1,
+ type => 'boolean',
+ description => "Enable/disable Numa.",
+ default => 0,
+ },
maxcpus => {
optional => 1,
type => 'integer',
optional => 1,
description => "Emulated CPU type.",
type => 'string',
- enum => [ qw(486 athlon pentium pentium2 pentium3 coreduo core2duo kvm32 kvm64 qemu32 qemu64 phenom Conroe Penryn Nehalem Westmere SandyBridge Haswell Opteron_G1 Opteron_G2 Opteron_G3 Opteron_G4 Opteron_G5 host) ],
+ enum => [ qw(486 athlon pentium pentium2 pentium3 coreduo core2duo kvm32 kvm64 qemu32 qemu64 phenom Conroe Penryn Nehalem Westmere SandyBridge Haswell Broadwell Opteron_G1 Opteron_G2 Opteron_G3 Opteron_G4 Opteron_G5 host) ],
default => 'kvm64',
},
parent => get_standard_option('pve-snapshot-name', {
maxLength => 40,
optional => 1,
},
+ smbios1 => {
+ description => "Specify SMBIOS type 1 fields.",
+ type => 'string', format => 'pve-qm-smbios1',
+ typetext => "[manufacturer=str][,product=str][,version=str][,serial=str] [,uuid=uuid][,sku=str][,family=str]",
+ maxLength => 256,
+ optional => 1,
+ },
};
# what about other qemu settings ?
my $MAX_USB_DEVICES = 5;
my $MAX_NETS = 32;
my $MAX_UNUSED_DISKS = 8;
-my $MAX_HOSTPCI_DEVICES = 2;
+my $MAX_HOSTPCI_DEVICES = 4;
my $MAX_SERIAL_PORTS = 4;
my $MAX_PARALLEL_PORTS = 3;
+my $MAX_NUMA = 8;
+
+my $numadesc = {
+ optional => 1,
+ type => 'string', format => 'pve-qm-numanode',
+ typetext => "cpus=<id[-id],memory=<mb>[[,hostnodes=<id[-id]>][,policy=<preferred|bind|interleave>]]",
+ description => "numa topology",
+};
+PVE::JSONSchema::register_standard_option("pve-qm-numanode", $numadesc);
+
+for (my $i = 0; $i < $MAX_NUMA; $i++) {
+ $confdesc->{"numa$i"} = $numadesc;
+}
my $nic_model_list = ['rtl8139', 'ne2k_pci', 'e1000', 'pcnet', 'virtio',
'ne2k_isa', 'i82551', 'i82557b', 'i82559er', 'vmxnet3'];
my $netdesc = {
optional => 1,
type => 'string', format => 'pve-qm-net',
- typetext => "MODEL=XX:XX:XX:XX:XX:XX [,bridge=<dev>][,rate=<mbps>][,tag=<vlanid>][,firewall=0|1]",
+ typetext => "MODEL=XX:XX:XX:XX:XX:XX [,bridge=<dev>][,queues=<nbqueues>][,rate=<mbps>][,tag=<vlanid>][,firewall=0|1]",
description => <<EODESCR,
Specify network devices.
my $hostpcidesc = {
optional => 1,
type => 'string', format => 'pve-qm-hostpci',
- typetext => "[host=]HOSTPCIDEVICE [,driver=kvm|vfio] [,rombar=on|off]",
+ typetext => "[host=]HOSTPCIDEVICE [,driver=kvm|vfio] [,rombar=on|off] [,pcie=0|1] [,x-vga=on|off]",
description => <<EODESCR,
Map host pci devices. HOSTPCIDEVICE syntax is:
The 'rombar' option determines whether or not the device's ROM will be visible in the guest's memory map (default is 'on').
-The 'driver' option is currently ignored.
-
Note: This option allows direct access to host hardware. So it is no longer possible to migrate such machines - use with special care.
Experimental: user reported problems with this option.
my $serialdesc = {
optional => 1,
type => 'string',
- pattern => '(/dev/ttyS\d+|socket)',
+ pattern => '(/dev/.+|socket)',
description => <<EODESCR,
-Create a serial device inside the VM (n is 0 to 3), and pass through a host serial device, or create a unix socket on the host side (use 'qm terminal' to open a terminal connection).
+Create a serial device inside the VM (n is 0 to 3), and pass through a host serial device (i.e. /dev/ttyS0), or create a unix socket on the host side (use 'qm terminal' to open a terminal connection).
Note: This option allows direct access to host hardware. So it is no longer possible to migrate such machines - use with special care.
return $res;
}
+sub machine_type_is_q35 {
+ my ($conf) = @_;
+
+ return $conf->{machine} && ($conf->{machine} =~ m/q35/) ? 1 : 0;
+}
+
+sub print_tabletdevice_full {
+ my ($conf) = @_;
+
+ my $q35 = machine_type_is_q35($conf);
+
+ # we use uhci for old VMs because tablet driver was buggy in older qemu
+ my $usbbus = $q35 ? "ehci" : "uhci";
+
+ return "usb-tablet,id=tablet,bus=$usbbus.0,port=1";
+}
+
sub print_drivedevice_full {
my ($storecfg, $conf, $vmid, $drive, $bridges) = @_;
if ($drive->{interface} eq 'virtio') {
my $pciaddr = print_pci_addr("$drive->{interface}$drive->{index}", $bridges);
$device = "virtio-blk-pci,drive=drive-$drive->{interface}$drive->{index},id=$drive->{interface}$drive->{index}$pciaddr";
+ $device .= ",iothread=iothread0" if $conf->{iothread};
} elsif ($drive->{interface} eq 'scsi') {
$maxdev = ($conf->{scsihw} && ($conf->{scsihw} !~ m/^lsi/)) ? 256 : 7;
my $controller = int($drive->{index} / $maxdev);
return $device;
}
+sub get_initiator_name {
+ my $initiator;
+
+ my $fh = IO::File->new('/etc/iscsi/initiatorname.iscsi') || return undef;
+ while (defined(my $line = <$fh>)) {
+ next if $line !~ m/^\s*InitiatorName\s*=\s*([\.\-:\w]+)/;
+ $initiator = $1;
+ last;
+ }
+ $fh->close();
+
+ return $initiator;
+}
+
sub print_drive_full {
my ($storecfg, $vmid, $drive) = @_;
$opts .= ",cache=none" if !$drive->{cache} && !drive_is_cdrom($drive);
+ my $detectzeroes = $drive->{discard} ? "unmap" : "on";
+ $opts .= ",detect-zeroes=$detectzeroes" if !drive_is_cdrom($drive);
+
my $pathinfo = $path ? "file=$path," : '';
return "${pathinfo}if=none,id=drive-$drive->{interface}$drive->{index}$opts";
my $extra = ($bootorder !~ m/n/) ? "romfile=," : '';
my $pciaddr = print_pci_addr("$netid", $bridges);
my $tmpstr = "$device,${extra}mac=$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)
+ my $vectors = $net->{queues} * 2 + 2;
+ $tmpstr .= ",vectors=$vectors,mq=on";
+ }
$tmpstr .= ",bootindex=$net->{bootindex}" if $net->{bootindex} ;
return $tmpstr;
}
my $vmname = $conf->{name} || "vm$vmid";
+ my $netdev = "";
+
if ($net->{bridge}) {
- return "type=tap,id=$netid,ifname=${ifname},script=/var/lib/qemu-server/pve-bridge,downscript=/var/lib/qemu-server/pve-bridgedown$vhostparam";
+ $netdev = "type=tap,id=$netid,ifname=${ifname},script=/var/lib/qemu-server/pve-bridge,downscript=/var/lib/qemu-server/pve-bridgedown$vhostparam";
} else {
- return "type=user,id=$netid,hostname=$vmname";
+ $netdev = "type=user,id=$netid,hostname=$vmname";
}
+
+ $netdev .= ",queues=$net->{queues}" if ($net->{queues} && $net->{model} eq 'virtio');
+
+ return $netdev;
}
sub drive_is_cdrom {
}
+sub parse_numa {
+ my ($data) = @_;
+
+ my $res = {};
+
+ foreach my $kvp (split(/,/, $data)) {
+
+ if ($kvp =~ m/^memory=(\S+)$/) {
+ $res->{memory} = $1;
+ } elsif ($kvp =~ m/^policy=(preferred|bind|interleave)$/) {
+ $res->{policy} = $1;
+ } elsif ($kvp =~ m/^cpus=(\d+)(-(\d+))?$/) {
+ $res->{cpus}->{start} = $1;
+ $res->{cpus}->{end} = $3;
+ } elsif ($kvp =~ m/^hostnodes=(\d+)(-(\d+))?$/) {
+ $res->{hostnodes}->{start} = $1;
+ $res->{hostnodes}->{end} = $3;
+ } else {
+ return undef;
+ }
+ }
+
+ return $res;
+}
+
sub parse_hostpci {
my ($value) = @_;
my $res = {};
foreach my $kv (@list) {
- if ($kv =~ m/^(host=)?([a-f0-9]{2}:[a-f0-9]{2}\.[a-f0-9])$/) {
+ if ($kv =~ m/^(host=)?([a-f0-9]{2}:[a-f0-9]{2})(\.([a-f0-9]))?$/) {
$found = 1;
- $res->{pciid} = $2;
+ if(defined($4)){
+ push @{$res->{pciid}}, { id => $2 , function => $4};
+
+ }else{
+ my $pcidevices = lspci($2);
+ $res->{pciid} = $pcidevices->{$2};
+ }
} elsif ($kv =~ m/^driver=(kvm|vfio)$/) {
$res->{driver} = $1;
} elsif ($kv =~ m/^rombar=(on|off)$/) {
$res->{rombar} = $1;
+ } elsif ($kv =~ m/^x-vga=(on|off)$/) {
+ $res->{'x-vga'} = $1;
+ } elsif ($kv =~ m/^pcie=(\d+)$/) {
+ $res->{pcie} = 1 if $1 == 1;
} else {
warn "unknown hostpci setting '$kv'\n";
}
$res->{macaddr} = $mac;
} elsif ($kvp =~ m/^bridge=(\S+)$/) {
$res->{bridge} = $1;
+ } elsif ($kvp =~ m/^queues=(\d+)$/) {
+ $res->{queues} = $1;
} elsif ($kvp =~ m/^rate=(\d+(\.\d+)?)$/) {
$res->{rate} = $1;
} elsif ($kvp =~ m/^tag=(\d+)$/) {
return $key;
}
+sub vm_is_volid_owner {
+ my ($storecfg, $vmid, $volid) = @_;
+
+ if ($volid !~ m|^/|) {
+ my ($path, $owner);
+ eval { ($path, $owner) = PVE::Storage::path($storecfg, $volid); };
+ if ($owner && ($owner == $vmid)) {
+ return 1;
+ }
+ }
+
+ return undef;
+}
+
+sub vmconfig_delete_pending_option {
+ my ($conf, $key) = @_;
+
+ delete $conf->{pending}->{$key};
+ my $pending_delete_hash = { $key => 1 };
+ foreach my $opt (PVE::Tools::split_list($conf->{pending}->{delete})) {
+ $pending_delete_hash->{$opt} = 1;
+ }
+ $conf->{pending}->{delete} = join(',', keys %$pending_delete_hash);
+}
+
+sub vmconfig_undelete_pending_option {
+ my ($conf, $key) = @_;
+
+ my $pending_delete_hash = {};
+ foreach my $opt (PVE::Tools::split_list($conf->{pending}->{delete})) {
+ $pending_delete_hash->{$opt} = 1;
+ }
+ delete $pending_delete_hash->{$key};
+
+ my @keylist = keys %$pending_delete_hash;
+ if (scalar(@keylist)) {
+ $conf->{pending}->{delete} = join(',', @keylist);
+ } else {
+ delete $conf->{pending}->{delete};
+ }
+}
+
+sub vmconfig_register_unused_drive {
+ my ($storecfg, $vmid, $conf, $drive) = @_;
+
+ if (!drive_is_cdrom($drive)) {
+ my $volid = $drive->{file};
+ if (vm_is_volid_owner($storecfg, $vmid, $volid)) {
+ add_unused_volume($conf, $volid, $vmid);
+ }
+ }
+}
+
+sub vmconfig_cleanup_pending {
+ my ($conf) = @_;
+
+ # remove pending changes when nothing changed
+ my $changes;
+ foreach my $opt (keys %{$conf->{pending}}) {
+ if (defined($conf->{$opt}) && ($conf->{pending}->{$opt} eq $conf->{$opt})) {
+ $changes = 1;
+ delete $conf->{pending}->{$opt};
+ }
+ }
+
+ # remove delete if option is not set
+ my $pending_delete_hash = {};
+ foreach my $opt (PVE::Tools::split_list($conf->{pending}->{delete})) {
+ if (defined($conf->{$opt})) {
+ $pending_delete_hash->{$opt} = 1;
+ } else {
+ $changes = 1;
+ }
+ }
+
+ my @keylist = keys %$pending_delete_hash;
+ if (scalar(@keylist)) {
+ $conf->{pending}->{delete} = join(',', @keylist);
+ } else {
+ delete $conf->{pending}->{delete};
+ }
+
+ return $changes;
+}
+
+my $valid_smbios1_options = {
+ manufacturer => '\S+',
+ product => '\S+',
+ version => '\S+',
+ serial => '\S+',
+ uuid => '[a-fA-F0-9]{8}(?:-[a-fA-F0-9]{4}){3}-[a-fA-F0-9]{12}',
+ sku => '\S+',
+ family => '\S+',
+};
+
+# smbios: [manufacturer=str][,product=str][,version=str][,serial=str][,uuid=uuid][,sku=str][,family=str]
+sub parse_smbios1 {
+ my ($data) = @_;
+
+ my $res = {};
+
+ foreach my $kvp (split(/,/, $data)) {
+ return undef if $kvp !~ m/^(\S+)=(.+)$/;
+ my ($k, $v) = split(/=/, $kvp);
+ return undef if !defined($k) || !defined($v);
+ return undef if !$valid_smbios1_options->{$k};
+ return undef if $v !~ m/^$valid_smbios1_options->{$k}$/;
+ $res->{$k} = $v;
+ }
+
+ return $res;
+}
+
+sub print_smbios1 {
+ my ($smbios1) = @_;
+
+ my $data = '';
+ foreach my $k (keys %$smbios1) {
+ next if !defined($smbios1->{$k});
+ next if !$valid_smbios1_options->{$k};
+ $data .= ',' if $data;
+ $data .= "$k=$smbios1->{$k}";
+ }
+ return $data;
+}
+
+PVE::JSONSchema::register_format('pve-qm-smbios1', \&verify_smbios1);
+sub verify_smbios1 {
+ my ($value, $noerr) = @_;
+
+ return $value if parse_smbios1($value);
+
+ return undef if $noerr;
+
+ die "unable to parse smbios (type 1) options\n";
+}
+
PVE::JSONSchema::register_format('pve-qm-bootdisk', \&verify_bootdisk);
sub verify_bootdisk {
my ($value, $noerr) = @_;
die "invalid boot disk '$value'\n";
}
+PVE::JSONSchema::register_format('pve-qm-numanode', \&verify_numa);
+sub verify_numa {
+ my ($value, $noerr) = @_;
+
+ return $value if parse_numa($value);
+
+ return undef if $noerr;
+
+ die "unable to parse numa options\n";
+}
+
PVE::JSONSchema::register_format('pve-qm-net', \&verify_net);
sub verify_net {
my ($value, $noerr) = @_;
my $res = {
digest => Digest::SHA::sha1_hex($raw),
snapshots => {},
+ pending => {},
};
$filename =~ m|/qemu-server/(\d+)\.conf$|
my $conf = $res;
my $descr = '';
+ my $section = '';
my @lines = split(/\n/, $raw);
foreach my $line (@lines) {
next if $line =~ m/^\s*$/;
- if ($line =~ m/^\[([a-z][a-z0-9_\-]+)\]\s*$/i) {
- my $snapname = $1;
+ if ($line =~ m/^\[PENDING\]\s*$/i) {
+ $section = 'pending';
$conf->{description} = $descr if $descr;
$descr = '';
- $conf = $res->{snapshots}->{$snapname} = {};
+ $conf = $res->{$section} = {};
+ next;
+
+ } elsif ($line =~ m/^\[([a-z][a-z0-9_\-]+)\]\s*$/i) {
+ $section = $1;
+ $conf->{description} = $descr if $descr;
+ $descr = '';
+ $conf = $res->{snapshots}->{$section} = {};
next;
}
my $key = $1;
my $value = $2;
$conf->{$key} = $value;
+ } elsif ($line =~ m/^delete:\s*(.*\S)\s*$/) {
+ my $value = $1;
+ if ($section eq 'pending') {
+ $conf->{delete} = $value; # we parse this later
+ } else {
+ warn "vm $vmid - propertry 'delete' is only allowed in [PENDING]\n";
+ }
} elsif ($line =~ m/^([a-z][a-z_]*\d*):\s*(\S+)\s*$/) {
my $key = $1;
my $value = $2;
my $used_volids = {};
my $cleanup_config = sub {
- my ($cref, $snapname) = @_;
+ my ($cref, $pending, $snapname) = @_;
foreach my $key (keys %$cref) {
next if $key eq 'digest' || $key eq 'description' || $key eq 'snapshots' ||
- $key eq 'snapstate';
+ $key eq 'snapstate' || $key eq 'pending';
my $value = $cref->{$key};
+ if ($key eq 'delete') {
+ die "propertry 'delete' is only allowed in [PENDING]\n"
+ if !$pending;
+ # fixme: check syntax?
+ next;
+ }
eval { $value = check_type($key, $value); };
die "unable to parse value of '$key' - $@" if $@;
};
&$cleanup_config($conf);
+
+ &$cleanup_config($conf->{pending}, 1);
+
foreach my $snapname (keys %{$conf->{snapshots}}) {
- &$cleanup_config($conf->{snapshots}->{$snapname}, $snapname);
+ die "internal error" if $snapname eq 'pending';
+ &$cleanup_config($conf->{snapshots}->{$snapname}, undef, $snapname);
}
# remove 'unusedX' settings if we re-add a volume
}
foreach my $key (sort keys %$conf) {
- next if $key eq 'digest' || $key eq 'description' || $key eq 'snapshots';
+ next if $key eq 'digest' || $key eq 'description' || $key eq 'pending' || $key eq 'snapshots';
$raw .= "$key: $conf->{$key}\n";
}
return $raw;
};
my $raw = &$generate_raw_config($conf);
+
+ if (scalar(keys %{$conf->{pending}})){
+ $raw .= "\n[PENDING]\n";
+ $raw .= &$generate_raw_config($conf->{pending});
+ }
+
foreach my $snapname (sort keys %{$conf->{snapshots}}) {
$raw .= "\n[$snapname]\n";
$raw .= &$generate_raw_config($conf->{snapshots}->{$snapname});
$qmpclient->queue_cmd($vmid, $statuscb, 'query-status');
}
- $qmpclient->queue_execute();
+ $qmpclient->queue_execute(undef, 1);
foreach my $vmid (keys %$list) {
next if $opt_vmid && ($vmid ne $opt_vmid);
my $have_ovz = -f '/proc/vz/vestat';
+ my $q35 = machine_type_is_q35($conf);
+
push @$cmd, '/usr/bin/kvm';
push @$cmd, '-id', $vmid;
push @$cmd, '-daemonize';
- $pciaddr = print_pci_addr("piix3", $bridges);
- push @$devices, '-device', "piix3-usb-uhci,id=uhci$pciaddr.0x2";
+ if ($conf->{smbios1}) {
+ push @$cmd, '-smbios', "type=1,$conf->{smbios1}";
+ }
- my $use_usb2 = 0;
- for (my $i = 0; $i < $MAX_USB_DEVICES; $i++) {
- next if !$conf->{"usb$i"};
- $use_usb2 = 1;
+ push @$cmd, '-object', "iothread,id=iothread0" if $conf->{iothread};
+
+ if ($q35) {
+ # the q35 chipset support native usb2, so we enable usb controller
+ # by default for this machine type
+ push @$devices, '-readconfig', '/usr/share/qemu-server/pve-q35.cfg';
+ } else {
+ $pciaddr = print_pci_addr("piix3", $bridges);
+ push @$devices, '-device', "piix3-usb-uhci,id=uhci$pciaddr.0x2";
+
+ my $use_usb2 = 0;
+ for (my $i = 0; $i < $MAX_USB_DEVICES; $i++) {
+ next if !$conf->{"usb$i"};
+ $use_usb2 = 1;
+ }
+ # include usb device config
+ push @$devices, '-readconfig', '/usr/share/qemu-server/pve-usb.cfg' if $use_usb2;
}
- # include usb device config
- push @$devices, '-readconfig', '/usr/share/qemu-server/pve-usb.cfg' if $use_usb2;
my $vga = $conf->{vga};
$tablet = 0 if $vga =~ m/^serial\d+$/; # disable if we use serial terminal (no vga card)
}
- push @$devices, '-device', 'usb-tablet,id=tablet,bus=uhci.0,port=1' if $tablet;
+ push @$devices, '-device', print_tabletdevice_full($conf) if $tablet;
# host pci devices
for (my $i = 0; $i < $MAX_HOSTPCI_DEVICES; $i++) {
- my $d = parse_hostpci($conf->{"hostpci$i"});
- next if !$d;
- $pciaddr = print_pci_addr("hostpci$i", $bridges);
- my $rombar = $d->{rombar} && $d->{rombar} eq 'off' ? ",rombar=0" : "";
- push @$devices, '-device', "pci-assign,host=$d->{pciid},id=hostpci$i$pciaddr$rombar";
+ my $d = parse_hostpci($conf->{"hostpci$i"});
+ next if !$d;
+
+ my $pcie = $d->{pcie};
+ if($pcie){
+ die "q35 machine model is not enabled" if !$q35;
+ $pciaddr = print_pcie_addr("hostpci$i");
+ }else{
+ $pciaddr = print_pci_addr("hostpci$i", $bridges);
+ }
+
+ my $rombar = $d->{rombar} && $d->{rombar} eq 'off' ? ",rombar=0" : "";
+ my $driver = $d->{driver} && $d->{driver} eq 'vfio' ? "vfio-pci" : "pci-assign";
+ my $xvga = $d->{'x-vga'} && $d->{'x-vga'} eq 'on' ? ",x-vga=on" : "";
+ if ($xvga && $xvga ne '') {
+ push @$cpuFlags, 'kvm=off';
+ $vga = 'none';
+ }
+ $driver = "vfio-pci" if $xvga ne '';
+ my $pcidevices = $d->{pciid};
+ my $multifunction = 1 if @$pcidevices > 1;
+
+ my $j=0;
+ foreach my $pcidevice (@$pcidevices) {
+
+ my $id = "hostpci$i";
+ $id .= ".$j" if $multifunction;
+ my $addr = $pciaddr;
+ $addr .= ".$j" if $multifunction;
+ my $devicestr = "$driver,host=$pcidevice->{id}.$pcidevice->{function},id=$id$addr";
+
+ if($j == 0){
+ $devicestr .= "$rombar$xvga";
+ $devicestr .= ",multifunction=on" if $multifunction;
+ }
+
+ push @$devices, '-device', $devicestr;
+ $j++;
+ }
}
# usb devices
my $cores = $conf->{cores} || 1;
my $maxcpus = $conf->{maxcpus} if $conf->{maxcpus};
+ my $total_cores = $sockets * $cores;
+ my $allowed_cores = $cpuinfo->{cpus};
+
+ die "MAX $allowed_cores cores allowed per VM on this node\n"
+ if ($allowed_cores < $total_cores);
+
if ($maxcpus) {
push @$cmd, '-smp', "cpus=$cores,maxcpus=$maxcpus";
} else {
# push @$cmd, '-cpu', "$cpu,enforce";
push @$cmd, '-cpu', $cpu;
+ my $memory = $conf->{memory} || $defaults->{memory};
+ push @$cmd, '-m', $memory;
+
+ if ($conf->{numa}) {
+
+ my $numa_totalmemory = undef;
+ for (my $i = 0; $i < $MAX_NUMA; $i++) {
+ next if !$conf->{"numa$i"};
+ my $numa = parse_numa($conf->{"numa$i"});
+ next if !$numa;
+ # memory
+ die "missing numa node$i memory value\n" if !$numa->{memory};
+ my $numa_memory = $numa->{memory};
+ $numa_totalmemory += $numa_memory;
+ my $numa_object = "memory-backend-ram,id=ram-node$i,size=$numa_memory"."M";
+
+ # cpus
+ my $cpus_start = $numa->{cpus}->{start};
+ die "missing numa node$i cpus\n" if !defined($cpus_start);
+ my $cpus_end = $numa->{cpus}->{end} if defined($numa->{cpus}->{end});
+ my $cpus = $cpus_start;
+ if (defined($cpus_end)) {
+ $cpus .= "-$cpus_end";
+ die "numa node$i : cpu range $cpus is incorrect\n" if $cpus_end <= $cpus_start;
+ }
+
+ # hostnodes
+ my $hostnodes_start = $numa->{hostnodes}->{start};
+ if (defined($hostnodes_start)) {
+ my $hostnodes_end = $numa->{hostnodes}->{end} if defined($numa->{hostnodes}->{end});
+ my $hostnodes = $hostnodes_start;
+ if (defined($hostnodes_end)) {
+ $hostnodes .= "-$hostnodes_end";
+ die "host node $hostnodes range is incorrect\n" if $hostnodes_end <= $hostnodes_start;
+ }
+
+ my $hostnodes_end_range = defined($hostnodes_end) ? $hostnodes_end : $hostnodes_start;
+ for (my $i = $hostnodes_start; $i <= $hostnodes_end_range; $i++ ) {
+ die "host numa node$i don't exist\n" if ! -d "/sys/devices/system/node/node$i/";
+ }
+
+ # policy
+ my $policy = $numa->{policy};
+ die "you need to define a policy for hostnode $hostnodes\n" if !$policy;
+ $numa_object .= ",host-nodes=$hostnodes,policy=$policy";
+ }
+
+ push @$cmd, '-object', $numa_object;
+ push @$cmd, '-numa', "node,nodeid=$i,cpus=$cpus,memdev=ram-node$i";
+ }
+
+ die "total memory for NUMA nodes must be equal to vm memory\n"
+ if $numa_totalmemory && $numa_totalmemory != $memory;
+
+ #if no custom tology, we split memory and cores across numa nodes
+ if(!$numa_totalmemory) {
+
+ my $numa_memory = ($memory / $sockets) . "M";
+
+ for (my $i = 0; $i < $sockets; $i++) {
+
+ my $cpustart = ($cores * $i);
+ my $cpuend = ($cpustart + $cores - 1) if $cores && $cores > 1;
+ my $cpus = $cpustart;
+ $cpus .= "-$cpuend" if $cpuend;
+
+ push @$cmd, '-object', "memory-backend-ram,size=$numa_memory,id=ram-node$i";
+ push @$cmd, '-numa', "node,nodeid=$i,cpus=$cpus,memdev=ram-node$i";
+ }
+ }
+ }
+
push @$cmd, '-S' if $conf->{freeze};
# set keyboard layout
#push @$cmd, '-soundhw', $soundhw if $soundhw;
if($conf->{agent}) {
- my $qgasocket = qga_socket($vmid);
+ my $qgasocket = qmp_socket($vmid, 1);
my $pciaddr = print_pci_addr("qga0", $bridges);
push @$devices, '-chardev', "socket,path=$qgasocket,server,nowait,id=qga0";
push @$devices, '-device', "virtio-serial,id=qga0$pciaddr";
$spice_port = PVE::Tools::next_spice_port();
- push @$cmd, '-spice', "tls-port=${spice_port},addr=127.0.0.1,tls-ciphers=DES-CBC3-SHA,seamless-migration=on";
+ push @$devices, '-spice', "tls-port=${spice_port},addr=127.0.0.1,tls-ciphers=DES-CBC3-SHA,seamless-migration=on";
- push @$cmd, '-device', "virtio-serial,id=spice$pciaddr";
- push @$cmd, '-chardev', "spicevmc,id=vdagent,name=vdagent";
- push @$cmd, '-device', "virtserialport,chardev=vdagent,name=com.redhat.spice.0";
+ push @$devices, '-device', "virtio-serial,id=spice$pciaddr";
+ push @$devices, '-chardev', "spicevmc,id=vdagent,name=vdagent";
+ push @$devices, '-device', "virtserialport,chardev=vdagent,name=com.redhat.spice.0";
}
# enable balloon by default, unless explicitly disabled
my $ahcicontroller = {};
my $scsihw = defined($conf->{scsihw}) ? $conf->{scsihw} : $defaults->{scsihw};
+ # Add iscsi initiator name if available
+ if (my $initiator = get_initiator_name()) {
+ push @$devices, '-iscsi', "initiator-name=$initiator";
+ }
+
foreach_drive($conf, sub {
my ($ds, $drive) = @_;
$ahcicontroller->{$controller}=1;
}
- push @$devices, '-drive',print_drive_full($storecfg, $vmid, $drive);
- push @$devices, '-device',print_drivedevice_full($storecfg, $conf, $vmid, $drive, $bridges);
+ my $drive_cmd = print_drive_full($storecfg, $vmid, $drive);
+ push @$devices, '-drive',$drive_cmd;
+ push @$devices, '-device', print_drivedevice_full($storecfg, $conf, $vmid, $drive, $bridges);
});
- push @$cmd, '-m', $conf->{memory} || $defaults->{memory};
-
for (my $i = 0; $i < $MAX_NETS; $i++) {
next if !$conf->{"net$i"};
my $d = parse_net($conf->{"net$i"});
push @$devices, '-device', $netdevicefull;
}
- #bridges
- while (my ($k, $v) = each %$bridges) {
- $pciaddr = print_pci_addr("pci.$k");
- unshift @$devices, '-device', "pci-bridge,id=pci.$k,chassis_nr=$k$pciaddr" if $k > 0;
+ if (!$q35) {
+ # add pci bridges
+ while (my ($k, $v) = each %$bridges) {
+ $pciaddr = print_pci_addr("pci.$k");
+ unshift @$devices, '-device', "pci-bridge,id=pci.$k,chassis_nr=$k$pciaddr" if $k > 0;
+ }
}
-
# hack: virtio with fairsched is unreliable, so we do not use fairsched
# when the VM uses virtio devices.
if (!$use_virtio && $have_ovz) {
}
sub qmp_socket {
- my ($vmid) = @_;
- return "${var_run_tmpdir}/$vmid.qmp";
-}
-
-sub qga_socket {
- my ($vmid) = @_;
- return "${var_run_tmpdir}/$vmid.qga";
+ my ($vmid, $qga) = @_;
+ my $sockettype = $qga ? 'qga' : 'qmp';
+ return "${var_run_tmpdir}/$vmid.$sockettype";
}
sub pidfile_name {
my ($vmid) = @_;
my $res = vm_mon_cmd($vmid, 'query-pci');
-
my $devices = {};
foreach my $pcibus (@$res) {
foreach my $device (@{$pcibus->{devices}}) {
next if !$device->{'qdev_id'};
- $devices->{$device->{'qdev_id'}} = $device;
+ $devices->{$device->{'qdev_id'}} = 1;
+ }
+ }
+
+ my $resblock = vm_mon_cmd($vmid, 'query-block');
+ foreach my $block (@$resblock) {
+ if($block->{device} =~ m/^drive-(\S+)/){
+ $devices->{$1} = 1;
+ }
+ }
+
+ my $resmice = vm_mon_cmd($vmid, 'query-mice');
+ foreach my $mice (@$resmice) {
+ if ($mice->{name} eq 'QEMU HID Tablet') {
+ $devices->{tablet} = 1;
+ last;
}
}
return 1 if !check_running($vmid);
- if ($deviceid eq 'tablet') {
- my $devicefull = "usb-tablet,id=tablet,bus=uhci.0,port=1";
- qemu_deviceadd($vmid, $devicefull);
- return 1;
- }
+ my $q35 = machine_type_is_q35($conf);
return 1 if !$conf->{hotplug};
my $devices_list = vm_devices_list($vmid);
return 1 if defined($devices_list->{$deviceid});
- qemu_bridgeadd($storecfg, $conf, $vmid, $deviceid); #add bridge if we need it for the device
+ if ($deviceid eq 'tablet') {
+ qemu_deviceadd($vmid, print_tabletdevice_full($conf));
+ return 1;
+ }
+
+ qemu_add_pci_bridge($storecfg, $conf, $vmid, $deviceid); # add PCI bridge if we need it for the device
if ($deviceid =~ m/^(virtio)(\d+)$/) {
return undef if !qemu_driveadd($storecfg, $vmid, $device);
}
if ($deviceid =~ m/^(scsi)(\d+)$/) {
- return 1 if ($conf->{scsihw} && ($conf->{scsihw} !~ m/^lsi/)); #virtio-scsi not yet support hotplug
return undef if !qemu_findorcreatescsihw($storecfg,$conf, $vmid, $device);
return undef if !qemu_driveadd($storecfg, $vmid, $device);
my $devicefull = print_drivedevice_full($storecfg, $conf, $vmid, $device);
}
}
- if ($deviceid =~ m/^(pci\.)(\d+)$/) {
+
+ if (!$q35 && $deviceid =~ m/^(pci\.)(\d+)$/) {
my $bridgeid = $2;
my $pciaddr = print_pci_addr($deviceid);
my $devicefull = "pci-bridge,id=pci.$bridgeid,chassis_nr=$bridgeid$pciaddr";
return 1 if !check_running ($vmid);
- if ($deviceid eq 'tablet') {
- qemu_devicedel($vmid, $deviceid);
- return 1;
- }
-
return 1 if !$conf->{hotplug};
my $devices_list = vm_devices_list($vmid);
return 1 if !defined($devices_list->{$deviceid});
+ if ($deviceid eq 'tablet') {
+ qemu_devicedel($vmid, $deviceid);
+ return 1;
+ }
+
die "can't unplug bootdisk" if $conf->{bootdisk} && $conf->{bootdisk} eq $deviceid;
if ($deviceid =~ m/^(virtio)(\d+)$/) {
return 1;
}
-sub qemu_bridgeadd {
+sub qemu_add_pci_bridge {
my ($storecfg, $conf, $vmid, $device) = @_;
my $bridges = {};
- my $bridgeid = undef;
+
+ my $bridgeid;
+
print_pci_addr($device, $bridges);
while (my ($k, $v) = each %$bridges) {
$bridgeid = $k;
}
- return if !$bridgeid || $bridgeid < 1;
+ return if !defined($bridgeid) || $bridgeid < 1;
+
my $bridge = "pci.$bridgeid";
my $devices_list = vm_devices_list($vmid);
- if(!defined($devices_list->{$bridge})) {
+ if (!defined($devices_list->{$bridge})) {
return undef if !vm_deviceplug($storecfg, $conf, $vmid, $bridge);
}
+
return 1;
}
sub qemu_cpu_hotplug {
my ($vmid, $conf, $cores) = @_;
- die "new cores config is not defined" if !$cores;
- die "you can't add more cores than maxcpus"
- if $conf->{maxcpus} && ($cores > $conf->{maxcpus});
- return if !check_running($vmid);
+ my $sockets = $conf->{sockets} || 1;
+ die "cpu hotplug only works with one socket\n"
+ if $sockets > 1;
- my $currentcores = $conf->{cores} if $conf->{cores};
- die "current cores is not defined" if !$currentcores;
- die "maxcpus is not defined" if !$conf->{maxcpus};
- raise_param_exc({ 'cores' => "online cpu unplug is not yet possible" })
- if($cores < $currentcores);
+ die "maxcpus is not defined\n"
+ if !$conf->{maxcpus};
+
+ die "you can't add more cores than maxcpus\n"
+ if $cores > $conf->{maxcpus};
+
+ my $currentcores = $conf->{cores} || 1;
+ die "online cpu unplug is not yet possible\n"
+ if $cores < $currentcores;
my $currentrunningcores = vm_mon_cmd($vmid, "query-cpus");
- raise_param_exc({ 'cores' => "cores number if running vm is different than configuration" })
- if scalar (@{$currentrunningcores}) != $currentcores;
+ die "cores number if running vm is different than configuration\n"
+ if scalar(@{$currentrunningcores}) != $currentcores;
- for(my $i = $currentcores; $i < $cores; $i++) {
+ for (my $i = $currentcores; $i < $cores; $i++) {
vm_mon_cmd($vmid, "cpu-add", id => int($i));
}
}
vm_mon_cmd($vmid, "delete-drive-snapshot", device => $deviceid, name => $snap);
}
-sub qga_freezefs {
- my ($vmid) = @_;
-
- #need to impplement call to qemu-ga
-}
-
-sub qga_unfreezefs {
- my ($vmid) = @_;
-
- #need to impplement call to qemu-ga
-}
-
sub set_migration_caps {
my ($vmid) = @_;
vm_mon_cmd_nocheck($vmid, "migrate-set-capabilities", capabilities => $cap_ref);
}
+# hotplug changes in [PENDING]
+# $selection hash can be used to only apply specified options, for
+# example: { cores => 1 } (only apply changed 'cores')
+# $errors ref is used to return error messages
+sub vmconfig_hotplug_pending {
+ my ($vmid, $conf, $storecfg, $selection, $errors) = @_;
+
+ my $defaults = load_defaults();
+
+ # commit values which do not have any impact on running VM first
+ # Note: those option cannot raise errors, we we do not care about
+ # $selection and always apply them.
+
+ my $add_error = sub {
+ my ($opt, $msg) = @_;
+ $errors->{$opt} = "hotplug problem - $msg";
+ };
+
+ my $changes = 0;
+ foreach my $opt (keys %{$conf->{pending}}) { # add/change
+ if ($opt eq 'name' || $opt eq 'hotplug' || $opt eq 'onboot' || $opt eq 'shares') {
+ $conf->{$opt} = $conf->{pending}->{$opt};
+ delete $conf->{pending}->{$opt};
+ $changes = 1;
+ }
+ }
+
+ if ($changes) {
+ update_config_nolock($vmid, $conf, 1);
+ $conf = load_config($vmid); # update/reload
+ }
+
+ my $hotplug = defined($conf->{hotplug}) ? $conf->{hotplug} : $defaults->{hotplug};
+
+ my @delete = PVE::Tools::split_list($conf->{pending}->{delete});
+ foreach my $opt (@delete) {
+ next if $selection && !$selection->{$opt};
+ eval {
+ if ($opt eq 'tablet') {
+ die "skip\n" if !$hotplug;
+ if ($defaults->{tablet}) {
+ vm_deviceplug($storecfg, $conf, $vmid, $opt);
+ } else {
+ vm_deviceunplug($vmid, $conf, $opt);
+ }
+ } elsif ($opt eq 'cores') {
+ die "skip\n" if !$hotplug;
+ qemu_cpu_hotplug($vmid, $conf, 1);
+ } else {
+ die "skip\n";
+ }
+ };
+ if (my $err = $@) {
+ &$add_error($opt, $err) if $err ne "skip\n";
+ } else {
+ # save new config if hotplug was successful
+ delete $conf->{$opt};
+ vmconfig_undelete_pending_option($conf, $opt);
+ update_config_nolock($vmid, $conf, 1);
+ $conf = load_config($vmid); # update/reload
+ }
+ }
+
+ foreach my $opt (keys %{$conf->{pending}}) {
+ next if $selection && !$selection->{$opt};
+ my $value = $conf->{pending}->{$opt};
+ eval {
+ if ($opt eq 'tablet') {
+ die "skip\n" if !$hotplug;
+ if ($value == 1) {
+ vm_deviceplug($storecfg, $conf, $vmid, $opt);
+ } elsif ($value == 0) {
+ vm_deviceunplug($vmid, $conf, $opt);
+ }
+ } elsif ($opt eq 'cores') {
+ die "skip\n" if !$hotplug;
+ qemu_cpu_hotplug($vmid, $conf, $value);
+ } elsif ($opt eq 'balloon') {
+ die "skip\n" if !(defined($conf->{shares}) && ($conf->{shares} == 0));
+ # allow manual ballooning if shares is set to zero
+ my $balloon = $conf->{pending}->{balloon} || $conf->{memory} || $defaults->{memory};
+ vm_mon_cmd($vmid, "balloon", value => $balloon*1024*1024);
+ } else {
+ die "skip\n"; # skip non-hot-pluggable options
+ }
+ };
+ if (my $err = $@) {
+ &$add_error($opt, $err) if $err ne "skip\n";
+ } else {
+ # save new config if hotplug was successful
+ $conf->{$opt} = $value;
+ delete $conf->{pending}->{$opt};
+ update_config_nolock($vmid, $conf, 1);
+ $conf = load_config($vmid); # update/reload
+ }
+ }
+}
+
+sub vmconfig_apply_pending {
+ my ($vmid, $conf, $storecfg) = @_;
+
+ # cold plug
+
+ my @delete = PVE::Tools::split_list($conf->{pending}->{delete});
+ foreach my $opt (@delete) { # delete
+ die "internal error" if $opt =~ m/^unused/;
+ $conf = load_config($vmid); # update/reload
+ if (!defined($conf->{$opt})) {
+ vmconfig_undelete_pending_option($conf, $opt);
+ update_config_nolock($vmid, $conf, 1);
+ } elsif (valid_drivename($opt)) {
+ vmconfig_register_unused_drive($storecfg, $vmid, $conf, parse_drive($opt, $conf->{$opt}));
+ vmconfig_undelete_pending_option($conf, $opt);
+ delete $conf->{$opt};
+ update_config_nolock($vmid, $conf, 1);
+ } else {
+ vmconfig_undelete_pending_option($conf, $opt);
+ delete $conf->{$opt};
+ update_config_nolock($vmid, $conf, 1);
+ }
+ }
+
+ $conf = load_config($vmid); # update/reload
+
+ foreach my $opt (keys %{$conf->{pending}}) { # add/change
+ $conf = load_config($vmid); # update/reload
+
+ if (defined($conf->{$opt}) && ($conf->{$opt} eq $conf->{pending}->{$opt})) {
+ # skip if nothing changed
+ } elsif (valid_drivename($opt)) {
+ vmconfig_register_unused_drive($storecfg, $vmid, $conf, parse_drive($opt, $conf->{$opt}))
+ if defined($conf->{$opt});
+ $conf->{$opt} = $conf->{pending}->{$opt};
+ } else {
+ $conf->{$opt} = $conf->{pending}->{$opt};
+ }
+
+ delete $conf->{pending}->{$opt};
+ update_config_nolock($vmid, $conf, 1);
+ }
+}
+
sub vm_start {
my ($storecfg, $vmid, $statefile, $skiplock, $migratedfrom, $paused, $forcemachine, $spice_ticket) = @_;
die "VM $vmid already running\n" if check_running($vmid, undef, $migratedfrom);
+ if (!$statefile && scalar(keys %{$conf->{pending}})) {
+ vmconfig_apply_pending($vmid, $conf, $storecfg);
+ $conf = load_config($vmid); # update/reload
+ }
+
my $defaults = load_defaults();
# set environment variable useful inside network script
for (my $i = 0; $i < $MAX_HOSTPCI_DEVICES; $i++) {
my $d = parse_hostpci($conf->{"hostpci$i"});
next if !$d;
- my $info = pci_device_info("0000:$d->{pciid}");
- die "IOMMU not present\n" if !check_iommu_support();
- die "no pci device info for device '$d->{pciid}'\n" if !$info;
- die "can't unbind pci device '$d->{pciid}'\n" if !pci_dev_bind_to_stub($info);
- die "can't reset pci device '$d->{pciid}'\n" if !pci_dev_reset($info);
+ my $pcidevices = $d->{pciid};
+ foreach my $pcidevice (@$pcidevices) {
+ my $pciid = $pcidevice->{id}.".".$pcidevice->{function};
+
+ my $info = pci_device_info("0000:$pciid");
+ die "IOMMU not present\n" if !check_iommu_support();
+ die "no pci device info for device '$pciid'\n" if !$info;
+
+ if ($d->{driver} && $d->{driver} eq "vfio") {
+ die "can't unbind/bind pci group to vfio '$pciid'\n" if !pci_dev_group_bind_to_vfio($pciid);
+ } else {
+ die "can't unbind/bind to stub pci device '$pciid'\n" if !pci_dev_bind_to_stub($info);
+ }
+
+ die "can't reset pci device '$pciid'\n" if $info->{has_fl_reset} and !pci_dev_reset($info);
+ }
}
PVE::Storage::activate_volumes($storecfg, $vollist);
if ($migratedfrom) {
eval {
- PVE::QemuServer::set_migration_caps($vmid);
+ set_migration_caps($vmid);
};
warn $@ if $@;
if ($spice_port) {
print "spice listens on port $spice_port\n";
if ($spice_ticket) {
- PVE::QemuServer::vm_mon_cmd_nocheck($vmid, "set_password", protocol => 'spice', password => $spice_ticket);
- PVE::QemuServer::vm_mon_cmd_nocheck($vmid, "expire_password", protocol => 'spice', time => "+30");
+ vm_mon_cmd_nocheck($vmid, "set_password", protocol => 'spice', password => $spice_ticket);
+ vm_mon_cmd_nocheck($vmid, "expire_password", protocol => 'spice', time => "+30");
}
}
eval {
die "VM $vmid not running\n" if !check_running($vmid, $nocheck);
my $sname = qmp_socket($vmid);
- if (-e $sname) {
+ if (-e $sname) { # test if VM is reasonambe new and supports qmp/qga
my $qmpclient = PVE::QMPClient->new();
$res = $qmpclient->cmd($vmid, $cmd, $timeout);
eval {
if ($shutdown) {
- $nocheck ? vm_mon_cmd_nocheck($vmid, "system_powerdown") : vm_mon_cmd($vmid, "system_powerdown");
-
+ if (!$nocheck && $conf->{agent}) {
+ vm_qmp_command($vmid, { execute => "guest-shutdown" }, $nocheck);
+ } else {
+ vm_qmp_command($vmid, { execute => "system_powerdown" }, $nocheck);
+ }
} else {
- $nocheck ? vm_mon_cmd_nocheck($vmid, "quit") : vm_mon_cmd($vmid, "quit");
+ vm_qmp_command($vmid, { execute => "quit" }, $nocheck);
}
};
my $err = $@;
return -d $testdir;
}
+sub pci_dev_bind_to_vfio {
+ my ($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";
+ my $info = pci_device_info($1);
+ pci_dev_bind_to_vfio($info) || die "Cannot bind $pciid to vfio\n";
+ }
+
+ return 1;
+}
+
sub print_pci_addr {
my ($id, $bridges) = @_;
vga1 => { bus => 0, addr => 24 },
vga2 => { bus => 0, addr => 25 },
vga3 => { bus => 0, addr => 26 },
+ hostpci2 => { bus => 0, addr => 27 },
+ hostpci3 => { bus => 0, addr => 28 },
#addr29 : usb-host (pve-usb.cfg)
'pci.1' => { bus => 0, addr => 30 },
'pci.2' => { bus => 0, addr => 31 },
}
+sub print_pcie_addr {
+ my ($id) = @_;
+
+ my $res = '';
+ my $devices = {
+ hostpci0 => { bus => "ich9-pcie-port-1", addr => 0 },
+ hostpci1 => { bus => "ich9-pcie-port-2", addr => 0 },
+ hostpci2 => { bus => "ich9-pcie-port-3", addr => 0 },
+ hostpci3 => { bus => "ich9-pcie-port-4", addr => 0 },
+ };
+
+ if (defined($devices->{$id}->{bus}) && defined($devices->{$id}->{addr})) {
+ my $addr = sprintf("0x%x", $devices->{$id}->{addr});
+ my $bus = $devices->{$id}->{bus};
+ $res = ",bus=$bus,addr=$addr";
+ }
+ return $res;
+
+}
+
# vzdump restore implementaion
sub tar_archive_read_firstfile {
my ($dev_id, $size, $devname) = ($1, $2, $3);
$devinfo->{$devname} = { size => $size, dev_id => $dev_id };
} elsif ($line =~ m/^CTIME: /) {
- # we correctly received the vma config, so we can disable
+ # we correctly received the vma config, so we can disable
# the timeout now for disk allocation (set to 10 minutes, so
# that we always timeout if something goes wrong)
alarm(600);
my $storecfg = cfs_read_file('storage.cfg');
# destroy existing data - keep empty config
- my $vmcfgfn = PVE::QemuServer::config_file($vmid);
+ my $vmcfgfn = config_file($vmid);
destroy_vm($storecfg, $vmid, 1) if -f $vmcfgfn;
my $tocmd = "/usr/lib/qemu-server/qmextract";
die "missing snapshot lock\n"
if !($conf->{lock} && $conf->{lock} eq 'snapshot');
+ my $has_machine_config = defined($conf->{machine});
+
my $snap = $conf->{snapshots}->{$snapname};
die "snapshot '$snapname' does not exist\n" if !defined($snap);
my $newconf = &$snapshot_apply_config($conf, $snap);
+ delete $newconf->{machine} if !$has_machine_config;
+
$newconf->{parent} = $snapname;
update_config_nolock($vmid, $newconf, 1);
};
sub snapshot_create {
- my ($vmid, $snapname, $save_vmstate, $freezefs, $comment) = @_;
+ my ($vmid, $snapname, $save_vmstate, $comment) = @_;
my $snap = &$snapshot_prepare($vmid, $snapname, $save_vmstate, $comment);
- $freezefs = $save_vmstate = 0 if !$snap->{vmstate}; # vm is not running
+ $save_vmstate = 0 if !$snap->{vmstate}; # vm is not running
- my $drivehash = {};
+ my $config = load_config($vmid);
my $running = check_running($vmid);
+ my $freezefs = $running && $config->{agent};
+ $freezefs = 0 if $snap->{vmstate}; # not needed if we save RAM
+
+ my $drivehash = {};
+
+ if ($freezefs) {
+ eval { vm_mon_cmd($vmid, "guest-fsfreeze-freeze"); };
+ warn "guest-fsfreeze-freeze problems - $@" if $@;
+ }
+
eval {
# create internal snapshots of all drives
}
};
- qga_freezefs($vmid) if $running && $freezefs;
-
foreach_drive($snap, sub {
my ($ds, $drive) = @_;
};
my $err = $@;
- eval { qga_unfreezefs($vmid) if $running && $freezefs; };
- warn $@ if $@;
+ if ($running) {
+ eval { vm_mon_cmd($vmid, "savevm-end") };
+ warn $@ if $@;
- eval { vm_mon_cmd($vmid, "savevm-end") if $running; };
- warn $@ if $@;
+ if ($freezefs) {
+ eval { vm_mon_cmd($vmid, "guest-fsfreeze-thaw"); };
+ warn "guest-fsfreeze-thaw problems - $@" if $@;
+ }
+
+ # savevm-end is async, we need to wait
+ for (;;) {
+ my $stat = vm_mon_cmd_nocheck($vmid, "query-savevm");
+ if (!$stat->{bytes}) {
+ last;
+ } else {
+ print "savevm not yet finished\n";
+ sleep(1);
+ next;
+ }
+ }
+ }
if ($err) {
warn "snapshot create failed: starting cleanup\n";
die "snapshot '$snapname' does not exist\n" if !defined($snap);
# remove parent refs
- &$unlink_parent($conf, $snap->{parent});
- foreach my $sn (keys %{$conf->{snapshots}}) {
- next if $sn eq $snapname;
- &$unlink_parent($conf->{snapshots}->{$sn}, $snap->{parent});
+ if (!$prepare) {
+ &$unlink_parent($conf, $snap->{parent});
+ foreach my $sn (keys %{$conf->{snapshots}}) {
+ next if $sn eq $snapname;
+ &$unlink_parent($conf->{snapshots}->{$sn}, $snap->{parent});
+ }
}
if ($remove_drive) {
}
sub qemu_drive_mirror {
- my ($vmid, $drive, $dst_volid, $vmiddst, $maxwait) = @_;
+ my ($vmid, $drive, $dst_volid, $vmiddst) = @_;
- my $count = 1;
+ my $count = 0;
my $old_len = 0;
my $frozen = undef;
+ my $maxwait = 120;
my $storecfg = PVE::Storage::config();
- my ($dst_storeid, $dst_volname) = PVE::Storage::parse_volume_id($dst_volid, 1);
+ my ($dst_storeid, $dst_volname) = PVE::Storage::parse_volume_id($dst_volid);
- if ($dst_storeid) {
- my $dst_scfg = PVE::Storage::storage_config($storecfg, $dst_storeid);
+ my $dst_scfg = PVE::Storage::storage_config($storecfg, $dst_storeid);
- my $format;
- if ($dst_volname =~ m/\.(raw|qcow2)$/){
- $format = $1;
- }
+ my $format;
+ if ($dst_volname =~ m/\.(raw|qcow2)$/){
+ $format = $1;
+ }
- my $dst_path = PVE::Storage::path($storecfg, $dst_volid);
+ my $dst_path = PVE::Storage::path($storecfg, $dst_volid);
- if ($format) {
- #fixme : sometime drive-mirror timeout, but works fine after.
- # (I have see the problem with big volume > 200GB), so we need to eval
- eval { vm_mon_cmd($vmid, "drive-mirror", timeout => 10, device => "drive-$drive", mode => "existing",
- sync => "full", target => $dst_path, format => $format); };
- } else {
- eval { vm_mon_cmd($vmid, "drive-mirror", timeout => 10, device => "drive-$drive", mode => "existing",
- sync => "full", target => $dst_path); };
- }
+ my $opts = { timeout => 10, device => "drive-$drive", mode => "existing", sync => "full", target => $dst_path };
+ $opts->{format} = $format if $format;
- eval {
- while (1) {
- my $stats = vm_mon_cmd($vmid, "query-block-jobs");
- my $stat = @$stats[0];
- die "mirroring job seem to have die. Maybe do you have bad sectors?" if !$stat;
- die "error job is not mirroring" if $stat->{type} ne "mirror";
-
- my $transferred = $stat->{offset};
- my $total = $stat->{len};
+ #fixme : sometime drive-mirror timeout, but works fine after.
+ # (I have see the problem with big volume > 200GB), so we need to eval
+ eval { vm_mon_cmd($vmid, "drive-mirror", %$opts); };
+ # ignore errors here
+
+ eval {
+ while (1) {
+ my $stats = vm_mon_cmd($vmid, "query-block-jobs");
+ my $stat = @$stats[0];
+ die "mirroring job seem to have die. Maybe do you have bad sectors?" if !$stat;
+ die "error job is not mirroring" if $stat->{type} ne "mirror";
+
+ my $busy = $stat->{busy};
+
+ if (my $total = $stat->{len}) {
+ my $transferred = $stat->{offset} || 0;
my $remaining = $total - $transferred;
my $percent = sprintf "%.2f", ($transferred * 100 / $total);
- print "transferred: $transferred bytes remaining: $remaining bytes total: $total bytes progression: $percent %\n";
+ print "transferred: $transferred bytes remaining: $remaining bytes total: $total bytes progression: $percent % busy: $busy\n";
+ }
+
+ if ($stat->{len} == $stat->{offset}) {
+ if ($busy eq 'false') {
- last if ($stat->{len} == $stat->{offset});
- if ($old_len == $stat->{offset}) {
- if ($maxwait && $count > $maxwait) {
- # if writes to disk occurs the disk needs to be freezed
- # to be able to complete the migration
- vm_suspend($vmid,1);
- $count = 0;
- $frozen = 1;
- } else {
- $count++ unless $frozen;
- }
- } elsif ($frozen) {
- vm_resume($vmid,1);
- $count = 0;
+ last if $vmiddst != $vmid;
+
+ # try to switch the disk if source and destination are on the same guest
+ eval { vm_mon_cmd($vmid, "block-job-complete", device => "drive-$drive") };
+ last if !$@;
+ die $@ if $@ !~ m/cannot be completed/;
}
- $old_len = $stat->{offset};
- sleep 1;
- }
- if ($vmiddst == $vmid) {
- # switch the disk if source and destination are on the same guest
- vm_mon_cmd($vmid, "block-job-complete", device => "drive-$drive");
+ if ($count > $maxwait) {
+ # if too much writes to disk occurs at the end of migration
+ #the disk needs to be freezed to be able to complete the migration
+ vm_suspend($vmid,1);
+ $frozen = 1;
+ }
+ $count ++
}
- };
- if (my $err = $@) {
- eval { vm_mon_cmd($vmid, "block-job-cancel", device => "drive-$drive"); };
- die "mirroring error: $err";
+ $old_len = $stat->{offset};
+ sleep 1;
}
- if ($vmiddst != $vmid) {
- # if we clone a disk for a new target vm, we don't switch the disk
- vm_mon_cmd($vmid, "block-job-cancel", device => "drive-$drive");
+ vm_resume($vmid, 1) if $frozen;
+
+ };
+ my $err = $@;
+
+ my $cancel_job = sub {
+ vm_mon_cmd($vmid, "block-job-cancel", device => "drive-$drive");
+ while (1) {
+ my $stats = vm_mon_cmd($vmid, "query-block-jobs");
+ my $stat = @$stats[0];
+ last if !$stat;
+ sleep 1;
}
+ };
+
+ if ($err) {
+ eval { &$cancel_job(); };
+ die "mirroring error: $err";
+ }
+
+ if ($vmiddst != $vmid) {
+ # if we clone a disk for a new target vm, we don't switch the disk
+ &$cancel_job(); # so we call block-job-cancel
}
}
if (!$full) {
print "create linked clone of drive $drivename ($drive->{file})\n";
- $newvolid = PVE::Storage::vdisk_clone($storecfg, $drive->{file}, $newvmid);
+ $newvolid = PVE::Storage::vdisk_clone($storecfg, $drive->{file}, $newvmid, $snapname);
push @$newvollist, $newvolid;
} else {
my ($storeid, $volname) = PVE::Storage::parse_volume_id($drive->{file});
my ($vmid) = @_;
my $cmd = { execute => 'query-machines', arguments => {} };
- my $res = PVE::QemuServer::vm_qmp_command($vmid, $cmd);
+ my $res = vm_qmp_command($vmid, $cmd);
my ($current, $default);
foreach my $e (@$res) {
return $current || $default || 'pc';
}
+sub lspci {
+
+ my $devices = {};
+
+ dir_glob_foreach("$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;
+ });
+
+ return $devices;
+}
+
1;