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;
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 $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.
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) = @_;
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 {
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;
}
+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 $res;
+}
+
+PVE::JSONSchema::register_format('pve-qm-smbios1', \&verify_smbios1);
+sub verify_smbios1 {
+ my ($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) = @_;
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"};
- $use_usb2 = 1;
+ if ($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" : "";
+ $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;
+ $j++;
+ }
}
# usb devices
push @$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 && $have_ovz) {
return 1 if !check_running($vmid);
+ my $q35 = machine_type_is_q35($conf);
+
if ($deviceid eq 'tablet') {
- my $devicefull = "usb-tablet,id=tablet,bus=uhci.0,port=1";
- qemu_deviceadd($vmid, $devicefull);
+ qemu_deviceadd($vmid, print_tabletdevice_full($conf));
return 1;
}
}
}
- if ($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";
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 !pci_dev_reset($info);
+ }
}
PVE::Storage::activate_volumes($storecfg, $vollist);
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, $bridges) = @_;
}
+sub print_pcie_addr {
+ my ($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 {
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;