[PATCH virt-v2v v2 0/2] input: -i vmx: Add support for NVMe devices
https://bugzilla.redhat.com/show_bug.cgi?id=2070530 Version 1 was here: https://listman.redhat.com/archives/libguestfs/2022-April/028560.html Version 2 fixes the bogus mapping of source device names as pointed out in Laszlo's review (and no longer touches target device names). The rest of the patch is the same. Rich.
This neutral refactoring changes the existing source device name calculation so that it can handle (in the next commit) NVMe device names. Existing source device names such as "hda" or "sda" are composed of a prefix ("hd", "sd") + alphabetic device index ("a"). NVMe devices have a namespace after the prefix and use a numeric index ("nvme0n1" == "nvme" + "0" + "n1"). By refactoring the calculation into a function we will be able to handle both. --- convert/convert_linux.ml | 22 ++++++++++++---------- 1 file changed, 12 insertions(+), 10 deletions(-) diff --git a/convert/convert_linux.ml b/convert/convert_linux.ml index 58a2a2a84d..d9397cfad1 100644 --- a/convert/convert_linux.ml +++ b/convert/convert_linux.ml @@ -1022,16 +1022,18 @@ let convert (g : G.guestfs) source inspect keep_serial_console _ = let map = List.mapi ( - fun i disk -> - let block_prefix_before_conversion = - match disk.s_controller with - | Some Source_IDE -> ide_block_prefix - | Some (Source_virtio_SCSI | Source_SCSI | Source_SATA) -> "sd" - | Some Source_virtio_blk -> "vd" - | None -> - (* This is basically a guess. It assumes the source used IDE. *) - ide_block_prefix in - let source_dev = block_prefix_before_conversion ^ drive_name i in + fun i { s_controller } -> + let device_name_before_conversion i = + match s_controller with + | None (* If we don't have a controller, guess. Assumes the + * source used IDE. + *) + | Some Source_IDE -> + ide_block_prefix ^ drive_name i + | Some (Source_virtio_SCSI | Source_SCSI | Source_SATA) -> + "sd" ^ drive_name i + | Some Source_virtio_blk -> "vd" ^ drive_name i in + let source_dev = device_name_before_conversion i in let target_dev = block_prefix_after_conversion ^ drive_name i in source_dev, target_dev ) source.s_disks in -- 2.35.1
On 04/08/22 13:11, Richard W.M. Jones wrote:
This neutral refactoring changes the existing source device name calculation so that it can handle (in the next commit) NVMe device names.
Existing source device names such as "hda" or "sda" are composed of a prefix ("hd", "sd") + alphabetic device index ("a"). NVMe devices have a namespace after the prefix and use a numeric index ("nvme0n1" == "nvme" + "0" + "n1"). By refactoring the calculation into a function we will be able to handle both. --- convert/convert_linux.ml | 22 ++++++++++++---------- 1 file changed, 12 insertions(+), 10 deletions(-)
diff --git a/convert/convert_linux.ml b/convert/convert_linux.ml index 58a2a2a84d..d9397cfad1 100644 --- a/convert/convert_linux.ml +++ b/convert/convert_linux.ml @@ -1022,16 +1022,18 @@ let convert (g : G.guestfs) source inspect keep_serial_console _ =
let map = List.mapi ( - fun i disk -> - let block_prefix_before_conversion = - match disk.s_controller with - | Some Source_IDE -> ide_block_prefix - | Some (Source_virtio_SCSI | Source_SCSI | Source_SATA) -> "sd" - | Some Source_virtio_blk -> "vd" - | None -> - (* This is basically a guess. It assumes the source used IDE. *) - ide_block_prefix in - let source_dev = block_prefix_before_conversion ^ drive_name i in + fun i { s_controller } -> + let device_name_before_conversion i = + match s_controller with + | None (* If we don't have a controller, guess. Assumes the + * source used IDE. + *) + | Some Source_IDE -> + ide_block_prefix ^ drive_name i + | Some (Source_virtio_SCSI | Source_SCSI | Source_SATA) -> + "sd" ^ drive_name i + | Some Source_virtio_blk -> "vd" ^ drive_name i in + let source_dev = device_name_before_conversion i in let target_dev = block_prefix_after_conversion ^ drive_name i in source_dev, target_dev ) source.s_disks in
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
We model NVMe devices in the source hypervisor. We currently assume that no one is using the namespaces feature of NVMe, ie. that each source device will appear in a Linux guest as /dev/nvme0n1, /dev/nvme1n1, etc. We could fix this if it is a problem, but it requires adjusting the current assumption for removable devices that slots are simple integers. The devices are mapped to virtio-blk, so in the target the device name has to change from /dev/nvme0 to /dev/vda (etc.) Reported-by: Ming Xie Fixes: https://bugzilla.redhat.com/show_bug.cgi?id=2070530 --- convert/convert_linux.ml | 5 +- convert/target_bus_assignment.ml | 4 +- input/parse_domain_from_vmx.ml | 26 +++++++++- lib/types.ml | 3 +- lib/types.mli | 2 +- tests/test-v2v-i-vmx-6.expected | 22 +++++++++ tests/test-v2v-i-vmx-6.vmx | 84 ++++++++++++++++++++++++++++++++ tests/test-v2v-i-vmx.sh | 4 +- 8 files changed, 141 insertions(+), 9 deletions(-) diff --git a/convert/convert_linux.ml b/convert/convert_linux.ml index d9397cfad1..255ba608ca 100644 --- a/convert/convert_linux.ml +++ b/convert/convert_linux.ml @@ -1032,7 +1032,10 @@ let convert (g : G.guestfs) source inspect keep_serial_console _ = ide_block_prefix ^ drive_name i | Some (Source_virtio_SCSI | Source_SCSI | Source_SATA) -> "sd" ^ drive_name i - | Some Source_virtio_blk -> "vd" ^ drive_name i in + | Some Source_virtio_blk -> "vd" ^ drive_name i + | Some Source_NVME -> + (* For NVMe assume no one is using namespaces. *) + sprintf "nvme%dn1" i in let source_dev = device_name_before_conversion i in let target_dev = block_prefix_after_conversion ^ drive_name i in source_dev, target_dev diff --git a/convert/target_bus_assignment.ml b/convert/target_bus_assignment.ml index 4b56a6e171..f8675cf29b 100644 --- a/convert/target_bus_assignment.ml +++ b/convert/target_bus_assignment.ml @@ -73,8 +73,8 @@ let rec target_bus_assignment source_disks source_removables guestcaps = | None -> ide_bus (* Wild guess, but should be safe. *) | Some Source_virtio_blk -> virtio_blk_bus | Some Source_IDE -> ide_bus - | Some (Source_virtio_SCSI | Source_SCSI | Source_SATA) -> - scsi_bus in + | Some (Source_virtio_SCSI | Source_SCSI | Source_SATA | + Source_NVME) -> scsi_bus in match r.s_removable_slot with | None -> diff --git a/input/parse_domain_from_vmx.ml b/input/parse_domain_from_vmx.ml index 730f1177f7..b812edeb81 100644 --- a/input/parse_domain_from_vmx.ml +++ b/input/parse_domain_from_vmx.ml @@ -103,6 +103,7 @@ let remote_file_exists uri path = let rec find_disks vmx vmx_source = find_scsi_disks vmx vmx_source + @ find_nvme_disks vmx vmx_source @ find_ide_disks vmx vmx_source (* Find all SCSI hard disks. @@ -129,6 +130,27 @@ and find_scsi_disks vmx vmx_source = get_scsi_controller_target is_scsi_controller_target scsi_device_types scsi_controller +(* Find all NVMe hard disks. + * + * In the VMX file: + * nvme0.pcislotnumber = "192" + * nvme0:0.fileName = "guest.vmdk" + *) +and find_nvme_disks vmx vmx_source = + let get_nvme_controller_target ns = + sscanf ns "nvme%d:%d" (fun c t -> c, t) + in + let is_nvme_controller_target ns = + try ignore (get_nvme_controller_target ns); true + with Scanf.Scan_failure _ | End_of_file | Failure _ -> false + in + let nvme_device_types = [ None ] in + let nvme_controller = Source_NVME in + + find_hdds vmx vmx_source + get_nvme_controller_target is_nvme_controller_target + nvme_device_types nvme_controller + (* Find all IDE hard disks. * * In the VMX file: @@ -153,12 +175,12 @@ and find_ide_disks vmx vmx_source = and find_hdds vmx vmx_source get_controller_target is_controller_target device_types controller = - (* Find namespaces matching '(ide|scsi)X:Y' with suitable deviceType. *) + (* Find namespaces matching '(ide|scsi|nvme)X:Y' with suitable deviceType. *) let hdds = Parse_vmx.select_namespaces ( function | [ns] -> - (* Check the namespace is '(ide|scsi)X:Y' *) + (* Check the namespace is '(ide|scsi|nvme)X:Y' *) if not (is_controller_target ns) then false else ( (* Check the deviceType is one we are looking for. *) diff --git a/lib/types.ml b/lib/types.ml index 5804c7c79f..92ed0e5251 100644 --- a/lib/types.ml +++ b/lib/types.ml @@ -56,7 +56,7 @@ and source_disk = { s_disk_id : int; s_controller : s_controller option; } -and s_controller = Source_IDE | Source_SATA | Source_SCSI | +and s_controller = Source_IDE | Source_SATA | Source_SCSI | Source_NVME | Source_virtio_blk | Source_virtio_SCSI and source_removable = { s_removable_type : s_removable_type; @@ -194,6 +194,7 @@ and string_of_controller = function | Source_IDE -> "ide" | Source_SATA -> "sata" | Source_SCSI -> "scsi" + | Source_NVME -> "nvme" | Source_virtio_blk -> "virtio-blk" | Source_virtio_SCSI -> "virtio-scsi" diff --git a/lib/types.mli b/lib/types.mli index dd2fe5925d..37238cd755 100644 --- a/lib/types.mli +++ b/lib/types.mli @@ -103,7 +103,7 @@ and source_disk = { } (** A source disk. *) -and s_controller = Source_IDE | Source_SATA | Source_SCSI | +and s_controller = Source_IDE | Source_SATA | Source_SCSI | Source_NVME | Source_virtio_blk | Source_virtio_SCSI (** Source disk controller. *) diff --git a/tests/test-v2v-i-vmx-6.expected b/tests/test-v2v-i-vmx-6.expected new file mode 100644 index 0000000000..1793b6b953 --- /dev/null +++ b/tests/test-v2v-i-vmx-6.expected @@ -0,0 +1,22 @@ + +Source guest information (--print-source option): + + source name: esx6.7-rhel8.6-nvme-disk +hypervisor type: vmware + VM genid: + memory: 2147483648 (bytes) + nr vCPUs: 1 + CPU vendor: + CPU model: + CPU topology: + CPU features: + firmware: bios + display: + sound: +disks: + 0 [nvme] +removable media: + +NICs: + Network "VM Network" mac: 00:50:56:ac:cf:51 [vmxnet3] + diff --git a/tests/test-v2v-i-vmx-6.vmx b/tests/test-v2v-i-vmx-6.vmx new file mode 100644 index 0000000000..9203c0cfcb --- /dev/null +++ b/tests/test-v2v-i-vmx-6.vmx @@ -0,0 +1,84 @@ +.encoding = "UTF-8" +config.version = "8" +virtualHW.version = "14" +nvram = "esx6.7-rhel8.6-nvme-disk.nvram" +pciBridge0.present = "TRUE" +svga.present = "TRUE" +pciBridge4.present = "TRUE" +pciBridge4.virtualDev = "pcieRootPort" +pciBridge4.functions = "8" +pciBridge5.present = "TRUE" +pciBridge5.virtualDev = "pcieRootPort" +pciBridge5.functions = "8" +pciBridge6.present = "TRUE" +pciBridge6.virtualDev = "pcieRootPort" +pciBridge6.functions = "8" +pciBridge7.present = "TRUE" +pciBridge7.virtualDev = "pcieRootPort" +pciBridge7.functions = "8" +vmci0.present = "TRUE" +hpet0.present = "TRUE" +floppy0.present = "FALSE" +svga.vramSize = "8388608" +memSize = "2048" +powerType.powerOff = "default" +powerType.suspend = "default" +powerType.reset = "default" +tools.upgrade.policy = "manual" +sched.cpu.units = "mhz" +sched.cpu.affinity = "all" +vm.createDate = "1648710632879834" +sata0.present = "TRUE" +nvme0.present = "TRUE" +nvme0:0.fileName = "nvme-disk.vmdk" +sched.nvme0:0.shares = "normal" +sched.nvme0:0.throughputCap = "off" +nvme0:0.present = "TRUE" +ethernet0.virtualDev = "vmxnet3" +ethernet0.networkName = "VM Network" +ethernet0.addressType = "vpx" +ethernet0.generatedAddress = "00:50:56:ac:cf:51" +ethernet0.uptCompatibility = "TRUE" +ethernet0.present = "TRUE" +sata0:0.startConnected = "FALSE" +sata0:0.deviceType = "cdrom-raw" +sata0:0.clientDevice = "TRUE" +sata0:0.fileName = "emptyBackingString" +sata0:0.present = "TRUE" +displayName = "esx6.7-rhel8.6-nvme-disk" +guestOS = "rhel8-64" +toolScripts.afterPowerOn = "TRUE" +toolScripts.afterResume = "TRUE" +toolScripts.beforeSuspend = "TRUE" +toolScripts.beforePowerOff = "TRUE" +uuid.bios = "42 2c 17 6d 0b 31 4e 16-82 97 f0 b4 99 0b 86 09" +vc.uuid = "50 2c fc c7 be fb b1 48-fc 7d 4b 39 b3 f5 de b5" +migrate.hostLog = "esx6.7-rhel8.6-nvme-disk-7585773e.hlog" +sched.cpu.min = "0" +sched.cpu.shares = "normal" +sched.mem.min = "0" +sched.mem.minSize = "0" +sched.mem.shares = "normal" +migrate.encryptionMode = "opportunistic" +numa.autosize.cookie = "10001" +numa.autosize.vcpu.maxPerVirtualNode = "1" +sched.swap.derivedName = "/vmfs/volumes/02155034-b2df70e7/esx6.7-rhel8.6-nvme-disk/esx6.7-rhel8.6-nvme-disk-34965f24.vswp" +uuid.location = "56 4d c5 90 b1 ca f0 32-64 99 32 a7 d8 97 27 3a" +nvme0:0.redo = "" +pciBridge0.pciSlotNumber = "17" +pciBridge4.pciSlotNumber = "21" +pciBridge5.pciSlotNumber = "22" +pciBridge6.pciSlotNumber = "23" +pciBridge7.pciSlotNumber = "24" +ethernet0.pciSlotNumber = "160" +vmci0.pciSlotNumber = "32" +sata0.pciSlotNumber = "33" +nvme0.pciSlotNumber = "192" +vmci0.id = "-1727298039" +monitor.phys_bits_used = "43" +vmotion.checkpointFBSize = "8388608" +vmotion.checkpointSVGAPrimarySize = "8388608" +cleanShutdown = "TRUE" +softPowerOff = "TRUE" +svga.guestBackedPrimaryAware = "TRUE" +tools.syncTime = "FALSE" \ No newline at end of file diff --git a/tests/test-v2v-i-vmx.sh b/tests/test-v2v-i-vmx.sh index db870bea05..d74ddfaaf8 100755 --- a/tests/test-v2v-i-vmx.sh +++ b/tests/test-v2v-i-vmx.sh @@ -39,14 +39,14 @@ rm -f test-v2v-i-vmx-*.actual # For the tests to succeed we need at least the fileName (VMDK input # files) to exist. -fns="BZ1308535_21disks.vmdk Fedora-20.vmdk RHEL-7.1-UEFI.vmdk Windows-7-x64.vmdk MSEdge-Win10_preview.vmdk" +fns="BZ1308535_21disks.vmdk Fedora-20.vmdk RHEL-7.1-UEFI.vmdk Windows-7-x64.vmdk MSEdge-Win10_preview.vmdk nvme-disk.vmdk" for fn in BZ1308535_21disks_{1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20}.vmdk; do fns="$fns $fn" done for fn in $fns; do qemu-img create -f vmdk $fn 512; done -for i in 1 2 3 4 5; do +for i in 1 2 3 4 5 6; do $VG virt-v2v --debug-gc \ -i vmx test-v2v-i-vmx-$i.vmx \ --print-source > test-v2v-i-vmx-$i.actual -- 2.35.1
On 04/08/22 13:11, Richard W.M. Jones wrote:
We model NVMe devices in the source hypervisor.
We currently assume that no one is using the namespaces feature of NVMe, ie. that each source device will appear in a Linux guest as /dev/nvme0n1, /dev/nvme1n1, etc. We could fix this if it is a problem, but it requires adjusting the current assumption for removable devices that slots are simple integers.
The devices are mapped to virtio-blk, so in the target the device name has to change from /dev/nvme0 to /dev/vda (etc.)
Reported-by: Ming Xie Fixes: https://bugzilla.redhat.com/show_bug.cgi?id=2070530 --- convert/convert_linux.ml | 5 +- convert/target_bus_assignment.ml | 4 +- input/parse_domain_from_vmx.ml | 26 +++++++++- lib/types.ml | 3 +- lib/types.mli | 2 +- tests/test-v2v-i-vmx-6.expected | 22 +++++++++ tests/test-v2v-i-vmx-6.vmx | 84 ++++++++++++++++++++++++++++++++ tests/test-v2v-i-vmx.sh | 4 +- 8 files changed, 141 insertions(+), 9 deletions(-)
diff --git a/convert/convert_linux.ml b/convert/convert_linux.ml index d9397cfad1..255ba608ca 100644 --- a/convert/convert_linux.ml +++ b/convert/convert_linux.ml @@ -1032,7 +1032,10 @@ let convert (g : G.guestfs) source inspect keep_serial_console _ = ide_block_prefix ^ drive_name i | Some (Source_virtio_SCSI | Source_SCSI | Source_SATA) -> "sd" ^ drive_name i - | Some Source_virtio_blk -> "vd" ^ drive_name i in + | Some Source_virtio_blk -> "vd" ^ drive_name i + | Some Source_NVME -> + (* For NVMe assume no one is using namespaces. *) + sprintf "nvme%dn1" i in
Ah right, we need %d / i here, not "drive_name i".
let source_dev = device_name_before_conversion i in let target_dev = block_prefix_after_conversion ^ drive_name i in source_dev, target_dev diff --git a/convert/target_bus_assignment.ml b/convert/target_bus_assignment.ml index 4b56a6e171..f8675cf29b 100644 --- a/convert/target_bus_assignment.ml +++ b/convert/target_bus_assignment.ml @@ -73,8 +73,8 @@ let rec target_bus_assignment source_disks source_removables guestcaps = | None -> ide_bus (* Wild guess, but should be safe. *) | Some Source_virtio_blk -> virtio_blk_bus | Some Source_IDE -> ide_bus - | Some (Source_virtio_SCSI | Source_SCSI | Source_SATA) -> - scsi_bus in + | Some (Source_virtio_SCSI | Source_SCSI | Source_SATA | + Source_NVME) -> scsi_bus in
match r.s_removable_slot with | None -> diff --git a/input/parse_domain_from_vmx.ml b/input/parse_domain_from_vmx.ml index 730f1177f7..b812edeb81 100644 --- a/input/parse_domain_from_vmx.ml +++ b/input/parse_domain_from_vmx.ml @@ -103,6 +103,7 @@ let remote_file_exists uri path =
let rec find_disks vmx vmx_source = find_scsi_disks vmx vmx_source + @ find_nvme_disks vmx vmx_source @ find_ide_disks vmx vmx_source
(* Find all SCSI hard disks. @@ -129,6 +130,27 @@ and find_scsi_disks vmx vmx_source = get_scsi_controller_target is_scsi_controller_target scsi_device_types scsi_controller
+(* Find all NVMe hard disks. + * + * In the VMX file: + * nvme0.pcislotnumber = "192" + * nvme0:0.fileName = "guest.vmdk" + *) +and find_nvme_disks vmx vmx_source = + let get_nvme_controller_target ns = + sscanf ns "nvme%d:%d" (fun c t -> c, t)
Did we just get lucky here that the NVMe naming scheme (device, namespace) happens to map quite cleanly to our internal (controller, target) scheme? I mean if there weren't namespaces, we could still use sscanf ns "nvme%d" (fun c -> c, 0) buf if there were *more* than two components in the NVMe identifier, we'd be in some trouble, would we not? (Now I think I understand we don't actually map the "target" we parse here back to a namespace, near the top of the patch; just asking about the internal tuple structure the existing disk collection code expects. ... More precisely, I think the namespace information is lost when we sort and renumber near the end of "find_hdds".)
+ in + let is_nvme_controller_target ns = + try ignore (get_nvme_controller_target ns); true + with Scanf.Scan_failure _ | End_of_file | Failure _ -> false + in + let nvme_device_types = [ None ] in + let nvme_controller = Source_NVME in + + find_hdds vmx vmx_source + get_nvme_controller_target is_nvme_controller_target + nvme_device_types nvme_controller + (* Find all IDE hard disks. * * In the VMX file: @@ -153,12 +175,12 @@ and find_ide_disks vmx vmx_source = and find_hdds vmx vmx_source get_controller_target is_controller_target device_types controller = - (* Find namespaces matching '(ide|scsi)X:Y' with suitable deviceType. *) + (* Find namespaces matching '(ide|scsi|nvme)X:Y' with suitable deviceType. *) let hdds = Parse_vmx.select_namespaces ( function | [ns] -> - (* Check the namespace is '(ide|scsi)X:Y' *) + (* Check the namespace is '(ide|scsi|nvme)X:Y' *) if not (is_controller_target ns) then false else ( (* Check the deviceType is one we are looking for. *) diff --git a/lib/types.ml b/lib/types.ml index 5804c7c79f..92ed0e5251 100644 --- a/lib/types.ml +++ b/lib/types.ml @@ -56,7 +56,7 @@ and source_disk = { s_disk_id : int; s_controller : s_controller option; } -and s_controller = Source_IDE | Source_SATA | Source_SCSI | +and s_controller = Source_IDE | Source_SATA | Source_SCSI | Source_NVME | Source_virtio_blk | Source_virtio_SCSI and source_removable = { s_removable_type : s_removable_type; @@ -194,6 +194,7 @@ and string_of_controller = function | Source_IDE -> "ide" | Source_SATA -> "sata" | Source_SCSI -> "scsi" + | Source_NVME -> "nvme" | Source_virtio_blk -> "virtio-blk" | Source_virtio_SCSI -> "virtio-scsi"
diff --git a/lib/types.mli b/lib/types.mli index dd2fe5925d..37238cd755 100644 --- a/lib/types.mli +++ b/lib/types.mli @@ -103,7 +103,7 @@ and source_disk = { } (** A source disk. *)
-and s_controller = Source_IDE | Source_SATA | Source_SCSI | +and s_controller = Source_IDE | Source_SATA | Source_SCSI | Source_NVME | Source_virtio_blk | Source_virtio_SCSI (** Source disk controller. *)
diff --git a/tests/test-v2v-i-vmx-6.expected b/tests/test-v2v-i-vmx-6.expected new file mode 100644 index 0000000000..1793b6b953 --- /dev/null +++ b/tests/test-v2v-i-vmx-6.expected @@ -0,0 +1,22 @@ + +Source guest information (--print-source option): + + source name: esx6.7-rhel8.6-nvme-disk +hypervisor type: vmware + VM genid: + memory: 2147483648 (bytes) + nr vCPUs: 1 + CPU vendor: + CPU model: + CPU topology: + CPU features: + firmware: bios + display: + sound: +disks: + 0 [nvme] +removable media: + +NICs: + Network "VM Network" mac: 00:50:56:ac:cf:51 [vmxnet3] + diff --git a/tests/test-v2v-i-vmx-6.vmx b/tests/test-v2v-i-vmx-6.vmx new file mode 100644 index 0000000000..9203c0cfcb --- /dev/null +++ b/tests/test-v2v-i-vmx-6.vmx @@ -0,0 +1,84 @@ +.encoding = "UTF-8" +config.version = "8" +virtualHW.version = "14" +nvram = "esx6.7-rhel8.6-nvme-disk.nvram" +pciBridge0.present = "TRUE" +svga.present = "TRUE" +pciBridge4.present = "TRUE" +pciBridge4.virtualDev = "pcieRootPort" +pciBridge4.functions = "8" +pciBridge5.present = "TRUE" +pciBridge5.virtualDev = "pcieRootPort" +pciBridge5.functions = "8" +pciBridge6.present = "TRUE" +pciBridge6.virtualDev = "pcieRootPort" +pciBridge6.functions = "8" +pciBridge7.present = "TRUE" +pciBridge7.virtualDev = "pcieRootPort" +pciBridge7.functions = "8" +vmci0.present = "TRUE" +hpet0.present = "TRUE" +floppy0.present = "FALSE" +svga.vramSize = "8388608" +memSize = "2048" +powerType.powerOff = "default" +powerType.suspend = "default" +powerType.reset = "default" +tools.upgrade.policy = "manual" +sched.cpu.units = "mhz" +sched.cpu.affinity = "all" +vm.createDate = "1648710632879834" +sata0.present = "TRUE" +nvme0.present = "TRUE" +nvme0:0.fileName = "nvme-disk.vmdk" +sched.nvme0:0.shares = "normal" +sched.nvme0:0.throughputCap = "off" +nvme0:0.present = "TRUE" +ethernet0.virtualDev = "vmxnet3" +ethernet0.networkName = "VM Network" +ethernet0.addressType = "vpx" +ethernet0.generatedAddress = "00:50:56:ac:cf:51" +ethernet0.uptCompatibility = "TRUE" +ethernet0.present = "TRUE" +sata0:0.startConnected = "FALSE" +sata0:0.deviceType = "cdrom-raw" +sata0:0.clientDevice = "TRUE" +sata0:0.fileName = "emptyBackingString" +sata0:0.present = "TRUE" +displayName = "esx6.7-rhel8.6-nvme-disk" +guestOS = "rhel8-64" +toolScripts.afterPowerOn = "TRUE" +toolScripts.afterResume = "TRUE" +toolScripts.beforeSuspend = "TRUE" +toolScripts.beforePowerOff = "TRUE" +uuid.bios = "42 2c 17 6d 0b 31 4e 16-82 97 f0 b4 99 0b 86 09" +vc.uuid = "50 2c fc c7 be fb b1 48-fc 7d 4b 39 b3 f5 de b5" +migrate.hostLog = "esx6.7-rhel8.6-nvme-disk-7585773e.hlog" +sched.cpu.min = "0" +sched.cpu.shares = "normal" +sched.mem.min = "0" +sched.mem.minSize = "0" +sched.mem.shares = "normal" +migrate.encryptionMode = "opportunistic" +numa.autosize.cookie = "10001" +numa.autosize.vcpu.maxPerVirtualNode = "1" +sched.swap.derivedName = "/vmfs/volumes/02155034-b2df70e7/esx6.7-rhel8.6-nvme-disk/esx6.7-rhel8.6-nvme-disk-34965f24.vswp" +uuid.location = "56 4d c5 90 b1 ca f0 32-64 99 32 a7 d8 97 27 3a" +nvme0:0.redo = "" +pciBridge0.pciSlotNumber = "17" +pciBridge4.pciSlotNumber = "21" +pciBridge5.pciSlotNumber = "22" +pciBridge6.pciSlotNumber = "23" +pciBridge7.pciSlotNumber = "24" +ethernet0.pciSlotNumber = "160" +vmci0.pciSlotNumber = "32" +sata0.pciSlotNumber = "33" +nvme0.pciSlotNumber = "192" +vmci0.id = "-1727298039" +monitor.phys_bits_used = "43" +vmotion.checkpointFBSize = "8388608" +vmotion.checkpointSVGAPrimarySize = "8388608" +cleanShutdown = "TRUE" +softPowerOff = "TRUE" +svga.guestBackedPrimaryAware = "TRUE" +tools.syncTime = "FALSE" \ No newline at end of file diff --git a/tests/test-v2v-i-vmx.sh b/tests/test-v2v-i-vmx.sh index db870bea05..d74ddfaaf8 100755 --- a/tests/test-v2v-i-vmx.sh +++ b/tests/test-v2v-i-vmx.sh @@ -39,14 +39,14 @@ rm -f test-v2v-i-vmx-*.actual # For the tests to succeed we need at least the fileName (VMDK input # files) to exist.
-fns="BZ1308535_21disks.vmdk Fedora-20.vmdk RHEL-7.1-UEFI.vmdk Windows-7-x64.vmdk MSEdge-Win10_preview.vmdk" +fns="BZ1308535_21disks.vmdk Fedora-20.vmdk RHEL-7.1-UEFI.vmdk Windows-7-x64.vmdk MSEdge-Win10_preview.vmdk nvme-disk.vmdk" for fn in BZ1308535_21disks_{1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20}.vmdk; do fns="$fns $fn" done
for fn in $fns; do qemu-img create -f vmdk $fn 512; done
-for i in 1 2 3 4 5; do +for i in 1 2 3 4 5 6; do $VG virt-v2v --debug-gc \ -i vmx test-v2v-i-vmx-$i.vmx \ --print-source > test-v2v-i-vmx-$i.actual
Reviewed-by: Laszlo Ersek <lersek@redhat.com> Thanks Laszlo
On Fri, Apr 08, 2022 at 05:00:07PM +0200, Laszlo Ersek wrote:
On 04/08/22 13:11, Richard W.M. Jones wrote:
+(* Find all NVMe hard disks. + * + * In the VMX file: + * nvme0.pcislotnumber = "192" + * nvme0:0.fileName = "guest.vmdk" + *) +and find_nvme_disks vmx vmx_source = + let get_nvme_controller_target ns = + sscanf ns "nvme%d:%d" (fun c t -> c, t)
Did we just get lucky here that the NVMe naming scheme (device, namespace) happens to map quite cleanly to our internal (controller, target) scheme?
I mean if there weren't namespaces, we could still use
sscanf ns "nvme%d" (fun c -> c, 0)
buf if there were *more* than two components in the NVMe identifier, we'd be in some trouble, would we not?
(Now I think I understand we don't actually map the "target" we parse here back to a namespace, near the top of the patch; just asking about the internal tuple structure the existing disk collection code expects. ... More precisely, I think the namespace information is lost when we sort and renumber near the end of "find_hdds".)
I only have an example of a guest with a single NVMe drive, so I've no idea how the second one would be named. But if it's the same as the SCSI examples then it'll be "nvme0:0", "nvme0:1", through "nvme0:15", then "nvme1:0", "nvme1:1", etc. I don't know what the numbers precisely mean. The only thing we use the 'c' and 't' numbers for is to order the disks, since the entries do not appear in order in the .vmx file. So really we don't care what the numbers are (but we do need to parse them as simple string sorting wouldn't work, because "scsi10:0" < "scsi1:0"). We only need the number of disks and their original order. The original namespace info is indeed lost. We will create the same number of virtio-blk devices on the target, and copy them in the same order. There's probably a lot of room for improvement here ... Rich. -- Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones Read my programming and virtualization blog: http://rwmj.wordpress.com virt-p2v converts physical machines to virtual machines. Boot with a live CD or over the network (PXE) and turn machines into KVM guests. http://libguestfs.org/virt-v2v
participants (2)
-
Laszlo Ersek -
Richard W.M. Jones