12:06 p.m.
Bugzilla: https://bugzilla.redhat.com/show_bug.cgi?id=1961107
This series advances in small steps, as follows:
- introduce the Standard_VGA data constructor (in addition to QXL,
Cirrus, ...);
- handle Standard_VGA (as a non-default option through the requested
guest caps) in the Linux and Windows guest conversions, and various
output modules;
- flip the default display choice from QXL to Standard_VGA if the
requested guest caps don't express a display preference;
- remove the QXL data constructor;
- clean up the test suite.
I'm posting the series as an RFC for two reasons:
- No actual guests have been converted for testing.
- I need further info (help) on the OVF, JSON and OpenStack output modes
because they are too tightly coupled with QXL, and/or I don't know how
to express "standard VGA" in them. (Watch out for the "Notes"
sections
in the relevant patches.)
However, the series should be complete enough to discuss the general
approach (and to convert some actual guests using the "QEMU cmdline" and
"libvirt XML" output formats).
(Side comment: today is a public holiday in Hungary, like in many other
countries. The fact that I'm posting this today does not imply that I
expect any feedback today (or soon after today). For personal reasons,
my work schedule has been very difficult to manage, and today has been a
rare day where I could actually sit down for a few hours
*uninterrupted*, and work on this. So the point of working on a holiday
was not to "be faster than others", but to make *any* progress worth
mentioning.)
Thanks,
Laszlo
Laszlo Ersek (10):
lib/types: introduce "Standard_VGA" constructor for
"guestcaps_video_type"
convert_linux: use "modesetting" X.org driver for Standard_VGA
convert/windows_virtio: restrict the warning with virtio-win.iso
absent
convert/windows_virtio: handle Standard_VGA in the requested guest
caps
output: handle Standard_VGA in the libvirt XML and QEMU cmdline
formats
convert_linux: flip default target display to Standard_VGA
convert/windows_virtio: flip default target display to Standard_VGA
lib/types: remove "QXL" constructor for "guestcaps_video_type"
tests: remove the fake Windows guest drivers for the QXL device
tests: remove the remaining QXL references (for completeness)
convert/convert_linux.ml | 7 ++-
convert/windows_virtio.ml | 48 +++++++-------------
lib/types.ml | 4 +-
lib/types.mli | 2 +-
output/create_json.ml | 2 +-
output/create_libvirt_xml.ml | 2 +-
output/openstack_image_properties.ml | 2 +-
output/output.ml | 5 +-
test-data/fake-virtio-win/Makefile.am | 16 -------
test-data/fake-virtio-win/drivers/amd64/Win2008R2/qxl.cat | 1 -
test-data/fake-virtio-win/drivers/amd64/Win2008R2/qxl.inf | 2 -
test-data/fake-virtio-win/drivers/amd64/Win2008R2/qxl.sys | 1 -
test-data/fake-virtio-win/drivers/amd64/Win2008R2/qxldd.dll | 1 -
test-data/fake-virtio-win/drivers/amd64/Win7/qxl.cat | 1 -
test-data/fake-virtio-win/drivers/amd64/Win7/qxl.inf | 2 -
test-data/fake-virtio-win/drivers/amd64/Win7/qxl.sys | 1 -
test-data/fake-virtio-win/drivers/amd64/Win7/qxldd.dll | 1 -
test-data/fake-virtio-win/drivers/i386/Win7/qxl.cat | 1 -
test-data/fake-virtio-win/drivers/i386/Win7/qxl.inf | 2 -
test-data/fake-virtio-win/drivers/i386/Win7/qxl.sys | 1 -
test-data/fake-virtio-win/drivers/i386/Win7/qxldd.dll | 1 -
test-data/fake-virtio-win/drivers/i386/WinXP/qxl.cat | 1 -
test-data/fake-virtio-win/drivers/i386/WinXP/qxl.inf | 2 -
test-data/fake-virtio-win/drivers/i386/WinXP/qxl.sys | 1 -
test-data/fake-virtio-win/drivers/i386/WinXP/qxldd.dll | 1 -
test-data/phony-guests/guests.xml.in | 2 +-
tests/test-v2v-i-ova.xml | 2 +-
tests/test-v2v-in-place.sh | 2 +-
tests/test-v2v-print-source.expected | 2 +-
tests/test-v2v-print-source.xml.in | 2 +-
tests/test-v2v-windows-conversion.sh | 2 +-
v2v/v2v_unit_tests.ml | 16 -------
32 files changed, 38 insertions(+), 98 deletions(-)
delete mode 100644 test-data/fake-virtio-win/drivers/amd64/Win2008R2/qxl.inf
delete mode 100644 test-data/fake-virtio-win/drivers/amd64/Win7/qxl.inf
delete mode 100644 test-data/fake-virtio-win/drivers/i386/Win7/qxl.inf
delete mode 100644 test-data/fake-virtio-win/drivers/i386/WinXP/qxl.inf
delete mode 100644 test-data/fake-virtio-win/drivers/amd64/Win2008R2/qxl.cat
delete mode 100644 test-data/fake-virtio-win/drivers/amd64/Win2008R2/qxl.sys
delete mode 100644 test-data/fake-virtio-win/drivers/amd64/Win2008R2/qxldd.dll
delete mode 100644 test-data/fake-virtio-win/drivers/amd64/Win7/qxl.cat
delete mode 100644 test-data/fake-virtio-win/drivers/amd64/Win7/qxl.sys
delete mode 100644 test-data/fake-virtio-win/drivers/amd64/Win7/qxldd.dll
delete mode 100644 test-data/fake-virtio-win/drivers/i386/Win7/qxl.cat
delete mode 100644 test-data/fake-virtio-win/drivers/i386/Win7/qxl.sys
delete mode 100644 test-data/fake-virtio-win/drivers/i386/Win7/qxldd.dll
delete mode 100644 test-data/fake-virtio-win/drivers/i386/WinXP/qxl.cat
delete mode 100644 test-data/fake-virtio-win/drivers/i386/WinXP/qxl.sys
delete mode 100644 test-data/fake-virtio-win/drivers/i386/WinXP/qxldd.dll
base-commit: 17e7cbffb9e5dad526dc8f63dc771aa15b306888
--
2.19.1.3.g30247aa5d201
12:06 p.m.
New subject: [virt-v2v RFC wave 2 01/10] lib/types: introduce "Standard_VGA" constructor for "guestcaps_video_type"
In order to replace QXL with Standard_VGA in small steps, bisectably,
first just introduce the Standard_VGA constructor.
Extend the string_of_video function at once ("stdvga"). The value returned
by this function is only used for debug logging (indirectly, via the
string_of_guestcaps and string_of_requested_guestcaps functions).
The virt-v2v build flags turn "non-exhaustive pattern matching" warnings
into errors. Match the new "Standard_VGA" constructor with a constant
false assertion wherever an expression of type "guestcaps_video_type" is
matched, in order to make this patch bisectable. These assertions will be
gradually replaced with actual logic in the rest of this patch set.
Bugzilla: https://bugzilla.redhat.com/show_bug.cgi?id=1961107
Signed-off-by: Laszlo Ersek <lersek(a)redhat.com>
---
convert/convert_linux.ml | 6 +++++-
convert/windows_virtio.ml | 2 ++
lib/types.ml | 3 ++-
lib/types.mli | 2 +-
output/create_json.ml | 1 +
output/create_libvirt_xml.ml | 1 +
output/openstack_image_properties.ml | 1 +
output/output.ml | 6 +++++-
8 files changed, 18 insertions(+), 4 deletions(-)
diff --git a/convert/convert_linux.ml b/convert/convert_linux.ml
index 057cf9dff7da..85d8c5ae0e92 100644
--- a/convert/convert_linux.ml
+++ b/convert/convert_linux.ml
@@ -829,7 +829,11 @@ let convert (g : G.guestfs) source inspect keep_serial_console rcaps
_ =
true
and configure_display_driver video =
- let video_driver = match video with QXL -> "qxl" | Cirrus ->
"cirrus" in
+ let video_driver =
+ match video with
+ | Standard_VGA -> assert false
+ | QXL -> "qxl"
+ | Cirrus -> "cirrus" in
let updated = ref false in
diff --git a/convert/windows_virtio.ml b/convert/windows_virtio.ml
index f843dae2d77f..1851fe2f3292 100644
--- a/convert/windows_virtio.ml
+++ b/convert/windows_virtio.ml
@@ -51,6 +51,7 @@ let rec install_drivers ((g, _) as reg) inspect rcaps =
if not (copy_drivers g inspect driverdir) then (
match rcaps with
+ | { rcaps_video = Some Standard_VGA } -> assert false
| { rcaps_block_bus = Some Virtio_blk | Some Virtio_SCSI }
| { rcaps_net_bus = Some Virtio_net }
| { rcaps_video = Some QXL } ->
@@ -159,6 +160,7 @@ let rec install_drivers ((g, _) as reg) inspect rcaps =
g#exists (driverdir // "qxl.inf") ||
g#exists (driverdir // "qxldod.inf") in
match rcaps.rcaps_video, has_qxl with
+ | Some Standard_VGA, _ -> assert false
| Some QXL, false ->
error (f_"there is no QXL driver for this version of Windows (%d.%d %s).
virt-v2v looks for this driver in %s")
inspect.i_major_version inspect.i_minor_version
diff --git a/lib/types.ml b/lib/types.ml
index 72d10d1ec774..0156d55954e9 100644
--- a/lib/types.ml
+++ b/lib/types.ml
@@ -429,7 +429,7 @@ and requested_guestcaps = {
}
and guestcaps_block_type = Virtio_blk | Virtio_SCSI | IDE
and guestcaps_net_type = Virtio_net | E1000 | RTL8139
-and guestcaps_video_type = QXL | Cirrus
+and guestcaps_video_type = Standard_VGA | QXL | Cirrus
and guestcaps_machine = I440FX | Q35 | Virt
let string_of_block_type = function
@@ -441,6 +441,7 @@ let string_of_net_type = function
| E1000 -> "e1000"
| RTL8139 -> "rtl8139"
let string_of_video = function
+ | Standard_VGA -> "stdvga"
| QXL -> "qxl"
| Cirrus -> "cirrus"
let string_of_machine = function
diff --git a/lib/types.mli b/lib/types.mli
index 0bae445d8393..19790f375035 100644
--- a/lib/types.mli
+++ b/lib/types.mli
@@ -293,7 +293,7 @@ and requested_guestcaps = {
and guestcaps_block_type = Virtio_blk | Virtio_SCSI | IDE
and guestcaps_net_type = Virtio_net | E1000 | RTL8139
-and guestcaps_video_type = QXL | Cirrus
+and guestcaps_video_type = Standard_VGA | QXL | Cirrus
and guestcaps_machine = I440FX | Q35 | Virt
val string_of_guestcaps : guestcaps -> string
diff --git a/output/create_json.ml b/output/create_json.ml
index 5fa2805a2dcd..aae4d09c59bd 100644
--- a/output/create_json.ml
+++ b/output/create_json.ml
@@ -204,6 +204,7 @@ let create_json_metadata source inspect
| RTL8139 -> "rtl8139" in
let video =
match guestcaps.gcaps_video with
+ | Standard_VGA -> assert false
| QXL -> "qxl"
| Cirrus -> "cirrus" in
let machine =
diff --git a/output/create_libvirt_xml.ml b/output/create_libvirt_xml.ml
index 64f1b0888cd6..ec1cd455ef13 100644
--- a/output/create_libvirt_xml.ml
+++ b/output/create_libvirt_xml.ml
@@ -421,6 +421,7 @@ let create_libvirt_xml ?pool source inspect
let video =
let video_model =
match guestcaps.gcaps_video with
+ | Standard_VGA -> assert false
| QXL -> e "model" [ "type", "qxl";
"ram", "65536" ] []
| Cirrus -> e "model" [ "type", "cirrus";
"vram", "9216" ] [] in
append_attr ("heads", "1") video_model;
diff --git a/output/openstack_image_properties.ml b/output/openstack_image_properties.ml
index 66f5399212c8..592da3e40e44 100644
--- a/output/openstack_image_properties.ml
+++ b/output/openstack_image_properties.ml
@@ -44,6 +44,7 @@ let create source inspect { target_buses; guestcaps; target_firmware }
=
| RTL8139 -> "rtl8139");
"hw_video_model",
(match guestcaps.gcaps_video with
+ | Standard_VGA -> assert false
| QXL -> "qxl"
| Cirrus -> "cirrus");
"hw_machine_type",
diff --git a/output/output.ml b/output/output.ml
index 9c2d8853bfef..93dbd4a62865 100644
--- a/output/output.ml
+++ b/output/output.ml
@@ -1539,7 +1539,11 @@ and qemu_finalize dir source inspect target_meta
"addr=127.0.0.1"]
);
arg "-vga"
- (match guestcaps.gcaps_video with Cirrus -> "cirrus" | QXL ->
"qxl")
+ (match guestcaps.gcaps_video with
+ | Standard_VGA -> assert false
+ | Cirrus -> "cirrus"
+ | QXL -> "qxl"
+ )
);
(* Add a sound card. *)
--
2.19.1.3.g30247aa5d201
12:06 p.m.
New subject: [virt-v2v RFC wave 2 02/10] convert_linux: use "modesetting" X.org driver for Standard_VGA
In a Fedora 34 guest using standard VGA video, and having no explicit
X.org config file, the X.org server logs the following (excerpt):
Markers: (--) probed, (**) from config file, (==) default setting,
(++) from command line, (!!) notice, (II) informational,
(WW) warning, (EE) error, (NI) not implemented, (??) unknown.
(==) Matched modesetting as autoconfigured driver 0
(==) Matched fbdev as autoconfigured driver 1
(==) Matched vesa as autoconfigured driver 2
(II) LoadModule: "modesetting"
(II) Loading /usr/lib64/xorg/modules/drivers/modesetting_drv.so
(II) LoadModule: "fbdev"
(II) Loading /usr/lib64/xorg/modules/drivers/fbdev_drv.so
(II) LoadModule: "vesa"
(II) Loading /usr/lib64/xorg/modules/drivers/vesa_drv.so
(II) Loading sub module "fbdevhw"
(II) LoadModule: "fbdevhw"
(II) Loading /usr/lib64/xorg/modules/libfbdevhw.so
(II) UnloadModule: "fbdev"
(II) Unloading fbdev
(II) UnloadSubModule: "fbdevhw"
(II) Unloading fbdevhw
(II) UnloadModule: "vesa"
(II) Unloading vesa
This tells us that the standard VGA device model is driven by the
"modesetting" driver. Update the "configure_display_driver" function
(for
Linux guests) accordingly.
Cc: Gerd Hoffmann <kraxel(a)redhat.com>
Bugzilla: https://bugzilla.redhat.com/show_bug.cgi?id=1961107
Signed-off-by: Laszlo Ersek <lersek(a)redhat.com>
---
convert/convert_linux.ml | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/convert/convert_linux.ml b/convert/convert_linux.ml
index 85d8c5ae0e92..15f68fd4f8db 100644
--- a/convert/convert_linux.ml
+++ b/convert/convert_linux.ml
@@ -831,7 +831,7 @@ let convert (g : G.guestfs) source inspect keep_serial_console rcaps _
=
and configure_display_driver video =
let video_driver =
match video with
- | Standard_VGA -> assert false
+ | Standard_VGA -> "modesetting"
| QXL -> "qxl"
| Cirrus -> "cirrus" in
--
2.19.1.3.g30247aa5d201
4:25 a.m.
New subject: [virt-v2v RFC wave 2 02/10] convert_linux: use "modesetting" X.org driver for Standard_VGA
Hi,
> (II) LoadModule: "modesetting"
> (II) LoadModule: "fbdev"
> (II) LoadModule: "vesa"
> (II) UnloadModule: "fbdev"
> (II) UnloadModule: "vesa"
This tells us that the standard VGA device model is driven by the
"modesetting" driver. Update the "configure_display_driver" function
(for
Linux guests) accordingly.
Correct.
Acked-by: Gerd Hoffmann <kraxel(a)redhat.com>
and configure_display_driver video =
let video_driver =
match video with
- | Standard_VGA -> assert false
+ | Standard_VGA -> "modesetting"
| QXL -> "qxl"
| Cirrus -> "cirrus" in
Oh. More updates needed.
cirrus is modesetting too, unless you use a *really* old distro.
kraxel@sirius ~/distgit/xorg-x11-drv-cirrus (rawhide)# cat dead.package
2014-07-06: retired driver, because it does not support KMS:
https://lists.fedoraproject.org/pipermail/devel/2013-August/188429.html
qxl is on a similar path. In case xorg-x11-drv-qxl.rpm is installed
xorg will prefer the qxl driver. In case the driver is not present xorg
will use the modesetting driver instead and it'll work equally fine.
I think RHEL-9 doesn't ship xorg-x11-drv-qxl.rpm any more.
virtio-gpu will also be handled by the modesetting driver.
So possibly you can just do 'let video_driver = "modesetting"'
take care & HTH,
Gerd
10:16 a.m.
New subject: [virt-v2v RFC wave 2 02/10] convert_linux: use "modesetting" X.org driver for Standard_VGA
On 11/02/21 10:25, Gerd Hoffmann wrote:
Hi,
>> (II) LoadModule: "modesetting"
>> (II) LoadModule: "fbdev"
>> (II) LoadModule: "vesa"
>> (II) UnloadModule: "fbdev"
>> (II) UnloadModule: "vesa"
> This tells us that the standard VGA device model is driven by the
> "modesetting" driver. Update the "configure_display_driver"
function (for
> Linux guests) accordingly.
Correct.
Acked-by: Gerd Hoffmann <kraxel(a)redhat.com>
> and configure_display_driver video =
> let video_driver =
> match video with
> - | Standard_VGA -> assert false
> + | Standard_VGA -> "modesetting"
> | QXL -> "qxl"
> | Cirrus -> "cirrus" in
Oh. More updates needed.
cirrus is modesetting too, unless you use a *really* old distro.
kraxel@sirius ~/distgit/xorg-x11-drv-cirrus (rawhide)# cat dead.package
2014-07-06: retired driver, because it does not support KMS:
https://lists.fedoraproject.org/pipermail/devel/2013-August/188429.html
qxl is on a similar path. In case xorg-x11-drv-qxl.rpm is installed
xorg will prefer the qxl driver. In case the driver is not present xorg
will use the modesetting driver instead and it'll work equally fine.
I think RHEL-9 doesn't ship xorg-x11-drv-qxl.rpm any more.
virtio-gpu will also be handled by the modesetting driver.
So possibly you can just do 'let video_driver = "modesetting"'
That might require a bunch of testing, plus I'd rather not mix in
unrelated changes... If we really want to drive Cirrus with a different
x.org driver, we should (perhaps, if at all) have a separate RHBZ about
that.
I will pick up your A-b for the "surgical" stdvga-related change though,
if that's OK with you.
Thanks!
Laszlo
take care & HTH,
Gerd
11:40 p.m.
New subject: [virt-v2v RFC wave 2 02/10] convert_linux: use "modesetting" X.org driver for Standard_VGA
Hi,
That might require a bunch of testing, plus I'd rather not mix
in
unrelated changes... If we really want to drive Cirrus with a different
x.org driver, we should (perhaps, if at all) have a separate RHBZ about
that.
I will pick up your A-b for the "surgical" stdvga-related change though,
if that's OK with you.
Sure, it's correct no matter what happens with the other devices.
take care,
Gerd
12:06 p.m.
New subject: [virt-v2v RFC wave 2 03/10] convert/windows_virtio: restrict the warning with virtio-win.iso absent
Consider the following two situations:
(a) For the Windows version being converted, the virtio-win.iso image does
not offer any guest drivers (IOW, the Windows version as a whole is
unsupported by virtio-win.iso).
(b) For the Windows version being converted, the virtio-win.iso image
offers *some* drivers (but not for all possible paravirtualized
devices).
There are three relevant device categories: block, network, display.
If the conversion specifically requests paravirt for a given device
category, and that particular paravirt driver is unavailable -- be that
due to reason (a) or (b) --, the conversion fails, with an error message.
This is expected.
If the conversion does not express any preference for paravirt in a given
device category, and that particular paravirt driver is unavailable -- be
that due to reason (a) or (b) --, a warning is emitted, and the conversion
continues, using an emulated device in that category. This is expected as
well.
However, if the conversion explicitly requests an emulated device model
for a given device category -- that is, if the conversion *forbids*
paravirtualization for a given device category --, and that particular
paravirt driver is unavailable anyway, then the behavior between (a) and
(b) differs. Under (a), a warning is emitted (incorrectly!), while under
(b), no warning is emitted.
Under condition (a), only issue the warning if the conversion permits
paravirtualization in at least one device category.
Fixes: 9ed9d048f2b296cd0ebbd638ac3a2c5c8b47fded
Bugzilla: https://bugzilla.redhat.com/show_bug.cgi?id=1961107
Signed-off-by: Laszlo Ersek <lersek(a)redhat.com>
---
convert/windows_virtio.ml | 12 +++++++-----
1 file changed, 7 insertions(+), 5 deletions(-)
diff --git a/convert/windows_virtio.ml b/convert/windows_virtio.ml
index 1851fe2f3292..11b34a03f876 100644
--- a/convert/windows_virtio.ml
+++ b/convert/windows_virtio.ml
@@ -59,12 +59,14 @@ let rec install_drivers ((g, _) as reg) inspect rcaps =
inspect.i_major_version inspect.i_minor_version inspect.i_arch
inspect.i_product_variant virtio_win
- | { rcaps_block_bus = (Some IDE | None);
+ | { rcaps_block_bus = ((Some IDE | None) as block_type);
rcaps_net_bus = ((Some E1000 | Some RTL8139 | None) as net_type);
- rcaps_video = (Some Cirrus | None) } ->
- warning (f_"there are no virtio drivers available for this version of Windows
(%d.%d %s %s). virt-v2v looks for drivers in %s\n\nThe guest will be configured to use
slower emulated devices.")
- inspect.i_major_version inspect.i_minor_version inspect.i_arch
- inspect.i_product_variant virtio_win;
+ rcaps_video = ((Some Cirrus | None) as video_type) } ->
+ if block_type = None || net_type = None || video_type = None then
+ warning (f_"there are no virtio drivers available for this version of
Windows (%d.%d %s %s). virt-v2v looks for drivers in %s\n\nThe guest will be configured
to use slower emulated devices.")
+ inspect.i_major_version inspect.i_minor_version inspect.i_arch
+ inspect.i_product_variant virtio_win
+ else ();
let net_type =
match net_type with
| Some model -> model
--
2.19.1.3.g30247aa5d201
3:52 a.m.
New subject: [virt-v2v RFC wave 2 03/10] convert/windows_virtio: restrict the warning with virtio-win.iso absent
On Mon, Nov 01, 2021 at 06:06:11PM +0100, Laszlo Ersek wrote:
Consider the following two situations:
(a) For the Windows version being converted, the virtio-win.iso image does
not offer any guest drivers (IOW, the Windows version as a whole is
unsupported by virtio-win.iso).
(b) For the Windows version being converted, the virtio-win.iso image
offers *some* drivers (but not for all possible paravirtualized
devices).
There are three relevant device categories: block, network, display.
If the conversion specifically requests paravirt for a given device
category, and that particular paravirt driver is unavailable -- be that
due to reason (a) or (b) --, the conversion fails, with an error message.
This is expected.
If the conversion does not express any preference for paravirt in a given
device category, and that particular paravirt driver is unavailable -- be
that due to reason (a) or (b) --, a warning is emitted, and the conversion
continues, using an emulated device in that category. This is expected as
well.
However, if the conversion explicitly requests an emulated device model
for a given device category -- that is, if the conversion *forbids*
paravirtualization for a given device category --, and that particular
paravirt driver is unavailable anyway, then the behavior between (a) and
(b) differs. Under (a), a warning is emitted (incorrectly!), while under
(b), no warning is emitted.
Under condition (a), only issue the warning if the conversion permits
paravirtualization in at least one device category.
In the current code (since removing --in-place) rcaps is always
{ None; None; None }, which means select the best possible driver for
each of block, bus and video. Since I think we should completely
remove rcaps, it only makes sense to think about this case, and not
worry about what would happen if rcaps contained anything else.
Fixes: 9ed9d048f2b296cd0ebbd638ac3a2c5c8b47fded
Bugzilla: https://bugzilla.redhat.com/show_bug.cgi?id=1961107
Signed-off-by: Laszlo Ersek <lersek(a)redhat.com>
---
convert/windows_virtio.ml | 12 +++++++-----
1 file changed, 7 insertions(+), 5 deletions(-)
diff --git a/convert/windows_virtio.ml b/convert/windows_virtio.ml
index 1851fe2f3292..11b34a03f876 100644
--- a/convert/windows_virtio.ml
+++ b/convert/windows_virtio.ml
@@ -59,12 +59,14 @@ let rec install_drivers ((g, _) as reg) inspect rcaps =
inspect.i_major_version inspect.i_minor_version inspect.i_arch
inspect.i_product_variant virtio_win
- | { rcaps_block_bus = (Some IDE | None);
+ | { rcaps_block_bus = ((Some IDE | None) as block_type);
rcaps_net_bus = ((Some E1000 | Some RTL8139 | None) as net_type);
- rcaps_video = (Some Cirrus | None) } ->
- warning (f_"there are no virtio drivers available for this version of Windows
(%d.%d %s %s). virt-v2v looks for drivers in %s\n\nThe guest will be configured to use
slower emulated devices.")
- inspect.i_major_version inspect.i_minor_version inspect.i_arch
- inspect.i_product_variant virtio_win;
+ rcaps_video = ((Some Cirrus | None) as video_type) } ->
+ if block_type = None || net_type = None || video_type = None then
+ warning (f_"there are no virtio drivers available for this version of
Windows (%d.%d %s %s). virt-v2v looks for drivers in %s\n\nThe guest will be configured
to use slower emulated devices.")
+ inspect.i_major_version inspect.i_minor_version inspect.i_arch
+ inspect.i_product_variant virtio_win
+ else ();
You don't need the else () clause here, OCaml will assume that.
The deeper problem with this patch is that it bypasses the matching
which the match statement is already able to do, ie:
| { rcaps_block_bus = None; rcaps_net_bus = None; rcaps_video = None } ->
warning (f_"...")
| { rcaps_block_bus = Some IDE; etc } ->
() (* ie. the "else ()" case *)
There's a bit more code that follows which might have to be factored out
(but read on).
let net_type =
match net_type with
| Some model -> model
The fundamental problem is that we should first remove rcaps, which
would simplify the whole thing.
I hesitate to submit patches for parts of the code where someone else
is working since it creates a mess of conflicts, but would you like me
to have a go at a patch for removing rcaps?
Rich.
--
Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones
Read my programming and virtualization blog: http://rwmj.wordpress.com
virt-df lists disk usage of guests without needing to install any
software inside the virtual machine. Supports Linux and Windows.
http://people.redhat.com/~rjones/virt-df/
10:45 a.m.
New subject: [virt-v2v RFC wave 2 03/10] convert/windows_virtio: restrict the warning with virtio-win.iso absent
On 11/02/21 09:52, Richard W.M. Jones wrote:
On Mon, Nov 01, 2021 at 06:06:11PM +0100, Laszlo Ersek wrote:
> Consider the following two situations:
>
> (a) For the Windows version being converted, the virtio-win.iso image does
> not offer any guest drivers (IOW, the Windows version as a whole is
> unsupported by virtio-win.iso).
>
> (b) For the Windows version being converted, the virtio-win.iso image
> offers *some* drivers (but not for all possible paravirtualized
> devices).
>
> There are three relevant device categories: block, network, display.
>
> If the conversion specifically requests paravirt for a given device
> category, and that particular paravirt driver is unavailable -- be that
> due to reason (a) or (b) --, the conversion fails, with an error message.
> This is expected.
>
> If the conversion does not express any preference for paravirt in a given
> device category, and that particular paravirt driver is unavailable -- be
> that due to reason (a) or (b) --, a warning is emitted, and the conversion
> continues, using an emulated device in that category. This is expected as
> well.
>
> However, if the conversion explicitly requests an emulated device model
> for a given device category -- that is, if the conversion *forbids*
> paravirtualization for a given device category --, and that particular
> paravirt driver is unavailable anyway, then the behavior between (a) and
> (b) differs. Under (a), a warning is emitted (incorrectly!), while under
> (b), no warning is emitted.
>
> Under condition (a), only issue the warning if the conversion permits
> paravirtualization in at least one device category.
In the current code (since removing --in-place) rcaps is always
{ None; None; None }, which means select the best possible driver for
each of block, bus and video. Since I think we should completely
remove rcaps, it only makes sense to think about this case, and not
worry about what would happen if rcaps contained anything else.
OK.
> Fixes: 9ed9d048f2b296cd0ebbd638ac3a2c5c8b47fded
> Bugzilla: https://bugzilla.redhat.com/show_bug.cgi?id=1961107
> Signed-off-by: Laszlo Ersek <lersek(a)redhat.com>
> ---
> convert/windows_virtio.ml | 12 +++++++-----
> 1 file changed, 7 insertions(+), 5 deletions(-)
>
> diff --git a/convert/windows_virtio.ml b/convert/windows_virtio.ml
> index 1851fe2f3292..11b34a03f876 100644
> --- a/convert/windows_virtio.ml
> +++ b/convert/windows_virtio.ml
> @@ -59,12 +59,14 @@ let rec install_drivers ((g, _) as reg) inspect rcaps =
> inspect.i_major_version inspect.i_minor_version inspect.i_arch
> inspect.i_product_variant virtio_win
>
> - | { rcaps_block_bus = (Some IDE | None);
> + | { rcaps_block_bus = ((Some IDE | None) as block_type);
> rcaps_net_bus = ((Some E1000 | Some RTL8139 | None) as net_type);
> - rcaps_video = (Some Cirrus | None) } ->
> - warning (f_"there are no virtio drivers available for this version of
Windows (%d.%d %s %s). virt-v2v looks for drivers in %s\n\nThe guest will be configured
to use slower emulated devices.")
> - inspect.i_major_version inspect.i_minor_version inspect.i_arch
> - inspect.i_product_variant virtio_win;
> + rcaps_video = ((Some Cirrus | None) as video_type) } ->
> + if block_type = None || net_type = None || video_type = None then
> + warning (f_"there are no virtio drivers available for this version of
Windows (%d.%d %s %s). virt-v2v looks for drivers in %s\n\nThe guest will be configured
to use slower emulated devices.")
> + inspect.i_major_version inspect.i_minor_version inspect.i_arch
> + inspect.i_product_variant virtio_win
> + else ();
You don't need the else () clause here, OCaml will assume that.
OK, thanks -- indeed, now that I double-checked, the book I use for
learning does explain this, I just didn't recall it (and when I looked
quickly while writing the patch, I missed that small paragraph).
The deeper problem with this patch is that it bypasses the matching
which the match statement is already able to do, ie:
| { rcaps_block_bus = None; rcaps_net_bus = None; rcaps_video = None } ->
warning (f_"...")
| { rcaps_block_bus = Some IDE; etc } ->
() (* ie. the "else ()" case *)
I disagree here; the code I wrote was very intentional.
The condition (aka "pattern") we need to handle, in general, is: "no
paravirt device requested in rcaps at all" [1].
That is, we have an original condition (a set of patterns) that says [0]:
- block is virtio-blk (+ maybe virtio-scsi), OR
- net is virtio-net, OR
- video is QXL.
Then, for [1], we need to handle the negation of [0], which yields (per
de Morgan):
- block hint is missing, or is IDE; AND
- net hint is missing, or is E1000 / RTL8139; AND
- video hint is missing, or is Cirrus.
The *primary* handling for this condition is that we output some basic
(non-paravirt, emulated) device configuration. The *secondary* handling
is that we emit a warning.
The pre-patch code does this alright, except the secondary handling --
the warning. Namely, in case *all* device preferences are explicitly
requested in rcaps, there is nothing to warn the user about. I.e., in
that case, the secondary handling (the warning) should be omitted. If
the user wants emulated devices in all categories, there is no
"degradation" to speak of. The primary handling (outputting the
particular set of guest caps) should stay the same.
This is what the patch fixes. It does not change the primary handling;
it only restricts the warning (the secondary handling) to the case when
at least one device category hint is omitted, and so the "degradation"
occurs, and deserves a mention to the user.
This cannot be implemented at the same level where the patterns are
currently listed, because once a pattern matches, other matches are not
attempted (to my knowledge). (In general, the pattern set should cover
every possibility, but no two patterns should overlap.)
The code you propose has two issues, in my opinion:
- It sets the "at least one device type hint missing in rcaps" case
*entirely apart*, and therefore that case would be handled *only* with a
warning. That's wrong: in that case, we *still* need the primary
handling, namely determining (and returning to the caller) a guest caps
tuple.
- The other issue is that the subcondition we want to associate with the
warning is "all device hints present" (alternatively, the negated form,
"at least one device hint missing). But the condition (pattern) you
wrote is something else: it says "all device hints missing". That's not
a condition we care about.
Put more simply, the "if / then / else" expression *subdivides* -- for
the sake of the secondary handling, i.e., for the warning -- the
existent condition [1]. Then, regardless of the branch taken in the "if
/ then / else", we still need to contine to the preexistent return
expression.
There's a bit more code that follows which might have to be factored out
Well yes -- if we factor out that that "tail code", then breaking the
logic into entirely separate patterns could be possible. But I still
think it would require a one-level-deeper nesting somewhere, and then
the "if / then / else" subdivision is not much different.
(but read on).
> let net_type =
> match net_type with
> | Some model -> model
The fundamental problem is that we should first remove rcaps, which
would simplify the whole thing.
I hesitate to submit patches for parts of the code where someone else
is working since it creates a mess of conflicts, but would you like me
to have a go at a patch for removing rcaps?
Yes, absolutely; please go ahead and do it. I'm happy to rebuild my
patches on top of that -- I won't really approach the new situation as a
rebase, with conflicts to resolve, but as a series of potential
individual cherry picks, and reimplementations from zero.
When removing code, my main question is always "granularity". I prefer
"finest granularity" in general, but that may not be right in this case
(or in the v2v projects in general). It depends quite a bit on reviewer
preference.
Thank you!
Laszlo
12:40 p.m.
New subject: [virt-v2v RFC wave 2 03/10] convert/windows_virtio: restrict the warning with virtio-win.iso absent
On Sat, Nov 06, 2021 at 04:45:33PM +0100, Laszlo Ersek wrote:
On 11/02/21 09:52, Richard W.M. Jones wrote:
> I hesitate to submit patches for parts of the code where someone else
> is working since it creates a mess of conflicts, but would you like me
> to have a go at a patch for removing rcaps?
Yes, absolutely; please go ahead and do it. I'm happy to rebuild my
patches on top of that -- I won't really approach the new situation as a
rebase, with conflicts to resolve, but as a series of potential
individual cherry picks, and reimplementations from zero.
I'll have a go at this, probably not going to be til Monday though.
When removing code, my main question is always
"granularity". I prefer
"finest granularity" in general, but that may not be right in this case
(or in the v2v projects in general). It depends quite a bit on reviewer
preference.
Finest granularity is usually good for me too.
I try to keep an eye out for what could be cherry-picked onto the
stable branch, thus splitting patches into bug fixes, refactoring, and
feature additions separately.
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
4:55 a.m.
New subject: [virt-v2v RFC wave 2 03/10] convert/windows_virtio: restrict the warning with virtio-win.iso absent
On 11/06/21 18:40, Richard W.M. Jones wrote:
On Sat, Nov 06, 2021 at 04:45:33PM +0100, Laszlo Ersek wrote:
> On 11/02/21 09:52, Richard W.M. Jones wrote:
>> I hesitate to submit patches for parts of the code where someone else
>> is working since it creates a mess of conflicts, but would you like me
>> to have a go at a patch for removing rcaps?
>
> Yes, absolutely; please go ahead and do it. I'm happy to rebuild my
> patches on top of that -- I won't really approach the new situation as a
> rebase, with conflicts to resolve, but as a series of potential
> individual cherry picks, and reimplementations from zero.
I'll have a go at this, probably not going to be til Monday though.
I'm done with the raw rebase (without handling the OVF, JSON and
OpenStack outputs).
I've noticed that the cirrus device model has become effectively dead
too, meaning both the Cirrus constructor of type "guestcaps_video_type",
and the Source_Cirrus constructor of type "source_video".
If I remove those constructors, then each type will be left with a
single constructor -- "guestcaps_video_type" will only have Standard_VGA
(not parametric), and "source_video" will have only "Source_other_video
of string" (basically: a string).
This suggests that "guestcaps_video_type" should be eliminated
altogether (its occurrences could be replaced by constants), and that
"source_video" should be replaced by plain "string".
(I'll go further: "source_video" looks unused beyond
"--print-source",
so it could be removed fully as well!)
Should I attempt doing this at the end of the series?
I'm asking now because these simplifications look technically possible
even before I start investigating the "OVF video device" topic. I expect
the latter to turn into an infinite mess, so if I can (or should) tack
the cirrus cleanup patches to the end of my series, I figure I'd like to
do that first.
Thanks!
Laszlo
5:08 a.m.
New subject: [virt-v2v RFC wave 2 03/10] convert/windows_virtio: restrict the warning with virtio-win.iso absent
On Tue, Nov 09, 2021 at 11:55:32AM +0100, Laszlo Ersek wrote:
On 11/06/21 18:40, Richard W.M. Jones wrote:
> On Sat, Nov 06, 2021 at 04:45:33PM +0100, Laszlo Ersek wrote:
>> On 11/02/21 09:52, Richard W.M. Jones wrote:
>>> I hesitate to submit patches for parts of the code where someone else
>>> is working since it creates a mess of conflicts, but would you like me
>>> to have a go at a patch for removing rcaps?
>>
>> Yes, absolutely; please go ahead and do it. I'm happy to rebuild my
>> patches on top of that -- I won't really approach the new situation as a
>> rebase, with conflicts to resolve, but as a series of potential
>> individual cherry picks, and reimplementations from zero.
>
> I'll have a go at this, probably not going to be til Monday though.
I'm done with the raw rebase (without handling the OVF, JSON and
OpenStack outputs).
I've noticed that the cirrus device model has become effectively dead
too, meaning both the Cirrus constructor of type "guestcaps_video_type",
and the Source_Cirrus constructor of type "source_video".
If I remove those constructors, then each type will be left with a
single constructor -- "guestcaps_video_type" will only have Standard_VGA
(not parametric), and "source_video" will have only "Source_other_video
of string" (basically: a string).
This suggests that "guestcaps_video_type" should be eliminated
altogether (its occurrences could be replaced by constants),
This bit certainly seems to make sense - if there's "only VGA" then
there's nothing that needs to be expressed by guestcaps_video_type.
Guestcaps is used to communicate from the conversion stage to the
output stage what the guest supports. Previous to your patch we had
to tell the output stage if the guest supported QXL or had to use
Cirrus. Presumably that's no longer needed.
and that
"source_video" should be replaced by plain "string".
(I'll go further: "source_video" looks unused beyond
"--print-source",
so it could be removed fully as well!)
Yes, --print-source is for debugging and troubleshooting virt-v2v. If
the information is not strictly required for v2v (and especially as it
is present elsewhere, eg in the source VMX or source libvirt XML) then
there's no need to print it.
Should I attempt doing this at the end of the series?
Seems entirely reasonable.
I'm asking now because these simplifications look technically
possible
even before I start investigating the "OVF video device" topic. I expect
the latter to turn into an infinite mess, so if I can (or should) tack
the cirrus cleanup patches to the end of my series, I figure I'd like to
do that first.
OVF is always "fun". Just remember it's not a standard, it's a
collection of formats, each targeted at a specific hypervisor,
sometimes several different formats for one hypervisor, that happen to
share some common XML elements some of the time.
Rich.
--
Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones
Read my programming and virtualization blog: http://rwmj.wordpress.com
Fedora Windows cross-compiler. Compile Windows programs, test, and
build Windows installers. Over 100 libraries supported.
http://fedoraproject.org/wiki/MinGW
5:18 a.m.
New subject: [virt-v2v RFC wave 2 03/10] convert/windows_virtio: restrict the warning with virtio-win.iso absent
On Tue, Nov 09, 2021 at 11:55:32AM +0100, Laszlo Ersek wrote:
I'm asking now because these simplifications look technically
possible
even before I start investigating the "OVF video device" topic. I expect
the latter to turn into an infinite mess, so if I can (or should) tack
the cirrus cleanup patches to the end of my series, I figure I'd like to
do that first.
On the QEMU side, assuming you don't care about guest OS dating back
from 1995, Cirrus should never be used under any circumstance [1].
Anywhere that might have used Cirrus should be switched to use VGA. So
if v2v does have any Cirrus related usage, I'd recommend to swap that
out as a priority.
Regards,
Daniel
[1] https://www.kraxel.org/blog/2014/10/qemu-using-cirrus-considered-harmful/
https://www.kraxel.org/blog/2019/09/display-devices-in-qemu/#tldr
--
|: https://berrange.com -o- https://www.flickr.com/photos/dberrange :|
|: https://libvirt.org -o- https://fstop138.berrange.com :|
|: https://entangle-photo.org -o- https://www.instagram.com/dberrange :|
7:02 a.m.
New subject: [virt-v2v RFC wave 2 03/10] convert/windows_virtio: restrict the warning with virtio-win.iso absent
On Tue, Nov 09, 2021 at 11:18:01AM +0000, Daniel P. Berrangé wrote:
On Tue, Nov 09, 2021 at 11:55:32AM +0100, Laszlo Ersek wrote:
> I'm asking now because these simplifications look technically possible
> even before I start investigating the "OVF video device" topic. I expect
> the latter to turn into an infinite mess, so if I can (or should) tack
> the cirrus cleanup patches to the end of my series, I figure I'd like to
> do that first.
On the QEMU side, assuming you don't care about guest OS dating back
from 1995, Cirrus should never be used under any circumstance [1].
Anywhere that might have used Cirrus should be switched to use VGA. So
if v2v does have any Cirrus related usage, I'd recommend to swap that
out as a priority.
Yup, that's what we're doing.
Rich.
Regards,
Daniel
[1] https://www.kraxel.org/blog/2014/10/qemu-using-cirrus-considered-harmful/
https://www.kraxel.org/blog/2019/09/display-devices-in-qemu/#tldr
--
|: https://berrange.com -o- https://www.flickr.com/photos/dberrange :|
|: https://libvirt.org -o- https://fstop138.berrange.com :|
|: https://entangle-photo.org -o- https://www.instagram.com/dberrange :|
--
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
12:06 p.m.
New subject: [virt-v2v RFC wave 2 04/10] convert/windows_virtio: handle Standard_VGA in the requested guest caps
Permit the conversion to request standard VGA display in the requested
guest capabilities structure.
Bugzilla: https://bugzilla.redhat.com/show_bug.cgi?id=1961107
Signed-off-by: Laszlo Ersek <lersek(a)redhat.com>
---
convert/windows_virtio.ml | 16 +++++++++-------
1 file changed, 9 insertions(+), 7 deletions(-)
diff --git a/convert/windows_virtio.ml b/convert/windows_virtio.ml
index 11b34a03f876..ee887fa8f715 100644
--- a/convert/windows_virtio.ml
+++ b/convert/windows_virtio.ml
@@ -51,7 +51,6 @@ let rec install_drivers ((g, _) as reg) inspect rcaps =
if not (copy_drivers g inspect driverdir) then (
match rcaps with
- | { rcaps_video = Some Standard_VGA } -> assert false
| { rcaps_block_bus = Some Virtio_blk | Some Virtio_SCSI }
| { rcaps_net_bus = Some Virtio_net }
| { rcaps_video = Some QXL } ->
@@ -61,7 +60,7 @@ let rec install_drivers ((g, _) as reg) inspect rcaps =
| { rcaps_block_bus = ((Some IDE | None) as block_type);
rcaps_net_bus = ((Some E1000 | Some RTL8139 | None) as net_type);
- rcaps_video = ((Some Cirrus | None) as video_type) } ->
+ rcaps_video = ((Some Standard_VGA | Some Cirrus | None) as video_type) } ->
if block_type = None || net_type = None || video_type = None then
warning (f_"there are no virtio drivers available for this version of
Windows (%d.%d %s %s). virt-v2v looks for drivers in %s\n\nThe guest will be configured
to use slower emulated devices.")
inspect.i_major_version inspect.i_minor_version inspect.i_arch
@@ -70,8 +69,12 @@ let rec install_drivers ((g, _) as reg) inspect rcaps =
let net_type =
match net_type with
| Some model -> model
- | None -> RTL8139 in
- (IDE, net_type, Cirrus, false, false, false, false)
+ | None -> RTL8139
+ and video_type =
+ match video_type with
+ | Some model -> model
+ | None -> Cirrus in
+ (IDE, net_type, video_type, false, false, false, false)
)
else (
(* Can we install the block driver? *)
@@ -162,7 +165,6 @@ let rec install_drivers ((g, _) as reg) inspect rcaps =
g#exists (driverdir // "qxl.inf") ||
g#exists (driverdir // "qxldod.inf") in
match rcaps.rcaps_video, has_qxl with
- | Some Standard_VGA, _ -> assert false
| Some QXL, false ->
error (f_"there is no QXL driver for this version of Windows (%d.%d %s).
virt-v2v looks for this driver in %s")
inspect.i_major_version inspect.i_minor_version
@@ -177,8 +179,8 @@ let rec install_drivers ((g, _) as reg) inspect rcaps =
| (Some QXL | None), true ->
QXL
- | Some Cirrus, _ ->
- Cirrus in
+ | Some (Cirrus | Standard_VGA as video_type), _ ->
+ video_type in
(* Did we install the miscellaneous drivers? *)
let virtio_rng_supported = g#exists (driverdir // "viorng.inf") in
--
2.19.1.3.g30247aa5d201
12:06 p.m.
New subject: [virt-v2v RFC wave 2 05/10] output: handle Standard_VGA in the libvirt XML and QEMU cmdline formats
Libvirt domain XML fragment (with the default video RAM size made
explicit):
<video>
<model type='vga' vram='16384' heads='1'/>
</video>
Example QEMU cmdline option:
-vga std
References:
- https://www.kraxel.org/blog/2019/09/display-devices-in-qemu/#VGA
- https://libvirt.org/formatdomain.html#video-devices
Bugzilla: https://bugzilla.redhat.com/show_bug.cgi?id=1961107
Signed-off-by: Laszlo Ersek <lersek(a)redhat.com>
---
Notes:
This is where further patches should be added for the OVF, JSON and
OpenStack output modes.
output/create_libvirt_xml.ml | 2 +-
output/output.ml | 2 +-
2 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/output/create_libvirt_xml.ml b/output/create_libvirt_xml.ml
index ec1cd455ef13..623755de7212 100644
--- a/output/create_libvirt_xml.ml
+++ b/output/create_libvirt_xml.ml
@@ -421,7 +421,7 @@ let create_libvirt_xml ?pool source inspect
let video =
let video_model =
match guestcaps.gcaps_video with
- | Standard_VGA -> assert false
+ | Standard_VGA -> e "model" [ "type", "vga";
"vram", "16384" ] []
| QXL -> e "model" [ "type", "qxl";
"ram", "65536" ] []
| Cirrus -> e "model" [ "type", "cirrus";
"vram", "9216" ] [] in
append_attr ("heads", "1") video_model;
diff --git a/output/output.ml b/output/output.ml
index 93dbd4a62865..b27d53059712 100644
--- a/output/output.ml
+++ b/output/output.ml
@@ -1540,7 +1540,7 @@ and qemu_finalize dir source inspect target_meta
);
arg "-vga"
(match guestcaps.gcaps_video with
- | Standard_VGA -> assert false
+ | Standard_VGA -> "std"
| Cirrus -> "cirrus"
| QXL -> "qxl"
)
--
2.19.1.3.g30247aa5d201
12:06 p.m.
New subject: [virt-v2v RFC wave 2 06/10] convert_linux: flip default target display to Standard_VGA
Policy change: if the requested guest caps do not express a preference for
any particular video type, pick Standard_VGA.
Bugzilla: https://bugzilla.redhat.com/show_bug.cgi?id=1961107
Signed-off-by: Laszlo Ersek <lersek(a)redhat.com>
---
convert/convert_linux.ml | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/convert/convert_linux.ml b/convert/convert_linux.ml
index 15f68fd4f8db..ae0bd39d1b2d 100644
--- a/convert/convert_linux.ml
+++ b/convert/convert_linux.ml
@@ -113,7 +113,7 @@ let convert (g : G.guestfs) source inspect keep_serial_console rcaps _
=
let video =
match rcaps.rcaps_video with
- | None -> QXL
+ | None -> Standard_VGA
| Some video -> video in
let block_type =
--
2.19.1.3.g30247aa5d201
12:06 p.m.
New subject: [virt-v2v RFC wave 2 07/10] convert/windows_virtio: flip default target display to Standard_VGA
Policy change: if the requested guest caps do not express a preference for
any particular video type, pick Standard_VGA. In detail, this means the
following:
- If the subject Windows version is unsupported by virtio-win.iso (either
entirely, or regarding just the QXL driver), then do not warn the user
that we pick "something else than QXL".
That's because we don't perceive Standard_VGA inferior to QXL.
- Pick Cirrus too only if specifically requested.
- Map the following patterns of the (rcaps.rcaps_video, has_qxl) tuple:
- (None, false)
- (None, true)
to Standard_VGA. All six other patterns preserve their previous
mappings.
Bugzilla: https://bugzilla.redhat.com/show_bug.cgi?id=1961107
Signed-off-by: Laszlo Ersek <lersek(a)redhat.com>
---
Notes:
This patch breaks the following test cases:
- test-v2v-o-glance.sh
- test-v2v-o-json.sh
- test-v2v-o-openstack.sh
That's because we still have the initial "assert false" expressions for
Standard_VGA in the OpenStack and JSON (... and OVF) producers, from
patch 'lib/types: introduce "Standard_VGA" constructor for
"guestcaps_video_type'.
convert/windows_virtio.ml | 13 +++++--------
tests/test-v2v-i-ova.xml | 2 +-
2 files changed, 6 insertions(+), 9 deletions(-)
diff --git a/convert/windows_virtio.ml b/convert/windows_virtio.ml
index ee887fa8f715..1a47a1500076 100644
--- a/convert/windows_virtio.ml
+++ b/convert/windows_virtio.ml
@@ -61,7 +61,7 @@ let rec install_drivers ((g, _) as reg) inspect rcaps =
| { rcaps_block_bus = ((Some IDE | None) as block_type);
rcaps_net_bus = ((Some E1000 | Some RTL8139 | None) as net_type);
rcaps_video = ((Some Standard_VGA | Some Cirrus | None) as video_type) } ->
- if block_type = None || net_type = None || video_type = None then
+ if block_type = None || net_type = None then
warning (f_"there are no virtio drivers available for this version of
Windows (%d.%d %s %s). virt-v2v looks for drivers in %s\n\nThe guest will be configured
to use slower emulated devices.")
inspect.i_major_version inspect.i_minor_version inspect.i_arch
inspect.i_product_variant virtio_win
@@ -73,7 +73,7 @@ let rec install_drivers ((g, _) as reg) inspect rcaps =
and video_type =
match video_type with
| Some model -> model
- | None -> Cirrus in
+ | None -> Standard_VGA in
(IDE, net_type, video_type, false, false, false, false)
)
else (
@@ -170,13 +170,10 @@ let rec install_drivers ((g, _) as reg) inspect rcaps =
inspect.i_major_version inspect.i_minor_version
inspect.i_arch virtio_win
- | None, false ->
- warning (f_"there is no QXL driver for this version of Windows (%d.%d %s).
virt-v2v looks for this driver in %s\n\nThe guest will be configured to use a basic VGA
display driver.")
- inspect.i_major_version inspect.i_minor_version
- inspect.i_arch virtio_win;
- Cirrus
+ | None, _ ->
+ Standard_VGA
- | (Some QXL | None), true ->
+ | Some QXL, true ->
QXL
| Some (Cirrus | Standard_VGA as video_type), _ ->
diff --git a/tests/test-v2v-i-ova.xml b/tests/test-v2v-i-ova.xml
index 30f52f557d9f..d7383905fdc0 100644
--- a/tests/test-v2v-i-ova.xml
+++ b/tests/test-v2v-i-ova.xml
@@ -39,7 +39,7 @@
<model type='virtio'/>
</interface>
<video>
- <model type='qxl' ram='65536' heads='1'/>
+ <model type='vga' vram='16384' heads='1'/>
</video>
<graphics type='vnc' autoport='yes' port='-1'/>
<rng model='virtio'>
--
2.19.1.3.g30247aa5d201
12:06 p.m.
New subject: [virt-v2v RFC wave 2 08/10] lib/types: remove "QXL" constructor for "guestcaps_video_type"
Removing the "QXL" constructor mostly needs the obvious fixup in the
various pattern matches. Worth highlighting is
"convert/windows_virtio.ml": the somewhat quirky QXL branches disappear,
and the resultant display selections match the Linux guest case (i.e.,
default: stdvga, or else stdvga/cirrus per requested guest caps).
Bugzilla: https://bugzilla.redhat.com/show_bug.cgi?id=1961107
Signed-off-by: Laszlo Ersek <lersek(a)redhat.com>
---
convert/convert_linux.ml | 1 -
convert/windows_virtio.ml | 29 ++++----------------
lib/types.ml | 3 +-
lib/types.mli | 2 +-
output/create_json.ml | 1 -
output/create_libvirt_xml.ml | 1 -
output/openstack_image_properties.ml | 1 -
output/output.ml | 1 -
8 files changed, 8 insertions(+), 31 deletions(-)
diff --git a/convert/convert_linux.ml b/convert/convert_linux.ml
index ae0bd39d1b2d..20fe68424413 100644
--- a/convert/convert_linux.ml
+++ b/convert/convert_linux.ml
@@ -832,7 +832,6 @@ let convert (g : G.guestfs) source inspect keep_serial_console rcaps _
=
let video_driver =
match video with
| Standard_VGA -> "modesetting"
- | QXL -> "qxl"
| Cirrus -> "cirrus" in
let updated = ref false in
diff --git a/convert/windows_virtio.ml b/convert/windows_virtio.ml
index 1a47a1500076..e69deab2fccc 100644
--- a/convert/windows_virtio.ml
+++ b/convert/windows_virtio.ml
@@ -52,15 +52,13 @@ let rec install_drivers ((g, _) as reg) inspect rcaps =
if not (copy_drivers g inspect driverdir) then (
match rcaps with
| { rcaps_block_bus = Some Virtio_blk | Some Virtio_SCSI }
- | { rcaps_net_bus = Some Virtio_net }
- | { rcaps_video = Some QXL } ->
+ | { rcaps_net_bus = Some Virtio_net } ->
error (f_"there are no virtio drivers available for this version of Windows
(%d.%d %s %s). virt-v2v looks for drivers in %s")
inspect.i_major_version inspect.i_minor_version inspect.i_arch
inspect.i_product_variant virtio_win
| { rcaps_block_bus = ((Some IDE | None) as block_type);
- rcaps_net_bus = ((Some E1000 | Some RTL8139 | None) as net_type);
- rcaps_video = ((Some Standard_VGA | Some Cirrus | None) as video_type) } ->
+ rcaps_net_bus = ((Some E1000 | Some RTL8139 | None) as net_type) } ->
if block_type = None || net_type = None then
warning (f_"there are no virtio drivers available for this version of
Windows (%d.%d %s %s). virt-v2v looks for drivers in %s\n\nThe guest will be configured
to use slower emulated devices.")
inspect.i_major_version inspect.i_minor_version inspect.i_arch
@@ -71,7 +69,7 @@ let rec install_drivers ((g, _) as reg) inspect rcaps =
| Some model -> model
| None -> RTL8139
and video_type =
- match video_type with
+ match rcaps.rcaps_video with
| Some model -> model
| None -> Standard_VGA in
(IDE, net_type, video_type, false, false, false, false)
@@ -159,25 +157,10 @@ let rec install_drivers ((g, _) as reg) inspect rcaps =
| Some net_type, _ ->
net_type in
- (* Can we install the QXL driver? *)
let video : guestcaps_video_type =
- let has_qxl =
- g#exists (driverdir // "qxl.inf") ||
- g#exists (driverdir // "qxldod.inf") in
- match rcaps.rcaps_video, has_qxl with
- | Some QXL, false ->
- error (f_"there is no QXL driver for this version of Windows (%d.%d %s).
virt-v2v looks for this driver in %s")
- inspect.i_major_version inspect.i_minor_version
- inspect.i_arch virtio_win
-
- | None, _ ->
- Standard_VGA
-
- | Some QXL, true ->
- QXL
-
- | Some (Cirrus | Standard_VGA as video_type), _ ->
- video_type in
+ match rcaps.rcaps_video with
+ | None -> Standard_VGA
+ | Some video_type -> video_type in
(* Did we install the miscellaneous drivers? *)
let virtio_rng_supported = g#exists (driverdir // "viorng.inf") in
diff --git a/lib/types.ml b/lib/types.ml
index 0156d55954e9..347f3e6c2672 100644
--- a/lib/types.ml
+++ b/lib/types.ml
@@ -429,7 +429,7 @@ and requested_guestcaps = {
}
and guestcaps_block_type = Virtio_blk | Virtio_SCSI | IDE
and guestcaps_net_type = Virtio_net | E1000 | RTL8139
-and guestcaps_video_type = Standard_VGA | QXL | Cirrus
+and guestcaps_video_type = Standard_VGA | Cirrus
and guestcaps_machine = I440FX | Q35 | Virt
let string_of_block_type = function
@@ -442,7 +442,6 @@ let string_of_net_type = function
| RTL8139 -> "rtl8139"
let string_of_video = function
| Standard_VGA -> "stdvga"
- | QXL -> "qxl"
| Cirrus -> "cirrus"
let string_of_machine = function
| I440FX -> "i440fx"
diff --git a/lib/types.mli b/lib/types.mli
index 19790f375035..43ec0f971a3c 100644
--- a/lib/types.mli
+++ b/lib/types.mli
@@ -293,7 +293,7 @@ and requested_guestcaps = {
and guestcaps_block_type = Virtio_blk | Virtio_SCSI | IDE
and guestcaps_net_type = Virtio_net | E1000 | RTL8139
-and guestcaps_video_type = Standard_VGA | QXL | Cirrus
+and guestcaps_video_type = Standard_VGA | Cirrus
and guestcaps_machine = I440FX | Q35 | Virt
val string_of_guestcaps : guestcaps -> string
diff --git a/output/create_json.ml b/output/create_json.ml
index aae4d09c59bd..ef866143d6d5 100644
--- a/output/create_json.ml
+++ b/output/create_json.ml
@@ -205,7 +205,6 @@ let create_json_metadata source inspect
let video =
match guestcaps.gcaps_video with
| Standard_VGA -> assert false
- | QXL -> "qxl"
| Cirrus -> "cirrus" in
let machine =
match guestcaps.gcaps_machine with
diff --git a/output/create_libvirt_xml.ml b/output/create_libvirt_xml.ml
index 623755de7212..9cf872861110 100644
--- a/output/create_libvirt_xml.ml
+++ b/output/create_libvirt_xml.ml
@@ -422,7 +422,6 @@ let create_libvirt_xml ?pool source inspect
let video_model =
match guestcaps.gcaps_video with
| Standard_VGA -> e "model" [ "type", "vga";
"vram", "16384" ] []
- | QXL -> e "model" [ "type", "qxl";
"ram", "65536" ] []
| Cirrus -> e "model" [ "type", "cirrus";
"vram", "9216" ] [] in
append_attr ("heads", "1") video_model;
e "video" [] [ video_model ] in
diff --git a/output/openstack_image_properties.ml b/output/openstack_image_properties.ml
index 592da3e40e44..cbf2e9533178 100644
--- a/output/openstack_image_properties.ml
+++ b/output/openstack_image_properties.ml
@@ -45,7 +45,6 @@ let create source inspect { target_buses; guestcaps; target_firmware }
=
"hw_video_model",
(match guestcaps.gcaps_video with
| Standard_VGA -> assert false
- | QXL -> "qxl"
| Cirrus -> "cirrus");
"hw_machine_type",
(match guestcaps.gcaps_machine with
diff --git a/output/output.ml b/output/output.ml
index b27d53059712..4218354f8330 100644
--- a/output/output.ml
+++ b/output/output.ml
@@ -1542,7 +1542,6 @@ and qemu_finalize dir source inspect target_meta
(match guestcaps.gcaps_video with
| Standard_VGA -> "std"
| Cirrus -> "cirrus"
- | QXL -> "qxl"
)
);
--
2.19.1.3.g30247aa5d201
12:06 p.m.
New subject: [virt-v2v RFC wave 2 09/10] tests: remove the fake Windows guest drivers for the QXL device
The QXL drivers no longer matter when converting Windows guests; remove
the fake images and their references from the test suite.
Bugzilla: https://bugzilla.redhat.com/show_bug.cgi?id=1961107
Signed-off-by: Laszlo Ersek <lersek(a)redhat.com>
---
test-data/fake-virtio-win/Makefile.am | 16 ----------------
test-data/fake-virtio-win/drivers/amd64/Win2008R2/qxl.inf | 2 --
test-data/fake-virtio-win/drivers/amd64/Win7/qxl.inf | 2 --
test-data/fake-virtio-win/drivers/i386/Win7/qxl.inf | 2 --
test-data/fake-virtio-win/drivers/i386/WinXP/qxl.inf | 2 --
test-data/fake-virtio-win/drivers/amd64/Win2008R2/qxl.cat | 1 -
test-data/fake-virtio-win/drivers/amd64/Win2008R2/qxl.sys | 1 -
test-data/fake-virtio-win/drivers/amd64/Win2008R2/qxldd.dll | 1 -
test-data/fake-virtio-win/drivers/amd64/Win7/qxl.cat | 1 -
test-data/fake-virtio-win/drivers/amd64/Win7/qxl.sys | 1 -
test-data/fake-virtio-win/drivers/amd64/Win7/qxldd.dll | 1 -
test-data/fake-virtio-win/drivers/i386/Win7/qxl.cat | 1 -
test-data/fake-virtio-win/drivers/i386/Win7/qxl.sys | 1 -
test-data/fake-virtio-win/drivers/i386/Win7/qxldd.dll | 1 -
test-data/fake-virtio-win/drivers/i386/WinXP/qxl.cat | 1 -
test-data/fake-virtio-win/drivers/i386/WinXP/qxl.sys | 1 -
test-data/fake-virtio-win/drivers/i386/WinXP/qxldd.dll | 1 -
tests/test-v2v-in-place.sh | 2 +-
tests/test-v2v-windows-conversion.sh | 2 +-
v2v/v2v_unit_tests.ml | 16 ----------------
20 files changed, 2 insertions(+), 54 deletions(-)
diff --git a/test-data/fake-virtio-win/Makefile.am
b/test-data/fake-virtio-win/Makefile.am
index 55b91a95a5e5..60c2d8245db9 100644
--- a/test-data/fake-virtio-win/Makefile.am
+++ b/test-data/fake-virtio-win/Makefile.am
@@ -38,10 +38,6 @@ drivers = \
drivers/amd64/Win2008R2/netkvm.cat \
drivers/amd64/Win2008R2/netkvm.inf \
drivers/amd64/Win2008R2/netkvm.sys \
- drivers/amd64/Win2008R2/qxl.cat \
- drivers/amd64/Win2008R2/qxldd.dll \
- drivers/amd64/Win2008R2/qxl.inf \
- drivers/amd64/Win2008R2/qxl.sys \
drivers/amd64/Win2008R2/vioscsi.cat \
drivers/amd64/Win2008R2/vioscsi.inf \
drivers/amd64/Win2008R2/vioscsi.sys \
@@ -75,10 +71,6 @@ drivers = \
drivers/amd64/Win7/netkvm.cat \
drivers/amd64/Win7/netkvm.inf \
drivers/amd64/Win7/netkvm.sys \
- drivers/amd64/Win7/qxl.cat \
- drivers/amd64/Win7/qxldd.dll \
- drivers/amd64/Win7/qxl.inf \
- drivers/amd64/Win7/qxl.sys \
drivers/amd64/Win7/vioscsi.cat \
drivers/amd64/Win7/vioscsi.inf \
drivers/amd64/Win7/vioscsi.sys \
@@ -121,10 +113,6 @@ drivers = \
drivers/i386/Win7/netkvm.cat \
drivers/i386/Win7/netkvm.inf \
drivers/i386/Win7/netkvm.sys \
- drivers/i386/Win7/qxl.cat \
- drivers/i386/Win7/qxldd.dll \
- drivers/i386/Win7/qxl.inf \
- drivers/i386/Win7/qxl.sys \
drivers/i386/Win7/vioscsi.cat \
drivers/i386/Win7/vioscsi.inf \
drivers/i386/Win7/vioscsi.sys \
@@ -152,10 +140,6 @@ drivers = \
drivers/i386/WinXP/netkvm.cat \
drivers/i386/WinXP/netkvm.inf \
drivers/i386/WinXP/netkvm.sys \
- drivers/i386/WinXP/qxl.cat \
- drivers/i386/WinXP/qxldd.dll \
- drivers/i386/WinXP/qxl.inf \
- drivers/i386/WinXP/qxl.sys \
drivers/i386/WinXP/viostor.cat \
drivers/i386/WinXP/viostor.inf \
drivers/i386/WinXP/viostor.sys
diff --git a/test-data/fake-virtio-win/drivers/amd64/Win2008R2/qxl.inf
b/test-data/fake-virtio-win/drivers/amd64/Win2008R2/qxl.inf
deleted file mode 100644
index 67e3fab0f3a7..000000000000
--- a/test-data/fake-virtio-win/drivers/amd64/Win2008R2/qxl.inf
+++ /dev/null
@@ -1,2 +0,0 @@
-[Version]
-DriverVer = 10/29/2013,6.1.0.10020
diff --git a/test-data/fake-virtio-win/drivers/amd64/Win7/qxl.inf
b/test-data/fake-virtio-win/drivers/amd64/Win7/qxl.inf
deleted file mode 100644
index 3310167e395f..000000000000
--- a/test-data/fake-virtio-win/drivers/amd64/Win7/qxl.inf
+++ /dev/null
@@ -1,2 +0,0 @@
-[Version]
-DriverVer = 07/17/2013,6.1.0.10018
diff --git a/test-data/fake-virtio-win/drivers/i386/Win7/qxl.inf
b/test-data/fake-virtio-win/drivers/i386/Win7/qxl.inf
deleted file mode 100644
index 3310167e395f..000000000000
--- a/test-data/fake-virtio-win/drivers/i386/Win7/qxl.inf
+++ /dev/null
@@ -1,2 +0,0 @@
-[Version]
-DriverVer = 07/17/2013,6.1.0.10018
diff --git a/test-data/fake-virtio-win/drivers/i386/WinXP/qxl.inf
b/test-data/fake-virtio-win/drivers/i386/WinXP/qxl.inf
deleted file mode 100644
index d1fe96b24e79..000000000000
--- a/test-data/fake-virtio-win/drivers/i386/WinXP/qxl.inf
+++ /dev/null
@@ -1,2 +0,0 @@
-[Version]
-DriverVer = 07/17/2013,5.1.0.10018
diff --git a/test-data/fake-virtio-win/drivers/amd64/Win2008R2/qxl.cat
b/test-data/fake-virtio-win/drivers/amd64/Win2008R2/qxl.cat
deleted file mode 100644
index acbc962012bc..000000000000
--- a/test-data/fake-virtio-win/drivers/amd64/Win2008R2/qxl.cat
+++ /dev/null
@@ -1 +0,0 @@
-fake
diff --git a/test-data/fake-virtio-win/drivers/amd64/Win2008R2/qxl.sys
b/test-data/fake-virtio-win/drivers/amd64/Win2008R2/qxl.sys
deleted file mode 100644
index acbc962012bc..000000000000
--- a/test-data/fake-virtio-win/drivers/amd64/Win2008R2/qxl.sys
+++ /dev/null
@@ -1 +0,0 @@
-fake
diff --git a/test-data/fake-virtio-win/drivers/amd64/Win2008R2/qxldd.dll
b/test-data/fake-virtio-win/drivers/amd64/Win2008R2/qxldd.dll
deleted file mode 100644
index acbc962012bc..000000000000
--- a/test-data/fake-virtio-win/drivers/amd64/Win2008R2/qxldd.dll
+++ /dev/null
@@ -1 +0,0 @@
-fake
diff --git a/test-data/fake-virtio-win/drivers/amd64/Win7/qxl.cat
b/test-data/fake-virtio-win/drivers/amd64/Win7/qxl.cat
deleted file mode 100644
index acbc962012bc..000000000000
--- a/test-data/fake-virtio-win/drivers/amd64/Win7/qxl.cat
+++ /dev/null
@@ -1 +0,0 @@
-fake
diff --git a/test-data/fake-virtio-win/drivers/amd64/Win7/qxl.sys
b/test-data/fake-virtio-win/drivers/amd64/Win7/qxl.sys
deleted file mode 100644
index acbc962012bc..000000000000
--- a/test-data/fake-virtio-win/drivers/amd64/Win7/qxl.sys
+++ /dev/null
@@ -1 +0,0 @@
-fake
diff --git a/test-data/fake-virtio-win/drivers/amd64/Win7/qxldd.dll
b/test-data/fake-virtio-win/drivers/amd64/Win7/qxldd.dll
deleted file mode 100644
index acbc962012bc..000000000000
--- a/test-data/fake-virtio-win/drivers/amd64/Win7/qxldd.dll
+++ /dev/null
@@ -1 +0,0 @@
-fake
diff --git a/test-data/fake-virtio-win/drivers/i386/Win7/qxl.cat
b/test-data/fake-virtio-win/drivers/i386/Win7/qxl.cat
deleted file mode 100644
index acbc962012bc..000000000000
--- a/test-data/fake-virtio-win/drivers/i386/Win7/qxl.cat
+++ /dev/null
@@ -1 +0,0 @@
-fake
diff --git a/test-data/fake-virtio-win/drivers/i386/Win7/qxl.sys
b/test-data/fake-virtio-win/drivers/i386/Win7/qxl.sys
deleted file mode 100644
index acbc962012bc..000000000000
--- a/test-data/fake-virtio-win/drivers/i386/Win7/qxl.sys
+++ /dev/null
@@ -1 +0,0 @@
-fake
diff --git a/test-data/fake-virtio-win/drivers/i386/Win7/qxldd.dll
b/test-data/fake-virtio-win/drivers/i386/Win7/qxldd.dll
deleted file mode 100644
index acbc962012bc..000000000000
--- a/test-data/fake-virtio-win/drivers/i386/Win7/qxldd.dll
+++ /dev/null
@@ -1 +0,0 @@
-fake
diff --git a/test-data/fake-virtio-win/drivers/i386/WinXP/qxl.cat
b/test-data/fake-virtio-win/drivers/i386/WinXP/qxl.cat
deleted file mode 100644
index acbc962012bc..000000000000
--- a/test-data/fake-virtio-win/drivers/i386/WinXP/qxl.cat
+++ /dev/null
@@ -1 +0,0 @@
-fake
diff --git a/test-data/fake-virtio-win/drivers/i386/WinXP/qxl.sys
b/test-data/fake-virtio-win/drivers/i386/WinXP/qxl.sys
deleted file mode 100644
index acbc962012bc..000000000000
--- a/test-data/fake-virtio-win/drivers/i386/WinXP/qxl.sys
+++ /dev/null
@@ -1 +0,0 @@
-fake
diff --git a/test-data/fake-virtio-win/drivers/i386/WinXP/qxldd.dll
b/test-data/fake-virtio-win/drivers/i386/WinXP/qxldd.dll
deleted file mode 100644
index acbc962012bc..000000000000
--- a/test-data/fake-virtio-win/drivers/i386/WinXP/qxldd.dll
+++ /dev/null
@@ -1 +0,0 @@
-fake
diff --git a/tests/test-v2v-in-place.sh b/tests/test-v2v-in-place.sh
index 9da8fa63a99b..c5fee2f9f591 100755
--- a/tests/test-v2v-in-place.sh
+++ b/tests/test-v2v-in-place.sh
@@ -95,7 +95,7 @@ mktest "is-file \"$firstboot_dir/firstboot.bat\""
true
mktest "is-dir \"$firstboot_dir/scripts\"" true
virtio_dir="/Windows/Drivers/VirtIO"
mktest "is-dir \"$virtio_dir\"" true
-for drv in netkvm qxl vioscsi viostor; do
+for drv in netkvm vioscsi viostor; do
for sfx in cat inf sys; do
mktest "is-file \"$virtio_dir/$drv.$sfx\"" true
done
diff --git a/tests/test-v2v-windows-conversion.sh b/tests/test-v2v-windows-conversion.sh
index aeab9a48b507..a4cf191d74b4 100755
--- a/tests/test-v2v-windows-conversion.sh
+++ b/tests/test-v2v-windows-conversion.sh
@@ -82,7 +82,7 @@ mktest "is-file \"$firstboot_dir/firstboot.bat\""
true
mktest "is-dir \"$firstboot_dir/scripts\"" true
virtio_dir="/Windows/Drivers/VirtIO"
mktest "is-dir \"$virtio_dir\"" true
-for drv in netkvm qxl vioscsi viostor; do
+for drv in netkvm vioscsi viostor; do
for sfx in cat inf sys; do
mktest "is-file \"$virtio_dir/$drv.$sfx\"" true
done
diff --git a/v2v/v2v_unit_tests.ml b/v2v/v2v_unit_tests.ml
index 584c5bb7b395..889f7998cf3b 100644
--- a/v2v/v2v_unit_tests.ml
+++ b/v2v/v2v_unit_tests.ml
@@ -603,15 +603,11 @@ let test_virtio_iso_path_matches_guest_os ctx =
"drivers/i386/Win2008/vioscsi.inf", Some win2k8_32;
"drivers/i386/Win7/viostor.inf", Some win7_32;
"drivers/i386/Win7/viostor.sys", Some win7_32;
- "drivers/i386/Win7/qxldd.dll", Some win7_32;
- "drivers/i386/Win7/qxl.sys", Some win7_32;
"drivers/i386/Win7/vioscsi.cat", Some win7_32;
"drivers/i386/Win7/netkvm.inf", Some win7_32;
"drivers/i386/Win7/netkvm.sys", Some win7_32;
"drivers/i386/Win7/viostor.cat", Some win7_32;
- "drivers/i386/Win7/qxl.inf", Some win7_32;
"drivers/i386/Win7/vioscsi.sys", Some win7_32;
- "drivers/i386/Win7/qxl.cat", Some win7_32;
"drivers/i386/Win7/netkvm.cat", Some win7_32;
"drivers/i386/Win7/vioscsi.inf", Some win7_32;
"drivers/i386/Win2003/viostor.inf", Some win2k3_32;
@@ -631,13 +627,9 @@ let test_virtio_iso_path_matches_guest_os ctx =
"drivers/i386/Win8/vioscsi.inf", Some win8_32;
"drivers/i386/WinXP/viostor.inf", Some winxp_32;
"drivers/i386/WinXP/viostor.sys", Some winxp_32;
- "drivers/i386/WinXP/qxldd.dll", Some winxp_32;
- "drivers/i386/WinXP/qxl.sys", Some winxp_32;
"drivers/i386/WinXP/netkvm.inf", Some winxp_32;
"drivers/i386/WinXP/netkvm.sys", Some winxp_32;
"drivers/i386/WinXP/viostor.cat", Some winxp_32;
- "drivers/i386/WinXP/qxl.inf", Some winxp_32;
- "drivers/i386/WinXP/qxl.cat", Some winxp_32;
"drivers/i386/WinXP/netkvm.cat", Some winxp_32;
"drivers/amd64/Win8.1/viostor.inf", Some win8_1_64;
"drivers/amd64/Win8.1/viostor.sys", Some win8_1_64;
@@ -659,15 +651,11 @@ let test_virtio_iso_path_matches_guest_os ctx =
"drivers/amd64/Win2008/vioscsi.inf", Some win2k8_64;
"drivers/amd64/Win7/viostor.inf", Some win7_64;
"drivers/amd64/Win7/viostor.sys", Some win7_64;
- "drivers/amd64/Win7/qxldd.dll", Some win7_64;
- "drivers/amd64/Win7/qxl.sys", Some win7_64;
"drivers/amd64/Win7/vioscsi.cat", Some win7_64;
"drivers/amd64/Win7/netkvm.inf", Some win7_64;
"drivers/amd64/Win7/netkvm.sys", Some win7_64;
"drivers/amd64/Win7/viostor.cat", Some win7_64;
- "drivers/amd64/Win7/qxl.inf", Some win7_64;
"drivers/amd64/Win7/vioscsi.sys", Some win7_64;
- "drivers/amd64/Win7/qxl.cat", Some win7_64;
"drivers/amd64/Win7/netkvm.cat", Some win7_64;
"drivers/amd64/Win7/vioscsi.inf", Some win7_64;
"drivers/amd64/Win2003/viostor.inf", Some win2k3_64;
@@ -696,15 +684,11 @@ let test_virtio_iso_path_matches_guest_os ctx =
"drivers/amd64/Win2012/vioscsi.inf", Some win2k12_64;
"drivers/amd64/Win2008R2/viostor.inf", Some win2k8r2_64;
"drivers/amd64/Win2008R2/viostor.sys", Some win2k8r2_64;
- "drivers/amd64/Win2008R2/qxldd.dll", Some win2k8r2_64;
- "drivers/amd64/Win2008R2/qxl.sys", Some win2k8r2_64;
"drivers/amd64/Win2008R2/vioscsi.cat", Some win2k8r2_64;
"drivers/amd64/Win2008R2/netkvm.inf", Some win2k8r2_64;
"drivers/amd64/Win2008R2/netkvm.sys", Some win2k8r2_64;
"drivers/amd64/Win2008R2/viostor.cat", Some win2k8r2_64;
- "drivers/amd64/Win2008R2/qxl.inf", Some win2k8r2_64;
"drivers/amd64/Win2008R2/vioscsi.sys", Some win2k8r2_64;
- "drivers/amd64/Win2008R2/qxl.cat", Some win2k8r2_64;
"drivers/amd64/Win2008R2/netkvm.cat", Some win2k8r2_64;
"drivers/amd64/Win2008R2/vioscsi.inf", Some win2k8r2_64;
"drivers/amd64/Win2012R2/viostor.inf", Some win2k12r2_64;
--
2.19.1.3.g30247aa5d201
12:06 p.m.
New subject: [virt-v2v RFC wave 2 10/10] tests: remove the remaining QXL references (for completeness)
Just for completeness' sake, change the display device from QXL to
standard VGA in:
- test-data/phony-guests/guests.xml.in
- tests/test-v2v-print-source.xml.in
This removes the last QXL occurrences in the source files.
Bugzilla: https://bugzilla.redhat.com/show_bug.cgi?id=1961107
Signed-off-by: Laszlo Ersek <lersek(a)redhat.com>
---
test-data/phony-guests/guests.xml.in | 2 +-
tests/test-v2v-print-source.expected | 2 +-
tests/test-v2v-print-source.xml.in | 2 +-
3 files changed, 3 insertions(+), 3 deletions(-)
diff --git a/test-data/phony-guests/guests.xml.in b/test-data/phony-guests/guests.xml.in
index 4139d04f689f..9f64c35cd80f 100644
--- a/test-data/phony-guests/guests.xml.in
+++ b/test-data/phony-guests/guests.xml.in
@@ -299,7 +299,7 @@
<model type='virtio'/>
</interface>
<video>
- <model type='qxl' ram='65536' vram='65536'
vgamem='16384' heads='1'/>
+ <model type='vga' vram='16384' heads='1'/>
</video>
</devices>
</domain>
diff --git a/tests/test-v2v-print-source.expected b/tests/test-v2v-print-source.expected
index afed30b5c46d..cfeade1c012d 100644
--- a/tests/test-v2v-print-source.expected
+++ b/tests/test-v2v-print-source.expected
@@ -9,7 +9,7 @@ hypervisor type: kvm
CPU features: acpi,apic,pae
firmware: unknown
display:
- video: qxl
+ video: vga
sound:
disks:
0 [virtio-blk]
diff --git a/tests/test-v2v-print-source.xml.in b/tests/test-v2v-print-source.xml.in
index 2f83fc73cb8a..d018d5fbcd27 100644
--- a/tests/test-v2v-print-source.xml.in
+++ b/tests/test-v2v-print-source.xml.in
@@ -27,7 +27,7 @@
<model type='virtio'/>
</interface>
<video>
- <model type='qxl' ram='65536' vram='65536'
vgamem='16384' heads='1'/>
+ <model type='vga' vram='16384' heads='1'/>
</video>
</devices>
</domain>
--
2.19.1.3.g30247aa5d201
4:06 a.m.
New subject: [virt-v2v RFC wave 2 10/10] tests: remove the remaining QXL references (for completeness)
The remaining patches make sense, thanks.
Rich.
--
Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones
Read my programming and virtualization blog: http://rwmj.wordpress.com
virt-top is 'top' for virtual machines. Tiny program with many
powerful monitoring features, net stats, disk stats, logging, etc.
http://people.redhat.com/~rjones/virt-top
1109
days inactive
1117
days old
21 comments
4 participants
participants (4)
-
Daniel P. Berrangé -
Gerd Hoffmann -
Laszlo Ersek -
Richard W.M. Jones