similar to: [PATCH 0/4] v2v: simplify Windows registry patching

Hi there, SUSE ships Virtual Machine Driver Pack for the virtio windows drivers. Get v2v and customize to discover them and use them if available. Cédric Bosdonnat (7): v2v: check next free oem%d.inf in /Windows/Inf v2v: extract controller offset discovery as a function customize: add support for pvvxsvc v2v: extract reusable parts of viostor regedits v2v: adapt the subkey in Enum

There are registry entries that are needed to add some other drivers. Extracting them into a function will help adding SUSE VMDP support. --- v2v/windows_virtio.ml | 311 ++++++++++++++++++++++++++++---------------------- 1 file changed, 176 insertions(+), 135 deletions(-) diff --git a/v2v/windows_virtio.ml b/v2v/windows_virtio.ml index 14ffc51..8a0b529 100644 --- a/v2v/windows_virtio.ml +++

On Tue, Apr 05, 2016 at 01:47:27PM +0200, Cédric Bosdonnat wrote: > + let oem_inf = set_free_oem_inf g root scsi_adapter_guid "viostor.inf" driverdir in Seems better if it was called *get_next*_free_oem_inf? > > (* There should be a key > * HKLM\SYSTEM\ControlSet001\Control\Class\<scsi_adapter_guid> > @@ -398,6 +378,28 @@ and

On Mon, Apr 03, 2017 at 07:44:45PM +0300, Pavel Butsykin wrote: > This patch fixes the problem of the first boot after installation virtio block > drivers on Windows 10 / Server 2016. The problem is related to the change in > the behavior of Windows-Kernel-Pnp. To fix we need to add a field "Version" to > HKLM/SYSTEM/DriverDatabase/DriverPackages/drv_inf_label then

A very miscellaneous set of cleanups to how we handle the Windows registry in virt-v2v, firstboot, and inspection code. This should all be straightforward non-controversial refactoring. Some highlights: - Add a new mllib Registry module containing various utility functions that are currently scattered all around. - Only compute the software/system hive paths once during inspection, and

[Partly using the mailing list as a backup again ...] This commit changes virt-v2v to make registry changes to all ControlSets, not just the CurrentControlSet. Not sure what difference if any this would make. Rich.

It seems that checking for oem%d.inf in the DeviceIds registry entry doesn't always list all oemXX.inf files. For example we may have oem1.inf free in the registry key, but used in another one. Also extract this into a separate function for later use to setup another driver. --- v2v/windows_virtio.ml | 52 ++++++++++++++++++++++++++------------------------- 1 file changed, 27 insertions(+),

On Tue, 2016-04-05 at 13:04 +0100, Richard W.M. Jones wrote: > On Tue, Apr 05, 2016 at 01:47:27PM +0200, Cédric Bosdonnat wrote: > > + let oem_inf = set_free_oem_inf g root scsi_adapter_guid > > "viostor.inf" driverdir in > > Seems better if it was called *get_next*_free_oem_inf? Yes, sounds better. > > (* There should be a key > > *

This installs packages with QEMU Guest Agent when converting Linux machine. The packages should be available on guest tools ISO. The patches work "as-is" but probably deserve some more attention: - it is "abusing" Winows_virtio code but renaming/refactoring everything to remove "windows" from the name and use "guest tools" seems like a lot of unnecesary

To add this support, the existing code searches for either the viostor or the VMDP (Virtual Machine Driver Pack) files and updates the registry accordingly. Note that VMDP's block driver pvvxblk depends on the ballooning driver pvvxbn. --- v2v/convert_windows.ml | 59 ++++++++++++++++++++------ v2v/windows_virtio.ml | 113 +++++++++++++++++++++++++++++++++++++++++-------- 2 files changed,

Hi there! I know it's been a while since I posted my first version of some patches. But here I have rebased them on top of Roman's work and added a few other ones. Cédric Bosdonnat (11): v2v: also search for windows virtio drivers in symlinks Update packagelist for SLES customize: fix windows firstboot script customize: change windows firstboot path customize: add support for

This patch is just for discussion. There are still a couple of issues that I'm trying to fix. One is that all of the test guests I have, even ones with static IPs, have multiple interfaces, some using DHCP, so the conditions for adding the Powershell script don't kick in. This makes testing very awkward. However a bigger issue is that I think the premise is wrong. In some registries

Hi all, v2 includes all comments from Pino and Richard. I also removed the mkdir /run/lvm in the init since it only failed in one case and couldn't reproduce it anymore. Cédric Bosdonnat (11): v2v: also search for windows virtio drivers in symlinks Update packagelist for SLES and openSUSE customize: fix windows firstboot script customize: change windows firstboot path customize:

v1 was here with much discussion: https://www.redhat.com/archives/libguestfs/2018-December/msg00048.html v2: - Fix the case where there are multiple interfaces. Note this does not preserve order correctly (see patch for comment on why that is a hard problem). - Preserve name servers. This patch is still for discussion only. I'd like to see what might be done to get this upstream

On Tue, Feb 09, 2016 at 12:20:35AM -0800, Joshua Pincus wrote: > Hi Rich, > > Thanks for the pointer. It looks like add_viostor_to_driver_database() is > the routine that I need to mimic. I'll try that out in the morning. I've > been staring at this problem for so long that my eyes are crossing. I should note that any doubled \\ that appear in the strings are really

Parallels proprietary hypervisor uses RDPMC as the hypercall instruction. As this instruction is supported since early P6 family, the drivers didn't even bother to check for the presence of the corresponding feature in CPUID. In QEMU/KVM, however, this instruction triggers #GP unless the VM is run with PMU (performance monitoring unit) enabled, which is often not the case (due to its impact

On Mon, Feb 08, 2016 at 04:50:09PM -0800, Joshua Pincus wrote: > Mr Jones, > > I recently came across your excellent 2010 blog entry, > https://rwmj.wordpress.com/2010/04/30/tip-install-a-device-driver-in-a-windows-vm/. > The blog entry was incredibly useful for Win7. I was wondering if you have > any similar insights for Win10? The CDDB is gone, and it's replacement, or

On Wed, Aug 17, 2016 at 08:59:43PM +0300, Roman Kagan wrote: > It will be used in new code in a followup patch. > > Signed-off-by: Roman Kagan <rkagan@virtuozzo.com> > --- > v2v/convert_windows.ml | 17 +++++++++-------- > 1 file changed, 9 insertions(+), 8 deletions(-) > > diff --git a/v2v/convert_windows.ml b/v2v/convert_windows.ml > index 55bb3ef..751a6ab

If a Windows boot fails and the user boots the last known good configuration, ControlSet001 will be marked as failed and no longer used. However, virt-v2v would only install viostor to ControlSet001, meaning it will fail to boot this guest after conversion. This patch looks up the current controlset and always installs registry keys to the correct one. Fixes RHBZ#644254 ---

Wait for driver installations fired by the PnP manager to complete before running firstboot scripts. The first patch is a minor refactoring to pave the way for the second patch. The latter contains the bulk of the changes as well as the description of the idea. Roman Kagan (2): v2v:windows: factor out getting CurrentControlSet v2v:windows: prevent conflicts with PnP on firstboot --- v1