Richard W.M. Jones [Tue, 13 Aug 2024 12:23:06 +0000 (13:23 +0100)]
virt-what-cvm.pod: Fix man page typo s390x-protvirt -> s390-protvirt
Reported-by: Yongkui Guo
Fixes: commit
037689fbe95e403b050c1eb736ebc8fdc2e601a5
(cherry picked from commit
963676c4dd4c2a9c070b76da6f8835ceb131dbe0)
Daniel P. Berrangé [Fri, 2 Aug 2024 15:07:46 +0000 (16:07 +0100)]
Add support for detecting protected virtualization on s390x
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
(cherry picked from commit
037689fbe95e403b050c1eb736ebc8fdc2e601a5)
Daniel P. Berrangé [Tue, 30 Jul 2024 09:46:46 +0000 (10:46 +0100)]
Fix CVM detection on Azure with TDX
The current TDX support was tested on Azure, however, since that time
they now block the CPUID leaf we were using. Instead it is required to
issue the Azure specific CPUID calls as we were already doing for SNP.
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
(cherry picked from commit
059cbff66740ef74cd663f88c5f96a80a8d6d6ea)
Richard W.M. Jones [Fri, 26 Jul 2024 15:05:12 +0000 (16:05 +0100)]
Add virt-what-cvm.pod to EXTRA_DIST
(Single hunk cherry pick from commit
6d1455766b)
Daniel P. Berrangé [Thu, 29 Jun 2023 16:51:06 +0000 (17:51 +0100)]
virt-what-cvm: rename 'azure-hcl' fact to 'hyperv-hcl'
Azure is a cloud service that uses the HyperV platform, so we
should refer to the fact as 'hyperv-hcl', not 'azure-hcl'.
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
(cherry picked from commit
94773022f76f994d7a9b37f59ba978bd28f30d1d)
Daniel P. Berrangé [Thu, 29 Jun 2023 16:51:05 +0000 (17:51 +0100)]
virt-what-cvm: drop TPM logic for detecting SNP on HyperV/Azure
Now we have proper CPUID detection, we no longer need the TPM
hacks.
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
(cherry picked from commit
82c0e9c469953a36f18db1e329629cecd950134a)
Daniel P. Berrangé [Thu, 29 Jun 2023 16:51:04 +0000 (17:51 +0100)]
virt-what-cvm: probe for SNP/HCL on HyperV/Azure via CPUID
When running a confidential VM on Azure (HyperV) we can probe
CPUID leaf 0x40000003 to detect if VM isolation is present,
and 0x4000000c to detect what kind of isolation is used.
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
(cherry picked from commit
bb0055b491501e16fca3ab61dc7a969effbf48f3)
Daniel P. Berrangé [Thu, 29 Jun 2023 16:51:03 +0000 (17:51 +0100)]
virt-what-cvm: support alternative cpuid leaf ordering
The HyperV CPUID leaf for reporting the vendor string has an
alternative ordering of ecx/edx.
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
(cherry picked from commit
15d3e4a92fd9c1490fb6f86b7ab3a2dff8364837)
Daniel P. Berrangé [Thu, 29 Jun 2023 16:51:02 +0000 (17:51 +0100)]
virt-what-cvm: check if hypervisor bit is set
Before doing any probes for a confidential VM, check that the
tool is running under a hypervisor, rather than bare metal
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
(cherry picked from commit
64718d09a504bd10c6ab973acc0779925482b62f)
Richard W.M. Jones [Tue, 30 May 2023 07:46:06 +0000 (08:46 +0100)]
docs: Add cross reference to virt-what-cvm(1) to virt-what(1)
(cherry picked from commit
52c833c7c6ede0b7fcefa7ad225206f410407eda)
Daniel P. Berrangé [Fri, 26 May 2023 11:39:03 +0000 (12:39 +0100)]
Introduce 'virt-what-cvm' program
The 'virt-what' program prints facts that reflect the hypervisor that
the guest is running under.
The new complementary 'virt-what-cvm' program prints facts that reflect
the confidential virtualization technology the guest is running under,
if any.
It is kept as a separate tool, rather than incorporating the facts into
'virt-what' output because it is considering a different aspect of the
virtualization. Furthermore there are specific security concerns around
the usage of facts reported by 'virt-what-cvm'.
The tool has been tested in a number of environments
* Azure confidential guest with AMD SEV-SNP (GA)
* Azure confidential guest with Intel TDX (technology preview)
* Fedora 37 QEMU/KVM guest with AMD SEV (GA)
* Fedora 37 QEMU/KVM guest with AMD SEV-ES (GA)
* Fedora 38 QEMU/KVM guest with AMD SEV-SNP + SVSM (devel snapshot)
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
(cherry picked from commit
22e33361e980ddefe08e2c68bf145943af8375f9)
Richard W.M. Jones [Tue, 2 Jul 2024 08:45:10 +0000 (08:45 +0000)]
Fix support for Hyper-V on Arm
This architecture lacks CPUID so we must fall back to looking
at DMI data instead.
Reported-by: Yuxin Sun
Thanks: Dan Berrange
Fixes: https://issues.redhat.com/browse/RHEL-45834
(cherry picked from commit
49a17ec8438d01280750d52ae661c8c5c9fe9d07)
Richard W.M. Jones [Wed, 28 Jun 2023 09:00:42 +0000 (10:00 +0100)]
Add support for CRI-O containers
Reported-by: Alessandro Taufer
Fixes: https://bugzilla.redhat.com/show_bug.cgi?id=2217407
Thanks: Xiaodai Wang
(cherry picked from commit
95d18c02038f44bc3e8ff264079f8b697eecf1bd)
Richard W.M. Jones [Wed, 7 Jun 2023 09:10:03 +0000 (10:10 +0100)]
nutanix: Don't match Nutanix-based baremetal systems
Nutanix ship baremetal systems which have modified firmware that
reports similar to:
Manufacturer: Nutanix
Product Name: NX-T00-4NL3-G5
This is not a virtual machine, but was recognised as such. Make sure
we match on 'Product Name: AHV' to identify only virtual machines.
Thanks: Scott Yeager, Shaoyu Shang
Fixes: https://bugzilla.redhat.com/show_bug.cgi?id=2211980
(cherry picked from commit
eeb7d8a1e7c73fc9c5b4e987d92a04136d61b07d)
Yongkui Guo [Mon, 30 Jan 2023 10:20:32 +0000 (18:20 +0800)]
Add support for Alibaba cloud on aarch64
This assumes you are running RHEL or another tenant which supports
ACPI. It probably won't work for guests using device tree.
Also we assume that the hypervisor is KVM. Since the hypervisor
indicates that it is RHEL 7, this is very likely to be true, but hard
to prove since there is no CPUID on Arm.
Fixes: https://bugzilla.redhat.com/show_bug.cgi?id=2165285
(cherry picked from commit
805604762468b754ea60e5dc176fa26f68d211b8)
Richard W.M. Jones [Thu, 5 Jan 2023 12:34:02 +0000 (12:34 +0000)]
Detect OCI containers
Podman using OCI containers sets container=oci in PID 1's environment.
Detect that and print "oci" fact.
This patch was originally contributed by Alessandro Valentini. I
modified it and added tests and documentation.
Reported-by: Alessandro Valentini
Fixes: https://bugzilla.redhat.com/show_bug.cgi?id=2155233
Richard W.M. Jones [Thu, 5 Jan 2023 12:48:55 +0000 (12:48 +0000)]
podman: Fix location of test file /proc/1/environ
Since it was at the wrong location, the file was not even used by the
test.
Fixes: commit
1df728aa4b ("Add podman support")
Richard W.M. Jones [Thu, 22 Dec 2022 10:34:03 +0000 (10:34 +0000)]
Move docker and podman tests up, add comments
Refactor to bring related tests closer together and make the comments
more uniform.
Richard W.M. Jones [Thu, 22 Dec 2022 10:31:47 +0000 (10:31 +0000)]
Rearrange lxc test to avoid use of cat
Thanks: Elias Probst <e.probst@ssc-services.de>
Richard W.M. Jones [Wed, 17 Aug 2022 09:57:19 +0000 (10:57 +0100)]
Version 1.25.
Xiao Liang [Tue, 24 May 2022 09:34:52 +0000 (17:34 +0800)]
detect AWS arm virtual instance as kvm
AWS arm virtual instance is KVM based hypervisor.
With this patch:
$ sudo virt-what
kvm
aws
$ cat /sys/devices/virtual/dmi/id/product_name
t4g.small
Without it:
$ sudo virt-what
aws
Signed-off-by: Xiao Liang <xiliang@redhat.com>
Richard W.M. Jones [Mon, 8 Aug 2022 16:13:11 +0000 (12:13 -0400)]
google-cloud: Print "kvm" fact on aarch64 instances
GCE instances are KVM with a non-QEMU userspace. On x86-64 we should
recognize this from the CPUID information, but aarch64 has no
equivalent to this. /proc/device-tree/hypervisor/compatible contains
something but RHEL does not use device tree. However the SMBIOS
tables exported by GCE do contain:
Manufacturer: KVM
so we can use that.
# uname -m
aarch64
# virt-what
google_cloud
kvm
This commit also updates the test.
Fixes: https://bugzilla.redhat.com/show_bug.cgi?id=2114997
Updates: commit
54b726fa7c9687b692162488e8934a426f9fa6c3
Thanks: Eric Auger, Yongkui Guo
Richard W.M. Jones [Wed, 3 Aug 2022 17:53:19 +0000 (17:53 +0000)]
tests: Add test for Google Cloud Engine
This was actually a GCE aarch64 instance, but I believe the
architecture makes no difference as all we're looking for is the
characteristic string in dmidecode output.
Thanks: Craig Donnelly
Fixes: https://bugzilla.redhat.com/show_bug.cgi?id=2114997
Updates: commit
8363fa9f0e2cbd970f34ffb7c7b23ed0ac2c374a
Richard W.M. Jones [Wed, 3 Aug 2022 17:37:23 +0000 (17:37 +0000)]
alibaba-cloud: Make the test more robust
This test worked by trying to make an HTTP connection to a path under
100.100.100.200, which if successful would indicate that it's an
Alibaba cloud instance. We only try the connection if SM-BIOS
information indicates the hardware OEM is Alibaba.
However 'make check' would hang if 100.100.100.200 doesn't immediately
reject the connection. I found that GCE is such an environment where
connecting to this address would hang rather than reject. GCE does
not have Alibaba hardware of course, but when doing 'make check' we
fake dmidecode data.
Make the test more robust by only trying to connect when we know we're
not in 'make check'.
Fixes: commit
b7cc3d93a613ef6d0ac5ccd6e32cc3d66e057243
Richard W.M. Jones [Thu, 23 Jun 2022 13:42:02 +0000 (14:42 +0100)]
Version 1.24.
Richard W.M. Jones [Thu, 23 Jun 2022 13:38:55 +0000 (14:38 +0100)]
illumos-lx: Fix test so it works under virtualization
Tests need to provide dummy sbin/virt-what-cpuid-helper, dmidecode and
uname. This ensures that if we are running the test suite under (eg)
KVM, that virt-what doesn't discover this.
The failure was:
./test.sh: test failed because output did not match expected
Expected output was:
----------------------------------------
illumos-lx
----------------------------------------
But the actual output of the program was:
----------------------------------------
illumos-lx
kvm
----------------------------------------
Fixes: commit
763cf1b1123136751b71f475610c5ac1eeaba0ad
Richard W.M. Jones [Thu, 23 Jun 2022 12:43:20 +0000 (13:43 +0100)]
Version 1.23.
Steve Mokris [Thu, 12 May 2022 16:02:46 +0000 (12:02 -0400)]
Detect Illumos (SmartOS/OmniOS) LX virtualization
Thanks for the feedback, Eric and Richard. In this version, I've expunged the redundant cats and included the Makefile.am.
Richard W.M. Jones [Thu, 14 Apr 2022 12:51:49 +0000 (13:51 +0100)]
Version 1.22.
Richard W.M. Jones [Thu, 14 Apr 2022 12:51:13 +0000 (13:51 +0100)]
Add test file to EXTRA_DIST
Fixes: commit
1df728aa4b1d2814265f9c86494f7d55ee0cf9af
Richard W.M. Jones [Thu, 14 Apr 2022 12:50:46 +0000 (13:50 +0100)]
Add test file to EXTRA_DIST
Fixes: commit
7e679278431aa1a188dfe423526dfea82fddcbd3
Richard W.M. Jones [Thu, 22 Jul 2021 07:58:16 +0000 (08:58 +0100)]
Display kvm/qemu fact for Nutanix AHV
Since these are based on KVM we should display those facts too.
Thanks: Paolo Bonzini
Richard W.M. Jones [Mon, 19 Jul 2021 17:05:14 +0000 (18:05 +0100)]
Add detection of Red Hat's hypervisor
Red Hat sets the following fields in SMBIOS information, allowing us
to detect that it's the branded Red Hat hypervisor.
Manufacturer: Red Hat
Product Name: KVM
Version: RHEL 7.6.0 PC (i440FX + PIIX, 1996)
Note that "RHEL 7.6.0" there is the machine type, which is not the
same as the hypervisor version (although the two have a relationship).
Thanks: Timothy Smith at IBM, Karen Noel
Richard W.M. Jones [Mon, 19 Jul 2021 14:14:32 +0000 (15:14 +0100)]
Add Google Cloud (not tested)
Thanks: Paolo Bonzini
Weisson [Fri, 2 Jul 2021 09:06:07 +0000 (17:06 +0800)]
use official product name for alibaba cloud bare-metal instances
Weisson [Tue, 1 Jun 2021 12:49:17 +0000 (20:49 +0800)]
Check for Alibaba Cloud X-Dragon Architecture
Thanks to Richard W.M. Jones' Super Powerful Bash Script
Weisson [Thu, 27 May 2021 09:55:57 +0000 (10:55 +0100)]
Add test for Alibaba Cloud Linux
Weisson [Thu, 27 May 2021 09:53:08 +0000 (10:53 +0100)]
Add support for Alibaba Cloud Linux
Richard W.M. Jones [Wed, 26 May 2021 10:31:24 +0000 (11:31 +0100)]
aws: Match multi-digit version number in BIOS information
See: https://listman.redhat.com/archives/virt-tools-list/2021-May/msg00011.html
Reported-by: Stuart Cooper
Razvan Corneliu [Mon, 10 May 2021 20:02:27 +0000 (21:02 +0100)]
Add support for VMware ESXi on ARM
Virt-what doesn't recognize VMware VMs for AARCH64/ARM (with licensing
implications in ESXi Fling for ARM and Fusion v.Next for ARM). I've
created and attached a patch that adds both support for recognizing
ESXi and the automated tests for it.
Richard W.M. Jones [Mon, 19 Apr 2021 11:56:58 +0000 (12:56 +0100)]
Version 1.21.
Richard W.M. Jones [Tue, 13 Apr 2021 08:35:07 +0000 (09:35 +0100)]
Simplify and fix invocation of cpuid.
Fixes a crash on some platforms identified by Yongkui Guo in
https://bugzilla.redhat.com/show_bug.cgi?id=1756381#c15
Jordan Webb [Mon, 13 Apr 2020 21:41:30 +0000 (21:41 +0000)]
Add podman support
Ville Skyttä [Sun, 1 Dec 2019 06:21:32 +0000 (08:21 +0200)]
docker: Lookup from /proc/self/cgroup too
Refs https://github.com/moby/moby/issues/18355
Ville Skyttä [Sun, 1 Dec 2019 05:58:04 +0000 (07:58 +0200)]
docker: Check for /.dockerenv too
/.dockerinit may no longer exist.
Ref https://github.com/moby/moby/issues/18355
willem van de velde [Tue, 22 Oct 2019 10:58:01 +0000 (11:58 +0100)]
Fix incorrect detection of MS Surfacebook 2 as a virtual machine.
Richard W.M. Jones [Fri, 4 Oct 2019 17:30:01 +0000 (18:30 +0100)]
helper: Fix KVM signature.
Thanks: Paolo Bonzini.
Richard W.M. Jones [Fri, 4 Oct 2019 13:25:55 +0000 (16:25 +0300)]
Add support for Nutanix Acropolis Hypervisor (AHV) (RHBZ#1756381).
Thanks: Cristian Seres for providing access to a guest.
Richard W.M. Jones [Fri, 4 Oct 2019 12:57:42 +0000 (15:57 +0300)]
Fix virt-what-cpuid-helper.
The value returned in %eax is the max_entry (eg. 0x40000000
if there are no further leafs). However it is not reliable.
In addition if there are multiple leafs we should probably
only print the highest one.
Also use uint32_t instead of unsigned int.
Thanks: Paolo Bonzini.
Richard W.M. Jones [Thu, 3 Oct 2019 09:21:12 +0000 (10:21 +0100)]
Version 1.20.
Richard W.M. Jones [Thu, 3 Oct 2019 09:13:43 +0000 (10:13 +0100)]
maintainer: Add rules for maintainers.
Adds:
make maintainer-check-extra-dist
make maintainer-commit
make maintainer-tag
Richard W.M. Jones [Thu, 3 Oct 2019 09:16:08 +0000 (10:16 +0100)]
build: Add .gitignore to tarball.
Mainly to keep maintainer-check-extra-dist happy.
Richard W.M. Jones [Thu, 3 Oct 2019 08:53:49 +0000 (09:53 +0100)]
Arrange tests in subdirectories.
This simplifies and isolates each test.
Richard W.M. Jones [Thu, 3 Oct 2019 08:08:19 +0000 (09:08 +0100)]
build: Remove *~ files on make clean.
Richard W.M. Jones [Wed, 31 Oct 2018 15:23:45 +0000 (15:23 +0000)]
aws: Add regression test for AWS on KVM on ARM architecture.
Richard W.M. Jones [Wed, 31 Oct 2018 19:04:22 +0000 (19:04 +0000)]
Version 1.19.
Richard W.M. Jones [Wed, 31 Oct 2018 15:23:45 +0000 (15:23 +0000)]
aws: Add regression test for AWS on Xen on x86_64 architecture.
Richard W.M. Jones [Wed, 31 Oct 2018 15:43:20 +0000 (15:43 +0000)]
aws: Add regression test for AWS on baremetal on x86_64 architecture.
Richard W.M. Jones [Wed, 31 Oct 2018 15:23:45 +0000 (15:23 +0000)]
aws: Add regression test for AWS on KVM on x86_64 architecture.
Richard W.M. Jones [Wed, 31 Oct 2018 15:35:34 +0000 (15:35 +0000)]
tests: Fix tests when run on AWS.
When running on AWS two of the tests failed. This was because the
tests did _not_ include a working dmidecode binary, hence the platform
dmidecode binary was being run instead, and that was detecting as AWS
owing to the previous commit. Fix this by supplying a working
dmidecode binary for those tests.
Richard W.M. Jones [Wed, 31 Oct 2018 15:04:24 +0000 (15:04 +0000)]
aws: Detect AWS from dmidecode information.
Move the AWS detection out from under Xen. AWS runs on KVM and
baremetal these days. The dmidecode information for KVM and baremetal
is a little bit different so we also need to adjust the test for that
case.
Typical SMBIOS info for Xen:
BIOS Information
Vendor: Xen
Version: 4.2.amazon
Typical SMBIOS info for KVM and baremetal:
BIOS Information
Vendor: Amazon EC2
Version: 1.0
Thanks: Vitaly Kuznetsov, Jon Masters.
Richard W.M. Jones [Tue, 28 Aug 2018 17:32:41 +0000 (18:32 +0100)]
Replace == with = since the former is a bash-ism.
Thanks: Eric Blake.
Fixes commit
d7fd8a7843030d2b1719353edfcd49dba3000122.
Jasper Lievisse Adriaanse [Wed, 25 Apr 2018 14:41:32 +0000 (16:41 +0200)]
Allow using sysctl, for example when /proc isn't available
Guido Günther [Sat, 27 Jan 2018 12:11:36 +0000 (13:11 +0100)]
Determine architecture via 'uname -m'
'uname -p' only gives unknown on x86_64, i386, arm6l (rpi) and aarch64
(scaleways).
Richard W.M. Jones [Tue, 17 Oct 2017 13:47:09 +0000 (14:47 +0100)]
Recognize ppc64le (little endian) virtualization (RHBZ#1147876).
Tested by Laurent Vivier, see:
https://bugzilla.redhat.com/show_bug.cgi?id=1147876#c35 & ff.
xiliang [Sat, 17 Jun 2017 16:33:28 +0000 (00:33 +0800)]
As xen pv guest can access cpuid from Intel CPUs started IvyBridge onwards have CPUID Faulting, added one more check in virt-what.
Guido Günther [Fri, 4 Aug 2017 15:02:08 +0000 (12:02 -0300)]
Remove bashisms
Use [ instead of [[ so we fall back to test if necessary:
http://pubs.opengroup.org/onlinepubs/
9699919799/utilities/test.html
Gbp-Pq: Name Remove-bashisms.patch
Jasper Lievisse Adriaanse [Thu, 10 Aug 2017 07:44:01 +0000 (08:44 +0100)]
Missing have_cpuinfo check.
Richard W.M. Jones [Mon, 31 Jul 2017 14:35:11 +0000 (15:35 +0100)]
Version 1.18.
Richard W.M. Jones [Mon, 31 Jul 2017 14:34:08 +0000 (15:34 +0100)]
Richard W.M. Jones [Mon, 31 Jul 2017 14:28:45 +0000 (15:28 +0100)]
Version 1.17.
Richard W.M. Jones [Mon, 31 Jul 2017 14:19:52 +0000 (15:19 +0100)]
ppc64: Test ${root}/proc/device-tree/hypervisor instead of real file.
Fixes commit
50153230d19401fe47fcfd19d7ca0a2623c1ff09.
Richard W.M. Jones [Mon, 31 Jul 2017 14:04:28 +0000 (15:04 +0100)]
tests: Fix ldoms test so it doesn't test the host.
Fixes commit
dc0c2972e178a6de9ee87501acca54447ea9c699.
Richard W.M. Jones [Mon, 31 Jul 2017 13:32:51 +0000 (14:32 +0100)]
Document ‘aws’ meaning Amazon Web Services cloud guest.
Updates commit
dc32228184622b28f412f527c0f353347db029b8.
Richard W.M. Jones [Mon, 31 Jul 2017 13:19:46 +0000 (14:19 +0100)]
Fix EXTRA_DIST to match actual contents of tests directory.
Richard W.M. Jones [Mon, 31 Jul 2017 13:14:06 +0000 (14:14 +0100)]
Version 1.16
Darren Kenny [Tue, 4 Jul 2017 07:26:01 +0000 (08:26 +0100)]
Add LDoms Support.
The current version of virt-what does not support discovery of any
Logical Domains (LDoms) variants. This patch adds that support.
Orabug:
26106813
Signed-off-by: Darren Kenny <darren.kenny@oracle.com>
Reviewed-By: Liam Merwick <liam.merwick@oracle.com>
Reviewed-By: Rashmi Narasimhan <rashmi.narasimhan@oracle.com>
Qi Guo [Flos] [Wed, 19 Jul 2017 15:39:29 +0000 (16:39 +0100)]
Update virt-what to detect AWS via dmidecode data.
Jasper Lievisse Adriaanse [Sat, 1 Jul 2017 11:12:46 +0000 (13:12 +0200)]
virt-what: recognize OpenBSD vmm
Hi,
Please find attached a patch for virt-what to recognize the OpenBSD VMM
hypervisor.
For more information on vmm, please refer to:
- http://man.openbsd.org/vmm.4
- https://www.openbsd.org/papers/asiabsdcon2017-vmm-slides.pdf
>From
837390b5afbb9253a62b15208d6fb4cbbbecabf1 Mon Sep 17 00:00:00 2001
From: Jasper Lievisse Adriaanse <j@jasper.la>
Date: Sat, 17 Jun 2017 21:17:39 +0200
Subject: [PATCH] vmm detection (OpenBSD hypervisor)
Daniel P. Berrange [Tue, 18 Jul 2017 12:45:09 +0000 (13:45 +0100)]
Update to detect QEMU via 'TCGTCGTCGTCG' CPUID
QEMU >= 2.10 will start reporting the string "TCGTCGTCGTCG"
for CPUID queries when running under TCG only (KVM signature
is unchanged)
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Thomas Huth [Tue, 4 Jul 2017 14:20:40 +0000 (16:20 +0200)]
Add support for KVM on System Z
KVM on System Z can be discovered in the same way as z/VM, by
looking at the "VM* Control Program" lines in /proc/sysinfo.
Signed-off-by: Thomas Huth <thuth@redhat.com>
Mark Huth [Wed, 5 Jul 2017 11:11:32 +0000 (21:11 +1000)]
Export internal PATH variable (RHBZ#1467878).
- Fixes virt-what failing if it's started without an environment, eg:
$ env -i /sbin/virt-what
... or with python subprocess.Popen, eg:
p = subprocess.Popen('/sbin/virt-what', shell=False, ... , env={})
Thomas Huth [Thu, 29 Jun 2017 10:27:42 +0000 (11:27 +0100)]
Detect KVM property when running on POWER (RHBZ#1455676).
QEMU on ppc64 adds an additional property in the device tree if
running with KVM, so you could check that property for the string
"linux,kvm" to see if KVM is enabled:
0000000: 6c69 6e75 782c 6b76 6d00 linux,kvm.
Richard W.M. Jones [Thu, 29 Jun 2017 10:22:21 +0000 (11:22 +0100)]
Revert "Print 'qemu' fact when running in qemu environment on POWER 9 (RHBZ#1455676)."
This reverts commit
15314ad0a6e3764f070b60f6df02a2f881b09fce.
Richard W.M. Jones [Thu, 29 Jun 2017 09:07:38 +0000 (10:07 +0100)]
Print 'qemu' fact when running in qemu environment on POWER 9 (RHBZ#1455676).
Thanks Nageswara R. Sastry
Richard W.M. Jones [Mon, 10 Oct 2016 14:15:41 +0000 (15:15 +0100)]
Add documentation for bhyve.
Leonardo Brondani Schenkel [Sat, 17 Sep 2016 12:59:52 +0000 (12:59 +0000)]
Add detection of bhyve (FreeBSD hypervisor)
Richard W.M. Jones [Mon, 22 Aug 2016 11:30:25 +0000 (12:30 +0100)]
Add detection of Red Hat Enterprise Virtualization hypervisor (RHBZ#1249438).
Thanks: Xianghua Chen
Richard W.M. Jones [Mon, 20 Jun 2016 12:14:53 +0000 (13:14 +0100)]
Update the previous commit to use <system>-<virt> standard for facts.
Also I renamed ppc64 to power, since that is the official IBM name.
Adrian Likins [Mon, 20 Jun 2016 11:57:27 +0000 (12:57 +0100)]
Add support for detecting ppc64 LPAR as virt guests.
https://bugzilla.redhat.com/show_bug.cgi?id=1072524#c13
Andrew Jones [Tue, 27 Oct 2015 17:16:04 +0000 (18:16 +0100)]
qemu/kvm: dmidecode: look for 'KVM'
We can check the product name for 'KVM' to determine if we're
accelerated. This only works for arm/aarch64 targets boot with
AAVMF and on RHEL x86 targets (which already have KVM in the cpuid
leaf), but it doesn't hurt to try on all.
Andrew Jones [Tue, 27 Oct 2015 17:16:03 +0000 (18:16 +0100)]
qemu/kvm: try dmidecode on all targets
Looking at "$dmi" on targets that don't support dmidecode is
harmless. Furthermore it works, not only for arm targets, but
also pc, so move it out of the 'if arm' block.
Andrew Jones [Tue, 27 Oct 2015 17:16:02 +0000 (18:16 +0100)]
trivial: virt-what.in doesn't use tabs
Richard W.M. Jones [Mon, 3 Aug 2015 11:23:13 +0000 (12:23 +0100)]
Add oVirt (RHBZ#1249438).
Andrew Jones [Tue, 22 Sep 2015 16:29:08 +0000 (18:29 +0200)]
trivial: comment fixup
dmidecode does now exist on non-PC architectures.
Andrew Jones [Tue, 22 Sep 2015 13:54:55 +0000 (15:54 +0200)]
Add QEMU/KVM detection for ACPI-boot ARM
Detect QEMU/KVM on ARM guests that have been boot with ACPI, i.e.
without device tree. When boot with ACPI SMBIOS is populated, so we
can use dmidecode.
Signed-off-by: Andrew Jones <drjones@redhat.com>
Richard W.M. Jones [Mon, 3 Aug 2015 11:23:05 +0000 (12:23 +0100)]
Update copyright years.
Richard W.M. Jones [Tue, 21 Apr 2015 11:10:44 +0000 (12:10 +0100)]
Version 1.15
Richard W.M. Jones [Tue, 21 Apr 2015 11:12:45 +0000 (12:12 +0100)]
xen-arm: Fix path in EXTRA_DIST.
Andrew Jones [Fri, 27 Mar 2015 20:19:00 +0000 (21:19 +0100)]
Add ARM support
Currently only detects LKVM, QEMU/KVM, and Xen. Also only works if
the guest is booted with DT (which is the only way that currently
exists).
Signed-off-by: Andrew Jones <drjones@redhat.com>