| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
| |
Signed-off-by: Michal Privoznik <michal.privoznik@gmail.com>
Signed-off-by: Sam James <sam@gentoo.org>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
In its upstream commit of v9.8.0-rc1~27 [1] libvirt introduced an
optional dependency on libnbd. Reflect this in corresponding
ebuilds. Now, pre-10.0.0 ebuilds do not have nbd USE flag. Thus,
it needs to be introduced to record the dependency properly.
NB, initially, the configure option was named 'libnbd' and only
after v9.9.0-rc1~108 [2] it was changed to 'nbdkit'. Hence the
discrepancy in libvirt-9.8.0-r3 and libvirt-9.9.0-r3 ebuilds.
1: https://gitlab.com/libvirt/libvirt/-/commit/40935b395c91a95a71cd66b9e2872346ac3a2145
2: https://gitlab.com/libvirt/libvirt/-/commit/7cbd8c42305735375b60abf8abc47838a7a050d4
Closes: https://bugs.gentoo.org/938346
Signed-off-by: Michal Privoznik <michal.privoznik@gmail.com>
Closes: https://github.com/gentoo/gentoo/pull/38274
Signed-off-by: Sam James <sam@gentoo.org>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Due to portage design whereby commands which fail aren't considered
failures unless you explicitly use `|| die`, a common footgun in bash
scripting propagates throughout the portage ecosystem:
the use of `cmd1 && cmd2` for conditional logic.
This python_check_deps function did such, and then handled the case
where `use test` was false by unconditionally ignoring the result of the
previous line by returning 0. Hence, python_check_deps could never
decide that an impl was unable to be used. As a result, if python 3.13
and 3.12 were both installed, but $(python_gen_any_dep ...) discovered
pytest installed solely for 3.12, portage would not reinstall pytest for
3.13 support whereas the eclass would select 3.13 as the preferred
(latest) python impl.
Fix this by correctly using bash, shunning `cmd1 && cmd2`, and instead
using `if cmd1; then cmd2; fi`, which returns correct return values
based on the return value of both cmd1 and cmd2, without requiring
hardcoded `return 0`s of any variety, unconditional or otherwise.
Fixes: bba723505f488b52bd593869b5b9a0df096ffbb4
Bug: https://bugs.gentoo.org/932652
Closes: https://bugs.gentoo.org/935849
Signed-off-by: Eli Schwartz <eschwartz@gentoo.org>
|
|
|
|
|
|
|
|
|
|
|
| |
Libvirt's use of Python doesn't really use new python features
since it still aims to build on stable distros like RHEL-8 where
only Python-3.6 is available. Thus, Python-3.13 is trivially
supported. Add it onto PYTHON_COMPAT list for all ebuilds.
Signed-off-by: Michal Privoznik <michal.privoznik@gmail.com>
Closes: https://github.com/gentoo/gentoo/pull/37497
Signed-off-by: Eli Schwartz <eschwartz@gentoo.org>
|
|
|
|
|
| |
Signed-off-by: Michal Privoznik <michal.privoznik@gmail.com>
Signed-off-by: Sam James <sam@gentoo.org>
|
|
|
|
|
|
|
|
|
|
|
| |
Upstream libvirt learned how to use nftables directly [1]. While
it did so as early as in 10.4.0 release, there were some critical
fixes merged in 10.5.0 release. Therefore, allow just the live
ebuild to pick up nftables as backend.
1: https://gitlab.com/libvirt/libvirt/-/commit/b89c4991daa0ee9371f10937fab3b03c5ffdabc6
Signed-off-by: Michal Privoznik <michal.privoznik@gmail.com>
Signed-off-by: Sam James <sam@gentoo.org>
|
|
|
|
|
|
|
|
|
|
|
| |
When AppArmor is enabled and sys-firmware/edk2-ovmf-bin is
installed then starting a guest under libvirt fails, because
libvirt assumed different paths for UEFI. A fix was merged
upstream so backport it.
Resolves: https://bugs.gentoo.org/911786
Signed-off-by: Michal Privoznik <michal.privoznik@gmail.com>
Signed-off-by: Sam James <sam@gentoo.org>
|
|
|
|
| |
Signed-off-by: Matthias Maier <tamiko@gentoo.org>
|
|
|
|
| |
Signed-off-by: Arthur Zamarin <arthurzam@gentoo.org>
|
|
|
|
| |
Signed-off-by: Sam James <sam@gentoo.org>
|
|
|
|
| |
Signed-off-by: Sam James <sam@gentoo.org>
|
|
|
|
| |
Signed-off-by: Sam James <sam@gentoo.org>
|
|
|
|
|
| |
Signed-off-by: Matoro Mahri <matoro_gentoo@matoro.tk>
Signed-off-by: Ionen Wolkens <ionen@gentoo.org>
|
|
|
|
|
| |
Signed-off-by: Matoro Mahri <matoro_gentoo@matoro.tk>
Signed-off-by: Ionen Wolkens <ionen@gentoo.org>
|
|
|
|
|
| |
Signed-off-by: Matoro Mahri <matoro_gentoo@matoro.tk>
Signed-off-by: Ionen Wolkens <ionen@gentoo.org>
|
|
|
|
| |
Signed-off-by: Arthur Zamarin <arthurzam@gentoo.org>
|
|
|
|
|
|
|
|
|
|
|
|
| |
As of its upstream commit v9.10.0-rc1~114 libvirt introduced its
own RPC generator written in python and also some tests for it.
But these require pytest. Therefore, generate corresponding
dependency if running tests.
Closes: https://bugs.gentoo.org/932652
Signed-off-by: Michal Privoznik <michal.privoznik@gmail.com>
Closes: https://github.com/gentoo/gentoo/pull/36814
Signed-off-by: Sam James <sam@gentoo.org>
|
|
|
|
|
|
|
|
|
|
|
| |
In libvirt-10.3.0 the virshtest was rewritten and unfortunately,
it caused a bug in which it tried to access user's $HOME.
Backport the upstream fix to make the test pass again.
Closes: https://bugs.gentoo.org/931109
Signed-off-by: Michal Privoznik <michal.privoznik@gmail.com>
Closes: https://github.com/gentoo/gentoo/pull/36584
Signed-off-by: Joonas Niilola <juippis@gentoo.org>
|
|
|
|
|
| |
Signed-off-by: Michal Privoznik <michal.privoznik@gmail.com>
Signed-off-by: Sam James <sam@gentoo.org>
|
|
|
|
|
|
|
|
|
|
|
|
| |
The fix made it into app-emulation/libvirt-10.2.0 release.
Backport the fix into anything older.
https://nvd.nist.gov/vuln/detail/CVE-2024-2494
Bug: https://bugs.gentoo.org/929966
Signed-off-by: Michal Privoznik <michal.privoznik@gmail.com>
Closes: https://github.com/gentoo/gentoo/pull/36242
Signed-off-by: Sam James <sam@gentoo.org>
|
|
|
|
|
|
|
|
|
|
|
| |
The fix made it into app-emulation/libvirt-10.1.0 release.
Backport the fix into anything older.
https://nvd.nist.gov/vuln/detail/CVE-2024-1441
Bug: https://bugs.gentoo.org/929965
Signed-off-by: Michal Privoznik <michal.privoznik@gmail.com>
Signed-off-by: Sam James <sam@gentoo.org>
|
|
|
|
|
| |
Signed-off-by: Michal Privoznik <michal.privoznik@gmail.com>
Signed-off-by: Sam James <sam@gentoo.org>
|
|
|
|
|
|
|
|
|
|
|
| |
Libvirt's use of Python doesn't really use new python features
since it still aims to build on stable distros like RHEL-8 where
only Python-3.6 is available. Thus, Python-3.12 is trivially
supported. Add it onto PYTHON_COMPAT list for all ebuilds.
Closes: https://bugs.gentoo.org/show_bug.cgi?id=929309
Signed-off-by: Michal Privoznik <michal.privoznik@gmail.com>
Signed-off-by: Sam James <sam@gentoo.org>
|
|
|
|
|
| |
Signed-off-by: Michael Mair-Keimberger <mmk@levelnine.at>
Signed-off-by: Conrad Kostecki <conikost@gentoo.org>
|
|
|
|
|
| |
Signed-off-by: Michal Privoznik <michal.privoznik@gmail.com>
Signed-off-by: Sam James <sam@gentoo.org>
|
|
|
|
| |
Signed-off-by: Sam James <sam@gentoo.org>
|
|
|
|
|
| |
Bug: https://bugs.gentoo.org/925459
Signed-off-by: Sam James <sam@gentoo.org>
|
|
|
|
|
| |
Signed-off-by: Matoro Mahri <matoro_gentoo@matoro.tk>
Signed-off-by: Ionen Wolkens <ionen@gentoo.org>
|
|
|
|
|
| |
Signed-off-by: Matoro Mahri <matoro_gentoo@matoro.tk>
Signed-off-by: Ionen Wolkens <ionen@gentoo.org>
|
|
|
|
|
|
|
|
|
|
|
|
| |
While technically app-emulation/virtiofsd is not a libvirt
dependency (app-emulation/qemu and app-emulation/virtiofsd can
function on their own, happily), it's handy (in some scenarios)
to drag it automatically in as a dependency. Invent "virtiofsd"
USE flag that controls this.
Bug: https://bugs.gentoo.org/913212
Signed-off-by: Michal Privoznik <michal.privoznik@gmail.com>
Signed-off-by: Sam James <sam@gentoo.org>
|
|
|
|
| |
Signed-off-by: Michał Górny <mgorny@gentoo.org>
|
|
|
|
|
| |
Signed-off-by: Michal Privoznik <michal.privoznik@gmail.com>
Signed-off-by: Sam James <sam@gentoo.org>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Starting from its upstream commit of v9.9.0-rc1~108 libvirt can
now use nbdkit to access network disks. The commit introduced new
meson option (-Dnbdkit), which is basically just a rename of an
old one (-Dlibnbd). We didn't expose the old one, but it makes
sense to export the new one for proper deps tracking.
But because nbdkit is available only on a small subset of arches,
the useflag has to be masked on other arches that doesn't have
sys-block/nbdkit.
Signed-off-by: Michal Privoznik <michal.privoznik@gmail.com>
Signed-off-by: Sam James <sam@gentoo.org>
|
|
|
|
| |
Signed-off-by: Arthur Zamarin <arthurzam@gentoo.org>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
As of its 2.12.0 release, libxml2 cleaned up header files which
rendered libvirt unable to compile. Backport the fix from
upstream repo.
After this, there are still some warnings about use of a
deprecated function, but those are harmless and we'll get fix
with new release.
Bug: https://bugs.gentoo.org/917516
Signed-off-by: Michal Privoznik <michal.privoznik@gmail.com>
Closes: https://github.com/gentoo/gentoo/pull/33929
Signed-off-by: Sam James <sam@gentoo.org>
|
|
|
|
|
|
|
|
|
|
|
|
| |
The upstream has ditched rpcgen provided by rpcsvc-proto in favor
of its own set of python scripts [1]. Reflect this change in the
live ebuild (there's no release with this change just yet).
1: https://gitlab.com/libvirt/libvirt/-/commit/a62486b95feed2cf17ce4adbe794a1ecff9ef22a
Signed-off-by: Michal Privoznik <michal.privoznik@gmail.com>
Closes: https://github.com/gentoo/gentoo/pull/33674
Signed-off-by: Sam James <sam@gentoo.org>
|
|
|
|
|
| |
Signed-off-by: Michal Privoznik <michal.privoznik@gmail.com>
Signed-off-by: Sam James <sam@gentoo.org>
|
|
|
|
| |
Signed-off-by: Mike Gilbert <floppym@gentoo.org>
|
|
|
|
| |
Signed-off-by: Sam James <sam@gentoo.org>
|
|
|
|
|
|
|
|
|
|
|
| |
A security flaw was identified in <app-emulation/libvirt-9.6.0
which can result int DoS. The upstream is fixed from 9.6.0.
Backport the fix to older versions found in portage.
Bug: https://bugs.gentoo.org/916497
Closes: https://github.com/gentoo/gentoo/pull/33604
Signed-off-by: Michal Privoznik <michal.privoznik@gmail.com>
Signed-off-by: Matthias Maier <tamiko@gentoo.org>
|
|
|
|
|
|
|
|
|
|
| |
The libvirt-9.4.0-do-not-use-sysconfig.patch does not apply
cleanly anymore (because of libvirt's upstream commit of
0f09223fc5a09f438b3e19567a8630672f108cb3). Rebase it and update
the live ebuild.
Signed-off-by: Michal Privoznik <michal.privoznik@gmail.com>
Signed-off-by: Matthias Maier <tamiko@gentoo.org>
|
|
|
|
|
| |
Signed-off-by: Michal Privoznik <michal.privoznik@gmail.com>
Signed-off-by: Matthias Maier <tamiko@gentoo.org>
|
|
|
|
| |
Signed-off-by: Sam James <sam@gentoo.org>
|
|
|
|
|
| |
Signed-off-by: Michal Privoznik <michal.privoznik@gmail.com>
Signed-off-by: Sam James <sam@gentoo.org>
|
|
|
|
| |
Signed-off-by: Arthur Zamarin <arthurzam@gentoo.org>
|
|
|
|
| |
Signed-off-by: Arthur Zamarin <arthurzam@gentoo.org>
|
|
|
|
|
|
|
|
| |
Libvirt 9.5.0 was released earlier this month.
Signed-off-by: Michal Privoznik <michal.privoznik@gmail.com>
Closes: https://github.com/gentoo/gentoo/pull/31836
Signed-off-by: Sam James <sam@gentoo.org>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
In 45065ac9faa6 the libvirt-9.5.0-fix-paths-for-apparmor.patch
was renamed to libvirt-9.4.0-... but the live ebuild still refers
to the 9.5.0. Ideally, the live ebuild would be just made to
apply the renamed patch, but there was some more movement in the
upstream since and thus the patch doesn't apply cleanly anyways.
Therefore, rebase it and name it libvirt-9.6.0-... since the
upstream development was made after 9.5.0 release.
Signed-off-by: Michal Privoznik <michal.privoznik@gmail.com>
Signed-off-by: Sam James <sam@gentoo.org>
|
|
|
|
|
|
|
|
|
| |
The swtpm dependency was added earlier (in 4fa9735e3f44) but only
for app-emulation/libvirt-9.4.0. The live ebuild was missed.
Bug: https://bugs.gentoo.org/776751
Signed-off-by: Michal Privoznik <michal.privoznik@gmail.com>
Signed-off-by: Sam James <sam@gentoo.org>
|
|
|
|
|
|
|
| |
Bug: https://bugs.gentoo.org/908805
Signed-off-by: Kai-Chun Ning <kaichun.ning@gmail.com>
Closes: https://github.com/gentoo/gentoo/pull/31575
Signed-off-by: Sam James <sam@gentoo.org>
|