This is a followup of https://github.com/containers/podman/pull/26048
It fixes `process-release.ps1` that was always looking for the amd64 release
zip file, even if `$env:PODMAN_ARCH` was set to arm64. With this fix it looks
for the right zip file.
It fixes `winmake.ps1` that, when the `-arch` param was not passed, set `$env:PODMAN_ARCH`
to the empty string instead of the local `$env:GOARCH`.
Signed-off-by: Mario Loriedo <mario.loriedo@gmail.com>
When container is being removed, podman iterates
through its exec sessions and checks whether exec
session pid is still alive.
The problem is that the pid can be reused for other processes,
so that it may not belong to exec session.
In this scenario podman may kill another process
This commit prevents it by doing following changes:
- Adds the PIDData string to ExecSession struct. This string
is used to store additional context for a PID to later verify
that the PID killed by the podman is really the one started by
it.
- Adds new package called pidhandle which implements the methods
generating the PIDData, and killing the PID with the PIDData
ensuring the right PID is killed by verifying the metadata.
The new code uses pidfd_open and name_to_handle_at when available.
It fallbacks to process start-time get using the gopsutil package.
Fixes: #25104
Signed-off-by: Jan Kaluza <jkaluza@redhat.com>
Winmake could only build amd64 artifacts (podman-remote, gvproxy,
win-sshproxy, podman.msi and podman-setup.exe).
This commit makes the necessary change to winmake so that it:
1) builds arm64 artifacts when executed on arm64
2) cross-compiles to arm64/amd64 with the `-architecture` parameter
It depends on https://github.com/containers/podman/pull/26023 that
removes the need to build `check.c` code (that is not used anyway).
Signed-off-by: Mario Loriedo <mario.loriedo@gmail.com>
This PR removes the file `check.c` and the instructions
to build it. The file was still built but the resulting
dll was not used by the installer anymore (see
https://github.com/containers/podman/pull/25237).
Signed-off-by: Mario Loriedo <mario.loriedo@gmail.com>
Locally building the windows installer requires to
build the documentation. And building documentation
requires Pandoc.
There is no pre-built binaries for Windows arm64
and this makes it complicated to build the Podman
Windows installer on Windows arm64.
To unlock this scenario we are adding a new winmake.ps1
target to build the documentation in a container (where
Pandoc is pre-installed).
Signed-off-by: Mario Loriedo <mario.loriedo@gmail.com>
Instead of hardcoding gvisor version in different build files, we get gvisor version from go.mod file.
Signed-off-by: Yevhen Vydolob <yvydolob@redhat.com>
It appears[1] that ginkgo binary can be easily built from the top-level
vendor, so it does not make sense to have a second copy in test/tools
(and a hassle of keeping the two in sync).
Inspired by [1], [2], and a run of make localtest, which shows:
> Ginkgo detected a version mismatch between the Ginkgo CLI and the version of Ginkgo imported by your packages:
> Ginkgo CLI Version:
> 2.22.1
> Mismatched package versions found:
> 2.23.3 used by podman, common, containers, parse, quadlet, rootlessport, abi, tunnel, libpod, events, file, shm, logs, annotations, libpod, utils, apiutil, auth, ctime, abi, expansion, utils, emulation, env, errorhandling, machine, compression, connection, define, ocipull, provider, proxyenv, qemu, command, shim, vmconfigs, rootless, signal, specgen, generate, kube, specgenutil, systemd, generate, notifyproxy, parser, timetype, trust, util, utils
[1]: af29bb5b6e (r2020246403)
[2]: e6c7ec94e2 (r2020230241)
Signed-off-by: Kir Kolyshkin <kolyshkin@gmail.com>
This amends commit 772ead253 ("Use simulated dual-stack binds when using WSL")
which started using linux tag but not for all files.
Alas, this breaks ginkgo run on Windows, so add an exclusion.
Signed-off-by: Kir Kolyshkin <kolyshkin@gmail.com>
As of commit d4c7ca39f ("update c/{buildah,common,image,storage} to
latest main") no dependencies use exclude_graphdriver_devicemapper tag.
This (hopefully) concludes the work of removing devicemapper graph
driver.
Signed-off-by: Kir Kolyshkin <kolyshkin@gmail.com>
Add a new target in winmake.ps1 to run unit tests and use
use it in a new cirrus task.
Fix machine_windows_test.go to make it work in CI machine.
Add the `!windows` tag on tests files that fail on Windows.
Signed-off-by: Mario Loriedo <mario.loriedo@gmail.com>
The current macos installer gvproxy version has a number of issues
around dns handling[1,2]. They should be fixed now in 0.8.3 as they
reverted the dns changes back to the 0.7.5 code. This should make it
work like it used to be. Our 5.3.2 installer ships the 0.7.5 version so
users of the podman installer never noticed that but we got plenty of
reports from brew users. This needs to be backported into v5.4 to make
sure we do not ship the broken version as part of the installer.
[1] https://github.com/containers/gvisor-tap-vsock/issues/462
[2] https://github.com/containers/gvisor-tap-vsock/issues/466
Signed-off-by: Paul Holzinger <pholzing@redhat.com>
The windows installer tests are command line / non interactive. To test as much as
possible the GUI / interactive scenario (that is what user do), update tests
need to use the installer with the default options. That's because when using the GUI
for an update, changing options is not possible.
Signed-off-by: Mario Loriedo <mario.loriedo@gmail.com>
Building the MSI hook on Windows
(`contrib/win-installer/podman-msihooks/check.c`)
currently requires MinGW. This commit updates the build
script so that, when MinGW is absent but the C compiler
included in Visual Studio BuildTools is installed, the
latter is used to build the MSI hook.
Other than that, `winmake.ps1` has a new `installertest`
target to run the Windows installer tests that are
currently verified by Cirrus CI.
Signed-off-by: Mario Loriedo <mario.loriedo@gmail.com>
Add the following targets in winmake.ps1:
- `installer`: builds the Windows installer
- `docs`: builds the documentation HTML pages
- `validatepr`: runs the Linux Makefile `.valiatepr`
target using podman.
- `lint`: faster validation that runs linter locally
Update of `build_windows.md`:
- removed every reference to the MSYS2 platform
- added instructions to build the installer and linters
Fix https://github.com/containers/podman/issues/21821 and https://github.com/containers/podman/issues/21911
Signed-off-by: Mario Loriedo <mario.loriedo@gmail.com>
For consistency with linux/osx makefile
I have added the win-gvproxy target as
an alias of win-sshproxy
[NO NEW TESTS NEEDED]
Signed-off-by: Mario Loriedo <mario.loriedo@gmail.com>
Moving from Go module v4 to v5 prepares us for public releases.
Move done using gomove [1] as with the v3 and v4 moves.
[1] https://github.com/KSubedi/gomove
Signed-off-by: Matt Heon <mheon@redhat.com>
This updates the Windows installer and macOS installer to ship the newer
binaries, as well as the vendored gvisor-tap-vsock code.
[NO NEW TESTS NEEDED]
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Also, de-duplicate power-shell variables and functions as they're
beginning to sprawl. This does not completely address all duplicates,
mainly those involved in the podman-machine CI workflow. So,
nothing under `contrib/win-installer` has been touched.
[NO NEW TESTS NEEDED]
Signed-off-by: Chris Evich <cevich@redhat.com>
Since gvisor-tap-vsock 0.7.1, the upstream project ships pre-built
Windows binaries for gvproxy and win-sshproxy. These binaries are built
with -Hwindowsgui as needed by podman.
This makes the same change in winmake.ps1, but I had to hardcode
gvisor-tap-vsock version in one more place.
[NO NEW TESTS NEEDED]
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Introduce a powershell script that mirrors Makefile capbilities on
Windows.
Syntax: ./winmake target [options]
[NO NEW TESTS NEEDED]
Signed-off-by: Ashley Cui <acui@redhat.com>
