| Age | Commit message (Collapse) | Author |
|---|
|
kernel: T9197: enable CONFIG_IWLMLD
|
|
Kernel: T9195: Update Linux Kernel to 6.18.44
|
|
Kernel 6.18 moved Intel BE200/BE-series (Wi-Fi 7) devices to the new
iwlmld op-mode driver: their dev_info entries in iwlwifi's pcie/drv.c
are guarded by IS_ENABLED(CONFIG_IWLMLD). The VyOS kernel config enables
IWLWIFI/IWLDVM/IWLMVM but not IWLMLD, so BE-series support is compiled
out and probing fails with error -22 (no wlan interface is created).
Enable IWLMLD as a module on x86 and arm64, matching the existing
IWLDVM/IWLMVM op-modes. The firmware these devices need is already
shipped by the linux-firmware package.
|
|
T9198: Update the accel-ppp-ng package version to 9654bb6
|
|
Update the accel-ppp-ng version to 9654bb6
Includes fixes for T8391
(Bonding interface permission errors with custom ethernet MTU)
|
|
Upgrade rolling release Kernel version.
This is not a routine bump: 6.18.39 through 6.18.41 are unbootable on Intel
Denverton (DNV), Ice Lake Xeon D (ICX-D/CDF) and Snowridge (SNR) platforms,
which covers a large part of the small-form-factor hardware VyOS is deployed
on. We moved to 6.18.41 in 97c2a87c0 (T9067).
Upstream b1b4efea05a5 ("serial: 8250_mid: Disable DMA for selected platforms",
in stable since v6.18.39) set dnv_board.setup and dnv_board.exit to
PTR_IF(false, ...) to work around a DMA errata, but left the three call sites
in mid8250_probe() and mid8250_remove() dereferencing them unguarded. Since
CONFIG_SERIAL_8250_MID=y in our defconfig, the driver is built in and the
resulting NULL function pointer call happens during early PCI probe, killing
the probing kworker with interrupts disabled and wedging the boot:
CPU: 0 UID: 0 PID: 11 Comm: kworker/0:1 Not tainted 6.18.41-vyos #1
Hardware name: Supermicro SYS-510D-10C-FN6P/X12SDV-10C-SP6F, BIOS 2.2a
Workqueue: events work_for_cpu_fn
RIP: 0010:0x0
Call Trace:
mid8250_probe.part.0+0x10c/0x230
local_pci_probe+0x41/0x90
work_for_cpu_fn+0x16/0x20
note: kworker/0:1[11] exited with irqs disabled
The missing NULL checks were added by 7fb13fd7e9a5 ("serial: 8250_mid: Fix
NULL function pointer dereference on DNV/ICX-D/SNR platforms"), which is in
stable as of v6.18.42, so any release from .42 onwards resolves this. Taking
.44 as the current 6.18.y tip.
All five patches under patches/kernel/ still apply: of the files they touch,
only net/ipv6/route.c saw upstream changes in this range (1db34097998c, in
__ip6_del_rt_siblings() around line 4011), which does not overlap the
rt6_link_filter()/rt6_score_route() hunks of 0001 around line 717.
Changelogs:
* 6.18.44 https://cdn.kernel.org/pub/linux/kernel/v6.x/ChangeLog-6.18.44
* 6.18.43 https://cdn.kernel.org/pub/linux/kernel/v6.x/ChangeLog-6.18.43
* 6.18.42 https://cdn.kernel.org/pub/linux/kernel/v6.x/ChangeLog-6.18.42
|
|
Kernel: T9181: modernize packaging and remove redundant work
|
|
Convert build-nat-rtsp.sh, build-ipt-netflow.sh, build-intel-nic.sh,
build-intel-qat.sh and build-linux-firmware.sh from fpm to debmake +
debuild, matching the jool/realtek packaging style.
|
|
build-kernel.sh parsed kernel_flavor out of defaults.toml itself via a
brittle awk one-liner, even though build.py already has that value from
its own defaults.toml read.
Pass kernel_flavor down as an env var instead, so defaults.toml is read
in exactly one place (build.py's main) and everything downstream flows
through kernel-vars or explicit parameters.
|
|
Rewrite build-jool.py, build-realtek-r8126.py and build-realtek-r8152.py
as build-*.sh, matching the Bash convention already used by every other
sub-dependency script in this directory (kernel-vars sourcing, debmake +
debuild packaging).
|
|
Accel-PPP-ng only links against a few VPP libraries, but the build triggered
a full VPP package rebuild (make pkg-deb) on every run, with no caching.
Skip that step when the required VPP libraries already exist from a
previous run.
Delete them to force a rebuild.
|
|
Kernel: T9170: support compressed modules when searching module firmware
|
|
Commit 21c7e661 ("Kernel: T5641: enable module compression to save disk
space") which enabled module compression using XZ missed that when we search
modules for firmware inclusion, the pattern was hardcoded to *.ko instead of
*.ko* - this has been fixed.
|
|
T8329: Make the Azure VF-naming helper script 'vyos_vf_name' available in initramfs
|
|
T9164: accept NOS-/VD- Jira keys in task-id check; pin NOS in coderabbit
|
|
podman: T7736: update netavark and aardvark-dns packages
|
|
VD tracker project renamed to NOS (2026-07) - PR titles and commit headlines
using the new NOS- key must stop being flagged invalid-task-id; legacy VD-
keys stay accepted, and CodeRabbit now resolves NOS alongside VD. Fleet
replication of canary vyos/vyos-1x#5379, whose adversarial review tightened
the regexes to ASCII [0-9] + literal-space separators.
🤖 Generated by [robots](https://vyos.io)
|
|
Package update required in lockstep mode to add support for renaming podman
macvlan interfaces.
|
|
T9147: Bump accel-ppp-ng version to 9874ea3
|
|
Kernel: T9067: Update Linux Kernel to 6.18.41
|
|
Upgrade rolling release Kernel version.
Changelogs:
* 6.18.41 https://lwn.net/Articles/1086227/
* 6.18.40 https://lwn.net/Articles/1084923/
* 6.18.39 https://lwn.net/Articles/1083464/
|
|
This version includes a fix for
T8391: Bonding interface permission errors with custom Ethernet MTU
|
|
build: T9140: include version and flavor information in the boot menu
|
|
|
|
kernel: T9138: enable support for Microsoft MANA NIC driver
|
|
|
|
iproute2: T8926: add custom build which matches Kernel 6.18
|
|
T9124: Fix smoketest file check
|
|
|
|
The prior expect of 'No such file or directory' can prevent smoketests from completing if test fails with that same error
|
|
Testsuite: T9021: answer question during test for bash_history migration
|
|
|
|
initramfs so that VFs with add events appearing in early stage can be renamed using this script
|
|
T9099: improve test framework - add safeguards
|
|
Kernel: T9103: fix arm64 syscalltbl path breaking linux-perf package build
|
|
Building linux-perf-* on a native arm64 build host failed with "No rule to
make target '.../unistd_64.h'" because arch/arm64/kernel/Makefile.syscalls
overrides syscalltbl with a path relative to the kernel root, but
tools/lib/perf/Makefile invokes the header generator from tools/lib/perf/
instead. Make the override absolute, matching the srctree-prefixed default.
|
|
BOOTLOADERchooseSerialConsole() sent fixed DOWN/ENTER keystrokes with blind
sleeps to navigate the post-install GRUB "Boot options" and "Select console
type" submenus. Under host load, a dropped keystroke could land navigation one
level too deep (e.g. "Select boot mode" instead of "Select console type").
GRUB submenus never time out on their own, so the VM sat there until the
unrelated 600s login wait in loginVM() expired, producing a confusing
pexpect.TIMEOUT far from the real cause.
Add child.expect() checks after each submenu transition to confirm the expected
menu actually rendered before sending the next keypress, so a misnavigation now
fails fast and points at the right step.
|
|
Test targets forward extra CLI arguments to their scripts via
$(filter-out $@,$(MAKECMDGOALS)) (e.g. make test -- --match interfaces_bond)
Those extra words are also goals as far as make is concerned, so after the test
recipe succeeded, make went on to "build" --match and interfaces_bond too,
falling through to the generic `%:` flavor rule and running
build-vyos-image --match
which errored and masked the fact that the smoketest itself had already passed.
Declare the trailing goals as phony no-ops when the primary target is one of
the test targets, so they no longer reach the `%:` rule.
The existing MATCH=<value> form is unaffected.
|
|
sbom: T9098: syft should run un squashfs instead of unpacked chroot
|
|
Kernel: T8868: Enable crash dump (kdump) and debug info in config
|
|
lb (live-build) binary runs binary_rootfs first, then binary_grub-efi.
binary_grub-efi temporarily installs EFI tooling and then runs:
"apt remove --auto-remove --purge --allow-remove-essential"
That cleanup is safe for the already-generated squashfs, but it mutates
build/chroot and can remove vyos-1x, breaking subsequent work that expects
chroot to remain intact for SBOM generation.
But why unpacking the squashfs?
In an ideal world we could call syft on the compressed squashfs file which is
supported. In our world, we do use a BCJ pre-filter chained with LZMA2 for
compressing the squashfs, which will increase the compression ratio without a
decompression penalty.
Difference: ~14.3 MB, ~2.7% smaller
This is unsupported by fyft which means we do need to unpack the squashfs first
before checking the files and generating the SBOM file.
|
|
openssl: T9083: fix APT detected package downgrade despite identical versions
|
|
VyOS rolling APT repo (packages.vyos.net/repositories/rolling) publishing its
own openssl/libssl3 packages, pinned at priority 600 (higher than Debian's 500)
causes a "poisoning" of the APT cache policy inside the chroot:
openssl:
Installed: 3.0.20-1~deb12u2
Candidate: 3.0.20-1~deb12u2
Version table:
* 3.0.20-1~deb12u2 500 <- debian bookworm + bookworm-security
* 3.0.20-1~deb12u2 600 <- packages.vyos.net rolling (same version string, higher pin)
* 3.0.17-1~deb12u2 500 <- bookworm-updates
This results in APT seeing a version-string tie between Debian's build and the
VyOS repo build of the same package and, because of the pin-priority swap,
classifies switching to the VyOS repo copy as a "downgrade".
This is fixed by appending a clear VyOS related marked to the package version,
preventing any possible downgrade detection.
Signed-off-by: Date Huang <tjjh89017@hotmail.com>
Co-authored-by: Christian Breunig <christian@breunig.cc>
|
|
Kernel: T5641: enable module compression to save disk space
|
|
Enabling the following options allows us to use `kdump-tools`:
- `CONFIG_CRASH_DUMP`
- `CONFIG_DEBUG_INFO`
- `CONFIG_PROC_VMCORE`
|
|
Kernel: T9067: Update Linux Kernel to 6.18.38
|
|
Option useful during development to not automatically clean the Linux Kernel
source folder of all its intermediate files.
|
|
After signing the Linux Kernel modules with the ephemeral key, proceed
by compressing the Kernel modules with xz to reduce the final ISO image
size.
Initial tests have shown a size reduction by 60MiB.
|
|
|
|
|
