| Age | Commit message (Collapse) | Author |
|---|
|
This version includes a fix for
T8391: Bonding interface permission errors with custom Ethernet MTU
|
|
|
|
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.
|
|
Kernel: T8868: Enable crash dump (kdump) and debug info in config
|
|
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.
|
|
|
|
|
|
|
|
for the fix for legacy hash algorithm loading
to make MS-CHAP work correctly again
|
|
Kernel: T8914: add support for 2.5G pluggables on BCM57810S
|
|
Kernel: T8605: net/l2tp: allow unmanaged tunnel setup without route to peer
|
|
Add the well-known JAMESMTL kernel module patch for bnx2x to advertise 2.5Gbit/s
capabilities on Broadcom NetXtreme2-X cards with BCM57810S chipset.
This is useful for ISP GPON access networks that use 2.5Gbit/s pluggables and
need the NIC to negotiate beyond 1000baseT/Full, avoiding the 940Mbit/s
practical cap on overprovisioned 1G services.
References:
* https://hack-gpon.org/broadcom-57810s/
* https://github.com/JAMESMTL/snippets/blob/dceb2fee74d80c66d/bnx2x/patches/bnx2x_warpcore_8727_2_5g_sgmii_txfault.patch
|
|
|
|
|
|
Update versions:
* igb v5.20.28
* ixgbe v6.4.4
* ixgbevf v5.3.36
* i40e v2.30.18
* ice v2.6.6
* iavf v4.13.35
|
|
|
|
The Kernel 6.18.33 now has an upstream fix for the fragnesia vulnerability
|
|
Kernel: T8880: reduce size of Kernel source TAR
|
|
* Kernel: T8846: consolidate config for CPU/Task time and stats accounting
* Kernel: T8879: add config section for hypervisor Virtual Socket Protocol
* Kernel: T8879: add missing Hyper-V drivers after 6.18 upgrade
Add drivers for Hyper-V hypervisor which got lost in transition from 6.6 to
6.18 Kernel upgrade.
|
|
Enable "make mrproper" before packaging kernel source tarball. After enabling
PWRU support in T8496, the auto-generated kernel source tarball grew from under
1 GB to ~4 GB, largely due to intermediate build artifacts (for example,
vmlinuz.o growing from 54 MB to 618 MB with additional debug symbols).
Since the tarball is intended to contain kernel sources and custom patches and
not intermediate objects, we now clean the tree with "make mrproper" before
creating the tarball. This keeps package size down while preserving
rebuildability.
|
|
This fixes the LPE https://github.com/0xdeadbeefnetwork/ssh-keysign-pwn
|
|
that allows unprivileged users to read files owned by any other user
|
|
Fragnesia is a universal Linux local privilege escalation exploit, discovered
with V12 by William Bowling with the V12 team. Fragnesia is a member of the
Dirty Frag vulnerability class. This is a separate bug in the ESP/XFRM from
dirtyfrag which has received its own patch. However, it is in the same surface
and the mitigation is the same as for dirtyfrag.
It abuses a logic bug in the Linux XFRM ESP-in-TCP subsystem to achieve
arbitrary byte writes into the kernel page cache of read-only files, without
requiring any race condition.
The technique extends the page-cache write bug class that includes Dirty Pipe:
when a TCP socket transitions to espintcp ULP mode after data has already been
spliced from a file into the receive queue, the kernel processes the queued
file pages as ESP ciphertext. The AES-GCM keystream byte at counter block
position 2, byte 0 is XORed directly into the cached file page. By selecting
the IV nonce to produce a desired keystream byte, any target byte in the file
can be set to any value — one byte per trigger invocation.
From: https://github.com/v12-security/pocs/blob/532994fc003a7/fragnesia/README.md
|
|
kernel: T8847: Enable PSI metric data
|
|
Enable the kernel [Pressure Stall Information][PSI] accounting feature.
This allows users to gather data on CPU and memory starvation. It's very
useful to help find cases where the system performance is impacted by
overloaded CPUs.
[PSI]: https://docs.kernel.org/accounting/psi.html
Signed-off-by: SuperQ <superq@gmail.com>
|
|
Kernel-created L2TPv3 tunnels (genetlink L2TP_CMD_TUNNEL_CREATE without
L2TP_ATTR_FD) used udp_sock_create() and kernel_connect(), which invoke
__ip4_datagram_connect() / __ip6_datagram_connect(). Those paths insist on
a successful FIB lookup at connect time. If no route to the configured
remote existed yet, tunnel and interface creation failed.
The data path already resolves routes on transmit (e.g. __ip_queue_xmit(),
inet6_csk_route_socket()). This change defers requiring a route until
packets are sent.
Details:
- UDP encapsulation: bind with udp_sock_create() after zeroing peer_udp_port,
then l2tp_udp_sk_set_peer() sets daddr/dport and socket "connected" state
without caching sk_dst from connect.
- IPv4 L2TP/IP (l2tp_ip): on -ENETUNREACH / -EHOSTUNREACH from
__ip4_datagram_connect(), l2tp_ip_connect_deferred() installs peer and
bind-table updates without a connect-time route.
- IPv6 L2TP/IP (l2tp_ip6): same for __ip6_datagram_connect(), including
IPv4-mapped peers and scope / bound-device checks aligned with the normal
connect path.
Forwarding still only happens once the FIB can reach the peer. Until then
outgoing packets follow the existing no-route drop path.
Assisted-by: Cursor:claude-4.6-opus
Signed-off-by: Christian Breunig <christian@breunig.cc>
|
|
The Intel iavf driver (formerly i40evf) is the Linux virtual function driver
for modern Intel Ethernet adapters supporting SR-IOV, including the X700/E800
series.
Renamed to iavf to support future devices, it replaced i40evf entirely by 2019.
The driver facilitates high-performance networking in virtualized environments.
This commit contains custom patches to make the driver build.
|
|
Instead of one build function per driver (ixgbe, i40, ice) which will anyways
result in the same Python code path to be executed - use one common helper
name: build_intel_nic
|
|
The current version of the Intel ice driver does not compile with a recent
LTS Kernel due to missing internal Kernel API adjustments in the Intel driver.
This commit contains custom patches to make the driver build.
|
|
The current version of the Intel i40e driver does not compile with a recent
LTS Kernel due to missing internal Kernel API adjustments in the Intel driver.
This commit contains custom patches to make the driver build.
|
|
|
|
|
|
It does not make sense to have the common VyOS functionality for e.g VXLAN,
GENEVE, PPP, WireGuard, filesystems crypto or module signing duplicated for
both arm64 and x86_64.
Split out common configuration parts to be defined only once.
|
|
* Forwared port all out-of-tree driver patches
* Add Intel QAT patch to make it compile for LInux 6.18
* Remove out-of-tree OpenVPN DCO module - now available upstream
|
|
|
|
|
|
|
|
|
|
|
|
The inotify support for overlayfs is no longer needed. Native upstream support
landed in kernel 4.16 (commit 31747eda41ef/764baba80168 in 2018, which made
overlayfs hash inodes by their lower inode so fsnotify works on overlay mounts),
and kernel 6.8 (commit bc2473c90fca in 2023) extended it so fsnotify generates
events for operations on the real underlying files of an overlay.
The patch was an out-of-tree workaround that never went upstream and predates
these solutions.
|
|
linux-kernel: T8506: Use scripts/kconfig/merge_config.sh for merging kernel config fragments
|
|
fragments
Using scripts/kconfig/merge_config.sh for merging config fragments provides
validation and insights compared to the previously used simple
concatenation
|
|
With this addition we can always sign a Kernel module later and ship it if
needed, without re-compiling the Kernel.
Kernel will report:
[ 1.223891] Loaded X.509 cert 'VyOS Networks Secure Boot Signer 2025 - linux: 6ca57e2add335babd08da69b48c70693edd2b037'
Issuer: CN = VyOS Networks Secure Boot CA
Validity
Not Before: Apr 26 09:07:06 2025 GMT
Not After : Apr 24 09:07:06 2035 GMT
Subject: CN = VyOS Networks Secure Boot Signer 2025 - linux
|
|
PWRU (https://github.com/cilium/pwru) is a very useful tool to debug complex networking issues on Linux, as it allows you to trace how packets travel through the kernel functions
|
|
|
