Thread (18 messages) flat view 18 messages, 3 authors, 8h ago

Re: [PATCH RFC v3 00/13] sysctl: add module aliases

From: Mauricio Faria de Oliveira <hidden>
Date: 2026-08-20 21:22:41
Also in: bpf, bridge, keyrings, linux-fsdevel, linux-kbuild, linux-rdma, linux-riscv, linux-s390, linux-sctp, lkml, lvs-devel, mptcp, netfilter-devel, virtualization

On 2026-08-20 09:51, Joel Granados wrote:
On Wed, Aug 19, 2026 at 03:16:13PM -0300, Mauricio Faria de Oliveira wrote:
quoted
This series adds 'sysctl:' aliases to modules that register sysctl tables; e.g.:

        $ modinfo ./mpls_router.ko  | grep sysctl:
        alias:          sysctl:*/net/mpls/conf/*/input
        alias:          sysctl:*/net/mpls/default_ttl
        alias:          sysctl:*/net/mpls/ip_ttl_propagate
        alias:          sysctl:*/net/mpls/platform_labels

It provides a trivial way to map /proc/sys files to modules (not trivial today),
and for userspace to handle nonexistent /proc/sys files (e.g., procps's sysctl
and systemd-sysctl applying tunables) with "modprobe sysctl:<...>" and a retry.

This is done almost automatically with register_sysctl(), register_net_sysctl()
and friends as wrappers of MODULE_SYSCTL_TABLE (similar to MODULE_DEVICE_TABLE),
which emits symbols for file2alias/modpost to find and parse the sysctl tables.

The big exception to 'almost' are sysctl tables and paths allocated or defined
at runtime (e.g., per-namespace or per-device), as all information is required
at build-time. Fortunately, such tables and paths are often based on 'templates'
which are static and can be used.

This is done by plumbing the template table/path as optional arguments (macros
with default values as default_gfp()), so not to create functions for all cases:

        register_sysctl(path, table [, table_tmpl [, path_tmpl]]);
        register_net_sysctl(net, path, table [, table_tmpl[, path_tmpl]]);
        register_net_sysctl_sz(net, path, table, size [, table_tmpl[, path_tmpl]]);
The "what" is described but I'm missing more clarity on the "why". Why
does this need to be trivial? Where is it that you will know the sysctl
file path of a module and not the module name or alias?
Thanks for looking at this.

The problem this feature addresses is sysctl settings not applied
because modules aren't yet loaded when systemd-sysctl/procps's sysctl
runs on boot, and the usage of /etc/modules as a workaround.

This can be addressed with a way for sysctl tools to load the module for
a non-existent sysctl file path.
For that, the mapping between a sysctl file path (known) and its module
(unknown) needs to be trivial.

It may not seem serious at first, but I've seen this consume significant
engineering time and impact production systems, in a previous technical
support job.
An example: cloud deployments with non-scalable network routing
performance due to nf_conntrack_max not applied after reboot or an
upgrade because the component which turned out to help with the
/etc/modules workaround had its placement changed from network router
nodes.
Additionally, sysctls are not module specific; they are a way to
read/write kernel variables. Putting a module specific aspect in the
function arguments (that needs to be ignored in non-module cases) seems
wrong. Why not use the module subsys to add sysctl alias instead of the
sysctl subsys?
Fair point. The current design has 2 reasons:

1) Use the 'path' argument (raised in [1]), only available in the
register sysctl functions.
2) Use a more implicit/transparent approach, instead of more
explicit/declarative approach. 

I guess that a different design could use MODULE_SYSCTL_TABLE() as
MODULE_DEVICE_TABLE(), declared per table instead of wrapped into
register sysctl functions.
However, it seems to require moving the value of the path argument (or
its template) into the macro and still referencing it in the function
(or its instantiation of the template), which adds obfuscation, to all
callers. 

What do you think?

[snip]
quoted
P.S.
====

I wrote proof-of-concept patches for procps sysctl and systemd-sysctl some time
ago, which worked as expected, i.e., successfully set sysctl tunables which did
not exist in /proc/sys, by running 'modprobe sysctl:/proc/sys/...' and retrying.
Do you have links to this? To the old one, not necessarily a new
version.
Sure, please see [2] and [3]. I did a quick rebase/refresh, but both
seem to work:

procps sysctl:

    $ strace -e execve -f -s 64 ./src/sysctl --modprobe=1 -w net.test=1
2>&1 | grep modprobe

    [pid 505958] execve("/bin/sh", ["sh", "-c", "--", "modprobe
sysctl:/proc/sys/net/test >/dev/null 2>&1"], 0x7ffc12efff80 /* 63 vars
*/) = 0

systemd-sysctl:

$ cat /etc/sysctl.d/net-test.conf
net.test = 1

    $ sudo strace -e execve -f -s 64 ./build/systemd-sysctl
--modprobe=true /etc/sysctl.d/net-test.conf  2>&1 | grep modprobe

[pid 505777] execve("/usr/sbin/modprobe", ["modprobe",
"sysctl:*/net/test"], 0x7ffd473d1418 /* 20 vars */) = 0

Thanks,

[1]
https://lore.kernel.org/linux-fsdevel/CAK7LNAR=7zgOiqTD9okXfZXroFH1yagMFsRuq0G-z6OfSUPLQg@mail.gmail.com/ (local)
[2] https://github.com/mfoliveira/procps/commits/sysctl-module-aliases/
[3] https://github.com/mfoliveira/systemd/commits/sysctl-module-aliases/
Best
quoted
Should this series eventually be merged, the patches will be submitted upstream
as well, for userspace to start consuming it.

Signed-off-by: Mauricio Faria de Oliveira <redacted>

Changes in v3:
- All issues were reported by sashiko-bot@kernel.org.
- Tested allmodconfig with option enabled on x86_64.
- Patch 2:
  - Fix typo ('s/options/option/').
- Patch 3:
  - Fix usage of __UNIQUE_ID().
  - Add __must_be_array() in MODULE_SYSCTL_TABLE().
- Patch 4:
  - Remove trailing backslash in register_sysctl().
- Patch 6:
  - Fix missing __register_net_sysctl_sz() in !CONFIG_SYSCTL.
- Patch 13:
  - Fix potential out-of-bounds read in do_sysctl_entry().
  - Add check for entry size of zero.
- Link to v2: https://lore.kernel.org/r/20260818-sysctl-module-aliases-v2-0-d5a69dae5798@igalia.com (local)

Changes in v2:
- This is based on the series submitted 4 years ago,
  with Originally-by: tags added in related patches.
- Link to v1: https://lore.kernel.org/linux-fsdevel/20220722022416.137548-1-mfo@canonical.com/ (local)

---
Mauricio Faria de Oliveira (13):
      keys, pidns, fs/verity, riscv/vector: reorder '#include <linux/sysctl.h>'
      proc: add config option SYSCTL_MODULE_ALIASES
      sysctl, mod_devicetable: add macro MODULE_SYSCTL_TABLE
      sysctl: add register_sysctl() wrapper for MODULE_SYSCTL_TABLE
      sysctl, parport: update register_sysctl() callers with template arguments
      sysctl, net: add register_net_sysctl{_sz}() wrappers for MODULE_SYSCTL_TABLE
      sysctl, net: update register_net_sysctl{_sz}() callers with template arguments
      sysctl, net: update register_net_sysctl_sz(ARRAY_SIZE(table_tmpl)) with template arguments
      sysctl, ipv6: update register_net_sysctl{_sz}() callers with template arguments
      sysctl, net: update register_net_sysctl_sz() edge case
      sysctl: unrandomize struct ctl_table.procname
      modpost: move addend_*_rel() calls into addend_rel()
      modpost: handle MODULE_SYSCTL_TABLE symbols

 arch/riscv/kernel/vector.c              |   1 +
 drivers/net/vrf.c                       |   3 +-
 drivers/parport/procfs.c                |  26 +++-
 fs/proc/Kconfig                         |  13 ++
 fs/verity/init.c                        |   1 +
 include/linux/key.h                     |   1 -
 include/linux/mod_devicetable.h         |   7 +
 include/linux/pid_namespace.h           |   1 +
 include/linux/sysctl.h                  | 110 ++++++++++++++-
 include/net/ipv6.h                      |   6 +-
 include/net/net_namespace.h             |  47 ++++++-
 net/bridge/br_netfilter_hooks.c         |   3 +-
 net/core/neighbour.c                    |  11 +-
 net/core/sysctl_net_core.c              |   3 +-
 net/ieee802154/6lowpan/reassembly.c     |   2 +-
 net/ipv4/devinet.c                      |  12 +-
 net/ipv4/ip_fragment.c                  |   3 +-
 net/ipv4/route.c                        |   3 +-
 net/ipv4/sysctl_net_ipv4.c              |   2 +-
 net/ipv4/xfrm4_policy.c                 |   3 +-
 net/ipv6/addrconf.c                     |   7 +-
 net/ipv6/icmp.c                         |   6 +-
 net/ipv6/netfilter/nf_conntrack_reasm.c |   4 +-
 net/ipv6/reassembly.c                   |   3 +-
 net/ipv6/route.c                        |  10 +-
 net/ipv6/sysctl_net_ipv6.c              |  13 +-
 net/ipv6/xfrm6_policy.c                 |   3 +-
 net/mpls/af_mpls.c                      |   9 +-
 net/mptcp/ctrl.c                        |   3 +-
 net/netfilter/ipvs/ip_vs_ctl.c          |   4 +-
 net/netfilter/ipvs/ip_vs_lblc.c         |   3 +-
 net/netfilter/ipvs/ip_vs_lblcr.c        |   3 +-
 net/netfilter/nf_conntrack_standalone.c |   5 +-
 net/netfilter/nf_log.c                  |   9 +-
 net/rds/tcp.c                           |   4 +-
 net/sctp/sysctl.c                       |   3 +-
 net/smc/smc_sysctl.c                    |   2 +-
 net/sysctl_net.c                        |  10 +-
 net/unix/sysctl_net_unix.c              |   3 +-
 net/vmw_vsock/af_vsock.c                |   4 +-
 net/xfrm/xfrm_sysctl.c                  |   2 +-
 scripts/mod/devicetable-offsets.c       |   6 +
 scripts/mod/file2alias.c                | 237 ++++++++++++++++++++++++++++++++
 scripts/mod/modpost.c                   |  46 ++++---
 scripts/mod/modpost.h                   |  25 ++++
 45 files changed, 575 insertions(+), 107 deletions(-)
---
base-commit: 2697ef8943c9985c14708a6429e21812693857b2
change-id: 20260818-sysctl-module-aliases-2f5801b1eb71

Best regards,
-- 
Mauricio Faria de Oliveira [off-list ref]
-- 
Mauricio
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help