superseded: [PATCH proxmox-ebpf v3 0/3] add proxmox-ebpf library

[Hannes Laimer <h.laimer@proxmox.com>]

superseded-by:
https://lore.proxmox.com/pve-devel/20261019142413.2255644-1-h.laimer@proxmox.com/

On 2026-09-09 12:41, Hannes Laimer wrote:
> proxmox-ebpf holds the eBPF programs and the code to load and drive
> them, grouped into subsystems. It is a library only, whoever needs a
> subsystem pulls in just that one through a cargo feature and calls it
> from their side. This series is the shared part, the actual
> subsystems come as their own series on top.
> 
> The core is the shared subsystem code. It takes care of the whole
> lifecycle, from loading and attaching per interface to pinning in
> bpffs and tearing down again. So nothing has to keep running, and each
> call picks up where the last one left off. Updates are handled too.
> The compiled object is hashed at build time and that hash plus a
> schema version are recorded per subsystem, so a call can tell what
> changed. A changed program is swapped onto the existing links without
> interrupting traffic, a changed map schema gets a teardown and
> rebuild. The BPF C code is also built natively against a small shim,
> so the parsing and packet building logic can be tested without a
> kernel. Packaging is debcargo like the other rust crates.
> 
> Changes carry a generation, a counter under /run. An apply notes its
> generation before it runs, so an older apply is dropped. So is a
> single change older than the last apply, that apply already read the
> change's input. A single change holds a lock while it reads and writes
> its entry and gets the generation of the last apply with it. That is
> all the lib does with the number. The subsystem decides when a change
> draws one, before or after it reads its input, and what an entry's
> generation means when a newer or an older one arrives. So the ordering
> rules live with the map they order.
> 
> The series applies on top of the scaffolding commit of the repo, which
> is not on the list since it carries the vendored vmlinux.h at ~167k
> lines. It is on my staff repo along with these three commits.
> 
> note: like for the dhcp series, this is only v3 due to me missing a
> problem and assuming it was fixed too fast. sorry for the noise also
> here :)
> 
> 
> proxmox-ebpf:
> 
> Hannes Laimer (3):
>   add the shared tc subsystem code
>   tests: add a native harness for the BPF C programs
>   debian: package the crate as a rust library
> 
>  .gitignore                     |   1 +
>  Cargo.toml                     |   7 +
>  Makefile                       |  54 ++
>  build.rs                       |  67 +++
>  debian/changelog               |   5 +
>  debian/control                 |  50 ++
>  debian/copyright               |  26 +
>  debian/debcargo.toml           |  13 +
>  debian/source/format           |   1 +
>  src/bpf-shim/bpf/bpf_endian.h  |  12 +
>  src/bpf-shim/bpf/bpf_helpers.h |  34 ++
>  src/bpf-shim/bpf_debug.h       |  10 +
>  src/bpf-shim/vmlinux.h         |  70 +++
>  src/lib.rs                     |   3 +
>  src/subsystem.rs               | 901 +++++++++++++++++++++++++++++++++
>  src/tc.rs                      | 257 ++++++++++
>  tests/common/mod.rs            | 209 ++++++++
>  tests/harness.rs               |  65 +++
>  18 files changed, 1785 insertions(+)
>  create mode 100644 Makefile
>  create mode 100644 debian/changelog
>  create mode 100644 debian/control
>  create mode 100644 debian/copyright
>  create mode 100644 debian/debcargo.toml
>  create mode 100644 debian/source/format
>  create mode 100644 src/bpf-shim/bpf/bpf_endian.h
>  create mode 100644 src/bpf-shim/bpf/bpf_helpers.h
>  create mode 100644 src/bpf-shim/bpf_debug.h
>  create mode 100644 src/bpf-shim/vmlinux.h
>  create mode 100644 src/subsystem.rs
>  create mode 100644 src/tc.rs
>  create mode 100644 tests/common/mod.rs
>  create mode 100644 tests/harness.rs
> 
> 
> Summary over all repositories:
>   18 files changed, 1785 insertions(+), 0 deletions(-)
>