* aya: enforce valid perf_event type & config combos
Add guardrails for when setting event type and config for perf_event
programs. The `PerfEventConfig` enum now defines the event `type` and
`config` of interest.
Remove public re-exports, and add idiomatic Rust types for:
- perf_hw_id => HardwareEvent
- perf_sw_ids => SoftwareEvent
- perf_hw_cache_id => HwCacheEvent
- perf_hw_cache_op_id => HwCacheOp
- perf_hw_cache_op_result_id => HwCacheResult
The motivation behind this is mainly for the `type` and `config` fields
of `bpf_link_info.perf_event.event`. The newly added enums are planned
to also be used in the `bpf_link_info` metadata.
Although `Breakpoint`/`PERF_TYPE_BREAKPOINT` variant exists, it is not
fully implemented. It's only usage at the moment is in link info.
* add breakpoint support to perf_event config interface
---------
Co-authored-by: Tyrone Wu <wudevelops@gmail.com>
For some reason, the aarch64 6.1 debian kernel was not compiled with
CONFIG_KALLSYMS_ALL=y, and the locations of globals are not available in
kallsyms. To attach breakpoints to these symbols in the test pipeline, we need
to read them from System.map and apply the kaslr offset to get their real
address. The System.map file is not provided in the kernel package by default,
so we need to extract it from the corresponding debug package.
- .github: pull the corresponding debug packages down as well as regular
kernels
- test: attach the perf_event_bp test breakpoint to the modprobe_path address
in kallsyms if present, or by applying the kaslr offset to the System.map
address if not found
- xtask: preferentially extract the System.map file from the debug package, if
available
- aya-obj: Generate userspace bindings for HW_BREAKPOINT_* options
- aya: Support PERF_TYPE_BREAKPOINT by adding an optional hardware
breakpoint configuration to PerfEvent::attach() and perf_event_open
functions.
- ebpf: Generate bindings for `struct bpf_perf_event_data` and update
the PerfEventContext to use them instead of a void pointer.
- xtask: Update codegen for constified anonymous HW_BREAKPOINT_* enums
When preparing the VM initramfs detect the `config-*` file that ships alongside
the vmlinuz/modules in each kernel archive and install it under `/boot` (both
as `/boot/config` and `/boot/config-<version>`). This makes the running
kernel’s configuration available inside the guest for the integration tests.
Bundle handling of Debian kernel archives into xtask so callers can pipe
the raw `.deb` paths straight into `cargo xtask integration-test vm …`.
The driver now extracts each archive into `<cache>/kernel-archives`,
locates the matching `vmlinuz-*`, `lib/modules/*`, and config files, and
feeds those into the initramfs build without requiring the user to
pre-run dpkg/tar. With this in place we drop
`.github/scripts/find_kernels.py`, simplify AGENTS.md/CI instructions to
use `find test/.tmp -name '*.deb'`, remove the gnu-tar requirement we no
longer need, and add `tar` as a workspace dependency for the extractor.
Recent changes[0][1] have broken compatibility with macOS; add a patch
to conditionally compile these snippets.
Patch and compile the source unconditionally; caching only the network
portion is good enough and less error prone.
[0] ae18b94099
[1] 97169cd6d9
Before this change, Aya supported only legacy BPF map definitions, which
are instances of the `bpf_map_def` struct and end up in the `maps` ELF
section.
This change introduces a BTF map definition for arrays, with custom
structs indicating the metadata of the map, which end up in the `.maps`
section.
Co-authored-by: Tamir Duberstein <tamird@gmail.com>
Enables creation of Map enum variants directly from MapData instances,
allowing user-space handles to pinned BPF maps without requiring the
original BPF object.
Supports multiple BPF map types.
Motivation:
- Simplifies accessing pinned maps from user space applications.
- Avoids full BPF reloads and potential deadlocks.
- Matches existing ergonomic APIs like LruHashMap::try_from.
- Keeps user code safe and idiomatic.
Closes https://github.com/aya-rs/aya/issues/1305.
Includes test coverage to validate the new API.
This change implements the `AsFd` trait for the `EbpfLogger` struct.
This allows obtaining a `BorrowedFd` from an `EbpfLogger` instance, which is safer than using `AsRawFd`.
This improves the ergonomics of using `EbpfLogger` with APIs that accept file descriptors.
Signed-off-by: Xiaobo Liu <cppcoffee@gmail.com>
Rather than support N async runtimes, push this to the user. The
relevant types (`PerfEventArrayBuffer` and `RingBuffer`) implement
`As{,Raw}Fd` which is sufficient with integration with tokio, smol, and
other async runtimes.
This doesn't get us to zero copy because the reserve/submit APIs do not
support DSTs for reasons I don't remember.
Now that it is unused in userspace, move `LOG_BUF_CAPACITY` to
`aya-log-ebpf` by making its type `LogValueLength` which obviates the
need for `log_value_length_sufficient`.
Implements TryFrom for FdLink for CgroupSkb, CgroupSock, CgroupSockAddr
and SockOps program types. This allows support for link pinning for
these program types, aligning with the documentation for FdLink.
Fixes: #739
Co-authored-by: Benjamin Barzen <bbarzen@amazon.com>
Cached probed for ProgramInfo fields instead of exposing it through
global FEATURE. Probing occurs on cache miss, which happens when first
accessing the field, *and* if the field is 0.
We have had loaded_links in the API as `#[doc(hidden)]` for a while.
I've been using it in bpfman and it's been fine. This commit does the
minimal work required to make the API stable.
We expose a `LinkInfo` type - similar to `ProgInfo` - which wraps the
generated type. In this case, `bpf_link_info`.
A few accessor functions have been added for `id`, `link_type` and
`program_id`. There are many more fields that could be (eventually)
made public.
As a convenience, `LinkInfo` can be retrieved from an existing FdLink
by using `FdLink::info()`.
Signed-off-by: Dave Tucker <dave@dtucker.co.uk>
Make the xz2 dependency optional to allow building without a C cross
compiler. This allows clippy.sh to be used on e.g. macOS more easily:
```
./clippy.sh --target x86_64-unknown-linux-gnu --exclude-features xz2
```
This removes octorust which takes absolutely ages to compile and also
requires native TLS libraries which in turn require a C toolchain. The
latter is a pain when cross compiling from macOS.
This reverts commit cc2da4a2a4.
The binding for BPF_ADD was missing from codegen.
Use BPF_.* to capture all of these to avoid missing bindings in future.
Signed-off-by: Dave Tucker <dave@dtucker.co.uk>
At present, `aya_build` will always use `+nightly` to build the
eBPF kernel. This is problematic in environments such as CI, where
tools always need to be installed first. Installing the current
nightly Rust toolchain gives you a new toolchain every day. This
poisones caches and makes CI jobs non-deterministic.
Resolves: #1226
Friday Ortiz
Friday Ortiz
Altug Bozkurt
Tamir Duberstein
Michal R
dependabot[bot]
tamird
Adam Schreck
Xiaobo Liu
Mehnaz Yunus
Tyrone Wu
Dave Tucker
dave-tucker
Michal Rostecki
Thomas Eizinger