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.
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.
clippy complained that cfg_attr is applied to the macro invocation and
therefore will not be expanded. This was a false-positive, however
when playing with cargo expand I did notice that the cfg and cfg_attr
section weren't propagating as I would expect them to.
Adding a meta matcher to the impl_try_from_map macro allows us to
remove the need for AsyncPerfEventArray to be in a separate invocation
of the macro while also making sure that attributes do get propagated
to the generated functions.
Signed-off-by: Dave Tucker <dave@dtucker.co.uk>
We have previously tried to import traits anonymously where possible but
enforcing this manually was hard.
Since Rust 1.83 clippy can now enforce this for us.
Signed-off-by: Dave Tucker <dave@dtucker.co.uk>
Change FromRawTracepointArgs::arg to return T rather than *const T which
seems to have been returning a dangling pointer.
Arguably this is not strictly necessary; edition 2024 seems to be
focused on increased strictness around unsafe code which doesn't unlock
new functionality for our users. That said, this work revealed an
apparent bug (see above) that we wouldn't otherwise catch due to
allow-by-default lints.
Per man 2 perf_event_open:
> RETURN VALUE
> On success, perf_event_open() returns the new file descriptor. On
> error, -1 is returned and errno is set to indicate the error.
Bake this into our syscalls so we stop using `_` so much which can hide
information loss. Remove the type parameter to SysResult.
Per man 2 bpf:
> RETURN VALUE
> For a successful call, the return value depends on the operation:
>
> BPF_MAP_CREATE
> The new file descriptor associated with the eBPF map.
>
> BPF_PROG_LOAD
> The new file descriptor associated with the eBPF program.
>
> All other commands
> Zero.
>
> On error, -1 is returned, and errno is set to indicate the error.
Bake this into our syscalls so we stop using `_` so much which can hide
information loss.
PerCpuHashMap was never returning MapError::KeyNotFound because
bpf_map_lookup_elem_per_cpu was replacing Ok(None) with
Ok(Some(zeroed_value)).
Update bpf_map_lookup_elem_per_cpu to map the Option value.
When `aya::obj` was migrated to be its own crate `aya-obj`, the `obj`
alias was created to preserve existing imports that relied on
`crate::obj`.
This resulted in 3 ways to import `aya-obj` objects:
- `use aya_obj::*`
- `use obj::*`
- `use crate::obj::*`
The `obj` alias is now removed to avoid confusion, and all `obj` imports
are funneled through `aya_obj`.
The loader should fill bss maps with zeros according to the size of the
ELF section.
Failure to do so yields weird verifier messages as follows:
```
cannot access ptr member ops with moff 0 in struct bpf_map with off 0 size 4
```
Reference to this in the cilium/ebpf code is here [1].
I could not find a reference in libbpf.
1: d0c8fc1937/elf_reader.go (L1159-L1165)
Signed-off-by: Dave Tucker <dave@dtucker.co.uk>
Adds detection for whether a field is available in `MapInfo`:
- For `map_type()`, we treturn new enum `MapType` instead of the integer
representation.
- For fields that can't be zero, we return `Option<NonZero*>` type.
- For `name_as_str()`, it now uses the feature probe `bpf_name()` to
detect if field is available.
Although the feature probe checks for program name, it can also be
used for map name since they were both introduced in the same commit.
Add conversion from u32 to program type, link type, and attach type.
Additionally, remove duplicate match statement for u32 conversion to
`BPF_MAP_TYPE_BLOOM_FILTER` & `BPF_MAP_TYPE_CGRP_STORAGE`.
New error `InvalidTypeBinding<T>` is created to represent when a
parsed/received value binding to a type is invalid.
This is used in the new conversions added here, and also replaces
`InvalidMapTypeError` in `TryFrom` for `bpf_map_type`.
Improves the existing integraiton tests for `loaded_programs()` and
`loaded_maps()` in consideration for older kernels:
- Opt for `SocketFilter` program in tests since XDP requires v4.8 and
fragments requires v5.18.
- For assertion tests, first perform the assertion, if the assertion
fails, then it checks the host kernel version to see if it is above
the minimum version requirement. If not, then continue with test,
otherwise fail.
For assertions that are skipped, they're logged in stderr which can
be observed with `-- --nocapture`.
This also fixes the `bpf_prog_get_info_by_fd()` call for kernels below
v4.15. If calling syscall on kernels below v4.15, it can produce an
`E2BIG` error because `check_uarg_tail_zero()` expects the entire
struct to all-zero bytes (which is caused from the map info).
Instead, we first attempt the syscall with the map info filled, if it
returns `E2BIG`, then perform syscall again with empty closure.
Also adds doc for which version a kernel feature was introduced for
better awareness.
The tests have been verified kernel versions:
- 4.13.0
- 4.15.0
- 6.1.0
Adding support for s390x (big endian architecture) and found that some
of the unit tests have structures and files implemented as byte arrays.
They are all coded as little endian and need a bug endian version to
work properly.
Signed-off-by: Billy McFall <22157057+Billy99@users.noreply.github.com>
There was a logic bug in the previously merged patch where we
set the correctly calculated max_entries size with the original.
To fix this and prevent regressions a unit test was added.
This highlighted that the original map definition needs to be
mutated in order for the max_entries change to be properly applied.
As such, this resize logic moved out of aya::sys into aya::maps
Signed-off-by: Dave Tucker <dave@dtucker.co.uk>
Both libbpf and cilium/ebpf have will set the max_entries of a
BPF_MAP_TYPE_PERF_EVENT_ARRAY to the number of online CPUs if
it was omitted at map definition time. This adds that same
logic to Aya.
Signed-off-by: Dave Tucker <dave@dtucker.co.uk>
The need for this type isn't specific to Miri; it is necessary on
toolchains containing https://github.com/rust-lang/rust/pull/124210 - it
just so happens that today this is nightly only, and so is Miri.
These only warn now:
```
test maps::hash_map::hash_map::tests::test_iter ... warning: integer-to-pointer cast
--> aya/src/maps/hash_map/hash_map.rs:304:15
|
304 | match unsafe { attr.__bindgen_anon_2.key } as *const T {
| ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ integer-to-pointer cast
|
= help: This program is using integer-to-pointer casts or (equivalently) `ptr::with_exposed_provenance`,
= help: which means that Miri might miss pointer bugs in this program.
= help: See https://doc.rust-lang.org/nightly/std/ptr/fn.with_exposed_provenance.html for more details on that operation.
= help: To ensure that Miri does not miss bugs in your program, use Strict Provenance APIs (https://doc.rust-lang.org/nightly/std/ptr/index.html#strict-provenance, https://crates.io/crates/sptr) instead.
= help: You can then pass the `-Zmiri-strict-provenance` flag to Miri, to ensure you are not relying on `with_exposed_provenance` semantics.
= help: Alternatively, the `-Zmiri-permissive-provenance` flag disables this warning.
= note: BACKTRACE on thread `maps::hash_map:`:
= note: inside `maps::hash_map::hash_map::tests::bpf_key::<i32>` at aya/src/maps/hash_map/hash_map.rs:304:15: 304:63
```
Adam Schreck
Tamir Duberstein
Omri Steiner
Dave Tucker
arctic-alpaca
aorhant
Tyrone Wu
Michal Rostecki
banditopazzo
Billy McFall
Zero King