Make profiling more amenable to other in-process engines (#11865)

Both the perfmap and jitdump protocols for profiling JITs require that
there's a single file with metadata per-process. This has the

Make profiling more amenable to other in-process engines (#11865)

Both the perfmap and jitdump protocols for profiling JITs require that
there's a single file with metadata per-process. This has the
unfortunate side effect of if there are multiple engines in the same
process there's no real way for them to coordinate when writing out
records to this file. In an attempt to at least try to synchronize this
commit updates these two implementations to issue a single `write`
syscall with the full contents of the record to a file opened in
`O_APPEND` mode.

If all the stars align and a partial write doesn't happen then that
means we're doing our best effort to cooperate with other engines in the
same process. If a partial write happens the decision in this PR is to
go ahead and retry the rest of the write (using the `write_all` helper).
This is a tradeoff where it makes the single-engine use case more robust
(partial writes are handled correctly) but the multi-engine use case
will produce corrupt files. It doesn't feel like a great tradeoff to say
that profiling can't be done in Wasmtime when a partial write would
happen for this niche use case of multiple engines, hence the direction
of this tradeoff.

Closes #11862

show more ...

Allow compiling debug-builtins in no_std environment to enable gdb jit (#11304)

* Allow compiling debug-builtins in no_std environment to enable gdb jit

- Make jit-debug crate #![no_std] when `std`

Allow compiling debug-builtins in no_std environment to enable gdb jit (#11304)

* Allow compiling debug-builtins in no_std environment to enable gdb jit

- Make jit-debug crate #![no_std] when `std` feature is not enabled
- Make gdb_jit_int module build in no_std environment
- Use std when perf_jitdebug is enabled because the perf_jitdebug module
does not run in no_std environment
- Change debug-builtins feature to only enable gdb_jit_int
- The perf_jitdebug feature is enabled in profiling mode only

Signed-off-by: Doru Blânzeanu <[email protected]>

* Remove feature listing from `default`

These are otherwise present via `std`.

* Shuffle around some features

* Enable `wasmtime-jit-debug/std` from wasmtime's `std` feature
* Only enable `perf_jitdump` from `profiling` as that's all that's
explicitly needed.

* Remove empty features list

* Cut down on some `#[cfg]` usage

* Update comments in gdb_jit_int module

* Add CI checks that `debug-builtins` builds on no_std

* Fix missing has_host_compiler_backend when debug-builtins enabled

Signed-off-by: Doru Blânzeanu <[email protected]>

---------

Signed-off-by: Doru Blânzeanu <[email protected]>
Co-authored-by: Alex Crichton <[email protected]>

show more ...

More clearly flag internal crates as such (#10963)

* More clearly flag internal crates as such

This commit is an attempt to more clearly flag internal crates in this
project as internal and not int

More clearly flag internal crates as such (#10963)

* More clearly flag internal crates as such

This commit is an attempt to more clearly flag internal crates in this
project as internal and not intended for external use. Specifically:

* Many crates are renamed from `wasmtime-foo` to
`wasmtime-internal-foo`.
* All of these crates now have `INTERNAL: ...` in their crates.io
description.
* All of these crates now have a warning at the top of their
documentation discouraging use.

This change is a result of rustsec/advisory-db#1999 where the goal is to
be crystal clear from a project perspective that usage of these crates
are highly discouraged and not supported. We'll still probably get such
advisories but we won't be considering them CVEs from the project itself
due to the internal nature of these crates and the discouraging
warnings.

Some concrete changes used here are:

* Inter-crate dependencies still use `wasmtime_foo` for naming and do
so with Cargo's package-renaming features.
* Crate renames are specified at the workspace level so the rename is
only in one locations and all other inherit it.
* Contribution documentation now has some brief guidelines about crate
organization.

* Update vet config

* Update checks for wasmtime-fiber

prtest:full

* Update publish script

* Another fiber rename

* Fix some doc tests

show more ...