LinkedGraph: support adding nodes and edges in arbitrary order
If an edge uses some not-yet-known node, we just leave the node's data empty, that data can be added later.
Use this support to avoid skipping edges in RetainedDepGraph.
This is continuation of https://github.com/rust-lang/rust/pull/152590, that PR just fixes the ICE, this PR also preserves all the edges in debug dumps.
This is also a minimized version of https://github.com/rust-lang/rust/pull/151821 with a smaller amount of data structure hacks.
core: make atomic primitives type aliases of `Atomic<T>`
Tracking issue: https://github.com/rust-lang/rust/issues/130539
This makes `AtomicI32` and friends type aliases of `Atomic<T>` by encoding their alignment requirements via the use of an internal `Storage` associated type. This is also used to encode that `AtomicBool` store a `u8` internally.
Modulo the `Send`/`Sync` implementations, this PR does not move any trait implementations, methods or associated functions – I'll leave that for another PR.
The manual `DynSend` implementation for `AtomicPtr` blocks the
auto-implementation for other atomic primitives since they forward to the same
`Atomic<T>` type now. This breakage cannot occur in user code as it depends on
`DynSend` being a custom auto-trait.
Use `HashStable` derive in more places
This applies `HashStable` derive in a couple more places. Also `stable_hasher` is declared for `HashStable_NoContext`.
If an edge uses some not-yet-known node, we just leave the node's data empty, that data can be added later.
Use this support to avoid skipping edges in DepGraphQuery
Use `scope` for `par_slice` instead of `join`
This uses `scope` instead of nested `join`s in `par_slice` so that each group of items are independent and do not end up blocking on another.
`get_resident_set_size` computed RSS by multiplying the number of pages
from `/proc/self/statm` with a hard-coded 4096-byte page size. This
produces incorrect results on systems where the runtime page size is
not 4 KiB.
Use `sysconf(_SC_PAGESIZE)` to determine the actual page size at runtime
so the RSS reported in `-Z time-passes` output is accurate across
platforms.
Tweak `VecCache` to improve performance
This has some tweaks to `VecCache` to improve performance.
- It saves a `compare_exchange` in `complete` using the new `put_unique` function.
- It removes bound checks on entries. These are instead checked in the `slot_index_exhaustive` test.
- `initialize_bucket` is outlined and tuned for that.
cc @Mark-Simulacrum
This simplifies the `PlaceholderReachability` `enum` by
replacing the case when no placeholders were reached with
a standard `Option::None`.
It also rewrites the API for `scc::Annotations` to be update-mut
rather than a more Functional programming style. This showed some slight
performance impact in early tests of the PR and definitely makes
the implementation simpler.
Don't leak sysroot crates through dependencies
Previously if a dependency of the current crate depended on a sysroot crate, then `extern crate` would in the current crate would pick the first loaded version of said sysroot crate even in case of an ambiguity. This is surprising and brittle. For `-Ldependency=` we already blocked this since rust-lang/rust#110229, but the fix didn't account for sysroot crates.
Should fix https://github.com/rust-lang/rust/issues/147966
There are two rustc_hash crates in the sysroot when building miri and
there is no way to disambiguate between the two. Instead use the
re-exports of it in the rustc_data_structures crate.
Add `Steal::risky_hack_borrow_mut`
I'm working on a rustc driver (Creusot) which needs to modify the MIR read by two queries, `mir_borrowck` and `check_liveness`, in different ways for each query. Both of these queries use `mir_promoted` to read the MIR, which is immutable (until it is stolen).
This adds an escape hatch so rustc drivers can mutate MIR for specific queries. And this removes `get_mut` which is unused and also unusable now that there's no way to get a `&mut Steal` from the rustc API.
Another approach may be to override the queries to modify the MIR after having read it from `mir_promoted`. However the implementation of queries is largely hidden, so I can't just copy their code to then modify it. A solution would be to parameterize the queries with callbacks which get instantiated with `mir_promoted` by default, but that seems more involved and ad hoc. That's why I'm proposing this smaller change instead.
It's common to import dependencies from the sysroot via `extern crate`
rather than use an explicit cargo dependency, when it's necessary to use
the same dependency version as used by rustc itself. However, this is
dangerous for crates.io crates, since rustc may not pull in the
dependency on some targets, or may pull in multiple versions. In both
cases, the `extern crate` fails to resolve.
To address this, re-export all such dependencies from the appropriate
`rustc_*` crates, and use this alias from crates which would otherwise
need to use `extern crate`.
Stuart Cook