| Age | Commit message (Collapse) | Author | Lines |
|---|
|
Most notably, this commit changes the `pub use crate::*;` in that file
to `use crate::*;`. This requires a lot of `use` items in other crates
to be adjusted, because everything defined within `rustc_span::*` was
also available via `rustc_span::source_map::*`, which is bizarre.
The commit also removes `SourceMap::span_to_relative_line_string`, which
is unused.
|
|
This fixes the changes brought to codegen tests when effect params are
added to libcore, by not attempting to monomorphize functions that get
the host param by being `const fn`.
|
|
Removes a piece of mutable state.
Follow up to #114578.
|
|
|
|
|
|
|
|
It doesn't seem worthwhile now that `MonoItem::size_estimate` is called
much less often.
|
|
This means we call `MonoItem::size_estimate` (which involves a query)
less often: just once per mono item, and then once more per inline item
placement. After that we can reuse the stored value as necessary. This
means `CodegenUnit::compute_size_estimate` is cheaper.
|
|
It replaces `(Linkage, Visibility)`, making the code nicer. Plus the
next commit will add another field.
|
|
|
|
- Rename `create_size_estimate` as `compute_size_estimate`, because that
makes more sense for the second and subsequent calls for each CGU.
- Change `CodegenUnit::size_estimate` from `Option<usize>` to `usize`.
We can still assert that `compute_size_estimate` is called first.
- Move the size estimation for `place_mono_items` inside the function,
for consistency with `merge_codegen_units`.
|
|
Because CGU merging relies on CGU sizes, but the CGU sizes before
inlining aren't accurate.
This requires tweaking how the sizes are updated during merging: if CGU
A and B both have an inlined function F, then `size(A + B)` will be a
little less than `size(A) + size(B)`, because `A + B` will only have one
copy of F. Also, the minimum CGU size is increased because it now has to
account for inlined functions.
This change doesn't have much effect on compile perf, but it makes
follow-on changes that involve more sophisticated reasoning about CGU
sizes much easier.
|
|
- Add more total and per-CGU measurements.
- Ensure CGUs are sorted by name before the first `debug_dump` calls,
for deterministic output.
- Print items within CGUs in sorted-by-name order, for deterministic
output.
- Add some assertions and comments clarifying sortedness of CGUs at
various points.
An example, before:
```
INITIAL PARTITIONING (5 CodegenUnits, max=29, min=1, max/min=29.0):
CodegenUnit scev95ysd7g4b0z estimated size 2:
- fn <() as std::process::Termination>::report [(External, Hidden)] [h082b15a6d07338dcE] estimated size 2
CodegenUnit 1j0frgtl72rsz24q estimated size 29:
- fn std::rt::lang_start::<()>::{closure#0} [(External, Hidden)] [h695c7b5d6a212565E] estimated size 17
- fn std::rt::lang_start::<()> [(External, Hidden)] [h4ca942948e9cb931E] estimated size 12
CodegenUnit 5dbzi1e5qm0d7kj2 estimated size 4:
- fn <[closure@std::rt::lang_start<()>::{closure#0}] as std::ops::FnOnce<()>>::call_once - shim [(External, Hidden)] [h24eaa44f03b2b233E] estimated size 1
- fn <fn() as std::ops::FnOnce<()>>::call_once - shim(fn()) [(External, Hidden)] [hf338f5339c3711acE] estimated size 1
- fn <[closure@std::rt::lang_start<()>::{closure#0}] as std::ops::FnOnce<()>>::call_once - shim(vtable) [(External, Hidden)] [h595d414cbb7651d5E] estimated size 1
- fn std::ptr::drop_in_place::<[closure@std::rt::lang_start<()>::{closure#0}]> - shim(None) [(External, Hidden)] [h17a19dcdb40600daE] estimated size 1
CodegenUnit 220m1mqa2mlbg7r3 estimated size 1:
- fn main [(External, Hidden)] [hb29587cdb6db5f42E] estimated size 1
CodegenUnit 4ulbh241f7tvyn7x estimated size 6:
- fn std::sys_common::backtrace::__rust_begin_short_backtrace::<fn(), ()> [(External, Hidden)] [h41dada2c21a1259dE] estimated size 6
```
and after:
```
INITIAL PARTITIONING (9 items, total_size=42; 5 CGUs, max_size=29, min_size=1, max_size/min_size=29.0):
- CGU[0] 1j0frgtl72rsz24q (2 items, size=29):
- fn std::rt::lang_start::<()> [(External, Hidden)] [h4ca942948e9cb931E] (size=12)
- fn std::rt::lang_start::<()>::{closure#0} [(External, Hidden)] [h695c7b5d6a212565E] (size=17)
- CGU[1] 220m1mqa2mlbg7r3 (1 items, size=1):
- fn main [(External, Hidden)] [hb29587cdb6db5f42E] (size=1)
- CGU[2] 4ulbh241f7tvyn7x (1 items, size=6):
- fn std::sys_common::backtrace::__rust_begin_short_backtrace::<fn(), ()> [(External, Hidden)] [h41dada2c21a1259dE] (size=6)
- CGU[3] 5dbzi1e5qm0d7kj2 (4 items, size=4):
- fn <[closure@std::rt::lang_start<()>::{closure#0}] as std::ops::FnOnce<()>>::call_once - shim(vtable) [(External, Hidden)] [h595d414cbb7651d5E] (size=1)
- fn <[closure@std::rt::lang_start<()>::{closure#0}] as std::ops::FnOnce<()>>::call_once - shim [(External, Hidden)] [h24eaa44f03b2b233E] (size=1)
- fn <fn() as std::ops::FnOnce<()>>::call_once - shim(fn()) [(External, Hidden)] [hf338f5339c3711acE] (size=1)
- fn std::ptr::drop_in_place::<[closure@std::rt::lang_start<()>::{closure#0}]> - shim(None) [(External, Hidden)] [h17a19dcdb40600daE] (size=1)
- CGU[4] scev95ysd7g4b0z (1 items, size=2):
- fn <() as std::process::Termination>::report [(External, Hidden)] [h082b15a6d07338dcE] (size=2)
```
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
spastorino noticed some silly expressions like `item_id.def_id.def_id`.
This commit renames several `def_id: OwnerId` fields as `owner_id`, so
those expressions become `item_id.owner_id.def_id`.
`item_id.owner_id.local_def_id` would be even clearer, but the use of
`def_id` for values of type `LocalDefId` is *very* widespread, so I left
that alone.
|
|
fix a ui test
use `into`
fix clippy ui test
fix a run-make-fulldeps test
implement `IntoQueryParam<DefId>` for `OwnerId`
use `OwnerId` for more queries
change the type of `ParentOwnerIterator::Item` to `(OwnerId, OwnerNode)`
|
|
|
|
This is no longer used only for debugging options (e.g. `-Zoutput-width`, `-Zallow-features`).
Rename it to be more clear.
|
|
It returns the previous work product or panics if there is none. This rename
makes the purpose of this method clearer.
|
|
|
|
|
|
In `CodegenUnit::items_in_deterministic_order`, there's a comment that
only local HirIds should be taken into account, but #90408 removed the
`as_local` call that sets others to None. Restoring that check fixes the
s390x hangs seen in [RHBZ 2058803].
[RHBZ 2058803]: https://bugzilla.redhat.com/show_bug.cgi?id=2058803
|
|
|
|
[code coverage] Fix missing dead code in modules that are never called
The issue here is that the logic used to determine which CGU to put the dead function stubs in doesn't handle cases where a module is never assigned to a CGU (which is what happens when all of the code in the module is dead).
The partitioning logic also caused issues in #85461 where inline functions were duplicated into multiple CGUs resulting in duplicate symbols.
This commit fixes the issue by removing the complex logic used to assign dead code stubs to CGUs and replaces it with a much simpler model: we pick one CGU to hold all the dead code stubs. We pick a CGU which has exported items which increases the likelihood the linker won't throw away our dead functions and we pick the smallest to minimize the impact on compilation times for crates with very large CGUs.
Fixes #91661
Fixes #86177
Fixes #85718
Fixes #79622
r? ```@tmandry```
cc ```@richkadel```
This PR is not urgent so please don't let it interrupt your holidays! 🎄 🎁
|
|
Implement `Ord`, `PartialOrd` for SpanData
|
|
The issue here is that the logic used to determine which CGU to put the
dead function stubs in doesn't handle cases where a module is never
assigned to a CGU.
The partitioning logic also caused issues in #85461 where inline
functions were duplicated into multiple CGUs resulting in duplicate
symbols.
This commit fixes the issue by removing the complex logic used to assign
dead code stubs to CGUs and replaces it with a much simplier model: we
pick one CGU to hold all the dead code stubs. We pick a CGU which has
exported items which increases the likelihood the linker won't throw
away our dead functions and we pick the smallest to minimize the impact
on compilation times for crates with very large CGUs.
Fixes #86177
Fixes #85718
Fixes #79622
|
|
Rollup of 7 pull requests
Successful merges:
- #91141 (Revert "Temporarily rename int_roundings functions to avoid conflicts")
- #91984 (Remove `in_band_lifetimes` from `rustc_middle`)
- #92028 (Sync portable-simd to fix libcore build for AVX-512 enabled targets)
- #92042 (Enable `#[thread_local]` for all windows-msvc targets)
- #92071 (Update example code for Vec::splice to change the length)
- #92077 (rustdoc: Remove unused `collapsed` field)
- #92081 (rustdoc: Remove unnecessary `need_backline` function)
Failed merges:
r? `@ghost`
`@rustbot` modify labels: rollup
|
|
See #91867
This was mostly straightforward. In several places, I take advantage
of the fact that lifetimes are non-hygenic: a macro declares the
'tcx' lifetime, which is then used in types passed in as macro
arguments.
|
|
|
|
|
|
polymorphization: shims and predicates
Supersedes #75737 and #75414. This pull request includes up some changes to polymorphization which hadn't landed previously and gets stage2 bootstrapping and the test suite passing when polymorphization is enabled. There are still issues with `type_id` and polymorphization to investigate but this should get polymorphization in a reasonable state to work on.
- #75737 and #75414 both worked but were blocked on having the rest of the test suite pass (with polymorphization enabled) with and without the PRs. It makes more sense to just land these so that the changes are in.
- #75737's changes remove the restriction of `InstanceDef::Item` on polymorphization, so that shims can now be polymorphized. This won't have much of an effect until polymorphization's analysis is more advanced, but it doesn't hurt.
- #75414's changes remove all logic which marks parameters as used based on their presence in predicates - given #75675, this will enable more polymorphization and avoid the symbol clashes that predicate logic previously sidestepped.
- Polymorphization now explicitly checks (and skips) foreign items, this is necessary for stage2 bootstrapping to work when polymorphization is enabled.
- The conditional determining the emission of a note adding context to a post-monomorphization error has been modified. Polymorphization results in `optimized_mir` running for shims during collection where that wouldn't happen previously, some errors are emitted during `optimized_mir` and these were considered post-monomorphization errors with the existing logic (more errors and shims have a `DefId` coming from the std crate, not the local crate), adding a note that resulted in tests failing. It isn't particularly feasible to change where polymorphization runs or prevent it from using `optimized_mir`, so it seemed more reasonable to not change the conditional.
- `characteristic_def_id_of_type` was being invoked during partitioning for self types of impl blocks which had projections that depended on the value of unused generic parameters of a function - this caused a ICE in a debuginfo test. If partitioning is enabled and the instance needs substitution then this is skipped. That test still fails for me locally, but not with an ICE, but it fails in a fresh checkout too, so 🤷♂️.
r? `@lcnr`
|
|
|
|
rustc adds notes to errors which happen post-monomorphization to
provide the user with helpful context (as these errors may rely on the
specific instantiations). To prevent this note being added where it is
not appropriate, the node is checked to originate outwith the current
crate. However, when polymorphization is enabled, this can result in
some errors (produced by `optimized_mir`) to occur earlier in
compilation than they normally would, during the collection of shims.
Some shims have ids that originate in the standard library, but these
should not receive the PME note, so instances for compiler-generated
functions no longer receive this note.
Signed-off-by: David Wood <david.wood@huawei.com>
|
|
|
|
This reverts commit 8176ab8bc18fdd7d3c2cf7f720c51166364c33a3.
|
|
This reverts commit d0ec85d3fb6d322496cb8f4bc1c21e19f23284ad.
|
|
|
|
|
|
Emit a diagnostic when the monomorphized item collector
encounters errors during a step of the recursive item collection.
These post-monomorphization errors otherwise only show the
erroneous expression without a trace, making them very obscure
and hard to pinpoint whenever they happen in dependencies.
|
|
|
|
non-local cases.
|
|
|
|
|
