Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Take priority into account within the pending queue #11032

Merged
merged 2 commits into from
Sep 13, 2022

Conversation

lqd
Copy link
Member

@lqd lqd commented Aug 30, 2022

This is the PR for the work discussed in this zulip thread and whose detailed description and some results are available here with graphs, summaries and raw data -- much of which was shown in the thread as well.

Units of works have a computed priority that is used in the dependency queue, so that higher priorities are dequeued sooner, as documented here.

This PR further applies that principle to the next step before being executed: if multiple pieces of work are waiting in the pending queue, we can sort that according to their priorities. Here as well, higher priorities should be scheduled sooner.

They are more often than not wider than pure chains of dependencies, and this should create more parallelism opportunities earlier in the pipeline: a high priority piece of work represents more future pieces of work down the line, and try to sustain CPU utilization longer (at the potential cost of this parallelism being distributed differently than today, between cargo invoking rustc and rustc's own codegen threads -- when applicable).

This is a scheduling tradeoff that behaves differently for each project, machine configuration, amount of available parallelism at a given point in time, etc, but seems to help more often than hinders: at low-core counts and with enough units of work to be done, so that there is jobserver token contention where choosing a "better" piece of work to work on next may be possible.

There's of course a bit of noise in the results linked above and 800 or so of the most popular crates.io crates is still a limited sample, but they're mostly meant to show a hopefully positive trend: while there are improvements and regressions, that trend looks to be more positive than negative, with the wins being more numerous and with higher amplitudes than the corresponding losses.

(A report on another scheduling experiment -- a superset of this PR, where I also simulate users manually giving a higher priority to syn, quote, serde_derive -- is available here and also improves this PR's results: the regressions are decreased in number and amplitude, whereas the improvements are bigger and more numerous. So that could be further work to iterate upon this one)

Since this mostly applies to clean builds, for low core counts, and with a sufficient number of dependencies to have some items in the pending queue, I feel this also applies well to CI use-cases (esp. on the free tiers).

Somewhat reassuring as well, and discussed in the thread but not in the report: I've also tried to make sure cargo and bootstrapping rustc are not negatively affected. cargo saw some improvements, and bootstrap stayed within today's variance of +/- 2 to 3%. Similarly, since 3y old versions of some tokio crates (0.2.0-alpha.1) were the most negatively affected, I've also checked that recent tokio versions (1.19) are not disproportionately impacted: their simple readme example, the more idiomatic mini-redis sample, and some of my friends' tokio projects were either unaffected or saw some interesting improvements.

And here's a cargo check -j2 graph to liven up this wall of text:

some results of cargo check -j2


I'm not a cargo expert so I'm not sure whether it would be preferable to integrate priorities deeper than just the dependency queue, and e.g. have Units contain a dedicated field or similar. So in the meantime I've done the simplest thing: just sort the pending queue and ask the units' priorities to the dep queue.

We could just as well have the priority recorded as part of the pending queue tuples themselves, or have that be some kind of priority queue/max heap instead of a Vec.

Let me know which you prefer, but it's in essence a very simple change as-is.

@rust-highfive
Copy link

Thanks for the pull request, and welcome! The Rust team is excited to review your changes, and you should hear from @weihanglo (or someone else) soon.

Please see the contribution instructions for more information.

@rust-highfive rust-highfive added the S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. label Aug 30, 2022
@lqd lqd force-pushed the priority_pending_queue branch from 649376e to ce441d4 Compare August 30, 2022 19:09
Copy link
Member

@weihanglo weihanglo left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The change is simple but straightforward. I cannot think of anything blocking this from getting merged. Thank you for putting so much time on these benchmarks!

// If multiple pieces of work are waiting in the pending queue, we can sort it according to
// their priorities: higher priorities should be scheduled sooner.
if self.pending_queue.len() > 1 {
self.pending_queue.sort_by(|(unit_a, _), (unit_b, _)| {
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Would it be better with sort_by_cached_key? Supposed cargo does some hashes here.

(Benchmark needed, though)

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

If we going to do a sort here anyway, then self.queue.dequeue() should not guarantee what order they items are returned in. In fact it should probably just return an iterator.

Also if we sort from lowest priority to highest priority, then when remove an item we can use the .pop() where we are now using self.pending_queue.remove(0);

Copy link
Member Author

@lqd lqd Sep 1, 2022

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@weihanglo Yeah, this is why I also mentioned whether you wanted a different structure where the priority would be part of the data stored in the pending queue.

Ultimately it won't make a huge difference, especially with the allocation involved in caching, but those Units are indeed pretty big and their hashing may not be super cheap, so it should still improve a bit indeed.

I've pushed a change into a sort_by_cached_key call.

Copy link
Member Author

@lqd lqd Sep 1, 2022

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@Eh2406 the sort will only apply if there are units in the pending queue here: so IIUC there can be cases (maybe when there are always jobserver tokens available, or at j1) where the units could be processed one by one, in the order in which they are dequeued from the dependency queue. For these cases, we'd still need to keep the max-priority ordering in self.queue.dequeue(), right ?

I've also pushed your suggestion to pull the high priority items from the end of the queue. (We could probably use the .pop() as a replacement for the !is_empty() check at the top of the loop, but I think the let chains to do that are still unstable)

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Yep. Using something like priority queue might be better. It is somehow a bit weird here
that we push the highest priority item from self.queue.dequeue(), and then immediately sort pending_queue in ascending order.

Anyway, we can always do it in a follow-up PR. I feel like we can merge it as-is. It would be wonderful if you can do another benchmark before r+. Thank you!

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Sure, what kind of benchmark would you like to see ?

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

No particular one. The same benchmark you did is good enough. Just want to leave a trace for how much performance gains Cargo benefits from this commit.

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

It is somehow a bit weird here that we push the highest priority item from self.queue.dequeue(), and then immediately sort pending_queue in ascending order.

It is, and if we want to keep a Vec for now instead of a max-heap, we can insert the unit directly at the correct index instead.

No particular one. The same benchmark you did is good enough

In the report there are 800 benchmarks (× 7 runs × 2 versions of cargo × 3 parallelism settings × 2 build modes), so re-running everything is not going to be possible soon unfortunately because it takes multiple days. I'll re-run some important ones, on a few different machines, e.g. tokio's mini-redis (I've been benchmarking that when pushing to this branch today, and it's still showing -9% at j4/6 for example) and some of the most positively and negatively affected crates.

I'll do those 2 things tomorrow.

@lqd lqd force-pushed the priority_pending_queue branch 3 times, most recently from 3f55dd4 to c685acd Compare September 1, 2022 21:10
lqd added 2 commits September 2, 2022 17:59
If multiple pieces of work are waiting in the pending queue, we can sort it according to
their priorities: higher priorities should be scheduled sooner.

They are more often than not wider than pure chains, and this should create more parallelism
opportunities earlier in the pipeline: a high priority piece of work represents more future
pieces of work down the line.

This is a scheduling tradeoff that behaves differently for each project, machine configuration,
amount of available parallelism at a given point in time, etc, but seems to help more often than
hinders, at low-core counts and with enough units of work to be done, so that there is jobserver
token contention where choosing a "better" piece of work to work on next is possible.
This cleans up the priority-sorted scheduling by removing the need
for a priority accessor that would hash the nodes, and allows inserting
in the queue at the correctly sorted position to remove the insert +
sort combination.
@lqd lqd force-pushed the priority_pending_queue branch from c685acd to 4ffe98a Compare September 2, 2022 21:15
@lqd
Copy link
Member Author

lqd commented Sep 2, 2022

Since we're keeping the pending queue as a vec for now, I've added the priority to its tuples: that removes the need for push/sort/hashing, we can more easily and efficiently insert the items at the correctly sorted position. Let me know if you like it better.

@weihanglo here are some benchmarks you asked for as well: I've re-ran some of the highest relative wins and losses from the report, as well as cargo, and tokio's mini-redis (to contrast with the old version of a tokio crate we see in the losses).

So here are the hyperfine results of 10 clean runs (with 2 additional warmup runs), for check and debug builds,

  • on linux, 24C/48T AMD EPYC 7401P: from j2 to j12
  • on mac, 8C kinda, M1: from j2 to j8
  • on windows, 4C/8T Intel i7 7700: from j2 to j8

for the crates:

(In these results, cargo-sorted is this PR, while cargo-baseline is cargo master from this morning.)

@weihanglo weihanglo added the T-cargo Team: Cargo label Sep 2, 2022
@weihanglo
Copy link
Member

@rfcbot merge

This keeps the pending_queue of unit jobs sorted, in order to make sure cargo always schedules a pending item with the highest priority first. More complete benchmarks are in this zulip discussion. A benchmark for the latest commit 4ffe98a is tracked here.

I feel like this PR just respects the existing priority scheduling mechanism. It doesn't invent anything new, so it might be safe to merge (at least the benchmark result looks generally good).

One other thing to enhance is using some data structure other than Vec for pending_queue. We can try it out within this PR, but I doubg it affects perf much more that this PR so far.

@rfcbot
Copy link
Collaborator

rfcbot commented Sep 2, 2022

Team member @weihanglo has proposed to merge this. The next step is review by the rest of the tagged team members:

No concerns currently listed.

Once a majority of reviewers approve (and at most 2 approvals are outstanding), this will enter its final comment period. If you spot a major issue that hasn't been raised at any point in this process, please speak up!

See this document for info about what commands tagged team members can give me.

@rfcbot rfcbot added proposed-final-comment-period An FCP proposal has started, but not yet signed off. disposition-merge FCP with intent to merge labels Sep 2, 2022
.map(|(key, _)| key.clone())
.max_by_key(|k| self.priority[k])?;
.map(|(key, _)| (key.clone(), self.priority[key]))
.max_by_key(|(_, priority)| *priority)?;
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This max_by_key is unnecessary as we are removing all and then sorting. If we change it to .next(), only the test will need to be updated.

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I can remove it if you want ? It doesn't seem to hurt today, and makes the order at least deterministic.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I got nerd sniped, but it doesn't matter. Leave it bee.

@rfcbot rfcbot added final-comment-period FCP — a period for last comments before action is taken and removed proposed-final-comment-period An FCP proposal has started, but not yet signed off. labels Sep 6, 2022
@rfcbot
Copy link
Collaborator

rfcbot commented Sep 6, 2022

🔔 This is now entering its final comment period, as per the review above. 🔔

@Eh2406
Copy link
Contributor

Eh2406 commented Sep 13, 2022

@bors r+

@bors
Copy link
Contributor

bors commented Sep 13, 2022

📌 Commit 4ffe98a has been approved by Eh2406

It is now in the queue for this repository.

@bors bors added S-waiting-on-bors Status: Waiting on bors to run and complete tests. Bors will change the label on completion. and removed S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. labels Sep 13, 2022
@bors
Copy link
Contributor

bors commented Sep 13, 2022

⌛ Testing commit 4ffe98a with merge 0825039...

@bors
Copy link
Contributor

bors commented Sep 13, 2022

☀️ Test successful - checks-actions
Approved by: Eh2406
Pushing 0825039 to master...

@bors bors merged commit 0825039 into rust-lang:master Sep 13, 2022
@lqd lqd deleted the priority_pending_queue branch September 13, 2022 18:46
weihanglo added a commit to weihanglo/rust that referenced this pull request Sep 13, 2022
10 commits in 646e9a0b9ea8354cc409d05f10e8dc752c5de78e..082503982ea0fb7a8fd72210427d43a2e2128a63
2022-09-02 14:29:28 +0000 to 2022-09-13 17:49:38 +0000
- Take priority into account within the pending queue (rust-lang/cargo#11032)
- fix(add): Clarify which version the features are added for (rust-lang/cargo#11075)
- doc: clarify config-relative paths for `--config <path>` (rust-lang/cargo#11079)
- Do not add home bin path to PATH if it's already there (rust-lang/cargo#11023)
- Don't use `for` loop on an `Option` (rust-lang/cargo#11081)
- Remove dead code (rust-lang/cargo#11080)
- Change progress indicator for sparse registries (rust-lang/cargo#11068)
- chore(ci): Ensure intradoc links are valid (rust-lang/cargo#11055)
- Cache index files based on contents hash (rust-lang/cargo#11044)
- fix: specifies the max length for crate name (rust-lang/cargo#11051)
bors added a commit to rust-lang-ci/rust that referenced this pull request Sep 13, 2022
Update cargo

10 commits in 646e9a0b9ea8354cc409d05f10e8dc752c5de78e..082503982ea0fb7a8fd72210427d43a2e2128a63 2022-09-02 14:29:28 +0000 to 2022-09-13 17:49:38 +0000
- Take priority into account within the pending queue (rust-lang/cargo#11032)
- fix(add): Clarify which version the features are added for (rust-lang/cargo#11075)
- doc: clarify config-relative paths for `--config <path>` (rust-lang/cargo#11079)
- Do not add home bin path to PATH if it's already there (rust-lang/cargo#11023)
- Don't use `for` loop on an `Option` (rust-lang/cargo#11081)
- Remove dead code (rust-lang/cargo#11080)
- Change progress indicator for sparse registries (rust-lang/cargo#11068)
- chore(ci): Ensure intradoc links are valid (rust-lang/cargo#11055)
- Cache index files based on contents hash (rust-lang/cargo#11044)
- fix: specifies the max length for crate name (rust-lang/cargo#11051)
@rfcbot rfcbot added finished-final-comment-period FCP complete to-announce and removed final-comment-period FCP — a period for last comments before action is taken labels Sep 16, 2022
@ehuss ehuss added this to the 1.65.0 milestone Sep 21, 2022
@SimonSapin
Copy link
Contributor

How are priorities determined? Can end users influence them?

@weihanglo
Copy link
Member

weihanglo commented Oct 31, 2022

Can end users influence them?

Unfortunately, no.
It is under consideration but may need more experiments. For today, you could use cargo build --timings to find the bottleneck of your build. And then prioritize manually by building that crate in advance via cargo build -p <slow-crate>.

How are priorities determined?

Still naive. Basically, every unit gets a same fixed cost1, and the priority of each is a sum of all the cost of its children plus its own2. We're welcome for more real world data and experiments!

Footnotes

  1. https://github.com/rust-lang/cargo/blob/2d04bcd1b0d6f40001a7b1762f78842851d0b334/src/cargo/core/compiler/job_queue.rs#L534-L538

  2. https://github.com/rust-lang/cargo/blob/996a6363ce4b9109d4ca757407dd6dcb4805c86f/src/cargo/util/dependency_queue.rs#L105-L112

wip-sync pushed a commit to NetBSD/pkgsrc-wip that referenced this pull request Nov 27, 2022
Pkgsrc changes:
 * We now manage to build for mipsel-unknown-netbsd, but despite the
   target spec saying cpu = "mips3", the compiler manages to emit 64-bit
   instructions which cause "illegal instruction" error.  Will need more
   work.  The mipsel-unknown-netbsd entry is commentd out since there
   is no 1.64.0 bootstrap.
 * Managed to retain the build of aarch64_be, llvm needed a patch to
   avoid use of neon instructions in the BE case (llvm doesn't support
   use of neon in BE mode).  Ref. patch to
   src/llvm-project/llvm/lib/Support/BLAKE3/blake3_impl.h.
 * The minimum gcc version is now 7.x, and that includes the cross-compiler
   for the targets.  For i386 this also needs to /usr/include/gcc-7 include
   files in the target root, because immintrin.h from gcc 5 is not
   compatible with gcc 7.x.  This applies for the targets where we build
   against a root from netbsd-8 (sparc64, powerpc, i386), and files/gcc-wrap
   gets a hack for this.
 * Pick up tweak for -latomic inclusion from
   rust-lang/rust#104220
   and
   rust-lang/rust#104572
 * Retain ability to do 32-bit NetBSD, by changing from 64 to 32 bit
   types in library/std/src/sys/unix/thread_parker/netbsd.rs.
 * I've struggled a bit to get the "openssl-src" build with -latomic
   where it's needed.  I introduce "NetBSD-generic32" system type and
   use it for the NetBSD mipsel target.  There is another attempt to
   do the same in the patch to vendor/openssl-sys/build/main.rs.
 * Bump bootstraps to 1.64.0, checksum updates.

Upstream changes:

Version 1.65.0 (2022-11-03)
==========================

Language
--------
- [Error on `as` casts of enums with `#[non_exhaustive]` variants]
  (rust-lang/rust#92744)
- [Stabilize `let else`](rust-lang/rust#93628)
- [Stabilize generic associated types (GATs)]
  (rust-lang/rust#96709)
- [Add lints `let_underscore_drop`, `let_underscore_lock`, and
  `let_underscore_must_use` from Clippy]
  (rust-lang/rust#97739)
- [Stabilize `break`ing from arbitrary labeled blocks ("label-break-value")]
  (rust-lang/rust#99332)
- [Uninitialized integers, floats, and raw pointers are now considered
  immediate UB](rust-lang/rust#98919).
  Usage of `MaybeUninit` is the correct way to work with uninitialized
  memory.
- [Stabilize raw-dylib for Windows x86_64, aarch64, and thumbv7a]
  (rust-lang/rust#99916)
- [Do not allow `Drop` impl on foreign ADTs]
  (rust-lang/rust#99576)

Compiler
--------
- [Stabilize -Csplit-debuginfo on Linux]
  (rust-lang/rust#98051)
- [Use niche-filling optimization even when multiple variants have
  data] (rust-lang/rust#94075)
- [Associated type projections are now verified to be well-formed
  prior to resolving the underlying type]
  (rust-lang/rust#99217)
- [Stringify non-shorthand visibility correctly]
  (rust-lang/rust#100350)
- [Normalize struct field types when unsizing]
  (rust-lang/rust#101831)
- [Update to LLVM 15](rust-lang/rust#99464)
- [Fix aarch64 call abi to correctly zeroext when needed]
  (rust-lang/rust#97800)
- [debuginfo: Generalize C++-like encoding for enums]
  (rust-lang/rust#98393)
- [Add `special_module_name` lint]
  (rust-lang/rust#94467)
- [Add support for generating unique profraw files by default when
  using `-C instrument-coverage`]
  (rust-lang/rust#100384)
- [Allow dynamic linking for iOS/tvOS targets]
  (rust-lang/rust#100636)

New targets:
- [Add armv4t-none-eabi as a tier 3 target]
  (rust-lang/rust#100244)
- [Add powerpc64-unknown-openbsd and riscv64-unknown-openbsd as tier 3 targets]
  (rust-lang/rust#101025)
- Refer to Rust's [platform support page][platform-support-doc] for more
  information on Rust's tiered platform support.

Libraries
---------
- [Don't generate `PartialEq::ne` in derive(PartialEq)]
  (rust-lang/rust#98655)
- [Windows RNG: Use `BCRYPT_RNG_ALG_HANDLE` by default]
  (rust-lang/rust#101325)
- [Forbid mixing `System` with direct system allocator calls]
  (rust-lang/rust#101394)
- [Document no support for writing to non-blocking stdio/stderr]
  (rust-lang/rust#101416)
- [`std::layout::Layout` size must not overflow `isize::MAX` when
  rounded up to `align`](rust-lang/rust#95295)
  This also changes the safety conditions on
  `Layout::from_size_align_unchecked`.

Stabilized APIs
---------------
- [`std::backtrace::Backtrace`]
  (https://doc.rust-lang.org/stable/std/backtrace/struct.Backtrace.html)
- [`Bound::as_ref`]
  (https://doc.rust-lang.org/stable/std/ops/enum.Bound.html#method.as_ref)
- [`std::io::read_to_string`]
  (https://doc.rust-lang.org/stable/std/io/fn.read_to_string.html)
- [`<*const T>::cast_mut`]
  (https://doc.rust-lang.org/stable/std/primitive.pointer.html#method.cast_mut)
- [`<*mut T>::cast_const`]
  (https://doc.rust-lang.org/stable/std/primitive.pointer.html#method.cast_const)

These APIs are now stable in const contexts:
- [`<*const T>::offset_from`]
  (https://doc.rust-lang.org/stable/std/primitive.pointer.html#method.offset_from)
- [`<*mut T>::offset_from`]
  (https://doc.rust-lang.org/stable/std/primitive.pointer.html#method.offset_from)

Cargo
-----
- [Apply GitHub fast path even for partial hashes]
  (rust-lang/cargo#10807)
- [Do not add home bin path to PATH if it's already there]
  (rust-lang/cargo#11023)
- [Take priority into account within the pending queue]
  (rust-lang/cargo#11032).
  This slightly optimizes job scheduling by Cargo, with typically
  small improvements on larger crate graph builds.

Compatibility Notes
-------------------
- [`std::layout::Layout` size must not overflow `isize::MAX` when
  rounded up to `align`] (rust-lang/rust#95295).
  This also changes the safety conditions on
  `Layout::from_size_align_unchecked`.
- [`PollFn` now only implements `Unpin` if the closure is `Unpin`]
  (rust-lang/rust#102737).
  This is a possible breaking change if users were relying on the
  blanket unpin implementation.  See discussion on the PR for
  details of why this change was made.
- [Drop ExactSizeIterator impl from std::char::EscapeAscii]
  (rust-lang/rust#99880)
  This is a backwards-incompatible change to the standard library's
  surface area, but is unlikely to affect real world usage.
- [Do not consider a single repeated lifetime eligible for elision
  in the return type] (rust-lang/rust#103450)
  This behavior was unintentionally changed in 1.64.0, and this
  release reverts that change by making this an error again.
- [Reenable disabled early syntax gates as future-incompatibility
  lints] (rust-lang/rust#99935)
- [Update the minimum external LLVM to 13]
  (rust-lang/rust#100460)
- [Don't duplicate file descriptors into stdio fds]
  (rust-lang/rust#101426)
- [Sunset RLS](rust-lang/rust#100863)
- [Deny usage of `#![cfg_attr(..., crate_type = ...)]` to set the
  crate type] (rust-lang/rust#99784)
  This strengthens the forward compatibility lint
  deprecated_cfg_attr_crate_type_name to deny.
- [`llvm-has-rust-patches` allows setting the build system to treat
  the LLVM as having Rust-specific patches]
  (rust-lang/rust#101072)
  This option may need to be set for distributions that are building
  Rust with a patched LLVM via `llvm-config`, not the built-in
  LLVM.

Internal Changes
----------------

These changes do not affect any public interfaces of Rust, but they represent
significant improvements to the performance or internals of rustc and related
tools.

- [Add `x.sh` and `x.ps1` shell scripts]
  (rust-lang/rust#99992)
- [compiletest: use target cfg instead of hard-coded tables]
  (rust-lang/rust#100260)
- [Use object instead of LLVM for reading bitcode from rlibs]
  (rust-lang/rust#98100)
- [Enable MIR inlining for optimized compilations]
  (rust-lang/rust#91743)
  This provides a 3-10% improvement in compiletimes for real world
  crates. See [perf results]
  (https://perf.rust-lang.org/compare.html?start=aedf78e56b2279cc869962feac5153b6ba7001ed&end=0075bb4fad68e64b6d1be06bf2db366c30bc75e1&stat=instructions:u).
netbsd-srcmastr pushed a commit to NetBSD/pkgsrc that referenced this pull request Jan 23, 2023
Pkgsrc changes:
 * pkglint cleanups, bump bootstrap kits to 1.65.0.
 * New target: mipsel-unknown-netbsd, for cpu=mips32 with soft-float.
 * Managed to retain the build of aarch64_be, llvm needed a patch to
   avoid use of neon instructions in the BE case (llvm doesn't support
   use of neon in BE mode).  Ref. patch to
   src/llvm-project/llvm/lib/Support/BLAKE3/blake3_impl.h.
   Also submitted upstream of LLVM to the BLAKE3 maintainers.
 * The minimum gcc version is now 7.x, and that includes the
   cross-compiler for the targets.  For i386 this also needs to
   /usr/include/gcc-7 include files in the target root, because
   immintrin.h from gcc 5 is not compatible with gcc 7.x.  This
   applies for the targets where we build against a root from netbsd-8
   (sparc64, powerpc, i386), and files/gcc-wrap gets a hack for this.
 * Pick up tweak for -latomic inclusion from
   rust-lang/rust#104220
   and
   rust-lang/rust#104572
 * Retain ability to do 32-bit NetBSD, by changing from 64 to 32 bit
   types in library/std/src/sys/unix/thread_parker/netbsd.rs.
 * I've tried to get the "openssl-src" build with -latomic where it's
   needed.  I've introduced the "NetBSD-generic32" system type and use
   it for the NetBSD mipsel target.  There is another attempt to do
   the same in the patch to vendor/openssl-sys/build/main.rs.


Upstream changes:

Version 1.66.1 (2023-01-10)
===========================

- Added validation of SSH host keys for git URLs in Cargo
  ([CVE-2022-46176](https://www.cve.org/CVERecord?id=CVE-2022-46176))


Version 1.66.0 (2022-12-15)
===========================

Language
--------
- [Permit specifying explicit discriminants on all `repr(Int)`
  enums](rust-lang/rust#95710)
  ```rust
  #[repr(u8)]
  enum Foo {
      A(u8) = 0,
      B(i8) = 1,
      C(bool) = 42,
  }
  ```
- [Allow transmutes between the same type differing only in
  lifetimes](rust-lang/rust#101520)
- [Change constant evaluation errors from a deny-by-default lint to a
  hard error](rust-lang/rust#102091)
- [Trigger `must_use` on `impl Trait` for
  supertraits](rust-lang/rust#102287) This
  makes `impl ExactSizeIterator` respect the existing `#[must_use]`
  annotation on `Iterator`.
- [Allow `..X` and `..=X` in
  patterns](rust-lang/rust#102275)
- [Uplift `clippy::for_loops_over_fallibles` lint into
  rustc](rust-lang/rust#99696)
- [Stabilize `sym` operands in inline
  assembly](rust-lang/rust#103168)
- [Update to Unicode 15](rust-lang/rust#101912)
- [Opaque types no longer imply lifetime
  bounds](rust-lang/rust#95474) This is a
  soundness fix which may break code that was erroneously relying on this
  behavior.

Compiler
--------
- [Add armv5te-none-eabi and thumbv5te-none-eabi tier 3
  targets](rust-lang/rust#101329)
  - Refer to Rust's [platform support page][platform-support-doc] for
    more information on Rust's tiered platform support.
- [Add support for linking against macOS universal
  libraries](rust-lang/rust#98736)

Libraries
---------
- [Fix `#[derive(Default)]` on a generic `#[default]` enum adding
  unnecessary `Default`
  bounds](rust-lang/rust#101040)
- [Update to Unicode 15](rust-lang/rust#101821)

Stabilized APIs
---------------
- [`proc_macro::Span::source_text`](https://doc.rust-lang.org/stable/proc_macro/struct.Span.html#method.source_text)
- [`uX::{checked_add_signed, overflowing_add_signed,
  saturating_add_signed,
  wrapping_add_signed}`](https://doc.rust-lang.org/stable/std/primitive.u8.html#method.checked_add_signed)
- [`iX::{checked_add_unsigned, overflowing_add_unsigned,
  saturating_add_unsigned,
  wrapping_add_unsigned}`](https://doc.rust-lang.org/stable/std/primitive.i8.html#method.checked_add_unsigned)
- [`iX::{checked_sub_unsigned, overflowing_sub_unsigned,
  saturating_sub_unsigned,
  wrapping_sub_unsigned}`](https://doc.rust-lang.org/stable/std/primitive.i8.html#method.checked_sub_unsigned)
- [`BTreeSet::{first, last, pop_first,
  pop_last}`](https://doc.rust-lang.org/stable/std/collections/struct.BTreeSet.html#method.first)
- [`BTreeMap::{first_key_value, last_key_value, first_entry, last_entry,
  pop_first,
  pop_last}`](https://doc.rust-lang.org/stable/std/collections/struct.BTreeMap.html#method.first_key_value)
- [Add `AsFd` implementations for stdio lock types on
  WASI.](rust-lang/rust#101768)
- [`impl TryFrom<Vec<T>> for Box<[T;
  N]>`](https://doc.rust-lang.org/stable/std/boxed/struct.Box.html#impl-TryFrom%3CVec%3CT%2C%20Global%3E%3E-for-Box%3C%5BT%3B%20N%5D%2C%20Global%3E)
- [`core::hint::black_box`](https://doc.rust-lang.org/stable/std/hint/fn.black_box.html)
- [`Duration::try_from_secs_{f32,f64}`](https://doc.rust-lang.org/stable/std/time/struct.Duration.html#method.try_from_secs_f32)
- [`Option::unzip`](https://doc.rust-lang.org/stable/std/option/enum.Option.html#method.unzip)
- [`std::os::fd`](https://doc.rust-lang.org/stable/std/os/fd/index.html)

Rustdoc
-------
- [Add Rustdoc warning for invalid HTML tags in the
  documentation](rust-lang/rust#101720)

Cargo
-----
- [Added `cargo remove` to remove dependencies from
  Cargo.toml](https://doc.rust-lang.org/nightly/cargo/commands/cargo-remove.html)
- [`cargo publish` now waits for the new version to be downloadable
  before exiting](rust-lang/cargo#11062)

See [detailed release notes](https://github.com/rust-lang/cargo/blob/master/CHANGELOG.md#cargo-166-2022-12-15) for more.

Compatibility Notes
-------------------
- [Only apply `ProceduralMasquerade` hack to older versions of
  `rental`](rust-lang/rust#94063)
- [Don't export `__heap_base` and `__data_end` on
  wasm32-wasi.](rust-lang/rust#102385)
- [Don't export `__wasm_init_memory` on
  WebAssembly.](rust-lang/rust#102426)
- [Only export `__tls_*` on
  wasm32-unknown-unknown.](rust-lang/rust#102440)
- [Don't link to `libresolv` in libstd on
  Darwin](rust-lang/rust#102766)
- [Update libstd's libc to 0.2.135 (to make `libstd` no longer pull in
  `libiconv.dylib` on
  Darwin)](rust-lang/rust#103277)
- [Opaque types no longer imply lifetime
  bounds](rust-lang/rust#95474)
  This is a soundness fix which may break code that was erroneously
  relying on this behavior.
- [Make `order_dependent_trait_objects` show up in future-breakage
  reports](rust-lang/rust#102635)
- [Change std::process::Command spawning to default to inheriting the
  parent's signal mask](rust-lang/rust#101077)

Internal Changes
----------------

These changes do not affect any public interfaces of Rust, but they
represent significant improvements to the performance or internals of
rustc and related tools.

- [Enable BOLT for LLVM
  compilation](rust-lang/rust#94381)
- [Enable LTO for
  rustc_driver.so](rust-lang/rust#101403)


Version 1.65.0 (2022-11-03)
==========================

Language
--------
- [Error on `as` casts of enums with `#[non_exhaustive]` variants]
  (rust-lang/rust#92744)
- [Stabilize `let else`](rust-lang/rust#93628)
- [Stabilize generic associated types (GATs)]
  (rust-lang/rust#96709)
- [Add lints `let_underscore_drop`, `let_underscore_lock`, and
  `let_underscore_must_use` from Clippy]
  (rust-lang/rust#97739)
- [Stabilize `break`ing from arbitrary labeled blocks ("label-break-value")]
  (rust-lang/rust#99332)
- [Uninitialized integers, floats, and raw pointers are now considered
  immediate UB](rust-lang/rust#98919).
  Usage of `MaybeUninit` is the correct way to work with uninitialized
  memory.
- [Stabilize raw-dylib for Windows x86_64, aarch64, and thumbv7a]
  (rust-lang/rust#99916)
- [Do not allow `Drop` impl on foreign ADTs]
  (rust-lang/rust#99576)

Compiler
--------
- [Stabilize -Csplit-debuginfo on Linux]
  (rust-lang/rust#98051)
- [Use niche-filling optimization even when multiple variants have
  data] (rust-lang/rust#94075)
- [Associated type projections are now verified to be well-formed
  prior to resolving the underlying type]
  (rust-lang/rust#99217)
- [Stringify non-shorthand visibility correctly]
  (rust-lang/rust#100350)
- [Normalize struct field types when unsizing]
  (rust-lang/rust#101831)
- [Update to LLVM 15](rust-lang/rust#99464)
- [Fix aarch64 call abi to correctly zeroext when needed]
  (rust-lang/rust#97800)
- [debuginfo: Generalize C++-like encoding for enums]
  (rust-lang/rust#98393)
- [Add `special_module_name` lint]
  (rust-lang/rust#94467)
- [Add support for generating unique profraw files by default when
  using `-C instrument-coverage`]
  (rust-lang/rust#100384)
- [Allow dynamic linking for iOS/tvOS targets]
  (rust-lang/rust#100636)

New targets:
- [Add armv4t-none-eabi as a tier 3 target]
  (rust-lang/rust#100244)
- [Add powerpc64-unknown-openbsd and riscv64-unknown-openbsd as tier 3 targets]
  (rust-lang/rust#101025)
- Refer to Rust's [platform support page][platform-support-doc] for more
  information on Rust's tiered platform support.

Libraries
---------
- [Don't generate `PartialEq::ne` in derive(PartialEq)]
  (rust-lang/rust#98655)
- [Windows RNG: Use `BCRYPT_RNG_ALG_HANDLE` by default]
  (rust-lang/rust#101325)
- [Forbid mixing `System` with direct system allocator calls]
  (rust-lang/rust#101394)
- [Document no support for writing to non-blocking stdio/stderr]
  (rust-lang/rust#101416)
- [`std::layout::Layout` size must not overflow `isize::MAX` when
  rounded up to `align`](rust-lang/rust#95295)
  This also changes the safety conditions on
  `Layout::from_size_align_unchecked`.

Stabilized APIs
---------------
- [`std::backtrace::Backtrace`]
  (https://doc.rust-lang.org/stable/std/backtrace/struct.Backtrace.html)
- [`Bound::as_ref`]
  (https://doc.rust-lang.org/stable/std/ops/enum.Bound.html#method.as_ref)
- [`std::io::read_to_string`]
  (https://doc.rust-lang.org/stable/std/io/fn.read_to_string.html)
- [`<*const T>::cast_mut`]
  (https://doc.rust-lang.org/stable/std/primitive.pointer.html#method.cast_mut)
- [`<*mut T>::cast_const`]
  (https://doc.rust-lang.org/stable/std/primitive.pointer.html#method.cast_const)

These APIs are now stable in const contexts:
- [`<*const T>::offset_from`]
  (https://doc.rust-lang.org/stable/std/primitive.pointer.html#method.offset_from)
- [`<*mut T>::offset_from`]
  (https://doc.rust-lang.org/stable/std/primitive.pointer.html#method.offset_from)

Cargo
-----
- [Apply GitHub fast path even for partial hashes]
  (rust-lang/cargo#10807)
- [Do not add home bin path to PATH if it's already there]
  (rust-lang/cargo#11023)
- [Take priority into account within the pending queue]
  (rust-lang/cargo#11032).
  This slightly optimizes job scheduling by Cargo, with typically
  small improvements on larger crate graph builds.

Compatibility Notes
-------------------
- [`std::layout::Layout` size must not overflow `isize::MAX` when
  rounded up to `align`] (rust-lang/rust#95295).
  This also changes the safety conditions on
  `Layout::from_size_align_unchecked`.
- [`PollFn` now only implements `Unpin` if the closure is `Unpin`]
  (rust-lang/rust#102737).
  This is a possible breaking change if users were relying on the
  blanket unpin implementation.  See discussion on the PR for
  details of why this change was made.
- [Drop ExactSizeIterator impl from std::char::EscapeAscii]
  (rust-lang/rust#99880)
  This is a backwards-incompatible change to the standard library's
  surface area, but is unlikely to affect real world usage.
- [Do not consider a single repeated lifetime eligible for elision
  in the return type] (rust-lang/rust#103450)
  This behavior was unintentionally changed in 1.64.0, and this
  release reverts that change by making this an error again.
- [Reenable disabled early syntax gates as future-incompatibility
  lints] (rust-lang/rust#99935)
- [Update the minimum external LLVM to 13]
  (rust-lang/rust#100460)
- [Don't duplicate file descriptors into stdio fds]
  (rust-lang/rust#101426)
- [Sunset RLS](rust-lang/rust#100863)
- [Deny usage of `#![cfg_attr(..., crate_type = ...)]` to set the
  crate type] (rust-lang/rust#99784)
  This strengthens the forward compatibility lint
  deprecated_cfg_attr_crate_type_name to deny.
- [`llvm-has-rust-patches` allows setting the build system to treat
  the LLVM as having Rust-specific patches]
  (rust-lang/rust#101072)
  This option may need to be set for distributions that are building
  Rust with a patched LLVM via `llvm-config`, not the built-in
  LLVM.

Internal Changes
----------------

These changes do not affect any public interfaces of Rust, but they represent
significant improvements to the performance or internals of rustc and related
tools.

- [Add `x.sh` and `x.ps1` shell scripts]
  (rust-lang/rust#99992)
- [compiletest: use target cfg instead of hard-coded tables]
  (rust-lang/rust#100260)
- [Use object instead of LLVM for reading bitcode from rlibs]
  (rust-lang/rust#98100)
- [Enable MIR inlining for optimized compilations]
  (rust-lang/rust#91743)
  This provides a 3-10% improvement in compiletimes for real world
  crates. See [perf results]
  (https://perf.rust-lang.org/compare.html?start=aedf78e56b2279cc869962feac5153b6ba7001ed&end=0075bb4fad68e64b6d1be06bf2db366c30bc75e1&stat=instructions:u).
RalfJung pushed a commit to RalfJung/rust-analyzer that referenced this pull request Apr 20, 2024
Update cargo

10 commits in 646e9a0b9ea8354cc409d05f10e8dc752c5de78e..082503982ea0fb7a8fd72210427d43a2e2128a63 2022-09-02 14:29:28 +0000 to 2022-09-13 17:49:38 +0000
- Take priority into account within the pending queue (rust-lang/cargo#11032)
- fix(add): Clarify which version the features are added for (rust-lang/cargo#11075)
- doc: clarify config-relative paths for `--config <path>` (rust-lang/cargo#11079)
- Do not add home bin path to PATH if it's already there (rust-lang/cargo#11023)
- Don't use `for` loop on an `Option` (rust-lang/cargo#11081)
- Remove dead code (rust-lang/cargo#11080)
- Change progress indicator for sparse registries (rust-lang/cargo#11068)
- chore(ci): Ensure intradoc links are valid (rust-lang/cargo#11055)
- Cache index files based on contents hash (rust-lang/cargo#11044)
- fix: specifies the max length for crate name (rust-lang/cargo#11051)
RalfJung pushed a commit to RalfJung/rust-analyzer that referenced this pull request Apr 27, 2024
Update cargo

10 commits in 646e9a0b9ea8354cc409d05f10e8dc752c5de78e..082503982ea0fb7a8fd72210427d43a2e2128a63 2022-09-02 14:29:28 +0000 to 2022-09-13 17:49:38 +0000
- Take priority into account within the pending queue (rust-lang/cargo#11032)
- fix(add): Clarify which version the features are added for (rust-lang/cargo#11075)
- doc: clarify config-relative paths for `--config <path>` (rust-lang/cargo#11079)
- Do not add home bin path to PATH if it's already there (rust-lang/cargo#11023)
- Don't use `for` loop on an `Option` (rust-lang/cargo#11081)
- Remove dead code (rust-lang/cargo#11080)
- Change progress indicator for sparse registries (rust-lang/cargo#11068)
- chore(ci): Ensure intradoc links are valid (rust-lang/cargo#11055)
- Cache index files based on contents hash (rust-lang/cargo#11044)
- fix: specifies the max length for crate name (rust-lang/cargo#11051)
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
disposition-merge FCP with intent to merge finished-final-comment-period FCP complete S-waiting-on-bors Status: Waiting on bors to run and complete tests. Bors will change the label on completion. T-cargo Team: Cargo
Projects
None yet
Development

Successfully merging this pull request may close these issues.

9 participants