[PAC] Encoder and hash (1/8) by jchlanda · Pull Request #159071 · rust-lang/rust

jchlanda · 2026-07-10T11:16:08Z

This patch implements Rust's equivalent of Clang's function pointer type discriminator computation used for pointer authentication. Compatibility with Clang is a primary design goal. For a given extern "C" function type, the discriminator produced by Rust must match the value computed by Clang so that function pointers can be exchanged safely between Rust and C code while preserving pointer authentication semantics.

The implementation mirrors Clang's behavior in ASTContext::encodeTypeForFunctionPointerAuth, ensuring that identical C-compatible function types produce identical discriminators. See: https://clang.llvm.org/doxygen/ASTContext_8cpp.html#abb1375e068e807917527842d05cadea3.

This is part 1 of a sequence of 8 PRs that together implement support for function pointer type discrimination:

Useful links:

Previous PAC work:
- pauthtest introduction: Introduce aarch64-unknown-linux-pauthtest target #155722
- Library support follow up: Library support for aarch64-unknown-linux-pauthtest target #156548
- Config follow up: Pointer authentication config and user facing options #156712
Project goal and tracking issue
Clang's implementation of ASTContext::encodeTypeForFunctionPointerAuth
LLVM's SpiHash

kovdan01

Would it be possible to add unit tests for hash computation like the ones present in llvm source tree at llvm/unittests/Support/SipHashTest.cpp?

View changes since this review

jchlanda · 2026-07-20T11:56:47Z

Would it be possible to add unit tests for hash computation like the ones present in llvm source tree at llvm/unittests/Support/SipHashTest.cpp?

View changes since this review

Done in: 5e05159

rustbot · 2026-07-24T18:07:24Z

Some changes occurred in src/doc/rustc/src/platform-support

cc @Noratrieb

rustbot · 2026-07-24T18:07:27Z

r? @wesleywiser

rustbot has assigned @wesleywiser.
They will have a look at your PR within the next two weeks and either review your PR or reassign to another reviewer.

Use r? to explicitly pick a reviewer

Why was this reviewer chosen?

The reviewer was selected based on:

Owners of files modified in this PR: compiler
compiler expanded to 74 candidates
Random selection from 18 candidates

jchlanda · 2026-07-24T18:07:27Z

r? compiler

This patch implements Rust's equivalent of Clang's function pointer type discriminator computation used in pointer authentication. Compatibility with Clang is a primary goal. The discriminator produced for a given external "C" function type must match the value computed by Clang so that function pointers can be exchanged safely between Rust and C code while preserving pointer authentication semantics. The implementation mirrors Clang's behavior in `ASTContext::encodeTypeForFunctionPointerAuth`, ensuring that identical C-compatible function types produce identical discriminators. See: <https://clang.llvm.org/doxygen/ASTContext_8cpp.html#abb1375e068e807917527842d05cadea3>.

rustbot · 2026-07-27T08:12:18Z

This PR was rebased onto a different main commit. Here's a range-diff highlighting what actually changed.

Rebasing is a normal part of keeping PRs up to date, so no action is needed—this note is just to help reviewers.

davidtwco · 2026-07-27T10:14:28Z

r? @davidtwco

davidtwco · 2026-07-27T10:16:43Z

@@ -0,0 +1,493 @@
+/*!


Use //! so that this is a documentation comment

View changes since the review

davidtwco · 2026-07-27T10:21:34Z

@@ -0,0 +1,142 @@
+// LLVM SipHash-2-4


Why can't we just call LLVM's definitions of these over FFI from rustc_codegen_llvm? That let us avoid the need for this to remain identical to LLVM's definition

View changes since the review

LLVM's implementation is currently a header-only C++ template and does not expose a callable C ABI function. Using it through FFI would require introducing a new wrapper API with a specific interface shape.

Given the comments in code, the header only was a deliberate design decision:

// This is a header-only implementation of SipHash. It lacks library // dependencies so it can be used from LLVM and compiler-rt.

I chose a direct Rust translation because the algorithm is small and stable, and it keeps the discriminator implementation self-contained. The implementation is easy to validate by using the same test vectors as LLVM's implementation.

davidtwco · 2026-07-27T10:22:20Z

+/// This is not a full semantic translation of Rust types. It is a lossy mapping
+/// that intentionally matches Clang's function pointer authentication encoding
+/// rules where Rust has a direct language-level equivalent.


These lines are duplicated

View changes since the review

davidtwco · 2026-07-27T10:24:24Z

+    let ty = canonicalize_c_type(tcx, ty);
+    match ty.kind() {
+        // C void / Rust ()
+        ty::Tuple(list) if list.is_empty() => ClangDiscTy::Void,


_ if ty.is_unit() is clearer

View changes since the review

davidtwco · 2026-07-27T10:29:34Z

+
+    loop {
+        match ty.kind() {
+            ty::Adt(def, args) if tcx.is_diagnostic_item(sym::Option, def.did()) => {


You should use the LangItem::Option, rather than the diagnostic item, given this isn't for a diagnostic.

View changes since the review

Done, had to add #[lang = "Option"] to minicore's Option.

davidtwco · 2026-07-27T10:30:44Z

+
+// Clang disc type.
+#[derive(Debug)]
+enum ClangDiscTy<'tcx> {


I'm not sure what we're gaining from having this intermediate representation - the mapping seems straightforward enough that you could just go direct from a Ty to a encoding

View changes since the review

The main benefit of this separation is to decouple Rust type handling from the encoding logic.

Clang's discriminator computation is based on its own normalized type categories, not on the full C type set. Rust needs to produce identical encodings for equivalent ABI types, so the implementation first normalizes Rust types into the corresponding Clang discriminator categories.

This separation allows the encoding logic to remain aligned with Clang's implementation, while the Rust-specific part is responsible only for the type translation: determining which Rust types map to which Clang discriminator categories. Without it Rust type matching and encoding would be intertwined, making it harder to keep the encoding logic identical between the two compilers.

For example, if Clang extends the encoding in the future (such as adding support for 128-bit types, which is in progress now), we can port the encoding changes directly and only need to extend the Rust type matcher to map the new Rust types into the corresponding Clang categories.

davidtwco · 2026-07-27T10:40:04Z

 All tests from `assembly-llvm`, `codegen-llvm`, `codegen-units`, `coverage`,
-`crashes`, `incremental`, `library`, `mir-opt`, `run-make`, `ui` and
-`ui-fulldeps` subsets are expected to pass.
+`crashes`, `incremental`, `library`, `mir-opt`, `run-make`, `rust_middle` `ui`


Suggested change

`crashes`, `incremental`, `library`, `mir-opt`, `run-make`, `rust_middle` `ui`

`crashes`, `incremental`, `library`, `mir-opt`, `run-make`, `rust_middle`, `ui`

View changes since the review

rustbot · 2026-07-27T10:51:14Z

Reminder, once the PR becomes ready for a review, use @rustbot ready.

jchlanda · 2026-07-28T07:58:46Z

@rustbot ready

scottmcm · 2026-07-28T16:04:01Z

+
+use super::*;
+
+// These tests mirror the SipHash tests from LLVM's Support/SipHashTest.cpp.


Are there licensing implications of this?

View changes since the review

I don't believe this introduces a licensing issue. LLVM appears to be using the reference vectors verbatim from the SipHash test vectors, which originate from: https://github.com/veorq/SipHash/blob/master/vectors.h

The tests here are just a trivial loop over those values to verify the implementation against the known reference outputs.

That said, I am not a lawyer and cannot provide any formal legal guarantee.

jchlanda force-pushed the jakub/pac_ty_disc_PR_1 branch 2 times, most recently from bca55be to 8cf67b9 Compare July 14, 2026 07:26

This comment has been minimized.

Sign in to view

jchlanda force-pushed the jakub/pac_ty_disc_PR_1 branch from 8cf67b9 to 7ee3f2c Compare July 14, 2026 08:51

This comment has been minimized.

Sign in to view

jchlanda force-pushed the jakub/pac_ty_disc_PR_1 branch from 7ee3f2c to fdd1189 Compare July 17, 2026 07:01

This comment has been minimized.

Sign in to view

jchlanda force-pushed the jakub/pac_ty_disc_PR_1 branch from d798449 to 4c1b0ac Compare July 17, 2026 09:46

kovdan01 reviewed Jul 20, 2026

View reviewed changes

jchlanda force-pushed the jakub/pac_ty_disc_PR_1 branch 2 times, most recently from 03337eb to 8b0e0ca Compare July 24, 2026 15:29

jchlanda changed the title ~~[WIP] 1 - Encoder and hash~~ [PAC] Encoder and hash (1/8) Jul 24, 2026

jchlanda marked this pull request as ready for review July 24, 2026 18:07

rustbot added S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. and removed S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. labels Jul 24, 2026

rustbot assigned wesleywiser Jul 24, 2026

rustbot assigned tiif and unassigned wesleywiser Jul 24, 2026

jchlanda force-pushed the jakub/pac_ty_disc_PR_1 branch from 8b0e0ca to 67fd305 Compare July 27, 2026 08:12

rustbot assigned davidtwco and unassigned tiif Jul 27, 2026

davidtwco requested changes Jul 27, 2026

View reviewed changes

rustbot removed the S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. label Jul 27, 2026

rustbot added the S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. label Jul 27, 2026

[PAC] Add unit tests for sip hash

43ed719

jchlanda force-pushed the jakub/pac_ty_disc_PR_1 branch from 67fd305 to 43ed719 Compare July 28, 2026 07:29

rustbot added S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. and removed S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. labels Jul 28, 2026

jchlanda mentioned this pull request Jul 28, 2026

AArch64 Pointer Authentication using aarch64-unknown-linux-pauthtest target on Linux ELF platforms rust-lang/rust-project-goals#618

Open

scottmcm reviewed Jul 28, 2026

View reviewed changes

	`crashes`, `incremental`, `library`, `mir-opt`, `run-make`, `rust_middle` `ui`
	`crashes`, `incremental`, `library`, `mir-opt`, `run-make`, `rust_middle`, `ui`


		use super::*;

		// These tests mirror the SipHash tests from LLVM's Support/SipHashTest.cpp.

Uh oh!

Uh oh!

Conversation

jchlanda commented Jul 10, 2026 • edited by rustbot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

kovdan01 left a comment • edited by rustbot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

jchlanda commented Jul 20, 2026

Uh oh!

rustbot commented Jul 24, 2026

Uh oh!

rustbot commented Jul 24, 2026

Uh oh!

jchlanda commented Jul 24, 2026

Uh oh!

rustbot commented Jul 27, 2026

Uh oh!

davidtwco commented Jul 27, 2026

Uh oh!

davidtwco Jul 27, 2026 • edited by rustbot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

davidtwco Jul 27, 2026 • edited by rustbot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

davidtwco Jul 27, 2026 • edited by rustbot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

davidtwco Jul 27, 2026 • edited by rustbot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

davidtwco Jul 27, 2026 • edited by rustbot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

davidtwco Jul 27, 2026 • edited by rustbot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

davidtwco Jul 27, 2026 • edited by rustbot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

rustbot commented Jul 27, 2026

Uh oh!

jchlanda commented Jul 28, 2026

Uh oh!

scottmcm Jul 28, 2026 • edited by rustbot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Reviewers

jchlanda commented Jul 10, 2026 •

edited by rustbot

Loading

kovdan01 left a comment •

edited by rustbot

Loading

davidtwco Jul 27, 2026 •

edited by rustbot

Loading

davidtwco Jul 27, 2026 •

edited by rustbot

Loading

davidtwco Jul 27, 2026 •

edited by rustbot

Loading

davidtwco Jul 27, 2026 •

edited by rustbot

Loading

davidtwco Jul 27, 2026 •

edited by rustbot

Loading

davidtwco Jul 27, 2026 •

edited by rustbot

Loading

davidtwco Jul 27, 2026 •

edited by rustbot

Loading

scottmcm Jul 28, 2026 •

edited by rustbot

Loading