[WIP] 4 - Static allocs by jchlanda · Pull Request #159081 · rust-lang/rust

jchlanda · 2026-07-10T13:16:15Z

The codegen now walks the layout of statically initialized types to find extern "C" function pointer fields, computes their type discriminators, and applies those discriminators when emitting authenticated function pointer relocations.

Also make sure that type discrimination is never applied to init/fini entries.

This is part 4 of a sequence of 8 PRs, that together aim to bring function pointer type discrimination support:

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

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>.

kovdan01 · 2026-07-20T10:10:10Z

+/// emitted at that offset.
+///
+/// Offsets are relative to the beginning of the allocation.
+pub(crate) struct FnPtrDiscriminatorAtOffset {


What is the established naming convention in rust for such kind of things? Should this be plural (discriminators)? Especially given the fact that we already have collect_fn_ptr_discriminators

View changes since the review

Sure, FnPtrDiscriminatorsAtOffset it is.

kovdan01 · 2026-07-20T10:16:37Z

+            let variant = def.non_enum_variant();
+
+            let Some((_, field)) = variant.fields.iter_enumerated().next() else {
+                return;


Could you please refer me to tests covering these early-return paths (from this case and from the below cases)? The logic by itself looks sound, but it definitely deserves its own test coverage (which I'm not sure I can correctly identify by myself in the new test suite from PR 8)

View changes since the review

Interesting, this is actually redundant. I initially though it would serve for something like:

#[repr(transparent)] struct Wrapper(extern "C" fn(i32)); impl Sync for Wrapper {} static T_WRAPPED_FN_PTR: Wrapper = Wrapper(g);

in https://github.com/jchlanda/rust/blob/fb279532dbf253c1f78dd02a3268bfdde8d6d7ad/tests/codegen-llvm/pauth/pauth-fn-ptr-type-discrimination-struct-members.rs#L1, but turns out struct arm handles it just fine. Removing.

While at it, I added a test case for Tupel which was missing (T_TUPLE, T_TUPLE_2 and test_8_tuple_members in the same file). Tests added to the 8th PR.

Thanks!

kovdan01 · 2026-07-22T09:38:46Z

+            .as_ref()
+            .and_then(|m| m.map.get(&Size::from_bytes(offset as u64)));
+
+        // Init/fini entries must not participate in function pointer type discrimination.


Nit: maybe worth slightly re-phrasing this - now it might be understood as "init/fini pointers are zero-discriminated", while in reality they receive their own special constant discriminator

View changes since the review

kovdan01 · 2026-07-22T09:42:12Z

        // FIXME: should we cache `const_alloc_to_llvm` to avoid repeating this for the
        // same `ConstAllocation`?
-        let cv = const_alloc_to_llvm(self, alloc.inner(), IsStatic::No, IsInitOrFini::No);
+        let cv = const_alloc_to_llvm(self, alloc.inner(), IsStatic::No, IsInitOrFini::No, None);


Could you please explain why fn_ptr_discriminators is guaranteed to be None here? Is ConstAllocation something which is guaranteed not to contain nested function pointers?

View changes since the review

Looked at this code again, and I think this is incorrect. Type discrimination should also be applied when computing addresses for statics; it just didn't come up in the existing tests.

From a quick look, this seems a little more involved because we don't have the necessary type information when calling static_addr_of. We'll likely need to either extend its API or compute the discriminator map at the call sites that still have access to the Rust type.

I've added an in-code FIXME to capture this, and if you're okay with it, I'd like to address it in a follow-up PR.

kovdan01 · 2026-07-22T09:46:56Z

    alloc: &Allocation,
    is_static: IsStatic,
    is_init_fini: IsInitOrFini,
+    fn_ptr_discriminators: Option<&FnPtrDiscriminatorsAtOffset>,


Nit: I'm not sure of a particular suggestion, but just as a thought - it might be worth somehow prefixing such things with ptrauth_ in order to (a) indicate that it's ptrauth-specific for readers wondering what's it all about (b) make it easier to grep ptrauth-related stuff.

If just adding prefix right now, the identifier would become too long though. So maybe the rest of the name could be somehow reconsidered.

I would appreciate if you could do a pass over all newly added data structures/variables/argument names which exist in "generic" non-ptrauth-specific context and ensure if in all such places it's clear that the corresponding item is ptrauth-specific

View changes since the review

For new fields that are extending already existing data structures, I'll add the ptrauth_ prefix.

I'll also do a follow-up PR to take another look at session.rs, as it doesn't really belong in this patch series. At that point I'll also review the names of the data structures/functions we're introducing. I agree that public-facing identifiers should be prefixed, but for local, throwaway-style types/funcs I'm less convinced. Since they're context-specific, they should be easier to reason about without additional prefixes.

This patch introduces the following: * Extends `FnAbi` (`callconv`) with a `ptrauth_type_discriminator` field. This field is only used when emitting pointer authentication call bundles. It is stored in `FnAbi` because the call site is not guaranteed to have access to an `Instance`, so the discriminator cannot always be computed on demand. * Adds support for `llvm.ptrauth.resign`. This intrinsic will be used when support for semantic transmute is added. * Performs a minor API redesign as groundwork for allowing call sites to modify schemas in place.

Also remove error messages/tests that used to guarded it.

The codegen now walks the layout of static initializer types to find extern "C" function pointer fields, computes their type discriminators, and applies those discriminators when emitting authenticated function pointer relocations. Also make sure that type discrimination is never applied to init/fini entries.

jchlanda force-pushed the jakub/pac_ty_disc_PR_4 branch 2 times, most recently from 9052b2a to 5b223de Compare July 14, 2026 07:27

This comment has been minimized.

Sign in to view

jchlanda force-pushed the jakub/pac_ty_disc_PR_4 branch from 5b223de to 45f422b Compare July 14, 2026 08:53

This comment has been minimized.

Sign in to view

jchlanda force-pushed the jakub/pac_ty_disc_PR_4 branch from 45f422b to 8659b72 Compare July 17, 2026 07:04

This comment has been minimized.

Sign in to view

jchlanda force-pushed the jakub/pac_ty_disc_PR_4 branch from 8659b72 to 8392cdf Compare July 17, 2026 08:21

jchlanda force-pushed the jakub/pac_ty_disc_PR_4 branch from 8392cdf to 5399226 Compare July 17, 2026 09:48

kovdan01 reviewed Jul 20, 2026

View reviewed changes

jchlanda force-pushed the jakub/pac_ty_disc_PR_4 branch 2 times, most recently from ef6b34d to e08ffe7 Compare July 20, 2026 15:19

kovdan01 reviewed Jul 22, 2026

View reviewed changes

jchlanda added 4 commits July 22, 2026 12:20

[PAC] Add unit tests for sip hash

03337eb

[PAC] Enable support for FPTR_TYPE_DISCR in ABI Version

af72799

Also remove error messages/tests that used to guarded it.

jchlanda force-pushed the jakub/pac_ty_disc_PR_4 branch from e08ffe7 to f77edea Compare July 22, 2026 12:41

Uh oh!

Uh oh!

Conversation

jchlanda commented Jul 10, 2026 • edited 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 Jul 20, 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!

jchlanda Jul 20, 2026

Choose a reason for hiding this comment

Uh oh!

kovdan01 Jul 20, 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!

jchlanda Jul 20, 2026

Choose a reason for hiding this comment

Uh oh!

kovdan01 Jul 22, 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!

jchlanda Jul 22, 2026

Choose a reason for hiding this comment

Uh oh!

kovdan01 Jul 22, 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!

jchlanda Jul 22, 2026

Choose a reason for hiding this comment

Uh oh!

kovdan01 Jul 22, 2026 • edited 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 Jul 22, 2026

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

jchlanda commented Jul 10, 2026 •

edited

Loading

kovdan01 Jul 20, 2026 •

edited by rustbot

Loading

kovdan01 Jul 20, 2026 •

edited by rustbot

Loading

kovdan01 Jul 22, 2026 •

edited by rustbot

Loading

kovdan01 Jul 22, 2026 •

edited by rustbot

Loading

kovdan01 Jul 22, 2026 •

edited

Loading