Show the offset, length and memory of uninit read errors by oli-obk · Pull Request #142673 · rust-lang/rust

oli-obk · 2025-06-18T14:08:00Z

I want to improve memory dumps in general. Not sure yet how to do so best within rust diagnostics, but in a perfect world I could generate a dummy in-memory file (that contains the rendered memory dump) that we then can then provide regular rustc Spans to. So we'd basically report normal diagnostics for them with squiggly lines and everything.

rustbot · 2025-06-18T14:08:07Z

Some changes occurred to the CTFE / Miri interpreter

cc @rust-lang/miri, @RalfJung, @oli-obk, @lcnr

Some changes occurred to the CTFE machinery

cc @RalfJung, @oli-obk, @lcnr

Some changes occurred to the CTFE / Miri interpreter

cc @rust-lang/miri

oli-obk · 2025-06-18T14:08:43Z

compiler/rustc_const_eval/src/interpret/operand.rs

-            throw_ub!(InvalidUninitBytes(None));
+            throw_ub!(InvalidUninitBytes(match op.to_op(self)?.as_mplace_or_imm() {
+                Left(mplace) => mplace.ptr().provenance.and_then(|prov| {
+                    let start = mplace.ptr().into_parts().1;
+                    let size = op.layout().size;
+                    let range = alloc_range(start, size);
+                    Some((prov.get_alloc_id()?, BadBytesAccess { access: range, bad: range }))
+                }),
+                Right(_) => None,
+            }))


This one doesn't show up in diagnostics, but it seemed good to change it, too

Why doesn't it show up in diagnostics? Aren't all the miri output diffs caused by exactly this?

Ah no, that diff is probably caused by the read_scalar change.

If this logic here can't be tested, I'd rather remove it, given that it is currently wrong due to how it uses into_parts.

#142839 gets rid of it

This is read_immediate, #142839 does nothing there...?

oh right, I mixed up things.

hmm. yea I think I just tried this one because I thought it should be hit somewhere. So I'll turn it into a span_bug

ok, digging into this got more interesting. I added two more commits. TLDR: we only ever use Immediate::Uninit for zsts or for uninit locals. But reading uninit locals only happens by converting them to Operand, which handles uninit immediates if processed by erroring.

That's kind of an accidental invariant though, I think... is it worth relying on?

It allows some simplifications. lmk what you think about the latest commits and I can either restore the previous zst/uninit design, or we keep it and revisit if we notice there's something better to do with it (or matthias manages to fuzz us examples of how that code is actually reachable for non-zsts)

ok, looks like mir opts use uninit a lot more, so I backed out that commit

oli-obk · 2025-06-18T17:44:52Z

grml. I haven't been able to run miri's dep tests in-tree in forever. Always sth about libc not found. Will try to bless the others and hope none of these need blessing

rustbot · 2025-06-20T07:18:44Z

The Miri subtree was changed

cc @rust-lang/miri

RalfJung · 2025-06-27T14:44:45Z

compiler/rustc_const_eval/src/interpret/operand.rs

-            throw_ub!(InvalidUninitBytes(None));
+            throw_ub!(InvalidUninitBytes(match op.to_op(self)?.as_mplace_or_imm() {
+                Left(mplace) => mplace.ptr().provenance.and_then(|prov| {
+                    let start = mplace.ptr().into_parts().1;
+                    let size = op.layout().size;
+                    let range = alloc_range(start, size);
+                    Some((prov.get_alloc_id()?, BadBytesAccess { access: range, bad: range }))
+                }),
+                Right(_) => None,
+            }))


Why doesn't it show up in diagnostics? Aren't all the miri output diffs caused by exactly this?

RalfJung · 2025-06-27T14:46:48Z

compiler/rustc_const_eval/src/interpret/operand.rs

-            throw_ub!(InvalidUninitBytes(None));
+            throw_ub!(InvalidUninitBytes(match op.to_op(self)?.as_mplace_or_imm() {
+                Left(mplace) => mplace.ptr().provenance.and_then(|prov| {
+                    let start = mplace.ptr().into_parts().1;


This is not correct -- depending on which type of provenance this is, start will be either relative to the allocation, or absolute. into_parts has a doc comment warning about this. :) Maybe we should rename it into_raw_parts or so to make it more clear that this API is somewhat dicey.

bors · 2025-06-27T20:16:45Z

☔ The latest upstream changes (presumably #143091) made this pull request unmergeable. Please resolve the merge conflicts.

RalfJung · 2025-06-28T12:42:00Z

@rustbot author

rustbot · 2025-06-28T12:42:04Z

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

rustbot · 2025-06-30T10:32:07Z

Some changes occurred to MIR optimizations

cc @rust-lang/wg-mir-opt

bors · 2025-06-30T15:23:41Z

☔ The latest upstream changes (presumably #142839) made this pull request unmergeable. Please resolve the merge conflicts.

oli-obk · 2025-07-01T11:13:41Z

@rustbot ready

src/tools/miri/tests/fail-dep/concurrency/libc_pthread_cond_double_destroy.stderr

RalfJung · 2025-07-17T20:01:22Z

(CI doesn't run as there is still a conflict)

oli-obk · 2025-07-17T20:03:38Z

yea was running locally to see if there were any issues after the rebase

RalfJung

r=me with the last nit resolved and when CI is green.

RalfJung · 2025-07-17T20:07:54Z

src/tools/miri/tests/fail/read_from_trivial_switch.rs

 //
 // See <https://github.com/rust-lang/miri/issues/4237>.

+//@ stderr-per-bitwidth


I'd prefer normalizing the output over having two stderr files.

oli-obk · 2025-07-18T09:46:20Z

@bors r=RalfJung

bors · 2025-07-18T09:46:22Z

📌 Commit 652ba27 has been approved by RalfJung

It is now in the queue for this repository.

Show the offset, length and memory of uninit read errors r? `@RalfJung` I want to improve memory dumps in general. Not sure yet how to do so best within rust diagnostics, but in a perfect world I could generate a dummy in-memory file (that contains the rendered memory dump) that we then can then provide regular rustc `Span`s to. So we'd basically report normal diagnostics for them with squiggly lines and everything.

Rollup of 9 pull requests Successful merges: - #138554 (Distinguish delim kind to decide whether to emit unexpected closing delimiter) - #142673 (Show the offset, length and memory of uninit read errors) - #142693 (More robustly deal with relaxed bounds and improve their diagnostics) - #143382 (stabilize `const_slice_reverse`) - #143928 (opt-dist: make llvm builds optional) - #143961 (Correct which exploit mitigations are enabled by default) - #144050 (Fix encoding of link_section and no_mangle cross crate) - #144059 (Refactor `CrateLoader` into the `CStore`) - #144123 (Generalize `unsize` and `unsize_into` destinations) r? `@ghost` `@rustbot` modify labels: rollup

Rollup merge of #142673 - oli-obk:uninit-read-mem, r=RalfJung Show the offset, length and memory of uninit read errors r? ``@RalfJung`` I want to improve memory dumps in general. Not sure yet how to do so best within rust diagnostics, but in a perfect world I could generate a dummy in-memory file (that contains the rendered memory dump) that we then can then provide regular rustc `Span`s to. So we'd basically report normal diagnostics for them with squiggly lines and everything.

Rollup of 9 pull requests Successful merges: - rust-lang/rust#138554 (Distinguish delim kind to decide whether to emit unexpected closing delimiter) - rust-lang/rust#142673 (Show the offset, length and memory of uninit read errors) - rust-lang/rust#142693 (More robustly deal with relaxed bounds and improve their diagnostics) - rust-lang/rust#143382 (stabilize `const_slice_reverse`) - rust-lang/rust#143928 (opt-dist: make llvm builds optional) - rust-lang/rust#143961 (Correct which exploit mitigations are enabled by default) - rust-lang/rust#144050 (Fix encoding of link_section and no_mangle cross crate) - rust-lang/rust#144059 (Refactor `CrateLoader` into the `CStore`) - rust-lang/rust#144123 (Generalize `unsize` and `unsize_into` destinations) r? `@ghost` `@rustbot` modify labels: rollup

Show the offset, length and memory of uninit read errors r? ``@RalfJung`` I want to improve memory dumps in general. Not sure yet how to do so best within rust diagnostics, but in a perfect world I could generate a dummy in-memory file (that contains the rendered memory dump) that we then can then provide regular rustc `Span`s to. So we'd basically report normal diagnostics for them with squiggly lines and everything.

…iaskrgr Rollup of 9 pull requests Successful merges: - rust-lang#138554 (Distinguish delim kind to decide whether to emit unexpected closing delimiter) - rust-lang#142673 (Show the offset, length and memory of uninit read errors) - rust-lang#142693 (More robustly deal with relaxed bounds and improve their diagnostics) - rust-lang#143382 (stabilize `const_slice_reverse`) - rust-lang#143928 (opt-dist: make llvm builds optional) - rust-lang#143961 (Correct which exploit mitigations are enabled by default) - rust-lang#144050 (Fix encoding of link_section and no_mangle cross crate) - rust-lang#144059 (Refactor `CrateLoader` into the `CStore`) - rust-lang#144123 (Generalize `unsize` and `unsize_into` destinations) r? `@ghost` `@rustbot` modify labels: rollup

Rollup of 9 pull requests Successful merges: - rust-lang/rust#138554 (Distinguish delim kind to decide whether to emit unexpected closing delimiter) - rust-lang/rust#142673 (Show the offset, length and memory of uninit read errors) - rust-lang/rust#142693 (More robustly deal with relaxed bounds and improve their diagnostics) - rust-lang/rust#143382 (stabilize `const_slice_reverse`) - rust-lang/rust#143928 (opt-dist: make llvm builds optional) - rust-lang/rust#143961 (Correct which exploit mitigations are enabled by default) - rust-lang/rust#144050 (Fix encoding of link_section and no_mangle cross crate) - rust-lang/rust#144059 (Refactor `CrateLoader` into the `CStore`) - rust-lang/rust#144123 (Generalize `unsize` and `unsize_into` destinations) r? `@ghost` `@rustbot` modify labels: rollup

rustbot assigned RalfJung Jun 18, 2025

rustbot added S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. T-compiler Relevant to the compiler team, which will review and decide on the PR/issue. labels Jun 18, 2025

oli-obk commented Jun 18, 2025

View reviewed changes

oli-obk force-pushed the uninit-read-mem branch from c41fdab to 6ffb2af Compare June 18, 2025 14:29

This comment has been minimized.

Sign in to view

oli-obk force-pushed the uninit-read-mem branch from 6ffb2af to a45e3aa Compare June 20, 2025 07:18

This comment has been minimized.

Sign in to view

oli-obk force-pushed the uninit-read-mem branch 2 times, most recently from ff2e26b to 0120c8e Compare June 23, 2025 08:34

RalfJung reviewed Jun 27, 2025

View reviewed changes

rustbot added S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. and removed S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. labels Jun 28, 2025

oli-obk force-pushed the uninit-read-mem branch 2 times, most recently from 3abcbd3 to 67c2e92 Compare June 30, 2025 10:32

This comment has been minimized.

Sign in to view

oli-obk force-pushed the uninit-read-mem branch from 67c2e92 to dd800e1 Compare June 30, 2025 12:52

This comment has been minimized.

Sign in to view

oli-obk force-pushed the uninit-read-mem branch from dd800e1 to 6d55968 Compare July 1, 2025 11:13

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 1, 2025

rustbot added S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. and removed S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. labels Jul 17, 2025

oli-obk mentioned this pull request Jul 17, 2025

Reusable output normalizations oli-obk/ui_test#337

Open

oli-obk force-pushed the uninit-read-mem branch from 7396378 to 4250ced Compare July 17, 2025 19:06

RalfJung reviewed Jul 17, 2025

View reviewed changes

src/tools/miri/tests/fail-dep/concurrency/libc_pthread_cond_double_destroy.stderr Outdated Show resolved Hide resolved

oli-obk force-pushed the uninit-read-mem branch from 4250ced to 456161f Compare July 17, 2025 19:59

Report the range of uninit bytes in CTFE errors

12eedaf

oli-obk force-pushed the uninit-read-mem branch from 456161f to 154f820 Compare July 17, 2025 20:03

RalfJung reviewed Jul 17, 2025

View reviewed changes

Show the memory of uninit reads

652ba27

oli-obk force-pushed the uninit-read-mem branch from 154f820 to 652ba27 Compare July 18, 2025 07:47

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-author Status: This is awaiting some action (such as code changes or more information) from the author. labels Jul 18, 2025

This was referenced Jul 18, 2025

Rollup of 10 pull requests #144138

Closed

Rollup of 9 pull requests #144145

Merged

bors merged commit b3827e4 into rust-lang:master Jul 19, 2025
11 checks passed

rustbot added this to the 1.90.0 milestone Jul 19, 2025

Uh oh!

Conversation

oli-obk commented Jun 18, 2025

Uh oh!

rustbot commented Jun 18, 2025

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

This comment has been minimized.

oli-obk commented Jun 18, 2025

Uh oh!

rustbot commented Jun 20, 2025

Uh oh!

This comment has been minimized.

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

bors commented Jun 27, 2025

Uh oh!

RalfJung commented Jun 28, 2025

Uh oh!

rustbot commented Jun 28, 2025

Uh oh!

rustbot commented Jun 30, 2025

Uh oh!

This comment has been minimized.

This comment has been minimized.

bors commented Jun 30, 2025

Uh oh!

oli-obk commented Jul 1, 2025

Uh oh!

Uh oh!

RalfJung commented Jul 17, 2025

Uh oh!

oli-obk commented Jul 17, 2025

Uh oh!

RalfJung left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

oli-obk commented Jul 18, 2025

Uh oh!

bors commented Jul 18, 2025

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants