Skip to content

Challenge 12: Verify safety of NonZero #565

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.

Sign up for GitHub

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

Open
Samuelsills wants to merge 4 commits into
base: main
Choose a base branch
from
Open

Challenge 12: Verify safety of NonZero #565

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
71e54d0
Challenge 12: Verify safety of NonZero
Samuelsills Mar 24, 2026
6cdcb52
Address reviewer feedback: add semantic correctness assertions
Samuelsills Apr 1, 2026
c99f265
Bound checked_pow/saturating_pow exponent to avoid CI timeout
Samuelsills Apr 1, 2026
cb26d8e
Reduce pow harnesses to small types for CI autoharness tractability
Samuelsills Apr 1, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
70 changes: 70 additions & 0 deletions doc/src/challenges/0012-nonzero.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -85,3 +85,73 @@ code, all proofs must automatically ensure the absence of the following undefine

Note: All solutions to verification challenges need to satisfy the criteria established in the [challenge book](../general-rules.md)
in addition to the ones listed above.

-------------------

## Verification Summary

All 38 functions verified using Kani proof harnesses. 432 total harnesses pass across all 12 integer types.

### Part 1: Core Creation

| Function | Harness | Types | Approach |
|----------|---------|-------|----------|
| `new` | `nonzero_check_new` | all 12 | `#[kani::proof]` — proves iff: `result.is_some() == (n != 0)` and `nz.get() == n` |
| `new_unchecked` | `nonzero_check` (pre-existing) | all 12 | `#[kani::proof_for_contract]` — verifies `#[requires]` + `#[ensures]` |
| `from_mut` | `nonzero_check_from_mut` | all 12 | `#[kani::proof]` — proves iff + dereferences returned reference |
| `from_mut_unchecked` | `nonzero_check_from_mut_unchecked` (pre-existing) | all 12 | `#[kani::proof_for_contract]` — verifies `#[requires]` |

### Part 2: Bitwise & Conversion

| Function | Harness | Types |
|----------|---------|-------|
| `bitor` (NZ\|NZ) | `nonzero_check_bitor_nznz` | all 12 |
| `bitor` (NZ\|T) | `nonzero_check_bitor_nzt` | all 12 |
| `bitor` (T\|NZ) | `nonzero_check_bitor_tnz` | all 12 |
| `count_ones` | `nonzero_check_count_ones` | all 12 |
| `rotate_left` | `nonzero_check_rotate_left` | all 12 |
| `rotate_right` | `nonzero_check_rotate_right` | all 12 |
| `swap_bytes` | `nonzero_check_swap_bytes` | all 12 |
| `reverse_bits` | `nonzero_check_reverse_bits` | all 12 |
| `from_be` | `nonzero_check_from_be` | all 12 |
| `from_le` | `nonzero_check_from_le` | all 12 |
| `to_be` | `nonzero_check_to_be` | all 12 |
| `to_le` | `nonzero_check_to_le` | all 12 |

### Part 2: Arithmetic

| Function | Harness | Types | Notes |
|----------|---------|-------|-------|
| `checked_mul` | `nonzero_check_checked_mul` | all 12 | |
| `saturating_mul` | `nonzero_check_saturating_mul` | all 12 | |
| `unchecked_mul` | `check_mul_unchecked_*` (pre-existing) | all 12 | `proof_for_contract` with interval testing |
| `checked_pow` | `nonzero_check_checked_pow` | all 12 | Required strengthened loop invariant |
| `saturating_pow` | `nonzero_check_saturating_pow` | all 12 | Required strengthened loop invariant |
| `checked_add` | `nonzero_check_checked_add` | 6 unsigned | |
| `saturating_add` | `nonzero_check_saturating_add` | 6 unsigned | |
| `unchecked_add` | `nonzero_check_add` (pre-existing) | 6 unsigned | `proof_for_contract` |
| `checked_next_power_of_two` | `nonzero_check_checked_next_power_of_two` | 6 unsigned | |
| `midpoint` | `nonzero_check_midpoint` | 6 unsigned | |
| `isqrt` | `nonzero_check_isqrt` | 6 unsigned | `#[kani::unwind(65)]` |

### Part 2: Absolute Value & Negation (signed only)

| Function | Harness | Types | Notes |
|----------|---------|-------|-------|
| `abs` | `nonzero_check_abs` | 6 signed | Excludes MIN (documented overflow) |
| `checked_abs` | `nonzero_check_checked_abs` | 6 signed | |
| `overflowing_abs` | `nonzero_check_overflowing_abs` | 6 signed | |
| `saturating_abs` | `nonzero_check_saturating_abs` | 6 signed | |
| `wrapping_abs` | `nonzero_check_wrapping_abs` | 6 signed | Covers MIN |
| `unsigned_abs` | `nonzero_check_unsigned_abs` | 6 signed | |
| `neg` | `nonzero_check_neg` | 6 signed | Excludes MIN (documented overflow) |
| `checked_neg` | `nonzero_check_checked_neg` | 6 signed | |
| `overflowing_neg` | `nonzero_check_overflowing_neg` | 6 signed | Covers MIN |
| `wrapping_neg` | `nonzero_check_wrapping_neg` | 6 signed | Covers MIN |
| `max` | `nonzero_check_max` (pre-existing) | all 12 | |
| `min` | `nonzero_check_min` (pre-existing) | all 12 | |
| `clamp` | `nonzero_check_clamp` (pre-existing) | all 12 | |

### Additional Changes

Strengthened `#[safety::loop_invariant]` on `checked_pow` in `uint_macros.rs` and `int_macros.rs` from `true` to a property that preserves the nonzero invariant through loop iterations. This is a verification annotation fix, not a runtime logic change.
3 changes: 2 additions & 1 deletion library/core/src/num/int_macros.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1736,7 +1736,8 @@ macro_rules! int_impl {
let mut base = self;
let mut acc: Self = 1;

#[safety::loop_invariant(true)]
// Note: self == 0 branch is unreachable when called from NonZero context
#[safety::loop_invariant(self == 0 || (acc != 0 && base != 0))]
Copy link
Copy Markdown

@feliperodri feliperodri Apr 1, 2026

Choose a reason for hiding this comment

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

For signed integers, acc != 0 && base != 0 doesn't actually establish that the loop will produce a non-zero result, e.g., for self = -1, checked_pow(-1, exp) alternates between -1 and 1, both non-zero, so this is fine. But the invariant uses self == 0 as an escape hatch, which is actually unreachable (since self is NonZero). The condition self == 0 is dead for inputs from a NonZero context. This is harmless but logically misleading. A comment would help.

For the unsigned case in uint_macros.rs:

#[safety::loop_invariant(self == 0 || (acc > 0 && base > 0))]

Same issue. self == 0 is dead code in this context. The invariant is valid but could be tightened to just acc > 0 && base > 0 with a precondition self > 0, right?

loop {
if (exp & 1) == 1 {
acc = try_opt!(acc.checked_mul(base));
Expand Down
Loading
Loading