Fix contract handling of promoted constants and constant static #3305
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
In preparation for the major change. Note: Need to add expected files to new tests.
|
@remi-delmas-3000 do you mind taking a look? |
Reading undefined bytes is UB. Any analysis that read uninitialized bytes is unsound, hence the value doesn't matter. |
|
Sorry, I accidentally rebased the branch. :( |
remi-delmas-3000
approved these changes
Jul 23, 2024
celinval
commented
Jul 23, 2024
tautschnig
added a commit
to tautschnig/kani
that referenced
this pull request
Aug 5, 2024
* Remove `gen_function_local_variable` and `initializer_fn_name`, the last uses of both of which were removed in model-checking#3305. * Remove `arg_count`, which was introduced in model-checking#3363, but not actually used. * Remove `insert_bb`, which was introduced in model-checking#3382, but not actually used. The toolchain upgrade to 2024-08-04 includes several bugfixes to dead-code analysis in rustc, explaining why we the recent PRs as listed above weren't flagged before for introducing dead code. Resolves: model-checking#3411
github-merge-queue bot
pushed a commit
that referenced
this pull request
Aug 5, 2024
* Remove `gen_function_local_variable` and `initializer_fn_name`, the last uses of both of which were removed in #3305. * Mark `arg_count`, which was introduced in #3363, as `allow(dead_code)` as it will soon be used. * Mark `insert_bb`, which was introduced in #3382, as `allow(dead_code)` as it will soon be used. The toolchain upgrade to 2024-08-04 includes several bugfixes to dead-code analysis in rustc, explaining why we the recent PRs as listed above weren't flagged before for introducing dead code. Resolves: #3411 By submitting this pull request, I confirm that my contribution is made under the terms of the Apache 2.0 and MIT licenses.
github-merge-queue bot
pushed a commit
that referenced
this pull request
Aug 9, 2024
## 0.54.0 ### Major Changes * We added support for slices in the `#[kani::modifies(...)]` clauses when using function contracts. * We introduce an `#[safety_constraint(...)]` attribute helper for the `Arbitrary` and `Invariant` macros. * We enabled support for concrete playback for harness that contains stubs or function contracts. * We added support for log2*, log10*, powif*, fma*, and sqrt* intrisincs. ### Breaking Changes * The `-Z ptr-to-ref-cast-checks` option has been removed, and pointer validity checks when casting raw pointers to references are now run by default. ## What's Changed * Make Kani reject mutable pointer casts if padding is incompatible and memory initialization is checked by @artemagvanian in #3332 * Fix visibility of some Kani intrinsics by @artemagvanian in #3323 * Function Contracts: Modify Slices by @pi314mm in #3295 * Support for disabling automatically generated pointer checks to avoid reinstrumentation by @artemagvanian in #3344 * Add support for global transformations by @artemagvanian in #3348 * Enable an `#[safety_constraint(...)]` attribute helper for the `Arbitrary` and `Invariant` macros by @adpaco-aws in #3283 * Fix contract handling of promoted constants and constant static by @celinval in #3305 * Bump CBMC Viewer to 3.9 by @tautschnig in #3373 * Update to CBMC version 6.1.1 by @tautschnig in #2995 * Define a struct-level `#[safety_constraint(...)]` attribute by @adpaco-aws in #3270 * Enable concrete playback for contract and stubs by @celinval in #3389 * Add code scanner tool by @celinval in #3120 * Enable contracts in associated functions by @celinval in #3363 * Enable log2*, log10* intrinsics by @tautschnig in #3001 * Enable powif* intrinsics by @tautschnig in #2999 * Enable fma* intrinsics by @tautschnig in #3002 * Enable sqrt* intrinsics by @tautschnig in #3000 * Remove assigns clause for ZST pointers by @carolynzech in #3417 * Instrumentation for delayed UB stemming from uninitialized memory by @artemagvanian in #3374 * Unify kani library and kani core logic by @jaisnan in #3333 * Stabilize pointer-to-reference cast validity checks by @artemagvanian in #3426 * Rust toolchain upgraded to `nightly-2024-08-07` by @jaisnan @qinheping @tautschnig @feliperodri ## New Contributors * @carolynzech made their first contribution in #3387 **Full Changelog**: kani-0.53.0...kani-0.54.0 By submitting this pull request, I confirm that my contribution is made under the terms of the Apache 2.0 and MIT licenses. Signed-off-by: Felipe R. Monteiro <felisous@amazon.com>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
When verifying contracts, CBMC initializes all static variables to non-deterministic values, except for those with constant types or with types / values annotated with
ID_C_no_nondet_initialization.Kani compiler never set these flags, which caused spurious failures when verification depended on promoted constants or constant static variables. This fix changes that.
First, I did a bit of refactoring since we may need to set this
Symbolproperty at a later time for static variables.I also got rid of the initialization function, since the allocation initialization can be done directly from an expression.
Then, I added the new property to the
Symboltype. In CBMC, this is a property of the type or expression. However, I decided to add it toSymbolto avoid having to add this attribute to all variants ofTypeandExpr.Resolves #3228
By submitting this pull request, I confirm that my contribution is made under the terms of the Apache 2.0 and MIT licenses.