More complete support for opaque pointers (required for LLVM 15+) #1085
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
kquick
approved these changes
May 4, 2023
|
I'm on vacation for the next two weeks, but I approve of the overall direction of this PR. Please feel free to merge without my review. |
RyanGlScott
force-pushed
the
llvm-15-take-two
branch
from
c6cca74 to
f0e445f
Compare
|
I was surprised to see that the CVC5 abduction-related tests fail on CI, as I observed these tests passing locally. But some further investigation reveals why: I was using CVC5 1.0.5 locally, the CI was using CVC5 1.0.2, and these tests are sensitive to the choice of CVC5 version. Agh! |
For now, we simply ignore the explicit `Type` field. In a subsequent commit, we will use this type rather than inspecting the pointee type of the pointer argument. See #1075.
For now, we simply ignore the explicit `Type` fields. In a subsequent commit, we will use these types rather than inspecting the pointee type of the pointer arguments. See #1075.
This adds a `PtrOpaqueType` data constructor to `MemType`, which encodes opaque pointer types in `crucible-llvm`. There also adds cases for `PtrOpaqueType` in several functions. See #1075. There is more work to be done to robustly support opaque pointers, but I will split this out into separate commits.
Currently, `uc-crux-llvm` does not support opaque pointers (see the discussion at #1075). For now, we make `uc-crux-llvm` error if given an opaque pointer as input. We also instruct the `uc-crux-llvm` test suite to invoke Clang in such a way that it will not produce opaque pointers.
Clang 15+ now treat `-Wint-conversion` warnings as fatal errors (see https://releases.llvm.org/15.0.0/tools/clang/docs/ReleaseNotes.html#improvements-to-clang-s-diagnostics), which causes some test cases in `uc-crux-llvm` to be spuriously rejected. This patch fixes the affected test cases by correcting some return types to be `int*` instead of `int` and adding a `uintptr_t` cast.
Clang 15+ now treat `-Wimplicit-int` warnings as fatal errors (see https://releases.llvm.org/15.0.0/tools/clang/docs/ReleaseNotes.html#improvements-to-clang-s-diagnostics), which causes some test cases in `uc-crux-llvm` to be spuriously rejected. This patch adds the necessary types to avoid these spurious errors.
…th Clang 15+ Clang 15+ now treat `-Wimplicit-function-declaration` warnings as a fatal error (see https://releases.llvm.org/15.0.0/tools/clang/docs/ReleaseNotes.html#improvements-to-clang-s-diagnostics), which causes some test cases in `crux-llvm` and `uc-crux-llvm` to be spuriously rejected. This patch adds the missing `#include`s needed to avoid these spurious errors. Fixes #1076.
This patch: 1. Makes the LLVM override quasiquoter aware of opaque pointer types, which are parsed like LLVM's `ptr` type. 2. Makes the LLVM override matching machinery treat opaque pointers as equal to non-opaque pointers. This means that most of the existing overrides that use non-opaque pointers will now also work for opaque pointers... 3. ...well, for _most_ overrides, that is. An exception to this rule are certain classes of polymorphic LLVM intrinsics, such as `llvm.lifetime.start.*`. Such an intrinsic would be named `llvm.lifetime.start.p0i8` with a non-opaque pointer, where the argument type is an `i8*`, but it would be named `llvm.lifetime.start.p0` with an opaque pointer, where the argument type is a `ptr`. We could devise some clever string hackery to match both `llvm.lifetime.start.p0i8` and `llvm.lifetime.start.p0` with the same override, but that seems like overkill. There aren't _too_ many intrinsics of this sort, so I've opted to just make copies of these intrinsics specialized to opaque pointers where appropriate. See #1075.
This tweaks the translator code to avoid inspecting pointers' pointee types, an approach that won't work for opaque pointers. See #1075. The following instructions' translations are adjusted in this patch: * `getelementptr` * `load` * `store` * `cmpxchg` * `atomicrmw`
…ing. This also requires bumping the `what4` submodule to bring in the changes from GaloisInc/what4#234, which allow `what4`'s test suite to build against `tasty-sugar-2.2.0.0`.
RyanGlScott
force-pushed
the
llvm-15-take-two
branch
from
f0e445f to
30e2622
Compare
|
I've rebased this on top of #1088, so the output of the CVC5 abduction-related tests should no longer be an issue. |
RyanGlScott
added a commit
to GaloisInc/saw-script
that referenced
this pull request
May 31, 2023
This commit bumps the `llvm-pretty` submodule to bring in a commit from GaloisInc/llvm-pretty#110 that adds additional fields to `ConstGEP` to represent the basis type and expression to use for offset calculations. This also bumps the `llvm-pretty-bc-parser` and `crucible` submodule to bring in corresponding changes from GaloisInc/llvm-pretty-bc-parser#221 and GaloisInc/crucible#1085, respectively. This change affects one use site in `heapster-saw`, which is easily adapted.
RyanGlScott
added a commit
to GaloisInc/saw-script
that referenced
this pull request
May 31, 2023
This patch adds support for LLVM 15 and 16 by adding support for opaque pointers, which are described in https://llvm.org/docs/OpaquePointers.html. I have also added a test case involving LLVM bitcode using opaque pointers to kick the tires and ensure that the basics work as expected. For the most part, this is a straightforward process, as most uses of pointer types in SAW already do not care about pointee types. There are some exceptions, however: * The `typeOfSetupValue` function, as well as several places that use this function, scrutinize pointee types of pointers, which would appear to fly in the face of opaque pointers. I attempt to explain in #1876 which this is actually OK for now (although a bit awkward). * The `llvm_boilerplate`/skeleton machinery does not support opaque pointers at all. See #1877. This patch also bumps the following submodules to bring in support for opaque pointers: * `llvm-pretty`: GaloisInc/llvm-pretty#110 * `llvm-pretty-bc-parser`: GaloisInc/llvm-pretty-bc-parser#221 * `crucible`: GaloisInc/crucible#1085 This also bumps the `what4` submodule to bring in the changes from GaloisInc/what4#234. This isn't necessary to support opaque pointers, but it _is_ necessary to support a build plan involving `tasty-sugar-2.2.*`, which `llvm-pretty-bc-parser`'s test suite now requires.
This was referenced May 31, 2023
RyanGlScott
added a commit
to GaloisInc/saw-script
that referenced
this pull request
Jun 1, 2023
This patch adds support for LLVM 15 and 16 by adding support for opaque pointers, which are described in https://llvm.org/docs/OpaquePointers.html. I have also added a test case involving LLVM bitcode using opaque pointers to kick the tires and ensure that the basics work as expected. For the most part, this is a straightforward process, as most uses of pointer types in SAW already do not care about pointee types. There are some exceptions, however: * The `typeOfSetupValue` function, as well as several places that use this function, scrutinize pointee types of pointers, which would appear to fly in the face of opaque pointers. I attempt to explain in #1876 which this is actually OK for now (although a bit awkward). * The `llvm_boilerplate`/skeleton machinery does not support opaque pointers at all. See #1877. * The `llvm_fresh_expanded_val` command does not support opaque pointers at all. See #1879. This patch also bumps the following submodules to bring in support for opaque pointers: * `llvm-pretty`: GaloisInc/llvm-pretty#110 * `llvm-pretty-bc-parser`: GaloisInc/llvm-pretty-bc-parser#221 * `crucible`: GaloisInc/crucible#1085 This also bumps the `what4` submodule to bring in the changes from GaloisInc/what4#234. This isn't necessary to support opaque pointers, but it _is_ necessary to support a build plan involving `tasty-sugar-2.2.*`, which `llvm-pretty-bc-parser`'s test suite now requires.
yav
pushed a commit
to GaloisInc/saw-script
that referenced
this pull request
Jun 16, 2023
This commit bumps the `llvm-pretty` submodule to bring in a commit from GaloisInc/llvm-pretty#110 that adds additional fields to `ConstGEP` to represent the basis type and expression to use for offset calculations. This also bumps the `llvm-pretty-bc-parser` and `crucible` submodule to bring in corresponding changes from GaloisInc/llvm-pretty-bc-parser#221 and GaloisInc/crucible#1085, respectively. This change affects one use site in `heapster-saw`, which is easily adapted.
yav
pushed a commit
to GaloisInc/saw-script
that referenced
this pull request
Jun 16, 2023
This patch adds support for LLVM 15 and 16 by adding support for opaque pointers, which are described in https://llvm.org/docs/OpaquePointers.html. I have also added a test case involving LLVM bitcode using opaque pointers to kick the tires and ensure that the basics work as expected. For the most part, this is a straightforward process, as most uses of pointer types in SAW already do not care about pointee types. There are some exceptions, however: * The `typeOfSetupValue` function, as well as several places that use this function, scrutinize pointee types of pointers, which would appear to fly in the face of opaque pointers. I attempt to explain in #1876 which this is actually OK for now (although a bit awkward). * The `llvm_boilerplate`/skeleton machinery does not support opaque pointers at all. See #1877. * The `llvm_fresh_expanded_val` command does not support opaque pointers at all. See #1879. This patch also bumps the following submodules to bring in support for opaque pointers: * `llvm-pretty`: GaloisInc/llvm-pretty#110 * `llvm-pretty-bc-parser`: GaloisInc/llvm-pretty-bc-parser#221 * `crucible`: GaloisInc/crucible#1085 This also bumps the `what4` submodule to bring in the changes from GaloisInc/what4#234. This isn't necessary to support opaque pointers, but it _is_ necessary to support a build plan involving `tasty-sugar-2.2.*`, which `llvm-pretty-bc-parser`'s test suite now requires.
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.
This is a collection of various tweaks to
crux-llvmthat allows it to fully support opaque pointers, an alternative representation of pointer types that is enabled by default in LLVM 15 or later. It also tweaksuc-crux-llvmto not support opaque pointers for now, as adding support for them inuc-crux-llvmwould be nontrivial. (See #1075 for more on this.) These changes suffice to makecrux-llvmanduc-crux-llvm's test suites pass with LLVM 15 and 16.I have split this patch up into self-contained commits, and for reviewing purposes, I'd recommend viewing each patch in isolation. The general themes of these patches are:
Adapting to
llvm-prettyAPI changes (from GaloisInc/llvm-pretty#110)Adapt to Load gaining an explicit typeAdapt to GEP/ConstGEP gaining explicit typescrux-llvmsupport for opaque pointerscrucible-llvm: Add PtrOpaqueType to MemTypecrucible-llvm: Make opaque pointers play nicely with overridescrucible-llvm: Avoid inspecting pointee types during translationuc-crux-llvmnon-support for opaque pointersuc-crux-llvm: Opaque pointers are unsupportedUpdating test cases to support Clang 15/16
uc-crux-llvm-test: Avoid -Wint-conversion errors with Clang 15+uc-crux-llvm-test: Avoid -Wimplicit-int errors with Clang 15+{uc,}-crux-llvm-test: Avoid -Wimplicit-function-declaration errors with Clang 15+crux-llvm-test: Accept new test output for Clang 15crux-llvm-test: Accept new test output for Clang 16{uc,}-crux-llvm: Document support for LLVM 16 in READMEsGeneral test suite/CI tweaks needed to support multiple LLVM ranges
Use tasty-sugar v2.2.0.0 for clang-range ranged parameter test filtering.Remove unneeded imports and functions in crux-llvm tests.CI: Regenerate cabal.GHC-*.config files