Latest bugfixes #1
Merged
Conversation
This file contains hidden or 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
The previous PR #5881 didn't do what it meant to and had a lot of debug code due to a mistaken commit. This reverts that commit and restores the setting of the build type flag
Update comment for sanitizer fix in #5851
…#5889) When DxilModule metadata is serialized, semantic strings for system values are canonicalized. If these strings are used anywhere, this can lead to a difference in behavior between an original DxilModule that has not undergone serialization and one that has been deserialized. This change canonicalizes the string as soon as a DxilSignatureElement is created, thereby removing the possibility of accidentally relying on the non-canonical string for a system value.
hcttest noexec will execute clang tests. Clang tests include both spirv lit tests and unit tests. So running hcttest spirv_only is redundant. Fixes #5890
Moved from bottom, skipped CompatibilityWithVk1p1.
Convert the moved shaders to lit FileCheck test.
When using an initializer list to initialize a class with base classes, the base classes should be initialized first, and then used as members of the child class. This seems to be something of a niche case, but it is one of the asserts that are failing for #5638.
In order to handle types such as `RWStructuredBuffer`, which has the method `operator[]()`, `operator[]` is treated as a special case by our access chain creation code - it's treated similarly to `ArraySubscriptExpr`. This has the unfortunate side effect that user-defined `operator[]` overloads are not called in all cases. To fix this, this PR adds a check for whether a type is user-defined before handling `operator[]` in `collectArrayStructIndices`. Part of the fix for #5638.
Convert from bottom of CodeGenSpirvTest.cpp. Skipped FileTest::CompatibilityWithVk1p1 as barrier for runCodeTest.
Moved from CompatibilityWit6hVk1p1
Convert the moved shaders to lit FileCheck test.
DXIL doesn't support the bswap intrinsic, normally a backend would lower this intrinsic to something else, in this case we don't have a backend so we just end up generating invalid DXIL. The solution here is to disable bswap matching when targeting DXIL. Fixes #5104
Just add the missing expected-note. This is for #5870
Convert first part of verify test shaders as lit FileCheck test. This is for #5870 CppErrorsHV2015 is skipped for cannot set hlsl 2015 thru dxc. CppErrors is skipped for non-expected warning when test with taef.
…5895) This commit fixes a crash in the compiler when lowering a groupshared variable with a multi-dimensional array type. The root cause of the bug was that we had a nested gep expression that could not be merged into a single gep because of an intervening addrspacecast. The `MultiDimArrayToOneDimArray` pass flattens the multi-dimension global variables to a single dimension. It relies on the `MergeGepUse` function to flatten any nested geps into a single gep that fully dereferences a scalar element. The fix is to modify the `MergeGepUse` function to look through addrspacecast instructions when trying to merge geps. We can now merge geps like gep(addrspacecast(gep(p0, gep_args0)) to p1*, gep_args1) into addrspacecast(gep(p0, gep_args0+gep_args1) to p1*) We also added a call to `removeDeadConstantUsers` before flattening multi-dimension globals because we can have some dead constants hanging around after merging geps and these constants should be ignored by the flattening pass.
Convert from bottom of CodeGenSpirvTest.cpp. Skipped FileTest::CompatibilityWithVk1p1 as barrier for runCodeTest.
Started from CompatibilityWit6hVk1p1
Convert the moved shaders to lit FileCheck test.
These tests are failing in our downstream builds after migrating to Lit due to CHECK-NOTs referencing variables before they are defined. Those CHECK-NOTs have now been replaced with an `--implicit-check-not` that checks that there are no `OpDecorate ... NoContraction` instructions other than those CHECKed in the test.
A few syntax errors fixed along the way: - hcttracei.py: `print` -> `print()`, `<>` -> `!=` - hcttest-system-values.py: `os.path.join()` close paren before arguments
CHECK-COUNT-# was not yet implemented in the version of FileCheck used in DXC, so these lines were not checked at all. Switching them to CHECKs caused the tests to fail, so they're now replaced with a combination of a CHECK and CHECK-NOT that pass. This was caught by the downstream test runner that uses a newer version of FileCheck which does check CHECK-COUNTs.
Convert from bottom of CodeGenSpirvTest.cpp. Skipped FileTest::CompatibilityWithVk1p1 as barrier for runCodeTest.
While theoratically required in pre-C++11 to avoid re-allocation upon call, C++11 guarantees that c_str() returns a pointer to the internal array so pre-calling c_str() is no longer required. llvm-svn: 242983 --------- Co-authored-by: Yaron Keren <yaron.keren@gmail.com>
Disable integer resource for SampleCmp/CmpLevel/CmpLevelZero. Fixes #5564
The default generated RUN line generated by ExtractIRForPassTest.py adds the desired pass to the options, but forgot to add the '-', so it would fail with difficult to diagnose error in dxopt during argument parsing. This adds the missing '-' so the RUN line doesn't have to be fixed in this way every time the script is used.
DxilRemoveDeadBlocks assumed it could run with an uninitialized DxilModule, because the test case had been captured from live state without the needed DXIL metadata. DxilRemoveDeadBlocks relies on DxilValueCache, which relies on code in DxilSimplify, which assumes we have a DxilModule if there are dxil ops. Without an initialized DxilModule with the correct low-precision mode, cached ops and types may be wrong. This change: - modifies DxilRemoveDeadBlocks to remove bSkipInit - updates the test to contain the metadata needed to initialize DxilModule
devshgraphicsprogramming merged commit 2ae9b2b into Devsh-Graphics-Programming:devshFixes AnastaZIuk pushed a commit that referenced this pull request Jul 24, 2025
…icrosoft#7401) Change the definition of the HLSL `dot4add_i8packed` and `dot4add_u8packed` intrinsics in `utils/hct/gen_intrin_main.txt` to simply spell out the return types, rather than saying that their return type is determined by their third argument. This prevents DXC from trying to give those functions declarations like declare i64 @"\01?dot4add_u8packed@hlsl@@YA_JII_J@Z"(i32, i32, i64 signext) #1 which seems to expect a 64-bit third argument and return value. These functions are not generic, and they have only one overload, so there is no need to use interesting `uComponentTypeId` values to get the right effects, and `HLSLExternalSource::MatchArguments` seems to get confused about how to treat argument types that affect the return types. Fixes microsoft#7400.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
8 participants
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.
No description provided.