Page MenuHomePhabricator
Feed Advanced Search

Wed, Jun 5

Anastasia added a comment to D54258: [Clang] Fix pretty printing of CUDA address spaces.

Merging in two weeks is fine for me.

Wed, Jun 5, 10:28 AM · Restricted Project, Restricted Project
Anastasia committed rGd97ea1bc1aca: [Clang] Fix pretty printing of CUDA address spaces (authored by Anastasia).
[Clang] Fix pretty printing of CUDA address spaces
Wed, Jun 5, 10:27 AM
Anastasia added a reviewer for D60193: [OpenCL] Added addrspace_cast operator: mantognini.
Wed, Jun 5, 9:22 AM
Anastasia added a comment to D60193: [OpenCL] Added addrspace_cast operator.

Ping! I would appreciate any feedback to this change please. Thanks you!

Wed, Jun 5, 9:22 AM
Anastasia created D62914: [Sema] Fix diagnostic for addr spaces in reference binding.
Wed, Jun 5, 9:21 AM
Anastasia abandoned D58060: Fix diagnostic for addr spaces in reference binding.

Due to my recent commits I had to changed this slightly so it is now under a new review: https://reviews.llvm.org/D62914

Wed, Jun 5, 9:21 AM
Anastasia committed rG7ca9b978c4f4: [OpenCL][PR42031] Prevent deducing addr space in type alias. (authored by Anastasia).
[OpenCL][PR42031] Prevent deducing addr space in type alias.
Wed, Jun 5, 7:48 AM
Anastasia committed rG5145b1e4421a: [Sema] Prevent binding incompatible addr space ref to temporaries (authored by Anastasia).
[Sema] Prevent binding incompatible addr space ref to temporaries
Wed, Jun 5, 7:03 AM

Mon, Jun 3

Anastasia added a comment to D62574: Initial draft of target-configurable address spaces..

This patch does not address the issue with the accessors
on Qualifiers (isAddressSpaceSupersetOf, compatiblyIncludes),
because I don't know how to solve it without breaking a ton of
rather nice encapsulation. Either, every mention of compatiblyIncludes
must be replaced with a call to a corresponding ASTContext method,
Qualifiers must have a handle to ASTContext, or ASTContext must be
passed to every such call. This revision mentions the issue in a comment:
https://reviews.llvm.org/D49294

I think using ASTContext helper is ok then.

Alright. Just to clarify, you mean the first option (using a new accessor on ASTContext and replacing all existing compatiblyIncludes with that)?

Mon, Jun 3, 11:31 AM · Restricted Project
Anastasia added inline comments to D62584: [OpenCL][PR42033] Deducing addr space with template parameter types.
Mon, Jun 3, 11:02 AM
Anastasia added a comment to D62584: [OpenCL][PR42033] Deducing addr space with template parameter types.

I think the right approach here is probably to make sure you're applying deduction during instantiation as well.

Mon, Jun 3, 10:55 AM
Anastasia updated the diff for D62584: [OpenCL][PR42033] Deducing addr space with template parameter types.
  • Added more test cases
  • Improved diagnostics to account for use of templates with addr spaces
Mon, Jun 3, 10:51 AM
Anastasia committed rG5099aef86964: [PR41567][Sema] Fixed cast kind in addr space conversions (authored by Anastasia).
[PR41567][Sema] Fixed cast kind in addr space conversions
Mon, Jun 3, 8:42 AM
Anastasia added inline comments to D60455: [SYCL] Implement SYCL device code outlining.
Mon, Jun 3, 7:55 AM · Restricted Project
Anastasia added a comment to D62591: [OpenCL][PR42031] Prevent deducing addr space in type alias..

@mantognini Good spot! Thanks!

Mon, Jun 3, 7:35 AM · Restricted Project
Anastasia updated the diff for D62591: [OpenCL][PR42031] Prevent deducing addr space in type alias..
  • Exclude pointee from early return in addr space deduction
  • Improved testing: added FileCheck and avoid testing unnecessary bits
Mon, Jun 3, 7:30 AM · Restricted Project
Anastasia added inline comments to D62413: [OpenCL][PR41727] Prevent ICE on global dtors.
Mon, Jun 3, 6:55 AM
Anastasia updated the diff for D62413: [OpenCL][PR41727] Prevent ICE on global dtors.
  • Guard OpenCL specific IR generation under OpenCL lang mode
Mon, Jun 3, 6:49 AM
Anastasia updated the diff for D62156: [Sema][PR41730] Diagnose addr space mismatch while constructing objects.

Changed FIXME to a regular comment, modified wording.

Mon, Jun 3, 4:37 AM
Anastasia added inline comments to D62156: [Sema][PR41730] Diagnose addr space mismatch while constructing objects.
Mon, Jun 3, 4:37 AM
Anastasia added inline comments to D60455: [SYCL] Implement SYCL device code outlining.
Mon, Jun 3, 3:01 AM · Restricted Project
Anastasia accepted D62580: [OpenCL] Use long instead of long long in x86 builtins.

LGTM! Thanks!

Mon, Jun 3, 3:00 AM · Restricted Project, Restricted Project

Thu, May 30

Anastasia added a comment to D62574: Initial draft of target-configurable address spaces..

This patch does not address the issue with the accessors
on Qualifiers (isAddressSpaceSupersetOf, compatiblyIncludes),
because I don't know how to solve it without breaking a ton of
rather nice encapsulation. Either, every mention of compatiblyIncludes
must be replaced with a call to a corresponding ASTContext method,
Qualifiers must have a handle to ASTContext, or ASTContext must be
passed to every such call. This revision mentions the issue in a comment:
https://reviews.llvm.org/D49294

Thu, May 30, 12:07 PM · Restricted Project
Anastasia committed rGf61b5481fd2c: [OpenCL] Fix OpenCL/SPIR version metadata in C++ mode. (authored by Anastasia).
[OpenCL] Fix OpenCL/SPIR version metadata in C++ mode.
Thu, May 30, 11:11 AM
Anastasia added inline comments to D60455: [SYCL] Implement SYCL device code outlining.
Thu, May 30, 10:53 AM · Restricted Project
Anastasia added inline comments to D62156: [Sema][PR41730] Diagnose addr space mismatch while constructing objects.
Thu, May 30, 7:46 AM
Anastasia updated the diff for D62156: [Sema][PR41730] Diagnose addr space mismatch while constructing objects.
  • Improved diagnostic
  • Added more comments
Thu, May 30, 7:46 AM
Anastasia created D62657: [OpenCL] Fix OpenCL/SPIR version metadata.
Thu, May 30, 6:12 AM · Restricted Project
Anastasia added a comment to D62580: [OpenCL] Use long instead of long long in x86 builtins.

Do you think it's possible to add a test?

Thu, May 30, 4:58 AM · Restricted Project, Restricted Project

Wed, May 29

Anastasia created D62591: [OpenCL][PR42031] Prevent deducing addr space in type alias..
Wed, May 29, 6:57 AM · Restricted Project
Anastasia accepted D62588: [OpenCL] Support logical vector operators in C++ mode.

LGTM! Thanks!

Wed, May 29, 6:53 AM · Restricted Project, Restricted Project
Anastasia created D62584: [OpenCL][PR42033] Deducing addr space with template parameter types.
Wed, May 29, 6:32 AM

Tue, May 28

Anastasia added a comment to D60455: [SYCL] Implement SYCL device code outlining.

Design question: since you are not aware what functions are to be run on a device while parsing them, at what point do you plan to diagnose the invalid behavior from the standard C++ i.e. using function pointers in kernel code is likely to cause issues?

We are going to use DeviceDiagBuilder and related infrastructure implemented in Clang to diagnose device side code errors in OpenMP/CUDA modes.
More details are in the comments here:
https://clang.llvm.org/doxygen/classclang_1_1Sema_1_1DeviceDiagBuilder.html#details

Tue, May 28, 3:41 AM · Restricted Project

Fri, May 24

Anastasia created D62413: [OpenCL][PR41727] Prevent ICE on global dtors.
Fri, May 24, 11:16 AM
Anastasia accepted D62377: [OpenCL] Add support for the cl_arm_integer_dot_product extensions.

LGTM! Thanks!

Fri, May 24, 7:00 AM · Restricted Project
Anastasia added inline comments to D62156: [Sema][PR41730] Diagnose addr space mismatch while constructing objects.
Fri, May 24, 7:00 AM
Anastasia updated the diff for D62156: [Sema][PR41730] Diagnose addr space mismatch while constructing objects.
  • Added candidates as not viable
  • Added dedicated diagnostic note
  • Added assert
Fri, May 24, 6:55 AM
Anastasia added a comment to D62377: [OpenCL] Add support for the cl_arm_integer_dot_product extensions.

This deserves at least a link to the published spec in the description!

Fri, May 24, 4:17 AM · Restricted Project
Anastasia updated the diff for D62335: [OpenCL][PR41963] Add overloads of old atomics with generic pointer type in C++ mode .

Fixed atomic_max and atomic_min. Thanks to David for pointing this out on the bug!

Fri, May 24, 2:42 AM

Thu, May 23

Anastasia created D62335: [OpenCL][PR41963] Add overloads of old atomics with generic pointer type in C++ mode .
Thu, May 23, 11:39 AM
Anastasia retitled D62156: [Sema][PR41730] Diagnose addr space mismatch while constructing objects from [Sema] Diagnose addr space mismatch while constructing objects to [Sema][PR41730] Diagnose addr space mismatch while constructing objects.
Thu, May 23, 11:23 AM
Anastasia updated the diff for D62156: [Sema][PR41730] Diagnose addr space mismatch while constructing objects.
  • Switched back to loop over all method qualifiers
  • Moved addr space check into AddOverloadCandidate
Thu, May 23, 10:38 AM
Anastasia accepted D62197: [OpenCL] Fix file-scope const sampler variable for 2.0.

Makes sense. LGTM! Thanks!

Thu, May 23, 4:45 AM · Restricted Project
Anastasia accepted D60763: Prototype OpenCL BIFs using Tablegen.

Please address the minor nitpicks suggested here in your final commit.

Thu, May 23, 4:37 AM · Restricted Project, Restricted Project
Anastasia added a comment to D60455: [SYCL] Implement SYCL device code outlining.

Design question: since you are not aware what functions are to be run on a device while parsing them, at what point do you plan to diagnose the invalid behavior from the standard C++ i.e. using function pointers in kernel code is likely to cause issues?

Thu, May 23, 4:28 AM · Restricted Project
Anastasia created D62299: [PR41567][Sema] Fixed cast kind in addr space conversions.
Thu, May 23, 3:30 AM · Restricted Project

Wed, May 22

Anastasia added inline comments to D60455: [SYCL] Implement SYCL device code outlining.
Wed, May 22, 10:05 AM · Restricted Project
Anastasia accepted D62208: [OpenCL] Enable queue_t and clk_event_t comparisons in C++ mode.

LGTM! Thanks!

Wed, May 22, 8:25 AM · Restricted Project, Restricted Project

Tue, May 21

Anastasia accepted D62181: [OpenCL] Support pipe keyword in C++ mode.

LGTM! Thanks!

Tue, May 21, 3:28 AM · Restricted Project
Anastasia added inline comments to D60763: Prototype OpenCL BIFs using Tablegen.
Tue, May 21, 3:24 AM · Restricted Project, Restricted Project

Mon, May 20

Anastasia added inline comments to D59988: [PR41276] Generate address space cast of 'this' for objects attributed by an address space in C++.
Mon, May 20, 11:32 AM · Restricted Project
Anastasia added inline comments to D62156: [Sema][PR41730] Diagnose addr space mismatch while constructing objects.
Mon, May 20, 11:21 AM
Anastasia created D62156: [Sema][PR41730] Diagnose addr space mismatch while constructing objects.
Mon, May 20, 11:21 AM
Anastasia updated the summary of D62156: [Sema][PR41730] Diagnose addr space mismatch while constructing objects.
Mon, May 20, 11:21 AM

May 10 2019

Anastasia added a comment to D60763: Prototype OpenCL BIFs using Tablegen.

The overall structure looks very clear now! I just have a couple of nitpicks! Thank you!

May 10 2019, 4:23 AM · Restricted Project, Restricted Project

May 9 2019

Anastasia committed rGe6cf6c78f88f: [OpenCL] Make global ctor init function a kernel (authored by Anastasia).
[OpenCL] Make global ctor init function a kernel
May 9 2019, 6:54 AM
Anastasia committed rG933e305ed909: [OpenCL] Switched CXX mode to be derived from C++17 (authored by Anastasia).
[OpenCL] Switched CXX mode to be derived from C++17
May 9 2019, 4:55 AM
Anastasia added a comment to D61488: [OpenCL] Make global ctor init function a kernel.

The comments could use a bit more polishing but nothing that justifies making another iteration IMHO. LGTM!

May 9 2019, 3:58 AM · Restricted Project
Anastasia committed rGeba9a6e08fcb: [SPIR] Simplified target checking. (authored by Anastasia).
[SPIR] Simplified target checking.
May 9 2019, 3:24 AM
Anastasia committed rG3cdf8981054b: [SPIR] Simplified target checking. (authored by Anastasia).
[SPIR] Simplified target checking.
May 9 2019, 3:16 AM

May 8 2019

Anastasia updated the diff for D61488: [OpenCL] Make global ctor init function a kernel.
  • Improved comments
  • Switched to SPIR kernel
May 8 2019, 8:56 AM · Restricted Project
Anastasia committed rG5b6dda33d122: [Sema][OpenCL] Make address space conversions a bit stricter. (authored by Anastasia).
[Sema][OpenCL] Make address space conversions a bit stricter.
May 8 2019, 7:24 AM
Anastasia updated the diff for D61506: [OpenCL] Switch to C++17.

Added HexFloat!

May 8 2019, 5:00 AM · Restricted Project
Anastasia added a comment to D58236: Make address space conversions a bit stricter..

This was accepted a while ago, but never landed. I don't have commit access; could someone commit it?

May 8 2019, 4:08 AM · Restricted Project, Restricted Project
Anastasia added a comment to D61506: [OpenCL] Switch to C++17.

Per the OpenCL C++ 1.0 specification, section 2:

The OpenCL C++ programming language is based on the ISO/IEC JTC1 SC22 WG21 N 3690 language specification (a.k.a. C++14 specification).

I think it would be reasonable to permit changing the base C++ standard in OpenCL C++ mode

Indeed! There should be an option to pick the version of C++ the user wants for OpenCL.

May 8 2019, 2:59 AM · Restricted Project

May 7 2019

Anastasia added inline comments to D59988: [PR41276] Generate address space cast of 'this' for objects attributed by an address space in C++.
May 7 2019, 7:44 AM · Restricted Project
Anastasia accepted D61639: Add Triple::isSPIR() to simplify code.

LGTM! Thanks!

May 7 2019, 7:26 AM · Restricted Project, Restricted Project
Anastasia committed rGd6865b7d71bc: [OpenCL] Prevent mangling kernel functions. (authored by Anastasia).
[OpenCL] Prevent mangling kernel functions.
May 7 2019, 7:21 AM
Anastasia added a comment to D61506: [OpenCL] Switch to C++17.

Per the OpenCL C++ 1.0 specification, section 2:

The OpenCL C++ programming language is based on the ISO/IEC JTC1 SC22 WG21 N 3690 language specification (a.k.a. C++14 specification).

I think it would be reasonable to permit changing the base C++ standard in OpenCL C++ mode, but we'd need a good reason to deviate from the behavior specified in the OpenCL C++ specification by default.

May 7 2019, 2:57 AM · Restricted Project

May 3 2019

Anastasia added a comment to D60379: Make precompiled headers reproducible.

Can this be closed now due to commit r358674?

May 3 2019, 7:58 AM · Restricted Project
Anastasia added inline comments to D59988: [PR41276] Generate address space cast of 'this' for objects attributed by an address space in C++.
May 3 2019, 7:46 AM · Restricted Project
Anastasia created D61506: [OpenCL] Switch to C++17.
May 3 2019, 7:05 AM · Restricted Project
Anastasia added inline comments to D61488: [OpenCL] Make global ctor init function a kernel.
May 3 2019, 5:02 AM · Restricted Project
Anastasia created D61488: [OpenCL] Make global ctor init function a kernel.
May 3 2019, 5:01 AM · Restricted Project
Anastasia updated the diff for D61318: [Sema] Prevent binding references with mismatching address spaces to temporaries.

Updated comments!

May 3 2019, 3:02 AM · Restricted Project
Anastasia added a comment to D57464: Generalize method overloading on addr spaces to C++.

I think I would lean towards the latter since it means less fudging around with a whole bunch of unrelated methods. Do @rjmccall or @rsmith have any further opinions on this?

Ok, I can change the patch to prototype this approach. I might need some example test cases though.

Alright!

Just to make sure of something here - are you waiting for me to provide some example cases? There hasn't been activity here in a while and I was just wondering if it was because you were waiting for this.

May 3 2019, 2:47 AM
Anastasia added a comment to D60454: [OpenCL] Prevent mangling kernel functions.

I think it would be more reasonable to just change getDeclLanguageLinkage to check for a kernel function.

May 3 2019, 2:39 AM · Restricted Project
Anastasia updated the diff for D60454: [OpenCL] Prevent mangling kernel functions.

Removed redundant C linkage checks!

May 3 2019, 2:31 AM · Restricted Project

May 2 2019

Anastasia committed rG26e095e84f0d: [OpenCL] Fix initialisation of this via pointer. (authored by Anastasia).
[OpenCL] Fix initialisation of this via pointer.
May 2 2019, 9:11 AM
Anastasia accepted D61319: [PR41674] [OpenCL] Fix initialisation of this via pointer.

LGTM! Thanks!

May 2 2019, 8:18 AM · Restricted Project
Anastasia updated the diff for D60454: [OpenCL] Prevent mangling kernel functions.

Extended FunctionDecl exten C logic to take kernel function into account.

May 2 2019, 8:15 AM · Restricted Project
Anastasia committed rG44bb0aa99414: [OpenCL] Deduce static data members to __global addr space. (authored by Anastasia).
[OpenCL] Deduce static data members to __global addr space.
May 2 2019, 7:39 AM
Anastasia updated the diff for D61318: [Sema] Prevent binding references with mismatching address spaces to temporaries.

Added extra overload for isAddressSpaceSupersetOf.

May 2 2019, 6:42 AM · Restricted Project
Anastasia added inline comments to D61274: [Sema][AST] Explicit visibility for OpenCL/CUDA kernels/variables.
May 2 2019, 6:11 AM · Restricted Project
Anastasia added inline comments to D61319: [PR41674] [OpenCL] Fix initialisation of this via pointer.
May 2 2019, 3:27 AM · Restricted Project

May 1 2019

Anastasia added a comment to D61304: [OpenCL][PR41609] Deduce static data members to __global addr space.

Presumably a similar rule would apply to thread-locals if you supported them.

May 1 2019, 2:08 AM · Restricted Project
Anastasia added inline comments to D61318: [Sema] Prevent binding references with mismatching address spaces to temporaries.
May 1 2019, 2:03 AM · Restricted Project
Anastasia added a comment to D61319: [PR41674] [OpenCL] Fix initialisation of this via pointer.

I would suggest an IR test, especially since what you want to achieve is addrspacecast in IR output. We use AST dump only if there is no other good way to test something. In fact we already have a test with similar logic that tests various OpenCL address spaces: test/CodeGenOpenCLCXX/addrspace-of-this.cl. I think it should be reasonable enough to extend it.

May 1 2019, 1:32 AM · Restricted Project

Apr 30 2019

Anastasia added a comment to D58060: Fix diagnostic for addr spaces in reference binding.

@rjmccall or @ebevhan do you have any more feedback for this patch?

Apr 30 2019, 9:33 AM
Anastasia created D61318: [Sema] Prevent binding references with mismatching address spaces to temporaries.
Apr 30 2019, 9:14 AM · Restricted Project
Anastasia added a comment to D61274: [Sema][AST] Explicit visibility for OpenCL/CUDA kernels/variables.

Seems reasonable for OpenCL kernels. You might want to add an AST dump test to check that the visibility is being set correctly in case it's being printed in AST.

Apr 30 2019, 5:00 AM · Restricted Project
Anastasia created D61304: [OpenCL][PR41609] Deduce static data members to __global addr space.
Apr 30 2019, 4:40 AM · Restricted Project

Apr 27 2019

Anastasia added a comment to D60455: [SYCL] Implement SYCL device code outlining.

I think potentially reusing OpenCL features is desirable since the device code of SYCL is largely OpenCL. However I don't think we are clear enough about the overall device compilation flow of SYCL and I can easily suggest a number of different approaches including those that don't modify compiler at all. :) I am afraid until we have the big picture clear it will be hard to make any sensible decisions.

Apr 27 2019, 12:19 AM · Restricted Project

Apr 26 2019

Anastasia added a comment to D60763: Prototype OpenCL BIFs using Tablegen.

I also think we could reduce the size of the tables.
To sum up how this is working right now:

  1. The isOpenCLBuiltin(char* functionName) funcion is called to determine if a function is part of OpenCL builtin functions. If so, it returns its associated pair (index, number of signatures) in the OpenCLBuiltinDecl, otherwise it returns 0.
  2. The OpenCLBuiltinDecl table is storing, for each OpenCL builtin function, the list of the possible signature associated for this function (e.g. cos can be "float cos(float)", "double cos(double)", ...). The available signatures are stored in the OpenCLSignature table. In the OpenCLBuiltinDecl are stored the indexes corresponding to the possible signature for each function.
  3. The OpenCLSignature is storing the possible signatures.

E.g.: For the prototype float cos(float):

  1. isOpenCLBuiltin("cos") is returning the pair (345, 18)
  2. OpenCLBuiltinDecl[345] to OpenCLBuiltinDecl[345+18] are the available signatures of the builtin function "cos". Let say OpenCLBuiltinDecl[346] is containing our "float cos(float)" prototype. OpenCLBuiltinDecl[346] is containing the pair (123, 2), indexing the OpenCLSignature table and how many entries we have to read.
  3. OpenCLSignature[123] is storing the return type of the function, so the "float" type. OpenCLSignature[124] is containing the type of the first argument, so the float type again. We are not looking further in the table because we are only looking for 2 types.

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

    In the "float cos(float)" prototype, the information about the "float" type is duplicated. Plus, this "float" type is also the same as in the "float sin(float)" function. A third table, storing the different types, would discard duplicated definitions of types. The OpenCLSignature would store indexes of the required types, and the third table the type itself. This third table would be called OpenCLTypes, and would be as:

    ` struct OpenCLTypes { A type (e.g.: float, int, ...) OpenCLTypeID ID; Size of the vector (if applicable) unsigned VectorWidth; 0 if the type is not a pointer unsigned isPointer; Address space of the pointer (if applicable) clang::LangAS AS; } ` and OpenCLSignature:

    ` struct OpenCLSignature { unsigned OpenCLTypesIndex } ` --------------------------------------------------------------------------------------- Another way to save space would be to group functions. The "sin" and "cos" functions are similar (identical, we could say), regarding their use/ signature. However, they have two distinct lists of signatures in the OpenCLBuiltinDecl table. The consequence would be that isOpenCLBuiltin("cos") and isOpenCLBuiltin("sin") would give the same output. Such group of functions could be created manually by adding an attribute in the OpenCLBuiltins.td file, or automatically generated in the ClangOpenCLBuiltinEmitter.cpp file. The first solution would however make the feature potentially less understandable/ more complex to add new functions. The second solution is complex to implement/ could require a lot of time to process.
Apr 26 2019, 9:04 PM · Restricted Project, Restricted Project

Apr 24 2019

Anastasia accepted D60764: Add clang cc1 option to generate OpenCL builtin functions.

LGTM! Thanks!

Apr 24 2019, 11:44 PM · Restricted Project

Apr 19 2019

Anastasia added a comment to D60455: [SYCL] Implement SYCL device code outlining.

SYCL is similar to OpenMP 5 for C++, where you use only C++ classes instead of #pragma. So it is quite C++-friendlier than OpenMP.

Apr 19 2019, 5:38 AM · Restricted Project

Apr 18 2019

Anastasia accepted D60835: [Serialization] Stable serialization order for OpenCLTypeExtMap and OpenCLDeclExtMap.

We could still go for something like the following but accept that some a random failure might happen not necessarily on a commit that introduces it?

I am not sure that this is needed. Non-deterministic tests are really annoying.

Yes. I think this might be a wider topic to address if we happen to have more issues of this kind.

Apr 18 2019, 6:58 AM · Restricted Project, Restricted Project
Anastasia added a comment to D60602: [InferAddressSpaces] Add AS parameter to the pass factory.

This doesn't fix anything broken in LLVM but I did check that it didn't break any of the existing tests either :).

I did think about adding a test but in this case testing the new parameter would require a full C++ app that makes use of the pass and grep'ing in the tests returned zero matches for similar tests.

Apr 18 2019, 6:38 AM · Restricted Project
Anastasia added a comment to D60602: [InferAddressSpaces] Add AS parameter to the pass factory.

The pass is used by the AMDGPU and NVPTX backends. All their tests are passing with this change.

Yes, but they were passing before? Or is this fixing any bug in the existing tests?

If the change is adding functionality it is a good practice to add a test demonstrating its use. Otherwise, someone can commit something later that break the functionality and we might not have any way to detect that it's broken.

Apr 18 2019, 6:18 AM · Restricted Project
Anastasia added a comment to D60602: [InferAddressSpaces] Add AS parameter to the pass factory.

The pass is used by the AMDGPU and NVPTX backends. All their tests are passing with this change.

Apr 18 2019, 6:15 AM · Restricted Project