[PATCH] D43841: Add an option to disable tail-call optimization for escaping blocks

This revision was automatically updated to reflect the committed changes. Closed by commit rL326530: Add an option to disable tail-call optimization for escaping blocks. (authored by ahatanak, committed by ). Herald added a subscriber: llvm-commits. Changed prior to commit: https://reviews.llvm

[PATCH] D39562: [CodeGen][ObjC] Fix an assertion failure caused by copy elision

ahatanak updated this revision to Diff 136895. ahatanak added a comment. Add a flag to OpaqeuValueExpr that indicates whether it is a unique reference to its subexpression. If the flag is set, IRGen will avoid binding the OVE and copying the result to the destination and instead just visit the O

[PATCH] D44095: [ObjC] Allow declaring __weak pointer fields in C structs in ObjC-ARC

ahatanak created this revision. ahatanak added reviewers: rjmccall, doug.gregor, rsmith. This patch makes changes that are needed to allow `__weak` fields in C structs when ARC is enabled. Repository: rC Clang https://reviews.llvm.org/D44095 Files: include/clang/AST/Decl.h include/clang

[PATCH] D39562: [CodeGen][ObjC] Fix an assertion failure caused by copy elision

ahatanak updated this revision to Diff 137250. ahatanak marked 3 inline comments as done. ahatanak added a comment. Address review comments. https://reviews.llvm.org/D39562 Files: include/clang/AST/Expr.h include/clang/AST/Stmt.h lib/CodeGen/CGExpr.cpp lib/CodeGen/CGExprAgg.cpp lib/Co

[PATCH] D39562: [CodeGen][ObjC] Fix an assertion failure caused by copy elision

ahatanak added inline comments. Comment at: lib/Sema/SemaPseudoObject.cpp:432 + if (capturedRHS->getType()->getAsCXXRecordDecl() && capturedRHS->isRValue()) +capturedRHS->setIsUnique(); + rjmccall wrote: > I think you can unconditionally set this here, actua

[PATCH] D39562: [CodeGen][ObjC] Fix an assertion failure caused by copy elision

ahatanak added a comment. Sorry, this patch is not ready yet. There are several regression tests that are failing because of the assert in setIsUnique. https://reviews.llvm.org/D39562 ___ cfe-commits mailing list cfe-commits@lists.llvm.org http://l

[PATCH] D39562: [CodeGen][ObjC] Fix an assertion failure caused by copy elision

ahatanak updated this revision to Diff 137418. ahatanak marked an inline comment as done. ahatanak added a comment. Don't set the IsUnique bit of an OpaqueValueExpr that is used as the result expression. https://reviews.llvm.org/D39562 Files: include/clang/AST/Expr.h include/clang/AST/Stmt

[PATCH] D39562: [CodeGen][ObjC] Fix an assertion failure caused by copy elision

ahatanak added inline comments. Comment at: lib/CodeGen/CGExpr.cpp:4815 +} + } + rjmccall wrote: > Oh! So it's an interesting point that the RHS might be used as the result > expression, which means its use might not really be unique anymore. It > happen

[PATCH] D39562: [CodeGen][ObjC] Fix an assertion failure caused by copy elision

ahatanak added a comment. I also add a call to setIsUnique(false) in OpaqueValueExpr's constructor, which fixed the failing tests I was seeing. https://reviews.llvm.org/D39562 ___ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llv

[PATCH] D44095: [ObjC] Allow declaring __weak pointer fields in C structs in ObjC-ARC

ahatanak updated this revision to Diff 137500. ahatanak marked 2 inline comments as done. ahatanak added a comment. Address review comments. https://reviews.llvm.org/D44095 Files: include/clang/AST/Decl.h include/clang/AST/DeclCXX.h include/clang/AST/Type.h lib/AST/ASTImporter.cpp lib

[PATCH] D44095: [ObjC] Allow declaring __weak pointer fields in C structs in ObjC-ARC

ahatanak added inline comments. Comment at: include/clang/AST/Decl.h:3631 +PassedIndirectly = true; + } + rjmccall wrote: > I feel like this flag should be set by Sema for C++ types that have to be > passed indirectly as well; it can then become the single

[PATCH] D44095: [ObjC] Allow declaring __weak pointer fields in C structs in ObjC-ARC

This revision was automatically updated to reflect the committed changes. Closed by commit rL327206: [ObjC] Allow declaring __weak pointer fields in C structs in ARC. (authored by ahatanak, committed by ). Herald added a subscriber: llvm-commits. Changed prior to commit: https://reviews.llvm.or

[PATCH] D44095: [ObjC] Allow declaring __weak pointer fields in C structs in ObjC-ARC

This revision was automatically updated to reflect the committed changes. Closed by commit rC327206: [ObjC] Allow declaring __weak pointer fields in C structs in ARC. (authored by ahatanak, committed by ). Repository: rC Clang https://reviews.llvm.org/D44095 Files: include/clang/AST/Decl.h

[PATCH] D36918: [Sema] Take into account the current context when checking the accessibility of a member function pointer

ahatanak reopened this revision. ahatanak added a comment. This revision is now accepted and ready to land. The patch got reverted in r325335 as it broke the Chromium build. I'm reopening this review. Repository: rC Clang https://reviews.llvm.org/D36918 ___

[PATCH] D36918: [Sema] Take into account the current context when checking the accessibility of a member function pointer

ahatanak updated this revision to Diff 138066. ahatanak added a comment. The new patch passes the calling context to CheckNonDependent so that Sema::CheckAddressOfMemberAccess uses the correct context to check accessibility of explicit template arguments. I initially tried moving the declaratio

[PATCH] D44095: [ObjC] Allow declaring __weak pointer fields in C structs in ObjC-ARC

ahatanak added a comment. This was reverted in r327294 because it caused module-enabled builders to fail after I moved CXXRecordDecl::CanPassInRegisters to RecordDecl. I'm recommitting this patch with fixes to ASTDeclReader and ASTWriter. Repository: rC Clang https://reviews.llvm.org/D44095

[PATCH] D36918: [Sema] Take into account the current context when checking the accessibility of a member function pointer

ahatanak added a comment. ping https://reviews.llvm.org/D36918 ___ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[PATCH] D39562: [CodeGen][ObjC] Fix an assertion failure caused by copy elision

This revision was automatically updated to reflect the committed changes. Closed by commit rL327939: [CodeGen] Ignore OpaqueValueExprs that are unique references to their (authored by ahatanak, committed by ). Herald added a subscriber: llvm-commits. Changed prior to commit: https://reviews.llv

[PATCH] D36918: [Sema] Take into account the current context when checking the accessibility of a member function pointer

ahatanak added a comment. ping Repository: rC Clang https://reviews.llvm.org/D36918 ___ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[PATCH] D36918: [Sema] Take into account the current context when checking the accessibility of a member function pointer

ahatanak added a comment. It doesn't look like it's possible to get the context needed to do accessibility check (CXXMethodDecl 'method' for 'C::overloadedMethod' in the test case) from 'Expr *OvlExpr' in CheckAddressOfMemberAccess. It's possible to get the class in which the member is defined

[PATCH] D45015: [Preprocessor] Allow libc++ to detect when aligned allocation is unavailable.

ahatanak added a comment. In https://reviews.llvm.org/D45015#1105388, @rsmith wrote: > In https://reviews.llvm.org/D45015#1105372, @vsapsai wrote: > > > What when compiler has `__builtin_operator_new`, > > `__builtin_operator_delete`? If I build libc++ tests with recent Clang > > which has thes

[PATCH] D45015: [Preprocessor] Allow libc++ to detect when aligned allocation is unavailable.

ahatanak added a comment. I think clang should error out or warn when aligned operator or builtin operator new/delete functions are used when they are not available (r306722 should have handled them). I'm also not sure not defining `__cpp_aligned_new` is sufficient. My understanding is that if

[PATCH] D45015: [Preprocessor] Allow libc++ to detect when aligned allocation is unavailable.

ahatanak added a comment. Could you elaborate on what kind of changes you are planning to make in libc++ after committing this patch? https://reviews.llvm.org/D45015 ___ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/cgi-

[PATCH] D45015: [Preprocessor] Allow libc++ to detect when aligned allocation is unavailable.

ahatanak added a comment. I see, thank you. clang front-end currently fails to issue a warning or error when an aligned allocation/deallocation functions are required but not available in a few cases (e.g., delete called from a deleting destructor, calls to operator or builtin operator new/del

[PATCH] D47757: [Sema] Diagnose unavailable aligned deallocation functions called from deleting destructors.

ahatanak created this revision. ahatanak added reviewers: rsmith, vsapsai, EricWF, erik.pilkington. Deallocation functions are called from deleting destructors that are declared virtual. Issue an error if an aligned deallocation function is selected but is not available. Repository: rC Clang

[PATCH] D45015: [Preprocessor] Allow libc++ to detect when aligned allocation is unavailable.

ahatanak added a comment. In https://reviews.llvm.org/D45015#1123097, @EricWF wrote: > In https://reviews.llvm.org/D45015#1121874, @ahatanak wrote: > > > I see, thank you. > > > > clang front-end currently fails to issue a warning or error when an aligned > > allocation/deallocation functions ar

[PATCH] D47757: [Sema] Produce diagnostics when unavailable aligned allocation/deallocation functions are called

ahatanak updated this revision to Diff 150211. ahatanak marked an inline comment as done. ahatanak retitled this revision from "[Sema] Diagnose unavailable aligned deallocation functions called from deleting destructors." to "[Sema] Produce diagnostics when unavailable aligned allocation/dealloca

[PATCH] D22391: [Sema] Add warning for implicitly casting a null constant to a non null pointer type

ahatanak updated this revision to Diff 150218. ahatanak marked an inline comment as done. ahatanak added a reviewer: dcoughlin. ahatanak set the repository for this revision to rC Clang. ahatanak added a comment. Sorry for the delay in responding. I've addressed Jordan's review comments. I had to

[PATCH] D22391: [Sema] Add warning for implicitly casting a null constant to a non null pointer type

ahatanak added inline comments. Comment at: lib/Sema/SemaExpr.cpp:7117 +if (E && S.checkNonNullExpr(E)) + return NullabilityKind::Nullable; + jordan_rose wrote: > This isn't quite correct, unfortunately. `(_Nonnull id)nil` should be > considered non-nul

[PATCH] D36918: [Sema] Take into account the current context when checking the accessibility of a member function pointer

This revision was automatically updated to reflect the committed changes. Closed by commit rC334569: [Sema] When the address of a member function is used as a template (authored by ahatanak, committed by ). Changed prior to commit: https://reviews.llvm.org/D36918?vs=148334&id=151102#toc Reposi

[PATCH] D53154: [CodeGen] Handle extern references to OBJC_CLASS_$_*

ahatanak added a comment. LGTM Repository: rC Clang https://reviews.llvm.org/D53154 ___ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[PATCH] D53860: [SemaCXX] Don't check base's dtor is accessible

ahatanak added a comment. http://wg21.link/p0968r0#2227 says: The destructor for each element of class type is potentially invoked (15.4 [class.dtor]) from the context where the aggregate initialization occurs. [Note: This provision ensures that destructors can be called for fully-constructed

[PATCH] D57438: [Sema][ObjC] Allow declaring ObjC pointer members in C++ unions under ARC

ahatanak created this revision. ahatanak added reviewers: rsmith, rjmccall, doug.gregor. ahatanak added a project: clang. Herald added subscribers: dexonsmith, jkorous. ObjC pointer members are currently not allowed in unions in either C or C++ mode. This patch lifts the restriction in C++ mode.

[PATCH] D57438: [Sema][ObjC] Allow declaring ObjC pointer members in C++ unions under ARC

ahatanak marked an inline comment as done. ahatanak added inline comments. Comment at: test/SemaObjCXX/arc-0x.mm:164 +union { + union { // expected-note 2 {{'S1' is implicitly deleted because variant field '' has a non-trivial}} expected-note 4 {{'S1' is implicitly dele

[PATCH] D57427: [CodeGen][ObjC] Fix assert on calling `__builtin_constant_p` with ObjC objects.

ahatanak added a comment. LGTM with one minor nit. Comment at: clang/test/CodeGenObjC/builtin-constant-p.m:1 +// RUN: %clang_cc1 -triple x86_64-apple-darwin10 -emit-llvm -O3 -o - %s | FileCheck %s + Since this is just testing front-end's IRGen, can you pass `

[PATCH] D54539: [CodeGen] Replace '@' characters in block descriptors' symbol names with '\1' on ELF targets.

ahatanak created this revision. ahatanak added reviewers: theraven, rjmccall. ahatanak added a project: clang. Herald added subscribers: dexonsmith, jkorous. This fixes a bug introduced in r340041 (see the discussion in https://reviews.llvm.org/D50783). '@' can't be used in block descriptors' sy

[PATCH] D54539: [CodeGen] Replace '@' characters in block descriptors' symbol names with '\1' on ELF targets.

ahatanak added a comment. In https://reviews.llvm.org/D54539#1298949, @rjmccall wrote: > This should not be changing the actual `@`-encoding value. If we want to use > the `@`-encoding in the symbol name, we can change it when naming the symbol, > provided that \1 is not legal in the `@`encodi

[PATCH] D47757: [Sema] Produce diagnostics when unavailable aligned allocation/deallocation functions are called

ahatanak updated this revision to Diff 176214. ahatanak added a comment. Herald added a subscriber: jkorous. Rebase & ping. I've reverted the changes I made to test/SemaCUDA/call-host-fn-from-device.cu since r342749 fixed the overload resolution bug. Repository: rC Clang CHANGES SINCE LAST

[PATCH] D47757: [Sema] Produce diagnostics when unavailable aligned allocation/deallocation functions are called

ahatanak added a comment. ping Repository: rC Clang CHANGES SINCE LAST ACTION https://reviews.llvm.org/D47757/new/ https://reviews.llvm.org/D47757 ___ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/li

[PATCH] D58514: Avoid needlessly copying blocks that initialize or are assigned to local auto variables to the heap

ahatanak updated this revision to Diff 188472. ahatanak marked an inline comment as done. ahatanak added a comment. Call `IgnoreParens` on the LHS too. Repository: rC Clang CHANGES SINCE LAST ACTION https://reviews.llvm.org/D58514/new/ https://reviews.llvm.org/D58514 Files: include/clan

[PATCH] D58514: Avoid needlessly copying blocks that initialize or are assigned to local auto variables to the heap

ahatanak closed this revision. ahatanak added a comment. Fixed in r355012. Repository: rC Clang CHANGES SINCE LAST ACTION https://reviews.llvm.org/D58514/new/ https://reviews.llvm.org/D58514 ___ cfe-commits mailing list cfe-commits@lists.llvm.

[PATCH] D58729: Emit boxed expression as a compile-time constant if the expression inside the parentheses is a string literal

ahatanak created this revision. ahatanak added reviewers: rjmccall, arphaman. ahatanak added a project: clang. Herald added subscribers: jdoerfert, dexonsmith, jkorous. clang currently emits an expression like `@("abc")` as a message send to `stringWithUTF8String`. This patch makes changes so tha

[PATCH] D58744: [CodeGen] Fix some broken IR generated by -fsanitize=unsigned-integer-overflow

ahatanak added inline comments. Comment at: clang/lib/CodeGen/CGExprScalar.cpp:2921 +atomicPHI->addIncoming(old, curBlock); +Builder.CreateCondBr(success, contBB, atomicOpBB); Builder.SetInsertPoint(contBB); Would passing `atomicPHI->getParent()` ins

[PATCH] D58744: [CodeGen] Fix some broken IR generated by -fsanitize=unsigned-integer-overflow

ahatanak added inline comments. Comment at: clang/test/CodeGen/sanitize-atomic-int-overflow.c:3 + +_Atomic(unsigned) atomic; + ahatanak wrote: > It's probably better to add some check strings here. I see this has been fixed in the updated patch. CHANGES SINCE L

[PATCH] D58744: [CodeGen] Fix some broken IR generated by -fsanitize=unsigned-integer-overflow

ahatanak accepted this revision. ahatanak added a comment. This revision is now accepted and ready to land. LGTM CHANGES SINCE LAST ACTION https://reviews.llvm.org/D58744/new/ https://reviews.llvm.org/D58744 ___ cfe-commits mailing list cfe-commi

[PATCH] D58729: Emit boxed expression as a compile-time constant if the expression inside the parentheses is a string literal

ahatanak marked an inline comment as done. ahatanak added inline comments. Comment at: lib/Sema/SemaExprObjC.cpp:534 +NSStringPointer, NSStringPointer); +return new (Context) ObjCBoxedExpr(SL, BoxedType, nullptr, SR); + } rjmccall wrote:

[PATCH] D58729: Emit boxed expression as a compile-time constant if the expression inside the parentheses is a string literal

ahatanak updated this revision to Diff 188794. ahatanak marked 4 inline comments as done. ahatanak added a comment. Address review comments. - Preserve sugar when creating an `ObjCBoxedExpr` in `Sema::BuildObjCBoxedExpr`. - Add a test case to test/SemaObjC/boxing-illegal.m that shows clang reject

[PATCH] D58729: Emit boxed expression as a compile-time constant if the expression inside the parentheses is a string literal

ahatanak added inline comments. Comment at: lib/Sema/SemaExprObjC.cpp:534 +NSStringPointer, NSStringPointer); +return new (Context) ObjCBoxedExpr(SL, BoxedType, nullptr, SR); + } rjmccall wrote: > ahatanak wrote: > > rjmccall wrote: > > >

[PATCH] D58729: Emit boxed expression as a compile-time constant if the expression inside the parentheses is a string literal

ahatanak updated this revision to Diff 189436. ahatanak marked 2 inline comments as done. ahatanak added a comment. If the string literal used for the boxed expression isn't a valid UTF-8 string, don't emit a compile-time constant. Repository: rC Clang CHANGES SINCE LAST ACTION https://rev

[PATCH] D58729: Emit boxed expression as a compile-time constant if the expression inside the parentheses is a string literal

ahatanak added inline comments. Comment at: test/SemaObjC/boxing-illegal.m:70 + s = @(L"abc"); // expected-error {{illegal type 'int *' used in a boxed expression}} + s = @("\pabc"); // expected-error {{illegal type 'unsigned char *' used in a boxed expression}} +} --

[PATCH] D58729: Emit boxed expression as a compile-time constant if the expression inside the parentheses is a string literal

ahatanak marked an inline comment as done. ahatanak added inline comments. Comment at: test/SemaObjC/boxing-illegal.m:70 + s = @(L"abc"); // expected-error {{illegal type 'int *' used in a boxed expression}} + s = @("\pabc"); // expected-error {{illegal type 'unsigned char *'

[PATCH] D58729: Emit boxed expression as a compile-time constant if the expression inside the parentheses is a string literal

ahatanak updated this revision to Diff 189630. ahatanak added a comment. Diagnose invalid UTF-8 strings in boxed expressions. Repository: rC Clang CHANGES SINCE LAST ACTION https://reviews.llvm.org/D58729/new/ https://reviews.llvm.org/D58729 Files: include/clang/AST/ExprObjC.h include

[PATCH] D58729: Emit boxed expression as a compile-time constant if the expression inside the parentheses is a string literal

ahatanak updated this revision to Diff 189657. ahatanak marked 2 inline comments as done. ahatanak added a comment. Improve diagnostic message. Repository: rC Clang CHANGES SINCE LAST ACTION https://reviews.llvm.org/D58729/new/ https://reviews.llvm.org/D58729 Files: include/clang/AST/Ex

[PATCH] D58729: Emit boxed expression as a compile-time constant if the expression inside the parentheses is a string literal

ahatanak updated this revision to Diff 189809. ahatanak marked an inline comment as done. ahatanak added a comment. Use the type of `NSStringPointer` in the diagnostic string. Repository: rC Clang CHANGES SINCE LAST ACTION https://reviews.llvm.org/D58729/new/ https://reviews.llvm.org/D5872

[PATCH] D58729: Emit boxed expression as a compile-time constant if the expression inside the parentheses is a string literal

ahatanak added inline comments. Comment at: include/clang/Basic/DiagnosticSemaKinds.td:922 +def warn_objc_boxing_invalid_utf8_string : Warning< + "string is ill-formed as UTF-8 and will become a null NSString* when boxed">, + InGroup; rjmccall wrote: > Might as

[PATCH] D58729: Emit boxed expression as a compile-time constant if the expression inside the parentheses is a string literal

ahatanak marked an inline comment as done. ahatanak added inline comments. Comment at: include/clang/Basic/DiagnosticSemaKinds.td:922 +def warn_objc_boxing_invalid_utf8_string : Warning< + "string is ill-formed as UTF-8 and will become a null NSString* when boxed">, + InGroup;

[PATCH] D58729: Emit boxed expression as a compile-time constant if the expression inside the parentheses is a string literal

This revision was automatically updated to reflect the committed changes. Closed by commit rL355662: [ObjC] Emit a boxed expression as a compile-time constant if the (authored by ahatanak, committed by ). Herald added a project: LLVM. Herald added a subscriber: llvm-commits. Changed prior to comm

[PATCH] D59234: [CodeGen][ObjC] Remove the leading 'l' from symbols for protocol metadata and protocol list

ahatanak created this revision. ahatanak added a reviewer: rjmccall. ahatanak added a project: clang. Herald added subscribers: dexonsmith, jkorous. The leading 'l' tells ld64 to remove the symbol name, which can make debugging difficult. rdar://problem/47256637 Repository: rC Clang https:/

[PATCH] D59234: [CodeGen][ObjC] Remove the leading 'l' from symbols for protocol metadata and protocol list

ahatanak added a comment. There are other symbols that starts with 'l' and have private linkage. I haven't made any changes to those symbols in this patch, but I think it's possible to remove the leading 'l' from those symbols and change the linkage to internal if we want to keep the symbol nam

[PATCH] D58514: Avoid needlessly copying blocks that initialize or are assigned to local auto variables to the heap

ahatanak added a comment. Seems like the chromium code is valid and shouldn't crash. John/Erik what do you think? The following code also crashes with this patch applied. typedef void (^BlockTy)(); BlockTy sb; __weak BlockTy wb; void foo(id a) { auto b = ^{ NSLog(@"foo %@", a);

[PATCH] D58514: Avoid needlessly copying blocks that initialize or are assigned to local auto variables to the heap

ahatanak added a comment. In D58514#1428495 , @dexonsmith wrote: > In D58514#1428434 , @ahatanak wrote: > > > Seems like the chromium code is valid and shouldn't crash. John/Erik what > > do you think? The followin

[PATCH] D58514: Avoid needlessly copying blocks that initialize or are assigned to local auto variables to the heap

ahatanak added a comment. Sorry, I misread the code. If you change the parameter type of `bar` to `BlockTy`, the code crashes. If the type is `id`, it doesn't crash because IRGen copies the block to the heap in `foo` before passing it to `bar`. Repository: rC Clang CHANGES SINCE LAST ACTION

[PATCH] D58514: Avoid needlessly copying blocks that initialize or are assigned to local auto variables to the heap

ahatanak added a comment. Do you mean copying the block to the heap before assigning it to `wb` and releasing it after the assignment inside `bar`? Wouldn't the block assigned to `wb` be deallocated after the release? Repository: rC Clang CHANGES SINCE LAST ACTION https://reviews.llvm.org

[PATCH] D59234: [CodeGen][ObjC] Remove the leading 'l' from symbols for protocol metadata and protocol list

ahatanak added a comment. Okay, I'll remove the leading 'l' from other symbols and make them internal in a separate patch. Repository: rC Clang CHANGES SINCE LAST ACTION https://reviews.llvm.org/D59234/new/ https://reviews.llvm.org/D59234 ___

[PATCH] D59234: [CodeGen][ObjC] Remove the leading 'l' from symbols for protocol metadata and protocol list

This revision was automatically updated to reflect the committed changes. Closed by commit rL356156: [CodeGen][ObjC] Remove the leading 'l' from symbols for protocol (authored by ahatanak, committed by ). Herald added a project: LLVM. Herald added a subscriber: llvm-commits. Changed prior to comm

[PATCH] D59656: [CodeGen][ObjC] Annotate calls to objc_retainAutoreleasedReturnValue with notail on x86-64

ahatanak created this revision. ahatanak added reviewers: rjmccall, erik.pilkington. ahatanak added a project: clang. Herald added subscribers: dexonsmith, jkorous. On x86-64, the epilogue code inserted before the tail jump blocks the autoreleased return optimization. ARC optimizer unconditional

[PATCH] D59656: [CodeGen][ObjC] Annotate calls to objc_retainAutoreleasedReturnValue with notail on x86-64

ahatanak added a comment. There are existing test cases that test notail isn't added to calls on targets that aren't x86-64 (for example, test/CodeGenObjC/arc-arm.m), so I didn't write a new one. Repository: rC Clang CHANGES SINCE LAST ACTION https://reviews.llvm.org/D59656/new/ https://

[PATCH] D59656: [CodeGen][ObjC] Annotate calls to objc_retainAutoreleasedReturnValue with notail on x86-64

ahatanak updated this revision to Diff 191761. ahatanak marked an inline comment as done. ahatanak added a comment. Rename function. Repository: rC Clang CHANGES SINCE LAST ACTION https://reviews.llvm.org/D59656/new/ https://reviews.llvm.org/D59656 Files: lib/CodeGen/CGObjC.cpp lib/Co

[PATCH] D59656: [CodeGen][ObjC] Annotate calls to objc_retainAutoreleasedReturnValue with notail on x86-64

This revision was automatically updated to reflect the committed changes. Closed by commit rL356705: [CodeGen][ObjC] Annotate calls to objc_retainAutoreleasedReturnValue (authored by ahatanak, committed by ). Herald added a project: LLVM. Herald added a subscriber: llvm-commits. Changed prior to

[PATCH] D59873: Add additional mangling for struct members of non trivial structs

ahatanak accepted this revision. ahatanak added a comment. This revision is now accepted and ready to land. LGTM Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D59873/new/ https://reviews.llvm.org/D59873 ___

[PATCH] D60301: [CodeGen][ObjC] Emit the retainRV marker as a module flag instead of named metadata.

ahatanak created this revision. ahatanak added reviewers: steven_wu, dexonsmith, rjmccall. ahatanak added a project: clang. Herald added a subscriber: jkorous. Repository: rC Clang https://reviews.llvm.org/D60301 Files: lib/CodeGen/CGObjC.cpp test/CodeGenObjC/arc-unsafeclaim.m Index: tes

[PATCH] D60301: [CodeGen][ObjC] Emit the retainRV marker as a module flag instead of named metadata.

ahatanak abandoned this revision. ahatanak added a comment. I'll write the summary and send the patch again. Repository: rC Clang CHANGES SINCE LAST ACTION https://reviews.llvm.org/D60301/new/ https://reviews.llvm.org/D60301 ___ cfe-commits ma

[PATCH] D60302: [CodeGen][ObjC] Emit the retainRV marker as a module flag instead of named metadata.

ahatanak created this revision. ahatanak added reviewers: steven_wu, dexonsmith, rjmccall. ahatanak added a project: clang. Herald added subscribers: jkorous, kristof.beyls, javed.absar, mehdi_amini. This fixes a bug which causes the ARC contract pass to not insert the assembly marker that is nee

[PATCH] D60302: [CodeGen][ObjC] Emit the retainRV marker as a module flag instead of named metadata.

ahatanak added a comment. Corresponding llvm patch is here: https://reviews.llvm.org/D60303. Repository: rC Clang CHANGES SINCE LAST ACTION https://reviews.llvm.org/D60302/new/ https://reviews.llvm.org/D60302 ___ cfe-commits mailing list cfe-c

[PATCH] D60302: [CodeGen][ObjC] Emit the retainRV marker as a module flag instead of named metadata.

ahatanak added a comment. The upgrader rewrites the named metadata into a module flag (see https://reviews.llvm.org/D60303). Repository: rC Clang CHANGES SINCE LAST ACTION https://reviews.llvm.org/D60302/new/ https://reviews.llvm.org/D60302 _

[PATCH] D60302: [CodeGen][ObjC] Emit the retainRV marker as a module flag instead of named metadata.

ahatanak added a comment. Thanks. Does this require changes to swift too? Repository: rC Clang CHANGES SINCE LAST ACTION https://reviews.llvm.org/D60302/new/ https://reviews.llvm.org/D60302 ___ cfe-commits mailing list cfe-commits@lists.llvm.o

[PATCH] D60161: Expose non-trivial struct helpers for Swift.

ahatanak added a comment. Just one minor comment, but otherwise the patch looks good to me. Comment at: clang/include/clang/CodeGen/CodeGenABITypes.h:92 +llvm::Function *generateNonTrivialCStructDefaultConstructor(CodeGenModule &GCM, +

[PATCH] D60302: [CodeGen][ObjC] Emit the retainRV marker as a module flag instead of named metadata.

This revision was automatically updated to reflect the committed changes. Closed by commit rL358048: [CodeGen][ObjC] Emit the retainRV marker as a module flag instead of (authored by ahatanak, committed by ). Herald added a project: LLVM. Herald added a subscriber: llvm-commits. Changed prior to

[PATCH] D60626: [clang] Aligned allocation is actually supported in macosx 10.13

ahatanak accepted this revision. ahatanak added a comment. This revision is now accepted and ready to land. LGTM Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D60626/new/ https://reviews.llvm.org/D60626 ___

[PATCH] D60736: [Sema][ObjC] Don't warn about a block implicitly retaining self if the block is marked noescape

ahatanak created this revision. ahatanak added reviewers: rjmccall, erik.pilkington, arphaman. ahatanak added a project: clang. Herald added subscribers: dexonsmith, jkorous. If the block implicitly referencing `self` doesn't escape, there is no risk of creating retain cycles, so clang shouldn't

[PATCH] D60736: [Sema][ObjC] Don't warn about a block implicitly retaining self if the block is marked noescape

ahatanak added a comment. Sorry, clang was failing to diagnose `self` referenced inside a c++ lambda that was nested inside a block. Repository: rC Clang CHANGES SINCE LAST ACTION https://reviews.llvm.org/D60736/new/ https://reviews.llvm.org/D60736 _

[PATCH] D60736: [Sema][ObjC] Don't warn about a block implicitly retaining self if the block is marked noescape

ahatanak updated this revision to Diff 195498. ahatanak added a comment. Keep a list of pairs of BlockDecl and SourceLocation and, after the body of an ObjC method is parsed, emit a diagnostic if a SourceLocation in the list is inside an escaping block. Repository: rC Clang CHANGES SINCE LA

[PATCH] D60736: [Sema][ObjC] Don't warn about a block implicitly retaining self if the block is marked noescape

ahatanak updated this revision to Diff 195589. ahatanak marked 2 inline comments as done. ahatanak added a comment. Remove a potentially expensive check. Repository: rC Clang CHANGES SINCE LAST ACTION https://reviews.llvm.org/D60736/new/ https://reviews.llvm.org/D60736 Files: include/cl

[PATCH] D60736: [Sema][ObjC] Don't warn about a block implicitly retaining self if the block is marked noescape

ahatanak added inline comments. Comment at: lib/Sema/SemaExpr.cpp:2580 +if (const BlockDecl *BD = CurContext->getInnerMostBlockDecl()) + if (!getDiagnostics().isIgnored(diag::warn_implicitly_retains_self, Loc)) +ImplicitlyRetainedSelfLocs.push_back({

[PATCH] D60736: [Sema][ObjC] Don't warn about a block implicitly retaining self if the block is marked noescape

ahatanak updated this revision to Diff 195601. ahatanak marked an inline comment as done. ahatanak added a comment. Rename function. Repository: rC Clang CHANGES SINCE LAST ACTION https://reviews.llvm.org/D60736/new/ https://reviews.llvm.org/D60736 Files: include/clang/AST/DeclBase.h

[PATCH] D60736: [Sema][ObjC] Don't warn about a block implicitly retaining self if the block is marked noescape

ahatanak updated this revision to Diff 195639. ahatanak added a comment. Diagnose implicitly retained self in `ActOnFinishFunctionBody`. Repository: rC Clang CHANGES SINCE LAST ACTION https://reviews.llvm.org/D60736/new/ https://reviews.llvm.org/D60736 Files: include/clang/AST/DeclBase.

[PATCH] D60736: [Sema][ObjC] Don't warn about a block implicitly retaining self if the block is marked noescape

This revision was automatically updated to reflect the committed changes. Closed by commit rL358624: [Sema][ObjC] Don't warn about an implicitly retained self if the (authored by ahatanak, committed by ). Herald added a project: LLVM. Herald added a subscriber: llvm-commits. Changed prior to comm

[PATCH] D60900: [ObjC][ARC] Check the basic block size before calling DominatorTree::dominate

ahatanak updated this revision to Diff 195951. ahatanak added a comment. Herald added a project: clang. Herald added a subscriber: cfe-commits. Fix the comment on why we have to check the size of the basic block before calling DominatorTree::dominate to compute the dominance relation between two

[PATCH] D61147: [Sema][ObjC] Add a flavor of -Wunused-parameter that doesn't diagnose unused parameters of ObjC methods

ahatanak created this revision. ahatanak added reviewers: erik.pilkington, rjmccall. ahatanak added a project: clang. Herald added subscribers: dexonsmith, jkorous. The new warning `-Wunused-parameter-non-objc-method` works exactly the same as `-Wunused-parameter ` except for unused parameters of

[PATCH] D61147: [Sema][ObjC] Add a flavor of -Wunused-parameter that doesn't diagnose unused parameters of ObjC methods

ahatanak added a comment. In D61147#1479468 , @erik.pilkington wrote: > > Yeah, I tend to think we should just suppress this unconditionally in > > Objective-C. > > IMO this warning still makes sense for normal functions, or methods that are > only decl

[PATCH] D61147: [Sema][ObjC] Add a flavor of -Wunused-parameter that doesn't diagnose unused parameters of ObjC methods

ahatanak added a comment. Yes, I guess we can make `-Wunused-parameter ` cover unused parameters of pure @implementation methods too since users can probably just remove them to silence the warning. Repository: rC Clang CHANGES SINCE LAST ACTION https://reviews.llvm.org/D61147/new/ https

[PATCH] D61147: [Sema][ObjC] Disable -Wunused-parameter for ObjC methods

ahatanak updated this revision to Diff 197469. ahatanak retitled this revision from "[Sema][ObjC] Add a flavor of -Wunused-parameter that doesn't diagnose unused parameters of ObjC methods" to "[Sema][ObjC] Disable -Wunused-parameter for ObjC methods". ahatanak edited the summary of this revision

[PATCH] D61147: [Sema][ObjC] Disable -Wunused-parameter for ObjC methods

ahatanak added a comment. In D61147#1479932 , @rjmccall wrote: > I would also recommend that you go fix the warning to never fire on virtual > C++ methods. I tried building clang/llvm with -Wunused-parameter turned on and there are lots of places wher

[PATCH] D61454: [CodeGen][ObjC] Remove the leading 'l_' from ObjC symbols and make private symbols in the __DATA segment internal.

ahatanak created this revision. ahatanak added reviewers: rjmccall, vsk. ahatanak added a project: clang. Herald added subscribers: dexonsmith, jkorous. This change is a continuation of https://reviews.llvm.org/D59234 and makes the symbols within __DATA visible. rdar://problem/48887111 Reposit

[PATCH] D61454: [CodeGen][ObjC] Remove the leading 'l_' from ObjC symbols and make private symbols in the __DATA segment internal.

ahatanak added inline comments. Comment at: lib/CodeGen/CGObjCMac.cpp:1978 +? llvm::GlobalVariable::InternalLinkage +: llvm::GlobalValue::PrivateLinkage); // FIXME. Fix section. rj

[PATCH] D61454: [CodeGen][ObjC] Remove the leading 'l_' from ObjC symbols and make private symbols in the __DATA segment internal.

ahatanak updated this revision to Diff 197877. ahatanak marked 6 inline comments as done. ahatanak edited the summary of this revision. ahatanak added a comment. Address review comments. Repository: rC Clang CHANGES SINCE LAST ACTION https://reviews.llvm.org/D61454/new/ https://reviews.llv

[PATCH] D61454: [CodeGen][ObjC] Remove the leading 'l_' from ObjC symbols and make private symbols in the __DATA segment internal.

ahatanak marked an inline comment as done. ahatanak added inline comments. Comment at: lib/CodeGen/CGObjCMac.cpp:3961 + // linkage so that the linker preserves the symbol name. + llvm::GlobalValue::LinkageTypes LT = ExplicitDataSegment || Section.empty() +

[PATCH] D61454: [CodeGen][ObjC] Remove the leading 'l_' from ObjC symbols and make private symbols in the __DATA segment internal.

ahatanak updated this revision to Diff 197918. ahatanak marked 2 inline comments as done. ahatanak added a comment. - Instead of passing a flag to `CreateMetadataVar` that indicates the section is in segment `__DATA`, just scan the section name string inside `CreateMetadataVar`. - Fix test case

[PATCH] D61147: [Sema][ObjC] Disable -Wunused-parameter for ObjC methods

This revision was automatically updated to reflect the committed changes. Closed by commit rL359864: [Sema][ObjC] Disable -Wunused-parameter for ObjC methods (authored by ahatanak, committed by ). Herald added a project: LLVM. Herald added a subscriber: llvm-commits. Changed prior to commit: ht

[PATCH] D61454: [CodeGen][ObjC] Remove the leading 'l_' from ObjC symbols and make private symbols in the __DATA segment internal.

ahatanak marked 2 inline comments as done. ahatanak added inline comments. Comment at: lib/CodeGen/CGObjCMac.cpp:3961 + // linkage so that the linker preserves the symbol name. + llvm::GlobalValue::LinkageTypes LT = + Section.empty() || Section.startswith("__DATA") ---

<    1   2   3   4   5   6   7   8   9   10   >