[PATCH] D133236: [clang-format] Use utf-8 for JSON object load

[Luke Drummond via Phabricator via cfe-commits]

ldrumm accepted this revision.
ldrumm added a comment.

LGTM


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D133236/new/

https://reviews.llvm.org/D133236

___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[PATCH] D133236: [clang-format] Use utf-8 for JSON object load

[Luke Drummond via Phabricator via cfe-commits]

ldrumm added a comment.

Do you have commit access?


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D133236/new/

https://reviews.llvm.org/D133236

___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[PATCH] D133236: [clang-format] Use utf-8 for JSON object load

[Luke Drummond via Phabricator via cfe-commits]

This revision was landed with ongoing or failed builds.
This revision was automatically updated to reflect the committed changes.
Closed by commit rG8f4e6cfe73d8: [clang-format] Use utf-8 for JSON object load 
(authored by kaitingwang, committed by ldrumm).

Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D133236/new/

https://reviews.llvm.org/D133236

Files:
  clang/tools/clang-format/clang-format.py


Index: clang/tools/clang-format/clang-format.py
===
--- clang/tools/clang-format/clang-format.py
+++ clang/tools/clang-format/clang-format.py
@@ -10,9 +10,9 @@
 # imap  :py3f /clang-format.py
 #   endif
 #
-# The if-elseif-endif conditional should pick either the python3 or python2 
+# The if-elseif-endif conditional should pick either the python3 or python2
 # integration depending on your vim setup.
-# 
+#
 # The first mapping enables clang-format for NORMAL and VISUAL mode, the second
 # mapping adds support for INSERT mode. Change "C-I" to another binding if you
 # need clang-format on a different key (C-I stands for Ctrl+i).
@@ -134,7 +134,7 @@
 )
   else:
 header, content = stdout.split(b'\n', 1)
-header = json.loads(header)
+header = json.loads(header.decode('utf-8'))
 # Strip off the trailing newline (added above).
 # This maintains trailing empty lines present in the buffer if
 # the -lines specification requests them to remain unchanged.


Index: clang/tools/clang-format/clang-format.py
===
--- clang/tools/clang-format/clang-format.py
+++ clang/tools/clang-format/clang-format.py
@@ -10,9 +10,9 @@
 # imap  :py3f /clang-format.py
 #   endif
 #
-# The if-elseif-endif conditional should pick either the python3 or python2 
+# The if-elseif-endif conditional should pick either the python3 or python2
 # integration depending on your vim setup.
-# 
+#
 # The first mapping enables clang-format for NORMAL and VISUAL mode, the second
 # mapping adds support for INSERT mode. Change "C-I" to another binding if you
 # need clang-format on a different key (C-I stands for Ctrl+i).
@@ -134,7 +134,7 @@
 )
   else:
 header, content = stdout.split(b'\n', 1)
-header = json.loads(header)
+header = json.loads(header.decode('utf-8'))
 # Strip off the trailing newline (added above).
 # This maintains trailing empty lines present in the buffer if
 # the -lines specification requests them to remain unchanged.
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[PATCH] D98411: [OpenCL] Respect calling convention for builtin

[Luke Drummond via Phabricator via cfe-commits]

ldrumm created this revision.
ldrumm added reviewers: Anastasia, mibintc, bader, yaxunl.
ldrumm added a project: clang.
Herald added a subscriber: wenlei.
ldrumm requested review of this revision.

`__translate_sampler_initializer` has a calling convention of
`spir_func`, but clang generated calls to it using the default CC.

Instruction Combining was lowering these mismatching calling conventions
to `store i1* undef` which itself was subsequently lowered to a trap
instruction by simplifyCFG resulting in runtime `SIGILL`

There are arguably two bugs here: but whether there's any wisdom in
converting an obviously invalid call into a runtime crash over aborting
with a sensible error message will require further discussion. So for
now it's enough to set the right calling convention on the runtime
helper.


Repository:
  rG LLVM Github Monorepo

https://reviews.llvm.org/D98411

Files:
  clang/lib/CodeGen/CodeGenModule.cpp
  clang/test/CodeGenOpenCL/sampler.cl


Index: clang/test/CodeGenOpenCL/sampler.cl
===
--- clang/test/CodeGenOpenCL/sampler.cl
+++ clang/test/CodeGenOpenCL/sampler.cl
@@ -39,7 +39,7 @@
   // Case 2b
   sampler_t smp = CLK_ADDRESS_CLAMP_TO_EDGE | CLK_NORMALIZED_COORDS_TRUE | 
CLK_FILTER_NEAREST;
   // CHECK: [[smp_ptr:%[A-Za-z0-9_\.]+]] = alloca %opencl.sampler_t 
addrspace(2)*
-  // CHECK: [[SAMP:%[0-9]+]] = call %opencl.sampler_t addrspace(2)* 
@__translate_sampler_initializer(i32 19)
+  // CHECK: [[SAMP:%[0-9]+]] = call spir_func %opencl.sampler_t addrspace(2)* 
@__translate_sampler_initializer(i32 19)
   // CHECK: store %opencl.sampler_t addrspace(2)* [[SAMP]], %opencl.sampler_t 
addrspace(2)** [[smp_ptr]]
 
   // Case 1b
@@ -56,12 +56,12 @@
 
   // Case 1a/2a
   fnc4smp(glb_smp);
-  // CHECK: [[SAMP:%[0-9]+]] = call %opencl.sampler_t addrspace(2)* 
@__translate_sampler_initializer(i32 35)
+  // CHECK: [[SAMP:%[0-9]+]] = call spir_func %opencl.sampler_t addrspace(2)* 
@__translate_sampler_initializer(i32 35)
   // CHECK: call spir_func void [[FUNCNAME]](%opencl.sampler_t addrspace(2)* 
[[SAMP]])
 
   // Case 1a/2c
   fnc4smp(glb_smp_const);
-  // CHECK: [[SAMP:%[0-9]+]] = call %opencl.sampler_t addrspace(2)* 
@__translate_sampler_initializer(i32 35)
+  // CHECK: [[SAMP:%[0-9]+]] = call spir_func %opencl.sampler_t addrspace(2)* 
@__translate_sampler_initializer(i32 35)
   // CHECK: call spir_func void [[FUNCNAME]](%opencl.sampler_t addrspace(2)* 
[[SAMP]])
 
   // Case 1c
@@ -70,12 +70,12 @@
   // CHECK: call spir_func void [[FUNCNAME]](%opencl.sampler_t addrspace(2)* 
[[SAMP]])
 
   fnc4smp(5);
-  // CHECK: [[SAMP:%[0-9]+]] = call %opencl.sampler_t addrspace(2)* 
@__translate_sampler_initializer(i32 5)
+  // CHECK: [[SAMP:%[0-9]+]] = call spir_func %opencl.sampler_t addrspace(2)* 
@__translate_sampler_initializer(i32 5)
   // CHECK: call spir_func void [[FUNCNAME]](%opencl.sampler_t addrspace(2)* 
[[SAMP]])
 
   const sampler_t const_smp = CLK_ADDRESS_CLAMP_TO_EDGE | 
CLK_NORMALIZED_COORDS_TRUE | CLK_FILTER_LINEAR;
   fnc4smp(const_smp);
-   // CHECK: [[CONST_SAMP:%[0-9]+]] = call %opencl.sampler_t addrspace(2)* 
@__translate_sampler_initializer(i32 35)
+   // CHECK: [[CONST_SAMP:%[0-9]+]] = call spir_func %opencl.sampler_t 
addrspace(2)* @__translate_sampler_initializer(i32 35)
   // CHECK: store %opencl.sampler_t addrspace(2)* [[CONST_SAMP]], 
%opencl.sampler_t addrspace(2)** [[CONST_SMP_PTR:%[a-zA-Z0-9]+]]
   fnc4smp(const_smp);
   // CHECK: [[SAMP:%[0-9]+]] = load %opencl.sampler_t addrspace(2)*, 
%opencl.sampler_t addrspace(2)** [[CONST_SMP_PTR]]
@@ -83,7 +83,7 @@
 
   constant sampler_t constant_smp = CLK_ADDRESS_CLAMP_TO_EDGE | 
CLK_NORMALIZED_COORDS_TRUE | CLK_FILTER_LINEAR;
   fnc4smp(constant_smp);
-  // CHECK: [[SAMP:%[0-9]+]] = call %opencl.sampler_t addrspace(2)* 
@__translate_sampler_initializer(i32 35)
+  // CHECK: [[SAMP:%[0-9]+]] = call spir_func %opencl.sampler_t addrspace(2)* 
@__translate_sampler_initializer(i32 35)
   // CHECK: call spir_func void [[FUNCNAME]](%opencl.sampler_t addrspace(2)* 
[[SAMP]])
 
   // TODO: enable sampler initialization with non-constant integer.
Index: clang/lib/CodeGen/CodeGenModule.cpp
===
--- clang/lib/CodeGen/CodeGenModule.cpp
+++ clang/lib/CodeGen/CodeGenModule.cpp
@@ -6263,15 +6263,17 @@
 
   return *SanStats;
 }
+
 llvm::Value *
 CodeGenModule::createOpenCLIntToSamplerConversion(const Expr *E,
   CodeGenFunction &CGF) {
   llvm::Constant *C = ConstantEmitter(CGF).emitAbstract(E, E->getType());
-  auto SamplerT = getOpenCLRuntime().getSamplerType(E->getType().getTypePtr());
-  auto FTy = llvm::FunctionType::get(SamplerT, {C->getType()}, false);
-  return CGF.Builder.CreateCall(CreateRuntimeFunction(FTy,
-"__translate_sampler_initializer"),
-{C});
+  auto *SamplerT = 
getOpenCLRuntime().getSamplerType(

[PATCH] D98411: [OpenCL] Respect calling convention for builtin

[Luke Drummond via Phabricator via cfe-commits]

ldrumm added a subscriber: svenvh.
ldrumm added a comment.

@Anastasia @svenvh is this already covered by https://reviews.llvm.org/D98039 ?


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D98411/new/

https://reviews.llvm.org/D98411

___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[PATCH] D98411: [OpenCL] Respect calling convention for builtin

[Luke Drummond via Phabricator via cfe-commits]

This revision was automatically updated to reflect the committed changes.
Closed by commit rGfcfd3fda7190: [OpenCL] Respect calling convention for 
builtin (authored by ldrumm).

Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D98411/new/

https://reviews.llvm.org/D98411

Files:
  clang/lib/CodeGen/CodeGenModule.cpp
  clang/test/CodeGenOpenCL/sampler.cl


Index: clang/test/CodeGenOpenCL/sampler.cl
===
--- clang/test/CodeGenOpenCL/sampler.cl
+++ clang/test/CodeGenOpenCL/sampler.cl
@@ -39,7 +39,7 @@
   // Case 2b
   sampler_t smp = CLK_ADDRESS_CLAMP_TO_EDGE | CLK_NORMALIZED_COORDS_TRUE | 
CLK_FILTER_NEAREST;
   // CHECK: [[smp_ptr:%[A-Za-z0-9_\.]+]] = alloca %opencl.sampler_t 
addrspace(2)*
-  // CHECK: [[SAMP:%[0-9]+]] = call %opencl.sampler_t addrspace(2)* 
@__translate_sampler_initializer(i32 19)
+  // CHECK: [[SAMP:%[0-9]+]] = call spir_func %opencl.sampler_t addrspace(2)* 
@__translate_sampler_initializer(i32 19)
   // CHECK: store %opencl.sampler_t addrspace(2)* [[SAMP]], %opencl.sampler_t 
addrspace(2)** [[smp_ptr]]
 
   // Case 1b
@@ -56,12 +56,12 @@
 
   // Case 1a/2a
   fnc4smp(glb_smp);
-  // CHECK: [[SAMP:%[0-9]+]] = call %opencl.sampler_t addrspace(2)* 
@__translate_sampler_initializer(i32 35)
+  // CHECK: [[SAMP:%[0-9]+]] = call spir_func %opencl.sampler_t addrspace(2)* 
@__translate_sampler_initializer(i32 35)
   // CHECK: call spir_func void [[FUNCNAME]](%opencl.sampler_t addrspace(2)* 
[[SAMP]])
 
   // Case 1a/2c
   fnc4smp(glb_smp_const);
-  // CHECK: [[SAMP:%[0-9]+]] = call %opencl.sampler_t addrspace(2)* 
@__translate_sampler_initializer(i32 35)
+  // CHECK: [[SAMP:%[0-9]+]] = call spir_func %opencl.sampler_t addrspace(2)* 
@__translate_sampler_initializer(i32 35)
   // CHECK: call spir_func void [[FUNCNAME]](%opencl.sampler_t addrspace(2)* 
[[SAMP]])
 
   // Case 1c
@@ -70,12 +70,12 @@
   // CHECK: call spir_func void [[FUNCNAME]](%opencl.sampler_t addrspace(2)* 
[[SAMP]])
 
   fnc4smp(5);
-  // CHECK: [[SAMP:%[0-9]+]] = call %opencl.sampler_t addrspace(2)* 
@__translate_sampler_initializer(i32 5)
+  // CHECK: [[SAMP:%[0-9]+]] = call spir_func %opencl.sampler_t addrspace(2)* 
@__translate_sampler_initializer(i32 5)
   // CHECK: call spir_func void [[FUNCNAME]](%opencl.sampler_t addrspace(2)* 
[[SAMP]])
 
   const sampler_t const_smp = CLK_ADDRESS_CLAMP_TO_EDGE | 
CLK_NORMALIZED_COORDS_TRUE | CLK_FILTER_LINEAR;
   fnc4smp(const_smp);
-   // CHECK: [[CONST_SAMP:%[0-9]+]] = call %opencl.sampler_t addrspace(2)* 
@__translate_sampler_initializer(i32 35)
+   // CHECK: [[CONST_SAMP:%[0-9]+]] = call spir_func %opencl.sampler_t 
addrspace(2)* @__translate_sampler_initializer(i32 35)
   // CHECK: store %opencl.sampler_t addrspace(2)* [[CONST_SAMP]], 
%opencl.sampler_t addrspace(2)** [[CONST_SMP_PTR:%[a-zA-Z0-9]+]]
   fnc4smp(const_smp);
   // CHECK: [[SAMP:%[0-9]+]] = load %opencl.sampler_t addrspace(2)*, 
%opencl.sampler_t addrspace(2)** [[CONST_SMP_PTR]]
@@ -83,7 +83,7 @@
 
   constant sampler_t constant_smp = CLK_ADDRESS_CLAMP_TO_EDGE | 
CLK_NORMALIZED_COORDS_TRUE | CLK_FILTER_LINEAR;
   fnc4smp(constant_smp);
-  // CHECK: [[SAMP:%[0-9]+]] = call %opencl.sampler_t addrspace(2)* 
@__translate_sampler_initializer(i32 35)
+  // CHECK: [[SAMP:%[0-9]+]] = call spir_func %opencl.sampler_t addrspace(2)* 
@__translate_sampler_initializer(i32 35)
   // CHECK: call spir_func void [[FUNCNAME]](%opencl.sampler_t addrspace(2)* 
[[SAMP]])
 
   // TODO: enable sampler initialization with non-constant integer.
Index: clang/lib/CodeGen/CodeGenModule.cpp
===
--- clang/lib/CodeGen/CodeGenModule.cpp
+++ clang/lib/CodeGen/CodeGenModule.cpp
@@ -6258,15 +6258,17 @@
 
   return *SanStats;
 }
+
 llvm::Value *
 CodeGenModule::createOpenCLIntToSamplerConversion(const Expr *E,
   CodeGenFunction &CGF) {
   llvm::Constant *C = ConstantEmitter(CGF).emitAbstract(E, E->getType());
-  auto SamplerT = getOpenCLRuntime().getSamplerType(E->getType().getTypePtr());
-  auto FTy = llvm::FunctionType::get(SamplerT, {C->getType()}, false);
-  return CGF.Builder.CreateCall(CreateRuntimeFunction(FTy,
-"__translate_sampler_initializer"),
-{C});
+  auto *SamplerT = 
getOpenCLRuntime().getSamplerType(E->getType().getTypePtr());
+  auto *FTy = llvm::FunctionType::get(SamplerT, {C->getType()}, false);
+  auto *Call = CGF.Builder.CreateCall(
+  CreateRuntimeFunction(FTy, "__translate_sampler_initializer"), {C});
+  Call->setCallingConv(Call->getCalledFunction()->getCallingConv());
+  return Call;
 }
 
 CharUnits CodeGenModule::getNaturalPointeeTypeAlignment(


Index: clang/test/CodeGenOpenCL/sampler.cl
===
--- clang/test/CodeGenOpenCL/sampler.cl
+++ clang/test/CodeGenOpenCL/sampler.cl
@@ -39,7 +3

[PATCH] D98411: [OpenCL] Respect calling convention for builtin

[Luke Drummond via Phabricator via cfe-commits]

ldrumm added a comment.

Given this fixes a runtime crash bug, should it be backported to the relevant 
release branch?


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D98411/new/

https://reviews.llvm.org/D98411

___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[PATCH] D151701: [HIP] Add missing __hip_atomic_fetch_sub support

[Luke Drummond via Phabricator via cfe-commits]

ldrumm created this revision.
ldrumm added reviewers: yaxunl, nikic, gandhi21299.
Herald added a subscriber: StephenFan.
Herald added a project: All.
ldrumm requested review of this revision.
Herald added a project: clang.

The rest of the fetch/op intrinsics were added in e13246a2ec3 
 but sub 
was conspicuous by its absence.


Repository:
  rG LLVM Github Monorepo

https://reviews.llvm.org/D151701

Files:
  clang/include/clang/Basic/Builtins.def
  clang/lib/AST/Expr.cpp
  clang/lib/CodeGen/CGAtomic.cpp
  clang/lib/Sema/SemaChecking.cpp
  clang/test/CodeGenCUDA/atomic-ops.cu

Index: clang/test/CodeGenCUDA/atomic-ops.cu
===
--- clang/test/CodeGenCUDA/atomic-ops.cu
+++ clang/test/CodeGenCUDA/atomic-ops.cu
@@ -6,6 +6,7 @@
 // CHECK: cmpxchg weak ptr {{%[0-9]+}}, i32 {{%[0-9]+}}, i32 {{%[0-9]+}} syncscope("singlethread-one-as") monotonic monotonic, align 4
 // CHECK: atomicrmw xchg ptr {{%[0-9]+}}, i32 {{%[0-9]+}} syncscope("singlethread-one-as")
 // CHECK: atomicrmw add ptr {{%[0-9]+}}, i32 {{%[0-9]+}} syncscope("singlethread-one-as")
+// CHECK: atomicrmw sub ptr {{%[0-9]+}}, i32 {{%[0-9]+}} syncscope("singlethread-one-as")
 // CHECK: atomicrmw and ptr {{%[0-9]+}}, i32 {{%[0-9]+}} syncscope("singlethread-one-as")
 // CHECK: atomicrmw or ptr {{%[0-9]+}}, i32 {{%[0-9]+}} syncscope("singlethread-one-as")
 // CHECK: atomicrmw xor ptr {{%[0-9]+}}, i32 {{%[0-9]+}} syncscope("singlethread-one-as")
@@ -18,6 +19,7 @@
   flag = __hip_atomic_compare_exchange_weak(ptr, &val, desired, __ATOMIC_RELAXED, __ATOMIC_RELAXED, __HIP_MEMORY_SCOPE_SINGLETHREAD);
   val = __hip_atomic_exchange(ptr, val, __ATOMIC_RELAXED, __HIP_MEMORY_SCOPE_SINGLETHREAD);
   val = __hip_atomic_fetch_add(ptr, val, __ATOMIC_RELAXED, __HIP_MEMORY_SCOPE_SINGLETHREAD);
+  val = __hip_atomic_fetch_sub(ptr, val, __ATOMIC_RELAXED, __HIP_MEMORY_SCOPE_SINGLETHREAD);
   val = __hip_atomic_fetch_and(ptr, val, __ATOMIC_RELAXED, __HIP_MEMORY_SCOPE_SINGLETHREAD);
   val = __hip_atomic_fetch_or(ptr, val, __ATOMIC_RELAXED, __HIP_MEMORY_SCOPE_SINGLETHREAD);
   val = __hip_atomic_fetch_xor(ptr, val, __ATOMIC_RELAXED, __HIP_MEMORY_SCOPE_SINGLETHREAD);
@@ -42,6 +44,7 @@
 // CHECK: cmpxchg weak ptr {{%[0-9]+}}, i32 {{%[0-9]+}}, i32 {{%[0-9]+}} syncscope("wavefront-one-as") monotonic monotonic, align 4
 // CHECK: atomicrmw xchg ptr {{%[0-9]+}}, i32 {{%[0-9]+}} syncscope("wavefront-one-as")
 // CHECK: atomicrmw add ptr {{%[0-9]+}}, i32 {{%[0-9]+}} syncscope("wavefront-one-as")
+// CHECK: atomicrmw sub ptr {{%[0-9]+}}, i32 {{%[0-9]+}} syncscope("wavefront-one-as")
 // CHECK: atomicrmw and ptr {{%[0-9]+}}, i32 {{%[0-9]+}} syncscope("wavefront-one-as")
 // CHECK: atomicrmw or ptr {{%[0-9]+}}, i32 {{%[0-9]+}} syncscope("wavefront-one-as")
 // CHECK: atomicrmw xor ptr {{%[0-9]+}}, i32 {{%[0-9]+}} syncscope("wavefront-one-as")
@@ -54,6 +57,7 @@
   flag = __hip_atomic_compare_exchange_weak(ptr, &val, desired, __ATOMIC_RELAXED, __ATOMIC_RELAXED, __HIP_MEMORY_SCOPE_WAVEFRONT);
   val = __hip_atomic_exchange(ptr, val, __ATOMIC_RELAXED, __HIP_MEMORY_SCOPE_WAVEFRONT);
   val = __hip_atomic_fetch_add(ptr, val, __ATOMIC_RELAXED, __HIP_MEMORY_SCOPE_WAVEFRONT);
+  val = __hip_atomic_fetch_sub(ptr, val, __ATOMIC_RELAXED, __HIP_MEMORY_SCOPE_WAVEFRONT);
   val = __hip_atomic_fetch_and(ptr, val, __ATOMIC_RELAXED, __HIP_MEMORY_SCOPE_WAVEFRONT);
   val = __hip_atomic_fetch_or(ptr, val, __ATOMIC_RELAXED, __HIP_MEMORY_SCOPE_WAVEFRONT);
   val = __hip_atomic_fetch_xor(ptr, val, __ATOMIC_RELAXED, __HIP_MEMORY_SCOPE_WAVEFRONT);
@@ -78,6 +82,7 @@
 // CHECK: cmpxchg weak ptr {{%[0-9]+}}, i32 {{%[0-9]+}}, i32 {{%[0-9]+}} syncscope("workgroup-one-as") monotonic monotonic, align 4
 // CHECK: atomicrmw xchg ptr {{%[0-9]+}}, i32 {{%[0-9]+}} syncscope("workgroup-one-as")
 // CHECK: atomicrmw add ptr {{%[0-9]+}}, i32 {{%[0-9]+}} syncscope("workgroup-one-as")
+// CHECK: atomicrmw sub ptr {{%[0-9]+}}, i32 {{%[0-9]+}} syncscope("workgroup-one-as")
 // CHECK: atomicrmw and ptr {{%[0-9]+}}, i32 {{%[0-9]+}} syncscope("workgroup-one-as")
 // CHECK: atomicrmw or ptr {{%[0-9]+}}, i32 {{%[0-9]+}} syncscope("workgroup-one-as")
 // CHECK: atomicrmw xor ptr {{%[0-9]+}}, i32 {{%[0-9]+}} syncscope("workgroup-one-as")
@@ -89,6 +94,7 @@
   flag = __hip_atomic_compare_exchange_weak(ptr, &val, desired, __ATOMIC_RELAXED, __ATOMIC_RELAXED, __HIP_MEMORY_SCOPE_WORKGROUP);
   val = __hip_atomic_exchange(ptr, val, __ATOMIC_RELAXED, __HIP_MEMORY_SCOPE_WORKGROUP);
   val = __hip_atomic_fetch_add(ptr, val, __ATOMIC_RELAXED, __HIP_MEMORY_SCOPE_WORKGROUP);
+  val = __hip_atomic_fetch_sub(ptr, val, __ATOMIC_RELAXED, __HIP_MEMORY_SCOPE_WORKGROUP);
   val = __hip_atomic_fetch_and(ptr, val, __ATOMIC_RELAXED, __HIP_MEMORY_SCOPE_WORKGROUP);
   val = __hip_atomic_fetch_or(ptr, val, __ATOMIC_RELAXED, __HIP_MEMORY_SCOPE_WORKGROUP);
   val = __hip_atomic_fetch_xor(ptr, val, __ATOMIC_RELAXED, __HIP_MEMORY_SCOPE_WORKG

[PATCH] D151701: [HIP] Add missing __hip_atomic_fetch_sub support

[Luke Drummond via Phabricator via cfe-commits]

ldrumm added a comment.

In D151701#4380666 , @yaxunl wrote:

> HIP did not add fetch/sub since fetch/sub x can be trivially implemented 
> through fetch/add -x and performance-wise equivalent.

There is existing isel for `global_atomic_sub` for RDNA targets which means we 
can avoid a subtraction. I also have a patch for the hip runtime ready to go 
that uses the this new builtin. It should shave off an extra instruction.

  __global__ void test_natural_sub(int *data, int rhs) {
__hip_atomic_fetch_sub(data, rhs, __ATOMIC_RELAXED, 
__HIP_MEMORY_SCOPE_WAVEFRONT);
  }
  
  __global__ void test_kernel_neg_add_sub(int *data, int rhs) {
__hip_atomic_fetch_add(data, -rhs, __ATOMIC_RELAXED, 
__HIP_MEMORY_SCOPE_WAVEFRONT);
  }

->

   <_Z16test_natural_subPii>:   

  
  s_clause 0x1   // 
: BFA10001
  s_load_dword s2, s[4:5], 0x8   // 
0004: F482 FA08
  s_load_dwordx2 s[0:1], s[4:5], null// 
000C: F4040002 FA00
  v_mov_b32_e32 v0, 0// 
0014: 7E000280
  s_waitcnt lgkmcnt(0)   // 
0018: BF8CC07F
  v_mov_b32_e32 v1, s2   // 
001C: 7E020202
  global_atomic_sub v0, v1, s[0:1]   // 
0020: DCCC8000 0100
  s_endpgm   // 
0028: BF81 
  
  
  0100 <_Z23test_kernel_neg_add_subPii>:
  s_clause 0x1   // 
0100: BFA10001
  s_load_dword s2, s[4:5], 0x8   // 
0104: F482 FA08
  s_load_dwordx2 s[0:1], s[4:5], null// 
010C: F4040002 FA00
  v_mov_b32_e32 v0, 0// 
0114: 7E000280
  s_waitcnt lgkmcnt(0)   // 
0118: BF8CC07F
  s_sub_i32 s2, 0, s2// 
011C: 81820280
  v_mov_b32_e32 v1, s2   // 
0120: 7E020202
  global_atomic_add v0, v1, s[0:1]   // 
0124: DCC88000 0100
  s_endpgm   // 
012C: BF81

The backend has isel for this instruction, but the frontend will never generate 
it. I think this improves things


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D151701/new/

https://reviews.llvm.org/D151701

___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[PATCH] D151701: [HIP] Add missing __hip_atomic_fetch_sub support

[Luke Drummond via Phabricator via cfe-commits]

This revision was automatically updated to reflect the committed changes.
Closed by commit rGe3fbede7f3fd: [HIP] Add missing __hip_atomic_fetch_sub 
support (authored by ldrumm).

Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D151701/new/

https://reviews.llvm.org/D151701

Files:
  clang/include/clang/Basic/Builtins.def
  clang/lib/AST/Expr.cpp
  clang/lib/CodeGen/CGAtomic.cpp
  clang/lib/Sema/SemaChecking.cpp
  clang/test/CodeGenCUDA/atomic-ops.cu

Index: clang/test/CodeGenCUDA/atomic-ops.cu
===
--- clang/test/CodeGenCUDA/atomic-ops.cu
+++ clang/test/CodeGenCUDA/atomic-ops.cu
@@ -6,6 +6,7 @@
 // CHECK: cmpxchg weak ptr {{%[0-9]+}}, i32 {{%[0-9]+}}, i32 {{%[0-9]+}} syncscope("singlethread-one-as") monotonic monotonic, align 4
 // CHECK: atomicrmw xchg ptr {{%[0-9]+}}, i32 {{%[0-9]+}} syncscope("singlethread-one-as")
 // CHECK: atomicrmw add ptr {{%[0-9]+}}, i32 {{%[0-9]+}} syncscope("singlethread-one-as")
+// CHECK: atomicrmw sub ptr {{%[0-9]+}}, i32 {{%[0-9]+}} syncscope("singlethread-one-as")
 // CHECK: atomicrmw and ptr {{%[0-9]+}}, i32 {{%[0-9]+}} syncscope("singlethread-one-as")
 // CHECK: atomicrmw or ptr {{%[0-9]+}}, i32 {{%[0-9]+}} syncscope("singlethread-one-as")
 // CHECK: atomicrmw xor ptr {{%[0-9]+}}, i32 {{%[0-9]+}} syncscope("singlethread-one-as")
@@ -18,6 +19,7 @@
   flag = __hip_atomic_compare_exchange_weak(ptr, &val, desired, __ATOMIC_RELAXED, __ATOMIC_RELAXED, __HIP_MEMORY_SCOPE_SINGLETHREAD);
   val = __hip_atomic_exchange(ptr, val, __ATOMIC_RELAXED, __HIP_MEMORY_SCOPE_SINGLETHREAD);
   val = __hip_atomic_fetch_add(ptr, val, __ATOMIC_RELAXED, __HIP_MEMORY_SCOPE_SINGLETHREAD);
+  val = __hip_atomic_fetch_sub(ptr, val, __ATOMIC_RELAXED, __HIP_MEMORY_SCOPE_SINGLETHREAD);
   val = __hip_atomic_fetch_and(ptr, val, __ATOMIC_RELAXED, __HIP_MEMORY_SCOPE_SINGLETHREAD);
   val = __hip_atomic_fetch_or(ptr, val, __ATOMIC_RELAXED, __HIP_MEMORY_SCOPE_SINGLETHREAD);
   val = __hip_atomic_fetch_xor(ptr, val, __ATOMIC_RELAXED, __HIP_MEMORY_SCOPE_SINGLETHREAD);
@@ -42,6 +44,7 @@
 // CHECK: cmpxchg weak ptr {{%[0-9]+}}, i32 {{%[0-9]+}}, i32 {{%[0-9]+}} syncscope("wavefront-one-as") monotonic monotonic, align 4
 // CHECK: atomicrmw xchg ptr {{%[0-9]+}}, i32 {{%[0-9]+}} syncscope("wavefront-one-as")
 // CHECK: atomicrmw add ptr {{%[0-9]+}}, i32 {{%[0-9]+}} syncscope("wavefront-one-as")
+// CHECK: atomicrmw sub ptr {{%[0-9]+}}, i32 {{%[0-9]+}} syncscope("wavefront-one-as")
 // CHECK: atomicrmw and ptr {{%[0-9]+}}, i32 {{%[0-9]+}} syncscope("wavefront-one-as")
 // CHECK: atomicrmw or ptr {{%[0-9]+}}, i32 {{%[0-9]+}} syncscope("wavefront-one-as")
 // CHECK: atomicrmw xor ptr {{%[0-9]+}}, i32 {{%[0-9]+}} syncscope("wavefront-one-as")
@@ -54,6 +57,7 @@
   flag = __hip_atomic_compare_exchange_weak(ptr, &val, desired, __ATOMIC_RELAXED, __ATOMIC_RELAXED, __HIP_MEMORY_SCOPE_WAVEFRONT);
   val = __hip_atomic_exchange(ptr, val, __ATOMIC_RELAXED, __HIP_MEMORY_SCOPE_WAVEFRONT);
   val = __hip_atomic_fetch_add(ptr, val, __ATOMIC_RELAXED, __HIP_MEMORY_SCOPE_WAVEFRONT);
+  val = __hip_atomic_fetch_sub(ptr, val, __ATOMIC_RELAXED, __HIP_MEMORY_SCOPE_WAVEFRONT);
   val = __hip_atomic_fetch_and(ptr, val, __ATOMIC_RELAXED, __HIP_MEMORY_SCOPE_WAVEFRONT);
   val = __hip_atomic_fetch_or(ptr, val, __ATOMIC_RELAXED, __HIP_MEMORY_SCOPE_WAVEFRONT);
   val = __hip_atomic_fetch_xor(ptr, val, __ATOMIC_RELAXED, __HIP_MEMORY_SCOPE_WAVEFRONT);
@@ -78,6 +82,7 @@
 // CHECK: cmpxchg weak ptr {{%[0-9]+}}, i32 {{%[0-9]+}}, i32 {{%[0-9]+}} syncscope("workgroup-one-as") monotonic monotonic, align 4
 // CHECK: atomicrmw xchg ptr {{%[0-9]+}}, i32 {{%[0-9]+}} syncscope("workgroup-one-as")
 // CHECK: atomicrmw add ptr {{%[0-9]+}}, i32 {{%[0-9]+}} syncscope("workgroup-one-as")
+// CHECK: atomicrmw sub ptr {{%[0-9]+}}, i32 {{%[0-9]+}} syncscope("workgroup-one-as")
 // CHECK: atomicrmw and ptr {{%[0-9]+}}, i32 {{%[0-9]+}} syncscope("workgroup-one-as")
 // CHECK: atomicrmw or ptr {{%[0-9]+}}, i32 {{%[0-9]+}} syncscope("workgroup-one-as")
 // CHECK: atomicrmw xor ptr {{%[0-9]+}}, i32 {{%[0-9]+}} syncscope("workgroup-one-as")
@@ -89,6 +94,7 @@
   flag = __hip_atomic_compare_exchange_weak(ptr, &val, desired, __ATOMIC_RELAXED, __ATOMIC_RELAXED, __HIP_MEMORY_SCOPE_WORKGROUP);
   val = __hip_atomic_exchange(ptr, val, __ATOMIC_RELAXED, __HIP_MEMORY_SCOPE_WORKGROUP);
   val = __hip_atomic_fetch_add(ptr, val, __ATOMIC_RELAXED, __HIP_MEMORY_SCOPE_WORKGROUP);
+  val = __hip_atomic_fetch_sub(ptr, val, __ATOMIC_RELAXED, __HIP_MEMORY_SCOPE_WORKGROUP);
   val = __hip_atomic_fetch_and(ptr, val, __ATOMIC_RELAXED, __HIP_MEMORY_SCOPE_WORKGROUP);
   val = __hip_atomic_fetch_or(ptr, val, __ATOMIC_RELAXED, __HIP_MEMORY_SCOPE_WORKGROUP);
   val = __hip_atomic_fetch_xor(ptr, val, __ATOMIC_RELAXED, __HIP_MEMORY_SCOPE_WORKGROUP);
@@ -112,6 +118,7 @@
 // CHECK: cmpxchg weak ptr {{%[0-9]+}}, i32 {{%[0-9]+}}, i32 {{%[0-9]+}} syncscope("agent-one-as") monotonic monotonic, align

[PATCH] D98411: [OpenCL] Respect calling convention for builtin

[Luke Drummond via Phabricator via cfe-commits]

ldrumm added a comment.

In D98411#2628592 , @Anastasia wrote:

> In D98411#2626613 , @ldrumm wrote:
>
>> Given this fixes a runtime crash bug, should it be backported to the 
>> relevant release branch?
>
> Sure, thanks!

https://bugs.llvm.org/show_bug.cgi?id=49605


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D98411/new/

https://reviews.llvm.org/D98411

___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[PATCH] D98411: [OpenCL] Respect calling convention for builtin

[Luke Drummond via Phabricator via cfe-commits]

ldrumm added inline comments.



Comment at: clang/lib/CodeGen/CodeGenModule.cpp:6268
+  auto *FTy = llvm::FunctionType::get(SamplerT, {C->getType()}, false);
+  auto *Call = CGF.Builder.CreateCall(
+  CreateRuntimeFunction(FTy, "__translate_sampler_initializer"), {C});

Anastasia wrote:
> Btw I think we should be using `CodeGenFunction::EmitRuntimeCall` here 
> instead because it sets the CC correctly through `SPIRABIInfo`.
> 
Thanks. I'll prepare a separate patch.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D98411/new/

https://reviews.llvm.org/D98411

___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[PATCH] D98705: [OpenCL] Prefer CodeGenFunction::EmitRuntimeCall

[Luke Drummond via Phabricator via cfe-commits]

ldrumm created this revision.
ldrumm added reviewers: Anastasia, bader.
ldrumm added a project: clang.
Herald added a subscriber: yaxunl.
ldrumm requested review of this revision.
Herald added a subscriber: cfe-commits.

CodeGenFunction::EmitRuntimeCall automatically sets the right calling  │
convention for the callee so we can avoid setting it ourselves


Repository:
  rG LLVM Github Monorepo

https://reviews.llvm.org/D98705

Files:
  clang/lib/CodeGen/CodeGenModule.cpp


Index: clang/lib/CodeGen/CodeGenModule.cpp
===
--- clang/lib/CodeGen/CodeGenModule.cpp
+++ clang/lib/CodeGen/CodeGenModule.cpp
@@ -6265,9 +6265,8 @@
   llvm::Constant *C = ConstantEmitter(CGF).emitAbstract(E, E->getType());
   auto *SamplerT = 
getOpenCLRuntime().getSamplerType(E->getType().getTypePtr());
   auto *FTy = llvm::FunctionType::get(SamplerT, {C->getType()}, false);
-  auto *Call = CGF.Builder.CreateCall(
+  auto *Call = CGF.EmitRuntimeCall(
   CreateRuntimeFunction(FTy, "__translate_sampler_initializer"), {C});
-  Call->setCallingConv(Call->getCalledFunction()->getCallingConv());
   return Call;
 }
 


Index: clang/lib/CodeGen/CodeGenModule.cpp
===
--- clang/lib/CodeGen/CodeGenModule.cpp
+++ clang/lib/CodeGen/CodeGenModule.cpp
@@ -6265,9 +6265,8 @@
   llvm::Constant *C = ConstantEmitter(CGF).emitAbstract(E, E->getType());
   auto *SamplerT = getOpenCLRuntime().getSamplerType(E->getType().getTypePtr());
   auto *FTy = llvm::FunctionType::get(SamplerT, {C->getType()}, false);
-  auto *Call = CGF.Builder.CreateCall(
+  auto *Call = CGF.EmitRuntimeCall(
   CreateRuntimeFunction(FTy, "__translate_sampler_initializer"), {C});
-  Call->setCallingConv(Call->getCalledFunction()->getCallingConv());
   return Call;
 }
 
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[PATCH] D98705: [OpenCL] Prefer CodeGenFunction::EmitRuntimeCall

[Luke Drummond via Phabricator via cfe-commits]

This revision was landed with ongoing or failed builds.
This revision was automatically updated to reflect the committed changes.
Closed by commit rG440f6bdf34f4: [OpenCL][NFCI] Prefer 
CodeGenFunction::EmitRuntimeCall (authored by ldrumm).

Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D98705/new/

https://reviews.llvm.org/D98705

Files:
  clang/lib/CodeGen/CodeGenModule.cpp


Index: clang/lib/CodeGen/CodeGenModule.cpp
===
--- clang/lib/CodeGen/CodeGenModule.cpp
+++ clang/lib/CodeGen/CodeGenModule.cpp
@@ -6265,9 +6265,8 @@
   llvm::Constant *C = ConstantEmitter(CGF).emitAbstract(E, E->getType());
   auto *SamplerT = 
getOpenCLRuntime().getSamplerType(E->getType().getTypePtr());
   auto *FTy = llvm::FunctionType::get(SamplerT, {C->getType()}, false);
-  auto *Call = CGF.Builder.CreateCall(
+  auto *Call = CGF.EmitRuntimeCall(
   CreateRuntimeFunction(FTy, "__translate_sampler_initializer"), {C});
-  Call->setCallingConv(Call->getCalledFunction()->getCallingConv());
   return Call;
 }
 


Index: clang/lib/CodeGen/CodeGenModule.cpp
===
--- clang/lib/CodeGen/CodeGenModule.cpp
+++ clang/lib/CodeGen/CodeGenModule.cpp
@@ -6265,9 +6265,8 @@
   llvm::Constant *C = ConstantEmitter(CGF).emitAbstract(E, E->getType());
   auto *SamplerT = getOpenCLRuntime().getSamplerType(E->getType().getTypePtr());
   auto *FTy = llvm::FunctionType::get(SamplerT, {C->getType()}, false);
-  auto *Call = CGF.Builder.CreateCall(
+  auto *Call = CGF.EmitRuntimeCall(
   CreateRuntimeFunction(FTy, "__translate_sampler_initializer"), {C});
-  Call->setCallingConv(Call->getCalledFunction()->getCallingConv());
   return Call;
 }
 
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[PATCH] D111315: Workaround build error for mingw-g++

[Luke Drummond via Phabricator via cfe-commits]

ldrumm created this revision.
ldrumm added reviewers: v.g.vassilev, rsmith.
ldrumm added a project: clang.
Herald added a subscriber: mstorsjo.
ldrumm requested review of this revision.

Workaround build error for mingw-g++

  

mingw-g++ does not correctly support the full std::errc namespace as
worded in the standard[1]. As such, we cannot reliably use all names
therein. This patch changes the use of
`std::errc::state_not_recoverable`, to use portable error codes from the
portable llvm::errc equivalent.

  

[1] https://gcc.gnu.org/bugzilla/show_bug.cgi?id=71444


Repository:
  rG LLVM Github Monorepo

https://reviews.llvm.org/D111315

Files:
  clang/lib/Interpreter/Interpreter.cpp


Index: clang/lib/Interpreter/Interpreter.cpp
===
--- clang/lib/Interpreter/Interpreter.cpp
+++ clang/lib/Interpreter/Interpreter.cpp
@@ -31,6 +31,7 @@
 
 #include "llvm/IR/Module.h"
 #include "llvm/Support/Host.h"
+#include "llvm/Support/Errc.h"
 
 using namespace clang;
 
@@ -47,14 +48,14 @@
   // failed. Extract that job from the Compilation.
   const driver::JobList &Jobs = Compilation->getJobs();
   if (!Jobs.size() || !isa(*Jobs.begin()))
-return llvm::createStringError(std::errc::state_not_recoverable,
+return llvm::createStringError(llvm::errc::io_error,
"Driver initialization failed. "
"Unable to create a driver job");
 
   // The one job we find should be to invoke clang again.
   const driver::Command *Cmd = cast(&(*Jobs.begin()));
   if (llvm::StringRef(Cmd->getCreator().getName()) != "clang")
-return llvm::createStringError(std::errc::state_not_recoverable,
+return llvm::createStringError(llvm::errc::io_error,
"Driver initialization failed");
 
   return &Cmd->getArguments();
@@ -89,13 +90,13 @@
   // Create the actual diagnostics engine.
   Clang->createDiagnostics();
   if (!Clang->hasDiagnostics())
-return llvm::createStringError(std::errc::state_not_recoverable,
+return llvm::createStringError(llvm::errc::not_enough_memory,
"Initialization failed. "
"Unable to create diagnostics engine");
 
   DiagsBuffer->FlushDiagnostics(Clang->getDiagnostics());
   if (!Success)
-return llvm::createStringError(std::errc::state_not_recoverable,
+return llvm::createStringError(std::errc::io_error,
"Initialization failed. "
"Unable to flush diagnostics");
 
@@ -106,7 +107,7 @@
   Clang->setTarget(TargetInfo::CreateTargetInfo(
   Clang->getDiagnostics(), Clang->getInvocation().TargetOpts));
   if (!Clang->hasTarget())
-return llvm::createStringError(std::errc::state_not_recoverable,
+return llvm::createStringError(llvm::errc::not_supported,
"Initialization failed. "
"Target is missing");
 


Index: clang/lib/Interpreter/Interpreter.cpp
===
--- clang/lib/Interpreter/Interpreter.cpp
+++ clang/lib/Interpreter/Interpreter.cpp
@@ -31,6 +31,7 @@
 
 #include "llvm/IR/Module.h"
 #include "llvm/Support/Host.h"
+#include "llvm/Support/Errc.h"
 
 using namespace clang;
 
@@ -47,14 +48,14 @@
   // failed. Extract that job from the Compilation.
   const driver::JobList &Jobs = Compilation->getJobs();
   if (!Jobs.size() || !isa(*Jobs.begin()))
-return llvm::createStringError(std::errc::state_not_recoverable,
+return llvm::createStringError(llvm::errc::io_error,
"Driver initialization failed. "
"Unable to create a driver job");
 
   // The one job we find should be to invoke clang again.
   const driver::Command *Cmd = cast(&(*Jobs.begin()));
   if (llvm::StringRef(Cmd->getCreator().getName()) != "clang")
-return llvm::createStringError(std::errc::state_not_recoverable,
+return llvm::createStringError(llvm::errc::io_error,
"Driver initialization failed");
 
   return &Cmd->getArguments();
@@ -89,13 +90,13 @@
   // Create the actual diagnostics engine.
   Clang->createDiagnostics();
   if (!Clang->hasDiagnostics())
-return llvm::createStringError(std::errc::state_not_recoverable,
+return llvm::createStringError(llvm::errc::not_enough_memory,
"Initialization failed. "
"Unable to create diagnostics engine");
 
   DiagsBuffer->FlushDiagnostics(Clang->getDiagnostics());
   if (!Success)
-return llvm::createStringError(std::errc::state_not_recoverable,
+return llvm::createStringError(std::errc::io_error,
"Initialization failed. "
"Unable to flush diagnostics");
 
@@ -106,7 

[PATCH] D111315: Workaround build error for mingw-g++

[Luke Drummond via Phabricator via cfe-commits]

ldrumm updated this revision to Diff 377864.

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D111315/new/

https://reviews.llvm.org/D111315

Files:
  clang/lib/Interpreter/Interpreter.cpp


Index: clang/lib/Interpreter/Interpreter.cpp
===
--- clang/lib/Interpreter/Interpreter.cpp
+++ clang/lib/Interpreter/Interpreter.cpp
@@ -31,6 +31,7 @@
 
 #include "llvm/IR/Module.h"
 #include "llvm/Support/Host.h"
+#include "llvm/Support/Errc.h"
 
 using namespace clang;
 
@@ -47,14 +48,14 @@
   // failed. Extract that job from the Compilation.
   const driver::JobList &Jobs = Compilation->getJobs();
   if (!Jobs.size() || !isa(*Jobs.begin()))
-return llvm::createStringError(std::errc::state_not_recoverable,
+return llvm::createStringError(llvm::errc::io_error,
"Driver initialization failed. "
"Unable to create a driver job");
 
   // The one job we find should be to invoke clang again.
   const driver::Command *Cmd = cast(&(*Jobs.begin()));
   if (llvm::StringRef(Cmd->getCreator().getName()) != "clang")
-return llvm::createStringError(std::errc::state_not_recoverable,
+return llvm::createStringError(llvm::errc::io_error,
"Driver initialization failed");
 
   return &Cmd->getArguments();
@@ -89,13 +90,13 @@
   // Create the actual diagnostics engine.
   Clang->createDiagnostics();
   if (!Clang->hasDiagnostics())
-return llvm::createStringError(std::errc::state_not_recoverable,
+return llvm::createStringError(llvm::errc::not_enough_memory,
"Initialization failed. "
"Unable to create diagnostics engine");
 
   DiagsBuffer->FlushDiagnostics(Clang->getDiagnostics());
   if (!Success)
-return llvm::createStringError(std::errc::state_not_recoverable,
+return llvm::createStringError(llvm::errc::io_error,
"Initialization failed. "
"Unable to flush diagnostics");
 
@@ -106,7 +107,7 @@
   Clang->setTarget(TargetInfo::CreateTargetInfo(
   Clang->getDiagnostics(), Clang->getInvocation().TargetOpts));
   if (!Clang->hasTarget())
-return llvm::createStringError(std::errc::state_not_recoverable,
+return llvm::createStringError(llvm::errc::not_supported,
"Initialization failed. "
"Target is missing");
 


Index: clang/lib/Interpreter/Interpreter.cpp
===
--- clang/lib/Interpreter/Interpreter.cpp
+++ clang/lib/Interpreter/Interpreter.cpp
@@ -31,6 +31,7 @@
 
 #include "llvm/IR/Module.h"
 #include "llvm/Support/Host.h"
+#include "llvm/Support/Errc.h"
 
 using namespace clang;
 
@@ -47,14 +48,14 @@
   // failed. Extract that job from the Compilation.
   const driver::JobList &Jobs = Compilation->getJobs();
   if (!Jobs.size() || !isa(*Jobs.begin()))
-return llvm::createStringError(std::errc::state_not_recoverable,
+return llvm::createStringError(llvm::errc::io_error,
"Driver initialization failed. "
"Unable to create a driver job");
 
   // The one job we find should be to invoke clang again.
   const driver::Command *Cmd = cast(&(*Jobs.begin()));
   if (llvm::StringRef(Cmd->getCreator().getName()) != "clang")
-return llvm::createStringError(std::errc::state_not_recoverable,
+return llvm::createStringError(llvm::errc::io_error,
"Driver initialization failed");
 
   return &Cmd->getArguments();
@@ -89,13 +90,13 @@
   // Create the actual diagnostics engine.
   Clang->createDiagnostics();
   if (!Clang->hasDiagnostics())
-return llvm::createStringError(std::errc::state_not_recoverable,
+return llvm::createStringError(llvm::errc::not_enough_memory,
"Initialization failed. "
"Unable to create diagnostics engine");
 
   DiagsBuffer->FlushDiagnostics(Clang->getDiagnostics());
   if (!Success)
-return llvm::createStringError(std::errc::state_not_recoverable,
+return llvm::createStringError(llvm::errc::io_error,
"Initialization failed. "
"Unable to flush diagnostics");
 
@@ -106,7 +107,7 @@
   Clang->setTarget(TargetInfo::CreateTargetInfo(
   Clang->getDiagnostics(), Clang->getInvocation().TargetOpts));
   if (!Clang->hasTarget())
-return llvm::createStringError(std::errc::state_not_recoverable,
+return llvm::createStringError(llvm::errc::not_supported,
"Initialization failed. "
"Target is missing");
 
___
cfe-commits mailing list
cfe-commits@lists.ll

[PATCH] D111315: Workaround build error for mingw-g++

[Luke Drummond via Phabricator via cfe-commits]

ldrumm added a comment.

In D111315#3048459 , @v.g.vassilev 
wrote:

> Can you map `state_not_recoverable` to `not_supported`?

Sure. I wasn't really sure that the io_error and not_enough_memory were 
appropriate. Thanks for the confirmation


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D111315/new/

https://reviews.llvm.org/D111315

___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[PATCH] D111315: Workaround build error for mingw-g++

[Luke Drummond via Phabricator via cfe-commits]

ldrumm updated this revision to Diff 377905.

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D111315/new/

https://reviews.llvm.org/D111315

Files:
  clang/lib/Interpreter/Interpreter.cpp


Index: clang/lib/Interpreter/Interpreter.cpp
===
--- clang/lib/Interpreter/Interpreter.cpp
+++ clang/lib/Interpreter/Interpreter.cpp
@@ -30,6 +30,7 @@
 #include "clang/Lex/PreprocessorOptions.h"
 
 #include "llvm/IR/Module.h"
+#include "llvm/Support/Errc.h"
 #include "llvm/Support/Host.h"
 
 using namespace clang;
@@ -47,14 +48,14 @@
   // failed. Extract that job from the Compilation.
   const driver::JobList &Jobs = Compilation->getJobs();
   if (!Jobs.size() || !isa(*Jobs.begin()))
-return llvm::createStringError(std::errc::state_not_recoverable,
+return llvm::createStringError(llvm::errc::not_supported,
"Driver initialization failed. "
"Unable to create a driver job");
 
   // The one job we find should be to invoke clang again.
   const driver::Command *Cmd = cast(&(*Jobs.begin()));
   if (llvm::StringRef(Cmd->getCreator().getName()) != "clang")
-return llvm::createStringError(std::errc::state_not_recoverable,
+return llvm::createStringError(llvm::errc::not_supported,
"Driver initialization failed");
 
   return &Cmd->getArguments();
@@ -89,13 +90,13 @@
   // Create the actual diagnostics engine.
   Clang->createDiagnostics();
   if (!Clang->hasDiagnostics())
-return llvm::createStringError(std::errc::state_not_recoverable,
+return llvm::createStringError(llvm::errc::not_supported,
"Initialization failed. "
"Unable to create diagnostics engine");
 
   DiagsBuffer->FlushDiagnostics(Clang->getDiagnostics());
   if (!Success)
-return llvm::createStringError(std::errc::state_not_recoverable,
+return llvm::createStringError(llvm::errc::not_supported,
"Initialization failed. "
"Unable to flush diagnostics");
 
@@ -106,7 +107,7 @@
   Clang->setTarget(TargetInfo::CreateTargetInfo(
   Clang->getDiagnostics(), Clang->getInvocation().TargetOpts));
   if (!Clang->hasTarget())
-return llvm::createStringError(std::errc::state_not_recoverable,
+return llvm::createStringError(llvm::errc::not_supported,
"Initialization failed. "
"Target is missing");
 


Index: clang/lib/Interpreter/Interpreter.cpp
===
--- clang/lib/Interpreter/Interpreter.cpp
+++ clang/lib/Interpreter/Interpreter.cpp
@@ -30,6 +30,7 @@
 #include "clang/Lex/PreprocessorOptions.h"
 
 #include "llvm/IR/Module.h"
+#include "llvm/Support/Errc.h"
 #include "llvm/Support/Host.h"
 
 using namespace clang;
@@ -47,14 +48,14 @@
   // failed. Extract that job from the Compilation.
   const driver::JobList &Jobs = Compilation->getJobs();
   if (!Jobs.size() || !isa(*Jobs.begin()))
-return llvm::createStringError(std::errc::state_not_recoverable,
+return llvm::createStringError(llvm::errc::not_supported,
"Driver initialization failed. "
"Unable to create a driver job");
 
   // The one job we find should be to invoke clang again.
   const driver::Command *Cmd = cast(&(*Jobs.begin()));
   if (llvm::StringRef(Cmd->getCreator().getName()) != "clang")
-return llvm::createStringError(std::errc::state_not_recoverable,
+return llvm::createStringError(llvm::errc::not_supported,
"Driver initialization failed");
 
   return &Cmd->getArguments();
@@ -89,13 +90,13 @@
   // Create the actual diagnostics engine.
   Clang->createDiagnostics();
   if (!Clang->hasDiagnostics())
-return llvm::createStringError(std::errc::state_not_recoverable,
+return llvm::createStringError(llvm::errc::not_supported,
"Initialization failed. "
"Unable to create diagnostics engine");
 
   DiagsBuffer->FlushDiagnostics(Clang->getDiagnostics());
   if (!Success)
-return llvm::createStringError(std::errc::state_not_recoverable,
+return llvm::createStringError(llvm::errc::not_supported,
"Initialization failed. "
"Unable to flush diagnostics");
 
@@ -106,7 +107,7 @@
   Clang->setTarget(TargetInfo::CreateTargetInfo(
   Clang->getDiagnostics(), Clang->getInvocation().TargetOpts));
   if (!Clang->hasTarget())
-return llvm::createStringError(std::errc::state_not_recoverable,
+return llvm::createStringError(llvm::errc::not_supported,
"Initialization failed. "
"Target is 

[PATCH] D111315: Workaround build error for mingw-g++

[Luke Drummond via Phabricator via cfe-commits]

This revision was landed with ongoing or failed builds.
This revision was automatically updated to reflect the committed changes.
Closed by commit rG6283d468e28b: Workaround build error for mingw-g++ (authored 
by ldrumm).

Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D111315/new/

https://reviews.llvm.org/D111315

Files:
  clang/lib/Interpreter/Interpreter.cpp


Index: clang/lib/Interpreter/Interpreter.cpp
===
--- clang/lib/Interpreter/Interpreter.cpp
+++ clang/lib/Interpreter/Interpreter.cpp
@@ -30,6 +30,7 @@
 #include "clang/Lex/PreprocessorOptions.h"
 
 #include "llvm/IR/Module.h"
+#include "llvm/Support/Errc.h"
 #include "llvm/Support/Host.h"
 
 using namespace clang;
@@ -47,14 +48,14 @@
   // failed. Extract that job from the Compilation.
   const driver::JobList &Jobs = Compilation->getJobs();
   if (!Jobs.size() || !isa(*Jobs.begin()))
-return llvm::createStringError(std::errc::state_not_recoverable,
+return llvm::createStringError(llvm::errc::not_supported,
"Driver initialization failed. "
"Unable to create a driver job");
 
   // The one job we find should be to invoke clang again.
   const driver::Command *Cmd = cast(&(*Jobs.begin()));
   if (llvm::StringRef(Cmd->getCreator().getName()) != "clang")
-return llvm::createStringError(std::errc::state_not_recoverable,
+return llvm::createStringError(llvm::errc::not_supported,
"Driver initialization failed");
 
   return &Cmd->getArguments();
@@ -89,13 +90,13 @@
   // Create the actual diagnostics engine.
   Clang->createDiagnostics();
   if (!Clang->hasDiagnostics())
-return llvm::createStringError(std::errc::state_not_recoverable,
+return llvm::createStringError(llvm::errc::not_supported,
"Initialization failed. "
"Unable to create diagnostics engine");
 
   DiagsBuffer->FlushDiagnostics(Clang->getDiagnostics());
   if (!Success)
-return llvm::createStringError(std::errc::state_not_recoverable,
+return llvm::createStringError(llvm::errc::not_supported,
"Initialization failed. "
"Unable to flush diagnostics");
 
@@ -106,7 +107,7 @@
   Clang->setTarget(TargetInfo::CreateTargetInfo(
   Clang->getDiagnostics(), Clang->getInvocation().TargetOpts));
   if (!Clang->hasTarget())
-return llvm::createStringError(std::errc::state_not_recoverable,
+return llvm::createStringError(llvm::errc::not_supported,
"Initialization failed. "
"Target is missing");
 


Index: clang/lib/Interpreter/Interpreter.cpp
===
--- clang/lib/Interpreter/Interpreter.cpp
+++ clang/lib/Interpreter/Interpreter.cpp
@@ -30,6 +30,7 @@
 #include "clang/Lex/PreprocessorOptions.h"
 
 #include "llvm/IR/Module.h"
+#include "llvm/Support/Errc.h"
 #include "llvm/Support/Host.h"
 
 using namespace clang;
@@ -47,14 +48,14 @@
   // failed. Extract that job from the Compilation.
   const driver::JobList &Jobs = Compilation->getJobs();
   if (!Jobs.size() || !isa(*Jobs.begin()))
-return llvm::createStringError(std::errc::state_not_recoverable,
+return llvm::createStringError(llvm::errc::not_supported,
"Driver initialization failed. "
"Unable to create a driver job");
 
   // The one job we find should be to invoke clang again.
   const driver::Command *Cmd = cast(&(*Jobs.begin()));
   if (llvm::StringRef(Cmd->getCreator().getName()) != "clang")
-return llvm::createStringError(std::errc::state_not_recoverable,
+return llvm::createStringError(llvm::errc::not_supported,
"Driver initialization failed");
 
   return &Cmd->getArguments();
@@ -89,13 +90,13 @@
   // Create the actual diagnostics engine.
   Clang->createDiagnostics();
   if (!Clang->hasDiagnostics())
-return llvm::createStringError(std::errc::state_not_recoverable,
+return llvm::createStringError(llvm::errc::not_supported,
"Initialization failed. "
"Unable to create diagnostics engine");
 
   DiagsBuffer->FlushDiagnostics(Clang->getDiagnostics());
   if (!Success)
-return llvm::createStringError(std::errc::state_not_recoverable,
+return llvm::createStringError(llvm::errc::not_supported,
"Initialization failed. "
"Unable to flush diagnostics");
 
@@ -106,7 +107,7 @@
   Clang->setTarget(TargetInfo::CreateTargetInfo(
   Clang->getDiagnostics(), Clang->getInvocation().TargetOpts));
   if (!Clang->hasTarget())
-return llvm::createS