summaryrefslogtreecommitdiffstats
path: root/dynamic-layers/clang-layer/recipes-devtools/clang/files/0001-OpenCL-Change-type-of-block-pointer-for-OpenCL.patch
diff options
context:
space:
mode:
Diffstat (limited to 'dynamic-layers/clang-layer/recipes-devtools/clang/files/0001-OpenCL-Change-type-of-block-pointer-for-OpenCL.patch')
-rw-r--r--dynamic-layers/clang-layer/recipes-devtools/clang/files/0001-OpenCL-Change-type-of-block-pointer-for-OpenCL.patch156
1 files changed, 156 insertions, 0 deletions
diff --git a/dynamic-layers/clang-layer/recipes-devtools/clang/files/0001-OpenCL-Change-type-of-block-pointer-for-OpenCL.patch b/dynamic-layers/clang-layer/recipes-devtools/clang/files/0001-OpenCL-Change-type-of-block-pointer-for-OpenCL.patch
new file mode 100644
index 00000000..1c491402
--- /dev/null
+++ b/dynamic-layers/clang-layer/recipes-devtools/clang/files/0001-OpenCL-Change-type-of-block-pointer-for-OpenCL.patch
@@ -0,0 +1,156 @@
1From 39a3ac0065c23d1e2d55dfd8792cc28a146a4307 Mon Sep 17 00:00:00 2001
2From: Alexey Bader <alexey.bader@intel.com>
3Date: Tue, 19 Feb 2019 15:19:06 +0000
4Subject: [PATCH 1/2] [OpenCL] Change type of block pointer for OpenCL
5
6Summary:
7
8For some reason OpenCL blocks in LLVM IR are represented as function pointers.
9These pointers do not point to any real function and never get called. Actually
10they point to some structure, which in turn contains pointer to the real block
11invoke function.
12This patch changes represntation of OpenCL blocks in LLVM IR from function
13pointers to pointers to `%struct.__block_literal_generic`.
14Such representation allows to avoid unnecessary bitcasts and simplifies
15further processing (e.g. translation to SPIR-V ) of the module for targets
16which do not support function pointers.
17
18Patch by: Alexey Sotkin.
19
20Reviewers: Anastasia, yaxunl, svenvh
21
22Reviewed By: Anastasia
23
24Subscribers: alexbatashev, cfe-commits
25
26Tags: #clang
27
28Differential Revision: https://reviews.llvm.org/D58277
29
30git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@354337 91177308-0d34-0410-b5e6-96231b3b80d8
31
32Upstream-Status: Backport
33[https://github.com/llvm-mirror/clang/commit/283f308bdb5893bab1f36791711346e746045f94]
34Signed-off-by: Anuj Mittal <anuj.mittal@intel.com>
35---
36 lib/CodeGen/CodeGenTypes.cpp | 4 +++-
37 test/CodeGenOpenCL/blocks.cl | 18 ++++++++----------
38 test/CodeGenOpenCL/cl20-device-side-enqueue.cl | 18 +++++++++---------
39 3 files changed, 20 insertions(+), 20 deletions(-)
40
41diff --git a/lib/CodeGen/CodeGenTypes.cpp b/lib/CodeGen/CodeGenTypes.cpp
42index 2acf1ac..93b3ebf 100644
43--- a/lib/CodeGen/CodeGenTypes.cpp
44+++ b/lib/CodeGen/CodeGenTypes.cpp
45@@ -637,7 +637,9 @@ llvm::Type *CodeGenTypes::ConvertType(QualType T) {
46
47 case Type::BlockPointer: {
48 const QualType FTy = cast<BlockPointerType>(Ty)->getPointeeType();
49- llvm::Type *PointeeType = ConvertTypeForMem(FTy);
50+ llvm::Type *PointeeType = CGM.getLangOpts().OpenCL
51+ ? CGM.getGenericBlockLiteralType()
52+ : ConvertTypeForMem(FTy);
53 unsigned AS = Context.getTargetAddressSpace(FTy);
54 ResultType = llvm::PointerType::get(PointeeType, AS);
55 break;
56diff --git a/test/CodeGenOpenCL/blocks.cl b/test/CodeGenOpenCL/blocks.cl
57index 675240c..19aacc3 100644
58--- a/test/CodeGenOpenCL/blocks.cl
59+++ b/test/CodeGenOpenCL/blocks.cl
60@@ -35,11 +35,10 @@ void foo(){
61 // SPIR: %[[block_captured:.*]] = getelementptr inbounds <{ i32, i32, i8 addrspace(4)*, i32 }>, <{ i32, i32, i8 addrspace(4)*, i32 }>* %[[block]], i32 0, i32 3
62 // SPIR: %[[i_value:.*]] = load i32, i32* %i
63 // SPIR: store i32 %[[i_value]], i32* %[[block_captured]],
64- // SPIR: %[[blk_ptr:.*]] = bitcast <{ i32, i32, i8 addrspace(4)*, i32 }>* %[[block]] to i32 ()*
65- // SPIR: %[[blk_gen_ptr:.*]] = addrspacecast i32 ()* %[[blk_ptr]] to i32 () addrspace(4)*
66- // SPIR: store i32 () addrspace(4)* %[[blk_gen_ptr]], i32 () addrspace(4)** %[[block_B:.*]],
67- // SPIR: %[[blk_gen_ptr:.*]] = load i32 () addrspace(4)*, i32 () addrspace(4)** %[[block_B]]
68- // SPIR: %[[block_literal:.*]] = bitcast i32 () addrspace(4)* %[[blk_gen_ptr]] to %struct.__opencl_block_literal_generic addrspace(4)*
69+ // SPIR: %[[blk_ptr:.*]] = bitcast <{ i32, i32, i8 addrspace(4)*, i32 }>* %[[block]] to %struct.__opencl_block_literal_generic*
70+ // SPIR: %[[blk_gen_ptr:.*]] = addrspacecast %struct.__opencl_block_literal_generic* %[[blk_ptr]] to %struct.__opencl_block_literal_generic addrspace(4)*
71+ // SPIR: store %struct.__opencl_block_literal_generic addrspace(4)* %[[blk_gen_ptr]], %struct.__opencl_block_literal_generic addrspace(4)** %[[block_B:.*]],
72+ // SPIR: %[[block_literal:.*]] = load %struct.__opencl_block_literal_generic addrspace(4)*, %struct.__opencl_block_literal_generic addrspace(4)** %[[block_B]]
73 // SPIR: %[[invoke_addr:.*]] = getelementptr inbounds %struct.__opencl_block_literal_generic, %struct.__opencl_block_literal_generic addrspace(4)* %[[block_literal]], i32 0, i32 2
74 // SPIR: %[[blk_gen_ptr:.*]] = bitcast %struct.__opencl_block_literal_generic addrspace(4)* %[[block_literal]] to i8 addrspace(4)*
75 // SPIR: %[[invoke_func_ptr:.*]] = load i8 addrspace(4)*, i8 addrspace(4)* addrspace(4)* %[[invoke_addr]]
76@@ -50,11 +49,10 @@ void foo(){
77 // AMDGCN: %[[block_captured:.*]] = getelementptr inbounds <{ i32, i32, i8*, i32 }>, <{ i32, i32, i8*, i32 }> addrspace(5)* %[[block]], i32 0, i32 3
78 // AMDGCN: %[[i_value:.*]] = load i32, i32 addrspace(5)* %i
79 // AMDGCN: store i32 %[[i_value]], i32 addrspace(5)* %[[block_captured]],
80- // AMDGCN: %[[blk_ptr:.*]] = bitcast <{ i32, i32, i8*, i32 }> addrspace(5)* %[[block]] to i32 () addrspace(5)*
81- // AMDGCN: %[[blk_gen_ptr:.*]] = addrspacecast i32 () addrspace(5)* %[[blk_ptr]] to i32 ()*
82- // AMDGCN: store i32 ()* %[[blk_gen_ptr]], i32 ()* addrspace(5)* %[[block_B:.*]],
83- // AMDGCN: %[[blk_gen_ptr:.*]] = load i32 ()*, i32 ()* addrspace(5)* %[[block_B]]
84- // AMDGCN: %[[block_literal:.*]] = bitcast i32 ()* %[[blk_gen_ptr]] to %struct.__opencl_block_literal_generic*
85+ // AMDGCN: %[[blk_ptr:.*]] = bitcast <{ i32, i32, i8*, i32 }> addrspace(5)* %[[block]] to %struct.__opencl_block_literal_generic addrspace(5)*
86+ // AMDGCN: %[[blk_gen_ptr:.*]] = addrspacecast %struct.__opencl_block_literal_generic addrspace(5)* %[[blk_ptr]] to %struct.__opencl_block_literal_generic*
87+ // AMDGCN: store %struct.__opencl_block_literal_generic* %[[blk_gen_ptr]], %struct.__opencl_block_literal_generic* addrspace(5)* %[[block_B:.*]],
88+ // AMDGCN: %[[block_literal:.*]] = load %struct.__opencl_block_literal_generic*, %struct.__opencl_block_literal_generic* addrspace(5)* %[[block_B]]
89 // AMDGCN: %[[invoke_addr:.*]] = getelementptr inbounds %struct.__opencl_block_literal_generic, %struct.__opencl_block_literal_generic* %[[block_literal]], i32 0, i32 2
90 // AMDGCN: %[[blk_gen_ptr:.*]] = bitcast %struct.__opencl_block_literal_generic* %[[block_literal]] to i8*
91 // AMDGCN: %[[invoke_func_ptr:.*]] = load i8*, i8** %[[invoke_addr]]
92diff --git a/test/CodeGenOpenCL/cl20-device-side-enqueue.cl b/test/CodeGenOpenCL/cl20-device-side-enqueue.cl
93index 4732194..8445016 100644
94--- a/test/CodeGenOpenCL/cl20-device-side-enqueue.cl
95+++ b/test/CodeGenOpenCL/cl20-device-side-enqueue.cl
96@@ -11,7 +11,7 @@ typedef struct {int a;} ndrange_t;
97
98 // For a block global variable, first emit the block literal as a global variable, then emit the block variable itself.
99 // COMMON: [[BL_GLOBAL:@__block_literal_global[^ ]*]] = internal addrspace(1) constant { i32, i32, i8 addrspace(4)* } { i32 {{[0-9]+}}, i32 {{[0-9]+}}, i8 addrspace(4)* addrspacecast (i8* bitcast (void (i8 addrspace(4)*, i8 addrspace(3)*)* [[INV_G:@[^ ]+]] to i8*) to i8 addrspace(4)*) }
100-// COMMON: @block_G = addrspace(1) constant void (i8 addrspace(3)*) addrspace(4)* addrspacecast (void (i8 addrspace(3)*) addrspace(1)* bitcast ({ i32, i32, i8 addrspace(4)* } addrspace(1)* [[BL_GLOBAL]] to void (i8 addrspace(3)*) addrspace(1)*) to void (i8 addrspace(3)*) addrspace(4)*)
101+// COMMON: @block_G = addrspace(1) constant %struct.__opencl_block_literal_generic addrspace(4)* addrspacecast (%struct.__opencl_block_literal_generic addrspace(1)* bitcast ({ i32, i32, i8 addrspace(4)* } addrspace(1)* [[BL_GLOBAL]] to %struct.__opencl_block_literal_generic addrspace(1)*) to %struct.__opencl_block_literal_generic addrspace(4)*)
102
103 // For anonymous blocks without captures, emit block literals as global variable.
104 // COMMON: [[BLG1:@__block_literal_global[^ ]*]] = internal addrspace(1) constant { i32, i32, i8 addrspace(4)* } { i32 {{[0-9]+}}, i32 {{[0-9]+}}, i8 addrspace(4)* addrspacecast (i8* bitcast (void (i8 addrspace(4)*, i8 addrspace(3)*)* {{@[^ ]+}} to i8*) to i8 addrspace(4)*) }
105@@ -77,9 +77,9 @@ kernel void device_side_enqueue(global int *a, global int *b, int i) {
106 // COMMON: [[DEF_Q:%[0-9]+]] = load %opencl.queue_t{{.*}}*, %opencl.queue_t{{.*}}** %default_queue
107 // COMMON: [[FLAGS:%[0-9]+]] = load i32, i32* %flags
108 // COMMON: store i8 addrspace(4)* addrspacecast (i8* bitcast (void (i8 addrspace(4)*)* [[INVL1:@__device_side_enqueue_block_invoke[^ ]*]] to i8*) to i8 addrspace(4)*), i8 addrspace(4)** %block.invoke
109- // B32: [[BL:%[0-9]+]] = bitcast <{ i32, i32, i8 addrspace(4)*, i32 addrspace(1)*, i32, i32 addrspace(1)* }>* %block to void ()*
110- // B64: [[BL:%[0-9]+]] = bitcast <{ i32, i32, i8 addrspace(4)*, i32 addrspace(1)*, i32 addrspace(1)*, i32 }>* %block to void ()*
111- // COMMON: [[BL_I8:%[0-9]+]] = addrspacecast void ()* [[BL]] to i8 addrspace(4)*
112+ // B32: [[BL:%[0-9]+]] = bitcast <{ i32, i32, i8 addrspace(4)*, i32 addrspace(1)*, i32, i32 addrspace(1)* }>* %block to %struct.__opencl_block_literal_generic*
113+ // B64: [[BL:%[0-9]+]] = bitcast <{ i32, i32, i8 addrspace(4)*, i32 addrspace(1)*, i32 addrspace(1)*, i32 }>* %block to %struct.__opencl_block_literal_generic*
114+ // COMMON: [[BL_I8:%[0-9]+]] = addrspacecast %struct.__opencl_block_literal_generic* [[BL]] to i8 addrspace(4)*
115 // COMMON-LABEL: call i32 @__enqueue_kernel_basic(
116 // COMMON-SAME: %opencl.queue_t{{.*}}* [[DEF_Q]], i32 [[FLAGS]], %struct.ndrange_t* byval [[NDR]]{{([0-9]+)?}},
117 // COMMON-SAME: i8 addrspace(4)* addrspacecast (i8* bitcast ({{.*}} [[INVLK1:[^ ]+_kernel]] to i8*) to i8 addrspace(4)*),
118@@ -95,8 +95,8 @@ kernel void device_side_enqueue(global int *a, global int *b, int i) {
119 // COMMON: [[WAIT_EVNT:%[0-9]+]] = addrspacecast %opencl.clk_event_t{{.*}}** %event_wait_list to %opencl.clk_event_t{{.*}}* addrspace(4)*
120 // COMMON: [[EVNT:%[0-9]+]] = addrspacecast %opencl.clk_event_t{{.*}}** %clk_event to %opencl.clk_event_t{{.*}}* addrspace(4)*
121 // COMMON: store i8 addrspace(4)* addrspacecast (i8* bitcast (void (i8 addrspace(4)*)* [[INVL2:@__device_side_enqueue_block_invoke[^ ]*]] to i8*) to i8 addrspace(4)*), i8 addrspace(4)** %block.invoke
122- // COMMON: [[BL:%[0-9]+]] = bitcast <{ i32, i32, i8 addrspace(4)*, i32{{.*}}, i32{{.*}}, i32{{.*}} }>* %block3 to void ()*
123- // COMMON: [[BL_I8:%[0-9]+]] = addrspacecast void ()* [[BL]] to i8 addrspace(4)*
124+ // COMMON: [[BL:%[0-9]+]] = bitcast <{ i32, i32, i8 addrspace(4)*, i32{{.*}}, i32{{.*}}, i32{{.*}} }>* %block3 to %struct.__opencl_block_literal_generic*
125+ // COMMON: [[BL_I8:%[0-9]+]] = addrspacecast %struct.__opencl_block_literal_generic* [[BL]] to i8 addrspace(4)*
126 // COMMON-LABEL: call i32 @__enqueue_kernel_basic_events
127 // COMMON-SAME: (%opencl.queue_t{{.*}}* [[DEF_Q]], i32 [[FLAGS]], %struct.ndrange_t* {{.*}}, i32 2, %opencl.clk_event_t{{.*}}* addrspace(4)* [[WAIT_EVNT]], %opencl.clk_event_t{{.*}}* addrspace(4)* [[EVNT]],
128 // COMMON-SAME: i8 addrspace(4)* addrspacecast (i8* bitcast ({{.*}} [[INVLK2:[^ ]+_kernel]] to i8*) to i8 addrspace(4)*),
129@@ -300,13 +300,13 @@ kernel void device_side_enqueue(global int *a, global int *b, int i) {
130 // Emits global block literal [[BLG8]] and invoke function [[INVG8]].
131 // The full type of these expressions are long (and repeated elsewhere), so we
132 // capture it as part of the regex for convenience and clarity.
133- // COMMON: store void () addrspace(4)* addrspacecast (void () addrspace(1)* bitcast ({ i32, i32, i8 addrspace(4)* } addrspace(1)* [[BLG8]] to void () addrspace(1)*) to void () addrspace(4)*), void () addrspace(4)** %block_A
134+ // COMMON: store %struct.__opencl_block_literal_generic addrspace(4)* addrspacecast (%struct.__opencl_block_literal_generic addrspace(1)* bitcast ({ i32, i32, i8 addrspace(4)* } addrspace(1)* [[BLG8]] to %struct.__opencl_block_literal_generic addrspace(1)*) to %struct.__opencl_block_literal_generic addrspace(4)*), %struct.__opencl_block_literal_generic addrspace(4)** %block_A
135 void (^const block_A)(void) = ^{
136 return;
137 };
138
139 // Emits global block literal [[BLG9]] and invoke function [[INVG9]].
140- // COMMON: store void (i8 addrspace(3)*) addrspace(4)* addrspacecast (void (i8 addrspace(3)*) addrspace(1)* bitcast ({ i32, i32, i8 addrspace(4)* } addrspace(1)* [[BLG9]] to void (i8 addrspace(3)*) addrspace(1)*) to void (i8 addrspace(3)*) addrspace(4)*), void (i8 addrspace(3)*) addrspace(4)** %block_B
141+ // COMMON: store %struct.__opencl_block_literal_generic addrspace(4)* addrspacecast (%struct.__opencl_block_literal_generic addrspace(1)* bitcast ({ i32, i32, i8 addrspace(4)* } addrspace(1)* [[BLG9]] to %struct.__opencl_block_literal_generic addrspace(1)*) to %struct.__opencl_block_literal_generic addrspace(4)*), %struct.__opencl_block_literal_generic addrspace(4)** %block_B
142 void (^const block_B)(local void *) = ^(local void *a) {
143 return;
144 };
145@@ -346,7 +346,7 @@ kernel void device_side_enqueue(global int *a, global int *b, int i) {
146 // COMMON: store i8 addrspace(4)* addrspacecast (i8* bitcast (void (i8 addrspace(4)*)* [[INVL3:@__device_side_enqueue_block_invoke[^ ]*]] to i8*) to i8 addrspace(4)*), i8 addrspace(4)** %block.invoke
147 // COMMON: [[DEF_Q:%[0-9]+]] = load %opencl.queue_t{{.*}}*, %opencl.queue_t{{.*}}** %default_queue
148 // COMMON: [[FLAGS:%[0-9]+]] = load i32, i32* %flags
149- // COMMON: [[BL_I8:%[0-9]+]] = addrspacecast void ()* {{.*}} to i8 addrspace(4)*
150+ // COMMON: [[BL_I8:%[0-9]+]] = addrspacecast %struct.__opencl_block_literal_generic* {{.*}} to i8 addrspace(4)*
151 // COMMON-LABEL: call i32 @__enqueue_kernel_basic(
152 // COMMON-SAME: %opencl.queue_t{{.*}}* [[DEF_Q]], i32 [[FLAGS]], %struct.ndrange_t* byval [[NDR]]{{([0-9]+)?}},
153 // COMMON-SAME: i8 addrspace(4)* addrspacecast (i8* bitcast ({{.*}} [[INVLK3:[^ ]+_kernel]] to i8*) to i8 addrspace(4)*),
154--
1551.8.3.1
156
generated by cgit v1.2.3-54-g00ecf (git 2.39.0) at 2024-04-25 09:34:28 +0000