This is an archive of the discontinued LLVM Phabricator instance.

Differential D150192

Allow clang to emit inrange metadata when generating code for array subscripts
Needs ReviewPublic

Authored by simeon-imgtec on May 9 2023, 4:14 AM.

Details

Reviewers
pcc
eli.friedman
jdoerfert
Summary

As out-of-bounds array accesses are undefined in C/C++, we can capture this restriction using the newly introduced inrange attribute on individual GEP indices. The underlying motivation is that it would enable less conservative inferences in (basic) alias analysis, in turn improving passes such as GVN and LICM.

As the inrange keyword is currently only supported on constant GEPs, we extend the support to GEP instructions, and mofiy clang to emit the inrange flag in an array subscript scenario where it is safe to do so.

Diff Detail

Event Timeline

simeon-imgtec created this revision.May 9 2023, 4:14 AM
simeon-imgtec created this object with visibility "All Users".
Herald added a project: Restricted Project. · View Herald TranscriptMay 9 2023, 4:14 AM
Herald added subscribers: mattd, asavonic, jdoerfert and 3 others. · View Herald Transcript
simeon-imgtec requested review of this revision.May 9 2023, 4:14 AM
Herald added a reviewer: jdoerfert. · View Herald TranscriptMay 9 2023, 4:14 AM
Herald added projects: Restricted Project, Restricted Project. · View Herald Transcript
Herald added subscribers: llvm-commits, cfe-commits, jplehr, sstefan1. · View Herald Transcript
Harbormaster completed remote builds in B230838: Diff 520653.May 9 2023, 5:25 AM
simeon-imgtec added a child revision: Restricted Differential Revision.May 9 2023, 5:28 AM
efriedma added a subscriber: efriedma.May 9 2023, 11:17 AM

From what I recall, "inrange" is actually more restrictive than the normal C/C++ array indexing rules. Specifically, the bits regarding comparisons. "inrange" was designed to allow splitting globals indexed using inrange.

That isn't to say that functionality like this isn't useful, but we probably need to call it something different. And I'm not sure an attribute on a GEP is actually the best way to represent the semantics. It might make sense to use a separate intrinsic to represent this, especially given the direction of https://discourse.llvm.org/t/rfc-replacing-getelementptr-with-ptradd/68699 .

shiva0217 added a subscriber: shiva0217.May 10 2023, 3:46 AM
In D150192#4330199, @efriedma wrote:

From what I recall, "inrange" is actually more restrictive than the normal C/C++ array indexing rules. Specifically, the bits regarding comparisons. "inrange" was designed to allow splitting globals indexed using inrange.

That isn't to say that functionality like this isn't useful, but we probably need to call it something different. And I'm not sure an attribute on a GEP is actually the best way to represent the semantics. It might make sense to use a separate intrinsic to represent this, especially given the direction of https://discourse.llvm.org/t/rfc-replacing-getelementptr-with-ptradd/68699 .

Hi Eli,

Very appreciate your inputs. ^^
Is there any idea how the intrinsic should look like and how it associated with GEP before ptradd arrive?

efriedma changed the visibility from "All Users" to "Public (No Login Required)".May 10 2023, 9:28 AM
efriedma added a comment.May 10 2023, 9:52 AM

See D115274

simeon mentioned this in D152275: Use memory region declaration intrinsic when generating code for array subscripts.Jun 6 2023, 7:51 AM

Large Diff

This large diff affects 164 files. Files without inline comments have been collapsed. Expand All Files

Revision Contents

PathSize
clang/
lib/
CodeGen/
11 lines
test/
CodeGen/
2 lines
2 lines
X86/
4 lines
10 lines
2 lines
2 lines
4 lines
2 lines
6 lines
6 lines
CodeGenCXX/
2 lines
2 lines
2 lines
4 lines
8 lines
2 lines
CodeGenObjC/
2 lines
CodeGenOpenCL/
2 lines
OpenMP/
4 lines
2 lines
16 lines
16 lines
16 lines
16 lines
16 lines
32 lines
32 lines
32 lines
16 lines
32 lines
32 lines
32 lines
8 lines
16 lines
8 lines
46 lines
56 lines
8 lines
8 lines
4 lines
4 lines
4 lines
12 lines
4 lines
4 lines
20 lines
4 lines
4 lines
12 lines
48 lines
48 lines
20 lines
60 lines
40 lines
16 lines
20 lines
8 lines
8 lines
24 lines
4 lines
24 lines
8 lines
4 lines
8 lines
12 lines
30 lines
8 lines
16 lines
8 lines
12 lines
8 lines
8 lines
8 lines
8 lines
8 lines
160 lines
32 lines
8 lines
2 lines
32 lines
8 lines
40 lines
98 lines
48 lines
98 lines
96 lines
10 lines
4 lines
40 lines
40 lines
24 lines
28 lines
16 lines
16 lines
26 lines
24 lines
28 lines
32 lines
16 lines
32 lines
88 lines
32 lines
48 lines
56 lines
48 lines
32 lines
48 lines
132 lines
16 lines
44 lines
120 lines
48 lines
56 lines
32 lines
32 lines
32 lines
60 lines
4 lines
4 lines
20 lines
24 lines
28 lines
16 lines
16 lines
20 lines
24 lines
10 lines
28 lines
16 lines
16 lines
16 lines
88 lines
40 lines
48 lines
56 lines
32 lines
32 lines
32 lines
132 lines
16 lines
72 lines
48 lines
56 lines
32 lines
32 lines
32 lines
16 lines
16 lines
3584 lines
SemaOpenCL/
2 lines
utils/
update_cc_test_checks/
Inputs/
4 lines
4 lines
8 lines
4 lines
llvm/
include/
llvm/
IR/
10 lines
8 lines
lib/
AsmParser/
6 lines
Bitcode/
Reader/
8 lines
Writer/
8 lines
IR/
5 lines
1 line
12 lines
Transforms/
InstCombine/
23 lines

Diff 520653

clang/lib/CodeGen/CGExpr.cpp

clang/test/CodeGen/2005-01-02-ConstantInits.c

clang/test/CodeGen/2010-07-14-ref-off-end.c

clang/test/CodeGen/X86/va-arg-sse.c

clang/test/CodeGen/builtin-align-array.c

clang/test/CodeGen/exprs.c

clang/test/CodeGen/matrix-type-operators.c

clang/test/CodeGen/ubsan-pointer-overflow.c

clang/test/CodeGen/unaligned-expr.c

clang/test/CodeGen/union-tbaa1.c

clang/test/CodeGen/vla.c

clang/test/CodeGenCXX/amdgcn-automatic-variable.cpp

clang/test/CodeGenCXX/compound-literals.cpp

clang/test/CodeGenCXX/cxx1z-decomposition.cpp

clang/test/CodeGenCXX/lambda-expressions.cpp

clang/test/CodeGenCXX/no-odr-use.cpp

clang/test/CodeGenCXX/pr45964-decomp-transform.cpp

clang/test/CodeGenObjC/arc.m

clang/test/CodeGenOpenCL/amdgcn-automatic-variable.cl

clang/test/OpenMP/align_clause_codegen.cpp

clang/test/OpenMP/declare_mapper_codegen.cpp

clang/test/OpenMP/distribute_firstprivate_codegen.cpp

clang/test/OpenMP/distribute_lastprivate_codegen.cpp

clang/test/OpenMP/distribute_parallel_for_firstprivate_codegen.cpp

clang/test/OpenMP/distribute_parallel_for_lastprivate_codegen.cpp

clang/test/OpenMP/distribute_parallel_for_private_codegen.cpp

clang/test/OpenMP/distribute_parallel_for_simd_firstprivate_codegen.cpp

clang/test/OpenMP/distribute_parallel_for_simd_lastprivate_codegen.cpp

clang/test/OpenMP/distribute_parallel_for_simd_private_codegen.cpp

clang/test/OpenMP/distribute_private_codegen.cpp

clang/test/OpenMP/distribute_simd_firstprivate_codegen.cpp

clang/test/OpenMP/distribute_simd_lastprivate_codegen.cpp

clang/test/OpenMP/distribute_simd_private_codegen.cpp

clang/test/OpenMP/for_firstprivate_codegen.cpp

clang/test/OpenMP/for_lastprivate_codegen.cpp

clang/test/OpenMP/for_private_codegen.cpp

clang/test/OpenMP/for_reduction_codegen.cpp

clang/test/OpenMP/for_reduction_codegen_UDR.cpp

clang/test/OpenMP/for_scan_codegen.cpp

clang/test/OpenMP/for_simd_scan_codegen.cpp

clang/test/OpenMP/irbuilder_simd_aligned.cpp

clang/test/OpenMP/master_taskloop_reduction_codegen.cpp

clang/test/OpenMP/master_taskloop_simd_reduction_codegen.cpp

clang/test/OpenMP/nvptx_distribute_parallel_generic_mode_codegen.cpp

clang/test/OpenMP/nvptx_parallel_codegen.cpp

clang/test/OpenMP/nvptx_parallel_for_codegen.cpp

clang/test/OpenMP/nvptx_target_codegen.cpp

clang/test/OpenMP/nvptx_target_parallel_codegen.cpp

clang/test/OpenMP/nvptx_target_parallel_num_threads_codegen.cpp

clang/test/OpenMP/nvptx_target_parallel_proc_bind_codegen.cpp

clang/test/OpenMP/nvptx_target_simd_codegen.cpp

clang/test/OpenMP/nvptx_target_teams_distribute_parallel_for_codegen.cpp

clang/test/OpenMP/nvptx_target_teams_distribute_parallel_for_simd_codegen.cpp

clang/test/OpenMP/nvptx_target_teams_distribute_simd_codegen.cpp

clang/test/OpenMP/ordered_codegen.cpp

clang/test/OpenMP/parallel_copyin_codegen.cpp

clang/test/OpenMP/parallel_firstprivate_codegen.cpp

clang/test/OpenMP/parallel_for_scan_codegen.cpp

clang/test/OpenMP/parallel_for_simd_scan_codegen.cpp

clang/test/OpenMP/parallel_master_taskloop_firstprivate_codegen.cpp

clang/test/OpenMP/parallel_master_taskloop_reduction_codegen.cpp

clang/test/OpenMP/parallel_master_taskloop_simd_firstprivate_codegen.cpp

clang/test/OpenMP/parallel_master_taskloop_simd_lastprivate_codegen.cpp

clang/test/OpenMP/parallel_master_taskloop_simd_reduction_codegen.cpp

clang/test/OpenMP/parallel_private_codegen.cpp

clang/test/OpenMP/parallel_reduction_codegen.cpp

clang/test/OpenMP/reduction_implicit_map.cpp

clang/test/OpenMP/sections_firstprivate_codegen.cpp

clang/test/OpenMP/sections_lastprivate_codegen.cpp

clang/test/OpenMP/sections_private_codegen.cpp

clang/test/OpenMP/sections_reduction_codegen.cpp

clang/test/OpenMP/single_firstprivate_codegen.cpp

clang/test/OpenMP/single_private_codegen.cpp

clang/test/OpenMP/target_codegen.cpp

clang/test/OpenMP/target_data_use_device_addr_codegen.cpp

clang/test/OpenMP/target_defaultmap_codegen_01.cpp

clang/test/OpenMP/target_firstprivate_codegen.cpp

clang/test/OpenMP/target_has_device_addr_codegen.cpp

clang/test/OpenMP/target_has_device_addr_codegen_01.cpp

clang/test/OpenMP/target_map_codegen_08.cpp

clang/test/OpenMP/target_map_codegen_18.inc

clang/test/OpenMP/target_map_member_expr_array_section_codegen.cpp

clang/test/OpenMP/target_parallel_codegen.cpp

clang/test/OpenMP/target_parallel_debug_codegen.cpp

clang/test/OpenMP/target_parallel_for_codegen.cpp

clang/test/OpenMP/target_parallel_for_debug_codegen.cpp

clang/test/OpenMP/target_parallel_for_simd_codegen.cpp

clang/test/OpenMP/target_private_codegen.cpp

clang/test/OpenMP/target_reduction_codegen.cpp

clang/test/OpenMP/target_teams_codegen.cpp

clang/test/OpenMP/target_teams_distribute_codegen.cpp

clang/test/OpenMP/target_teams_distribute_collapse_codegen.cpp

clang/test/OpenMP/target_teams_distribute_dist_schedule_codegen.cpp

clang/test/OpenMP/target_teams_distribute_firstprivate_codegen.cpp

clang/test/OpenMP/target_teams_distribute_lastprivate_codegen.cpp

clang/test/OpenMP/target_teams_distribute_parallel_for_codegen.cpp

clang/test/OpenMP/target_teams_distribute_parallel_for_collapse_codegen.cpp

clang/test/OpenMP/target_teams_distribute_parallel_for_dist_schedule_codegen.cpp

clang/test/OpenMP/target_teams_distribute_parallel_for_firstprivate_codegen.cpp

clang/test/OpenMP/target_teams_distribute_parallel_for_lastprivate_codegen.cpp

clang/test/OpenMP/target_teams_distribute_parallel_for_private_codegen.cpp

clang/test/OpenMP/target_teams_distribute_parallel_for_schedule_codegen.cpp

clang/test/OpenMP/target_teams_distribute_parallel_for_simd_codegen.cpp

clang/test/OpenMP/target_teams_distribute_parallel_for_simd_collapse_codegen.cpp

clang/test/OpenMP/target_teams_distribute_parallel_for_simd_dist_schedule_codegen.cpp

clang/test/OpenMP/target_teams_distribute_parallel_for_simd_firstprivate_codegen.cpp

clang/test/OpenMP/target_teams_distribute_parallel_for_simd_lastprivate_codegen.cpp

clang/test/OpenMP/target_teams_distribute_parallel_for_simd_private_codegen.cpp

clang/test/OpenMP/target_teams_distribute_parallel_for_simd_schedule_codegen.cpp

clang/test/OpenMP/target_teams_distribute_private_codegen.cpp

clang/test/OpenMP/target_teams_distribute_reduction_codegen.cpp

clang/test/OpenMP/target_teams_distribute_simd_codegen.cpp

clang/test/OpenMP/target_teams_distribute_simd_collapse_codegen.cpp

clang/test/OpenMP/target_teams_distribute_simd_dist_schedule_codegen.cpp

clang/test/OpenMP/target_teams_distribute_simd_firstprivate_codegen.cpp

clang/test/OpenMP/target_teams_distribute_simd_lastprivate_codegen.cpp

clang/test/OpenMP/target_teams_distribute_simd_private_codegen.cpp

clang/test/OpenMP/task_codegen.cpp

clang/test/OpenMP/taskloop_reduction_codegen.cpp

clang/test/OpenMP/taskloop_simd_reduction_codegen.cpp

clang/test/OpenMP/teams_distribute_codegen.cpp

clang/test/OpenMP/teams_distribute_collapse_codegen.cpp

clang/test/OpenMP/teams_distribute_dist_schedule_codegen.cpp

clang/test/OpenMP/teams_distribute_firstprivate_codegen.cpp

clang/test/OpenMP/teams_distribute_lastprivate_codegen.cpp

clang/test/OpenMP/teams_distribute_parallel_for_codegen.cpp

clang/test/OpenMP/teams_distribute_parallel_for_collapse_codegen.cpp

clang/test/OpenMP/teams_distribute_parallel_for_copyin_codegen.cpp

clang/test/OpenMP/teams_distribute_parallel_for_dist_schedule_codegen.cpp

clang/test/OpenMP/teams_distribute_parallel_for_firstprivate_codegen.cpp

clang/test/OpenMP/teams_distribute_parallel_for_lastprivate_codegen.cpp

clang/test/OpenMP/teams_distribute_parallel_for_private_codegen.cpp

clang/test/OpenMP/teams_distribute_parallel_for_schedule_codegen.cpp

clang/test/OpenMP/teams_distribute_parallel_for_simd_codegen.cpp

clang/test/OpenMP/teams_distribute_parallel_for_simd_collapse_codegen.cpp

clang/test/OpenMP/teams_distribute_parallel_for_simd_dist_schedule_codegen.cpp

clang/test/OpenMP/teams_distribute_parallel_for_simd_firstprivate_codegen.cpp

clang/test/OpenMP/teams_distribute_parallel_for_simd_lastprivate_codegen.cpp

clang/test/OpenMP/teams_distribute_parallel_for_simd_private_codegen.cpp

clang/test/OpenMP/teams_distribute_parallel_for_simd_schedule_codegen.cpp

clang/test/OpenMP/teams_distribute_private_codegen.cpp

clang/test/OpenMP/teams_distribute_simd_codegen.cpp

clang/test/OpenMP/teams_distribute_simd_collapse_codegen.cpp

clang/test/OpenMP/teams_distribute_simd_dist_schedule_codegen.cpp

clang/test/OpenMP/teams_distribute_simd_firstprivate_codegen.cpp

clang/test/OpenMP/teams_distribute_simd_lastprivate_codegen.cpp

clang/test/OpenMP/teams_distribute_simd_private_codegen.cpp

clang/test/OpenMP/teams_firstprivate_codegen.cpp

clang/test/OpenMP/teams_private_codegen.cpp

clang/test/OpenMP/threadprivate_codegen.cpp

clang/test/SemaOpenCL/block-array-capturing.cl

clang/test/utils/update_cc_test_checks/Inputs/check-attributes.cpp.funcattrs.expected

clang/test/utils/update_cc_test_checks/Inputs/check-attributes.cpp.plain.expected

clang/test/utils/update_cc_test_checks/Inputs/generated-funcs.c.generated.expected

clang/test/utils/update_cc_test_checks/Inputs/generated-funcs.c.no-generated.expected

llvm/include/llvm/IR/Instructions.h

llvm/include/llvm/IR/Operator.h

llvm/lib/AsmParser/LLParser.cpp

llvm/lib/Bitcode/Reader/BitcodeReader.cpp

llvm/lib/Bitcode/Writer/BitcodeWriter.cpp

llvm/lib/IR/AsmWriter.cpp

llvm/lib/IR/Instruction.cpp

llvm/lib/IR/Instructions.cpp

llvm/lib/Transforms/InstCombine/InstructionCombining.cpp