aboutsummaryrefslogtreecommitdiff
path: root/lib/Target/AMDGPU
diff options
context:
space:
mode:
Diffstat (limited to 'lib/Target/AMDGPU')
-rw-r--r--lib/Target/AMDGPU/AMDGPU.h2
-rw-r--r--lib/Target/AMDGPU/AMDGPUAliasAnalysis.cpp64
-rw-r--r--lib/Target/AMDGPU/AMDGPUAliasAnalysis.h2
-rw-r--r--lib/Target/AMDGPU/AMDGPUISelDAGToDAG.cpp6
4 files changed, 41 insertions, 33 deletions
diff --git a/lib/Target/AMDGPU/AMDGPU.h b/lib/Target/AMDGPU/AMDGPU.h
index 796766d94622..2b49c2ea88e1 100644
--- a/lib/Target/AMDGPU/AMDGPU.h
+++ b/lib/Target/AMDGPU/AMDGPU.h
@@ -229,7 +229,7 @@ struct AMDGPUAS {
enum : unsigned {
// The maximum value for flat, generic, local, private, constant and region.
- MAX_COMMON_ADDRESS = 5,
+ MAX_AMDGPU_ADDRESS = 6,
GLOBAL_ADDRESS = 1, ///< Address space for global memory (RAT0, VTX0).
CONSTANT_ADDRESS = 4, ///< Address space for constant memory (VTX2)
diff --git a/lib/Target/AMDGPU/AMDGPUAliasAnalysis.cpp b/lib/Target/AMDGPU/AMDGPUAliasAnalysis.cpp
index ef4b69d09d9f..974fbcb87191 100644
--- a/lib/Target/AMDGPU/AMDGPUAliasAnalysis.cpp
+++ b/lib/Target/AMDGPU/AMDGPUAliasAnalysis.cpp
@@ -50,47 +50,51 @@ void AMDGPUAAWrapperPass::getAnalysisUsage(AnalysisUsage &AU) const {
AMDGPUAAResult::ASAliasRulesTy::ASAliasRulesTy(AMDGPUAS AS_, Triple::ArchType Arch_)
: Arch(Arch_), AS(AS_) {
// These arrarys are indexed by address space value
- // enum elements 0 ... to 5
- static const AliasResult ASAliasRulesPrivIsZero[6][6] = {
- /* Private Global Constant Group Flat Region*/
- /* Private */ {MayAlias, NoAlias , NoAlias , NoAlias , MayAlias, NoAlias},
- /* Global */ {NoAlias , MayAlias, NoAlias , NoAlias , MayAlias, NoAlias},
- /* Constant */ {NoAlias , NoAlias , MayAlias, NoAlias , MayAlias, NoAlias},
- /* Group */ {NoAlias , NoAlias , NoAlias , MayAlias, MayAlias, NoAlias},
- /* Flat */ {MayAlias, MayAlias, MayAlias, MayAlias, MayAlias, MayAlias},
- /* Region */ {NoAlias , NoAlias , NoAlias , NoAlias , MayAlias, MayAlias}
+ // enum elements 0 ... to 6
+ static const AliasResult ASAliasRulesPrivIsZero[7][7] = {
+ /* Private Global Constant Group Flat Region Constant 32-bit */
+ /* Private */ {MayAlias, NoAlias , NoAlias , NoAlias , MayAlias, NoAlias , NoAlias},
+ /* Global */ {NoAlias , MayAlias, MayAlias, NoAlias , MayAlias, NoAlias , MayAlias},
+ /* Constant */ {NoAlias , MayAlias, MayAlias, NoAlias , MayAlias, NoAlias , MayAlias},
+ /* Group */ {NoAlias , NoAlias , NoAlias , MayAlias, MayAlias, NoAlias , NoAlias},
+ /* Flat */ {MayAlias, MayAlias, MayAlias, MayAlias, MayAlias, MayAlias, MayAlias},
+ /* Region */ {NoAlias , NoAlias , NoAlias , NoAlias , MayAlias, MayAlias, NoAlias},
+ /* Constant 32-bit */ {NoAlias , MayAlias, MayAlias, NoAlias , MayAlias, NoAlias , MayAlias}
};
- static const AliasResult ASAliasRulesGenIsZero[6][6] = {
- /* Flat Global Region Group Constant Private */
- /* Flat */ {MayAlias, MayAlias, MayAlias, MayAlias, MayAlias, MayAlias},
- /* Global */ {MayAlias, MayAlias, NoAlias , NoAlias , NoAlias , NoAlias},
- /* Constant */ {MayAlias, NoAlias , MayAlias, NoAlias , NoAlias, NoAlias},
- /* Group */ {MayAlias, NoAlias , NoAlias , MayAlias, NoAlias , NoAlias},
- /* Region */ {MayAlias, NoAlias , NoAlias , NoAlias, MayAlias, NoAlias},
- /* Private */ {MayAlias, NoAlias , NoAlias , NoAlias , NoAlias , MayAlias}
+ static const AliasResult ASAliasRulesGenIsZero[7][7] = {
+ /* Flat Global Region Group Constant Private Constant 32-bit */
+ /* Flat */ {MayAlias, MayAlias, MayAlias, MayAlias, MayAlias, MayAlias, MayAlias},
+ /* Global */ {MayAlias, MayAlias, NoAlias , NoAlias , MayAlias, NoAlias , MayAlias},
+ /* Region */ {MayAlias, NoAlias , NoAlias , NoAlias, MayAlias, NoAlias , MayAlias},
+ /* Group */ {MayAlias, NoAlias , NoAlias , MayAlias, NoAlias , NoAlias , NoAlias},
+ /* Constant */ {MayAlias, MayAlias, MayAlias, NoAlias , NoAlias, NoAlias , MayAlias},
+ /* Private */ {MayAlias, NoAlias , NoAlias , NoAlias , NoAlias , MayAlias, NoAlias},
+ /* Constant 32-bit */ {MayAlias, MayAlias, MayAlias, NoAlias , MayAlias, NoAlias , NoAlias}
};
- assert(AS.MAX_COMMON_ADDRESS <= 5);
+ static_assert(AMDGPUAS::MAX_AMDGPU_ADDRESS <= 6, "Addr space out of range");
if (AS.FLAT_ADDRESS == 0) {
- assert(AS.GLOBAL_ADDRESS == 1 &&
- AS.REGION_ADDRESS == 2 &&
- AS.LOCAL_ADDRESS == 3 &&
- AS.CONSTANT_ADDRESS == 4 &&
- AS.PRIVATE_ADDRESS == 5);
+ assert(AS.GLOBAL_ADDRESS == 1 &&
+ AS.REGION_ADDRESS == 2 &&
+ AS.LOCAL_ADDRESS == 3 &&
+ AS.CONSTANT_ADDRESS == 4 &&
+ AS.PRIVATE_ADDRESS == 5 &&
+ AS.CONSTANT_ADDRESS_32BIT == 6);
ASAliasRules = &ASAliasRulesGenIsZero;
} else {
- assert(AS.PRIVATE_ADDRESS == 0 &&
- AS.GLOBAL_ADDRESS == 1 &&
- AS.CONSTANT_ADDRESS == 2 &&
- AS.LOCAL_ADDRESS == 3 &&
- AS.FLAT_ADDRESS == 4 &&
- AS.REGION_ADDRESS == 5);
+ assert(AS.PRIVATE_ADDRESS == 0 &&
+ AS.GLOBAL_ADDRESS == 1 &&
+ AS.CONSTANT_ADDRESS == 2 &&
+ AS.LOCAL_ADDRESS == 3 &&
+ AS.FLAT_ADDRESS == 4 &&
+ AS.REGION_ADDRESS == 5 &&
+ AS.CONSTANT_ADDRESS_32BIT == 6);
ASAliasRules = &ASAliasRulesPrivIsZero;
}
}
AliasResult AMDGPUAAResult::ASAliasRulesTy::getAliasResult(unsigned AS1,
unsigned AS2) const {
- if (AS1 > AS.MAX_COMMON_ADDRESS || AS2 > AS.MAX_COMMON_ADDRESS) {
+ if (AS1 > AS.MAX_AMDGPU_ADDRESS || AS2 > AS.MAX_AMDGPU_ADDRESS) {
if (Arch == Triple::amdgcn)
report_fatal_error("Pointer address space out of range");
return AS1 == AS2 ? MayAlias : NoAlias;
diff --git a/lib/Target/AMDGPU/AMDGPUAliasAnalysis.h b/lib/Target/AMDGPU/AMDGPUAliasAnalysis.h
index 645a38af753c..09ad51d5e42f 100644
--- a/lib/Target/AMDGPU/AMDGPUAliasAnalysis.h
+++ b/lib/Target/AMDGPU/AMDGPUAliasAnalysis.h
@@ -63,7 +63,7 @@ private:
private:
Triple::ArchType Arch;
AMDGPUAS AS;
- const AliasResult (*ASAliasRules)[6][6];
+ const AliasResult (*ASAliasRules)[7][7];
} ASAliasRules;
};
diff --git a/lib/Target/AMDGPU/AMDGPUISelDAGToDAG.cpp b/lib/Target/AMDGPU/AMDGPUISelDAGToDAG.cpp
index f25f4d4693ea..7cb0e12a6809 100644
--- a/lib/Target/AMDGPU/AMDGPUISelDAGToDAG.cpp
+++ b/lib/Target/AMDGPU/AMDGPUISelDAGToDAG.cpp
@@ -1451,7 +1451,11 @@ bool AMDGPUDAGToDAGISel::SelectSMRD(SDValue Addr, SDValue &SBase,
SDValue &Offset, bool &Imm) const {
SDLoc SL(Addr);
- if (CurDAG->isBaseWithConstantOffset(Addr)) {
+ // A 32-bit (address + offset) should not cause unsigned 32-bit integer
+ // wraparound, because s_load instructions perform the addition in 64 bits.
+ if ((Addr.getValueType() != MVT::i32 ||
+ Addr->getFlags().hasNoUnsignedWrap()) &&
+ CurDAG->isBaseWithConstantOffset(Addr)) {
SDValue N0 = Addr.getOperand(0);
SDValue N1 = Addr.getOperand(1);
generated by cgit v1.2.3 (git 2.25.1) at 2025-11-01 03:54:59 +0000