Summary: The bit order of the has_vec and longtbtable bits in the traceback table generated by the XL compiler flipped at some point after v12.1. This is different from the definition is the AIX header debug.h. The change in the XL compiler that caused the deviation from the OS header definition was unintentional. Since both orderings are extant and the XL compiler runtime also expects the ordering defined by the OS, we will correct the output from LLVM to match the defined ordering given by the OS (which is also consistent with the Assembler Language Reference). Mitigation for traceback tables encoded with the wrong ordering is required for either ordering. Reviewers: XingXue, HubertTong Differential Revision: https://reviews.llvm.org/D105487
219 lines
12 KiB
LLVM
219 lines
12 KiB
LLVM
; RUN: llc -verify-machineinstrs -mtriple powerpc-ibm-aix-xcoff -mcpu=pwr4 \
|
|
; RUN: -mattr=-altivec -xcoff-traceback-table=true < %s | \
|
|
; RUN: FileCheck --check-prefixes=CHECK-ASM,COMMON %s
|
|
|
|
; RUN: llc -verify-machineinstrs -mtriple powerpc-ibm-aix-xcoff -function-sections \
|
|
; RUN: -mcpu=pwr4 -mattr=-altivec < %s | \
|
|
; RUN: FileCheck --check-prefixes=CHECK-FUNC,COMMON %s
|
|
|
|
|
|
%struct.S = type { i32, i32 }
|
|
%struct.D = type { float, double }
|
|
%struct.SD = type { %struct.S*, %struct.D }
|
|
|
|
@__const.main.s = private unnamed_addr constant %struct.S { i32 10, i32 20 }, align 4
|
|
@__const.main.d = private unnamed_addr constant %struct.D { float 1.000000e+01, double 2.000000e+01 }, align 8
|
|
|
|
define double @_Z10add_structifd1SP2SD1Di(i32 %value, float %fvalue, double %dvalue, %struct.S* byval(%struct.S) align 4 %s, %struct.SD* %dp, %struct.D* byval(%struct.D) align 4 %0, i32 %v2) #0 {
|
|
entry:
|
|
%d = alloca %struct.D, align 8
|
|
%value.addr = alloca i32, align 4
|
|
%fvalue.addr = alloca float, align 4
|
|
%dvalue.addr = alloca double, align 8
|
|
%dp.addr = alloca %struct.SD*, align 4
|
|
%v2.addr = alloca i32, align 4
|
|
%1 = bitcast %struct.D* %d to i8*
|
|
%2 = bitcast %struct.D* %0 to i8*
|
|
call void @llvm.memcpy.p0i8.p0i8.i32(i8* align 8 %1, i8* align 4 %2, i32 16, i1 false)
|
|
store i32 %value, i32* %value.addr, align 4
|
|
store float %fvalue, float* %fvalue.addr, align 4
|
|
store double %dvalue, double* %dvalue.addr, align 8
|
|
store %struct.SD* %dp, %struct.SD** %dp.addr, align 4
|
|
store i32 %v2, i32* %v2.addr, align 4
|
|
%3 = load double, double* %dvalue.addr, align 8
|
|
%4 = load float, float* %fvalue.addr, align 4
|
|
%conv = fpext float %4 to double
|
|
%add = fadd double %3, %conv
|
|
%5 = load i32, i32* %value.addr, align 4
|
|
%conv1 = sitofp i32 %5 to double
|
|
%add2 = fadd double %add, %conv1
|
|
%i1 = getelementptr inbounds %struct.S, %struct.S* %s, i32 0, i32 0
|
|
%6 = load i32, i32* %i1, align 4
|
|
%conv3 = sitofp i32 %6 to double
|
|
%add4 = fadd double %add2, %conv3
|
|
%7 = load %struct.SD*, %struct.SD** %dp.addr, align 4
|
|
%d5 = getelementptr inbounds %struct.SD, %struct.SD* %7, i32 0, i32 1
|
|
%d1 = getelementptr inbounds %struct.D, %struct.D* %d5, i32 0, i32 1
|
|
%8 = load double, double* %d1, align 8
|
|
%add6 = fadd double %add4, %8
|
|
%f1 = getelementptr inbounds %struct.D, %struct.D* %d, i32 0, i32 0
|
|
%9 = load float, float* %f1, align 8
|
|
%conv7 = fpext float %9 to double
|
|
%add8 = fadd double %add6, %conv7
|
|
%10 = load i32, i32* %v2.addr, align 4
|
|
%conv9 = sitofp i32 %10 to double
|
|
%add10 = fadd double %add8, %conv9
|
|
ret double %add10
|
|
}
|
|
|
|
declare void @llvm.memcpy.p0i8.p0i8.i32(i8* noalias nocapture writeonly, i8* noalias nocapture readonly, i32, i1 immarg) #1
|
|
|
|
define i32 @main() {
|
|
entry:
|
|
%retval = alloca i32, align 4
|
|
%s = alloca %struct.S, align 4
|
|
%d = alloca %struct.D, align 8
|
|
%sd = alloca %struct.SD, align 8
|
|
%agg.tmp = alloca %struct.S, align 4
|
|
%agg.tmp4 = alloca %struct.D, align 8
|
|
store i32 0, i32* %retval, align 4
|
|
%0 = bitcast %struct.S* %s to i8*
|
|
call void @llvm.memcpy.p0i8.p0i8.i32(i8* align 4 %0, i8* align 4 bitcast (%struct.S* @__const.main.s to i8*), i32 8, i1 false)
|
|
%1 = bitcast %struct.D* %d to i8*
|
|
call void @llvm.memcpy.p0i8.p0i8.i32(i8* align 8 %1, i8* align 8 bitcast (%struct.D* @__const.main.d to i8*), i32 16, i1 false)
|
|
%sp = getelementptr inbounds %struct.SD, %struct.SD* %sd, i32 0, i32 0
|
|
store %struct.S* %s, %struct.S** %sp, align 8
|
|
%d1 = getelementptr inbounds %struct.SD, %struct.SD* %sd, i32 0, i32 1
|
|
%f1 = getelementptr inbounds %struct.D, %struct.D* %d1, i32 0, i32 0
|
|
store float 1.000000e+02, float* %f1, align 8
|
|
%d2 = getelementptr inbounds %struct.SD, %struct.SD* %sd, i32 0, i32 1
|
|
%d13 = getelementptr inbounds %struct.D, %struct.D* %d2, i32 0, i32 1
|
|
store double 2.000000e+02, double* %d13, align 8
|
|
%2 = bitcast %struct.S* %agg.tmp to i8*
|
|
%3 = bitcast %struct.S* %s to i8*
|
|
call void @llvm.memcpy.p0i8.p0i8.i32(i8* align 4 %2, i8* align 4 %3, i32 8, i1 false)
|
|
%4 = bitcast %struct.D* %agg.tmp4 to i8*
|
|
%5 = bitcast %struct.D* %d to i8*
|
|
call void @llvm.memcpy.p0i8.p0i8.i32(i8* align 8 %4, i8* align 8 %5, i32 16, i1 false)
|
|
%call = call double @_Z10add_structifd1SP2SD1Di(i32 1, float 2.000000e+00, double 3.000000e+00, %struct.S* byval(%struct.S) align 4 %agg.tmp, %struct.SD* %sd, %struct.D* byval(%struct.D) align 4 %agg.tmp4, i32 7)
|
|
%add = fadd double %call, 1.000000e+00
|
|
%conv = fptosi double %add to i32
|
|
ret i32 %conv
|
|
}
|
|
|
|
define double @_Z7add_bari1SfdP2SD1Di(i32 %value, %struct.S* byval(%struct.S) align 4 %s, float %fvalue, double %dvalue, %struct.SD* %dp, %struct.D* byval(%struct.D) align 4 %0, i32 %v2) #0 {
|
|
entry:
|
|
%d = alloca %struct.D, align 8
|
|
%value.addr = alloca i32, align 4
|
|
%fvalue.addr = alloca float, align 4
|
|
%dvalue.addr = alloca double, align 8
|
|
%dp.addr = alloca %struct.SD*, align 4
|
|
%v2.addr = alloca i32, align 4
|
|
%1 = bitcast %struct.D* %d to i8*
|
|
%2 = bitcast %struct.D* %0 to i8*
|
|
call void @llvm.memcpy.p0i8.p0i8.i32(i8* align 8 %1, i8* align 4 %2, i32 16, i1 false)
|
|
store i32 %value, i32* %value.addr, align 4
|
|
store float %fvalue, float* %fvalue.addr, align 4
|
|
store double %dvalue, double* %dvalue.addr, align 8
|
|
store %struct.SD* %dp, %struct.SD** %dp.addr, align 4
|
|
store i32 %v2, i32* %v2.addr, align 4
|
|
%3 = load double, double* %dvalue.addr, align 8
|
|
%4 = load float, float* %fvalue.addr, align 4
|
|
%conv = fpext float %4 to double
|
|
%add = fadd double %3, %conv
|
|
%5 = load i32, i32* %value.addr, align 4
|
|
%conv1 = sitofp i32 %5 to double
|
|
%add2 = fadd double %add, %conv1
|
|
%i1 = getelementptr inbounds %struct.S, %struct.S* %s, i32 0, i32 0
|
|
%6 = load i32, i32* %i1, align 4
|
|
%conv3 = sitofp i32 %6 to double
|
|
%add4 = fadd double %add2, %conv3
|
|
%7 = load %struct.SD*, %struct.SD** %dp.addr, align 4
|
|
%d5 = getelementptr inbounds %struct.SD, %struct.SD* %7, i32 0, i32 1
|
|
%d1 = getelementptr inbounds %struct.D, %struct.D* %d5, i32 0, i32 1
|
|
%8 = load double, double* %d1, align 8
|
|
%add6 = fadd double %add4, %8
|
|
%f1 = getelementptr inbounds %struct.D, %struct.D* %d, i32 0, i32 0
|
|
%9 = load float, float* %f1, align 8
|
|
%conv7 = fpext float %9 to double
|
|
%add8 = fadd double %add6, %conv7
|
|
%10 = load i32, i32* %v2.addr, align 4
|
|
%conv9 = sitofp i32 %10 to double
|
|
%add10 = fadd double %add8, %conv9
|
|
ret double %add10
|
|
}
|
|
|
|
; CHECK-ASM-LABEL: ._Z10add_structifd1SP2SD1Di:{{[[:space:]] *}}# %bb.0:
|
|
; CHECK-FUNC-LABEL: csect ._Z10add_structifd1SP2SD1Di[PR],2{{[[:space:]] *}}# %bb.0:
|
|
; COMMON-NEXT: lwz 4, L..C0(2)
|
|
; COMMON-NEXT: stfs 1, -24(1)
|
|
; COMMON-NEXT: lfs 0, 0(4)
|
|
; COMMON-NEXT: lwz 4, 56(1)
|
|
; COMMON: fsub 0, 2, 0
|
|
; COMMON-NEXT: stw 9, -36(1)
|
|
; COMMON-NEXT: fadd 1, 1, 0
|
|
; COMMON-NEXT: blr
|
|
; COMMON-NEXT: L.._Z10add_structifd1SP2SD1Di0:
|
|
; COMMON-NEXT: .vbyte 4, 0x00000000 # Traceback table begin
|
|
; COMMON-NEXT: .byte 0x00 # Version = 0
|
|
; COMMON-NEXT: .byte 0x09 # Language = CPlusPlus
|
|
; COMMON-NEXT: .byte 0x22 # -IsGlobaLinkage, -IsOutOfLineEpilogOrPrologue
|
|
; COMMON-NEXT: # +HasTraceBackTableOffset, -IsInternalProcedure
|
|
; COMMON-NEXT: # -HasControlledStorage, -IsTOCless
|
|
; COMMON-NEXT: # +IsFloatingPointPresent
|
|
; COMMON-NEXT: # -IsFloatingPointOperationLogOrAbortEnabled
|
|
; COMMON-NEXT: .byte 0x40 # -IsInterruptHandler, +IsFunctionNamePresent, -IsAllocaUsed
|
|
; COMMON-NEXT: # OnConditionDirective = 0, -IsCRSaved, -IsLRSaved
|
|
; COMMON-NEXT: .byte 0x80 # +IsBackChainStored, -IsFixup, NumOfFPRsSaved = 0
|
|
; COMMON-NEXT: .byte 0x00 # -HasExtensionTable, -HasVectorInfo, NumOfGPRsSaved = 0
|
|
; COMMON-NEXT: .byte 0x05 # NumberOfFixedParms = 5
|
|
; COMMON-NEXT: .byte 0x05 # NumberOfFPParms = 2, +HasParmsOnStack
|
|
; COMMON-NEXT: .vbyte 4, 0x58000000 # Parameter type = i, f, d, i, i, i, i
|
|
; CHECK-ASM-NEXT: .vbyte 4, L.._Z10add_structifd1SP2SD1Di0-._Z10add_structifd1SP2SD1Di # Function size
|
|
; CHECK-FUNC-NEXT: .vbyte 4, L.._Z10add_structifd1SP2SD1Di0-._Z10add_structifd1SP2SD1Di[PR] # Function size
|
|
; COMMON-NEXT: .vbyte 2, 0x001a # Function name len = 26
|
|
; COMMON-NEXT: .byte "_Z10add_structifd1SP2SD1Di" # Function Name
|
|
; COMMON-NEXT: # -- End function
|
|
|
|
|
|
; CHECK-ASM-LABEL: .main:{{[[:space:]] *}}# %bb.0:
|
|
; CHECK-FUNC-LABEL: .csect .main[PR],2{{[[:space:]] *}}# %bb.0
|
|
; COMMON-NEXT: mflr 0
|
|
; COMMON-NEXT: stw 0, 8(1)
|
|
; COMMON: mtlr 0
|
|
; COMMON-NEXT: blr
|
|
; COMMON-NEXT: L..main0:
|
|
; COMMON-NEXT: .vbyte 4, 0x00000000 # Traceback table begin
|
|
; COMMON-NEXT: .byte 0x00 # Version = 0
|
|
; COMMON-NEXT: .byte 0x09 # Language = CPlusPlus
|
|
; COMMON-NEXT: .byte 0x22 # -IsGlobaLinkage, -IsOutOfLineEpilogOrPrologue
|
|
; COMMON-NEXT: # +HasTraceBackTableOffset, -IsInternalProcedure
|
|
; COMMON-NEXT: # -HasControlledStorage, -IsTOCless
|
|
; COMMON-NEXT: # +IsFloatingPointPresent
|
|
; COMMON-NEXT: # -IsFloatingPointOperationLogOrAbortEnabled
|
|
; COMMON-NEXT: .byte 0x41 # -IsInterruptHandler, +IsFunctionNamePresent, -IsAllocaUsed
|
|
; COMMON-NEXT: # OnConditionDirective = 0, -IsCRSaved, +IsLRSaved
|
|
; COMMON-NEXT: .byte 0x80 # +IsBackChainStored, -IsFixup, NumOfFPRsSaved = 0
|
|
; COMMON-NEXT: .byte 0x00 # -HasExtensionTable, -HasVectorInfo, NumOfGPRsSaved = 0
|
|
; COMMON-NEXT: .byte 0x00 # NumberOfFixedParms = 0
|
|
; COMMON-NEXT: .byte 0x01 # NumberOfFPParms = 0, +HasParmsOnStack
|
|
; CHECK-ASM-NEXT: .vbyte 4, L..main0-.main # Function size
|
|
; CHECK-FUNC-NEXT: .vbyte 4, L..main0-.main[PR] # Function size
|
|
; COMMON-NEXT: .vbyte 2, 0x0004 # Function name len = 4
|
|
; COMMON-NEXT: .byte "main" # Function Name
|
|
; COMMON-NEXT: # -- End function
|
|
|
|
|
|
; CHECK-ASM-LABEL: ._Z7add_bari1SfdP2SD1Di:{{[[:space:]] *}}# %bb.0:
|
|
; CHECK-FUNC-LABEL: .csect ._Z7add_bari1SfdP2SD1Di[PR],2{{[[:space:]] *}}# %bb.0:
|
|
; COMMON: .vbyte 4, 0x00000000 # Traceback table begin
|
|
; COMMON-NEXT: .byte 0x00 # Version = 0
|
|
; COMMON-NEXT: .byte 0x09 # Language = CPlusPlus
|
|
; COMMON-NEXT: .byte 0x22 # -IsGlobaLinkage, -IsOutOfLineEpilogOrPrologue
|
|
; COMMON-NEXT: # +HasTraceBackTableOffset, -IsInternalProcedure
|
|
; COMMON-NEXT: # -HasControlledStorage, -IsTOCless
|
|
; COMMON-NEXT: # +IsFloatingPointPresent
|
|
; COMMON-NEXT: # -IsFloatingPointOperationLogOrAbortEnabled
|
|
; COMMON-NEXT: .byte 0x40 # -IsInterruptHandler, +IsFunctionNamePresent, -IsAllocaUsed
|
|
; COMMON-NEXT: # OnConditionDirective = 0, -IsCRSaved, -IsLRSaved
|
|
; COMMON-NEXT: .byte 0x80 # +IsBackChainStored, -IsFixup, NumOfFPRsSaved = 0
|
|
; COMMON-NEXT: .byte 0x00 # -HasExtensionTable, -HasVectorInfo, NumOfGPRsSaved = 0
|
|
; COMMON-NEXT: .byte 0x05 # NumberOfFixedParms = 5
|
|
; COMMON-NEXT: .byte 0x05 # NumberOfFPParms = 2, +HasParmsOnStack
|
|
; COMMON-NEXT: .vbyte 4, 0x16000000 # Parameter type = i, i, i, f, d, i, i
|
|
; CHECK-ASM-NEXT: .vbyte 4, L.._Z7add_bari1SfdP2SD1Di0-._Z7add_bari1SfdP2SD1Di # Function size
|
|
; CHECK-FUNC-NEXT: .vbyte 4, L.._Z7add_bari1SfdP2SD1Di0-._Z7add_bari1SfdP2SD1Di[PR] # Function size
|
|
; COMMON-NEXT: .vbyte 2, 0x0016 # Function name len = 22
|
|
; COMMON-NEXT: .byte "_Z7add_bari1SfdP2SD1Di" # Function Name
|
|
; COMMON-NEXT: # -- End function
|