RPCS3/llvm-mirror
1
0
Fork 0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2024-11-23 19:23:23 +01:00
Code Issues Projects Releases Wiki Activity
9d5c56d7ba
llvm-mirror/test/Assembler
History
Sanjay Patel fd69991264 [IR] redefine 'UnsafeAlgebra' / 'reassoc' fast-math-flags and add 'trans' fast-math-flag 
As discussed on llvm-dev:
http://lists.llvm.org/pipermail/llvm-dev/2016-November/107104.html
and again more recently:
http://lists.llvm.org/pipermail/llvm-dev/2017-October/118118.html

...this is a step in cleaning up our fast-math-flags implementation in IR to better match
the capabilities of both clang's user-visible flags and the backend's flags for SDNode.

As proposed in the above threads, we're replacing the 'UnsafeAlgebra' bit (which had the 
'umbrella' meaning that all flags are set) with a new bit that only applies to algebraic 
reassociation - 'AllowReassoc'.

We're also adding a bit to allow approximations for library functions called 'ApproxFunc' 
(this was initially proposed as 'libm' or similar).

...and we're out of bits. 7 bits ought to be enough for anyone, right? :) FWIW, I did 
look at getting this out of SubclassOptionalData via SubclassData (spacious 16-bits), 
but that's apparently already used for other purposes. Also, I don't think we can just 
add a field to FPMathOperator because Operator is not intended to be instantiated. 
We'll defer movement of FMF to another day.

We keep the 'fast' keyword. I thought about removing that, but seeing IR like this:
%f.fast = fadd reassoc nnan ninf nsz arcp contract afn float %op1, %op2
...made me think we want to keep the shortcut synonym.

Finally, this change is binary incompatible with existing IR as seen in the 
compatibility tests. This statement:
"Newer releases can ignore features from older releases, but they cannot miscompile 
them. For example, if nsw is ever replaced with something else, dropping it would be 
a valid way to upgrade the IR." 
( http://llvm.org/docs/DeveloperPolicy.html#ir-backwards-compatibility )
...provides the flexibility we want to make this change without requiring a new IR 
version. Ie, we're not loosening the FP strictness of existing IR. At worst, we will 
fail to optimize some previously 'fast' code because it's no longer recognized as 
'fast'. This should get fixed as we audit/squash all of the uses of 'isFast()'.

Note: an inter-dependent clang commit to use the new API name should closely follow 
commit.

Differential Revision: https://reviews.llvm.org/D39304

llvm-svn: 317488
2017-11-06 16:27:15 +00:00
..
2002-03-08-NameCollision2.ll
2002-03-08-NameCollision.ll
2002-04-07-HexFloatConstants.ll
2002-04-07-InfConstant.ll
2002-04-29-NameBinding.ll
2002-05-02-InvalidForwardRef.ll
2002-07-14-OpaqueType.ll
2002-07-25-QuoteInString.ll
2002-07-25-ReturnPtrFunction.ll
2002-07-31-SlashInString.ll
2002-08-15-CastAmbiguity.ll
2002-08-15-ConstantExprProblem.ll
2002-08-15-UnresolvedGlobalReference.ll
2002-08-16-ConstExprInlined.ll
2002-08-19-BytecodeReader.ll
2002-08-22-DominanceProblem.ll
2002-10-08-LargeArrayPerformance.ll
2002-10-13-ConstantEncodingProblem.ll
2002-12-15-GlobalResolve.ll
2003-01-30-UnsignedString.ll
2003-04-15-ConstantInitAssertion.ll
2003-04-25-UnresolvedGlobalReference.ll
2003-05-03-BytecodeReaderProblem.ll
2003-05-12-MinIntProblem.ll
2003-05-15-AssemblerProblem.ll
2003-05-15-SwitchBug.ll
2003-05-21-ConstantShiftExpr.ll
2003-05-21-EmptyStructTest.ll
2003-05-21-MalformedShiftCrash.ll
2003-05-21-MalformedStructCrash.ll
2003-08-20-ConstantExprGEP-Fold.ll
2003-08-21-ConstantExprCast-Fold.ll
2003-11-05-ConstantExprShift.ll
2003-11-11-ImplicitRename.ll Make shell redirection construct portable 2017-07-12 13:24:46 +00:00
2003-11-12-ConstantExprCast.ll
2003-11-24-SymbolTableCrash.ll
2004-01-11-getelementptrfolding.ll
2004-01-20-MaxLongLong.ll
2004-02-01-NegativeZero.ll
2004-02-27-SelfUseAssertError.ll
2004-03-07-FunctionAddressAlignment.ll
2004-03-30-UnclosedFunctionCrash.ll
2004-04-04-GetElementPtrIndexTypes.ll
2004-06-07-VerifierBug.ll
2004-10-22-BCWriterUndefBug.ll
2004-11-28-InvalidTypeCrash.ll
2005-01-03-FPConstantDisassembly.ll
2005-01-31-CallingAggregateFunction.ll
2005-05-05-OpaqueUndefValues.ll
2005-12-21-ZeroInitVector.ll
2006-09-28-CrashOnInvalid.ll
2006-12-09-Cast-To-Bool.ll
2007-01-02-Undefined-Arg-Type.ll
2007-01-05-Cmp-ConstExpr.ll
2007-01-16-CrashOnBadCast2.ll
2007-01-16-CrashOnBadCast.ll
2007-03-18-InvalidNumberedVar.ll
2007-03-19-NegValue.ll
2007-04-20-AlignedLoad.ll
2007-04-20-AlignedStore.ll
2007-04-25-AssemblerFoldExternWeak.ll
2007-05-21-Escape.ll
2007-07-19-ParamAttrAmbiguity.ll
2007-08-06-AliasInvalid.ll
2007-09-10-AliasFwdRef.ll
2007-09-29-GC.ll
2007-11-26-AttributeOverload.ll Make shell redirection construct portable 2017-07-12 13:24:46 +00:00
2007-12-11-AddressSpaces.ll
2008-01-11-VarargAttrs.ll
2008-02-18-IntPointerCrash.ll
2008-07-10-APInt.ll
2008-09-02-FunctionNotes2.ll
2008-09-02-FunctionNotes.ll
2008-09-29-RetAttr.ll
2008-10-14-QuoteInName.ll
2009-02-01-UnnamedForwardRef.ll
2009-02-28-CastOpc.ll
2009-02-28-StripOpaqueName.ll
2009-03-24-ZextConstantExpr.ll
2009-07-24-ZeroArgGEP.ll
2010-02-05-FunctionLocalMetadataBecomesNull.ll
addrspacecast-alias.ll
aggregate-constant-values.ll
aggregate-return-single-value.ll
alias-redefinition.ll
alias-use-list-order.ll
align-inst-alloca.ll
align-inst-load.ll
align-inst-store.ll
align-inst.ll
alignstack.ll
alloca-addrspace0.ll Allow DataLayout to specify addrspace for allocas. 2017-04-10 22:27:50 +00:00
alloca-addrspace-elems.ll Fix assembler for alloca of multiple elements in non-zero addr space 2017-10-14 03:23:18 +00:00
alloca-addrspace-parse-error-0.ll Allow DataLayout to specify addrspace for allocas. 2017-04-10 22:27:50 +00:00
alloca-addrspace-parse-error-1.ll Allow DataLayout to specify addrspace for allocas. 2017-04-10 22:27:50 +00:00
alloca-invalid-type-2.ll
alloca-invalid-type.ll
alloca-size-one.ll
anon-functions.ll
atomic.ll Enhance synchscope representation 2017-07-11 22:23:00 +00:00
attribute-builtin.ll
auto_upgrade_intrinsics.ll Support arbitrary address space pointers in masked gather/scatter intrinsics. 2017-05-03 12:28:54 +00:00
auto_upgrade_nvvm_intrinsics.ll [NVPTX] Auto-upgrade some NVPTX intrinsics to LLVM target-generic code. 2017-01-21 01:00:32 +00:00
autoupgrade-thread-pointer.ll
bcwrap.ll
call-invalid-1.ll
comment.ll
ConstantExprFold.ll
ConstantExprFoldCast.ll
ConstantExprFoldSelect.ll
ConstantExprNoFold.ll
datalayout-alloca-addrspace-mismatch-0.ll Allow DataLayout to specify addrspace for allocas. 2017-04-10 22:27:50 +00:00
datalayout-alloca-addrspace-mismatch-1.ll Allow DataLayout to specify addrspace for allocas. 2017-04-10 22:27:50 +00:00
datalayout-alloca-addrspace-mismatch-2.ll Allow DataLayout to specify addrspace for allocas. 2017-04-10 22:27:50 +00:00
datalayout-alloca-addrspace.ll Allow DataLayout to specify addrspace for allocas. 2017-04-10 22:27:50 +00:00
debug-info.ll [DebugInfo] Emit address space with DW_AT_address_class attribute for pointer and reference types 2017-03-08 23:55:44 +00:00
dicompileunit.ll DebugInfo: Add flag to CU to disable emission of inline debug info into the skeleton CU 2016-08-24 18:29:49 +00:00
dicompositetype-members.ll
diexpression.ll Parse and print DIExpressions inline to ease IR and MIR testing 2017-08-23 20:31:27 +00:00
difile-escaped-chars.ll
diglobalvariable.ll [DI] Every DIGlobalVariable should have a type. 2017-08-16 13:39:07 +00:00
DIGlobalVariableExpression.ll Parse and print DIExpressions inline to ease IR and MIR testing 2017-08-23 20:31:27 +00:00
diimportedentity.ll Debug Info: Add a file: field to DIImportedEntity. 2017-07-19 00:09:54 +00:00
dilexicalblock.ll
dilocalvariable-arg-large.ll
dilocalvariable.ll DebugInfo: preparation to implement DW_AT_alignment 2016-10-20 00:13:12 +00:00
dilocation.ll
DIMacroFile.ll [LLParser] Make the line field of DIMacro(File) optional. 2016-12-22 00:29:00 +00:00
dimodule.ll
dinamespace.ll Remove line and file from DINamespace. 2017-04-28 22:25:46 +00:00
diobjcproperty.ll
distinct-mdnode.ll
disubprogram.ll Add support for DW_TAG_thrown_type. 2017-04-26 22:56:44 +00:00
disubrange-empty-array.ll
disubroutinetype.ll
ditemplateparameter.ll
ditype-large-values.ll DebugInfo: change alignment type from uint64_t to uint32_t to save space. 2016-10-18 14:31:22 +00:00
dllimport-dsolocal-diag.ll Represent runtime preemption in the IR. 2017-10-26 15:00:26 +00:00
drop-debug-info.ll
externally-initialized.ll
extractvalue-invalid-idx.ll
extractvalue-no-idx.ll
fast-math-flags.ll [IR] redefine 'UnsafeAlgebra' / 'reassoc' fast-math-flags and add 'trans' fast-math-flag 2017-11-06 16:27:15 +00:00
flags.ll
generic-debug-node.ll
getelementptr_invalid_ptr.ll
getelementptr_struct.ll
getelementptr_vec_ce2.ll [LLParser] Parse vector GEP constant expression correctly 2016-12-21 18:29:47 +00:00
getelementptr_vec_ce.ll [LLParser] Parse vector GEP constant expression correctly 2016-12-21 18:29:47 +00:00
getelementptr_vec_idx1.ll
getelementptr_vec_idx2.ll
getelementptr_vec_idx3.ll
getelementptr_vec_idx4.ll
getelementptr_vec_struct.ll
getelementptr.ll IR: Introduce inrange attribute on getelementptr indices. 2016-11-10 22:34:55 +00:00
getInt.ll
global-addrspace-forwardref.ll
globalvariable-attributes.ll [IR] Allow attributes with global variables 2017-05-11 12:28:08 +00:00
gv-invalid-type.ll
half-constprop.ll
half-conv.ll
half.ll
huge-array.ll
ifunc-asm.ll
ifunc-dsolocal-daig.ll Represent runtime preemption in the IR. 2017-10-26 15:00:26 +00:00
ifunc-use-list-order.ll
inalloca.ll
incorrect-tdep-attrs-parsing.ll
inline-asm-clobber.ll
insertextractvalue.ll
insertvalue-invalid-idx.ll
insertvalue-invalid-type-1.ll
insertvalue-invalid-type.ll
internal-hidden-alias.ll
internal-hidden-function.ll
internal-hidden-variable.ll
internal-protected-alias.ll
internal-protected-function.ll
internal-protected-variable.ll
invalid_cast2.ll
invalid_cast3.ll
invalid_cast4.ll
invalid_cast.ll
invalid-alias-mismatched-explicit-type.ll
invalid-attrgrp.ll
invalid-comdat2.ll
invalid-comdat.ll
invalid-datalayout1.ll
invalid-datalayout2.ll
invalid-datalayout3.ll
invalid-datalayout4.ll
invalid-datalayout5.ll
invalid-datalayout6.ll
invalid-datalayout7.ll
invalid-datalayout8.ll
invalid-datalayout9.ll
invalid-datalayout10.ll
invalid-datalayout11.ll
invalid-datalayout12.ll
invalid-datalayout13.ll
invalid-datalayout14.ll
invalid-datalayout15.ll
invalid-datalayout16.ll
invalid-datalayout17.ll
invalid-datalayout18.ll
invalid-datalayout19.ll
invalid-datalayout20.ll
invalid-datalayout21.ll
invalid-datalayout22.ll
invalid-datalayout23.ll
invalid-datalayout24.ll
invalid-datalayout-alloca-addrspace.ll Allow DataLayout to specify addrspace for allocas. 2017-04-10 22:27:50 +00:00
invalid-debug-info-version.ll
invalid-dicompileunit-emissionkind-bad.ll
invalid-dicompileunit-language-bad.ll
invalid-dicompileunit-language-overflow.ll
invalid-dicompileunit-missing-language.ll
invalid-dicompileunit-null-file.ll
invalid-dicompileunit-uniqued.ll
invalid-dicompositetype-missing-tag.ll
invalid-diderivedtype-missing-basetype.ll
invalid-diderivedtype-missing-tag.ll
invalid-dienumerator-missing-name.ll
invalid-dienumerator-missing-value.ll
invalid-diexpression-large.ll
invalid-diexpression-verify.ll Parse and print DIExpressions inline to ease IR and MIR testing 2017-08-23 20:31:27 +00:00
invalid-difile-missing-directory.ll
invalid-difile-missing-filename.ll
invalid-diglobalvariable-empty-name.ll
invalid-diglobalvariable-missing-name.ll
invalid-diimportedentity-missing-scope.ll
invalid-diimportedentity-missing-tag.ll
invalid-dilexicalblock-missing-scope.ll
invalid-dilexicalblock-null-scope.ll
invalid-dilexicalblockfile-missing-discriminator.ll
invalid-dilexicalblockfile-missing-scope.ll
invalid-dilexicalblockfile-null-scope.ll
invalid-dilocalvariable-arg-large.ll
invalid-dilocalvariable-arg-negative.ll
invalid-dilocalvariable-missing-scope.ll
invalid-dilocalvariable-null-scope.ll
invalid-dilocation-field-bad.ll
invalid-dilocation-field-twice.ll
invalid-dilocation-missing-scope-2.ll
invalid-dilocation-missing-scope.ll
invalid-dilocation-null-scope.ll
invalid-dilocation-overflow-column.ll
invalid-dilocation-overflow-line.ll
invalid-dinamespace-missing-namespace.ll
invalid-disubprogram-uniqued-definition.ll
invalid-disubrange-count-large.ll
invalid-disubrange-count-missing.ll
invalid-disubrange-count-negative.ll
invalid-disubrange-lowerBound-max.ll
invalid-disubrange-lowerBound-min.ll
invalid-disubroutinetype-missing-types.ll
invalid-ditemplatetypeparameter-missing-type.ll
invalid-ditemplatevalueparameter-missing-value.ll
invalid-fp80hex.ll
invalid-fwdref1.ll
invalid-fwdref2.ll
invalid-generic-debug-node-tag-bad.ll
invalid-generic-debug-node-tag-missing.ll
invalid-generic-debug-node-tag-overflow.ll
invalid-generic-debug-node-tag-wrong-type.ll
invalid-gep-mismatched-explicit-type.ll
invalid-gep-missing-explicit-type.ll
invalid-hexint.ll
invalid-inline-constraint.ll
invalid-inttype.ll Fix IntegerType::MAX_INT_BITS value 2016-10-04 12:43:46 +00:00
invalid-label.ll
invalid-landingpad.ll
invalid-load-mismatched-explicit-type.ll
invalid-load-missing-explicit-type.ll
invalid-mdnode-badref.ll
invalid-mdnode-vector2.ll
invalid-mdnode-vector.ll
invalid-metadata-attachment-has-type.ll
invalid-metadata-function-local-attachments.ll
invalid-metadata-function-local-complex-1.ll
invalid-metadata-function-local-complex-2.ll
invalid-metadata-function-local-complex-3.ll
invalid-metadata-has-type.ll
invalid-name2.ll
invalid-name.ll
invalid-safestack-param.ll
invalid-safestack-return.ll
invalid-specialized-mdnode.ll
invalid-untyped-metadata.ll
invalid-uselistorder_bb-missing-bb.ll
invalid-uselistorder_bb-missing-body.ll
invalid-uselistorder_bb-missing-func.ll
invalid-uselistorder_bb-not-bb.ll
invalid-uselistorder_bb-not-func.ll
invalid-uselistorder_bb-numbered.ll
invalid-uselistorder-function-between-blocks.ll
invalid-uselistorder-function-missing-named.ll
invalid-uselistorder-function-missing-numbered.ll
invalid-uselistorder-global-missing.ll
invalid-uselistorder-indexes-duplicated.ll
invalid-uselistorder-indexes-empty.ll
invalid-uselistorder-indexes-one.ll
invalid-uselistorder-indexes-ordered.ll
invalid-uselistorder-indexes-range.ll
invalid-uselistorder-indexes-toofew.ll
invalid-uselistorder-indexes-toomany.ll
invalid-uselistorder-type.ll
large-comdat.ll
local-unnamed-addr.ll
max-inttype.ll Fix IntegerType::MAX_INT_BITS value 2016-10-04 12:43:46 +00:00
metadata-decl.ll
metadata-function-local.ll
metadata-null-operands.ll
metadata.ll
missing-tbaa.ll
MultipleReturnValueType.ll
musttail-invalid-1.ll
musttail-invalid-2.ll
musttail.ll
named-metadata.ll
no-mdstring-upgrades.ll
numbered-values.ll
private-hidden-alias.ll
private-hidden-function.ll
private-hidden-variable.ll
private-protected-alias.ll
private-protected-function.ll
private-protected-variable.ll
return-column.s test: add an additional cfi_return_column test 2017-07-30 21:30:54 +00:00
select.ll
short-hexpair.ll
source-filename-backslash.ll
source-filename.ll
tls-models.ll
token.ll
unnamed-addr.ll
unnamed-alias.ll
unnamed-comdat.ll
unnamed.ll
unsized-recursive-type.ll
uselistorder_bb.ll
uselistorder.ll
vbool-cmp.ll
vector-cmp.ll
vector-select.ll
vector-shift.ll
x86mmx.ll