Derive GEP index type from Data Layout

[delena]

Summary

In the current version InstCombiner “normalizes” GEPs and extends Index operand to the pointer width.

It works fine if you can convert pointer to integer for address calculation and all registered targets do this.
The target I’m working on has very restricted ISA for the pointer calculation. Hal suggested to retrieve information for GEP index width from Data Layout.
http://lists.llvm.org/pipermail/llvm-dev/2018-January/120416.html

I added the interface to Data Layout and I changed the InstCombiner.
I know that I didn't touch all GEP creation points, but all changes that you see in the review are covered by our internal test system.

[craig.topper]

lib/Transforms/InstCombine/InstructionCombining.cpp

1887 ↗

(On Diff #129964)

Should this assert message be updated since it not guaranteed to be pointer width now?

1927 ↗

(On Diff #129964)

Same with this assert.

[delena]

Fixed 2 "assert" messages.

[delena]

Added tests for a data layout, where pointer is wider than the largest supported integer type.

[craig.topper]

lib/Analysis/ConstantFolding.cpp

812 ↗

(On Diff #130138)

Can you put curly braces after this if and the for loop below it to help with readability? I know the for loop didn't have them before, but I feel like it should have. I tend to think that if the inner scope is curly braced, the outer scope should be too.

[theraven]

I don't like this patch as is, for several reasons.

1. It's adding a hack that assumes that the offset should be the width of the widest integer operation. This is probably true in most cases (it is for us), but if we're going to introduce the idea that an address offset is distinct from the size of the pointer then we should do it properly and add that to the TargetInfo string explicitly (defaulting to the same size, if not specified).
2. We're computing the correct width every time it's requested, which looks expensive. TargetInfo should store the width for each address space and, for non-vector types, not have to do any calculation to determine the kind of integer to return.
3. It fixes only around 20% of the places that we've found that assume that the size and range of the pointer are the same.

[delena]

Updated, following Craig's comments.

[delena]

In D42123#981602, @theraven wrote:

I don't like this patch as is, for several reasons.

1. It's adding a hack that assumes that the offset should be the width of the widest integer operation. This is probably true in most cases (it is for us), but if we're going to introduce the idea that an address offset is distinct from the size of the pointer then we should do it properly and add that to the TargetInfo string explicitly (defaulting to the same size, if not specified).

So you propose to extend Data Layout string and add index size to it, right? It was one of options that Hal suggested. Ok.

2. We're computing the correct width every time it's requested, which looks expensive. TargetInfo should store the width for each address space and, for non-vector types, not have to do any calculation to determine the kind of integer to return.

We calculated getIntPtrType() anyway, getIndexType() is not more expensive. If I extend TargetInfo, the extension will be optional and all other targets will calculate getIntPtrType() anyway.

3. It fixes only around 20% of the places that we've found that assume that the size and range of the pointer are the same.

I can't derive all places from your code. You can show me them all, one-by-one, or we'll fix more places gradually on top of this patch.

[theraven]

In D42123#982063, @delena wrote:

In D42123#981602, @theraven wrote:

I don't like this patch as is, for several reasons.

1. It's adding a hack that assumes that the offset should be the width of the widest integer operation. This is probably true in most cases (it is for us), but if we're going to introduce the idea that an address offset is distinct from the size of the pointer then we should do it properly and add that to the TargetInfo string explicitly (defaulting to the same size, if not specified).

So you propose to extend Data Layout string and add index size to it, right? It was one of options that Hal suggested. Ok.

Yes, if we're going to fix this upstream, let's fix it properly.

2. We're computing the correct width every time it's requested, which looks expensive. TargetInfo should store the width for each address space and, for non-vector types, not have to do any calculation to determine the kind of integer to return.

We calculated getIntPtrType() anyway, getIndexType() is not more expensive. If I extend TargetInfo, the extension will be optional and all other targets will calculate getIntPtrType() anyway.

If you read it from the DataLayout string, you'll either construct it at that parsing time from the specified version or from the default version.

3. It fixes only around 20% of the places that we've found that assume that the size and range of the pointer are the same.

I can't derive all places from your code. You can show me them all, one-by-one, or we'll fix more places gradually on top of this patch.

Greping the code for all uses of getPointerBaseSize should show them all, but I can send you a list.

[delena]

I want to deprecate SCEVs for pointers if the index size is not equal to pointer size.
What do you think?

bool ScalarEvolution::isSCEVable(Type *Ty) const {
  if (Ty->isIntegerTy())
    return true;
  if (Ty->isPointerTy()) {
    // Pointer can't be scevable if index type and pointer type have different
    // width.
    const DataLayout& DL = getDataLayout();
    if (DL.getIndexTypeSizeInBits(Ty) == DL.getPointerTypeSizeInBits(Ty))
      return true;
  }
  return false;

}

[theraven]

In D42123#987589, @delena wrote:

I want to deprecate SCEVs for pointers if the index size is not equal to pointer size.
What do you think?

This will mean that you don't get a load of loop optimisations. I think that's a pretty big hammer. There's no reason why SCEV can't work here - we use it and have a bunch of patches against it to make it work in this context. Please take a look at our code and see how much of it is applicable to you.

[delena]

I looked at your code:
/ Return the size in bits of the specified type, for which isSCEVable must
/ return true.
uint64_t ScalarEvolution::getTypeSizeInBits(Type *Ty) const {

assert(isSCEVable(Ty) && "Type is not SCEVable!");
const DataLayout &DL = getDataLayout();
if (PointerType *PT = dyn_cast<PointerType>(Ty))
  return DL.getPointerBaseSizeInBits(PT->getPointerAddressSpace());
return DL.getTypeSizeInBits(Ty);

}
I can't say that size of pointer is smaller that it is. I can't truncate pointer to integer in order to expand all SCEV expressions.

[delena]

Added index width specification to the DataLayout. Updated the langref.
Fixed Pointer vs Index sizes in the code.
Added more tests.

[sanjoy]

../lib/Analysis/ScalarEvolution.cpp

3667 ↗

(On Diff #131955)

Generally speaking; the SCEV changes need to be tested.

3675 ↗

(On Diff #131955)

I don't think this is a correct place to make this change -- the size of a pointer is the size of a pointer. I think you need to change the SCEV corresponding to GEP(Ptr, Idx) to be "sext(Ptr) + Idx" or "Ptr + sext(Idx)" depending on their relative sizes.

[delena]

../lib/Analysis/ScalarEvolution.cpp

3667 ↗

(On Diff #131955)

I added several tests that go through the SCEV. Looks ok right now. I can't say that I cover all corner cases, but we can do further changes gradually, there is no impact on in-tree targets. If you see something specific that requires more testing now, please let me know.

3675 ↗

(On Diff #131955)

I can't create SCEV expressions with ptr+ind, it will fail with assertion on different types.

[Ayal]

../lib/Analysis/ScalarEvolution.cpp

3675 ↗

(On Diff #131955)

Elena, Sanjoy's thought above to change SCEV to be "sext(Ptr) + Idx" or "Ptr + sext(Idx)" will bring the two addends to be of the same type, i.e., of the larger type. The challenge in your case is lack of target support for integer addition of pointer-sized integers; which seems similar to CHERI's case. Except CHERI pointers (or capabilities) hold in addition to a standard-sized address additional information, such that the latter can be stripped out for SCEV purposes (IIUC - @theraven please correct if needed); whereas in your case the address itself is larger than a standard-sized integer. Perhaps for your case too the pointer can be stripped down to standard-sized integers to leverage SCEV's capabilities on "legal" types, which seems to be what your patch is doing, coupled with separate logic that deals with the stripped out bits(?).

[theraven]

../lib/Analysis/ScalarEvolution.cpp

3675 ↗

(On Diff #131955)

That's pretty much the case for us: our pointers are 128 bits, but have a 64-bit range (64 bits of metadata). We have modified DataLayout to explicitly understand that pointer size and range are different (in a slightly hacky way, which we should improve before we think about upstreaming). In scalar evolution, we always use the pointer's range as the type.

We don't support arbitrary integer operations on pointers and in our back end we have added some new MVTs to represent non-integer pointer types. Our architecture provides pointer + integer addressing modes.

I believe that, in the motivating example for this change, the existing ScalarEvolution code is correct: it should use pointer-sized integers, because otherwise the analyses are likely to be wrong in some exciting corner cases.

We have addressed this by adding explicit PTRADD SelectionDAG nodes, which perform pointer + integer addition. For complex addressing modes, we end up with (ptradd base (some complex integer ops)). This works well as long as the underlying hardware supports address register + integer register addressing, which I presume is the case for Intel (it is for all Harvard architectures that I've come across).

If you are targeting an architecture for which pointer operations and integer operations are not the same, then you should follow the same approach: in the back end, lower pointers to some non-integer type and match pointer operations with different patterns to integer ones. We have a bunch of SelectionDAG and TableGen patches that make this work well, which we'd be happy to upstream.

[delena]

" > We have addressed this by adding explicit PTRADD SelectionDAG nodes, which perform pointer + integer addition. For complex addressing modes, we end up with (ptradd base (some complex integer ops)). This works well as long as the underlying hardware supports address register + integer register addressing, which I presume is the case for Intel (it is for all Harvard architectures that I've come across)."

Yes, we also added ADDPTR node for SelectionDAG and we have more changes related to the special pointer type. Apparently, the codegen does not work with MVT::Ptr.
We can try to upstream the part of DAG builder, that makes ADDPTR from GEP.

[delena]

@theraven , the latest uploaded version is aligned with what you implemented out of the tree. Could you, please, take a look?

[theraven]

Two very small nits (which I'd be happy to see fixed after commit, but might be easier to fix first), but otherwise it looks like a significantly cleaned-up version of what we have.

Thank you very much for working on this! Our next merge will be a little bit painful, but subsequent ones should be a lot easier.

Are you planning on upstreaming your ADDPTR SelectionDAG stuff? We have added PTRADD, INTTOPTR and PTRTOINT nodes and if they're useful to someone apart from us then we can upstream them.

../include/llvm/IR/DataLayout.h

357 ↗

(On Diff #131955)

Please can we not have a default for AS? We've added defaults for other things like this because they were existing APIs and we didn't want to have to update all of the callers at once, but all of the callers of this are already being updated and so should specify the correct AS.

../lib/Analysis/ScalarEvolution.cpp

3675 ↗

(On Diff #131955)

I agree with @sanjoy that this isn't the correct place for this change, but it does happen to be the least disruptive place for the change. The correct solution is probably to rename this method to something like getTypeArithmeticSizeInBits so that it's clear that it's returning a size as a proxy for a range and not a storage size.

[sanjoy]

../lib/Analysis/ScalarEvolution.cpp

3667 ↗

(On Diff #131955)

I don't see these tests in this current version of the patch.

3675 ↗

(On Diff #131955)

Given what you said, the right fix seems to be to truncate 128 bit pointers to 64 bits in getSCEV instead of lying about the pointer's size. SCEV calls into other parts of LLVM like ValueTracking, and other parts of LLVM call into SCEV (invdvars, lsr, scev-aa etc.) and I'm worried that a discrepancy like this (pointer size = 64 in SCEV but 128 elsewhere) will cause bugs.

[delena]

I don't see these tests in this current version of the patch.

All tests that you see *-custom.ll" go through the scev calculations.

[theraven]

../lib/Analysis/ScalarEvolution.cpp

3675 ↗

(On Diff #131955)

Truncating for us would be absolutely the wrong thing, because it changes the semantics (throws away all of the bounds metadata, falls back to some per-environment bounds which may not even allow access to this address). In our case, we have a difference between the size and the range.

All of the places we've seen this used in ScalarEvolution, it cares about the range, not the storage size, it just happens that on most architectures these are the same (because a 32-bit pointer has a range of 2^32 bytes, a 64-bit pointer has a range of 2^64 bytes).

The correct solution is to either rename this method something like getTypeArithmeticSizeInBits and update all of the callers, or add a new getTypeArithmeticSizeInBits method and update all of the callers (I don't believe that there are any that care about the storage size, but I might have missed one).

We've been running with a change that's semantically identical to @delena's for a few years and have not encountered any miscompilations as a result (building a whole of the FreeBSD base system, a bunch of benchmarks, and a few other large programs), so I'm fairly confident that it's safe.

[arichardson]

Thank you very much for working on this. It will make our future upstream merges much easier.

../docs/LangRef.rst

1913 ↗

(On Diff #131955)

fourth parameter

../include/llvm/IR/DataLayout.h

357 ↗

(On Diff #131955)

Yes, please remove the default value here. We have run into lots of issues due to using the size of AS0 instead of the correct one.

../lib/Transforms/InstCombine/InstructionCombining.cpp

1511 ↗

(On Diff #131955)

Index width may not be the same width as pointer width

[delena]

Updated according to the latest comments.

[delena]

Added more tests with custom data layout.

llvm/trunk/docs/LangRef.rst

``sret``
     This indicates that the pointer parameter specifies the address of a
     structure that is the return value of the function in the source
     program. This pointer must be guaranteed by the caller to be valid:
     loads and stores to the structure may be assumed by the callee not
     to trap and to be properly aligned. This is not a valid attribute
     for return values.
 
 .. _attr_align:
-    
 ``align <n>``
     This indicates that the pointer value may be assumed by the optimizer to
     have the specified alignment.
 
     Note that this attribute has additional semantics when combined with the
     ``byval`` attribute.
 
 .. _noalias:
``S<size>``
     promotion of stack variables is limited to the natural stack
     alignment to avoid dynamic stack realignment. The stack alignment
     must be a multiple of 8-bits. If omitted, the natural stack
     alignment defaults to "unspecified", which does not prevent any
     alignment promotions.
 ``A<address space>``
     Specifies the address space of  objects created by '``alloca``'.
     Defaults to the default address space of 0.
-``p[n]:<size>:<abi>:<pref>``
+``p[n]:<size>:<abi>:<pref>:<idx>``
     This specifies the *size* of a pointer and its ``<abi>`` and
-    ``<pref>``\erred alignments for address space ``n``. All sizes are in
-    bits. The address space, ``n``, is optional, and if not specified,
+    ``<pref>``\erred alignments for address space ``n``. The fourth parameter
+    ``<idx>`` is a size of index that used for address calculation. If not
+    specified, the default index size is equal to the pointer size. All sizes
+    are in bits. The address space, ``n``, is optional, and if not specified,
     denotes the default address space 0. The value of ``n`` must be
     in the range [1,2^23).
 ``i<size>:<abi>:<pref>``
     This specifies the alignment for an integer type of a given bit
     ``<size>``. The value of ``<size>`` must be in the range [1,2^23).
 ``v<size>:<abi>:<pref>``
     This specifies the alignment for a vector type of a given bit
     ``<size>``.

 .. _fastmath:
 
 Fast-Math Flags
 ---------------
 
 LLVM IR floating-point operations (:ref:`fadd <i_fadd>`,
 :ref:`fsub <i_fsub>`, :ref:`fmul <i_fmul>`, :ref:`fdiv <i_fdiv>`,
 :ref:`frem <i_frem>`, :ref:`fcmp <i_fcmp>`) and :ref:`call <i_call>`
-may use the following flags to enable otherwise unsafe 
+may use the following flags to enable otherwise unsafe
 floating-point transformations.
 
 ``nnan``
    No NaNs - Allow optimizations to assume the arguments and result are not
    NaN. Such optimizations are required to retain defined behavior over
    NaNs, but the value of the result is undefined.
 
 ``ninf``
``arcp``
    argument rather than perform division.
 
 ``contract``
    Allow floating-point contraction (e.g. fusing a multiply followed by an
    addition into a fused multiply-and-add).
 
 ``afn``
    Approximate functions - Allow substitution of approximate calculations for
-   functions (sin, log, sqrt, etc). See floating-point intrinsic definitions 
-   for places where this can apply to LLVM's intrinsic math functions. 
+   functions (sin, log, sqrt, etc). See floating-point intrinsic definitions
+   for places where this can apply to LLVM's intrinsic math functions.
 
 ``reassoc``
-   Allow reassociation transformations for floating-point instructions. 
+   Allow reassociation transformations for floating-point instructions.
    This may dramatically change results in floating point.
 
 ``fast``
    This flag implies all of the others.
 
 .. _uselistorder:
 
 Use-list Order Directives

 
 The two arguments to the '``frem``' instruction must be :ref:`floating
 point <t_floating>` or :ref:`vector <t_vector>` of floating point values.
 Both arguments must have identical types.
 
 Semantics:
 """"""""""
 
-Return the same value as a libm '``fmod``' function but without trapping or 
+Return the same value as a libm '``fmod``' function but without trapping or
 setting ``errno``.
 
-The remainder has the same sign as the dividend. This instruction can also 
+The remainder has the same sign as the dividend. This instruction can also
 take any number of :ref:`fast-math flags <fastmath>`, which are optimization
 hints to enable otherwise unsafe floating-point optimizations:
 
 Example:
 """"""""
 
 .. code-block:: text
 

 If the ``isvolatile`` parameter is ``true``, the ``llvm.memset`` call is
 a :ref:`volatile operation <volatile>`. The detailed access behavior is not
 very cleanly specified and it is unwise to depend on it.
 
 Semantics:
 """"""""""
 
 The '``llvm.memset.*``' intrinsics fill "len" bytes of memory starting
-at the destination location. 
+at the destination location.
 
 '``llvm.sqrt.*``' Intrinsic
 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
 
 Syntax:
 """""""
 
 This is an overloaded intrinsic. You can use ``llvm.sqrt`` on any

 """"""""""
 
 The argument and return value are floating-point numbers of the same type.
 
 Semantics:
 """"""""""
 
 Return the same value as a corresponding libm '``sqrt``' function but without
-trapping or setting ``errno``. For types specified by IEEE-754, the result 
+trapping or setting ``errno``. For types specified by IEEE-754, the result
 matches a conforming libm implementation.
 
-When specified with the fast-math-flag 'afn', the result may be approximated 
+When specified with the fast-math-flag 'afn', the result may be approximated
 using a less accurate calculation.
 
 '``llvm.powi.*``' Intrinsic
 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
 
 Syntax:
 """""""
 

 The argument and return value are floating-point numbers of the same type.
 
 Semantics:
 """"""""""
 
 Return the same value as a corresponding libm '``sin``' function but without
 trapping or setting ``errno``.
 
-When specified with the fast-math-flag 'afn', the result may be approximated 
+When specified with the fast-math-flag 'afn', the result may be approximated
 using a less accurate calculation.
 
 '``llvm.cos.*``' Intrinsic
 ^^^^^^^^^^^^^^^^^^^^^^^^^^
 
 Syntax:
 """""""
 

 The argument and return value are floating-point numbers of the same type.
 
 Semantics:
 """"""""""
 
 Return the same value as a corresponding libm '``cos``' function but without
 trapping or setting ``errno``.
 
-When specified with the fast-math-flag 'afn', the result may be approximated 
+When specified with the fast-math-flag 'afn', the result may be approximated
 using a less accurate calculation.
 
 '``llvm.pow.*``' Intrinsic
 ^^^^^^^^^^^^^^^^^^^^^^^^^^
 
 Syntax:
 """""""
 

 The arguments and return value are floating-point numbers of the same type.
 
 Semantics:
 """"""""""
 
 Return the same value as a corresponding libm '``pow``' function but without
 trapping or setting ``errno``.
 
-When specified with the fast-math-flag 'afn', the result may be approximated 
+When specified with the fast-math-flag 'afn', the result may be approximated
 using a less accurate calculation.
 
 '``llvm.exp.*``' Intrinsic
 ^^^^^^^^^^^^^^^^^^^^^^^^^^
 
 Syntax:
 """""""
 

 The argument and return value are floating-point numbers of the same type.
 
 Semantics:
 """"""""""
 
 Return the same value as a corresponding libm '``exp``' function but without
 trapping or setting ``errno``.
 
-When specified with the fast-math-flag 'afn', the result may be approximated 
+When specified with the fast-math-flag 'afn', the result may be approximated
 using a less accurate calculation.
 
 '``llvm.exp2.*``' Intrinsic
 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
 
 Syntax:
 """""""
 

 The argument and return value are floating-point numbers of the same type.
 
 Semantics:
 """"""""""
 
 Return the same value as a corresponding libm '``exp2``' function but without
 trapping or setting ``errno``.
 
-When specified with the fast-math-flag 'afn', the result may be approximated 
+When specified with the fast-math-flag 'afn', the result may be approximated
 using a less accurate calculation.
 
 '``llvm.log.*``' Intrinsic
 ^^^^^^^^^^^^^^^^^^^^^^^^^^
 
 Syntax:
 """""""
 

 The argument and return value are floating-point numbers of the same type.
 
 Semantics:
 """"""""""
 
 Return the same value as a corresponding libm '``log``' function but without
 trapping or setting ``errno``.
 
-When specified with the fast-math-flag 'afn', the result may be approximated 
+When specified with the fast-math-flag 'afn', the result may be approximated
 using a less accurate calculation.
 
 '``llvm.log10.*``' Intrinsic
 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
 
 Syntax:
 """""""
 

 The argument and return value are floating-point numbers of the same type.
 
 Semantics:
 """"""""""
 
 Return the same value as a corresponding libm '``log10``' function but without
 trapping or setting ``errno``.
 
-When specified with the fast-math-flag 'afn', the result may be approximated 
+When specified with the fast-math-flag 'afn', the result may be approximated
 using a less accurate calculation.
 
 '``llvm.log2.*``' Intrinsic
 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
 
 Syntax:
 """""""
 

 The argument and return value are floating-point numbers of the same type.
 
 Semantics:
 """"""""""
 
 Return the same value as a corresponding libm '``log2``' function but without
 trapping or setting ``errno``.
 
-When specified with the fast-math-flag 'afn', the result may be approximated 
+When specified with the fast-math-flag 'afn', the result may be approximated
 using a less accurate calculation.
 
 '``llvm.fma.*``' Intrinsic
 ^^^^^^^^^^^^^^^^^^^^^^^^^^
 
 Syntax:
 """""""
 

 The arguments and return value are floating-point numbers of the same type.
 
 Semantics:
 """"""""""
 
 Return the same value as a corresponding libm '``fma``' function but without
 trapping or setting ``errno``.
 
-When specified with the fast-math-flag 'afn', the result may be approximated 
+When specified with the fast-math-flag 'afn', the result may be approximated
 using a less accurate calculation.
 
 '``llvm.fabs.*``' Intrinsic
 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
 
 Syntax:
 """""""
 

 Lowering:
 """""""""
 
 In the most general case call to the '``llvm.memset.element.unordered.atomic.*``' is
 lowered to a call to the symbol ``__llvm_memset_element_unordered_atomic_*``. Where '*'
 is replaced with an actual element size.
 
 The optimizer is allowed to inline the memory assignment when it's profitable to do so.
-

llvm/trunk/include/llvm/CodeGen/BasicTTIImpl.h

unsigned getEstimatedNumberOfCaseClusters(const SwitchInst &SI,
     unsigned N = SI.getNumCases();
     const TargetLoweringBase *TLI = getTLI();
     const DataLayout &DL = this->getDataLayout();
 
     JumpTableSize = 0;
     bool IsJTAllowed = TLI->areJTsAllowed(SI.getParent()->getParent());
 
     // Early exit if both a jump table and bit test are not allowed.
-    if (N < 1 || (!IsJTAllowed && DL.getPointerSizeInBits() < N))
+    if (N < 1 || (!IsJTAllowed && DL.getIndexSizeInBits(0u) < N))
       return N;
 
     APInt MaxCaseVal = SI.case_begin()->getCaseValue()->getValue();
     APInt MinCaseVal = MaxCaseVal;
     for (auto CI : SI.cases()) {
       const APInt &CaseVal = CI.getCaseValue()->getValue();
       if (CaseVal.sgt(MaxCaseVal))
         MaxCaseVal = CaseVal;
       if (CaseVal.slt(MinCaseVal))
         MinCaseVal = CaseVal;
     }
 
     // Check if suitable for a bit test
-    if (N <= DL.getPointerSizeInBits()) {
+    if (N <= DL.getIndexSizeInBits(0u)) {
       SmallPtrSet<const BasicBlock *, 4> Dests;
       for (auto I : SI.cases())
         Dests.insert(I.getCaseSuccessor());
 
       if (TLI->isSuitableForBitTests(Dests.size(), N, MinCaseVal, MaxCaseVal,
                                      DL))
         return 1;
     }

llvm/trunk/include/llvm/CodeGen/TargetLowering.h

virtual bool areJTsAllowed(const Function *Fn) const {
     return isOperationLegalOrCustom(ISD::BR_JT, MVT::Other) ||
            isOperationLegalOrCustom(ISD::BRIND, MVT::Other);
   }
 
   /// Check whether the range [Low,High] fits in a machine word.
   bool rangeFitsInWord(const APInt &Low, const APInt &High,
                        const DataLayout &DL) const {
     // FIXME: Using the pointer type doesn't seem ideal.
-    uint64_t BW = DL.getPointerSizeInBits();
+    uint64_t BW = DL.getIndexSizeInBits(0u);
     uint64_t Range = (High - Low).getLimitedValue(UINT64_MAX - 1) + 1;
     return Range <= BW;
   }
 
   /// Return true if lowering to a jump table is suitable for a set of case
   /// clusters which may contain \p NumCases cases, \p Range range of values.
   /// FIXME: This function check the maximum table size and density, but the
   /// minimum size is not checked. It would be nice if the minimum size is

llvm/trunk/include/llvm/IR/DataLayout.h

 ///
 /// \note The unusual order of elements in the structure attempts to reduce
 /// padding and make the structure slightly more cache friendly.
 struct PointerAlignElem {
   unsigned ABIAlign;
   unsigned PrefAlign;
   uint32_t TypeByteWidth;
   uint32_t AddressSpace;
+  uint32_t IndexWidth;
 
   /// Initializer
   static PointerAlignElem get(uint32_t AddressSpace, unsigned ABIAlign,
-                              unsigned PrefAlign, uint32_t TypeByteWidth);
+                              unsigned PrefAlign, uint32_t TypeByteWidth,
+                              uint32_t IndexWidth);
 
   bool operator==(const PointerAlignElem &rhs) const;
 };
 
 /// \brief A parsed version of the target data layout string in and methods for
 /// querying it.
 ///
 /// The target data layout string is specified *by the target* - a frontend
private:
   /// well-defined bitwise representation.
   SmallVector<unsigned, 8> NonIntegralAddressSpaces;
 
   void setAlignment(AlignTypeEnum align_type, unsigned abi_align,
                     unsigned pref_align, uint32_t bit_width);
   unsigned getAlignmentInfo(AlignTypeEnum align_type, uint32_t bit_width,
                             bool ABIAlign, Type *Ty) const;
   void setPointerAlignment(uint32_t AddrSpace, unsigned ABIAlign,
-                           unsigned PrefAlign, uint32_t TypeByteWidth);
+                           unsigned PrefAlign, uint32_t TypeByteWidth,
+                           uint32_t IndexWidth);
 
   /// Internal helper method that returns requested alignment for type.
   unsigned getAlignment(Type *Ty, bool abi_or_pref) const;
 
   /// Parses a target data specification string. Assert if the string is
   /// malformed.
   void parseSpecifier(StringRef LayoutDescription);
 
public:
   /// the backends/clients are updated.
   unsigned getPointerPrefAlignment(unsigned AS = 0) const;
 
   /// Layout pointer size
   /// FIXME: The defaults need to be removed once all of
   /// the backends/clients are updated.
   unsigned getPointerSize(unsigned AS = 0) const;
 
+  // Index size used for address calculation.
+  unsigned getIndexSize(unsigned AS) const;
+
   /// Return the address spaces containing non-integral pointers.  Pointers in
   /// this address space don't have a well-defined bitwise representation.
   ArrayRef<unsigned> getNonIntegralAddressSpaces() const {
     return NonIntegralAddressSpaces;
   }
 
   bool isNonIntegralPointerType(PointerType *PT) const {
     ArrayRef<unsigned> NonIntegralSpaces = getNonIntegralAddressSpaces();
     return find(NonIntegralSpaces, PT->getAddressSpace()) !=
            NonIntegralSpaces.end();
   }
 
   bool isNonIntegralPointerType(Type *Ty) const {
     auto *PTy = dyn_cast<PointerType>(Ty);
     return PTy && isNonIntegralPointerType(PTy);
   }
 
   /// Layout pointer size, in bits
   /// FIXME: The defaults need to be removed once all of
   /// the backends/clients are updated.
   unsigned getPointerSizeInBits(unsigned AS = 0) const {
     return getPointerSize(AS) * 8;
   }
 
+  /// Size in bits of index used for address calculation in getelementptr.
+  unsigned getIndexSizeInBits(unsigned AS) const {
+    return getIndexSize(AS) * 8;
+  }
+
   /// Layout pointer size, in bits, based on the type.  If this function is
   /// called with a pointer type, then the type size of the pointer is returned.
   /// If this function is called with a vector of pointers, then the type size
   /// of the pointer is returned.  This should only be called with a pointer or
   /// vector of pointers.
   unsigned getPointerTypeSizeInBits(Type *) const;
 
+  /// Layout size of the index used in GEP calculation.
+  /// The function should be called with pointer or vector of pointers type.
+  unsigned getIndexTypeSizeInBits(Type *Ty) const;
+
   unsigned getPointerTypeSize(Type *Ty) const {
     return getPointerTypeSizeInBits(Ty) / 8;
   }
 
   /// Size examples:
   ///
   /// Type        SizeInBits  StoreSizeInBits  AllocSizeInBits[*]
   /// ----        ----------  ---------------  ---------------
Type *getLargestLegalIntType(LLVMContext &C) const {
     unsigned LargestSize = getLargestLegalIntTypeSizeInBits();
     return (LargestSize == 0) ? nullptr : Type::getIntNTy(C, LargestSize);
   }
 
   /// \brief Returns the size of largest legal integer type size, or 0 if none
   /// are set.
   unsigned getLargestLegalIntTypeSizeInBits() const;
 
+  /// \brief Returns the type of a GEP index.
+  /// If it was not specified explicitly, it will be the integer type of the
+  /// pointer width - IntPtrType.
+  Type *getIndexType(Type *PtrTy) const;
+
   /// \brief Returns the offset from the beginning of the type for the specified
   /// indices.
   ///
   /// Note that this takes the element type, not the pointer type.
   /// This is used to implement getelementptr.
   int64_t getIndexedOffsetInType(Type *ElemTy, ArrayRef<Value *> Indices) const;
 
   /// \brief Returns a StructLayout object, indicating the alignment of the

llvm/trunk/lib/Analysis/ConstantFolding.cpp

 } // end anonymous namespace
 
 /// If this constant is a constant offset from a global, return the global and
 /// the constant. Because of constantexprs, this function is recursive.
 bool llvm::IsConstantOffsetFromGlobal(Constant *C, GlobalValue *&GV,
                                       APInt &Offset, const DataLayout &DL) {
   // Trivial case, constant is the global.
   if ((GV = dyn_cast<GlobalValue>(C))) {
-    unsigned BitWidth = DL.getPointerTypeSizeInBits(GV->getType());
+    unsigned BitWidth = DL.getIndexTypeSizeInBits(GV->getType());
     Offset = APInt(BitWidth, 0);
     return true;
   }
 
   // Otherwise, if this isn't a constant expr, bail out.
   auto *CE = dyn_cast<ConstantExpr>(C);
   if (!CE) return false;
 
   // Look through ptr->int and ptr->ptr casts.
   if (CE->getOpcode() == Instruction::PtrToInt ||
       CE->getOpcode() == Instruction::BitCast)
     return IsConstantOffsetFromGlobal(CE->getOperand(0), GV, Offset, DL);
 
   // i32* getelementptr ([5 x i32]* @a, i32 0, i32 5)
   auto *GEP = dyn_cast<GEPOperator>(CE);
   if (!GEP)
     return false;
 
-  unsigned BitWidth = DL.getPointerTypeSizeInBits(GEP->getType());
+  unsigned BitWidth = DL.getIndexTypeSizeInBits(GEP->getType());
   APInt TmpOffset(BitWidth, 0);
 
   // If the base isn't a global+constant, we aren't either.
   if (!IsConstantOffsetFromGlobal(CE->getOperand(0), GV, TmpOffset, DL))
     return false;
 
   // Otherwise, add any offset that our operands provide.
   if (!GEP->accumulateConstantOffset(DL, TmpOffset))
Constant *SymbolicallyEvaluateGEP(const GEPOperator *GEP,
   if (!Ptr->getType()->isPointerTy())
     return nullptr;
 
   Type *IntPtrTy = DL.getIntPtrType(Ptr->getType());
 
   // If this is a constant expr gep that is effectively computing an
   // "offsetof", fold it into 'cast int Size to T*' instead of 'gep 0, 0, 12'
   for (unsigned i = 1, e = Ops.size(); i != e; ++i)
       if (!isa<ConstantInt>(Ops[i])) {
 
         // If this is "gep i8* Ptr, (sub 0, V)", fold this as:
         // "inttoptr (sub (ptrtoint Ptr), V)"
         if (Ops.size() == 2 && ResElemTy->isIntegerTy(8)) {
           auto *CE = dyn_cast<ConstantExpr>(Ops[1]);
           assert((!CE || CE->getType() == IntPtrTy) &&
                  "CastGEPIndices didn't canonicalize index types!");
           if (CE && CE->getOpcode() == Instruction::Sub &&
               CE->getOperand(0)->isNullValue()) {
             Constant *Res = ConstantExpr::getPtrToInt(Ptr, CE->getType());
             Res = ConstantExpr::getSub(Res, CE->getOperand(1));
             Res = ConstantExpr::getIntToPtr(Res, ResTy);
             if (auto *FoldedRes = ConstantFoldConstant(Res, DL, TLI))
               Res = FoldedRes;
             return Res;
           }
         }
         return nullptr;
       }
 
   unsigned BitWidth = DL.getTypeSizeInBits(IntPtrTy);
   APInt Offset =
       APInt(BitWidth,
             DL.getIndexedOffsetInType(
                 SrcElemTy,
                 makeArrayRef((Value * const *)Ops.data() + 1, Ops.size() - 1)));
   Ptr = StripPtrCastKeepAS(Ptr, SrcElemTy);

llvm/trunk/lib/Analysis/InlineCost.cpp

void CallAnalyzer::disableLoadElimination() {
   }
 }
 
 /// \brief Accumulate a constant GEP offset into an APInt if possible.
 ///
 /// Returns false if unable to compute the offset for any reason. Respects any
 /// simplified values known during the analysis of this callsite.
 bool CallAnalyzer::accumulateGEPOffset(GEPOperator &GEP, APInt &Offset) {
-  unsigned IntPtrWidth = DL.getPointerTypeSizeInBits(GEP.getType());
+  unsigned IntPtrWidth = DL.getIndexTypeSizeInBits(GEP.getType());
   assert(IntPtrWidth == Offset.getBitWidth());
 
   for (gep_type_iterator GTI = gep_type_begin(GEP), GTE = gep_type_end(GEP);
        GTI != GTE; ++GTI) {
     ConstantInt *OpC = dyn_cast<ConstantInt>(GTI.getOperand());
     if (!OpC)
       if (Constant *SimpleOp = SimplifiedValues.lookup(GTI.getOperand()))
         OpC = dyn_cast<ConstantInt>(SimpleOp);

 /// accumulates the total constant offset applied in the returned constant. It
 /// returns 0 if V is not a pointer, and returns the constant '0' if there are
 /// no constant offsets applied.
 ConstantInt *CallAnalyzer::stripAndComputeInBoundsConstantOffsets(Value *&V) {
   if (!V->getType()->isPointerTy())
     return nullptr;
 
   unsigned AS = V->getType()->getPointerAddressSpace();
-  unsigned IntPtrWidth = DL.getPointerSizeInBits(AS);
+  unsigned IntPtrWidth = DL.getIndexSizeInBits(AS);
   APInt Offset = APInt::getNullValue(IntPtrWidth);
 
   // Even though we don't look through PHI nodes, we could be called on an
   // instruction in an unreachable block, which may be on a cycle.
   SmallPtrSet<Value *, 4> Visited;
   Visited.insert(V);
   do {
     if (GEPOperator *GEP = dyn_cast<GEPOperator>(V)) {

llvm/trunk/lib/Analysis/InstructionSimplify.cpp

if (Ty->isSized()) {
       uint64_t TyAllocSize = Q.DL.getTypeAllocSize(Ty);
       // getelementptr P, N -> P if P points to a type of zero size.
       if (TyAllocSize == 0)
         return Ops[0];
 
       // The following transforms are only safe if the ptrtoint cast
       // doesn't truncate the pointers.
       if (Ops[1]->getType()->getScalarSizeInBits() ==
-          Q.DL.getPointerSizeInBits(AS)) {
+          Q.DL.getIndexSizeInBits(AS)) {
         auto PtrToIntOrZero = [GEPTy](Value *P) -> Value * {
           if (match(P, m_Zero()))
             return Constant::getNullValue(GEPTy);
           Value *Temp;
           if (match(P, m_PtrToInt(m_Value(Temp))))
             if (Temp->getType() == GEPTy)
               return Temp;
           return nullptr;
if (Ty->isSized()) {
             return R;
       }
     }
   }
 
   if (Q.DL.getTypeAllocSize(LastType) == 1 &&
       all_of(Ops.slice(1).drop_back(1),
              [](Value *Idx) { return match(Idx, m_Zero()); })) {
-    unsigned PtrWidth =
-        Q.DL.getPointerSizeInBits(Ops[0]->getType()->getPointerAddressSpace());
-    if (Q.DL.getTypeSizeInBits(Ops.back()->getType()) == PtrWidth) {
-      APInt BasePtrOffset(PtrWidth, 0);
+    unsigned IdxWidth =
+        Q.DL.getIndexSizeInBits(Ops[0]->getType()->getPointerAddressSpace());
+    if (Q.DL.getTypeSizeInBits(Ops.back()->getType()) == IdxWidth) {
+      APInt BasePtrOffset(IdxWidth, 0);
       Value *StrippedBasePtr =
           Ops[0]->stripAndAccumulateInBoundsConstantOffsets(Q.DL,
                                                             BasePtrOffset);
 
       // gep (gep V, C), (sub 0, V) -> C
       if (match(Ops.back(),
                 m_Sub(m_Zero(), m_PtrToInt(m_Specific(StrippedBasePtr))))) {
         auto *CI = ConstantInt::get(GEPTy->getContext(), BasePtrOffset);

llvm/trunk/lib/Analysis/Loads.cpp

if (KnownDerefBytes.uge(Size))
       if (!CheckForNonNull || isKnownNonZero(V, DL, 0, nullptr, CtxI, DT))
         return isAligned(V, Align, DL);
   }
 
   // For GEPs, determine if the indexing lands within the allocated object.
   if (const GEPOperator *GEP = dyn_cast<GEPOperator>(V)) {
     const Value *Base = GEP->getPointerOperand();
 
-    APInt Offset(DL.getPointerTypeSizeInBits(GEP->getType()), 0);
+    APInt Offset(DL.getIndexTypeSizeInBits(GEP->getType()), 0);
     if (!GEP->accumulateConstantOffset(DL, Offset) || Offset.isNegative() ||
         !Offset.urem(APInt(Offset.getBitWidth(), Align)).isMinValue())
       return false;
 
     // If the base pointer is dereferenceable for Offset+Size bytes, then the
     // GEP (== Base + Offset) is dereferenceable for Size bytes.  If the base
     // pointer is aligned to Align bytes, and the Offset is divisible by Align
     // then the GEP (== Base + Offset == k_0 * Align + k_1 * Align) is also
bool llvm::isDereferenceableAndAlignedPointer(const Value *V, unsigned Align,
   if (Align == 0)
     Align = DL.getABITypeAlignment(Ty);
 
   if (!Ty->isSized())
     return false;
 
   SmallPtrSet<const Value *, 32> Visited;
   return ::isDereferenceableAndAlignedPointer(
-      V, Align, APInt(DL.getTypeSizeInBits(VTy), DL.getTypeStoreSize(Ty)), DL,
+      V, Align, APInt(DL.getIndexTypeSizeInBits(VTy), DL.getTypeStoreSize(Ty)), DL,
       CtxI, DT, Visited);
 }
 
 bool llvm::isDereferenceablePointer(const Value *V, const DataLayout &DL,
                                     const Instruction *CtxI,
                                     const DominatorTree *DT) {
   return isDereferenceableAndAlignedPointer(V, 1, DL, CtxI, DT);
 }

llvm/trunk/lib/Analysis/LoopAccessAnalysis.cpp

bool llvm::isConsecutiveAccess(Value *A, Value *B, const DataLayout &DL,
   // Make sure that A and B are different pointers.
   if (PtrA == PtrB)
     return false;
 
   // Make sure that A and B have the same type if required.
   if (CheckType && PtrA->getType() != PtrB->getType())
     return false;
 
-  unsigned PtrBitWidth = DL.getPointerSizeInBits(ASA);
+  unsigned IdxWidth = DL.getIndexSizeInBits(ASA);
   Type *Ty = cast<PointerType>(PtrA->getType())->getElementType();
-  APInt Size(PtrBitWidth, DL.getTypeStoreSize(Ty));
+  APInt Size(IdxWidth, DL.getTypeStoreSize(Ty));
 
-  APInt OffsetA(PtrBitWidth, 0), OffsetB(PtrBitWidth, 0);
+  APInt OffsetA(IdxWidth, 0), OffsetB(IdxWidth, 0);
   PtrA = PtrA->stripAndAccumulateInBoundsConstantOffsets(DL, OffsetA);
   PtrB = PtrB->stripAndAccumulateInBoundsConstantOffsets(DL, OffsetB);
 
   //  OffsetDelta = OffsetB - OffsetA;
   const SCEV *OffsetSCEVA = SE.getConstant(OffsetA);
   const SCEV *OffsetSCEVB = SE.getConstant(OffsetB);
   const SCEV *OffsetDeltaSCEV = SE.getMinusSCEV(OffsetSCEVB, OffsetSCEVA);
   const SCEVConstant *OffsetDeltaC = dyn_cast<SCEVConstant>(OffsetDeltaSCEV);

llvm/trunk/lib/Analysis/ScalarEvolution.cpp

bool ScalarEvolution::isSCEVable(Type *Ty) const {
   // Integers and pointers are always SCEVable.
   return Ty->isIntegerTy() || Ty->isPointerTy();
 }
 
 /// Return the size in bits of the specified type, for which isSCEVable must
 /// return true.
 uint64_t ScalarEvolution::getTypeSizeInBits(Type *Ty) const {
   assert(isSCEVable(Ty) && "Type is not SCEVable!");
+  if (Ty->isPointerTy())
+    return getDataLayout().getIndexTypeSizeInBits(Ty);
   return getDataLayout().getTypeSizeInBits(Ty);
 }
 
 /// Return a type with the same bitwidth as the given type and which represents
 /// how SCEV will treat the given type, for which isSCEVable must return
 /// true. For pointer types, this is the pointer-sized integer type.
 Type *ScalarEvolution::getEffectiveSCEVType(Type *Ty) const {
   assert(isSCEVable(Ty) && "Type is not SCEVable!");

llvm/trunk/lib/Analysis/ValueTracking.cpp

static cl::opt<unsigned> DomConditionsMaxUses("dom-conditions-max-uses",
                                               cl::Hidden, cl::init(20));
 
 /// Returns the bitwidth of the given scalar or pointer type. For vector types,
 /// returns the element type's bitwidth.
 static unsigned getBitWidth(Type *Ty, const DataLayout &DL) {
   if (unsigned BitWidth = Ty->getScalarSizeInBits())
     return BitWidth;
 
-  return DL.getPointerTypeSizeInBits(Ty);
+  return DL.getIndexTypeSizeInBits(Ty);
 }
 
 namespace {
 
 // Simplifying using an assume can only be done in a particular control-flow
 // context (the context instruction provides that context). If an assume and
 // the context instruction are not in the same block then the DT helps in
 // figuring out if we can use it.
case Instruction::IntToPtr:
     LLVM_FALLTHROUGH;
   case Instruction::ZExt:
   case Instruction::Trunc: {
     Type *SrcTy = I->getOperand(0)->getType();
 
     unsigned SrcBitWidth;
     // Note that we handle pointer operands here because of inttoptr/ptrtoint
     // which fall through here.
-    SrcBitWidth = Q.DL.getTypeSizeInBits(SrcTy->getScalarType());
+    Type *ScalarTy = SrcTy->getScalarType();
+    SrcBitWidth = ScalarTy->isPointerTy() ?
+      Q.DL.getIndexTypeSizeInBits(ScalarTy) :
+      Q.DL.getTypeSizeInBits(ScalarTy);
 
     assert(SrcBitWidth && "SrcBitWidth can't be zero");
     Known = Known.zextOrTrunc(SrcBitWidth);
     computeKnownBits(I->getOperand(0), Known, Depth + 1, Q);
     Known = Known.zextOrTrunc(BitWidth);
     // Any top bits are known to be zero.
     if (BitWidth > SrcBitWidth)
       Known.Zero.setBitsFrom(SrcBitWidth);
void computeKnownBits(const Value *V, KnownBits &Known, unsigned Depth,
                       const Query &Q) {
   assert(V && "No Value?");
   assert(Depth <= MaxDepth && "Limit Search Depth");
   unsigned BitWidth = Known.getBitWidth();
 
   assert((V->getType()->isIntOrIntVectorTy(BitWidth) ||
           V->getType()->isPtrOrPtrVectorTy()) &&
          "Not integer or pointer type!");
-  assert(Q.DL.getTypeSizeInBits(V->getType()->getScalarType()) == BitWidth &&
-         "V and Known should have same BitWidth");
+
+  Type *ScalarTy = V->getType()->getScalarType();
+  unsigned ExpectedWidth = ScalarTy->isPointerTy() ?
+    Q.DL.getIndexTypeSizeInBits(ScalarTy) : Q.DL.getTypeSizeInBits(ScalarTy);
+  assert(ExpectedWidth == BitWidth && "V and Known should have same BitWidth");
   (void)BitWidth;
+  (void)ExpectedWidth;
 
   const APInt *C;
   if (match(V, m_APInt(C))) {
     // We know all of the bits for a scalar constant or a splat vector constant!
     Known.One = *C;
     Known.Zero = ~Known.One;
     return;
   }

 static unsigned ComputeNumSignBitsImpl(const Value *V, unsigned Depth,
                                        const Query &Q) {
   assert(Depth <= MaxDepth && "Limit Search Depth");
 
   // We return the minimum number of sign bits that are guaranteed to be present
   // in V, so for undef we have to conservatively return 1.  We don't have the
   // same behavior for poison though -- that's a FIXME today.
 
-  unsigned TyBits = Q.DL.getTypeSizeInBits(V->getType()->getScalarType());
+  Type *ScalarTy = V->getType()->getScalarType();
+  unsigned TyBits = ScalarTy->isPointerTy() ?
+    Q.DL.getIndexTypeSizeInBits(ScalarTy) :
+    Q.DL.getTypeSizeInBits(ScalarTy);
+
   unsigned Tmp, Tmp2;
   unsigned FirstAnswer = 1;
 
   // Note that ConstantInt is handled by the general computeKnownBits case
   // below.
 
   if (Depth == MaxDepth)
     return 1;  // Limit search depth.
Value *llvm::FindInsertedValue(Value *V, ArrayRef<unsigned> idx_range,
   // or load instruction)
   return nullptr;
 }
 
 /// Analyze the specified pointer to see if it can be expressed as a base
 /// pointer plus a constant offset. Return the base and offset to the caller.
 Value *llvm::GetPointerBaseWithConstantOffset(Value *Ptr, int64_t &Offset,
                                               const DataLayout &DL) {
-  unsigned BitWidth = DL.getPointerTypeSizeInBits(Ptr->getType());
+  unsigned BitWidth = DL.getIndexTypeSizeInBits(Ptr->getType());
   APInt ByteOffset(BitWidth, 0);
 
   // We walk up the defs but use a visited set to handle unreachable code. In
   // that case, we stop after accumulating the cycle once (not that it
   // matters).
   SmallPtrSet<Value *, 16> Visited;
   while (Visited.insert(Ptr).second) {
     if (Ptr->getType()->isVectorTy())
       break;
 
     if (GEPOperator *GEP = dyn_cast<GEPOperator>(Ptr)) {
       // If one of the values we have visited is an addrspacecast, then
       // the pointer type of this GEP may be different from the type
       // of the Ptr parameter which was passed to this function.  This
       // means when we construct GEPOffset, we need to use the size
       // of GEP's pointer type rather than the size of the original
       // pointer type.
-      APInt GEPOffset(DL.getPointerTypeSizeInBits(Ptr->getType()), 0);
+      APInt GEPOffset(DL.getIndexTypeSizeInBits(Ptr->getType()), 0);
       if (!GEP->accumulateConstantOffset(DL, GEPOffset))
         break;
 
       ByteOffset += GEPOffset.getSExtValue();
 
       Ptr = GEP->getPointerOperand();
     } else if (Operator::getOpcode(Ptr) == Instruction::BitCast ||
                Operator::getOpcode(Ptr) == Instruction::AddrSpaceCast) {

llvm/trunk/lib/CodeGen/CodeGenPrepare.cpp

bool CodeGenPrepare::optimizeCallInst(CallInst *CI, bool &ModifiedDT) {
   if (TLI && TLI->shouldAlignPointerArgs(CI, MinSize, PrefAlign)) {
     for (auto &Arg : CI->arg_operands()) {
       // We want to align both objects whose address is used directly and
       // objects whose address is used in casts and GEPs, though it only makes
       // sense for GEPs if the offset is a multiple of the desired alignment and
       // if size - offset meets the size threshold.
       if (!Arg->getType()->isPointerTy())
         continue;
-      APInt Offset(DL->getPointerSizeInBits(
+      APInt Offset(DL->getIndexSizeInBits(
                        cast<PointerType>(Arg->getType())->getAddressSpace()),
                    0);
       Value *Val = Arg->stripAndAccumulateInBoundsConstantOffsets(*DL, Offset);
       uint64_t Offset2 = Offset.getLimitedValue();
       if ((Offset2 & (PrefAlign-1)) != 0)
         continue;
       AllocaInst *AI;
       if ((AI = dyn_cast<AllocaInst>(Val)) && AI->getAlignment() < PrefAlign &&

llvm/trunk/lib/CodeGen/SelectionDAG/SelectionDAG.cpp

 
 /// InferPtrAlignment - Infer alignment of a load / store address. Return 0 if
 /// it cannot be inferred.
 unsigned SelectionDAG::InferPtrAlignment(SDValue Ptr) const {
   // If this is a GlobalAddress + cst, return the alignment.
   const GlobalValue *GV;
   int64_t GVOffset = 0;
   if (TLI->isGAPlusOffset(Ptr.getNode(), GV, GVOffset)) {
-    unsigned PtrWidth = getDataLayout().getPointerTypeSizeInBits(GV->getType());
-    KnownBits Known(PtrWidth);
+    unsigned IdxWidth = getDataLayout().getIndexTypeSizeInBits(GV->getType());
+    KnownBits Known(IdxWidth);
     llvm::computeKnownBits(GV, Known, getDataLayout());
     unsigned AlignBits = Known.countMinTrailingZeros();
     unsigned Align = AlignBits ? 1 << std::min(31U, AlignBits) : 0;
     if (Align)
       return MinAlign(Align, GVOffset);
   }
 
   // If this is a direct reference to a stack slot, use information about the

llvm/trunk/lib/CodeGen/SelectionDAG/SelectionDAGBuilder.cpp

if (StructType *StTy = GTI.getStructTypeOrNull()) {
         SDNodeFlags Flags;
         if (int64_t(Offset) >= 0 && cast<GEPOperator>(I).isInBounds())
           Flags.setNoUnsignedWrap(true);
 
         N = DAG.getNode(ISD::ADD, dl, N.getValueType(), N,
                         DAG.getConstant(Offset, dl, N.getValueType()), Flags);
       }
     } else {
-      MVT PtrTy =
-          DAG.getTargetLoweringInfo().getPointerTy(DAG.getDataLayout(), AS);
-      unsigned PtrSize = PtrTy.getSizeInBits();
-      APInt ElementSize(PtrSize, DL->getTypeAllocSize(GTI.getIndexedType()));
+      unsigned IdxSize = DAG.getDataLayout().getIndexSizeInBits(AS);
+      MVT IdxTy = MVT::getIntegerVT(IdxSize);
+      APInt ElementSize(IdxSize, DL->getTypeAllocSize(GTI.getIndexedType()));
 
       // If this is a scalar constant or a splat vector of constants,
       // handle it quickly.
       const auto *CI = dyn_cast<ConstantInt>(Idx);
       if (!CI && isa<ConstantDataVector>(Idx) &&
           cast<ConstantDataVector>(Idx)->getSplatValue())
         CI = cast<ConstantInt>(cast<ConstantDataVector>(Idx)->getSplatValue());
 
       if (CI) {
         if (CI->isZero())
           continue;
-        APInt Offs = ElementSize * CI->getValue().sextOrTrunc(PtrSize);
+        APInt Offs = ElementSize * CI->getValue().sextOrTrunc(IdxSize);
         LLVMContext &Context = *DAG.getContext();
         SDValue OffsVal = VectorWidth ?
-          DAG.getConstant(Offs, dl, EVT::getVectorVT(Context, PtrTy, VectorWidth)) :
-          DAG.getConstant(Offs, dl, PtrTy);
+          DAG.getConstant(Offs, dl, EVT::getVectorVT(Context, IdxTy, VectorWidth)) :
+          DAG.getConstant(Offs, dl, IdxTy);
 
         // In an inbouds GEP with an offset that is nonnegative even when
         // interpreted as signed, assume there is no unsigned overflow.
         SDNodeFlags Flags;
         if (Offs.isNonNegative() && cast<GEPOperator>(I).isInBounds())
           Flags.setNoUnsignedWrap(true);
 
         N = DAG.getNode(ISD::ADD, dl, N.getValueType(), N, OffsVal, Flags);

llvm/trunk/lib/IR/DataLayout.cpp

 }
 
 //===----------------------------------------------------------------------===//
 // PointerAlignElem, PointerAlign support
 //===----------------------------------------------------------------------===//
 
 PointerAlignElem
 PointerAlignElem::get(uint32_t AddressSpace, unsigned ABIAlign,
-                      unsigned PrefAlign, uint32_t TypeByteWidth) {
+                      unsigned PrefAlign, uint32_t TypeByteWidth,
+                      uint32_t IndexWidth) {
   assert(ABIAlign <= PrefAlign && "Preferred alignment worse than ABI!");
   PointerAlignElem retval;
   retval.AddressSpace = AddressSpace;
   retval.ABIAlign = ABIAlign;
   retval.PrefAlign = PrefAlign;
   retval.TypeByteWidth = TypeByteWidth;
+  retval.IndexWidth = IndexWidth;
   return retval;
 }
 
 bool
 PointerAlignElem::operator==(const PointerAlignElem &rhs) const {
   return (ABIAlign == rhs.ABIAlign
           && AddressSpace == rhs.AddressSpace
           && PrefAlign == rhs.PrefAlign
-          && TypeByteWidth == rhs.TypeByteWidth);
+          && TypeByteWidth == rhs.TypeByteWidth
+          && IndexWidth == rhs.IndexWidth);
 }
 
 //===----------------------------------------------------------------------===//
 //                       DataLayout Class Implementation
 //===----------------------------------------------------------------------===//
 
 const char *DataLayout::getManglingComponent(const Triple &T) {
   if (T.isOSBinFormatMachO())
void DataLayout::reset(StringRef Desc) {
   ManglingMode = MM_None;
   NonIntegralAddressSpaces.clear();
 
   // Default alignments
   for (const LayoutAlignElem &E : DefaultAlignments) {
     setAlignment((AlignTypeEnum)E.AlignType, E.ABIAlign, E.PrefAlign,
                  E.TypeBitWidth);
   }
-  setPointerAlignment(0, 8, 8, 8);
+  setPointerAlignment(0, 8, 8, 8, 8);
 
   parseSpecifier(Desc);
 }
 
 /// Checked version of split, to ensure mandatory subparts.
 static std::pair<StringRef, StringRef> split(StringRef Str, char Separator) {
   assert(!Str.empty() && "parse error, string can't be empty here");
   std::pair<StringRef, StringRef> Split = Str.split(Separator);
case 'p': {
         report_fatal_error(
             "Missing alignment specification for pointer in datalayout string");
       Split = split(Rest, ':');
       unsigned PointerABIAlign = inBytes(getInt(Tok));
       if (!isPowerOf2_64(PointerABIAlign))
         report_fatal_error(
             "Pointer ABI alignment must be a power of 2");
 
+      // Size of index used in GEP for address calculation.
+      // The parameter is optional. By default it is equal to size of pointer.
+      unsigned IndexSize = PointerMemSize;
+
       // Preferred alignment.
       unsigned PointerPrefAlign = PointerABIAlign;
       if (!Rest.empty()) {
         Split = split(Rest, ':');
         PointerPrefAlign = inBytes(getInt(Tok));
         if (!isPowerOf2_64(PointerPrefAlign))
           report_fatal_error(
             "Pointer preferred alignment must be a power of 2");
-      }
 
+        // Now read the index. It is the second optional parameter here.
+        if (!Rest.empty()) {
+          Split = split(Rest, ':');
+          IndexSize = inBytes(getInt(Tok));
+          if (!IndexSize)
+            report_fatal_error("Invalid index size of 0 bytes");
+        }
+      }
       setPointerAlignment(AddrSpace, PointerABIAlign, PointerPrefAlign,
-                          PointerMemSize);
+                          PointerMemSize, IndexSize);
       break;
     }
     case 'i':
     case 'v':
     case 'f':
     case 'a': {
       AlignTypeEnum AlignType;
       switch (Specifier) {

 DataLayout::findPointerLowerBound(uint32_t AddressSpace) {
   return std::lower_bound(Pointers.begin(), Pointers.end(), AddressSpace,
                           [](const PointerAlignElem &A, uint32_t AddressSpace) {
     return A.AddressSpace < AddressSpace;
   });
 }
 
 void DataLayout::setPointerAlignment(uint32_t AddrSpace, unsigned ABIAlign,
-                                     unsigned PrefAlign,
-                                     uint32_t TypeByteWidth) {
+                                     unsigned PrefAlign, uint32_t TypeByteWidth,
+                                     uint32_t IndexWidth) {
   if (PrefAlign < ABIAlign)
     report_fatal_error(
         "Preferred alignment cannot be less than the ABI alignment");
 
   PointersTy::iterator I = findPointerLowerBound(AddrSpace);
   if (I == Pointers.end() || I->AddressSpace != AddrSpace) {
     Pointers.insert(I, PointerAlignElem::get(AddrSpace, ABIAlign, PrefAlign,
-                                             TypeByteWidth));
+                                             TypeByteWidth, IndexWidth));
   } else {
     I->ABIAlign = ABIAlign;
     I->PrefAlign = PrefAlign;
     I->TypeByteWidth = TypeByteWidth;
+    I->IndexWidth = IndexWidth;
   }
 }
 
 /// getAlignmentInfo - Return the alignment (either ABI if ABIInfo = true or
 /// preferred if ABIInfo = false) the layout wants for the specified datatype.
 unsigned DataLayout::getAlignmentInfo(AlignTypeEnum AlignType,
                                       uint32_t BitWidth, bool ABIInfo,
                                       Type *Ty) const {

 
 unsigned DataLayout::getPointerTypeSizeInBits(Type *Ty) const {
   assert(Ty->isPtrOrPtrVectorTy() &&
          "This should only be called with a pointer or pointer vector type");
   Ty = Ty->getScalarType();
   return getPointerSizeInBits(cast<PointerType>(Ty)->getAddressSpace());
 }
 
+unsigned DataLayout::getIndexSize(unsigned AS) const {
+  PointersTy::const_iterator I = findPointerLowerBound(AS);
+  if (I == Pointers.end() || I->AddressSpace != AS) {
+    I = findPointerLowerBound(0);
+    assert(I->AddressSpace == 0);
+  }
+  return I->IndexWidth;
+}
+
+unsigned DataLayout::getIndexTypeSizeInBits(Type *Ty) const {
+  assert(Ty->isPtrOrPtrVectorTy() &&
+         "This should only be called with a pointer or pointer vector type");
+  Ty = Ty->getScalarType();
+  return getIndexSizeInBits(cast<PointerType>(Ty)->getAddressSpace());
+}
+
 /*!
   \param abi_or_pref Flag that determines which alignment is returned. true
   returns the ABI alignment, false returns the preferred alignment.
   \param Ty The underlying type for which alignment is determined.
 
   Get the ABI (\a abi_or_pref == true) or preferred alignment (\a abi_or_pref
   == false) for the requested type \a Ty.
  */

 unsigned DataLayout::getPreferredTypeAlignmentShift(Type *Ty) const {
   unsigned Align = getPrefTypeAlignment(Ty);
   assert(!(Align & (Align-1)) && "Alignment is not a power of two!");
   return Log2_32(Align);
 }
 
 IntegerType *DataLayout::getIntPtrType(LLVMContext &C,
                                        unsigned AddressSpace) const {
-  return IntegerType::get(C, getPointerSizeInBits(AddressSpace));
+  return IntegerType::get(C, getIndexSizeInBits(AddressSpace));
 }
 
 Type *DataLayout::getIntPtrType(Type *Ty) const {
   assert(Ty->isPtrOrPtrVectorTy() &&
          "Expected a pointer or pointer vector type.");
-  unsigned NumBits = getPointerTypeSizeInBits(Ty);
+  unsigned NumBits = getIndexTypeSizeInBits(Ty);
   IntegerType *IntTy = IntegerType::get(Ty->getContext(), NumBits);
   if (VectorType *VecTy = dyn_cast<VectorType>(Ty))
     return VectorType::get(IntTy, VecTy->getNumElements());
   return IntTy;
 }
 
 Type *DataLayout::getSmallestLegalIntType(LLVMContext &C, unsigned Width) const {
   for (unsigned LegalIntWidth : LegalIntWidths)
     if (Width <= LegalIntWidth)
       return Type::getIntNTy(C, LegalIntWidth);
   return nullptr;
 }
 
 unsigned DataLayout::getLargestLegalIntTypeSizeInBits() const {
   auto Max = std::max_element(LegalIntWidths.begin(), LegalIntWidths.end());
   return Max != LegalIntWidths.end() ? *Max : 0;
 }
 
+Type *DataLayout::getIndexType(Type *Ty) const {
+  assert(Ty->isPtrOrPtrVectorTy() &&
+         "Expected a pointer or pointer vector type.");
+  unsigned NumBits = getIndexTypeSizeInBits(Ty);
+  IntegerType *IntTy = IntegerType::get(Ty->getContext(), NumBits);
+  if (VectorType *VecTy = dyn_cast<VectorType>(Ty))
+    return VectorType::get(IntTy, VecTy->getNumElements());
+  return IntTy;
+}
+
 int64_t DataLayout::getIndexedOffsetInType(Type *ElemTy,
                                            ArrayRef<Value *> Indices) const {
   int64_t Result = 0;
 
   generic_gep_type_iterator<Value* const*>
     GTI = gep_type_begin(ElemTy, Indices),
     GTE = gep_type_end(ElemTy, Indices);
   for (; GTI != GTE; ++GTI) {

llvm/trunk/lib/IR/Operator.cpp

Type *GEPOperator::getResultElementType() const {
   if (auto *I = dyn_cast<GetElementPtrInst>(this))
     return I->getResultElementType();
   return cast<GetElementPtrConstantExpr>(this)->getResultElementType();
 }
 
 bool GEPOperator::accumulateConstantOffset(const DataLayout &DL,
                                            APInt &Offset) const {
   assert(Offset.getBitWidth() ==
-             DL.getPointerSizeInBits(getPointerAddressSpace()) &&
-         "The offset must have exactly as many bits as our pointer.");
+             DL.getIndexSizeInBits(getPointerAddressSpace()) &&
+         "The offset bit width does not match DL specification.");
 
   for (gep_type_iterator GTI = gep_type_begin(this), GTE = gep_type_end(this);
        GTI != GTE; ++GTI) {
     ConstantInt *OpC = dyn_cast<ConstantInt>(GTI.getOperand());
     if (!OpC)
       return false;
     if (OpC->isZero())
       continue;

llvm/trunk/lib/IR/Value.cpp

 }
 
 const Value *
 Value::stripAndAccumulateInBoundsConstantOffsets(const DataLayout &DL,
                                                  APInt &Offset) const {
   if (!getType()->isPointerTy())
     return this;
 
-  assert(Offset.getBitWidth() == DL.getPointerSizeInBits(cast<PointerType>(
+  assert(Offset.getBitWidth() == DL.getIndexSizeInBits(cast<PointerType>(
                                      getType())->getAddressSpace()) &&
-         "The offset must have exactly as many bits as our pointer.");
+         "The offset bit width does not match the DL specification.");
 
   // Even though we don't look through PHI nodes, we could be called on an
   // instruction in an unreachable block, which may be on a cycle.
   SmallPtrSet<const Value *, 4> Visited;
   Visited.insert(this);
   const Value *V = this;
   do {
     if (auto *GEP = dyn_cast<GEPOperator>(V)) {

llvm/trunk/lib/Transforms/InstCombine/InstCombineCasts.cpp

 Instruction *InstCombiner::visitPtrToInt(PtrToIntInst &CI) {
   // If the destination integer type is not the intptr_t type for this target,
   // do a ptrtoint to intptr_t then do a trunc or zext.  This allows the cast
   // to be exposed to other transforms.
 
   Type *Ty = CI.getType();
   unsigned AS = CI.getPointerAddressSpace();
 
-  if (Ty->getScalarSizeInBits() == DL.getPointerSizeInBits(AS))
+  if (Ty->getScalarSizeInBits() == DL.getIndexSizeInBits(AS))
     return commonPointerCastTransforms(CI);
 
   Type *PtrTy = DL.getIntPtrType(CI.getContext(), AS);
   if (Ty->isVectorTy()) // Handle vectors of pointers.
     PtrTy = VectorType::get(PtrTy, Ty->getVectorNumElements());
 
   Value *P = Builder.CreatePtrToInt(CI.getOperand(0), PtrTy);
   return CastInst::CreateIntegerCast(P, Ty, /*isSigned=*/false);

 static Instruction *foldBitCastBitwiseLogic(BitCastInst &BitCast,
                                             InstCombiner::BuilderTy &Builder) {
   Type *DestTy = BitCast.getType();
   BinaryOperator *BO;
   if (!DestTy->isIntOrIntVectorTy() ||
       !match(BitCast.getOperand(0), m_OneUse(m_BinOp(BO))) ||
       !BO->isBitwiseLogicOp())
     return nullptr;
-  
   // FIXME: This transform is restricted to vector types to avoid backend
   // problems caused by creating potentially illegal operations. If a fix-up is
   // added to handle that situation, we can remove this check.
   if (!DestTy->isVectorTy() || !BO->getType()->isVectorTy())
     return nullptr;
-  
   Value *X;
   if (match(BO->getOperand(0), m_OneUse(m_BitCast(m_Value(X)))) &&
       X->getType() == DestTy && !isa<Constant>(X)) {
     // bitcast(logic(bitcast(X), Y)) --> logic'(X, bitcast(Y))
     Value *CastedOp1 = Builder.CreateBitCast(BO->getOperand(1), DestTy);
     return BinaryOperator::Create(BO->getOpcode(), X, CastedOp1);
   }
 

llvm/trunk/lib/Transforms/InstCombine/InstCombineCompares.cpp

static Value *rewriteGEPAsOffset(Value *Start, Value *Base,
   // Perform all the substitutions. This is a bit tricky because we can
   // have cycles in our use-def chains.
   // 1. Create the PHI nodes without any incoming values.
   // 2. Create all the other values.
   // 3. Add the edges for the PHI nodes.
   // 4. Emit GEPs to get the original pointers.
   // 5. Remove the original instructions.
   Type *IndexType = IntegerType::get(
-      Base->getContext(), DL.getPointerTypeSizeInBits(Start->getType()));
+      Base->getContext(), DL.getIndexTypeSizeInBits(Start->getType()));
 
   DenseMap<Value *, Value *> NewInsts;
   NewInsts[Base] = ConstantInt::getNullValue(IndexType);
 
   // Create the new PHI nodes, without adding any incoming values.
   for (Value *Val : Explored) {
     if (Val == Base)
       continue;

 }
 
 /// Looks through GEPs, IntToPtrInsts and PtrToIntInsts in order to express
 /// the input Value as a constant indexed GEP. Returns a pair containing
 /// the GEPs Pointer and Index.
 static std::pair<Value *, Value *>
 getAsConstantIndexedAddress(Value *V, const DataLayout &DL) {
   Type *IndexType = IntegerType::get(V->getContext(),
-                                     DL.getPointerTypeSizeInBits(V->getType()));
+                                     DL.getIndexTypeSizeInBits(V->getType()));
 
   Constant *Index = ConstantInt::getNullValue(IndexType);
   while (true) {
     if (GEPOperator *GEP = dyn_cast<GEPOperator>(V)) {
       // We accept only inbouds GEPs here to exclude the possibility of
       // overflow.
       if (!GEP->isInBounds())
         break;

 Instruction *InstCombiner::foldICmpUsingKnownBits(ICmpInst &I) {
   Value *Op0 = I.getOperand(0), *Op1 = I.getOperand(1);
   Type *Ty = Op0->getType();
   ICmpInst::Predicate Pred = I.getPredicate();
 
   // Get scalar or pointer size.
   unsigned BitWidth = Ty->isIntOrIntVectorTy()
                           ? Ty->getScalarSizeInBits()
-                          : DL.getTypeSizeInBits(Ty->getScalarType());
+                          : DL.getIndexTypeSizeInBits(Ty->getScalarType());
 
   if (!BitWidth)
     return nullptr;
 
   KnownBits Op0Known(BitWidth);
   KnownBits Op1Known(BitWidth);
 
   if (SimplifyDemandedBits(&I, 0,

llvm/trunk/lib/Transforms/InstCombine/InstructionCombining.cpp

Type *InstCombiner::FindElementAtOffset(PointerType *PtrTy, int64_t Offset,
                                         SmallVectorImpl<Value *> &NewIndices) {
   Type *Ty = PtrTy->getElementType();
   if (!Ty->isSized())
     return nullptr;
 
   // Start with the index over the outer type.  Note that the type size
   // might be zero (even if the offset isn't zero) if the indexed type
   // is something like [0 x {int, int}]
-  Type *IntPtrTy = DL.getIntPtrType(PtrTy);
+  Type *IndexTy = DL.getIndexType(PtrTy);
   int64_t FirstIdx = 0;
   if (int64_t TySize = DL.getTypeAllocSize(Ty)) {
     FirstIdx = Offset/TySize;
     Offset -= FirstIdx*TySize;
 
     // Handle hosts where % returns negative instead of values [0..TySize).
     if (Offset < 0) {
       --FirstIdx;
       Offset += TySize;
       assert(Offset >= 0);
     }
     assert((uint64_t)Offset < (uint64_t)TySize && "Out of range offset");
   }
 
-  NewIndices.push_back(ConstantInt::get(IntPtrTy, FirstIdx));
+  NewIndices.push_back(ConstantInt::get(IndexTy, FirstIdx));
 
   // Index into the types.  If we fail, set OrigBase to null.
   while (Offset) {
     // Indexing into tail padding between struct/array elements.
     if (uint64_t(Offset * 8) >= DL.getTypeSizeInBits(Ty))
       return nullptr;
 
     if (StructType *STy = dyn_cast<StructType>(Ty)) {
       const StructLayout *SL = DL.getStructLayout(STy);
       assert(Offset < (int64_t)SL->getSizeInBytes() &&
              "Offset must stay within the indexed type");
 
       unsigned Elt = SL->getElementContainingOffset(Offset);
       NewIndices.push_back(ConstantInt::get(Type::getInt32Ty(Ty->getContext()),
                                             Elt));
 
       Offset -= SL->getElementOffset(Elt);
       Ty = STy->getElementType(Elt);
     } else if (ArrayType *AT = dyn_cast<ArrayType>(Ty)) {
       uint64_t EltSize = DL.getTypeAllocSize(AT->getElementType());
       assert(EltSize && "Cannot index into a zero-sized array");
-      NewIndices.push_back(ConstantInt::get(IntPtrTy,Offset/EltSize));
+      NewIndices.push_back(ConstantInt::get(IndexTy,Offset/EltSize));
       Offset %= EltSize;
       Ty = AT->getElementType();
     } else {
       // Otherwise, we can't index into the middle of this atomic type, bail.
       return nullptr;
     }
   }
 
if (Value *V = SimplifyGEPInst(GEP.getSourceElementType(), Ops,
                                  SQ.getWithInstruction(&GEP)))
     return replaceInstUsesWith(GEP, V);
 
   Value *PtrOp = GEP.getOperand(0);
 
   // Eliminate unneeded casts for indices, and replace indices which displace
   // by multiples of a zero size type with zero.
   bool MadeChange = false;
-  Type *IntPtrTy =
-    DL.getIntPtrType(GEP.getPointerOperandType()->getScalarType());
+
+  // Index width may not be the same width as pointer width.
+  // Data layout chooses the right type based on supported integer types.
+  Type *NewScalarIndexTy =
+      DL.getIndexType(GEP.getPointerOperandType()->getScalarType());
 
   gep_type_iterator GTI = gep_type_begin(GEP);
   for (User::op_iterator I = GEP.op_begin() + 1, E = GEP.op_end(); I != E;
        ++I, ++GTI) {
     // Skip indices into struct types.
     if (GTI.isStruct())
       continue;
 
-    // Index type should have the same width as IntPtr
     Type *IndexTy = (*I)->getType();
-    Type *NewIndexType = IndexTy->isVectorTy() ?
-      VectorType::get(IntPtrTy, IndexTy->getVectorNumElements()) : IntPtrTy;
+    Type *NewIndexType =
+        IndexTy->isVectorTy()
+            ? VectorType::get(NewScalarIndexTy, IndexTy->getVectorNumElements())
+            : NewScalarIndexTy;
 
     // If the element type has zero size then any index over it is equivalent
     // to an index of zero, so replace it with zero if it is not zero already.
     Type *EltTy = GTI.getIndexedType();
     if (EltTy->isSized() && DL.getTypeAllocSize(EltTy) == 0)
       if (!isa<Constant>(*I) || !cast<Constant>(*I)->isNullValue()) {
         *I = Constant::getNullValue(NewIndexType);
         MadeChange = true;
if (!Indices.empty())
                  : GetElementPtrInst::Create(Src->getSourceElementType(),
                                              Src->getOperand(0), Indices,
                                              GEP.getName());
   }
 
   if (GEP.getNumIndices() == 1) {
     unsigned AS = GEP.getPointerAddressSpace();
     if (GEP.getOperand(1)->getType()->getScalarSizeInBits() ==
-        DL.getPointerSizeInBits(AS)) {
+        DL.getIndexSizeInBits(AS)) {
       Type *Ty = GEP.getSourceElementType();
       uint64_t TyAllocSize = DL.getTypeAllocSize(Ty);
 
       bool Matched = false;
       uint64_t C;
       Value *V = nullptr;
       if (TyAllocSize == 1) {
         V = GEP.getOperand(1);
if (HasZeroPointerIndex) {
       // Transform things like:
       // %t = getelementptr i32* bitcast ([2 x i32]* %str to i32*), i32 %V
       // into:  %t1 = getelementptr [2 x i32]* %str, i32 0, i32 %V; bitcast
       Type *SrcElTy = StrippedPtrTy->getElementType();
       Type *ResElTy = GEP.getSourceElementType();
       if (SrcElTy->isArrayTy() &&
           DL.getTypeAllocSize(SrcElTy->getArrayElementType()) ==
               DL.getTypeAllocSize(ResElTy)) {
-        Type *IdxType = DL.getIntPtrType(GEP.getType());
+        Type *IdxType = DL.getIndexType(GEP.getType());
         Value *Idx[2] = { Constant::getNullValue(IdxType), GEP.getOperand(1) };
         Value *NewGEP =
             GEP.isInBounds()
                 ? Builder.CreateInBoundsGEP(nullptr, StrippedPtr, Idx,
                                             GEP.getName())
                 : Builder.CreateGEP(nullptr, StrippedPtr, Idx, GEP.getName());
 
         // V and GEP are both pointer types --> BitCast
if (HasZeroPointerIndex) {
         // factor.
         uint64_t ResSize = DL.getTypeAllocSize(ResElTy);
         uint64_t SrcSize = DL.getTypeAllocSize(SrcElTy);
         if (ResSize && SrcSize % ResSize == 0) {
           Value *Idx = GEP.getOperand(1);
           unsigned BitWidth = Idx->getType()->getPrimitiveSizeInBits();
           uint64_t Scale = SrcSize / ResSize;
 
-          // Earlier transforms ensure that the index has type IntPtrType, which
-          // considerably simplifies the logic by eliminating implicit casts.
-          assert(Idx->getType() == DL.getIntPtrType(GEP.getType()) &&
-                 "Index not cast to pointer width?");
+          // Earlier transforms ensure that the index has the right type
+          // according to Data Layout, which considerably simplifies the
+          // logic by eliminating implicit casts.
+          assert(Idx->getType() == DL.getIndexType(GEP.getType()) &&
+                 "Index type does not match the Data Layout preferences");
 
           bool NSW;
           if (Value *NewIdx = Descale(Idx, APInt(BitWidth, Scale), NSW)) {
             // Successfully decomposed Idx as NewIdx * Scale, form a new GEP.
             // If the multiplication NewIdx * Scale may overflow then the new
             // GEP may not be "inbounds".
             Value *NewGEP =
                 GEP.isInBounds() && NSW
if (HasZeroPointerIndex) {
         uint64_t ResSize = DL.getTypeAllocSize(ResElTy);
         uint64_t ArrayEltSize =
             DL.getTypeAllocSize(SrcElTy->getArrayElementType());
         if (ResSize && ArrayEltSize % ResSize == 0) {
           Value *Idx = GEP.getOperand(1);
           unsigned BitWidth = Idx->getType()->getPrimitiveSizeInBits();
           uint64_t Scale = ArrayEltSize / ResSize;
 
-          // Earlier transforms ensure that the index has type IntPtrType, which
-          // considerably simplifies the logic by eliminating implicit casts.
-          assert(Idx->getType() == DL.getIntPtrType(GEP.getType()) &&
-                 "Index not cast to pointer width?");
+          // Earlier transforms ensure that the index has the right type
+          // according to the Data Layout, which considerably simplifies
+          // the logic by eliminating implicit casts.
+          assert(Idx->getType() == DL.getIndexType(GEP.getType()) &&
+                 "Index type does not match the Data Layout preferences");
 
           bool NSW;
           if (Value *NewIdx = Descale(Idx, APInt(BitWidth, Scale), NSW)) {
             // Successfully decomposed Idx as NewIdx * Scale, form a new GEP.
             // If the multiplication NewIdx * Scale may overflow then the new
             // GEP may not be "inbounds".
-            Value *Off[2] = {
-                Constant::getNullValue(DL.getIntPtrType(GEP.getType())),
-                NewIdx};
+            Type *IndTy = DL.getIndexType(GEP.getType());
+            Value *Off[2] = {Constant::getNullValue(IndTy), NewIdx};
 
             Value *NewGEP = GEP.isInBounds() && NSW
                                 ? Builder.CreateInBoundsGEP(
                                       SrcElTy, StrippedPtr, Off, GEP.getName())
                                 : Builder.CreateGEP(SrcElTy, StrippedPtr, Off,
                                                     GEP.getName());
             // The NewGEP must be pointer typed, so must the old one -> BitCast
             return CastInst::CreatePointerBitCastOrAddrSpaceCast(NewGEP,
Instruction *InstCombiner::visitGetElementPtrInst(GetElementPtrInst &GEP) {
   /// See if we can simplify:
   ///   X = bitcast A* to B*
   ///   Y = gep X, <...constant indices...>
   /// into a gep of the original struct.  This is important for SROA and alias
   /// analysis of unions.  If "A" is also a bitcast, wait for A/X to be merged.
   if (BitCastInst *BCI = dyn_cast<BitCastInst>(PtrOp)) {
     Value *Operand = BCI->getOperand(0);
     PointerType *OpType = cast<PointerType>(Operand->getType());
-    unsigned OffsetBits = DL.getPointerTypeSizeInBits(GEP.getType());
+    unsigned OffsetBits = DL.getIndexTypeSizeInBits(GEP.getType());
     APInt Offset(OffsetBits, 0);
     if (!isa<BitCastInst>(Operand) &&
         GEP.accumulateConstantOffset(DL, Offset)) {
 
       // If this GEP instruction doesn't move the pointer, just replace the GEP
       // with a bitcast of the real input to the dest type.
       if (!Offset) {
         // If the bitcast is of an allocation, and the allocation will be
if (!isa<BitCastInst>(Operand) &&
         if (NGEP->getType()->getPointerAddressSpace() != GEP.getAddressSpace())
           return new AddrSpaceCastInst(NGEP, GEP.getType());
         return new BitCastInst(NGEP, GEP.getType());
       }
     }
   }
 
   if (!GEP.isInBounds()) {
-    unsigned PtrWidth =
-        DL.getPointerSizeInBits(PtrOp->getType()->getPointerAddressSpace());
-    APInt BasePtrOffset(PtrWidth, 0);
+    unsigned IdxWidth =
+        DL.getIndexSizeInBits(PtrOp->getType()->getPointerAddressSpace());
+    APInt BasePtrOffset(IdxWidth, 0);
     Value *UnderlyingPtrOp =
             PtrOp->stripAndAccumulateInBoundsConstantOffsets(DL,
                                                              BasePtrOffset);
     if (auto *AI = dyn_cast<AllocaInst>(UnderlyingPtrOp)) {
       if (GEP.accumulateConstantOffset(DL, BasePtrOffset) &&
           BasePtrOffset.isNonNegative()) {
-        APInt AllocSize(PtrWidth, DL.getTypeAllocSize(AI->getAllocatedType()));
+        APInt AllocSize(IdxWidth, DL.getTypeAllocSize(AI->getAllocatedType()));
         if (BasePtrOffset.ule(AllocSize)) {
           return GetElementPtrInst::CreateInBounds(
               PtrOp, makeArrayRef(Ops).slice(1), GEP.getName());
         }
       }
     }
   }
 

llvm/trunk/lib/Transforms/Scalar/SROA.cpp

for (LoadInst *LI : Loads) {
     uint64_t PartOffset = 0, PartSize = Offsets.Splits.front();
     int Idx = 0, Size = Offsets.Splits.size();
     for (;;) {
       auto *PartTy = Type::getIntNTy(Ty->getContext(), PartSize * 8);
       auto AS = LI->getPointerAddressSpace();
       auto *PartPtrTy = PartTy->getPointerTo(AS);
       LoadInst *PLoad = IRB.CreateAlignedLoad(
           getAdjustedPtr(IRB, DL, BasePtr,
-                         APInt(DL.getPointerSizeInBits(AS), PartOffset),
+                         APInt(DL.getIndexSizeInBits(AS), PartOffset),
                          PartPtrTy, BasePtr->getName() + "."),
           getAdjustedAlignment(LI, PartOffset, DL), /*IsVolatile*/ false,
           LI->getName());
       PLoad->copyMetadata(*LI, LLVMContext::MD_mem_parallel_loop_access); 
 
       // Append this load onto the list of split loads so we can find it later
       // to rewrite the stores.
       SplitLoads.push_back(PLoad);
for (User *LU : LI->users()) {
         uint64_t PartOffset = Idx == 0 ? 0 : Offsets.Splits[Idx - 1];
         auto *PartPtrTy =
             PLoad->getType()->getPointerTo(SI->getPointerAddressSpace());
 
         auto AS = SI->getPointerAddressSpace();
         StoreInst *PStore = IRB.CreateAlignedStore(
             PLoad,
             getAdjustedPtr(IRB, DL, StoreBasePtr,
-                           APInt(DL.getPointerSizeInBits(AS), PartOffset),
+                           APInt(DL.getIndexSizeInBits(AS), PartOffset),
                            PartPtrTy, StoreBasePtr->getName() + "."),
             getAdjustedAlignment(SI, PartOffset, DL), /*IsVolatile*/ false);
         PStore->copyMetadata(*LI, LLVMContext::MD_mem_parallel_loop_access);
         DEBUG(dbgs() << "      +" << PartOffset << ":" << *PStore << "\n");
       }
 
       // We want to immediately iterate on any allocas impacted by splitting
       // this store, and we have to track any promotable alloca (indicated by
for (;;) {
       LoadInst *PLoad;
       if (SplitLoads) {
         PLoad = (*SplitLoads)[Idx];
       } else {
         IRB.SetInsertPoint(LI);
         auto AS = LI->getPointerAddressSpace();
         PLoad = IRB.CreateAlignedLoad(
             getAdjustedPtr(IRB, DL, LoadBasePtr,
-                           APInt(DL.getPointerSizeInBits(AS), PartOffset),
+                           APInt(DL.getIndexSizeInBits(AS), PartOffset),
                            LoadPartPtrTy, LoadBasePtr->getName() + "."),
             getAdjustedAlignment(LI, PartOffset, DL), /*IsVolatile*/ false,
             LI->getName());
       }
 
       // And store this partition.
       IRB.SetInsertPoint(SI);
       auto AS = SI->getPointerAddressSpace();
       StoreInst *PStore = IRB.CreateAlignedStore(
           PLoad,
           getAdjustedPtr(IRB, DL, StoreBasePtr,
-                         APInt(DL.getPointerSizeInBits(AS), PartOffset),
+                         APInt(DL.getIndexSizeInBits(AS), PartOffset),
                          StorePartPtrTy, StoreBasePtr->getName() + "."),
           getAdjustedAlignment(SI, PartOffset, DL), /*IsVolatile*/ false);
 
       // Now build a new slice for the alloca.
       NewSlices.push_back(
           Slice(BaseOffset + PartOffset, BaseOffset + PartOffset + PartSize,
                 &PStore->getOperandUse(PStore->getPointerOperandIndex()),
                 /*IsSplittable*/ false));

llvm/trunk/lib/Transforms/Scalar/SeparateConstOffsetFromGEP.cpp

void SeparateConstOffsetFromGEP::swapGEPOperand(GetElementPtrInst *First,
                                                 GetElementPtrInst *Second) {
   Value *Offset1 = First->getOperand(1);
   Value *Offset2 = Second->getOperand(1);
   First->setOperand(1, Offset2);
   Second->setOperand(1, Offset1);
 
   // We changed p+o+c to p+c+o, p+c may not be inbound anymore.
   const DataLayout &DAL = First->getModule()->getDataLayout();
-  APInt Offset(DAL.getPointerSizeInBits(
+  APInt Offset(DAL.getIndexSizeInBits(
                    cast<PointerType>(First->getType())->getAddressSpace()),
                0);
   Value *NewBase =
       First->stripAndAccumulateInBoundsConstantOffsets(DAL, Offset);
   uint64_t ObjectSize;
   if (!getObjectSize(NewBase, ObjectSize, DAL, TLI) ||
      Offset.ugt(ObjectSize)) {
     First->setIsInBounds(false);
     Second->setIsInBounds(false);
   } else
     First->setIsInBounds(true);
 }

llvm/trunk/lib/Transforms/Utils/Local.cpp

if (auto *CI = dyn_cast<CastInst>(&I)) {
     // No-op casts are irrelevant for debug info.
     MetadataAsValue *CastSrc = wrapMD(I.getOperand(0));
     for (auto *DII : DbgUsers) {
       DII->setOperand(0, CastSrc);
       DEBUG(dbgs() << "SALVAGE: " << *DII << '\n');
     }
   } else if (auto *GEP = dyn_cast<GetElementPtrInst>(&I)) {
     unsigned BitWidth =
-        M.getDataLayout().getPointerSizeInBits(GEP->getPointerAddressSpace());
+        M.getDataLayout().getIndexSizeInBits(GEP->getPointerAddressSpace());
     // Rewrite a constant GEP into a DIExpression.  Since we are performing
     // arithmetic to compute the variable's *value* in the DIExpression, we
     // need to mark the expression with a DW_OP_stack_value.
     APInt Offset(BitWidth, 0);
     if (GEP->accumulateConstantOffset(M.getDataLayout(), Offset))
       for (auto *DII : DbgUsers)
         applyOffset(DII, Offset.getSExtValue());
   } else if (auto *BI = dyn_cast<BinaryOperator>(&I)) {
void llvm::copyRangeMetadata(const DataLayout &DL, const LoadInst &OldLI,
 
   // Give up unless it is converted to a pointer where there is a single very
   // valuable mapping we can do reliably.
   // FIXME: It would be nice to propagate this in more ways, but the type
   // conversions make it hard.
   if (!NewTy->isPointerTy())
     return;
 
-  unsigned BitWidth = DL.getTypeSizeInBits(NewTy);
+  unsigned BitWidth = DL.getIndexTypeSizeInBits(NewTy);
   if (!getConstantRangeFromMetadata(*N).contains(APInt(BitWidth, 0))) {
     MDNode *NN = MDNode::get(OldLI.getContext(), None);
     NewLI.setMetadata(LLVMContext::MD_nonnull, NN);
   }
 }
 
 namespace {
 

llvm/trunk/lib/Transforms/Vectorize/LoadStoreVectorizer.cpp

bool Vectorizer::isConsecutiveAccess(Value *A, Value *B) {
   if (PtrA == PtrB ||
       DL.getTypeStoreSize(PtrATy) != DL.getTypeStoreSize(PtrBTy) ||
       DL.getTypeStoreSize(PtrATy->getScalarType()) !=
           DL.getTypeStoreSize(PtrBTy->getScalarType()))
     return false;
 
   APInt Size(PtrBitWidth, DL.getTypeStoreSize(PtrATy));
 
-  APInt OffsetA(PtrBitWidth, 0), OffsetB(PtrBitWidth, 0);
+  unsigned IdxWidth = DL.getIndexSizeInBits(ASA);
+  APInt OffsetA(IdxWidth, 0), OffsetB(IdxWidth, 0);
   PtrA = PtrA->stripAndAccumulateInBoundsConstantOffsets(DL, OffsetA);
   PtrB = PtrB->stripAndAccumulateInBoundsConstantOffsets(DL, OffsetB);
 
   APInt OffsetDelta = OffsetB - OffsetA;
 
   // Check if they are based on the same pointer. That makes the offsets
   // sufficient.
   if (PtrA == PtrB)

llvm/trunk/test/Transforms/InstCombine/gep-custom-dl.ll

+; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
+; RUN: opt < %s -instcombine -S | FileCheck %s
+
+target datalayout = "e-m:m-p:40:64:64:32-i32:32-i16:16-i8:8-n32"
+
+%struct.B = type { double }
+%struct.A = type { %struct.B, i32, i32 }
+%struct.C = type { [7 x i8] }
+
+
+@Global = constant [10 x i8] c"helloworld"
+
+
+; Test that two array indexing geps fold
+define i32* @test1(i32* %I) {
+; CHECK-LABEL: @test1(
+; CHECK-NEXT:    [[B:%.*]] = getelementptr i32, i32* [[I:%.*]], i32 21
+; CHECK-NEXT:    ret i32* [[B]]
+;
+  %A = getelementptr i32, i32* %I, i8 17
+  %B = getelementptr i32, i32* %A, i16 4
+  ret i32* %B
+}
+
+; Test that two getelementptr insts fold
+define i32* @test2({ i32 }* %I) {
+; CHECK-LABEL: @test2(
+; CHECK-NEXT:    [[B:%.*]] = getelementptr { i32 }, { i32 }* [[I:%.*]], i32 1, i32 0
+; CHECK-NEXT:    ret i32* [[B]]
+;
+  %A = getelementptr { i32 }, { i32 }* %I, i32 1
+  %B = getelementptr { i32 }, { i32 }* %A, i32 0, i32 0
+  ret i32* %B
+}
+
+define void @test3(i8 %B) {
+; This should be turned into a constexpr instead of being an instruction
+; CHECK-LABEL: @test3(
+; CHECK-NEXT:    store i8 [[B:%.*]], i8* getelementptr inbounds ([10 x i8], [10 x i8]* @Global, i32 0, i32 4), align 1
+; CHECK-NEXT:    ret void
+;
+  %A = getelementptr [10 x i8], [10 x i8]* @Global, i32 0, i32 4
+  store i8 %B, i8* %A
+  ret void
+}
+
+%as1_ptr_struct = type { i32 addrspace(1)* }
+%as2_ptr_struct = type { i32 addrspace(2)* }
+
+@global_as2 = addrspace(2) global i32 zeroinitializer
+@global_as1_as2_ptr = addrspace(1) global %as2_ptr_struct { i32 addrspace(2)* @global_as2 }
+
+; This should be turned into a constexpr instead of being an instruction
+define void @test_evaluate_gep_nested_as_ptrs(i32 addrspace(2)* %B) {
+; CHECK-LABEL: @test_evaluate_gep_nested_as_ptrs(
+; CHECK-NEXT:    store i32 addrspace(2)* [[B:%.*]], i32 addrspace(2)* addrspace(1)* getelementptr inbounds (%as2_ptr_struct, [[AS2_PTR_STRUCT:%.*]] addrspace(1)* @global_as1_as2_ptr, i32 0, i32 0), align 8
+; CHECK-NEXT:    ret void
+;
+  %A = getelementptr %as2_ptr_struct, %as2_ptr_struct addrspace(1)* @global_as1_as2_ptr, i32 0, i32 0
+  store i32 addrspace(2)* %B, i32 addrspace(2)* addrspace(1)* %A
+  ret void
+}
+
+@arst = addrspace(1) global [4 x i8 addrspace(2)*] zeroinitializer
+
+define void @test_evaluate_gep_as_ptrs_array(i8 addrspace(2)* %B) {
+; CHECK-LABEL: @test_evaluate_gep_as_ptrs_array(
+; CHECK-NEXT:    store i8 addrspace(2)* [[B:%.*]], i8 addrspace(2)* addrspace(1)* getelementptr inbounds ([4 x i8 addrspace(2)*], [4 x i8 addrspace(2)*] addrspace(1)* @arst, i32 0, i32 2), align 16
+; CHECK-NEXT:    ret void
+;
+
+  %A = getelementptr [4 x i8 addrspace(2)*], [4 x i8 addrspace(2)*] addrspace(1)* @arst, i16 0, i16 2
+  store i8 addrspace(2)* %B, i8 addrspace(2)* addrspace(1)* %A
+  ret void
+}
+
+define i32* @test4(i32* %I, i32 %C, i32 %D) {
+; CHECK-LABEL: @test4(
+; CHECK-NEXT:    [[A:%.*]] = getelementptr i32, i32* [[I:%.*]], i32 [[C:%.*]]
+; CHECK-NEXT:    [[B:%.*]] = getelementptr i32, i32* [[A]], i32 [[D:%.*]]
+; CHECK-NEXT:    ret i32* [[B]]
+;
+  %A = getelementptr i32, i32* %I, i32 %C
+  %B = getelementptr i32, i32* %A, i32 %D
+  ret i32* %B
+}
+
+
+define i1 @test5({ i32, i32 }* %x, { i32, i32 }* %y) {
+; CHECK-LABEL: @test5(
+; CHECK-NEXT:    [[TMP_4:%.*]] = icmp eq { i32, i32 }* [[X:%.*]], [[Y:%.*]]
+; CHECK-NEXT:    ret i1 [[TMP_4]]
+;
+  %tmp.1 = getelementptr { i32, i32 }, { i32, i32 }* %x, i32 0, i32 1
+  %tmp.3 = getelementptr { i32, i32 }, { i32, i32 }* %y, i32 0, i32 1
+  ;; seteq x, y
+  %tmp.4 = icmp eq i32* %tmp.1, %tmp.3
+  ret i1 %tmp.4
+}
+
+%S = type { i32, [ 100 x i32] }
+
+define <2 x i1> @test6(<2 x i32> %X, <2 x %S*> %P) nounwind {
+; CHECK-LABEL: @test6(
+; CHECK-NEXT:    [[C:%.*]] = icmp eq <2 x i32> [[X:%.*]], <i32 1073741823, i32 1073741823>
+; CHECK-NEXT:    ret <2 x i1> [[C]]
+;
+  %A = getelementptr inbounds %S, <2 x %S*> %P, <2 x i32> zeroinitializer, <2 x i32> <i32 1, i32 1>, <2 x i32> %X
+  %B = getelementptr inbounds %S, <2 x %S*> %P, <2 x i32> <i32 0, i32 0>, <2 x i32> <i32 0, i32 0>
+  %C = icmp eq <2 x i32*> %A, %B
+  ret <2 x i1> %C
+}
+
+@G = external global [3 x i8]
+define i8* @test7(i16 %Idx) {
+; CHECK-LABEL: @test7(
+; CHECK-NEXT:    [[ZE_IDX:%.*]] = zext i16 [[IDX:%.*]] to i32
+; CHECK-NEXT:    [[TMP:%.*]] = getelementptr [3 x i8], [3 x i8]* @G, i32 0, i32 [[ZE_IDX]]
+; CHECK-NEXT:    ret i8* [[TMP]]
+;
+  %ZE_Idx = zext i16 %Idx to i32
+  %tmp = getelementptr i8, i8* getelementptr ([3 x i8], [3 x i8]* @G, i32 0, i32 0), i32 %ZE_Idx
+  ret i8* %tmp
+}
+
+
+; Test folding of constantexpr geps into normal geps.
+@Array = external global [40 x i32]
+define i32 *@test8(i32 %X) {
+; CHECK-LABEL: @test8(
+; CHECK-NEXT:    [[A:%.*]] = getelementptr [40 x i32], [40 x i32]* @Array, i32 0, i32 [[X:%.*]]
+; CHECK-NEXT:    ret i32* [[A]]
+;
+  %A = getelementptr i32, i32* getelementptr ([40 x i32], [40 x i32]* @Array, i32 0, i32 0), i32 %X
+  ret i32* %A
+}
+
+define i32 *@test9(i32 *%base, i8 %ind) {
+; CHECK-LABEL: @test9(
+; CHECK-NEXT:    [[TMP1:%.*]] = sext i8 [[IND:%.*]] to i32
+; CHECK-NEXT:    [[RES:%.*]] = getelementptr i32, i32* [[BASE:%.*]], i32 [[TMP1]]
+; CHECK-NEXT:    ret i32* [[RES]]
+;
+  %res = getelementptr i32, i32 *%base, i8 %ind
+  ret i32* %res
+}
+
+define i32 @test10() {
+; CHECK-LABEL: @test10(
+; CHECK-NEXT:    ret i32 8
+;
+  %A = getelementptr { i32, double }, { i32, double }* null, i32 0, i32 1
+  %B = ptrtoint double* %A to i32
+  ret i32 %B
+}

llvm/trunk/test/Transforms/InstCombine/icmp-custom-dl.ll

+; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
+; RUN: opt < %s -instcombine -S | FileCheck %s
+
+target datalayout = "e-p:40:64:64:32-p1:16:16:16-p2:32:32:32-p3:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-s0:64:64-f80:128:128-n8:16:32:64"
+
+declare i32 @test58_d(i64 )
+
+define i1 @test59(i8* %foo) {
+; CHECK-LABEL: @test59(
+; CHECK-NEXT:    [[GEP1:%.*]] = getelementptr inbounds i8, i8* [[FOO:%.*]], i32 8
+; CHECK-NEXT:    [[TMP1:%.*]] = ptrtoint i8* [[GEP1]] to i32
+; CHECK-NEXT:    [[USE:%.*]] = zext i32 [[TMP1]] to i64
+; CHECK-NEXT:    [[CALL:%.*]] = call i32 @test58_d(i64 [[USE]])
+; CHECK-NEXT:    ret i1 true
+;
+  %bit = bitcast i8* %foo to i32*
+  %gep1 = getelementptr inbounds i32, i32* %bit, i64 2
+  %gep2 = getelementptr inbounds i8, i8* %foo, i64 10
+  %cast1 = bitcast i32* %gep1 to i8*
+  %cmp = icmp ult i8* %cast1, %gep2
+  %use = ptrtoint i8* %cast1 to i64
+  %call = call i32 @test58_d(i64 %use)
+  ret i1 %cmp
+}
+
+define i1 @test59_as1(i8 addrspace(1)* %foo) {
+; CHECK-LABEL: @test59_as1(
+; CHECK-NEXT:    [[GEP1:%.*]] = getelementptr inbounds i8, i8 addrspace(1)* [[FOO:%.*]], i16 8
+; CHECK-NEXT:    [[TMP1:%.*]] = ptrtoint i8 addrspace(1)* [[GEP1]] to i16
+; CHECK-NEXT:    [[USE:%.*]] = zext i16 [[TMP1]] to i64
+; CHECK-NEXT:    [[CALL:%.*]] = call i32 @test58_d(i64 [[USE]])
+; CHECK-NEXT:    ret i1 true
+;
+  %bit = bitcast i8 addrspace(1)* %foo to i32 addrspace(1)*
+  %gep1 = getelementptr inbounds i32, i32 addrspace(1)* %bit, i64 2
+  %gep2 = getelementptr inbounds i8, i8 addrspace(1)* %foo, i64 10
+  %cast1 = bitcast i32 addrspace(1)* %gep1 to i8 addrspace(1)*
+  %cmp = icmp ult i8 addrspace(1)* %cast1, %gep2
+  %use = ptrtoint i8 addrspace(1)* %cast1 to i64
+  %call = call i32 @test58_d(i64 %use)
+  ret i1 %cmp
+}
+
+define i1 @test60(i8* %foo, i64 %i, i64 %j) {
+; CHECK-LABEL: @test60(
+; CHECK-NEXT:    [[TMP1:%.*]] = trunc i64 [[I:%.*]] to i32
+; CHECK-NEXT:    [[TMP2:%.*]] = trunc i64 [[J:%.*]] to i32
+; CHECK-NEXT:    [[GEP1_IDX:%.*]] = shl nuw i32 [[TMP1]], 2
+; CHECK-NEXT:    [[TMP3:%.*]] = icmp slt i32 [[GEP1_IDX]], [[TMP2]]
+; CHECK-NEXT:    ret i1 [[TMP3]]
+;
+  %bit = bitcast i8* %foo to i32*
+  %gep1 = getelementptr inbounds i32, i32* %bit, i64 %i
+  %gep2 = getelementptr inbounds i8, i8* %foo, i64 %j
+  %cast1 = bitcast i32* %gep1 to i8*
+  %cmp = icmp ult i8* %cast1, %gep2
+  ret i1 %cmp
+}
+
+define i1 @test60_as1(i8 addrspace(1)* %foo, i64 %i, i64 %j) {
+; CHECK-LABEL: @test60_as1(
+; CHECK-NEXT:    [[TMP1:%.*]] = trunc i64 [[I:%.*]] to i16
+; CHECK-NEXT:    [[TMP2:%.*]] = trunc i64 [[J:%.*]] to i16
+; CHECK-NEXT:    [[GEP1_IDX:%.*]] = shl nuw i16 [[TMP1]], 2
+; CHECK-NEXT:    [[TMP3:%.*]] = icmp slt i16 [[GEP1_IDX]], [[TMP2]]
+; CHECK-NEXT:    ret i1 [[TMP3]]
+;
+  %bit = bitcast i8 addrspace(1)* %foo to i32 addrspace(1)*
+  %gep1 = getelementptr inbounds i32, i32 addrspace(1)* %bit, i64 %i
+  %gep2 = getelementptr inbounds i8, i8 addrspace(1)* %foo, i64 %j
+  %cast1 = bitcast i32 addrspace(1)* %gep1 to i8 addrspace(1)*
+  %cmp = icmp ult i8 addrspace(1)* %cast1, %gep2
+  ret i1 %cmp
+}
+
+; Same as test60, but look through an addrspacecast instead of a
+; bitcast. This uses the same sized addrspace.
+define i1 @test60_addrspacecast(i8* %foo, i64 %i, i64 %j) {
+; CHECK-LABEL: @test60_addrspacecast(
+; CHECK-NEXT:    [[TMP1:%.*]] = trunc i64 [[J:%.*]] to i32
+; CHECK-NEXT:    [[I_TR:%.*]] = trunc i64 [[I:%.*]] to i32
+; CHECK-NEXT:    [[TMP2:%.*]] = shl i32 [[I_TR]], 2
+; CHECK-NEXT:    [[TMP3:%.*]] = icmp slt i32 [[TMP2]], [[TMP1]]
+; CHECK-NEXT:    ret i1 [[TMP3]]
+;
+  %bit = addrspacecast i8* %foo to i32 addrspace(3)*
+  %gep1 = getelementptr inbounds i32, i32 addrspace(3)* %bit, i64 %i
+  %gep2 = getelementptr inbounds i8, i8* %foo, i64 %j
+  %cast1 = addrspacecast i32 addrspace(3)* %gep1 to i8*
+  %cmp = icmp ult i8* %cast1, %gep2
+  ret i1 %cmp
+}
+
+define i1 @test60_addrspacecast_smaller(i8* %foo, i16 %i, i64 %j) {
+; CHECK-LABEL: @test60_addrspacecast_smaller(
+; CHECK-NEXT:    [[GEP1_IDX:%.*]] = shl nuw i16 [[I:%.*]], 2
+; CHECK-NEXT:    [[TMP1:%.*]] = trunc i64 [[J:%.*]] to i16
+; CHECK-NEXT:    [[TMP2:%.*]] = icmp slt i16 [[GEP1_IDX]], [[TMP1]]
+; CHECK-NEXT:    ret i1 [[TMP2]]
+;
+  %bit = addrspacecast i8* %foo to i32 addrspace(1)*
+  %gep1 = getelementptr inbounds i32, i32 addrspace(1)* %bit, i16 %i
+  %gep2 = getelementptr inbounds i8, i8* %foo, i64 %j
+  %cast1 = addrspacecast i32 addrspace(1)* %gep1 to i8*
+  %cmp = icmp ult i8* %cast1, %gep2
+  ret i1 %cmp
+}
+
+define i1 @test60_addrspacecast_larger(i8 addrspace(1)* %foo, i32 %i, i16 %j) {
+; CHECK-LABEL: @test60_addrspacecast_larger(
+; CHECK-NEXT:    [[I_TR:%.*]] = trunc i32 [[I:%.*]] to i16
+; CHECK-NEXT:    [[TMP1:%.*]] = shl i16 [[I_TR]], 2
+; CHECK-NEXT:    [[TMP2:%.*]] = icmp slt i16 [[TMP1]], [[J:%.*]]
+; CHECK-NEXT:    ret i1 [[TMP2]]
+;
+  %bit = addrspacecast i8 addrspace(1)* %foo to i32 addrspace(2)*
+  %gep1 = getelementptr inbounds i32, i32 addrspace(2)* %bit, i32 %i
+  %gep2 = getelementptr inbounds i8, i8 addrspace(1)* %foo, i16 %j
+  %cast1 = addrspacecast i32 addrspace(2)* %gep1 to i8 addrspace(1)*
+  %cmp = icmp ult i8 addrspace(1)* %cast1, %gep2
+  ret i1 %cmp
+}
+
+define i1 @test61(i8* %foo, i64 %i, i64 %j) {
+; CHECK-LABEL: @test61(
+; CHECK-NEXT:    [[BIT:%.*]] = bitcast i8* [[FOO:%.*]] to i32*
+; CHECK-NEXT:    [[TMP1:%.*]] = trunc i64 [[I:%.*]] to i32
+; CHECK-NEXT:    [[GEP1:%.*]] = getelementptr i32, i32* [[BIT]], i32 [[TMP1]]
+; CHECK-NEXT:    [[TMP2:%.*]] = trunc i64 [[J:%.*]] to i32
+; CHECK-NEXT:    [[GEP2:%.*]] = getelementptr i8, i8* [[FOO]], i32 [[TMP2]]
+; CHECK-NEXT:    [[CAST1:%.*]] = bitcast i32* [[GEP1]] to i8*
+; CHECK-NEXT:    [[CMP:%.*]] = icmp ugt i8* [[GEP2]], [[CAST1]]
+; CHECK-NEXT:    ret i1 [[CMP]]
+;
+  %bit = bitcast i8* %foo to i32*
+  %gep1 = getelementptr i32, i32* %bit, i64 %i
+  %gep2 = getelementptr  i8,  i8* %foo, i64 %j
+  %cast1 = bitcast i32* %gep1 to i8*
+  %cmp = icmp ult i8* %cast1, %gep2
+  ret i1 %cmp
+; Don't transform non-inbounds GEPs.
+}
+
+define i1 @test61_as1(i8 addrspace(1)* %foo, i16 %i, i16 %j) {
+; CHECK-LABEL: @test61_as1(
+; CHECK-NEXT:    [[BIT:%.*]] = bitcast i8 addrspace(1)* [[FOO:%.*]] to i32 addrspace(1)*
+; CHECK-NEXT:    [[GEP1:%.*]] = getelementptr i32, i32 addrspace(1)* [[BIT]], i16 [[I:%.*]]
+; CHECK-NEXT:    [[GEP2:%.*]] = getelementptr i8, i8 addrspace(1)* [[FOO]], i16 [[J:%.*]]
+; CHECK-NEXT:    [[CAST1:%.*]] = bitcast i32 addrspace(1)* [[GEP1]] to i8 addrspace(1)*
+; CHECK-NEXT:    [[CMP:%.*]] = icmp ugt i8 addrspace(1)* [[GEP2]], [[CAST1]]
+; CHECK-NEXT:    ret i1 [[CMP]]
+;
+  %bit = bitcast i8 addrspace(1)* %foo to i32 addrspace(1)*
+  %gep1 = getelementptr i32, i32 addrspace(1)* %bit, i16 %i
+  %gep2 = getelementptr i8, i8 addrspace(1)* %foo, i16 %j
+  %cast1 = bitcast i32 addrspace(1)* %gep1 to i8 addrspace(1)*
+  %cmp = icmp ult i8 addrspace(1)* %cast1, %gep2
+  ret i1 %cmp
+; Don't transform non-inbounds GEPs.
+}
+
+define i1 @test62(i8* %a) {
+; CHECK-LABEL: @test62(
+; CHECK-NEXT:    ret i1 true
+;
+  %arrayidx1 = getelementptr inbounds i8, i8* %a, i64 1
+  %arrayidx2 = getelementptr inbounds i8, i8* %a, i64 10
+  %cmp = icmp slt i8* %arrayidx1, %arrayidx2
+  ret i1 %cmp
+}
+
+define i1 @test62_as1(i8 addrspace(1)* %a) {
+; CHECK-LABEL: @test62_as1(
+; CHECK-NEXT:    ret i1 true
+;
+  %arrayidx1 = getelementptr inbounds i8, i8 addrspace(1)* %a, i64 1
+  %arrayidx2 = getelementptr inbounds i8, i8 addrspace(1)* %a, i64 10
+  %cmp = icmp slt i8 addrspace(1)* %arrayidx1, %arrayidx2
+  ret i1 %cmp
+}
+
+
+; Variation of the above with an ashr
+define i1 @icmp_and_ashr_multiuse(i32 %X) {
+; CHECK-LABEL: @icmp_and_ashr_multiuse(
+; CHECK-NEXT:    [[AND:%.*]] = and i32 [[X:%.*]], 240
+; CHECK-NEXT:    [[AND2:%.*]] = and i32 [[X]], 496
+; CHECK-NEXT:    [[TOBOOL:%.*]] = icmp ne i32 [[AND]], 224
+; CHECK-NEXT:    [[TOBOOL2:%.*]] = icmp ne i32 [[AND2]], 432
+; CHECK-NEXT:    [[AND3:%.*]] = and i1 [[TOBOOL]], [[TOBOOL2]]
+; CHECK-NEXT:    ret i1 [[AND3]]
+;
+  %shr = ashr i32 %X, 4
+  %and = and i32 %shr, 15
+  %and2 = and i32 %shr, 31 ; second use of the shift
+  %tobool = icmp ne i32 %and, 14
+  %tobool2 = icmp ne i32 %and2, 27
+  %and3 = and i1 %tobool, %tobool2
+  ret i1 %and3
+}
+
+define i1 @icmp_lshr_and_overshift(i8 %X) {
+; CHECK-LABEL: @icmp_lshr_and_overshift(
+; CHECK-NEXT:    [[TOBOOL:%.*]] = icmp ugt i8 [[X:%.*]], 31
+; CHECK-NEXT:    ret i1 [[TOBOOL]]
+;
+  %shr = lshr i8 %X, 5
+  %and = and i8 %shr, 15
+  %tobool = icmp ne i8 %and, 0
+  ret i1 %tobool
+}
+
+; We shouldn't simplify this because the and uses bits that are shifted in.
+define i1 @icmp_ashr_and_overshift(i8 %X) {
+; CHECK-LABEL: @icmp_ashr_and_overshift(
+; CHECK-NEXT:    [[SHR:%.*]] = ashr i8 [[X:%.*]], 5
+; CHECK-NEXT:    [[AND:%.*]] = and i8 [[SHR]], 15
+; CHECK-NEXT:    [[TOBOOL:%.*]] = icmp ne i8 [[AND]], 0
+; CHECK-NEXT:    ret i1 [[TOBOOL]]
+;
+  %shr = ashr i8 %X, 5
+  %and = and i8 %shr, 15
+  %tobool = icmp ne i8 %and, 0
+  ret i1 %tobool
+}
+
+; PR16244
+define i1 @test71(i8* %x) {
+; CHECK-LABEL: @test71(
+; CHECK-NEXT:    ret i1 false
+;
+  %a = getelementptr i8, i8* %x, i64 8
+  %b = getelementptr inbounds i8, i8* %x, i64 8
+  %c = icmp ugt i8* %a, %b
+  ret i1 %c
+}
+
+define i1 @test71_as1(i8 addrspace(1)* %x) {
+; CHECK-LABEL: @test71_as1(
+; CHECK-NEXT:    ret i1 false
+;
+  %a = getelementptr i8, i8 addrspace(1)* %x, i64 8
+  %b = getelementptr inbounds i8, i8 addrspace(1)* %x, i64 8
+  %c = icmp ugt i8 addrspace(1)* %a, %b
+  ret i1 %c
+}
+

llvm/trunk/test/Transforms/LoopIdiom/struct-custom-dl.ll

+; RUN: opt -basicaa -loop-idiom < %s -S | FileCheck %s
+target datalayout = "e-p:40:64:64:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-s0:64:64-f80:128:128-n8:16:32:64"
+
+%struct.foo = type { i32, i32 }
+%struct.foo1 = type { i32, i32, i32 }
+%struct.foo2 = type { i32, i16, i16 }
+
+;void bar1(foo_t *f, unsigned n) {
+;  for (unsigned i = 0; i < n; ++i) {
+;    f[i].a = 0;
+;    f[i].b = 0;
+;  }
+;}
+define void @bar1(%struct.foo* %f, i32 %n) nounwind ssp {
+entry:
+  %cmp1 = icmp eq i32 %n, 0
+  br i1 %cmp1, label %for.end, label %for.body.preheader
+
+for.body.preheader:                               ; preds = %entry
+  br label %for.body
+
+for.body:                                         ; preds = %for.body.preheader, %for.body
+  %indvars.iv = phi i32 [ 0, %for.body.preheader ], [ %indvars.iv.next, %for.body ]
+  %a = getelementptr inbounds %struct.foo, %struct.foo* %f, i32 %indvars.iv, i32 0
+  store i32 0, i32* %a, align 4
+  %b = getelementptr inbounds %struct.foo, %struct.foo* %f, i32 %indvars.iv, i32 1
+  store i32 0, i32* %b, align 4
+  %indvars.iv.next = add nuw nsw i32 %indvars.iv, 1
+  %exitcond = icmp ne i32 %indvars.iv.next, %n
+  br i1 %exitcond, label %for.body, label %for.end.loopexit
+
+for.end.loopexit:                                 ; preds = %for.body
+  br label %for.end
+
+for.end:                                          ; preds = %for.end.loopexit, %entry
+  ret void
+; CHECK-LABEL: @bar1(
+; CHECK: call void @llvm.memset
+; CHECK-NOT: store
+}
+
+;void bar2(foo_t *f, unsigned n) {
+;  for (unsigned i = 0; i < n; ++i) {
+;    f[i].b = 0;
+;    f[i].a = 0;
+;  }
+;}
+define void @bar2(%struct.foo* %f, i32 %n) nounwind ssp {
+entry:
+  %cmp1 = icmp eq i32 %n, 0
+  br i1 %cmp1, label %for.end, label %for.body.preheader
+
+for.body.preheader:                               ; preds = %entry
+  br label %for.body
+
+for.body:                                         ; preds = %for.body.preheader, %for.body
+  %indvars.iv = phi i32 [ 0, %for.body.preheader ], [ %indvars.iv.next, %for.body ]
+  %b = getelementptr inbounds %struct.foo, %struct.foo* %f, i32 %indvars.iv, i32 1
+  store i32 0, i32* %b, align 4
+  %a = getelementptr inbounds %struct.foo, %struct.foo* %f, i32 %indvars.iv, i32 0
+  store i32 0, i32* %a, align 4
+  %indvars.iv.next = add nuw nsw i32 %indvars.iv, 1
+  %exitcond = icmp ne i32 %indvars.iv.next, %n
+  br i1 %exitcond, label %for.body, label %for.end.loopexit
+
+for.end.loopexit:                                 ; preds = %for.body
+  br label %for.end
+
+for.end:                                          ; preds = %for.end.loopexit, %entry
+  ret void
+; CHECK-LABEL: @bar2(
+; CHECK: call void @llvm.memset
+; CHECK-NOT: store
+}
+
+;void bar3(foo_t *f, unsigned n) {
+;  for (unsigned i = n; i > 0; --i) {
+;    f[i].a = 0;
+;    f[i].b = 0;
+;  }
+;}
+define void @bar3(%struct.foo* nocapture %f, i32 %n) nounwind ssp {
+entry:
+  %cmp1 = icmp eq i32 %n, 0
+  br i1 %cmp1, label %for.end, label %for.body.preheader
+
+for.body.preheader:                               ; preds = %entry
+  br label %for.body
+
+for.body:                                         ; preds = %for.body.preheader, %for.body
+  %indvars.iv = phi i32 [ %n, %for.body.preheader ], [ %indvars.iv.next, %for.body ]
+  %a = getelementptr inbounds %struct.foo, %struct.foo* %f, i32 %indvars.iv, i32 0
+  store i32 0, i32* %a, align 4
+  %b = getelementptr inbounds %struct.foo, %struct.foo* %f, i32 %indvars.iv, i32 1
+  store i32 0, i32* %b, align 4
+  %dec = add i32 %indvars.iv, -1
+  %cmp = icmp eq i32 %dec, 0
+  %indvars.iv.next = add nsw i32 %indvars.iv, -1
+  br i1 %cmp, label %for.end.loopexit, label %for.body
+
+for.end.loopexit:                                 ; preds = %for.body
+  br label %for.end
+
+for.end:                                          ; preds = %for.end.loopexit, %entry
+  ret void
+; CHECK-LABEL: @bar3(
+; CHECK: call void @llvm.memset
+; CHECK-NOT: store
+}
+
+;void bar4(foo_t *f, unsigned n) {
+;  for (unsigned i = 0; i < n; ++i) {
+;    f[i].a = 0;
+;    f[i].b = 1;
+;  }
+;}
+define void @bar4(%struct.foo* nocapture %f, i32 %n) nounwind ssp {
+entry:
+  %cmp1 = icmp eq i32 %n, 0
+  br i1 %cmp1, label %for.end, label %for.body.preheader
+
+for.body.preheader:                               ; preds = %entry
+  br label %for.body
+
+for.body:                                         ; preds = %for.body.preheader, %for.body
+  %indvars.iv = phi i32 [ 0, %for.body.preheader ], [ %indvars.iv.next, %for.body ]
+  %a = getelementptr inbounds %struct.foo, %struct.foo* %f, i32 %indvars.iv, i32 0
+  store i32 0, i32* %a, align 4
+  %b = getelementptr inbounds %struct.foo, %struct.foo* %f, i32 %indvars.iv, i32 1
+  store i32 1, i32* %b, align 4
+  %indvars.iv.next = add nuw nsw i32 %indvars.iv, 1
+  %exitcond = icmp ne i32 %indvars.iv.next, %n
+  br i1 %exitcond, label %for.body, label %for.end.loopexit
+
+for.end.loopexit:                                 ; preds = %for.body
+  br label %for.end
+
+for.end:                                          ; preds = %for.end.loopexit, %entry
+  ret void
+; CHECK-LABEL: @bar4(
+; CHECK-NOT: call void @llvm.memset
+}
+
+;void bar5(foo1_t *f, unsigned n) {
+;  for (unsigned i = 0; i < n; ++i) {
+;    f[i].a = 0;
+;    f[i].b = 0;
+;  }
+;}
+define void @bar5(%struct.foo1* nocapture %f, i32 %n) nounwind ssp {
+entry:
+  %cmp1 = icmp eq i32 %n, 0
+  br i1 %cmp1, label %for.end, label %for.body.preheader
+
+for.body.preheader:                               ; preds = %entry
+  br label %for.body
+
+for.body:                                         ; preds = %for.body.preheader, %for.body
+  %indvars.iv = phi i32 [ 0, %for.body.preheader ], [ %indvars.iv.next, %for.body ]
+  %a = getelementptr inbounds %struct.foo1, %struct.foo1* %f, i32 %indvars.iv, i32 0
+  store i32 0, i32* %a, align 4
+  %b = getelementptr inbounds %struct.foo1, %struct.foo1* %f, i32 %indvars.iv, i32 1
+  store i32 0, i32* %b, align 4
+  %indvars.iv.next = add nuw nsw i32 %indvars.iv, 1
+  %exitcond = icmp ne i32 %indvars.iv.next, %n
+  br i1 %exitcond, label %for.body, label %for.end.loopexit
+
+for.end.loopexit:                                 ; preds = %for.body
+  br label %for.end
+
+for.end:                                          ; preds = %for.end.loopexit, %entry
+  ret void
+; CHECK-LABEL: @bar5(
+; CHECK-NOT: call void @llvm.memset
+}
+
+;void bar6(foo2_t *f, unsigned n) {
+;  for (unsigned i = 0; i < n; ++i) {
+;    f[i].a = 0;
+;    f[i].b = 0;
+;    f[i].c = 0;
+;  }
+;}
+define void @bar6(%struct.foo2* nocapture %f, i32 %n) nounwind ssp {
+entry:
+  %cmp1 = icmp eq i32 %n, 0
+  br i1 %cmp1, label %for.end, label %for.body.preheader
+
+for.body.preheader:                               ; preds = %entry
+  br label %for.body
+
+for.body:                                         ; preds = %for.body.preheader, %for.body
+  %indvars.iv = phi i32 [ 0, %for.body.preheader ], [ %indvars.iv.next, %for.body ]
+  %a = getelementptr inbounds %struct.foo2, %struct.foo2* %f, i32 %indvars.iv, i32 0
+  store i32 0, i32* %a, align 4
+  %b = getelementptr inbounds %struct.foo2, %struct.foo2* %f, i32 %indvars.iv, i32 1
+  store i16 0, i16* %b, align 4
+  %c = getelementptr inbounds %struct.foo2, %struct.foo2* %f, i32 %indvars.iv, i32 2
+  store i16 0, i16* %c, align 2
+  %indvars.iv.next = add nuw nsw i32 %indvars.iv, 1
+  %exitcond = icmp ne i32 %indvars.iv.next, %n
+  br i1 %exitcond, label %for.body, label %for.end.loopexit
+
+for.end.loopexit:                                 ; preds = %for.body
+  br label %for.end
+
+for.end:                                          ; preds = %for.end.loopexit, %entry
+  ret void
+; CHECK-LABEL: @bar6(
+; CHECK: call void @llvm.memset
+; CHECK-NOT: store
+}

llvm/trunk/test/Transforms/LoopIdiom/unroll-custom-dl.ll

+; RUN: opt -basicaa -loop-idiom < %s -S | FileCheck %s
+target datalayout = "e-p:64:64:64:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-s0:64:64-f80:128:128-n8:16:32:64"
+
+; CHECK: @.memset_pattern = private unnamed_addr constant [4 x i32] [i32 2, i32 2, i32 2, i32 2], align 16
+
+target triple = "x86_64-apple-darwin10.0.0"
+
+;void test(int *f, unsigned n) {
+;  for (unsigned i = 0; i < 2 * n; i += 2) {
+;    f[i] = 0;
+;    f[i+1] = 0;
+;  }
+;}
+define void @test(i32* %f, i32 %n) nounwind ssp {
+entry:
+  %0 = shl i32 %n, 1
+  %cmp1 = icmp eq i32 %0, 0
+  br i1 %cmp1, label %for.end, label %for.body.preheader
+
+for.body.preheader:                               ; preds = %entry
+  br label %for.body
+
+for.body:                                         ; preds = %for.body.preheader, %for.body
+  %indvars.iv = phi i32 [ 0, %for.body.preheader ], [ %indvars.iv.next, %for.body ]
+  %arrayidx = getelementptr inbounds i32, i32* %f, i32 %indvars.iv
+  store i32 0, i32* %arrayidx, align 4
+  %1 = or i32 %indvars.iv, 1
+  %arrayidx2 = getelementptr inbounds i32, i32* %f, i32 %1
+  store i32 0, i32* %arrayidx2, align 4
+  %indvars.iv.next = add nuw nsw i32 %indvars.iv, 2
+  %cmp = icmp ult i32 %indvars.iv.next, %0
+  br i1 %cmp, label %for.body, label %for.end.loopexit
+
+for.end.loopexit:                                 ; preds = %for.body
+  br label %for.end
+
+for.end:                                          ; preds = %for.end.loopexit, %entry
+  ret void
+; CHECK-LABEL: @test(
+; CHECK: call void @llvm.memset
+; CHECK-NOT: store
+}
+
+;void test_pattern(int *f, unsigned n) {
+;  for (unsigned i = 0; i < 2 * n; i += 2) {
+;    f[i] = 2;
+;    f[i+1] = 2;
+;  }
+;}
+define void @test_pattern(i32* %f, i32 %n) nounwind ssp {
+entry:
+  %mul = shl i32 %n, 1
+  %cmp1 = icmp eq i32 %mul, 0
+  br i1 %cmp1, label %for.end, label %for.body.preheader
+
+for.body.preheader:                               ; preds = %entry
+  br label %for.body
+
+for.body:                                         ; preds = %for.body.preheader, %for.body
+  %indvars.iv = phi i32 [ 0, %for.body.preheader ], [ %indvars.iv.next, %for.body ]
+  %arrayidx = getelementptr inbounds i32, i32* %f, i32 %indvars.iv
+  store i32 2, i32* %arrayidx, align 4
+  %x1 = or i32 %indvars.iv, 1
+  %arrayidx2 = getelementptr inbounds i32, i32* %f, i32 %x1
+  store i32 2, i32* %arrayidx2, align 4
+  %indvars.iv.next = add nuw nsw i32 %indvars.iv, 2
+  %cmp = icmp ult i32 %indvars.iv.next, %mul
+  br i1 %cmp, label %for.body, label %for.end.loopexit
+
+for.end.loopexit:                                 ; preds = %for.body
+  br label %for.end
+
+for.end:                                          ; preds = %for.end.loopexit, %entry
+  ret void
+; CHECK-LABEL: @test_pattern(
+; CHECK: call void @memset_pattern16
+; CHECK-NOT: store
+}

llvm/trunk/test/Transforms/PhaseOrdering/scev-custom-dl.ll

+; RUN: opt -O3 -S -analyze -scalar-evolution < %s | FileCheck %s
+
+target datalayout = "e-m:m-p:40:64:64:32-i32:32-i16:16-i8:8-n32"
+
+;
+; This file contains phase ordering tests for scalar evolution.
+; Test that the standard passes don't obfuscate the IR so scalar evolution can't
+; recognize expressions.
+
+; CHECK: test1
+; The loop body contains two increments by %div.
+; Make sure that 2*%div is recognizable, and not expressed as a bit mask of %d.
+; CHECK: -->  {%p,+,(8 * (%d /u 4))}
+define void @test1(i32 %d, i32* %p) nounwind uwtable ssp {
+entry:
+  %div = udiv i32 %d, 4
+  br label %for.cond
+
+for.cond:                                         ; preds = %for.inc, %entry
+  %p.addr.0 = phi i32* [ %p, %entry ], [ %add.ptr1, %for.inc ]
+  %i.0 = phi i32 [ 0, %entry ], [ %inc, %for.inc ]
+  %cmp = icmp ne i32 %i.0, 64
+  br i1 %cmp, label %for.body, label %for.end
+
+for.body:                                         ; preds = %for.cond
+  store i32 0, i32* %p.addr.0, align 4
+  %add.ptr = getelementptr inbounds i32, i32* %p.addr.0, i32 %div
+  store i32 1, i32* %add.ptr, align 4
+  %add.ptr1 = getelementptr inbounds i32, i32* %add.ptr, i32 %div
+  br label %for.inc
+
+for.inc:                                          ; preds = %for.body
+  %inc = add i32 %i.0, 1
+  br label %for.cond
+
+for.end:                                          ; preds = %for.cond
+  ret void
+}
+
+; CHECK: test1a
+; Same thing as test1, but it is even more tempting to fold 2 * (%d /u 2)
+; CHECK: -->  {%p,+,(8 * (%d /u 2))}
+define void @test1a(i32 %d, i32* %p) nounwind uwtable ssp {
+entry:
+  %div = udiv i32 %d, 2
+  br label %for.cond
+
+for.cond:                                         ; preds = %for.inc, %entry
+  %p.addr.0 = phi i32* [ %p, %entry ], [ %add.ptr1, %for.inc ]
+  %i.0 = phi i32 [ 0, %entry ], [ %inc, %for.inc ]
+  %cmp = icmp ne i32 %i.0, 64
+  br i1 %cmp, label %for.body, label %for.end
+
+for.body:                                         ; preds = %for.cond
+  store i32 0, i32* %p.addr.0, align 4
+  %add.ptr = getelementptr inbounds i32, i32* %p.addr.0, i32 %div
+  store i32 1, i32* %add.ptr, align 4
+  %add.ptr1 = getelementptr inbounds i32, i32* %add.ptr, i32 %div
+  br label %for.inc
+
+for.inc:                                          ; preds = %for.body
+  %inc = add i32 %i.0, 1
+  br label %for.cond
+
+for.end:                                          ; preds = %for.cond
+  ret void
+}

llvm/trunk/test/Transforms/SimplifyCFG/switch_create-custom-dl.ll

+; RUN: opt -S -simplifycfg < %s | FileCheck %s
+target datalayout="p:40:64:64:32"
+
+declare void @foo1()
+
+declare void @foo2()
+
+define void @test1(i32 %V) {
+        %C1 = icmp eq i32 %V, 4         ; <i1> [#uses=1]
+        %C2 = icmp eq i32 %V, 17                ; <i1> [#uses=1]
+        %CN = or i1 %C1, %C2            ; <i1> [#uses=1]
+        br i1 %CN, label %T, label %F
+T:              ; preds = %0
+        call void @foo1( )
+        ret void
+F:              ; preds = %0
+        call void @foo2( )
+        ret void
+; CHECK-LABEL: @test1(
+; CHECK:  switch i32 %V, label %F [
+; CHECK:    i32 17, label %T
+; CHECK:    i32 4, label %T
+; CHECK:  ]
+}
+
+define void @test1_ptr(i32* %V) {
+        %C1 = icmp eq i32* %V, inttoptr (i32 4 to i32*)
+        %C2 = icmp eq i32* %V, inttoptr (i32 17 to i32*)
+        %CN = or i1 %C1, %C2            ; <i1> [#uses=1]
+        br i1 %CN, label %T, label %F
+T:              ; preds = %0
+        call void @foo1( )
+        ret void
+F:              ; preds = %0
+        call void @foo2( )
+        ret void
+; CHECK-LABEL: @test1_ptr(
+; DL:  %magicptr = ptrtoint i32* %V to i32
+; DL:  switch i32 %magicptr, label %F [
+; DL:    i32 17, label %T
+; DL:    i32 4, label %T
+; DL:  ]
+}
+
+define void @test1_ptr_as1(i32 addrspace(1)* %V) {
+        %C1 = icmp eq i32 addrspace(1)* %V, inttoptr (i32 4 to i32 addrspace(1)*)
+        %C2 = icmp eq i32 addrspace(1)* %V, inttoptr (i32 17 to i32 addrspace(1)*)
+        %CN = or i1 %C1, %C2            ; <i1> [#uses=1]
+        br i1 %CN, label %T, label %F
+T:              ; preds = %0
+        call void @foo1( )
+        ret void
+F:              ; preds = %0
+        call void @foo2( )
+        ret void
+; CHECK-LABEL: @test1_ptr_as1(
+; DL:  %magicptr = ptrtoint i32 addrspace(1)* %V to i16
+; DL:  switch i16 %magicptr, label %F [
+; DL:    i16 17, label %T
+; DL:    i16 4, label %T
+; DL:  ]
+}
+
+define void @test2(i32 %V) {
+        %C1 = icmp ne i32 %V, 4         ; <i1> [#uses=1]
+        %C2 = icmp ne i32 %V, 17                ; <i1> [#uses=1]
+        %CN = and i1 %C1, %C2           ; <i1> [#uses=1]
+        br i1 %CN, label %T, label %F
+T:              ; preds = %0
+        call void @foo1( )
+        ret void
+F:              ; preds = %0
+        call void @foo2( )
+        ret void
+; CHECK-LABEL: @test2(
+; CHECK:  switch i32 %V, label %T [
+; CHECK:    i32 17, label %F
+; CHECK:    i32 4, label %F
+; CHECK:  ]
+}
+
+define void @test3(i32 %V) {
+        %C1 = icmp eq i32 %V, 4         ; <i1> [#uses=1]
+        br i1 %C1, label %T, label %N
+N:              ; preds = %0
+        %C2 = icmp eq i32 %V, 17                ; <i1> [#uses=1]
+        br i1 %C2, label %T, label %F
+T:              ; preds = %N, %0
+        call void @foo1( )
+        ret void
+F:              ; preds = %N
+        call void @foo2( )
+        ret void
+
+; CHECK-LABEL: @test3(
+; CHECK: switch i32 %V, label %F [
+; CHECK:     i32 4, label %T
+; CHECK:     i32 17, label %T
+; CHECK:   ]
+}
+
+
+
+define i32 @test4(i8 zeroext %c) nounwind ssp noredzone {
+entry:
+  %cmp = icmp eq i8 %c, 62
+  br i1 %cmp, label %lor.end, label %lor.lhs.false
+
+lor.lhs.false:                                    ; preds = %entry
+  %cmp4 = icmp eq i8 %c, 34
+  br i1 %cmp4, label %lor.end, label %lor.rhs
+
+lor.rhs:                                          ; preds = %lor.lhs.false
+  %cmp8 = icmp eq i8 %c, 92
+  br label %lor.end
+
+lor.end:                                          ; preds = %lor.rhs, %lor.lhs.false, %entry
+  %0 = phi i1 [ true, %lor.lhs.false ], [ true, %entry ], [ %cmp8, %lor.rhs ]
+  %lor.ext = zext i1 %0 to i32
+  ret i32 %lor.ext
+
+; CHECK-LABEL: @test4(
+; CHECK:  switch i8 %c, label %lor.rhs [
+; CHECK:    i8 62, label %lor.end
+; CHECK:    i8 34, label %lor.end
+; CHECK:    i8 92, label %lor.end
+; CHECK:  ]
+}
+
+define i32 @test5(i8 zeroext %c) nounwind ssp noredzone {
+entry:
+  switch i8 %c, label %lor.rhs [
+    i8 62, label %lor.end
+    i8 34, label %lor.end
+    i8 92, label %lor.end
+  ]
+
+lor.rhs:                                          ; preds = %entry
+  %V = icmp eq i8 %c, 92
+  br label %lor.end
+
+lor.end:                                          ; preds = %entry, %entry, %entry, %lor.rhs
+  %0 = phi i1 [ true, %entry ], [ %V, %lor.rhs ], [ true, %entry ], [ true, %entry ]
+  %lor.ext = zext i1 %0 to i32
+  ret i32 %lor.ext
+; CHECK-LABEL: @test5(
+; CHECK:  switch i8 %c, label %lor.rhs [
+; CHECK:    i8 62, label %lor.end
+; CHECK:    i8 34, label %lor.end
+; CHECK:    i8 92, label %lor.end
+; CHECK:  ]
+}
+
+
+define i1 @test6({ i32, i32 }* %I) {
+entry:
+        %tmp.1.i = getelementptr { i32, i32 }, { i32, i32 }* %I, i64 0, i32 1         ; <i32*> [#uses=1]
+        %tmp.2.i = load i32, i32* %tmp.1.i           ; <i32> [#uses=6]
+        %tmp.2 = icmp eq i32 %tmp.2.i, 14               ; <i1> [#uses=1]
+        br i1 %tmp.2, label %shortcirc_done.4, label %shortcirc_next.0
+shortcirc_next.0:               ; preds = %entry
+        %tmp.6 = icmp eq i32 %tmp.2.i, 15               ; <i1> [#uses=1]
+        br i1 %tmp.6, label %shortcirc_done.4, label %shortcirc_next.1
+shortcirc_next.1:               ; preds = %shortcirc_next.0
+        %tmp.11 = icmp eq i32 %tmp.2.i, 16              ; <i1> [#uses=1]
+        br i1 %tmp.11, label %shortcirc_done.4, label %shortcirc_next.2
+shortcirc_next.2:               ; preds = %shortcirc_next.1
+        %tmp.16 = icmp eq i32 %tmp.2.i, 17              ; <i1> [#uses=1]
+        br i1 %tmp.16, label %shortcirc_done.4, label %shortcirc_next.3
+shortcirc_next.3:               ; preds = %shortcirc_next.2
+        %tmp.21 = icmp eq i32 %tmp.2.i, 18              ; <i1> [#uses=1]
+        br i1 %tmp.21, label %shortcirc_done.4, label %shortcirc_next.4
+shortcirc_next.4:               ; preds = %shortcirc_next.3
+        %tmp.26 = icmp eq i32 %tmp.2.i, 19              ; <i1> [#uses=1]
+        br label %UnifiedReturnBlock
+shortcirc_done.4:               ; preds = %shortcirc_next.3, %shortcirc_next.2, %shortcirc_next.1, %shortcirc_next.0, %entry
+        br label %UnifiedReturnBlock
+UnifiedReturnBlock:             ; preds = %shortcirc_done.4, %shortcirc_next.4
+        %UnifiedRetVal = phi i1 [ %tmp.26, %shortcirc_next.4 ], [ true, %shortcirc_done.4 ]             ; <i1> [#uses=1]
+        ret i1 %UnifiedRetVal
+
+; CHECK-LABEL: @test6(
+; CHECK: %tmp.2.i.off = add i32 %tmp.2.i, -14
+; CHECK: %switch = icmp ult i32 %tmp.2.i.off, 6
+}
+
+define void @test7(i8 zeroext %c, i32 %x) nounwind ssp noredzone {
+entry:
+  %cmp = icmp ult i32 %x, 32
+  %cmp4 = icmp eq i8 %c, 97
+  %or.cond = or i1 %cmp, %cmp4
+  %cmp9 = icmp eq i8 %c, 99
+  %or.cond11 = or i1 %or.cond, %cmp9
+  br i1 %or.cond11, label %if.then, label %if.end
+
+if.then:                                          ; preds = %entry
+  tail call void @foo1() nounwind noredzone
+  ret void
+
+if.end:                                           ; preds = %entry
+  ret void
+
+; CHECK-LABEL: @test7(
+; CHECK:   %cmp = icmp ult i32 %x, 32
+; CHECK:   br i1 %cmp, label %if.then, label %switch.early.test
+; CHECK: switch.early.test:
+; CHECK:   switch i8 %c, label %if.end [
+; CHECK:     i8 99, label %if.then
+; CHECK:     i8 97, label %if.then
+; CHECK:   ]
+}
+
+define i32 @test8(i8 zeroext %c, i32 %x, i1 %C) nounwind ssp noredzone {
+entry:
+  br i1 %C, label %N, label %if.then
+N:
+  %cmp = icmp ult i32 %x, 32
+  %cmp4 = icmp eq i8 %c, 97
+  %or.cond = or i1 %cmp, %cmp4
+  %cmp9 = icmp eq i8 %c, 99
+  %or.cond11 = or i1 %or.cond, %cmp9
+  br i1 %or.cond11, label %if.then, label %if.end
+
+if.then:                                          ; preds = %entry
+  %A = phi i32 [0, %entry], [42, %N]
+  tail call void @foo1() nounwind noredzone
+  ret i32 %A
+
+if.end:                                           ; preds = %entry
+  ret i32 0
+
+; CHECK-LABEL: @test8(
+; CHECK: switch.early.test:
+; CHECK:   switch i8 %c, label %if.end [
+; CHECK:     i8 99, label %if.then
+; CHECK:     i8 97, label %if.then
+; CHECK:   ]
+; CHECK:   %A = phi i32 [ 0, %entry ], [ 42, %switch.early.test ], [ 42, %N ], [ 42, %switch.early.test ]
+}
+
+;; This is "Example 7" from http://blog.regehr.org/archives/320
+define i32 @test9(i8 zeroext %c) nounwind ssp noredzone {
+entry:
+  %cmp = icmp ult i8 %c, 33
+  br i1 %cmp, label %lor.end, label %lor.lhs.false
+
+lor.lhs.false:                                    ; preds = %entry
+  %cmp4 = icmp eq i8 %c, 46
+  br i1 %cmp4, label %lor.end, label %lor.lhs.false6
+
+lor.lhs.false6:                                   ; preds = %lor.lhs.false
+  %cmp9 = icmp eq i8 %c, 44
+  br i1 %cmp9, label %lor.end, label %lor.lhs.false11
+
+lor.lhs.false11:                                  ; preds = %lor.lhs.false6
+  %cmp14 = icmp eq i8 %c, 58
+  br i1 %cmp14, label %lor.end, label %lor.lhs.false16
+
+lor.lhs.false16:                                  ; preds = %lor.lhs.false11
+  %cmp19 = icmp eq i8 %c, 59
+  br i1 %cmp19, label %lor.end, label %lor.lhs.false21
+
+lor.lhs.false21:                                  ; preds = %lor.lhs.false16
+  %cmp24 = icmp eq i8 %c, 60
+  br i1 %cmp24, label %lor.end, label %lor.lhs.false26
+
+lor.lhs.false26:                                  ; preds = %lor.lhs.false21
+  %cmp29 = icmp eq i8 %c, 62
+  br i1 %cmp29, label %lor.end, label %lor.lhs.false31
+
+lor.lhs.false31:                                  ; preds = %lor.lhs.false26
+  %cmp34 = icmp eq i8 %c, 34
+  br i1 %cmp34, label %lor.end, label %lor.lhs.false36
+
+lor.lhs.false36:                                  ; preds = %lor.lhs.false31
+  %cmp39 = icmp eq i8 %c, 92
+  br i1 %cmp39, label %lor.end, label %lor.rhs
+
+lor.rhs:                                          ; preds = %lor.lhs.false36
+  %cmp43 = icmp eq i8 %c, 39
+  br label %lor.end
+
+lor.end:                                          ; preds = %lor.rhs, %lor.lhs.false36, %lor.lhs.false31, %lor.lhs.false26, %lor.lhs.false21, %lor.lhs.false16, %lor.lhs.false11, %lor.lhs.false6, %lor.lhs.false, %entry
+  %0 = phi i1 [ true, %lor.lhs.false36 ], [ true, %lor.lhs.false31 ], [ true, %lor.lhs.false26 ], [ true, %lor.lhs.false21 ], [ true, %lor.lhs.false16 ], [ true, %lor.lhs.false11 ], [ true, %lor.lhs.false6 ], [ true, %lor.lhs.false ], [ true, %entry ], [ %cmp43, %lor.rhs ]
+  %conv46 = zext i1 %0 to i32
+  ret i32 %conv46
+
+; CHECK-LABEL: @test9(
+; CHECK:   %cmp = icmp ult i8 %c, 33
+; CHECK:   br i1 %cmp, label %lor.end, label %switch.early.test
+
+; CHECK: switch.early.test:
+; CHECK:   switch i8 %c, label %lor.rhs [
+; CHECK:     i8 92, label %lor.end
+; CHECK:     i8 62, label %lor.end
+; CHECK:     i8 60, label %lor.end
+; CHECK:     i8 59, label %lor.end
+; CHECK:     i8 58, label %lor.end
+; CHECK:     i8 46, label %lor.end
+; CHECK:     i8 44, label %lor.end
+; CHECK:     i8 34, label %lor.end
+; CHECK:     i8 39, label %lor.end
+; CHECK:   ]
+}
+
+define i32 @test10(i32 %mode, i1 %Cond) {
+  %A = icmp ne i32 %mode, 0
+  %B = icmp ne i32 %mode, 51
+  %C = and i1 %A, %B
+  %D = and i1 %C, %Cond
+  br i1 %D, label %T, label %F
+T:
+  ret i32 123
+F:
+  ret i32 324
+
+; CHECK-LABEL: @test10(
+; CHECK:  br i1 %Cond, label %switch.early.test, label %F
+; CHECK:switch.early.test:
+; CHECK:  switch i32 %mode, label %T [
+; CHECK:    i32 51, label %F
+; CHECK:    i32 0, label %F
+; CHECK:  ]
+}
+
+; PR8780
+define i32 @test11(i32 %bar) nounwind {
+entry:
+  %cmp = icmp eq i32 %bar, 4
+  %cmp2 = icmp eq i32 %bar, 35
+  %or.cond = or i1 %cmp, %cmp2
+  %cmp5 = icmp eq i32 %bar, 53
+  %or.cond1 = or i1 %or.cond, %cmp5
+  %cmp8 = icmp eq i32 %bar, 24
+  %or.cond2 = or i1 %or.cond1, %cmp8
+  %cmp11 = icmp eq i32 %bar, 23
+  %or.cond3 = or i1 %or.cond2, %cmp11
+  %cmp14 = icmp eq i32 %bar, 55
+  %or.cond4 = or i1 %or.cond3, %cmp14
+  %cmp17 = icmp eq i32 %bar, 12
+  %or.cond5 = or i1 %or.cond4, %cmp17
+  %cmp20 = icmp eq i32 %bar, 35
+  %or.cond6 = or i1 %or.cond5, %cmp20
+  br i1 %or.cond6, label %if.then, label %if.end
+
+if.then:                                          ; preds = %entry
+  br label %return
+
+if.end:                                           ; preds = %entry
+  br label %return
+
+return:                                           ; preds = %if.end, %if.then
+  %retval.0 = phi i32 [ 1, %if.then ], [ 0, %if.end ]
+  ret i32 %retval.0
+
+; CHECK-LABEL: @test11(
+; CHECK: switch i32 %bar, label %if.end [
+; CHECK:   i32 55, label %return
+; CHECK:   i32 53, label %return
+; CHECK:   i32 35, label %return
+; CHECK:   i32 24, label %return
+; CHECK:   i32 23, label %return
+; CHECK:   i32 12, label %return
+; CHECK:   i32 4, label %return
+; CHECK: ]
+}
+
+define void @test12() nounwind {
+entry:
+  br label %bb49.us.us
+
+bb49.us.us:
+  %A = icmp eq i32 undef, undef
+  br i1 %A, label %bb55.us.us, label %malformed
+
+bb48.us.us:
+  %B = icmp ugt i32 undef, undef
+  br i1 %B, label %bb55.us.us, label %bb49.us.us
+
+bb55.us.us:
+  br label %bb48.us.us
+
+malformed:
+  ret void
+; CHECK-LABEL: @test12(
+
+}
+
+; test13 - handle switch formation with ult.
+define void @test13(i32 %x) nounwind ssp noredzone {
+entry:
+  %cmp = icmp ult i32 %x, 2
+  br i1 %cmp, label %if.then, label %lor.lhs.false3
+
+lor.lhs.false3:                                   ; preds = %lor.lhs.false
+  %cmp5 = icmp eq i32 %x, 3
+  br i1 %cmp5, label %if.then, label %lor.lhs.false6
+
+lor.lhs.false6:                                   ; preds = %lor.lhs.false3
+  %cmp8 = icmp eq i32 %x, 4
+  br i1 %cmp8, label %if.then, label %lor.lhs.false9
+
+lor.lhs.false9:                                   ; preds = %lor.lhs.false6
+  %cmp11 = icmp eq i32 %x, 6
+  br i1 %cmp11, label %if.then, label %if.end
+
+if.then:                                          ; preds = %lor.lhs.false9, %lor.lhs.false6, %lor.lhs.false3, %lor.lhs.false, %entry
+  call void @foo1() noredzone
+  br label %if.end
+
+if.end:                                           ; preds = %if.then, %lor.lhs.false9
+  ret void
+; CHECK-LABEL: @test13(
+; CHECK:  switch i32 %x, label %if.end [
+; CHECK:     i32 6, label %if.then
+; CHECK:     i32 4, label %if.then
+; CHECK:     i32 3, label %if.then
+; CHECK:     i32 1, label %if.then
+; CHECK:     i32 0, label %if.then
+; CHECK:   ]
+}
+
+; test14 - handle switch formation with ult.
+define void @test14(i32 %x) nounwind ssp noredzone {
+entry:
+  %cmp = icmp ugt i32 %x, 2
+  br i1 %cmp, label %lor.lhs.false3, label %if.then
+
+lor.lhs.false3:                                   ; preds = %lor.lhs.false
+  %cmp5 = icmp ne i32 %x, 3
+  br i1 %cmp5, label %lor.lhs.false6, label %if.then
+
+lor.lhs.false6:                                   ; preds = %lor.lhs.false3
+  %cmp8 = icmp ne i32 %x, 4
+  br i1 %cmp8, label %lor.lhs.false9, label %if.then
+
+lor.lhs.false9:                                   ; preds = %lor.lhs.false6
+  %cmp11 = icmp ne i32 %x, 6
+  br i1 %cmp11, label %if.end, label %if.then
+
+if.then:                                          ; preds = %lor.lhs.false9, %lor.lhs.false6, %lor.lhs.false3, %lor.lhs.false, %entry
+  call void @foo1() noredzone
+  br label %if.end
+
+if.end:                                           ; preds = %if.then, %lor.lhs.false9
+  ret void
+; CHECK-LABEL: @test14(
+; CHECK:  switch i32 %x, label %if.end [
+; CHECK:     i32 6, label %if.then
+; CHECK:     i32 4, label %if.then
+; CHECK:     i32 3, label %if.then
+; CHECK:     i32 1, label %if.then
+; CHECK:     i32 0, label %if.then
+; CHECK:   ]
+}
+
+; Don't crash on ginormous ranges.
+define void @test15(i128 %x) nounwind {
+  %cmp = icmp ugt i128 %x, 2
+  br i1 %cmp, label %if.end, label %lor.false
+
+lor.false:
+  %cmp2 = icmp ne i128 %x, 100000000000000000000
+  br i1 %cmp2, label %if.end, label %if.then
+
+if.then:
+  call void @foo1() noredzone
+  br label %if.end
+
+if.end:
+  ret void
+
+; CHECK-LABEL: @test15(
+; CHECK-NOT: switch
+; CHECK: ret void
+}
+
+; PR8675
+; rdar://5134905
+define zeroext i1 @test16(i32 %x) nounwind {
+entry:
+; CHECK-LABEL: @test16(
+; CHECK: %x.off = add i32 %x, -1
+; CHECK: %switch = icmp ult i32 %x.off, 3
+  %cmp.i = icmp eq i32 %x, 1
+  br i1 %cmp.i, label %lor.end, label %lor.lhs.false
+
+lor.lhs.false:
+  %cmp.i2 = icmp eq i32 %x, 2
+  br i1 %cmp.i2, label %lor.end, label %lor.rhs
+
+lor.rhs:
+  %cmp.i1 = icmp eq i32 %x, 3
+  br label %lor.end
+
+lor.end:
+  %0 = phi i1 [ true, %lor.lhs.false ], [ true, %entry ], [ %cmp.i1, %lor.rhs ]
+  ret i1 %0
+}
+
+; Check that we don't turn an icmp into a switch where it's not useful.
+define void @test17(i32 %x, i32 %y) {
+  %cmp = icmp ult i32 %x, 3
+  %switch = icmp ult i32 %y, 2
+  %or.cond775 = or i1 %cmp, %switch
+  br i1 %or.cond775, label %lor.lhs.false8, label %return
+
+lor.lhs.false8:
+  tail call void @foo1()
+  ret void
+
+return:
+  ret void
+
+; CHECK-LABEL: @test17(
+; CHECK-NOT: switch.early.test
+; CHECK-NOT: switch i32
+; CHECK: ret void
+}
+
+define void @test18(i32 %arg) {
+bb:
+  %tmp = and i32 %arg, -2
+  %tmp1 = icmp eq i32 %tmp, 8
+  %tmp2 = icmp eq i32 %arg, 10
+  %tmp3 = or i1 %tmp1, %tmp2
+  %tmp4 = icmp eq i32 %arg, 11
+  %tmp5 = or i1 %tmp3, %tmp4
+  %tmp6 = icmp eq i32 %arg, 12
+  %tmp7 = or i1 %tmp5, %tmp6
+  br i1 %tmp7, label %bb19, label %bb8
+
+bb8:                                              ; preds = %bb
+  %tmp9 = add i32 %arg, -13
+  %tmp10 = icmp ult i32 %tmp9, 2
+  %tmp11 = icmp eq i32 %arg, 16
+  %tmp12 = or i1 %tmp10, %tmp11
+  %tmp13 = icmp eq i32 %arg, 17
+  %tmp14 = or i1 %tmp12, %tmp13
+  %tmp15 = icmp eq i32 %arg, 18
+  %tmp16 = or i1 %tmp14, %tmp15
+  %tmp17 = icmp eq i32 %arg, 15
+  %tmp18 = or i1 %tmp16, %tmp17
+  br i1 %tmp18, label %bb19, label %bb20
+
+bb19:                                             ; preds = %bb8, %bb
+  tail call void @foo1()
+  br label %bb20
+
+bb20:                                             ; preds = %bb19, %bb8
+  ret void
+
+; CHECK-LABEL: @test18(
+; CHECK: %arg.off = add i32 %arg, -8
+; CHECK: icmp ult i32 %arg.off, 11
+}
+
+define void @PR26323(i1 %tobool23, i32 %tmp3) {
+entry:
+  %tobool5 = icmp ne i32 %tmp3, 0
+  %neg14 = and i32 %tmp3, -2
+  %cmp17 = icmp ne i32 %neg14, -1
+  %or.cond = and i1 %tobool5, %tobool23
+  %or.cond1 = and i1 %cmp17, %or.cond
+  br i1 %or.cond1, label %if.end29, label %if.then27
+
+if.then27:                                        ; preds = %entry
+  call void @foo1()
+  unreachable
+
+if.end29:                                         ; preds = %entry
+  ret void
+}
+
+; CHECK-LABEL: define void @PR26323(
+; CHECK:  %tobool5 = icmp ne i32 %tmp3, 0
+; CHECK:  %neg14 = and i32 %tmp3, -2
+; CHECK:  %cmp17 = icmp ne i32 %neg14, -1
+; CHECK:  %or.cond = and i1 %tobool5, %tobool23
+; CHECK:  %or.cond1 = and i1 %cmp17, %or.cond
+; CHECK:  br i1 %or.cond1, label %if.end29, label %if.then27
+
+; Form a switch when and'ing a negated power of two
+; CHECK-LABEL: define void @test19
+; CHECK: switch i32 %arg, label %else [
+; CHECK: i32 32, label %if
+; CHECK: i32 13, label %if
+; CHECK: i32 12, label %if
+define void @test19(i32 %arg) {
+  %and = and i32 %arg, -2
+  %cmp1 = icmp eq i32 %and, 12
+  %cmp2 = icmp eq i32 %arg, 32
+  %pred = or i1 %cmp1, %cmp2
+  br i1 %pred, label %if, label %else
+
+if:
+  call void @foo1()
+  ret void
+
+else:
+  ret void
+}
+
+; Since %cmp1 is always false, a switch is never formed
+; CHECK-LABEL: define void @test20
+; CHECK-NOT: switch
+; CHECK: ret void
+define void @test20(i32 %arg) {
+  %and = and i32 %arg, -2
+  %cmp1 = icmp eq i32 %and, 13
+  %cmp2 = icmp eq i32 %arg, 32
+  %pred = or i1 %cmp1, %cmp2
+  br i1 %pred, label %if, label %else
+
+if:
+  call void @foo1()
+  ret void
+
+else:
+  ret void
+}
+
+; Form a switch when or'ing a power of two
+; CHECK-LABEL: define void @test21
+; CHECK: i32 32, label %else
+; CHECK: i32 13, label %else
+; CHECK: i32 12, label %else
+define void @test21(i32 %arg) {
+  %and = or i32 %arg, 1
+  %cmp1 = icmp ne i32 %and, 13
+  %cmp2 = icmp ne i32 %arg, 32
+  %pred = and i1 %cmp1, %cmp2
+  br i1 %pred, label %if, label %else
+
+if:
+  call void @foo1()
+  ret void
+
+else:
+  ret void
+}
+
+; Since %cmp1 is always false, a switch is never formed
+; CHECK-LABEL: define void @test22
+; CHECK-NOT: switch
+; CHECK: ret void
+define void @test22(i32 %arg) {
+  %and = or i32 %arg, 1
+  %cmp1 = icmp ne i32 %and, 12
+  %cmp2 = icmp ne i32 %arg, 32
+  %pred = and i1 %cmp1, %cmp2
+  br i1 %pred, label %if, label %else
+
+if:
+  call void @foo1()
+  ret void
+
+else:
+  ret void
+}
 No newline at end of file