Re: [llvm-commits] [llvm-gcc-4.2] r46381 - in /llvm-gcc-4.2/trunk/gcc: llvm-abi.h llvm-internal.h llvm-types.cpp

[Duncan Sands]

Hi Evan,

> Can you get me a test case? I need to see what gcc is doing. Thanks,

Sure.  Zero sized fields in structs may not be very useful, but it is
still legal to use them, in particular you can take their address,
like this:

struct Z {};

struct Y {
int i;
struct Z z;
};

void *f(struct Y *y) {
  return &y->z;
}

This used to work before your patch.  For that matter, llvm-gcc-4.2
itself fails to build here (linux x86-32; crashes compiling libgcov.c) for
what looks to be the same reason.  Also, don't forget that Ada can have
fields of variable size in various places in a struct, and these can be
represented by zero-size LLVM fields (this kind of struct doesn't get
passed to functions by copy though).

Can you please explain why you made this change (a testcase would be
good).

Best wishes,

Duncan.
___
llvm-commits mailing list
llvm-commits@cs.uiuc.edu
http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits

Re: [llvm-commits] [llvm-gcc-4.2] r46381 - in /llvm-gcc-4.2/trunk/gcc: llvm-abi.h llvm-internal.h llvm-types.cpp

[Bill Wendling]

On Jan 27, 2008, at 1:03 AM, Duncan Sands wrote:

> Hi Evan,
>
>> Can you get me a test case? I need to see what gcc is doing. Thanks,
>
> Sure.  Zero sized fields in structs may not be very useful, but it is
> still legal to use them, in particular you can take their address,
> like this:
>
> struct Z {};
>
> struct Y {
> int i;
> struct Z z;
> };
>
> void *f(struct Y *y) {
>   return &y->z;
> }
>
> This used to work before your patch.  For that matter, llvm-gcc-4.2
> itself fails to build here (linux x86-32; crashes compiling  
> libgcov.c) for
> what looks to be the same reason.  Also, don't forget that Ada can  
> have
> fields of variable size in various places in a struct, and these  
> can be
> represented by zero-size LLVM fields (this kind of struct doesn't get
> passed to functions by copy though).
>
> Can you please explain why you made this change (a testcase would be
> good).
>
I'm running into this problem as well -- it crashing while compiling  
gcov.c on my PPC G4.

-bw
___
llvm-commits mailing list
llvm-commits@cs.uiuc.edu
http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits

Re: [llvm-commits] [llvm-gcc-4.2] r46381 - in /llvm-gcc-4.2/trunk/gcc: llvm-abi.h llvm-internal.h llvm-types.cpp

[Duncan Sands]

> I'm running into this problem as well -- it crashing while compiling  
> gcov.c on my PPC G4.

While r46381 seems the obvious culprit, to tell the truth I didn't check
whether reverting this patch fixes the problem - so I could be wrong.

Ciao,

Duncan.
___
llvm-commits mailing list
llvm-commits@cs.uiuc.edu
http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits

Re: [llvm-commits] [llvm-gcc-4.2] r46412 - in /llvm-gcc-4.2/trunk/gcc: config/i386/llvm-i386-target.h config/i386/llvm-i386.cpp llvm-abi.h llvm-convert.cpp

[Dale Johannesen]

On Jan 26, 2008, at 11:18 PM, Chris Lattner wrote:

> Author: lattner
> Date: Sun Jan 27 01:18:20 2008
> New Revision: 46412
>
> URL: http://llvm.org/viewvc/llvm-project?rev=46412&view=rev
> Log:
> Change x86-32 aggregate passing code to pass many aggregates "by  
> pieces"
> instead of passing them with byval.  This allows us to get  
> significantly
> better code without affecting the ABI (because x86-32 passes stuff  
> on the
> stack).  Most significantly, we pass _Complex as a two scalar fp  
> values
> now.  For example, before we compiled ccosf to:

Not everything is passed on the stack with -msseregparm, which doesn't  
work now but should.  In that mode a _Complex is not passed the same  
as two scalars.

> _ccosf:
>   pushl   %ebp
>   movl%esp, %ebp
>   pushl   %edi
>   pushl   %esi
>   subl$16, %esp
>   movss   12(%ebp), %xmm0
>   xorps   LCPI1_0, %xmm0
>   movss   8(%ebp), %xmm1
>   movss   %xmm0, -16(%ebp)
>   movss   %xmm1, -12(%ebp)
>   leal-16(%ebp), %esi
>   movl$2, %ecx
>   movl%esp, %edi
>   rep;movsl
>   call_ccoshf
>   addl$16, %esp
>   popl%esi
>   popl%edi
>   popl%ebp
>   ret
>
> now we get:
>
> _ccosf:
>   pushl   %ebp
>   movl%esp, %ebp
>   subl$8, %esp
>   movss   8(%ebp), %xmm0
>   movss   %xmm0, 4(%esp)
>   movss   12(%ebp), %xmm0
>   xorps   LCPI1_0, %xmm0
>   movss   %xmm0, (%esp)
>   call_ccoshf
>   addl$8, %esp
>   popl%ebp
>   ret
>
> Evan, please review this.  Thanks!
>
>
> Modified:
>llvm-gcc-4.2/trunk/gcc/config/i386/llvm-i386-target.h
>llvm-gcc-4.2/trunk/gcc/config/i386/llvm-i386.cpp
>llvm-gcc-4.2/trunk/gcc/llvm-abi.h
>llvm-gcc-4.2/trunk/gcc/llvm-convert.cpp
>
> Modified: llvm-gcc-4.2/trunk/gcc/config/i386/llvm-i386-target.h
> URL: 
> http://llvm.org/viewvc/llvm-project/llvm-gcc-4.2/trunk/gcc/config/i386/llvm-i386-target.h?rev=46412&r1=46411&r2=46412&view=diff
>
> = 
> = 
> = 
> = 
> = 
> = 
> = 
> = 
> ==
> --- llvm-gcc-4.2/trunk/gcc/config/i386/llvm-i386-target.h (original)
> +++ llvm-gcc-4.2/trunk/gcc/config/i386/llvm-i386-target.h Sun Jan 27  
> 01:18:20 2008
> @@ -62,20 +62,24 @@
> }   \
>   }
>
> -extern bool llvm_x86_should_pass_aggregate_in_memory(tree);
> +#ifdef LLVM_ABI_H
> +extern bool llvm_x86_should_pass_aggregate_in_memory(tree, const  
> Type *);
>
> -#define LLVM_SHOULD_PASS_AGGREGATE_USING_BYVAL_ATTR(X)  \
> -  llvm_x86_should_pass_aggregate_in_memory(X)
> +#define LLVM_SHOULD_PASS_AGGREGATE_USING_BYVAL_ATTR(X, TY)  \
> +  llvm_x86_should_pass_aggregate_in_memory(X, TY)
>
>
> -#ifdef LLVM_ABI_H
> extern bool
> -llvm_x86_64_should_pass_aggregate_in_mixed_regs(tree,
> +llvm_x86_64_should_pass_aggregate_in_mixed_regs(tree, const Type *Ty,
> +std::vector Type*>&);
> +extern bool
> +llvm_x86_32_should_pass_aggregate_in_mixed_regs(tree, const Type *Ty,
> std::vector Type*>&);
>
> -#define LLVM_SHOULD_PASS_AGGREGATE_IN_MIXED_REGS(T, E)  \
> -  (TARGET_64BIT &&  \
> -   llvm_x86_64_should_pass_aggregate_in_mixed_regs((T), (E)))
> +#define LLVM_SHOULD_PASS_AGGREGATE_IN_MIXED_REGS(T, TY,  
> E)\
> +   
> (TARGET_64BIT ? \
> +   llvm_x86_64_should_pass_aggregate_in_mixed_regs((T), (TY),  
> (E)) :  \
> +   llvm_x86_32_should_pass_aggregate_in_mixed_regs((T), (TY), (E)))
> #endif /* LLVM_ABI_H */
>
> /* LLVM LOCAL end (ENTIRE FILE!)  */
>
> Modified: llvm-gcc-4.2/trunk/gcc/config/i386/llvm-i386.cpp
> URL: 
> http://llvm.org/viewvc/llvm-project/llvm-gcc-4.2/trunk/gcc/config/i386/llvm-i386.cpp?rev=46412&r1=46411&r2=46412&view=diff
>
> = 
> = 
> = 
> = 
> = 
> = 
> = 
> = 
> ==
> --- llvm-gcc-4.2/trunk/gcc/config/i386/llvm-i386.cpp (original)
> +++ llvm-gcc-4.2/trunk/gcc/config/i386/llvm-i386.cpp Sun Jan 27  
> 01:18:20 2008
> @@ -691,7 +691,7 @@
>
> /* Target hook for llvm-abi.h. It returns true if an aggregate of the
>specified type should be passed in memory. */
> -bool llvm_x86_should_pass_aggregate_in_memory(tree TreeType) {
> +bool llvm_x86_should_pass_aggregate_in_memory(tree TreeType, const  
> Type *Ty) {
>   enum machine_mode Mode = ix86_getNaturalModeForType(TreeType);
>   HOST_WIDE_INT Bytes =
> (Mode == BLKmode) ? int_size_in_bytes(TreeType) : (int)  
> GET_MODE_SIZE(Mode);
> @@ -706,8 +706,10 @@
> if (llvm_x86_is_all_integer_types(Ty))
>   return false;
>   }
> -  if (!TARGET_64BIT)
> -return true;
> +  if (!TARGET_64BIT) {
> +std::vector Elts;
> +return ! 
> llvm_x86_32_should_pass_aggregate_in_mixed_regs(TreeT

[llvm-commits] [llvm-gcc-4.2] r46415 - /llvm-gcc-4.2/trunk/gcc/llvm-types.cpp

[Duncan Sands]

Author: baldrick
Date: Sun Jan 27 12:11:13 2008
New Revision: 46415

URL: http://llvm.org/viewvc/llvm-project?rev=46415&view=rev
Log:
Do not mark a function readnone or readonly if it has
a byval parameter.

Modified:
llvm-gcc-4.2/trunk/gcc/llvm-types.cpp

Modified: llvm-gcc-4.2/trunk/gcc/llvm-types.cpp
URL: 
http://llvm.org/viewvc/llvm-project/llvm-gcc-4.2/trunk/gcc/llvm-types.cpp?rev=46415&r1=46414&r2=46415&view=diff

==
--- llvm-gcc-4.2/trunk/gcc/llvm-types.cpp (original)
+++ llvm-gcc-4.2/trunk/gcc/llvm-types.cpp Sun Jan 27 12:11:13 2008
@@ -1219,14 +1219,19 @@
   DeclArgs = TREE_CHAIN(DeclArgs);
   }
   
-  // If we see a byval argument and if the function is 'readnone' we have to
-  // demote the function to being 'readonly' instead.  Not doing so would allow
-  // optimizers to delete stores into the argument that is passed into the
-  // function.
-  if (HasByVal && Attrs[0].index == 0 &&
-  (Attrs[0].attrs & ParamAttr::ReadNone)) {
-Attrs[0].attrs &= ~ParamAttr::ReadNone;
-Attrs[0].attrs |= ParamAttr::ReadOnly;
+  // If there is a byval argument then it is not safe to mark the function
+  // 'readnone' or 'readonly': gcc permits a 'const' or 'pure' function to
+  // write to struct arguments passed by value, but in LLVM this becomes a
+  // write through the byval pointer argument, which LLVM does not allow for
+  // readonly/readnone functions.
+  if (HasByVal && Attrs[0].index == 0) {
+uint16_t &RAttrs = Attrs[0].attrs;
+RAttrs &= ~(ParamAttr::ReadNone | ParamAttr::ReadOnly);
+if (RAttrs == ParamAttr::None) {
+  for (unsigned i = 1, e = Attrs.size(); i < e ; ++i)
+Attrs[i-1] = Attrs[i];
+  Attrs.pop_back();
+}
   }
 
   // If the argument list ends with a void type node, it isn't vararg.


___
llvm-commits mailing list
llvm-commits@cs.uiuc.edu
http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits

[llvm-commits] [llvm] r46416 - in /llvm/trunk: lib/Transforms/Utils/InlineFunction.cpp test/Transforms/Inline/byval2.ll

[Duncan Sands]

Author: baldrick
Date: Sun Jan 27 12:12:58 2008
New Revision: 46416

URL: http://llvm.org/viewvc/llvm-project?rev=46416&view=rev
Log:
Revert r46393: readonly/readnone functions are no
longer allowed to write through byval arguments.

Modified:
llvm/trunk/lib/Transforms/Utils/InlineFunction.cpp
llvm/trunk/test/Transforms/Inline/byval2.ll

Modified: llvm/trunk/lib/Transforms/Utils/InlineFunction.cpp
URL: 
http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Transforms/Utils/InlineFunction.cpp?rev=46416&r1=46415&r2=46416&view=diff

==
--- llvm/trunk/lib/Transforms/Utils/InlineFunction.cpp (original)
+++ llvm/trunk/lib/Transforms/Utils/InlineFunction.cpp Sun Jan 27 12:12:58 2008
@@ -240,15 +240,12 @@
  E = CalledFunc->arg_end(); I != E; ++I, ++AI, ++ArgNo) {
   Value *ActualArg = *AI;
   
-  // When byval arguments are inlined, we need to make the copy implied
-  // by them explicit.  It is tempting to think that this is not needed if
-  // the callee is readonly, because the callee doesn't modify the struct.
-  // However this would be wrong: readonly means that any writes the callee
-  // performs are not visible to the caller.  But writes by the callee to
-  // an argument passed byval are by definition not visible to the caller!
-  // Since we allow this kind of readonly function, there needs to be an
-  // explicit copy in order to keep the writes invisible after inlining.
-  if (CalledFunc->paramHasAttr(ArgNo+1, ParamAttr::ByVal)) {
+  // When byval arguments actually inlined, we need to make the copy 
implied
+  // by them explicit.  However, we don't do this if the callee is readonly
+  // or readnone, because the copy would be unneeded: the callee doesn't
+  // modify the struct.
+  if (CalledFunc->paramHasAttr(ArgNo+1, ParamAttr::ByVal) &&
+  !CalledFunc->onlyReadsMemory()) {
 const Type *AggTy = cast(I->getType())->getElementType();
 const Type *VoidPtrTy = PointerType::getUnqual(Type::Int8Ty);
 

Modified: llvm/trunk/test/Transforms/Inline/byval2.ll
URL: 
http://llvm.org/viewvc/llvm-project/llvm/trunk/test/Transforms/Inline/byval2.ll?rev=46416&r1=46415&r2=46416&view=diff

==
--- llvm/trunk/test/Transforms/Inline/byval2.ll (original)
+++ llvm/trunk/test/Transforms/Inline/byval2.ll Sun Jan 27 12:12:58 2008
@@ -1,7 +1,7 @@
-; RUN: llvm-as < %s | opt -inline | llvm-dis | grep {llvm.memcpy}
+; RUN: llvm-as < %s | opt -inline | llvm-dis | not grep {llvm.memcpy}
 
-; Inlining a byval struct should cause an explicit copy 
-; into an alloca even if the function is readonly
+; Inlining a byval struct should NOT cause an explicit copy 
+; into an alloca if the function is readonly
 
%struct.ss = type { i32, i64 }
 @.str = internal constant [10 x i8] c"%d, %lld\0A\00"  ; <[10 x i8]*> 
[#uses=1]


___
llvm-commits mailing list
llvm-commits@cs.uiuc.edu
http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits

[llvm-commits] [llvm] r46414 - in /llvm/trunk: lib/CodeGen/SelectionDAG/DAGCombiner.cpp test/CodeGen/X86/fp-in-intregs.ll

[Chris Lattner]

Author: lattner
Date: Sun Jan 27 11:42:27 2008
New Revision: 46414

URL: http://llvm.org/viewvc/llvm-project?rev=46414&view=rev
Log:
Implement some dag combines that allow doing fneg/fabs/fcopysign in integer
registers if used by a bitconvert or using a bitconvert.  This allows us to
avoid constant pool loads and use cheaper integer instructions when the
values come from or end up in integer regs anyway.  For example, we now 
compile CodeGen/X86/fp-in-intregs.ll to:

_test1:
movl$2147483648, %eax
xorl4(%esp), %eax
ret
_test2:
movl$1065353216, %eax
orl 4(%esp), %eax
andl$3212836864, %eax
ret

Instead of:
_test1:
movss   4(%esp), %xmm0
xorps   LCPI2_0, %xmm0
movd%xmm0, %eax
ret
_test2:
movss   4(%esp), %xmm0
andps   LCPI3_0, %xmm0
movss   LCPI3_1, %xmm1
andps   LCPI3_2, %xmm1
orps%xmm0, %xmm1
movd%xmm1, %eax
ret

bitconverts can happen due to various calling conventions that require
fp values to passed in integer regs in some cases, e.g. when returning
a complex.


Added:
llvm/trunk/test/CodeGen/X86/fp-in-intregs.ll
Modified:
llvm/trunk/lib/CodeGen/SelectionDAG/DAGCombiner.cpp

Modified: llvm/trunk/lib/CodeGen/SelectionDAG/DAGCombiner.cpp
URL: 
http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/CodeGen/SelectionDAG/DAGCombiner.cpp?rev=46414&r1=46413&r2=46414&view=diff

==
--- llvm/trunk/lib/CodeGen/SelectionDAG/DAGCombiner.cpp (original)
+++ llvm/trunk/lib/CodeGen/SelectionDAG/DAGCombiner.cpp Sun Jan 27 11:42:27 2008
@@ -1190,7 +1190,7 @@
 return DAG.getNode(ISD::SUB, VT, DAG.getConstant(0, VT), N0);
   // If we know the sign bits of both operands are zero, strength reduce to a
   // udiv instead.  Handles (X&15) /s 4 -> X&15 >> 2
-  uint64_t SignBit = 1ULL << (MVT::getSizeInBits(VT)-1);
+  uint64_t SignBit = MVT::getIntVTSignBit(VT);
   if (DAG.MaskedValueIsZero(N1, SignBit) &&
   DAG.MaskedValueIsZero(N0, SignBit))
 return DAG.getNode(ISD::UDIV, N1.getValueType(), N0, N1);
@@ -1306,7 +1306,7 @@
 return DAG.getNode(ISD::SREM, VT, N0, N1);
   // If we know the sign bits of both operands are zero, strength reduce to a
   // urem instead.  Handles (X & 0x0FFF) %s 16 -> X&15
-  uint64_t SignBit = 1ULL << (MVT::getSizeInBits(VT)-1);
+  uint64_t SignBit = MVT::getIntVTSignBit(VT);
   if (DAG.MaskedValueIsZero(N1, SignBit) &&
   DAG.MaskedValueIsZero(N0, SignBit))
 return DAG.getNode(ISD::UREM, VT, N0, N1);
@@ -3276,6 +3276,57 @@
 }
   }
   
+  // Fold bitconvert(fneg(x)) -> xor(bitconvert(x), signbit)
+  // Fold bitconvert(fabs(x)) -> and(bitconvert(x), ~signbit)
+  // This often reduces constant pool loads.
+  if ((N0.getOpcode() == ISD::FNEG || N0.getOpcode() == ISD::FABS) &&
+  N0.Val->hasOneUse() && MVT::isInteger(VT) && !MVT::isVector(VT)) {
+SDOperand NewConv = DAG.getNode(ISD::BIT_CONVERT, VT, N0.getOperand(0));
+AddToWorkList(NewConv.Val);
+
+uint64_t SignBit = MVT::getIntVTSignBit(VT);
+if (N0.getOpcode() == ISD::FNEG)
+  return DAG.getNode(ISD::XOR, VT, NewConv, DAG.getConstant(SignBit, VT));
+assert(N0.getOpcode() == ISD::FABS);
+return DAG.getNode(ISD::AND, VT, NewConv, DAG.getConstant(~SignBit, VT));
+  }
+  
+  // Fold bitconvert(fcopysign(cst, x)) -> bitconvert(x)&sign | cst&~sign'
+  // Note that we don't handle copysign(x,cst) because this can always be 
folded
+  // to an fneg or fabs.
+  if (N0.getOpcode() == ISD::FCOPYSIGN && N0.Val->hasOneUse() &&
+  isa(N0.getOperand(0))) {
+unsigned OrigXWidth = MVT::getSizeInBits(N0.getOperand(1).getValueType());
+SDOperand X = DAG.getNode(ISD::BIT_CONVERT, 
MVT::getIntegerType(OrigXWidth),
+  N0.getOperand(1));
+AddToWorkList(X.Val);
+
+// If X has a different width than the result/lhs, sext it or truncate it.
+unsigned VTWidth = MVT::getSizeInBits(VT);
+if (OrigXWidth < VTWidth) {
+  X = DAG.getNode(ISD::SIGN_EXTEND, VT, X);
+  AddToWorkList(X.Val);
+} else if (OrigXWidth > VTWidth) {
+  // To get the sign bit in the right place, we have to shift it right
+  // before truncating.
+  X = DAG.getNode(ISD::SRL, X.getValueType(), X, 
+  DAG.getConstant(OrigXWidth-VTWidth, X.getValueType()));
+  AddToWorkList(X.Val);
+  X = DAG.getNode(ISD::TRUNCATE, VT, X);
+  AddToWorkList(X.Val);
+}
+
+uint64_t SignBit = MVT::getIntVTSignBit(VT);
+X = DAG.getNode(ISD::AND, VT, X, DAG.getConstant(SignBit, VT));
+AddToWorkList(X.Val);
+
+SDOperand Cst = DAG.getNode(ISD::BIT_CONVERT, VT, N0.getOperand(0));
+Cst = DAG.getNode(ISD::AND, VT, Cst, DAG.getConstant(~SignBit, VT));
+AddToWorkList(Cst.Val);
+
+return DAG.getNode(ISD::OR, VT, X, Cst);
+  }
+  
   return SDOperand();
 }
 
@@ -3732,6 +3783,19 @@
   if (

[llvm-commits] [llvm] r46417 - in /llvm/trunk: lib/Transforms/Scalar/LoopUnroll.cpp test/Transforms/LoopUnroll/2007-11-05-Crash.ll

[Nick Lewycky]

Author: nicholas
Date: Sun Jan 27 12:35:00 2008
New Revision: 46417

URL: http://llvm.org/viewvc/llvm-project?rev=46417&view=rev
Log:
Be more careful modifying the use_list while also iterating through it.

Modified:
llvm/trunk/lib/Transforms/Scalar/LoopUnroll.cpp
llvm/trunk/test/Transforms/LoopUnroll/2007-11-05-Crash.ll

Modified: llvm/trunk/lib/Transforms/Scalar/LoopUnroll.cpp
URL: 
http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Transforms/Scalar/LoopUnroll.cpp?rev=46417&r1=46416&r2=46417&view=diff

==
--- llvm/trunk/lib/Transforms/Scalar/LoopUnroll.cpp (original)
+++ llvm/trunk/lib/Transforms/Scalar/LoopUnroll.cpp Sun Jan 27 12:35:00 2008
@@ -365,8 +365,9 @@
   // be updated specially after unrolling all the way.
   if (*BB != LatchBlock)
 for (Value::use_iterator UI = (*BB)->use_begin(), UE = 
(*BB)->use_end();
- UI != UE; ++UI) {
+ UI != UE;) {
   Instruction *UseInst = cast(*UI);
+  ++UI;
   if (isa(UseInst) && !L->contains(UseInst->getParent())) {
 PHINode *phi = cast(UseInst);
 Value *Incoming = phi->getIncomingValueForBlock(*BB);

Modified: llvm/trunk/test/Transforms/LoopUnroll/2007-11-05-Crash.ll
URL: 
http://llvm.org/viewvc/llvm-project/llvm/trunk/test/Transforms/LoopUnroll/2007-11-05-Crash.ll?rev=46417&r1=46416&r2=46417&view=diff

==
--- llvm/trunk/test/Transforms/LoopUnroll/2007-11-05-Crash.ll (original)
+++ llvm/trunk/test/Transforms/LoopUnroll/2007-11-05-Crash.ll Sun Jan 27 
12:35:00 2008
@@ -1,5 +1,7 @@
 ; RUN: llvm-as < %s | opt -disable-output -loop-unroll
 ; PR1770
+; PR1947
+
%struct.cl_engine = type { i32, i16, i32, i8**, i8**, i8*, i8*, i8*, 
i8*, i8*, i8*, i8* }
%struct.cl_limits = type { i32, i32, i32, i32, i16, i64 }
%struct.cli_ac_alt = type { i8, i8*, i16, i16, %struct.cli_ac_alt* }


___
llvm-commits mailing list
llvm-commits@cs.uiuc.edu
http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits

Re: [llvm-commits] [llvm-gcc-4.2] r46415 - /llvm-gcc-4.2/trunk/gcc/llvm-types.cpp

[Chris Lattner]

> URL: http://llvm.org/viewvc/llvm-project?rev=46415&view=rev
> Log:
> Do not mark a function readnone or readonly if it has
> a byval parameter.

Thanks Duncan:

> +uint16_t &RAttrs = Attrs[0].attrs;
> +RAttrs &= ~(ParamAttr::ReadNone | ParamAttr::ReadOnly);
> +if (RAttrs == ParamAttr::None) {
> +  for (unsigned i = 1, e = Attrs.size(); i < e ; ++i)
> +Attrs[i-1] = Attrs[i];
> +  Attrs.pop_back();
> +}

How about:

if (RAttrs == ParamAttr::None)
   Attrs.erase(Attrs.begin());

?

-Chris

___
llvm-commits mailing list
llvm-commits@cs.uiuc.edu
http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits

Re: [llvm-commits] [llvm-gcc-4.2] r46415 - /llvm-gcc-4.2/trunk/gcc/llvm-types.cpp

[Duncan Sands]

> How about:
> 
> if (RAttrs == ParamAttr::None)
>Attrs.erase(Attrs.begin());

Now you see that I really don't know anything about std::vector :)
I will change it.

Ciao,

Duncan.
___
llvm-commits mailing list
llvm-commits@cs.uiuc.edu
http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits

[llvm-commits] [llvm-gcc-4.2] r46418 - /llvm-gcc-4.2/trunk/gcc/llvm-types.cpp

[Duncan Sands]

Author: baldrick
Date: Sun Jan 27 12:58:34 2008
New Revision: 46418

URL: http://llvm.org/viewvc/llvm-project?rev=46418&view=rev
Log:
Baldrick discovers "erase".

Modified:
llvm-gcc-4.2/trunk/gcc/llvm-types.cpp

Modified: llvm-gcc-4.2/trunk/gcc/llvm-types.cpp
URL: 
http://llvm.org/viewvc/llvm-project/llvm-gcc-4.2/trunk/gcc/llvm-types.cpp?rev=46418&r1=46417&r2=46418&view=diff

==
--- llvm-gcc-4.2/trunk/gcc/llvm-types.cpp (original)
+++ llvm-gcc-4.2/trunk/gcc/llvm-types.cpp Sun Jan 27 12:58:34 2008
@@ -1227,11 +1227,8 @@
   if (HasByVal && Attrs[0].index == 0) {
 uint16_t &RAttrs = Attrs[0].attrs;
 RAttrs &= ~(ParamAttr::ReadNone | ParamAttr::ReadOnly);
-if (RAttrs == ParamAttr::None) {
-  for (unsigned i = 1, e = Attrs.size(); i < e ; ++i)
-Attrs[i-1] = Attrs[i];
-  Attrs.pop_back();
-}
+if (RAttrs == ParamAttr::None)
+  Attrs.erase(Attrs.begin());
   }
 
   // If the argument list ends with a void type node, it isn't vararg.


___
llvm-commits mailing list
llvm-commits@cs.uiuc.edu
http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits

[llvm-commits] [llvm] r46419 - /llvm/trunk/include/llvm/CodeGen/BreakCriticalMachineEdge.h

[Owen Anderson]

Author: resistor
Date: Sun Jan 27 13:51:03 2008
New Revision: 46419

URL: http://llvm.org/viewvc/llvm-project?rev=46419&view=rev
Log:
Fixes for BreakCriticalMachineCodeEdge by Fernando.

Modified:
llvm/trunk/include/llvm/CodeGen/BreakCriticalMachineEdge.h

Modified: llvm/trunk/include/llvm/CodeGen/BreakCriticalMachineEdge.h
URL: 
http://llvm.org/viewvc/llvm-project/llvm/trunk/include/llvm/CodeGen/BreakCriticalMachineEdge.h?rev=46419&r1=46418&r2=46419&view=diff

==
--- llvm/trunk/include/llvm/CodeGen/BreakCriticalMachineEdge.h (original)
+++ llvm/trunk/include/llvm/CodeGen/BreakCriticalMachineEdge.h Sun Jan 27 
13:51:03 2008
@@ -33,13 +33,14 @@
 
   // insert the new block into the machine function.
   src->getParent()->getBasicBlockList().insert(src->getParent()->end(),
-   crit_mbb);
+ crit_mbb);
 
   // insert a unconditional branch linking the new block to dst
   const TargetMachine& TM = src->getParent()->getTarget();
   const TargetInstrInfo* TII = TM.getInstrInfo();
   std::vector emptyConditions;
-  TII->InsertBranch(*crit_mbb, dst, (MachineBasicBlock*)0, emptyConditions);
+  TII->InsertBranch(*crit_mbb, dst, (MachineBasicBlock*)0, 
+emptyConditions);
 
   // modify every branch in src that points to dst to point to the new
   // machine basic block instead:
@@ -48,18 +49,18 @@
   while (mii != src->begin()) {
 mii--;
 // if there are no more branches, finish the loop
-if (!TII->isTerminatorInstr(mii->getOpcode())) {
+if (!mii->getDesc().isTerminator()) {
   break;
 }
-
+
 // Scan the operands of this branch, replacing any uses of dst with
 // crit_mbb.
 for (unsigned i = 0, e = mii->getNumOperands(); i != e; ++i) {
   MachineOperand & mo = mii->getOperand(i);
   if (mo.isMachineBasicBlock() &&
-  mo.getMachineBasicBlock() == dst) {
+  mo.getMBB() == dst) {
 found_branch = true;
-mo.setMachineBasicBlock(crit_mbb);
+mo.setMBB(crit_mbb);
   }
 }
   }
@@ -68,7 +69,8 @@
   // I am inserting too many gotos, but I am trusting that the asm printer
   // will optimize the unnecessary gotos.
   if(!found_branch) {
-TII->InsertBranch(*src, crit_mbb, (MachineBasicBlock*)0, emptyConditions);
+TII->InsertBranch(*src, crit_mbb, (MachineBasicBlock*)0, 
+  emptyConditions);
   }
 
   /// Change all the phi functions in dst, so that the incoming block be
@@ -77,29 +79,28 @@
 /// the first instructions are always phi functions.
 if(mii->getOpcode() != TargetInstrInfo::PHI)
   break;
-
+
 // Find the operands corresponding to the source block
 std::vector toRemove;
 unsigned reg = 0;
 for (unsigned u = 0; u != mii->getNumOperands(); ++u)
   if (mii->getOperand(u).isMachineBasicBlock() &&
-  mii->getOperand(u).getMachineBasicBlock() == src) {
+  mii->getOperand(u).getMBB() == src) {
 reg = mii->getOperand(u-1).getReg();
 toRemove.push_back(u-1);
   }
-
 // Remove all uses of this MBB
 for (std::vector::reverse_iterator I = toRemove.rbegin(),
  E = toRemove.rend(); I != E; ++I) {
   mii->RemoveOperand(*I+1);
   mii->RemoveOperand(*I);
 }
-
+
 // Add a single use corresponding to the new MBB
-mii->addRegOperand(reg, false);
-mii->addMachineBasicBlockOperand(crit_mbb);
+mii->addOperand(MachineOperand::CreateReg(reg, false));
+mii->addOperand(MachineOperand::CreateMBB(crit_mbb));
   }
-  
+
   return crit_mbb;
 }
 


___
llvm-commits mailing list
llvm-commits@cs.uiuc.edu
http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits

[llvm-commits] [llvm] r46420 - /llvm/trunk/lib/CodeGen/RegAllocLocal.cpp

[Owen Anderson]

Author: resistor
Date: Sun Jan 27 16:00:00 2008
New Revision: 46420

URL: http://llvm.org/viewvc/llvm-project?rev=46420&view=rev
Log:
Reg alloc doesn't really need LiveVariables.

Modified:
llvm/trunk/lib/CodeGen/RegAllocLocal.cpp

Modified: llvm/trunk/lib/CodeGen/RegAllocLocal.cpp
URL: 
http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/CodeGen/RegAllocLocal.cpp?rev=46420&r1=46419&r2=46420&view=diff

==
--- llvm/trunk/lib/CodeGen/RegAllocLocal.cpp (original)
+++ llvm/trunk/lib/CodeGen/RegAllocLocal.cpp Sun Jan 27 16:00:00 2008
@@ -19,7 +19,6 @@
 #include "llvm/CodeGen/MachineInstr.h"
 #include "llvm/CodeGen/MachineFrameInfo.h"
 #include "llvm/CodeGen/MachineRegisterInfo.h"
-#include "llvm/CodeGen/LiveVariables.h"
 #include "llvm/CodeGen/RegAllocRegistry.h"
 #include "llvm/Target/TargetInstrInfo.h"
 #include "llvm/Target/TargetMachine.h"
@@ -51,7 +50,6 @@
 MachineFunction *MF;
 const MRegisterInfo *MRI;
 const TargetInstrInfo *TII;
-LiveVariables *LV;
 
 // StackSlotForVirtReg - Maps virtual regs to the frame index where these
 // values are spilled.
@@ -148,7 +146,6 @@
 }
 
 virtual void getAnalysisUsage(AnalysisUsage &AU) const {
-  AU.addRequired();
   AU.addRequiredID(PHIEliminationID);
   AU.addRequiredID(TwoAddressInstructionPassID);
   MachineFunctionPass::getAnalysisUsage(AU);
@@ -497,9 +494,8 @@
 Ops.push_back(OpNum);
 if (MachineInstr* FMI = TII->foldMemoryOperand(MI, Ops, FrameIndex)) {
   ++NumFolded;
-  // Since we changed the address of MI, make sure to update live variables
-  // to know that the new instruction has the properties of the old one.
-  LV->instructionChanged(MI, FMI);
+  // Update kill/dead flags.
+  FMI->copyKillDeadInfo(MI);
   return MBB.insert(MBB.erase(MI), FMI);
 }
 
@@ -778,8 +774,6 @@
 // Finally, if this is a noop copy instruction, zap it.
 unsigned SrcReg, DstReg;
 if (TII.isMoveInstr(*MI, SrcReg, DstReg) && SrcReg == DstReg) {
-  LV->removeVirtualRegistersKilled(MI);
-  LV->removeVirtualRegistersDead(MI);
   MBB.erase(MI);
 }
   }
@@ -821,7 +815,6 @@
   TM = &Fn.getTarget();
   MRI = TM->getRegisterInfo();
   TII = TM->getInstrInfo();
-  LV = &getAnalysis();
 
   PhysRegsUsed.assign(MRI->getNumRegs(), -1);
   


___
llvm-commits mailing list
llvm-commits@cs.uiuc.edu
http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits

Re: [llvm-commits] [llvm-gcc-4.2] r46381 - in /llvm-gcc-4.2/trunk/gcc: llvm-abi.h llvm-internal.h llvm-types.cpp

[Evan Cheng]

Sorry I have been away from my computer for the last two days. I'll  
look at this later tonight.

Evan

On Jan 27, 2008, at 10:02 AM, Duncan Sands <[EMAIL PROTECTED]> wrote:

>> I'm running into this problem as well -- it crashing while compiling
>> gcov.c on my PPC G4.
>
> While r46381 seems the obvious culprit, to tell the truth I didn't  
> check
> whether reverting this patch fixes the problem - so I could be wrong.
>
> Ciao,
>
> Duncan.
> ___
> llvm-commits mailing list
> llvm-commits@cs.uiuc.edu
> http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits
___
llvm-commits mailing list
llvm-commits@cs.uiuc.edu
http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits

Re: [llvm-commits] [llvm-gcc-4.2] r46412 - in /llvm-gcc-4.2/trunk/gcc: config/i386/llvm-i386-target.h config/i386/llvm-i386.cpp llvm-abi.h llvm-convert.cpp

[Chris Lattner]

On Jan 27, 2008, at 10:09 AM, Dale Johannesen wrote:
> On Jan 26, 2008, at 11:18 PM, Chris Lattner wrote:
>> Author: lattner
>> Date: Sun Jan 27 01:18:20 2008
>> New Revision: 46412
>>
>> URL: http://llvm.org/viewvc/llvm-project?rev=46412&view=rev
>> Log:
>> Change x86-32 aggregate passing code to pass many aggregates "by
>> pieces"
>> instead of passing them with byval.  This allows us to get
>> significantly
>> better code without affecting the ABI (because x86-32 passes stuff
>> on the
>> stack).  Most significantly, we pass _Complex as a two scalar fp
>> values
>> now.  For example, before we compiled ccosf to:
>
> Not everything is passed on the stack with -msseregparm, which doesn't
> work now but should.  In that mode a _Complex is not passed the same
> as two scalars.

Right, in that case, the scalars need to be marked with the inreg  
attribute though.  In this case, they never will be marked with that,  
so I don't think this could cause a problem.  The corollary is when  
you use -mregparm, which marks integer registers as inreg.

-Chris
___
llvm-commits mailing list
llvm-commits@cs.uiuc.edu
http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits

[llvm-commits] [llvm] r46421 - /llvm/trunk/tools/llvm-ld/llvm-ld.cpp

[Chris Lattner]

Author: lattner
Date: Sun Jan 27 16:58:59 2008
New Revision: 46421

URL: http://llvm.org/viewvc/llvm-project?rev=46421&view=rev
Log:
Add support for frameworks.  Patch by Shantonu Sen!

Modified:
llvm/trunk/tools/llvm-ld/llvm-ld.cpp

Modified: llvm/trunk/tools/llvm-ld/llvm-ld.cpp
URL: 
http://llvm.org/viewvc/llvm-project/llvm/trunk/tools/llvm-ld/llvm-ld.cpp?rev=46421&r1=46420&r2=46421&view=diff

==
--- llvm/trunk/tools/llvm-ld/llvm-ld.cpp (original)
+++ llvm/trunk/tools/llvm-ld/llvm-ld.cpp Sun Jan 27 16:58:59 2008
@@ -55,10 +55,18 @@
   cl::desc("Specify a library search path"),
   cl::value_desc("directory"));
 
+static cl::list FrameworkPaths("F", cl::Prefix,
+  cl::desc("Specify a framework search path"),
+  cl::value_desc("directory"));
+
 static cl::list Libraries("l", cl::Prefix,
   cl::desc("Specify libraries to link to"),
   cl::value_desc("library prefix"));
 
+static cl::list Frameworks("framework",
+  cl::desc("Specify frameworks to link to"),
+  cl::value_desc("framework"));
+
 // Options to control the linking, optimization, and code gen processes
 static cl::opt LinkAsLibrary("link-as-library",
   cl::desc("Link the .bc files together as a library, not an executable"));
@@ -287,6 +295,8 @@
 ///  OutputFilename  - The name of the file to generate.
 ///  NativeLinkItems - The native libraries, files, code with which to link
 ///  LibPaths- The list of directories in which to find libraries.
+///  FrameworksPaths - The list of directories in which to find frameworks.
+///  Frameworks  - The list of frameworks (dynamic libraries)
 ///  gcc - The pathname to use for GGC.
 ///  envp- A copy of the process's current environment.
 ///
@@ -331,10 +341,12 @@
   args.push_back(OutputFilename);
   args.push_back(InputFilename);
 
-  // Add in the library paths
+  // Add in the library and framework paths
   for (unsigned index = 0; index < LibPaths.size(); index++) {
-args.push_back("-L");
-args.push_back(LibPaths[index]);
+args.push_back("-L" + LibPaths[index]);
+  }
+  for (unsigned index = 0; index < FrameworkPaths.size(); index++) {
+args.push_back("-F" + FrameworkPaths[index]);
   }
 
   // Add the requested options
@@ -350,6 +362,11 @@
 args.push_back(LinkItems[index].first);
 }
 
+  // Add in frameworks to link.
+  for (unsigned index = 0; index < Frameworks.size(); index++) {
+args.push_back("-framework");
+args.push_back(Frameworks[index]);
+  }
   
   // Now that "args" owns all the std::strings for the arguments, call the 
c_str
   // method to get the underlying string array.  We do this game so that the


___
llvm-commits mailing list
llvm-commits@cs.uiuc.edu
http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits

[llvm-commits] [llvm] r46422 - /llvm/trunk/lib/CodeGen/SelectionDAG/DAGCombiner.cpp

[Chris Lattner]

Author: lattner
Date: Sun Jan 27 17:21:58 2008
New Revision: 46422

URL: http://llvm.org/viewvc/llvm-project?rev=46422&view=rev
Log:
fix a crash on CodeGen/X86/vector-rem.ll

Modified:
llvm/trunk/lib/CodeGen/SelectionDAG/DAGCombiner.cpp

Modified: llvm/trunk/lib/CodeGen/SelectionDAG/DAGCombiner.cpp
URL: 
http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/CodeGen/SelectionDAG/DAGCombiner.cpp?rev=46422&r1=46421&r2=46422&view=diff

==
--- llvm/trunk/lib/CodeGen/SelectionDAG/DAGCombiner.cpp (original)
+++ llvm/trunk/lib/CodeGen/SelectionDAG/DAGCombiner.cpp Sun Jan 27 17:21:58 2008
@@ -1306,10 +1306,12 @@
 return DAG.getNode(ISD::SREM, VT, N0, N1);
   // If we know the sign bits of both operands are zero, strength reduce to a
   // urem instead.  Handles (X & 0x0FFF) %s 16 -> X&15
-  uint64_t SignBit = MVT::getIntVTSignBit(VT);
-  if (DAG.MaskedValueIsZero(N1, SignBit) &&
-  DAG.MaskedValueIsZero(N0, SignBit))
-return DAG.getNode(ISD::UREM, VT, N0, N1);
+  if (!MVT::isVector(VT)) {
+uint64_t SignBit = MVT::getIntVTSignBit(VT);
+if (DAG.MaskedValueIsZero(N1, SignBit) &&
+DAG.MaskedValueIsZero(N0, SignBit))
+  return DAG.getNode(ISD::UREM, VT, N0, N1);
+  }
   
   // If X/C can be simplified by the division-by-constant logic, lower
   // X%C to the equivalent of X-X/C*C.


___
llvm-commits mailing list
llvm-commits@cs.uiuc.edu
http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits

[llvm-commits] [llvm] r46423 - /llvm/trunk/test/CodeGen/ARM/fparith.ll

[Chris Lattner]

Author: lattner
Date: Sun Jan 27 17:26:37 2008
New Revision: 46423

URL: http://llvm.org/viewvc/llvm-project?rev=46423&view=rev
Log:
Update this test.  Due to dag combiner improvements, we now compile
f7/f11 to:

_f7:
eor r0, r0, #2, 2 @ -2147483648
bx lr
_f11:
bic r0, r0, #2, 2 @ -2147483648
bx lr

instead of:

_f7:
fmsr s0, r0
fnegs s0, s0
fmrs r0, s0
bx lr

_f11:
fmsr s0, r0
fabss s0, s0
fmrs r0, s0
bx lr


Modified:
llvm/trunk/test/CodeGen/ARM/fparith.ll

Modified: llvm/trunk/test/CodeGen/ARM/fparith.ll
URL: 
http://llvm.org/viewvc/llvm-project/llvm/trunk/test/CodeGen/ARM/fparith.ll?rev=46423&r1=46422&r2=46423&view=diff

==
--- llvm/trunk/test/CodeGen/ARM/fparith.ll (original)
+++ llvm/trunk/test/CodeGen/ARM/fparith.ll Sun Jan 27 17:26:37 2008
@@ -3,7 +3,7 @@
 ; RUN: grep faddd %t
 ; RUN: grep fmuls %t
 ; RUN: grep fmuld %t
-; RUN: grep fnegs %t
+; RUN: grep eor %t
 ; RUN: grep fnegd %t
 ; RUN: grep fdivs %t
 ; RUN: grep fdivd %t


___
llvm-commits mailing list
llvm-commits@cs.uiuc.edu
http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits

[llvm-commits] [llvm] r46425 - /llvm/trunk/docs/Makefile

[Chris Lattner]

Author: lattner
Date: Sun Jan 27 17:43:26 2008
New Revision: 46425

URL: http://llvm.org/viewvc/llvm-project?rev=46425&view=rev
Log:
Always for PACKAGEVERSION to 'mainline' for teh llvm web page.

Modified:
llvm/trunk/docs/Makefile

Modified: llvm/trunk/docs/Makefile
URL: 
http://llvm.org/viewvc/llvm-project/llvm/trunk/docs/Makefile?rev=46425&r1=46424&r2=46425&view=diff

==
--- llvm/trunk/docs/Makefile (original)
+++ llvm/trunk/docs/Makefile Sun Jan 27 17:43:26 2008
@@ -13,15 +13,12 @@
 ifdef BUILD_FOR_WEBSITE
 PROJ_OBJ_DIR = .
 DOXYGEN = doxygen
-# Extract version number from the AC_INT line in configure.ac
-# AC_INIT([[llvm]],[[2.1svn]],[EMAIL PROTECTED])
-PACKAGE_VERSION = $(shell grep AC_INIT ../autoconf/configure.ac | sed -e 
's/^[^0-9]*\([0-9_.a-zA-Z-]*\).*/\1/' )
 
 $(PROJ_OBJ_DIR)/doxygen.cfg: doxygen.cfg.in
cat $< | sed \
  -e 's/@abs_top_srcdir@/../g' \
  -e 's/@DOT@/dot/g' \
- -e 's/@PACKAGE_VERSION@/$(PACKAGE_VERSION)/' \
+ -e 's/@PACKAGE_VERSION@/mainline/' \
  -e 's/@abs_top_builddir@/../g' > $@
 endif
 


___
llvm-commits mailing list
llvm-commits@cs.uiuc.edu
http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits

[llvm-commits] [llvm] r46424 - /llvm/trunk/lib/CodeGen/SelectionDAG/DAGCombiner.cpp

[Chris Lattner]

Author: lattner
Date: Sun Jan 27 17:32:17 2008
New Revision: 46424

URL: http://llvm.org/viewvc/llvm-project?rev=46424&view=rev
Log:
Fix PowerPC/./2007-10-18-PtrArithmetic.ll

Modified:
llvm/trunk/lib/CodeGen/SelectionDAG/DAGCombiner.cpp

Modified: llvm/trunk/lib/CodeGen/SelectionDAG/DAGCombiner.cpp
URL: 
http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/CodeGen/SelectionDAG/DAGCombiner.cpp?rev=46424&r1=46423&r2=46424&view=diff

==
--- llvm/trunk/lib/CodeGen/SelectionDAG/DAGCombiner.cpp (original)
+++ llvm/trunk/lib/CodeGen/SelectionDAG/DAGCombiner.cpp Sun Jan 27 17:32:17 2008
@@ -1190,10 +1190,12 @@
 return DAG.getNode(ISD::SUB, VT, DAG.getConstant(0, VT), N0);
   // If we know the sign bits of both operands are zero, strength reduce to a
   // udiv instead.  Handles (X&15) /s 4 -> X&15 >> 2
-  uint64_t SignBit = MVT::getIntVTSignBit(VT);
-  if (DAG.MaskedValueIsZero(N1, SignBit) &&
-  DAG.MaskedValueIsZero(N0, SignBit))
-return DAG.getNode(ISD::UDIV, N1.getValueType(), N0, N1);
+  if (!MVT::isVector(VT)) {
+uint64_t SignBit = MVT::getIntVTSignBit(VT);
+if (DAG.MaskedValueIsZero(N1, SignBit) &&
+DAG.MaskedValueIsZero(N0, SignBit))
+  return DAG.getNode(ISD::UDIV, N1.getValueType(), N0, N1);
+  }
   // fold (sdiv X, pow2) -> simple ops after legalize
   if (N1C && N1C->getValue() && !TLI.isIntDivCheap() &&
   (isPowerOf2_64(N1C->getSignExtended()) || 
@@ -3297,7 +3299,8 @@
   // Note that we don't handle copysign(x,cst) because this can always be 
folded
   // to an fneg or fabs.
   if (N0.getOpcode() == ISD::FCOPYSIGN && N0.Val->hasOneUse() &&
-  isa(N0.getOperand(0))) {
+  isa(N0.getOperand(0)) &&
+  MVT::isInteger(VT) && !MVT::isVector(VT)) {
 unsigned OrigXWidth = MVT::getSizeInBits(N0.getOperand(1).getValueType());
 SDOperand X = DAG.getNode(ISD::BIT_CONVERT, 
MVT::getIntegerType(OrigXWidth),
   N0.getOperand(1));
@@ -3787,7 +3790,9 @@
 
   // Transform fneg(bitconvert(x)) -> bitconvert(x^sign) to avoid loading
   // constant pool values.
-  if (N0.getOpcode() == ISD::BIT_CONVERT && N0.Val->hasOneUse()) {
+  if (N0.getOpcode() == ISD::BIT_CONVERT && N0.Val->hasOneUse() &&
+  MVT::isInteger(N0.getOperand(0).getValueType()) &&
+  !MVT::isVector(N0.getOperand(0).getValueType())) {
 SDOperand Int = N0.getOperand(0);
 MVT::ValueType IntVT = Int.getValueType();
 if (MVT::isInteger(IntVT) && !MVT::isVector(IntVT)) {
@@ -3819,7 +3824,9 @@
   
   // Transform fabs(bitconvert(x)) -> bitconvert(x&~sign) to avoid loading
   // constant pool values.
-  if (N0.getOpcode() == ISD::BIT_CONVERT && N0.Val->hasOneUse()) {
+  if (N0.getOpcode() == ISD::BIT_CONVERT && N0.Val->hasOneUse() &&
+  MVT::isInteger(N0.getOperand(0).getValueType()) &&
+  !MVT::isVector(N0.getOperand(0).getValueType())) {
 SDOperand Int = N0.getOperand(0);
 MVT::ValueType IntVT = Int.getValueType();
 if (MVT::isInteger(IntVT) && !MVT::isVector(IntVT)) {


___
llvm-commits mailing list
llvm-commits@cs.uiuc.edu
http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits

Re: [llvm-commits] [llvm-gcc-4.2] r46381 - in /llvm-gcc-4.2/trunk/gcc: llvm-abi.h llvm-internal.h llvm-types.cpp

[Evan Cheng]

On Jan 27, 2008, at 1:03 AM, Duncan Sands <[EMAIL PROTECTED]> wrote:

> Hi Evan,
>
>> Can you get me a test case? I need to see what gcc is doing. Thanks,
>
> Sure.  Zero sized fields in structs may not be very useful, but it is
> still legal to use them, in particular you can take their address,
> like this:
>
> struct Z {};
>
> struct Y {
>int i;
>struct Z z;
> };
>
> void *f(struct Y *y) {
>  return &y->z;
> }
>
> This used to work before your patch.  For that matter, llvm-gcc-4.2
> itself fails to build here (linux x86-32; crashes compiling  
> libgcov.c) for
> what looks to be the same reason.  Also, don't forget that Ada can  
> have
> fields of variable size in various places in a struct, and these can  
> be
> represented by zero-size LLVM fields (this kind of struct doesn't get
> passed to functions by copy though).
>
> Can you please explain why you made this change (a testcase would be
> good)

I did check in two test cases in a subsequent commit. Those used to  
choke llvm-gcc. I'll look at this today. It's obvious not all zero  
sized aggregates are not created equal.

Evan

>
>
> Best wishes,
>
> Duncan.
___
llvm-commits mailing list
llvm-commits@cs.uiuc.edu
http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits

[llvm-commits] [llvm] r46427 - in /llvm/trunk: lib/Transforms/Scalar/SCCP.cpp test/Transforms/SCCP/2008-01-27-UndefCorrelate.ll

[Chris Lattner]

Author: lattner
Date: Sun Jan 27 18:32:30 2008
New Revision: 46427

URL: http://llvm.org/viewvc/llvm-project?rev=46427&view=rev
Log:
Fix PR1938 by forcing the code that uses an undefined value to branch one
way or the other.  Rewriting the code itself prevents subsequent analysis
passes from making contradictory conclusions about the code that could 
cause an infeasible path to be made feasible.

Added:
llvm/trunk/test/Transforms/SCCP/2008-01-27-UndefCorrelate.ll
Modified:
llvm/trunk/lib/Transforms/Scalar/SCCP.cpp

Modified: llvm/trunk/lib/Transforms/Scalar/SCCP.cpp
URL: 
http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Transforms/Scalar/SCCP.cpp?rev=46427&r1=46426&r2=46427&view=diff

==
--- llvm/trunk/lib/Transforms/Scalar/SCCP.cpp (original)
+++ llvm/trunk/lib/Transforms/Scalar/SCCP.cpp Sun Jan 27 18:32:30 2008
@@ -1313,15 +1313,30 @@
   continue;
 }
 
-// If the edge to the first successor isn't thought to be feasible yet, 
mark
-// it so now.
-if (KnownFeasibleEdges.count(Edge(BB, TI->getSuccessor(0
+// If the edge to the second successor isn't thought to be feasible yet,
+// mark it so now.  We pick the second one so that this goes to some
+// enumerated value in a switch instead of going to the default 
destination.
+if (KnownFeasibleEdges.count(Edge(BB, TI->getSuccessor(1
   continue;
 
 // Otherwise, it isn't already thought to be feasible.  Mark it as such now
 // and return.  This will make other blocks reachable, which will allow new
 // values to be discovered and existing ones to be moved in the lattice.
-markEdgeExecutable(BB, TI->getSuccessor(0));
+markEdgeExecutable(BB, TI->getSuccessor(1));
+
+// This must be a conditional branch of switch on undef.  At this point,
+// force the old terminator to branch to the first successor.  This is
+// required because we are now influencing the dataflow of the function 
with
+// the assumption that this edge is taken.  If we leave the branch 
condition
+// as undef, then further analysis could think the undef went another way
+// leading to an inconsistent set of conclusions.
+if (BranchInst *BI = dyn_cast(TI)) {
+  BI->setCondition(ConstantInt::getFalse());
+} else {
+  SwitchInst *SI = cast(TI);
+  SI->setCondition(SI->getCaseValue(1));
+}
+
 return true;
   }
 

Added: llvm/trunk/test/Transforms/SCCP/2008-01-27-UndefCorrelate.ll
URL: 
http://llvm.org/viewvc/llvm-project/llvm/trunk/test/Transforms/SCCP/2008-01-27-UndefCorrelate.ll?rev=46427&view=auto

==
--- llvm/trunk/test/Transforms/SCCP/2008-01-27-UndefCorrelate.ll (added)
+++ llvm/trunk/test/Transforms/SCCP/2008-01-27-UndefCorrelate.ll Sun Jan 27 
18:32:30 2008
@@ -0,0 +1,36 @@
+; RUN: llvm-as < %s | opt -sccp | llvm-dis | grep undef | count 1
+; PR1938
+
+define i32 @main() {
+entry:
+   br label %bb
+
+bb:
+   %indvar = phi i32 [ 0, %entry ], [ %k, %bb.backedge ]
+   %k = add i32 %indvar, 1
+   br i1 undef, label %cond_true, label %cond_false
+
+cond_true:
+   %tmp97 = icmp slt i32 %k, 10
+   br i1 %tmp97, label %bb.backedge, label %bb12
+
+bb.backedge:
+   br label %bb
+
+cond_false:
+   %tmp9 = icmp slt i32 %k, 10
+   br i1 %tmp9, label %bb.backedge, label %bb12
+
+bb12:
+   %tmp14 = icmp eq i32 %k, 10
+   br i1 %tmp14, label %cond_next18, label %cond_true17
+
+cond_true17:
+   tail call void @abort( )
+   unreachable
+
+cond_next18:
+   ret i32 0
+}
+
+declare void @abort()


___
llvm-commits mailing list
llvm-commits@cs.uiuc.edu
http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits

[llvm-commits] [llvm] r46428 - /llvm/trunk/docs/LangRef.html

[Chris Lattner]

Author: lattner
Date: Sun Jan 27 18:36:27 2008
New Revision: 46428

URL: http://llvm.org/viewvc/llvm-project?rev=46428&view=rev
Log:
make handling of overflow and undefined results much more clear.  
Patch by Eli Friedman, thanks!

Modified:
llvm/trunk/docs/LangRef.html

Modified: llvm/trunk/docs/LangRef.html
URL: 
http://llvm.org/viewvc/llvm-project/llvm/trunk/docs/LangRef.html?rev=46428&r1=46427&r2=46428&view=diff

==
--- llvm/trunk/docs/LangRef.html (original)
+++ llvm/trunk/docs/LangRef.html Sun Jan 27 18:36:27 2008
@@ -2051,6 +2051,11 @@
 Semantics:
 The value produced is the integer or floating point sum of the two
 operands.
+If an integer sum has unsigned overflow, the result returned is the
+mathematical result modulo 2n, where n is the bit width of
+the result.
+Because LLVM integers use a two's complement representation, this
+instruction is appropriate for both signed and unsigned integers.
 Example:
    = add i32 4, %var  ; yields {i32}:result = 4 
+ %var
 
@@ -2076,6 +2081,11 @@
 Semantics:
 The value produced is the integer or floating point difference of
 the two operands.
+If an integer difference has unsigned overflow, the result returned is the
+mathematical result modulo 2n, where n is the bit width of
+the result.
+Because LLVM integers use a two's complement representation, this
+instruction is appropriate for both signed and unsigned integers.
 Example:
 
    = sub i32 4, %var  ; yields {i32}:result = 4 - 
%var
@@ -2101,9 +2111,15 @@
 Semantics:
 The value produced is the integer or floating point product of the
 two operands.
-Because the operands are the same width, the result of an integer
-multiplication is the same whether the operands should be deemed unsigned or
-signed.
+If the result of an integer multiplication has unsigned overflow,
+the result returned is the mathematical result modulo 
+2n, where n is the bit width of the result.
+Because LLVM integers use a two's complement representation, and the
+result is the same width as the operands, this instruction returns the
+correct result for both signed and unsigned integers.  If a full product
+(e.g. i32xi32->i64) is needed, the operands
+should be sign-extended or zero-extended as appropriate to the
+width of the full product.
 Example:
    = mul i32 4, %var  ; yields {i32}:result = 4 
* %var
 
@@ -2124,9 +2140,10 @@
 types. This instruction can also take vector versions 
 of the values in which case the elements must be integers.
 Semantics:
-The value produced is the unsigned integer quotient of the two operands. 
This
-instruction always performs an unsigned division operation, regardless of 
-whether the arguments are unsigned or not.
+The value produced is the unsigned integer quotient of the two operands.
+Note that unsigned integer division and signed integer division are distinct
+operations; for signed integer division, use 'sdiv'.
+Division by zero leads to undefined behavior.
 Example:
    = udiv i32 4, %var  ; yields {i32}:result = 4 
/ %var
 
@@ -2147,9 +2164,12 @@
 types. This instruction can also take vector versions 
 of the values in which case the elements must be integers.
 Semantics:
-The value produced is the signed integer quotient of the two operands. This
-instruction always performs a signed division operation, regardless of whether
-the arguments are signed or not.
+The value produced is the signed integer quotient of the two operands.
+Note that signed integer division and unsigned integer division are distinct
+operations; for unsigned integer division, use 'udiv'.
+Division by zero leads to undefined behavior. Overflow also leads to
+undefined behavior; this is a rare case, but can occur, for example,
+by doing a 32-bit division of -2147483648 by -1.
 Example:
    = sdiv i32 4, %var  ; yields {i32}:result = 4 
/ %var
 
@@ -2194,6 +2214,9 @@
 This instruction returns the unsigned integer remainder of a 
division.
 This instruction always performs an unsigned division to get the remainder,
 regardless of whether the arguments are unsigned or not.
+Note that unsigned integer remainder and signed integer remainder are
+distinct operations; for signed integer remainder, use 'srem'.
+Taking the remainder of a division by zero leads to undefined behavior.
 Example:
    = urem i32 4, %var  ; yields {i32}:result = 4 
% %var
 
@@ -2225,6 +2248,14 @@
 Math Forum. For a table of how this is implemented in various languages,
 please see http://en.wikipedia.org/wiki/Modulo_operation";>
 Wikipedia: modulo operation.
+Note that signed integer remainder and unsigned integer remainder are
+distinct operations; for unsigned integer remainder, use 'urem'.
+Taking the remainder of a division by zero leads to undefined behavior.
+Overflow also leads to undefined behavior; this is a rare case, but can occur,
+for example, by taking the remaind

[llvm-commits] [llvm] r46429 - in /llvm/trunk: lib/Transforms/Scalar/InstructionCombining.cpp test/Transforms/InstCombine/2008-01-27-FloatSelect.ll

[Chris Lattner]

Author: lattner
Date: Sun Jan 27 18:58:18 2008
New Revision: 46429

URL: http://llvm.org/viewvc/llvm-project?rev=46429&view=rev
Log:
Fix PR1932 by disabling an xform invalid for fdiv.

Added:
llvm/trunk/test/Transforms/InstCombine/2008-01-27-FloatSelect.ll
Modified:
llvm/trunk/lib/Transforms/Scalar/InstructionCombining.cpp

Modified: llvm/trunk/lib/Transforms/Scalar/InstructionCombining.cpp
URL: 
http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Transforms/Scalar/InstructionCombining.cpp?rev=46429&r1=46428&r2=46429&view=diff

==
--- llvm/trunk/lib/Transforms/Scalar/InstructionCombining.cpp (original)
+++ llvm/trunk/lib/Transforms/Scalar/InstructionCombining.cpp Sun Jan 27 
18:58:18 2008
@@ -2494,14 +2494,15 @@
   if (isa(Op1))
 return ReplaceInstUsesWith(I, Op1);
 
-  // Handle cases involving: div X, (select Cond, Y, Z)
+  // Handle cases involving: [su]div X, (select Cond, Y, Z)
+  // This does not apply for fdiv.
   if (SelectInst *SI = dyn_cast(Op1)) {
-// div X, (Cond ? 0 : Y) -> div X, Y.  If the div and the select are in the
-// same basic block, then we replace the select with Y, and the condition 
-// of the select with false (if the cond value is in the same BB).  If the
-// select has uses other than the div, this allows them to be simplified
-// also. Note that div X, Y is just as good as div X, 0 (undef)
-if (Constant *ST = dyn_cast(SI->getOperand(1)))
+// [su]div X, (Cond ? 0 : Y) -> div X, Y.  If the div and the select are in
+// the same basic block, then we replace the select with Y, and the
+// condition of the select with false (if the cond value is in the same 
BB).
+// If the select has uses other than the div, this allows them to be
+// simplified also. Note that div X, Y is just as good as div X, 0 (undef)
+if (ConstantInt *ST = dyn_cast(SI->getOperand(1)))
   if (ST->isNullValue()) {
 Instruction *CondI = dyn_cast(SI->getOperand(0));
 if (CondI && CondI->getParent() == I.getParent())
@@ -2513,8 +2514,8 @@
 return &I;
   }
 
-// Likewise for: div X, (Cond ? Y : 0) -> div X, Y
-if (Constant *ST = dyn_cast(SI->getOperand(2)))
+// Likewise for: [su]div X, (Cond ? Y : 0) -> div X, Y
+if (ConstantInt *ST = dyn_cast(SI->getOperand(2)))
   if (ST->isNullValue()) {
 Instruction *CondI = dyn_cast(SI->getOperand(0));
 if (CondI && CondI->getParent() == I.getParent())

Added: llvm/trunk/test/Transforms/InstCombine/2008-01-27-FloatSelect.ll
URL: 
http://llvm.org/viewvc/llvm-project/llvm/trunk/test/Transforms/InstCombine/2008-01-27-FloatSelect.ll?rev=46429&view=auto

==
--- llvm/trunk/test/Transforms/InstCombine/2008-01-27-FloatSelect.ll (added)
+++ llvm/trunk/test/Transforms/InstCombine/2008-01-27-FloatSelect.ll Sun Jan 27 
18:58:18 2008
@@ -0,0 +1,7 @@
+; RUN: llvm-as < %s | opt -instcombine | llvm-dis | grep select
+
+define double @fold(i1 %a, double %b) {
+%s = select i1 %a, double 0., double 1.
+%c = fdiv double %b, %s
+ret double %c
+}


___
llvm-commits mailing list
llvm-commits@cs.uiuc.edu
http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits

[llvm-commits] [llvm] r46431 - in /llvm/trunk: include/llvm/Instructions.h lib/Transforms/Scalar/InstructionCombining.cpp lib/VMCore/Instructions.cpp test/Transforms/InstCombine/2008-01-21-MismatchedC

[Nick Lewycky]

Author: nicholas
Date: Sun Jan 27 21:48:02 2008
New Revision: 46431

URL: http://llvm.org/viewvc/llvm-project?rev=46431&view=rev
Log:
Handle some more combinations of extend and icmp. Fixes PR1940.

Added:

llvm/trunk/test/Transforms/InstCombine/2008-01-21-MismatchedCastAndCompare.ll
Modified:
llvm/trunk/include/llvm/Instructions.h
llvm/trunk/lib/Transforms/Scalar/InstructionCombining.cpp
llvm/trunk/lib/VMCore/Instructions.cpp

Modified: llvm/trunk/include/llvm/Instructions.h
URL: 
http://llvm.org/viewvc/llvm-project/llvm/trunk/include/llvm/Instructions.h?rev=46431&r1=46430&r2=46431&view=diff

==
--- llvm/trunk/include/llvm/Instructions.h (original)
+++ llvm/trunk/include/llvm/Instructions.h Sun Jan 27 21:48:02 2008
@@ -642,6 +642,18 @@
   /// @brief Return the signed version of the predicate.
   static Predicate getSignedPredicate(Predicate pred);
 
+  /// For example, EQ->EQ, SLE->ULE, UGT->UGT, etc.
+  /// @returns the predicate that would be the result if the operand were
+  /// regarded as unsigned.
+  /// @brief Return the unsigned version of the predicate
+  Predicate getUnsignedPredicate() const {
+return getUnsignedPredicate(getPredicate());
+  }
+
+  /// This is a static version that you can use without an instruction.
+  /// @brief Return the unsigned version of the predicate.
+  static Predicate getUnsignedPredicate(Predicate pred);
+
   /// isEquality - Return true if this predicate is either EQ or NE.  This also
   /// tests for commutativity.
   static bool isEquality(Predicate P) {

Modified: llvm/trunk/lib/Transforms/Scalar/InstructionCombining.cpp
URL: 
http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Transforms/Scalar/InstructionCombining.cpp?rev=46431&r1=46430&r2=46431&view=diff

==
--- llvm/trunk/lib/Transforms/Scalar/InstructionCombining.cpp (original)
+++ llvm/trunk/lib/Transforms/Scalar/InstructionCombining.cpp Sun Jan 27 
21:48:02 2008
@@ -5820,18 +5820,22 @@
 if (RHSCIOp->getType() != LHSCIOp->getType()) 
   return 0;
 
-// If the signedness of the two compares doesn't agree (i.e. one is a sext
+// If the signedness of the two casts doesn't agree (i.e. one is a sext
 // and the other is a zext), then we can't handle this.
 if (CI->getOpcode() != LHSCI->getOpcode())
   return 0;
 
-// Likewise, if the signedness of the [sz]exts and the compare don't 
match, 
-// then we can't handle this.
-if (isSignedExt != isSignedCmp && !ICI.isEquality())
-  return 0;
-
-// Okay, just insert a compare of the reduced operands now!
-return new ICmpInst(ICI.getPredicate(), LHSCIOp, RHSCIOp);
+// Deal with equality cases early.
+if (ICI.isEquality())
+  return new ICmpInst(ICI.getPredicate(), LHSCIOp, RHSCIOp);
+
+// A signed comparison of sign extended values simplifies into a
+// signed comparison.
+if (isSignedCmp && isSignedExt)
+  return new ICmpInst(ICI.getPredicate(), LHSCIOp, RHSCIOp);
+
+// The other three cases all fold into an unsigned comparison.
+return new ICmpInst(ICI.getUnsignedPredicate(), LHSCIOp, RHSCIOp);
   }
 
   // If we aren't dealing with a constant on the RHS, exit early

Modified: llvm/trunk/lib/VMCore/Instructions.cpp
URL: 
http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/VMCore/Instructions.cpp?rev=46431&r1=46430&r2=46431&view=diff

==
--- llvm/trunk/lib/VMCore/Instructions.cpp (original)
+++ llvm/trunk/lib/VMCore/Instructions.cpp Sun Jan 27 21:48:02 2008
@@ -2429,6 +2429,19 @@
   }
 }
 
+ICmpInst::Predicate ICmpInst::getUnsignedPredicate(Predicate pred) {
+  switch (pred) {
+default: assert(! "Unknown icmp predicate!");
+case ICMP_EQ: case ICMP_NE: 
+case ICMP_UGT: case ICMP_ULT: case ICMP_UGE: case ICMP_ULE: 
+   return pred;
+case ICMP_SGT: return ICMP_UGT;
+case ICMP_SLT: return ICMP_ULT;
+case ICMP_SGE: return ICMP_UGE;
+case ICMP_SLE: return ICMP_ULE;
+  }
+}
+
 bool ICmpInst::isSignedPredicate(Predicate pred) {
   switch (pred) {
 default: assert(! "Unknown icmp predicate!");

Added: 
llvm/trunk/test/Transforms/InstCombine/2008-01-21-MismatchedCastAndCompare.ll
URL: 
http://llvm.org/viewvc/llvm-project/llvm/trunk/test/Transforms/InstCombine/2008-01-21-MismatchedCastAndCompare.ll?rev=46431&view=auto

==
--- 
llvm/trunk/test/Transforms/InstCombine/2008-01-21-MismatchedCastAndCompare.ll 
(added)
+++ 
llvm/trunk/test/Transforms/InstCombine/2008-01-21-MismatchedCastAndCompare.ll 
Sun Jan 27 21:48:02 2008
@@ -0,0 +1,17 @@
+; RUN: llvm-as < %s | opt -instcombine | notcast
+; RUN: llvm-as < %s | opt -instcombine | not grep {icmp s}
+; PR1940
+
+define i1 @test1(i8 %A, i8 %B) {
+%a = zext i8 %A to i32
+%b 

[llvm-commits] [llvm] r46432 - /llvm/trunk/Makefile.rules

[Chris Lattner]

Author: lattner
Date: Sun Jan 27 22:18:41 2008
New Revision: 46432

URL: http://llvm.org/viewvc/llvm-project?rev=46432&view=rev
Log:
reduce duplicate -I flags passed to the compiler, cleaning up the VERBOSE
output.  Patch contributed by Sam Bishop!

Modified:
llvm/trunk/Makefile.rules

Modified: llvm/trunk/Makefile.rules
URL: 
http://llvm.org/viewvc/llvm-project/llvm/trunk/Makefile.rules?rev=46432&r1=46431&r2=46432&view=diff

==
--- llvm/trunk/Makefile.rules (original)
+++ llvm/trunk/Makefile.rules Sun Jan 27 22:18:41 2008
@@ -454,12 +454,11 @@
 LD.Flags  += -L$(LibDir) -L$(LLVMLibDir) 
 CPP.BaseFlags += -D_GNU_SOURCE -D__STDC_LIMIT_MACROS
 # All -I flags should go here, so that they don't confuse llvm-config.
-CPP.Flags += -I$(PROJ_OBJ_DIR) -I$(PROJ_SRC_DIR) \
--I$(PROJ_OBJ_ROOT)/include \
--I$(PROJ_SRC_ROOT)/include \
--I$(LLVM_OBJ_ROOT)/include \
--I$(LLVM_SRC_ROOT)/include \
-$(CPP.BaseFlags)
+CPP.Flags += $(patsubst %,-I%/include,$(sort \
+$(PROJ_OBJ_DIR) $(PROJ_SRC_DIR) \
+$(PROJ_OBJ_ROOT) $(PROJ_SRC_ROOT) \
+$(LLVM_OBJ_ROOT) $(LLVM_SRC_ROOT))) \
+$(CPP.BaseFlags)
 
 Compile.C = $(CC) $(CPP.Flags) $(C.Flags) $(CompileCommonOpts) -c
 LTCompile.C   = $(LIBTOOL) $(LibTool.Flags) --mode=compile $(Compile.C)


___
llvm-commits mailing list
llvm-commits@cs.uiuc.edu
http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits

Re: [llvm-commits] Makefile.rules patch: remove redundant -I compiler flags

[Chris Lattner]

On Jan 15, 2008, at 9:07 AM, Sam Bishop wrote:

> The attached patch makes VERBOSE builds easier to wade through.   
> It's likely
> to also permute the PROJ_* and LLVM_* includes, but I don't think that
> should matter.  I've done a full-tree compile and everything seems  
> to be
> working.

Applied, thanks!
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20080121/057641.html

-Chris
___
llvm-commits mailing list
llvm-commits@cs.uiuc.edu
http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits

Re: [llvm-commits] [llvm] r46394 - in /llvm/trunk/lib: Target/PowerPC/PPCAsmPrinter.cpp Transforms/Scalar/SimplifyCFG.cpp

[Chris Lattner]

>>> +  BuildMI(MBB, MBB.end(), TM.getInstrInfo()->get(PPC::NOP));
>>> +  }
>>
>> Is there any reason to actually make a machineinstr here?  It seems
>> like it would be possible to just emit "nop" to the .s file.  The JIT
>> doesn't need this.
>>
> I initially had that, but I thought that it would be better to do
> things though the "print machine instruction" function just so I got
> the syntax correct. But nop is probably safe to emit like that.

"nop" is valid on arm, thanks.

-Chris

___
llvm-commits mailing list
llvm-commits@cs.uiuc.edu
http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits

[llvm-commits] [llvm] r46433 - /llvm/trunk/lib/Transforms/IPO/SimplifyLibCalls.cpp

[Chris Lattner]

Author: lattner
Date: Sun Jan 27 22:41:43 2008
New Revision: 46433

URL: http://llvm.org/viewvc/llvm-project?rev=46433&view=rev
Log:
Transform calls to memcpy into llvm.memcpy calls, patch by Eli Friedman.

Modified:
llvm/trunk/lib/Transforms/IPO/SimplifyLibCalls.cpp

Modified: llvm/trunk/lib/Transforms/IPO/SimplifyLibCalls.cpp
URL: 
http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Transforms/IPO/SimplifyLibCalls.cpp?rev=46433&r1=46432&r2=46433&view=diff

==
--- llvm/trunk/lib/Transforms/IPO/SimplifyLibCalls.cpp (original)
+++ llvm/trunk/lib/Transforms/IPO/SimplifyLibCalls.cpp Sun Jan 27 22:41:43 2008
@@ -919,6 +919,36 @@
   }
 } memcmpOptimizer;
 
+/// This LibCallOptimization will simplify a call to the memcpy library
+/// function.  It simply converts them into calls to llvm.memcpy.*;
+/// the resulting call should be optimized later.
+/// @brief Simplify the memcpy library function.
+struct VISIBILITY_HIDDEN MemCpyOptimization : public LibCallOptimization {
+public:
+  MemCpyOptimization() : LibCallOptimization("memcpy",
+  "Number of 'memcpy' calls simplified") {}
+
+  /// @brief Make sure that the "memcpy" function has the right prototype
+  virtual bool ValidateCalledFunction(const Function *F, SimplifyLibCalls 
&SLC){
+const FunctionType *FT = F->getFunctionType();
+const Type* voidPtr = PointerType::getUnqual(Type::Int8Ty);
+return FT->getReturnType() == voidPtr && FT->getNumParams() == 3 &&
+   FT->getParamType(0) == voidPtr &&
+   FT->getParamType(1) == voidPtr &&
+   FT->getParamType(2) == SLC.getIntPtrType();
+  }
+
+  /// @brief Perform the memcpy optimization
+  virtual bool OptimizeCall(CallInst *CI, SimplifyLibCalls &SLC) {
+Value *MemcpyOps[] = {
+  CI->getOperand(1), CI->getOperand(2), CI->getOperand(3),
+  ConstantInt::get(Type::Int32Ty, 1)   // align = 1 always.
+};
+new CallInst(SLC.get_memcpy(), MemcpyOps, MemcpyOps + 4, "", CI);
+// memcpy always returns the destination
+return ReplaceCallWith(CI, CI->getOperand(1));
+  }
+} MemCpyOptimizer;
 
 /// This LibCallOptimization will simplify a call to the memcpy library
 /// function by expanding it out to a single store of size 0, 1, 2, 4, or 8


___
llvm-commits mailing list
llvm-commits@cs.uiuc.edu
http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits

[llvm-commits] [test-suite] r46434 - in /test-suite/trunk/MultiSource/Benchmarks/VersaBench: ./ README.txt beamformer/ beamformer/Makefile beamformer/beamformer.c dbms/

[Chris Lattner]

Author: lattner
Date: Sun Jan 27 23:17:19 2008
New Revision: 46434

URL: http://llvm.org/viewvc/llvm-project?rev=46434&view=rev
Log:
add a benchmark from 'VersaBench'.

Added:
test-suite/trunk/MultiSource/Benchmarks/VersaBench/
test-suite/trunk/MultiSource/Benchmarks/VersaBench/README.txt
test-suite/trunk/MultiSource/Benchmarks/VersaBench/beamformer/
test-suite/trunk/MultiSource/Benchmarks/VersaBench/beamformer/Makefile
test-suite/trunk/MultiSource/Benchmarks/VersaBench/beamformer/beamformer.c
test-suite/trunk/MultiSource/Benchmarks/VersaBench/dbms/

Added: test-suite/trunk/MultiSource/Benchmarks/VersaBench/README.txt
URL: 
http://llvm.org/viewvc/llvm-project/test-suite/trunk/MultiSource/Benchmarks/VersaBench/README.txt?rev=46434&view=auto

==
--- test-suite/trunk/MultiSource/Benchmarks/VersaBench/README.txt (added)
+++ test-suite/trunk/MultiSource/Benchmarks/VersaBench/README.txt Sun Jan 27 
23:17:19 2008
@@ -0,0 +1,2 @@
+Obtained from:
+http://cag.csail.mit.edu/versabench/benchmarks.html

Added: test-suite/trunk/MultiSource/Benchmarks/VersaBench/beamformer/Makefile
URL: 
http://llvm.org/viewvc/llvm-project/test-suite/trunk/MultiSource/Benchmarks/VersaBench/beamformer/Makefile?rev=46434&view=auto

==
--- test-suite/trunk/MultiSource/Benchmarks/VersaBench/beamformer/Makefile 
(added)
+++ test-suite/trunk/MultiSource/Benchmarks/VersaBench/beamformer/Makefile Sun 
Jan 27 23:17:19 2008
@@ -0,0 +1,10 @@
+LEVEL = ../../../..
+
+PROG = beamformer
+ifdef LARGE_PROBLEM_SIZE
+RUN_OPTIONS = -i 400
+else
+RUN_OPTIONS = -i 140
+endif
+include $(LEVEL)/MultiSource/Makefile.multisrc
+

Added: 
test-suite/trunk/MultiSource/Benchmarks/VersaBench/beamformer/beamformer.c
URL: 
http://llvm.org/viewvc/llvm-project/test-suite/trunk/MultiSource/Benchmarks/VersaBench/beamformer/beamformer.c?rev=46434&view=auto

==
--- test-suite/trunk/MultiSource/Benchmarks/VersaBench/beamformer/beamformer.c 
(added)
+++ test-suite/trunk/MultiSource/Benchmarks/VersaBench/beamformer/beamformer.c 
Sun Jan 27 23:17:19 2008
@@ -0,0 +1,402 @@
+/*
+ * Copyright (c) 2003 David Maze
+ *
+ * Permission  is hereby  granted,  free  of  charge, to  any  person
+ * obtaining a  copy of  this software  and  associated documentation
+ * files   (the  "Software"),  to   deal  in  the   Software  without
+ * restriction, including without  limitation the rights to use, copy,
+ * modify, merge, publish,  distribute, sublicense, and/or sell copies
+ * of  the Software,  and to  permit persons to  whom the  Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above  copyright notice  and this  permission notice  shall be
+ * included in all copies or substantial portions of the Software.
+ *
+ * THE  SOFTWARE IS  PROVIDED "AS IS",  WITHOUT WARRANTY OF  ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING  BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY,   FITNESS   FOR   A   PARTICULARPURPOSEAND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM,  DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF  CONTRACT, TORT OR OTHERWISE, ARISING FROM,  OUT OF OR IN
+ * CONNECTION  WITH THE SOFTWARE OR  THE USE OR OTHER  DEALINGS IN THE
+ * SOFTWARE.  
+ */
+
+/*
+ * beamformer.c: Standalone beam-former reference implementation
+ * David Maze <[EMAIL PROTECTED]>
+ * $Id: beamformer.c,v 1.5 2003/11/07 08:47:00 thies Exp $
+ */
+
+/* Modified by: Rodric M. Rabbah 06-03-04 */
+
+#include 
+#include 
+#include 
+#include 
+
+/* 
+This implementation is derived from the StreamIt implementation,
+rather than from the PCA VSIPL-based implementation.  It is intended
+to be easier to synchronize our reference implementation with our
+StreamIt implementation, not have dependencies on extra libraries, and
+generally be a fairer comparison with the StreamIt code (that is,
+equivalent to our other benchmarks).
+
+This version gets consistent output with the StreamIt version when
+compiled with RANDOM_INPUTS and RANDOM_WEIGHTS.  The algorithm should
+be equivalent to the SERIALIZED implementation.
+
+FLAGS: the StreamIt TemplateBeamFormer can be built with COARSE
+defined or not, with SERIALIZED defined or not, and with RANDOM_INPUTS
+defined or not.  Defining COARSE changes the algorithm to add two
+decimation stages; this presently isn't implemented.  SERIALIZED
+changes where in the graph printing happens; in the non-SERIALIZED
+version printing happens in the detector stage, but this doesn't have
+a deterministic order.  RANDOM_INPUTS and RANDOM_WEIGHTS change
+weights at several points in the code, where "RANDOM" is "something
+the author made up" rather than "determined via rand()".  The default
+flags are both RANDOM_WEIGHTS and RANDOM_INPU

[llvm-commits] [test-suite] r46435 [1/4] - /test-suite/trunk/MultiSource/Benchmarks/VersaBench/dbms/

[Chris Lattner]

Author: lattner
Date: Sun Jan 27 23:23:39 2008
New Revision: 46435

URL: http://llvm.org/viewvc/llvm-project?rev=46435&view=rev
Log:
add a VersaBench benchmark.

Added:
test-suite/trunk/MultiSource/Benchmarks/VersaBench/dbms/Makefile
test-suite/trunk/MultiSource/Benchmarks/VersaBench/dbms/calcMetricsData.c
test-suite/trunk/MultiSource/Benchmarks/VersaBench/dbms/chooseEntry.c
test-suite/trunk/MultiSource/Benchmarks/VersaBench/dbms/clearLine.c
test-suite/trunk/MultiSource/Benchmarks/VersaBench/dbms/closeFiles.c
test-suite/trunk/MultiSource/Benchmarks/VersaBench/dbms/consistent.c
test-suite/trunk/MultiSource/Benchmarks/VersaBench/dbms/createDataObject.c
test-suite/trunk/MultiSource/Benchmarks/VersaBench/dbms/createIndexEntry.c
test-suite/trunk/MultiSource/Benchmarks/VersaBench/dbms/createIndexNode.c
test-suite/trunk/MultiSource/Benchmarks/VersaBench/dbms/dataManagement.h
test-suite/trunk/MultiSource/Benchmarks/VersaBench/dbms/dataObject.h
test-suite/trunk/MultiSource/Benchmarks/VersaBench/dbms/dbms.ref.in
test-suite/trunk/MultiSource/Benchmarks/VersaBench/dbms/dbms.train.in
test-suite/trunk/MultiSource/Benchmarks/VersaBench/dbms/delete.c
test-suite/trunk/MultiSource/Benchmarks/VersaBench/dbms/delete.h
test-suite/trunk/MultiSource/Benchmarks/VersaBench/dbms/deleteDataObject.c
test-suite/trunk/MultiSource/Benchmarks/VersaBench/dbms/deleteEntry.c
test-suite/trunk/MultiSource/Benchmarks/VersaBench/dbms/deleteIndexEntry.c
test-suite/trunk/MultiSource/Benchmarks/VersaBench/dbms/deleteIndexNode.c
test-suite/trunk/MultiSource/Benchmarks/VersaBench/dbms/errorMessage.c
test-suite/trunk/MultiSource/Benchmarks/VersaBench/dbms/errorMessage.h
test-suite/trunk/MultiSource/Benchmarks/VersaBench/dbms/getDeleteCommand.c
test-suite/trunk/MultiSource/Benchmarks/VersaBench/dbms/getDeleteCommand.h
test-suite/trunk/MultiSource/Benchmarks/VersaBench/dbms/getFloat.c
test-suite/trunk/MultiSource/Benchmarks/VersaBench/dbms/getFloat.h
test-suite/trunk/MultiSource/Benchmarks/VersaBench/dbms/getInitCommand.c
test-suite/trunk/MultiSource/Benchmarks/VersaBench/dbms/getInitCommand.h
test-suite/trunk/MultiSource/Benchmarks/VersaBench/dbms/getInsertCommand.c
test-suite/trunk/MultiSource/Benchmarks/VersaBench/dbms/getInsertCommand.h
test-suite/trunk/MultiSource/Benchmarks/VersaBench/dbms/getInt.c
test-suite/trunk/MultiSource/Benchmarks/VersaBench/dbms/getInt.h
test-suite/trunk/MultiSource/Benchmarks/VersaBench/dbms/getKeyAttribute.c
test-suite/trunk/MultiSource/Benchmarks/VersaBench/dbms/getKeyAttribute.h
test-suite/trunk/MultiSource/Benchmarks/VersaBench/dbms/getNextCommandCode.c
test-suite/trunk/MultiSource/Benchmarks/VersaBench/dbms/getNextCommandCode.h
test-suite/trunk/MultiSource/Benchmarks/VersaBench/dbms/getNonKeyAttribute.c
test-suite/trunk/MultiSource/Benchmarks/VersaBench/dbms/getNonKeyAttribute.h
test-suite/trunk/MultiSource/Benchmarks/VersaBench/dbms/getQueryCommand.c
test-suite/trunk/MultiSource/Benchmarks/VersaBench/dbms/getQueryCommand.h
test-suite/trunk/MultiSource/Benchmarks/VersaBench/dbms/getString.c
test-suite/trunk/MultiSource/Benchmarks/VersaBench/dbms/index.h
test-suite/trunk/MultiSource/Benchmarks/VersaBench/dbms/indexKey.h
test-suite/trunk/MultiSource/Benchmarks/VersaBench/dbms/initMetricsData.c
test-suite/trunk/MultiSource/Benchmarks/VersaBench/dbms/insert.c
test-suite/trunk/MultiSource/Benchmarks/VersaBench/dbms/insert.h
test-suite/trunk/MultiSource/Benchmarks/VersaBench/dbms/insertEntry.c
test-suite/trunk/MultiSource/Benchmarks/VersaBench/dbms/insertEntry.h
test-suite/trunk/MultiSource/Benchmarks/VersaBench/dbms/keyUnion.c
test-suite/trunk/MultiSource/Benchmarks/VersaBench/dbms/main.c
test-suite/trunk/MultiSource/Benchmarks/VersaBench/dbms/metrics.h
test-suite/trunk/MultiSource/Benchmarks/VersaBench/dbms/openFiles.c
test-suite/trunk/MultiSource/Benchmarks/VersaBench/dbms/openFiles.h
test-suite/trunk/MultiSource/Benchmarks/VersaBench/dbms/outputMetricsData.c
test-suite/trunk/MultiSource/Benchmarks/VersaBench/dbms/outputQuery.c
test-suite/trunk/MultiSource/Benchmarks/VersaBench/dbms/partitionEntries.c
test-suite/trunk/MultiSource/Benchmarks/VersaBench/dbms/penalty.c
test-suite/trunk/MultiSource/Benchmarks/VersaBench/dbms/query.c
test-suite/trunk/MultiSource/Benchmarks/VersaBench/dbms/query.h
test-suite/trunk/MultiSource/Benchmarks/VersaBench/dbms/setMetricsData.c
test-suite/trunk/MultiSource/Benchmarks/VersaBench/dbms/splitNode.c
test-suite/trunk/MultiSource/Benchmarks/VersaBench/dbms/splitNode.h
test-suite/trunk/MultiSource/Benchmarks/VersaBench/dbms/timer.c
test-suite/trunk/MultiSource/Benchmarks/VersaBench/dbms/updateMetricsData.c
test-suite/trunk/MultiSource/Benchmarks/VersaBench/dbms/valid.c
test-suite/trunk/MultiSource/Benchmarks/VersaBench/dbms/volume.c

Ad

[llvm-commits] [test-suite] r46436 - in /test-suite/trunk/MultiSource/Benchmarks/VersaBench/8b10b: ./ Makefile calc.c calc.h input.txt main.c testbench.c testbench.h

[Chris Lattner]

Author: lattner
Date: Sun Jan 27 23:35:03 2008
New Revision: 46436

URL: http://llvm.org/viewvc/llvm-project?rev=46436&view=rev
Log:
add the 8b10b benchmark, hacked to take enough time to run to be interesting.

Added:
test-suite/trunk/MultiSource/Benchmarks/VersaBench/8b10b/
test-suite/trunk/MultiSource/Benchmarks/VersaBench/8b10b/Makefile
test-suite/trunk/MultiSource/Benchmarks/VersaBench/8b10b/calc.c
test-suite/trunk/MultiSource/Benchmarks/VersaBench/8b10b/calc.h
test-suite/trunk/MultiSource/Benchmarks/VersaBench/8b10b/input.txt   (with 
props)
test-suite/trunk/MultiSource/Benchmarks/VersaBench/8b10b/main.c
test-suite/trunk/MultiSource/Benchmarks/VersaBench/8b10b/testbench.c
test-suite/trunk/MultiSource/Benchmarks/VersaBench/8b10b/testbench.h

Added: test-suite/trunk/MultiSource/Benchmarks/VersaBench/8b10b/Makefile
URL: 
http://llvm.org/viewvc/llvm-project/test-suite/trunk/MultiSource/Benchmarks/VersaBench/8b10b/Makefile?rev=46436&view=auto

==
--- test-suite/trunk/MultiSource/Benchmarks/VersaBench/8b10b/Makefile (added)
+++ test-suite/trunk/MultiSource/Benchmarks/VersaBench/8b10b/Makefile Sun Jan 
27 23:35:03 2008
@@ -0,0 +1,10 @@
+LEVEL = ../../../..
+
+PROG = 8b10b
+ifdef SMALL_PROBLEM_SIZE
+RUN_OPTIONS = input.txt 2000
+else
+RUN_OPTIONS = input.txt 2
+endif
+include $(LEVEL)/MultiSource/Makefile.multisrc
+

Added: test-suite/trunk/MultiSource/Benchmarks/VersaBench/8b10b/calc.c
URL: 
http://llvm.org/viewvc/llvm-project/test-suite/trunk/MultiSource/Benchmarks/VersaBench/8b10b/calc.c?rev=46436&view=auto

==
--- test-suite/trunk/MultiSource/Benchmarks/VersaBench/8b10b/calc.c (added)
+++ test-suite/trunk/MultiSource/Benchmarks/VersaBench/8b10b/calc.c Sun Jan 27 
23:35:03 2008
@@ -0,0 +1,261 @@
+/*
+ * Copyright (c) 2002 David Wentzlaff
+ *
+ * Permission  is hereby  granted,  free  of  charge, to  any  person
+ * obtaining a  copy of  this software  and  associated documentation
+ * files   (the  "Software"),  to   deal  in  the   Software  without
+ * restriction, including without  limitation the rights to use, copy,
+ * modify, merge, publish,  distribute, sublicense, and/or sell copies
+ * of  the Software,  and to  permit persons to  whom the  Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above  copyright notice  and this  permission notice  shall be
+ * included in all copies or substantial portions of the Software.
+ *
+ * THE  SOFTWARE IS  PROVIDED "AS IS",  WITHOUT WARRANTY OF  ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING  BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY,   FITNESS   FOR   A   PARTICULARPURPOSEAND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM,  DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF  CONTRACT, TORT OR OTHERWISE, ARISING FROM,  OUT OF OR IN
+ * CONNECTION  WITH THE SOFTWARE OR  THE USE OR OTHER  DEALINGS IN THE
+ * SOFTWARE.  
+ */
+
+//This is part of David Wentzlaff's Masters Thesis
+//This file is Copyright David Wentzlaff 2002 All Rights Reserved.
+//
+// Filename : calc.c
+// Date : 07/31/2002
+
+/* Modified by: Rodric M. Rabbah 06-03-04 */
+
+#include 
+
+#include "calc.h"
+
+//note in all of these things, we assume that the K bit is the
+//high order bit
+
+//also note that every thing here is little endian 
+//and the IBM systems journal paper is big endian.
+//so that means that D.1 is 1 which lives in
+//locattion 16 of the table not 1!
+
+#define ENCODE5B(DATA, FLIP, PLUS, MINUS, X)   
(((DATA)<<4)&0x3e0)|(((DATA)<<1)&2)|FLIP)<<3)|((PLUS)<<2)|((MINUS)<<1)|((X)))<<16)
+#define ENCODE3B(DATA, FLIP, PLUS, MINUS, X)   
((DATA)&1)|(((DATA)<<1)&0x1c)|FLIP)<<3)|((PLUS)<<2)|((MINUS)<<1)|((X)))<<16)
+
+#define GET_FLIP_5(DATA)   (((DATA)>>(16+3))&1)
+#define GET_PLUS_5(DATA)   (((DATA)>>(16+2))&1)
+#define GET_MINUS_5(DATA)  (((DATA)>>(16+1))&1)
+#define GET_X_5(DATA)  (((DATA)>>(16))&1)
+
+#define GET_FLIP_3(DATA)   (((DATA)>>(16+3))&1)
+#define GET_PLUS_3(DATA)   (((DATA)>>(16+2))&1)
+#define GET_MINUS_3(DATA)  (((DATA)>>(16+1))&1)
+#define GET_X_3(DATA)  (((DATA)>>(16))&1)
+
+#define B5_MASK0x3e2
+#define B3_MASK0x1d
+
+//setup the table for the 5b/6b part
+unsigned int lookupTable5B[64] = 
+{
+   /* 0 */ ENCODE5B(0x18,1,1,0,0),
+   /* 1 */ ENCODE5B(0x1b,1,0,1,0),
+   /* 2 */ ENCODE5B(0x06,1,1,0,0),
+   /* 3 */ ENCODE5B(0x0c,1,1,0,0),
+   /* 4 */ ENCODE5B(0x0a,1,1,0,0),
+   /* 5 */ ENCODE5B(0x0b,0,0,0,1),
+   /* 6 */ ENCODE5B(0x0d,0,0,0,1),
+   /* 7 */ ENCODE5B(0x0e,0,0,0,1),
+   /* 8 */ ENCODE5B(0x12,1,1,0,0),
+   /* 9 */ ENCODE5B(0x13,0,0,0,1),
+   /* 10 */ ENCODE5B(0x15,0,0,0,1),
+   /* 11 */ ENCODE5B(0x16,0,0,0,1),
+   /* 12 */ ENCODE5B(0x19,0,0,0,1),
+   /* 13 */ ENCODE5B(0x1a,0,0,0,1),
+   /* 14 */ ENCODE5

[llvm-commits] [test-suite] r46438 - in /test-suite/trunk/MultiSource/Benchmarks/VersaBench/bmm: ./ Makefile bmm.c

[Chris Lattner]

Author: lattner
Date: Mon Jan 28 00:10:47 2008
New Revision: 46438

URL: http://llvm.org/viewvc/llvm-project?rev=46438&view=rev
Log:
add the bmm benchmark, hacked to not produce huge output nor use huge input.

Added:
test-suite/trunk/MultiSource/Benchmarks/VersaBench/bmm/
test-suite/trunk/MultiSource/Benchmarks/VersaBench/bmm/Makefile
test-suite/trunk/MultiSource/Benchmarks/VersaBench/bmm/bmm.c

Added: test-suite/trunk/MultiSource/Benchmarks/VersaBench/bmm/Makefile
URL: 
http://llvm.org/viewvc/llvm-project/test-suite/trunk/MultiSource/Benchmarks/VersaBench/bmm/Makefile?rev=46438&view=auto

==
--- test-suite/trunk/MultiSource/Benchmarks/VersaBench/bmm/Makefile (added)
+++ test-suite/trunk/MultiSource/Benchmarks/VersaBench/bmm/Makefile Mon Jan 28 
00:10:47 2008
@@ -0,0 +1,14 @@
+LEVEL = ../../../..
+
+PROG = bmm
+ifdef LARGE_PROBLEM_SIZE
+RUN_OPTIONS = 1024 1024
+else
+ifdef SMALL_PROBLEM_SIZE
+RUN_OPTIONS = 128 32
+else
+RUN_OPTIONS = 1024 64
+endif
+endif
+include $(LEVEL)/MultiSource/Makefile.multisrc
+

Added: test-suite/trunk/MultiSource/Benchmarks/VersaBench/bmm/bmm.c
URL: 
http://llvm.org/viewvc/llvm-project/test-suite/trunk/MultiSource/Benchmarks/VersaBench/bmm/bmm.c?rev=46438&view=auto

==
--- test-suite/trunk/MultiSource/Benchmarks/VersaBench/bmm/bmm.c (added)
+++ test-suite/trunk/MultiSource/Benchmarks/VersaBench/bmm/bmm.c Mon Jan 28 
00:10:47 2008
@@ -0,0 +1,246 @@
+/**
+
+SOFTWARE LICENSE AGREEMENT NOTICE
+   ---
+
+IT IS  A BREACH OF THIS  LICENSE AGREEMENT TO REMOVE  THIS NOTICE FROM
+THE  FILE  OR SOFTWARE,  OR  ANY MODIFIED  VERSIONS  OF  THIS FILE  OR
+SOFTWARE   OR   DERIVATIVE  WORKS.
+___
+
+Copyright Notices/Identification  of Licensor(s) of  Original Software
+in the File
+
+Copyright (C) 1994 Hewlett-Packard Company
+
+All rights reserved by the foregoing, respectively.
+___
+
+Copyright Notices/Identification of Subsequent
+Licensor(s)/Contributors of Derivative Works
+
+Copyright  
+
+
+All rights reserved by the foregoing, respectively.
+___
+
+The code contained in this  file, including both binary and source [if
+released  by  the  owner(s)]   (hereafter,  Software)  is  subject  to
+copyright  by the  respective Licensor(s)  and ownership  remains with
+such  Licensor(s).  The  Licensor(s) of  the original  Software remain
+free  to  license  their  respective proprietary  Software  for  other
+purposes  that are independent  and separate  from this  file, without
+obligation to any party.
+
+Licensor(s) grant(s)  you (hereafter, Licensee)  a license to  use the
+Software for  academic, research and internal  business purposes only,
+without a  fee.  "Internal business purposes" means  that Licensee may
+install, use and execute the Software for the purpose of designing and
+evaluating  products.   Licensee  may  submit proposals  for  research
+support, and receive funding  from private and Government sponsors for
+continued development, support and maintenance of the Software for the
+purposes permitted herein.
+
+Licensee may also disclose results obtained by executing the Software,
+as well as algorithms embodied therein.  Licensee may redistribute the
+Software to third parties provided that the copyright notices and this
+License Agreement  Notice statement are  reproduced on all  copies and
+that  no charge is  associated with  such copies.  No patent  or other
+intellectual property license is granted or implied by this Agreement,
+and this  Agreement does not  license any acts except  those expressly
+recited.
+
+Licensee may modify the Software  to make derivative works (as defined
+in Section 101 of Title  17, U.S. Code) (hereafter, Derivative Works),
+as  necessary for  its own  academic, research  and  internal business
+purposes.   Title  to  copyrights  and  other  proprietary  rights  in
+Derivative  Works  created by  Licensee  shall  be  owned by  Licensee
+subject,  however,  to the  underlying  ownership  interest(s) of  the
+Licensor(s)  in the  copyrights and  other proprietary  rights  in the
+original Software.   All the same  rights and licenses  granted herein
+and  all  other  terms  and  conditions contained  in  this  Agreement
+pertaining to the Software shall continue to apply to any parts of the
+Software  included in  Derivative Works.   Licensee's  Derivative Work
+should clearly notify users that it  is a modified version and not the
+original Software distributed by the Licensor(s).
+
+If  Licensee wants  to make  its Derivative  Works 

[llvm-commits] [test-suite] r46439 - in /test-suite/trunk/MultiSource/Benchmarks: Makefile VersaBench/Makefile VersaBench/README.txt

[Chris Lattner]

Author: lattner
Date: Mon Jan 28 00:14:28 2008
New Revision: 46439

URL: http://llvm.org/viewvc/llvm-project?rev=46439&view=rev
Log:
run versabench by default

Added:
test-suite/trunk/MultiSource/Benchmarks/VersaBench/Makefile
Modified:
test-suite/trunk/MultiSource/Benchmarks/Makefile
test-suite/trunk/MultiSource/Benchmarks/VersaBench/README.txt

Modified: test-suite/trunk/MultiSource/Benchmarks/Makefile
URL: 
http://llvm.org/viewvc/llvm-project/test-suite/trunk/MultiSource/Benchmarks/Makefile?rev=46439&r1=46438&r2=46439&view=diff

==
--- test-suite/trunk/MultiSource/Benchmarks/Makefile (original)
+++ test-suite/trunk/MultiSource/Benchmarks/Makefile Mon Jan 28 00:14:28 2008
@@ -4,6 +4,6 @@
 PARALLEL_DIRS := Fhourstones Fhourstones-3.1 \
  McCat Olden OptimizerEval Ptrdist llubenchmark \
  sim FreeBench MallocBench Prolangs-C Prolangs-C++ SciMark2-C\
-mediabench ASCI_Purple MiBench tramp3d-v4
+mediabench ASCI_Purple MiBench VersaBench tramp3d-v4
 
 include $(LEVEL)/Makefile.programs

Added: test-suite/trunk/MultiSource/Benchmarks/VersaBench/Makefile
URL: 
http://llvm.org/viewvc/llvm-project/test-suite/trunk/MultiSource/Benchmarks/VersaBench/Makefile?rev=46439&view=auto

==
--- test-suite/trunk/MultiSource/Benchmarks/VersaBench/Makefile (added)
+++ test-suite/trunk/MultiSource/Benchmarks/VersaBench/Makefile Mon Jan 28 
00:14:28 2008
@@ -0,0 +1,6 @@
+# MultiSource/VersaBench Makefile:  Build all subdirectories automatically
+
+LEVEL = ../../..
+PARALLEL_DIRS = 8b10b beamformer bmm dbms ecbdes
+
+include $(LEVEL)/Makefile.programs

Modified: test-suite/trunk/MultiSource/Benchmarks/VersaBench/README.txt
URL: 
http://llvm.org/viewvc/llvm-project/test-suite/trunk/MultiSource/Benchmarks/VersaBench/README.txt?rev=46439&r1=46438&r2=46439&view=diff

==
--- test-suite/trunk/MultiSource/Benchmarks/VersaBench/README.txt (original)
+++ test-suite/trunk/MultiSource/Benchmarks/VersaBench/README.txt Mon Jan 28 
00:14:28 2008
@@ -1,2 +1,5 @@
 Obtained from:
 http://cag.csail.mit.edu/versabench/benchmarks.html
+
+This contains a subset of the programs and some are hacked to run
+long enough.


___
llvm-commits mailing list
llvm-commits@cs.uiuc.edu
http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits

[llvm-commits] [test-suite] r46440 - in /test-suite/trunk/MultiSource/Benchmarks/Trimaran: ./ enc-3des/ enc-3des/Makefile enc-3des/des.c enc-3des/des.h

[Chris Lattner]

Author: lattner
Date: Mon Jan 28 00:27:00 2008
New Revision: 46440

URL: http://llvm.org/viewvc/llvm-project?rev=46440&view=rev
Log:
Add a benchmark from the trimaran collection

Added:
test-suite/trunk/MultiSource/Benchmarks/Trimaran/
test-suite/trunk/MultiSource/Benchmarks/Trimaran/enc-3des/
test-suite/trunk/MultiSource/Benchmarks/Trimaran/enc-3des/Makefile
test-suite/trunk/MultiSource/Benchmarks/Trimaran/enc-3des/des.c
test-suite/trunk/MultiSource/Benchmarks/Trimaran/enc-3des/des.h

Added: test-suite/trunk/MultiSource/Benchmarks/Trimaran/enc-3des/Makefile
URL: 
http://llvm.org/viewvc/llvm-project/test-suite/trunk/MultiSource/Benchmarks/Trimaran/enc-3des/Makefile?rev=46440&view=auto

==
--- test-suite/trunk/MultiSource/Benchmarks/Trimaran/enc-3des/Makefile (added)
+++ test-suite/trunk/MultiSource/Benchmarks/Trimaran/enc-3des/Makefile Mon Jan 
28 00:27:00 2008
@@ -0,0 +1,14 @@
+LEVEL = ../../../..
+
+PROG = enc-3des
+ifdef LARGE_PROBLEM_SIZE
+RUN_OPTIONS = 200
+else
+ifdef SMALL_PROBLEM_SIZE
+RUN_OPTIONS = 10
+else
+RUN_OPTIONS = 100
+endif
+endif
+include $(LEVEL)/MultiSource/Makefile.multisrc
+

Added: test-suite/trunk/MultiSource/Benchmarks/Trimaran/enc-3des/des.c
URL: 
http://llvm.org/viewvc/llvm-project/test-suite/trunk/MultiSource/Benchmarks/Trimaran/enc-3des/des.c?rev=46440&view=auto

==
--- test-suite/trunk/MultiSource/Benchmarks/Trimaran/enc-3des/des.c (added)
+++ test-suite/trunk/MultiSource/Benchmarks/Trimaran/enc-3des/des.c Mon Jan 28 
00:27:00 2008
@@ -0,0 +1,606 @@
+/*
+ *  FIPS-46-3 compliant 3DES implementation
+ *
+ *  Copyright (C) 2001-2003  Christophe Devine
+ *
+ *  This program is free software; you can redistribute it and/or modify
+ *  it under the terms of the GNU General Public License as published by
+ *  the Free Software Foundation; either version 2 of the License, or
+ *  (at your option) any later version.
+ *
+ *  This program is distributed in the hope that it will be useful,
+ *  but WITHOUT ANY WARRANTY; without even the implied warranty of
+ *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ *  GNU General Public License for more details.
+ *
+ *  You should have received a copy of the GNU General Public License
+ *  along with this program; if not, write to the Free Software
+ *  Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+ */
+
+#include "des.h"
+
+/* the eight DES S-boxes */
+
+uint32 SB1[64] =
+{
+0x01010400, 0x, 0x0001, 0x01010404,
+0x01010004, 0x00010404, 0x0004, 0x0001,
+0x0400, 0x01010400, 0x01010404, 0x0400,
+0x01000404, 0x01010004, 0x0100, 0x0004,
+0x0404, 0x01000400, 0x01000400, 0x00010400,
+0x00010400, 0x0101, 0x0101, 0x01000404,
+0x00010004, 0x0104, 0x0104, 0x00010004,
+0x, 0x0404, 0x00010404, 0x0100,
+0x0001, 0x01010404, 0x0004, 0x0101,
+0x01010400, 0x0100, 0x0100, 0x0400,
+0x01010004, 0x0001, 0x00010400, 0x0104,
+0x0400, 0x0004, 0x01000404, 0x00010404,
+0x01010404, 0x00010004, 0x0101, 0x01000404,
+0x0104, 0x0404, 0x00010404, 0x01010400,
+0x0404, 0x01000400, 0x01000400, 0x,
+0x00010004, 0x00010400, 0x, 0x01010004
+};
+
+static uint32 SB2[64] =
+{
+0x80108020, 0x80008000, 0x8000, 0x00108020,
+0x0010, 0x0020, 0x80100020, 0x80008020,
+0x8020, 0x80108020, 0x80108000, 0x8000,
+0x80008000, 0x0010, 0x0020, 0x80100020,
+0x00108000, 0x00100020, 0x80008020, 0x,
+0x8000, 0x8000, 0x00108020, 0x8010,
+0x00100020, 0x8020, 0x, 0x00108000,
+0x8020, 0x80108000, 0x8010, 0x8020,
+0x, 0x00108020, 0x80100020, 0x0010,
+0x80008020, 0x8010, 0x80108000, 0x8000,
+0x8010, 0x80008000, 0x0020, 0x80108020,
+0x00108020, 0x0020, 0x8000, 0x8000,
+0x8020, 0x80108000, 0x0010, 0x8020,
+0x00100020, 0x80008020, 0x8020, 0x00100020,
+0x00108000, 0x, 0x80008000, 0x8020,
+0x8000, 0x80100020, 0x80108020, 0x00108000
+};
+
+static uint32 SB3[64] =
+{
+0x0208, 0x08020200, 0x, 0x08020008,
+0x08000200, 0x, 0x00020208, 0x08000200,
+0x00020008, 0x0808, 0x0808, 0x0002,
+0x08020208, 0x00020008, 0x0802, 0x0208,
+0x0800, 0x0008, 0x08020200, 0x0200,
+0x00020200, 0x0802, 0x08020008, 0x00020208,
+0x08000208, 0x00020200, 0x0002, 0x08000208,
+0x0008, 0x08020208, 0x0200, 0x0800,
+0x08020200, 0x0800, 0x00020008, 0x0208,
+0x0002, 0x08020200, 0x08000200, 0x,
+0x0200, 0x00020008, 0x08020208, 0x08000200,
+0x0808, 0x0200, 0x, 0x08020008,

[llvm-commits] [test-suite] r46441 - in /test-suite/trunk/MultiSource/Benchmarks/Trimaran/enc-md5: ./ Makefile md5.c md5.h

[Chris Lattner]

Author: lattner
Date: Mon Jan 28 00:36:10 2008
New Revision: 46441

URL: http://llvm.org/viewvc/llvm-project?rev=46441&view=rev
Log:
add an adapted version of md5

Added:
test-suite/trunk/MultiSource/Benchmarks/Trimaran/enc-md5/
test-suite/trunk/MultiSource/Benchmarks/Trimaran/enc-md5/Makefile
test-suite/trunk/MultiSource/Benchmarks/Trimaran/enc-md5/md5.c
test-suite/trunk/MultiSource/Benchmarks/Trimaran/enc-md5/md5.h

Added: test-suite/trunk/MultiSource/Benchmarks/Trimaran/enc-md5/Makefile
URL: 
http://llvm.org/viewvc/llvm-project/test-suite/trunk/MultiSource/Benchmarks/Trimaran/enc-md5/Makefile?rev=46441&view=auto

==
--- test-suite/trunk/MultiSource/Benchmarks/Trimaran/enc-md5/Makefile (added)
+++ test-suite/trunk/MultiSource/Benchmarks/Trimaran/enc-md5/Makefile Mon Jan 
28 00:36:10 2008
@@ -0,0 +1,14 @@
+LEVEL = ../../../..
+
+PROG = enc-md5
+ifdef LARGE_PROBLEM_SIZE
+RUN_OPTIONS = 50
+else
+ifdef SMALL_PROBLEM_SIZE
+RUN_OPTIONS = 1
+else
+RUN_OPTIONS = 10
+endif
+endif
+include $(LEVEL)/MultiSource/Makefile.multisrc
+

Added: test-suite/trunk/MultiSource/Benchmarks/Trimaran/enc-md5/md5.c
URL: 
http://llvm.org/viewvc/llvm-project/test-suite/trunk/MultiSource/Benchmarks/Trimaran/enc-md5/md5.c?rev=46441&view=auto

==
--- test-suite/trunk/MultiSource/Benchmarks/Trimaran/enc-md5/md5.c (added)
+++ test-suite/trunk/MultiSource/Benchmarks/Trimaran/enc-md5/md5.c Mon Jan 28 
00:36:10 2008
@@ -0,0 +1,293 @@
+/*
+ *  RFC 1321 compliant MD5 implementation
+ *
+ *  Copyright (C) 2001-2003  Christophe Devine
+ *
+ *  This program is free software; you can redistribute it and/or modify
+ *  it under the terms of the GNU General Public License as published by
+ *  the Free Software Foundation; either version 2 of the License, or
+ *  (at your option) any later version.
+ *
+ *  This program is distributed in the hope that it will be useful,
+ *  but WITHOUT ANY WARRANTY; without even the implied warranty of
+ *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ *  GNU General Public License for more details.
+ *
+ *  You should have received a copy of the GNU General Public License
+ *  along with this program; if not, write to the Free Software
+ *  Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+ */
+
+#include 
+
+#include "md5.h"
+
+#define GET_UINT32(n,b,i)   \
+{   \
+(n) = ( (uint32) (b)[(i)]   )   \
+| ( (uint32) (b)[(i) + 1] <<  8 )   \
+| ( (uint32) (b)[(i) + 2] << 16 )   \
+| ( (uint32) (b)[(i) + 3] << 24 );  \
+}
+
+#define PUT_UINT32(n,b,i)   \
+{   \
+(b)[(i)] = (uint8) ( (n)   );   \
+(b)[(i) + 1] = (uint8) ( (n) >>  8 );   \
+(b)[(i) + 2] = (uint8) ( (n) >> 16 );   \
+(b)[(i) + 3] = (uint8) ( (n) >> 24 );   \
+}
+
+void md5_starts( md5_context *ctx )
+{
+ctx->total[0] = 0;
+ctx->total[1] = 0;
+
+ctx->state[0] = 0x67452301;
+ctx->state[1] = 0xEFCDAB89;
+ctx->state[2] = 0x98BADCFE;
+ctx->state[3] = 0x10325476;
+}
+
+void md5_process( md5_context *ctx, uint8 data[64] )
+{
+uint32 X[16], A, B, C, D;
+
+GET_UINT32( X[0],  data,  0 );
+GET_UINT32( X[1],  data,  4 );
+GET_UINT32( X[2],  data,  8 );
+GET_UINT32( X[3],  data, 12 );
+GET_UINT32( X[4],  data, 16 );
+GET_UINT32( X[5],  data, 20 );
+GET_UINT32( X[6],  data, 24 );
+GET_UINT32( X[7],  data, 28 );
+GET_UINT32( X[8],  data, 32 );
+GET_UINT32( X[9],  data, 36 );
+GET_UINT32( X[10], data, 40 );
+GET_UINT32( X[11], data, 44 );
+GET_UINT32( X[12], data, 48 );
+GET_UINT32( X[13], data, 52 );
+GET_UINT32( X[14], data, 56 );
+GET_UINT32( X[15], data, 60 );
+
+#define S(x,n) ((x << n) | ((x & 0x) >> (32 - n)))
+
+#define P(a,b,c,d,k,s,t)\
+{   \
+a += F(b,c,d) + X[k] + t; a = S(a,s) + b;   \
+}
+
+A = ctx->state[0];
+B = ctx->state[1];
+C = ctx->state[2];
+D = ctx->state[3];
+
+#define F(x,y,z) (z ^ (x & (y ^ z)))
+
+P( A, B, C, D,  0,  7, 0xD76AA478 );
+P( D, A, B, C,  1, 12, 0xE8C7B756 );
+P( C, D, A, B,  2, 17, 0x242070DB );
+P( B, C, D, A,  3, 22, 0xC1BDCEEE );
+P( A, B, C, D,  4,  7, 0xF57C0FAF );
+P( D, A, B, C,  5, 12, 0x4787C62A );
+P( C, D, A, B,  6, 17, 0xA8304613 );
+P( B, C, D, A,  7, 22, 0xFD469501 );
+P( A, B, C, D,  8,  7, 0x698098D8 );
+P( D, A, B, C,  9, 12, 0x8B44F7AF );
+P( C, D, A, B, 10, 17, 0x5BB1 );
+P( B, C, D, A, 11, 22, 0x895CD7BE );
+P( A, B, C, D, 12,  7, 0x6B901122 );
+P( D, A, B, C, 13, 12, 0xFD987193 );
+P( C, D, A, B, 14, 17, 0xA67943

[llvm-commits] [test-suite] r46442 - in /test-suite/trunk/MultiSource/Benchmarks/Trimaran/enc-pc1: ./ Makefile pc1cod.c

[Chris Lattner]

Author: lattner
Date: Mon Jan 28 00:52:21 2008
New Revision: 46442

URL: http://llvm.org/viewvc/llvm-project?rev=46442&view=rev
Log:
add a really scary benchmark

Added:
test-suite/trunk/MultiSource/Benchmarks/Trimaran/enc-pc1/
test-suite/trunk/MultiSource/Benchmarks/Trimaran/enc-pc1/Makefile
test-suite/trunk/MultiSource/Benchmarks/Trimaran/enc-pc1/pc1cod.c

Added: test-suite/trunk/MultiSource/Benchmarks/Trimaran/enc-pc1/Makefile
URL: 
http://llvm.org/viewvc/llvm-project/test-suite/trunk/MultiSource/Benchmarks/Trimaran/enc-pc1/Makefile?rev=46442&view=auto

==
--- test-suite/trunk/MultiSource/Benchmarks/Trimaran/enc-pc1/Makefile (added)
+++ test-suite/trunk/MultiSource/Benchmarks/Trimaran/enc-pc1/Makefile Mon Jan 
28 00:52:21 2008
@@ -0,0 +1,14 @@
+LEVEL = ../../../..
+
+PROG = enc-pc1
+ifdef LARGE_PROBLEM_SIZE
+RUN_OPTIONS = 2000
+else
+ifdef SMALL_PROBLEM_SIZE
+RUN_OPTIONS = 100
+else
+RUN_OPTIONS = 500
+endif
+endif
+include $(LEVEL)/MultiSource/Makefile.multisrc
+

Added: test-suite/trunk/MultiSource/Benchmarks/Trimaran/enc-pc1/pc1cod.c
URL: 
http://llvm.org/viewvc/llvm-project/test-suite/trunk/MultiSource/Benchmarks/Trimaran/enc-pc1/pc1cod.c?rev=46442&view=auto

==
--- test-suite/trunk/MultiSource/Benchmarks/Trimaran/enc-pc1/pc1cod.c (added)
+++ test-suite/trunk/MultiSource/Benchmarks/Trimaran/enc-pc1/pc1cod.c Mon Jan 
28 00:52:21 2008
@@ -0,0 +1,333 @@
+ /* File PC1COD.c */ 
+ /* written in Borland Turbo C 2.0 on PC */ 
+ /* PC1 Cipher Algorithm ( Pukall Cipher 1 ) */ 
+ /* By Alexander PUKALL 1991 */ 
+ /* free code no restriction to use */ 
+ /* please include the name of the Author in the final software */ 
+ /* the Key is 256 bits */ 
+ /* Tested with Turbo C 2.0 for DOS and Microsoft Visual C++ 5.0 for Win 32 */ 
+
+ /* Note that PC1COD256.c is the encryption routine */ 
+ /* PC1DEC256.c is the decryption routine */ 
+ /* Only the K zone change in the two routines */ 
+ /* You can create a single routine with the two parts in it */ 
+
+#include 
+#include 
+unsigned short  ax, bx, cx, dx, si, tmp, x1a2, x1a0[16], res, i,
+inter, cfc, cfd, compte;
+
+unsigned char   cle[32];   /* les variables sont definies de facon
+ * globale */
+
+unsigned char   buff[32];
+
+short   c;
+
+int c1, count;
+
+short   d, e;
+
+FILE * in;
+
+
+assemble() 
+{
+  
+  
+  x1a0[0] = (cle[0] * 256) + cle[1];
+  
+  code();
+  
+  inter = res;
+  
+  
+  x1a0[1] = x1a0[0] ^ ((cle[2] * 256) + cle[3]);
+  
+  code();
+  
+  inter = inter ^ res;
+  
+  
+  x1a0[2] = x1a0[1] ^ ((cle[4] * 256) + cle[5]);
+  
+  code();
+  
+  inter = inter ^ res;
+  
+  
+  x1a0[3] = x1a0[2] ^ ((cle[6] * 256) + cle[7]);
+  
+  code();
+  
+  inter = inter ^ res;
+  
+  
+  x1a0[4] = x1a0[3] ^ ((cle[8] * 256) + cle[9]);
+  
+  code();
+  
+  inter = inter ^ res;
+  
+  
+  x1a0[5] = x1a0[4] ^ ((cle[10] * 256) + cle[11]);
+  
+  code();
+  
+  inter = inter ^ res;
+  
+  
+  x1a0[6] = x1a0[5] ^ ((cle[12] * 256) + cle[13]);
+  
+  code();
+  
+  inter = inter ^ res;
+  
+  
+  x1a0[7] = x1a0[6] ^ ((cle[14] * 256) + cle[15]);
+  
+  code();
+  
+  inter = inter ^ res;
+  
+  
+  x1a0[8] = x1a0[7] ^ ((cle[16] * 256) + cle[17]);
+  
+  code();
+  
+  inter = inter ^ res;
+  
+  
+  x1a0[9] = x1a0[8] ^ ((cle[18] * 256) + cle[19]);
+  
+  code();
+  
+  inter = inter ^ res;
+  
+  
+  x1a0[10] = x1a0[9] ^ ((cle[20] * 256) + cle[21]);
+  
+  code();
+  
+  inter = inter ^ res;
+  
+  
+  x1a0[11] = x1a0[10] ^ ((cle[22] * 256) + cle[23]);
+  
+  code();
+  
+  inter = inter ^ res;
+  
+  
+  x1a0[12] = x1a0[11] ^ ((cle[24] * 256) + cle[25]);
+  
+  code();
+  
+  inter = inter ^ res;
+  
+  
+  x1a0[13] = x1a0[12] ^ ((cle[26] * 256) + cle[27]);
+  
+  code();
+  
+  inter = inter ^ res;
+  
+  
+  x1a0[14] = x1a0[13] ^ ((cle[28] * 256) + cle[29]);
+  
+  code();
+  
+  inter = inter ^ res;
+  
+  
+  x1a0[15] = x1a0[14] ^ ((cle[30] * 256) + cle[31]);
+  
+  code();
+  
+  inter = inter ^ res;
+  
+  
+  i = 0;
+  
+  return (0);
+  
+} 
+
+code() 
+{
+  
+  dx = x1a2 + i;
+  
+  ax = x1a0[i];
+  
+  cx = 0x015a;
+  
+  bx = 0x4e35;
+  
+  
+  tmp = ax;
+  
+  ax = si;
+  
+  si = tmp;
+  
+  
+  tmp = ax;
+  
+  ax = dx;
+  
+  dx = tmp;
+  
+  
+  if (ax != 0)
+
+  {
+
+ax = ax * bx;
+
+  } 
+  
+  tmp = ax;
+  
+  ax = cx;
+  
+  cx = tmp;
+  
+  
+  if (ax != 0)
+
+  {
+
+ax = ax * si;
+
+cx = ax + cx;
+
+  } 
+  
+  tmp = ax;
+  
+  ax = si;
+  
+  si = tmp;
+  
+  ax = ax * bx;
+  
+  dx = cx + dx;
+  
+  
+  ax = ax + 1;
+  
+  
+  x1a2 = dx;
+  
+  x1a0[i] = ax;
+  
+  
+  res = ax ^ dx;
+  
+  i = i + 1;
+  
+  return (0);
+  
+} 
+
+/* Use Linear congruential PRNG */
+int my_rand_r(int *seedp)
+{
+  /* Knuth & Lewis */
+  unsigned x = *seedp * 1664525 + 1013904223;
+  *seedp = x;
+  return (x >> 

[llvm-commits] [test-suite] r46443 - in /test-suite/trunk/MultiSource/Benchmarks/Trimaran/enc-rc4: ./ Makefile rc4.c rc4.h

[Chris Lattner]

Author: lattner
Date: Mon Jan 28 00:56:33 2008
New Revision: 46443

URL: http://llvm.org/viewvc/llvm-project?rev=46443&view=rev
Log:
add a new benchmark.

Added:
test-suite/trunk/MultiSource/Benchmarks/Trimaran/enc-rc4/
test-suite/trunk/MultiSource/Benchmarks/Trimaran/enc-rc4/Makefile
test-suite/trunk/MultiSource/Benchmarks/Trimaran/enc-rc4/rc4.c
test-suite/trunk/MultiSource/Benchmarks/Trimaran/enc-rc4/rc4.h

Added: test-suite/trunk/MultiSource/Benchmarks/Trimaran/enc-rc4/Makefile
URL: 
http://llvm.org/viewvc/llvm-project/test-suite/trunk/MultiSource/Benchmarks/Trimaran/enc-rc4/Makefile?rev=46443&view=auto

==
--- test-suite/trunk/MultiSource/Benchmarks/Trimaran/enc-rc4/Makefile (added)
+++ test-suite/trunk/MultiSource/Benchmarks/Trimaran/enc-rc4/Makefile Mon Jan 
28 00:56:33 2008
@@ -0,0 +1,14 @@
+LEVEL = ../../../..
+
+PROG = enc-rc4
+ifdef LARGE_PROBLEM_SIZE
+RUN_OPTIONS = 100
+else
+ifdef SMALL_PROBLEM_SIZE
+RUN_OPTIONS = 2000
+else
+RUN_OPTIONS = 20
+endif
+endif
+include $(LEVEL)/MultiSource/Makefile.multisrc
+

Added: test-suite/trunk/MultiSource/Benchmarks/Trimaran/enc-rc4/rc4.c
URL: 
http://llvm.org/viewvc/llvm-project/test-suite/trunk/MultiSource/Benchmarks/Trimaran/enc-rc4/rc4.c?rev=46443&view=auto

==
--- test-suite/trunk/MultiSource/Benchmarks/Trimaran/enc-rc4/rc4.c (added)
+++ test-suite/trunk/MultiSource/Benchmarks/Trimaran/enc-rc4/rc4.c Mon Jan 28 
00:56:33 2008
@@ -0,0 +1,153 @@
+/*
+ *  An implementation of the ARC4 algorithm
+ *
+ *  Copyright (C) 2001-2003  Christophe Devine
+ *
+ *  This program is free software; you can redistribute it and/or modify
+ *  it under the terms of the GNU General Public License as published by
+ *  the Free Software Foundation; either version 2 of the License, or
+ *  (at your option) any later version.
+ *
+ *  This program is distributed in the hope that it will be useful,
+ *  but WITHOUT ANY WARRANTY; without even the implied warranty of
+ *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ *  GNU General Public License for more details.
+ *
+ *  You should have received a copy of the GNU General Public License
+ *  along with this program; if not, write to the Free Software
+ *  Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+ */
+
+#include "rc4.h"
+
+void rc4_setup( struct rc4_state *s, unsigned char *key,  int length )
+{
+int i, j, k, *m, a;
+
+s->x = 0;
+s->y = 0;
+m = s->m;
+
+for( i = 0; i < 256; i++ )
+{
+m[i] = i;
+}
+
+j = k = 0;
+
+for( i = 0; i < 256; i++ )
+{
+a = m[i];
+j = (unsigned char) ( j + a + key[k] );
+m[i] = m[j]; m[j] = a;
+if( ++k >= length ) k = 0;
+}
+}
+
+void rc4_crypt( struct rc4_state *s, unsigned char *data, int length )
+{ 
+int i, x, y, *m, a, b;
+
+x = s->x;
+y = s->y;
+m = s->m;
+
+for( i = 0; i < length; i++ )
+{
+x = (unsigned char) ( x + 1 ); a = m[x];
+y = (unsigned char) ( y + a );
+m[x] = b = m[y];
+m[y] = a;
+data[i] ^= m[(unsigned char) ( a + b )];
+}
+
+s->x = x;
+s->y = y;
+}
+
+#include 
+#include 
+
+/*
+ * ARC4 tests vectors from OpenSSL (crypto/rc4/rc4test.c)
+ */
+
+static unsigned char keys[7][30]={
+{8,0x01,0x23,0x45,0x67,0x89,0xab,0xcd,0xef},
+{8,0x01,0x23,0x45,0x67,0x89,0xab,0xcd,0xef},
+{8,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00},
+{4,0xef,0x01,0x23,0x45},
+{8,0x01,0x23,0x45,0x67,0x89,0xab,0xcd,0xef},
+{4,0xef,0x01,0x23,0x45},
+};
+
+static unsigned char data_len[7]={8,8,8,20,28,10};
+static unsigned char data[7][30]={
+{0x01,0x23,0x45,0x67,0x89,0xab,0xcd,0xef,0xff},
+{0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0xff},
+{0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0xff},
+{0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+   0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+   0x00,0x00,0x00,0x00,0xff},
+{0x12,0x34,0x56,0x78,0x9A,0xBC,0xDE,0xF0,
+   0x12,0x34,0x56,0x78,0x9A,0xBC,0xDE,0xF0,
+   0x12,0x34,0x56,0x78,0x9A,0xBC,0xDE,0xF0,
+   0x12,0x34,0x56,0x78,0xff},
+{0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0xff},
+{0},
+};
+
+static unsigned char output[7][30]={
+{0x75,0xb7,0x87,0x80,0x99,0xe0,0xc5,0x96,0x00},
+{0x74,0x94,0xc2,0xe7,0x10,0x4b,0x08,0x79,0x00},
+{0xde,0x18,0x89,0x41,0xa3,0x37,0x5d,0x3a,0x00},
+{0xd6,0xa1,0x41,0xa7,0xec,0x3c,0x38,0xdf,
+ 0xbd,0x61,0x5a,0x11,0x62,0xe1,0xc7,0xba,
+ 0x36,0xb6,0x78,0x58,0x00},
+{0x66,0xa0,0x94,0x9f,0x8a,0xf7,0xd6,0x89,
+ 0x1f,0x7f,0x83,0x2b,0xa8,0x33,0xc0,0x0c,
+ 0x89,0x2e,0xbe,0x30,0x14,0x3c,0xe2,0x87,
+ 0x40,0x01,0x1e,0xcf,0x00},
+{0xd6,0xa1,0x41,0x

[llvm-commits] [test-suite] r46445 - /test-suite/trunk/MultiSource/Benchmarks/VersaBench/beamformer/Makefile

[Nick Lewycky]

Author: nicholas
Date: Mon Jan 28 01:16:22 2008
New Revision: 46445

URL: http://llvm.org/viewvc/llvm-project?rev=46445&view=rev
Log:
Make this test a little more forgiving.

Modified:
test-suite/trunk/MultiSource/Benchmarks/VersaBench/beamformer/Makefile

Modified: test-suite/trunk/MultiSource/Benchmarks/VersaBench/beamformer/Makefile
URL: 
http://llvm.org/viewvc/llvm-project/test-suite/trunk/MultiSource/Benchmarks/VersaBench/beamformer/Makefile?rev=46445&r1=46444&r2=46445&view=diff

==
--- test-suite/trunk/MultiSource/Benchmarks/VersaBench/beamformer/Makefile 
(original)
+++ test-suite/trunk/MultiSource/Benchmarks/VersaBench/beamformer/Makefile Mon 
Jan 28 01:16:22 2008
@@ -1,6 +1,7 @@
 LEVEL = ../../../..
 
 PROG = beamformer
+FP_TOLERANCE = 0.01
 ifdef LARGE_PROBLEM_SIZE
 RUN_OPTIONS = -i 400
 else


___
llvm-commits mailing list
llvm-commits@cs.uiuc.edu
http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits

[llvm-commits] [test-suite] r46447 - in /test-suite/trunk/MultiSource/Benchmarks/Trimaran/netbench-url: packet.c packet.h url.c

[Chris Lattner]

Author: lattner
Date: Mon Jan 28 01:23:41 2008
New Revision: 46447

URL: http://llvm.org/viewvc/llvm-project?rev=46447&view=rev
Log:
make this program not read random memory

Modified:
test-suite/trunk/MultiSource/Benchmarks/Trimaran/netbench-url/packet.c
test-suite/trunk/MultiSource/Benchmarks/Trimaran/netbench-url/packet.h
test-suite/trunk/MultiSource/Benchmarks/Trimaran/netbench-url/url.c

Modified: test-suite/trunk/MultiSource/Benchmarks/Trimaran/netbench-url/packet.c
URL: 
http://llvm.org/viewvc/llvm-project/test-suite/trunk/MultiSource/Benchmarks/Trimaran/netbench-url/packet.c?rev=46447&r1=46446&r2=46447&view=diff

==
--- test-suite/trunk/MultiSource/Benchmarks/Trimaran/netbench-url/packet.c 
(original)
+++ test-suite/trunk/MultiSource/Benchmarks/Trimaran/netbench-url/packet.c Mon 
Jan 28 01:23:41 2008
@@ -22,25 +22,20 @@
 
 /* Traces automatically drop the TCP checksum and urgent pointer, so header is 
36 bytes */
 #define HEADER_SIZE 36
-int packets_processed = 0;
 
 #ifdef CONSTANT_PACKET
 unsigned int packet_index = 0;
 #endif
 
 char *
-get_next_packet ()
+get_next_packet (int packet_number)
 {
+  packet_number %= MAX_PACKETS;
   char *packet;
   unsigned int packet_length;
   
-  if (packets_processed == MAX_PACKETS)
-{
-  fprintf (stderr, "All packets (%d) processed, recompile with larger 
MAX_PACKETS \n", MAX_PACKETS);
-  exit (1);
-}
   
-  packet_length = (packet_lengths[packets_processed]);
+  packet_length = (packet_lengths[packet_number]);
   if (packet_length < 40)
 {
   /* Should never happen */
@@ -61,10 +56,9 @@
 }
 
   /* Copy the header information */
-  memcpy ((void *)packet, (void *)headers[packets_processed], HEADER_SIZE);
+  memcpy ((void *)packet, (void *)headers[packet_number], HEADER_SIZE);
 #endif
 
-  packets_processed ++;
   return packet;
 }
 
@@ -72,7 +66,6 @@
 unsigned int 
 packet_size (unsigned int packet_number)
 {
-  if (packet_number == MAX_PACKETS)
-return (packet_lengths[packets_processed]);
-  else return (packet_lengths[packet_number]);
+  packet_number %= MAX_PACKETS;
+  return (packet_lengths[packet_number]);
 }

Modified: test-suite/trunk/MultiSource/Benchmarks/Trimaran/netbench-url/packet.h
URL: 
http://llvm.org/viewvc/llvm-project/test-suite/trunk/MultiSource/Benchmarks/Trimaran/netbench-url/packet.h?rev=46447&r1=46446&r2=46447&view=diff

==
--- test-suite/trunk/MultiSource/Benchmarks/Trimaran/netbench-url/packet.h 
(original)
+++ test-suite/trunk/MultiSource/Benchmarks/Trimaran/netbench-url/packet.h Mon 
Jan 28 01:23:41 2008
@@ -13,9 +13,9 @@
  * and improve what you give them.
  *
  */
-#define MAX_PACKETS 1500
+#define MAX_PACKETS 100
 
-char *get_next_packet ();
+char *get_next_packet (int);
 
 /* For testing...*/
 unsigned int packet_size (unsigned int);

Modified: test-suite/trunk/MultiSource/Benchmarks/Trimaran/netbench-url/url.c
URL: 
http://llvm.org/viewvc/llvm-project/test-suite/trunk/MultiSource/Benchmarks/Trimaran/netbench-url/url.c?rev=46447&r1=46446&r2=46447&view=diff

==
--- test-suite/trunk/MultiSource/Benchmarks/Trimaran/netbench-url/url.c 
(original)
+++ test-suite/trunk/MultiSource/Benchmarks/Trimaran/netbench-url/url.c Mon Jan 
28 01:23:41 2008
@@ -113,7 +113,7 @@
 {
   long original_checksum, checksum;
   
-  packet = get_next_packet();
+  packet = get_next_packet(i);
   
   /* We only have to check the checksum to make sure nothing is changed.. 
*/
 #ifdef LITTLE_ENDIAN


___
llvm-commits mailing list
llvm-commits@cs.uiuc.edu
http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits

[llvm-commits] [test-suite] r46448 - in /test-suite/trunk/MultiSource/Benchmarks: Makefile Trimaran/Makefile Trimaran/README.txt

[Chris Lattner]

Author: lattner
Date: Mon Jan 28 01:26:07 2008
New Revision: 46448

URL: http://llvm.org/viewvc/llvm-project?rev=46448&view=rev
Log:
Build trimaran benchmarks by default

Added:
test-suite/trunk/MultiSource/Benchmarks/Trimaran/Makefile
test-suite/trunk/MultiSource/Benchmarks/Trimaran/README.txt
Modified:
test-suite/trunk/MultiSource/Benchmarks/Makefile

Modified: test-suite/trunk/MultiSource/Benchmarks/Makefile
URL: 
http://llvm.org/viewvc/llvm-project/test-suite/trunk/MultiSource/Benchmarks/Makefile?rev=46448&r1=46447&r2=46448&view=diff

==
--- test-suite/trunk/MultiSource/Benchmarks/Makefile (original)
+++ test-suite/trunk/MultiSource/Benchmarks/Makefile Mon Jan 28 01:26:07 2008
@@ -4,6 +4,6 @@
 PARALLEL_DIRS := Fhourstones Fhourstones-3.1 \
  McCat Olden OptimizerEval Ptrdist llubenchmark \
  sim FreeBench MallocBench Prolangs-C Prolangs-C++ SciMark2-C\
-mediabench ASCI_Purple MiBench VersaBench tramp3d-v4
+mediabench ASCI_Purple MiBench Trimaran VersaBench tramp3d-v4
 
 include $(LEVEL)/Makefile.programs

Added: test-suite/trunk/MultiSource/Benchmarks/Trimaran/Makefile
URL: 
http://llvm.org/viewvc/llvm-project/test-suite/trunk/MultiSource/Benchmarks/Trimaran/Makefile?rev=46448&view=auto

==
--- test-suite/trunk/MultiSource/Benchmarks/Trimaran/Makefile (added)
+++ test-suite/trunk/MultiSource/Benchmarks/Trimaran/Makefile Mon Jan 28 
01:26:07 2008
@@ -0,0 +1,7 @@
+# MultiSource/Trimaran Makefile:  Build all subdirectories automatically
+
+LEVEL = ../../..
+PARALLEL_DIRS = enc-3des enc-md5 enc-pc1 enc-rc4 netbench-crc netbench-url
+
+
+include $(LEVEL)/Makefile.programs

Added: test-suite/trunk/MultiSource/Benchmarks/Trimaran/README.txt
URL: 
http://llvm.org/viewvc/llvm-project/test-suite/trunk/MultiSource/Benchmarks/Trimaran/README.txt?rev=46448&view=auto

==
--- test-suite/trunk/MultiSource/Benchmarks/Trimaran/README.txt (added)
+++ test-suite/trunk/MultiSource/Benchmarks/Trimaran/README.txt Mon Jan 28 
01:26:07 2008
@@ -0,0 +1,3 @@
+this is obtained from here.  One program was elided (ipchains)
+because it was linux specific.
+http://www.trimaran.org/benchmarks.shtml


___
llvm-commits mailing list
llvm-commits@cs.uiuc.edu
http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits