Changes in directory llvm/lib/Target/X86:
X86CodeEmitter.cpp updated: 1.125 -> 1.126
---
Log message:
JIT large code model support.
---
Diffs of the changes: (+13 -7)
X86CodeEmitter.cpp | 20 +---
1 files changed, 13 insertions(+), 7 deletions(-)
Index: llvm/lib/Target/X
Changes in directory llvm/lib/VMCore:
ConstantFolding.cpp updated: 1.108 -> 1.109
---
Log message:
Can't make these fail now with just BitCast. Previous failures must have
been in conjunction with something else. By right, they should just be
BitCasts.
---
Diffs of the changes: (+2 -2)
Co
> Index: llvm/lib/VMCore/ConstantFolding.cpp
> diff -u llvm/lib/VMCore/ConstantFolding.cpp:1.107 llvm/lib/VMCore/
> ConstantFolding.cpp:1.108
> --- llvm/lib/VMCore/ConstantFolding.cpp:1.107 Sun Dec 3 23:19:34 2006
> +++ llvm/lib/VMCore/ConstantFolding.cpp Mon Dec 4 21:30:09 2006
> @@ -777,
Changes in directory llvm/lib/Target/X86:
X86ATTAsmPrinter.cpp updated: 1.76 -> 1.77
---
Log message:
Asm printing bug.
---
Diffs of the changes: (+4 -6)
X86ATTAsmPrinter.cpp | 10 --
1 files changed, 4 insertions(+), 6 deletions(-)
Index: llvm/lib/Target/X86/X86ATTAsmPrinter.cp
Changes in directory llvm-test/MultiSource/Benchmarks/McCat/09-vor:
vor.c updated: 1.1 -> 1.2
---
Log message:
Make this test return 0 reliably.
---
Diffs of the changes: (+2 -2)
vor.c |4 ++--
1 files changed, 2 insertions(+), 2 deletions(-)
Index: llvm-test/MultiSource/Benchmarks/M
Changes in directory llvm-test/MultiSource/Benchmarks/McCat/05-eks:
main.c updated: 1.1 -> 1.2
---
Log message:
Make this program return 0 reliably.
---
Diffs of the changes: (+2 -1)
main.c |3 ++-
1 files changed, 2 insertions(+), 1 deletion(-)
Index: llvm-test/MultiSource/Benchmark
Changes in directory llvm/lib/Target/X86:
X86CodeEmitter.cpp updated: 1.124 -> 1.125
X86ISelDAGToDAG.cpp updated: 1.133 -> 1.134
X86ISelLowering.cpp updated: 1.300 -> 1.301
X86InstrInfo.h updated: 1.60 -> 1.61
X86InstrX86-64.td updated: 1.9 -> 1.10
X86JITInfo.cpp updated: 1.32 -> 1.33
X86Relocat
Changes in directory llvm/lib/Target/X86:
README-X86-64.txt updated: 1.1 -> 1.2
---
Log message:
Update
---
Diffs of the changes: (+10 -0)
README-X86-64.txt | 10 ++
1 files changed, 10 insertions(+)
Index: llvm/lib/Target/X86/README-X86-64.txt
diff -u llvm/lib/Target/X86/README
Changes in directory llvm/lib/VMCore:
ConstantFolding.cpp updated: 1.107 -> 1.108
---
Log message:
Remove the last inferred casts from VMCore.
---
Diffs of the changes: (+7 -10)
ConstantFolding.cpp | 17 +++--
1 files changed, 7 insertions(+), 10 deletions(-)
Index: llvm/li
Changes in directory llvm/lib/VMCore:
Constants.cpp updated: 1.179 -> 1.180
---
Log message:
Implement getPointerCast.
---
Diffs of the changes: (+10 -0)
Constants.cpp | 10 ++
1 files changed, 10 insertions(+)
Index: llvm/lib/VMCore/Constants.cpp
diff -u llvm/lib/VMCore/Consta
Changes in directory llvm/lib/VMCore:
Instructions.cpp updated: 1.51 -> 1.52
---
Log message:
Implement createPointerCast.
---
Diffs of the changes: (+25 -0)
Instructions.cpp | 25 +
1 files changed, 25 insertions(+)
Index: llvm/lib/VMCore/Instructions.cpp
diff
Changes in directory llvm/lib/Transforms/Scalar:
InstructionCombining.cpp updated: 1.558 -> 1.559
---
Log message:
add an instcombine xform. This speeds up 462.libquantum from 9.78s to
7.48s. This regression is due to unforseen consequences of the cast patch.
---
Diffs of the changes: (+17
Changes in directory llvm/test/Regression/Transforms/InstCombine:
shl-trunc.ll added (r1.1)
---
Log message:
new testcase
---
Diffs of the changes: (+7 -0)
shl-trunc.ll |7 +++
1 files changed, 7 insertions(+)
Index: llvm/test/Regression/Transforms/InstCombine/shl-trunc.ll
diff -
Changes in directory llvm/lib/Transforms/Scalar:
SCCP.cpp updated: 1.137 -> 1.138
---
Log message:
SCCP does not handle Packed Type properly. Disable Packed Type handling
for now.
---
Diffs of the changes: (+17 -1)
SCCP.cpp | 18 +-
1 files changed, 17 insertions(+), 1 de
Changes in directory llvm/test/Regression/Transforms/SCCP:
2006-12-04-PackedType.ll added (r1.1)
---
Log message:
SCCP does not handle Packed Type properly. Disable Packed Type handling
for now.
---
Diffs of the changes: (+149 -0)
2006-12-04-PackedType.ll | 149 +++
Changes in directory llvm/include/llvm:
Constants.h updated: 1.101 -> 1.102
InstrTypes.h updated: 1.51 -> 1.52
---
Log message:
Add cast creation functions for Pointer Casts, either BitCast or PtrToInt
---
Diffs of the changes: (+23 -1)
Constants.h |8 +++-
InstrTypes.h | 16 +++
Changes in directory llvm/lib/Target/X86:
X86ISelLowering.cpp updated: 1.299 -> 1.300
---
Log message:
Add support for 64-bit 'r' constraint. Patch by by Rafael ~Avila de EspÃndol!
This fixes PR1029: http://llvm.org/PR1029 and
Regression/CodeGen/X86/x86-64-asm.ll
---
Diffs of the changes:
Changes in directory llvm/test/Regression/CodeGen/X86:
x86-64-asm.ll added (r1.1)
---
Log message:
Testcase for PR1029: http://llvm.org/PR1029 , patch by Rafael vila de
EspÃndola
---
Diffs of the changes: (+15 -0)
x86-64-asm.ll | 15 +++
1 files changed, 15 insertions(+)
Changes in directory llvm-test/External/SPEC:
Makefile.spec updated: 1.59 -> 1.60
---
Log message:
Implement a new feature: output file filtering. A makefile can override
the default filter (cat) to contort the test's output. This is necessary
for things like 183.equake that put out the program
Changes in directory llvm-test/SingleSource/UnitTests:
2006-12-04-DynAllocAndRestore.cpp added (r1.1)
---
Log message:
Make sure that restore stack does the right thing.
---
Diffs of the changes: (+47 -0)
2006-12-04-DynAllocAndRestore.cpp | 47 ++
1 fil
Changes in directory llvm-test/External/SPEC/CFP2000/183.equake:
Makefile updated: 1.5 -> 1.6
---
Log message:
Provide an output file filter for this test because 183.equake prints out
the program name in its messages. The program name has a different suffix
for native, llc, jit, and cbe. The f
On Mon, 2006-12-04 at 10:45 -0800, Chris Lattner wrote:
> On Dec 4, 2006, at 10:40 AM, Reid Spencer wrote:
>
> > On Mon, 2006-12-04 at 10:29 -0800, Chris Lattner wrote:
> >
> >> I thought only 32- and 64-bit integers were allowed?
> >
> > I thought we agreed on any size for sequential type indices
Changes in directory llvm/lib/Target/PowerPC:
PPCISelLowering.cpp updated: 1.239 -> 1.240
---
Log message:
Restoration of the stack pointer after a deallocation of a alloca was not
updating the SP link.
---
Diffs of the changes: (+28 -1)
PPCISelLowering.cpp | 29 +
Changes in directory llvm/lib/VMCore:
Constants.cpp updated: 1.178 -> 1.179
---
Log message:
Separate getCompare from get and getCompareTy from getTy.
---
Diffs of the changes: (+17 -16)
Constants.cpp | 33 +
1 files changed, 17 insertions(+), 16 deletions
Changes in directory llvm/include/llvm:
Constants.h updated: 1.100 -> 1.101
---
Log message:
Separate the get and getTy functions for ICmp and FCmp into getCompare
and getCompareTy. Also, fix some incorrect comments.
---
Diffs of the changes: (+12 -7)
Constants.h | 19 ---
Changes in directory llvm/lib/Analysis:
ScalarEvolution.cpp updated: 1.62 -> 1.63
---
Log message:
Fix comment grammaro
---
Diffs of the changes: (+2 -2)
ScalarEvolution.cpp |4 ++--
1 files changed, 2 insertions(+), 2 deletions(-)
Index: llvm/lib/Analysis/ScalarEvolution.cpp
diff -u
Changes in directory llvm/docs:
LangRef.html updated: 1.176 -> 1.177
---
Log message:
We only support 32-bit or 64-bit sequential type indices for GEP. Document
it as so.
---
Diffs of the changes: (+5 -5)
LangRef.html | 10 +-
1 files changed, 5 insertions(+), 5 deletions(-)
In
Changes in directory llvm/lib/Analysis/IPA:
CallGraph.cpp updated: 1.59 -> 1.60
---
Log message:
Add a comment and fix a memory leak. Thanks to Vikram for pointing this out.
---
Diffs of the changes: (+5 -4)
CallGraph.cpp |9 +
1 files changed, 5 insertions(+), 4 deletions(-)
Done.
On Mon, 2006-12-04 at 11:19 -0800, Chris Lattner wrote:
> >
> > For PR950: http://llvm.org/PR950 : For ICmp and FCmp constant
> > expressions, put the predicate outiside the parentheses to match
> > what llvm-upgrade generates.
>
> Please document the syntax of these constant exprs in:
Implemented. I'll start using them.
Reid.
On Mon, 2006-12-04 at 11:11 -0800, Chris Lattner wrote:
> Reid,
>
> Most of your uses of getInferredCast fall into one of these categories:
>
> 1. They can only be one type of cast.
> 2. They can be a [bitcast or trunc] or [bitcast or zext] or [bitcast
Changes in directory llvm/test/Regression/DebugInfo:
2006-11-30-NoCompileUnit.cpp updated: 1.1 -> 1.2
2006-11-30-Pubnames.cpp updated: 1.2 -> 1.3
---
Log message:
These pass on Linux now.
---
Diffs of the changes: (+4 -6)
2006-11-30-NoCompileUnit.cpp |8 +++-
2006-11-30-Pubnames.cp
Changes in directory llvm-test:
configure updated: 1.41 -> 1.42
---
Log message:
If F95 fails the sanity check then don't set its definition. This works
around false positives if gfortran's f95 is discovered instead of NAG's
---
Diffs of the changes: (+9 -0)
configure |9 +
1 f
Changes in directory llvm-test/autoconf:
configure.ac updated: 1.39 -> 1.40
---
Log message:
If F95 fails the sanity check then don't set its definition. This works
around false positives if gfortran's f95 is discovered instead of NAG's
---
Diffs of the changes: (+5 -0)
configure.ac |5
Changes in directory llvm/autoconf/m4:
sanity_check.m4 updated: 1.2 -> 1.3
---
Log message:
Expose the sanity.
---
Diffs of the changes: (+2 -0)
sanity_check.m4 |2 ++
1 files changed, 2 insertions(+)
Index: llvm/autoconf/m4/sanity_check.m4
diff -u llvm/autoconf/m4/sanity_check.m4:1.
Changes in directory llvm/include/llvm:
InstrTypes.h updated: 1.50 -> 1.51
Constants.h updated: 1.99 -> 1.100
---
Log message:
Implement new cast creation functions for both instructions and constant
expressions. These will get used to reduce clutter as we replace various
calls to createInferre
Changes in directory llvm/test/Regression/CodeGen/Generic:
debug-info.ll updated: 1.2 -> 1.3
---
Log message:
Fix the intrinsic debug call to make this match current definition.
---
Diffs of the changes: (+2 -2)
debug-info.ll |4 ++--
1 files changed, 2 insertions(+), 2 deletions(-)
Changes in directory llvm/include/llvm/Analysis:
ScalarEvolutionExpander.h updated: 1.8 -> 1.9
---
Log message:
Unclutter this by using new cast creation functions.
---
Diffs of the changes: (+2 -8)
ScalarEvolutionExpander.h | 10 ++
1 files changed, 2 insertions(+), 8 deletions(
Changes in directory llvm/lib/VMCore:
Constants.cpp updated: 1.177 -> 1.178
Instructions.cpp updated: 1.50 -> 1.51
---
Log message:
Implement new cast creation functions for both instructions and constant
expressions. These will get used to reduce clutter as we replace various
calls to createIn
Changes in directory llvm/test/Regression/CodeGen/X86:
2006-11-28-Memcpy.ll updated: 1.4 -> 1.5
---
Log message:
Make this test succeed on both Darwin and Linux
---
Diffs of the changes: (+1 -1)
2006-11-28-Memcpy.ll |2 +-
1 files changed, 1 insertion(+), 1 deletion(-)
Index: llvm/te
>
> For PR950: http://llvm.org/PR950 : For ICmp and FCmp constant
> expressions, put the predicate outiside the parentheses to match
> what llvm-upgrade generates.
Please document the syntax of these constant exprs in:
http://llvm.org/docs/LangRef.html#constantexprs
-Chris
___
Changes in directory llvm/docs:
LangRef.html updated: 1.175 -> 1.176
---
Log message:
Document the icmp and fcmp constant expressions.
---
Diffs of the changes: (+10 -4)
LangRef.html | 14 ++
1 files changed, 10 insertions(+), 4 deletions(-)
Index: llvm/docs/LangRef.html
di
Changes in directory llvm-www/releases/1.8:
index.html updated: 1.1 -> 1.2
---
Log message:
Oops. wrong index.html. Fixed now.
---
Diffs of the changes: (+1 -88)
index.html | 89 -
1 files changed, 1 insertion(+), 88 deletions(-
Changes in directory llvm-www/releases/1.9:
index.html added (r1.1)
---
Log message:
Auto redirect
---
Diffs of the changes: (+7 -0)
index.html |7 +++
1 files changed, 7 insertions(+)
Index: llvm-www/releases/1.9/index.html
diff -c /dev/null llvm-www/releases/1.9/index.html:1.1
> Value *visitTruncateExpr(SCEVTruncateExpr *S) {
>Value *V = expand(S->getOperand());
> - return CastInst::createInferredCast(V, S->getType(), "tmp.",
> InsertPt);
> + Instruction::CastOps Opcode = (V->getType()-
> >getPrimitiveSizeInBits() ==
> + S->getType()->g
Changes in directory llvm-www/releases/1.8:
index.html added (r1.1)
---
Log message:
Auto redirect
---
Diffs of the changes: (+94 -0)
index.html | 94 +
1 files changed, 94 insertions(+)
Index: llvm-www/releases/1.8/index.html
Reid,
Most of your uses of getInferredCast fall into one of these categories:
1. They can only be one type of cast.
2. They can be a [bitcast or trunc] or [bitcast or zext] or [bitcast
or sext].
The former should be eliminated obviously, so I want to talk about
the later.
As a specific exa
> --- llvm/lib/VMCore/ConstantFolding.cpp:1.105 Fri Dec 1 13:50:54 2006
> +++ llvm/lib/VMCore/ConstantFolding.cpp Sun Dec 3 20:45:43 2006
> @@ -777,7 +777,8 @@
> uint64_t V =
>DoubleToBits(cast(CP->getOperand(i))-
> >getValue());
> Constant *C = ConstantInt::
Changes in directory llvm/test/Regression/Transforms/LowerSetJmp:
2005-05-05-OldUses.ll (r1.2) removed
---
Log message:
Purpose of test is lost in upgrade.
---
Diffs of the changes: (+0 -0)
0 files changed
___
llvm-commits mailing list
llvm-com
Changes in directory llvm/lib/Target/X86:
X86TargetMachine.cpp updated: 1.128 -> 1.129
---
Log message:
Non-darwin gcc should default to static relocation to match gcc.
---
Diffs of the changes: (+1 -1)
X86TargetMachine.cpp |2 +-
1 files changed, 1 insertion(+), 1 deletion(-)
Index:
On Dec 4, 2006, at 10:40 AM, Reid Spencer wrote:
> On Mon, 2006-12-04 at 10:29 -0800, Chris Lattner wrote:
>
>> I thought only 32- and 64-bit integers were allowed?
>
> I thought we agreed on any size for sequential type indices. See
> PR950.
> I codified the rules after we agreed on them.
The
Changes in directory llvm/lib/VMCore:
Constants.cpp updated: 1.176 -> 1.177
---
Log message:
Reduce the size of the ExprMapKeyType class.
---
Diffs of the changes: (+3 -3)
Constants.cpp |6 +++---
1 files changed, 3 insertions(+), 3 deletions(-)
Index: llvm/lib/VMCore/Constants.cpp
d
Reid, I don't like this series of patches. You're replacing one evil
with another. Please endeavor to figure out what type of cast is
actually needed and put it in, rather than changing one sort of
inferred cast into another.
> Index: llvm/lib/Analysis/ScalarEvolution.cpp
> diff -u llvm/l
Changes in directory llvm/test/Regression/CodeGen/X86:
2006-12-04-X86-64-Bitcast.ll updated: 1.1 -> 1.2
---
Log message:
XFAIL until PR1033: http://llvm.org/PR1033 is fixed.
---
Diffs of the changes: (+1 -0)
2006-12-04-X86-64-Bitcast.ll |1 +
1 files changed, 1 insertion(+)
Index: l
On Mon, 2006-12-04 at 10:29 -0800, Chris Lattner wrote:
> I thought only 32- and 64-bit integers were allowed?
I thought we agreed on any size for sequential type indices. See PR950.
I codified the rules after we agreed on them.
Reid
signature.asc
Description: This is a digitally signed messag
On Dec 4, 2006, at 10:33 AM, Reid Spencer wrote:
> On Mon, 2006-12-04 at 10:22 -0800, Chris Lattner wrote:
>>>// If this is another cast that can be eliminated, it isn't
>>> codegen either.
>>>if (const CastInst *CI = dyn_cast(V))
>>> -if (isEliminableCastPair(CI, CastInst::getCastOpc
> Index: llvm/docs/LangRef.html
> diff -u llvm/docs/LangRef.html:1.174 llvm/docs/LangRef.html:1.175
> --- llvm/docs/LangRef.html:1.174 Mon Nov 27 15:53:28 2006
> +++ llvm/docs/LangRef.htmlSun Dec 3 10:53:48 2006
> @@ -2491,9 +2491,9 @@
> provided depend on the type of the first pointer a
On Mon, 2006-12-04 at 10:22 -0800, Chris Lattner wrote:
> >// If this is another cast that can be eliminated, it isn't
> > codegen either.
> >if (const CastInst *CI = dyn_cast(V))
> > -if (isEliminableCastPair(CI, CastInst::getCastOpcode(V, Ty),
> > Ty, TD))
> > +if (isEliminabl
Changes in directory llvm/test/Regression/CodeGen/X86:
2006-12-04-X86-64-Bitcast.ll added (r1.1)
---
Log message:
Test case for PR1033: http://llvm.org/PR1033 , x86-64 code gen fails on bitcast.
---
Diffs of the changes: (+12 -0)
2006-12-04-X86-64-Bitcast.ll | 12
1 files ch
> Index: llvm/lib/Transforms/Scalar/InstructionCombining.cpp
> diff -u llvm/lib/Transforms/Scalar/InstructionCombining.cpp:1.557
> llvm/lib/Transforms/Scalar/InstructionCombining.cpp:1.558
> --- llvm/lib/Transforms/Scalar/InstructionCombining.cpp:1.557 Fri
> Dec 1 18:13:08 2006
> +++ llvm/lib/
Changes in directory llvm/test/Regression/CodeGen/X86:
x86-64-mem.ll updated: 1.2 -> 1.3
---
Log message:
Fix test.
---
Diffs of the changes: (+8 -8)
x86-64-mem.ll | 16
1 files changed, 8 insertions(+), 8 deletions(-)
Index: llvm/test/Regression/CodeGen/X86/x86-64-mem.
Changes in directory llvm/lib/Analysis:
ScalarEvolutionExpander.cpp updated: 1.6 -> 1.7
---
Log message:
Fix 80 cols violation
---
Diffs of the changes: (+3 -2)
ScalarEvolutionExpander.cpp |5 +++--
1 files changed, 3 insertions(+), 2 deletions(-)
Index: llvm/lib/Analysis/ScalarEvolu
Changes in directory llvm/lib/Analysis:
ScalarEvolution.cpp updated: 1.61 -> 1.62
---
Log message:
Fix inferred casts.
---
Diffs of the changes: (+9 -6)
ScalarEvolution.cpp | 15 +--
1 files changed, 9 insertions(+), 6 deletions(-)
Index: llvm/lib/Analysis/ScalarEvolution.c
Changes in directory llvm/test/Regression/Transforms/ADCE:
2005-02-17-PHI-Invoke-Crash.ll updated: 1.2 -> 1.3
---
Log message:
Remove debug info.
---
Diffs of the changes: (+0 -5)
2005-02-17-PHI-Invoke-Crash.ll |5 -
1 files changed, 5 deletions(-)
Index: llvm/test/Regression/Tran
Changes in directory llvm/test/Regression/Debugger:
.cvsignore (r1.1) removed
dg.exp (r1.3) removed
funccall.c (r1.4) removed
funccall.ll (r1.15) removed
---
Log message:
Moving to DebugInfo.
---
Diffs of the changes: (+0 -0)
0 files changed
__
Changes in directory llvm/test/Regression/DebugInfo:
funccall.c added (r1.1)
funccall.ll added (r1.1)
---
Log message:
Fixed to use current debug info.
---
Diffs of the changes: (+183 -0)
funccall.c | 17 ++
funccall.ll | 166 +
Changes in directory llvm/test/Regression/CodeGen/Generic:
DebugStuff.ll (r1.6) removed
---
Log message:
Duplicate of Regression/Debugger/funccall.ll.
---
Diffs of the changes: (+0 -0)
0 files changed
___
llvm-commits mailing list
llvm-commits@
Changes in directory nightlytest-serverside:
NightlyTestAccept.php updated: 1.60 -> 1.61
---
Log message:
Hack link until log path fixed.
---
Diffs of the changes: (+1 -1)
NightlyTestAccept.php |2 +-
1 files changed, 1 insertion(+), 1 deletion(-)
Index: nightlytest-serverside/Nightly
67 matches
Mail list logo
