Rewrite the upgrade of GEP indices to be a little less obtuse. This fixes
test/Assembler/2002-08-19-BytecodeReader.llx
test/Assembler/2003-08-21-ConstantExprCast-Fold.llx
test/Assembler/2004-01-11-getelementptrfolding.llx
llvm-svn: 36095
For the short CALL/INVOKE syntax, the signedness of the result type is two
extractions away from the type argument because its a POINTER to function
type, not a function type.
llvm-svn: 35735
Make the sys::Path::getFileStatus function more efficient by having it
return a pointer to the FileStatus structure rather than copy it. Adjust
uses of the function accordingly. Also, fix some memory issues in sys::Path.
llvm-svn: 35476
Make Signedness information pervasive throughout all types and values.
There is no easy way to get around this. Because the GEP instruction can
index through an arbitrarily complex value structure, it is necessary to
keep track of signedness information throughout that structure. This change
makes Signedness a full class, capable of representing Signedness in
arbitrarily shaped types. The class is then used throughout llvm-upgrade to
track signedness and differentiate between globals, locals, and functions
based on their signedness.
For PR1243:
This patch also removes bogus warnings about renaming internal globals. It
now only emits such warnings when renaming non-internal globals because
they may affect linkage.
llvm-svn: 35234
last entry stored in the map could be retrieved for a given integer type.
Propagating the sign information required an invasive change to ensure that
all ValueRef (ValID) instances get the right sign information as well. Also,
put in some assertions to ensure the RenameMap always gives us out the type
that is expected.
This fixes PR1256 and
test/Assembler/2007-03-14-UgpradeLocalSignless.ll
llvm-svn: 35112
reading bytecode.
2. The interpreter can delete the ModuleProvider and replace it with
another so don't depend on it being around after the EE is created.
3. Don't just run llvm_shutdown on exit but actually delete the EE as well.
This cleans up a vast amount of memory (but not all) that EE retained
through exit.
llvm-svn: 34888
Since CompilerDriver.cpp revision 1.44, llvmc must be linked against LLVMCore.a (llvm::Module::~Module) and LLVMBCReader.a
(llvm::getBytecodeModuleProvider).
llvm-svn: 34196
Make llvm-ld more gccld-like by having it run the same set of passes. The
delta was probably due to lack of llvm-ld being maintained. Just another
reason to have only one optimizing linker in in LLVM.
llvm-svn: 34058
Always rename, never give a redef error. We could check for collapsed type
planes and generate an error if that's not the cause, but the 99.9999
percentile case will be that its the result of collapsed type planes. So,
rather than doing an expensive check, just rename.
llvm-svn: 34047
Some changes to get the smbd.ll test case working:
1. Move the logic for CSRETCC->sret attribute out of the ResolveDefinitions
code and into getExistingValue. This resolves it much earlier and works
in function scope as well.
2. Fix handling of CSRETCC->sret for the store instruction.
3. Rewrite the code for handling renaming to factor in linkage types.
4. Rename a structure filed for a PATypeInfo* so it doesn't get confused
with a field for a Type*.
llvm-svn: 34043
Allow @ before identifer names. Recognize the i1, i8, i16, i32, i64 keywords
as type names corresponding to bool, ubyte, ushort, uint, and ulong
respectively. While these aren't LLVM 1.9 constructs, permitting the syntax
allows post-1.9 assembly files to be upgraded.
llvm-svn: 34023
api's look like this:
ModuleProvider *getBytecodeModuleProvider(
const std::string &Filename, ///< Name of file to be read
BCDecompressor_t *BCDC = Compressor::decompressToNewBuffer,
std::string* ErrMsg = 0, ///< Optional error message holder
BytecodeHandler* H = 0 ///< Optional handler for reader events
);
This is ugly, but allows a client to say:
getBytecodeModuleProvider("foo", 0);
If they do this, there is no dependency on the compression libraries, saving
codesize.
llvm-svn: 34012
This patch replaces the SymbolTable class with ValueSymbolTable which does
not support types planes. This means that all symbol names in LLVM must now
be unique. The patch addresses the necessary changes to deal with this and
removes code no longer needed as a result. This completes the bulk of the
changes for this PR. Some cleanup patches will follow.
llvm-svn: 33918
Make llvm-upgrade valgrind clean. Deleting type instances is a really
nasty thing to do to LLVM. This was a hold-over from the re-write.
This fixes test/Assembler/2007-01-22-UpgradeTypeMapInvalidMemory.ll and
potentially many other bugs.
llvm-svn: 33858
1. Allow -- as well as - options (Bill Wendling)
2. Pass unrecognized options to opt and let it handle the errors, if any
(Chris Lattner).
llvm-svn: 33798
Step 1: Copy gccas functionality to opt. This endows opt with a new
-std-compile-opts option to get the set of optimization passes that
gccas used. It also adds -disable-inlining and -disable-opt which
both apply only if -std-compile-opts is given. The -strip-debug option
was also removed. It just makes sure that "-strip" gets done early and
is mostly there for compatibility with gccas. Finally, a new
-verify-each option will cause the verify pass to be run after each pass.
llvm-svn: 33786
This feature is needed in order to support shifts of more than 255 bits
on large integer types. This changes the syntax for llvm assembly to
make shl, ashr and lshr instructions look like a binary operator:
shl i32 %X, 1
instead of
shl i32 %X, i8 1
Additionally, this should help a few passes perform additional optimizations.
llvm-svn: 33776
When an unresolved definition is found, check to see if it is only unresolved
because the csretcc was upgraded to the sret param attribute. Such changes
change the function type and lead to unresolved definitions. In such cases, just
cast the function to the type expected by the CallInst. That is, cast to the
version of the function that has the sret param attribute.
llvm-svn: 33623
1. New parameter attribute called 'inreg'. It has meaning "place this
parameter in registers, if possible". This is some generalization of
gcc's regparm(n) attribute. It's currently used only in X86-32 backend.
2. Completely rewritten CC handling/lowering code inside X86 backend.
Merged stdcall + c CCs and fastcall + fast CC.
3. Dropped CSRET CC. We cannot add struct return variant for each
target-specific CC (e.g. stdcall + csretcc and so on).
4. Instead of CSRET CC introduced 'sret' parameter attribute. Setting in
on first attribute has meaning 'This is hidden pointer to structure
return. Handle it gently'.
5. Fixed small bug in llvm-extract + add new feature to
FunctionExtraction pass, which relinks all internal-linkaged callees
from deleted function to external linkage. This will allow further
linking everything together.
NOTEs: 1. Documentation will be updated soon.
2. llvm-upgrade should be improved to translate csret => sret.
Before this, there will be some unexpected test fails.
llvm-svn: 33597
When a value is found to have the same name as another, try harder to
disambiguate when its a type plane collapse issue and when it isn't. We
traverse the type to see if it contains an integer. If it does not then
we issue the error because it can't be resulting from integer type planes
collapsing. Otherwise we just rename it, even if that's a bit of overkill.
llvm-svn: 33594
PR645
PR761
PR1082
PR1122
Completely rewrite llvm-upgrade. This should be its final design. Any future
changes will use this same design. The changes involve the following:
1. Make this work very much like the 1.9 AsmParser
2. Retain old upgrades dating back to release 1.2 time frame.
3. Merge in some of the upgrades between 1.9 and 2.0 (e.g. icmp/fcmp).
4. Attach a Signedness value (Signless, Unsigned, Signed) to every type,
Value, Constant, Instruction, and list of those things in the Parser.
Use these to make signedness decisions for instruction upgrades.
5. Implement unique name upgrade for function values and global values.
6. Identify rename cases that might cause problems and warn about them.
For example: renaming a global variable with external linkage.
7. Generate a 2.0 IR using VMCore. This is necessary for numerous
reasons and has the advantage that it never goes out of date.
8. Use the AsmPrinter to make the output nice.
9. Clean up error and warning messages from 1.9 form.
llvm-svn: 33531
instcombine to clean up the code, which makes more code suitable for SRoA.
This helps C++ code in particular, e.g. speeding up tramp3d by 31%.
llvm-svn: 33235
Increment the counter after the second use, not after the first use. This
fixes PR1113.
Also, rename some classes for simplicity and to more naturally be
reminscient of LLVM 1.9. This in preparation for additional classes that
will provide a scaled down model of the LLVM 1.9 IR.
llvm-svn: 33226
rename Type::getIntegralTypeMask to Type::getIntegerTypeMask.
This makes naming much more consistent. For example, there are now no longer any
instances of IntegerType that are not considered isInteger! :)
llvm-svn: 33225
like a lot, its really only two changes:
1. Move stuff that should be private to .y out of the .h file.
2. Make all semantic values pointers.
This cleans up the silly destroy methods and changes them to delete. It
also moves the TypeInfo and ValueInfo classes into the .y where we can
modify them more readily because they have no clients any more.
This shouldn't result in any functional changes in llvm-upgrade.
llvm-svn: 33211
Implement upgrading of the varargs intrinsics. These must now have i8*
arguments. This patch bitcasts arguments to i8* and fixes the prototypes.
llvm-svn: 33163
Implement the arbitrary bit-width integer feature. The feature allows
integers of any bitwidth (up to 64) to be defined instead of just 1, 8,
16, 32, and 64 bit integers.
This change does several things:
1. Introduces a new Derived Type, IntegerType, to represent the number of
bits in an integer. The Type classes SubclassData field is used to
store the number of bits. This allows 2^23 bits in an integer type.
2. Removes the five integer Type::TypeID values for the 1, 8, 16, 32 and
64-bit integers. These are replaced with just IntegerType which is not
a primitive any more.
3. Adjust the rest of LLVM to account for this change.
Note that while this incremental change lays the foundation for arbitrary
bit-width integers, LLVM has not yet been converted to actually deal with
them in any significant way. Most optimization passes, for example, will
still only deal with the byte-width integer types. Future increments
will rectify this situation.
llvm-svn: 33113
Reid Spencer