Dale Johannesen
|
bb811a2445
Add X86 MMX type to bitcode and Type.
|
15 years ago |
Dan Gohman
|
70c2fc0823
Discard metadata produced by LLVM 2.7. The value enumeration it used
|
15 years ago |
Chris Lattner
|
61c70e98ac
remove unions from LLVM IR. They are severely buggy and not
|
15 years ago |
Bill Wendling
|
55ae515f9d
Create the new linker type "linker_private_weak_def_auto".
|
15 years ago |
Chris Lattner
|
2359e5193a
avoid undefined behavior negating minint.
|
15 years ago |
Dan Gohman
|
17aa92c92a
Make NamedMDNode not be a subclass of Value, and simplify the interface
|
15 years ago |
Dan Gohman
|
29d2716766
Add some debug output to help diagnose PR7689.
|
15 years ago |
Dan Gohman
|
872814ae04
Disallow null as a named metadata operand.
|
15 years ago |
Dan Gohman
|
19538d1e9f
Add support for remapping metadata kind IDs when reading in a
|
15 years ago |
Bill Wendling
|
5e721d7682
Implement the "linker_private_weak" linkage type. This will be used for
|
15 years ago |
Bill Wendling
|
07d3177117
Revert r107205 and r107207.
|
15 years ago |
Bill Wendling
|
207855cff9
Introducing the "linker_weak" linkage type. This will be used for Objective-C
|
15 years ago |
Gabor Greif
|
fc91e7d9cb
resort to ArgOperand API
|
15 years ago |
Devang Patel
|
6209869f83
Speedup bitcode writer. Do not walk all values for all functions to emit function local metadata. In one testcase, probably worst case scenario, the 70x speed up is seen.
|
15 years ago |
Dan Gohman
|
a37dd3e6a5
Bitcode support for allocas with arbitrary array size types.
|
15 years ago |
Dan Gohman
|
c2f500aaed
Don't flush the raw_ostream in llvm::WriteBitcodeToFile; it's at
|
15 years ago |
Dan Gohman
|
d928fb670f
Don't special-case stdout in llvm::WriteBitcodeToFile; just consider
|
15 years ago |
Eric Christopher
|
551754c495
Revert 101465, it broke internal OpenGL testing.
|
15 years ago |
Gabor Greif
|
4ec2258ffb
reapply r101434
|
15 years ago |
Gabor Greif
|
607a7ab3da
back out r101423 and r101397, they break llvm-gcc self-host on darwin10
|
15 years ago |
Gabor Greif
|
2ff961f668
reapply r101364, which has been backed out in r101368
|
15 years ago |
Gabor Greif
|
9ee1720811
back out r101364, as it trips the linux nightlybot on some clang C++ tests
|
15 years ago |
Gabor Greif
|
165dac08d1
rotate CallInst operands, i.e. move callee to the back
|
15 years ago |
Chris Lattner
|
a6245247e9
Add special case bitcode support for DebugLoc. This avoids
|
15 years ago |
Gabor Greif
|
c9f7500d17
Finally land the InvokeInst operand reordering.
|
15 years ago |
Gabor Greif
|
1cde4af157
backing out r99170 because it still fails on clang-x86_64-darwin10-fnt
|
15 years ago |
Gabor Greif
|
9b1061e2e3
Now that hopefully all direct accesses to InvokeInst operands are fixed
|
15 years ago |
Gabor Greif
|
cc52ed0c4f
back out r98957, it broke http://smooshlab.apple.com:8010/builders/clang-x86_64-darwin10-fnt/builds/703 in the nightly test suite
|
15 years ago |
Gabor Greif
|
f4f10e3779
Recommit r80858 again (which has been backed out in r80871).
|
15 years ago |
Chris Lattner
|
93b122d3c4
reapply r98656 unmodified, which exposed the asmprinter not
|
15 years ago |