summaryrefslogtreecommitdiffstats
path: root/llvm/lib
Commit message (Collapse)AuthorAgeFilesLines
* Do not add unreachable code to a natural loop!Chris Lattner2003-10-221-1/+4
| | | | llvm-svn: 9377
* Make sure to print labels on nodes without namesChris Lattner2003-10-221-1/+7
| | | | llvm-svn: 9376
* If the basic block has no name, make sure to print the % number of itChris Lattner2003-10-221-1/+7
| | | | llvm-svn: 9375
* Implement the Function::viewCFG* methods, for use in a debugger. Also, theChris Lattner2003-10-221-0/+144
| | | | | | -print-cfg pass now lives here. llvm-svn: 9374
* Add comments to describe what these functions actually do.Misha Brukman2003-10-221-0/+7
| | | | llvm-svn: 9370
* * Use <cmath> instead of <math.h>Misha Brukman2003-10-221-12/+12
| | | | | | * Order #includes according to LLVM coding standards llvm-svn: 9369
* Removed completely duplicated function comment (an identical one appears later).Misha Brukman2003-10-221-19/+0
| | | | llvm-svn: 9368
* No, really, order the #includes correctly.Misha Brukman2003-10-221-1/+1
| | | | llvm-svn: 9367
* Update the 'used' flag correctlyChris Lattner2003-10-221-6/+7
| | | | llvm-svn: 9366
* Loop over the module, not the symbol table. This makes the code handleChris Lattner2003-10-221-21/+18
| | | | | | unused external functions again llvm-svn: 9365
* Implement FunctionResolve/2003-10-21-GlobalResolveHack.llChris Lattner2003-10-221-2/+34
| | | | llvm-svn: 9363
* * Fix order of #includes to follow style guideMisha Brukman2003-10-221-39/+27
| | | | | | | | * It's no longer a BasicBlock pass: update comment on run() method * Fix placement of braces to be consistent * Delete extraneous whitespace llvm-svn: 9361
* Converted tabs to spaces.Misha Brukman2003-10-221-1/+1
| | | | llvm-svn: 9360
* Fix bug: FunctionResolve/2003-10-21-GlobalTypeDifference.llChris Lattner2003-10-211-55/+19
| | | | llvm-svn: 9359
* Fix bug: Linker/2003-10-21-ConflictingTypesTolerance.llChris Lattner2003-10-211-13/+10
| | | | llvm-svn: 9357
* Fix message to make more sense and confuse Chris lessChris Lattner2003-10-211-2/+7
| | | | llvm-svn: 9354
* Make sure to add a type for va_next operands, which may not other wiseChris Lattner2003-10-211-2/+2
| | | | | | occur in the module llvm-svn: 9336
* Fix problem which effected linking apacheChris Lattner2003-10-212-3/+7
| | | | llvm-svn: 9335
* Hrm, unbreak stuph :(Chris Lattner2003-10-212-1/+2
| | | | llvm-svn: 9334
* Fix preselection/lowerswitches bugChris Lattner2003-10-211-0/+3
| | | | llvm-svn: 9333
* Pull the PHI special case into it's own visit* methodChris Lattner2003-10-211-9/+12
| | | | llvm-svn: 9332
* The lastOp operand is never usedChris Lattner2003-10-211-20/+11
| | | | llvm-svn: 9331
* Added LLVM copyright notice.John Criswell2003-10-214-0/+36
| | | | llvm-svn: 9324
* Added LLVM copyright header.John Criswell2003-10-2143-0/+297
| | | | llvm-svn: 9321
* Preselection is _not_ a basicblock pass, because it adds global variables toChris Lattner2003-10-211-101/+22
| | | | | | | | the module. This change converts it from being a basic block pass to being a simple pass. This allows elimination of the annotation and simplification of the logic for moving constants into global variables. llvm-svn: 9320
* When opcodes like ADD were split into reg. and immed. versions (ADDi and ADDr),Vikram S. Adve2003-10-211-1/+2
| | | | | | this code wasn't fixed correctly so it missed copy operations that used ADDi. llvm-svn: 9318
* Why does g++ not even generate a warning when you miss a break statement?Vikram S. Adve2003-10-211-0/+1
| | | | llvm-svn: 9317
* Implement the new varargs instructions and intrinsics.Vikram S. Adve2003-10-211-24/+22
| | | | llvm-svn: 9316
* Added LLVM copyright notice to Makefiles.John Criswell2003-10-2031-0/+248
| | | | llvm-svn: 9312
* Convert this code from using annotations to using a local mapChris Lattner2003-10-203-45/+25
| | | | llvm-svn: 9310
* Goodbye, AddRegNumToValuesBrian Gaeke2003-10-202-10/+0
| | | | llvm-svn: 9309
* Remove using declarationChris Lattner2003-10-201-13/+11
| | | | llvm-svn: 9307
* Added LLVM project notice to the top of every C++ source file.John Criswell2003-10-20209-1/+1463
| | | | | | Header files will be on the way. llvm-svn: 9298
* "Function"[0] == 'F', not 'M'Chris Lattner2003-10-201-8/+8
| | | | llvm-svn: 9297
* Minor leftover fixups from replaceMachineCodeForFunction () change.Brian Gaeke2003-10-202-2/+3
| | | | llvm-svn: 9295
* Make replaceMachineCodeForFunction() return void.Brian Gaeke2003-10-201-1/+1
| | | | llvm-svn: 9289
* Make replaceMachineCodeForFunction return void.Brian Gaeke2003-10-202-4/+2
| | | | llvm-svn: 9288
* Apparently the dependencies are wrong for this file, so it didn't rebuild itChris Lattner2003-10-201-1/+1
| | | | | | when changing Instruction.def. :( llvm-svn: 9286
* Reorder for minor efficiency gainChris Lattner2003-10-201-1/+1
| | | | llvm-svn: 9285
* Emit x86 instructions for: A = B op C, where A and B are 16-bit registers,Chris Lattner2003-10-202-0/+54
| | | | | | | | | | | | | | | | | | | | | | C is a constant which can be sign-extended from 8 bits without value loss, and op is one of: add, sub, imul, and, or, xor. This allows the JIT to emit the one byte version of the constant instead of the two or 4 byte version. Because these instructions are very common, this can save a LOT of code space. For example, I sampled two benchmarks, 176.gcc and 254.gap. BM Old New Reduction 176.gcc 2673621 2548962 4.89% 254.gap 498261 475104 4.87% Note that while the percentage is not spectacular, this did eliminate 124.6 _KILOBYTES_ of codespace from gcc. Not bad. Note that this doesn't effect the llc version at all, because the assembler already does this optimization. llvm-svn: 9284
* Hrm, a relic from the past. How cute :)Chris Lattner2003-10-201-1/+1
| | | | llvm-svn: 9283
* Further cleanups and simplificationsChris Lattner2003-10-201-32/+17
| | | | llvm-svn: 9282
* Eliminate code for pointer size and endianness emulation.Chris Lattner2003-10-202-160/+21
| | | | llvm-svn: 9281
* The Grammar Police was here.Misha Brukman2003-10-201-1/+1
| | | | llvm-svn: 9280
* * Rename X86::IMULr16 -> X86::IMULrr16Chris Lattner2003-10-204-14/+44
| | | | | | | * Implement R1 = R2 * C where R1 and R2 are 32 or 16 bits. This avoids an extra copy into a register, reducing register pressure. llvm-svn: 9278
* Eliminate unused classChris Lattner2003-10-191-5/+0
| | | | llvm-svn: 9270
* Change the Opcode enum for PHI nodes from "Instruction::PHINode" to ↵Chris Lattner2003-10-1912-17/+17
| | | | | | "Instruction::PHI" to be more consistent with the other instructions. llvm-svn: 9269
* * Multiplications by 2^X are turned into shifts. This factors code out of theChris Lattner2003-10-191-95/+201
| | | | | | | | | | | getelementptr code path for use by other code paths (like malloc and alloca). * Optimize comparisons with zero * Generate neg, not, inc, and dec instructions, when possible. This gives some code size wins, which might translate into performance. We'll see tommorow in the nightly tester. llvm-svn: 9267
* Add some new instructions. WheeeChris Lattner2003-10-191-1/+21
| | | | llvm-svn: 9266
* .string adds an implicit zero at the end. This is not what we wanted.Chris Lattner2003-10-191-1/+1
| | | | | | This fixes PR#44. llvm-svn: 9252
OpenPOWER on IntegriCloud