summaryrefslogtreecommitdiffstats
path: root/llvm/lib
Commit message (Collapse)AuthorAgeFilesLines
* Turn a memcpy from string constant into a series of stores of constant values.Evan Cheng2006-02-151-47/+69
| | | | llvm-svn: 26219
* fix a bunch of alpha regressions. see bug 709Andrew Lenharth2006-02-151-6/+6
| | | | llvm-svn: 26218
* new testChris Lattner2006-02-151-0/+12
| | | | llvm-svn: 26217
* Should not combine ISD::LOCATIONs until we have scheme to remove fromJim Laskey2006-02-152-36/+2
| | | | | | MachineDebugInfo tables. llvm-svn: 26216
* Checking the wrong value. This caused us to emit silly code likeChris Lattner2006-02-151-1/+1
| | | | | | | Y = seteq bool X, true instead of just using X :) llvm-svn: 26215
* Convert this over to work with the new makefilesChris Lattner2006-02-153-181/+2617
| | | | llvm-svn: 26206
* Sparc actually *DOES* have a directive for emitting zeros. In fact, it requiresChris Lattner2006-02-151-1/+1
| | | | | | | | | | | | it, because this: .bss X: .byte 0 results in the assembler warning: "initialization in bss segment". Annoying. llvm-svn: 26204
* random lexer change to test the makefile updating stuffChris Lattner2006-02-153-0/+3
| | | | llvm-svn: 26203
* Fix SingleSource/Regression/C/2004-08-12-InlinerAndAllocas.c on Sparc.Chris Lattner2006-02-151-1/+22
| | | | | | | | The ABI specifies that there is a register save area at the bottom of the stack, which means the actual used pointer needs to be an offset from the subtracted value. llvm-svn: 26202
* Remove an entry.Evan Cheng2006-02-151-5/+0
| | | | llvm-svn: 26197
* Use .zerofill on x86/darwin.Evan Cheng2006-02-151-15/+22
| | | | llvm-svn: 26196
* Lower memcpy with small constant size operand into a series of load / storeEvan Cheng2006-02-151-8/+34
| | | | | | ops. llvm-svn: 26195
* more refactoring, no functionality change.Chris Lattner2006-02-151-12/+11
| | | | llvm-svn: 26194
* cvtsd2ss / cvtss2sd encoding bug.Evan Cheng2006-02-151-4/+4
| | | | llvm-svn: 26193
* movaps, movapd encoding bug.Evan Cheng2006-02-151-8/+8
| | | | llvm-svn: 26192
* pull some code out into a functionChris Lattner2006-02-151-18/+28
| | | | llvm-svn: 26191
* Canonicalize inner loops before outer loops. Inner loop canonicalizationChris Lattner2006-02-141-4/+5
| | | | | | | | can provide work for the outer loop to canonicalize. This fixes a case that breaks unswitching. llvm-svn: 26189
* Doh again!Evan Cheng2006-02-141-1/+1
| | | | llvm-svn: 26188
* When splitting exit edges to canonicalize loops, make sure to put the newChris Lattner2006-02-141-18/+20
| | | | | | | | block in the appropriate loop nest. Third time is the charm, right? llvm-svn: 26187
* new noteChris Lattner2006-02-141-0/+5
| | | | llvm-svn: 26186
* If we have zero initialized data with external linkage, use .zerofill toChris Lattner2006-02-141-4/+11
| | | | | | | | | | | | | emit it (instead of .space), saving a bit of space in the .o file. For example: int foo[100]; int bar[100] = {}; when compiled with C++ or -fno-common results in shrinkage from 1160 to 360 bytes of space. The X86 backend can also do this on darwin. llvm-svn: 26185
* Using wrong DW_FORM.Jim Laskey2006-02-141-1/+1
| | | | llvm-svn: 26184
* Don't special case XS, XD prefixes.Evan Cheng2006-02-141-14/+9
| | | | llvm-svn: 26183
* Bug fix: XS, XD prefixes were being emitted twice.Evan Cheng2006-02-141-9/+9
| | | | | | XMM registers were not being handled. llvm-svn: 26182
* Make sure that weak functions are aligned properlyChris Lattner2006-02-141-2/+1
| | | | llvm-svn: 26181
* DuhEvan Cheng2006-02-141-3/+0
| | | | llvm-svn: 26180
* Remove -disable-x86-sseEvan Cheng2006-02-141-10/+2
| | | | llvm-svn: 26179
* add an assertChris Lattner2006-02-141-1/+2
| | | | llvm-svn: 26178
* Keep to < 80 colsEvan Cheng2006-02-141-6/+6
| | | | llvm-svn: 26177
* Missed a break so memcpy cases fell through to memset. Doh.Evan Cheng2006-02-141-1/+1
| | | | llvm-svn: 26176
* Fixed a build breakage.Evan Cheng2006-02-141-14/+23
| | | | llvm-svn: 26175
* Rename maxStoresPerMemSet to maxStoresPerMemset, etc.Evan Cheng2006-02-143-5/+5
| | | | llvm-svn: 26174
* Add a entry.Evan Cheng2006-02-141-0/+17
| | | | llvm-svn: 26173
* Set maxStoresPerMemSet to 16. Ditto for maxStoresPerMemCpy andEvan Cheng2006-02-141-3/+5
| | | | | | maxStoresPerMemMove. Although the last one is not used. llvm-svn: 26172
* Expand memset dst, c, size to a series of stores if size falls below theEvan Cheng2006-02-141-4/+127
| | | | | | target specific theshold, e.g. 16 for x86. llvm-svn: 26171
* Enable SSE (for the right subtargets)Evan Cheng2006-02-141-6/+3
| | | | llvm-svn: 26169
* now that libcalls don't suck, we can remove this hackChris Lattner2006-02-141-6/+2
| | | | llvm-svn: 26164
* adjust to new style of handling lexer changesChris Lattner2006-02-142-111/+466
| | | | llvm-svn: 26162
* Use statistics to keep track of what flavors of loops we are unswitchingChris Lattner2006-02-141-7/+19
| | | | llvm-svn: 26157
* Fix a latent bug in the call sequence handling stuff. Some targets (e.g. x86)Chris Lattner2006-02-141-4/+11
| | | | | | create these nodes with flag results. Remember that we legalized them. llvm-svn: 26156
* Implement Instcombine/and.ll:test34Chris Lattner2006-02-131-0/+12
| | | | llvm-svn: 26155
* If any of the sign extended bits are demanded, the input sign bit is demandedChris Lattner2006-02-131-3/+9
| | | | | | | | for a sign extension. This fixes InstCombine/2006-02-13-DemandedMiscompile.ll and Ptrdist/bc. llvm-svn: 26152
* Another hack due to allowing multiple symbols with the same name.Chris Lattner2006-02-131-11/+52
| | | | llvm-svn: 26150
* Another work around for the 'symbols with different types can have the sameChris Lattner2006-02-131-0/+4
| | | | | | name' issue. llvm-svn: 26149
* improved zap discoveryAndrew Lenharth2006-02-131-0/+33
| | | | llvm-svn: 26148
* Rename to better reflect usage (current and planned.)Jim Laskey2006-02-132-3/+4
| | | | llvm-svn: 26145
* Completely rewrite libcall insertion by the legalizer, providing theChris Lattner2006-02-131-198/+217
| | | | | | | | | | | | | | | | following handy-dandy properties: 1. it is always correct now 2. it is much faster than before 3. it is easier to understand This implementation builds off of the recent simplifications of the legalizer that made it single-pass instead of iterative. This fixes JM/lencod, JM/ldecod, and CodeGen/Generic/2006-02-12-InsertLibcall.ll (at least on PPC). llvm-svn: 26144
* Switch targets over to using SelectionDAG::getCALLSEQ_START to createChris Lattner2006-02-134-12/+8
| | | | | | CALLSEQ_START nodes. llvm-svn: 26143
* Switch to using getCALLSEQ_START instead of using our own creation callsChris Lattner2006-02-131-4/+4
| | | | llvm-svn: 26142
* Be careful not to request or look at bits shifted in from outside the sizeChris Lattner2006-02-131-3/+9
| | | | | | of the input. This fixes the mediabench/gsm/toast failure last night. llvm-svn: 26138
OpenPOWER on IntegriCloud