- Jan 29, 2009
-
-
Evan Cheng authored
A slight compile time optimization. If the caller knows there isn't a free register getReg() should not call getFreeReg(). llvm-svn: 63263
-
Dale Johannesen authored
functions. Currently omitted: memcpy, memmove, memset. llvm-svn: 63259
-
Evan Cheng authored
llvm-svn: 63255
-
Mon P Wang authored
llvm-svn: 63252
-
- Jan 28, 2009
-
-
Bill Wendling authored
llvm-svn: 63247
-
Bill Wendling authored
llvm-svn: 63245
-
Dan Gohman authored
llvm-svn: 63244
-
Dan Gohman authored
accessed via SDNode::getNode. llvm-svn: 63240
-
Dan Gohman authored
llvm-svn: 63239
-
Dale Johannesen authored
classes (those that reasonably have a DebugLoc associated with them). llvm-svn: 63236
-
Dan Gohman authored
llvm-svn: 63234
-
Devang Patel authored
Do not forget to derived type while constructing an array type. llvm-svn: 63233
-
Chris Lattner authored
after the others. llvm-svn: 63227
-
Chris Lattner authored
llvm-svn: 63222
-
Mon P Wang authored
llvm-svn: 63217
-
Dan Gohman authored
llvm-svn: 63214
-
Chris Lattner authored
Patch by Piotr Rak! llvm-svn: 63213
-
Dan Gohman authored
new isOperationLegalOrCustom, which does what isOperationLegal previously did. Update a bunch of callers to use isOperationLegalOrCustom instead of isOperationLegal. In some case it wasn't obvious which behavior is desired; when in doubt I changed then to isOperationLegalOrCustom as that preserves their previous behavior. This is for the second half of PR3376. llvm-svn: 63212
-
Duncan Sands authored
llvm-svn: 63199
-
Duncan Sands authored
llvm-svn: 63198
-
Duncan Sands authored
deleting the escape analysis pass. llvm-svn: 63197
-
Evan Cheng authored
The memory alignment requirement on some of the mov{h|l}p{d|s} patterns are 16-byte. That is overly strict. These instructions read / write f64 memory locations without alignment requirement. llvm-svn: 63195
-
Mon P Wang authored
llvm-svn: 63194
-
Mon P Wang authored
llvm-svn: 63193
-
Mikhail Glushenkov authored
llvm-svn: 63174
-
Mikhail Glushenkov authored
llvm-svn: 63173
-
Mikhail Glushenkov authored
Adds new option properties 'multi_val', 'one_or_more' and 'zero_or_one'. llvm-svn: 63172
-
Mikhail Glushenkov authored
llvm-svn: 63171
-
Dan Gohman authored
llvm-svn: 63170
-
Dan Gohman authored
shift amounts, to avoid implicitly assuming that target architectures will ignore the high bits. llvm-svn: 63169
-
Bill Wendling authored
llvm-svn: 63164
-
Evan Cheng authored
llvm-svn: 63161
-
Dale Johannesen authored
anyone will pay attention. llvm-svn: 63155
-
Devang Patel authored
llvm-svn: 63154
-
Dale Johannesen authored
llvm-svn: 63152
-
Dale Johannesen authored
llvm-svn: 63151
-
Bill Wendling authored
llvm-svn: 63149
-
- Jan 27, 2009
-
-
Anton Korobeynikov authored
mergeable string section. I don't see any bad impact of such decision (rather then placing it into mergeable const section, as it was before), but at least Darwin linker won't complain anymore. The problem in LLVM is that we don't have special type for string constants (like gcc does). Even more, we have two separate types: ConstatArray for non-null strings and ConstantAggregateZero for null stuff.... It's a bit weird :) llvm-svn: 63142
-
Dan Gohman authored
llvm-svn: 63139
-
Dale Johannesen authored
llvm-svn: 63138
-