- Oct 27, 2008
-
-
David Greene authored
Fix PR2634. Create new virtual registers from spills early so that we can give it the same stack slot as the spilled interval if it is folded. This prevents the fold/unfold code from pointing to the wrong register. llvm-svn: 58255
-
Steve Naroff authored
llvm-svn: 58253
-
Sebastian Redl authored
llvm-svn: 58252
-
Matthijs Kooijman authored
llvm-svn: 58250
-
Duncan Sands authored
(and a bunch of other node types). While there, I added a doNotCSE predicate and used it to reduce code duplication (some of the duplicated code was wrong...). This fixes ARM/cse-libcalls.ll when using LegalizeTypes. llvm-svn: 58249
-
Zhongxing Xu authored
llvm-svn: 58248
-
Zhongxing Xu authored
llvm-svn: 58247
-
Duncan Sands authored
worklist twice: UpdateNodeOperands could morph a new node into a node already on the worklist. We would then recalculate the NodeId for this existing node and add it to the worklist. The testcase is ARM/cse-libcalls.ll, the problem showing up once UpdateNodeOperands is taught to do CSE for calls. llvm-svn: 58246
-
Zhongxing Xu authored
llvm-svn: 58245
-
Douglas Gregor authored
When destroying a translation unit, deallocate its owned declarations in reverse order, because there may be dependencies among the declarations. llvm-svn: 58244
-
Zhongxing Xu authored
llvm-svn: 58243
-
Steve Naroff authored
This simplifies debug of this particular diagnostic (and removes some code clutter). llvm-svn: 58242
-
rdar://problem/6315646Steve Naroff authored
Fix <rdar://problem/6315646> clang on xcode: error: invalid operands to binary expression ('id<NSTableViewDelegate>' and 'XCExtendedArrayController *'). There is still a bug here (as the FIXME in the test case indicates). Prior to this patch, the bug would generate an error. Now, we simply do nothing (which is less harmful until we can get it right). The complete bug fix will require changing ASTContext::mergeTypes(), which I'd like to defer for now. llvm-svn: 58241
-
Torok Edwin authored
If it is not part of a loop it is obviously invariant wrt to all loops. llvm-svn: 58240
-
Gabor Greif authored
llvm-svn: 58239
-
Torok Edwin authored
that need it to require it by ID. llvm-svn: 58238
-
Bill Wendling authored
llvm-svn: 58237
-
Zhongxing Xu authored
llvm-svn: 58236
-
Zhongxing Xu authored
llvm-svn: 58235
-
Zhongxing Xu authored
llvm-svn: 58234
-
Zhongxing Xu authored
Only loc::MemRegionVal can be modified. This avoids crashing in RegionStore when a function pointer is used as an argument. llvm-svn: 58233
-
Duncan Sands authored
codegen infrastructure, by default. Please report any breakage to the mailing lists. llvm-svn: 58232
-
Nick Lewycky authored
llvm-svn: 58231
-
Evan Cheng authored
For now, don't split live intervals around x87 stack register barriers. FpGET_ST0_80 must be right after a call instruction (and ADJCALLSTACKUP) so we need to find a way to prevent reload of x87 registers between them. llvm-svn: 58230
-
Chris Lattner authored
LargeBlockInfo, we can now dramatically simplify their implementation and speed them up at the same time. Now the code has time proportional to the number of uses of the alloca, not the size of the block. This also eliminates code that tried to batch up different allocas which are used in the same blocks, and eliminates the 'retry list' logic which was baroque and no unneccesary. In addition to being a speedup for crazy cases, this is also a nice cleanup: PromoteMemoryToRegister.cpp | 270 +++++++++++++++----------------------------- 1 file changed, 96 insertions(+), 174 deletions(-) llvm-svn: 58229
-
Chris Lattner authored
llvm-svn: 58228
-
Chris Lattner authored
a trivial dense map. Use this in RewriteSingleStoreAlloca to avoid aggressively rescanning blocks over and over again. This fixes PR2925, speeding up mem2reg on the testcase in that bug from 4.56s to 0.02s in a debug build on my machine. llvm-svn: 58227
-
Chris Lattner authored
llvm-svn: 58226
-
Dale Johannesen authored
150, based on llvm-test measurements. llvm-svn: 58225
-
Chris Lattner authored
llvm-svn: 58224
-
Chris Lattner authored
darwin or AIX abis. This fixes PR2904. llvm-svn: 58222
-
Nick Lewycky authored
Note that -check-exit-code was on by default while -append-exit-code is not. llvm-svn: 58221
-
Chris Lattner authored
t.c:1:13: error: cannot codegen this designators yet int a[10] = {2, 4, [8]=9, 10}; ^~~~~~~~~~~~~~~~~ llvm-svn: 58220
-
Chris Lattner authored
llvm-svn: 58218
-
Chris Lattner authored
support as far as I know. llvm-svn: 58217
-
Chris Lattner authored
objc mode. llvm-svn: 58216
-
Chris Lattner authored
struct foo Y[10] = { [4] .arr [2] 4 // expected-error {{expected '=' or another designator}} }; because the "missing equals" extension only is valid if there is exactly one array designator. llvm-svn: 58215
-
Chris Lattner authored
fix an obscure memory leak. llvm-svn: 58213
-
- Oct 26, 2008
-
-
Chris Lattner authored
improve diagnostic for a malformed field designator. llvm-svn: 58212
-
Chris Lattner authored
easier to understand and hack on, no functionality change. llvm-svn: 58210
-