- Feb 04, 2012
-
-
Nick Lewycky authored
llvm-svn: 149725
-
Devang Patel authored
llvm-svn: 149724
-
Nick Kledzik authored
llvm-svn: 149723
-
Jakob Stoklund Olesen authored
llvm-svn: 149722
-
Nico Weber authored
Idea by Jean-Daniel Dupas. llvm-svn: 149721
-
Nick Kledzik authored
Remove __attribute__((weak)) on function prototype. It has a different meaning on prototypes then it does on definitions. It is not needed on the prototype and causes build failures for static codegen llvm-svn: 149720
-
Eli Friedman authored
llvm-svn: 149719
-
- Feb 03, 2012
-
-
Eli Friedman authored
Still left: explicit captures in lambdas need to cause implicit capture, and I need to take a look at the diagnostics for some cases. llvm-svn: 149718
-
Jakob Stoklund Olesen authored
llvm-svn: 149717
-
Chad Rosier authored
llvm-svn: 149716
-
Jakob Stoklund Olesen authored
Calls that use register mask operands don't have implicit defs for returned values. The register mask operand handles the call clobber, but it always behaves like a set of dead defs. Add live implicit defs for any implicitly defined physregs that are actually used. llvm-svn: 149715
-
Chad Rosier authored
llvm-svn: 149714
-
Brendon Cahoon authored
When adding the {-1, -1} entry to the DFAStateInputTable, we need to increment the index used to populate the DFAStateEntryTable. Otherwise, the entry table will be off by one for each transition after the {-1, -1} entry. PR11908. llvm-svn: 149713
-
Chad Rosier authored
llvm-svn: 149712
-
Howard Hinnant authored
__gxx_personality_v0 and __cxa_call_unexpected are complete on darwin and I *think* linux (not positive), but still unimplemented on arm. llvm-svn: 149711
-
Johnny Chen authored
llvm-svn: 149710
-
Jakob Stoklund Olesen authored
llvm-svn: 149709
-
Jakob Stoklund Olesen authored
SelectionDAG has 4 different ways of passing physreg defs to users. Collect all of the uses at the same time, and pass all of them to MI->setPhysRegsDeadExcept() to mark the remaining defs dead. The setPhysRegsDeadExcept() function will soon add the required implicit-defs to instructions with register mask operands. llvm-svn: 149708
-
Johnny Chen authored
llvm-svn: 149707
-
Chad Rosier authored
llvm-svn: 149706
-
Andrew Trick authored
llvm-svn: 149705
-
Chad Rosier authored
llvm-svn: 149704
-
Douglas Gregor authored
size. Otherwise, we can end up with bogus layouts. llvm-svn: 149703
-
Howard Hinnant authored
Pushed optimization back up. Crash disappeared with compiler upgrade. Assumed to be due to compiler bug. llvm-svn: 149702
-
Howard Hinnant authored
llvm-svn: 149701
-
Enrico Granata authored
Fixing issues where synthetic children providers for STL containers std::list and std::map where not doing their job properly llvm-svn: 149700
-
Duncan Sands authored
llvm-svn: 149698
-
Douglas Gregor authored
template without a corresponding parameter pack, don't immediately substitute the alias template. This is under discussion in the C++ committee, and may become ill-formed, but for now we match GCC. llvm-svn: 149697
-
Argyrios Kyrtzidis authored
llvm-svn: 149696
-
Hans Wennborg authored
Also, in C, call this a C11 extension rather than a GNU extension. llvm-svn: 149695
-
Nadav Rotem authored
The type-legalizer often scalarizes code. One of the common patterns is extract-and-truncate. In this patch we optimize this pattern and convert the sequence into extract op of a narrow type. This allows the BUILD_VECTOR dag optimizations to construct efficient shuffle operations in many cases. llvm-svn: 149692
-
Alexey Samsonov authored
AddressSanitizer: simplify RoundUpToPowerOfTwo/clz; add a couple of CHECKs for real_X calls. Patch by timurrrr@google.com llvm-svn: 149687
-
Alexey Samsonov authored
AddressSanitizer: Replace __attribute__ with macro (for Win compatibility). Patch by timurrrr@google.com llvm-svn: 149686
-
Douglas Gregor authored
template. Such pack expansions can easily fail at template instantiation time, if the expanded parameter packs are of the wrong length. Fixes <rdar://problem/10040867>, PR9021, and the example that came up today at Going Native. llvm-svn: 149685
-
Greg Clayton authored
lldb.SBValueList now exposes the len() method and also allows item access: lldb.SBValueList[<int>] - where <int> is an integer index into the list, returns a single lldb.SBValue which might be empty if the index is out of range lldb.SBValueList[<str>] - where <str> is the name to look for, returns a list() of lldb.SBValue objects with any matching values (the list might be empty if nothing matches) lldb.SBValueList[<re>] - where <re> is a compiles regular expression, returns a list of lldb.SBValue objects for containing any matches or a empty list if nothing matches lldb.SBFrame now exposes: lldb.SBFrame.variables => SBValueList of all variables that are in scope lldb.SBFrame.vars => see lldb.SBFrame.variables lldb.SBFrame.locals => SBValueList of all variables that are locals in the current frame lldb.SBFrame.arguments => SBValueList of all variables that are arguments in the current frame lldb.SBFrame.args => see lldb.SBFrame.arguments lldb.SBFrame.statics => SBValueList of all static variables lldb.SBFrame.registers => SBValueList of all registers for the current frame lldb.SBFrame.regs => see lldb.SBFrame.registers Combine any of the above properties with the new lldb.SBValueList functionality and now you can do: y = lldb.frame.vars['rect.origin.y'] or vars = lldb.frame.vars for i in range len(vars): print vars[i] Also expose "lldb.SBFrame.var(<str>)" where <str> can be en expression path for any variable or child within the variable. This makes it easier to get a value from the current frame like "rect.origin.y". The resulting value is also not a constant result as expressions will return, but a live value that will continue to track the current value for the variable expression path. lldb.SBValue now exposes: lldb.SBValue.unsigned => unsigned integer for the value lldb.SBValue.signed => a signed integer for the value llvm-svn: 149684
-
Craig Topper authored
llvm-svn: 149683
-
Anna Zaks authored
llvm-svn: 149682
-
Cameron Zwarich authored
disassembler, just like the generic disassembler. llvm-svn: 149681
-
Craig Topper authored
Remove unnecessary qualification on 256-bit vector handling in LowerBUILD_VECTOR. Condition was already guaranteed by earlier code. llvm-svn: 149680
-
rdar://problem/10799325Bob Wilson authored
That llvm change removed the -trap-func backend option, so that using -ftrap-function with clang would cause the backend to complain. Fix it by adding the trap function name to the CodeGenOptions and passing it through to the TargetOptions. llvm-svn: 149679
-