Howard Hinnant [Fri, 6 Jan 2012 20:39:47 +0000 (20:39 +0000)]
__gxx_personality_v0 update. This is completely untested code. But my brain is scrambled and I wanted to get it checked in. Code review from anyone who knows anything at all about personality functions would be much appreciated. What is in here is borrowed heavily from llvm/examples/ExceptionDemo/ExceptionDemo.cpp. There are some calls to abort() that should probably be changed to return an error code instead. There may be encodings under readEncodedPointer that need to be implemented. And my handling of type_info is almost a complete guess.
llvm-svn: 147680
Chad Rosier [Fri, 6 Jan 2012 20:11:59 +0000 (20:11 +0000)]
Initializing to false makes better sense. Thanks, David.
llvm-svn: 147679
Chad Rosier [Fri, 6 Jan 2012 20:05:14 +0000 (20:05 +0000)]
Fix uninitialized variable warning.
llvm-svn: 147678
Eli Friedman [Fri, 6 Jan 2012 20:03:09 +0000 (20:03 +0000)]
Revert r147655; it's breaking the compiler_rt build on OSX.
llvm-svn: 147677
Chad Rosier [Fri, 6 Jan 2012 20:02:49 +0000 (20:02 +0000)]
Fix uninitialized variable warning.
llvm-svn: 147676
Chad Rosier [Fri, 6 Jan 2012 19:59:58 +0000 (19:59 +0000)]
Fix uninitialized variable warning.
llvm-svn: 147675
Kostya Serebryany [Fri, 6 Jan 2012 19:44:11 +0000 (19:44 +0000)]
[asan] do not use new/delete for the internal thread structure
llvm-svn: 147674
Eric Christopher [Fri, 6 Jan 2012 19:35:04 +0000 (19:35 +0000)]
Fix a leak I noticed while reviewing the accelerator table changes. Passes
lldb testsuite.
rdar://
10652330
llvm-svn: 147673
DeLesley Hutchins [Fri, 6 Jan 2012 19:16:50 +0000 (19:16 +0000)]
Thread safety analysis: added support for trylock attribute.
llvm-svn: 147672
Kostya Serebryany [Fri, 6 Jan 2012 19:11:09 +0000 (19:11 +0000)]
[asan] move more code into OS-specific files
llvm-svn: 147671
DeLesley Hutchins [Fri, 6 Jan 2012 18:36:09 +0000 (18:36 +0000)]
Added LocalVariableMap
llvm-svn: 147670
Joerg Sonnenberger [Fri, 6 Jan 2012 18:32:26 +0000 (18:32 +0000)]
__FLT_EVAL_METHOD__ should be 1 on NetBSD/i386, since it defaults to
"double" rounding.
llvm-svn: 147669
Sean Callanan [Fri, 6 Jan 2012 18:24:47 +0000 (18:24 +0000)]
Fixed a bug where the DWARF location expression
parser was creating malformed resuls. When the
location of a variable is computed by reading a
register and adding an offset, we shouldn't say
that the variable's value is located in that
register. This was confusing the expression
parser when trying to read a variable captured
by a block.
llvm-svn: 147668
Kostya Serebryany [Fri, 6 Jan 2012 18:09:21 +0000 (18:09 +0000)]
[asan] cleanup: remove the SIGILL-related code (compiler part)
llvm-svn: 147667
Fariborz Jahanian [Fri, 6 Jan 2012 18:07:23 +0000 (18:07 +0000)]
objc++: some declarations related to atomic
properties of c++ object types with non-trivial
assignment copy. Not used yet. // rdar://6137845
llvm-svn: 147666
Kostya Serebryany [Fri, 6 Jan 2012 18:02:04 +0000 (18:02 +0000)]
[asan] cleanup: remove the SIGILL-related code (rt part)
llvm-svn: 147665
Jakub Staszak [Fri, 6 Jan 2012 17:44:30 +0000 (17:44 +0000)]
Silence GCC warnings.
llvm-svn: 147664
Kostya Serebryany [Fri, 6 Jan 2012 17:35:27 +0000 (17:35 +0000)]
[asan] mention -fno-omit-frame-pointer in the docs
llvm-svn: 147663
Douglas Gregor [Fri, 6 Jan 2012 17:19:32 +0000 (17:19 +0000)]
When inferring a submodule ID during module creation, look up the
include stack to find the first file that is known to be part of the
module. This copes with situations where the module map doesn't
completely specify all of the headers that are involved in the module,
which can come up when there are very strange #include_next chains
(e.g., with weird compiler/stdlib headers like stdarg.h or float.h).
llvm-svn: 147662
Anna Zaks [Fri, 6 Jan 2012 17:11:23 +0000 (17:11 +0000)]
[analyzer] Correct the c++-analyzer symlink
As Matt pointed out, this should be just a link to 'ccc-analyzer'.
llvm-svn: 147661
Douglas Gregor [Fri, 6 Jan 2012 16:59:53 +0000 (16:59 +0000)]
Stash Decl's TopLevelDeclInObjCContainer and ModulePrivate bits
into the two unused lower bits of the NextDeclInContext link, dropping
the number of bits in Decl down to 32, and saving 8 bytes per
declaration on x86-64.
llvm-svn: 147660
Richard Smith [Fri, 6 Jan 2012 16:39:00 +0000 (16:39 +0000)]
C++11 generalized constant expressions: implement checking and diagnostics for
pointer-arithmetic-related undefined behavior and unspecified results. We
continue to fold such values, but now notice they aren't constant expressions.
llvm-svn: 147659
Douglas Gregor [Fri, 6 Jan 2012 16:22:39 +0000 (16:22 +0000)]
Introduce a "Hidden" bit into Decl, to track whether that declaration
is hidden from name lookup. The previous hack of tweaking the
ModulePrivate bit when loading a declaration from a hidden submodule
was brittle.
Note that we now have 34 bits in Decl. I'll fix that next.
llvm-svn: 147658
Douglas Gregor [Fri, 6 Jan 2012 16:09:53 +0000 (16:09 +0000)]
Teach DeclContext deserialization to pay attention to endianness, from
David Fang and Takumi Nakamura. Fixes many PCH failures on PowerPC.
llvm-svn: 147657
David Chisnall [Fri, 6 Jan 2012 12:20:19 +0000 (12:20 +0000)]
If we are compiling with -fno-builtin then don't do constant folding of
builtins.
This fixes PR11711.
llvm-svn: 147655
Eli Bendersky [Fri, 6 Jan 2012 07:49:17 +0000 (07:49 +0000)]
Fix typo in string
llvm-svn: 147654
John McCall [Fri, 6 Jan 2012 05:06:35 +0000 (05:06 +0000)]
Fix the mangling of class template arguments in a particular
dependent case. Thanks to Jason Merrill for pointing this out.
llvm-svn: 147653
Rafael Espindola [Fri, 6 Jan 2012 04:54:01 +0000 (04:54 +0000)]
Improvements to the uninitialized variable warning: Check if the constructor
call is elidable or if the constructor is trivial instead of checking if it
is user declared.
llvm-svn: 147652
Eric Christopher [Fri, 6 Jan 2012 04:35:23 +0000 (04:35 +0000)]
As part of the ongoing work in finalizing the accelerator tables, extend
the debug type accelerator tables to contain the tag and a flag
stating whether or not a compound type is a complete type.
rdar://
10652330
llvm-svn: 147651
Eli Friedman [Fri, 6 Jan 2012 03:05:34 +0000 (03:05 +0000)]
More lambda work. Fixes a minor bug Richard pointed out, makes lookup for lambda parameters work correctly, recording more information into the AST.
llvm-svn: 147650
Richard Smith [Fri, 6 Jan 2012 02:30:50 +0000 (02:30 +0000)]
David Blaikie and Chandler would like us to diagnose
int f();
in function scopes under -Wvexing-parse, so now we do.
llvm-svn: 147649
Anna Zaks [Fri, 6 Jan 2012 02:19:07 +0000 (02:19 +0000)]
[analyzer] c++-analyzer should be a symlink.
This fixes a regression from r147643.
llvm-svn: 147648
Kostya Serebryany [Fri, 6 Jan 2012 02:12:25 +0000 (02:12 +0000)]
[asan] move more stuff to OS-specific files
llvm-svn: 147647
Greg Clayton [Fri, 6 Jan 2012 02:01:06 +0000 (02:01 +0000)]
<rdar://problem/
10652336>
Fixed a crasher when trying to load an expression prefix file:
% touch /tmp/carp.txt
% xcrun lldb
(lldb) settings set target.expr-prefix /tmp/carp.txt
Segmentation fault
llvm-svn: 147646
Anna Zaks [Fri, 6 Jan 2012 01:54:08 +0000 (01:54 +0000)]
[analyzer] Update the docs to reflect that gcc is not the only default
compiler option.
llvm-svn: 147645
Anna Zaks [Fri, 6 Jan 2012 01:54:05 +0000 (01:54 +0000)]
[analyzer] scan-build: If we are on MacOS, use clang as the default
compiler.
llvm-svn: 147644
Anna Zaks [Fri, 6 Jan 2012 01:54:02 +0000 (01:54 +0000)]
[analyzer] Skip --serialize-diagnostic when running scan-build.
Otherwise, the analyzer will try to analyze the serialized diagnostic
file as if it were a source file.
llvm-svn: 147643
Richard Smith [Fri, 6 Jan 2012 01:31:20 +0000 (01:31 +0000)]
Tweak to r147599 for PR10828: Move the check from the parser into sema, and use
the Semantic Powers to only warn on class types (or dependent types), where the
constructor or destructor could do something interesting.
llvm-svn: 147642
Eli Friedman [Fri, 6 Jan 2012 01:23:10 +0000 (01:23 +0000)]
Minor refactoring of sentinel warning on blocks. Add a test for this warning.
llvm-svn: 147641
Greg Clayton [Fri, 6 Jan 2012 00:47:38 +0000 (00:47 +0000)]
<rdar://problem/
10649734>
Fixed an issue where the python interpreter could deadlock LLDB.
llvm-svn: 147640
Johnny Chen [Fri, 6 Jan 2012 00:46:12 +0000 (00:46 +0000)]
http://llvm.org/bugs/show_bug.cgi?id=11619
Allow creating SBData values from arrays or primitives in Python
Patch submitted by Enrico Granata.
llvm-svn: 147639
Greg Clayton [Fri, 6 Jan 2012 00:43:54 +0000 (00:43 +0000)]
<rdar://problem/
10647191>
Removed an extra call to close that was causing problems and also
now use the Host::File class to open the file.
llvm-svn: 147638
Johnny Chen [Fri, 6 Jan 2012 00:35:38 +0000 (00:35 +0000)]
http://llvm.org/bugs/show_bug.cgi?id=11618
lldb::SBValue::AddressOf does not work on dereferenced registers in synthetic children provider
Patch submitted by Enrico Granata.
llvm-svn: 147637
Fariborz Jahanian [Fri, 6 Jan 2012 00:29:35 +0000 (00:29 +0000)]
objc++: sythesize a helper function to be used
for copying atomic properties of c++ objects
with non-trivial copy assignment in setters/getters.
Not yet used. // rdar://6137845
llvm-svn: 147636
Argyrios Kyrtzidis [Fri, 6 Jan 2012 00:22:09 +0000 (00:22 +0000)]
Fix a memory leak of PragmaNamespaces, rdar://
10611796.
llvm-svn: 147635
Greg Clayton [Fri, 6 Jan 2012 00:17:16 +0000 (00:17 +0000)]
We finalized on the new .apple_types accelerator table format where we don't
emit fully qualified names, so now we make sure the DW_TAG values match and
still lookup using the basename.
llvm-svn: 147634
Johnny Chen [Fri, 6 Jan 2012 00:05:01 +0000 (00:05 +0000)]
http://llvm.org/bugs/show_bug.cgi?id=11715
comma at end of enumerator list
llvm-svn: 147633
Rafael Espindola [Thu, 5 Jan 2012 23:59:42 +0000 (23:59 +0000)]
Fix build on minix. Patch by Ben Gras!
llvm-svn: 147632
Eli Friedman [Thu, 5 Jan 2012 23:59:40 +0000 (23:59 +0000)]
Address Richard's review comments on r147561 (Evaluate support for address-of-label differences).
llvm-svn: 147631
Dan Gohman [Thu, 5 Jan 2012 23:58:56 +0000 (23:58 +0000)]
Fix SpeculativelyExecuteBB to either speculate all or none of the phis
present in the bottom of the CFG triangle, as the transformation isn't
ever valuable if the branch can't be eliminated.
Also, unify some heuristics between SimplifyCFG's multiple
if-converters, for consistency.
This fixes rdar://
10627242.
llvm-svn: 147630
Johnny Chen [Thu, 5 Jan 2012 23:51:37 +0000 (23:51 +0000)]
Fix indentation.
llvm-svn: 147629
Kostya Serebryany [Thu, 5 Jan 2012 23:50:34 +0000 (23:50 +0000)]
[asan] use dl_iterate_phdr for pre-symbolization on linux instead of parsing /proc/self/maps
llvm-svn: 147628
Johnny Chen [Thu, 5 Jan 2012 23:50:14 +0000 (23:50 +0000)]
Add back the file I wrongly deleted in r147613.
llvm-svn: 147627
Douglas Gregor [Thu, 5 Jan 2012 23:49:36 +0000 (23:49 +0000)]
Always allocate an extra 8 bytes for a deserialized declaration, since
some code in Clang expects 8-byte alignment of declarations.
llvm-svn: 147626
Eli Friedman [Thu, 5 Jan 2012 23:03:32 +0000 (23:03 +0000)]
PR11705, part 2: globalopt shouldn't put inttoptr/ptrtoint operations into global initializers if there's an implied extension or truncation.
llvm-svn: 147625
Rafael Espindola [Thu, 5 Jan 2012 23:02:01 +0000 (23:02 +0000)]
Link symbols with different visibilities according to the rules in the
System V Application Binary Interface. This lets us use
-fvisibility-inlines-hidden with LTO.
Fixes PR11697.
llvm-svn: 147624
Dan Gohman [Thu, 5 Jan 2012 22:54:35 +0000 (22:54 +0000)]
Revert r56315. When the instruction to speculate is a load, this
code can incorrectly move the load across a store. This never
happens in practice today, but only because the current
heuristics accidentally preclude it.
llvm-svn: 147623
Ted Kremenek [Thu, 5 Jan 2012 22:47:47 +0000 (22:47 +0000)]
After further discussion, rename attribute 'objc_disable_automatic_synthesis' to 'objc_requires_property_definitions'.
llvm-svn: 147622
Sean Callanan [Thu, 5 Jan 2012 22:35:40 +0000 (22:35 +0000)]
Fixed a problem in our local Clang's method for
performing Objective-C instance variable lookup.
Previously, it only completed the derived class
that was the beginning of the search. Now, as
it walks up the superclass chain looking for the
ivar, it completes each superclass in turn.
Also added a testcase covering this issue.
llvm-svn: 147621
Eli Friedman [Thu, 5 Jan 2012 22:34:08 +0000 (22:34 +0000)]
Tweak the fix to PR8977: an empty expression-list represents value initialization, not default initialization. Fixes PR11712.
llvm-svn: 147620
Douglas Gregor [Thu, 5 Jan 2012 22:33:30 +0000 (22:33 +0000)]
Don't seed the ASTWriter's declaration -> ID mapping with the IDs of
each deserialized declaration, since that information is already
available in each declaration.
llvm-svn: 147619
Benjamin Kramer [Thu, 5 Jan 2012 22:31:37 +0000 (22:31 +0000)]
Kill ObjectCodeEmitter and BinaryObject, they were unused and superseded by MC.
llvm-svn: 147618
Douglas Gregor [Thu, 5 Jan 2012 22:27:05 +0000 (22:27 +0000)]
When we deserialize a declaration from a module file, allocate extra
storage for the global declaration ID. Declarations that are parsed
(rather than deserialized) are unaffected, so the number of
declarations that pay this cost tends to be relatively small (since
relatively few declarations are ever deserialized).
This replaces a largish DenseMap within the AST reader. It's not
strictly a win in terms of memory use---not every declaration was
added to that DenseMap in the first place---but it's cleaner to have
this information available for every deserialized declaration, so that
future clients can rely on it.
llvm-svn: 147617
Nick Lewycky [Thu, 5 Jan 2012 22:21:45 +0000 (22:21 +0000)]
SCCCaptured is trivially false on entry to this loop and not modified inside it.
Eliminate the dead test for it on each loop iteration. No functionality change.
llvm-svn: 147616
Rafael Espindola [Thu, 5 Jan 2012 22:07:43 +0000 (22:07 +0000)]
Remove the old ELF writer.
llvm-svn: 147615
Douglas Gregor [Thu, 5 Jan 2012 21:55:30 +0000 (21:55 +0000)]
When creating declarations that are deserialized from an module file,
go through a central allocation routine
Decl::AllocateDeserializedDecl(). No actual functionality change (yet).
llvm-svn: 147614
Johnny Chen [Thu, 5 Jan 2012 21:48:15 +0000 (21:48 +0000)]
Fix incomplete commit of llvm.org/viewvc/llvm-project?rev=147609&view=rev:
This patch combines common code from Linux and FreeBSD into
a new POSIX platform. It also contains fixes for 64bit FreeBSD.
The patch is based on changes by Mark Peek <mp@FreeBSD.org> and
"K. Macy" <kmacy@freebsd.org> in their github repo located at
https://github.com/fbsd/lldb.
llvm-svn: 147613
Argyrios Kyrtzidis [Thu, 5 Jan 2012 21:36:25 +0000 (21:36 +0000)]
When loading an AST file, set SourceManager::MainFileID to the main file of the AST file,
as suggested by Tom Honermann.
llvm-svn: 147612
Danil Malyshev [Thu, 5 Jan 2012 21:16:14 +0000 (21:16 +0000)]
A small re-factored JIT/MCJIT::getPointerToNamedFunction(), so it could be called with the base class.
llvm-svn: 147610
Johnny Chen [Thu, 5 Jan 2012 19:17:38 +0000 (19:17 +0000)]
This patch combines common code from Linux and FreeBSD into
a new POSIX platform. It also contains fixes for 64bit FreeBSD.
The patch is based on changes by Mark Peek <mp@FreeBSD.org> and
"K. Macy" <kmacy@freebsd.org> in their github repo located at
https://github.com/fbsd/lldb.
llvm-svn: 147609
Sebastian Pop [Thu, 5 Jan 2012 18:28:46 +0000 (18:28 +0000)]
revert r147542 after comments from Joerg Sonnenberger
llvm-svn: 147608
Chris Lattner [Thu, 5 Jan 2012 17:35:07 +0000 (17:35 +0000)]
allow clients of SmallSet to specify their own comparison function for the set.
Patch by Stepan Dyatkovskiy!
llvm-svn: 147607
Chandler Carruth [Thu, 5 Jan 2012 11:25:47 +0000 (11:25 +0000)]
Remove an unused variable.
llvm-svn: 147605
Chandler Carruth [Thu, 5 Jan 2012 11:05:55 +0000 (11:05 +0000)]
Prevent a DAGCombine from firing where there are two uses of
a combined-away node and the result of the combine isn't substantially
smaller than the input, it's just canonicalized. This is the first part
of a significant (7%) performance gain for Snappy's hot decompression
loop.
llvm-svn: 147604
Chandler Carruth [Thu, 5 Jan 2012 11:05:47 +0000 (11:05 +0000)]
Cleanup and FileCheck-ize a test.
llvm-svn: 147603
Craig Topper [Thu, 5 Jan 2012 08:56:10 +0000 (08:56 +0000)]
Mark scalar FMA4 instructions as ignoring the VEX.L bit.
llvm-svn: 147602
Victor Umansky [Thu, 5 Jan 2012 08:46:19 +0000 (08:46 +0000)]
Peephole optimization of ptest-conditioned branch in X86 arch. Performs instruction combining of sequences generated by ptestz/ptestc intrinsics to ptest+jcc pair for SSE and AVX.
Testing: passed 'make check' including LIT tests for all sequences being handled (both SSE and AVX)
Reviewers: Evan Cheng, David Blaikie, Bruno Lopes, Elena Demikhovsky, Chad Rosier, Anton Korobeynikov
llvm-svn: 147601
Eli Bendersky [Thu, 5 Jan 2012 08:18:41 +0000 (08:18 +0000)]
test commit (verifyiing commit access)
llvm-svn: 147600
Richard Smith [Thu, 5 Jan 2012 04:12:21 +0000 (04:12 +0000)]
PR10828: Produce a warning when a no-arguments function is declared in block
scope, when no other indication is provided that the user intended to declare a
function rather than a variable.
Remove some false positives from the existing 'parentheses disambiguated as a
function' warning by suppressing it when the declaration is marked as 'typedef'
or 'extern'.
Add a new warning group -Wvexing-parse containing both of these warnings.
The new warning is enabled by default; despite a number of false positives (and
one bug) in clang's test-suite, I have only found genuine bugs with it when
running it over a significant quantity of real C++ code.
llvm-svn: 147599
Greg Clayton [Thu, 5 Jan 2012 04:00:49 +0000 (04:00 +0000)]
Bumped Xcode project versions for lldb-101 and debugserver-160.
llvm-svn: 147597
Greg Clayton [Thu, 5 Jan 2012 03:57:59 +0000 (03:57 +0000)]
Added code in the Host layer that can report system log messages
so that we don't have "fprintf (stderr, ...)" calls sprinkled everywhere.
Changed all needed locations over to using this.
For non-darwin, we log to stderr only. On darwin, we log to stderr _and_
to ASL (Apple System Log facility). This will allow GUI apps to have a place
for these error and warning messages to go, and also allows the command line
apps to log directly to the terminal.
llvm-svn: 147596
Eli Friedman [Thu, 5 Jan 2012 03:35:19 +0000 (03:35 +0000)]
More lambda work. Tweak the Sema interface slightly. Start adding the pieces to build the lambda class and its call operator. Create an actual scope for the lambda body.
llvm-svn: 147595
Andrew Trick [Thu, 5 Jan 2012 02:52:11 +0000 (02:52 +0000)]
Minor postra scheduler cleanup. It could result in more precise antidependence latency on ARM in exceedingly rare cases.
llvm-svn: 147594
Bill Wendling [Thu, 5 Jan 2012 02:13:20 +0000 (02:13 +0000)]
Replace the uint64_t -> double convertion algorithm with one that's more efficient.
This small bit of ASM code is sufficient to do what the old algorithm did:
movq %rax, %xmm0
punpckldq (c0), %xmm0 // c0: (uint4){ 0x43300000U, 0x45300000U, 0U, 0U }
subpd (c1), %xmm0 // c1: (double2){ 0x1.0p52, 0x1.0p52 * 0x1.0p32 }
#ifdef __SSE3__
haddpd %xmm0, %xmm0
#else
pshufd $0x4e, %xmm0, %xmm1
addpd %xmm1, %xmm0
#endif
It's arguably faster. One caveat, the 'haddpd' instruction isn't very fast on
all processors.
<rdar://problem/7719814>
llvm-svn: 147593
Sean Callanan [Thu, 5 Jan 2012 02:00:14 +0000 (02:00 +0000)]
Fixed a potential hang while trying to execute
a function in the inferior.
llvm-svn: 147592
Chad Rosier [Thu, 5 Jan 2012 01:56:11 +0000 (01:56 +0000)]
[driver] Add support for passing -lazy_framework/-lazy_library to the linker.
rdar://
10630328
llvm-svn: 147591
Johnny Chen [Thu, 5 Jan 2012 01:26:01 +0000 (01:26 +0000)]
Add comment explaining the default constructor (ArchSpec) used in CreateTarget().
llvm-svn: 147590
Eli Friedman [Thu, 5 Jan 2012 01:16:29 +0000 (01:16 +0000)]
Change casting slightly to avoid warnings about casting away const.
llvm-svn: 147589
Douglas Gregor [Thu, 5 Jan 2012 01:11:47 +0000 (01:11 +0000)]
When we're performing name lookup for a tag, we still allow ourselves
to see hidden declarations because every tag lookup is effectively a
redeclaration lookup. For example, image that
struct foo;
is declared in a submodule that is known but hasn't been imported. If
someone later writes
struct foo *foo_p;
then "struct foo" is either a reference or a redeclaration. To keep
the redeclaration chains sound, we treat it like a redeclaration for
name-lookup purposes.
llvm-svn: 147588
Sean Callanan [Thu, 5 Jan 2012 01:11:09 +0000 (01:11 +0000)]
Fixed a dangling pointer bug associated with the
result variable on a "finish" statement. The
ownership of the result value was not being properly
assigned to the newly-created persistent result
variable; now it is.
llvm-svn: 147587
Kostya Serebryany [Thu, 5 Jan 2012 01:07:27 +0000 (01:07 +0000)]
[asan] move {linux,mac}-specific code from asan_thread.cc to asan_{linux,mac}.cc; also add asan_procmaps.h which I forgot to add on previous commit.
llvm-svn: 147586
Andrew Trick [Thu, 5 Jan 2012 01:01:01 +0000 (01:01 +0000)]
comment cleanup
llvm-svn: 147585
Devang Patel [Thu, 5 Jan 2012 00:51:28 +0000 (00:51 +0000)]
Do not hard code asm variant number.
llvm-svn: 147583
Eli Friedman [Thu, 5 Jan 2012 00:49:17 +0000 (00:49 +0000)]
Minor cleanup; no functional change.
llvm-svn: 147582
Kostya Serebryany [Thu, 5 Jan 2012 00:44:33 +0000 (00:44 +0000)]
[asan] implement our own /proc/self/maps reader and use it on linux instead of sysinfo.h
llvm-svn: 147581
Benjamin Kramer [Thu, 5 Jan 2012 00:43:34 +0000 (00:43 +0000)]
FileCheck hygiene.
llvm-svn: 147580
Jakob Stoklund Olesen [Thu, 5 Jan 2012 00:26:57 +0000 (00:26 +0000)]
Reapply r146997, "Heed spill slot alignment on ARM."
Now that canRealignStack() understands frozen reserved registers, it is
safe to use it for aligned spill instructions.
It will only return true if the registers reserved at the beginning of
register allocation allow for dynamic stack realignment.
<rdar://problem/
10625436>
llvm-svn: 147579
Jakob Stoklund Olesen [Thu, 5 Jan 2012 00:26:52 +0000 (00:26 +0000)]
Avoid reserving an ARM base pointer during register allocation.
Once register allocation has started the reserved registers are frozen.
Fix the ARM canRealignStack() hook to respect the frozen register state.
Now the hook returns false if register allocation was started with frame
pointer elimination enabled.
It also returns false if register allocation started without a reserved
base pointer, and stack realignment would require a base pointer. This
bug was breaking oggenc on armv6.
No test case, an upcoming patch will use this functionality to realign
the stack for spill slots when possible.
llvm-svn: 147578
Jakob Stoklund Olesen [Thu, 5 Jan 2012 00:26:49 +0000 (00:26 +0000)]
Freeze reserved registers before starting register allocation.
The register allocators don't currently support adding reserved
registers while they are running. Extend the MRI API to keep track of
the set of reserved registers when register allocation started.
Target hooks like hasFP() and needsStackRealignment() can look at this
set to avoid reserving more registers during register allocation.
llvm-svn: 147577
Argyrios Kyrtzidis [Thu, 5 Jan 2012 00:19:03 +0000 (00:19 +0000)]
Sanity checks in SourceManager::getFileEntryForID() and SourceManager::getFileEntryForSLocEntry()
to make sure we do not crash. rdar://
10594186
llvm-svn: 147576