Home - Waterfall Grid T-Grid Console Builders Recent Builds Buildslaves Changesources - JSON API - About

Console View


Categories: Cygwin Linux Linux cross Linux fast Windows
Legend:   Passed Failed Warnings Failed Again Running Exception Offline No data

Cygwin Linux Linux cross Linux fast Windows
r270953
David Majnemer
[CodeGen] Don't crash when sizeof(long) != 4 for some intrins

_InterlockedIncrement and _InterlockedDecrement have 'long' in their
prototypes.  We assumed 'long' was the same size as an i32 which is
incorrect for other targets.

This fixes PR27892.
r270952
David Majnemer
[Intrin.h] Sort the __read[fg]s intrinsics

No functional change is intended.
r270951
Zachary Turner
[codeview,pdb] Try really hard to conserve memory when reading.

PDBs can be extremely large.  We're already mapping the entire
PDB into the process's address space, but to make matters worse
the blocks of the PDB are not arranged contiguously.  So, when
we have something like an array or a string embedded into the
stream, we have to make a copy.  Since it's convenient to use
traditional data structures to iterate and manipulate these
records, we need the memory to be contiguous.

As a result of this, we were using roughly twice as much memory
as the file size of the PDB, because every stream was copied
out and re-stitched together contiguously.

This patch addresses this by improving the MappedBlockStream
to allocate from a BumpPtrAllocator only when a read requires
a discontiguous read.  Furthermore, it introduces some data
structures backed by a stream which can iterate over both
fixed and variable length records of a PDB.  Since everything

(snip)
  • cmake-llvm-x86_64-linux: built llvm failed -  stdio
r270950
Lang Hames
[Support] Rename unconvertibleErrorCode to inconvertibleErrorCode.

Based on a totally scientific, 30 second google search "in-" appears to be the
preferred prefix.
r270949
Derek Bruening
[esan] Fix workingset-memset test failure

Fixes an esan workingset-memset test failure by switching to malloc to
avoid a shadow mapping issue with mmap in certain situations that will be
fully fixed separately.
r270948
Lang Hames
[Support] Add a StringError convenience class to Error.h

StringError can be used to represent Errors that aren't recoverable based on
the error type, but that have a useful error message that can be reported to
the user or logged.
r270947
Kostya Serebryany
[libFuzzer] make check-fuzzer a bit faster
r270946
Michael Zolotukhin
[LoopUnrollAnalyzer] Bail out instead of dying with assert when facing huge index.

This fixes PR27902.
r270945
Kostya Serebryany
[libFuzzer] make OOM-handling more portable. Instead of sending a signal to the main fuzzing thread, print the message in the getrusage thread and exit.
r270944
Mehdi Amini
ValueMapper: fix typo in minor optimization on constant mapping (NFC)

If every operands of a constant are mapping to themselves, and the
type does not change, we have an early exit as acknowledged in the
comment:

  // Otherwise, we have some other constant to remap.  Start by checking to see
  // if all operands have an identity remapping.

However instead of checking for identity the code was checking if the
operands were mapped to the constant itself, which is rarely true.

As a consequence, the coverage report showed that the early exit was
never taken.
r270943
Rui Ueyama
pdbdump: print out the name of the stream 0.

Differential Revision: http://reviews.llvm.org/D20712
r270942
Kostya Serebryany
[libFuzzer] more refactoring: make sure CurrentUnitData is awlays a valid pointer to read from
r270941
Greg Clayton
Add a new "lldb" log channel named "demangle". If we have crashes that are related to demangling, we now can enable this logging and we will be able to reproduce demangler crashes (usually due to overflowing the stack) without needing someone's project.

<rdar://problem/25221899>
r270940
NAKAMURA Takumi
OpenMPClause.h: Fix r270882. [-Wdocumentation]
r270939
Jim Ingham
Lock the access to the BreakpointLocationCollection.

I was investigating an odd crash in lldb when the breakpoint site
goes to bump the hit counts of the locations it implements.  I noticed
that the BreakpointLocationCollection wasn't locking itself for access and
modification.  I don't see how that can cause the crash I'm seeing, but still
this is the right thing to do...

<rdar://problem/25178205>
r270938
Jim Ingham
Don't cache the stret/vrs. non-stret code pointer as static data in the runtime.
It belongs in the instance, since then when you change architectures it can be adjusted
appropriately.

<rdar://problem/26308079>
r270937
Rui Ueyama
pdbdump: Add -raw-all to enable all -raw-* flags.

Differential Revision: http://reviews.llvm.org/D20707
r270936
Hans Wennborg
Sort my entry in CODE_OWNERS.TXT
r270935
Mitch Bodart
[CodeGen] Fix problem with X86 byte registers in CriticalAntiDepBreaker

CriticalAntiDepBreaker was not correctly tracking defs of the high X86 byte
registers, leading to incorrect use of a busy register to break an
antidependence.

Fixes pr27681, and its duplicates pr27580, pr27804.

Differential Revision: http://reviews.llvm.org/D20456
r270934
Rui Ueyama
Fix typo.
r270933
Easwaran Raman
Attach profile summary in IR based instrumentation pass.

Differential revision: http://reviews.llvm.org/D20655
r270932
Greg Clayton
With -gmodules, we have been having a harder time always finding a type when we need one.

We have seen cases where we have been unable to find an argument type for a function, or we find one from another language, and then we try to create a function type by calling:

lldb_private::ClangASTContext::CreateFunctionType(clang::ASTContext*, lldb_private::CompilerType const&, lldb_private::CompilerType const*, unsigned int, bool, unsigned int)

This fix will ensure that all arguments to lldb_private::ClangASTContext::CreateFunctionType() are in order by checking:
- AST is valid
- if arguments are specified we have a valid argument array
- return type is valid
- return type is a clang type
- all argument types are valid
- all argument types are clang types

If any of these fail, we return an invalid CompilerType. If we don't return an invalid type, clang will crash anyway, and LLDB must not crash even in the presence of bad or missing debug info.

<rdar://problem/25172715>
r270931
Justin Bogner
SDAG: Use an Optional<> instead of a sigil value. NFC

This just makes it a bit more clear that we don't intend to use a
deleted node for anything here.
r270930
Xinliang David Li
Add missing test case changes
r270929
Kostya Serebryany
[libFuzzer] more refactoring around CurrentUnit. Also add a threading test on which we currently have a race (when reporting bugs from multiple threads)