summaryrefslogtreecommitdiffstats
path: root/lldb/source/Plugins
Commit message (Collapse)AuthorAgeFilesLines
...
* Revert r343342 together with LLVM commit 347490.Luke Cheeseman2018-11-232-3/+3
| | | | llvm-svn: 347491
* Remove unused variable. NFC.Benjamin Kramer2018-11-191-1/+0
| | | | llvm-svn: 347188
* Revert "Implement basic DidAttach and DidLaunch for DynamicLoaderWindowsDYLD"Zachary Turner2018-11-181-44/+2
| | | | | | | This breaks many tests on Windows, which now all fail with an error such as "Unable to read memory at address <xxxxxxxx>". llvm-svn: 347174
* [Cocoa] Prefer llvm::SignExtend64. Pointed out by zturner.Davide Italiano2018-11-161-2/+1
| | | | llvm-svn: 347087
* [SymbolFile] Remove unused function. NFCI.Davide Italiano2018-11-161-18/+0
| | | | llvm-svn: 347084
* [lldb] NFC: Remove the extra ';'Henry Wong2018-11-161-1/+1
| | | | | | | | | | | | | | | | | | | | Summary: Remove extra `;` to eliminate the following pedantic warning. ``` warning: extra ';' [-Wpedantic] ``` Reviewers: shafik Reviewed By: shafik Subscribers: abidh, lldb-commits Differential Revision: https://reviews.llvm.org/D54528 llvm-svn: 347047
* Don't use uniform initialization syntax.Zachary Turner2018-11-163-11/+21
| | | | llvm-svn: 347020
* [NativePDB] Rewrite the PdbSymUid to use our own custom namespacing scheme.Zachary Turner2018-11-1613-411/+435
| | | | | | | | | | | | | | | | | | | | | | | Originally we created our 64-bit UID scheme by using the first byte as sort of a "tag" to represent what kind of symbol this was, and we re-used the PDB_SymType enumeration for this. For native pdb support, this is not really the right abstraction layer, because what we really want is something that tells us *how* to find the symbol. This means, specifically, is in the globals stream / public stream / module stream / TPI stream / etc, and for whichever one it is in, where is it within that stream? A good example of why the old namespacing scheme was insufficient is that it is more or less impossible to create a uid for a field list member of a class/struction/union/enum that tells you how to locate the original record. With this new scheme, the first byte is no longer a PDB_SymType enum but a new enum created specifically to identify where in the PDB this record lives. This gives us much better flexibility in what kinds of symbols the uids can identify. llvm-svn: 347018
* Implement basic DidAttach and DidLaunch for DynamicLoaderWindowsDYLDNathan Lanza2018-11-151-2/+44
| | | | | | | | | | | | | | | Summary: This commit implements basic DidAttach and DidLaunch for the windows DynamicLoader plugin which allow us to load shared libraries from the inferior. Reviewers: sas, zturner Reviewed By: zturner Differential Revision: https://reviews.llvm.org/D54544 llvm-svn: 346994
* Add setting to require hardware breakpoints.Jonas Devlieghere2018-11-151-2/+4
| | | | | | | | | | | | | | | | When debugging read-only memory we cannot use software breakpoint. We already have support for hardware breakpoints and users can specify them with `-H`. However, there's no option to force LLDB to use hardware breakpoints internally, for example while stepping. This patch adds a setting target.require-hardware-breakpoint that forces LLDB to always use hardware breakpoints. Because hardware breakpoints are a limited resource and can fail to resolve, this patch also extends error handling in thread plans, where breakpoints are used for stepping. Differential revision: https://reviews.llvm.org/D54221 llvm-svn: 346920
* Fix some compilation failures introduced in recent patches.Zachary Turner2018-11-142-4/+5
| | | | | | | | | | This fixes two compilation failures: 1) Designated initializers are C++20. We can't use them in LLVM. 2) thread_result_t is not a pointer type on all platforms, so returning nullptr is an error. llvm-svn: 346873
* [LLDB] - Recommit r346848 "[LLDB] - Support the single file split DWARF.".George Rimar2018-11-145-17/+50
| | | | | | | | | | | | | | | | | | | | | | | | | Test cases were updated to not use the local compilation dir which is different between development pc and build bots. Original commit message: [LLDB] - Support the single file split DWARF. DWARF5 spec describes a single file split dwarf case (when .dwo sections are in the .o files). Problem is that LLDB does not work correctly in that case. The issue is that, for example, both .debug_info and .debug_info.dwo has the same type: eSectionTypeDWARFDebugInfo. And when code searches section by type it might find the regular debug section and not the .dwo one. The patch fixes that. With it, LLDB is able to work with output compiled with -gsplit-dwarf=single flag correctly. Differential revision: https://reviews.llvm.org/D52403 llvm-svn: 346855
* Revert r346848 "[LLDB] - Support the single file split DWARF."George Rimar2018-11-145-50/+17
| | | | | | | It broke BB: http://green.lab.llvm.org/green/job/lldb-cmake/12522/testReport/junit/LLDB/Breakpoint/single_file_split_dwarf_test/ llvm-svn: 346853
* Fix a crash when parsing incorrect DWARFPavel Labath2018-11-141-3/+3
| | | | | | | | | | | | | | | | | | | | | | | | | | | Summary: While parsing a childless compile unit DIE we could crash if the DIE was followed by any extra data (such as a superfluous end-of-children marker). This happened because the break-on-depth=0 check was performed only when parsing the null DIE, which was not correct because with a childless root DIE, we could reach the end of the unit without ever encountering the null DIE. If the compile unit contribution ended directly after the CU DIE, everything would be fine as we would terminate parsing due to reaching EOF. However, if the contribution contained extra data (perhaps a superfluous end-of-children marker), we would crash because we would treat that data as the begging of another compile unit. This fixes the crash by moving the depth=0 check to a more generic place, and also adds a regression test. Reviewers: clayborg, jankratochvil, JDevlieghere Subscribers: lldb-commits Differential Revision: https://reviews.llvm.org/D54417 llvm-svn: 346849
* [LLDB] - Support the single file split DWARF.George Rimar2018-11-145-17/+50
| | | | | | | | | | | | | | | | | | DWARF5 spec describes a single file split dwarf case (when .dwo sections are in the .o files). Problem is that LLDB does not work correctly in that case. The issue is that, for example, both .debug_info and .debug_info.dwo has the same type: eSectionTypeDWARFDebugInfo. And when code searches section by type it might find the regular debug section and not the .dwo one. The patch fixes that. With it, LLDB is able to work with output compiled with -gsplit-dwarf=single flag correctly. Differential revision: https://reviews.llvm.org/D52296 llvm-svn: 346848
* Suppress a "-Wliteral-conversion" compiler warning.Haojian Wu2018-11-141-1/+1
| | | | | error: implicit conversion from 'double' to 'uint64_t' (aka 'unsigned long') changes value from -0 to 0 [-Werror,-Wliteral-conversion] llvm-svn: 346841
* Fix the "make_unique is ambiguous" compiler error.Haojian Wu2018-11-141-2/+2
| | | | llvm-svn: 346839
* Fix a bug in the parsing of the LC_BUILD_VERSION Mach-O load command.Adrian Prantl2018-11-131-16/+20
| | | | | | | | | LC_BUILD_VERSION records are of variable length. The original code would use uninitialized memory when the size of a record was exactly 24. rdar://problem/46032185 llvm-svn: 346812
* [NativePDB] Add support for S_CONSTANT records.Zachary Turner2018-11-132-8/+134
| | | | | | | | | | | | | | | | | clang-cl does not emit these, but MSVC does, so we need to be able to handle them. Because clang-cl does not generate them, it was a bit hard to write a test. So what I had to do was get an PDB file with some S_CONSTANT records in using cl and link, dump it using llvm-pdbutil dump -globals -sym-data to get the bytes of the records, generate the same object file using clang-cl but with -S to emit an assembly file, and replace all the S_LDATA32 records with the bytes of the S_CONSTANT records. This way, we can compile the file using llvm-mc and link it with lld-link. Differential Revision: https://reviews.llvm.org/D54452 llvm-svn: 346787
* [NativePDB] Improved support for nested type reconstruction.Zachary Turner2018-11-131-4/+73
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | In a previous patch, we pre-processed the TPI stream in order to build the reverse mapping from nested type -> parent type so that we could accurately reconstruct a DeclContext hierarchy. However, there were some issues. An LF_NESTTYPE record is really just a typedef, so although it happens to be used to indicate the name of the nested type and referring to the global record which defines the type, it is also used for every other kind of nested typedef. When we rebuild the DeclContext hierarchy, we want it to be as accurate as possible, which means that if we have something like: struct A { struct B {}; using C = B; }; We don't want to create two CXXRecordDecls in the AST each with the exact same definition. We just want to create one for B and then define C as an alias to B. Previously, however, it would not be able to distinguish between the two cases and it would treat A::B and A::C as being two classes each with separate definitions. We address the first half of improving the pre-processing logic so that only actual definitions are treated this way. Later, in a followup patch, we can handle the case of nested typedefs since we're already going to be enumerating the field list anyway and this patch introduces the general framework for distinguishing between the two cases. Differential Revision: https://reviews.llvm.org/D54357 llvm-svn: 346786
* [Cocoa] Implement formatter for the new NSDate representation.Davide Italiano2018-11-131-1/+63
| | | | | | <rdar://problem/46002786> llvm-svn: 346783
* Add GDB remote packet reproducer.Jonas Devlieghere2018-11-139-183/+775
| | | | llvm-svn: 346780
* Since ABI's now hold a process WP, they should be handedJim Ingham2018-11-1312-48/+12
| | | | | | | | out one per process rather than keeping a single global instance. Differential Revision: https://reviews.llvm.org/D54460 llvm-svn: 346775
* [lldb] Add synthetic frontend for _NSCallStackArrayKuba Mracek2018-11-122-1/+33
| | | | | | | | An Obj-C array type _NSCallStackArray is used in NSException backtraces. This patch adds a synthetic frontend for _NSCallStackArray, which now correctly returns frame PCs. Differential Revision: https://reviews.llvm.org/D44081 llvm-svn: 346708
* Re-land "Extract construction of DataBufferLLVM into FileSystem"Jonas Devlieghere2018-11-126-18/+14
| | | | | | This fixes some UB in isLocal detected by the sanitized bot. llvm-svn: 346707
* [lldb] Extract more fields from NSException valuesKuba Mracek2018-11-121-8/+37
| | | | | | | | This patch teaches LLDB about more fields on NSException Obj-C objects, specifically we can now retrieve the "name" and "reason" of an NSException. The goal is to eventually be able to have SB API that can provide details about the currently thrown/caught/processed exception. Differential Revision: https://reviews.llvm.org/D43884 llvm-svn: 346695
* Revert "Extract construction of DataBufferLLVM into FileSystem"Davide Italiano2018-11-126-14/+18
| | | | | | It broke the lldb sanitizer bots. llvm-svn: 346694
* [lldb] Refactor ObjC/NSException.cpp (cleanup, avoid code duplication). NFC.Kuba Mracek2018-11-121-75/+47
| | | | | | | | | | - Refactor reading of NSException fields into ExtractFields method to avoid code duplication. - Remove "m_child_ptr" field, as it's not used anywhere. - Clang-format. Differential Revision: https://reviews.llvm.org/D44073 llvm-svn: 346679
* Fix an unused variable warning. NFCAlexander Kornienko2018-11-121-0/+2
| | | | llvm-svn: 346651
* Remove header grouping comments.Jonas Devlieghere2018-11-11303-1095/+2
| | | | | | | | This patch removes the comments grouping header includes. They were added after running IWYU over the LLDB codebase. However they add little value, are often outdates and burdensome to maintain. llvm-svn: 346626
* Remove comments after header includes.Jonas Devlieghere2018-11-1113-16/+14
| | | | | | | | | | This patch removes the comments following the header includes. They were added after running IWYU over the LLDB codebase. However they add little value, are often outdates and burdensome to maintain. Differential revision: https://reviews.llvm.org/D54385 llvm-svn: 346625
* Extract construction of DataBufferLLVM into FileSystemJonas Devlieghere2018-11-106-18/+14
| | | | | | | | | | This moves construction of data buffers into the FileSystem class. Like some of the previous refactorings we don't translate the path yet because the functionality hasn't been landed in LLVM yet. Differential revision: https://reviews.llvm.org/D54272 llvm-svn: 346598
* Unbreak the linux bot from the previous commit. Fred needed to useJason Molenda2018-11-101-0/+6
| | | | | | | | | some of the macros from mach/exc_resource.h to decode EXC_RESOURCE, but that header doesn't exist on non-apple platforms and StopInfoMachException.cpp needs to build on those systems. EXC_RESOURCE won't be decoded when lldb is built on non-darwin systems. llvm-svn: 346573
* Enable listening for EXC_RESOURCE events, and format machJason Molenda2018-11-101-0/+45
| | | | | | | | | | | | | | | | | | | | event as a thread stop reason if we receive one, using some macros to decode the payload. Patch originally written by Fred Riss, with a few small changes by myself. Writing a test for this is a little tricky because the mach exception data interpretation relies on header macros or function calls - it may change over time and writing a gdb_remote_client test for this would break as older encoding interpretation is changed. I'll tak with Fred about this more, but neither of us has been thrilled with the kind of tests we could write for it. <rdar://problem/13097323>, <rdar://problem/40144456> llvm-svn: 346571
* Work with a gdb-remote target that doesn't handle theJason Molenda2018-11-091-0/+5
| | | | | | | | | | | | | | | | | | | | | | | | | qWatchpointSupportInfo packet correctly. In GDBRemoteCommunicationClient::GetWatchpointSupportInfo, if the response to qWatchpointSupportInfo does not include the 'num' field, then we did not get an answer we understood, mark this target as not supporting that packet. In Target.cpp, rename the very confusingly named CheckIfWatchpointsExhausted to CheckIfWatchpointsSupported, and check the error status returned by Process::GetWatchpointSupportInfo. If we cannot determine what the number of supported watchpoints are, assume that they will work. We'll handle the failure later when we try to create/enable the watchpoint if the Z2 packet isn't supported. Add a gdb_remote_client test case. <rdar://problem/42621432> llvm-svn: 346561
* [NativePDB] Fix completion of enum types.Zachary Turner2018-11-091-0/+1
| | | | | | | | | | | | | This was originally submitted in a patch which fixed two unrelated bugs at the same time. This portion of the fix was reverted because it broke several other things. However, the fix employed originally was totally wrong, and attempted to change something in the ValueObject printer when actually the bug was in the NativePDB plugin. We need to mark forward enum decls as having external storage, otherwise we won't be asked to complete them when the time comes. This patch implements the proper fix, and updates tests accordingly. llvm-svn: 346517
* [NativePDB] Add support for bitfield records.Zachary Turner2018-11-091-4/+20
| | | | | | | | | Bitfields are represented as LF_MEMBER records whose TypeIndex points to an LF_BITFIELD record that describes the bit width, bit offset, and underlying type of the bitfield. All we need to do is resolve these when resolving record types. llvm-svn: 346511
* Resubmit "Fix bug in PE/COFF plugin."Zachary Turner2018-11-091-1/+5
| | | | | | | | | | The original commit was actually 2 unrelated bug fixes, but it turns out the second bug fix wasn't quite correct, so the entire patch was reverted. Resubmitting this half of the patch by itself, then will follow up with a new patch which fixes the rest of the issue in a more appropriate way. llvm-svn: 346505
* Revert "Fix bug in PE/COFF plugin and ValueObjectVariable."Davide Italiano2018-11-081-5/+1
| | | | | | It breaks some tests on MacOS. llvm-svn: 346444
* [lldb] Fix initialization order warnings.Jorge Gorbe Moya2018-11-081-1/+1
| | | | | | | | | | | | | | | Moved the declaration of m_kind below the declaration of cvclass, cvunion and cvenum. This order is necessary because in one of the constructors the initialization of m_kind depends on the value of cvclass. third_party/llvm/llvm/tools/lldb/source/Plugins/SymbolFile/NativePDB/PdbUtil.cpp:50:7: error: field 'cvclass' will be initialized after field 'm_kind' [-Werror,-Wreorder] : cvclass(std::move(c)), ^ third_party/llvm/llvm/tools/lldb/source/Plugins/SymbolFile/NativePDB/PdbUtil.cpp:51:14: error: field 'cvclass' is uninitialized when used here [-Werror,-Wuninitialized] m_kind(cvclass.Kind == TypeRecordKind::Struct ? Struct : Class) {} llvm-svn: 346435
* Fix bug in PE/COFF plugin and ValueObjectVariable.Zachary Turner2018-11-081-1/+5
| | | | | | | | | | | | | | | | | | There are two bugs here. The first is that MSVC and clang-cl emit their bss section under the name '.data' instead of '.bss' but with the size and file offset set to 0. ObjectFilePECOFF didn't handle this, and would only recognize a section as bss if it was actually called '.bss'. The effect of this is that if we tried to print the value of a variable that lived in BSS we would fail. The second bug is that ValueObjectVariable was only returning the forward type, which is insufficient to print the value of an enum. So we bump this up to the layout type. Differential Revision: https://reviews.llvm.org/D54241 llvm-svn: 346430
* [NativePDB] Higher fidelity reconstruction of AST from Debug Info.Zachary Turner2018-11-084-18/+221
| | | | | | | | | | | | | | | | | | | | | | | | | In order to accurately put a type into the correct location in the AST we construct from debug info, we need to be able to determine what DeclContext (namespace, global, nested class, etc) that it goes into. PDB doesn't contain this mapping. It does, however, contain the reverse mapping. That is, for a given class type T, you can determine all classes Q1, Q2, ..., Qn that are nested inside of T. We need to know, for a given class type Q, what type T is it nested inside of. This patch builds this map as a pre-processing step when we first load the PDB by scanning every type. Initial tests show that while this can be slow in debug builds of LLDB, it is quite fast in release builds (less than 2 seconds for a ~1GB PDB, and it only needs to happen once). Furthermore, having this pre-processing step in place allows us to repurpose it for building up other kinds of indexing to it down the line. For the time being, this gives us very accurate reconstruction of the DeclContext hierarchy. Differential Revision: https://reviews.llvm.org/D54216 llvm-svn: 346429
* Refactor ClangASTContext::AddEnumerationValueToEnumerationType() to remove ↵Shafik Yaghmour2018-11-083-11/+4
| | | | | | | | | | redundant parameter which can be calculated from other parameter. rdar://problem/43822994 Differential Revision: https://reviews.llvm.org/D54003 llvm-svn: 346428
* [FileSystem] Add convenience method to check for directories.Jonas Devlieghere2018-11-087-19/+18
| | | | | | | | | | | Replace calls to LLVM's is_directory with calls to LLDB's FileSytem class. For this I introduced a new convenience method that, like the other methods, takes either a path or filespec. This still uses the LLVM functions under the hood. Differential revision: https://reviews.llvm.org/D54135 llvm-svn: 346375
* Add a break to avoid an unannotated fall-through.Eric Christopher2018-11-071-0/+1
| | | | llvm-svn: 346294
* Compiler warning pointed out a mistake - fix it.Jason Molenda2018-11-071-1/+1
| | | | llvm-svn: 346283
* [ObjectFileELF] Fix misaligned read/writes caught by UBSan.Davide Italiano2018-11-061-2/+3
| | | | llvm-svn: 346244
* CPlusPlusLanguage: Use new demangler API to implement type substitutionPavel Labath2018-11-061-62/+80
| | | | | | | | | | | | | | | | Summary: Now that llvm demangler supports more generic customization, we can implement type substitution directly on top of this API. This will allow us to remove the specialized hooks which were added to the demangler to support this use case. Reviewers: sgraenitz, erik.pilkington, JDevlieghere Subscribers: lldb-commits Differential Revision: https://reviews.llvm.org/D54074 llvm-svn: 346233
* Add parens to silence warningBenjamin Kramer2018-11-061-1/+1
| | | | | | MSVCUndecoratedNameParser.cpp:25:36: warning: '&&' within '||' [-Wlogical-op-parentheses] llvm-svn: 346220
* [PDB] Introduce `MSVCUndecoratedNameParser`Aleksandr Urakov2018-11-068-121/+310
| | | | | | | | | | | | This patch introduces the simple MSVCUndecoratedNameParser. It is needed for parsing names of PDB symbols corresponding to template instantiations. For example, for the name `operator<<A>'::`2'::B::operator> we can't just split the name with :: (as it is implemented for now) to retrieve its scopes. This parser processes such names in a more correct way. Differential Revision: https://reviews.llvm.org/D52461 llvm-svn: 346213
OpenPOWER on IntegriCloud