summaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
...
* [CGDebugInfo] Emit subprograms for decls when AT_tail_call is understood ↵Vedant Kumar2019-11-196-31/+72
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | (reland with fixes) Currently, clang emits subprograms for declared functions when the target debugger or DWARF standard is known to support entry values (DW_OP_entry_value & the GNU equivalent). Treat DW_AT_tail_call the same way to allow debuggers to follow cross-TU tail calls. Pre-patch debug session with a cross-TU tail call: ``` * frame #0: 0x0000000100000fa4 main`target at b.c:4:3 [opt] frame #1: 0x0000000100000f99 main`main at a.c:8:10 [opt] ``` Post-patch (note that the tail-calling frame, "helper", is visible): ``` * frame #0: 0x0000000100000fa4 main`target at b.c:4:3 [opt] frame #1: 0x0000000100000f80 main`helper [opt] [artificial] frame #2: 0x0000000100000f99 main`main at a.c:8:10 [opt] ``` This was reverted in 5b9a072c because it attached declaration subprograms to inlinable builtin calls, which interacted badly with the MergeICmps pass. The fix is to not attach declarations to builtins. rdar://46577651 Differential Revision: https://reviews.llvm.org/D69743
* [docs] Remove dangling parenthesis from documentationTim Northover2019-11-191-1/+1
| | | | Patch by leiteg.
* Add a key method to Sema to optimize debug info sizeReid Kleckner2019-11-192-1/+7
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | It turns out that the debug info describing the Sema class is an appreciable percentage of the total object file size of objects in Sema. By adding a key function, clang is able to optimize the debug info size by emitting a forward declaration in TUs that do not define the key function. On Windows, with clang-cl, these are the total sizes of object files in Sema before and after this change, compiling with optimizations and debug info: before: 335,012 KB after: 278,116 KB delta: -56,896 KB percent: -17.0% The effect on link time was negligible, despite having ~56MB less input. On Linux, with clang, these are the same sizes using DWARF -g and optimizations: before: 603,756 KB after: 515,340 KB delta: -88,416 KB percent: -14.6% I didn't use type units, DWARF-5, fission, or any other special flags. Reviewed By: thakis Differential Revision: https://reviews.llvm.org/D70340
* [profile] Unbreak Fuchsia/Windows after D68351Vedant Kumar2019-11-191-3/+3
| | | | | | Continuous mode is not yet supported on Fuchsia/Windows, however an error should not be reported unless the user attempted to actually enable continuous mode.
* The patch is the compiler error specific on the compile error on CMVCdiggerlin2019-11-191-3/+5
| | | | | | | | | | | | | | | | | | | | | | | SUMMARY: CMVC has a compiler error on the const uint64_t OffsetToRaw = is64Bit() ? toSection64(Sec)->FileOffsetToRawData : toSection32(Sec)->FileOffsetToRawData; while gcc compiler do not have the problem. I have to change the code to uint64_t OffsetToRaw; if (is64Bit()) OffsetToRaw = toSection64(Sec)->FileOffsetToRawData; else OffsetToRaw = toSection32(Sec)->FileOffsetToRawData; Reviewers: Sean Fertile Subscribers: rupprecht, seiyai,hiraditya Differential Revision: https://reviews.llvm.org/D70255
* fixe leak found by asan build botTyker2019-11-191-1/+3
|
* [DebugInfo] Describe size of spilled values in call site paramsVedant Kumar2019-11-192-2/+6
| | | | | | | | A call site parameter description of a memory operand needs to unambiguously convey the size of the operand to prevent incorrect entry value evaluation. Thanks for David Stenberg for pointing this issue out!
* llvm/ObjCARC: Eliminate inlined AutoreleaseRV callsDuncan P. N. Exon Smith2019-11-193-73/+438
| | | | | | | | | | | | | | | | | | | | | | | | | Pair up inlined AutoreleaseRV calls with their matching RetainRV or ClaimRV. - RetainRV cancels out AutoreleaseRV. Delete both instructions. - ClaimRV is a peephole for RetainRV+Release. Delete AutoreleaseRV and replace ClaimRV with Release. This avoids problems where more aggressive inlining triggers memory regressions. This patch is happy to skip over non-callable instructions and non-ARC intrinsics looking for the pair. It is likely sound to also skip over opaque function calls, but that's harder to reason about, and it's not relevant to the goal here: if there's an opaque function call splitting up a pair, it's very unlikely that a handshake would have happened dynamically without inlining. Note that this patch also subsumes the previous logic that looked backwards from ReleaseRV. https://reviews.llvm.org/D70370 rdar://problem/46509586
* [SLP] fix miscompile on min/max reductions with extra uses (PR43948) (2nd try)Sanjay Patel2019-11-193-13/+28
| | | | | | | | | | | | | | | | | | | | | | The 1st attempt was reverted because it revealed an existing bug where we could produce invalid IR (use of value before definition). That should be fixed with: rG39de82ecc9c2 The bug manifests as replacing a reduction operand with an undef value. The problem appears to be limited to cases where a min/max reduction has extra uses of the compare operand to the select. In the general case, we are tracking "ExternallyUsedValues" and an "IgnoreList" of the reduction operations, but those may not apply to the final compare+select in a min/max reduction. For that, we use replaceAllUsesWith (RAUW) to ensure that the new vectorized reduction values are transferred to all subsequent users. Differential Revision: https://reviews.llvm.org/D70148
* [nfc][libomptarget] Remove casts of string literals to char*JonChesterfield2019-11-192-5/+4
|
* MTE: add more unchecked instructions.Evgenii Stepanov2019-11-192-7/+152
| | | | | | | | | | | | | | Summary: In particular, 1- and 2-byte loads and stores ignore the pointer tag when using SP as the base register. Reviewers: pcc, ostannard Subscribers: hiraditya, llvm-commits Tags: #llvm Differential Revision: https://reviews.llvm.org/D70341
* test-release.sh: Update to fetch source from GitHubTom Stellard2019-11-191-60/+52
| | | | | | | | | | | | | | | | Summary: This also changes the test-release.sh script to build using the monorepo layout instead of copying sub-projects into llvm/tools or llvm/projects. Reviewers: jdoerfert, hans Reviewed By: hans Subscribers: hans, llvm-commits Tags: #llvm Differential Revision: https://reviews.llvm.org/D70353
* [ADT][Expensive checks] Create a std::random_device seed only once when ↵Alex Lorenz2019-11-191-8/+21
| | | | | | | | | | | shuffling before sorting This speeds up the build of compiler-rt with an expensive checks enabled clang by an order of 1 or 2 magnitudes on my machine. I was hoping this would also fix the 'large.test' libFuzzer timeout on the expensive checks bot on green dragon http://lab.llvm.org:8080/green/job/clang-stage1-cmake-RA-expensive/, but the fuzzer test still takes too long to compile because of other IR/MIR verification inefficiencies. Differential Revision: https://reviews.llvm.org/D70288
* Add a "Using LLDB" section to the welcome page of the websiteAdrian Prantl2019-11-194-17/+31
| | | | | | | | | | | | | | | This is an attempt to feature the user-facing resources more prominently on the LLDB website by calling out the tutorial and the GDB command map wight on the start page. I also moved the "Why a new debugger" section to the "Goals" subpage. Given that LLDB's first release is almost a decade in the past now, the title is a bit of an anachronism. Lastly, I moved the Architecture sub-page from "use" to "resources", since end-users do not care about the source code layout. Differential Revision: https://reviews.llvm.org/D70449
* [ARM] MVE interleaving load and stores.David Green2019-11-199-3874/+1119
| | | | | | | | | | | | | | | | | | | | Now that we have the intrinsics, we can add VLD2/4 and VST2/4 lowering for MVE. This works the same way as Neon, recognising the load/shuffles combination and converting them into intrinsics in a pre-isel pass, which just calls getMaxSupportedInterleaveFactor, lowerInterleavedLoad and lowerInterleavedStore. The main difference to Neon is that we do not have a VLD3 instruction. Otherwise most of the code works very similarly, with just some minor differences in the form of the intrinsics to work around. VLD3 is disabled by making isLegalInterleavedAccessType return false for those cases. We may need some other future adjustments, such as VLD4 take up half the available registers so should maybe cost more. This patch should get the basics in though. Differential Revision: https://reviews.llvm.org/D69392
* [ARM] Add and update a lot of VLDn tests. NFCDavid Green2019-11-198-509/+12017
|
* implement printing out raw section data of xcoff objectfile for llvm-objdumpdiggerlin2019-11-193-8/+107
| | | | | | | | | | | SUMMARY: implement printing out raw section data of xcoff objectfile for llvm-objdump and option -D --disassemble-all option for llvm-objdump Reviewers: Sean Fertile Subscribers: rupprecht, seiyai,hiraditya Differential Revision: https://reviews.llvm.org/D70255
* [C-index] Fix test when using Debug target & MSVC STLAlexandre Ganea2019-11-191-1/+1
| | | | | | Avoids a deadlock in "clang/test/Index/crash-recovery-modules.m" when building with the MSVC STL & _ITERATOR_DEBUG_LEVEL == 2 (meaning a DEBUG build) Differential Revision: https://reviews.llvm.org/D69959
* [LLDB] Fix formatting in the driver (NFC)Jonas Devlieghere2019-11-191-2/+1
|
* [Symbolizers] On Darwin compute function offset when possible.Dan Liew2019-11-194-3/+107
| | | | | | | | | | | | | | | | | | | | | | | | | | | Summary: The sanitizer symbolizers support printing the function offset (difference between pc and function start) of a stackframe using the `%q` format specifier. Unfortunately this didn't actually work because neither the atos or dladdr symbolizer set the `AddressInfo::function_offset` field. This patch teaches both symbolizers to try to compute the function offset. In the case of the atos symbolizer, atos might not report the function offset (e.g. it reports a source location instead) so in this case it fallsback to using `dladdr()` to compute the function offset. Two test cases are included. rdar://problem/56695185 Reviewers: kubamracek, yln Subscribers: #sanitizers, llvm-commits Tags: #sanitizers, #llvm Differential Revision: https://reviews.llvm.org/D69549
* clang: Add -fconvergent-functions flagMatt Arsenault2019-11-196-3/+19
| | | | | | | | The CUDA builtin library is apparently compiled in C++ mode, so the assumption of convergent needs to be made in a typically non-SPMD language. The functions in the library should still be assumed convergent. Currently they are not, which is potentially incorrect and this happens to work after the library is linked.
* [ELF] Disallow out-of-range section group indices after D70146Fangrui Song2019-11-191-0/+2
| | | | | Exposed by invalid/sht-group-wrong-section.test http://45.33.8.238/win/2613/step_9.txt
* [OPENMP]Rename function, NFC.Alexey Bataev2019-11-193-16/+16
| | | | Change the name of the CGOpenMPRuntime::emitOMPIfClause to CGOpenMPRuntime::emitIfClause.
* [nfc][libomptarget] Write amdgcn macros in terms of compiler intrinsicsJonChesterfield2019-11-191-3/+3
|
* [NFC] Refactor representation of materialized temporariesTyker2019-11-1946-175/+333
| | | | | | | | | | | | | | | Summary: this patch refactor representation of materialized temporaries to prevent an issue raised by rsmith in https://reviews.llvm.org/D63640#inline-612718 Reviewers: rsmith, martong, shafik Reviewed By: rsmith Subscribers: thakis, sammccall, ilya-biryukov, rnkovacs, arphaman, cfe-commits Tags: #clang Differential Revision: https://reviews.llvm.org/D69360
* [libTooling] Extend `buildASTFromCodeWithArgs` to take files argument.Yitzhak Mandelbaum2019-11-192-2/+9
| | | | | | | | | | | | | | | | | Summary: Adds an optional parameter to `buildASTFromCodeWithArgs` that allows the user to pass additional files that the main code needs to compile. This change makes `buildASTFromCodeWithArgs` consistent with `runToolOnCodeWithArgs`. Patch by Alexey Eremin. Reviewers: gribozavr Subscribers: cfe-commits Tags: #clang Differential Revision: https://reviews.llvm.org/D70175
* [FileCheck] Use lit's internal shell for the test suiteJoel E. Denny2019-11-191-0/+5
| | | | | | | | | | | | | | | | | | An advantage is that there are less portability concerns when writing tests. For example, `-u` is not supported by all implementations of `env`, but lit's internal shell provides its own `env` that supports `-u`. A disadvantage is that some shell constructs, such as parentheses, are not supported, but FileCheck's test suite currently doesn't require such constructs. For comparison, lit configures its test suite in the same manner. See `llvm/utils/lit/tests/lit.cfg`. Reviewed By: rnk Differential Revision: https://reviews.llvm.org/D70278
* [ELF] Improve --gc-sections compatibility with GNU ld regarding section groupsFangrui Song2019-11-195-2/+129
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Based on D70020 by serge-sans-paille. The ELF spec says: > Furthermore, there may be internal references among these sections that would not make sense if one of the sections were removed or replaced by a duplicate from another object. Therefore, such groups must be included or omitted from the linked object as a unit. A section cannot be a member of more than one group. GNU ld has 2 behaviors that we don't have: - Group members (nextInSectionGroup != nullptr) are subject to garbage collection. This includes non-SHF_ALLOC SHT_NOTE sections. In particular, discarding non-SHF_ALLOC SHT_NOTE sections is an expected behavior by the Annobin project. See https://developers.redhat.com/blog/2018/02/20/annobin-storing-information-binaries/ for more information. - Groups members are retained or discarded as a unit. Members may have internal references that are not expressed as SHF_LINK_ORDER, relocations, etc. It seems that we should be more conservative here: if a section is marked live, mark all the other member within the group. Both behaviors are reasonable. This patch implements them. A new field InputSectionBase::nextInSectionGroup tracks the next member within a group. on ELF64, this increases sizeof(InputSectionBase) froms 144 to 152. InputSectionBase::dependentSections tracks section dependencies, which is used by both --gc-sections and /DISCARD/. We can't overload it for the "next member" semantic, because we should allow /DISCARD/ to discard sections independent of --gc-sections (GNU ld behavior). This behavior may be reasonably used by `/DISCARD/ : { *(.ARM.exidx*) }` or `/DISCARD/ : { *(.note*) }` (new test `linkerscript/discard-group.s`). Reviewed By: ruiu Differential Revision: https://reviews.llvm.org/D70146
* Fix openmp on PowerPC64-BE-ELFv2 ABI on FreeBSD.AndreyChurbanov2019-11-192-15/+15
| | | | | | Patch by adalava (Alfredo Dal'Ava J.nior) Differential Revision: https://reviews.llvm.org/D67190
* [NFC] Attempting to fix sphinx build failure with badly encoded characters ↵Mitchell Balan2019-11-191-1/+1
| | | | from incoming patch.
* gn build: Merge 7fe9435dc88LLVM GN Syncbot2019-11-191-0/+1
|
* Work on cleaning up denormal mode handlingMatt Arsenault2019-11-1910-13/+131
| | | | | | | | | | | | | | | | | | | | | | Cleanup handling of the denormal-fp-math attribute. Consolidate places checking the allowed names in one place. This is in preparation for introducing FP type specific variants of the denormal-fp-mode attribute. AMDGPU will switch to using this in place of the current hacky use of subtarget features for the denormal mode. Introduce a new header for dealing with FP modes. The constrained intrinsic classes define related enums that should also be moved into this header for uses in other contexts. The verifier could use a check to make sure the denorm-fp-mode attribute is sane, but there currently isn't one. Currently, DAGCombiner incorrectly asssumes non-IEEE behavior by default in the one current user. Clang must be taught to start emitting this attribute by default to avoid regressions when this is switched to assume ieee behavior if the attribute isn't present.
* [cmake] Disable GCC 9's -Winit-list-lifetime warning in ArrayRefPavel Labath2019-11-191-0/+10
| | | | | | | | | | | | | | | | | | | | | | | | Summary: This is a new warning which fires when one stores a reference to the initializer_list contents in a way which may outlive the initializer_list which it came from. In llvm this warning is triggered whenever someone uses the initializer_list ArrayRef constructor. This is indeed a dangerous thing to do (I myself was bitten by that at least once), but it is not more dangerous than calling other ArrayRef constructors with temporary objects -- something which we are used to and have accepted as a tradeoff for ArrayRef's efficiency. Currently, this warnings generates so much output that it completely obscures any actionable warnings, so this patch disables it. Reviewers: rnk, aaron.ballman Subscribers: mgorny, llvm-commits Tags: #llvm Differential Revision: https://reviews.llvm.org/D70122
* [clang][NFC] Make various uses of Regex constThomas Preud'homme2019-11-196-43/+43
| | | | | | | | | | | | | | | | The const-correctness of match() was fixed in rL372764, which allows uses of Regex objects to be const in cases they couldn't be before. This patch tightens up the const-ness of Regex in various such cases. Reviewers: thopre Reviewed By: thopre Subscribers: cfe-commits Tags: #clang Differential Revision: https://reviews.llvm.org/D68155
* [NFC] Attempting to fix sphinx build failure with badly encoded characters ↵Mitchell Balan2019-11-191-1/+1
| | | | from incoming patch.
* [AIX][XCOFF] Write Function descriptors and TOC base to data sectionjasonliu2019-11-192-6/+125
| | | | | | This patch implements writing function descriptors and TOC base into data section, and also add function descriptors(both csect and label) and TOC base symbols to the symbol table.
* [clang-tidy] Give readability-redundant-member-init an option ↵Mitchell Balan2019-11-195-8/+91
| | | | | | | | | | | | | | | | | | | IgnoreBaseInCopyConstructors to avoid breaking code with gcc -Werror=extra Summary: readability-redundant-member-init removes redundant / unnecessary member and base class initialization. Unfortunately for the specific case of a copy constructor's initialization of a base class, gcc at strict warning levels warns if "base class is not initialized in the copy constructor of a derived class". This patch adds an option `IgnoreBaseInCopyConstructors` defaulting to 0 (thus maintaining current behavior by default) to skip the specific case of removal of redundant base class initialization in the copy constructor. Enabling this option enables the resulting code to continue to compile successfully under `gcc -Werror=extra`. New test cases `WithCopyConstructor1` and `WithCopyConstructor2` in clang-tools-extra/test/clang-tidy/readability-redundant-member-init.cpp show that it removes redundant members even from copy constructors. Reviewers: malcolm.parsons, alexfh, hokein, aaron.ballman, lebedev.ri Patch by: poelmanc Subscribers: mgehre, lebedev.ri, cfe-commits Tags: #clang, #clang-tools-extra Differential revision: https://reviews.llvm.org/D69145
* [SLP] fix insertion point for min/max reductionSanjay Patel2019-11-192-3/+18
| | | | | | As discussed in D70148 (and caused a revert of the original commit): if we insert at the select, then we can produce invalid IR because the replacement for the compare may have uses before the select.
* gn build: Merge 765b1250f68LLVM GN Syncbot2019-11-192-0/+2
|
* Fixup AVR tests to reflect changes in addend format in llvm-objdumpDavid Bozier2019-11-198-14/+14
| | | | | | | | | | | | | | Summary: Changes to llvm-objdump made in D69997 Reviewers: thakis, jhenderson, grimar Reviewed By: thakis Subscribers: dylanmckay, Jim, llvm-commits Tags: #llvm Differential Revision: https://reviews.llvm.org/D70438
* [SLP] add test for reduction miscompile; NFCSanjay Patel2019-11-191-0/+30
| | | | See D70148 for discussion.
* [lldb] Add logging to IRExecutionUnit::GetStaticInitializersRaphael Isemann2019-11-191-6/+21
|
* [ARM,MVE] Add intrinsics for scalar shifts.Simon Tatham2019-11-197-23/+606
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This fills in the small family of MVE intrinsics that have nothing to do with vectors: they implement bit-shift operations on 32- or 64-bit values held in one or two general-purpose registers. Most of these shift operations saturate if shifting left, and round to nearest if shifting right, although LSLL and ASRL behave like ordinary shifts. When these instructions take a variable shift count in a register, they pay attention to its sign, so that (for example) LSLL or UQRSHLL will shift left if given a positive number but right if given a negative one. That makes even LSLL and ASRL different enough from standard LLVM IR shift semantics that I couldn't see any better alternative than to simply model the whole family as a set of MVE-specific IR intrinsics. (The //immediate// forms of LSLL and ASRL, on the other hand, do behave exactly like a standard IR shift of a 64-bit value. In fact, those forms don't have ACLE intrinsics defined at all, because you can just write an ordinary C shift operation if you want one of those.) The 64-bit shifts have to be instruction-selected in C++, because they deliver two output values. But the 32-bit ones are simple enough that I could write a DAG isel pattern directly into each Instruction record. Reviewers: ostannard, MarkMurrayARM, dmgreen Reviewed By: dmgreen Subscribers: kristof.beyls, hiraditya, cfe-commits, llvm-commits Tags: #clang, #llvm Differential Revision: https://reviews.llvm.org/D70319
* [clangd] Show values of more expressions on hoverSam McCall2019-11-192-14/+83
| | | | | | | | | | Reviewers: kadircet Subscribers: ilya-biryukov, MaskRay, jkorous, arphaman, usaxena95, cfe-commits Tags: #clang Differential Revision: https://reviews.llvm.org/D70359
* AMDGPU: Refactor treatment of denormal modeMatt Arsenault2019-11-1919-90/+153
| | | | | | | | | | | Start moving towards treating this as a property of the calling convention, and not the subtarget. The default denormal mode should not be part of the subtarget, and be moved into a separate function attribute. This patch is still NFC. The denormal mode remains as a subtarget feature for now, but make the necessary changes to switch to using an attribute.
* AMDGPU: Be explicit about denormal mode in MIR testsMatt Arsenault2019-11-198-873/+1965
| | | | | | | Start checking the machine function in GlobalISel instead of the target directly. This temporarily breaks fcanonicalize selection in GlobalISel.
* [clang-tidy] DefaultOperatorNewCheck test fixes.Balázs Kéri2019-11-192-11/+11
|
* [clangd] Untangle Hover from XRefs, move into own file.Sam McCall2019-11-1913-1961/+2025
| | | | | | | | | | | | | | | | | | | | | | | | Summary: This is mostly mechanical, with a few exceptions: - getDeducedType moved into AST.h where it belongs. It now takes ASTContext instead of ParsedAST, and avoids using the preprocessor. - hover now uses SelectionTree directly rather than via getDeclAtPosition helper - hover on 'auto' used to find the decl that contained the 'auto' and use that to set Kind and documentation for the hover result. Now we use targetDecl() to find the decl matching the deduced type instead. This changes tests, e.g. 'variable' -> class for auto on lambdas. I think this is better, but the motivation was to avoid depending on the internals of DeducedTypeVisitor. This functionality is removed from the visitor. Reviewers: kadircet Subscribers: mgorny, ilya-biryukov, MaskRay, jkorous, arphaman, usaxena95, cfe-commits Tags: #clang Differential Revision: https://reviews.llvm.org/D70357
* [lldb][NFC] Early exit in IRExecutionUnit::GetStaticInitializersRaphael Isemann2019-11-191-24/+28
|
* DAG: Add function context to isFMAFasterThanFMulAndFAddMatt Arsenault2019-11-1919-28/+64
| | | | | | | | AMDGPU needs to know the FP mode for the function to answer this correctly when this is removed from the subtarget. AArch64 had to make this more complicated by using this from an IR hook, so add an IR typed overload.
OpenPOWER on IntegriCloud