Changes in x86AssemblyInspectionEngine.cpp

4871dfc6 - [LLDB][NFC][Reliability] Fix uninitialized variables from Coverity scan. Part 2

Mon, 25 Jul 2022 22:52:21 +0000

[LLDB][NFC][Reliability] Fix uninitialized variables from Coverity scan. Part 2Improve LLDB reliability by fixing the following "uninitialized variables" static code inspection warnings fromscan.coverity.com:1476275, 1274012, 1455035, 1364789, 14542821467483, 1406152, 1406255, 1454837, 14544161467446, 1462022, 1461909, 1420566, 13272281367767, 1431254, 1467299, 1312678, 14317801454731, 1490403Differential Revision: https://reviews.llvm.org/D130528 List of files: /llvm-project-15.0.7/lldb/source/Plugins/UnwindAssembly/x86/x86AssemblyInspectionEngine.cpp

9877159d - Revert "[LLDB][NFC][Reliability] Fix uninitialized variables from Coverity scan. Part 2"

Tue, 26 Jul 2022 01:23:19 +0000

Revert "[LLDB][NFC][Reliability] Fix uninitialized variables from Coverity scan. Part 2"This reverts commit b9aedd94e6796e4b4866ab4c091b736b3db58cb7. List of files: /llvm-project-15.0.7/lldb/source/Plugins/UnwindAssembly/x86/x86AssemblyInspectionEngine.cpp

b9aedd94 - [LLDB][NFC][Reliability] Fix uninitialized variables from Coverity scan. Part 2

Mon, 25 Jul 2022 22:52:21 +0000

459cfa5e - [LLDB][NFC][Reliability] Fix uninitialized variables from Coverity scan

Wed, 20 Jul 2022 20:32:53 +0000

[LLDB][NFC][Reliability] Fix uninitialized variables from Coverity scanImprove LLDB reliability by fixing the following "uninitialized variables" static code inspection warnings fromscan.coverity.com:1094796 1095721 1095728 1095737 10957411095756 1095779 1095789 1095805 12145521229457 1232475 1274006 1274010 12934271364800 1364802 1364804 1364812 13648161374902 1374909 1384975 1399312 14204511431704 1454230 1454554 1454615 14545791454594 1454832 1457759 1458696 14619091467658 1487814 1487830 1487845Differential Revision: https://reviews.llvm.org/D130098 List of files: /llvm-project-15.0.7/lldb/source/Plugins/UnwindAssembly/x86/x86AssemblyInspectionEngine.cpp

80814287 - [lldb][NFC] Fix all formatting errors in .cpp file headers

Fri, 24 Jan 2020 07:23:27 +0000

[lldb][NFC] Fix all formatting errors in .cpp file headersSummary:A *.cpp file header in LLDB (and in LLDB) should like this:```//===-- TestUtilities.cpp -------------------------------------------------===//```However in LLDB most of our source files have arbitrary changes to this format andthese changes are spreading through LLDB as folks usually just use the existingsource files as templates for their new files (most notably the unnecessaryeditor language indicator `-*- C++ -*-` is spreading and in every reviewsomeone is pointing out that this is wrong, resulting in people pointing out that thisis done in the same way in other files).This patch removes most of these inconsistencies including the editor language indicators,all the different missing/additional '-' characters, files that center the file name, missingtrailing `===//` (mostly caused by clang-format breaking the line).Reviewers: aprantl, espindola, jfb, shafik, JDevlieghereReviewed By: JDevlieghereSubscribers: dexonsmith, wuzish, emaste, sdardis, nemanjai, kbarton, MaskRay, atanasyan, arphaman, jfb, abidh, jsji, JDevlieghere, usaxena95, lldb-commitsTags: #lldbDifferential Revision: https://reviews.llvm.org/D73258 List of files: /llvm-project-15.0.7/lldb/source/Plugins/UnwindAssembly/x86/x86AssemblyInspectionEngine.cpp

50703223 - Fix the unwinding plan augmentation from x86 assembly

Thu, 10 Oct 2019 13:23:09 +0000

Fix the unwinding plan augmentation from x86 assemblyUnwind plan augmentation should compute the plan row at offset x fromthe instruction before offset x, but currently we compute it from theinstruction at offset x. Note that this behavior is a regressionintroduced when moving the x86 assembly inspection engine to its ownfile(https://github.com/llvm/llvm-project/commit/1c9858b298d79ce82c45a2954096718b39550109#diff-375a2be066db6f34bb9a71442c9b71fcL913);the original version handled this properly by copying the previousinstruction out before advancing the instruction pointer.The relevant bug with more info is here: https://bugs.llvm.org/show_bug.cgi?id=43561Differential Revision: https://reviews.llvm.org/D68454Patch by Jaroslav Sevcik .llvm-svn: 374342 List of files: /llvm-project-15.0.7/lldb/source/Plugins/UnwindAssembly/x86/x86AssemblyInspectionEngine.cpp

3fd917d8 - Support Linux signal return trampolines in frame initialization

Fri, 19 Jul 2019 14:05:55 +0000

Support Linux signal return trampolines in frame initializationSummary:Add __kernel_rt_sigreturn to the list of trap handlers for Linux (it'sused as such on aarch64 at least), and __restore_rt as well (used onx86_64).Skip decrement-and-recompute for trap handlers inInitializeNonZerothFrame, as signal dispatch may point the child frame'sreturn address to the start of the return trampoline.Parse the 'S' flag for signal handlers from eh_frame augmentation, andpropagate it to the unwind plan.Reviewers: labath, jankratochvil, compnerd, jfb, jasonmolendaReviewed By: jasonmolendaSubscribers: clayborg, MaskRay, wuzish, nemanjai, kbarton, jrtc27, atanasyan, jsji, javed.absar, kristof.beyls, lldb-commitsTags: #lldbDifferential Revision: https://reviews.llvm.org/D63667llvm-svn: 366580 List of files: /llvm-project-15.0.7/lldb/source/Plugins/UnwindAssembly/x86/x86AssemblyInspectionEngine.cpp

ff918fb4 - Fix -Wsign-compare by explicit cast after r362557

Wed, 05 Jun 2019 01:49:06 +0000

Fix -Wsign-compare by explicit cast after r362557llvm-svn: 362570 List of files: /llvm-project-15.0.7/lldb/source/Plugins/UnwindAssembly/x86/x86AssemblyInspectionEngine.cpp

c93b9958 - Call abs to avoid signed/unsigned comparison warning.

Tue, 04 Jun 2019 22:46:20 +0000

Call abs to avoid signed/unsigned comparison warning.llvm-svn: 362557 List of files: /llvm-project-15.0.7/lldb/source/Plugins/UnwindAssembly/x86/x86AssemblyInspectionEngine.cpp

dbb4322e - [lldb] Fix out-of-bounds read after c3ea7c66fec021867e005ad1b02f3c7e80feaa85

Tue, 04 Jun 2019 15:27:19 +0000

[lldb] Fix out-of-bounds read after c3ea7c66fec021867e005ad1b02f3c7e80feaa85"Add support for mid-function epilogues on x86 that end in a non-local jump."Detected by asan.llvm-svn: 362510 List of files: /llvm-project-15.0.7/lldb/source/Plugins/UnwindAssembly/x86/x86AssemblyInspectionEngine.cpp

c3ea7c66 - Add support for mid-function epilogues on x86 that end in a non-local jump.

Mon, 03 Jun 2019 22:34:12 +0000

Add support for mid-function epilogues on x86 that end in a non-local jump.The x86 assembly inspection engine has code to support detecting amid-function epilogue that ends in a RET instruction; add support for recognizing an epilogue that ends in a JMP, and add a check that theunwind state has been restored to the original stack setup; reinstatethe post-prologue unwind state after this JMP instruction.The assembly inspection engine used for other architectures, UnwindAssemblyInstEmulation, detects mid-function epilogues by tracking branch instructions within the function and "forwards"the current unwind state to the targets of the branches. Ifan epilogue unwinds the stack and exits, followed by a branchtarget, we get back to the correct unwind state. The x86 unwinder should move to this same algorithm, or possibly evenlook at implementing an x86 instruction emulation plugin andget UnwindAssemblyInstEmulation to work for x86 too. I addeda branch instruction recognizier method that will be necessaryif we want to switch the algorithm.Differential Revision: https://reviews.llvm.org/D62764 llvm-svn: 362456 List of files: /llvm-project-15.0.7/lldb/source/Plugins/UnwindAssembly/x86/x86AssemblyInspectionEngine.cpp

796ac80b - Use std::make_shared in LLDB (NFC)

Mon, 11 Feb 2019 23:13:08 +0000

Use std::make_shared in LLDB (NFC)Unlike std::make_unique, which is only available since C++14,std::make_shared is available since C++11. Not only is std::make_shareda lot more readable compared to ::reset(new), it also performs a singleheap allocation for the object and control block.Differential revision: https://reviews.llvm.org/D57990llvm-svn: 353764 List of files: /llvm-project-15.0.7/lldb/source/Plugins/UnwindAssembly/x86/x86AssemblyInspectionEngine.cpp

d1307ec4 - Fix x86 return pattern detection

Sun, 10 Feb 2019 15:41:53 +0000

Fix x86 return pattern detectionSummary: Replace 0xc9 (LEAVE) with 0xcb (RETF) in ret_pattern_p(). Also put 0xc3 first, since it is the most common form and will match first.Reviewers: jasonmolendaReviewed By: jasonmolendaSubscribers: labath, lldb-commitsTags: #lldbDifferential Revision: https://reviews.llvm.org/D57928llvm-svn: 353643 List of files: /llvm-project-15.0.7/lldb/source/Plugins/UnwindAssembly/x86/x86AssemblyInspectionEngine.cpp

823c66b6 - [x64] Process the B field of the REX prefix correctly for the PUSH and POP

Wed, 06 Feb 2019 08:48:30 +0000

[x64] Process the B field of the REX prefix correctly for the PUSH and POPinstructionsSummary: This patch makes `x86AssemblyInspectionEngine` to process zero value ofthe `B` field of the `REX` prefix in a correct way for `PUSH` and `POP`instructions. MSVC sometimes emits `pushq %rbp` instruction as `0x40 0x55`, andit was not parsed correctly before.Reviewers: jasonmolenda, labathReviewed By: jasonmolenda, labathSubscribers: abidh, lldb-commitsTags: #lldbDifferential Revision: https://reviews.llvm.org/D57745llvm-svn: 353281 List of files: /llvm-project-15.0.7/lldb/source/Plugins/UnwindAssembly/x86/x86AssemblyInspectionEngine.cpp

2946cd70 - Update the file headers across all of the LLVM projects in the monorepo

Sat, 19 Jan 2019 08:50:56 +0000

Update the file headers across all of the LLVM projects in the monorepoto reflect the new license.We understand that people may be surprised that we're moving the headerentirely to discuss the new license. We checked this carefully with theFoundation's lawyer and we believe this is the correct approach.Essentially, all code in the project is now made available by the LLVMproject under our new license, so you will see that the license headersinclude that license only. Some of our contributors have contributedcode under our old license, and accordingly, we have retained a copy ofour old license notice in the top-level files in each project andrepository.llvm-svn: 351636 List of files: /llvm-project-15.0.7/lldb/source/Plugins/UnwindAssembly/x86/x86AssemblyInspectionEngine.cpp

a6682a41 - Simplify Boolean expressions

Sat, 15 Dec 2018 00:15:33 +0000

Simplify Boolean expressionsThis patch simplifies boolean expressions acorss LLDB. It was generatedusing clang-tidy with the following command:run-clang-tidy.py -checks='-*,readability-simplify-boolean-expr' -format -fix $PWDDifferential revision: https://reviews.llvm.org/D55584llvm-svn: 349215 List of files: /llvm-project-15.0.7/lldb/source/Plugins/UnwindAssembly/x86/x86AssemblyInspectionEngine.cpp

4538ed3b - [x86] Fix issues with a realigned stack in MSVC compiled applications

Tue, 30 Oct 2018 10:07:08 +0000

[x86] Fix issues with a realigned stack in MSVC compiled applicationsSummary:This patch fixes issues with a stack realignment.MSVC maintains two frame pointers (`ebx` and `ebp`) for a realigned stack - oneis used for access to function parameters, while another is used for access tolocals. To support this the patch:- adds an alternative frame pointer (`ebx`);- considers stack realignment instructions (e.g. `and esp, -32`);- along with CFA (Canonical Frame Address) which point to the position next to the saved return address (or to the first parameter on the stack) introduces AFA (Aligned Frame Address) which points to the position of the stack pointer right after realignment. AFA is used for access to registers saved after the realignment (see the test);Here is an example of the code with the realignment:```struct __declspec(align(256)) OverAligned { char c;};void foo(int foo_arg) { OverAligned oa_foo = { 1 }; auto aaa_foo = 1234;}void bar(int bar_arg) { OverAligned oa_bar = { 2 }; auto aaa_bar = 5678; foo(1111);}int main() { bar(2222); return 0;}```and here is the `bar` disassembly:```push ebxmov ebx, espsub esp, 8and esp, -100hadd esp, 4push ebpmov ebp, [ebx+4]mov [esp+4], ebpmov ebp, espsub esp, 200hmov byte ptr [ebp-200h], 2mov dword ptr [ebp-4], 5678push 1111 ; foo_argcall j_?foo@@YAXH@Z ; foo(int)add esp, 4mov esp, ebppop ebpmov esp, ebxpop ebxretn```Reviewers: labath, zturner, jasonmolenda, stella.stamenovaReviewed By: jasonmolendaSubscribers: abidh, lldb-commitsTags: #lldbDifferential Revision: https://reviews.llvm.org/D53435llvm-svn: 345577 List of files: /llvm-project-15.0.7/lldb/source/Plugins/UnwindAssembly/x86/x86AssemblyInspectionEngine.cpp

05097246 - Reflow paragraphs in comments.

Mon, 30 Apr 2018 16:49:04 +0000

Reflow paragraphs in comments.This is intended as a clean up after the big clang-format commit(r280751), which unfortunately resulted in many of the commentparagraphs in LLDB being very hard to read.FYI, the script I used was:import textwrapimport commandsimport osimport sysimport retmp = "%s.tmp"%sys.argv[1]out = open(tmp, "w+")with open(sys.argv[1], "r") as f: header = "" text = "" comment = re.compile(r'^( *//) ([^ ].*)$') special = re.compile(r'^((([A-Z]+[: ])|([0-9]+ )).*)|(.*;)$') for line in f: match = comment.match(line) if match and not special.match(match.group(2)): # skip intentionally short comments. if not text and len(match.group(2)) < 40: out.write(line) continue if text: text += " " + match.group(2) else: header = match.group(1) text = match.group(2) continue if text: filled = textwrap.wrap(text, width=(78-len(header)), break_long_words=False) for l in filled: out.write(header+" "+l+'\n') text = "" out.write(line)os.rename(tmp, sys.argv[1])Differential Revision: https://reviews.llvm.org/D46144llvm-svn: 331197 List of files: /llvm-project-15.0.7/lldb/source/Plugins/UnwindAssembly/x86/x86AssemblyInspectionEngine.cpp

f6208045 - The x86 instruction unwinder can be asked to disassemble non-instruction

Sat, 08 Jul 2017 00:12:15 +0000

The x86 instruction unwinder can be asked to disassemble non-instruction blocks of memory, and if the final bytes of that block look like a longx86 instruction, it can cause the llvm disassembler to read past the endof the buffer. Use the maximum allowed instruction length that we passto the llvm disassembler as a way to limit this to the size of the buffer.An example of how to trigger this is when lldb does a function call, itputs a breakpoint on the beginning of main() and uses that as the returnaddress from the function call. When we stop at that location, lldb maytry to find the first frame up the stack. Because this is on the firstinstruction of a function, it will get the word-size value at the stackpointer and assume that this was the caller's pc value. But this is randomstack memory and could point to anything - an object in memory, somethingin the data section, whatever. And if we have a symbol for that thing,we'll try to disassemble it.This was leading to infrequent crashes in customer scenarios; figured outwhat was happening with address sanitizer. llvm-svn: 307454 List of files: /llvm-project-15.0.7/lldb/source/Plugins/UnwindAssembly/x86/x86AssemblyInspectionEngine.cpp

8996346b - [UnwindAssembly/x86] Add support for "lea imm(%ebp), %esp" pattern

Thu, 29 Jun 2017 12:40:13 +0000

[UnwindAssembly/x86] Add support for "lea imm(%ebp), %esp" patternSummary:The instruction pattern:and $-16, %espsub $imm, %esp...lea imm(%ebp), %espappears when the compiler is realigning the stack (for example inmain(), or almost everywhere with -mstackrealign switch). The "and"instruction is very difficult to model, but that's not necessary, asthese frames are always %ebp-based (the compiler also needs a way torestore the original %esp). Therefore the plans we were generating forthese function were almost correct already. The only place we were doingit wrong were the last instructions of the epilogue (usually just"ret"), where we had to revert to %esp-based unwinding, as the %ebp hadbeen popped already.This was wrong because our "distance of esp from cfa" counter had pickedup the "sub" instruction (and incremented the counter) but it had notseen that the register was reset by the "lea" instruction.This patch fixes that shortcoming, and adds a test for handlingfunctions like this.I have not been able to tickle the compiler into producing a 64-bitfunction with this pattern, but I don't see a reason why it couldn'tproduce it, if it chose to, so I add a x86_64 test as well.Reviewers: jasonmolenda, tberghammerSubscribers: lldb-commitsDifferential Revision: https://reviews.llvm.org/D34750llvm-svn: 306666 List of files: /llvm-project-15.0.7/lldb/source/Plugins/UnwindAssembly/x86/x86AssemblyInspectionEngine.cpp