[llvm-bugs] Issue 9019 in oss-fuzz: llvm/clang-fuzzer: Stack-overflow in clang::DiagnosticIDs::isUnrecoverable
Comment #1 on issue 9019 by ClusterFuzz-External: llvm/clang-fuzzer: Stack-overflow in clang::DiagnosticIDs::isUnrecoverable https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=9019#c1 ClusterFuzz has detected this issue as fixed in range 201806260706:201806270010. Detailed report: https://oss-fuzz.com/testcase?key=5106466699608064 Project: llvm Fuzzer: libFuzzer_llvm_clang-fuzzer Fuzz target binary: clang-fuzzer Job Type: libfuzzer_asan_llvm Platform Id: linux Crash Type: Stack-overflow Crash Address: 0x7ffc66484dc8 Crash State: clang::DiagnosticIDs::isUnrecoverable clang::DiagnosticIDs::ProcessDiag clang::DiagnosticsEngine::EmitCurrentDiagnostic Sanitizer: address (ASAN) Regressed: https://oss-fuzz.com/revisions?job=libfuzzer_asan_llvm&range=201801030610:201801040618 Fixed: https://oss-fuzz.com/revisions?job=libfuzzer_asan_llvm&range=201806260706:201806270010 Reproducer Testcase: https://oss-fuzz.com/download?testcase_id=5106466699608064 See https://github.com/google/oss-fuzz/blob/master/docs/reproducing.md for more information. If you suspect that the result above is incorrect, try re-doing that job on the test case report page. -- You received this message because: 1. You were specifically CC'd on the issue You may adjust your notification preferences at: https://bugs.chromium.org/hosting/settings Reply to this email to add a comment. ___ llvm-bugs mailing list llvm-bugs@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] Issue 9032 in oss-fuzz: llvm/clang-fuzzer: Stack-overflow in ConvertDeclSpecToType
Comment #2 on issue 9032 by ClusterFuzz-External: llvm/clang-fuzzer: Stack-overflow in ConvertDeclSpecToType https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=9032#c2 ClusterFuzz has detected this issue as fixed in range 201806260706:201806270010. Detailed report: https://oss-fuzz.com/testcase?key=575576908672 Project: llvm Fuzzer: libFuzzer_llvm_clang-fuzzer Fuzz target binary: clang-fuzzer Job Type: libfuzzer_asan_llvm Platform Id: linux Crash Type: Stack-overflow Crash Address: 0x7ffe322adf98 Crash State: ConvertDeclSpecToType GetDeclSpecTypeForDeclarator clang::Sema::GetTypeForDeclarator Sanitizer: address (ASAN) Regressed: https://oss-fuzz.com/revisions?job=libfuzzer_asan_llvm&range=201806202312:201806210752 Fixed: https://oss-fuzz.com/revisions?job=libfuzzer_asan_llvm&range=201806260706:201806270010 Reproducer Testcase: https://oss-fuzz.com/download?testcase_id=575576908672 See https://github.com/google/oss-fuzz/blob/master/docs/reproducing.md for more information. If you suspect that the result above is incorrect, try re-doing that job on the test case report page. -- You received this message because: 1. You were specifically CC'd on the issue You may adjust your notification preferences at: https://bugs.chromium.org/hosting/settings Reply to this email to add a comment. ___ llvm-bugs mailing list llvm-bugs@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] Issue 9032 in oss-fuzz: llvm/clang-fuzzer: Stack-overflow in ConvertDeclSpecToType
Updates: Labels: ClusterFuzz-Verified Status: Verified Comment #3 on issue 9032 by ClusterFuzz-External: llvm/clang-fuzzer: Stack-overflow in ConvertDeclSpecToType https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=9032#c3 ClusterFuzz testcase 575576908672 is verified as fixed, so closing issue as verified. If this is incorrect, please file a bug on https://github.com/google/oss-fuzz/issues/new -- You received this message because: 1. You were specifically CC'd on the issue You may adjust your notification preferences at: https://bugs.chromium.org/hosting/settings Reply to this email to add a comment. ___ llvm-bugs mailing list llvm-bugs@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] Issue 3954 in oss-fuzz: llvm/clang-fuzzer: ASSERT: D.isPastIdentifier() && "Haven't past the location of the identifier yet?"
Comment #8 on issue 3954 by ClusterFuzz-External: llvm/clang-fuzzer: ASSERT: D.isPastIdentifier() && "Haven't past the location of the identifier yet?" https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=3954#c8 ClusterFuzz has detected this issue as fixed in range 201806260706:201806270010. Detailed report: https://oss-fuzz.com/testcase?key=4843324394438656 Project: llvm Fuzzer: libFuzzer_llvm_clang-fuzzer Fuzz target binary: clang-fuzzer Job Type: libfuzzer_asan_llvm Platform Id: linux Crash Type: ASSERT Crash Address: Crash State: D.isPastIdentifier() && "Haven't past the location of the identifier yet?" clang::Parser::ParseDirectDeclarator clang::Parser::ParseDeclaratorInternal Sanitizer: address (ASAN) Fixed: https://oss-fuzz.com/revisions?job=libfuzzer_asan_llvm&range=201806260706:201806270010 Reproducer Testcase: https://oss-fuzz.com/download?testcase_id=4843324394438656 See https://github.com/google/oss-fuzz/blob/master/docs/reproducing.md for more information. If you suspect that the result above is incorrect, try re-doing that job on the test case report page. -- You received this message because: 1. You were specifically CC'd on the issue You may adjust your notification preferences at: https://bugs.chromium.org/hosting/settings Reply to this email to add a comment. ___ llvm-bugs mailing list llvm-bugs@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] Issue 9118 in oss-fuzz: llvm/clang-fuzzer: Stack-overflow in clang::APValue::swap
Comment #1 on issue 9118 by ClusterFuzz-External: llvm/clang-fuzzer: Stack-overflow in clang::APValue::swap https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=9118#c1 ClusterFuzz has detected this issue as fixed in range 201806260706:201806270010. Detailed report: https://oss-fuzz.com/testcase?key=6030335270715392 Project: llvm Fuzzer: libFuzzer_llvm_clang-fuzzer Fuzz target binary: clang-fuzzer Job Type: libfuzzer_asan_llvm Platform Id: linux Crash Type: Stack-overflow Crash Address: 0x7ffc416997d8 Crash State: clang::APValue::swap IntExprEvaluator::Success clang::StmtVisitorBasebool>::Visit Sanitizer: address (ASAN) Fixed: https://oss-fuzz.com/revisions?job=libfuzzer_asan_llvm&range=201806260706:201806270010 Reproducer Testcase: https://oss-fuzz.com/download?testcase_id=6030335270715392 See https://github.com/google/oss-fuzz/blob/master/docs/reproducing.md for more information. If you suspect that the result above is incorrect, try re-doing that job on the test case report page. -- You received this message because: 1. You were specifically CC'd on the issue You may adjust your notification preferences at: https://bugs.chromium.org/hosting/settings Reply to this email to add a comment. ___ llvm-bugs mailing list llvm-bugs@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] Issue 9067 in oss-fuzz: llvm/clang-fuzzer: Stack-overflow in clang::FunctionProtoType::getExtProtoInfo
Comment #1 on issue 9067 by ClusterFuzz-External: llvm/clang-fuzzer: Stack-overflow in clang::FunctionProtoType::getExtProtoInfo https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=9067#c1 ClusterFuzz has detected this issue as fixed in range 201806260706:201806270010. Detailed report: https://oss-fuzz.com/testcase?key=6565067356110848 Project: llvm Fuzzer: libFuzzer_llvm_clang-fuzzer Fuzz target binary: clang-fuzzer Job Type: libfuzzer_asan_llvm Platform Id: linux Crash Type: Stack-overflow Crash Address: 0x7fff838f6c08 Crash State: clang::FunctionProtoType::getExtProtoInfo clang::FunctionProtoType::Profile llvm::ContextualFoldingSetclang::ASTContext&>::NodeEq Sanitizer: address (ASAN) Fixed: https://oss-fuzz.com/revisions?job=libfuzzer_asan_llvm&range=201806260706:201806270010 Reproducer Testcase: https://oss-fuzz.com/download?testcase_id=6565067356110848 See https://github.com/google/oss-fuzz/blob/master/docs/reproducing.md for more information. If you suspect that the result above is incorrect, try re-doing that job on the test case report page. -- You received this message because: 1. You were specifically CC'd on the issue You may adjust your notification preferences at: https://bugs.chromium.org/hosting/settings Reply to this email to add a comment. ___ llvm-bugs mailing list llvm-bugs@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] Issue 3954 in oss-fuzz: llvm/clang-fuzzer: ASSERT: D.isPastIdentifier() && "Haven't past the location of the identifier yet?"
Updates: Labels: ClusterFuzz-Verified Status: Verified Comment #9 on issue 3954 by ClusterFuzz-External: llvm/clang-fuzzer: ASSERT: D.isPastIdentifier() && "Haven't past the location of the identifier yet?" https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=3954#c9 ClusterFuzz testcase 4843324394438656 is verified as fixed, so closing issue as verified. If this is incorrect, please file a bug on https://github.com/google/oss-fuzz/issues/new -- You received this message because: 1. You were specifically CC'd on the issue You may adjust your notification preferences at: https://bugs.chromium.org/hosting/settings Reply to this email to add a comment. ___ llvm-bugs mailing list llvm-bugs@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] Issue 9019 in oss-fuzz: llvm/clang-fuzzer: Stack-overflow in clang::DiagnosticIDs::isUnrecoverable
Updates: Labels: ClusterFuzz-Verified Status: Verified Comment #2 on issue 9019 by ClusterFuzz-External: llvm/clang-fuzzer: Stack-overflow in clang::DiagnosticIDs::isUnrecoverable https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=9019#c2 ClusterFuzz testcase 5106466699608064 is verified as fixed, so closing issue as verified. If this is incorrect, please file a bug on https://github.com/google/oss-fuzz/issues/new -- You received this message because: 1. You were specifically CC'd on the issue You may adjust your notification preferences at: https://bugs.chromium.org/hosting/settings Reply to this email to add a comment. ___ llvm-bugs mailing list llvm-bugs@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] Issue 9118 in oss-fuzz: llvm/clang-fuzzer: Stack-overflow in clang::APValue::swap
Updates: Labels: ClusterFuzz-Verified Status: Verified Comment #2 on issue 9118 by ClusterFuzz-External: llvm/clang-fuzzer: Stack-overflow in clang::APValue::swap https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=9118#c2 ClusterFuzz testcase 6030335270715392 is verified as fixed, so closing issue as verified. If this is incorrect, please file a bug on https://github.com/google/oss-fuzz/issues/new -- You received this message because: 1. You were specifically CC'd on the issue You may adjust your notification preferences at: https://bugs.chromium.org/hosting/settings Reply to this email to add a comment. ___ llvm-bugs mailing list llvm-bugs@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] Issue 9067 in oss-fuzz: llvm/clang-fuzzer: Stack-overflow in clang::FunctionProtoType::getExtProtoInfo
Updates: Labels: ClusterFuzz-Verified Status: Verified Comment #2 on issue 9067 by ClusterFuzz-External: llvm/clang-fuzzer: Stack-overflow in clang::FunctionProtoType::getExtProtoInfo https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=9067#c2 ClusterFuzz testcase 6565067356110848 is verified as fixed, so closing issue as verified. If this is incorrect, please file a bug on https://github.com/google/oss-fuzz/issues/new -- You received this message because: 1. You were specifically CC'd on the issue You may adjust your notification preferences at: https://bugs.chromium.org/hosting/settings Reply to this email to add a comment. ___ llvm-bugs mailing list llvm-bugs@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] [Bug 37881] lld should give warning for unknown -z option
https://bugs.llvm.org/show_bug.cgi?id=37881 George Rimar changed: What|Removed |Added Status|NEW |RESOLVED Resolution|--- |FIXED --- Comment #2 from George Rimar --- It was fixed by Rui in r335712. -- You are receiving this mail because: You are on the CC list for the bug.___ llvm-bugs mailing list llvm-bugs@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] [Bug 36768] Linker script OVERLAYs not supported
https://bugs.llvm.org/show_bug.cgi?id=36768 George Rimar changed: What|Removed |Added Resolution|--- |FIXED Status|ASSIGNED|RESOLVED --- Comment #12 from George Rimar --- r335714 -- You are receiving this mail because: You are on the CC list for the bug.___ llvm-bugs mailing list llvm-bugs@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] [Bug 37777] internal llvm error while building libcxx_fuzzer_x86_64
https://bugs.llvm.org/show_bug.cgi?id=3 Matthias Krüger changed: What|Removed |Added Resolution|--- |FIXED Status|NEW |RESOLVED --- Comment #1 from Matthias Krüger --- Looks like this has been fixed, I can't reproduce anymore. -- You are receiving this mail because: You are on the CC list for the bug.___ llvm-bugs mailing list llvm-bugs@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] [Bug 37950] New: ExecutionContext::GetByteOrder() always returns endian::InlHostByteOrder()
https://bugs.llvm.org/show_bug.cgi?id=37950
Bug ID: 37950
Summary: ExecutionContext::GetByteOrder() always returns
endian::InlHostByteOrder()
Product: lldb
Version: unspecified
Hardware: PC
OS: Linux
Status: NEW
Severity: normal
Priority: P
Component: All Bugs
Assignee: lldb-...@lists.llvm.org
Reporter: ramana.venka...@gmail.com
CC: llvm-bugs@lists.llvm.org
Created attachment 20471
--> https://bugs.llvm.org/attachment.cgi?id=20471&action=edit
Patch
lldb::ByteOrder ExecutionContext::GetByteOrder() const {
if (m_target_sp && m_target_sp->GetArchitecture().IsValid())
m_target_sp->GetArchitecture().GetByteOrder();
if (m_process_sp)
m_process_sp->GetByteOrder();
return endian::InlHostByteOrder();
}
As can be seen from the above piece of code, the byte order returned is always
endian::InlHostByteOrder(), which is not wrong.
--
You are receiving this mail because:
You are on the CC list for the bug.___
llvm-bugs mailing list
llvm-bugs@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] [Bug 37569] IsPowerOfTwo not treating INT_MIN as a negative number
https://bugs.llvm.org/show_bug.cgi?id=37569 Simon Pilgrim changed: What|Removed |Added Component|new bugs|Common Code Generator Code Fixed By Commit(s)||335717, 335719 Status|NEW |RESOLVED Resolution|--- |FIXED Product|new-bugs|libraries --- Comment #1 from Simon Pilgrim --- rL335717 fixed the issue by preventing the combine. rL335719 improved scalar/vector codegen by folding: SDIV(X, MIN_SIGNED) -> SELECT(X == MIN_SIGNED, 1, 0) -- You are receiving this mail because: You are on the CC list for the bug.___ llvm-bugs mailing list llvm-bugs@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] [Bug 37951] New: error: use of undefined value '%.split.split.split.[...].split.spl171 with "-bounds-checking -aggressive-instcombine -loop-unswitch -loop-rotate -licm -sroa -simple-loop-u
https://bugs.llvm.org/show_bug.cgi?id=37951 Bug ID: 37951 Summary: error: use of undefined value '%.split.split.split.[...].split.spl171 with "-bounds-checking -aggressive-instcombine -loop-unswitch -loop-rotate -licm -sroa -simple-loop-unswitch" Product: new-bugs Version: unspecified Hardware: PC OS: Linux Status: NEW Severity: enhancement Priority: P Component: new bugs Assignee: unassignedb...@nondot.org Reporter: mikael.hol...@ericsson.com CC: llvm-bugs@lists.llvm.org Created attachment 20472 --> https://bugs.llvm.org/attachment.cgi?id=20472&action=edit reproducer Running opt -S -o - bbi-15719.ll -bounds-checking -aggressive-instcombine -loop-unswitch -loop-rotate -licm -sroa -simple-loop-unswitch | opt -S gives opt: :1729:29: error: use of undefined value '%.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.spl171' %a.1.0 = phi i16 [ undef, %.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.split.spl171 ], [ %_tmp548, %bb1 ] Looking at the output from the first opt invocation there is indeed a use of the mentioned basic block at line 1729 that isn't defined anywhere. What's funny though is that this isn't caught by the verifier, so just doing opt -S -o - bbi-15719.ll -bounds-checking -aggressive-instcombine -loop-unswitch -loop-rotate -licm -sroa -simple-loop-unswitch - verify-each passes. Found during fuzz testing with random passes on random input. -- You are receiving this mail because: You are on the CC list for the bug.___ llvm-bugs mailing list llvm-bugs@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] [Bug 37952] New: Miscompile with -basicaa-recphi=1 -gvn
https://bugs.llvm.org/show_bug.cgi?id=37952 Bug ID: 37952 Summary: Miscompile with -basicaa-recphi=1 -gvn Product: new-bugs Version: trunk Hardware: PC OS: Linux Status: NEW Severity: enhancement Priority: P Component: new bugs Assignee: unassignedb...@nondot.org Reporter: bjorn.a.petters...@ericsson.com CC: llvm-bugs@lists.llvm.org Created attachment 20473 --> https://bugs.llvm.org/attachment.cgi?id=20473&action=edit reproducer The attached llvm-lit adapted test case fails for the RUN line using "-basicaa-recphi=1 -gvn". Afaict it is a miscompile. Note that -basicaa-recphi seems to have been defaulted to 0 since it was introduced in 2015 (r242320). If it turns out that the fault is related to that option, maybe the easy solution is to remove it? The problem was detected when doing fuzzy testing using randomized options (normally we don't use -basicaa-recphi=1). -- You are receiving this mail because: You are on the CC list for the bug.___ llvm-bugs mailing list llvm-bugs@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] [Bug 37953] New: [meta] [debugify] fix -check-debugify failures
https://bugs.llvm.org/show_bug.cgi?id=37953 Bug ID: 37953 Summary: [meta] [debugify] fix -check-debugify failures Product: new-bugs Version: trunk Hardware: PC OS: Windows NT Status: NEW Severity: enhancement Priority: P Component: new bugs Assignee: unassignedb...@nondot.org Reporter: greg.bedw...@sony.com CC: anastasis.gra...@gmail.com, dav...@freebsd.org, f...@flametop.co.uk, llvm-bugs@lists.llvm.org, paul_robin...@playstation.sony.com, v...@apple.com Depends on: 37741, 37942 This is a meta bug for tracking work related to fixing all known failures from debugify ( https://reviews.llvm.org/D40512 ). The motivation for this is in order to provide the best possible user experience when debugging optimized code by ensuring that optimization passes preserve debug information wherever possible. >From Bug 37942, Comment 1: > An error report from -check-debugify should always be considered a compiler > bug. It would be cumbersome if we couldn't rely on this -- so much so that > we should consider relaxing -check-debugify rather than allowing it to emit > false positives. Referenced Bugs: https://bugs.llvm.org/show_bug.cgi?id=37741 [Bug 37741] variable information lost from debug info after combine redundant instructions https://bugs.llvm.org/show_bug.cgi?id=37942 [Bug 37942] [debugify] fails from Combine redundant instructions and SROA -- You are receiving this mail because: You are on the CC list for the bug.___ llvm-bugs mailing list llvm-bugs@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] [Bug 37954] New: [debugify] mem2reg fail (ERROR: Instruction with empty DebugLoc)
https://bugs.llvm.org/show_bug.cgi?id=37954 Bug ID: 37954 Summary: [debugify] mem2reg fail (ERROR: Instruction with empty DebugLoc) Product: new-bugs Version: trunk Hardware: PC OS: Windows NT Status: NEW Severity: enhancement Priority: P Component: new bugs Assignee: unassignedb...@nondot.org Reporter: greg.bedw...@sony.com CC: llvm-bugs@lists.llvm.org Blocks: 37953 Created attachment 20474 --> https://bugs.llvm.org/attachment.cgi?id=20474&action=edit xx.ll Testcase from Stephen Wilks's work on auto test-case generation/reduction. The test-case originally failed in the loop vectorizer at -O3, but I spotted the mem2reg failure while trying to get the minimal set of passes to reproduce. $ clang --version clang version 7.0.0 (trunk 335708) (llvm/trunk 335721) Target: x86_64-pc-windows-msvc Thread model: posix InstalledDir: e:\work\upstream-llvm\build-vs2015-native-ninja\bin $ clang loop_vectorization.cpp -S -emit-llvm -O0 -Xclang -disable-O0-optnone -o xx.ll $ opt xx.ll -disable-output -debugify-each -mem2reg ERROR: Instruction with empty DebugLoc in function ?baz@@YAXPEAHH@Z -- %bar.addr.0 = phi i32 [ %bar, %entry ], [ %dec, %while.body ], !dbg !28 WARNING: Missing line 1 WARNING: Missing line 2 WARNING: Missing line 3 WARNING: Missing line 4 WARNING: Missing line 6 WARNING: Missing line 8 WARNING: Missing line 11 WARNING: Missing line 12 WARNING: Missing line 13 CheckFunctionDebugify [Promote Memory to Register]: FAIL CheckFunctionDebugify [Module Verifier]: PASS Referenced Bugs: https://bugs.llvm.org/show_bug.cgi?id=37953 [Bug 37953] [meta] [debugify] fix -check-debugify failures -- You are receiving this mail because: You are on the CC list for the bug.___ llvm-bugs mailing list llvm-bugs@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] [Bug 37955] New: [debugify] loop-vectorize fail
https://bugs.llvm.org/show_bug.cgi?id=37955
Bug ID: 37955
Summary: [debugify] loop-vectorize fail
Product: new-bugs
Version: trunk
Hardware: PC
OS: Windows NT
Status: NEW
Severity: enhancement
Priority: P
Component: new bugs
Assignee: unassignedb...@nondot.org
Reporter: greg.bedw...@sony.com
CC: anastasis.gra...@gmail.com, dav...@freebsd.org,
llvm-bugs@lists.llvm.org, v...@apple.com
Blocks: 37953
Created attachment 20475
--> https://bugs.llvm.org/attachment.cgi?id=20475&action=edit
.ll file
>From the same source example as in Bug 37954, from Stephen Wilks's work to
automatically find debugify failures.
$ clang -v
clang version 7.0.0 (trunk 335708) (llvm/trunk 335721)
Target: x86_64-pc-windows-msvc
Thread model: posix
InstalledDir: e:\work\upstream-llvm\build-vs2015-native-ninja\bin
$ cat loop_vectorization.cpp
void baz(int *foo, int bar) {
while (bar--)
foo[bar] = bar;
}
$ clang loop_vectorization.cpp -S -emit-llvm -O0 -Xclang -disable-O0-optnone -o
- | opt -S -O3 -opt-bisect-limit=59 -o loop-vectorize-input.ll 2>nul
$ opt loop-vectorize-input.ll -disable-output -debugify-each -loop-vectorize
ERROR: Instruction with empty DebugLoc in function ?baz@@YAXPEAHH@Z -- %index
= phi i64 [ 0, %vector.ph ], [ %index.next, %vector.body ]
ERROR: Instruction with empty DebugLoc in function ?baz@@YAXPEAHH@Z --
%vec.ind = phi <4 x i64> [ %induction, %vector.ph ], [ %vec.ind.next,
%vector.body ]
ERROR: Instruction with empty DebugLoc in function ?baz@@YAXPEAHH@Z --
%step.add = add <4 x i64> %vec.ind,
ERROR: Instruction with empty DebugLoc in function ?baz@@YAXPEAHH@Z --
%index.next = add i64 %index, 8
ERROR: Instruction with empty DebugLoc in function ?baz@@YAXPEAHH@Z --
%vec.ind.next = add <4 x i64> %step.add,
ERROR: Instruction with empty DebugLoc in function ?baz@@YAXPEAHH@Z -- %15 =
icmp eq i64 %index.next, %n.vec
ERROR: Instruction with empty DebugLoc in function ?baz@@YAXPEAHH@Z -- br i1
%15, label %middle.block, label %vector.body, !llvm.loop !30
ERROR: Instruction with empty DebugLoc in function ?baz@@YAXPEAHH@Z -- %cmp.n
= icmp eq i64 %0, %n.vec
ERROR: Instruction with empty DebugLoc in function ?baz@@YAXPEAHH@Z --
%bc.resume.val = phi i64 [ %ind.end, %middle.block ], [ %0,
%while.body.preheader ]
CheckFunctionDebugify [Loop Vectorization]: FAIL
CheckFunctionDebugify [Module Verifier]: PASS
$
Referenced Bugs:
https://bugs.llvm.org/show_bug.cgi?id=37953
[Bug 37953] [meta] [debugify] fix -check-debugify failures
--
You are receiving this mail because:
You are on the CC list for the bug.___
llvm-bugs mailing list
llvm-bugs@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] [Bug 37956] New: [debugify] deadargelim fail
https://bugs.llvm.org/show_bug.cgi?id=37956
Bug ID: 37956
Summary: [debugify] deadargelim fail
Product: new-bugs
Version: trunk
Hardware: PC
OS: Windows NT
Status: NEW
Severity: enhancement
Priority: P
Component: new bugs
Assignee: unassignedb...@nondot.org
Reporter: greg.bedw...@sony.com
CC: llvm-bugs@lists.llvm.org
Blocks: 37953
Created attachment 20476
--> https://bugs.llvm.org/attachment.cgi?id=20476&action=edit
ll reproducer
$ clang -v
clang version 7.0.0 (trunk 335708) (llvm/trunk 335721)
Target: x86_64-pc-windows-msvc
Thread model: posix
InstalledDir: e:\work\upstream-llvm\build-vs2015-native-ninja\bin
$ cat dead_argument_elimination.cpp
struct Foo {} *ff;
static Foo bar() { return *ff; }
Foo baz() { return bar(); }
$ clang dead_argument_elimination.cpp -O0 -Xclang -disable-O0-optnone -S
-emit-llvm -o - | opt -sroa -ipsccp -S -o deadargelim-input.ll
$ opt deadargelim-input.ll -debugify-each -deadargelim -disable-output
ERROR: Instruction with empty DebugLoc in function ?bar@@YA?AUFoo@@XZ -- ret
void
WARNING: Missing line 4
CheckModuleDebugify [Dead Argument Elimination]: FAIL
CheckFunctionDebugify [Module Verifier]: PASS
CheckFunctionDebugify [Module Verifier]: PASS
Referenced Bugs:
https://bugs.llvm.org/show_bug.cgi?id=37953
[Bug 37953] [meta] [debugify] fix -check-debugify failures
--
You are receiving this mail because:
You are on the CC list for the bug.___
llvm-bugs mailing list
llvm-bugs@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] [Bug 37958] New: PatchableFunction crashes on empty function
https://bugs.llvm.org/show_bug.cgi?id=37958
Bug ID: 37958
Summary: PatchableFunction crashes on empty function
Product: libraries
Version: trunk
Hardware: PC
OS: All
Status: NEW
Severity: enhancement
Priority: P
Component: Backend: X86
Assignee: unassignedb...@nondot.org
Reporter: apilipe...@azulsystems.com
CC: llvm-bugs@lists.llvm.org
I don't think that this is a problem for real world code, but it confuses
bugpoint and makes it reduce crashes down to empty method.
foo.ll:
define void @foo() "patchable-function"="prologue-short-redirect" {
unreachable
}
llc foo.ll
Stack dump:
0. Program arguments: /local/apilipenko/orca-2/build/buildR/bin/llc foo.ll
1. Running pass 'Function Pass Manager' on module 'foo.ll'.
2. Running pass 'Implement the 'patchable-function' attribute' on function
'@foo'
#0 0x7f34f54eff9a llvm::sys::PrintStackTrace(llvm::raw_ostream&)
(/local/apilipenko/orca-2/build/buildR/bin/../lib/liborca.so.0.dev+0x603f9a)
#1 0x7f34f54ee8da llvm::sys::RunSignalHandlers()
(/local/apilipenko/orca-2/build/buildR/bin/../lib/liborca.so.0.dev+0x6028da)
#2 0x7f34f54eea09 SignalHandler(int)
(/local/apilipenko/orca-2/build/buildR/bin/../lib/liborca.so.0.dev+0x602a09)
#3 0x7f34f4cdf130 __restore_rt (/lib64/libpthread.so.0+0xf130)
#4 0x7f34f57aad44 llvm::MachineInstr::getMF() const
(/local/apilipenko/orca-2/build/buildR/bin/../lib/liborca.so.0.dev+0x8bed44)
#5 0x7f34f575cd4c
llvm::ilist_traits::removeNodeFromList(llvm::MachineInstr*)
(/local/apilipenko/orca-2/build/buildR/bin/../lib/liborca.so.0.dev+0x870d4c)
#6 0x7f34f57ab03f llvm::MachineInstr::eraseFromParent()
(/local/apilipenko/orca-2/build/buildR/bin/../lib/liborca.so.0.dev+0x8bf03f)
#7 0x7f34f584b53f (anonymous
namespace)::PatchableFunction::runOnMachineFunction(llvm::MachineFunction&)
(/local/apilipenko/orca-2/build/buildR/bin/../lib/liborca.so.0.dev+0x95f53f)
#8 0x7f34f57a6ff9 llvm::MachineFunctionPass::runOnFunction(llvm::Function&)
(/local/apilipenko/orca-2/build/buildR/bin/../lib/liborca.so.0.dev+0x8baff9)
#9 0x7f34f55d0dba llvm::FPPassManager::runOnFunction(llvm::Function&)
(/local/apilipenko/orca-2/build/buildR/bin/../lib/liborca.so.0.dev+0x6e4dba)
#10 0x7f34f55d1614 llvm::legacy::PassManagerImpl::run(llvm::Module&)
(/local/apilipenko/orca-2/build/buildR/bin/../lib/liborca.so.0.dev+0x6e5614)
#11 0x004123d4 main
(/local/apilipenko/orca-2/build/buildR/bin/llc+0x4123d4)
#12 0x7f34f4100af5 __libc_start_main (/lib64/libc.so.6+0x21af5)
#13 0x004128fd _start
(/local/apilipenko/orca-2/build/buildR/bin/llc+0x4128fd)
Segmentation fault (core dumped)
PatchableFunction finds the first instruction which is code generated (skipping
DBG_LABEL, DBG_VALUE, KILL, etc.) and replaces it with PATCHABLE_OP ,
, , where and are the first instruction
it found. If the function is empty it can't do this trick.
MachineBasicBlock::iterator FirstActualI = FirstMBB.begin();
for (; doesNotGeneratecode(*FirstActualI); ++FirstActualI)
assert(FirstActualI != FirstMBB.end());
... uses FirstActualI ...
Locally for my bugpoint run I fixed it by ignoring such functions but it seem
to violate the contract of "patchable-function".
--
You are receiving this mail because:
You are on the CC list for the bug.___
llvm-bugs mailing list
llvm-bugs@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] [Bug 37957] New: Assertion failure in clang::Expr::ClassifyImpl(clang::ASTContext &, clang::SourceLocation *) const: getValueKind() == VK_XValue
https://bugs.llvm.org/show_bug.cgi?id=37957
Bug ID: 37957
Summary: Assertion failure in
clang::Expr::ClassifyImpl(clang::ASTContext &,
clang::SourceLocation *) const: getValueKind() ==
VK_XValue
Product: clang
Version: unspecified
Hardware: PC
OS: Linux
Status: NEW
Severity: enhancement
Priority: P
Component: Frontend
Assignee: unassignedclangb...@nondot.org
Reporter: ale...@google.com
CC: llvm-bugs@lists.llvm.org
$ cat test-clang__Expr__ClassifyImpl.cc
typedef long a;
class b {
public:
operator a();
};
struct c;
class C {
public:
c operator[](int);
};
struct c {
b d[];
};
C e;
void f() {
b g;
0 > e[0].d[g];
}
$ clang-check test-clang__Expr__ClassifyImpl.cc -- -std=c++11
assertion failed at llvm/tools/clang/lib/AST/ExprClassification.cpp:58 in Cl
clang::Expr::ClassifyImpl(clang::ASTContext &, clang::SourceLocation *) const:
getValueKind() == VK_XValue
@ 0x556be2ad5026 __assert_fail
@ 0x556be18551bc clang::Expr::ClassifyImpl()
@ 0x556be0f09a0f clang::Sema::AddConversionCandidate()
@ 0x556be0efc3e4 IsUserDefinedConversion()
@ 0x556be0f25316 TryUserDefinedConversion()
@ 0x556be0ef5f9e TryImplicitConversion()
@ 0x556be0f05801 TryCopyInitialization()
@ 0x556be0f0b21f clang::Sema::AddBuiltinCandidate()
@ 0x556be0f0fbb3 (anonymous
namespace)::BuiltinOperatorOverloadBuilder::addGenericBinaryArithmeticOverloads()
@ 0x556be0f0ebb5 clang::Sema::AddBuiltinOperatorCandidates()
@ 0x556be0f1d5ee clang::Sema::CreateOverloadedBinOp()
@ 0x556be10d29bd BuildOverloadedBinOp()
@ 0x556be109b889 clang::Sema::ActOnBinOp()
@ 0x556be0cd85a4 clang::Parser::ParseRHSOfBinaryExpression()
@ 0x556be0cd75bc clang::Parser::ParseAssignmentExpression()
@ 0x556be0cd749e clang::Parser::ParseExpression()
@ 0x556be0c94de7 clang::Parser::ParseExprStatement()
@ 0x556be0c93dec
clang::Parser::ParseStatementOrDeclarationAfterAttributes()
@ 0x556be0c9397e clang::Parser::ParseStatementOrDeclaration()
@ 0x556be0c9adb5 clang::Parser::ParseCompoundStatementBody()
@ 0x556be0c9b69a clang::Parser::ParseFunctionStatementBody()
@ 0x556be0c7f4e0 clang::Parser::ParseFunctionDefinition()
@ 0x556be0d063bc clang::Parser::ParseDeclGroup()
@ 0x556be0c7e4bf clang::Parser::ParseDeclOrFunctionDefInternal()
@ 0x556be0c7dd7c clang::Parser::ParseDeclarationOrFunctionDefinition()
@ 0x556be0c7cabc clang::Parser::ParseExternalDeclaration()
@ 0x556be0c7bebb clang::Parser::ParseTopLevelDecl()
@ 0x556be0c77195 clang::ParseAST()
@ 0x556be0ac6883 clang::FrontendAction::Execute()
--
You are receiving this mail because:
You are on the CC list for the bug.___
llvm-bugs mailing list
llvm-bugs@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] [Bug 37959] New: [debugify] globalopt fail
https://bugs.llvm.org/show_bug.cgi?id=37959
Bug ID: 37959
Summary: [debugify] globalopt fail
Product: new-bugs
Version: trunk
Hardware: PC
OS: Windows NT
Status: NEW
Severity: enhancement
Priority: P
Component: new bugs
Assignee: unassignedb...@nondot.org
Reporter: greg.bedw...@sony.com
CC: llvm-bugs@lists.llvm.org
Blocks: 37953
Created attachment 20477
--> https://bugs.llvm.org/attachment.cgi?id=20477&action=edit
ll reproducer
$ clang -v
clang version 7.0.0 (trunk 335708) (llvm/trunk 335721)
Target: x86_64-pc-windows-msvc
Thread model: posix
InstalledDir: e:\work\upstream-llvm\build-vs2015-native-ninja\bin
$ cat global_variable_optimizer.cpp
static int foo;
int bar() {
foo = 5;
return foo;
}
$ clang global_variable_optimizer.cpp -O0 -Xclang -disable-O0-optnone -S
-emit-llvm -o globalopt-input.ll
$ opt globalopt-input.ll -disable-output -debugify-each -globalopt
ERROR: Instruction with empty DebugLoc in function ?bar@@YAHXZ -- store i1
true, i1* @foo
ERROR: Instruction with empty DebugLoc in function ?bar@@YAHXZ -- %.b = load
i1, i1* @foo
ERROR: Instruction with empty DebugLoc in function ?bar@@YAHXZ -- %0 = select
i1 %.b, i32 5, i32 0
WARNING: Missing line 1
WARNING: Missing line 2
CheckModuleDebugify [Global Variable Optimizer]: FAIL
CheckFunctionDebugify [Module Verifier]: PASS
$
Referenced Bugs:
https://bugs.llvm.org/show_bug.cgi?id=37953
[Bug 37953] [meta] [debugify] fix -check-debugify failures
--
You are receiving this mail because:
You are on the CC list for the bug.___
llvm-bugs mailing list
llvm-bugs@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] [Bug 37960] New: [debugify] ipsccp fail
https://bugs.llvm.org/show_bug.cgi?id=37960
Bug ID: 37960
Summary: [debugify] ipsccp fail
Product: new-bugs
Version: trunk
Hardware: PC
OS: Windows NT
Status: NEW
Severity: enhancement
Priority: P
Component: new bugs
Assignee: unassignedb...@nondot.org
Reporter: greg.bedw...@sony.com
CC: anastasis.gra...@gmail.com, dav...@freebsd.org,
llvm-bugs@lists.llvm.org, v...@apple.com
Blocks: 37953
Created attachment 20478
--> https://bugs.llvm.org/attachment.cgi?id=20478&action=edit
ll reproducer
Silly testcase, but raising on the basis that no input (no matter how silly)
should be causing debugify fails.
$ clang -v
clang version 7.0.0 (trunk 335708) (llvm/trunk 335721)
Target: x86_64-pc-windows-msvc
Thread model: posix
InstalledDir: e:\work\upstream-llvm\build-vs2015-native-ninja\bin
$ cat interprocedural_sparse_conditional_constant_propagation.cpp
static void foo() {}
void bar() { foo; }
$ clang interprocedural_sparse_conditional_constant_propagation.cpp -O0 -Xclang
-disable-O0-optnone -S -emit-llvm -o ipsccp-input.ll
interprocedural_sparse_conditional_constant_propagation.cpp:2:14: warning:
expression result unused [-Wunused-value]
void bar() { foo; }
^~~
1 warning generated.
$ opt ipsccp-input.ll -disable-output -debugify-each -ipsccp
ERROR: Instruction with empty DebugLoc in function ?foo@@YAXXZ -- unreachable
WARNING: Missing line 2
CheckModuleDebugify [Interprocedural Sparse Conditional Constant Propagation]:
FAIL
CheckFunctionDebugify [Module Verifier]: PASS
CheckFunctionDebugify [Module Verifier]: PASS
$
Referenced Bugs:
https://bugs.llvm.org/show_bug.cgi?id=37953
[Bug 37953] [meta] [debugify] fix -check-debugify failures
--
You are receiving this mail because:
You are on the CC list for the bug.___
llvm-bugs mailing list
llvm-bugs@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] [Bug 37961] New: [debugify] simplifycfg fail
https://bugs.llvm.org/show_bug.cgi?id=37961
Bug ID: 37961
Summary: [debugify] simplifycfg fail
Product: new-bugs
Version: trunk
Hardware: PC
OS: Windows NT
Status: NEW
Severity: enhancement
Priority: P
Component: new bugs
Assignee: unassignedb...@nondot.org
Reporter: greg.bedw...@sony.com
CC: anastasis.gra...@gmail.com, dav...@freebsd.org,
llvm-bugs@lists.llvm.org, v...@apple.com
Blocks: 37953
Created attachment 20479
--> https://bugs.llvm.org/attachment.cgi?id=20479&action=edit
ll reproducer
$ clang -v
clang version 7.0.0 (trunk 335708) (llvm/trunk 335721)
Target: x86_64-pc-windows-msvc
Thread model: posix
InstalledDir: e:\work\upstream-llvm\build-vs2015-native-ninja\bin
$ cat simplify_the_cfg.cpp
void foo(int bar, int baz) {
for (;; bar++)
for (; bar; bar++)
if (baz)
;
}
$ clang simplify_the_cfg.cpp -O0 -Xclang -disable-O0-optnone -S -emit-llvm -o -
| opt -simplifycfg -instcombine -S -o simplifycfg-input.ll
$ opt simplifycfg-input.ll -disable-output -debugify-each -simplifycfg
ERROR: Instruction with empty DebugLoc in function ?foo@@YAXHH@Z -- %0 = load
i32, i32* %bar.addr, align 4, !dbg !30
ERROR: Instruction with empty DebugLoc in function ?foo@@YAXHH@Z -- %inc4 =
add nsw i32 %0, 1, !dbg !30
WARNING: Missing line 9
WARNING: Missing line 10
WARNING: Missing line 11
WARNING: Missing line 12
WARNING: Missing line 13
WARNING: Missing line 14
WARNING: Missing variable 5
WARNING: Missing variable 6
WARNING: Missing variable 7
WARNING: Missing variable 8
CheckFunctionDebugify [Simplify the CFG]: FAIL
CheckFunctionDebugify [Module Verifier]: PASS
$
Referenced Bugs:
https://bugs.llvm.org/show_bug.cgi?id=37953
[Bug 37953] [meta] [debugify] fix -check-debugify failures
--
You are receiving this mail because:
You are on the CC list for the bug.___
llvm-bugs mailing list
llvm-bugs@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] [Bug 37807] -fsplit-stack problems with split -> nosplit call
https://bugs.llvm.org/show_bug.cgi?id=37807 Than McIntosh changed: What|Removed |Added Resolution|--- |FIXED Status|NEW |RESOLVED --- Comment #1 from Than McIntosh --- Fix submitted in r335604. Closing this bug out now... -- You are receiving this mail because: You are on the CC list for the bug.___ llvm-bugs mailing list llvm-bugs@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] [Bug 37962] New: [debugify] callsite-splitting fail
https://bugs.llvm.org/show_bug.cgi?id=37962
Bug ID: 37962
Summary: [debugify] callsite-splitting fail
Product: new-bugs
Version: trunk
Hardware: PC
OS: Windows NT
Status: NEW
Severity: enhancement
Priority: P
Component: new bugs
Assignee: unassignedb...@nondot.org
Reporter: greg.bedw...@sony.com
CC: anastasis.gra...@gmail.com, dav...@freebsd.org,
llvm-bugs@lists.llvm.org, v...@apple.com
Blocks: 37953
Created attachment 20480
--> https://bugs.llvm.org/attachment.cgi?id=20480&action=edit
ll reproducer
$ clang -v
clang version 7.0.0 (trunk 335708) (llvm/trunk 335721)
Target: x86_64-pc-windows-msvc
Thread model: posix
InstalledDir: e:\work\upstream-llvm\build-vs2015-native-ninja\bin
$ cat call-site_splitting.cpp
int foo(int aa, int bb) {
return aa + bb;
}
void bar(int& cc, int dd) {
cc = foo(cc && 2 | cc, dd);
}
$ clang call-site_splitting.cpp -O0 -Xclang -disable-O0-optnone -emit-llvm -S
-o - | opt -instcombine -S -o callsite-splitting-input.ll
$ opt callsite-splitting-input.ll -disable-output -debugify-each
-callsite-splitting
CheckFunctionDebugify [Call-site splitting]: PASS
CheckFunctionDebugify [Module Verifier]: PASS
ERROR: Instruction with empty DebugLoc in function ?bar@@YAXAEAHH@Z --
%phi.call = phi i32 [ %call1, %land.rhs.split ], [ %call2, %entry.split ]
WARNING: Missing line 7
CheckFunctionDebugify [Call-site splitting]: FAIL
CheckFunctionDebugify [Module Verifier]: PASS
$
Referenced Bugs:
https://bugs.llvm.org/show_bug.cgi?id=37953
[Bug 37953] [meta] [debugify] fix -check-debugify failures
--
You are receiving this mail because:
You are on the CC list for the bug.___
llvm-bugs mailing list
llvm-bugs@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] [Bug 37964] New: [debugify] gvn fail
https://bugs.llvm.org/show_bug.cgi?id=37964
Bug ID: 37964
Summary: [debugify] gvn fail
Product: new-bugs
Version: trunk
Hardware: PC
OS: Windows NT
Status: NEW
Severity: enhancement
Priority: P
Component: new bugs
Assignee: unassignedb...@nondot.org
Reporter: greg.bedw...@sony.com
CC: anastasis.gra...@gmail.com, dav...@freebsd.org,
llvm-bugs@lists.llvm.org, v...@apple.com
Blocks: 37953
Created attachment 20482
--> https://bugs.llvm.org/attachment.cgi?id=20482&action=edit
ll reproducer
$ clang -v
clang version 7.0.0 (trunk 335708) (llvm/trunk 335721)
Target: x86_64-pc-windows-msvc
Thread model: posix
InstalledDir: e:\work\upstream-llvm\build-vs2015-native-ninja\bin
$ cat global_value_numbering.cpp
int foo;
char bar();
void baz() {
for (;;)
for (; foo; foo = bar())
;
}
$ clang global_value_numbering.cpp -O0 -Xclang -disable-O0-optnone -emit-llvm
-S -o - | opt -O3 -opt-bisect-limit=51 -S -o gvn-input.ll 2>nul
$ opt -disable-output -debugify-each -gvn gvn-input.ll
ERROR: Instruction with empty DebugLoc in function ?baz@@YAXXZ -- %.pr1 = phi
i32 [ 0, %for.cond ], [ %conv, %for.cond.loopexit.loopexit ]
CheckFunctionDebugify [Global Value Numbering]: FAIL
CheckFunctionDebugify [Module Verifier]: PASS
Referenced Bugs:
https://bugs.llvm.org/show_bug.cgi?id=37953
[Bug 37953] [meta] [debugify] fix -check-debugify failures
--
You are receiving this mail because:
You are on the CC list for the bug.___
llvm-bugs mailing list
llvm-bugs@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] [Bug 37938] Missed opportunity for BTR instruction
https://bugs.llvm.org/show_bug.cgi?id=37938 Craig Topper changed: What|Removed |Added Fixed By Commit(s)||r335754 Status|NEW |RESOLVED Resolution|--- |FIXED --- Comment #3 from Craig Topper --- Fixed in r335754 -- You are receiving this mail because: You are on the CC list for the bug.___ llvm-bugs mailing list llvm-bugs@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] [Bug 37966] New: [debugify] jump-threading fail
https://bugs.llvm.org/show_bug.cgi?id=37966
Bug ID: 37966
Summary: [debugify] jump-threading fail
Product: new-bugs
Version: trunk
Hardware: PC
OS: Windows NT
Status: NEW
Severity: enhancement
Priority: P
Component: new bugs
Assignee: unassignedb...@nondot.org
Reporter: greg.bedw...@sony.com
CC: llvm-bugs@lists.llvm.org
Blocks: 37953
Created attachment 20483
--> https://bugs.llvm.org/attachment.cgi?id=20483&action=edit
ll reproducer
$ clang -v
clang version 7.0.0 (trunk 335708) (llvm/trunk 335721)
Target: x86_64-pc-windows-msvc
Thread model: posix
InstalledDir: e:\work\upstream-llvm\build-vs2015-native-ninja\bin
$ cat jump_threading.cpp
void foo(int);
void bar(int& aa, int& bb, int cc) {
int dd[3];
while (cc)
dd[cc] = 2;
while (aa < 5) {
if(*dd)
bb = 0;
foo(bb);
}
}
$ clang jump_threading.cpp -O0 -Xclang -disable-O0-optnone -emit-llvm -S -o - |
opt -O3 -opt-bisect-limit=39 -S -o jump_threading-input.ll 2>nul
$ opt -disable-output -debugify-each -jump-threading jump_threading-input.ll
ERROR: Instruction with empty DebugLoc in function ?bar@@YAXAEAH0H@Z -- br
label %while.body
WARNING: Missing line 7
WARNING: Missing line 8
WARNING: Missing line 16
WARNING: Missing line 18
WARNING: Missing line 19
WARNING: Missing line 20
WARNING: Missing line 26
WARNING: Missing line 28
WARNING: Missing line 29
WARNING: Missing line 31
WARNING: Missing line 37
WARNING: Missing line 38
CheckFunctionDebugify [Jump Threading]: FAIL
CheckFunctionDebugify [Module Verifier]: PASS
Referenced Bugs:
https://bugs.llvm.org/show_bug.cgi?id=37953
[Bug 37953] [meta] [debugify] fix -check-debugify failures
--
You are receiving this mail because:
You are on the CC list for the bug.___
llvm-bugs mailing list
llvm-bugs@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] [Bug 37967] New: [debugify] memcpyopt fail
https://bugs.llvm.org/show_bug.cgi?id=37967
Bug ID: 37967
Summary: [debugify] memcpyopt fail
Product: new-bugs
Version: trunk
Hardware: PC
OS: Windows NT
Status: NEW
Severity: enhancement
Priority: P
Component: new bugs
Assignee: unassignedb...@nondot.org
Reporter: greg.bedw...@sony.com
CC: llvm-bugs@lists.llvm.org
Blocks: 37953
Created attachment 20484
--> https://bugs.llvm.org/attachment.cgi?id=20484&action=edit
ll reproducer
$ clang -v
clang version 7.0.0 (trunk 335708) (llvm/trunk 335721)
Target: x86_64-pc-windows-msvc
Thread model: posix
InstalledDir: e:\work\upstream-llvm\build-vs2015-native-ninja\bin
$ cat memcpy_optimization.cpp
struct Foo {
long f0;
long f1;
} *a;
int bar(Foo, Foo, int&c, Foo) {
c = bar(a[0], a[0], c, a[0]);
return c;
}
$ clang -target x86_64-unknown-unknown memcpy_optimization.cpp -O0 -Xclang
-disable-O0-optnone -emit-llvm -S -o memcpyopt-input.ll
$ opt -debugify-each -disable-output -instcombine -memcpyopt memcpyopt-input.ll
WARNING: Missing line 1
WARNING: Missing line 2
WARNING: Missing line 7
WARNING: Missing line 8
WARNING: Missing line 9
WARNING: Missing line 10
WARNING: Missing line 11
WARNING: Missing line 12
WARNING: Missing line 13
WARNING: Missing line 14
WARNING: Missing line 15
WARNING: Missing line 16
WARNING: Missing line 19
WARNING: Missing line 21
WARNING: Missing line 23
WARNING: Missing line 24
WARNING: Missing line 26
WARNING: Missing line 28
WARNING: Missing line 29
WARNING: Missing line 30
WARNING: Missing line 32
WARNING: Missing line 34
WARNING: Missing line 39
WARNING: Missing line 47
WARNING: Missing line 48
WARNING: Missing variable 1
WARNING: Missing variable 2
CheckFunctionDebugify [Combine redundant instructions]: PASS
ERROR: Instruction with empty DebugLoc in function _Z3bar3FooS_RiS_ --
%tmpcast = bitcast i8* %1 to %struct.Foo*
CheckFunctionDebugify [MemCpy Optimization]: FAIL
CheckFunctionDebugify [Module Verifier]: PASS
Referenced Bugs:
https://bugs.llvm.org/show_bug.cgi?id=37953
[Bug 37953] [meta] [debugify] fix -check-debugify failures
--
You are receiving this mail because:
You are on the CC list for the bug.___
llvm-bugs mailing list
llvm-bugs@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] [Bug 37573] AArch64 MachineOutliner incorrectly outlines use of r16/r17
https://bugs.llvm.org/show_bug.cgi?id=37573 Jessica Paquette changed: What|Removed |Added Fixed By Commit(s)||r335758 Resolution|--- |FIXED Assignee|unassignedb...@nondot.org |jpaque...@apple.com Status|NEW |RESOLVED --- Comment #11 from Jessica Paquette --- Fixed in r335758. https://reviews.llvm.org/rL335758 -- You are receiving this mail because: You are on the CC list for the bug.___ llvm-bugs mailing list llvm-bugs@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] [Bug 37968] New: Modules crash-on-invalid due to struct declaration in one module, struct definition in another, but no direct dependency between the two
https://bugs.llvm.org/show_bug.cgi?id=37968
Bug ID: 37968
Summary: Modules crash-on-invalid due to struct declaration in
one module, struct definition in another, but no
direct dependency between the two
Product: clang
Version: trunk
Hardware: PC
OS: Linux
Status: NEW
Severity: enhancement
Priority: P
Component: Modules
Assignee: unassignedclangb...@nondot.org
Reporter: dblai...@gmail.com
CC: dgre...@apple.com, llvm-bugs@lists.llvm.org,
richard-l...@metafoo.co.uk
This one is a bit involved to reproduce - at least I haven't been able to get
it down below 3 module interface units and a module implementation unit as
follows:
a.cppm:
export module a;
export struct a {
virtual ~a();
};
b.cppm:
export module b;
import a;
export struct b : a {
};
x.cppm:
export module x;
struct a;
export struct x {
a *m;
x();
}
x.cpp:
module x;
import b;
import a; // Buggily required import, likely (will be filed separately)
y::y() { new b(); }
compiled as:
$ clang++ -fmodules-ts -c -std=c++17 -Xclang -fmodules-codegen --precompile
a.cppm
$ clang++ -fmodules-ts -c -std=c++17 -Xclang -fmodules-codegen --precompile
b.cppm -fmodule-file=a.pcm
$ clang++ -fmodules-ts -c -std=c++17 -Xclang -fmodules-codegen --precompile
FruitBowl.cppm -fmodule-file=x.pcm
clang++-tot -fmodules-ts -c -std=c++17 x.cpp -fmodule-file=a.pcm
-fmodule-file=b.pcm -fmodule-file=x.pcm
The crash trace is as follows (apologies for any bad line wrapping):
clang-5.0:
/usr/local/google/home/blaikie/dev/llvm/src/tools/clang/lib/AST/Decl.cpp:1422:
clang::LinkageInfo clang::LinkageComputer::getLVFo
rDecl(const clang::NamedDecl *, clang::LVComputationKind): Assertion
`D->getCachedLinkage() == LV.getLinkage()' failed.
Stack dump:
0. Program arguments:
/usr/local/google/home/blaikie/dev/llvm/build/default/bin/clang-5.0 -cc1
-triple x86_64-unknown-linux-gnu -emit-o
bj -mrelax-all -disable-free -main-file-name FruitBowl.cpp -mrelocation-model
static -mthread-model posix -mdisable-fp-elim -fmath-errno -ma
sm-verbose -mconstructor-aliases -munwind-tables -fuse-init-array -target-cpu
x86-64 -dwarf-column-info -debugger-tuning=gdb -coverage-notes
-file
/usr/local/google/home/blaikie/dev/scratch/modules_cached_linkage_crash/FruitBowl.gcno
-resource-dir /usr/local/google/home/blaikie/de
v/llvm/build/default/lib/clang/7.0.0 -internal-isystem
/usr/lib/gcc/x86_64-linux-gnu/7.3.0/../../../../include/c++/7.3.0
-internal-isystem /
usr/lib/gcc/x86_64-linux-gnu/7.3.0/../../../../include/x86_64-linux-gnu/c++/7.3.0
-internal-isystem /usr/lib/gcc/x86_64-linux-gnu/7.3.0/../.
./../../include/x86_64-linux-gnu/c++/7.3.0 -internal-isystem
/usr/lib/gcc/x86_64-linux-gnu/7.3.0/../../../../include/c++/7.3.0/backward -int
ernal-isystem /usr/local/include -internal-isystem
/usr/local/google/home/blaikie/dev/llvm/build/default/lib/clang/7.0.0/include
-internal-e
xternc-isystem /usr/include/x86_64-linux-gnu -internal-externc-isystem /include
-internal-externc-isystem /usr/include -std=c++17 -fdeprecat
ed-macro -fdebug-compilation-dir
/usr/local/google/home/blaikie/dev/scratch/modules_cached_linkage_crash
-ferror-limit 19 -fmessage-length 1
40 -fmodules-ts -fno-implicit-modules -fmodule-file=AbstractFruit.pcm
-fmodule-file=Apple.pcm -fmodule-file=FruitBowl.pcm -fobjc-runtime=gcc
-fcxx-exceptions -fexceptions -fdiagnostics-show-option -fcolor-diagnostics -o
FruitBowl.o -x c++ FruitBowl.cpp
1. parser at end of file
2. Per-file LLVM IR generation
3.
/usr/local/google/home/blaikie/dev/scratch/modules_cached_linkage_crash/Apple.cppm:3:15:
Generating code for declaration 'b::~b'
#0 0x06ddf5a9 llvm::sys::PrintStackTrace(llvm::raw_ostream&)
/usr/local/google/home/blaikie/dev/llvm/src/lib/Support/Unix/Signals.in
c:490:11
#1 0x06ddf759 PrintStackTraceSignalHandler(void*)
/usr/local/google/home/blaikie/dev/llvm/src/lib/Support/Unix/Signals.inc:554:1
#2 0x06ddda46 llvm::sys::RunSignalHandlers()
/usr/local/google/home/blaikie/dev/llvm/src/lib/Support/Signals.cpp:66:5
#3 0x06ddfd87 SignalHandler(int)
/usr/local/google/home/blaikie/dev/llvm/src/lib/Support/Unix/Signals.inc:353:1
#4 0x7f2b1af8e0c0 __restore_rt
(/lib/x86_64-linux-gnu/libpthread.so.0+0x110c0)
#5 0x7f2b19b21fcf gsignal (/lib/x86_64-linux-gnu/libc.so.6+0x32fcf)
#6 0x7f2b19b233fa abort (/lib/x86_64-linux-gnu/libc.so.6+0x343fa)
#7 0x7f2b19b1ae37 (/lib/x86_64-linux-gnu/libc.so.6+0x2be37)
#8 0x7f2b19b1aee2 (/lib/x86_64-linux-gnu/libc.so.6+0x2bee2)
#9 0x0a920ff9 clang::LinkageComputer::getLVForDecl(clang::NamedDecl
const*, clang::LVComputationKind) /usr/local/google/home/blaikie
/dev/llvm/src/tools/clang/lib/AST/Decl.cpp:1422:5
#10 0x0a922593
clang::LinkageComputer::getDeclLinkageAndVisibility(clang::NamedDecl const*)
/u
[llvm-bugs] [Bug 37969] New: Incorrect ownership of implicit declarations of things like operator delete
https://bugs.llvm.org/show_bug.cgi?id=37969
Bug ID: 37969
Summary: Incorrect ownership of implicit declarations of things
like operator delete
Product: clang
Version: trunk
Hardware: PC
OS: Linux
Status: NEW
Severity: enhancement
Priority: P
Component: Modules
Assignee: unassignedclangb...@nondot.org
Reporter: dblai...@gmail.com
CC: dgre...@apple.com, llvm-bugs@lists.llvm.org,
richard-l...@metafoo.co.uk
foo.cppm:
export module foo
void f(int *x) {
delete x;
}
bar.cpp:
void operator delete(void*)noexcept(true);
$ clang++ -fmodules-ts -c -std=c++17 --precompile foo.cppm
$ clang++ -fmodules-ts -c -std=c++17 bar.cpp -fmodule-file=foo.pcm
bar.cpp:2:6: error: declaration of 'operator delete' in the global module
follows declaration in module foo
void operator delete(void *)noexcept(true);
^
note: previous declaration is here
1 error generated.
--
You are receiving this mail because:
You are on the CC list for the bug.___
llvm-bugs mailing list
llvm-bugs@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] [Bug 37957] Assertion failure in clang::Expr::ClassifyImpl(clang::ASTContext &, clang::SourceLocation *) const: getValueKind() == VK_XValue
https://bugs.llvm.org/show_bug.cgi?id=37957 Richard Smith changed: What|Removed |Added Resolution|--- |FIXED Status|NEW |RESOLVED --- Comment #1 from Richard Smith --- Fixed in r335779. -- You are receiving this mail because: You are on the CC list for the bug.___ llvm-bugs mailing list llvm-bugs@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] [Bug 37970] New: clang non-deterministic DEBUG_VALUE output, armv7a-linux-androideabi -mthumb -O2 -g
https://bugs.llvm.org/show_bug.cgi?id=37970 Bug ID: 37970 Summary: clang non-deterministic DEBUG_VALUE output, armv7a-linux-androideabi -mthumb -O2 -g Product: libraries Version: trunk Hardware: All OS: All Status: NEW Severity: normal Priority: P Component: Backend: ARM Assignee: unassignedb...@nondot.org Reporter: c...@google.com CC: llvm-bugs@lists.llvm.org Created attachment 20485 --> https://bugs.llvm.org/attachment.cgi?id=20485&action=edit output 1.s and 2.s have different @DEBUG_VALUE sequences To reproduce, modify and run the run.sh script in the test.zip. Repeated compilations of the same input file produced different @DEBUG_VALUE sequences in 1.s and 2.s. The problem occurred with -target armv7a-linux-androideabi -mthumb -O2 -g. Other targets, -O1, or -g1 seemed to produce deterministic output. I traced back to llvm change https://reviews.llvm.org/rL321862, which started to have such problem. I am not sure if that change is buggy or just exposed an older bug. The test case was reduced from Android open source code. It is required in Android release process to be able to generate identical output from the same input and compiler. We cannot release Android without a fix or reversing rL321862. -- You are receiving this mail because: You are on the CC list for the bug.___ llvm-bugs mailing list llvm-bugs@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] [Bug 37971] New: lld --fix-cortex-a53-843419 is broken
https://bugs.llvm.org/show_bug.cgi?id=37971 Bug ID: 37971 Summary: lld --fix-cortex-a53-843419 is broken Product: lld Version: unspecified Hardware: PC OS: Linux Status: NEW Severity: normal Priority: P Component: ELF Assignee: unassignedb...@nondot.org Reporter: pho...@chromium.org CC: llvm-bugs@lists.llvm.org See failure in https://fuchsia-review.googlesource.com/170083: ld.lld: /b/s/w/ir/kitchen-workdir/llvm-project/llvm/include/llvm/Support/Casting.h:106: static bool llvm::isa_impl_cl::doit(const From *) [To = lld::elf::InputSection, From = const lld::elf::SectionBase *]: Assertion `Val && "isa<> used on a null pointer"' failed. -- You are receiving this mail because: You are on the CC list for the bug.___ llvm-bugs mailing list llvm-bugs@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] Issue 9142 in oss-fuzz: llvm/llvm-opt-fuzzer--x86_64-loop_predication: ASSERT: i < getNumArgOperands() && "Out of bounds!"
Status: New Owner: CC: k...@google.com, masc...@google.com, jdevlieg...@apple.com, igm...@gmail.com, llvm-b...@lists.llvm.org, j...@chromium.org, v...@apple.com, mitchphi...@outlook.com, xpl...@gmail.com, akils...@apple.com Labels: ClusterFuzz Stability-Memory-AddressSanitizer Reproducible Engine-libfuzzer Proj-llvm Reported-2018-06-28 Type: Bug New issue 9142 by ClusterFuzz-External: llvm/llvm-opt-fuzzer--x86_64-loop_predication: ASSERT: i < getNumArgOperands() && "Out of bounds!" https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=9142 Detailed report: https://oss-fuzz.com/testcase?key=5475599811346432 Project: llvm Fuzzer: libFuzzer_llvm_llvm-opt-fuzzer--x86_64-loop_predication Fuzz target binary: llvm-opt-fuzzer--x86_64-loop_predication Job Type: libfuzzer_asan_llvm Platform Id: linux Crash Type: ASSERT Crash Address: Crash State: i < getNumArgOperands() && "Out of bounds!" llvm::UpgradeIntrinsicCall BitcodeReader::materialize Sanitizer: address (ASAN) Regressed: https://oss-fuzz.com/revisions?job=libfuzzer_asan_llvm&range=201806270836:201806271936 Reproducer Testcase: https://oss-fuzz.com/download?testcase_id=5475599811346432 Issue filed automatically. See https://github.com/google/oss-fuzz/blob/master/docs/reproducing.md for more information. When you fix this bug, please * mention the fix revision(s). * state whether the bug was a short-lived regression or an old bug in any stable releases. * add any other useful information. This information can help downstream consumers. If you need to contact the OSS-Fuzz team with a question, concern, or any other feedback, please file an issue at https://github.com/google/oss-fuzz/issues. -- You received this message because: 1. You were specifically CC'd on the issue You may adjust your notification preferences at: https://bugs.chromium.org/hosting/settings Reply to this email to add a comment. ___ llvm-bugs mailing list llvm-bugs@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] [Bug 37972] New: Building LLVM crashes computer when linking libLTO.so
https://bugs.llvm.org/show_bug.cgi?id=37972 Bug ID: 37972 Summary: Building LLVM crashes computer when linking libLTO.so Product: clang Version: 6.0 Hardware: PC OS: Linux Status: NEW Severity: normal Priority: P Component: -New Bugs Assignee: unassignedclangb...@nondot.org Reporter: mattreecebent...@gmail.com CC: llvm-bugs@lists.llvm.org [ 61%] Linking CXX shared library ../../lib/libLTO.so At this point the computer will (after several minutes) freeze, no mouse movement, nothing. I don't experience this on any other application or build. Computer: xubuntu 18 gcc 7.3 Core2 e8500 4GB ram 1.1GB swap file -- You are receiving this mail because: You are on the CC list for the bug.___ llvm-bugs mailing list llvm-bugs@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] Issue 9144 in oss-fuzz: llvm/llvm-isel-fuzzer--x86_64-O2: ASSERT: VT == MVT::i8 && "SetCC type must be 8-bit integer"
Status: New Owner: CC: k...@google.com, masc...@google.com, jdevlieg...@apple.com, igm...@gmail.com, llvm-b...@lists.llvm.org, j...@chromium.org, v...@apple.com, mitchphi...@outlook.com, xpl...@gmail.com, akils...@apple.com Labels: ClusterFuzz Stability-Memory-AddressSanitizer Reproducible Engine-libfuzzer Proj-llvm Reported-2018-06-28 Type: Bug New issue 9144 by ClusterFuzz-External: llvm/llvm-isel-fuzzer--x86_64-O2: ASSERT: VT == MVT::i8 && "SetCC type must be 8-bit integer" https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=9144 Detailed report: https://oss-fuzz.com/testcase?key=6323977889775616 Project: llvm Fuzzer: libFuzzer_llvm_llvm-isel-fuzzer--x86_64-O2 Fuzz target binary: llvm-isel-fuzzer--x86_64-O2 Job Type: libfuzzer_asan_llvm Platform Id: linux Crash Type: ASSERT Crash Address: Crash State: VT == MVT::i8 && "SetCC type must be 8-bit integer" llvm::X86TargetLowering::LowerSETCC llvm::X86TargetLowering::LowerSELECT Sanitizer: address (ASAN) Regressed: https://oss-fuzz.com/revisions?job=libfuzzer_asan_llvm&range=201806270836:201806271936 Reproducer Testcase: https://oss-fuzz.com/download?testcase_id=6323977889775616 Issue filed automatically. See https://github.com/google/oss-fuzz/blob/master/docs/reproducing.md for more information. When you fix this bug, please * mention the fix revision(s). * state whether the bug was a short-lived regression or an old bug in any stable releases. * add any other useful information. This information can help downstream consumers. If you need to contact the OSS-Fuzz team with a question, concern, or any other feedback, please file an issue at https://github.com/google/oss-fuzz/issues. -- You received this message because: 1. You were specifically CC'd on the issue You may adjust your notification preferences at: https://bugs.chromium.org/hosting/settings Reply to this email to add a comment. ___ llvm-bugs mailing list llvm-bugs@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] [Bug 37973] New: Wrong code generation for CTLZ pattern
https://bugs.llvm.org/show_bug.cgi?id=37973
Bug ID: 37973
Summary: Wrong code generation for CTLZ pattern
Product: new-bugs
Version: trunk
Hardware: PC
OS: Linux
Status: NEW
Severity: normal
Priority: P
Component: new bugs
Assignee: unassignedb...@nondot.org
Reporter: ovm...@gmail.com
CC: llvm-bugs@lists.llvm.org
The motivating example:
int lzcnt(int x)
{
int count = 0;
while (x > 0) {
count++;
x = x >> 1;
}
return count;
}
int main()
{
int x = 1;
int y = lzcnt(x);
printf("count = %d\n", y);
return 0;
}
This code is compiled as: clang test.c -O3 -march=core-avx2
The pattern is recognized and the output looks like: "count = 1". This is
wrong since the number of leading zeros for x = 1 is 31.
Inside test.ll we can see:
define dso_local i32 @main() local_unnamed_addr #1 {
entry:
%call1 = tail call i32 (i8*, ...) @printf(i8* getelementptr inbounds
([13 x i8], [13 x i8]* @.str, i64 0, i64 0),
i32 1)
ret i32 0
}
So, in my opinion, the problem is in the result substitution (i32 1 in
@printf), this is the result of lzcnt (number of leading 1 position), and not
quantity of leading zeros.
The same example but with scanf("%d", &x) in the main function produces the
following output for 1 as input: "count = 32". This output is also wrong.
Inside test.ll we can see:
%1 = load i32, i32* %x, align 4, !tbaa !3
%cmp4.i = icmp sgt i32 %1, 0
%2 = call i32 @llvm.ctlz.i32(i32 %1, i1 false) #5, !range !2
%3 = sub nsw i32 32, %2
Possibly it should be: %3 = sub nsw i32 31, %2
--
You are receiving this mail because:
You are on the CC list for the bug.___
llvm-bugs mailing list
llvm-bugs@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
