[llvm-bugs] [Bug 40946] New: Missed Loop vectorization for global loop variants
https://bugs.llvm.org/show_bug.cgi?id=40946
Bug ID: 40946
Summary: Missed Loop vectorization for global loop variants
Product: tools
Version: trunk
Hardware: PC
OS: Linux
Status: NEW
Severity: enhancement
Priority: P
Component: opt
Assignee: unassignedb...@nondot.org
Reporter: balarishi.bhog...@amd.com
CC: llvm-bugs@lists.llvm.org
If the loop variant variables are functional parameters of point type, then
loop vectorizations happens.
But, When loop variant variables are global, loop vectorization is not taking
place.
===
// Flag: -O3 -fno-unroll-loops -c
char *a, *b, *c;
void loop_not_vectorized()
{
for (int i = 0; i < 1800; i++) // Loop not vectorized
{
a[i] = b[i] * c[i];
}
}
===
// Flag: -O3 -fno-unroll-loops -c
void loop_vectorized(char *a, char *b, char *c)
{
for (int i = 0; i < 1800; i++) // Loop vectorized
{
a[i] = b[i] * c[i];
}
}
===
--
You are receiving this mail because:
You are on the CC list for the bug.___
llvm-bugs mailing list
llvm-bugs@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] [Bug 40947] New: LLDB compiled by gcc doesn't evaluate expressions on 32 bit
https://bugs.llvm.org/show_bug.cgi?id=40947 Bug ID: 40947 Summary: LLDB compiled by gcc doesn't evaluate expressions on 32 bit Product: lldb Version: 5.0 Hardware: PC OS: Linux Status: NEW Severity: normal Priority: P Component: All Bugs Assignee: lldb-...@lists.llvm.org Reporter: m.akse...@samsung.com CC: llvm-bugs@lists.llvm.org I'm using LLDB cross-compiled by gcc on ARM. When I'm trying to evaluate a simple expression (e.g. x+1), debugger gives an error message: "Expression can't be run, because there is no JIT compiled function". I think I've found the reason. In function lldb_private::IRExecutionUnit::GetRunnableInfo there is a cast operator from void* (32 bit in my case) to lldb::addr_t (64 bit). Pointer-to-integer conversion is implementation defined as per the C99 standard and gcc performs signed extension here. Further it compares with unsigned number (uintptr_t) in IRExecutionUnit::GetRemoteAddressForLocal and address can't be found. -- You are receiving this mail because: You are on the CC list for the bug.___ llvm-bugs mailing list llvm-bugs@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] [Bug 40948] New: Add help text for the --coverage option
https://bugs.llvm.org/show_bug.cgi?id=40948 Bug ID: 40948 Summary: Add help text for the --coverage option Product: clang Version: unspecified Hardware: All OS: All Status: NEW Severity: enhancement Priority: P Component: Documentation Assignee: unassignedclangb...@nondot.org Reporter: mcastelluc...@mozilla.com CC: llvm-bugs@lists.llvm.org, richard-l...@metafoo.co.uk -- You are receiving this mail because: You are on the CC list for the bug.___ llvm-bugs mailing list llvm-bugs@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] [Bug 40949] New: opt cannot terminate with "opt -mem2reg -loop-unswitch -reg2mem -licm -structurizecfg -reg2mem -licm"
https://bugs.llvm.org/show_bug.cgi?id=40949
Bug ID: 40949
Summary: opt cannot terminate with "opt -mem2reg -loop-unswitch
-reg2mem -licm -structurizecfg -reg2mem -licm"
Product: libraries
Version: trunk
Hardware: PC
OS: Linux
Status: NEW
Severity: normal
Priority: P
Component: Scalar Optimizations
Assignee: unassignedb...@nondot.org
Reporter: csz...@163.com
CC: llvm-bugs@lists.llvm.org
Created attachment 21558
--> https://bugs.llvm.org/attachment.cgi?id=21558&action=edit
.bc file of the source code
For this test case, opt cannot terminate. I have tested it using godbolt, and
the opt 6/7/truck versions all have this problem. The outputs of opt 6/7/truck
are as follows.
opt truck:
/compiler-explorer/c-preload/compiler-wrapper: line 21: 8512 Killed
"$@"
Compiler returned: 137
opt 7:
compiler-explorer/c-preload/compiler-wrapper: line 21: 13372 Killed
"$@"
Compiler returned: 137
opt 6:
/compiler-explorer/c-preload/compiler-wrapper: line 21: 18355 Killed
"$@"
Compiler returned: 137
The followings are my test environment and test steps.
$clang -v
clang version 9.0.0 (trunk 355281)
Target: x86_64-unknown-linux-gnu
Thread model: posix
InstalledDir: /home/jack-zhou/Documents/llvm/llvm_truck/llvm/build4/bin
Found candidate GCC installation: /usr/lib/gcc/i686-linux-gnu/8
Found candidate GCC installation: /usr/lib/gcc/x86_64-linux-gnu/5
Found candidate GCC installation: /usr/lib/gcc/x86_64-linux-gnu/5.5.0
Found candidate GCC installation: /usr/lib/gcc/x86_64-linux-gnu/6
Found candidate GCC installation: /usr/lib/gcc/x86_64-linux-gnu/6.5.0
Found candidate GCC installation: /usr/lib/gcc/x86_64-linux-gnu/7
Found candidate GCC installation: /usr/lib/gcc/x86_64-linux-gnu/7.3.0
Found candidate GCC installation: /usr/lib/gcc/x86_64-linux-gnu/8
Selected GCC installation: /usr/lib/gcc/x86_64-linux-gnu/7.3.0
Candidate multilib: .;@m64
Candidate multilib: 32;@m32
Candidate multilib: x32;@mx32
Selected multilib: .;@m64
$clang -O3 -c -emit-llvm -mllvm -disable-llvm-optzns small.c -o small.bc
$opt -mem2reg -loop-unswitch -reg2mem -licm -structurizecfg -reg2mem -licm
small.bc -o small-opt.bc
The following is the source code of small.c
#include "csmith.h"
a() {
int32_t b[8];
uint32_t c;
uint64_t d = 4073709551610;
for (;;) {
uint8_t e;
for (; 8;)
for (;; c++) {
if (b)
return 0;
if (d)
break;
}
}
}
--
You are receiving this mail because:
You are on the CC list for the bug.___
llvm-bugs mailing list
llvm-bugs@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] [Bug 40950] New: opt cannot terminate with "opt -loop-rotate -loop-extract"
https://bugs.llvm.org/show_bug.cgi?id=40950
Bug ID: 40950
Summary: opt cannot terminate with "opt -loop-rotate
-loop-extract"
Product: libraries
Version: trunk
Hardware: PC
OS: Linux
Status: NEW
Severity: enhancement
Priority: P
Component: Scalar Optimizations
Assignee: unassignedb...@nondot.org
Reporter: csz...@163.com
CC: llvm-bugs@lists.llvm.org
Created attachment 21559
--> https://bugs.llvm.org/attachment.cgi?id=21559&action=edit
.bc file of the source code
For this test case, opt cannot terminate. I have tested it using godbolt, and
the opt 6/7/truck versions all have this problem. The outputs of opt 6/7/truck
are as follows.
opt truck:
compiler-explorer/c-preload/compiler-wrapper: line 21: 27482 Killed
"$@"
Compiler returned: 137
opt 7:
compiler-explorer/c-preload/compiler-wrapper: line 21: 32256 Killed
"$@"
Compiler returned: 137
opt 6:
LLVM ERROR: out of memory
/compiler-explorer/c-preload/compiler-wrapper: line 21: 30493 Killed
"$@"
Compiler returned: 137
The followings are my test environment and test steps.
$clang -v
clang version 9.0.0 (trunk 355281)
Target: x86_64-unknown-linux-gnu
Thread model: posix
InstalledDir: /home/jack-zhou/Documents/llvm/llvm_truck/llvm/build4/bin
Found candidate GCC installation: /usr/lib/gcc/i686-linux-gnu/8
Found candidate GCC installation: /usr/lib/gcc/x86_64-linux-gnu/5
Found candidate GCC installation: /usr/lib/gcc/x86_64-linux-gnu/5.5.0
Found candidate GCC installation: /usr/lib/gcc/x86_64-linux-gnu/6
Found candidate GCC installation: /usr/lib/gcc/x86_64-linux-gnu/6.5.0
Found candidate GCC installation: /usr/lib/gcc/x86_64-linux-gnu/7
Found candidate GCC installation: /usr/lib/gcc/x86_64-linux-gnu/7.3.0
Found candidate GCC installation: /usr/lib/gcc/x86_64-linux-gnu/8
Selected GCC installation: /usr/lib/gcc/x86_64-linux-gnu/7.3.0
Candidate multilib: .;@m64
Candidate multilib: 32;@m32
Candidate multilib: x32;@mx32
Selected multilib: .;@m64
$clang -O3 -c -emit-llvm -mllvm -disable-llvm-optzns small.c -o small.bc
$opt -loop-rotate -loop-extract -o small-opt.bc
In this step, opt will consume all RAM for a computer with 16GB RAM and not
terminate.
The followings are the source code of small.c and IR
int a;
c(d) {
for (; a;)
if (d)
break;
}
--
; ModuleID = 'small.bc'
source_filename = "small.c"
target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128"
target triple = "x86_64-unknown-linux-gnu"
@a = common dso_local global i32 0, align 4
; Function Attrs: nounwind uwtable
define dso_local i32 @c(i32) #0 {
%2 = alloca i32, align 4
%3 = alloca i32, align 4
store i32 %0, i32* %3, align 4, !tbaa !2
br label %4
; :4: ; preds = %11, %1
%5 = load i32, i32* @a, align 4, !tbaa !2
%6 = icmp ne i32 %5, 0
br i1 %6, label %7, label %12
; :7: ; preds = %4
%8 = load i32, i32* %3, align 4, !tbaa !2
%9 = icmp ne i32 %8, 0
br i1 %9, label %10, label %11
; :10: ; preds = %7
br label %12
; :11: ; preds = %7
br label %4
; :12: ; preds = %10, %4
%13 = load i32, i32* %2, align 4
ret i32 %13
}
attributes #0 = { nounwind uwtable
"correctly-rounded-divide-sqrt-fp-math"="false" "disable-tail-calls"="false"
"less-precise-fpmad"="false" "min-legal-vector-width"="0"
"no-frame-pointer-elim"="false" "no-infs-fp-math"="false"
"no-jump-tables"="false" "no-nans-fp-math"="false"
"no-signed-zeros-fp-math"="false" "no-trapping-math"="false"
"stack-protector-buffer-size"="8" "target-cpu"="x86-64"
"target-features"="+fxsr,+mmx,+sse,+sse2,+x87" "unsafe-fp-math"="false"
"use-soft-float"="false" }
!llvm.module.flags = !{!0}
!llvm.ident = !{!1}
!0 = !{i32 1, !"wchar_size", i32 4}
!1 = !{!"clang version 9.0.0 (trunk 355281)"}
!2 = !{!3, !3, i64 0}
!3 = !{!"int", !4, i64 0}
!4 = !{!"omnipotent char", !5, i64 0}
!5 = !{!"Simple C/C++ TBAA"}
--
You are receiving this mail because:
You are on the CC list for the bug.___
llvm-bugs mailing list
llvm-bugs@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] [Bug 40951] New: llvm/test/Transforms/SLPVectorizer/X86/phi.ll test is failing with release_80 branch
https://bugs.llvm.org/show_bug.cgi?id=40951 Bug ID: 40951 Summary: llvm/test/Transforms/SLPVectorizer/X86/phi.ll test is failing with release_80 branch Product: Test Suite Version: 8.0 Hardware: PC OS: Linux Status: NEW Severity: release blocker Priority: P Component: lit Assignee: unassignedb...@nondot.org Reporter: balarishi.bhog...@amd.com CC: dan...@zuster.org, llvm-bugs@lists.llvm.org With release_80 branch, facing the below failure for llvm/test/Transforms/SLPVectorizer/X86/phi.ll: FAIL: LLVM :: Transforms/SLPVectorizer/X86/phi.ll (27008 of 29439) TEST 'LLVM :: Transforms/SLPVectorizer/X86/phi.ll' FAILED Script: -- : 'RUN: at line 2'; /home/amd/JENKINS/workspace/CPUPC_AMD_LLVM-8-0-0-ORG_Branch/BUILD_DIR/bin/opt < /home/amd/JENKINS/workspace/CPUPC_AMD_LLVM-8-0-0-ORG_Branch/llvm/test/Transforms/SLPVectorizer/X86/phi.ll -basicaa -slp-vectorizer -slp-threshold=-100 -dce -S -mtriple=i386-apple-macosx10.8.0 -mcpu=corei7-avx | /home/amd/JENKINS/workspace/CPUPC_AMD_LLVM-8-0-0-ORG_Branch/BUILD_DIR/bin/FileCheck /home/amd/JENKINS/workspace/CPUPC_AMD_LLVM-8-0-0-ORG_Branch/llvm/test/Transforms/SLPVectorizer/X86/phi.ll -- Exit Code: 1 Command Output (stderr): -- /home/amd/JENKINS/workspace/CPUPC_AMD_LLVM-8-0-0-ORG_Branch/llvm/test/Transforms/SLPVectorizer/X86/phi.ll:147:15: error: CHECK-NEXT: expected string not found in input ; CHECK-NEXT: [[REORDER_SHUFFLE:%.*]] = shufflevector <4 x float> [[TMP2]], <4 x float> undef, <4 x i32> ^ :52:2: note: scanning from here %3 = extractelement <4 x float> %2, i32 0 ^ :52:2: note: with variable "TMP2" equal to "%2" %3 = extractelement <4 x float> %2, i32 0 ^ -- Please let me know if additional details are required. -- You are receiving this mail because: You are on the CC list for the bug.___ llvm-bugs mailing list llvm-bugs@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] [Bug 40331] [meta] 8.0.0 Release Blockers
https://bugs.llvm.org/show_bug.cgi?id=40331 Bug 40331 depends on bug 40891, which changed state. Bug 40891 Summary: [X86] DAG combine infinite loop caused by combineTruncatedArithmetic trying to truncate a bitcast of a build vector of constants https://bugs.llvm.org/show_bug.cgi?id=40891 What|Removed |Added Status|CONFIRMED |RESOLVED Resolution|--- |FIXED -- You are receiving this mail because: You are on the CC list for the bug.___ llvm-bugs mailing list llvm-bugs@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] [Bug 40891] [X86] DAG combine infinite loop caused by combineTruncatedArithmetic trying to truncate a bitcast of a build vector of constants
https://bugs.llvm.org/show_bug.cgi?id=40891 Hans Wennborg changed: What|Removed |Added Resolution|--- |FIXED Status|CONFIRMED |RESOLVED --- Comment #8 from Hans Wennborg --- Merged to release_80 in r355310. -- You are receiving this mail because: You are on the CC list for the bug.___ llvm-bugs mailing list llvm-bugs@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] [Bug 40909] Cherry-pick r352465 to 8.0
https://bugs.llvm.org/show_bug.cgi?id=40909 Hans Wennborg changed: What|Removed |Added Status|NEW |RESOLVED Resolution|--- |FIXED --- Comment #2 from Hans Wennborg --- Merged to release_80 in r355311. -- You are receiving this mail because: You are on the CC list for the bug.___ llvm-bugs mailing list llvm-bugs@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] [Bug 40331] [meta] 8.0.0 Release Blockers
https://bugs.llvm.org/show_bug.cgi?id=40331 Bug 40331 depends on bug 40909, which changed state. Bug 40909 Summary: Cherry-pick r352465 to 8.0 https://bugs.llvm.org/show_bug.cgi?id=40909 What|Removed |Added Status|NEW |RESOLVED Resolution|--- |FIXED -- You are receiving this mail because: You are on the CC list for the bug.___ llvm-bugs mailing list llvm-bugs@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] [Bug 40662] [AMDGPU][MC] Enable lds_direct operand for v_readfirstlane_b32, v_readlane_b32 and v_writelane_b32
https://bugs.llvm.org/show_bug.cgi?id=40662 Dmitry changed: What|Removed |Added Resolution|--- |FIXED Fixed By Commit(s)||355312 Status|NEW |RESOLVED -- You are receiving this mail because: You are on the CC list for the bug.___ llvm-bugs mailing list llvm-bugs@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] [Bug 40331] [meta] 8.0.0 Release Blockers
https://bugs.llvm.org/show_bug.cgi?id=40331 Bug 40331 depends on bug 40907, which changed state. Bug 40907 Summary: Cherry-pick r355136 to 8.0 branch https://bugs.llvm.org/show_bug.cgi?id=40907 What|Removed |Added Status|NEW |RESOLVED Resolution|--- |FIXED -- You are receiving this mail because: You are on the CC list for the bug.___ llvm-bugs mailing list llvm-bugs@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] [Bug 40907] Cherry-pick r355136 to 8.0 branch
https://bugs.llvm.org/show_bug.cgi?id=40907 Hans Wennborg changed: What|Removed |Added Resolution|--- |FIXED Status|NEW |RESOLVED --- Comment #2 from Hans Wennborg --- Merged in r355313. -- You are receiving this mail because: You are on the CC list for the bug.___ llvm-bugs mailing list llvm-bugs@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] [Bug 40952] New: clang crashes if -save-temps is used with -fropi/-frwpi (ARM)
https://bugs.llvm.org/show_bug.cgi?id=40952 Bug ID: 40952 Summary: clang crashes if -save-temps is used with -fropi/-frwpi (ARM) Product: clang Version: trunk Hardware: PC OS: Windows NT Status: NEW Severity: enhancement Priority: P Component: Driver Assignee: unassignedclangb...@nondot.org Reporter: konstantin.schw...@hightec-rt.com CC: llvm-bugs@lists.llvm.org, neeil...@live.com, richard-l...@metafoo.co.uk If clang is invoked with both -save-temps and -fropi/-frwpi, the integrated assembler hits an assertion: llvm-project/clang/tools/driver/cc1as_main.cpp:381: bool ExecuteAssembler((anonymous namespace)::AssemblerInvocation &, clang::DiagnosticsEngine &): Assertion `Opts.RelocationModel == "dynamic-no-pic" && "Invalid PIC model!"' 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 https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] [Bug 40953] New: Single-line version of // clang-format off
https://bugs.llvm.org/show_bug.cgi?id=40953 Bug ID: 40953 Summary: Single-line version of // clang-format off Product: clang Version: trunk Hardware: PC OS: Linux Status: NEW Severity: enhancement Priority: P Component: Formatter Assignee: unassignedclangb...@nondot.org Reporter: dr...@jwdt.org CC: djas...@google.com, kli...@google.com, llvm-bugs@lists.llvm.org With the current commands // clang-format off and // clang-format on, it involves some overhead to disable the formatting only on a single line of code, since I need at least a second line containing only the command // clang-format on I would suggest to add another special comment // clang-format ignore-line, which leaves the current line of code unchanged, except that it indents it to the current brace level. -- You are receiving this mail because: You are on the CC list for the bug.___ llvm-bugs mailing list llvm-bugs@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] [Bug 32021] Clang-CL can not compile TraceLoggingProvider.h
https://bugs.llvm.org/show_bug.cgi?id=32021 Stephen Kelly changed: What|Removed |Added Status|REOPENED|RESOLVED Resolution|--- |FIXED --- Comment #10 from Stephen Kelly --- Thanks for checking and updating. I think that's good enough. -- You are receiving this mail because: You are on the CC list for the bug.___ llvm-bugs mailing list llvm-bugs@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] Issue 11790 in oss-fuzz: llvm/llvm-opt-fuzzer--x86_64-simplifycfg: Out-of-memory in llvm_llvm-opt-fuzzer--x86_64-simplifycfg
Updates: Labels: Deadline-Approaching Comment #4 on issue 11790 by sheriff...@chromium.org: llvm/llvm-opt-fuzzer--x86_64-simplifycfg: Out-of-memory in llvm_llvm-opt-fuzzer--x86_64-simplifycfg https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=11790#c4 This bug is approaching its deadline for being fixed, and will be automatically derestricted within 7 days. If a fix is planned within 2 weeks after the deadline has passed, a grace extension can be granted. - Your friendly Sheriffbot -- 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 https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] [Bug 40954] New: LICM is moving an fdiv instruction out from under a control dependence
https://bugs.llvm.org/show_bug.cgi?id=40954 Bug ID: 40954 Summary: LICM is moving an fdiv instruction out from under a control dependence Product: libraries Version: trunk Hardware: PC OS: Windows NT Status: NEW Severity: normal Priority: P Component: Loop Optimizer Assignee: unassignedb...@nondot.org Reporter: eric.schwe...@pgroup.com CC: llvm-bugs@lists.llvm.org Created attachment 21560 --> https://bugs.llvm.org/attachment.cgi?id=21560&action=edit reproducer of LICM moving an fdiv instruction In the attached reproducer, the instruction %33 is control dependent upon the branch in block L.LB2_498. When we run the following command, % opt -licm bug.ll -o bug.llvm and look at the output file, we see that the fdiv becomes %28 and has been moved to block L.LB2_555, which is a predecessor of L.LB2_498. This code motion introduces a floating-point exception into the flow path which was previously guarded by the conditional branch. -- You are receiving this mail because: You are on the CC list for the bug.___ llvm-bugs mailing list llvm-bugs@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] [Bug 40955] New: lists.llvm.org accepts passwords and is not requiring an SSL connection.
https://bugs.llvm.org/show_bug.cgi?id=40955 Bug ID: 40955 Summary: lists.llvm.org accepts passwords and is not requiring an SSL connection. Product: Website Version: unspecified Hardware: PC OS: All Status: NEW Severity: normal Priority: P Component: General Website Assignee: unassignedb...@nondot.org Reporter: m...@sqlby.me CC: llvm-bugs@lists.llvm.org, m...@sqlby.me Inital report from Jonny Grant: This page accepts passwords without being on a secure connection. Can llvm buy the SSL certificate and simply redirect from http? http://lists.llvm.org/cgi-bin/mailman/options/cfe-dev Interestingly. There does seem to be an SSL certificate on the server: https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-dev the problem is that: 1. The http server doesn't refresh to https 2. the main clang page links to http https://clang.llvm.org/get_involved.html 3. the mailmain https page even links back to plain http archive! http://lists.llvm.org/pipermail/cfe-dev/ 4. Even entering https://lists.llvm.org/ HTTPS redirects back to http://lists.llvm.org/mailman/listinfo ! Fix is pretty easy, take mailman off http server http://lists.llvm.org/ and put a 302 redirect back to https://lists.llvm.org/ Cheers, Jonny -- You are receiving this mail because: You are on the CC list for the bug.___ llvm-bugs mailing list llvm-bugs@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] [Bug 40926] opt crashes with "with "opt -bdce -flattencfg" : Assertion `Val && "isa<> used on a null pointer"' failed.
https://bugs.llvm.org/show_bug.cgi?id=40926 Sanjay Patel changed: What|Removed |Added CC||spatel+l...@rotateright.com Status|NEW |RESOLVED Resolution|--- |DUPLICATE --- Comment #1 from Sanjay Patel --- Like bug 40925, this doesn't need -bdce to show the bug. The bug is visible by sending just the correct output from -bdce to -flattencfg (which might be an unused/dead pass at this point). *** This bug has been marked as a duplicate of bug 40925 *** -- You are receiving this mail because: You are on the CC list for the bug.___ llvm-bugs mailing list llvm-bugs@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] [Bug 40927] opt crashes with "opt -gvn -flattencfg" : Assertion `Val && "isa<> used on a null pointer"' failed.
https://bugs.llvm.org/show_bug.cgi?id=40927 Sanjay Patel changed: What|Removed |Added Resolution|--- |DUPLICATE Status|NEW |RESOLVED CC||spatel+l...@rotateright.com --- Comment #1 from Sanjay Patel --- Like bug 40925, this doesn't need any passes besides -flattencfg to show the bug. The bug is visible by sending just the correct IR after -gvn to -flattencfg (which might be an unused/dead pass at this point). *** This bug has been marked as a duplicate of bug 40925 *** -- You are receiving this mail because: You are on the CC list for the bug.___ llvm-bugs mailing list llvm-bugs@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] [Bug 40928] opt crashes with "opt -instcombine -flattencfg"
https://bugs.llvm.org/show_bug.cgi?id=40928 Sanjay Patel changed: What|Removed |Added Resolution|--- |DUPLICATE CC||spatel+l...@rotateright.com Status|NEW |RESOLVED --- Comment #1 from Sanjay Patel --- Like bug 40925, this doesn't need any passes besides -flattencfg to show the bug. The bug is visible by sending just the correct IR after -instcombine to -flattencfg (which might be an unused/dead pass at this point). *** This bug has been marked as a duplicate of bug 40925 *** -- You are receiving this mail because: You are on the CC list for the bug.___ llvm-bugs mailing list llvm-bugs@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] [Bug 40929] opt crashes with "opt -constprop -structurizecfg -flattencfg"
https://bugs.llvm.org/show_bug.cgi?id=40929 Sanjay Patel changed: What|Removed |Added Status|NEW |RESOLVED Resolution|--- |DUPLICATE CC||spatel+l...@rotateright.com --- Comment #1 from Sanjay Patel --- Like bug 40925, this doesn't need any passes besides -flattencfg to show the bug. The bug is visible by sending just the correct IR after -constprop -structurizecfg to -flattencfg (which might be an unused/dead pass at this point). *** This bug has been marked as a duplicate of bug 40925 *** -- You are receiving this mail because: You are on the CC list for the bug.___ llvm-bugs mailing list llvm-bugs@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] [Bug 40956] New: Inline Assembly requires more registers (x86)
https://bugs.llvm.org/show_bug.cgi?id=40956 Bug ID: 40956 Summary: Inline Assembly requires more registers (x86) Product: libraries Version: 8.0 Hardware: PC OS: All Status: NEW Severity: normal Priority: P Component: Backend: X86 Assignee: unassignedb...@nondot.org Reporter: cmucchie...@wallix.com CC: craig.top...@gmail.com, llvm-bugs@lists.llvm.org, llvm-...@redking.me.uk, spatel+l...@rotateright.com Created attachment 21561 --> https://bugs.llvm.org/attachment.cgi?id=21561&action=edit Failure in inline compilation I have an error while compilingC with inline assembly on x86 32bits with a code that use 5 registers, I have the following error: error: inline assembly requires more registers than available at line X I have reduced the problem to this specific code (see attached IR file). With this code, the error is always triggered. By moving the line %9 = alloca i32, align 4 before the branch instruction, the code compile without errors. By moving the line %8 = alloca i32, align 4 the error is displayed 2 times. -- You are receiving this mail because: You are on the CC list for the bug.___ llvm-bugs mailing list llvm-bugs@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] [Bug 40933] opt crashes with "opt -mem2reg -inline -early-cse-memssa -flattencfg"
https://bugs.llvm.org/show_bug.cgi?id=40933 Sanjay Patel changed: What|Removed |Added Resolution|--- |DUPLICATE Status|NEW |RESOLVED CC||spatel+l...@rotateright.com --- Comment #1 from Sanjay Patel --- Like bug 40925, this doesn't need any passes besides -flattencfg to show the bug. The bug is visible by sending just the correct IR after -mem2reg -inline -early-cse-memssa to -flattencfg (which might be an unused/dead pass at this point). *** This bug has been marked as a duplicate of bug 40925 *** -- You are receiving this mail because: You are on the CC list for the bug.___ llvm-bugs mailing list llvm-bugs@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] [Bug 40784] Crash in MC with alias of alias.
https://bugs.llvm.org/show_bug.cgi?id=40784 Sam Clegg changed: What|Removed |Added Resolution|--- |FIXED Status|NEW |RESOLVED --- Comment #1 from Sam Clegg --- Fixed in rL354694 -- You are receiving this mail because: You are on the CC list for the bug.___ llvm-bugs mailing list llvm-bugs@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] [Bug 40957] New: blockaddresses compare equal
https://bugs.llvm.org/show_bug.cgi?id=40957 Bug ID: 40957 Summary: blockaddresses compare equal Product: libraries Version: trunk Hardware: PC OS: Linux Status: NEW Severity: enhancement Priority: P Component: Interprocedural Optimizations Assignee: unassignedb...@nondot.org Reporter: ndesaulni...@google.com CC: jykni...@google.com, llvm-bugs@lists.llvm.org, llvm-...@ndave.org, srhi...@google.com GCC and Clang disagree on whether different GNU-C local labels compare equal: https://godbolt.org/z/1JuiSL. I *think* Clang is correct in the second case, but it's not as clear cut to me in the first case. -- You are receiving this mail because: You are on the CC list for the bug.___ llvm-bugs mailing list llvm-bugs@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] [Bug 40957] blockaddresses compare equal
https://bugs.llvm.org/show_bug.cgi?id=40957 Nick Desaulniers changed: What|Removed |Added Status|NEW |RESOLVED Resolution|--- |WONTFIX --- Comment #1 from Nick Desaulniers --- The langref section on basic blocks mentions this is undefined behavior: https://llvm.org/docs/LangRef.html#id1024 -- You are receiving this mail because: You are on the CC list for the bug.___ llvm-bugs mailing list llvm-bugs@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] [Bug 36559] Error in type legalising on v6m
https://bugs.llvm.org/show_bug.cgi?id=36559 Dave Green changed: What|Removed |Added Resolution|--- |FIXED Status|NEW |RESOLVED --- Comment #1 from Dave Green --- No longer reproduces from the code here. -- You are receiving this mail because: You are on the CC list for the bug.___ llvm-bugs mailing list llvm-bugs@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] [Bug 40958] New: [GISEL] Error legalising neon type at -O0 on AArch64.
https://bugs.llvm.org/show_bug.cgi?id=40958
Bug ID: 40958
Summary: [GISEL] Error legalising neon type at -O0 on AArch64.
Product: libraries
Version: trunk
Hardware: PC
OS: Linux
Status: NEW
Severity: enhancement
Priority: P
Component: GlobalISel
Assignee: unassignedb...@nondot.org
Reporter: david.gr...@arm.com
CC: llvm-bugs@lists.llvm.org, matthew.arsena...@amd.com,
quentin.colom...@gmail.com
This code:
typedef __attribute__((neon_vector_type(4))) float float32x4_t;
typedef struct {
float32x4_t b;
float32x4_t c;
} d;
void e() {
d f = {0,0};
g(f);
}
Fails during legalisation under:
clang -O0 -target aarch64-linux-gnu -mcpu=cortex-a57 -c 20150303004327-180.c
-mllvm -verify-machineinstrs
With incorrect element types on:
%12:_(<4 x s32>) = G_BUILD_VECTOR %28:_(s64), %30:_(s64)
Without the verify-machineinstrs, it runs into an illegal register copy during
reg alloc.
Bisecting this led back to rL353121, but it may be a knock on effect. I don't
read global isel debug output very well yet. Perhaps it's the narrowing of the
extract that's wrong, not the combining?
The same thing in IR:
; Function Attrs: noinline nounwind optnone
define dso_local void @e() #0 {
entry:
%f = alloca %struct.d, align 16
%0 = bitcast %struct.d* %f to i8*
call void @llvm.memset.p0i8.i64(i8* align 16 %0, i8 0, i64 32, i1 false)
%1 = bitcast %struct.d* %f to [2 x <4 x float>]*
%2 = load [2 x <4 x float>], [2 x <4 x float>]* %1, align 16
%call = call i32 bitcast (i32 (...)* @g to i32 ([2 x <4 x float>])*)([2 x <4
x float>] %2)
ret void
}
--
You are receiving this mail because:
You are on the CC list for the bug.___
llvm-bugs mailing list
llvm-bugs@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] [Bug 40931] opt crashes with "opt -jump-threading -consthoist"
https://bugs.llvm.org/show_bug.cgi?id=40931 Sanjay Patel changed: What|Removed |Added Resolution|--- |DUPLICATE Status|NEW |RESOLVED CC||spatel+l...@rotateright.com --- Comment #1 from Sanjay Patel --- I don't mean to discourage your fuzzing, but it would help to filter out and/or wait for a response to likely duplicates such as this test. Like bug 40930, this doesn't need any passes besides -consthoist to show the bug. The bug is visible by sending the output IR after -jump-threading to -consthoist. *** This bug has been marked as a duplicate of bug 40930 *** -- You are receiving this mail because: You are on the CC list for the bug.___ llvm-bugs mailing list llvm-bugs@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] [Bug 40932] opt crashes with "opt -inline -consthoist"
https://bugs.llvm.org/show_bug.cgi?id=40932 Sanjay Patel changed: What|Removed |Added Status|NEW |RESOLVED Resolution|--- |DUPLICATE CC||spatel+l...@rotateright.com --- Comment #1 from Sanjay Patel --- I don't mean to discourage your fuzzing, but it would help to filter out and/or wait for a response to likely duplicates such as this test. Like bug 40930, this doesn't need any passes besides -consthoist to show the bug. The bug is visible by sending the output IR after -inline to -consthoist. *** This bug has been marked as a duplicate of bug 40930 *** -- You are receiving this mail because: You are on the CC list for the bug.___ llvm-bugs mailing list llvm-bugs@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] [Bug 40934] opt crashes with "opt -mem2reg -jump-threading -consthoist"
https://bugs.llvm.org/show_bug.cgi?id=40934 Sanjay Patel changed: What|Removed |Added Resolution|--- |DUPLICATE Status|NEW |RESOLVED CC||spatel+l...@rotateright.com --- Comment #1 from Sanjay Patel --- I don't mean to discourage your fuzzing, but it would help to filter out and/or wait for a response to likely duplicates such as this test. Like bug 40930, this doesn't need any passes besides -consthoist to show the bug. The bug is visible by sending the output IR after -mem2reg -jump-threading to -consthoist. *** This bug has been marked as a duplicate of bug 40930 *** -- You are receiving this mail because: You are on the CC list for the bug.___ llvm-bugs mailing list llvm-bugs@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] [Bug 40959] New: Passing memory references to inline assembler does not accept 'c' modifier on ARM32
https://bugs.llvm.org/show_bug.cgi?id=40959
Bug ID: 40959
Summary: Passing memory references to inline assembler does not
accept 'c' modifier on ARM32
Product: new-bugs
Version: 8.0
Hardware: PC
OS: Linux
Status: NEW
Severity: normal
Priority: P
Component: new bugs
Assignee: unassignedb...@nondot.org
Reporter: frank.mehn...@kernkonzept.com
CC: htmldevelo...@gmail.com, llvm-bugs@lists.llvm.org
Consider the following example:
static void foo()
{
}
int main(void)
{
asm volatile (".pushsection \".data\"\n"
".long %c[foo]\n"
".popsection\n"
:
: [foo] "i" (&foo));
return 0;
}
The intention is to code the address of foo() into the data segment. This code
does not compile with
clang-8 --target=arm-linux-gnueabi -O2 -o bar bar.c -fno-pic
but it successfully compiles if the 'c' modifier is omitted:
asm volatile (".pushsection \".data\"\n"
".long %[foo]\n"
".popsection\n"
Same code on ARM64 compiles fine:
asm volatile (".pushsection \".data\"\n"
".8byte %c[foo]\n"
".popsection\n"
I believe that specifying the 'c' modifier is correct. Gcc always demans on it
and does not compile without the 'c' modifier.
--
You are receiving this mail because:
You are on the CC list for the bug.___
llvm-bugs mailing list
llvm-bugs@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] [Bug 40930] opt crashes with "opt -consthoist"
https://bugs.llvm.org/show_bug.cgi?id=40930 Sanjay Patel changed: What|Removed |Added Status|NEW |RESOLVED Fixed By Commit(s)||r355337 Resolution|--- |FIXED CC||spatel+l...@rotateright.com --- Comment #4 from Sanjay Patel --- Should be fixed with: https://reviews.llvm.org/rL355337 -- You are receiving this mail because: You are on the CC list for the bug.___ llvm-bugs mailing list llvm-bugs@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] [Bug 40960] New: Merge r355227 and r355228 to 8.0
https://bugs.llvm.org/show_bug.cgi?id=40960 Bug ID: 40960 Summary: Merge r355227 and r355228 to 8.0 Product: libraries Version: trunk Hardware: PC OS: Windows NT Status: NEW Severity: enhancement Priority: P Component: Backend: X86 Assignee: unassignedb...@nondot.org Reporter: craig.top...@gmail.com CC: craig.top...@gmail.com, llvm-bugs@lists.llvm.org, llvm-...@redking.me.uk, spatel+l...@rotateright.com Hans, is it possible to merge this to 8.0. I know we're getting pretty late. ISPC folks tell me they are seeing a regression on some test cases with 8.0 that these commits fix. The gather/scatter intrinsics have been misdeclared for a long time probably since they were added years ago. My reduced test case fails on previous versions llvm. So I think that means that ISPC is hitting some other optimization change that made them more susceptible to this issue. -- You are receiving this mail because: You are on the CC list for the bug.___ llvm-bugs mailing list llvm-bugs@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] [Bug 40961] New: loop vectorizer unable to reason with trip count bounds
https://bugs.llvm.org/show_bug.cgi?id=40961
Bug ID: 40961
Summary: loop vectorizer unable to reason with trip count
bounds
Product: libraries
Version: trunk
Hardware: Other
OS: All
Status: NEW
Severity: enhancement
Priority: P
Component: Loop Optimizer
Assignee: unassignedb...@nondot.org
Reporter: ois...@gmail.com
CC: llvm-bugs@lists.llvm.org
I see bugs 38280, 37423, but this one is simpler still. Consider the following:
void topup(int a[], unsigned long i)
{
for (; i < 16; i++) {
a[i] = 1;
}
}
The result, compiled with -O -march=haswell
- loop is vectorized despite the small absolute limit on trip count
- vectorized part has loop step of 256 (!) and is actually unreachable
--
You are receiving this mail because:
You are on the CC list for the bug.___
llvm-bugs mailing list
llvm-bugs@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] [Bug 25648] Specifying invalid architecture for Linux as target calls unreachable
https://bugs.llvm.org/show_bug.cgi?id=25648 Florian Hahn changed: What|Removed |Added Status|NEW |RESOLVED Resolution|--- |FIXED CC||florian_h...@apple.com --- Comment #1 from Florian Hahn --- Current trunk prints clang-9: error: unknown target triple 'invalid-unknown-linux', please use -triple or -arch -- You are receiving this mail because: You are on the CC list for the bug.___ llvm-bugs mailing list llvm-bugs@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] [Bug 40954] LICM is moving an fdiv instruction out from under a control dependence
https://bugs.llvm.org/show_bug.cgi?id=40954 Eli Friedman changed: What|Removed |Added Resolution|--- |DUPLICATE CC||efrie...@quicinc.com Status|NEW |RESOLVED --- Comment #1 from Eli Friedman --- *** This bug has been marked as a duplicate of bug 8100 *** -- You are receiving this mail because: You are on the CC list for the bug.___ llvm-bugs mailing list llvm-bugs@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] [Bug 40576] Clang should turn destroying delete off by default.
https://bugs.llvm.org/show_bug.cgi?id=40576 Eric Fiselier changed: What|Removed |Added Resolution|--- |INVALID Status|NEW |RESOLVED --- Comment #3 from Eric Fiselier --- @Richard is right. Clang is right. There is no bug here. -- You are receiving this mail because: You are on the CC list for the bug.___ llvm-bugs mailing list llvm-bugs@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
