[llvm-bugs] [Bug 33807] New: [Polly] Importing schedules and memory access updates that change the dependences not possible
https://bugs.llvm.org/show_bug.cgi?id=33807 Bug ID: 33807 Summary: [Polly] Importing schedules and memory access updates that change the dependences not possible Product: Polly Version: unspecified Hardware: PC OS: Linux Status: NEW Severity: enhancement Priority: P Component: Optimizer Assignee: polly-...@googlegroups.com Reporter: tob...@grosser.es CC: llvm-bugs@lists.llvm.org Traditionally we only allowed schedules and memory access relations to be imported which do not change the data dependences. However, with DeLICM it now becomes possible to perform transformations that enable more freedom in the data dependences. We should probably teach the jscop import to only check for RAW dependences and to also take the possibly updated memory accesses into account. -- 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 33779] [Polly] Failure in FAIL: consumer-typeset.execution_time
https://bugs.llvm.org/show_bug.cgi?id=33779 Tobias Grosser changed: What|Removed |Added Status|NEW |RESOLVED Resolution|--- |FIXED --- Comment #1 from Tobias Grosser --- This has been resolved in r308113. -- 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 33197] polly: runtime errors in isl / imath.c (asan/ubsan)
https://bugs.llvm.org/show_bug.cgi?id=33197 Tobias Grosser changed: What|Removed |Added Resolution|--- |FIXED Status|NEW |RESOLVED CC||tob...@grosser.es --- Comment #2 from Tobias Grosser --- This has meanwhile being upstreamed to imath and isl and also integrated back into Polly. Please reopen if this issue still persists. Thank you Michael for addressing 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 http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
[llvm-bugs] [Bug 33808] New: PowerPC64: Code built with -mcpu=power9 hits select issue
https://bugs.llvm.org/show_bug.cgi?id=33808
Bug ID: 33808
Summary: PowerPC64: Code built with -mcpu=power9 hits select
issue
Product: libraries
Version: trunk
Hardware: PC
OS: Linux
Status: NEW
Severity: enhancement
Priority: P
Component: Backend: PowerPC
Assignee: unassignedb...@nondot.org
Reporter: an...@samba.org
CC: echri...@gmail.com, hfin...@anl.gov,
kit.bar...@gmail.com, llvm-bugs@lists.llvm.org,
nemanja.i@gmail.com
I'm hitting an error with the following test case:
typedef struct {
short b, c, d, e;
} f;
char g;
f *h;
int i;
void j(void)
{
short k;
for (; i; i++) {
h->b = h->c = h->d = k = 257 * g;
h->e = k;
h++;
}
}
# clang -O2 -mcpu=power9 test.i
fatal error: error in backend: Cannot select: 0x127a99d10: v8i16 = BUILD_VECTOR
Constant:i32<257>, Constant:i32<257>, Constant:i32<257>, Constant:i32<257>,
Constant:i32<257>, Constant:i32<257>, Constant:i32<257>,
Constant:i32<257>
0x127a99c40: i32 = Constant<257>
0x127a99c40: i32 = Constant<257>
0x127a99c40: i32 = Constant<257>
0x127a99c40: i32 = Constant<257>
0x127a99c40: i32 = Constant<257>
0x127a99c40: i32 = Constant<257>
0x127a99c40: i32 = Constant<257>
0x127a99c40: i32 = Constant<257>
In function: j
clang-3.9: error: clang frontend command failed with exit code 70 (use -v to
see invocation)
clang version 5.0.0 (git://github.com/llvm-mirror/clang.git
3c1756a95139610e837bfd6c992d45950f18db75)
(https://github.com/llvm-mirror/llvm.git
40d67727c7971b006a0ccd28a2f8992e22b9ea26)
Target: powerpc64le-unknown-linux-gnu
--
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 33809] New: [clang-format/C++] version 5.0.0 (trunk?) adds empty line after 'public:' where previous versions don't
https://bugs.llvm.org/show_bug.cgi?id=33809
Bug ID: 33809
Summary: [clang-format/C++] version 5.0.0 (trunk?) adds empty
line after 'public:' where previous versions don't
Product: clang
Version: trunk
Hardware: Macintosh
OS: MacOS X
Status: NEW
Severity: normal
Priority: P
Component: Formatter
Assignee: unassignedclangb...@nondot.org
Reporter: christian.ridderst...@gmail.com
CC: djas...@google.com, kli...@google.com,
llvm-bugs@lists.llvm.org
I noticed a difference in behaviour between clang-format v. 5.0.0 (trunk?)
compared to earlier versions (3.8, 3.9 and 4.0). The recent version adds an
empty line where previous versions didn't. The line is added just before an
access specifier, or perhaps it's due to the preceeding inline function
definition, see the example below. Anyway, the result is unexpected.
This is on macOS Sierra 10.12.5 with clang-format from homebrew.
Example: Illustration that version 5.0.0 adds an empty line. This empty line is
_not_ added by version 3.8, 3.9 or 4.0 when tested on the same file and with
the same configuration file.
$ cat .clang_format
BreakBeforeBraces: Custom
BraceWrapping:
AfterFunction: true
$ export F=test_1.h
$ diff $F <( clang-format@3.8 $F )
$ diff $F <( clang-format@3.9 $F )
$ diff $F <( clang-format@4.0 $F )
$ diff $F <( clang-format@5.0 $F )
10a11
>
$ diff -y -W90 $F <( clang-format@5.0 $F )
/* /*
Result of using clang-format on this code Result of using clang-format on
this code
depends on version of clang-format: depends on version of
clang-format:
- v 3.8, 3.9, 4.0: Code is unchanged- v 3.8, 3.9, 4.0: Code is
unchanged
- v 5.0: Adds empty line before 'private:' - v 5.0: Adds empty line before
'private:'
*/ */
class A { class A {
public: public:
int f() const { return 0; } int f() const { return 0; }
>
private:private:
}; };
--
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 33810] New: -fmodules crashes CodeGenModule::getMangledName(), called recursively
https://bugs.llvm.org/show_bug.cgi?id=33810 Bug ID: 33810 Summary: -fmodules crashes CodeGenModule::getMangledName(), called recursively Product: clang Version: trunk Hardware: PC OS: Linux Status: NEW Severity: normal Priority: P Component: -New Bugs Assignee: unassignedclangb...@nondot.org Reporter: geek4ci...@gmail.com CC: llvm-bugs@lists.llvm.org http://lab.llvm.org:8011/builders/clang-x86_64-linux-selfhost-modules/builds/7503 http://bb.pgr.jp/builders/bootstrap-clang-libcxx-lld-i686-linux r308041 (and r308040) began crashing. As far as I investigated, lazily-resolved decls would call CodeGenModule::getMangledName() recursively, but it is not re-entrant. StringRef &FoundStr = MangledDeclNames[CanonicalGD]; if (!FoundStr.empty()) return FoundStr; ... return FoundStr = Result.first->first(); When (MapVector)MangledDeclName grew (realloc), &FoundStr might not point valid element, to cause memory corruption. I think; return MangledDeclNames[CanonicalGD] = Result.first->first(); would work, but I'm not sure it'd be the right resolution. -- 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 33811] New: clang www pages not updating
https://bugs.llvm.org/show_bug.cgi?id=33811 Bug ID: 33811 Summary: clang www pages not updating Product: Website Version: unspecified Hardware: All OS: All Status: NEW Severity: normal Priority: P Component: General Website Assignee: unassignedb...@nondot.org Reporter: richard-l...@metafoo.co.uk CC: llvm-bugs@lists.llvm.org Commit r308099 from yesterday changed the source of clang.llvm.org/cxx_status.html, but that page has not updated yet. The page is also missing its sidebar, somehow the is being served rather than replaced. -- 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 33765] Replacing instructions in instcombine violates dominance relation
https://bugs.llvm.org/show_bug.cgi?id=33765 Davide Italiano changed: What|Removed |Added Status|NEW |RESOLVED Resolution|--- |FIXED --- Comment #5 from Davide Italiano --- r308144 should help. -- 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 33691] Recognise vector rotations with non-splat constants
https://bugs.llvm.org/show_bug.cgi?id=33691 Andrew Zhogin changed: What|Removed |Added Resolution|--- |FIXED Status|ASSIGNED|RESOLVED --- Comment #1 from Andrew Zhogin --- Fixed in r308150. -- 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 33812] New: ThinLTO importing function with inline ASM can't lead to duplicate symbols
https://bugs.llvm.org/show_bug.cgi?id=33812 Bug ID: 33812 Summary: ThinLTO importing function with inline ASM can't lead to duplicate symbols Product: new-bugs Version: unspecified Hardware: PC OS: All Status: NEW Severity: enhancement Priority: P Component: new bugs Assignee: unassignedb...@nondot.org Reporter: joker@gmail.com CC: llvm-bugs@lists.llvm.org Created attachment 18803 --> https://bugs.llvm.org/attachment.cgi?id=18803&action=edit File with inline asm For example the two attached files illustrate the issue. The inline ASM is declaring a global symbol and after importing and inlining we end up with the symbol redeclared. On Mac I see: $ clang test_caller.cpp test_asm.ll -flto=thin duplicate symbol __dtrace_probe in: /var/folders/yy/prvy9rf51xvbr26mdg9x00q1vtv3gs/T/cc-892b93.o/0.thinlto.o /var/folders/yy/prvy9rf51xvbr26mdg9x00q1vtv3gs/T/cc-892b93.o/1.thinlto.o ld: 1 duplicate symbol for architecture x86_64 clang: error: linker command failed with exit code 1 (use -v to see invocation) Conservatively: we could disable the importing for functions that contains inline ASM. -- 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
