Bug#818380: clang-3.7: segfaults on s390x

2016-03-19 Thread Daniel Stender
Package: clang-3.7
Version: 1:3.7.1-1+b1
Severity: serious
Justification: segfault

Hi,

packaging a new upstream version of afl we've discovered that there's a problem 
with current
clang-3.7 on s390x (2.08b-3 build against clang-3.7/llvm-3.7-dev today) [1].

Probe case (provided by Michael Zalewski) tested on s390x porterbox:


$ echo 'int main() { printf("%p\n", __builtin_return_address(0)); return 0; }' 
> test.c
$ clang-3.7 -g -O2 -fPIE -fPIC test.c
test.c:1:14: warning: implicitly declaring library function 'printf' with type 
'int (const char *, ...)'
int main() { printf("%p\n", __builtin_return_address(0)); return 0; }
 ^
test.c:1:14: note: include the header  or explicitly provide a 
declaration for 'printf'
0  libLLVM-3.7.so.1 0x03fffbf1007e 
llvm::sys::PrintStackTrace(llvm::raw_ostream&) + 38
1  libLLVM-3.7.so.1 0x03fffbf0f2be
2  libLLVM-3.7.so.1 0x03976dcc
3  libLLVM-3.7.so.1 0x03fffc6b2e3a 
llvm::SystemZTargetLowering::LowerOperation(llvm::SDValue, llvm::SelectionDAG&) 
const + 74
4  libLLVM-3.7.so.1 0x03fffbd2c13e
5  libLLVM-3.7.so.1 0x03fffbd2ecea llvm::SelectionDAG::Legalize() + 418
6  libLLVM-3.7.so.1 0x03fffbe3af70 
llvm::SelectionDAGISel::CodeGenAndEmitDAG() + 616
7  libLLVM-3.7.so.1 0x03fffbe3f26e 
llvm::SelectionDAGISel::SelectAllBasicBlocks(llvm::Function const&) + 1126
8  libLLVM-3.7.so.1 0x03fffbe40cfe 
llvm::SelectionDAGISel::runOnMachineFunction(llvm::MachineFunction&) + 1078
9  libLLVM-3.7.so.1 0x03fffb500710 
llvm::FPPassManager::runOnFunction(llvm::Function&) + 584
10 libLLVM-3.7.so.1 0x03fffb500a0e 
llvm::FPPassManager::runOnModule(llvm::Module&) + 54
11 libLLVM-3.7.so.1 0x03fffb5002be 
llvm::legacy::PassManagerImpl::run(llvm::Module&) + 990
12 clang0x804f274c 
clang::EmitBackendOutput(clang::DiagnosticsEngine&, clang::CodeGenOptions 
const&, clang::TargetOptions const&, clang::LangOptions const&, 
llvm::StringRef, llvm::Module*, clang::BackendAction, llvm::raw_pwrite_stream*) 
+ 2700
13 clang0x804da5c8
14 clang0x806fdcac clang::ParseAST(clang::Sema&, bool, 
bool) + 868
15 clang0x802ff7f6 clang::FrontendAction::Execute() + 622
16 clang0x802d1fd6 
clang::CompilerInstance::ExecuteAction(clang::FrontendAction&) + 398
17 clang0x802b58aa 
clang::ExecuteCompilerInvocation(clang::CompilerInstance*) + 2346
18 clang0x802af932 cc1_main(llvm::ArrayRef, 
char const*, void*) + 2754
19 clang0x802abc4c main + 1380
20 libc.so.60x03fffa8be4c0 __libc_start_main + 272
21 clang0x802ae6d6
Stack dump:
0.  Program arguments: /usr/lib/llvm-3.7/bin/clang -cc1 -triple 
s390x-ibm-linux-gnu -emit-obj -disable-free -disable-llvm-verifier 
-main-file-name test.c -mrelocation-model pic -pic-level 2
-mthread-model posix -fmath-errno -masm-verbose -mconstructor-aliases -fuse-
init-array -target-cpu z10 -target-linker-version 2.26 
-momit-leaf-frame-pointer -g -dwarf-column-info -resource-dir 
/usr/lib/llvm-3.7/bin/../lib/clang/3.7.1 -internal-isystem /usr/local/include
-internal-isystem /usr/lib/llvm-3.7/bin/../lib/clang/3.7.1/include 
-internal-externc-isystem /usr/include/s390x-linux-gnu 
-internal-externc-isystem /include -internal-externc-isystem /usr/include -O2 
-fdebug-compilation-dir /home/stender -ferror-limit 19 -fmessage-length
105 -mstackrealign -fno-signed-char -fobjc-runtime=gcc 
-fdiagnostics-show-option -fcolor-diagnostics -vectorize-loops -vectorize-slp 
-o /tmp/test-ea55f5.o -x c test.c 
1.   parser at end of file
2.  Code generation
3.  Running pass 'Function Pass Manager' on module 'test.c'.
4.  Running pass 'SystemZ DAG->DAG Pattern Instruction Selection' on 
function '@main'
clang: error: unable to execute command: Segmentation fault
clang: error: clang frontend command failed due to signal (use -v to see 
invocation)
Debian clang version 3.7.1-2 (tags/RELEASE_371/final) (based on LLVM 3.7.1)
Target: s390x-ibm-linux-gnu
Thread model: posix
clang: note: diagnostic msg: PLEASE submit a bug report to 
http://bugs.debian.org/ and include the crash backtrace, preprocessed source, 
and associated run script.
clang: note: diagnostic msg: 

PLEASE ATTACH THE FOLLOWING FILES TO THE BUG REPORT:
Preprocessed source(s) and associated run script(s) are located at:
clang: note: diagnostic msg: /tmp/test-e9248d.c
clang: note: diagnostic msg: /tmp/test-e9248d.sh
clang: note: diagnostic msg:


BTW, this is a regression from clang-3.6 (2.08b-1 and -2 have been build
against clang-3.6/llvm-3.6-dev yesterday with same result) [2], so please
reassign or clone that bug report if found suitable.

Thanks,
DS

[1] https://bugs.debian.org/818288 (afl: FTBFS on s390x)

[2] https://buildd.debian.org/status/logs.php?pkg=afl&arch=s390x

-- System Information:
Debian Release: stretch/sid
  APT prefers testing
  APT policy: (500

Bug#818380: clang-3.7: segfaults on s390x

2016-03-19 Thread Daniel Stender
On 16.03.2016 17:12, Sylvestre Ledru wrote:
> I don't think I will be able to manage that.
> I will need help to deal with this issue.
> 
> Please try if 3.8 has the same issue?
> 
> Thanks,
> Sylvestre

Yes, that's pretty awful :-/

... positive also for clang-3.8.

Thanks,
Dan

-- 
4096R/DF5182C8
http://www.danielstender.com/blog/



Bug#818380: clang-3.7: segfaults on s390x

2016-03-19 Thread Sylvestre Ledru
Hello,
Le 16/03/2016 à 17:05, Daniel Stender a écrit :
> Package: clang-3.7
> Version: 1:3.7.1-1+b1
> Severity: serious
> Justification: segfault
>
> Hi,
>
> packaging a new upstream version of afl we've discovered that there's a 
> problem with current
> clang-3.7 on s390x (2.08b-3 build against clang-3.7/llvm-3.7-dev today) [1].
I don't think I will be able to manage that.
I will need help to deal with this issue.

Please try if 3.8 has the same issue?

Thanks,
Sylvestre



Bug#818380: clang-3.7: segfaults on s390x

2016-03-19 Thread Daniel Stender
On 16.03.2016 18:25, Sylvestre Ledru wrote:
> Thanks for testing
> 
> Did you report a bug upstream?
> Thanks
> Sylvestre

Not to llvm ...

Dan

-- 
4096R/DF5182C8
http://www.danielstender.com/blog/



Bug#818380: clang-3.7: segfaults on s390x

2016-03-19 Thread Sylvestre Ledru
Le 16/03/2016 à 18:22, Daniel Stender a écrit :
> On 16.03.2016 17:12, Sylvestre Ledru wrote:
>> I don't think I will be able to manage that.
>> I will need help to deal with this issue.
>>
>> Please try if 3.8 has the same issue?
>>
>> Thanks,
>> Sylvestre
> Yes, that's pretty awful :-/
>
> ... positive also for clang-3.8.
>
>
Thanks for testing

Did you report a bug upstream?
Thanks
Sylvestre