[Bug sanitizer/59106] New: Failure to link against static libasan

[Bug sanitizer/59106] New: Failure to link against static libasan

[y.gribov at samsung dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59106

            Bug ID: 59106
           Summary: Failure to link against static libasan
           Product: gcc
           Version: 4.9.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: sanitizer
          Assignee: unassigned at gcc dot gnu.org
          Reporter: y.gribov at samsung dot com
                CC: dodji at gcc dot gnu.org, dvyukov at gcc dot gnu.org,
                    e.gavrin at samsung dot com, jakub at gcc dot gnu.org,
                    kcc at gcc dot gnu.org, tetra2005 at gmail dot com,
                    v.garbuzov at samsung dot com
            Target: x86_64

When linking sanitized C executable against static libasan I see linker errors:
 $ cat repro.i
 int main() {
   return 0;
 }
 $ /home/ygribov/install/gcc-master/bin/gcc -g repro.i -fsanitize=address
-static-libasan

/home/ygribov/install/gcc-master/lib/gcc/x86_64-unknown-linux-gnu/4.9.0/../../../../lib64/libasan.a(asan_thread.o):(.data.rel.ro._ZTIN6__asan17AsanThreadContextE[_ZTIN6__asan17AsanThreadContextE]+0x0):
undefined reference to `vtable for __cxxabiv1::__si_class_type_info'
 ...
 collect2: error: ld returned 1 exit status

This is trunk gcc, x86_64.

[Bug sanitizer/59106] Failure to link against static libasan

[y.gribov at samsung dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59106

--- Comment #1 from Yury Gribov <y.gribov at samsung dot com> ---
Created attachment 31206
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=31206&action=edit
Proposed patch

Does this look sane?

[Bug sanitizer/59106] Failure to link against static libasan

[kcc at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59106

--- Comment #2 from Kostya Serebryany <kcc at gcc dot gnu.org> ---
We should be building libsanitizer with -fno-rtti

[Bug sanitizer/59106] Failure to link against static libasan

[y.gribov at samsung dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59106

--- Comment #3 from Yury Gribov <y.gribov at samsung dot com> ---
(In reply to Kostya Serebryany from comment #2)
> We should be building libsanitizer with -fno-rtti

Ubsan seems to require it:
 /home/ygribov/gcc/gcc-master/libsanitizer/ubsan/ubsan_type_hash.cc:121:63:
error: ‘dynamic_cast’ not permitted with -fno-rtti
         dynamic_cast<const abi::__si_class_type_info*>(Derived))
                                                               ^
>From gcc-bugs-return-434447-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org Wed Nov 13 11:20:19 2013
Return-Path: <gcc-bugs-return-434447-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org>
Delivered-To: listarch-gcc-bugs@gcc.gnu.org
Received: (qmail 13578 invoked by alias); 13 Nov 2013 11:20:19 -0000
Mailing-List: contact gcc-bugs-help@gcc.gnu.org; run by ezmlm
Precedence: bulk
List-Id: <gcc-bugs.gcc.gnu.org>
List-Archive: <http://gcc.gnu.org/ml/gcc-bugs/>
List-Post: <mailto:gcc-bugs@gcc.gnu.org>
List-Help: <mailto:gcc-bugs-help@gcc.gnu.org>
Sender: gcc-bugs-owner@gcc.gnu.org
Delivered-To: mailing list gcc-bugs@gcc.gnu.org
Received: (qmail 13519 invoked by uid 48); 13 Nov 2013 11:20:15 -0000
From: "samsonov at google dot com" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug sanitizer/59106] Failure to link against static libasan
Date: Wed, 13 Nov 2013 11:20:00 -0000
X-Bugzilla-Reason: CC
X-Bugzilla-Type: changed
X-Bugzilla-Watch-Reason: None
X-Bugzilla-Product: gcc
X-Bugzilla-Component: sanitizer
X-Bugzilla-Version: 4.9.0
X-Bugzilla-Keywords:
X-Bugzilla-Severity: normal
X-Bugzilla-Who: samsonov at google dot com
X-Bugzilla-Status: UNCONFIRMED
X-Bugzilla-Priority: P3
X-Bugzilla-Assigned-To: unassigned at gcc dot gnu.org
X-Bugzilla-Target-Milestone: ---
X-Bugzilla-Flags:
X-Bugzilla-Changed-Fields: cc
Message-ID: <bug-59106-4-PeMn4revta@http.gcc.gnu.org/bugzilla/>
In-Reply-To: <bug-59106-4@http.gcc.gnu.org/bugzilla/>
References: <bug-59106-4@http.gcc.gnu.org/bugzilla/>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: 7bit
X-Bugzilla-URL: http://gcc.gnu.org/bugzilla/
Auto-Submitted: auto-generated
MIME-Version: 1.0
X-SW-Source: 2013-11/txt/msg01224.txt.bz2
Content-length: 529

http://gcc.gnu.org/bugzilla/show_bug.cgi?idY106

Alexey Samsonov <samsonov at google dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |samsonov at google dot com

--- Comment #4 from Alexey Samsonov <samsonov at google dot com> ---
Yeah, in LLVM/Clang we build C++-specific part of UBSan with -frtti (and link
it to C++ programs only), and the rest sanitizers with -fno-rtti.

[Bug sanitizer/59106] Failure to link against static libasan

[kcc at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59106

--- Comment #5 from Kostya Serebryany <kcc at gcc dot gnu.org> ---
a small part of ubsan does indeed require rtti. 
See ubsan/CMakeLists.txt (UBSAN_SOURCES vs UBSAN_CXX_SOURCES)
The rest of libsanitizer does not and should use -fno-rtti. 

OMG, I hate having so many different build systems.
Long term this will be a maintenance nightmare to support them.

[Bug sanitizer/59106] Failure to link against static libasan

[y.gribov at samsung dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59106

Yury Gribov <y.gribov at samsung dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
  Attachment #31206|0                           |1
        is obsolete|                            |

--- Comment #6 from Yury Gribov <y.gribov at samsung dot com> ---
Created attachment 31207
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=31207&action=edit
New patch

(In reply to Kostya Serebryany from comment #5)
> a small part of ubsan does indeed require rtti. 
> See ubsan/CMakeLists.txt (UBSAN_SOURCES vs UBSAN_CXX_SOURCES)

This is going to be fun because automake does not seem to allow to specify
per-file compile flags.

> The rest of libsanitizer does not and should use -fno-rtti. 

Patch attached (without ubsan support for now).

[Bug sanitizer/59106] Failure to link against static libasan

[kcc at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59106

--- Comment #7 from Kostya Serebryany <kcc at gcc dot gnu.org> ---
Patch looks ok. 
Please send it in a usual way and someone else (Jakub?) will need to approve
it.

[Bug sanitizer/59106] Failure to link against static libasan

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59106

--- Comment #8 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
(In reply to Alexey Samsonov from comment #4)
> Yeah, in LLVM/Clang we build C++-specific part of UBSan with -frtti (and
> link it to C++ programs only), and the rest sanitizers with -fno-rtti.

Can't we just add asm (".weak _Z..."); to the ubsan source that needs C++ and
perhaps check if those symbols are non-NULL, otherwise simply fail the dynamic
type checking?
Then libubsan wouldn't require libstdc++, but, supposedly if you call the ubsan
C++ diagnostics, you'd have libstdc++ or libsupc++ linked in and it would DTRT.

[Bug sanitizer/59106] Failure to link against static libasan

[kcc at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59106

Kostya Serebryany <kcc at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |nlewycky at google dot com

--- Comment #9 from Kostya Serebryany <kcc at gcc dot gnu.org> ---
>> Can't we just add asm (".weak _Z...");
The current scheme is more portable 
(e.g. I am not sure if Windows supports weak symbols at all)

[Bug sanitizer/59106] Failure to link against static libasan

[y.gribov at samsung dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59106

Yury Gribov <y.gribov at samsung dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
  Attachment #31207|0                           |1
        is obsolete|                            |

--- Comment #10 from Yury Gribov <y.gribov at samsung dot com> ---
Created attachment 31208
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=31208&action=edit
Patch with a test

Added a test.

[Bug sanitizer/59106] Failure to link against static libasan

[kcc at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59106

--- Comment #11 from Kostya Serebryany <kcc at gcc dot gnu.org> ---
Trying to build chrome with gcc's asan...
The build barks as in this bug -- we do need to build libsanitizer (at least
asan) with -fno-rtti

[Bug sanitizer/59106] Failure to link against static libasan

[y.gribov at samsung dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59106

--- Comment #12 from Yury Gribov <y.gribov at samsung dot com> ---
(In reply to Kostya Serebryany from comment #11)
> The build barks as in this bug

Did it work with the patch?

[Bug sanitizer/59106] Failure to link against static libasan

[kcc at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59106

--- Comment #13 from Kostya Serebryany <kcc at gcc dot gnu.org> ---
> Did it work with the patch?
Yes!

[Bug sanitizer/59106] Failure to link against static libasan

[y.gribov at samsung dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59106

--- Comment #14 from Yury Gribov <y.gribov at samsung dot com> ---
Sent to gcc-patches: http://gcc.gnu.org/ml/gcc-patches/2013-11/msg01990.html

As for libubsan, it does not seem to support static link anyway:

 /home/ygribov/install/gcc-master/bin/gcc -g repro.i -fsanitize=undefined
-static-libubsan
 ...
 /sanitizer_linux_libcdep.cc:71: undefined reference to `pthread_getattr_np'
 /sanitizer_linux_libcdep.cc:74: undefined reference to `pthread_attr_getstack'
 /sanitizer_linux_libcdep.cc:85: undefined reference to `dlsym'
 ...

Once it does, we can go with the same per-file flags hack that we do e.g. in
libstdc++-v3/src/Makefile.am.

[Bug sanitizer/59106] Failure to link against static libasan

[kcc at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59106

--- Comment #15 from Kostya Serebryany <kcc at gcc dot gnu.org> ---
(In reply to Yury Gribov from comment #14)
> Sent to gcc-patches: http://gcc.gnu.org/ml/gcc-patches/2013-11/msg01990.html
> 
> As for libubsan, it does not seem to support static link anyway:

you need -ldl -lpthread (asan has it already)

> 
>  /home/ygribov/install/gcc-master/bin/gcc -g repro.i -fsanitize=undefined
> -static-libubsan
>  ...
>  /sanitizer_linux_libcdep.cc:71: undefined reference to `pthread_getattr_np'
>  /sanitizer_linux_libcdep.cc:74: undefined reference to
> `pthread_attr_getstack'
>  /sanitizer_linux_libcdep.cc:85: undefined reference to `dlsym'
>  ...
> 
> Once it does, we can go with the same per-file flags hack that we do e.g. in
> libstdc++-v3/src/Makefile.am.

[Bug sanitizer/59106] Failure to link against static libasan

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59106

--- Comment #16 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
See my http://gcc.gnu.org/ml/gcc-patches/2013-11/msg01874.html, that should fix
that -ldl -lpthread already.

[Bug sanitizer/59106] Failure to link against static libasan

[ygribov at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59106

--- Comment #17 from ygribov at gcc dot gnu.org ---
Author: ygribov
Date: Mon Nov 18 08:03:16 2013
New Revision: 204934

URL: http://gcc.gnu.org/viewcvs?rev=204934&root=gcc&view=rev
Log:
libsanitizer:
2013-11-18  Yury Gribov  <y.gribov@samsung.com>

    PR sanitizer/59106
    * asan/Makefile.am (AM_CXXFLAGS): Add -fno-rtti.
    * interception/Makefile.am (AM_CXXFLAGS): Likewise.
    * lsan/Makefile.am (AM_CXXFLAGS): Likewise.
    * sanitizer_common/Makefile.am (AM_CXXFLAGS): Likewise.
    * tsan/Makefile.am (AM_CXXFLAGS): Likewise.
    * asan/Makefile.in: Regenerate.
    * interception/Makefile.in: Regenerate.
    * tsan/Makefile.in: Regenerate.
    * lsan/Makefile.in: Regenerate.
    * sanitizer_common/Makefile.in: Regenerate.

gcc/testsuite:
2013-11-18  Yury Gribov  <y.gribov@samsung.com>

    PR sanitizer/59106
    * c-c++-common/asan/pr59106.c: New test.

Modified:
    trunk/gcc/testsuite/ChangeLog
    trunk/libsanitizer/ChangeLog
    trunk/libsanitizer/asan/Makefile.am
    trunk/libsanitizer/asan/Makefile.in
    trunk/libsanitizer/interception/Makefile.am
    trunk/libsanitizer/interception/Makefile.in
    trunk/libsanitizer/lsan/Makefile.am
    trunk/libsanitizer/lsan/Makefile.in
    trunk/libsanitizer/sanitizer_common/Makefile.am
    trunk/libsanitizer/sanitizer_common/Makefile.in
    trunk/libsanitizer/tsan/Makefile.am
    trunk/libsanitizer/tsan/Makefile.in

[Bug sanitizer/59106] Failure to link against static libasan

[y.gribov at samsung dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59106

--- Comment #18 from Yury Gribov <y.gribov at samsung dot com> ---
Created attachment 31233
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=31233&action=edit
Libubsan -fno-rtti patch

(In reply to Jakub Jelinek from comment #16)
> See my http://gcc.gnu.org/ml/gcc-patches/2013-11/msg01874.html, that should
> fix that -ldl -lpthread already.

Nice, seems to work.

How about this patch to fix -fno-rtti issue for libubsan then?

[Bug sanitizer/59106] Failure to link against static libasan

[ygribov at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59106

--- Comment #19 from ygribov at gcc dot gnu.org ---
Author: ygribov
Date: Thu Nov 28 12:20:23 2013
New Revision: 205482

URL: http://gcc.gnu.org/viewcvs?rev=205482&root=gcc&view=rev
Log:
2013-11-28  Jakub Jelinek  <jakub@redhat.com>
        Yury Gribov  <y.gribov@samsung.com>

    PR sanitizer/59106
    * ubsan/Makefile.am (AM_CXXFLAGS): Disable -frtti for files that
    don't need it.
    * ubsan/Makefile.in: Regenerated.

Modified:
    trunk/libsanitizer/ChangeLog
    trunk/libsanitizer/ubsan/Makefile.am
    trunk/libsanitizer/ubsan/Makefile.in

[Bug sanitizer/59106] Failure to link against static libasan

[y.gribov at samsung dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59106

Yury Gribov <y.gribov at samsung dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |RESOLVED
         Resolution|---                         |FIXED

--- Comment #20 from Yury Gribov <y.gribov at samsung dot com> ---
All issues fixed, closing the bug.