From: Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
To: Jakub Jelinek <jakub@redhat.com>
Cc: Dodji Seketeli <dodji@redhat.com>, Wei Mi <wmi@google.com>,
Mike Stump <mikestump@comcast.net>,
GCC Patches <gcc-patches@gcc.gnu.org>,
David Li <davidxl@google.com>,
Diego Novillo <dnovillo@google.com>,
Kostya Serebryany <kcc@google.com>,
Dodji Seketeli <dseketel@redhat.com>,
Alexander Potapenko <glider@google.com>,
Evgeniy Stepanov <eugenis@google.com>,
Alexey Samsonov <samsonov@google.com>,
Dmitry Vyukov <dvyukov@google.com>
Subject: Re: [PATCH] asan unit tests from llvm lit-test incremental changes
Date: Thu, 13 Dec 2012 07:44:00 -0000 [thread overview]
Message-ID: <CAGQ9bdzDncOHruvnAk14LCRjfXa2s4=d5nnXHX2+aenuYB0i+Q@mail.gmail.com> (raw)
In-Reply-To: <20121212213013.GH2315@tucnak.redhat.com>
On Thu, Dec 13, 2012 at 1:30 AM, Jakub Jelinek <jakub@redhat.com> wrote:
> On Wed, Dec 12, 2012 at 10:16:49PM +0100, Dodji Seketeli wrote:
>> Independently of this review, I think it's be interesting to hear
>> Kostya's voice on:
>>
>> Jakub Jelinek <jakub@redhat.com> writes:
>>
>> > 2) In large-func-test-1.C, I had to stop matching the backtrace after
>> > _Znw[jm], because libasan is using the fast but inaccurate backtrace,
>> > and while the tests can be easily tweaked to compile with
>> > -fno-omit-frame-pointer, we definitely can't rely on libstdc++.so to be
>> > built with that option. Most likely it isn't.
The tests should be built with -fno-omit-frame-pointer and we don't
need that for libstdc++.so.
This is how it works in LLVM.
asan's interceptors are written in such a way that they don't care
if libstdc++.so or the asan run-time have frame pointers.
>>> I repeat that I think
>> > that at least for Linux libasan should use the _Unwind* based backtrace
>> > at least for the fatal functions (__asan_report* etc.),
We are discussing it from time to time.
Sometimes, if e.g. an error happens inside a qsort callback,
the fp-based unwinder fails to unwind through libc, while _Unwind would work.
I was opposed to this sometime ago because _Unwind often produced
buggy stack traces on Ubuntu Lucid (the version we cared about).
I also vaguely remember some problems with _Unwind* depending on
malloc (or maybe that's something else?)
Now we mostly care about Ubuntu Precise and we need to test whether
_Unwind produces good enough results there before switching.
unwinding on malloc/free should keep using the fp-based unwinder, at
least by default.
We'll be tracking the issue in
https://code.google.com/p/address-sanitizer/issues/detail?id=137
>> and perhaps for
>> > these malloc wrappers like ::operator new, ::operator new[] and their
>> > const std::nothrow_t& variants libasan could intercept them, call
>> > malloc and if that returns NULL, call the original corresponding function
>> > so that it deals with exceptions, new handler etc.
Hmm.. Why's that?
Calling libc's malloc or libstdc++'s operator new in asan run-time is
really a bad idea.
asan's allocator should never return 0 anyway, it should simply crash.
I don't think we want to support new handler at all.
>
> Yeah, I'd appreciate that too.
>
>> and on:
>>
>> > 3) deep-thread-stack-1.C fails for me right now with some libasan assertion,
>> > Kostya, can you please look at that?
>> > AsanThread *t = asanThreadRegistry().GetCurrent();
>> > CHECK(t);
>> > where it failed on the CHECK, because t was NULL. I've skipped the test for
>> > now.
>>
>> [...]
>
> This one is for the testcase solved right now already by the -lasan -lpthread
> linking instead of just -lpthread (and driver adding -lasan afterwards).
> We'll need to think about how to tweak the driver to add -lasan early on the
> command line, before user passed -l* options.
>>
>> > --- gcc/testsuite/g++.dg/asan/deep-tail-call-1.C.jj 2012-12-04 20:24:10.000000000 +0100
>> > +++ gcc/testsuite/g++.dg/asan/deep-tail-call-1.C 2012-12-05 11:01:48.600443834 +0100
>> > @@ -1,21 +1,22 @@
>> > -// { dg-do run }
>> > +// { dg-do run }
>> > // { dg-options "-fno-omit-frame-pointer -fno-optimize-sibling-calls" }
>> > // { dg-additional-options "-mno-omit-leaf-frame-pointer" { target { i?86-*-* x86_64-*-* } } }
>> > -// { dg-shouldfail "asan" }
>> > +// { dg-shouldfail "asan" }
>> >
>> > int global[10];
>> > void __attribute__((noinline)) call4(int i) { global[i+10]++; }
>> > void __attribute__((noinline)) call3(int i) { call4(i); }
>> > void __attribute__((noinline)) call2(int i) { call3(i); }
>> > void __attribute__((noinline)) call1(int i) { call2(i); }
>> > -int main(int argc, char **argv) {
>> > - call1(argc);
>> > +volatile int one = 1;
>>
>> Just curious, why do we need this variable to be volatile, especially
>> since the test is compiled without optimization?
>
> asan.exp tests are torture tests, they iterate over several -O* options,
> unless explicitly dg-skip-if skipped. It could be non-volatile with
> asm volatile ("" : : : "memory");
> or asm volatile ("" : "+m" (one)); or similar too, sure.
> I just don't want to rely on argc being one, and the compiler shouldn't know
> that one is 1 in the test.
>
>> [...]
>>
>> The patch looks OK to me in any case.
>
> Thanks.
>
> Jakub
--kcc
next prev parent reply other threads:[~2012-12-13 7:44 UTC|newest]
Thread overview: 41+ messages / expand[flat|nested] mbox.gz Atom feed top
2012-11-28 9:15 [PATCH] asan unit tests from llvm lit-test Wei Mi
2012-11-28 10:10 ` Konstantin Serebryany
2012-11-28 10:25 ` Jakub Jelinek
2012-11-28 10:41 ` Konstantin Serebryany
2012-11-28 11:03 ` Jakub Jelinek
2012-11-28 11:14 ` Konstantin Serebryany
2012-11-29 20:59 ` [PATCH] asan_test.cc from llvm Jakub Jelinek
2012-11-30 9:35 ` Konstantin Serebryany
2012-11-30 10:22 ` Jakub Jelinek
2012-11-30 10:55 ` Konstantin Serebryany
2012-11-30 14:52 ` Jakub Jelinek
2012-11-30 16:06 ` Jakub Jelinek
[not found] ` <CAKOQZ8y70goUL91pQJt_S=8W+Dn5VTZ5oRphvGuFwMMh41mkLg@mail.gmail.com>
2012-11-30 16:34 ` Jakub Jelinek
2012-12-03 7:07 ` Konstantin Serebryany
2012-12-03 9:18 ` Jakub Jelinek
2012-12-03 9:52 ` Konstantin Serebryany
2012-12-03 11:05 ` Jakub Jelinek
2012-12-03 11:42 ` Konstantin Serebryany
2012-11-28 11:25 ` [PATCH] asan unit tests from llvm lit-test Jakub Jelinek
2012-11-28 11:39 ` Konstantin Serebryany
2012-11-28 10:14 ` Jakub Jelinek
2012-11-30 21:05 ` Wei Mi
2012-12-03 7:16 ` Konstantin Serebryany
2012-12-03 11:01 ` Jakub Jelinek
2012-12-03 18:33 ` Wei Mi
2012-12-03 18:49 ` Konstantin Serebryany
2012-12-03 19:44 ` Jakub Jelinek
2012-12-03 19:09 ` Mike Stump
2012-12-03 19:37 ` Jakub Jelinek
2012-12-03 19:50 ` Mike Stump
[not found] ` <CAN=P9pgjjq66KS2DVkuOSeH2ejQPDcyKhwz5MdKyE3RB64E=xw@mail.gmail.com>
2012-12-04 7:34 ` Jakub Jelinek
2012-12-04 18:01 ` Wei Mi
2012-12-05 12:29 ` [PATCH] asan unit tests from llvm lit-test incremental changes Jakub Jelinek
2012-12-12 21:32 ` Dodji Seketeli
2012-12-12 21:31 ` Jakub Jelinek
2012-12-13 7:44 ` Konstantin Serebryany [this message]
2012-12-13 8:37 ` Jakub Jelinek
2012-12-13 10:23 ` Konstantin Serebryany
2012-12-13 15:22 ` Jakub Jelinek
2012-12-05 23:29 ` [asan] Fix up dg-set-target-env-var Jakub Jelinek
2012-12-06 0:23 ` Mike Stump
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to='CAGQ9bdzDncOHruvnAk14LCRjfXa2s4=d5nnXHX2+aenuYB0i+Q@mail.gmail.com' \
--to=konstantin.s.serebryany@gmail.com \
--cc=davidxl@google.com \
--cc=dnovillo@google.com \
--cc=dodji@redhat.com \
--cc=dseketel@redhat.com \
--cc=dvyukov@google.com \
--cc=eugenis@google.com \
--cc=gcc-patches@gcc.gnu.org \
--cc=glider@google.com \
--cc=jakub@redhat.com \
--cc=kcc@google.com \
--cc=mikestump@comcast.net \
--cc=samsonov@google.com \
--cc=wmi@google.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).