Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

cdd_done() generates heap-use-after-free error for "CDD timed predecessor static test" #36

Open
magoorden opened this issue Jul 27, 2022 · 0 comments

Comments

@magoorden
Copy link
Contributor

From #31.

I noticed that I did not deinitialized the CDD library in the new test case. So I added in my code locally that

TEST_CASE("CDD timed predecessor static test")
{
 ...
 cdd expected = cdd_remove_negative(expected1 | expected2);

  REQUIRE(cdd_equiv(result, expected));
  cdd_done();
}

But now I get an unexpected 'heap-use-after-free' error from AddressSanitizer. It has something to do with cdd_remove_negative, since when I remove that one from the shown line above, the error is no longer present.

The full error description is

==10537==ERROR: AddressSanitizer: heap-use-after-free on address 0x7fb6edb73800 at pc 0x55d1e3892d5b bp 0x7ffd7b9d57f0 sp 0x7ffd7b9d57e0
WRITE of size 8 at 0x7fb6edb73800 thread T0
    #0 0x55d1e3892d5a in cdd_rec_deref /home/martijn/Documents/UCDD/src/kernel.c:439
    #1 0x55d1e3879e0b in cdd::~cdd() /home/martijn/Documents/UCDD/src/cppext.cpp:44
    #2 0x55d1e3822e06 in DOCTEST_ANON_FUNC_7 /home/martijn/Documents/UCDD/test/test_cdd.cpp:932
    #3 0x55d1e37ee076 in doctest::Context::run() /home/martijn/Documents/UCDD/libs/doctest/include/doctest/doctest.h:6727
    #4 0x55d1e37efe6e in main /home/martijn/Documents/UCDD/libs/doctest/include/doctest/doctest.h:6805
    #5 0x7fb6f123b082 in __libc_start_main ../csu/libc-start.c:308
    #6 0x55d1e37b76dd in _start (/home/martijn/Documents/UCDD/cmake-build-debug/test/test_cdd+0x1c6dd)

0x7fb6edb73800 is located 0 bytes inside of 160000-byte region [0x7fb6edb73800,0x7fb6edb9a900)
freed by thread T0 here:
    #0 0x7fb6f186240f in __interceptor_free ../../../../src/libsanitizer/asan/asan_malloc_linux.cc:122
    #1 0x55d1e3891f59 in cdd_done /home/martijn/Documents/UCDD/src/kernel.c:286
    #2 0x55d1e3822dfa in DOCTEST_ANON_FUNC_7 /home/martijn/Documents/UCDD/test/test_cdd.cpp:937
    #3 0x55d1e37ee076 in doctest::Context::run() /home/martijn/Documents/UCDD/libs/doctest/include/doctest/doctest.h:6727
    #4 0x55d1e37efe6e in main /home/martijn/Documents/UCDD/libs/doctest/include/doctest/doctest.h:6805
    #5 0x7fb6f123b082 in __libc_start_main ../csu/libc-start.c:308

previously allocated by thread T0 here:
    #0 0x7fb6f1862808 in __interceptor_malloc ../../../../src/libsanitizer/asan/asan_malloc_linux.cc:144
    #1 0x55d1e38919a1 in cdd_init /home/martijn/Documents/UCDD/src/kernel.c:197
    #2 0x55d1e3821afe in DOCTEST_ANON_FUNC_7 /home/martijn/Documents/UCDD/test/test_cdd.cpp:891
    #3 0x55d1e37ee076 in doctest::Context::run() /home/martijn/Documents/UCDD/libs/doctest/include/doctest/doctest.h:6727
    #4 0x55d1e37efe6e in main /home/martijn/Documents/UCDD/libs/doctest/include/doctest/doctest.h:6805
    #5 0x7fb6f123b082 in __libc_start_main ../csu/libc-start.c:308

I guess that cdd_done() frees all cdd instances, but then you have one expected left, which results in double-free. Try putting your expected into a separate scope instead, like:

  {  
     cdd expected = cdd_remove_negative(expected1 | expected2);
     REQUIRE(cdd_equiv(result, expected));
  }
  cdd_done();

This will guarantee that expected is destroyed before cdd_done(). I am not entirely sure why cdd_done() is needed, looks like another hack with global variables :-(


Doing this will push the error message further back into the test case. It only works when I do

TEST_CASE("CDD timed predecessor static test")
{
    cdd_init(100000, 10000, 10000);
    cdd_add_clocks(4);
    cdd_add_bddvar(3);
    {
      // Actual content of the test case.
    }
    cdd_done();
}

This is not what I expected of using cdd_done(), see other test cases (like "CDD intersection with size 3") where there is no need for putting the content of the test case within a name space.

magoorden added a commit to Ecdar/UCDD that referenced this issue Jul 27, 2022
To prevent memory issues, the actual test case is placed into its own namespace, see UPPAALModelChecker#31 and UPPAALModelChecker#36.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

1 participant