Bug 510292

After upgrading LTP testsuite version to 20250930 (tracked in bug 510169) munmap01 syscall test started failing.  It however turns out that this testcase was substantially rewritten and the failure is expected:

> $ make -j"$(nproc)" ltpchecks TESTS=munmap01
> [ ... stuff deleted ... ]
> VALGRIND=/home/mcermak/WORK/valgrind/valgrind/vg-in-place \
>   /home/mcermak/WORK/valgrind/valgrind/auxprogs/ltp-tester.sh
> Running individual syscall tests specified in the TESTS env var ...
> [1/1] Testing munmap01 ...
> munmap01: unempty log2.filtered:
> ==1900711==
> ==1900711== Process terminating with default action of signal 11 (SIGSEGV): dumping core
> ==1900711==  Access not within mapped region at address 0x483F000
> ==1900711==    at 0x401F9A: run (munmap01.c:46)
> ==1900711==    by 0x40D42F: fork_testrun.isra.0 (tst_test.c:1669)
> ==1900711==    by 0x40F894: tst_run_tcases (tst_test.c:2041)
> ==1900711==    by 0x401DCD: main (tst_test.h:738)
> ==1900711==  If you believe this happened as a result of a stack
> ==1900711==  overflow in your program's main thread (unlikely but
> ==1900711==  possible), you can try to increase the size of the
> ==1900711==  main thread stack using the --main-stacksize= flag.
> ==1900711==  The main thread stack size used in this run was 8388608.
> ==1900724==
> ==1900724== Process terminating with default action of signal 11 (SIGSEGV): dumping core
> ==1900724==  Access not within mapped region at address 0x4840000
> ==1900724==    at 0x401F9A: run (munmap01.c:46)
> ==1900724==    by 0x40D42F: fork_testrun.isra.0 (tst_test.c:1669)
> ==1900724==    by 0x40F894: tst_run_tcases (tst_test.c:2041)
> ==1900724==    by 0x401DCD: main (tst_test.h:738)
> ==1900724==  If you believe this happened as a result of a stack
> ==1900724==  overflow in your program's main thread (unlikely but
> ==1900724==  possible), you can try to increase the size of the
> ==1900724==  main thread stack using the --main-stacksize= flag.
> ==1900724==  The main thread stack size used in this run was 8388608.
>
> Brief LTP test results summary
> -----------------------------------------
> FAIL: 1
> -----------------------------------------

Looking closer at the new version of the test it turns out that this test now intentionally tries to write to an unmaped region, and considers the resulting segfault an expected outcome.

> static void run(void)
> {
>       int status;
>
>       SAFE_MUNMAP(map_addr, map_len);
>       map_base = NULL;
>
>       /*
>        * Check whether further reference is possible to the unmapped memory
>        * region by writing to the first byte of region with some arbitrary
>        * number.
>        */
>       if (!SAFE_FORK()) {
>               *map_addr = 50; 
>               _exit(0);
>       }
>
>       SAFE_WAIT(&status);
>       if (WIFSIGNALED(status) && WTERMSIG(status) == SIGSEGV) {
>               tst_res(TPASS, "Child was unable to access unmapped memory");
>               return;
>       }
>       tst_res(TFAIL, "Child succeeds to refer unmapped memory region");
> }

This false positive needs to be silenced via auxprogs/filters/munmap01.