Is Valgrind supposed to work inside Flatpak sandbox?

[Jiří Janoušek]

Hello Flatpak friends,

I have this buggy sample C file from The Valgrind Quick Start Guide:

------- leak.c -----
#include <stdlib.h>
void f(void) {
  int* x = malloc(10 * sizeof(int));
  x[10] = 0;
}
int main(void) {
  f();
  return 0;
}
------ EOF ------

When I run Valgrind on a host (Fedora 29), I get the expected output:

host$ gcc -g -O0 leak.c -o leak
host$ valgrind --leak-check=yes ./leak
==10191== Memcheck, a memory error detector
==10191== Copyright (C) 2002-2017, and GNU GPL'd, by Julian Seward et al.
==10191== Using Valgrind-3.14.0 and LibVEX; rerun with -h for copyright info
==10191== Command: ./leak
==10191==
==10191== Invalid write of size 4
==10191==    at 0x401144: f (leak.c:4)
==10191==    by 0x401155: main (leak.c:7)
==10191==  Address 0x4a3b068 is 0 bytes after a block of size 40 alloc'd
==10191==    at 0x483880B: malloc (vg_replace_malloc.c:299)
==10191==    by 0x401137: f (leak.c:3)
==10191==    by 0x401155: main (leak.c:7)
==10191==
==10191==
==10191== HEAP SUMMARY:
==10191==     in use at exit: 40 bytes in 1 blocks
==10191==   total heap usage: 1 allocs, 0 frees, 40 bytes allocated
==10191==
==10191== 40 bytes in 1 blocks are definitely lost in loss record 1 of 1
==10191==    at 0x483880B: malloc (vg_replace_malloc.c:299)
==10191==    by 0x401137: f (leak.c:3)
==10191==    by 0x401155: main (leak.c:7)
==10191==
==10191== LEAK SUMMARY:
==10191==    definitely lost: 40 bytes in 1 blocks
==10191==    indirectly lost: 0 bytes in 0 blocks
==10191==      possibly lost: 0 bytes in 0 blocks
==10191==    still reachable: 0 bytes in 0 blocks
==10191==         suppressed: 0 bytes in 0 blocks
==10191==
==10191== For counts of detected and suppressed errors, rerun with: -v
==10191== ERROR SUMMARY: 2 errors from 2 contexts (suppressed: 0 from 0)

However, when I run Valgrind inside a Flatpak sandbox (flatpak
1.0.6-3.fc29), I get only nonsense:

host$ flatpak run --devel --filesystem=$PWD org.gnome.Sdk//3.30
sh-4.4$ rm -f leak; gcc -g -O0 leak.c -o leak
sh-4.4$ valgrind --leak-check=yes ./leak
==11== Memcheck, a memory error detector
==11== Copyright (C) 2002-2017, and GNU GPL'd, by Julian Seward et al.
==11== Using Valgrind-3.13.0 and LibVEX; rerun with -h for copyright info
==11== Command: ./leak
==11==
==11== Conditional jump or move depends on uninitialised value(s)
==11==    at 0x401D248: ??? (in /usr/lib/x86_64-linux-gnu/ld-2.27.so)
==11==    by 0x40081DC: ??? (in /usr/lib/x86_64-linux-gnu/ld-2.27.so)
==11==
==11== Conditional jump or move depends on uninitialised value(s)
==11==    at 0x401D46E: ??? (in /usr/lib/x86_64-linux-gnu/ld-2.27.so)
==11==    by 0x2D34365F3638782E: ???
==11==    by 0x6E672D78756E696B: ???
==11==    by 0x6972676C61762F74: ???
==11==    by 0x65727067762F646D: ???
==11==    by 0x726F635F64616F6B: ???
==11==    by 0x2D3436646D612D64: ???
==11==    by 0x6F732E78756E696B: ???
==11==
==11== Conditional jump or move depends on uninitialised value(s)
==11==    at 0x401D46E: ??? (in /usr/lib/x86_64-linux-gnu/ld-2.27.so)
==11==    by 0x400B580: ??? (in /usr/lib/x86_64-linux-gnu/ld-2.27.so)
==11==
==11==
==11== HEAP SUMMARY:
==11==     in use at exit: 0 bytes in 0 blocks
==11==   total heap usage: 0 allocs, 0 frees, 0 bytes allocated
==11==
==11== All heap blocks were freed -- no leaks are possible
==11==
==11== For counts of detected and suppressed errors, rerun with: -v
==11== Use --track-origins=yes to see where uninitialised values come from
==11== ERROR SUMMARY: 6 errors from 3 contexts (suppressed: 0 from 0)

Are there any magic spells to make Valgrind work?

Best regards,

Jiri Janousek