[Qemu-devel] possible leak in bios-tables-test.c

[Qemu-devel] possible leak in bios-tables-test.c

[Igor Mammedov]

When running test under valgrind it spits out a bunch of still reachable
allocations. most of them are glib g_test related but there is one that
belong to the test itself. I've found why it happens so reporting in hope
that someone else would be able to track it down.

reproducer:

make && make tests/bios-tables-test && QTEST_QEMU_BINARY=./x86_64-softmmu/qemu-system-x86_64 \
valgrind --tool=memcheck --leak-check=full --show-leak-kinds=all tests/bios-tables-test

==52347== 5 bytes in 1 blocks are still reachable in loss record 4 of 27
==52347==    at 0x4C29BC3: malloc (vg_replace_malloc.c:299)
==52347==    by 0x529F68D: g_malloc (in /usr/lib64/libglib-2.0.so.0.5600.1)
==52347==    by 0x52B869E: g_strdup (in /usr/lib64/libglib-2.0.so.0.5600.1)
==52347==    by 0x52CD151: g_get_tmp_dir (in /usr/lib64/libglib-2.0.so.0.5600.1)
==52347==    by 0x528664D: ??? (in /usr/lib64/libglib-2.0.so.0.5600.1)
==52347==    by 0x528733E: g_file_open_tmp (in /usr/lib64/libglib-2.0.so.0.5600.1)
==52347==    by 0x112484: dump_aml_files (bios-tables-test.c:272)
==52347==    by 0x112E68: test_acpi_asl (bios-tables-test.c:432)
==52347==    by 0x113BE3: test_acpi_one (bios-tables-test.c:649)
==52347==    by 0x113CB1: test_acpi_piix4_tcg (bios-tables-test.c:676)
==52347==    by 0x52C0290: ??? (in /usr/lib64/libglib-2.0.so.0.5600.1)
==52347==    by 0x52C014A: ??? (in /usr/lib64/libglib-2.0.so.0.5600.1)