On 8/2/21 10:44 AM, Yonghong Song wrote:
>
>
> On 7/31/21 7:32 AM, Hengqi Chen wrote:
>> Add test for btf__load_vmlinux_btf/btf__load_module_btf APIs. It first
>> checks that if btrfs module BTF exists, if yes, load module BTF and
>> check symbol existence.
>>
>> Signed-off-by: Hengqi Chen <hengqi.chen@xxxxxxxxx>
>> ---
>> .../selftests/bpf/prog_tests/btf_module.c | 31 +++++++++++++++++++
>> 1 file changed, 31 insertions(+)
>> create mode 100644 tools/testing/selftests/bpf/prog_tests/btf_module.c
>>
>> diff --git a/tools/testing/selftests/bpf/prog_tests/btf_module.c b/tools/testing/selftests/bpf/prog_tests/btf_module.c
>> new file mode 100644
>> index 000000000000..cad1314e3356
>> --- /dev/null
>> +++ b/tools/testing/selftests/bpf/prog_tests/btf_module.c
>> @@ -0,0 +1,31 @@
>> +// SPDX-License-Identifier: GPL-2.0
>> +/* Copyright (c) 2021 Hengqi Chen */
>> +
>> +#include <test_progs.h>
>> +#include <bpf/btf.h>
>> +
>> +static const char *module_path = "/sys/kernel/btf/btrfs";
>> +static const char *module_name = "btrfs";
>> +
>> +void test_btf_module()
>> +{
>> + struct btf *vmlinux_btf, *module_btf;
>> + __s32 type_id;
>> +
>> + if (access(module_path, F_OK))
>> + return;
>> +
>> + vmlinux_btf = btf__load_vmlinux_btf();
>> + if (!ASSERT_OK_PTR(vmlinux_btf, "could not load vmlinux BTF"))
>> + return;
>> +
>> + module_btf = btf__load_module_btf(module_name, vmlinux_btf);
>> + if (!ASSERT_OK_PTR(module_btf, "could not load module BTF"))
>> + return;
>
> Should we do `btf__free(vmlinux_btf)` before `return`?
> From implementation perspective, maybe use "goto" so we have one
> place to do `btf__free(vmlinux_btf)`.
>
Didn't notice that ASSERT macro does not abort the execution.
Will add cleanup code.
>> +
>> + type_id = btf__find_by_name(module_btf, "btrfs_file_open");
>> + ASSERT_GT(type_id, 0, "func btrfs_file_open not found");
>> +
>> + btf__free(module_btf);
>> + btf__free(vmlinux_btf);
>> +}
>>
