Re: [PATCH bpf-next v5 8/8] resolve_btfids: Change in-place update with raw binary output

[Ihor Solodrai]

On 12/18/25 4:53 PM, bot+bpf-ci@xxxxxxxxxx wrote:
>> diff --git a/tools/bpf/resolve_btfids/main.c b/tools/bpf/resolve_btfids/main.c
>> index e721e20a2bbd..7127396c5f50 100644
>> --- a/tools/bpf/resolve_btfids/main.c
>> +++ b/tools/bpf/resolve_btfids/main.c
> 
> [ ... ]
> 
>> @@ -587,6 +563,19 @@ static int load_btf(struct object *obj)
>>  	obj->base_btf = base_btf;
>>  	obj->btf = btf;
>>
>> +	if (obj->base_btf && obj->distill_base) {
>> +		err = btf__distill_base(obj->btf, &base_btf, &btf);
>> +		if (err) {
>> +			pr_err("FAILED to distill base BTF: %s\n", strerror(errno));
>> +			goto out_err;
>> +		}
>> +
>> +		btf__free(obj->btf);
>> +		btf__free(obj->base_btf);
>> +		obj->btf = btf;
>> +		obj->base_btf = base_btf;
>> +	}
>> +
>>  	return 0;
>>
>>  out_err:
>> +	btf__free(base_btf);
>> +	btf__free(btf);
>>  	return err;
>>  }
> 
> Can this lead to a double free if btf__distill_base() fails?
> 
> At the point of calling btf__distill_base(), the local variables base_btf
> and btf have already been assigned to obj->base_btf and obj->btf. If
> btf__distill_base() fails, it does not modify its output parameters, so
> base_btf and btf still point to the same memory as obj->base_btf and
> obj->btf.
> 
> The out_err path then frees base_btf and btf (the original BTF objects).
> But obj->base_btf and obj->btf still point to this freed memory. When
> load_btf() returns an error to main(), the cleanup at the out: label
> calls btf__free(obj.base_btf) and btf__free(obj.btf), which would free
> the same memory a second time.
> 
> The execution path would be:
>   main() -> load_btf() -> btf__distill_base() fails
>   -> out_err frees base_btf and btf
>   -> return err to main()
>   -> main() goto out
>   -> btf__free(obj.base_btf) and btf__free(obj.btf) -> double free
> 
> Perhaps the out_err path should set obj->base_btf = NULL and obj->btf =
> NULL after freeing, or load_btf() should clear obj->base_btf/obj->btf
> before going to out_err?

I confirmed this is indeed a double free, although in a corner case.

I agree, we should set obj->btf and obj->btf_base to NULL at out_err.

Sending a v6 in a bit.

> 
> 
> ---
> AI reviewed your patch. Please fix the bug or email reply why it's not a bug.
> See: https://github.com/kernel-patches/vmtest/blob/master/ci/claude/README.md
> 
> CI run summary: https://github.com/kernel-patches/bpf/actions/runs/20355860150