On 9/2/24 15:16, Alan Maguire wrote:
> On 02/09/2024 12:11, Viktor Malik wrote:
>> Coverity reports a possible memleak in `dwarf_cus__create_cu` due to
>> `cu` not being freed for the case when it was allocated but
>> `cu__set_common` failed. Fix the leak.
>>
>> Signed-off-by: Viktor Malik <vmalik@xxxxxxxxxx>
>> ---
>> dwarf_loader.c | 4 +++-
>> 1 file changed, 3 insertions(+), 1 deletion(-)
>>
>> diff --git a/dwarf_loader.c b/dwarf_loader.c
>> index 04f5637..ad36e3c 100644
>> --- a/dwarf_loader.c
>> +++ b/dwarf_loader.c
>> @@ -3378,8 +3378,10 @@ static struct dwarf_cu *dwarf_cus__create_cu(struct dwarf_cus *dcus, Dwarf_Die *
>> */
>> const char *name = attr_string(cu_die, DW_AT_name, dcus->conf);
>> struct cu *cu = cu__new(name ?: "", pointer_size, dcus->build_id, dcus->build_id_len, dcus->filename, dcus->conf->use_obstack);
>> - if (cu == NULL || cu__set_common(cu, dcus->conf, dcus->mod, dcus->elf) != 0)
>> + if (cu == NULL || cu__set_common(cu, dcus->conf, dcus->mod, dcus->elf) != 0) {
>> + cu__delete(cu);
>> return NULL;
>> + }
>>
>> struct dwarf_cu *dcu = dwarf_cu__new(cu);
>>
>
> Good catch! Looks like there's another few instances of cu leaks like
> this later on in __cus__load_debug_types():
>
>
> if (*cup == NULL) {
> struct cu *cu;
>
> cu = cu__new("", pointer_size, build_id,
> build_id_len, filename,
> conf->use_obstack);
> if (cu == NULL ||
> cu__set_common(cu, conf, mod, elf) != 0) {
> need cu__delete() here ->
> return DWARF_CB_ABORT;
> }
>
> if (dwarf_cu__init(dcup, cu) != 0)
> need cu__delete() here ->
> return DWARF_CB_ABORT;
>
>
> Could you respin to fix them too? Thanks!
>
Sure, I'll check those and send v2 tomorrow.
Viktor
