On 13.06.2023 18:03, Anthony PERARD wrote:
> On Mon, Jun 12, 2023 at 01:46:40PM +0200, Jan Beulich wrote:
>> The function returns immediately after the enclosing if().
>>
>> Coverity ID: 1532314
>> Fixes: bd7a29c3d0b9 ("tools/libs/ctrl: fix xc_core_arch_map_p2m() to support
>> linear p2m table")
>> Signed-off-by: Jan Beulich <jbeul...@suse.com>
>>
>> --- a/tools/libs/guest/xg_core_x86.c
>> +++ b/tools/libs/guest/xg_core_x86.c
>> @@ -210,7 +210,6 @@ xc_core_arch_map_p2m_list_rw(xc_interfac
>> }
>>
>> munmap(ptes, n_pages * PAGE_SIZE);
>> - ptes = NULL;
>> off = p2m_vaddr & ((mask >> shift) << shift);
>> }
>
> Do we have to remove this assignment? What if someone adds code later
> and reuse the content of the variable `ptes`? Or what if someone adds
> codes after the loop, and handle an error with `goto out`, we would have
> a double-munmap().
Imo it would be at that time when the assignment wants (re)adding. I
fully agree with the tool, and I expect Misra (if it was applied to
the tool stack as well) would demand the very same change.
Jan
