Re: [PATCH v2] mm: initialize return of vm_insert_pages

[Tom Rix]

On 7/3/20 12:04 PM, Andrew Morton wrote:
> On Fri,  3 Jul 2020 08:53:54 -0700 trix@xxxxxxxxxx wrote:
>
>> From: Tom Rix <trix@xxxxxxxxxx>
>>
>> clang static analysis reports a garbage return
>>
>> In file included from mm/memory.c:84:
>> mm/memory.c:1612:2: warning: Undefined or garbage value returned to caller [core.uninitialized.UndefReturn]
>>         return err;
>>         ^~~~~~~~~~
>>
>> The setting of err depends on a loop executing.
>> So initialize err.
>>
>> ...
>>
>> --- a/mm/memory.c
>> +++ b/mm/memory.c
>> @@ -1601,7 +1601,7 @@ int vm_insert_pages(struct vm_area_struct *vma, unsigned long addr,
>>  	return insert_pages(vma, addr, pages, num, vma->vm_page_prot);
>>  #else
>>  	unsigned long idx = 0, pgcount = *num;
>> -	int err;
>> +	int err = -EINVAL;
>>  
>>  	for (; idx < pgcount; ++idx) {
>>  		err = vm_insert_page(vma, addr + (PAGE_SIZE * idx), pages[idx]);
> If a caller were to ask vm_insert_pages() to insert zero pages, I
> suspect we should just return zero and set *num to zero.

That would be v1 of the patch.

For the loop to not execute *num/pgcount is initially 0, and *num = pgcount - idx is 0 on exit.

Â

>