> On Jun 21, 2019, at 5:44 AM, Kirill A. Shutemov <kirill@xxxxxxxxxxxxx> wrote:
>
> On Thu, Jun 13, 2019 at 10:57:45AM -0700, Song Liu wrote:
>> @@ -419,6 +419,11 @@ static struct page *follow_pmd_mask(struct vm_area_struct *vma,
>> put_page(page);
>> if (pmd_none(*pmd))
>> return no_page_table(vma, flags);
>> + } else { /* flags & FOLL_SPLIT_PMD */
>> + spin_unlock(ptl);
>> + ret = 0;
>> + split_huge_pmd(vma, pmd, address);
>> + pte_alloc(mm, pmd);
>
> pte_alloc() can fail and the failure should be propogated to the caller.
Good catch! Fixing it now.
Thanks,
Song
