On Mon, Aug 21, 2017 at 02:06:45PM -0700, David Rientjes wrote:
> If there are large numbers of hugepages to iterate while reading
> /proc/pid/smaps, the page walk never does cond_resched(). On archs
> without split pmd locks, there can be significant and observable
> contention on mm->page_table_lock which cause lengthy delays without
> rescheduling.
>
> Always reschedule in smaps_pte_range() if necessary since the pagewalk
> iteration can be expensive.
>
> Signed-off-by: David Rientjes <rient...@google.com>
> ---
> fs/proc/task_mmu.c | 5 +++--
> 1 file changed, 3 insertions(+), 2 deletions(-)
>
> diff --git a/fs/proc/task_mmu.c b/fs/proc/task_mmu.c
> --- a/fs/proc/task_mmu.c
> +++ b/fs/proc/task_mmu.c
> @@ -599,11 +599,11 @@ static int smaps_pte_range(pmd_t *pmd, unsigned long
> addr, unsigned long end,
> if (ptl) {
> smaps_pmd_entry(pmd, addr, walk);
> spin_unlock(ptl);
> - return 0;
> + goto out;
> }
>
> if (pmd_trans_unstable(pmd))
> - return 0;
> + goto out;
> /*
> * The mmap_sem held all the way back in m_start() is what
> * keeps khugepaged out of here and from collapsing things
> @@ -613,6 +613,7 @@ static int smaps_pte_range(pmd_t *pmd, unsigned long
> addr, unsigned long end,
> for (; addr != end; pte++, addr += PAGE_SIZE)
> smaps_pte_entry(pte, addr, walk);
> pte_unmap_unlock(pte - 1, ptl);
> +out:
> cond_resched();
> return 0;
> }
Maybe just call cond_resched() at the beginning of the function and don't
bother with gotos?
--
Kirill A. Shutemov
