Re: [PATCH] mm: disable preemption in apply_to_pte_range

[Jeremy Fitzhardinge <jeremy@goop.org>]

Jeremy Fitzhardinge wrote:
> commit 79d9c90453a7bc9e7613ae889a97ff6b44ab8380

Scratch that.  This instead.
    J

    mm: disable preemption in apply_to_pte_range
    
    Lazy mmu mode needs preemption disabled, so if we're apply to
    init_mm (which doesn't require any pte locks), then explicitly
    disable preemption.  (Do it unconditionally after checking we've
    successfully done the allocation to simplify the error handling.)
    
    Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>

diff --git a/mm/memory.c b/mm/memory.c
index baa999e..b80cc31 100644
--- a/mm/memory.c
+++ b/mm/memory.c
@@ -1718,6 +1718,7 @@ static int apply_to_pte_range(struct mm_struct *mm, pmd_t *pmd,
 
 	BUG_ON(pmd_huge(*pmd));
 
+	preempt_disable();
 	arch_enter_lazy_mmu_mode();
 
 	token = pmd_pgtable(*pmd);
@@ -1729,6 +1730,7 @@ static int apply_to_pte_range(struct mm_struct *mm, pmd_t *pmd,
 	} while (pte++, addr += PAGE_SIZE, addr != end);
 
 	arch_leave_lazy_mmu_mode();
+	preempt_enable();
 
 	if (mm != &init_mm)
 		pte_unmap_unlock(pte-1, ptl);