ARM: make_coherent: fix problems with highpte, part 1
update_mmu_cache() is called with a page table already mapped. We
call make_coherent(), which then calls adjust_pte() which wants to
map other page tables. This causes kmap_atomic() to BUG() because
the slot its trying to use is already taken.
Since do_adjust_pte() modifies the page tables, we are also missing
any form of locking, so we're risking corrupting the page tables.
Fix this by using pte_offset_map_nested(), and taking the pte page
table lock around do_adjust_pte().
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
diff --git a/arch/arm/mm/fault-armv.c b/arch/arm/mm/fault-armv.c
index 7a8efe1..8e9bc51 100644
--- a/arch/arm/mm/fault-armv.c
+++ b/arch/arm/mm/fault-armv.c
@@ -67,6 +67,7 @@
static int adjust_pte(struct vm_area_struct *vma, unsigned long address)
{
+ spinlock_t *ptl;
pgd_t *pgd;
pmd_t *pmd;
pte_t *pte;
@@ -80,11 +81,19 @@
if (pmd_none_or_clear_bad(pmd))
return 0;
- pte = pte_offset_map(pmd, address);
+ /*
+ * This is called while another page table is mapped, so we
+ * must use the nested version. This also means we need to
+ * open-code the spin-locking.
+ */
+ ptl = pte_lockptr(vma->vm_mm, pmd);
+ pte = pte_offset_map_nested(pmd, address);
+ spin_lock(ptl);
ret = do_adjust_pte(vma, address, pte);
- pte_unmap(pte);
+ spin_unlock(ptl);
+ pte_unmap_nested(pte);
return ret;
}
