i386: clear_fixmap() should not use set_pte() From: "Jan Beulich" While not strictly required with the current code (as the upper half of page table entries generated by __set_fixmap() cannot be non-zero due to the second parameter of this function being 'unsigned long'), the use of set_pte() in __set_fixmap() in the context of clear_fixmap() is still improper with CONFIG_X86_PAE (see the respective comment in include/asm-i386/pgtable-3level.h) and would turn into a bug if that second parameter ever gets changed to a 64-bit type. Signed-off-by: Jan Beulich Signed-off-by: Andi Kleen --- arch/i386/mm/pgtable.c | 7 +++++-- 1 files changed, 5 insertions(+), 2 deletions(-) Index: linux/arch/i386/mm/pgtable.c =================================================================== --- linux.orig/arch/i386/mm/pgtable.c +++ linux/arch/i386/mm/pgtable.c @@ -95,8 +95,11 @@ static void set_pte_pfn(unsigned long va return; } pte = pte_offset_kernel(pmd, vaddr); - /* stored as-is, to permit clearing entries */ - set_pte(pte, pfn_pte(pfn, flags)); + if (pgprot_val(flags)) + /* stored as-is, to permit clearing entries */ + set_pte(pte, pfn_pte(pfn, flags)); + else + pte_clear(&init_mm, vaddr, pte); /* * It's enough to flush this one mapping.