Message ID | 20200523195718.1479419-1-rppt@linux.ibm.com (mailing list archive) |
---|---|
State | New, archived |
Headers | show |
Series | sparc32: fix page table traversal in srmmu_nocache_init() | expand |
On Sat, May 23, 2020 at 12:57 PM Mike Rapoport <rppt@linux.ibm.com> wrote: > > Remove __nocache_fix() for p4d_offset() and pud_offset() and keep it only > for PUD and lower levels. Davem - I took this directly, since it's a fix for the patch Andrew sent and I merged that way. Just so you know. Linus
diff --git a/arch/sparc/mm/srmmu.c b/arch/sparc/mm/srmmu.c index e9f7af32da07..a8c2f2615fc6 100644 --- a/arch/sparc/mm/srmmu.c +++ b/arch/sparc/mm/srmmu.c @@ -331,8 +331,8 @@ static void __init srmmu_nocache_init(void) while (vaddr < srmmu_nocache_end) { pgd = pgd_offset_k(vaddr); - p4d = p4d_offset(__nocache_fix(pgd), vaddr); - pud = pud_offset(__nocache_fix(p4d), vaddr); + p4d = p4d_offset(pgd, vaddr); + pud = pud_offset(p4d, vaddr); pmd = pmd_offset(__nocache_fix(pud), vaddr); pte = pte_offset_kernel(__nocache_fix(pmd), vaddr);
The srmmu_nocache_init() uses __nocache_fix() macro to add an offset to page table entry to access srmmu_nocache_pool. But since sparc32 has only three actual page table levels, pgd, p4d and pud are essentially the same thing and pgd_offset() and p4d_offset() are no-ops, the __nocache_fix() should be done only at PUD level. Remove __nocache_fix() for p4d_offset() and pud_offset() and keep it only for PUD and lower levels. Fixes: c2bc26f7ca1f ("sparc32: use PUD rather than PGD to get PMD in srmmu_nocache_init()") Signed-off-by: Mike Rapoport <rppt@linux.ibm.com> Cc: David S. Miller <davem@davemloft.net> Cc: Anatoly Pugachev <matorola@gmail.com> Cc: <stable@vger.kernel.org> --- arch/sparc/mm/srmmu.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)