Re: [PATCH v3] arm64/mm: fix incorrect file_map_count for invalid pmd

From: Denys Vlasenko
Date: Mon Nov 28 2022 - 11:28:12 EST


On 11/21/22 19:18, Will Deacon wrote:
On Mon, Nov 21, 2022 at 03:36:08PM +0800, Liu Shixin wrote:
The page table check trigger BUG_ON() unexpectedly when split hugepage:

------------[ cut here ]------------
kernel BUG at mm/page_table_check.c:119!
Internal error: Oops - BUG: 00000000f2000800 [#1] SMP
Dumping ftrace buffer:
(ftrace buffer empty)
Modules linked in:
CPU: 7 PID: 210 Comm: transhuge-stres Not tainted 6.1.0-rc3+ #748
Hardware name: linux,dummy-virt (DT)
pstate: 20000005 (nzCv daif -PAN -UAO -TCO -DIT -SSBS BTYPE=--)
pc : page_table_check_set.isra.0+0x398/0x468
lr : page_table_check_set.isra.0+0x1c0/0x468
[...]
Call trace:
page_table_check_set.isra.0+0x398/0x468
__page_table_check_pte_set+0x160/0x1c0
__split_huge_pmd_locked+0x900/0x1648
__split_huge_pmd+0x28c/0x3b8
unmap_page_range+0x428/0x858
unmap_single_vma+0xf4/0x1c8
zap_page_range+0x2b0/0x410
madvise_vma_behavior+0xc44/0xe78
do_madvise+0x280/0x698
__arm64_sys_madvise+0x90/0xe8
invoke_syscall.constprop.0+0xdc/0x1d8
do_el0_svc+0xf4/0x3f8
el0_svc+0x58/0x120
el0t_64_sync_handler+0xb8/0xc0
el0t_64_sync+0x19c/0x1a0
[...]

On arm64, pmd_leaf() will return true even if the pmd is invalid due to
pmd_present_invalid() check. So in pmdp_invalidate() the file_map_count
will not only decrease once but also increase once. Then in set_pte_at(),
the file_map_count increase again, and so trigger BUG_ON() unexpectedly.

Add !pmd_present_invalid() check in pmd_user_accessible_page() to fix the
problem.

Fixes: 42b2547137f5 ("arm64/mm: enable ARCH_SUPPORTS_PAGE_TABLE_CHECK")
Reported-by: Denys Vlasenko <dvlasenk@xxxxxxxxxx>
Signed-off-by: Liu Shixin <liushixin2@xxxxxxxxxx>
Acked-by: Pasha Tatashin <pasha.tatashin@xxxxxxxxxx>
Acked-by: David Hildenbrand <david@xxxxxxxxxx>
Reviewed-by: Kefeng Wang <wangkefeng.wang@xxxxxxxxxx>
---
v1->v2: Update comment and optimize the code by moving p?d_valid() at
first place suggested by Mark.
v2->v3: Replace pmd_valid() with pmd_present_invalid() suggested by Will.

arch/arm64/include/asm/pgtable.h | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/arch/arm64/include/asm/pgtable.h b/arch/arm64/include/asm/pgtable.h
index edf6625ce965..17afb09f386f 100644
--- a/arch/arm64/include/asm/pgtable.h
+++ b/arch/arm64/include/asm/pgtable.h
@@ -863,7 +863,7 @@ static inline bool pte_user_accessible_page(pte_t pte)
static inline bool pmd_user_accessible_page(pmd_t pmd)
{
- return pmd_leaf(pmd) && (pmd_user(pmd) || pmd_user_exec(pmd));
+ return pmd_leaf(pmd) && !pmd_present_invalid(pmd) && (pmd_user(pmd) || pmd_user_exec(pmd));
}

Acked-by: Will Deacon <will@xxxxxxxxxx>

But please see my comment on v2 about pud_user_exec() for the PUD case.

Can you be more specific? Do you ask for pud_user_exec() to be defined
and used here? Or something else?

Until this patch lands, amd64 PAGE_TABLE_CHECK + THP remains broken...