On 21.12.23 22:58, Andrew Morton wrote:
On Thu, 21 Dec 2023 12:32:04 +0100 David Hildenbrand <david@xxxxxxxxxx> wrote:
On 21.12.23 12:23, Oliver Sang wrote:
hi, Andrew Morton,
On Wed, Dec 20, 2023 at 02:11:35PM -0800, Andrew Morton wrote:
On Tue, 19 Dec 2023 23:46:50 +0800 kernel test robot <oliver.sang@xxxxxxxxx> wrote:
Hello,
kernel test robot noticed "kernel_BUG_at_mm/memory.c" on:
commit: bbcbf2a3f05f74f9d268eab57abbdce6a65a94ad ("mm: convert ksm_might_need_to_copy() to work on folios")
I assume this is a bisection result, so it's quite repeatable?
yes, we bisect to this commit, it's quite repeatable:
ddd06bb63d9793ce bbcbf2a3f05f74f9d268eab57ab
---------------- ---------------------------
fail:runs %reproduction fail:runs
| | |
:6 100% 6:6 dmesg.Kernel_panic-not_syncing:Fatal_exception
:6 100% 6:6 dmesg.RIP:do_swap_page
:6 100% 6:6 dmesg.invalid_opcode:#[##]
:6 100% 6:6 dmesg.kernel_BUG_at_mm/memory.c
Can you try with the snipped I sent? Please let me know if you need a
full patch for testing purposes.
I think a full patch would be better, please.
From b82e309096abde6c0f24bba50a281e8d3855c132 Mon Sep 17 00:00:00 2001
From: David Hildenbrand <david@xxxxxxxxxx>
Date: Thu, 21 Dec 2023 23:04:30 +0100
Subject: [PATCH] Fixup: mm: convert ksm_might_need_to_copy() to work on folios
We must only adjust the page if the folio changed. Otherwise, if we
had a large folio in the swapcache and the folio didn't change, we'd
suddenly change the page to-be-mapped.
Reported-by: kernel test robot <oliver.sang@xxxxxxxxx>
Closes: https://lore.kernel.org/oe-lkp/202312192319.fa8f5709-oliver.sang@xxxxxxxxx
Signed-off-by: David Hildenbrand <david@xxxxxxxxxx>
---
mm/memory.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/mm/memory.c b/mm/memory.c
index 149f779910fd5..2f9668d357f5c 100644
--- a/mm/memory.c
+++ b/mm/memory.c
@@ -3952,7 +3952,8 @@ vm_fault_t do_swap_page(struct vm_fault *vmf)
folio = swapcache;
goto out_page;
}
- page = folio_page(folio, 0);
+ if (folio != swapcache)
+ page = folio_page(folio, 0);
/*
* If we want to map a page that's in the swapcache writable, we
--
2.43.0
--
Cheers,
David / dhildenb
