On 08/10/2016 09:13 AM, Kirill A. Shutemov wrote:
On Wed, Aug 10, 2016 at 07:33:38AM -0700, Laura Abbott wrote:
Hi,

There have been several reports[1] of assertions tripping when using
tcpdump on the latest master:

[ 1013.718212] device wlp2s0 entered promiscuous mode
[ 1013.736003] page:ffffea0004380000 count:2 mapcount:0 mapping:
(null) index:0x0 compound_mapcount: 0
[ 1013.736013] flags: 0x17ffffc0004000(head)
[ 1013.736017] page dumped because: VM_BUG_ON_PAGE(!PageLocked(page))
[ 1013.736044] ------------[ cut here ]------------
[ 1013.736091] kernel BUG at mm/rmap.c:1288!

The patch below should do the trick.

From 8026e3a2cecb7cdd3a63ebc266fb359ef7ec965b Mon Sep 17 00:00:00 2001
From: "Kirill A. Shutemov" <[email protected]>
Date: Wed, 10 Aug 2016 18:51:54 +0300
Subject: [PATCH] mm, rmap: fix false positive VM_BUG() in page_add_file_rmap()

PageTransCompound() doesn't distinguish THP from from any other type of
compound pages. This can lead to false-positive VM_BUG_ON() in
page_add_file_rmap() if called on compound page from a driver[1].

I think we can exclude such cases by checking if the page belong to a
mapping.

The VM_BUG_ON_PAGE() is downgraded to VM_WARN_ON_ONCE(). This path
should not cause any harm to non-THP page, but good to know if we step
on anything else.

[1] http://lkml.kernel.org/r/[email protected]

Signed-off-by: Kirill A. Shutemov <[email protected]>
Reported-by: Laura Abbott <[email protected]>

Tested-by: Laura Abbott <[email protected]>

---
 mm/rmap.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/mm/rmap.c b/mm/rmap.c
index eee844997bd8..f071d6f7a986 100644
--- a/mm/rmap.c
+++ b/mm/rmap.c
@@ -1286,8 +1286,9 @@ void page_add_file_rmap(struct page *page, bool compound)
                else
                        __inc_node_page_state(page, NR_FILE_PMDMAPPED);
        } else {
-               if (PageTransCompound(page)) {
-                       VM_BUG_ON_PAGE(!PageLocked(page), page);
+               if (PageTransCompound(page) && page_mapping(page)) {
+                       VM_WARN_ON_ONCE(!PageLocked(page));
+
                        SetPageDoubleMap(compound_head(page));
                        if (PageMlocked(page))
                                clear_page_mlock(compound_head(page));


Reply via email to