FAILED: patch "[PATCH] mm: add VM_BUG_ON_PAGE() to page_mapcount()" failed to apply to 4.4-stable tree

The patch below does not apply to the 4.4-stable tree.
If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <[email protected]>.

thanks,

greg k-h

------------------ original commit in Linus's tree ------------------

From 1d148e218a0d0566b1c06f2f45f1436d53b049b2 Mon Sep 17 00:00:00 2001
From: "Wang, Yalin" <[email protected]>
Date: Wed, 11 Feb 2015 15:24:48 -0800
Subject: [PATCH] mm: add VM_BUG_ON_PAGE() to page_mapcount()

Add VM_BUG_ON_PAGE() for slab pages. _mapcount is an union with slab struct in struct page, so we must avoid accessing _mapcount if this page is a slab page. Also remove the unneeded bracket.

index 8dd4fde9d2e5..c6bf813a6b3d 100644
@@ -484,7 +484,8 @@ static inline void page_mapcount_reset(struct page *page)
 
 static inline int page_mapcount(struct page *page)
 {
-	return atomic_read(&(page)->_mapcount) + 1;
+	VM_BUG_ON_PAGE(PageSlab(page), page);
+	return atomic_read(&page->_mapcount) + 1;
 }
 
 static inline int page_count(struct page *page)

Please don't.

The patch known to cause trouble and going to be effectively reverted:

https://lore.kernel.org/r/159032779896.957378.7852761411265662220.stgit@buzz

If you have it in any other stable queue, please drop it.

 Kirill A. Shutemov

Oops, I ment to say that 6988f31d558a ("mm: remove VM_BUG_ON(PageSlab()) from page_mapcount()") did not apply to the 4.4 tree and needs a working backport.

I put in the wrong id as that is the commit that 6988f31d558a fixes.

So, if you could provide a working backport, that would be wonderful.

thanks,

greg k-h

The patch below should do. I've dropped the comments. They are not relevant for v4.4.

From e0568884b6bab1658d4e97d38b22b25c2e6eddeb Mon Sep 17 00:00:00 2001
From: Konstantin Khlebnikov <[email protected]>
Date: Wed, 27 May 2020 22:20:47 -0700
Subject: [PATCH] mm: remove VM_BUG_ON(PageSlab()) from page_mapcount()

Replace superfluous VM_BUG_ON() with comment about correct usage.

Technically reverts commit 1d148e218a0d ("mm: add VM_BUG_ON_PAGE() to page_mapcount()"), but context lines have changed.

Function isolate_migratepages_block() runs some checks out of lru_lock when choose pages for migration. After checking PageLRU() it checks extra page references by comparing page_count() and page_mapcount(). Between these two checks page could be removed from lru, freed and taken by slab.

As a result this race triggers VM_BUG_ON(PageSlab()) in page_mapcount(). Race window is tiny. For certain workload this happens around once a year.

    page:ffffea0105ca9380 count:1 mapcount:0 mapping:ffff88ff7712c180 index:0x0 compound_mapcount: 0
    flags: 0x500000000008100(slab|head)
    raw: 0500000000008100 dead000000000100 dead000000000200 ffff88ff7712c180
    raw: 0000000000000000 0000000080200020 00000001ffffffff 0000000000000000
    page dumped because: VM_BUG_ON_PAGE(PageSlab(page))
    ------------[ cut here ]------------
    kernel BUG at ./include/linux/mm.h:628!
    invalid opcode: 0000 [#1] SMP NOPTI
    CPU: 77 PID: 504 Comm: kcompactd1 Tainted: G        W         4.19.109-27 #1
    Hardware name: Yandex T175-N41-Y3N/MY81-EX0-Y3N, BIOS R05 06/20/2019
    RIP: 0010:isolate_migratepages_block+0x986/0x9b0

The code in isolate_migratepages_block() was added in commit 119d6d59dcc0 ("mm, compaction: avoid isolating pinned pages") before adding VM_BUG_ON into page_mapcount().

This race has been predicted in 2015 by Vlastimil Babka (see link below).

---
 include/linux/mm.h | 1 -
 1 file changed, 1 deletion(-)

index 69fa3df9e712..03cf5526e445 100644
@@ -446,7 +446,6 @@ static inline void page_mapcount_reset(struct page *page)
 
 static inline int page_mapcount(struct page *page)
 {
-	VM_BUG_ON_PAGE(PageSlab(page), page);
 	return atomic_read(&page->_mapcount) + 1;
 }
 

 Kirill A. Shutemov

That worked, thanks!

greg k-h