* [merged] page_alloc-fix-freeing-non-compound-pages.patch removed from -mm tree
@ 2020-10-14 21:03 akpm
0 siblings, 0 replies; only message in thread
From: akpm @ 2020-10-14 21:03 UTC (permalink / raw)
To: hughd, mm-commits, npiggin, peterz, rppt, willy
The patch titled
Subject: mm/page_alloc.c: fix freeing non-compound pages
has been removed from the -mm tree. Its filename was
page_alloc-fix-freeing-non-compound-pages.patch
This patch was dropped because it was merged into mainline or a subsystem tree
------------------------------------------------------
From: "Matthew Wilcox (Oracle)" <willy@infradead.org>
Subject: mm/page_alloc.c: fix freeing non-compound pages
Here is a very rare race which leaks memory:
Page P0 is allocated to the page cache. Page P1 is free.
Thread A Thread B Thread C
find_get_entry():
xas_load() returns P0
Removes P0 from page cache
P0 finds its buddy P1
alloc_pages(GFP_KERNEL, 1) returns P0
P0 has refcount 1
page_cache_get_speculative(P0)
P0 has refcount 2
__free_pages(P0)
P0 has refcount 1
put_page(P0)
P1 is not freed
Fix this by freeing all the pages in __free_pages() that won't be freed
by the call to put_page(). It's usually not a good idea to split a page,
but this is a very unlikely scenario.
Link: https://lkml.kernel.org/r/20200926213919.26642-1-willy@infradead.org
Fixes: e286781d5f2e ("mm: speculative page references")
Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Acked-by: Mike Rapoport <rppt@linux.ibm.com>
Cc: Nick Piggin <npiggin@gmail.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
---
lib/Kconfig.debug | 9 ++++++++
lib/Makefile | 1
lib/test_free_pages.c | 42 ++++++++++++++++++++++++++++++++++++++++
mm/page_alloc.c | 3 ++
4 files changed, 55 insertions(+)
--- a/lib/Kconfig.debug~page_alloc-fix-freeing-non-compound-pages
+++ a/lib/Kconfig.debug
@@ -2367,6 +2367,15 @@ config TEST_HMM
If unsure, say N.
+config TEST_FREE_PAGES
+ tristate "Test freeing pages"
+ help
+ Test that a memory leak does not occur due to a race between
+ freeing a block of pages and a speculative page reference.
+ Loading this module is safe if your kernel has the bug fixed.
+ If the bug is not fixed, it will leak gigabytes of memory and
+ probably OOM your system.
+
config TEST_FPU
tristate "Test floating point operations in kernel space"
depends on X86 && !KCOV_INSTRUMENT_ALL
--- a/lib/Makefile~page_alloc-fix-freeing-non-compound-pages
+++ a/lib/Makefile
@@ -101,6 +101,7 @@ obj-$(CONFIG_TEST_BLACKHOLE_DEV) += test
obj-$(CONFIG_TEST_MEMINIT) += test_meminit.o
obj-$(CONFIG_TEST_LOCKUP) += test_lockup.o
obj-$(CONFIG_TEST_HMM) += test_hmm.o
+obj-$(CONFIG_TEST_FREE_PAGES) += test_free_pages.o
#
# CFLAGS for compiling floating point code inside the kernel. x86/Makefile turns
--- /dev/null
+++ a/lib/test_free_pages.c
@@ -0,0 +1,42 @@
+// SPDX-License-Identifier: GPL-2.0+
+/*
+ * test_free_pages.c: Check that free_pages() doesn't leak memory
+ * Copyright (c) 2020 Oracle
+ * Author: Matthew Wilcox <willy@infradead.org>
+ */
+
+#include <linux/gfp.h>
+#include <linux/mm.h>
+#include <linux/module.h>
+
+static void test_free_pages(gfp_t gfp)
+{
+ unsigned int i;
+
+ for (i = 0; i < 1000 * 1000; i++) {
+ unsigned long addr = __get_free_pages(gfp, 3);
+ struct page *page = virt_to_page(addr);
+
+ /* Simulate page cache getting a speculative reference */
+ get_page(page);
+ free_pages(addr, 3);
+ put_page(page);
+ }
+}
+
+static int m_in(void)
+{
+ test_free_pages(GFP_KERNEL);
+ test_free_pages(GFP_KERNEL | __GFP_COMP);
+
+ return 0;
+}
+
+static void m_ex(void)
+{
+}
+
+module_init(m_in);
+module_exit(m_ex);
+MODULE_AUTHOR("Matthew Wilcox <willy@infradead.org>");
+MODULE_LICENSE("GPL");
--- a/mm/page_alloc.c~page_alloc-fix-freeing-non-compound-pages
+++ a/mm/page_alloc.c
@@ -4952,6 +4952,9 @@ void __free_pages(struct page *page, uns
{
if (put_page_testzero(page))
free_the_page(page, order);
+ else if (!PageHead(page))
+ while (order-- > 0)
+ free_the_page(page + (1 << order), order);
}
EXPORT_SYMBOL(__free_pages);
_
Patches currently in -mm which might be from willy@infradead.org are
xarray-add-xa_get_order.patch
xarray-add-xas_split.patch
xarray-add-xas_split-fix-2.patch
xarray-add-xas_split-fix-3patch.patch
mm-filemap-fix-storing-to-a-thp-shadow-entry.patch
mm-filemap-fix-page-cache-removal-for-arbitrary-sized-thps.patch
mm-memory-remove-page-fault-assumption-of-compound-page-size.patch
mm-memory-remove-page-fault-assumption-of-compound-page-size-fix.patch
mm-page_owner-change-split_page_owner-to-take-a-count.patch
mm-huge_memory-fix-page_trans_huge_mapcount-assumption-of-thp-size.patch
mm-huge_memory-fix-can_split_huge_page-assumption-of-thp-size.patch
mm-rmap-fix-assumptions-of-thp-size.patch
mm-truncate-fix-truncation-for-pages-of-arbitrary-size.patch
mm-page-writeback-support-tail-pages-in-wait_for_stable_page.patch
mm-vmscan-allow-arbitrary-sized-pages-to-be-paged-out.patch
fs-add-a-filesystem-flag-for-thps.patch
fs-do-not-update-nr_thps-for-mappings-which-support-thps.patch
mm-readahead-add-define_readahead.patch
mm-readahead-make-page_cache_ra_unbounded-take-a-readahead_control.patch
mm-readahead-make-do_page_cache_ra-take-a-readahead_control.patch
mm-readahead-add-page_cache_sync_ra-and-page_cache_async_ra.patch
mm-rename-page_order-to-buddy_order.patch
ramfs-fix-nommu-mmap-with-gaps-in-the-page-cache.patch
harden-autofs-ioctl-table.patch
mm-update-the-documentation-for-vfree.patch
^ permalink raw reply [flat|nested] only message in thread
only message in thread, other threads:[~2020-10-14 21:03 UTC | newest]
Thread overview: (only message) (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2020-10-14 21:03 [merged] page_alloc-fix-freeing-non-compound-pages.patch removed from -mm tree akpm
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).