Revision 462a8e08e0e6287e5ce13187257edbf24213ed03 authored by David Chen on 09 February 2023, 17:48:28 UTC, committed by Linus Torvalds on 12 February 2023, 18:30:05 UTC
When we upgraded our kernel, we started seeing some page corruption like
the following consistently:

  BUG: Bad page state in process ganesha.nfsd  pfn:1304ca
  page:0000000022261c55 refcount:0 mapcount:-128 mapping:0000000000000000 index:0x0 pfn:0x1304ca
  flags: 0x17ffffc0000000()
  raw: 0017ffffc0000000 ffff8a513ffd4c98 ffffeee24b35ec08 0000000000000000
  raw: 0000000000000000 0000000000000001 00000000ffffff7f 0000000000000000
  page dumped because: nonzero mapcount
  CPU: 0 PID: 15567 Comm: ganesha.nfsd Kdump: loaded Tainted: P    B      O      5.10.158-1.nutanix.20221209.el7.x86_64 #1
  Hardware name: VMware, Inc. VMware Virtual Platform/440BX Desktop Reference Platform, BIOS 6.00 04/05/2016
  Call Trace:
   dump_stack+0x74/0x96
   bad_page.cold+0x63/0x94
   check_new_page_bad+0x6d/0x80
   rmqueue+0x46e/0x970
   get_page_from_freelist+0xcb/0x3f0
   ? _cond_resched+0x19/0x40
   __alloc_pages_nodemask+0x164/0x300
   alloc_pages_current+0x87/0xf0
   skb_page_frag_refill+0x84/0x110
   ...

Sometimes, it would also show up as corruption in the free list pointer
and cause crashes.

After bisecting the issue, we found the issue started from commit
e320d3012d25 ("mm/page_alloc.c: fix freeing non-compound pages"):

	if (put_page_testzero(page))
		free_the_page(page, order);
	else if (!PageHead(page))
		while (order-- > 0)
			free_the_page(page + (1 << order), order);

So the problem is the check PageHead is racy because at this point we
already dropped our reference to the page.  So even if we came in with
compound page, the page can already be freed and PageHead can return
false and we will end up freeing all the tail pages causing double free.

Fixes: e320d3012d25 ("mm/page_alloc.c: fix freeing non-compound pages")
Link: https://lore.kernel.org/lkml/BYAPR02MB448855960A9656EEA81141FC94D99@BYAPR02MB4488.namprd02.prod.outlook.com/
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: stable@vger.kernel.org
Signed-off-by: Chunwei Chen <david.chen@nutanix.com>
Reviewed-by: Vlastimil Babka <vbabka@suse.cz>
Reviewed-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
1 parent f339c25
Raw File
resource_kunit.c
// SPDX-License-Identifier: GPL-2.0+
/*
 * Test cases for API provided by resource.c and ioport.h
 */

#include <kunit/test.h>
#include <linux/ioport.h>
#include <linux/kernel.h>
#include <linux/string.h>

#define R0_START	0x0000
#define R0_END		0xffff
#define R1_START	0x1234
#define R1_END		0x2345
#define R2_START	0x4567
#define R2_END		0x5678
#define R3_START	0x6789
#define R3_END		0x789a
#define R4_START	0x2000
#define R4_END		0x7000

static struct resource r0 = { .start = R0_START, .end = R0_END };
static struct resource r1 = { .start = R1_START, .end = R1_END };
static struct resource r2 = { .start = R2_START, .end = R2_END };
static struct resource r3 = { .start = R3_START, .end = R3_END };
static struct resource r4 = { .start = R4_START, .end = R4_END };

struct result {
	struct resource *r1;
	struct resource *r2;
	struct resource r;
	bool ret;
};

static struct result results_for_union[] = {
	{
		.r1 = &r1, .r2 = &r0, .r.start = R0_START, .r.end = R0_END, .ret = true,
	}, {
		.r1 = &r2, .r2 = &r0, .r.start = R0_START, .r.end = R0_END, .ret = true,
	}, {
		.r1 = &r3, .r2 = &r0, .r.start = R0_START, .r.end = R0_END, .ret = true,
	}, {
		.r1 = &r4, .r2 = &r0, .r.start = R0_START, .r.end = R0_END, .ret = true,
	}, {
		.r1 = &r2, .r2 = &r1, .ret = false,
	}, {
		.r1 = &r3, .r2 = &r1, .ret = false,
	}, {
		.r1 = &r4, .r2 = &r1, .r.start = R1_START, .r.end = R4_END, .ret = true,
	}, {
		.r1 = &r2, .r2 = &r3, .ret = false,
	}, {
		.r1 = &r2, .r2 = &r4, .r.start = R4_START, .r.end = R4_END, .ret = true,
	}, {
		.r1 = &r3, .r2 = &r4, .r.start = R4_START, .r.end = R3_END, .ret = true,
	},
};

static struct result results_for_intersection[] = {
	{
		.r1 = &r1, .r2 = &r0, .r.start = R1_START, .r.end = R1_END, .ret = true,
	}, {
		.r1 = &r2, .r2 = &r0, .r.start = R2_START, .r.end = R2_END, .ret = true,
	}, {
		.r1 = &r3, .r2 = &r0, .r.start = R3_START, .r.end = R3_END, .ret = true,
	}, {
		.r1 = &r4, .r2 = &r0, .r.start = R4_START, .r.end = R4_END, .ret = true,
	}, {
		.r1 = &r2, .r2 = &r1, .ret = false,
	}, {
		.r1 = &r3, .r2 = &r1, .ret = false,
	}, {
		.r1 = &r4, .r2 = &r1, .r.start = R4_START, .r.end = R1_END, .ret = true,
	}, {
		.r1 = &r2, .r2 = &r3, .ret = false,
	}, {
		.r1 = &r2, .r2 = &r4, .r.start = R2_START, .r.end = R2_END, .ret = true,
	}, {
		.r1 = &r3, .r2 = &r4, .r.start = R3_START, .r.end = R4_END, .ret = true,
	},
};

static void resource_do_test(struct kunit *test, bool ret, struct resource *r,
			     bool exp_ret, struct resource *exp_r,
			     struct resource *r1, struct resource *r2)
{
	KUNIT_EXPECT_EQ_MSG(test, ret, exp_ret, "Resources %pR %pR", r1, r2);
	KUNIT_EXPECT_EQ_MSG(test, r->start, exp_r->start, "Start elements are not equal");
	KUNIT_EXPECT_EQ_MSG(test, r->end, exp_r->end, "End elements are not equal");
}

static void resource_do_union_test(struct kunit *test, struct result *r)
{
	struct resource result;
	bool ret;

	memset(&result, 0, sizeof(result));
	ret = resource_union(r->r1, r->r2, &result);
	resource_do_test(test, ret, &result, r->ret, &r->r, r->r1, r->r2);

	memset(&result, 0, sizeof(result));
	ret = resource_union(r->r2, r->r1, &result);
	resource_do_test(test, ret, &result, r->ret, &r->r, r->r2, r->r1);
}

static void resource_test_union(struct kunit *test)
{
	struct result *r = results_for_union;
	unsigned int i = 0;

	do {
		resource_do_union_test(test, &r[i]);
	} while (++i < ARRAY_SIZE(results_for_union));
}

static void resource_do_intersection_test(struct kunit *test, struct result *r)
{
	struct resource result;
	bool ret;

	memset(&result, 0, sizeof(result));
	ret = resource_intersection(r->r1, r->r2, &result);
	resource_do_test(test, ret, &result, r->ret, &r->r, r->r1, r->r2);

	memset(&result, 0, sizeof(result));
	ret = resource_intersection(r->r2, r->r1, &result);
	resource_do_test(test, ret, &result, r->ret, &r->r, r->r2, r->r1);
}

static void resource_test_intersection(struct kunit *test)
{
	struct result *r = results_for_intersection;
	unsigned int i = 0;

	do {
		resource_do_intersection_test(test, &r[i]);
	} while (++i < ARRAY_SIZE(results_for_intersection));
}

static struct kunit_case resource_test_cases[] = {
	KUNIT_CASE(resource_test_union),
	KUNIT_CASE(resource_test_intersection),
	{}
};

static struct kunit_suite resource_test_suite = {
	.name = "resource",
	.test_cases = resource_test_cases,
};
kunit_test_suite(resource_test_suite);

MODULE_LICENSE("GPL");
back to top