HWPOISON: add fs/device filters

Filesystem data/metadata present the most tricky-to-isolate pages. It requires careful code review and stress testing to get them right. The fs/device filter helps to target the stress tests to some specific filesystem pages. The filter condition is block device's major/minor numbers: - corrupt-filter-dev-major - corrupt-filter-dev-minor When specified (non -1), only page cache pages that belong to that device will be poisoned. The filters are checked reliably on the locked and refcounted page. Haicheng: clear PG_hwpoison and drop bad page count if filter not OK AK: Add documentation CC: Haicheng Li <haicheng.li@intel.com> CC: Nick Piggin <npiggin@suse.de> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com> Signed-off-by: Andi Kleen <ak@linux.intel.com>

HWPOISON: add fs/device filters
Filesystem data/metadata present the most tricky-to-isolate pages. It requires careful code review and stress testing to get them right. The fs/device filter helps to target the stress tests to some specific filesystem pages. The filter condition is block device's major/minor numbers: - corrupt-filter-dev-major - corrupt-filter-dev-minor When specified (non -1), only page cache pages that belong to that device will be poisoned. The filters are checked reliably on the locked and refcounted page. Haicheng: clear PG_hwpoison and drop bad page count if filter not OK AK: Add documentation CC: Haicheng Li <haicheng.li@intel.com> CC: Nick Piggin <npiggin@suse.de> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com> Signed-off-by: Andi Kleen <ak@linux.intel.com>
Wu Fengguang · Andi Kleen
1 parent 138ce286eb
Showing 4 changed files with 72 additions and 0 deletions Side-by-side Diff
Documentation/vm/hwpoison.txt
mm/hwpoison-inject.c
mm/internal.h
mm/memory-failure.c
@@ -115,6 +115,13 @@
 Note these injection interfaces are not stable and might change between
 kernel versions
  
+corrupt-filter-dev-major
+corrupt-filter-dev-minor
+
+Only handle memory failures to pages associated with the file system defined
+by block device major/minor.  -1U is the wildcard value.
+This should be only used for testing with artificial injection.
+
 Architecture specific MCE injector
  
 x86 has mce-inject, mce-test
@@ -3,6 +3,7 @@
 #include <linux/debugfs.h>
 #include <linux/kernel.h>
 #include <linux/mm.h>
+#include "internal.h"
  
 static struct dentry *hwpoison_dir;
  
@@ -51,6 +52,16 @@
  
 	dentry = debugfs_create_file("unpoison-pfn", 0600, hwpoison_dir,
 				     NULL, &unpoison_fops);
+	if (!dentry)
+		goto fail;
+
+	dentry = debugfs_create_u32("corrupt-filter-dev-major", 0600,
+				    hwpoison_dir, &hwpoison_filter_dev_major);
+	if (!dentry)
+		goto fail;
+
+	dentry = debugfs_create_u32("corrupt-filter-dev-minor", 0600,
+				    hwpoison_dir, &hwpoison_filter_dev_minor);
 	if (!dentry)
 		goto fail;
  
@@ -250,4 +250,7 @@
 #define ZONE_RECLAIM_SOME	0
 #define ZONE_RECLAIM_SUCCESS	1
 #endif
+
+extern u32 hwpoison_filter_dev_major;
+extern u32 hwpoison_filter_dev_minor;
@@ -48,6 +48,50 @@
  
 atomic_long_t mce_bad_pages __read_mostly = ATOMIC_LONG_INIT(0);
  
+u32 hwpoison_filter_dev_major = ~0U;
+u32 hwpoison_filter_dev_minor = ~0U;
+EXPORT_SYMBOL_GPL(hwpoison_filter_dev_major);
+EXPORT_SYMBOL_GPL(hwpoison_filter_dev_minor);
+
+static int hwpoison_filter_dev(struct page *p)
+{
+	struct address_space *mapping;
+	dev_t dev;
+
+	if (hwpoison_filter_dev_major == ~0U &&
+	    hwpoison_filter_dev_minor == ~0U)
+		return 0;
+
+	/*
+	 * page_mapping() does not accept slab page
+	 */
+	if (PageSlab(p))
+		return -EINVAL;
+
+	mapping = page_mapping(p);
+	if (mapping == NULL || mapping->host == NULL)
+		return -EINVAL;
+
+	dev = mapping->host->i_sb->s_dev;
+	if (hwpoison_filter_dev_major != ~0U &&
+	    hwpoison_filter_dev_major != MAJOR(dev))
+		return -EINVAL;
+	if (hwpoison_filter_dev_minor != ~0U &&
+	    hwpoison_filter_dev_minor != MINOR(dev))
+		return -EINVAL;
+
+	return 0;
+}
+
+int hwpoison_filter(struct page *p)
+{
+	if (hwpoison_filter_dev(p))
+		return -EINVAL;
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(hwpoison_filter);
+
 /*
  * Send all the processes who have the page mapped an ``action optional''
  * signal.
@@ -842,6 +886,13 @@
 		printk(KERN_ERR "MCE %#lx: just unpoisoned\n", pfn);
 		res = 0;
 		goto out;
+	}
+	if (hwpoison_filter(p)) {
+		if (TestClearPageHWPoison(p))
+			atomic_long_dec(&mce_bad_pages);
+		unlock_page(p);
+		put_page(p);
+		return 0;
 	}
  
 	wait_on_page_writeback(p);
...	...	@@ -115,6 +115,13 @@
115	115	Note these injection interfaces are not stable and might change between
116	116	kernel versions
117	117
	118	+corrupt-filter-dev-major
	119	+corrupt-filter-dev-minor
	120	+
	121	+Only handle memory failures to pages associated with the file system defined
	122	+by block device major/minor. -1U is the wildcard value.
	123	+This should be only used for testing with artificial injection.
	124	+
118	125	Architecture specific MCE injector
119	126
120	127	x86 has mce-inject, mce-test
...	...	@@ -3,6 +3,7 @@
3	3	#include <linux/debugfs.h>
4	4	#include <linux/kernel.h>
5	5	#include <linux/mm.h>
	6	+#include "internal.h"
6	7
7	8	static struct dentry *hwpoison_dir;
8	9
...	...	@@ -51,6 +52,16 @@
51	52
52	53	dentry = debugfs_create_file("unpoison-pfn", 0600, hwpoison_dir,
53	54	NULL, &unpoison_fops);
	55	+ if (!dentry)
	56	+ goto fail;
	57	+
	58	+ dentry = debugfs_create_u32("corrupt-filter-dev-major", 0600,
	59	+ hwpoison_dir, &hwpoison_filter_dev_major);
	60	+ if (!dentry)
	61	+ goto fail;
	62	+
	63	+ dentry = debugfs_create_u32("corrupt-filter-dev-minor", 0600,
	64	+ hwpoison_dir, &hwpoison_filter_dev_minor);
54	65	if (!dentry)
55	66	goto fail;
56	67
...	...	@@ -250,4 +250,7 @@
250	250	#define ZONE_RECLAIM_SOME 0
251	251	#define ZONE_RECLAIM_SUCCESS 1
252	252	#endif
	253	+
	254	+extern u32 hwpoison_filter_dev_major;
	255	+extern u32 hwpoison_filter_dev_minor;
...	...	@@ -48,6 +48,50 @@
48	48
49	49	atomic_long_t mce_bad_pages __read_mostly = ATOMIC_LONG_INIT(0);
50	50
	51	+u32 hwpoison_filter_dev_major = ~0U;
	52	+u32 hwpoison_filter_dev_minor = ~0U;
	53	+EXPORT_SYMBOL_GPL(hwpoison_filter_dev_major);
	54	+EXPORT_SYMBOL_GPL(hwpoison_filter_dev_minor);
	55	+
	56	+static int hwpoison_filter_dev(struct page *p)
	57	+{
	58	+ struct address_space *mapping;
	59	+ dev_t dev;
	60	+
	61	+ if (hwpoison_filter_dev_major == ~0U &&
	62	+ hwpoison_filter_dev_minor == ~0U)
	63	+ return 0;
	64	+
	65	+ /*
	66	+ * page_mapping() does not accept slab page
	67	+ */
	68	+ if (PageSlab(p))
	69	+ return -EINVAL;
	70	+
	71	+ mapping = page_mapping(p);
	72	+ if (mapping == NULL \|\| mapping->host == NULL)
	73	+ return -EINVAL;
	74	+
	75	+ dev = mapping->host->i_sb->s_dev;
	76	+ if (hwpoison_filter_dev_major != ~0U &&
	77	+ hwpoison_filter_dev_major != MAJOR(dev))
	78	+ return -EINVAL;
	79	+ if (hwpoison_filter_dev_minor != ~0U &&
	80	+ hwpoison_filter_dev_minor != MINOR(dev))
	81	+ return -EINVAL;
	82	+
	83	+ return 0;
	84	+}
	85	+
	86	+int hwpoison_filter(struct page *p)
	87	+{
	88	+ if (hwpoison_filter_dev(p))
	89	+ return -EINVAL;
	90	+
	91	+ return 0;
	92	+}
	93	+EXPORT_SYMBOL_GPL(hwpoison_filter);
	94	+
51	95	/*
52	96	* Send all the processes who have the page mapped an ``action optional''
53	97	* signal.
...	...	@@ -842,6 +886,13 @@
842	886	printk(KERN_ERR "MCE %#lx: just unpoisoned\n", pfn);
843	887	res = 0;
844	888	goto out;
	889	+ }
	890	+ if (hwpoison_filter(p)) {
	891	+ if (TestClearPageHWPoison(p))
	892	+ atomic_long_dec(&mce_bad_pages);
	893	+ unlock_page(p);
	894	+ put_page(p);
	895	+ return 0;
845	896	}
846	897
847	898	wait_on_page_writeback(p);