block: implement CONFIG_DEBUG_BLOCK_EXT_DEVT

Extended devt introduces non-contiguos device numbers. This patch implements a debug option which forces most devt allocations to be from the extended area and spreads them out. This is enabled by default if DEBUG_KERNEL is set and achieves... 1. Detects code paths in kernel or userland which expect predetermined consecutive device numbers. 2. When something goes wrong, avoid corruption as adding to the minor of earlier partition won't lead to the wrong but valid device. Signed-off-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <jens.axboe@oracle.com>

block: implement CONFIG_DEBUG_BLOCK_EXT_DEVT
Extended devt introduces non-contiguos device numbers. This patch implements a debug option which forces most devt allocations to be from the extended area and spreads them out. This is enabled by default if DEBUG_KERNEL is set and achieves... 1. Detects code paths in kernel or userland which expect predetermined consecutive device numbers. 2. When something goes wrong, avoid corruption as adding to the minor of earlier partition won't lead to the wrong but valid device. Signed-off-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Tejun Heo · Jens Axboe
1 parent f615b48cc7
Showing 4 changed files with 63 additions and 3 deletions Side-by-side Diff
block/genhd.c
drivers/ide/ide-disk.c
drivers/scsi/sd.c
lib/Kconfig.debug
@@ -299,6 +299,38 @@
 static struct kobj_map *bdev_map;
  
 /**
+ * blk_mangle_minor - scatter minor numbers apart
+ * @minor: minor number to mangle
+ *
+ * Scatter consecutively allocated @minor number apart if MANGLE_DEVT
+ * is enabled.  Mangling twice gives the original value.
+ *
+ * RETURNS:
+ * Mangled value.
+ *
+ * CONTEXT:
+ * Don't care.
+ */
+static int blk_mangle_minor(int minor)
+{
+#ifdef CONFIG_DEBUG_BLOCK_EXT_DEVT
+	int i;
+
+	for (i = 0; i < MINORBITS / 2; i++) {
+		int low = minor & (1 << i);
+		int high = minor & (1 << (MINORBITS - 1 - i));
+		int distance = MINORBITS - 1 - 2 * i;
+
+		minor ^= low | high;	/* clear both bits */
+		low <<= distance;	/* swap the positions */
+		high >>= distance;
+		minor |= low | high;	/* and set */
+	}
+#endif
+	return minor;
+}
+
+/**
  * blk_alloc_devt - allocate a dev_t for a partition
  * @part: partition to allocate dev_t for
  * @gfp_mask: memory allocation flag
@@ -339,7 +371,7 @@
 		return -EBUSY;
 	}
  
-	*devt = MKDEV(BLOCK_EXT_MAJOR, idx);
+	*devt = MKDEV(BLOCK_EXT_MAJOR, blk_mangle_minor(idx));
 	return 0;
 }
  
@@ -361,7 +393,7 @@
  
 	if (MAJOR(devt) == BLOCK_EXT_MAJOR) {
 		mutex_lock(&ext_devt_mutex);
-		idr_remove(&ext_devt_idr, MINOR(devt));
+		idr_remove(&ext_devt_idr, blk_mangle_minor(MINOR(devt)));
 		mutex_unlock(&ext_devt_mutex);
 	}
 }
@@ -473,7 +505,7 @@
 		struct hd_struct *part;
  
 		mutex_lock(&ext_devt_mutex);
-		part = idr_find(&ext_devt_idr, MINOR(devt));
+		part = idr_find(&ext_devt_idr, blk_mangle_minor(MINOR(devt)));
 		if (part && get_disk(part_to_disk(part))) {
 			*partno = part->partno;
 			disk = part_to_disk(part);
@@ -42,7 +42,13 @@
 #include <asm/div64.h>
  
 #define IDE_DISK_PARTS		(1 << PARTN_BITS)
+
+#if !defined(CONFIG_DEBUG_BLOCK_EXT_DEVT)
 #define IDE_DISK_MINORS		IDE_DISK_PARTS
+#else
+#define IDE_DISK_MINORS		1
+#endif
+
 #define IDE_DISK_EXT_MINORS	(IDE_DISK_PARTS - IDE_DISK_MINORS)
  
 struct ide_disk_obj {
@@ -87,7 +87,13 @@
 MODULE_ALIAS_SCSI_DEVICE(TYPE_RBC);
  
 #define SD_PARTS	64
+
+#if !defined(CONFIG_DEBUG_BLOCK_EXT_DEVT)
 #define SD_MINORS	16
+#else
+#define SD_MINORS	1
+#endif
+
 #define SD_EXT_MINORS	(SD_PARTS - SD_MINORS)
  
 static int  sd_revalidate_disk(struct gendisk *);
@@ -624,6 +624,22 @@
  
 	  Say N if you are unsure.
  
+config DEBUG_BLOCK_EXT_DEVT
+        bool "Force extended block device numbers and spread them"
+	depends on DEBUG_KERNEL
+	depends on BLOCK
+	default y
+	help
+	  Conventionally, block device numbers are allocated from
+	  predetermined contiguous area.  However, extended block area
+	  may introduce non-contiguous block device numbers.  This
+	  option forces most block device numbers to be allocated from
+	  the extended space and spreads them to discover kernel or
+	  userland code paths which assume predetermined contiguous
+	  device number allocation.
+
+	  Say N if you are unsure.
+
 config LKDTM
 	tristate "Linux Kernel Dump Test Tool Module"
 	depends on DEBUG_KERNEL
...	...	@@ -299,6 +299,38 @@
299	299	static struct kobj_map *bdev_map;
300	300
301	301	/**
	302	+ * blk_mangle_minor - scatter minor numbers apart
	303	+ * @minor: minor number to mangle
	304	+ *
	305	+ * Scatter consecutively allocated @minor number apart if MANGLE_DEVT
	306	+ * is enabled. Mangling twice gives the original value.
	307	+ *
	308	+ * RETURNS:
	309	+ * Mangled value.
	310	+ *
	311	+ * CONTEXT:
	312	+ * Don't care.
	313	+ */
	314	+static int blk_mangle_minor(int minor)
	315	+{
	316	+#ifdef CONFIG_DEBUG_BLOCK_EXT_DEVT
	317	+ int i;
	318	+
	319	+ for (i = 0; i < MINORBITS / 2; i++) {
	320	+ int low = minor & (1 << i);
	321	+ int high = minor & (1 << (MINORBITS - 1 - i));
	322	+ int distance = MINORBITS - 1 - 2 * i;
	323	+
	324	+ minor ^= low \| high; /* clear both bits */
	325	+ low <<= distance; /* swap the positions */
	326	+ high >>= distance;
	327	+ minor \|= low \| high; /* and set */
	328	+ }
	329	+#endif
	330	+ return minor;
	331	+}
	332	+
	333	+/**
302	334	* blk_alloc_devt - allocate a dev_t for a partition
303	335	* @part: partition to allocate dev_t for
304	336	* @gfp_mask: memory allocation flag
...	...	@@ -339,7 +371,7 @@
339	371	return -EBUSY;
340	372	}
341	373
342		- *devt = MKDEV(BLOCK_EXT_MAJOR, idx);
	374	+ *devt = MKDEV(BLOCK_EXT_MAJOR, blk_mangle_minor(idx));
343	375	return 0;
344	376	}
345	377
...	...	@@ -361,7 +393,7 @@
361	393
362	394	if (MAJOR(devt) == BLOCK_EXT_MAJOR) {
363	395	mutex_lock(&ext_devt_mutex);
364		- idr_remove(&ext_devt_idr, MINOR(devt));
	396	+ idr_remove(&ext_devt_idr, blk_mangle_minor(MINOR(devt)));
365	397	mutex_unlock(&ext_devt_mutex);
366	398	}
367	399	}
...	...	@@ -473,7 +505,7 @@
473	505	struct hd_struct *part;
474	506
475	507	mutex_lock(&ext_devt_mutex);
476		- part = idr_find(&ext_devt_idr, MINOR(devt));
	508	+ part = idr_find(&ext_devt_idr, blk_mangle_minor(MINOR(devt)));
477	509	if (part && get_disk(part_to_disk(part))) {
478	510	*partno = part->partno;
479	511	disk = part_to_disk(part);
...	...	@@ -42,7 +42,13 @@
42	42	#include <asm/div64.h>
43	43
44	44	#define IDE_DISK_PARTS (1 << PARTN_BITS)
	45	+
	46	+#if !defined(CONFIG_DEBUG_BLOCK_EXT_DEVT)
45	47	#define IDE_DISK_MINORS IDE_DISK_PARTS
	48	+#else
	49	+#define IDE_DISK_MINORS 1
	50	+#endif
	51	+
46	52	#define IDE_DISK_EXT_MINORS (IDE_DISK_PARTS - IDE_DISK_MINORS)
47	53
48	54	struct ide_disk_obj {
...	...	@@ -87,7 +87,13 @@
87	87	MODULE_ALIAS_SCSI_DEVICE(TYPE_RBC);
88	88
89	89	#define SD_PARTS 64
	90	+
	91	+#if !defined(CONFIG_DEBUG_BLOCK_EXT_DEVT)
90	92	#define SD_MINORS 16
	93	+#else
	94	+#define SD_MINORS 1
	95	+#endif
	96	+
91	97	#define SD_EXT_MINORS (SD_PARTS - SD_MINORS)
92	98
93	99	static int sd_revalidate_disk(struct gendisk *);
...	...	@@ -624,6 +624,22 @@
624	624
625	625	Say N if you are unsure.
626	626
	627	+config DEBUG_BLOCK_EXT_DEVT
	628	+ bool "Force extended block device numbers and spread them"
	629	+ depends on DEBUG_KERNEL
	630	+ depends on BLOCK
	631	+ default y
	632	+ help
	633	+ Conventionally, block device numbers are allocated from
	634	+ predetermined contiguous area. However, extended block area
	635	+ may introduce non-contiguous block device numbers. This
	636	+ option forces most block device numbers to be allocated from
	637	+ the extended space and spreads them to discover kernel or
	638	+ userland code paths which assume predetermined contiguous
	639	+ device number allocation.
	640	+
	641	+ Say N if you are unsure.
	642	+
627	643	config LKDTM
628	644	tristate "Linux Kernel Dump Test Tool Module"
629	645	depends on DEBUG_KERNEL