mm: add a basic debugging framework for memory initialisation

Boot initialisation is very complex, with significant numbers of architecture-specific routines, hooks and code ordering. While significant amounts of the initialisation is architecture-independent, it trusts the data received from the architecture layer. This is a mistake, and has resulted in a number of difficult-to-diagnose bugs. This patchset adds some validation and tracing to memory initialisation. It also introduces a few basic defensive measures. The validation code can be explicitly disabled for embedded systems. This patch: Add additional debugging and verification code for memory initialisation. Once enabled, the verification checks are always run and when required additional debugging information may be outputted via a mminit_loglevel= command-line parameter. The verification code is placed in a new file mm/mm_init.c. Ideally other mm initialisation code will be moved here over time. Signed-off-by: Mel Gorman <mel@csn.ul.ie> Cc: Christoph Lameter <cl@linux-foundation.org> Cc: Andy Whitcroft <apw@shadowen.org> Cc: Ingo Molnar <mingo@elte.hu> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

mm: add a basic debugging framework for memory initialisation
Boot initialisation is very complex, with significant numbers of architecture-specific routines, hooks and code ordering. While significant amounts of the initialisation is architecture-independent, it trusts the data received from the architecture layer. This is a mistake, and has resulted in a number of difficult-to-diagnose bugs. This patchset adds some validation and tracing to memory initialisation. It also introduces a few basic defensive measures. The validation code can be explicitly disabled for embedded systems. This patch: Add additional debugging and verification code for memory initialisation. Once enabled, the verification checks are always run and when required additional debugging information may be outputted via a mminit_loglevel= command-line parameter. The verification code is placed in a new file mm/mm_init.c. Ideally other mm initialisation code will be moved here over time. Signed-off-by: Mel Gorman <mel@csn.ul.ie> Cc: Christoph Lameter <cl@linux-foundation.org> Cc: Andy Whitcroft <apw@shadowen.org> Cc: Ingo Molnar <mingo@elte.hu> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Mel Gorman · Linus Torvalds
1 parent 9483a578df
Showing 6 changed files with 79 additions and 9 deletions Side-by-side Diff
Documentation/kernel-parameters.txt
lib/Kconfig.debug
mm/Makefile
mm/internal.h
mm/mm_init.c
mm/page_alloc.c
@@ -1225,6 +1225,14 @@
  
 	mga=		[HW,DRM]
  
+	mminit_loglevel=
+			[KNL] When CONFIG_DEBUG_MEMORY_INIT is set, this
+			parameter allows control of the logging verbosity for
+			the additional memory initialisation checks. A value
+			of 0 disables mminit logging and a level of 4 will
+			log everything. Information is printed at KERN_DEBUG
+			so loglevel=8 may also need to be specified.
+
 	mousedev.tap_time=
 			[MOUSE] Maximum time between finger touching and
 			leaving touchpad surface for touch to be considered
@@ -505,6 +505,18 @@
  
 	  If unsure, say N.
  
+config DEBUG_MEMORY_INIT
+	bool "Debug memory initialisation" if EMBEDDED
+	default !EMBEDDED
+	help
+	  Enable this for additional checks during memory initialisation.
+	  The sanity checks verify aspects of the VM such as the memory model
+	  and other information provided by the architecture. Verbose
+	  information will be printed at KERN_DEBUG loglevel depending
+	  on the mminit_loglevel= command-line option.
+
+	  If unsure, say Y
+
 config DEBUG_LIST
 	bool "Debug linked list manipulation"
 	depends on DEBUG_KERNEL
@@ -26,6 +26,7 @@
 obj-$(CONFIG_TINY_SHMEM) += tiny-shmem.o
 obj-$(CONFIG_SLOB) += slob.o
 obj-$(CONFIG_SLAB) += slab.o
+obj-$(CONFIG_DEBUG_MEMORY_INIT) += mm_init.o
 obj-$(CONFIG_SLUB) += slub.o
 obj-$(CONFIG_MEMORY_HOTPLUG) += memory_hotplug.o
 obj-$(CONFIG_FS_XIP) += filemap_xip.o
@@ -59,5 +59,32 @@
 #define __paginginit __init
 #endif
  
+/* Memory initialisation debug and verification */
+enum mminit_level {
+	MMINIT_WARNING,
+	MMINIT_VERIFY,
+	MMINIT_TRACE
+};
+
+#ifdef CONFIG_DEBUG_MEMORY_INIT
+
+extern int mminit_loglevel;
+
+#define mminit_dprintk(level, prefix, fmt, arg...) \
+do { \
+	if (level < mminit_loglevel) { \
+		printk(level <= MMINIT_WARNING ? KERN_WARNING : KERN_DEBUG); \
+		printk(KERN_CONT "mminit::" prefix " " fmt, ##arg); \
+	} \
+} while (0)
+
+#else
+
+static inline void mminit_dprintk(enum mminit_level level,
+				const char *prefix, const char *fmt, ...)
+{
+}
+
+#endif /* CONFIG_DEBUG_MEMORY_INIT */
 #endif
+/*
+ * mm_init.c - Memory initialisation verification and debugging
+ *
+ * Copyright 2008 IBM Corporation, 2008
+ * Author Mel Gorman <mel@csn.ul.ie>
+ *
+ */
+#include <linux/kernel.h>
+#include <linux/init.h>
+
+int __meminitdata mminit_loglevel;
+
+static __init int set_mminit_loglevel(char *str)
+{
+	get_option(&str, &mminit_loglevel);
+	return 0;
+}
+early_param("mminit_loglevel", set_mminit_loglevel);
@@ -2975,7 +2975,8 @@
 void __init push_node_boundaries(unsigned int nid,
 		unsigned long start_pfn, unsigned long end_pfn)
 {
-	printk(KERN_DEBUG "Entering push_node_boundaries(%u, %lu, %lu)\n",
+	mminit_dprintk(MMINIT_TRACE, "zoneboundary",
+			"Entering push_node_boundaries(%u, %lu, %lu)\n",
 			nid, start_pfn, end_pfn);
  
 	/* Initialise the boundary for this node if necessary */
@@ -2993,7 +2994,8 @@
 static void __meminit account_node_boundary(unsigned int nid,
 		unsigned long *start_pfn, unsigned long *end_pfn)
 {
-	printk(KERN_DEBUG "Entering account_node_boundary(%u, %lu, %lu)\n",
+	mminit_dprintk(MMINIT_TRACE, "zoneboundary",
+			"Entering account_node_boundary(%u, %lu, %lu)\n",
 			nid, *start_pfn, *end_pfn);
  
 	/* Return if boundary information has not been provided */
@@ -3368,8 +3370,8 @@
 			PAGE_ALIGN(size * sizeof(struct page)) >> PAGE_SHIFT;
 		if (realsize >= memmap_pages) {
 			realsize -= memmap_pages;
-			printk(KERN_DEBUG
-				"  %s zone: %lu pages used for memmap\n",
+			mminit_dprintk(MMINIT_TRACE, "memmap_init",
+				"%s zone: %lu pages used for memmap\n",
 				zone_names[j], memmap_pages);
 		} else
 			printk(KERN_WARNING
@@ -3379,7 +3381,8 @@
 		/* Account for reserved pages */
 		if (j == 0 && realsize > dma_reserve) {
 			realsize -= dma_reserve;
-			printk(KERN_DEBUG "  %s zone: %lu pages reserved\n",
+			mminit_dprintk(MMINIT_TRACE, "memmap_init",
+					"%s zone: %lu pages reserved\n",
 					zone_names[0], dma_reserve);
 		}
  
@@ -3520,10 +3523,11 @@
 {
 	int i;
  
-	printk(KERN_DEBUG "Entering add_active_range(%d, %#lx, %#lx) "
-			  "%d entries of %d used\n",
-			  nid, start_pfn, end_pfn,
-			  nr_nodemap_entries, MAX_ACTIVE_REGIONS);
+	mminit_dprintk(MMINIT_TRACE, "memory_register",
+			"Entering add_active_range(%d, %#lx, %#lx) "
+			"%d entries of %d used\n",
+			nid, start_pfn, end_pfn,
+			nr_nodemap_entries, MAX_ACTIVE_REGIONS);
  
 	/* Merge with existing active regions if possible */
 	for (i = 0; i < nr_nodemap_entries; i++) {
...	...	@@ -1225,6 +1225,14 @@
1225	1225
1226	1226	mga= [HW,DRM]
1227	1227
	1228	+ mminit_loglevel=
	1229	+ [KNL] When CONFIG_DEBUG_MEMORY_INIT is set, this
	1230	+ parameter allows control of the logging verbosity for
	1231	+ the additional memory initialisation checks. A value
	1232	+ of 0 disables mminit logging and a level of 4 will
	1233	+ log everything. Information is printed at KERN_DEBUG
	1234	+ so loglevel=8 may also need to be specified.
	1235	+
1228	1236	mousedev.tap_time=
1229	1237	[MOUSE] Maximum time between finger touching and
1230	1238	leaving touchpad surface for touch to be considered
...	...	@@ -505,6 +505,18 @@
505	505
506	506	If unsure, say N.
507	507
	508	+config DEBUG_MEMORY_INIT
	509	+ bool "Debug memory initialisation" if EMBEDDED
	510	+ default !EMBEDDED
	511	+ help
	512	+ Enable this for additional checks during memory initialisation.
	513	+ The sanity checks verify aspects of the VM such as the memory model
	514	+ and other information provided by the architecture. Verbose
	515	+ information will be printed at KERN_DEBUG loglevel depending
	516	+ on the mminit_loglevel= command-line option.
	517	+
	518	+ If unsure, say Y
	519	+
508	520	config DEBUG_LIST
509	521	bool "Debug linked list manipulation"
510	522	depends on DEBUG_KERNEL
...	...	@@ -26,6 +26,7 @@
26	26	obj-$(CONFIG_TINY_SHMEM) += tiny-shmem.o
27	27	obj-$(CONFIG_SLOB) += slob.o
28	28	obj-$(CONFIG_SLAB) += slab.o
	29	+obj-$(CONFIG_DEBUG_MEMORY_INIT) += mm_init.o
29	30	obj-$(CONFIG_SLUB) += slub.o
30	31	obj-$(CONFIG_MEMORY_HOTPLUG) += memory_hotplug.o
31	32	obj-$(CONFIG_FS_XIP) += filemap_xip.o
...	...	@@ -59,5 +59,32 @@
59	59	#define __paginginit __init
60	60	#endif
61	61
	62	+/* Memory initialisation debug and verification */
	63	+enum mminit_level {
	64	+ MMINIT_WARNING,
	65	+ MMINIT_VERIFY,
	66	+ MMINIT_TRACE
	67	+};
	68	+
	69	+#ifdef CONFIG_DEBUG_MEMORY_INIT
	70	+
	71	+extern int mminit_loglevel;
	72	+
	73	+#define mminit_dprintk(level, prefix, fmt, arg...) \
	74	+do { \
	75	+ if (level < mminit_loglevel) { \
	76	+ printk(level <= MMINIT_WARNING ? KERN_WARNING : KERN_DEBUG); \
	77	+ printk(KERN_CONT "mminit::" prefix " " fmt, ##arg); \
	78	+ } \
	79	+} while (0)
	80	+
	81	+#else
	82	+
	83	+static inline void mminit_dprintk(enum mminit_level level,
	84	+ const char prefix, const char fmt, ...)
	85	+{
	86	+}
	87	+
	88	+#endif /* CONFIG_DEBUG_MEMORY_INIT */
62	89	#endif
	1	+/*
	2	+ * mm_init.c - Memory initialisation verification and debugging
	3	+ *
	4	+ * Copyright 2008 IBM Corporation, 2008
	5	+ * Author Mel Gorman <mel@csn.ul.ie>
	6	+ *
	7	+ */
	8	+#include <linux/kernel.h>
	9	+#include <linux/init.h>
	10	+
	11	+int __meminitdata mminit_loglevel;
	12	+
	13	+static __init int set_mminit_loglevel(char *str)
	14	+{
	15	+ get_option(&str, &mminit_loglevel);
	16	+ return 0;
	17	+}
	18	+early_param("mminit_loglevel", set_mminit_loglevel);
...	...	@@ -2975,7 +2975,8 @@
2975	2975	void __init push_node_boundaries(unsigned int nid,
2976	2976	unsigned long start_pfn, unsigned long end_pfn)
2977	2977	{
2978		- printk(KERN_DEBUG "Entering push_node_boundaries(%u, %lu, %lu)\n",
	2978	+ mminit_dprintk(MMINIT_TRACE, "zoneboundary",
	2979	+ "Entering push_node_boundaries(%u, %lu, %lu)\n",
2979	2980	nid, start_pfn, end_pfn);
2980	2981
2981	2982	/* Initialise the boundary for this node if necessary */
...	...	@@ -2993,7 +2994,8 @@
2993	2994	static void __meminit account_node_boundary(unsigned int nid,
2994	2995	unsigned long start_pfn, unsigned long end_pfn)
2995	2996	{
2996		- printk(KERN_DEBUG "Entering account_node_boundary(%u, %lu, %lu)\n",
	2997	+ mminit_dprintk(MMINIT_TRACE, "zoneboundary",
	2998	+ "Entering account_node_boundary(%u, %lu, %lu)\n",
2997	2999	nid, start_pfn, end_pfn);
2998	3000
2999	3001	/* Return if boundary information has not been provided */
...	...	@@ -3368,8 +3370,8 @@
3368	3370	PAGE_ALIGN(size * sizeof(struct page)) >> PAGE_SHIFT;
3369	3371	if (realsize >= memmap_pages) {
3370	3372	realsize -= memmap_pages;
3371		- printk(KERN_DEBUG
3372		- " %s zone: %lu pages used for memmap\n",
	3373	+ mminit_dprintk(MMINIT_TRACE, "memmap_init",
	3374	+ "%s zone: %lu pages used for memmap\n",
3373	3375	zone_names[j], memmap_pages);
3374	3376	} else
3375	3377	printk(KERN_WARNING
...	...	@@ -3379,7 +3381,8 @@
3379	3381	/* Account for reserved pages */
3380	3382	if (j == 0 && realsize > dma_reserve) {
3381	3383	realsize -= dma_reserve;
3382		- printk(KERN_DEBUG " %s zone: %lu pages reserved\n",
	3384	+ mminit_dprintk(MMINIT_TRACE, "memmap_init",
	3385	+ "%s zone: %lu pages reserved\n",
3383	3386	zone_names[0], dma_reserve);
3384	3387	}
3385	3388
...	...	@@ -3520,10 +3523,11 @@
3520	3523	{
3521	3524	int i;
3522	3525
3523		- printk(KERN_DEBUG "Entering add_active_range(%d, %#lx, %#lx) "
3524		- "%d entries of %d used\n",
3525		- nid, start_pfn, end_pfn,
3526		- nr_nodemap_entries, MAX_ACTIVE_REGIONS);
	3526	+ mminit_dprintk(MMINIT_TRACE, "memory_register",
	3527	+ "Entering add_active_range(%d, %#lx, %#lx) "
	3528	+ "%d entries of %d used\n",
	3529	+ nid, start_pfn, end_pfn,
	3530	+ nr_nodemap_entries, MAX_ACTIVE_REGIONS);
3527	3531
3528	3532	/* Merge with existing active regions if possible */
3529	3533	for (i = 0; i < nr_nodemap_entries; i++) {