[PATCH] drop-pagecache

Add /proc/sys/vm/drop_caches. When written to, this will cause the kernel to discard as much pagecache and/or reclaimable slab objects as it can. THis operation requires root permissions. It won't drop dirty data, so the user should run `sync' first. Caveats: a) Holds inode_lock for exorbitant amounts of time. b) Needs to be taught about NUMA nodes: propagate these all the way through so the discarding can be controlled on a per-node basis. This is a debugging feature: useful for getting consistent results between filesystem benchmarks. We could possibly put it under a config option, but it's less than 300 bytes. Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] drop-pagecache
Add /proc/sys/vm/drop_caches. When written to, this will cause the kernel to discard as much pagecache and/or reclaimable slab objects as it can. THis operation requires root permissions. It won't drop dirty data, so the user should run `sync' first. Caveats: a) Holds inode_lock for exorbitant amounts of time. b) Needs to be taught about NUMA nodes: propagate these all the way through so the discarding can be controlled on a per-node basis. This is a debugging feature: useful for getting consistent results between filesystem benchmarks. We could possibly put it under a config option, but it's less than 300 bytes. Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Andrew Morton · Linus Torvalds
1 parent bec6b0c89b
Showing 9 changed files with 107 additions and 5 deletions Side-by-side Diff
Documentation/filesystems/proc.txt
Documentation/sysctl/vm.txt
fs/Makefile
fs/drop_caches.c
include/linux/mm.h
include/linux/sysctl.h
kernel/sysctl.c
mm/truncate.c
mm/vmscan.c
@@ -1302,6 +1302,23 @@
 unnecessary page faults in thrashing situation. The unit of the value is
 second. The value would be useful to tune thrashing behavior.
  
+drop_caches
+-----------
+
+Writing to this will cause the kernel to drop clean caches, dentries and
+inodes from memory, causing that memory to become free.
+
+To free pagecache:
+	echo 1 > /proc/sys/vm/drop_caches
+To free dentries and inodes:
+	echo 2 > /proc/sys/vm/drop_caches
+To free pagecache, dentries and inodes:
+	echo 3 > /proc/sys/vm/drop_caches
+
+As this is a non-destructive operation and dirty objects are not freeable, the
+user should run `sync' first.
+
+
 2.5 /proc/sys/dev - Device specific parameters
 ----------------------------------------------
  
@@ -26,12 +26,13 @@
 - min_free_kbytes
 - laptop_mode
 - block_dump
+- drop-caches
  
 ==============================================================
  
 dirty_ratio, dirty_background_ratio, dirty_expire_centisecs,
 dirty_writeback_centisecs, vfs_cache_pressure, laptop_mode,
-block_dump, swap_token_timeout:
+block_dump, swap_token_timeout, drop-caches:
  
 See Documentation/filesystems/proc.txt
  
@@ -10,7 +10,7 @@
 		ioctl.o readdir.o select.o fifo.o locks.o dcache.o inode.o \
 		attr.o bad_inode.o file.o filesystems.o namespace.o aio.o \
 		seq_file.o xattr.o libfs.o fs-writeback.o mpage.o direct-io.o \
-		ioprio.o pnode.o
+		ioprio.o pnode.o drop_caches.o
  
 obj-$(CONFIG_INOTIFY)		+= inotify.o
 obj-$(CONFIG_EPOLL)		+= eventpoll.o
+/*
+ * Implement the manual drop-all-pagecache function
+ */
+
+#include <linux/kernel.h>
+#include <linux/mm.h>
+#include <linux/fs.h>
+#include <linux/writeback.h>
+#include <linux/sysctl.h>
+#include <linux/gfp.h>
+
+/* A global variable is a bit ugly, but it keeps the code simple */
+int sysctl_drop_caches;
+
+static void drop_pagecache_sb(struct super_block *sb)
+{
+	struct inode *inode;
+
+	spin_lock(&inode_lock);
+	list_for_each_entry(inode, &sb->s_inodes, i_sb_list) {
+		if (inode->i_state & (I_FREEING|I_WILL_FREE))
+			continue;
+		invalidate_inode_pages(inode->i_mapping);
+	}
+	spin_unlock(&inode_lock);
+}
+
+void drop_pagecache(void)
+{
+	struct super_block *sb;
+
+	spin_lock(&sb_lock);
+restart:
+	list_for_each_entry(sb, &super_blocks, s_list) {
+		sb->s_count++;
+		spin_unlock(&sb_lock);
+		down_read(&sb->s_umount);
+		if (sb->s_root)
+			drop_pagecache_sb(sb);
+		up_read(&sb->s_umount);
+		spin_lock(&sb_lock);
+		if (__put_super_and_need_restart(sb))
+			goto restart;
+	}
+	spin_unlock(&sb_lock);
+}
+
+void drop_slab(void)
+{
+	int nr_objects;
+
+	do {
+		nr_objects = shrink_slab(1000, GFP_KERNEL, 1000);
+	} while (nr_objects > 10);
+}
+
+int drop_caches_sysctl_handler(ctl_table *table, int write,
+	struct file *file, void __user *buffer, size_t *length, loff_t *ppos)
+{
+	proc_dointvec_minmax(table, write, file, buffer, length, ppos);
+	if (write) {
+		if (sysctl_drop_caches & 1)
+			drop_pagecache();
+		if (sysctl_drop_caches & 2)
+			drop_slab();
+	}
+	return 0;
+}
@@ -1036,6 +1036,13 @@
 /* /proc/<pid>/oom_adj set to -17 protects from the oom-killer */
 #define OOM_DISABLE -17
  
+int drop_caches_sysctl_handler(struct ctl_table *, int, struct file *,
+					void __user *, size_t *, loff_t *);
+int shrink_slab(unsigned long scanned, gfp_t gfp_mask,
+			unsigned long lru_pages);
+void drop_pagecache(void);
+void drop_slab(void);
+
 #endif /* __KERNEL__ */
 #endif /* _LINUX_MM_H */
@@ -180,6 +180,7 @@
 	VM_VFS_CACHE_PRESSURE=26, /* dcache/icache reclaim pressure */
 	VM_LEGACY_VA_LAYOUT=27, /* legacy/compatibility virtual address space layout */
 	VM_SWAP_TOKEN_TIMEOUT=28, /* default time for token time out */
+	VM_DROP_PAGECACHE=29,	/* int: nuke lots of pagecache */
 };
  
  
@@ -68,6 +68,7 @@
 extern int printk_ratelimit_jiffies;
 extern int printk_ratelimit_burst;
 extern int pid_max_min, pid_max_max;
+extern int sysctl_drop_caches;
  
 #if defined(CONFIG_X86_LOCAL_APIC) && defined(CONFIG_X86)
 int unknown_nmi_panic;
@@ -772,6 +773,15 @@
 		.maxlen		= sizeof(sysctl_lowmem_reserve_ratio),
 		.mode		= 0644,
 		.proc_handler	= &lowmem_reserve_ratio_sysctl_handler,
+		.strategy	= &sysctl_intvec,
+	},
+	{
+		.ctl_name	= VM_DROP_PAGECACHE,
+		.procname	= "drop_caches",
+		.data		= &sysctl_drop_caches,
+		.maxlen		= sizeof(int),
+		.mode		= 0644,
+		.proc_handler	= drop_caches_sysctl_handler,
 		.strategy	= &sysctl_intvec,
 	},
 	{
@@ -249,7 +249,6 @@
 				break;
 		}
 		pagevec_release(&pvec);
-		cond_resched();
 	}
 	return ret;
 }
@@ -180,8 +180,7 @@
  *
  * Returns the number of slab objects which we shrunk.
  */
-static int shrink_slab(unsigned long scanned, gfp_t gfp_mask,
-			unsigned long lru_pages)
+int shrink_slab(unsigned long scanned, gfp_t gfp_mask, unsigned long lru_pages)
 {
 	struct shrinker *shrinker;
 	int ret = 0;
...	...	@@ -1302,6 +1302,23 @@
1302	1302	unnecessary page faults in thrashing situation. The unit of the value is
1303	1303	second. The value would be useful to tune thrashing behavior.
1304	1304
	1305	+drop_caches
	1306	+-----------
	1307	+
	1308	+Writing to this will cause the kernel to drop clean caches, dentries and
	1309	+inodes from memory, causing that memory to become free.
	1310	+
	1311	+To free pagecache:
	1312	+ echo 1 > /proc/sys/vm/drop_caches
	1313	+To free dentries and inodes:
	1314	+ echo 2 > /proc/sys/vm/drop_caches
	1315	+To free pagecache, dentries and inodes:
	1316	+ echo 3 > /proc/sys/vm/drop_caches
	1317	+
	1318	+As this is a non-destructive operation and dirty objects are not freeable, the
	1319	+user should run `sync' first.
	1320	+
	1321	+
1305	1322	2.5 /proc/sys/dev - Device specific parameters
1306	1323	----------------------------------------------
1307	1324
...	...	@@ -26,12 +26,13 @@
26	26	- min_free_kbytes
27	27	- laptop_mode
28	28	- block_dump
	29	+- drop-caches
29	30
30	31	==============================================================
31	32
32	33	dirty_ratio, dirty_background_ratio, dirty_expire_centisecs,
33	34	dirty_writeback_centisecs, vfs_cache_pressure, laptop_mode,
34		-block_dump, swap_token_timeout:
	35	+block_dump, swap_token_timeout, drop-caches:
35	36
36	37	See Documentation/filesystems/proc.txt
37	38
...	...	@@ -10,7 +10,7 @@
10	10	ioctl.o readdir.o select.o fifo.o locks.o dcache.o inode.o \
11	11	attr.o bad_inode.o file.o filesystems.o namespace.o aio.o \
12	12	seq_file.o xattr.o libfs.o fs-writeback.o mpage.o direct-io.o \
13		- ioprio.o pnode.o
	13	+ ioprio.o pnode.o drop_caches.o
14	14
15	15	obj-$(CONFIG_INOTIFY) += inotify.o
16	16	obj-$(CONFIG_EPOLL) += eventpoll.o
	1	+/*
	2	+ * Implement the manual drop-all-pagecache function
	3	+ */
	4	+
	5	+#include <linux/kernel.h>
	6	+#include <linux/mm.h>
	7	+#include <linux/fs.h>
	8	+#include <linux/writeback.h>
	9	+#include <linux/sysctl.h>
	10	+#include <linux/gfp.h>
	11	+
	12	+/* A global variable is a bit ugly, but it keeps the code simple */
	13	+int sysctl_drop_caches;
	14	+
	15	+static void drop_pagecache_sb(struct super_block *sb)
	16	+{
	17	+ struct inode *inode;
	18	+
	19	+ spin_lock(&inode_lock);
	20	+ list_for_each_entry(inode, &sb->s_inodes, i_sb_list) {
	21	+ if (inode->i_state & (I_FREEING\|I_WILL_FREE))
	22	+ continue;
	23	+ invalidate_inode_pages(inode->i_mapping);
	24	+ }
	25	+ spin_unlock(&inode_lock);
	26	+}
	27	+
	28	+void drop_pagecache(void)
	29	+{
	30	+ struct super_block *sb;
	31	+
	32	+ spin_lock(&sb_lock);
	33	+restart:
	34	+ list_for_each_entry(sb, &super_blocks, s_list) {
	35	+ sb->s_count++;
	36	+ spin_unlock(&sb_lock);
	37	+ down_read(&sb->s_umount);
	38	+ if (sb->s_root)
	39	+ drop_pagecache_sb(sb);
	40	+ up_read(&sb->s_umount);
	41	+ spin_lock(&sb_lock);
	42	+ if (__put_super_and_need_restart(sb))
	43	+ goto restart;
	44	+ }
	45	+ spin_unlock(&sb_lock);
	46	+}
	47	+
	48	+void drop_slab(void)
	49	+{
	50	+ int nr_objects;
	51	+
	52	+ do {
	53	+ nr_objects = shrink_slab(1000, GFP_KERNEL, 1000);
	54	+ } while (nr_objects > 10);
	55	+}
	56	+
	57	+int drop_caches_sysctl_handler(ctl_table *table, int write,
	58	+ struct file file, void __user buffer, size_t length, loff_t ppos)
	59	+{
	60	+ proc_dointvec_minmax(table, write, file, buffer, length, ppos);
	61	+ if (write) {
	62	+ if (sysctl_drop_caches & 1)
	63	+ drop_pagecache();
	64	+ if (sysctl_drop_caches & 2)
	65	+ drop_slab();
	66	+ }
	67	+ return 0;
	68	+}
...	...	@@ -1036,6 +1036,13 @@
1036	1036	/* /proc/<pid>/oom_adj set to -17 protects from the oom-killer */
1037	1037	#define OOM_DISABLE -17
1038	1038
	1039	+int drop_caches_sysctl_handler(struct ctl_table , int, struct file ,
	1040	+ void __user , size_t , loff_t *);
	1041	+int shrink_slab(unsigned long scanned, gfp_t gfp_mask,
	1042	+ unsigned long lru_pages);
	1043	+void drop_pagecache(void);
	1044	+void drop_slab(void);
	1045	+
1039	1046	#endif /* __KERNEL__ */
1040	1047	#endif /* _LINUX_MM_H */
...	...	@@ -180,6 +180,7 @@
180	180	VM_VFS_CACHE_PRESSURE=26, /* dcache/icache reclaim pressure */
181	181	VM_LEGACY_VA_LAYOUT=27, /* legacy/compatibility virtual address space layout */
182	182	VM_SWAP_TOKEN_TIMEOUT=28, /* default time for token time out */
	183	+ VM_DROP_PAGECACHE=29, /* int: nuke lots of pagecache */
183	184	};
184	185
185	186
...	...	@@ -68,6 +68,7 @@
68	68	extern int printk_ratelimit_jiffies;
69	69	extern int printk_ratelimit_burst;
70	70	extern int pid_max_min, pid_max_max;
	71	+extern int sysctl_drop_caches;
71	72
72	73	#if defined(CONFIG_X86_LOCAL_APIC) && defined(CONFIG_X86)
73	74	int unknown_nmi_panic;
...	...	@@ -772,6 +773,15 @@
772	773	.maxlen = sizeof(sysctl_lowmem_reserve_ratio),
773	774	.mode = 0644,
774	775	.proc_handler = &lowmem_reserve_ratio_sysctl_handler,
	776	+ .strategy = &sysctl_intvec,
	777	+ },
	778	+ {
	779	+ .ctl_name = VM_DROP_PAGECACHE,
	780	+ .procname = "drop_caches",
	781	+ .data = &sysctl_drop_caches,
	782	+ .maxlen = sizeof(int),
	783	+ .mode = 0644,
	784	+ .proc_handler = drop_caches_sysctl_handler,
775	785	.strategy = &sysctl_intvec,
776	786	},
777	787	{
...	...	@@ -249,7 +249,6 @@
249	249	break;
250	250	}
251	251	pagevec_release(&pvec);
252		- cond_resched();
253	252	}
254	253	return ret;
255	254	}
...	...	@@ -180,8 +180,7 @@
180	180	*
181	181	* Returns the number of slab objects which we shrunk.
182	182	*/
183		-static int shrink_slab(unsigned long scanned, gfp_t gfp_mask,
184		- unsigned long lru_pages)
	183	+int shrink_slab(unsigned long scanned, gfp_t gfp_mask, unsigned long lru_pages)
185	184	{
186	185	struct shrinker *shrinker;
187	186	int ret = 0;