[RFC PATCH v2 2/2] mm: tuning hardcoded reserved memory

From: Andrew Shewmaker
Date: Wed Feb 27 2013 - 16:53:13 EST


Add a rootuser_reserve_pages knob to allow admins of large memory
systems running with overcommit disabled to change the hardcoded
memory reserve to something other than 3%.

Signed-off-by: Andrew Shewmaker <agshew@xxxxxxxxx>

---

Patch based off of mmotm git tree as of February 27th.

I set rootuser_reserve pages to be a default of 1000, and I suppose
I should have initialzed similarly to the way min_free_kbytes is,
scaling it with the size of the box. However, I wanted to get a
simple version of this patch out for feedback to see if it has any
chance of acceptance or if I need to take an entirely different
approach.

Any feedback will be appreciated!

Documentation/sysctl/vm.txt | 9 +++++++++
include/linux/mm.h | 2 ++
kernel/sysctl.c | 8 ++++++++
mm/mmap.c | 30 +++++++++++++++++++++++-------
4 files changed, 42 insertions(+), 7 deletions(-)

diff --git a/Documentation/sysctl/vm.txt b/Documentation/sysctl/vm.txt
index 078701f..3a71de9 100644
--- a/Documentation/sysctl/vm.txt
+++ b/Documentation/sysctl/vm.txt
@@ -51,6 +51,7 @@ Currently, these files are in /proc/sys/vm:
- page-cluster
- panic_on_oom
- percpu_pagelist_fraction
+- rootuser_reserve_pages
- stat_interval
- swappiness
- vfs_cache_pressure
@@ -628,6 +629,14 @@ the high water marks for each per cpu page list.

==============================================================

+rootuser_reserve_pages
+
+The number of free pages left in the system that should be reserved for users
+with the capability cap_sys_admin. The default falue is 3% of total system
+memory. Changing this takes effect whenever an application requests memory.
+
+==============================================================
+
stat_interval

The time interval between which vm statistics are updated. The default
diff --git a/include/linux/mm.h b/include/linux/mm.h
index 66e2f7c..af7b39f 100644
--- a/include/linux/mm.h
+++ b/include/linux/mm.h
@@ -1677,6 +1677,8 @@ int in_gate_area_no_mm(unsigned long addr);

int drop_caches_sysctl_handler(struct ctl_table *, int,
void __user *, size_t *, loff_t *);
+int rootuser_reserve_pages_sysctl_handler(struct ctl_table *, int,
+ void __user *, size_t *, loff_t *);
unsigned long shrink_slab(struct shrink_control *shrink,
unsigned long nr_pages_scanned,
unsigned long lru_pages);
diff --git a/kernel/sysctl.c b/kernel/sysctl.c
index c88878d..cd1987e 100644
--- a/kernel/sysctl.c
+++ b/kernel/sysctl.c
@@ -96,6 +96,7 @@
/* External variables not in a header file. */
extern int sysctl_overcommit_memory;
extern int sysctl_overcommit_ratio;
+extern int sysctl_rootuser_reserve_pages;
extern int max_threads;
extern int suid_dumpable;
#ifdef CONFIG_COREDUMP
@@ -1413,6 +1414,13 @@ static struct ctl_table vm_table[] = {
.extra2 = &one,
},
#endif
+ {
+ .procname = "rootuser_reserve_pages",
+ .data = &sysctl_rootuser_reserve_pages,
+ .maxlen = sizeof(sysctl_rootuser_reserve_pages),
+ .mode = 0644,
+ .proc_handler = rootuser_reserve_pages_sysctl_handler,
+ },
{ }
};

diff --git a/mm/mmap.c b/mm/mmap.c
index d1e4124..b58af97 100644
--- a/mm/mmap.c
+++ b/mm/mmap.c
@@ -32,6 +32,7 @@
#include <linux/khugepaged.h>
#include <linux/uprobes.h>
#include <linux/rbtree_augmented.h>
+#include <linux/sysctl.h>

#include <asm/uaccess.h>
#include <asm/cacheflush.h>
@@ -83,6 +84,7 @@ EXPORT_SYMBOL(vm_get_page_prot);
int sysctl_overcommit_memory __read_mostly = OVERCOMMIT_GUESS; /* heuristic overcommit */
int sysctl_overcommit_ratio __read_mostly = 50; /* default is 50% */
int sysctl_max_map_count __read_mostly = DEFAULT_MAX_MAP_COUNT;
+int sysctl_rootuser_reserve_pages __read_mostly = 1000;
/*
* Make sure vm_committed_as in one cacheline and not cacheline shared with
* other variables. It can be updated by several CPUs frequently.
@@ -165,7 +167,7 @@ int __vm_enough_memory(struct mm_struct *mm, long pages, int cap_sys_admin)
* Leave the last 3% for root
*/
if (!cap_sys_admin)
- free -= free / 32;
+ free -= sysctl_rootuser_reserve_pages;

if (free > pages)
return 0;
@@ -179,9 +181,9 @@ int __vm_enough_memory(struct mm_struct *mm, long pages, int cap_sys_admin)
* Leave the last 3% for root
*/
if (!cap_sys_admin)
- allowed -= allowed / 32;
+ allowed -= sysctl_rootuser_reserve_pages;
allowed += total_swap_pages;

if (percpu_counter_read_positive(&vm_committed_as) < allowed)
return 0;
error:
@@ -3052,3 +3049,22 @@ void __init mmap_init(void)
ret = percpu_counter_init(&vm_committed_as, 0);
VM_BUG_ON(ret);
}
+
+/*
+ * rootuser_reserve_pages_sysctl_handler - just a wrapper around proc_dointvec_minmax() so
+ * that we can cap the number of pages to the current number of free pages.
+ */
+int rootuser_reserve_pages_sysctl_handler(ctl_table *table, int write,
+ void __user *buffer, size_t *length, loff_t *ppos)
+{
+ unsigned long free;
+
+ proc_dointvec(table, write, buffer, length, ppos);
+
+ if (write) {
+ free = global_page_state(NR_FREE_PAGES);
+ if (sysctl_rootuser_reserve_pages > free)
+ sysctl_rootuser_reserve_pages = free;
+ }
+ return 0;
+}
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/