From: Rik van Riel <riel@redhat.com>
the patch below (softly) enforces RLIMIT_RSS in the 2.6 kernel, it has been
tested by Pavel and seems to work ok for his workload.
---
include/linux/init_task.h | 2 ++
include/linux/sched.h | 1 +
include/linux/swap.h | 4 ++--
kernel/sys.c | 8 ++++++++
mm/rmap.c | 18 +++++++++++++++++-
mm/vmscan.c | 12 ++++++++----
6 files changed, 38 insertions(+), 7 deletions(-)
diff -puN include/linux/init_task.h~vm-rss-limit-enforcement include/linux/init_task.h
--- 25/include/linux/init_task.h~vm-rss-limit-enforcement 2004-01-29 17:44:04.000000000 -0800
+++ 25-akpm/include/linux/init_task.h 2004-01-29 19:11:49.000000000 -0800
@@ -2,6 +2,7 @@
#define _LINUX__INIT_TASK_H
#include <linux/file.h>
+#include <linux/resource.h>
#define INIT_FILES \
{ \
@@ -41,6 +42,7 @@
.page_table_lock = SPIN_LOCK_UNLOCKED, \
.mmlist = LIST_HEAD_INIT(name.mmlist), \
.default_kioctx = INIT_KIOCTX(name.default_kioctx, name), \
+ .rlimit_rss = RLIM_INFINITY \
}
#define INIT_SIGNALS(sig) { \
diff -puN include/linux/sched.h~vm-rss-limit-enforcement include/linux/sched.h
--- 25/include/linux/sched.h~vm-rss-limit-enforcement 2004-01-29 17:44:04.000000000 -0800
+++ 25-akpm/include/linux/sched.h 2004-01-29 19:12:09.000000000 -0800
@@ -205,6 +205,7 @@ struct mm_struct {
unsigned long arg_start, arg_end, env_start, env_end;
unsigned long rss, total_vm, locked_vm;
unsigned long def_flags;
+ unsigned long rlimit_rss;
cpumask_t cpu_vm_mask;
unsigned long saved_auxv[40]; /* for /proc/PID/auxv */
diff -puN include/linux/swap.h~vm-rss-limit-enforcement include/linux/swap.h
--- 25/include/linux/swap.h~vm-rss-limit-enforcement 2004-01-29 17:44:04.000000000 -0800
+++ 25-akpm/include/linux/swap.h 2004-01-29 17:44:04.000000000 -0800
@@ -179,7 +179,7 @@ extern int vm_swappiness;
/* linux/mm/rmap.c */
#ifdef CONFIG_MMU
-int FASTCALL(page_referenced(struct page *));
+int FASTCALL(page_referenced(struct page *, int *));
struct pte_chain *FASTCALL(page_add_rmap(struct page *, pte_t *,
struct pte_chain *));
void FASTCALL(page_remove_rmap(struct page *, pte_t *));
@@ -188,7 +188,7 @@ int FASTCALL(try_to_unmap(struct page *)
/* linux/mm/shmem.c */
extern int shmem_unuse(swp_entry_t entry, struct page *page);
#else
-#define page_referenced(page) TestClearPageReferenced(page)
+#define page_referenced(page, _x) TestClearPageReferenced(page)
#define try_to_unmap(page) SWAP_FAIL
#endif /* CONFIG_MMU */
diff -puN kernel/sys.c~vm-rss-limit-enforcement kernel/sys.c
--- 25/kernel/sys.c~vm-rss-limit-enforcement 2004-01-29 17:44:04.000000000 -0800
+++ 25-akpm/kernel/sys.c 2004-01-29 19:11:51.000000000 -0800
@@ -1308,6 +1308,14 @@ asmlinkage long sys_setrlimit(unsigned i
if (retval)
return retval;
+ /* The rlimit is specified in bytes, convert to pages for mm. */
+ if (resource == RLIMIT_RSS && current->mm) {
+ unsigned long pages = RLIM_INFINITY;
+ if (new_rlim.rlim_cur != RLIM_INFINITY)
+ pages = new_rlim.rlim_cur >> PAGE_SHIFT;
+ current->mm->rlimit_rss = pages;
+ }
+
*old_rlim = new_rlim;
return 0;
}
diff -puN mm/rmap.c~vm-rss-limit-enforcement mm/rmap.c
--- 25/mm/rmap.c~vm-rss-limit-enforcement 2004-01-29 17:44:04.000000000 -0800
+++ 25-akpm/mm/rmap.c 2004-01-29 19:12:20.000000000 -0800
@@ -104,6 +104,7 @@ pte_chain_encode(struct pte_chain *pte_c
/**
* page_referenced - test if the page was referenced
* @page: the page to test
+ * @rsslimit: set if the process(es) using the page is(are) over RSS limit.
*
* Quick test_and_clear_referenced for all mappings to a page,
* returns the number of processes which referenced the page.
@@ -111,9 +112,13 @@ pte_chain_encode(struct pte_chain *pte_c
*
* If the page has a single-entry pte_chain, collapse that back to a PageDirect
* representation. This way, it's only done under memory pressure.
+ *
+ * The pte_chain_lock() is sufficient to pin down mm_structs while we examine
+ * them.
*/
-int page_referenced(struct page * page)
+int page_referenced(struct page *page, int *rsslimit)
{
+ struct mm_struct * mm;
struct pte_chain *pc;
int referenced = 0;
@@ -127,10 +132,17 @@ int page_referenced(struct page * page)
pte_t *pte = rmap_ptep_map(page->pte.direct);
if (ptep_test_and_clear_young(pte))
referenced++;
+
+ mm = ptep_to_mm(pte);
+ if (mm->rss > mm->rlimit_rss)
+ *rsslimit = 1;
rmap_ptep_unmap(pte);
} else {
int nr_chains = 0;
+ /* We clear it if any task using the page is under its limit. */
+ *rsslimit = 1;
+
/* Check all the page tables mapping this page. */
for (pc = page->pte.chain; pc; pc = pte_chain_next(pc)) {
int i;
@@ -142,6 +154,10 @@ int page_referenced(struct page * page)
p = rmap_ptep_map(pte_paddr);
if (ptep_test_and_clear_young(p))
referenced++;
+
+ mm = ptep_to_mm(p);
+ if (mm->rss < mm->rlimit_rss)
+ *rsslimit = 0;
rmap_ptep_unmap(p);
nr_chains++;
}
diff -puN mm/vmscan.c~vm-rss-limit-enforcement mm/vmscan.c
--- 25/mm/vmscan.c~vm-rss-limit-enforcement 2004-01-29 17:44:04.000000000 -0800
+++ 25-akpm/mm/vmscan.c 2004-01-29 19:12:39.000000000 -0800
@@ -250,6 +250,7 @@ shrink_list(struct list_head *page_list,
LIST_HEAD(ret_pages);
struct pagevec freed_pvec;
int pgactivate = 0;
+ int over_rsslimit = 0;
int ret = 0;
cond_resched();
@@ -278,8 +279,8 @@ shrink_list(struct list_head *page_list,
goto keep_locked;
pte_chain_lock(page);
- referenced = page_referenced(page);
- if (referenced && page_mapping_inuse(page)) {
+ referenced = page_referenced(page, &over_rsslimit);
+ if (referenced && page_mapping_inuse(page) && !over_rsslimit) {
/* In active use or really unfreeable. Activate it. */
pte_chain_unlock(page);
goto activate_locked;
@@ -597,6 +598,7 @@ refill_inactive_zone(struct zone *zone,
long mapped_ratio;
long distress;
long swap_tendency;
+ int over_rsslimit = 0;
lru_add_drain();
pgmoved = 0;
@@ -657,13 +659,15 @@ refill_inactive_zone(struct zone *zone,
list_del(&page->lru);
if (page_mapped(page)) {
pte_chain_lock(page);
- if (page_mapped(page) && page_referenced(page)) {
+ if (page_mapped(page) &&
+ page_referenced(page, &over_rsslimit) &&
+ !over_rsslimit) {
pte_chain_unlock(page);
list_add(&page->lru, &l_active);
continue;
}
pte_chain_unlock(page);
- if (!reclaim_mapped) {
+ if (!reclaim_mapped && !over_rsslimit) {
list_add(&page->lru, &l_active);
continue;
}
_