[Devel] [PATCH RHEL7 COMMIT] ub: introduce ub_total_pages
Konstantin Khorenko
khorenko at virtuozzo.com
Tue Jun 9 07:45:42 PDT 2015
The commit is pushed to "branch-rh7-3.10.0-123.1.2-ovz" and will appear at https://src.openvz.org/scm/ovz/vzkernel.git
after rh7-3.10.0-123.1.2.vz7.5.11
------>
commit 9dcb75795efc68f4726863d8ba9a46be7646de73
Author: Vladimir Davydov <vdavydov at parallels.com>
Date: Tue Jun 9 18:45:42 2015 +0400
ub: introduce ub_total_pages
Now we do not need that hackish mem_cgroup_total_pages, which returns
the maximal number of pages allowed by the limit set upon the top level
cgroup the current resides in. Now we can simply get the limit from the
memory cgroup that is bound to a beancounter using beancounter.memory.
This patch therefore reworks mem_cgroup_total_pages, allowing to pass a
cgroup of interest to it, and introduces ub_total_pages, which returns
the limit of a given beancounter.
Signed-off-by: Vladimir Davydov <vdavydov at parallels.com>
Reviewed-by: Cyrill Gorcunov <gorcunov at openvz.org>
Related to patchset:
https://jira.sw.ru/browse/PSBM-34035
Vladimir Davydov (6):
cgroup: add function to lookup a cgroup by pathname
cgroup: make cgroup_kernel_attach safe against cgroup removal
ub: rename functions for reading/writing cgroup resources
ub: rework connection between ub and mem/blkio cgroups
memcg: pass struct mem_cgroup instead of cgroup to ub-related methods
ub: introduce ub_total_pages
---
fs/proc/base.c | 5 +++--
include/bc/beancounter.h | 1 +
include/linux/memcontrol.h | 6 ------
kernel/bc/beancounter.c | 13 +++++++++++++
kernel/bc/io_acct.c | 2 +-
mm/memcontrol.c | 28 ++--------------------------
mm/shmem.c | 7 +++----
7 files changed, 23 insertions(+), 39 deletions(-)
diff --git a/fs/proc/base.c b/fs/proc/base.c
index 360169a..4a39156 100644
--- a/fs/proc/base.c
+++ b/fs/proc/base.c
@@ -449,9 +449,10 @@ static int proc_oom_score(struct task_struct *task, char *buffer)
unsigned long totalpages = totalram_pages + total_swap_pages;
unsigned long points = 0;
struct mem_cgroup *memcg = NULL;
+ struct user_beancounter *ub = get_exec_ub();
- if (!ve_is_super(get_exec_env())) {
- totalpages = min(totalpages, mem_cgroup_total_pages(true));
+ if (ub != get_ub0()) {
+ totalpages = min(totalpages, ub_total_pages(ub, true));
memcg = OOM_BADNESS_DUMMY_MEMCG;
}
diff --git a/include/bc/beancounter.h b/include/bc/beancounter.h
index 6d55052..340d097 100644
--- a/include/bc/beancounter.h
+++ b/include/bc/beancounter.h
@@ -317,6 +317,7 @@ extern void ub_sync_memcg(struct user_beancounter *ub);
extern void ub_page_stat(struct user_beancounter *ub,
const nodemask_t *nodemask,
unsigned long *pages);
+extern unsigned long ub_total_pages(struct user_beancounter *ub, bool swap);
extern const char *ub_rnames[];
/*
diff --git a/include/linux/memcontrol.h b/include/linux/memcontrol.h
index ae28833..5e0e7c1 100644
--- a/include/linux/memcontrol.h
+++ b/include/linux/memcontrol.h
@@ -128,7 +128,6 @@ extern void mem_cgroup_print_oom_info(struct mem_cgroup *memcg,
struct task_struct *p);
extern void mem_cgroup_replace_page_cache(struct page *oldpage,
struct page *newpage);
-unsigned long mem_cgroup_total_pages(bool swap);
#ifdef CONFIG_MEMCG_SWAP
extern int do_swap_account;
@@ -402,11 +401,6 @@ static inline void mem_cgroup_replace_page_cache(struct page *oldpage,
struct page *newpage)
{
}
-
-static inline unsigned long mem_cgroup_total_pages(swap)
-{
- return ULONG_MAX;
-}
#endif /* CONFIG_MEMCG */
#if !defined(CONFIG_MEMCG) || !defined(CONFIG_DEBUG_VM)
diff --git a/kernel/bc/beancounter.c b/kernel/bc/beancounter.c
index 7d20f73..81f5518 100644
--- a/kernel/bc/beancounter.c
+++ b/kernel/bc/beancounter.c
@@ -216,6 +216,8 @@ extern int mem_cgroup_apply_beancounter(struct mem_cgroup *memcg,
struct user_beancounter *ub);
extern void mem_cgroup_get_nr_pages(struct mem_cgroup *memcg, int nid,
unsigned long *pages);
+extern unsigned long mem_cgroup_total_pages(struct mem_cgroup *memcg,
+ bool swap);
/*
* Update memcg limits according to beancounter configuration.
@@ -259,6 +261,17 @@ void ub_page_stat(struct user_beancounter *ub, const nodemask_t *nodemask,
css_put(css);
}
+unsigned long ub_total_pages(struct user_beancounter *ub, bool swap)
+{
+ struct cgroup_subsys_state *css;
+ unsigned long ret;
+
+ css = ub_get_mem_css(ub);
+ ret = mem_cgroup_total_pages(mem_cgroup_from_cont(css->cgroup), swap);
+ css_put(css);
+ return ret;
+}
+
void init_beancounter_precharge(struct user_beancounter *ub, int resource)
{
/* limit maximum precharge with one half of current resource excess */
diff --git a/kernel/bc/io_acct.c b/kernel/bc/io_acct.c
index 50f8fb7..db3926d 100644
--- a/kernel/bc/io_acct.c
+++ b/kernel/bc/io_acct.c
@@ -135,7 +135,7 @@ int ub_dirty_limits(unsigned long *pbackground,
if (!dirty_ratio)
return 0;
- available_memory = mem_cgroup_total_pages(false);
+ available_memory = ub_total_pages(ub, false);
if (available_memory == ULONG_MAX || available_memory == 0)
return 0;
diff --git a/mm/memcontrol.c b/mm/memcontrol.c
index bda7d42..b883404 100644
--- a/mm/memcontrol.c
+++ b/mm/memcontrol.c
@@ -1655,36 +1655,12 @@ void mem_cgroup_note_oom_kill(struct mem_cgroup *root_memcg,
css_put(&memcg_to_put->css);
}
-/*
- * XXX: This function returns the limit of the topmost memory cgroup the
- * current process belongs to (ULONG_MAX if unlimited). It is used to get the
- * number of memory pages available to a Virtuozzo container. Here we
- * implicitly assume that each container lives in its own top level memory
- * cgroup. If it is changed, this function must be reworked. E.g. we could
- * assign a memory cgroup to each ve or beancounter cgroup and get the memory
- * cgroup of a container from get_exec_env() or get_exec_ub().
- *
- * If @swap is true, this function returns the total number of memory + swap
- * pages available.
- */
-unsigned long mem_cgroup_total_pages(bool swap)
+unsigned long mem_cgroup_total_pages(struct mem_cgroup *memcg, bool swap)
{
- struct mem_cgroup *memcg_to_put, *memcg, *parent;
- unsigned long long limit = RESOURCE_MAX;
-
- memcg_to_put = memcg = try_get_mem_cgroup_from_mm(current->mm);
- if (!memcg || memcg == root_mem_cgroup)
- goto out;
-
- while ((parent = parent_mem_cgroup(memcg)) &&
- parent != root_mem_cgroup)
- memcg = parent;
+ unsigned long long limit;
limit = swap ? res_counter_read_u64(&memcg->memsw, RES_LIMIT) :
res_counter_read_u64(&memcg->res, RES_LIMIT);
-out:
- if (memcg_to_put)
- css_put(&memcg_to_put->css);
if (limit == RESOURCE_MAX)
return ULONG_MAX;
return min_t(unsigned long long, ULONG_MAX, limit >> PAGE_SHIFT);
diff --git a/mm/shmem.c b/mm/shmem.c
index ca7f12a..0ae9991 100644
--- a/mm/shmem.c
+++ b/mm/shmem.c
@@ -106,13 +106,12 @@ enum sgp_type {
#ifdef CONFIG_TMPFS
static unsigned long tmpfs_ram_pages(void)
{
- unsigned long memcg_rampages;
+ struct user_beancounter *ub = get_exec_ub();
- if (ve_is_super(get_exec_env()))
+ if (ub == get_ub0())
return totalram_pages;
- memcg_rampages = mem_cgroup_total_pages(false);
- return min(totalram_pages, memcg_rampages);
+ return min(totalram_pages, ub_total_pages(ub, false));
}
static unsigned long shmem_default_max_blocks(void)
More information about the Devel
mailing list