2005-04-16 22:20:36 +00:00
|
|
|
/*
|
|
|
|
|
* linux/mm/oom_kill.c
|
|
|
|
|
*
|
|
|
|
|
* Copyright (C) 1998,2000 Rik van Riel
|
|
|
|
|
* Thanks go out to Claus Fischer for some serious inspiration and
|
|
|
|
|
* for goading me into coding this file...
|
|
|
|
|
*
|
|
|
|
|
* The routines in this file are used to kill a process when
|
[PATCH] cpusets: oom_kill tweaks
This patch series extends the use of the cpuset attribute 'mem_exclusive'
to support cpuset configurations that:
1) allow GFP_KERNEL allocations to come from a potentially larger
set of memory nodes than GFP_USER allocations, and
2) can constrain the oom killer to tasks running in cpusets in
a specified subtree of the cpuset hierarchy.
Here's an example usage scenario. For a few hours or more, a large NUMA
system at a University is to be divided in two halves, with a bunch of student
jobs running in half the system under some form of batch manager, and with a
big research project running in the other half. Each of the student jobs is
placed in a small cpuset, but should share the classic Unix time share
facilities, such as buffered pages of files in /bin and /usr/lib. The big
research project wants no interference whatsoever from the student jobs, and
has highly tuned, unusual memory and i/o patterns that intend to make full use
of all the main memory on the nodes available to it.
In this example, we have two big sibling cpusets, one of which is further
divided into a more dynamic set of child cpusets.
We want kernel memory allocations constrained by the two big cpusets, and user
allocations constrained by the smaller child cpusets where present. And we
require that the oom killer not operate across the two halves of this system,
or else the first time a student job runs amuck, the big research project will
likely be first inline to get shot.
Tweaking /proc/<pid>/oom_adj is not ideal -- if the big research project
really does run amuck allocating memory, it should be shot, not some other
task outside the research projects mem_exclusive cpuset.
I propose to extend the use of the 'mem_exclusive' flag of cpusets to manage
such scenarios. Let memory allocations for user space (GFP_USER) be
constrained by a tasks current cpuset, but memory allocations for kernel space
(GFP_KERNEL) by constrained by the nearest mem_exclusive ancestor of the
current cpuset, even though kernel space allocations will still _prefer_ to
remain within the current tasks cpuset, if memory is easily available.
Let the oom killer be constrained to consider only tasks that are in
overlapping mem_exclusive cpusets (it won't help much to kill a task that
normally cannot allocate memory on any of the same nodes as the ones on which
the current task can allocate.)
The current constraints imposed on setting mem_exclusive are unchanged. A
cpuset may only be mem_exclusive if its parent is also mem_exclusive, and a
mem_exclusive cpuset may not overlap any of its siblings memory nodes.
This patch was presented on linux-mm in early July 2005, though did not
generate much feedback at that time. It has been built for a variety of
arch's using cross tools, and built, booted and tested for function on SN2
(ia64).
There are 4 patches in this set:
1) Some minor cleanup, and some improvements to the code layout
of one routine to make subsequent patches cleaner.
2) Add another GFP flag - __GFP_HARDWALL. It marks memory
requests for USER space, which are tightly confined by the
current tasks cpuset.
3) Now memory requests (such as KERNEL) that not marked HARDWALL can
if short on memory, look in the potentially larger pool of memory
defined by the nearest mem_exclusive ancestor cpuset of the current
tasks cpuset.
4) Finally, modify the oom killer to skip any task whose mem_exclusive
cpuset doesn't overlap ours.
Patch (1), the one time I looked on an SN2 (ia64) build, actually saved 32
bytes of kernel text space. Patch (2) has no affect on the size of kernel
text space (it just adds a preprocessor flag). Patches (3) and (4) added
about 600 bytes each of kernel text space, mostly in kernel/cpuset.c, which
matters only if CONFIG_CPUSET is enabled.
This patch:
This patch applies a few comment and code cleanups to mm/oom_kill.c prior to
applying a few small patches to improve cpuset management of memory placement.
The comment changed in oom_kill.c was seriously misleading. The code layout
change in select_bad_process() makes room for adding another condition on
which a process can be spared the oom killer (see the subsequent
cpuset_nodes_overlap patch for this addition).
Also a couple typos and spellos that bugged me, while I was here.
This patch should have no material affect.
Signed-off-by: Paul Jackson <pj@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-09-06 22:18:09 +00:00
|
|
|
* we're seriously out of memory. This gets called from __alloc_pages()
|
|
|
|
|
* in mm/page_alloc.c when we really run out of memory.
|
2005-04-16 22:20:36 +00:00
|
|
|
*
|
|
|
|
|
* Since we won't call these routines often (on a well-configured
|
|
|
|
|
* machine) this file will double as a 'coding guide' and a signpost
|
|
|
|
|
* for newbie kernel hackers. It features several pointers to major
|
|
|
|
|
* kernel subsystems and hints as to where to find out what things do.
|
|
|
|
|
*/
|
|
|
|
|
|
|
|
|
|
#include <linux/mm.h>
|
|
|
|
|
#include <linux/sched.h>
|
|
|
|
|
#include <linux/swap.h>
|
|
|
|
|
#include <linux/timex.h>
|
|
|
|
|
#include <linux/jiffies.h>
|
2005-09-06 22:18:13 +00:00
|
|
|
#include <linux/cpuset.h>
|
2005-04-16 22:20:36 +00:00
|
|
|
|
|
|
|
|
/* #define DEBUG */
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* oom_badness - calculate a numeric value for how bad this task has been
|
|
|
|
|
* @p: task struct of which task we should calculate
|
[PATCH] cpusets: oom_kill tweaks
This patch series extends the use of the cpuset attribute 'mem_exclusive'
to support cpuset configurations that:
1) allow GFP_KERNEL allocations to come from a potentially larger
set of memory nodes than GFP_USER allocations, and
2) can constrain the oom killer to tasks running in cpusets in
a specified subtree of the cpuset hierarchy.
Here's an example usage scenario. For a few hours or more, a large NUMA
system at a University is to be divided in two halves, with a bunch of student
jobs running in half the system under some form of batch manager, and with a
big research project running in the other half. Each of the student jobs is
placed in a small cpuset, but should share the classic Unix time share
facilities, such as buffered pages of files in /bin and /usr/lib. The big
research project wants no interference whatsoever from the student jobs, and
has highly tuned, unusual memory and i/o patterns that intend to make full use
of all the main memory on the nodes available to it.
In this example, we have two big sibling cpusets, one of which is further
divided into a more dynamic set of child cpusets.
We want kernel memory allocations constrained by the two big cpusets, and user
allocations constrained by the smaller child cpusets where present. And we
require that the oom killer not operate across the two halves of this system,
or else the first time a student job runs amuck, the big research project will
likely be first inline to get shot.
Tweaking /proc/<pid>/oom_adj is not ideal -- if the big research project
really does run amuck allocating memory, it should be shot, not some other
task outside the research projects mem_exclusive cpuset.
I propose to extend the use of the 'mem_exclusive' flag of cpusets to manage
such scenarios. Let memory allocations for user space (GFP_USER) be
constrained by a tasks current cpuset, but memory allocations for kernel space
(GFP_KERNEL) by constrained by the nearest mem_exclusive ancestor of the
current cpuset, even though kernel space allocations will still _prefer_ to
remain within the current tasks cpuset, if memory is easily available.
Let the oom killer be constrained to consider only tasks that are in
overlapping mem_exclusive cpusets (it won't help much to kill a task that
normally cannot allocate memory on any of the same nodes as the ones on which
the current task can allocate.)
The current constraints imposed on setting mem_exclusive are unchanged. A
cpuset may only be mem_exclusive if its parent is also mem_exclusive, and a
mem_exclusive cpuset may not overlap any of its siblings memory nodes.
This patch was presented on linux-mm in early July 2005, though did not
generate much feedback at that time. It has been built for a variety of
arch's using cross tools, and built, booted and tested for function on SN2
(ia64).
There are 4 patches in this set:
1) Some minor cleanup, and some improvements to the code layout
of one routine to make subsequent patches cleaner.
2) Add another GFP flag - __GFP_HARDWALL. It marks memory
requests for USER space, which are tightly confined by the
current tasks cpuset.
3) Now memory requests (such as KERNEL) that not marked HARDWALL can
if short on memory, look in the potentially larger pool of memory
defined by the nearest mem_exclusive ancestor cpuset of the current
tasks cpuset.
4) Finally, modify the oom killer to skip any task whose mem_exclusive
cpuset doesn't overlap ours.
Patch (1), the one time I looked on an SN2 (ia64) build, actually saved 32
bytes of kernel text space. Patch (2) has no affect on the size of kernel
text space (it just adds a preprocessor flag). Patches (3) and (4) added
about 600 bytes each of kernel text space, mostly in kernel/cpuset.c, which
matters only if CONFIG_CPUSET is enabled.
This patch:
This patch applies a few comment and code cleanups to mm/oom_kill.c prior to
applying a few small patches to improve cpuset management of memory placement.
The comment changed in oom_kill.c was seriously misleading. The code layout
change in select_bad_process() makes room for adding another condition on
which a process can be spared the oom killer (see the subsequent
cpuset_nodes_overlap patch for this addition).
Also a couple typos and spellos that bugged me, while I was here.
This patch should have no material affect.
Signed-off-by: Paul Jackson <pj@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-09-06 22:18:09 +00:00
|
|
|
* @uptime: current uptime in seconds
|
2005-04-16 22:20:36 +00:00
|
|
|
*
|
|
|
|
|
* The formula used is relatively simple and documented inline in the
|
|
|
|
|
* function. The main rationale is that we want to select a good task
|
|
|
|
|
* to kill when we run out of memory.
|
|
|
|
|
*
|
|
|
|
|
* Good in this context means that:
|
|
|
|
|
* 1) we lose the minimum amount of work done
|
|
|
|
|
* 2) we recover a large amount of memory
|
|
|
|
|
* 3) we don't kill anything innocent of eating tons of memory
|
|
|
|
|
* 4) we want to kill the minimum amount of processes (one)
|
|
|
|
|
* 5) we try to kill the process the user expects us to kill, this
|
|
|
|
|
* algorithm has been meticulously tuned to meet the principle
|
|
|
|
|
* of least surprise ... (be careful when you change it)
|
|
|
|
|
*/
|
|
|
|
|
|
|
|
|
|
unsigned long badness(struct task_struct *p, unsigned long uptime)
|
|
|
|
|
{
|
|
|
|
|
unsigned long points, cpu_time, run_time, s;
|
2006-04-19 05:20:38 +00:00
|
|
|
struct mm_struct *mm;
|
|
|
|
|
struct task_struct *child;
|
2005-04-16 22:20:36 +00:00
|
|
|
|
2006-04-19 05:20:38 +00:00
|
|
|
task_lock(p);
|
|
|
|
|
mm = p->mm;
|
|
|
|
|
if (!mm) {
|
|
|
|
|
task_unlock(p);
|
2005-04-16 22:20:36 +00:00
|
|
|
return 0;
|
2006-04-19 05:20:38 +00:00
|
|
|
}
|
2005-04-16 22:20:36 +00:00
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* The memory size of the process is the basis for the badness.
|
|
|
|
|
*/
|
2006-04-19 05:20:38 +00:00
|
|
|
points = mm->total_vm;
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* After this unlock we can no longer dereference local variable `mm'
|
|
|
|
|
*/
|
|
|
|
|
task_unlock(p);
|
2005-04-16 22:20:36 +00:00
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* Processes which fork a lot of child processes are likely
|
[PATCH] OOM kill: children accounting
In the badness() calculation, there's currently this piece of code:
/*
* Processes which fork a lot of child processes are likely
* a good choice. We add the vmsize of the children if they
* have an own mm. This prevents forking servers to flood the
* machine with an endless amount of children
*/
list_for_each(tsk, &p->children) {
struct task_struct *chld;
chld = list_entry(tsk, struct task_struct, sibling);
if (chld->mm = p->mm && chld->mm)
points += chld->mm->total_vm;
}
The intention is clear: If some server (apache) keeps spawning new children
and we run OOM, we want to kill the father rather than picking a child.
This -- to some degree -- also helps a bit with getting fork bombs under
control, though I'd consider this a desirable side-effect rather than a
feature.
There's one problem with this: No matter how many or few children there are,
if just one of them misbehaves, and all others (including the father) do
everything right, we still always kill the whole family. This hits in real
life; whether it's javascript in konqueror resulting in kdeinit (and thus the
whole KDE session) being hit or just a classical server that spawns children.
Sidenote: The killer does kill all direct children as well, not only the
selected father, see oom_kill_process().
The idea in attached patch is that we do want to account the memory
consumption of the (direct) children to the father -- however not fully.
This maintains the property that fathers with too many children will still
very likely be picked, whereas a single misbehaving child has the chance to
be picked by the OOM killer.
In the patch I account only half (rounded up) of the children's vm_size to
the parent. This means that if one child eats more mem than the rest of
the family, it will be picked, otherwise it's still the father and thus the
whole family that gets selected.
This is heuristics -- we could debate whether accounting for a fourth would
be better than for half of it. Or -- if people would consider it worth the
trouble -- make it a sysctl. For now I sticked to accounting for half,
which should IMHO be a significant improvement.
The patch does one more thing: As users tend to be irritated by the choice
of killed processes (mainly because the children are killed first, despite
some of them having a very low OOM score), I added some more output: The
selected (father) process will be reported first and it's oom_score printed
to syslog.
Description:
Only account for half of children's vm size in oom score calculation
This should still give the parent enough point in case of fork bombs. If
any child however has more than 50% of the vm size of all children
together, it'll get a higher score and be elected.
This patch also makes the kernel display the oom_score.
Signed-off-by: Kurt Garloff <garloff@suse.de>
Cc: Rik van Riel <riel@redhat.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-02-21 02:27:51 +00:00
|
|
|
* a good choice. We add half the vmsize of the children if they
|
2005-04-16 22:20:36 +00:00
|
|
|
* have an own mm. This prevents forking servers to flood the
|
[PATCH] OOM kill: children accounting
In the badness() calculation, there's currently this piece of code:
/*
* Processes which fork a lot of child processes are likely
* a good choice. We add the vmsize of the children if they
* have an own mm. This prevents forking servers to flood the
* machine with an endless amount of children
*/
list_for_each(tsk, &p->children) {
struct task_struct *chld;
chld = list_entry(tsk, struct task_struct, sibling);
if (chld->mm = p->mm && chld->mm)
points += chld->mm->total_vm;
}
The intention is clear: If some server (apache) keeps spawning new children
and we run OOM, we want to kill the father rather than picking a child.
This -- to some degree -- also helps a bit with getting fork bombs under
control, though I'd consider this a desirable side-effect rather than a
feature.
There's one problem with this: No matter how many or few children there are,
if just one of them misbehaves, and all others (including the father) do
everything right, we still always kill the whole family. This hits in real
life; whether it's javascript in konqueror resulting in kdeinit (and thus the
whole KDE session) being hit or just a classical server that spawns children.
Sidenote: The killer does kill all direct children as well, not only the
selected father, see oom_kill_process().
The idea in attached patch is that we do want to account the memory
consumption of the (direct) children to the father -- however not fully.
This maintains the property that fathers with too many children will still
very likely be picked, whereas a single misbehaving child has the chance to
be picked by the OOM killer.
In the patch I account only half (rounded up) of the children's vm_size to
the parent. This means that if one child eats more mem than the rest of
the family, it will be picked, otherwise it's still the father and thus the
whole family that gets selected.
This is heuristics -- we could debate whether accounting for a fourth would
be better than for half of it. Or -- if people would consider it worth the
trouble -- make it a sysctl. For now I sticked to accounting for half,
which should IMHO be a significant improvement.
The patch does one more thing: As users tend to be irritated by the choice
of killed processes (mainly because the children are killed first, despite
some of them having a very low OOM score), I added some more output: The
selected (father) process will be reported first and it's oom_score printed
to syslog.
Description:
Only account for half of children's vm size in oom score calculation
This should still give the parent enough point in case of fork bombs. If
any child however has more than 50% of the vm size of all children
together, it'll get a higher score and be elected.
This patch also makes the kernel display the oom_score.
Signed-off-by: Kurt Garloff <garloff@suse.de>
Cc: Rik van Riel <riel@redhat.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-02-21 02:27:51 +00:00
|
|
|
* machine with an endless amount of children. In case a single
|
|
|
|
|
* child is eating the vast majority of memory, adding only half
|
|
|
|
|
* to the parents will make the child our kill candidate of choice.
|
2005-04-16 22:20:36 +00:00
|
|
|
*/
|
2006-04-19 05:20:38 +00:00
|
|
|
list_for_each_entry(child, &p->children, sibling) {
|
|
|
|
|
task_lock(child);
|
|
|
|
|
if (child->mm != mm && child->mm)
|
|
|
|
|
points += child->mm->total_vm/2 + 1;
|
|
|
|
|
task_unlock(child);
|
2005-04-16 22:20:36 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* CPU time is in tens of seconds and run time is in thousands
|
|
|
|
|
* of seconds. There is no particular reason for this other than
|
|
|
|
|
* that it turned out to work very well in practice.
|
|
|
|
|
*/
|
|
|
|
|
cpu_time = (cputime_to_jiffies(p->utime) + cputime_to_jiffies(p->stime))
|
|
|
|
|
>> (SHIFT_HZ + 3);
|
|
|
|
|
|
|
|
|
|
if (uptime >= p->start_time.tv_sec)
|
|
|
|
|
run_time = (uptime - p->start_time.tv_sec) >> 10;
|
|
|
|
|
else
|
|
|
|
|
run_time = 0;
|
|
|
|
|
|
|
|
|
|
s = int_sqrt(cpu_time);
|
|
|
|
|
if (s)
|
|
|
|
|
points /= s;
|
|
|
|
|
s = int_sqrt(int_sqrt(run_time));
|
|
|
|
|
if (s)
|
|
|
|
|
points /= s;
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* Niced processes are most likely less important, so double
|
|
|
|
|
* their badness points.
|
|
|
|
|
*/
|
|
|
|
|
if (task_nice(p) > 0)
|
|
|
|
|
points *= 2;
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* Superuser processes are usually more important, so we make it
|
|
|
|
|
* less likely that we kill those.
|
|
|
|
|
*/
|
|
|
|
|
if (cap_t(p->cap_effective) & CAP_TO_MASK(CAP_SYS_ADMIN) ||
|
|
|
|
|
p->uid == 0 || p->euid == 0)
|
|
|
|
|
points /= 4;
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* We don't want to kill a process with direct hardware access.
|
|
|
|
|
* Not only could that mess up the hardware, but usually users
|
|
|
|
|
* tend to only have this flag set on applications they think
|
|
|
|
|
* of as important.
|
|
|
|
|
*/
|
|
|
|
|
if (cap_t(p->cap_effective) & CAP_TO_MASK(CAP_SYS_RAWIO))
|
|
|
|
|
points /= 4;
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* Adjust the score by oomkilladj.
|
|
|
|
|
*/
|
|
|
|
|
if (p->oomkilladj) {
|
|
|
|
|
if (p->oomkilladj > 0)
|
|
|
|
|
points <<= p->oomkilladj;
|
|
|
|
|
else
|
|
|
|
|
points >>= -(p->oomkilladj);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
#ifdef DEBUG
|
|
|
|
|
printk(KERN_DEBUG "OOMkill: task %d (%s) got %d points\n",
|
|
|
|
|
p->pid, p->comm, points);
|
|
|
|
|
#endif
|
|
|
|
|
return points;
|
|
|
|
|
}
|
|
|
|
|
|
2006-02-21 02:27:52 +00:00
|
|
|
/*
|
|
|
|
|
* Types of limitations to the nodes from which allocations may occur
|
|
|
|
|
*/
|
|
|
|
|
#define CONSTRAINT_NONE 1
|
|
|
|
|
#define CONSTRAINT_MEMORY_POLICY 2
|
|
|
|
|
#define CONSTRAINT_CPUSET 3
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* Determine the type of allocation constraint.
|
|
|
|
|
*/
|
|
|
|
|
static inline int constrained_alloc(struct zonelist *zonelist, gfp_t gfp_mask)
|
|
|
|
|
{
|
|
|
|
|
#ifdef CONFIG_NUMA
|
|
|
|
|
struct zone **z;
|
|
|
|
|
nodemask_t nodes = node_online_map;
|
|
|
|
|
|
|
|
|
|
for (z = zonelist->zones; *z; z++)
|
|
|
|
|
if (cpuset_zone_allowed(*z, gfp_mask))
|
|
|
|
|
node_clear((*z)->zone_pgdat->node_id,
|
|
|
|
|
nodes);
|
|
|
|
|
else
|
|
|
|
|
return CONSTRAINT_CPUSET;
|
|
|
|
|
|
|
|
|
|
if (!nodes_empty(nodes))
|
|
|
|
|
return CONSTRAINT_MEMORY_POLICY;
|
|
|
|
|
#endif
|
|
|
|
|
|
|
|
|
|
return CONSTRAINT_NONE;
|
|
|
|
|
}
|
|
|
|
|
|
2005-04-16 22:20:36 +00:00
|
|
|
/*
|
|
|
|
|
* Simple selection loop. We chose the process with the highest
|
|
|
|
|
* number of 'points'. We expect the caller will lock the tasklist.
|
|
|
|
|
*
|
|
|
|
|
* (not docbooked, we don't want this one cluttering up the manual)
|
|
|
|
|
*/
|
[PATCH] OOM kill: children accounting
In the badness() calculation, there's currently this piece of code:
/*
* Processes which fork a lot of child processes are likely
* a good choice. We add the vmsize of the children if they
* have an own mm. This prevents forking servers to flood the
* machine with an endless amount of children
*/
list_for_each(tsk, &p->children) {
struct task_struct *chld;
chld = list_entry(tsk, struct task_struct, sibling);
if (chld->mm = p->mm && chld->mm)
points += chld->mm->total_vm;
}
The intention is clear: If some server (apache) keeps spawning new children
and we run OOM, we want to kill the father rather than picking a child.
This -- to some degree -- also helps a bit with getting fork bombs under
control, though I'd consider this a desirable side-effect rather than a
feature.
There's one problem with this: No matter how many or few children there are,
if just one of them misbehaves, and all others (including the father) do
everything right, we still always kill the whole family. This hits in real
life; whether it's javascript in konqueror resulting in kdeinit (and thus the
whole KDE session) being hit or just a classical server that spawns children.
Sidenote: The killer does kill all direct children as well, not only the
selected father, see oom_kill_process().
The idea in attached patch is that we do want to account the memory
consumption of the (direct) children to the father -- however not fully.
This maintains the property that fathers with too many children will still
very likely be picked, whereas a single misbehaving child has the chance to
be picked by the OOM killer.
In the patch I account only half (rounded up) of the children's vm_size to
the parent. This means that if one child eats more mem than the rest of
the family, it will be picked, otherwise it's still the father and thus the
whole family that gets selected.
This is heuristics -- we could debate whether accounting for a fourth would
be better than for half of it. Or -- if people would consider it worth the
trouble -- make it a sysctl. For now I sticked to accounting for half,
which should IMHO be a significant improvement.
The patch does one more thing: As users tend to be irritated by the choice
of killed processes (mainly because the children are killed first, despite
some of them having a very low OOM score), I added some more output: The
selected (father) process will be reported first and it's oom_score printed
to syslog.
Description:
Only account for half of children's vm size in oom score calculation
This should still give the parent enough point in case of fork bombs. If
any child however has more than 50% of the vm size of all children
together, it'll get a higher score and be elected.
This patch also makes the kernel display the oom_score.
Signed-off-by: Kurt Garloff <garloff@suse.de>
Cc: Rik van Riel <riel@redhat.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-02-21 02:27:51 +00:00
|
|
|
static struct task_struct *select_bad_process(unsigned long *ppoints)
|
2005-04-16 22:20:36 +00:00
|
|
|
{
|
|
|
|
|
struct task_struct *g, *p;
|
|
|
|
|
struct task_struct *chosen = NULL;
|
|
|
|
|
struct timespec uptime;
|
[PATCH] OOM kill: children accounting
In the badness() calculation, there's currently this piece of code:
/*
* Processes which fork a lot of child processes are likely
* a good choice. We add the vmsize of the children if they
* have an own mm. This prevents forking servers to flood the
* machine with an endless amount of children
*/
list_for_each(tsk, &p->children) {
struct task_struct *chld;
chld = list_entry(tsk, struct task_struct, sibling);
if (chld->mm = p->mm && chld->mm)
points += chld->mm->total_vm;
}
The intention is clear: If some server (apache) keeps spawning new children
and we run OOM, we want to kill the father rather than picking a child.
This -- to some degree -- also helps a bit with getting fork bombs under
control, though I'd consider this a desirable side-effect rather than a
feature.
There's one problem with this: No matter how many or few children there are,
if just one of them misbehaves, and all others (including the father) do
everything right, we still always kill the whole family. This hits in real
life; whether it's javascript in konqueror resulting in kdeinit (and thus the
whole KDE session) being hit or just a classical server that spawns children.
Sidenote: The killer does kill all direct children as well, not only the
selected father, see oom_kill_process().
The idea in attached patch is that we do want to account the memory
consumption of the (direct) children to the father -- however not fully.
This maintains the property that fathers with too many children will still
very likely be picked, whereas a single misbehaving child has the chance to
be picked by the OOM killer.
In the patch I account only half (rounded up) of the children's vm_size to
the parent. This means that if one child eats more mem than the rest of
the family, it will be picked, otherwise it's still the father and thus the
whole family that gets selected.
This is heuristics -- we could debate whether accounting for a fourth would
be better than for half of it. Or -- if people would consider it worth the
trouble -- make it a sysctl. For now I sticked to accounting for half,
which should IMHO be a significant improvement.
The patch does one more thing: As users tend to be irritated by the choice
of killed processes (mainly because the children are killed first, despite
some of them having a very low OOM score), I added some more output: The
selected (father) process will be reported first and it's oom_score printed
to syslog.
Description:
Only account for half of children's vm size in oom score calculation
This should still give the parent enough point in case of fork bombs. If
any child however has more than 50% of the vm size of all children
together, it'll get a higher score and be elected.
This patch also makes the kernel display the oom_score.
Signed-off-by: Kurt Garloff <garloff@suse.de>
Cc: Rik van Riel <riel@redhat.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-02-21 02:27:51 +00:00
|
|
|
*ppoints = 0;
|
2005-04-16 22:20:36 +00:00
|
|
|
|
|
|
|
|
do_posix_clock_monotonic_gettime(&uptime);
|
[PATCH] cpusets: oom_kill tweaks
This patch series extends the use of the cpuset attribute 'mem_exclusive'
to support cpuset configurations that:
1) allow GFP_KERNEL allocations to come from a potentially larger
set of memory nodes than GFP_USER allocations, and
2) can constrain the oom killer to tasks running in cpusets in
a specified subtree of the cpuset hierarchy.
Here's an example usage scenario. For a few hours or more, a large NUMA
system at a University is to be divided in two halves, with a bunch of student
jobs running in half the system under some form of batch manager, and with a
big research project running in the other half. Each of the student jobs is
placed in a small cpuset, but should share the classic Unix time share
facilities, such as buffered pages of files in /bin and /usr/lib. The big
research project wants no interference whatsoever from the student jobs, and
has highly tuned, unusual memory and i/o patterns that intend to make full use
of all the main memory on the nodes available to it.
In this example, we have two big sibling cpusets, one of which is further
divided into a more dynamic set of child cpusets.
We want kernel memory allocations constrained by the two big cpusets, and user
allocations constrained by the smaller child cpusets where present. And we
require that the oom killer not operate across the two halves of this system,
or else the first time a student job runs amuck, the big research project will
likely be first inline to get shot.
Tweaking /proc/<pid>/oom_adj is not ideal -- if the big research project
really does run amuck allocating memory, it should be shot, not some other
task outside the research projects mem_exclusive cpuset.
I propose to extend the use of the 'mem_exclusive' flag of cpusets to manage
such scenarios. Let memory allocations for user space (GFP_USER) be
constrained by a tasks current cpuset, but memory allocations for kernel space
(GFP_KERNEL) by constrained by the nearest mem_exclusive ancestor of the
current cpuset, even though kernel space allocations will still _prefer_ to
remain within the current tasks cpuset, if memory is easily available.
Let the oom killer be constrained to consider only tasks that are in
overlapping mem_exclusive cpusets (it won't help much to kill a task that
normally cannot allocate memory on any of the same nodes as the ones on which
the current task can allocate.)
The current constraints imposed on setting mem_exclusive are unchanged. A
cpuset may only be mem_exclusive if its parent is also mem_exclusive, and a
mem_exclusive cpuset may not overlap any of its siblings memory nodes.
This patch was presented on linux-mm in early July 2005, though did not
generate much feedback at that time. It has been built for a variety of
arch's using cross tools, and built, booted and tested for function on SN2
(ia64).
There are 4 patches in this set:
1) Some minor cleanup, and some improvements to the code layout
of one routine to make subsequent patches cleaner.
2) Add another GFP flag - __GFP_HARDWALL. It marks memory
requests for USER space, which are tightly confined by the
current tasks cpuset.
3) Now memory requests (such as KERNEL) that not marked HARDWALL can
if short on memory, look in the potentially larger pool of memory
defined by the nearest mem_exclusive ancestor cpuset of the current
tasks cpuset.
4) Finally, modify the oom killer to skip any task whose mem_exclusive
cpuset doesn't overlap ours.
Patch (1), the one time I looked on an SN2 (ia64) build, actually saved 32
bytes of kernel text space. Patch (2) has no affect on the size of kernel
text space (it just adds a preprocessor flag). Patches (3) and (4) added
about 600 bytes each of kernel text space, mostly in kernel/cpuset.c, which
matters only if CONFIG_CPUSET is enabled.
This patch:
This patch applies a few comment and code cleanups to mm/oom_kill.c prior to
applying a few small patches to improve cpuset management of memory placement.
The comment changed in oom_kill.c was seriously misleading. The code layout
change in select_bad_process() makes room for adding another condition on
which a process can be spared the oom killer (see the subsequent
cpuset_nodes_overlap patch for this addition).
Also a couple typos and spellos that bugged me, while I was here.
This patch should have no material affect.
Signed-off-by: Paul Jackson <pj@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-09-06 22:18:09 +00:00
|
|
|
do_each_thread(g, p) {
|
|
|
|
|
unsigned long points;
|
|
|
|
|
int releasing;
|
|
|
|
|
|
2005-04-16 22:20:36 +00:00
|
|
|
/* skip the init task with pid == 1 */
|
[PATCH] cpusets: oom_kill tweaks
This patch series extends the use of the cpuset attribute 'mem_exclusive'
to support cpuset configurations that:
1) allow GFP_KERNEL allocations to come from a potentially larger
set of memory nodes than GFP_USER allocations, and
2) can constrain the oom killer to tasks running in cpusets in
a specified subtree of the cpuset hierarchy.
Here's an example usage scenario. For a few hours or more, a large NUMA
system at a University is to be divided in two halves, with a bunch of student
jobs running in half the system under some form of batch manager, and with a
big research project running in the other half. Each of the student jobs is
placed in a small cpuset, but should share the classic Unix time share
facilities, such as buffered pages of files in /bin and /usr/lib. The big
research project wants no interference whatsoever from the student jobs, and
has highly tuned, unusual memory and i/o patterns that intend to make full use
of all the main memory on the nodes available to it.
In this example, we have two big sibling cpusets, one of which is further
divided into a more dynamic set of child cpusets.
We want kernel memory allocations constrained by the two big cpusets, and user
allocations constrained by the smaller child cpusets where present. And we
require that the oom killer not operate across the two halves of this system,
or else the first time a student job runs amuck, the big research project will
likely be first inline to get shot.
Tweaking /proc/<pid>/oom_adj is not ideal -- if the big research project
really does run amuck allocating memory, it should be shot, not some other
task outside the research projects mem_exclusive cpuset.
I propose to extend the use of the 'mem_exclusive' flag of cpusets to manage
such scenarios. Let memory allocations for user space (GFP_USER) be
constrained by a tasks current cpuset, but memory allocations for kernel space
(GFP_KERNEL) by constrained by the nearest mem_exclusive ancestor of the
current cpuset, even though kernel space allocations will still _prefer_ to
remain within the current tasks cpuset, if memory is easily available.
Let the oom killer be constrained to consider only tasks that are in
overlapping mem_exclusive cpusets (it won't help much to kill a task that
normally cannot allocate memory on any of the same nodes as the ones on which
the current task can allocate.)
The current constraints imposed on setting mem_exclusive are unchanged. A
cpuset may only be mem_exclusive if its parent is also mem_exclusive, and a
mem_exclusive cpuset may not overlap any of its siblings memory nodes.
This patch was presented on linux-mm in early July 2005, though did not
generate much feedback at that time. It has been built for a variety of
arch's using cross tools, and built, booted and tested for function on SN2
(ia64).
There are 4 patches in this set:
1) Some minor cleanup, and some improvements to the code layout
of one routine to make subsequent patches cleaner.
2) Add another GFP flag - __GFP_HARDWALL. It marks memory
requests for USER space, which are tightly confined by the
current tasks cpuset.
3) Now memory requests (such as KERNEL) that not marked HARDWALL can
if short on memory, look in the potentially larger pool of memory
defined by the nearest mem_exclusive ancestor cpuset of the current
tasks cpuset.
4) Finally, modify the oom killer to skip any task whose mem_exclusive
cpuset doesn't overlap ours.
Patch (1), the one time I looked on an SN2 (ia64) build, actually saved 32
bytes of kernel text space. Patch (2) has no affect on the size of kernel
text space (it just adds a preprocessor flag). Patches (3) and (4) added
about 600 bytes each of kernel text space, mostly in kernel/cpuset.c, which
matters only if CONFIG_CPUSET is enabled.
This patch:
This patch applies a few comment and code cleanups to mm/oom_kill.c prior to
applying a few small patches to improve cpuset management of memory placement.
The comment changed in oom_kill.c was seriously misleading. The code layout
change in select_bad_process() makes room for adding another condition on
which a process can be spared the oom killer (see the subsequent
cpuset_nodes_overlap patch for this addition).
Also a couple typos and spellos that bugged me, while I was here.
This patch should have no material affect.
Signed-off-by: Paul Jackson <pj@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-09-06 22:18:09 +00:00
|
|
|
if (p->pid == 1)
|
|
|
|
|
continue;
|
|
|
|
|
if (p->oomkilladj == OOM_DISABLE)
|
|
|
|
|
continue;
|
2005-09-06 22:18:13 +00:00
|
|
|
/* If p's nodes don't overlap ours, it won't help to kill p. */
|
|
|
|
|
if (!cpuset_excl_nodes_overlap(p))
|
|
|
|
|
continue;
|
|
|
|
|
|
[PATCH] cpusets: oom_kill tweaks
This patch series extends the use of the cpuset attribute 'mem_exclusive'
to support cpuset configurations that:
1) allow GFP_KERNEL allocations to come from a potentially larger
set of memory nodes than GFP_USER allocations, and
2) can constrain the oom killer to tasks running in cpusets in
a specified subtree of the cpuset hierarchy.
Here's an example usage scenario. For a few hours or more, a large NUMA
system at a University is to be divided in two halves, with a bunch of student
jobs running in half the system under some form of batch manager, and with a
big research project running in the other half. Each of the student jobs is
placed in a small cpuset, but should share the classic Unix time share
facilities, such as buffered pages of files in /bin and /usr/lib. The big
research project wants no interference whatsoever from the student jobs, and
has highly tuned, unusual memory and i/o patterns that intend to make full use
of all the main memory on the nodes available to it.
In this example, we have two big sibling cpusets, one of which is further
divided into a more dynamic set of child cpusets.
We want kernel memory allocations constrained by the two big cpusets, and user
allocations constrained by the smaller child cpusets where present. And we
require that the oom killer not operate across the two halves of this system,
or else the first time a student job runs amuck, the big research project will
likely be first inline to get shot.
Tweaking /proc/<pid>/oom_adj is not ideal -- if the big research project
really does run amuck allocating memory, it should be shot, not some other
task outside the research projects mem_exclusive cpuset.
I propose to extend the use of the 'mem_exclusive' flag of cpusets to manage
such scenarios. Let memory allocations for user space (GFP_USER) be
constrained by a tasks current cpuset, but memory allocations for kernel space
(GFP_KERNEL) by constrained by the nearest mem_exclusive ancestor of the
current cpuset, even though kernel space allocations will still _prefer_ to
remain within the current tasks cpuset, if memory is easily available.
Let the oom killer be constrained to consider only tasks that are in
overlapping mem_exclusive cpusets (it won't help much to kill a task that
normally cannot allocate memory on any of the same nodes as the ones on which
the current task can allocate.)
The current constraints imposed on setting mem_exclusive are unchanged. A
cpuset may only be mem_exclusive if its parent is also mem_exclusive, and a
mem_exclusive cpuset may not overlap any of its siblings memory nodes.
This patch was presented on linux-mm in early July 2005, though did not
generate much feedback at that time. It has been built for a variety of
arch's using cross tools, and built, booted and tested for function on SN2
(ia64).
There are 4 patches in this set:
1) Some minor cleanup, and some improvements to the code layout
of one routine to make subsequent patches cleaner.
2) Add another GFP flag - __GFP_HARDWALL. It marks memory
requests for USER space, which are tightly confined by the
current tasks cpuset.
3) Now memory requests (such as KERNEL) that not marked HARDWALL can
if short on memory, look in the potentially larger pool of memory
defined by the nearest mem_exclusive ancestor cpuset of the current
tasks cpuset.
4) Finally, modify the oom killer to skip any task whose mem_exclusive
cpuset doesn't overlap ours.
Patch (1), the one time I looked on an SN2 (ia64) build, actually saved 32
bytes of kernel text space. Patch (2) has no affect on the size of kernel
text space (it just adds a preprocessor flag). Patches (3) and (4) added
about 600 bytes each of kernel text space, mostly in kernel/cpuset.c, which
matters only if CONFIG_CPUSET is enabled.
This patch:
This patch applies a few comment and code cleanups to mm/oom_kill.c prior to
applying a few small patches to improve cpuset management of memory placement.
The comment changed in oom_kill.c was seriously misleading. The code layout
change in select_bad_process() makes room for adding another condition on
which a process can be spared the oom killer (see the subsequent
cpuset_nodes_overlap patch for this addition).
Also a couple typos and spellos that bugged me, while I was here.
This patch should have no material affect.
Signed-off-by: Paul Jackson <pj@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-09-06 22:18:09 +00:00
|
|
|
/*
|
|
|
|
|
* This is in the process of releasing memory so for wait it
|
|
|
|
|
* to finish before killing some other task by mistake.
|
|
|
|
|
*/
|
|
|
|
|
releasing = test_tsk_thread_flag(p, TIF_MEMDIE) ||
|
|
|
|
|
p->flags & PF_EXITING;
|
|
|
|
|
if (releasing && !(p->flags & PF_DEAD))
|
|
|
|
|
return ERR_PTR(-1UL);
|
|
|
|
|
if (p->flags & PF_SWAPOFF)
|
|
|
|
|
return p;
|
|
|
|
|
|
|
|
|
|
points = badness(p, uptime.tv_sec);
|
[PATCH] OOM kill: children accounting
In the badness() calculation, there's currently this piece of code:
/*
* Processes which fork a lot of child processes are likely
* a good choice. We add the vmsize of the children if they
* have an own mm. This prevents forking servers to flood the
* machine with an endless amount of children
*/
list_for_each(tsk, &p->children) {
struct task_struct *chld;
chld = list_entry(tsk, struct task_struct, sibling);
if (chld->mm = p->mm && chld->mm)
points += chld->mm->total_vm;
}
The intention is clear: If some server (apache) keeps spawning new children
and we run OOM, we want to kill the father rather than picking a child.
This -- to some degree -- also helps a bit with getting fork bombs under
control, though I'd consider this a desirable side-effect rather than a
feature.
There's one problem with this: No matter how many or few children there are,
if just one of them misbehaves, and all others (including the father) do
everything right, we still always kill the whole family. This hits in real
life; whether it's javascript in konqueror resulting in kdeinit (and thus the
whole KDE session) being hit or just a classical server that spawns children.
Sidenote: The killer does kill all direct children as well, not only the
selected father, see oom_kill_process().
The idea in attached patch is that we do want to account the memory
consumption of the (direct) children to the father -- however not fully.
This maintains the property that fathers with too many children will still
very likely be picked, whereas a single misbehaving child has the chance to
be picked by the OOM killer.
In the patch I account only half (rounded up) of the children's vm_size to
the parent. This means that if one child eats more mem than the rest of
the family, it will be picked, otherwise it's still the father and thus the
whole family that gets selected.
This is heuristics -- we could debate whether accounting for a fourth would
be better than for half of it. Or -- if people would consider it worth the
trouble -- make it a sysctl. For now I sticked to accounting for half,
which should IMHO be a significant improvement.
The patch does one more thing: As users tend to be irritated by the choice
of killed processes (mainly because the children are killed first, despite
some of them having a very low OOM score), I added some more output: The
selected (father) process will be reported first and it's oom_score printed
to syslog.
Description:
Only account for half of children's vm size in oom score calculation
This should still give the parent enough point in case of fork bombs. If
any child however has more than 50% of the vm size of all children
together, it'll get a higher score and be elected.
This patch also makes the kernel display the oom_score.
Signed-off-by: Kurt Garloff <garloff@suse.de>
Cc: Rik van Riel <riel@redhat.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-02-21 02:27:51 +00:00
|
|
|
if (points > *ppoints || !chosen) {
|
[PATCH] cpusets: oom_kill tweaks
This patch series extends the use of the cpuset attribute 'mem_exclusive'
to support cpuset configurations that:
1) allow GFP_KERNEL allocations to come from a potentially larger
set of memory nodes than GFP_USER allocations, and
2) can constrain the oom killer to tasks running in cpusets in
a specified subtree of the cpuset hierarchy.
Here's an example usage scenario. For a few hours or more, a large NUMA
system at a University is to be divided in two halves, with a bunch of student
jobs running in half the system under some form of batch manager, and with a
big research project running in the other half. Each of the student jobs is
placed in a small cpuset, but should share the classic Unix time share
facilities, such as buffered pages of files in /bin and /usr/lib. The big
research project wants no interference whatsoever from the student jobs, and
has highly tuned, unusual memory and i/o patterns that intend to make full use
of all the main memory on the nodes available to it.
In this example, we have two big sibling cpusets, one of which is further
divided into a more dynamic set of child cpusets.
We want kernel memory allocations constrained by the two big cpusets, and user
allocations constrained by the smaller child cpusets where present. And we
require that the oom killer not operate across the two halves of this system,
or else the first time a student job runs amuck, the big research project will
likely be first inline to get shot.
Tweaking /proc/<pid>/oom_adj is not ideal -- if the big research project
really does run amuck allocating memory, it should be shot, not some other
task outside the research projects mem_exclusive cpuset.
I propose to extend the use of the 'mem_exclusive' flag of cpusets to manage
such scenarios. Let memory allocations for user space (GFP_USER) be
constrained by a tasks current cpuset, but memory allocations for kernel space
(GFP_KERNEL) by constrained by the nearest mem_exclusive ancestor of the
current cpuset, even though kernel space allocations will still _prefer_ to
remain within the current tasks cpuset, if memory is easily available.
Let the oom killer be constrained to consider only tasks that are in
overlapping mem_exclusive cpusets (it won't help much to kill a task that
normally cannot allocate memory on any of the same nodes as the ones on which
the current task can allocate.)
The current constraints imposed on setting mem_exclusive are unchanged. A
cpuset may only be mem_exclusive if its parent is also mem_exclusive, and a
mem_exclusive cpuset may not overlap any of its siblings memory nodes.
This patch was presented on linux-mm in early July 2005, though did not
generate much feedback at that time. It has been built for a variety of
arch's using cross tools, and built, booted and tested for function on SN2
(ia64).
There are 4 patches in this set:
1) Some minor cleanup, and some improvements to the code layout
of one routine to make subsequent patches cleaner.
2) Add another GFP flag - __GFP_HARDWALL. It marks memory
requests for USER space, which are tightly confined by the
current tasks cpuset.
3) Now memory requests (such as KERNEL) that not marked HARDWALL can
if short on memory, look in the potentially larger pool of memory
defined by the nearest mem_exclusive ancestor cpuset of the current
tasks cpuset.
4) Finally, modify the oom killer to skip any task whose mem_exclusive
cpuset doesn't overlap ours.
Patch (1), the one time I looked on an SN2 (ia64) build, actually saved 32
bytes of kernel text space. Patch (2) has no affect on the size of kernel
text space (it just adds a preprocessor flag). Patches (3) and (4) added
about 600 bytes each of kernel text space, mostly in kernel/cpuset.c, which
matters only if CONFIG_CPUSET is enabled.
This patch:
This patch applies a few comment and code cleanups to mm/oom_kill.c prior to
applying a few small patches to improve cpuset management of memory placement.
The comment changed in oom_kill.c was seriously misleading. The code layout
change in select_bad_process() makes room for adding another condition on
which a process can be spared the oom killer (see the subsequent
cpuset_nodes_overlap patch for this addition).
Also a couple typos and spellos that bugged me, while I was here.
This patch should have no material affect.
Signed-off-by: Paul Jackson <pj@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-09-06 22:18:09 +00:00
|
|
|
chosen = p;
|
[PATCH] OOM kill: children accounting
In the badness() calculation, there's currently this piece of code:
/*
* Processes which fork a lot of child processes are likely
* a good choice. We add the vmsize of the children if they
* have an own mm. This prevents forking servers to flood the
* machine with an endless amount of children
*/
list_for_each(tsk, &p->children) {
struct task_struct *chld;
chld = list_entry(tsk, struct task_struct, sibling);
if (chld->mm = p->mm && chld->mm)
points += chld->mm->total_vm;
}
The intention is clear: If some server (apache) keeps spawning new children
and we run OOM, we want to kill the father rather than picking a child.
This -- to some degree -- also helps a bit with getting fork bombs under
control, though I'd consider this a desirable side-effect rather than a
feature.
There's one problem with this: No matter how many or few children there are,
if just one of them misbehaves, and all others (including the father) do
everything right, we still always kill the whole family. This hits in real
life; whether it's javascript in konqueror resulting in kdeinit (and thus the
whole KDE session) being hit or just a classical server that spawns children.
Sidenote: The killer does kill all direct children as well, not only the
selected father, see oom_kill_process().
The idea in attached patch is that we do want to account the memory
consumption of the (direct) children to the father -- however not fully.
This maintains the property that fathers with too many children will still
very likely be picked, whereas a single misbehaving child has the chance to
be picked by the OOM killer.
In the patch I account only half (rounded up) of the children's vm_size to
the parent. This means that if one child eats more mem than the rest of
the family, it will be picked, otherwise it's still the father and thus the
whole family that gets selected.
This is heuristics -- we could debate whether accounting for a fourth would
be better than for half of it. Or -- if people would consider it worth the
trouble -- make it a sysctl. For now I sticked to accounting for half,
which should IMHO be a significant improvement.
The patch does one more thing: As users tend to be irritated by the choice
of killed processes (mainly because the children are killed first, despite
some of them having a very low OOM score), I added some more output: The
selected (father) process will be reported first and it's oom_score printed
to syslog.
Description:
Only account for half of children's vm size in oom score calculation
This should still give the parent enough point in case of fork bombs. If
any child however has more than 50% of the vm size of all children
together, it'll get a higher score and be elected.
This patch also makes the kernel display the oom_score.
Signed-off-by: Kurt Garloff <garloff@suse.de>
Cc: Rik van Riel <riel@redhat.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-02-21 02:27:51 +00:00
|
|
|
*ppoints = points;
|
2005-04-16 22:20:36 +00:00
|
|
|
}
|
[PATCH] cpusets: oom_kill tweaks
This patch series extends the use of the cpuset attribute 'mem_exclusive'
to support cpuset configurations that:
1) allow GFP_KERNEL allocations to come from a potentially larger
set of memory nodes than GFP_USER allocations, and
2) can constrain the oom killer to tasks running in cpusets in
a specified subtree of the cpuset hierarchy.
Here's an example usage scenario. For a few hours or more, a large NUMA
system at a University is to be divided in two halves, with a bunch of student
jobs running in half the system under some form of batch manager, and with a
big research project running in the other half. Each of the student jobs is
placed in a small cpuset, but should share the classic Unix time share
facilities, such as buffered pages of files in /bin and /usr/lib. The big
research project wants no interference whatsoever from the student jobs, and
has highly tuned, unusual memory and i/o patterns that intend to make full use
of all the main memory on the nodes available to it.
In this example, we have two big sibling cpusets, one of which is further
divided into a more dynamic set of child cpusets.
We want kernel memory allocations constrained by the two big cpusets, and user
allocations constrained by the smaller child cpusets where present. And we
require that the oom killer not operate across the two halves of this system,
or else the first time a student job runs amuck, the big research project will
likely be first inline to get shot.
Tweaking /proc/<pid>/oom_adj is not ideal -- if the big research project
really does run amuck allocating memory, it should be shot, not some other
task outside the research projects mem_exclusive cpuset.
I propose to extend the use of the 'mem_exclusive' flag of cpusets to manage
such scenarios. Let memory allocations for user space (GFP_USER) be
constrained by a tasks current cpuset, but memory allocations for kernel space
(GFP_KERNEL) by constrained by the nearest mem_exclusive ancestor of the
current cpuset, even though kernel space allocations will still _prefer_ to
remain within the current tasks cpuset, if memory is easily available.
Let the oom killer be constrained to consider only tasks that are in
overlapping mem_exclusive cpusets (it won't help much to kill a task that
normally cannot allocate memory on any of the same nodes as the ones on which
the current task can allocate.)
The current constraints imposed on setting mem_exclusive are unchanged. A
cpuset may only be mem_exclusive if its parent is also mem_exclusive, and a
mem_exclusive cpuset may not overlap any of its siblings memory nodes.
This patch was presented on linux-mm in early July 2005, though did not
generate much feedback at that time. It has been built for a variety of
arch's using cross tools, and built, booted and tested for function on SN2
(ia64).
There are 4 patches in this set:
1) Some minor cleanup, and some improvements to the code layout
of one routine to make subsequent patches cleaner.
2) Add another GFP flag - __GFP_HARDWALL. It marks memory
requests for USER space, which are tightly confined by the
current tasks cpuset.
3) Now memory requests (such as KERNEL) that not marked HARDWALL can
if short on memory, look in the potentially larger pool of memory
defined by the nearest mem_exclusive ancestor cpuset of the current
tasks cpuset.
4) Finally, modify the oom killer to skip any task whose mem_exclusive
cpuset doesn't overlap ours.
Patch (1), the one time I looked on an SN2 (ia64) build, actually saved 32
bytes of kernel text space. Patch (2) has no affect on the size of kernel
text space (it just adds a preprocessor flag). Patches (3) and (4) added
about 600 bytes each of kernel text space, mostly in kernel/cpuset.c, which
matters only if CONFIG_CPUSET is enabled.
This patch:
This patch applies a few comment and code cleanups to mm/oom_kill.c prior to
applying a few small patches to improve cpuset management of memory placement.
The comment changed in oom_kill.c was seriously misleading. The code layout
change in select_bad_process() makes room for adding another condition on
which a process can be spared the oom killer (see the subsequent
cpuset_nodes_overlap patch for this addition).
Also a couple typos and spellos that bugged me, while I was here.
This patch should have no material affect.
Signed-off-by: Paul Jackson <pj@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-09-06 22:18:09 +00:00
|
|
|
} while_each_thread(g, p);
|
2005-04-16 22:20:36 +00:00
|
|
|
return chosen;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* We must be careful though to never send SIGKILL a process with
|
|
|
|
|
* CAP_SYS_RAW_IO set, send SIGTERM instead (but it's unlikely that
|
|
|
|
|
* we select a process with CAP_SYS_RAW_IO set).
|
|
|
|
|
*/
|
2006-02-21 02:27:52 +00:00
|
|
|
static void __oom_kill_task(task_t *p, const char *message)
|
2005-04-16 22:20:36 +00:00
|
|
|
{
|
|
|
|
|
if (p->pid == 1) {
|
|
|
|
|
WARN_ON(1);
|
|
|
|
|
printk(KERN_WARNING "tried to kill init!\n");
|
|
|
|
|
return;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
task_lock(p);
|
|
|
|
|
if (!p->mm || p->mm == &init_mm) {
|
|
|
|
|
WARN_ON(1);
|
|
|
|
|
printk(KERN_WARNING "tried to kill an mm-less task!\n");
|
|
|
|
|
task_unlock(p);
|
|
|
|
|
return;
|
|
|
|
|
}
|
|
|
|
|
task_unlock(p);
|
2006-02-21 02:27:52 +00:00
|
|
|
printk(KERN_ERR "%s: Killed process %d (%s).\n",
|
|
|
|
|
message, p->pid, p->comm);
|
2005-04-16 22:20:36 +00:00
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* We give our sacrificial lamb high priority and access to
|
|
|
|
|
* all the memory it needs. That way it should be able to
|
|
|
|
|
* exit() and clear out its resources quickly...
|
|
|
|
|
*/
|
|
|
|
|
p->time_slice = HZ;
|
|
|
|
|
set_tsk_thread_flag(p, TIF_MEMDIE);
|
|
|
|
|
|
|
|
|
|
force_sig(SIGKILL, p);
|
|
|
|
|
}
|
|
|
|
|
|
2006-04-19 05:20:44 +00:00
|
|
|
static int oom_kill_task(task_t *p, const char *message)
|
2005-04-16 22:20:36 +00:00
|
|
|
{
|
2006-04-19 05:20:44 +00:00
|
|
|
struct mm_struct *mm;
|
2005-04-16 22:20:36 +00:00
|
|
|
task_t * g, * q;
|
|
|
|
|
|
2006-04-19 05:20:44 +00:00
|
|
|
mm = p->mm;
|
|
|
|
|
|
|
|
|
|
/* WARNING: mm may not be dereferenced since we did not obtain its
|
|
|
|
|
* value from get_task_mm(p). This is OK since all we need to do is
|
|
|
|
|
* compare mm to q->mm below.
|
|
|
|
|
*
|
|
|
|
|
* Furthermore, even if mm contains a non-NULL value, p->mm may
|
|
|
|
|
* change to NULL at any time since we do not hold task_lock(p).
|
|
|
|
|
* However, this is of no concern to us.
|
|
|
|
|
*/
|
|
|
|
|
|
|
|
|
|
if (mm == NULL || mm == &init_mm)
|
|
|
|
|
return 1;
|
2005-04-16 22:20:36 +00:00
|
|
|
|
2006-02-21 02:27:52 +00:00
|
|
|
__oom_kill_task(p, message);
|
2005-04-16 22:20:36 +00:00
|
|
|
/*
|
|
|
|
|
* kill all processes that share the ->mm (i.e. all threads),
|
|
|
|
|
* but are in a different thread group
|
|
|
|
|
*/
|
|
|
|
|
do_each_thread(g, q)
|
|
|
|
|
if (q->mm == mm && q->tgid != p->tgid)
|
2006-02-21 02:27:52 +00:00
|
|
|
__oom_kill_task(q, message);
|
2005-04-16 22:20:36 +00:00
|
|
|
while_each_thread(g, q);
|
|
|
|
|
|
2006-04-19 05:20:44 +00:00
|
|
|
return 0;
|
2005-04-16 22:20:36 +00:00
|
|
|
}
|
|
|
|
|
|
2006-04-19 05:20:44 +00:00
|
|
|
static int oom_kill_process(struct task_struct *p, unsigned long points,
|
|
|
|
|
const char *message)
|
2005-04-16 22:20:36 +00:00
|
|
|
{
|
|
|
|
|
struct task_struct *c;
|
|
|
|
|
struct list_head *tsk;
|
|
|
|
|
|
[PATCH] OOM kill: children accounting
In the badness() calculation, there's currently this piece of code:
/*
* Processes which fork a lot of child processes are likely
* a good choice. We add the vmsize of the children if they
* have an own mm. This prevents forking servers to flood the
* machine with an endless amount of children
*/
list_for_each(tsk, &p->children) {
struct task_struct *chld;
chld = list_entry(tsk, struct task_struct, sibling);
if (chld->mm = p->mm && chld->mm)
points += chld->mm->total_vm;
}
The intention is clear: If some server (apache) keeps spawning new children
and we run OOM, we want to kill the father rather than picking a child.
This -- to some degree -- also helps a bit with getting fork bombs under
control, though I'd consider this a desirable side-effect rather than a
feature.
There's one problem with this: No matter how many or few children there are,
if just one of them misbehaves, and all others (including the father) do
everything right, we still always kill the whole family. This hits in real
life; whether it's javascript in konqueror resulting in kdeinit (and thus the
whole KDE session) being hit or just a classical server that spawns children.
Sidenote: The killer does kill all direct children as well, not only the
selected father, see oom_kill_process().
The idea in attached patch is that we do want to account the memory
consumption of the (direct) children to the father -- however not fully.
This maintains the property that fathers with too many children will still
very likely be picked, whereas a single misbehaving child has the chance to
be picked by the OOM killer.
In the patch I account only half (rounded up) of the children's vm_size to
the parent. This means that if one child eats more mem than the rest of
the family, it will be picked, otherwise it's still the father and thus the
whole family that gets selected.
This is heuristics -- we could debate whether accounting for a fourth would
be better than for half of it. Or -- if people would consider it worth the
trouble -- make it a sysctl. For now I sticked to accounting for half,
which should IMHO be a significant improvement.
The patch does one more thing: As users tend to be irritated by the choice
of killed processes (mainly because the children are killed first, despite
some of them having a very low OOM score), I added some more output: The
selected (father) process will be reported first and it's oom_score printed
to syslog.
Description:
Only account for half of children's vm size in oom score calculation
This should still give the parent enough point in case of fork bombs. If
any child however has more than 50% of the vm size of all children
together, it'll get a higher score and be elected.
This patch also makes the kernel display the oom_score.
Signed-off-by: Kurt Garloff <garloff@suse.de>
Cc: Rik van Riel <riel@redhat.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-02-21 02:27:51 +00:00
|
|
|
printk(KERN_ERR "Out of Memory: Kill process %d (%s) score %li and "
|
|
|
|
|
"children.\n", p->pid, p->comm, points);
|
2005-04-16 22:20:36 +00:00
|
|
|
/* Try to kill a child first */
|
|
|
|
|
list_for_each(tsk, &p->children) {
|
|
|
|
|
c = list_entry(tsk, struct task_struct, sibling);
|
|
|
|
|
if (c->mm == p->mm)
|
|
|
|
|
continue;
|
2006-04-19 05:20:44 +00:00
|
|
|
if (!oom_kill_task(c, message))
|
|
|
|
|
return 0;
|
2005-04-16 22:20:36 +00:00
|
|
|
}
|
2006-02-21 02:27:52 +00:00
|
|
|
return oom_kill_task(p, message);
|
2005-04-16 22:20:36 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* oom_kill - kill the "best" process when we run out of memory
|
|
|
|
|
*
|
|
|
|
|
* If we run out of memory, we have the choice between either
|
|
|
|
|
* killing a random task (bad), letting the system crash (worse)
|
|
|
|
|
* OR try to be smart about which process to kill. Note that we
|
|
|
|
|
* don't have to be perfect here, we just have to be good.
|
|
|
|
|
*/
|
2006-02-21 02:27:52 +00:00
|
|
|
void out_of_memory(struct zonelist *zonelist, gfp_t gfp_mask, int order)
|
2005-04-16 22:20:36 +00:00
|
|
|
{
|
2006-02-21 02:27:52 +00:00
|
|
|
task_t *p;
|
2006-03-01 00:59:19 +00:00
|
|
|
unsigned long points = 0;
|
2005-04-16 22:20:36 +00:00
|
|
|
|
2005-07-08 00:56:06 +00:00
|
|
|
if (printk_ratelimit()) {
|
|
|
|
|
printk("oom-killer: gfp_mask=0x%x, order=%d\n",
|
|
|
|
|
gfp_mask, order);
|
2006-02-01 11:05:51 +00:00
|
|
|
dump_stack();
|
2005-07-08 00:56:06 +00:00
|
|
|
show_mem();
|
|
|
|
|
}
|
2005-06-22 00:14:56 +00:00
|
|
|
|
2006-01-14 21:21:06 +00:00
|
|
|
cpuset_lock();
|
2005-04-16 22:20:36 +00:00
|
|
|
read_lock(&tasklist_lock);
|
2006-02-21 02:27:52 +00:00
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* Check if there were limitations on the allocation (only relevant for
|
|
|
|
|
* NUMA) that may require different handling.
|
|
|
|
|
*/
|
|
|
|
|
switch (constrained_alloc(zonelist, gfp_mask)) {
|
|
|
|
|
case CONSTRAINT_MEMORY_POLICY:
|
2006-04-19 05:20:44 +00:00
|
|
|
oom_kill_process(current, points,
|
2006-02-21 02:27:52 +00:00
|
|
|
"No available memory (MPOL_BIND)");
|
|
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
case CONSTRAINT_CPUSET:
|
2006-04-19 05:20:44 +00:00
|
|
|
oom_kill_process(current, points,
|
2006-02-21 02:27:52 +00:00
|
|
|
"No available memory in cpuset");
|
|
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
case CONSTRAINT_NONE:
|
2005-04-16 22:20:36 +00:00
|
|
|
retry:
|
2006-02-21 02:27:52 +00:00
|
|
|
/*
|
|
|
|
|
* Rambo mode: Shoot down a process and hope it solves whatever
|
|
|
|
|
* issues we may have.
|
|
|
|
|
*/
|
|
|
|
|
p = select_bad_process(&points);
|
2005-04-16 22:20:36 +00:00
|
|
|
|
2006-02-21 02:27:52 +00:00
|
|
|
if (PTR_ERR(p) == -1UL)
|
|
|
|
|
goto out;
|
2005-04-16 22:20:36 +00:00
|
|
|
|
2006-02-21 02:27:52 +00:00
|
|
|
/* Found nothing?!?! Either we hang forever, or we panic. */
|
|
|
|
|
if (!p) {
|
|
|
|
|
read_unlock(&tasklist_lock);
|
|
|
|
|
cpuset_unlock();
|
|
|
|
|
panic("Out of memory and no killable processes...\n");
|
|
|
|
|
}
|
2005-04-16 22:20:36 +00:00
|
|
|
|
2006-04-19 05:20:44 +00:00
|
|
|
if (oom_kill_process(p, points, "Out of memory"))
|
2006-02-21 02:27:52 +00:00
|
|
|
goto retry;
|
|
|
|
|
|
|
|
|
|
break;
|
|
|
|
|
}
|
2005-04-16 22:20:36 +00:00
|
|
|
|
2006-02-21 02:27:52 +00:00
|
|
|
out:
|
2006-03-02 10:54:28 +00:00
|
|
|
read_unlock(&tasklist_lock);
|
2006-01-14 21:21:06 +00:00
|
|
|
cpuset_unlock();
|
2005-04-16 22:20:36 +00:00
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* Give "p" a good chance of killing itself before we
|
2006-01-08 09:01:05 +00:00
|
|
|
* retry to allocate memory unless "p" is current
|
2005-04-16 22:20:36 +00:00
|
|
|
*/
|
2006-01-08 09:01:05 +00:00
|
|
|
if (!test_thread_flag(TIF_MEMDIE))
|
2006-03-02 10:54:28 +00:00
|
|
|
schedule_timeout_uninterruptible(1);
|
2005-04-16 22:20:36 +00:00
|
|
|
}
|