Resource Groups... formerly CKRM
Subsystems (5)
cover Resource Groups... formerly CKRM
Andrew,
CKRM has gone through a major overhaul by removing some of the complexity, cutting down on features and moving portions to userspace.
We have posted ckrm last week as an RFC:
infrastructure:
http://www.ussg.iu.edu/hypermail/linux/kernel/0604.2/1297.html numtasks controller: http://www.ussg.iu.edu/hypermail/linux/kernel/0604.2/1311.html CPU controller: http://www.ussg.iu.edu/hypermail/linux/kernel/0604.2/1310.html
Diffstat of this patchset (including the numtasks controller that follows) is:
23 files changed, 2466 insertions(+), 5 deletions(-)
including Documentaion and comments.
This patchset will be followed with two controllers:
- a simple controller, numtasks to control number of tasks
- CPU controller, to control CPU resource.
Please consider these for inclusion in -mm tree.
--
Changes since last post:
- CKRM has been replaced with Resource Groups
- 'class' has been replaced with resource group
- Limit the depth of the hierarchy.
---------
Patch Descriptions:
01/12 - controller_support
This patch defines data structures for defining a resource group and resource controller.
Provides register/unregister functions for a controller. Provides utility functions to get a controller's data structure.
02/12 - resource_group_support
Provides functions to alloc and free a user defined resource group. Provides utility macro to walk through the resource group hierarchy.
03/12 - shares_support
Provides functions to set/get shares of a specific resource of a resource group.
Defines a teardown function that is intended to be called when user disables resource group (by umount of the configfs component).
04/12 - tasksupport
Adds logic to support adding/removing task to/from a resource group Provides interface to set a task's resource group.
05/12 - tasksupport_fork_exit_init
Initializes and clears Resource Group specific information in a task at fork() and exit(). Inititalizes Resource Groups (called from start_kernel)
06/12: task_procsupport
Adds an interface in /proc to get the name of a resource group to which a task is attached.
07/12 - user_interface
Create the configfs component for managing Resource groups. Hooks up with configfs. Provides functions for creating and deleting resource groups.
08/12 - user_interface_attr_support
Adds the basic attribute store and show functions.
09/12 - user_interface_stats
Adds attr_store and attr_show support for stats file.
10/12 - user_interface_shares
Adds attr_store and attr_show support for shares file.
11/12 - user_interface_members
Adds attr_store and attr_show support for members file.
12/12 - res_groups_docs
Documentation describing important Resource Groups components such as resource group, shares, controllers, and the configfs based user interface. --
Thanks & Regards,
chandra
Signature · 4 lines
----------------------------------------------------------------------
Chandra Seetharaman | Be careful what you choose....
- [email protected] | .......you may get it.
----------------------------------------------------------------------PATCH 1/12 Register/Unregister interface for Controllers
01/12 - controller_support
This patch defines data structures for defining a resource group and resource controller.
Provides register/unregister functions for a controller. Provides utility functions to get a controller's data structure. --
include/linux/res_group.h | 75 +++++++++++++++ include/linux/res_group_rc.h | 68 +++++++++++++ init/Kconfig | 14 ++ kernel/Makefile | 1 kernel/res_group/Makefile | 1 kernel/res_group/local.h | 10 ++ kernel/res_group/res_group.c | 214 +++++++++++++++++++++++++++++++++++++++++++ 7 files changed, 383 insertions(+)
include/linux/res_group.h new file +75 -0
@@ -0,0 +1,75 @@ +/* +* res_group.h - Header file to be used by Resource Groups +* +* Copyright (C) Hubertus Franke, IBM Corp. 2003, 2004 +* (C) Shailabh Nagar, IBM Corp. 2003, 2004 +* (C) Chandra Seetharaman, IBM Corp. 2003, 2004, 2005 +* +* Provides data structures, macros and kernel APIs +* +* More details at http://ckrm.sf.net +* +* This program is free software; you can redistribute it and/or modify +* it under the terms of the GNU General Public License as published by +* the Free Software Foundation; either version 2 of the License, or +* (at your option) any later version. +* +*/ + +#ifndef _LINUX_RES_GROUP_H +#define _LINUX_RES_GROUP_H + +#ifdef CONFIG_RES_GROUPS +#include <linux/spinlock.h> +#include <linux/list.h> +#include <linux/kref.h> + +#define SHARE_UNCHANGED (-1)/* implicitly specified by userspace, +* never stored in a resource group' +* shares struct; never displayed */ +#define SHARE_UNSUPPORTED (-2)/* If the resource controller doesn't +* support user changing a shares value +* it sets the corresponding share +* value to UNSUPPORTED when it returns +* the newly allocated shares data +* structure */ +#define SHARE_DONT_CARE (-3) + +#define SHARE_DEFAULT_DIVISOR (100) + +#define MAX_RES_CTLRS 8/* max # of resource controllers */ + +#define NO_RES_GROUP NULL +#define NO_SHARE NULL +#define NO_RES_ID MAX_RES_CTLRS/* Invalid ID */ + +/* +* Share quantities are a child's fraction of the parent's resource +* specified by a divisor in the parent and a dividend in the child. +* +* Shares are represented as a relative quantity between parent and child +* to simplify locking when propagating modifications to the shares of a +* resource group. Only the parent and the children of the modified +* resource group need to be locked. +*/ +struct res_shares { +}; + +/* +* Class is the grouping of tasks with shares of each resource that has +* registered a resource controller (see include/linux/res_group_rc.h). +*/ +struct resource_group { +int depth;/* depth of this resource group. root == 0 */ +spinlock_t group_lock;/* protects task_list, shares and children +* When grabbing group_lock in a hierarchy, +* grab parent's group_lock first. +* If resource controller uses a resource_group +* specific lock, grab group_lock before +* grabbing resource specific lock */ +struct res_shares *shares[MAX_RES_CTLRS ];/* resource shares */ +struct list_head group_list;/* entry in system-wide list */ +}; + +#endif /* CONFIG_RES_GROUPS */ +#endif /* _LINUX_RES_GROUP_H */
include/linux/res_group_rc.h new file +68 -0
68 lines not on this page · raw message
kernel/res_group/res_group.c new file +214 -0
214 lines not on this page · raw message
kernel/Makefile +1 -0
@@ -38,6 +38,7 @@ obj-$(CONFIG_GENERIC_HARDIRQS) += irq/ obj-$(CONFIG_SECCOMP) += seccomp.o obj-$(CONFIG_RCU_TORTURE_TEST) += rcutorture.o obj-$(CONFIG_RELAY) += relay.o +obj-$(CONFIG_RES_GROUPS) += res_group/ ifneq ($(CONFIG_SCHED_NO_NO_OMIT_FRAME_POINTER),y) # According to Alan Modra <[email protected]>, the -fno-omit-frame-pointer is
kernel/res_group/Makefile new file +1 -0
@@ -0,0 +1 @@ +obj-y = res_group.o
init/Kconfig +14 -0
@@ -150,6 +150,20 @@ config BSD_PROCESS_ACCT_V3 for processing it. A preliminary version of these tools is available at <http://www.physik3.uni-rostock.de/tim/kernel/utils/acct/>. +menu "Resource Groups" + +config RES_GROUPS + bool "Resource Groups" + depends on EXPERIMENTAL + help + Resource Groups is a framework for controlling and monitoring + resource allocation of user-defined groups of tasks. For more + information, please visit http://ckrm.sf.net. + + If you say Y here, enable the Resource Group File System and at least + one of the resource controllers below. Say N if you are unsure. + +endmenu config SYSCTL bool "Sysctl support" ---help---
kernel/res_group/local.h new file +10 -0
@@ -0,0 +1,10 @@ +/* +* Contains function definitions that are local to the Resource Groups. +* NOT to be included by controllers. +*/ + +#include <linux/res_group_rc.h> + +extern struct res_controller *get_controller_by_name (const char *); +extern struct res_controller *get_controller_by_id (unsigned int ); +extern void put_controller (struct res_controller *);
Signature · 4 lines
----------------------------------------------------------------------
Chandra Seetharaman | Be careful what you choose....
- [email protected] | .......you may get it.
----------------------------------------------------------------------PATCH 2/12 Resource group creation/deletion
In reply to Chandra Seetharaman02/12 - resource_group_support
Provides functions to alloc and free a user defined resource group. Provides utility macro to walk through the resource group hierarchy. --
include/linux/res_group.h | 9 ++ include/linux/res_group_rc.h | 9 ++ kernel/res_group/local.h | 4 + kernel/res_group/res_group.c | 167 +++++++++++++++++++++++++++++++++++++++---- 4 files changed, 176 insertions(+), 13 deletions(-)
include/linux/res_group.h +9 -0
@@ -38,6 +38,7 @@#define SHARE_DEFAULT_DIVISOR (100)#define MAX_RES_CTLRS 8/* max # of resource controllers */ +#define MAX_DEPTH 5/* max depth of hierarchy supported */ #define NO_RES_GROUP NULL#define NO_SHARE NULL
@@ -60,6 +61,8 @@ struct res_shares {* registered a resource controller (see include/linux/res_group_rc.h). */ struct resource_group { +const char *name; +struct kref ref;int depth;/* depth of this resource group. root == 0 */ spinlock_t group_lock;/* protects task_list, shares and children * When grabbing group_lock in a hierarchy,
@@ -69,6 +72,12 @@ struct resource_group {* grabbing resource specific lock */ struct res_shares *shares[MAX_RES_CTLRS ];/* resource shares */ struct list_head group_list;/* entry in system-wide list */ + +struct list_head task_list;/* this resource groups's tasks */ + +struct resource_group *parent; +struct list_head siblings;/* entry in list of siblings */ + struct list_head children; /* head of children */ };#endif /* CONFIG_RES_GROUPS */
kernel/res_group/res_group.c +154 -13
224 lines not on this page · raw message
include/linux/res_group_rc.h +9 -0
@@ -65,4 +65,13 @@ struct res_controller {extern int register_controller (struct res_controller *);extern int unregister_controller (struct res_controller *); +extern struct resource_group default_res_group; +static inline int is_res_group_root (const struct resource_group *rgroup) +{ +return (rgroup == &default_res_group); +} + +#define for_each_child (child, parent) \ + list_for_each_entry(child, &parent->children, siblings) +#endif /* _LINUX_RES_GROUP_RC_H */
kernel/res_group/local.h +4 -0
@@ -8,3 +8,7 @@extern struct res_controller *get_controller_by_name (const char *);extern struct res_controller *get_controller_by_id (unsigned int );extern void put_controller (struct res_controller *); +extern struct resource_group *alloc_res_group (struct resource_group *, +const char *); +extern int free_res_group (struct resource_group *); +extern void release_res_group (struct kref *);
Signature · 4 lines
----------------------------------------------------------------------
Chandra Seetharaman | Be careful what you choose....
- [email protected] | .......you may get it.
----------------------------------------------------------------------PATCH 3/12 Share Handling
In reply to Chandra Seetharaman03/12 - shares_support
Provides functions to set/get shares of a specific resource of a resource group.
Defines a teardown function that is intended to be called when user disables resource group (by umount of the configfs component). --
include/linux/res_group.h | 14 ++ include/linux/res_group_rc.h | 10 + kernel/res_group/Makefile | 2 kernel/res_group/local.h | 6 + kernel/res_group/res_group.c | 24 ++++ kernel/res_group/shares.c | 242 +++++++++++++++++++++++++++++++++++++++++++ 6 files changed, 297 insertions(+), 1 deletion(-)
include/linux/res_group.h +14 -0
@@ -54,6 +54,20 @@* resource group need to be locked. */ struct res_shares { +/* shares only set by userspace */ +int min_shares;/* minimun fraction of parent's resources allowed */ +int max_shares;/* maximum fraction of parent's resources allowed */ +int child_shares_divisor;/* >= 1, may not be DONT_CARE */ + +/* +* share values invisible to userspace. adjusted when userspace +* sets shares +*/ +int unused_min_shares; +/* 0 <= unused_min_shares <= (child_shares_divisor - +* Sum of min_shares of children) +*/ +int cur_max_shares;/* max(children's max_shares). need better name */ };/*
kernel/res_group/shares.c new file +242 -0
242 lines not on this page · raw message
kernel/res_group/Makefile +1 -1
@@ -1 +1 @@ -obj-y = res_group.o +obj-y = res_group.o shares.o
kernel/res_group/res_group.c +24 -0
@@ -173,6 +173,8 @@ static void do_free_shares_struct(structif (shares ==NULL )return ; +set_shares_to_default (rgroup, ctlr); +spin_lock (&rgroup->group_lock); rgroup->shares[ctlr->ctlr_id] =NULL ;spin_unlock (&rgroup->group_lock);
@@ -345,6 +347,26 @@ int unregister_controller(struct res_conreturn remove_controller (ctlr); } +/* +* Bring the state of Resource Groups to the initial state. Only shares +* of the default resource group need to be changed to default values. +* At this point no user-defined resource groups should exist. +*/ +void res_group_teardown (void ) +{ +int i; +struct res_controller *ctlr; + +BUG_ON (num_res_groups !=0 ); +for (i =0 ; i <MAX_RES_CTLRS ; i++) { + ctlr =get_controller_by_id (i); +if (ctlr) { +set_shares_to_default (&default_res_group, ctlr); +put_controller (ctlr); + } + } +} +EXPORT_SYMBOL_GPL (register_controller);EXPORT_SYMBOL_GPL (unregister_controller);EXPORT_SYMBOL_GPL (get_controller_by_name);
@@ -353,3 +375,5 @@ EXPORT_SYMBOL_GPL(put_controller);EXPORT_SYMBOL_GPL (alloc_res_group);EXPORT_SYMBOL_GPL (free_res_group);EXPORT_SYMBOL_GPL (default_res_group); +EXPORT_SYMBOL_GPL (set_controller_shares); +EXPORT_SYMBOL_GPL (res_group_teardown);
include/linux/res_group_rc.h +10 -0
@@ -74,4 +74,14 @@ static inline int is_res_group_root(cons#define for_each_child (child, parent) \ list_for_each_entry(child, &parent->children, siblings) +/* Get controller specific shares structure for the given resource group */ +static inline struct res_shares *get_controller_shares ( +struct resource_group *rgroup,struct res_controller *ctlr) +{ +if (rgroup && ctlr) +return rgroup->shares[ctlr->ctlr_id]; +else +return NO_SHARE ; +} +#endif /* _LINUX_RES_GROUP_RC_H */
kernel/res_group/local.h +6 -0
@@ -12,3 +12,9 @@ extern struct resource_group *alloc_res_ const char *);extern int free_res_group (struct resource_group *);extern void release_res_group (struct kref *); +extern int set_controller_shares (struct resource_group *, +struct res_controller *,const struct res_shares *); +/* Set shares for the given resource group and resource to default values */ +extern void set_shares_to_default (struct resource_group *, +struct res_controller *); +extern void res_group_teardown (void );
Signature · 4 lines
----------------------------------------------------------------------
Chandra Seetharaman | Be careful what you choose....
- [email protected] | .......you may get it.
----------------------------------------------------------------------PATCH 6/12 Add proc interface to get resource group info of task
In reply to Chandra Seetharaman06/12: task_procsupport
Adds an interface in /proc to get the name of a resource group to which a task is attached.
--
fs/proc/base.c | 19 +++++++++++++++++++ include/linux/res_group.h | 2 ++ kernel/res_group/task.c | 32 +++++++++++++++++++++++++++++++- 3 files changed, 52 insertions(+), 1 deletion(-)
fs/proc/base.c +19 -0
@@ -70,6 +70,7 @@#include <linux/ptrace.h> #include <linux/seccomp.h> #include <linux/cpuset.h> +#include <linux/res_group.h> #include <linux/audit.h> #include <linux/poll.h> #include "internal.h"
@@ -115,6 +116,9 @@ enum pid_directory_inos {#ifdef CONFIG_CPUSETS PROC_TGID_CPUSET,#endif +#ifdef CONFIG_RES_GROUPS + PROC_TGID_RES_GROUP, +#endif #ifdef CONFIG_SECURITY PROC_TGID_ATTR , PROC_TGID_ATTR_CURRENT,
@@ -156,6 +160,9 @@ enum pid_directory_inos {#ifdef CONFIG_CPUSETS PROC_TID_CPUSET,#endif +#ifdef CONFIG_RES_GROUPS + PROC_TID_RES_GROUP, +#endif #ifdef CONFIG_SECURITY PROC_TID_ATTR , PROC_TID_ATTR_CURRENT,
@@ -219,6 +226,9 @@ static struct pid_entry tgid_base_stuff[#ifdef CONFIG_CPUSETS E(PROC_TGID_CPUSET, "cpuset", S_IFREG|S_IRUGO),#endif +#ifdef CONFIG_RES_GROUPS +E (PROC_TGID_RES_GROUP ,"res_group" ,S_IFREG |S_IRUGO ), +#endifE (PROC_TGID_OOM_SCORE ,"oom_score" ,S_IFREG |S_IRUGO ),E (PROC_TGID_OOM_ADJUST ,"oom_adj" ,S_IFREG |S_IRUGO |S_IWUSR ), #ifdef CONFIG_AUDITSYSCALL
@@ -261,6 +271,9 @@ static struct pid_entry tid_base_stuff[]#ifdef CONFIG_CPUSETS E(PROC_TID_CPUSET, "cpuset", S_IFREG|S_IRUGO),#endif +#ifdef CONFIG_RES_GROUPS +E (PROC_TID_RES_GROUP ,"res_group" ,S_IFREG |S_IRUGO ), +#endifE (PROC_TID_OOM_SCORE ,"oom_score" ,S_IFREG |S_IRUGO ),E (PROC_TID_OOM_ADJUST ,"oom_adj" ,S_IFREG |S_IRUGO |S_IWUSR ), #ifdef CONFIG_AUDITSYSCALL
@@ -1823,6 +1836,12 @@ static struct dentry *proc_pident_lookup inode->i_fop = &proc_cpuset_operations;break ;#endif +#ifdef CONFIG_RES_GROUPS +case PROC_TID_RES_GROUP : +case PROC_TGID_RES_GROUP : + inode->i_fop = &proc_res_group_operations; +break ; +#endif case PROC_TID_OOM_SCORE :case PROC_TGID_OOM_SCORE : inode->i_fop = &proc_info_file_operations;
kernel/res_group/task.c +31 -1
@@ -13,7 +13,8 @@* (at your option) any later version. * */ -#include <linux/sched.h> +#include <linux/seq_file.h> +#include <linux/proc_fs.h> #include <linux/module.h> #include "local.h"
@@ -194,4 +195,33 @@ next_task:kref_put (&rgroup->ref, release_res_group); } +static int proc_res_group_show (struct seq_file *m,void *v) +{ +struct task_struct *tsk = m->private; +struct resource_group *rgroup = tsk->res_group; + +if (!rgroup) +return -EINVAL ; + +kref_get (&rgroup->ref); +seq_puts (m,"/" ); +if (!is_res_group_root (rgroup)) +seq_puts (m, rgroup->name); +seq_putc (m,'\n' ); +kref_put (&rgroup->ref, release_res_group); +return 0 ; +} + +static int res_group_open (struct inode *inode,struct file *file) +{ +struct task_struct *tsk =PROC_I (inode)->task; +return single_open (file, proc_res_group_show, tsk); +} + +struct file_operations proc_res_group_operations = { + .open = res_group_open, + .read = seq_read, + .llseek = seq_lseek, + .release = single_release, +};EXPORT_SYMBOL_GPL (set_res_group);
include/linux/res_group.h +2 -0
@@ -94,6 +94,8 @@ struct resource_group {struct list_head children;/* head of children */ }; +extern struct file_operations proc_res_group_operations; +extern void init_task_res_group (struct task_struct *);extern void clear_task_res_group (struct task_struct *);extern void res_group_init (void );
Signature · 4 lines
----------------------------------------------------------------------
Chandra Seetharaman | Be careful what you choose....
- [email protected] | .......you may get it.
----------------------------------------------------------------------PATCH 4/12 Add task logic to resource groups
In reply to Chandra Seetharaman04/12 - tasksupport
Adds logic to support adding/removing task to/from a resource group Provides interface to set a task's resource group.
--
include/linux/sched.h | 4 + kernel/res_group/Makefile | 2 kernel/res_group/local.h | 1 kernel/res_group/res_group.c | 25 +++++++ kernel/res_group/task.c | 145 +++++++++++++++++++++++++++++++++++++++++++ 5 files changed, 176 insertions(+), 1 deletion(-)
include/linux/sched.h +4 -0
@@ -888,6 +888,10 @@ struct task_struct {* cache last used pipe for splice */ struct pipe_inode_info *splice_pipe; +#ifdef CONFIG_RES_GROUPS +struct resource_group *res_group; +struct list_head member_list;/* list of tasks in the resource group */ +#endif /* CONFIG_RES_GROUPS */ };static inline pid_t process_group (struct task_struct *tsk)
kernel/res_group/task.c new file +145 -0
145 lines not on this page · raw message
kernel/res_group/Makefile +1 -1
@@ -1 +1 @@ -obj-y = res_group.o shares.o +obj-y = res_group.o shares.o task.o
kernel/res_group/res_group.c +25 -0
@@ -251,6 +251,27 @@ static int add_controller(struct res_conreturn ret; }/* +* Helper function to move all tasks in a resource group to/from the +* registering/unregistering resource controller. +* +* Assumes ctlr is valid and rgroup is initialized with resource +* controller's shares. +*/ +static void move_tasks (struct resource_group *rgroup, +struct res_controller *ctlr, +struct res_shares *from,struct res_shares *to) +{ +struct task_struct *tsk; + +if (!ctlr->move_task) +return ; +spin_lock (&rgroup->group_lock); +list_for_each_entry (tsk, &rgroup->task_list, member_list) + ctlr->move_task (tsk, from, to); +spin_unlock (&rgroup->group_lock); +} + +/* * Interface for registering a resource controller. * * Returns the 0 on success, -errno for failure.
@@ -287,6 +308,8 @@ int register_controller(struct res_contrkref_get (&rgroup->ref);read_unlock (&res_group_lock);do_alloc_shares_struct (rgroup, ctlr); +move_tasks (rgroup, ctlr,NO_SHARE , + rgroup->shares[ctlr->ctlr_id]);if (prev_rgroup)kref_put (&prev_rgroup->ref, release_res_group); prev_rgroup = rgroup;
@@ -333,6 +356,8 @@ int unregister_controller(struct res_conlist_for_each_entry_reverse (rgroup, &res_groups, group_list) {kref_get (&rgroup->ref);read_unlock (&res_group_lock); +move_tasks (rgroup, ctlr, rgroup->shares[ctlr->ctlr_id], +NO_SHARE );do_free_shares_struct (rgroup, ctlr);if (prev_rgroup)kref_put (&prev_rgroup->ref, release_res_group);
kernel/res_group/local.h +1 -0
@@ -18,3 +18,4 @@ extern int set_controller_shares(structextern void set_shares_to_default (struct resource_group *,struct res_controller *);extern void res_group_teardown (void ); +extern int set_res_group (pid_t ,struct resource_group *);
Signature · 4 lines
----------------------------------------------------------------------
Chandra Seetharaman | Be careful what you choose....
- [email protected] | .......you may get it.
----------------------------------------------------------------------PATCH 7/12 Resource Groups Configfs Subsystem (RGCS)
In reply to Chandra Seetharaman07/12 - user_interface
Create the configfs component for managing Resource groups. Hooks up with configfs. Provides functions for creating and deleting resource groups. --
init/Kconfig | 12 +++ kernel/res_group/Makefile | 1 kernel/res_group/rgcs.c | 162 ++++++++++++++++++++++++++++++++++++++++++++++ 3 files changed, 175 insertions(+)
init/Kconfig +12 -0
@@ -163,6 +163,18 @@ config RES_GROUPS If you say Y here, enable the Resource Group File System and at least one of the resource controllers below. Say N if you are unsure. +config RGCS + tristate "Resource Groups Configfs Subsystem" + depends on RES_GROUPS + select CONFIGFS_FS + default m + help + RGCS is the User Interface for Resource Groups. Compiling it as + a module permits users to only load RGCS if they intend to use it. + + Say M if unsure, Y to save on module loading. N doesn't make sense + when Resource Groups has been configured. + endmenu config SYSCTL bool "Sysctl support"
kernel/res_group/rgcs.c new file +162 -0
162 lines not on this page · raw message
kernel/res_group/Makefile +1 -0
@@ -1 +1,2 @@ obj-y = res_group.o shares.o task.o +obj-$(CONFIG_RGCS) += rgcs.o
Signature · 4 lines
----------------------------------------------------------------------
Chandra Seetharaman | Be careful what you choose....
- [email protected] | .......you may get it.
----------------------------------------------------------------------PATCH 8/12 Add attribute support to RGCS
In reply to Chandra Seetharaman08/12 - user_interface_attr_support
Adds the basic attribute store and show functions. --
rgcs.c | 53 ++++++++++++++++++++++++++++++++++++++++++++++++++++- 1 files changed, 52 insertions(+), 1 deletion(-)
kernel/res_group/rgcs.c +52 -1
@@ -14,13 +14,21 @@* as published by the Free Software Foundation. * */ +#include <linux/ctype.h> #include <linux/module.h> #include <linux/configfs.h> +#include <linux/parser.h> #include "local.h" static struct configfs_subsystem rgcs_subsys;static struct config_item_type rgcs_item_type; +struct rgroup_attribute { +struct configfs_attribute configfs_attr; +ssize_t (*show)(struct resource_group *,char *); +int (*store)(struct resource_group *,const char *); +}; +struct rgcs_rgroup {char *name;struct resource_group *core;
@@ -58,6 +66,42 @@ static inline struct resource_group *itereturn group_to_res_group (to_config_group (item)); } +static ssize_t rgcs_attr_show (struct config_item *item, +struct configfs_attribute *attr,char *buf) +{ +struct rgroup_attribute *rgroup_attr; +struct resource_group *rgroup =item_to_res_group (item); + + rgroup_attr = container_of(attr, struct rgroup_attribute, + configfs_attr); +return rgroup_attr->show (rgroup, buf); +} + +static ssize_t rgcs_attr_store (struct config_item *item, +struct configfs_attribute *attr,const char *buf, +size_t count) +{ +char *filtered_buf, *p; +ssize_t rc; +struct rgroup_attribute *rgroup_attr; +struct resource_group *rgroup =item_to_res_group (item); + + rgroup_attr = container_of(attr, struct rgroup_attribute, + configfs_attr); + filtered_buf =kzalloc (count +1 ,GFP_KERNEL ); +if (!filtered_buf) +return -ENOMEM ; +strncpy (filtered_buf, buf, count); +for (p = filtered_buf;isprint (*p); ++p) + ; + *p ='\0' ; + rc = rgroup_attr->store (rgroup, filtered_buf); +kfree (filtered_buf); +if (rc) +return rc; +return count; +} +/* * This is the function that is called when a 'mkdir' command * is issued under our filesystem
@@ -119,17 +163,24 @@ static void rgcs_rgroup_release_item(str }static struct configfs_item_operations rgcs_rgroup_item_ops = { - .release = rgcs_rgroup_release_item, + .release = rgcs_rgroup_release_item, + .show_attribute = rgcs_attr_show, + .store_attribute = rgcs_attr_store, };static struct configfs_group_operations rgcs_rgroup_group_ops = { .make_group = make_rgcs_rgroup, }; +static struct configfs_attribute *rgroup_attrs[] = { +NULL +}; +static struct config_item_type rgcs_item_type = { .ct_owner =THIS_MODULE , .ct_item_ops = &rgcs_rgroup_item_ops, .ct_group_ops = &rgcs_rgroup_group_ops, + .ct_attrs = rgroup_attrs };static struct configfs_subsystem rgcs_subsys = {
Signature · 4 lines
----------------------------------------------------------------------
Chandra Seetharaman | Be careful what you choose....
- [email protected] | .......you may get it.
----------------------------------------------------------------------PATCH 9/12 Add stats file support to RGCS
In reply to Chandra Seetharaman09/12 - user_interface_stats
Adds attr_store and attr_show support for stats file. --
rgcs.c | 112 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++-- 1 files changed, 110 insertions(+), 2 deletions(-)
kernel/res_group/rgcs.c +110 -2
@@ -20,8 +20,102 @@#include <linux/parser.h> #include "local.h" -static struct configfs_subsystem rgcs_subsys; -static struct config_item_type rgcs_item_type; +#define RES_STRING "res" + +static ssize_t show_stats (struct resource_group *rgroup,char *buf) +{ +int i, j =0 , rc =0 ; +size_t buf_size =PAGE_SIZE -1 ;/* allow only PAGE_SIZE # of bytes */ +struct res_controller *ctlr; +struct res_shares *shares; + +for (i =0 ; i <MAX_RES_CTLRS ; i++, j =0 ) { +if (buf_size <=0 ) +break ; + ctlr =get_controller_by_id (i); +if (!ctlr) +continue ; + shares =get_controller_shares (rgroup, ctlr); +if (shares && ctlr->show_stats) + j = ctlr->show_stats (shares, buf, buf_size); +put_controller (ctlr); + rc += j; + buf += j; + buf_size -= j; + } +if (i <MAX_RES_CTLRS ) + rc = -ENOSPC ; +return rc; +} + +enum parse_token_t { + parse_res_type, parse_err +}; + +static match_table_t parse_tokens = { + {parse_res_type,RES_STRING "=%s" }, + {parse_err,NULL } +}; + +static int stats_parse (const char *options, +char **resname,char **remaining_line) +{ +char *p, *str; +int rc = -EINVAL ; + +if (!options) +return -EINVAL ; + +while ((p =strsep ((char **)&options,"," )) !=NULL ) { +substring_t args[MAX_OPT_ARGS ]; +int token; + +if (!*p) +continue ; + token =match_token (p, parse_tokens, args); +if (token == parse_res_type) { + *resname =match_strdup (args); + str = p +strlen (p) +1 ; + *remaining_line =kmalloc (strlen (str) +1 ,GFP_KERNEL ); +if (*remaining_line ==NULL ) { +kfree (*resname); + *resname =NULL ; + rc = -ENOMEM ; + }else { +strcpy (*remaining_line, str); + rc =0 ; + } +break ; + } + } +return rc; +} + +static int reset_stats (struct resource_group *rgroup,const char *str) +{ +int rc; +char *resname =NULL , *statstr =NULL ; +struct res_controller *ctlr; +struct res_shares *shares; + + rc =stats_parse (str, &resname, &statstr); +if (rc) +return rc; + + ctlr =get_controller_by_name (resname); +if (!ctlr) { + rc = -EINVAL ; + goto done; + } + shares =get_controller_shares (rgroup, ctlr); +if (shares && ctlr->reset_stats) + rc = ctlr->reset_stats (shares, statstr); +put_controller (ctlr); +done: +kfree (resname); +kfree (statstr); +return rc; +}struct rgroup_attribute {struct configfs_attribute configfs_attr;
@@ -29,6 +123,19 @@ struct rgroup_attribute {int (*store)(struct resource_group *,const char *); }; +struct rgroup_attribute stats_attr = { + .configfs_attr = { + .ca_name ="stats" , + .ca_owner =THIS_MODULE , + .ca_mode =S_IRUGO |S_IWUSR + }, + .show = show_stats, + .store = reset_stats +}; + +static struct configfs_subsystem rgcs_subsys; +static struct config_item_type rgcs_item_type; +struct rgcs_rgroup {char *name;struct resource_group *core;
@@ -173,6 +280,7 @@ static struct configfs_group_operations };static struct configfs_attribute *rgroup_attrs[] = { + &stats_attr.configfs_attr,NULL };
Signature · 4 lines
----------------------------------------------------------------------
Chandra Seetharaman | Be careful what you choose....
- [email protected] | .......you may get it.
----------------------------------------------------------------------PATCH 10/12 Add shares file support to RGCS
In reply to Chandra Seetharaman10/12 - user_interface_shares
Adds attr_store and attr_show support for shares file. --
rgcs.c | 136 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 1 files changed, 136 insertions(+)
kernel/res_group/rgcs.c +136 -0
@@ -21,6 +21,9 @@#include "local.h" #define RES_STRING "res" +#define MIN_SHARES_STRING "min_shares" +#define MAX_SHARES_STRING "max_shares" +#define CHILD_SHARES_DIVISOR_STRING "child_shares_divisor"static ssize_t show_stats (struct resource_group *rgroup,char *buf) {
@@ -117,6 +120,128 @@ done:return rc; } + +enum share_token_t { +MIN_SHARES_TOKEN , +MAX_SHARES_TOKEN , +CHILD_SHARES_DIVISOR_TOKEN , +RESOURCE_TYPE_TOKEN , +ERROR_TOKEN +}; + +/* Token matching for parsing input to this magic file */ +static match_table_t shares_tokens = { + {RESOURCE_TYPE_TOKEN ,RES_STRING "=%s" }, + {MIN_SHARES_TOKEN ,MIN_SHARES_STRING "=%d" }, + {MAX_SHARES_TOKEN ,MAX_SHARES_STRING "=%d" }, + {CHILD_SHARES_DIVISOR_TOKEN ,CHILD_SHARES_DIVISOR_STRING "=%d" }, + {ERROR_TOKEN ,NULL } +}; + +static int shares_parse (const char *options,char **resname, +struct res_shares *shares) +{ +char *p; +int option, rc = -EINVAL ; + + *resname =NULL ; +if (!options) + goto done; +while ((p =strsep ((char **)&options,"," )) !=NULL ) { +substring_t args[MAX_OPT_ARGS ]; +int token; + +if (!*p) +continue ; + token =match_token (p, shares_tokens, args); +switch (token) { +case RESOURCE_TYPE_TOKEN : +if (*resname) + goto done; + *resname =match_strdup (args); +break ; +case MIN_SHARES_TOKEN : +if (match_int (args, &option)) + goto done; + shares->min_shares = option; +break ; +case MAX_SHARES_TOKEN : +if (match_int (args, &option)) + goto done; + shares->max_shares = option; +break ; +case CHILD_SHARES_DIVISOR_TOKEN : +if (match_int (args, &option)) + goto done; + shares->child_shares_divisor = option; +break ; +default : + goto done; + } + } + rc =0 ; +done: +if (rc) { +kfree (*resname); + *resname =NULL ; + } +return rc; +} + +static int set_shares (struct resource_group *rgroup,const char *str) +{ +char *resname =NULL ; +int rc; +struct res_controller *ctlr; +struct res_shares shares = { + .min_shares =SHARE_UNCHANGED , + .max_shares =SHARE_UNCHANGED , + .child_shares_divisor =SHARE_UNCHANGED , + }; + + rc =shares_parse (str, &resname, &shares); +if (!rc) { + ctlr =get_controller_by_name (resname); +if (ctlr) { + rc =set_controller_shares (rgroup, ctlr, &shares); +put_controller (ctlr); + }else + rc = -EINVAL ; +kfree (resname); + } +return rc; +} + +static ssize_t show_shares (struct resource_group *rgroup,char *buf) +{ +int i; +ssize_t j, rc =0 , bufsize =PAGE_SIZE ; +struct res_shares *shares; +struct res_controller *ctlr; + +for (i =0 ; i <MAX_RES_CTLRS ; i++) { + ctlr =get_controller_by_id (i); +if (!ctlr) +continue ; + shares =get_controller_shares (rgroup, ctlr); +if (shares) { +if (bufsize <=0 ) +break ; + j =snprintf (buf, bufsize,"%s=%s,%s=%d,%s=%d,%s=%d\n" , +RES_STRING , ctlr->name, +MIN_SHARES_STRING , shares->min_shares, +MAX_SHARES_STRING , shares->max_shares, +CHILD_SHARES_DIVISOR_STRING , + shares->child_shares_divisor); + rc += j; buf += j; bufsize -= j; + } +put_controller (ctlr); + } +if (i <MAX_RES_CTLRS ) + rc = -ENOSPC ; +return rc; +} +struct rgroup_attribute {struct configfs_attribute configfs_attr;ssize_t (*show)(struct resource_group *,char *);
@@ -133,6 +258,16 @@ struct rgroup_attribute stats_attr = { .store = reset_stats }; +struct rgroup_attribute shares_attr = { + .configfs_attr = { + .ca_name ="shares" , + .ca_owner =THIS_MODULE , + .ca_mode =S_IRUGO |S_IWUSR + }, + .show = show_shares, + .store = set_shares +}; +static struct configfs_subsystem rgcs_subsys;static struct config_item_type rgcs_item_type;
@@ -281,6 +416,7 @@ static struct configfs_group_operationsstatic struct configfs_attribute *rgroup_attrs[] = { &stats_attr.configfs_attr, + &shares_attr.configfs_attr,NULL };
Signature · 4 lines
----------------------------------------------------------------------
Chandra Seetharaman | Be careful what you choose....
- [email protected] | .......you may get it.
----------------------------------------------------------------------PATCH 11/12 Add members file support to RGCS
In reply to Chandra Seetharaman11/12 - user_interface_members
Adds attr_store and attr_show support for members file. --
rgcs.c | 49 +++++++++++++++++++++++++++++++++++++++++++++++++ 1 files changed, 49 insertions(+)
kernel/res_group/rgcs.c +49 -0
@@ -242,6 +242,43 @@ static ssize_t show_shares(struct resourreturn rc; } +/* +* Given a buffer with a pid in it, add the task with that pid to the +* resource group. Ignores entire buffer after the first pid is parsed. +*/ +static int add_member (struct resource_group *rgroup,const char *str) +{ +pid_t pid; + + pid = (pid_t )simple_strtol (str,NULL ,0 ); +if (pid <=0 ) +return -EINVAL ;/* Not a valid pid */ +return set_res_group (pid, rgroup); +} + +/* +* Lists pids of tasks that belong to the given rgroup. +*/ +static ssize_t show_members (struct resource_group *rgroup,char *buf) +{ +ssize_t i, rc =0 , bufsize =PAGE_SIZE ; +struct task_struct *tsk; + +spin_lock (&rgroup->group_lock); +list_for_each_entry (tsk, &rgroup->task_list, member_list) { +if (bufsize <=0 ) { + rc = -ENOSPC ; +break ; + } +if (!tsk->pid)/* Ignore swappers */ +continue ; + i =snprintf (buf, bufsize,"%ld\n" , (long )tsk->pid); + buf += i; rc += i; bufsize -= i; + } +spin_unlock (&rgroup->group_lock); +return rc; +} +struct rgroup_attribute {struct configfs_attribute configfs_attr;ssize_t (*show)(struct resource_group *,char *);
@@ -268,6 +305,17 @@ struct rgroup_attribute shares_attr = { .store = set_shares }; +struct rgroup_attribute members_attr = { + .configfs_attr = { + .ca_name ="members" , + .ca_owner =THIS_MODULE , + .ca_mode =S_IRUGO |S_IWUSR + }, + .show = show_members, + .store = add_member +}; + +static struct configfs_subsystem rgcs_subsys;static struct config_item_type rgcs_item_type;
@@ -417,6 +465,7 @@ static struct configfs_group_operationsstatic struct configfs_attribute *rgroup_attrs[] = { &stats_attr.configfs_attr, &shares_attr.configfs_attr, + &members_attr.configfs_attr,NULL };
Signature · 4 lines
----------------------------------------------------------------------
Chandra Seetharaman | Be careful what you choose....
- [email protected] | .......you may get it.
----------------------------------------------------------------------PATCH 12/12 Documentation for Resource Groups
In reply to Chandra Seetharaman12/12 - res_groups_docs
Documentation describing important Resource Groups components such as resource group, shares, controllers, and the configfs based user interface. --
res_groups_basics | 71 +++++++++++++++++++++++++++++++++++++++++++++++++++++ res_groups_install | 29 +++++++++++++++++++++ res_groups_usage | 53 +++++++++++++++++++++++++++++++++++++++ 3 files changed, 153 insertions(+)
Documentation/res_groups/res_groups_basics new file +71 -0
@@ -0,0 +1,71 @@ +Resource Groups Basics +------------- +A brief review of Resoruce Groups concepts and terminology will help make +installation and testing easier. For more details, please visit +http://ckrm.sf.net. + +Concept: +User defines a resource group, attaches some amount of resources to +the resource group, and associates tasks with that resource group. Tasks +belonging to that resource group will be bound by the amount of resource +that is assigned to that group. + +User interface for the Resource Groups is through configfs. Resource group +is depicted as as a directory. Hierarchy of resource groups can be +created in which children of a resource group share resources allotted to +the parent. Tasks can be attached to any resource group which is at any level. +There is no correlation between parent-child relationship of tasks and +the parent-child relationship of the resource groups they belong to. + +During fork(), resource group is inherited by a task. A privileged user can +reassign a task to any resource group. + +Characteristics of a resource group can be accessed/changed through the +following files under the directory representing the resource group: + +shares: allows changing shares of different resources managed by the + resource group +stats: shows statistics associated with each resources managed by the + resource group +members: allows assignment of tasks to a resource group and shows tasks + that are assigned to a resource group. + +Resource allocations of a resource group is resource group to the amount +of resources available to the resource group's parent. +Resource allocations for a resource group is controlled by the parameters: + +min_shares: Minimum amount shares that can be allocated by a resource group. A + special value of DONT_CARE(-3) means that there is no minimum + shares of a resource is specified. This resource group may not get + any resource if the system is running short on resources. +max_shares: Specifies the maximum amount of resource that is allowed to be + allocated by a resource group. A special value DONT_CARE(-3) mean + that there is no specific limit is specified, this resource group + can get all the resources available. +child_shares_divisor: total guarantee that is allowed among the children of + this resource group. In other words, the sum of "guarantee"s of all + children of this resource group cannot exit this number. + +Any of these parameter can have a special value, UNSUPPORTED(-2) meaning +that the specific controller does not support this parameter. User's +request to change the value will be ignored. + +None of this parameters neither absolute nor have any units associated with +them. These are just numbers that are used to calculate the absolute number +of resource available for a specific resource group. + +Absolute amount of minimum shares available for a resource group is +calculuated as: + + absolute minimum shares = (parent's absolute amount of resource) * + (resource group's min_shares / parent's child_shares_divisor) + +Maximum shares is also calculated in the same way. + +Root resource group is allocated all the resources available in the system. +In other words absolute amount of resource available for the root resource +group is same as that of what is available in the system. + +min_shares and max_shares are related to the parent, these values on +the root resource group have no effect. Similarly, child_shares_divisor +is related to children, and hence it has no effect on leaf resource groups.
Documentation/res_groups/res_groups_install new file +29 -0
@@ -0,0 +1,29 @@ +1. Configure + +Select appropriate configuration options: + + Enable configfs filesystem: + File systems ---> + Pseudo filesystems ---> + <M> Userspace-driven configuration filesystem (EXPERIMENTAL) + + Enable Resource Group components: + General Setup ---> + Resource Groups ---> + [*] Resource Groups + <M> Resource Groups Configfs Subsystem (User API) + [*] Number of Tasks Resource Manager + + +2. Build, boot the kernel + +3. Enable rgcs + + # insmod <patchestree>/fs/configfs/configfs.ko # if compiled as module + # insmod <patchedtree>/kernel/res_config/rgcs.ko # if compiled in as module + # mount -t configfs none /config + + This will create the directory /config/res_groups which is the root of + classes. + +4. Work with class hierarchy as explained in the file res_groups_usage
Documentation/res_groups/res_groups_usage new file +53 -0
53 lines not on this page · raw message
Signature · 4 lines
----------------------------------------------------------------------
Chandra Seetharaman | Be careful what you choose....
- [email protected] | .......you may get it.
----------------------------------------------------------------------PATCH 5/12 Init and clear resource group info in task
In reply to Chandra Seetharaman05/12 - tasksupport_fork_exit_init
Initializes and clears Resource Group specific information in a task at fork() and exit().
Inititalizes Resource Groups (called from start_kernel)
--
include/linux/res_group.h | 7 +++++ init/main.c | 2 + kernel/exit.c | 2 + kernel/fork.c | 2 + kernel/res_group/local.h | 1 kernel/res_group/res_group.c | 11 +++++++++ kernel/res_group/task.c | 52 +++++++++++++++++++++++++++++++++++++++++++ 7 files changed, 77 insertions(+)
kernel/exit.c +2 -0
@@ -35,6 +35,7 @@#include <linux/futex.h> #include <linux/compat.h> #include <linux/pipe_fs_i.h> +#include <linux/res_group.h> #include <asm/uaccess.h> #include <asm/unistd.h>
@@ -731,6 +732,7 @@ static void exit_notify(struct task_strustruct task_struct *t;struct list_head ptrace_dead, *_p, *_n; +clear_task_res_group (tsk);if (signal_pending (tsk) && !(tsk->signal->flags &SIGNAL_GROUP_EXIT ) && !thread_group_empty (tsk)) {/*
kernel/fork.c +2 -0
@@ -44,6 +44,7 @@#include <linux/rmap.h> #include <linux/acct.h> #include <linux/cn_proc.h> +#include <linux/res_group.h> #include <asm/pgtable.h> #include <asm/pgalloc.h>
@@ -1215,6 +1216,7 @@ static task_t *copy_process(unsigned lon total_forks++;spin_unlock (¤t->sighand->siglock);write_unlock_irq (&tasklist_lock); +init_task_res_group (p);proc_fork_connector (p);return p;
init/main.c +2 -0
@@ -47,6 +47,7 @@#include <linux/rmap.h> #include <linux/mempolicy.h> #include <linux/key.h> +#include <linux/res_group.h> #include <asm/io.h> #include <asm/bugs.h>
@@ -541,6 +542,7 @@ asmlinkage void __init start_kernel(voidproc_root_init ();#endif cpuset_init (); +res_group_init ();check_bugs ();
kernel/res_group/res_group.c +11 -0
@@ -231,6 +231,7 @@ int free_res_group(struct resource_groupreturn -EBUSY ; }spin_unlock (&rgroup->group_lock); +move_tasks_to_parent (rgroup);kref_put (&rgroup->ref, release_res_group);return 0 ; }
@@ -392,6 +393,16 @@ void res_group_teardown(void) } } +void res_group_init (void ) +{ +write_lock (&res_group_lock); +list_add_tail (&default_res_group.group_list, &res_groups); +write_unlock (&res_group_lock); +kref_init (&default_res_group.ref); + init_task.res_group = &default_res_group; +init_task_res_group (&init_task); +} +EXPORT_SYMBOL_GPL (register_controller);EXPORT_SYMBOL_GPL (unregister_controller);EXPORT_SYMBOL_GPL (get_controller_by_name);
kernel/res_group/task.c +52 -0
56 lines not on this page · raw message
include/linux/res_group.h +7 -0
@@ -94,5 +94,12 @@ struct resource_group {struct list_head children;/* head of children */ }; +extern void init_task_res_group (struct task_struct *); +extern void clear_task_res_group (struct task_struct *); +extern void res_group_init (void ); +#else /* CONFIG_RES_GROUPS */ +static inline void init_task_res_group (struct task_struct *tsk) { } +static inline void clear_task_res_group (struct task_struct *tsk) { } +static inline void res_group_init (void ) { }#endif /* CONFIG_RES_GROUPS */ #endif /* _LINUX_RES_GROUP_H */
kernel/res_group/local.h +1 -0
@@ -19,3 +19,4 @@ extern void set_shares_to_default(struct struct res_controller *);extern void res_group_teardown (void );extern int set_res_group (pid_t ,struct resource_group *); +extern void move_tasks_to_parent (struct resource_group *);
Signature · 4 lines
----------------------------------------------------------------------
Chandra Seetharaman | Be careful what you choose....
- [email protected] | .......you may get it.
----------------------------------------------------------------------