2006-07-14 11:24:40 +04:00
/*
* taskstats . c - Export per - task statistics to userland
*
* Copyright ( C ) Shailabh Nagar , IBM Corp . 2006
* ( C ) Balbir Singh , IBM Corp . 2006
*
* 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 .
*
* This program is distributed in the hope that it will be useful ,
* but WITHOUT ANY WARRANTY ; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE . See the
* GNU General Public License for more details .
*
*/
# include <linux/kernel.h>
# include <linux/taskstats_kern.h>
2006-10-01 10:28:55 +04:00
# include <linux/tsacct_kern.h>
2006-07-14 11:24:41 +04:00
# include <linux/delayacct.h>
2006-07-14 11:24:47 +04:00
# include <linux/cpumask.h>
# include <linux/percpu.h>
include cleanup: Update gfp.h and slab.h includes to prepare for breaking implicit slab.h inclusion from percpu.h
percpu.h is included by sched.h and module.h and thus ends up being
included when building most .c files. percpu.h includes slab.h which
in turn includes gfp.h making everything defined by the two files
universally available and complicating inclusion dependencies.
percpu.h -> slab.h dependency is about to be removed. Prepare for
this change by updating users of gfp and slab facilities include those
headers directly instead of assuming availability. As this conversion
needs to touch large number of source files, the following script is
used as the basis of conversion.
http://userweb.kernel.org/~tj/misc/slabh-sweep.py
The script does the followings.
* Scan files for gfp and slab usages and update includes such that
only the necessary includes are there. ie. if only gfp is used,
gfp.h, if slab is used, slab.h.
* When the script inserts a new include, it looks at the include
blocks and try to put the new include such that its order conforms
to its surrounding. It's put in the include block which contains
core kernel includes, in the same order that the rest are ordered -
alphabetical, Christmas tree, rev-Xmas-tree or at the end if there
doesn't seem to be any matching order.
* If the script can't find a place to put a new include (mostly
because the file doesn't have fitting include block), it prints out
an error message indicating which .h file needs to be added to the
file.
The conversion was done in the following steps.
1. The initial automatic conversion of all .c files updated slightly
over 4000 files, deleting around 700 includes and adding ~480 gfp.h
and ~3000 slab.h inclusions. The script emitted errors for ~400
files.
2. Each error was manually checked. Some didn't need the inclusion,
some needed manual addition while adding it to implementation .h or
embedding .c file was more appropriate for others. This step added
inclusions to around 150 files.
3. The script was run again and the output was compared to the edits
from #2 to make sure no file was left behind.
4. Several build tests were done and a couple of problems were fixed.
e.g. lib/decompress_*.c used malloc/free() wrappers around slab
APIs requiring slab.h to be added manually.
5. The script was run on all .h files but without automatically
editing them as sprinkling gfp.h and slab.h inclusions around .h
files could easily lead to inclusion dependency hell. Most gfp.h
inclusion directives were ignored as stuff from gfp.h was usually
wildly available and often used in preprocessor macros. Each
slab.h inclusion directive was examined and added manually as
necessary.
6. percpu.h was updated not to include slab.h.
7. Build test were done on the following configurations and failures
were fixed. CONFIG_GCOV_KERNEL was turned off for all tests (as my
distributed build env didn't work with gcov compiles) and a few
more options had to be turned off depending on archs to make things
build (like ipr on powerpc/64 which failed due to missing writeq).
* x86 and x86_64 UP and SMP allmodconfig and a custom test config.
* powerpc and powerpc64 SMP allmodconfig
* sparc and sparc64 SMP allmodconfig
* ia64 SMP allmodconfig
* s390 SMP allmodconfig
* alpha SMP allmodconfig
* um on x86_64 SMP allmodconfig
8. percpu.h modifications were reverted so that it could be applied as
a separate patch and serve as bisection point.
Given the fact that I had only a couple of failures from tests on step
6, I'm fairly confident about the coverage of this conversion patch.
If there is a breakage, it's likely to be something in one of the arch
headers which should be easily discoverable easily on most builds of
the specific arch.
Signed-off-by: Tejun Heo <tj@kernel.org>
Guess-its-ok-by: Christoph Lameter <cl@linux-foundation.org>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Lee Schermerhorn <Lee.Schermerhorn@hp.com>
2010-03-24 11:04:11 +03:00
# include <linux/slab.h>
Add cgroupstats
This patch is inspired by the discussion at
http://lkml.org/lkml/2007/4/11/187 and implements per cgroup statistics
as suggested by Andrew Morton in http://lkml.org/lkml/2007/4/11/263. The
patch is on top of 2.6.21-mm1 with Paul's cgroups v9 patches (forward
ported)
This patch implements per cgroup statistics infrastructure and re-uses
code from the taskstats interface. A new set of cgroup operations are
registered with commands and attributes. It should be very easy to
*extend* per cgroup statistics, by adding members to the cgroupstats
structure.
The current model for cgroupstats is a pull, a push model (to post
statistics on interesting events), should be very easy to add. Currently
user space requests for statistics by passing the cgroup file
descriptor. Statistics about the state of all the tasks in the cgroup
is returned to user space.
TODO's/NOTE:
This patch provides an infrastructure for implementing cgroup statistics.
Based on the needs of each controller, we can incrementally add more statistics,
event based support for notification of statistics, accumulation of taskstats
into cgroup statistics in the future.
Sample output
# ./cgroupstats -C /cgroup/a
sleeping 2, blocked 0, running 1, stopped 0, uninterruptible 0
# ./cgroupstats -C /cgroup/
sleeping 154, blocked 0, running 0, stopped 0, uninterruptible 0
If the approach looks good, I'll enhance and post the user space utility for
the same
Feedback, comments, test results are always welcome!
[akpm@linux-foundation.org: build fix]
Signed-off-by: Balbir Singh <balbir@linux.vnet.ibm.com>
Cc: Paul Menage <menage@google.com>
Cc: Jay Lan <jlan@engr.sgi.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-10-19 10:39:44 +04:00
# include <linux/cgroupstats.h>
# include <linux/cgroup.h>
# include <linux/fs.h>
# include <linux/file.h>
2006-07-14 11:24:40 +04:00
# include <net/genetlink.h>
2011-07-27 03:09:06 +04:00
# include <linux/atomic.h>
2006-07-14 11:24:40 +04:00
2006-07-14 11:24:47 +04:00
/*
* Maximum length of a cpumask that can be specified in
* the TASKSTATS_CMD_ATTR_REGISTER / DEREGISTER_CPUMASK attribute
*/
# define TASKSTATS_CPUMASK_MAXLEN (100+6*NR_CPUS)
2008-07-25 12:48:55 +04:00
static DEFINE_PER_CPU ( __u32 , taskstats_seqnum ) ;
2006-07-14 11:24:40 +04:00
static int family_registered ;
2006-12-07 07:33:20 +03:00
struct kmem_cache * taskstats_cache ;
2006-07-14 11:24:40 +04:00
static struct genl_family family = {
. id = GENL_ID_GENERATE ,
. name = TASKSTATS_GENL_NAME ,
. version = TASKSTATS_GENL_VERSION ,
. maxattr = TASKSTATS_CMD_ATTR_MAX ,
} ;
2010-02-18 11:14:31 +03:00
static const struct nla_policy taskstats_cmd_get_policy [ TASKSTATS_CMD_ATTR_MAX + 1 ] = {
2006-07-14 11:24:40 +04:00
[ TASKSTATS_CMD_ATTR_PID ] = { . type = NLA_U32 } ,
[ TASKSTATS_CMD_ATTR_TGID ] = { . type = NLA_U32 } ,
2006-07-14 11:24:47 +04:00
[ TASKSTATS_CMD_ATTR_REGISTER_CPUMASK ] = { . type = NLA_STRING } ,
[ TASKSTATS_CMD_ATTR_DEREGISTER_CPUMASK ] = { . type = NLA_STRING } , } ;
2010-02-18 11:14:31 +03:00
static const struct nla_policy cgroupstats_cmd_get_policy [ CGROUPSTATS_CMD_ATTR_MAX + 1 ] = {
Add cgroupstats
This patch is inspired by the discussion at
http://lkml.org/lkml/2007/4/11/187 and implements per cgroup statistics
as suggested by Andrew Morton in http://lkml.org/lkml/2007/4/11/263. The
patch is on top of 2.6.21-mm1 with Paul's cgroups v9 patches (forward
ported)
This patch implements per cgroup statistics infrastructure and re-uses
code from the taskstats interface. A new set of cgroup operations are
registered with commands and attributes. It should be very easy to
*extend* per cgroup statistics, by adding members to the cgroupstats
structure.
The current model for cgroupstats is a pull, a push model (to post
statistics on interesting events), should be very easy to add. Currently
user space requests for statistics by passing the cgroup file
descriptor. Statistics about the state of all the tasks in the cgroup
is returned to user space.
TODO's/NOTE:
This patch provides an infrastructure for implementing cgroup statistics.
Based on the needs of each controller, we can incrementally add more statistics,
event based support for notification of statistics, accumulation of taskstats
into cgroup statistics in the future.
Sample output
# ./cgroupstats -C /cgroup/a
sleeping 2, blocked 0, running 1, stopped 0, uninterruptible 0
# ./cgroupstats -C /cgroup/
sleeping 154, blocked 0, running 0, stopped 0, uninterruptible 0
If the approach looks good, I'll enhance and post the user space utility for
the same
Feedback, comments, test results are always welcome!
[akpm@linux-foundation.org: build fix]
Signed-off-by: Balbir Singh <balbir@linux.vnet.ibm.com>
Cc: Paul Menage <menage@google.com>
Cc: Jay Lan <jlan@engr.sgi.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-10-19 10:39:44 +04:00
[ CGROUPSTATS_CMD_ATTR_FD ] = { . type = NLA_U32 } ,
} ;
2006-07-14 11:24:47 +04:00
struct listener {
struct list_head list ;
pid_t pid ;
2006-07-14 11:24:47 +04:00
char valid ;
2006-07-14 11:24:40 +04:00
} ;
2006-07-14 11:24:47 +04:00
struct listener_list {
struct rw_semaphore sem ;
struct list_head list ;
} ;
static DEFINE_PER_CPU ( struct listener_list , listener_array ) ;
enum actions {
REGISTER ,
DEREGISTER ,
CPU_DONT_CARE
} ;
2006-07-14 11:24:40 +04:00
static int prepare_reply ( struct genl_info * info , u8 cmd , struct sk_buff * * skbp ,
2006-12-07 07:36:55 +03:00
size_t size )
2006-07-14 11:24:40 +04:00
{
struct sk_buff * skb ;
void * reply ;
/*
* If new attributes are added , please revisit this allocation
*/
2006-11-15 06:44:52 +03:00
skb = genlmsg_new ( size , GFP_KERNEL ) ;
2006-07-14 11:24:40 +04:00
if ( ! skb )
return - ENOMEM ;
if ( ! info ) {
2010-12-08 19:42:22 +03:00
int seq = this_cpu_inc_return ( taskstats_seqnum ) - 1 ;
2006-07-14 11:24:40 +04:00
2006-11-15 06:46:02 +03:00
reply = genlmsg_put ( skb , 0 , seq , & family , 0 , cmd ) ;
2006-07-14 11:24:40 +04:00
} else
2006-11-15 06:46:02 +03:00
reply = genlmsg_put_reply ( skb , info , & family , 0 , cmd ) ;
2006-07-14 11:24:40 +04:00
if ( reply = = NULL ) {
nlmsg_free ( skb ) ;
return - EINVAL ;
}
* skbp = skb ;
return 0 ;
}
2006-07-14 11:24:47 +04:00
/*
* Send taskstats data in @ skb to listener with nl_pid @ pid
*/
2009-07-10 13:51:34 +04:00
static int send_reply ( struct sk_buff * skb , struct genl_info * info )
2006-07-14 11:24:40 +04:00
{
2007-04-26 06:08:35 +04:00
struct genlmsghdr * genlhdr = nlmsg_data ( nlmsg_hdr ( skb ) ) ;
2006-07-14 11:24:47 +04:00
void * reply = genlmsg_data ( genlhdr ) ;
2006-07-14 11:24:40 +04:00
int rc ;
rc = genlmsg_end ( skb , reply ) ;
if ( rc < 0 ) {
nlmsg_free ( skb ) ;
return rc ;
}
2009-07-10 13:51:34 +04:00
return genlmsg_reply ( skb , info ) ;
2006-07-14 11:24:40 +04:00
}
2006-07-14 11:24:47 +04:00
/*
* Send taskstats data in @ skb to listeners registered for @ cpu ' s exit data
*/
2006-12-07 07:36:51 +03:00
static void send_cpu_listeners ( struct sk_buff * skb ,
struct listener_list * listeners )
2006-07-14 11:24:47 +04:00
{
2007-04-26 06:08:35 +04:00
struct genlmsghdr * genlhdr = nlmsg_data ( nlmsg_hdr ( skb ) ) ;
2006-07-14 11:24:47 +04:00
struct listener * s , * tmp ;
struct sk_buff * skb_next , * skb_cur = skb ;
void * reply = genlmsg_data ( genlhdr ) ;
2006-07-30 14:03:11 +04:00
int rc , delcount = 0 ;
2006-07-14 11:24:47 +04:00
rc = genlmsg_end ( skb , reply ) ;
if ( rc < 0 ) {
nlmsg_free ( skb ) ;
2006-07-30 14:03:11 +04:00
return ;
2006-07-14 11:24:47 +04:00
}
rc = 0 ;
2006-07-14 11:24:47 +04:00
down_read ( & listeners - > sem ) ;
2006-07-30 14:03:11 +04:00
list_for_each_entry ( s , & listeners - > list , list ) {
2006-07-14 11:24:47 +04:00
skb_next = NULL ;
if ( ! list_is_last ( & s - > list , & listeners - > list ) ) {
skb_next = skb_clone ( skb_cur , GFP_KERNEL ) ;
2006-07-30 14:03:11 +04:00
if ( ! skb_next )
2006-07-14 11:24:47 +04:00
break ;
}
2009-07-10 13:51:34 +04:00
rc = genlmsg_unicast ( & init_net , skb_cur , s - > pid ) ;
2006-07-30 14:03:11 +04:00
if ( rc = = - ECONNREFUSED ) {
2006-07-14 11:24:47 +04:00
s - > valid = 0 ;
delcount + + ;
2006-07-14 11:24:47 +04:00
}
skb_cur = skb_next ;
}
2006-07-14 11:24:47 +04:00
up_read ( & listeners - > sem ) ;
2006-07-14 11:24:47 +04:00
2006-07-30 14:03:11 +04:00
if ( skb_cur )
nlmsg_free ( skb_cur ) ;
2006-07-14 11:24:47 +04:00
if ( ! delcount )
2006-07-30 14:03:11 +04:00
return ;
2006-07-14 11:24:47 +04:00
/* Delete invalidated entries */
down_write ( & listeners - > sem ) ;
list_for_each_entry_safe ( s , tmp , & listeners - > list , list ) {
if ( ! s - > valid ) {
list_del ( & s - > list ) ;
kfree ( s ) ;
}
}
up_write ( & listeners - > sem ) ;
2006-07-14 11:24:47 +04:00
}
2010-10-28 02:34:44 +04:00
static void fill_stats ( struct task_struct * tsk , struct taskstats * stats )
2006-07-14 11:24:40 +04:00
{
2006-12-07 07:36:54 +03:00
memset ( stats , 0 , sizeof ( * stats ) ) ;
2006-07-14 11:24:40 +04:00
/*
* Each accounting subsystem adds calls to its functions to
* fill in relevant parts of struct taskstsats as follows
*
2006-07-30 14:03:10 +04:00
* per - task - foo ( stats , tsk ) ;
2006-07-14 11:24:40 +04:00
*/
2006-07-30 14:03:10 +04:00
delayacct_add_tsk ( stats , tsk ) ;
2006-10-01 10:28:55 +04:00
/* fill in basic acct fields */
2006-07-14 11:24:41 +04:00
stats - > version = TASKSTATS_VERSION ;
2007-07-16 10:40:48 +04:00
stats - > nvcsw = tsk - > nvcsw ;
stats - > nivcsw = tsk - > nivcsw ;
2006-10-01 10:28:55 +04:00
bacct_add_tsk ( stats , tsk ) ;
2006-07-14 11:24:41 +04:00
2006-10-01 10:28:58 +04:00
/* fill in extended acct fields */
xacct_add_tsk ( stats , tsk ) ;
2010-10-28 02:34:44 +04:00
}
2006-10-01 10:28:58 +04:00
2010-10-28 02:34:44 +04:00
static int fill_stats_for_pid ( pid_t pid , struct taskstats * stats )
{
struct task_struct * tsk ;
2006-07-14 11:24:40 +04:00
2010-10-28 02:34:44 +04:00
rcu_read_lock ( ) ;
tsk = find_task_by_vpid ( pid ) ;
if ( tsk )
get_task_struct ( tsk ) ;
rcu_read_unlock ( ) ;
if ( ! tsk )
return - ESRCH ;
fill_stats ( tsk , stats ) ;
put_task_struct ( tsk ) ;
return 0 ;
2006-07-14 11:24:40 +04:00
}
2010-10-28 02:34:44 +04:00
static int fill_stats_for_tgid ( pid_t tgid , struct taskstats * stats )
2006-07-14 11:24:40 +04:00
{
2010-10-28 02:34:44 +04:00
struct task_struct * tsk , * first ;
2006-07-14 11:24:44 +04:00
unsigned long flags ;
2006-10-28 21:38:54 +04:00
int rc = - ESRCH ;
2006-07-14 11:24:40 +04:00
2006-07-14 11:24:44 +04:00
/*
* Add additional stats from live tasks except zombie thread group
* leaders who are already counted with the dead tasks
*/
2006-10-28 21:38:54 +04:00
rcu_read_lock ( ) ;
2010-10-28 02:34:44 +04:00
first = find_task_by_vpid ( tgid ) ;
2006-07-14 11:24:44 +04:00
2006-10-28 21:38:54 +04:00
if ( ! first | | ! lock_task_sighand ( first , & flags ) )
goto out ;
2006-07-14 11:24:44 +04:00
2006-10-28 21:38:54 +04:00
if ( first - > signal - > stats )
memcpy ( stats , first - > signal - > stats , sizeof ( * stats ) ) ;
2006-12-07 07:36:54 +03:00
else
memset ( stats , 0 , sizeof ( * stats ) ) ;
2006-10-28 21:38:49 +04:00
2006-10-28 21:38:54 +04:00
tsk = first ;
2006-07-14 11:24:40 +04:00
do {
2006-10-28 21:38:54 +04:00
if ( tsk - > exit_state )
2006-07-14 11:24:44 +04:00
continue ;
2006-07-14 11:24:40 +04:00
/*
2006-07-14 11:24:44 +04:00
* Accounting subsystem can call its functions here to
2006-07-14 11:24:40 +04:00
* fill in relevant parts of struct taskstsats as follows
*
2006-07-14 11:24:44 +04:00
* per - task - foo ( stats , tsk ) ;
2006-07-14 11:24:40 +04:00
*/
2006-07-14 11:24:44 +04:00
delayacct_add_tsk ( stats , tsk ) ;
2006-07-14 11:24:41 +04:00
2007-07-16 10:40:48 +04:00
stats - > nvcsw + = tsk - > nvcsw ;
stats - > nivcsw + = tsk - > nivcsw ;
2006-07-14 11:24:40 +04:00
} while_each_thread ( first , tsk ) ;
2006-07-14 11:24:41 +04:00
2006-10-28 21:38:54 +04:00
unlock_task_sighand ( first , & flags ) ;
rc = 0 ;
out :
rcu_read_unlock ( ) ;
stats - > version = TASKSTATS_VERSION ;
2006-07-14 11:24:40 +04:00
/*
2007-10-20 01:10:43 +04:00
* Accounting subsystems can also add calls here to modify
2006-07-14 11:24:44 +04:00
* fields of taskstats .
2006-07-14 11:24:40 +04:00
*/
2006-10-28 21:38:54 +04:00
return rc ;
2006-07-14 11:24:44 +04:00
}
static void fill_tgid_exit ( struct task_struct * tsk )
{
unsigned long flags ;
2006-10-28 21:38:53 +04:00
spin_lock_irqsave ( & tsk - > sighand - > siglock , flags ) ;
2006-07-14 11:24:44 +04:00
if ( ! tsk - > signal - > stats )
goto ret ;
/*
* Each accounting subsystem calls its functions here to
* accumalate its per - task stats for tsk , into the per - tgid structure
*
* per - task - foo ( tsk - > signal - > stats , tsk ) ;
*/
delayacct_add_tsk ( tsk - > signal - > stats , tsk ) ;
ret :
2006-10-28 21:38:53 +04:00
spin_unlock_irqrestore ( & tsk - > sighand - > siglock , flags ) ;
2006-07-14 11:24:44 +04:00
return ;
2006-07-14 11:24:40 +04:00
}
2009-01-01 02:42:28 +03:00
static int add_del_listener ( pid_t pid , const struct cpumask * mask , int isadd )
2006-07-14 11:24:47 +04:00
{
struct listener_list * listeners ;
2011-06-28 03:18:11 +04:00
struct listener * s , * tmp , * s2 ;
2006-07-14 11:24:47 +04:00
unsigned int cpu ;
2006-07-14 11:24:44 +04:00
2009-01-01 02:42:28 +03:00
if ( ! cpumask_subset ( mask , cpu_possible_mask ) )
2006-07-14 11:24:47 +04:00
return - EINVAL ;
if ( isadd = = REGISTER ) {
2009-01-01 02:42:28 +03:00
for_each_cpu ( cpu , mask ) {
2011-08-04 03:21:04 +04:00
s = kmalloc_node ( sizeof ( struct listener ) ,
GFP_KERNEL , cpu_to_node ( cpu ) ) ;
2006-07-14 11:24:47 +04:00
if ( ! s )
goto cleanup ;
2011-08-04 03:21:04 +04:00
2006-07-14 11:24:47 +04:00
s - > pid = pid ;
2006-07-14 11:24:47 +04:00
s - > valid = 1 ;
2006-07-14 11:24:47 +04:00
listeners = & per_cpu ( listener_array , cpu ) ;
down_write ( & listeners - > sem ) ;
2011-08-04 03:21:04 +04:00
list_for_each_entry ( s2 , & listeners - > list , list ) {
2011-08-04 03:21:05 +04:00
if ( s2 - > pid = = pid & & s2 - > valid )
2011-08-04 03:21:04 +04:00
goto exists ;
2011-06-28 03:18:11 +04:00
}
2006-07-14 11:24:47 +04:00
list_add ( & s - > list , & listeners - > list ) ;
2011-06-28 03:18:11 +04:00
s = NULL ;
2011-08-04 03:21:04 +04:00
exists :
2006-07-14 11:24:47 +04:00
up_write ( & listeners - > sem ) ;
2011-08-04 03:21:04 +04:00
kfree ( s ) ; /* nop if NULL */
2006-07-14 11:24:47 +04:00
}
return 0 ;
}
/* Deregister or cleanup */
cleanup :
2009-01-01 02:42:28 +03:00
for_each_cpu ( cpu , mask ) {
2006-07-14 11:24:47 +04:00
listeners = & per_cpu ( listener_array , cpu ) ;
down_write ( & listeners - > sem ) ;
list_for_each_entry_safe ( s , tmp , & listeners - > list , list ) {
if ( s - > pid = = pid ) {
list_del ( & s - > list ) ;
kfree ( s ) ;
break ;
}
}
up_write ( & listeners - > sem ) ;
}
return 0 ;
}
2009-01-01 02:42:28 +03:00
static int parse ( struct nlattr * na , struct cpumask * mask )
2006-07-14 11:24:47 +04:00
{
char * data ;
int len ;
int ret ;
if ( na = = NULL )
return 1 ;
len = nla_len ( na ) ;
if ( len > TASKSTATS_CPUMASK_MAXLEN )
return - E2BIG ;
if ( len < 1 )
return - EINVAL ;
data = kmalloc ( len , GFP_KERNEL ) ;
if ( ! data )
return - ENOMEM ;
nla_strlcpy ( data , na , len ) ;
2008-12-13 13:50:25 +03:00
ret = cpulist_parse ( data , mask ) ;
2006-07-14 11:24:47 +04:00
kfree ( data ) ;
return ret ;
}
taskstats: use better ifdef for alignment
Commit 4be2c95d ("taskstats: pad taskstats netlink response for aligment
issues on ia64") added a null field to align the taskstats structure but
the discussion centered around ia64. The issue exists on other platforms
with inefficient unaligned access and adding them piecemeal would be an
unmaintainable mess.
This patch uses Dave Miller's suggestion of using a combination of
CONFIG_64BIT && !CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS to determine
whether alignment is needed.
Note that this will cause breakage on those platforms with applications
like iotop which had hard-coded offsets into the packet to access the
taskstats structure.
The message seen on systems without the alignment fixes looks like: kernel
unaligned access to 0xe000023879dca9bc, ip=0xa000000100133d10
The addresses may vary but resolve to locations inside __delayacct_add_tsk.
iotop makes what I'd call unreasonable assumptions about the contents of a
netlink genetlink packet containing generic attributes. They're typed and
have headers that specify value lengths, so the client can (should)
identify and skip the ones the client doesn't understand.
The kernel, as of version 2.6.36, presented a packet like so:
+--------------------------------+
| genlmsghdr - 4 bytes |
+--------------------------------+
| NLA header - 4 bytes | /* Aggregate header */
+-+------------------------------+
| | NLA header - 4 bytes | /* PID header */
| +------------------------------+
| | pid/tgid - 4 bytes |
| +------------------------------+
| | NLA header - 4 bytes | /* stats header */
| + -----------------------------+ <- oops. aligned on 4 byte boundary
| | struct taskstats - 328 bytes |
+-+------------------------------+
The iotop code expects that the kernel will behave as it did then,
assuming that the packet format is set in stone. The format is set in
stone, but the packet offsets are not. There's nothing in the packet
format that guarantees that the packet will be sent in exactly the same
way. The attribute contents are set (or versioned) and the aggregate
contents are set but they can be anywhere in the packet.
The issue here isn't that an unaligned structure gets passed to userspace,
it's that the NLA infrastructure has something of a weakness: The 4 byte
attribute header may force the payload to be unaligned. The taskstats
structure is created at an unaligned location and then 64-bit values are
operated on inside the kernel, so the unaligned access warnings gets
spewed everywhere.
It's possible to use the unaligned access API to operate on the structure
in the kernel but it seems like a wasted effort to work around userspace
code that isn't following the packet format. Any new additions would also
need the be worked around. It's a maintenance nightmare.
The conclusion of the earlier discussion seemed to be "ok fine, if we have
to break it, don't break it on arches that don't have the problem." Dave
pointed out that the unaligned access problem doesn't only exist on ia64,
but also on other 64-bit arches that don't have efficient unaligned access
and it should be fixed there as well. The committed version of the patch
and this addition keep with the conclusion of that discussion not to break
it unnecessarily, which the pid padding and the packet padding fixes did
do. x86_64 and powerpc don't suffer this problem so they shouldn't suffer
the solution. Other 64-bit architectures do and will, though.
Signed-off-by: Jeff Mahoney <jeffm@suse.com>
Reported-by: David S. Miller <davem@davemloft.net>
Acked-by: David S. Miller <davem@davemloft.net>
Cc: Dan Carpenter <error27@gmail.com>
Cc: Balbir Singh <balbir@in.ibm.com>
Cc: Florian Mickler <florian@mickler.org>
Cc: Guillaume Chazarain <guichaz@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2011-01-13 04:00:48 +03:00
# if defined(CONFIG_64BIT) && !defined(CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS)
2010-12-22 04:24:30 +03:00
# define TASKSTATS_NEEDS_PADDING 1
# endif
2006-12-07 07:36:54 +03:00
static struct taskstats * mk_reply ( struct sk_buff * skb , int type , u32 pid )
2006-12-07 07:36:53 +03:00
{
2006-12-07 07:36:54 +03:00
struct nlattr * na , * ret ;
2006-12-07 07:36:53 +03:00
int aggr ;
2006-12-07 07:36:55 +03:00
aggr = ( type = = TASKSTATS_TYPE_PID )
? TASKSTATS_TYPE_AGGR_PID
: TASKSTATS_TYPE_AGGR_TGID ;
2006-12-07 07:36:53 +03:00
2010-12-22 04:24:30 +03:00
/*
* The taskstats structure is internally aligned on 8 byte
* boundaries but the layout of the aggregrate reply , with
* two NLA headers and the pid ( each 4 bytes ) , actually
* force the entire structure to be unaligned . This causes
* the kernel to issue unaligned access warnings on some
* architectures like ia64 . Unfortunately , some software out there
* doesn ' t properly unroll the NLA packet and assumes that the start
* of the taskstats structure will always be 20 bytes from the start
* of the netlink payload . Aligning the start of the taskstats
* structure breaks this software , which we don ' t want . So , for now
* the alignment only happens on architectures that require it
* and those users will have to update to fixed versions of those
* packages . Space is reserved in the packet only when needed .
* This ifdef should be removed in several years e . g . 2012 once
* we can be confident that fixed versions are installed on most
* systems . We add the padding before the aggregate since the
* aggregate is already a defined type .
*/
# ifdef TASKSTATS_NEEDS_PADDING
if ( nla_put ( skb , TASKSTATS_TYPE_NULL , 0 , NULL ) < 0 )
goto err ;
# endif
2006-12-07 07:36:53 +03:00
na = nla_nest_start ( skb , aggr ) ;
2006-12-07 07:36:55 +03:00
if ( ! na )
goto err ;
2010-12-22 04:24:30 +03:00
if ( nla_put ( skb , type , sizeof ( pid ) , & pid ) < 0 )
2006-12-07 07:36:54 +03:00
goto err ;
ret = nla_reserve ( skb , TASKSTATS_TYPE_STATS , sizeof ( struct taskstats ) ) ;
if ( ! ret )
goto err ;
2006-12-07 07:36:53 +03:00
nla_nest_end ( skb , na ) ;
2006-12-07 07:36:54 +03:00
return nla_data ( ret ) ;
err :
return NULL ;
2006-12-07 07:36:53 +03:00
}
Add cgroupstats
This patch is inspired by the discussion at
http://lkml.org/lkml/2007/4/11/187 and implements per cgroup statistics
as suggested by Andrew Morton in http://lkml.org/lkml/2007/4/11/263. The
patch is on top of 2.6.21-mm1 with Paul's cgroups v9 patches (forward
ported)
This patch implements per cgroup statistics infrastructure and re-uses
code from the taskstats interface. A new set of cgroup operations are
registered with commands and attributes. It should be very easy to
*extend* per cgroup statistics, by adding members to the cgroupstats
structure.
The current model for cgroupstats is a pull, a push model (to post
statistics on interesting events), should be very easy to add. Currently
user space requests for statistics by passing the cgroup file
descriptor. Statistics about the state of all the tasks in the cgroup
is returned to user space.
TODO's/NOTE:
This patch provides an infrastructure for implementing cgroup statistics.
Based on the needs of each controller, we can incrementally add more statistics,
event based support for notification of statistics, accumulation of taskstats
into cgroup statistics in the future.
Sample output
# ./cgroupstats -C /cgroup/a
sleeping 2, blocked 0, running 1, stopped 0, uninterruptible 0
# ./cgroupstats -C /cgroup/
sleeping 154, blocked 0, running 0, stopped 0, uninterruptible 0
If the approach looks good, I'll enhance and post the user space utility for
the same
Feedback, comments, test results are always welcome!
[akpm@linux-foundation.org: build fix]
Signed-off-by: Balbir Singh <balbir@linux.vnet.ibm.com>
Cc: Paul Menage <menage@google.com>
Cc: Jay Lan <jlan@engr.sgi.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-10-19 10:39:44 +04:00
static int cgroupstats_user_cmd ( struct sk_buff * skb , struct genl_info * info )
{
int rc = 0 ;
struct sk_buff * rep_skb ;
struct cgroupstats * stats ;
struct nlattr * na ;
size_t size ;
u32 fd ;
struct file * file ;
int fput_needed ;
na = info - > attrs [ CGROUPSTATS_CMD_ATTR_FD ] ;
if ( ! na )
return - EINVAL ;
fd = nla_get_u32 ( info - > attrs [ CGROUPSTATS_CMD_ATTR_FD ] ) ;
file = fget_light ( fd , & fput_needed ) ;
2007-11-15 04:00:37 +03:00
if ( ! file )
return 0 ;
Add cgroupstats
This patch is inspired by the discussion at
http://lkml.org/lkml/2007/4/11/187 and implements per cgroup statistics
as suggested by Andrew Morton in http://lkml.org/lkml/2007/4/11/263. The
patch is on top of 2.6.21-mm1 with Paul's cgroups v9 patches (forward
ported)
This patch implements per cgroup statistics infrastructure and re-uses
code from the taskstats interface. A new set of cgroup operations are
registered with commands and attributes. It should be very easy to
*extend* per cgroup statistics, by adding members to the cgroupstats
structure.
The current model for cgroupstats is a pull, a push model (to post
statistics on interesting events), should be very easy to add. Currently
user space requests for statistics by passing the cgroup file
descriptor. Statistics about the state of all the tasks in the cgroup
is returned to user space.
TODO's/NOTE:
This patch provides an infrastructure for implementing cgroup statistics.
Based on the needs of each controller, we can incrementally add more statistics,
event based support for notification of statistics, accumulation of taskstats
into cgroup statistics in the future.
Sample output
# ./cgroupstats -C /cgroup/a
sleeping 2, blocked 0, running 1, stopped 0, uninterruptible 0
# ./cgroupstats -C /cgroup/
sleeping 154, blocked 0, running 0, stopped 0, uninterruptible 0
If the approach looks good, I'll enhance and post the user space utility for
the same
Feedback, comments, test results are always welcome!
[akpm@linux-foundation.org: build fix]
Signed-off-by: Balbir Singh <balbir@linux.vnet.ibm.com>
Cc: Paul Menage <menage@google.com>
Cc: Jay Lan <jlan@engr.sgi.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-10-19 10:39:44 +04:00
2007-11-15 04:00:37 +03:00
size = nla_total_size ( sizeof ( struct cgroupstats ) ) ;
Add cgroupstats
This patch is inspired by the discussion at
http://lkml.org/lkml/2007/4/11/187 and implements per cgroup statistics
as suggested by Andrew Morton in http://lkml.org/lkml/2007/4/11/263. The
patch is on top of 2.6.21-mm1 with Paul's cgroups v9 patches (forward
ported)
This patch implements per cgroup statistics infrastructure and re-uses
code from the taskstats interface. A new set of cgroup operations are
registered with commands and attributes. It should be very easy to
*extend* per cgroup statistics, by adding members to the cgroupstats
structure.
The current model for cgroupstats is a pull, a push model (to post
statistics on interesting events), should be very easy to add. Currently
user space requests for statistics by passing the cgroup file
descriptor. Statistics about the state of all the tasks in the cgroup
is returned to user space.
TODO's/NOTE:
This patch provides an infrastructure for implementing cgroup statistics.
Based on the needs of each controller, we can incrementally add more statistics,
event based support for notification of statistics, accumulation of taskstats
into cgroup statistics in the future.
Sample output
# ./cgroupstats -C /cgroup/a
sleeping 2, blocked 0, running 1, stopped 0, uninterruptible 0
# ./cgroupstats -C /cgroup/
sleeping 154, blocked 0, running 0, stopped 0, uninterruptible 0
If the approach looks good, I'll enhance and post the user space utility for
the same
Feedback, comments, test results are always welcome!
[akpm@linux-foundation.org: build fix]
Signed-off-by: Balbir Singh <balbir@linux.vnet.ibm.com>
Cc: Paul Menage <menage@google.com>
Cc: Jay Lan <jlan@engr.sgi.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-10-19 10:39:44 +04:00
2007-11-15 04:00:37 +03:00
rc = prepare_reply ( info , CGROUPSTATS_CMD_NEW , & rep_skb ,
size ) ;
if ( rc < 0 )
goto err ;
Add cgroupstats
This patch is inspired by the discussion at
http://lkml.org/lkml/2007/4/11/187 and implements per cgroup statistics
as suggested by Andrew Morton in http://lkml.org/lkml/2007/4/11/263. The
patch is on top of 2.6.21-mm1 with Paul's cgroups v9 patches (forward
ported)
This patch implements per cgroup statistics infrastructure and re-uses
code from the taskstats interface. A new set of cgroup operations are
registered with commands and attributes. It should be very easy to
*extend* per cgroup statistics, by adding members to the cgroupstats
structure.
The current model for cgroupstats is a pull, a push model (to post
statistics on interesting events), should be very easy to add. Currently
user space requests for statistics by passing the cgroup file
descriptor. Statistics about the state of all the tasks in the cgroup
is returned to user space.
TODO's/NOTE:
This patch provides an infrastructure for implementing cgroup statistics.
Based on the needs of each controller, we can incrementally add more statistics,
event based support for notification of statistics, accumulation of taskstats
into cgroup statistics in the future.
Sample output
# ./cgroupstats -C /cgroup/a
sleeping 2, blocked 0, running 1, stopped 0, uninterruptible 0
# ./cgroupstats -C /cgroup/
sleeping 154, blocked 0, running 0, stopped 0, uninterruptible 0
If the approach looks good, I'll enhance and post the user space utility for
the same
Feedback, comments, test results are always welcome!
[akpm@linux-foundation.org: build fix]
Signed-off-by: Balbir Singh <balbir@linux.vnet.ibm.com>
Cc: Paul Menage <menage@google.com>
Cc: Jay Lan <jlan@engr.sgi.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-10-19 10:39:44 +04:00
2007-11-15 04:00:37 +03:00
na = nla_reserve ( rep_skb , CGROUPSTATS_TYPE_CGROUP_STATS ,
sizeof ( struct cgroupstats ) ) ;
stats = nla_data ( na ) ;
memset ( stats , 0 , sizeof ( * stats ) ) ;
Add cgroupstats
This patch is inspired by the discussion at
http://lkml.org/lkml/2007/4/11/187 and implements per cgroup statistics
as suggested by Andrew Morton in http://lkml.org/lkml/2007/4/11/263. The
patch is on top of 2.6.21-mm1 with Paul's cgroups v9 patches (forward
ported)
This patch implements per cgroup statistics infrastructure and re-uses
code from the taskstats interface. A new set of cgroup operations are
registered with commands and attributes. It should be very easy to
*extend* per cgroup statistics, by adding members to the cgroupstats
structure.
The current model for cgroupstats is a pull, a push model (to post
statistics on interesting events), should be very easy to add. Currently
user space requests for statistics by passing the cgroup file
descriptor. Statistics about the state of all the tasks in the cgroup
is returned to user space.
TODO's/NOTE:
This patch provides an infrastructure for implementing cgroup statistics.
Based on the needs of each controller, we can incrementally add more statistics,
event based support for notification of statistics, accumulation of taskstats
into cgroup statistics in the future.
Sample output
# ./cgroupstats -C /cgroup/a
sleeping 2, blocked 0, running 1, stopped 0, uninterruptible 0
# ./cgroupstats -C /cgroup/
sleeping 154, blocked 0, running 0, stopped 0, uninterruptible 0
If the approach looks good, I'll enhance and post the user space utility for
the same
Feedback, comments, test results are always welcome!
[akpm@linux-foundation.org: build fix]
Signed-off-by: Balbir Singh <balbir@linux.vnet.ibm.com>
Cc: Paul Menage <menage@google.com>
Cc: Jay Lan <jlan@engr.sgi.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-10-19 10:39:44 +04:00
2007-11-15 04:00:37 +03:00
rc = cgroupstats_build ( stats , file - > f_dentry ) ;
if ( rc < 0 ) {
nlmsg_free ( rep_skb ) ;
goto err ;
Add cgroupstats
This patch is inspired by the discussion at
http://lkml.org/lkml/2007/4/11/187 and implements per cgroup statistics
as suggested by Andrew Morton in http://lkml.org/lkml/2007/4/11/263. The
patch is on top of 2.6.21-mm1 with Paul's cgroups v9 patches (forward
ported)
This patch implements per cgroup statistics infrastructure and re-uses
code from the taskstats interface. A new set of cgroup operations are
registered with commands and attributes. It should be very easy to
*extend* per cgroup statistics, by adding members to the cgroupstats
structure.
The current model for cgroupstats is a pull, a push model (to post
statistics on interesting events), should be very easy to add. Currently
user space requests for statistics by passing the cgroup file
descriptor. Statistics about the state of all the tasks in the cgroup
is returned to user space.
TODO's/NOTE:
This patch provides an infrastructure for implementing cgroup statistics.
Based on the needs of each controller, we can incrementally add more statistics,
event based support for notification of statistics, accumulation of taskstats
into cgroup statistics in the future.
Sample output
# ./cgroupstats -C /cgroup/a
sleeping 2, blocked 0, running 1, stopped 0, uninterruptible 0
# ./cgroupstats -C /cgroup/
sleeping 154, blocked 0, running 0, stopped 0, uninterruptible 0
If the approach looks good, I'll enhance and post the user space utility for
the same
Feedback, comments, test results are always welcome!
[akpm@linux-foundation.org: build fix]
Signed-off-by: Balbir Singh <balbir@linux.vnet.ibm.com>
Cc: Paul Menage <menage@google.com>
Cc: Jay Lan <jlan@engr.sgi.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-10-19 10:39:44 +04:00
}
2009-07-10 13:51:34 +04:00
rc = send_reply ( rep_skb , info ) ;
2007-11-15 04:00:37 +03:00
Add cgroupstats
This patch is inspired by the discussion at
http://lkml.org/lkml/2007/4/11/187 and implements per cgroup statistics
as suggested by Andrew Morton in http://lkml.org/lkml/2007/4/11/263. The
patch is on top of 2.6.21-mm1 with Paul's cgroups v9 patches (forward
ported)
This patch implements per cgroup statistics infrastructure and re-uses
code from the taskstats interface. A new set of cgroup operations are
registered with commands and attributes. It should be very easy to
*extend* per cgroup statistics, by adding members to the cgroupstats
structure.
The current model for cgroupstats is a pull, a push model (to post
statistics on interesting events), should be very easy to add. Currently
user space requests for statistics by passing the cgroup file
descriptor. Statistics about the state of all the tasks in the cgroup
is returned to user space.
TODO's/NOTE:
This patch provides an infrastructure for implementing cgroup statistics.
Based on the needs of each controller, we can incrementally add more statistics,
event based support for notification of statistics, accumulation of taskstats
into cgroup statistics in the future.
Sample output
# ./cgroupstats -C /cgroup/a
sleeping 2, blocked 0, running 1, stopped 0, uninterruptible 0
# ./cgroupstats -C /cgroup/
sleeping 154, blocked 0, running 0, stopped 0, uninterruptible 0
If the approach looks good, I'll enhance and post the user space utility for
the same
Feedback, comments, test results are always welcome!
[akpm@linux-foundation.org: build fix]
Signed-off-by: Balbir Singh <balbir@linux.vnet.ibm.com>
Cc: Paul Menage <menage@google.com>
Cc: Jay Lan <jlan@engr.sgi.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-10-19 10:39:44 +04:00
err :
2007-11-15 04:00:37 +03:00
fput_light ( file , fput_needed ) ;
Add cgroupstats
This patch is inspired by the discussion at
http://lkml.org/lkml/2007/4/11/187 and implements per cgroup statistics
as suggested by Andrew Morton in http://lkml.org/lkml/2007/4/11/263. The
patch is on top of 2.6.21-mm1 with Paul's cgroups v9 patches (forward
ported)
This patch implements per cgroup statistics infrastructure and re-uses
code from the taskstats interface. A new set of cgroup operations are
registered with commands and attributes. It should be very easy to
*extend* per cgroup statistics, by adding members to the cgroupstats
structure.
The current model for cgroupstats is a pull, a push model (to post
statistics on interesting events), should be very easy to add. Currently
user space requests for statistics by passing the cgroup file
descriptor. Statistics about the state of all the tasks in the cgroup
is returned to user space.
TODO's/NOTE:
This patch provides an infrastructure for implementing cgroup statistics.
Based on the needs of each controller, we can incrementally add more statistics,
event based support for notification of statistics, accumulation of taskstats
into cgroup statistics in the future.
Sample output
# ./cgroupstats -C /cgroup/a
sleeping 2, blocked 0, running 1, stopped 0, uninterruptible 0
# ./cgroupstats -C /cgroup/
sleeping 154, blocked 0, running 0, stopped 0, uninterruptible 0
If the approach looks good, I'll enhance and post the user space utility for
the same
Feedback, comments, test results are always welcome!
[akpm@linux-foundation.org: build fix]
Signed-off-by: Balbir Singh <balbir@linux.vnet.ibm.com>
Cc: Paul Menage <menage@google.com>
Cc: Jay Lan <jlan@engr.sgi.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-10-19 10:39:44 +04:00
return rc ;
}
2010-10-28 02:34:44 +04:00
static int cmd_attr_register_cpumask ( struct genl_info * info )
2006-07-14 11:24:40 +04:00
{
2009-01-01 02:42:28 +03:00
cpumask_var_t mask ;
2010-10-28 02:34:44 +04:00
int rc ;
2009-01-01 02:42:28 +03:00
if ( ! alloc_cpumask_var ( & mask , GFP_KERNEL ) )
return - ENOMEM ;
rc = parse ( info - > attrs [ TASKSTATS_CMD_ATTR_REGISTER_CPUMASK ] , mask ) ;
2006-07-14 11:24:47 +04:00
if ( rc < 0 )
2010-10-28 02:34:44 +04:00
goto out ;
rc = add_del_listener ( info - > snd_pid , mask , REGISTER ) ;
out :
free_cpumask_var ( mask ) ;
return rc ;
}
static int cmd_attr_deregister_cpumask ( struct genl_info * info )
{
cpumask_var_t mask ;
int rc ;
2006-07-14 11:24:47 +04:00
2010-10-28 02:34:44 +04:00
if ( ! alloc_cpumask_var ( & mask , GFP_KERNEL ) )
return - ENOMEM ;
2009-01-01 02:42:28 +03:00
rc = parse ( info - > attrs [ TASKSTATS_CMD_ATTR_DEREGISTER_CPUMASK ] , mask ) ;
2006-07-14 11:24:47 +04:00
if ( rc < 0 )
2010-10-28 02:34:44 +04:00
goto out ;
rc = add_del_listener ( info - > snd_pid , mask , DEREGISTER ) ;
out :
2009-01-01 02:42:28 +03:00
free_cpumask_var ( mask ) ;
2010-10-28 02:34:44 +04:00
return rc ;
}
2010-12-22 04:24:30 +03:00
static size_t taskstats_packet_size ( void )
{
size_t size ;
size = nla_total_size ( sizeof ( u32 ) ) +
nla_total_size ( sizeof ( struct taskstats ) ) + nla_total_size ( 0 ) ;
# ifdef TASKSTATS_NEEDS_PADDING
size + = nla_total_size ( 0 ) ; /* Padding for alignment */
# endif
return size ;
}
2010-10-28 02:34:44 +04:00
static int cmd_attr_pid ( struct genl_info * info )
{
struct taskstats * stats ;
struct sk_buff * rep_skb ;
size_t size ;
u32 pid ;
int rc ;
2006-07-14 11:24:40 +04:00
2010-12-22 04:24:30 +03:00
size = taskstats_packet_size ( ) ;
2006-07-14 11:24:40 +04:00
2006-12-07 07:36:55 +03:00
rc = prepare_reply ( info , TASKSTATS_CMD_NEW , & rep_skb , size ) ;
2006-07-14 11:24:40 +04:00
if ( rc < 0 )
return rc ;
2006-12-07 07:36:54 +03:00
rc = - EINVAL ;
2010-10-28 02:34:44 +04:00
pid = nla_get_u32 ( info - > attrs [ TASKSTATS_CMD_ATTR_PID ] ) ;
stats = mk_reply ( rep_skb , TASKSTATS_TYPE_PID , pid ) ;
if ( ! stats )
2006-07-14 11:24:40 +04:00
goto err ;
2010-10-28 02:34:44 +04:00
rc = fill_stats_for_pid ( pid , stats ) ;
2010-10-28 02:34:44 +04:00
if ( rc < 0 )
goto err ;
2009-07-10 13:51:34 +04:00
return send_reply ( rep_skb , info ) ;
2006-07-14 11:24:40 +04:00
err :
nlmsg_free ( rep_skb ) ;
return rc ;
}
2010-10-28 02:34:44 +04:00
static int cmd_attr_tgid ( struct genl_info * info )
{
struct taskstats * stats ;
struct sk_buff * rep_skb ;
size_t size ;
u32 tgid ;
int rc ;
2010-12-22 04:24:30 +03:00
size = taskstats_packet_size ( ) ;
2010-10-28 02:34:44 +04:00
rc = prepare_reply ( info , TASKSTATS_CMD_NEW , & rep_skb , size ) ;
if ( rc < 0 )
return rc ;
rc = - EINVAL ;
tgid = nla_get_u32 ( info - > attrs [ TASKSTATS_CMD_ATTR_TGID ] ) ;
stats = mk_reply ( rep_skb , TASKSTATS_TYPE_TGID , tgid ) ;
if ( ! stats )
goto err ;
2010-10-28 02:34:44 +04:00
rc = fill_stats_for_tgid ( tgid , stats ) ;
2010-10-28 02:34:44 +04:00
if ( rc < 0 )
goto err ;
return send_reply ( rep_skb , info ) ;
err :
nlmsg_free ( rep_skb ) ;
return rc ;
}
static int taskstats_user_cmd ( struct sk_buff * skb , struct genl_info * info )
{
if ( info - > attrs [ TASKSTATS_CMD_ATTR_REGISTER_CPUMASK ] )
return cmd_attr_register_cpumask ( info ) ;
else if ( info - > attrs [ TASKSTATS_CMD_ATTR_DEREGISTER_CPUMASK ] )
return cmd_attr_deregister_cpumask ( info ) ;
else if ( info - > attrs [ TASKSTATS_CMD_ATTR_PID ] )
return cmd_attr_pid ( info ) ;
else if ( info - > attrs [ TASKSTATS_CMD_ATTR_TGID ] )
return cmd_attr_tgid ( info ) ;
else
return - EINVAL ;
}
2006-12-07 07:36:52 +03:00
static struct taskstats * taskstats_tgid_alloc ( struct task_struct * tsk )
{
struct signal_struct * sig = tsk - > signal ;
struct taskstats * stats ;
if ( sig - > stats | | thread_group_empty ( tsk ) )
goto ret ;
/* No problem if kmem_cache_zalloc() fails */
stats = kmem_cache_zalloc ( taskstats_cache , GFP_KERNEL ) ;
spin_lock_irq ( & tsk - > sighand - > siglock ) ;
if ( ! sig - > stats ) {
sig - > stats = stats ;
stats = NULL ;
}
spin_unlock_irq ( & tsk - > sighand - > siglock ) ;
if ( stats )
kmem_cache_free ( taskstats_cache , stats ) ;
ret :
return sig - > stats ;
}
2006-07-14 11:24:40 +04:00
/* Send pid data out on exit */
2006-12-07 07:36:51 +03:00
void taskstats_exit ( struct task_struct * tsk , int group_dead )
2006-07-14 11:24:40 +04:00
{
int rc ;
2006-12-07 07:36:51 +03:00
struct listener_list * listeners ;
2006-12-07 07:36:54 +03:00
struct taskstats * stats ;
2006-07-14 11:24:40 +04:00
struct sk_buff * rep_skb ;
size_t size ;
int is_thread_group ;
2006-10-31 09:07:15 +03:00
if ( ! family_registered )
2006-07-14 11:24:40 +04:00
return ;
/*
* Size includes space for nested attributes
*/
2010-12-22 04:24:30 +03:00
size = taskstats_packet_size ( ) ;
2006-07-14 11:24:40 +04:00
2006-12-07 07:36:52 +03:00
is_thread_group = ! ! taskstats_tgid_alloc ( tsk ) ;
2006-10-31 09:07:15 +03:00
if ( is_thread_group ) {
/* PID + STATS + TGID + STATS */
size = 2 * size ;
/* fill the tsk->signal->stats structure */
fill_tgid_exit ( tsk ) ;
}
2010-12-08 19:42:22 +03:00
listeners = __this_cpu_ptr ( & listener_array ) ;
2006-12-07 07:36:51 +03:00
if ( list_empty ( & listeners - > list ) )
return ;
2006-12-07 07:36:55 +03:00
rc = prepare_reply ( NULL , TASKSTATS_CMD_NEW , & rep_skb , size ) ;
2006-07-14 11:24:40 +04:00
if ( rc < 0 )
2006-12-07 07:36:54 +03:00
return ;
2006-07-14 11:24:40 +04:00
2006-12-07 07:36:54 +03:00
stats = mk_reply ( rep_skb , TASKSTATS_TYPE_PID , tsk - > pid ) ;
if ( ! stats )
2006-12-07 07:36:55 +03:00
goto err ;
2006-07-14 11:24:40 +04:00
2010-10-28 02:34:44 +04:00
fill_stats ( tsk , stats ) ;
2006-07-14 11:24:40 +04:00
/*
2006-07-14 11:24:44 +04:00
* Doesn ' t matter if tsk is the leader or the last group member leaving
2006-07-14 11:24:40 +04:00
*/
2006-12-07 07:36:53 +03:00
if ( ! is_thread_group | | ! group_dead )
2006-07-14 11:24:44 +04:00
goto send ;
2006-07-14 11:24:40 +04:00
2006-12-07 07:36:54 +03:00
stats = mk_reply ( rep_skb , TASKSTATS_TYPE_TGID , tsk - > tgid ) ;
if ( ! stats )
2006-12-07 07:36:55 +03:00
goto err ;
2006-12-07 07:36:54 +03:00
memcpy ( stats , tsk - > signal - > stats , sizeof ( * stats ) ) ;
2006-07-14 11:24:40 +04:00
2006-07-14 11:24:44 +04:00
send :
2006-12-07 07:36:51 +03:00
send_cpu_listeners ( rep_skb , listeners ) ;
2006-07-14 11:24:44 +04:00
return ;
2006-12-07 07:36:55 +03:00
err :
2006-07-14 11:24:40 +04:00
nlmsg_free ( rep_skb ) ;
}
static struct genl_ops taskstats_ops = {
. cmd = TASKSTATS_CMD_GET ,
2006-07-14 11:24:47 +04:00
. doit = taskstats_user_cmd ,
2006-07-14 11:24:40 +04:00
. policy = taskstats_cmd_get_policy ,
2011-09-20 04:04:37 +04:00
. flags = GENL_ADMIN_PERM ,
2006-07-14 11:24:40 +04:00
} ;
Add cgroupstats
This patch is inspired by the discussion at
http://lkml.org/lkml/2007/4/11/187 and implements per cgroup statistics
as suggested by Andrew Morton in http://lkml.org/lkml/2007/4/11/263. The
patch is on top of 2.6.21-mm1 with Paul's cgroups v9 patches (forward
ported)
This patch implements per cgroup statistics infrastructure and re-uses
code from the taskstats interface. A new set of cgroup operations are
registered with commands and attributes. It should be very easy to
*extend* per cgroup statistics, by adding members to the cgroupstats
structure.
The current model for cgroupstats is a pull, a push model (to post
statistics on interesting events), should be very easy to add. Currently
user space requests for statistics by passing the cgroup file
descriptor. Statistics about the state of all the tasks in the cgroup
is returned to user space.
TODO's/NOTE:
This patch provides an infrastructure for implementing cgroup statistics.
Based on the needs of each controller, we can incrementally add more statistics,
event based support for notification of statistics, accumulation of taskstats
into cgroup statistics in the future.
Sample output
# ./cgroupstats -C /cgroup/a
sleeping 2, blocked 0, running 1, stopped 0, uninterruptible 0
# ./cgroupstats -C /cgroup/
sleeping 154, blocked 0, running 0, stopped 0, uninterruptible 0
If the approach looks good, I'll enhance and post the user space utility for
the same
Feedback, comments, test results are always welcome!
[akpm@linux-foundation.org: build fix]
Signed-off-by: Balbir Singh <balbir@linux.vnet.ibm.com>
Cc: Paul Menage <menage@google.com>
Cc: Jay Lan <jlan@engr.sgi.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-10-19 10:39:44 +04:00
static struct genl_ops cgroupstats_ops = {
. cmd = CGROUPSTATS_CMD_GET ,
. doit = cgroupstats_user_cmd ,
. policy = cgroupstats_cmd_get_policy ,
} ;
2006-07-14 11:24:40 +04:00
/* Needed early in initialization */
void __init taskstats_init_early ( void )
{
2006-07-14 11:24:47 +04:00
unsigned int i ;
2007-05-07 01:49:57 +04:00
taskstats_cache = KMEM_CACHE ( taskstats , SLAB_PANIC ) ;
2006-07-14 11:24:47 +04:00
for_each_possible_cpu ( i ) {
INIT_LIST_HEAD ( & ( per_cpu ( listener_array , i ) . list ) ) ;
init_rwsem ( & ( per_cpu ( listener_array , i ) . sem ) ) ;
}
2006-07-14 11:24:40 +04:00
}
static int __init taskstats_init ( void )
{
int rc ;
rc = genl_register_family ( & family ) ;
if ( rc )
return rc ;
rc = genl_register_ops ( & family , & taskstats_ops ) ;
if ( rc < 0 )
goto err ;
Add cgroupstats
This patch is inspired by the discussion at
http://lkml.org/lkml/2007/4/11/187 and implements per cgroup statistics
as suggested by Andrew Morton in http://lkml.org/lkml/2007/4/11/263. The
patch is on top of 2.6.21-mm1 with Paul's cgroups v9 patches (forward
ported)
This patch implements per cgroup statistics infrastructure and re-uses
code from the taskstats interface. A new set of cgroup operations are
registered with commands and attributes. It should be very easy to
*extend* per cgroup statistics, by adding members to the cgroupstats
structure.
The current model for cgroupstats is a pull, a push model (to post
statistics on interesting events), should be very easy to add. Currently
user space requests for statistics by passing the cgroup file
descriptor. Statistics about the state of all the tasks in the cgroup
is returned to user space.
TODO's/NOTE:
This patch provides an infrastructure for implementing cgroup statistics.
Based on the needs of each controller, we can incrementally add more statistics,
event based support for notification of statistics, accumulation of taskstats
into cgroup statistics in the future.
Sample output
# ./cgroupstats -C /cgroup/a
sleeping 2, blocked 0, running 1, stopped 0, uninterruptible 0
# ./cgroupstats -C /cgroup/
sleeping 154, blocked 0, running 0, stopped 0, uninterruptible 0
If the approach looks good, I'll enhance and post the user space utility for
the same
Feedback, comments, test results are always welcome!
[akpm@linux-foundation.org: build fix]
Signed-off-by: Balbir Singh <balbir@linux.vnet.ibm.com>
Cc: Paul Menage <menage@google.com>
Cc: Jay Lan <jlan@engr.sgi.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-10-19 10:39:44 +04:00
rc = genl_register_ops ( & family , & cgroupstats_ops ) ;
if ( rc < 0 )
goto err_cgroup_ops ;
2006-07-14 11:24:40 +04:00
family_registered = 1 ;
2011-03-24 02:43:27 +03:00
pr_info ( " registered taskstats version %d \n " , TASKSTATS_GENL_VERSION ) ;
2006-07-14 11:24:40 +04:00
return 0 ;
Add cgroupstats
This patch is inspired by the discussion at
http://lkml.org/lkml/2007/4/11/187 and implements per cgroup statistics
as suggested by Andrew Morton in http://lkml.org/lkml/2007/4/11/263. The
patch is on top of 2.6.21-mm1 with Paul's cgroups v9 patches (forward
ported)
This patch implements per cgroup statistics infrastructure and re-uses
code from the taskstats interface. A new set of cgroup operations are
registered with commands and attributes. It should be very easy to
*extend* per cgroup statistics, by adding members to the cgroupstats
structure.
The current model for cgroupstats is a pull, a push model (to post
statistics on interesting events), should be very easy to add. Currently
user space requests for statistics by passing the cgroup file
descriptor. Statistics about the state of all the tasks in the cgroup
is returned to user space.
TODO's/NOTE:
This patch provides an infrastructure for implementing cgroup statistics.
Based on the needs of each controller, we can incrementally add more statistics,
event based support for notification of statistics, accumulation of taskstats
into cgroup statistics in the future.
Sample output
# ./cgroupstats -C /cgroup/a
sleeping 2, blocked 0, running 1, stopped 0, uninterruptible 0
# ./cgroupstats -C /cgroup/
sleeping 154, blocked 0, running 0, stopped 0, uninterruptible 0
If the approach looks good, I'll enhance and post the user space utility for
the same
Feedback, comments, test results are always welcome!
[akpm@linux-foundation.org: build fix]
Signed-off-by: Balbir Singh <balbir@linux.vnet.ibm.com>
Cc: Paul Menage <menage@google.com>
Cc: Jay Lan <jlan@engr.sgi.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-10-19 10:39:44 +04:00
err_cgroup_ops :
genl_unregister_ops ( & family , & taskstats_ops ) ;
2006-07-14 11:24:40 +04:00
err :
genl_unregister_family ( & family ) ;
return rc ;
}
/*
* late initcall ensures initialization of statistics collection
* mechanisms precedes initialization of the taskstats interface
*/
late_initcall ( taskstats_init ) ;