linux/kernel/acct.c

655 lines
17 KiB
C
Raw Normal View History

/*
* linux/kernel/acct.c
*
* BSD Process Accounting for Linux
*
* Author: Marco van Wieringen <mvw@planets.elm.net>
*
* Some code based on ideas and code from:
* Thomas K. Dyas <tdyas@eden.rutgers.edu>
*
* This file implements BSD-style process accounting. Whenever any
* process exits, an accounting record of type "struct acct" is
* written to the file specified with the acct() system call. It is
* up to user-level programs to do useful things with the accounting
* log. The kernel just provides the raw accounting information.
*
* (C) Copyright 1995 - 1997 Marco van Wieringen - ELM Consultancy B.V.
*
* Plugged two leaks. 1) It didn't return acct_file into the free_filps if
* the file happened to be read-only. 2) If the accounting was suspended
* due to the lack of space it happily allowed to reopen it and completely
* lost the old acct_file. 3/10/98, Al Viro.
*
* Now we silently close acct_file on attempt to reopen. Cleaned sys_acct().
* XTerms and EMACS are manifestations of pure evil. 21/10/98, AV.
*
* Fixed a nasty interaction with with sys_umount(). If the accointing
* was suspeneded we failed to stop it on umount(). Messy.
* Another one: remount to readonly didn't stop accounting.
* Question: what should we do if we have CAP_SYS_ADMIN but not
* CAP_SYS_PACCT? Current code does the following: umount returns -EBUSY
* unless we are messing with the root. In that case we are getting a
* real mess with do_remount_sb(). 9/11/98, AV.
*
* Fixed a bunch of races (and pair of leaks). Probably not the best way,
* but this one obviously doesn't introduce deadlocks. Later. BTW, found
* one race (and leak) in BSD implementation.
* OK, that's better. ANOTHER race and leak in BSD variant. There always
* is one more bug... 10/11/98, AV.
*
* Oh, fsck... Oopsable SMP race in do_process_acct() - we must hold
* ->mmap_sem to walk the vma list of current->mm. Nasty, since it leaks
* a struct file opened for write. Fixed. 2/6/2000, AV.
*/
#include <linux/mm.h>
#include <linux/slab.h>
#include <linux/acct.h>
#include <linux/capability.h>
#include <linux/file.h>
#include <linux/tty.h>
#include <linux/security.h>
#include <linux/vfs.h>
#include <linux/jiffies.h>
#include <linux/times.h>
#include <linux/syscalls.h>
[PATCH] saner handling of auto_acct_off() and DQUOT_OFF() in umount The way we currently deal with quota and process accounting that might keep vfsmount busy at umount time is inherently broken; we try to turn them off just in case (not quite correctly, at that) and a) pray umount doesn't fail (otherwise they'll stay turned off) b) pray nobody doesn anything funny just as we turn quota off Moreover, LSM provides hooks for doing the same sort of broken logics. The proper way to deal with that is to introduce the second kind of reference to vfsmount. Semantics: - when the last normal reference is dropped, all special ones are converted to normal ones and if there had been any, cleanup is done. - normal reference can be cloned into a special one - special reference can be converted to normal one; that's a no-op if we'd already passed the point of no return (i.e. mntput() had converted special references to normal and started cleanup). The way it works: e.g. starting process accounting converts the vfsmount reference pinned by the opened file into special one and turns it back to normal when it gets shut down; acct_auto_close() is done when no normal references are left. That way it does *not* obstruct umount(2) and it silently gets turned off when the last normal reference to vfsmount is gone. Which is exactly what we want... The same should be done by LSM module that holds some internal references to vfsmount and wants to shut them down on umount - it should make them special and security_sb_umount_close() will be called exactly when the last normal reference to vfsmount is gone. quota handling is even simpler - we don't use normal file IO anymore, so there's no need to hold vfsmounts at all. DQUOT_OFF() is done from deactivate_super(), where it really belongs. Signed-off-by: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-11-07 22:13:39 +00:00
#include <linux/mount.h>
#include <asm/uaccess.h>
#include <asm/div64.h>
#include <linux/blkdev.h> /* sector_div */
#include <linux/pid_namespace.h>
/*
* These constants control the amount of freespace that suspend and
* resume the process accounting system, and the time delay between
* each check.
* Turned into sysctl-controllable parameters. AV, 12/11/98
*/
int acct_parm[3] = {4, 2, 30};
#define RESUME (acct_parm[0]) /* >foo% free space - resume */
#define SUSPEND (acct_parm[1]) /* <foo% free space - suspend */
#define ACCT_TIMEOUT (acct_parm[2]) /* foo second timeout between checks */
/*
* External references and all of the globals.
*/
static void do_acct_process(struct bsd_acct_struct *acct,
struct pid_namespace *ns, struct file *);
/*
* This structure is used so that all the data protected by lock
* can be placed in the same cache line as the lock. This primes
* the cache line to have the data after getting the lock.
*/
bsdacct: rename acct_gbls to bsd_acct_struct After I fixed access to task->tgid in kernel/acct.c, Oleg pointed out some bad side effects with this accounting vs pid namespaces interaction. I.e. when some task in pid namespace sets this accounting up, this blocks all the others from doing the same. Restricting this to init namespace only could help, but didn't look a graceful solution. So here is the approach to make this accounting work with pid namespaces properly. The idea is simple - when a task dies it accounts itself in each namespace it is visible from and which set the accounting up. For example here are the commands run and the output of lastcomm from init and sub namespaces: init_ns# accton pacct sub_ns# accton pacct (this is a different file - sub ns is run in a chroot-ed environment) init_ns# cat /dev/null sub_ns# ls /dev/null init_ns# accton sub_ns# accton sub_ns# lastcomm -f pacct ls 0 [136,0] 0.00 secs Thu May 15 10:30 accton 0 [136,0] 0.00 secs Thu May 15 10:30 init_ns# lastcomm -f pacct accton root pts/0 0.00 secs Thu May 15 14:30 << got from sub cat root pts/1 0.00 secs Thu May 15 14:30 ls root pts/0 0.00 secs Thu May 15 14:30 << got from sub accton root pts/1 0.00 secs Thu May 15 14:30 That was the summary, the details are in patches. This patch: It will be visible in pid_namespace.h file, so fix its name to look better outside the acct.c file. Signed-off-by: Pavel Emelyanov <xemul@openvz.org> Cc: Balbir Singh <balbir@in.ibm.com> Cc: "Eric W. Biederman" <ebiederm@xmission.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-07-25 08:48:42 +00:00
struct bsd_acct_struct {
int active;
unsigned long needcheck;
struct file *file;
struct pid_namespace *ns;
struct list_head list;
};
static DEFINE_SPINLOCK(acct_lock);
static LIST_HEAD(acct_list);
/*
* Check the amount of free space and suspend/resume accordingly.
*/
static int check_free_space(struct bsd_acct_struct *acct, struct file *file)
{
struct kstatfs sbuf;
int res;
int act;
u64 resume;
u64 suspend;
spin_lock(&acct_lock);
res = acct->active;
if (!file || time_is_before_jiffies(acct->needcheck))
goto out;
spin_unlock(&acct_lock);
/* May block */
if (vfs_statfs(&file->f_path, &sbuf))
return res;
suspend = sbuf.f_blocks * SUSPEND;
resume = sbuf.f_blocks * RESUME;
do_div(suspend, 100);
do_div(resume, 100);
if (sbuf.f_bavail <= suspend)
act = -1;
else if (sbuf.f_bavail >= resume)
act = 1;
else
act = 0;
/*
* If some joker switched acct->file under us we'ld better be
* silent and _not_ touch anything.
*/
spin_lock(&acct_lock);
if (file != acct->file) {
if (act)
res = act>0;
goto out;
}
if (acct->active) {
if (act < 0) {
acct->active = 0;
printk(KERN_INFO "Process accounting paused\n");
}
} else {
if (act > 0) {
acct->active = 1;
printk(KERN_INFO "Process accounting resumed\n");
}
}
acct->needcheck = jiffies + ACCT_TIMEOUT*HZ;
res = acct->active;
out:
spin_unlock(&acct_lock);
return res;
}
/*
* Close the old accounting file (if currently open) and then replace
* it with file (if non-NULL).
*
* NOTE: acct_lock MUST be held on entry and exit.
*/
static void acct_file_reopen(struct bsd_acct_struct *acct, struct file *file,
struct pid_namespace *ns)
{
struct file *old_acct = NULL;
struct pid_namespace *old_ns = NULL;
if (acct->file) {
old_acct = acct->file;
old_ns = acct->ns;
acct->active = 0;
acct->file = NULL;
acct->ns = NULL;
list_del(&acct->list);
}
if (file) {
acct->file = file;
acct->ns = ns;
acct->needcheck = jiffies + ACCT_TIMEOUT*HZ;
acct->active = 1;
list_add(&acct->list, &acct_list);
}
if (old_acct) {
mnt_unpin(old_acct->f_path.mnt);
spin_unlock(&acct_lock);
do_acct_process(acct, old_ns, old_acct);
filp_close(old_acct, NULL);
spin_lock(&acct_lock);
}
}
static int acct_on(struct filename *pathname)
[PATCH] saner handling of auto_acct_off() and DQUOT_OFF() in umount The way we currently deal with quota and process accounting that might keep vfsmount busy at umount time is inherently broken; we try to turn them off just in case (not quite correctly, at that) and a) pray umount doesn't fail (otherwise they'll stay turned off) b) pray nobody doesn anything funny just as we turn quota off Moreover, LSM provides hooks for doing the same sort of broken logics. The proper way to deal with that is to introduce the second kind of reference to vfsmount. Semantics: - when the last normal reference is dropped, all special ones are converted to normal ones and if there had been any, cleanup is done. - normal reference can be cloned into a special one - special reference can be converted to normal one; that's a no-op if we'd already passed the point of no return (i.e. mntput() had converted special references to normal and started cleanup). The way it works: e.g. starting process accounting converts the vfsmount reference pinned by the opened file into special one and turns it back to normal when it gets shut down; acct_auto_close() is done when no normal references are left. That way it does *not* obstruct umount(2) and it silently gets turned off when the last normal reference to vfsmount is gone. Which is exactly what we want... The same should be done by LSM module that holds some internal references to vfsmount and wants to shut them down on umount - it should make them special and security_sb_umount_close() will be called exactly when the last normal reference to vfsmount is gone. quota handling is even simpler - we don't use normal file IO anymore, so there's no need to hold vfsmounts at all. DQUOT_OFF() is done from deactivate_super(), where it really belongs. Signed-off-by: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-11-07 22:13:39 +00:00
{
struct file *file;
struct vfsmount *mnt;
struct pid_namespace *ns;
struct bsd_acct_struct *acct = NULL;
[PATCH] saner handling of auto_acct_off() and DQUOT_OFF() in umount The way we currently deal with quota and process accounting that might keep vfsmount busy at umount time is inherently broken; we try to turn them off just in case (not quite correctly, at that) and a) pray umount doesn't fail (otherwise they'll stay turned off) b) pray nobody doesn anything funny just as we turn quota off Moreover, LSM provides hooks for doing the same sort of broken logics. The proper way to deal with that is to introduce the second kind of reference to vfsmount. Semantics: - when the last normal reference is dropped, all special ones are converted to normal ones and if there had been any, cleanup is done. - normal reference can be cloned into a special one - special reference can be converted to normal one; that's a no-op if we'd already passed the point of no return (i.e. mntput() had converted special references to normal and started cleanup). The way it works: e.g. starting process accounting converts the vfsmount reference pinned by the opened file into special one and turns it back to normal when it gets shut down; acct_auto_close() is done when no normal references are left. That way it does *not* obstruct umount(2) and it silently gets turned off when the last normal reference to vfsmount is gone. Which is exactly what we want... The same should be done by LSM module that holds some internal references to vfsmount and wants to shut them down on umount - it should make them special and security_sb_umount_close() will be called exactly when the last normal reference to vfsmount is gone. quota handling is even simpler - we don't use normal file IO anymore, so there's no need to hold vfsmounts at all. DQUOT_OFF() is done from deactivate_super(), where it really belongs. Signed-off-by: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-11-07 22:13:39 +00:00
/* Difference from BSD - they don't do O_APPEND */
file = file_open_name(pathname, O_WRONLY|O_APPEND|O_LARGEFILE, 0);
[PATCH] saner handling of auto_acct_off() and DQUOT_OFF() in umount The way we currently deal with quota and process accounting that might keep vfsmount busy at umount time is inherently broken; we try to turn them off just in case (not quite correctly, at that) and a) pray umount doesn't fail (otherwise they'll stay turned off) b) pray nobody doesn anything funny just as we turn quota off Moreover, LSM provides hooks for doing the same sort of broken logics. The proper way to deal with that is to introduce the second kind of reference to vfsmount. Semantics: - when the last normal reference is dropped, all special ones are converted to normal ones and if there had been any, cleanup is done. - normal reference can be cloned into a special one - special reference can be converted to normal one; that's a no-op if we'd already passed the point of no return (i.e. mntput() had converted special references to normal and started cleanup). The way it works: e.g. starting process accounting converts the vfsmount reference pinned by the opened file into special one and turns it back to normal when it gets shut down; acct_auto_close() is done when no normal references are left. That way it does *not* obstruct umount(2) and it silently gets turned off when the last normal reference to vfsmount is gone. Which is exactly what we want... The same should be done by LSM module that holds some internal references to vfsmount and wants to shut them down on umount - it should make them special and security_sb_umount_close() will be called exactly when the last normal reference to vfsmount is gone. quota handling is even simpler - we don't use normal file IO anymore, so there's no need to hold vfsmounts at all. DQUOT_OFF() is done from deactivate_super(), where it really belongs. Signed-off-by: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-11-07 22:13:39 +00:00
if (IS_ERR(file))
return PTR_ERR(file);
if (!S_ISREG(file_inode(file)->i_mode)) {
[PATCH] saner handling of auto_acct_off() and DQUOT_OFF() in umount The way we currently deal with quota and process accounting that might keep vfsmount busy at umount time is inherently broken; we try to turn them off just in case (not quite correctly, at that) and a) pray umount doesn't fail (otherwise they'll stay turned off) b) pray nobody doesn anything funny just as we turn quota off Moreover, LSM provides hooks for doing the same sort of broken logics. The proper way to deal with that is to introduce the second kind of reference to vfsmount. Semantics: - when the last normal reference is dropped, all special ones are converted to normal ones and if there had been any, cleanup is done. - normal reference can be cloned into a special one - special reference can be converted to normal one; that's a no-op if we'd already passed the point of no return (i.e. mntput() had converted special references to normal and started cleanup). The way it works: e.g. starting process accounting converts the vfsmount reference pinned by the opened file into special one and turns it back to normal when it gets shut down; acct_auto_close() is done when no normal references are left. That way it does *not* obstruct umount(2) and it silently gets turned off when the last normal reference to vfsmount is gone. Which is exactly what we want... The same should be done by LSM module that holds some internal references to vfsmount and wants to shut them down on umount - it should make them special and security_sb_umount_close() will be called exactly when the last normal reference to vfsmount is gone. quota handling is even simpler - we don't use normal file IO anymore, so there's no need to hold vfsmounts at all. DQUOT_OFF() is done from deactivate_super(), where it really belongs. Signed-off-by: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-11-07 22:13:39 +00:00
filp_close(file, NULL);
return -EACCES;
}
if (!file->f_op->write) {
filp_close(file, NULL);
return -EIO;
}
ns = task_active_pid_ns(current);
if (ns->bacct == NULL) {
acct = kzalloc(sizeof(struct bsd_acct_struct), GFP_KERNEL);
if (acct == NULL) {
filp_close(file, NULL);
return -ENOMEM;
}
}
spin_lock(&acct_lock);
if (ns->bacct == NULL) {
ns->bacct = acct;
acct = NULL;
}
mnt = file->f_path.mnt;
mnt_pin(mnt);
acct_file_reopen(ns->bacct, file, ns);
spin_unlock(&acct_lock);
[PATCH] saner handling of auto_acct_off() and DQUOT_OFF() in umount The way we currently deal with quota and process accounting that might keep vfsmount busy at umount time is inherently broken; we try to turn them off just in case (not quite correctly, at that) and a) pray umount doesn't fail (otherwise they'll stay turned off) b) pray nobody doesn anything funny just as we turn quota off Moreover, LSM provides hooks for doing the same sort of broken logics. The proper way to deal with that is to introduce the second kind of reference to vfsmount. Semantics: - when the last normal reference is dropped, all special ones are converted to normal ones and if there had been any, cleanup is done. - normal reference can be cloned into a special one - special reference can be converted to normal one; that's a no-op if we'd already passed the point of no return (i.e. mntput() had converted special references to normal and started cleanup). The way it works: e.g. starting process accounting converts the vfsmount reference pinned by the opened file into special one and turns it back to normal when it gets shut down; acct_auto_close() is done when no normal references are left. That way it does *not* obstruct umount(2) and it silently gets turned off when the last normal reference to vfsmount is gone. Which is exactly what we want... The same should be done by LSM module that holds some internal references to vfsmount and wants to shut them down on umount - it should make them special and security_sb_umount_close() will be called exactly when the last normal reference to vfsmount is gone. quota handling is even simpler - we don't use normal file IO anymore, so there's no need to hold vfsmounts at all. DQUOT_OFF() is done from deactivate_super(), where it really belongs. Signed-off-by: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-11-07 22:13:39 +00:00
mntput(mnt); /* it's pinned, now give up active reference */
kfree(acct);
[PATCH] saner handling of auto_acct_off() and DQUOT_OFF() in umount The way we currently deal with quota and process accounting that might keep vfsmount busy at umount time is inherently broken; we try to turn them off just in case (not quite correctly, at that) and a) pray umount doesn't fail (otherwise they'll stay turned off) b) pray nobody doesn anything funny just as we turn quota off Moreover, LSM provides hooks for doing the same sort of broken logics. The proper way to deal with that is to introduce the second kind of reference to vfsmount. Semantics: - when the last normal reference is dropped, all special ones are converted to normal ones and if there had been any, cleanup is done. - normal reference can be cloned into a special one - special reference can be converted to normal one; that's a no-op if we'd already passed the point of no return (i.e. mntput() had converted special references to normal and started cleanup). The way it works: e.g. starting process accounting converts the vfsmount reference pinned by the opened file into special one and turns it back to normal when it gets shut down; acct_auto_close() is done when no normal references are left. That way it does *not* obstruct umount(2) and it silently gets turned off when the last normal reference to vfsmount is gone. Which is exactly what we want... The same should be done by LSM module that holds some internal references to vfsmount and wants to shut them down on umount - it should make them special and security_sb_umount_close() will be called exactly when the last normal reference to vfsmount is gone. quota handling is even simpler - we don't use normal file IO anymore, so there's no need to hold vfsmounts at all. DQUOT_OFF() is done from deactivate_super(), where it really belongs. Signed-off-by: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-11-07 22:13:39 +00:00
return 0;
}
/**
* sys_acct - enable/disable process accounting
* @name: file name for accounting records or NULL to shutdown accounting
*
* Returns 0 for success or negative errno values for failure.
*
* sys_acct() is the only system call needed to implement process
* accounting. It takes the name of the file where accounting records
* should be written. If the filename is NULL, accounting will be
* shutdown.
*/
SYSCALL_DEFINE1(acct, const char __user *, name)
{
int error = 0;
if (!capable(CAP_SYS_PACCT))
return -EPERM;
if (name) {
struct filename *tmp = getname(name);
[PATCH] saner handling of auto_acct_off() and DQUOT_OFF() in umount The way we currently deal with quota and process accounting that might keep vfsmount busy at umount time is inherently broken; we try to turn them off just in case (not quite correctly, at that) and a) pray umount doesn't fail (otherwise they'll stay turned off) b) pray nobody doesn anything funny just as we turn quota off Moreover, LSM provides hooks for doing the same sort of broken logics. The proper way to deal with that is to introduce the second kind of reference to vfsmount. Semantics: - when the last normal reference is dropped, all special ones are converted to normal ones and if there had been any, cleanup is done. - normal reference can be cloned into a special one - special reference can be converted to normal one; that's a no-op if we'd already passed the point of no return (i.e. mntput() had converted special references to normal and started cleanup). The way it works: e.g. starting process accounting converts the vfsmount reference pinned by the opened file into special one and turns it back to normal when it gets shut down; acct_auto_close() is done when no normal references are left. That way it does *not* obstruct umount(2) and it silently gets turned off when the last normal reference to vfsmount is gone. Which is exactly what we want... The same should be done by LSM module that holds some internal references to vfsmount and wants to shut them down on umount - it should make them special and security_sb_umount_close() will be called exactly when the last normal reference to vfsmount is gone. quota handling is even simpler - we don't use normal file IO anymore, so there's no need to hold vfsmounts at all. DQUOT_OFF() is done from deactivate_super(), where it really belongs. Signed-off-by: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-11-07 22:13:39 +00:00
if (IS_ERR(tmp))
return (PTR_ERR(tmp));
error = acct_on(tmp);
putname(tmp);
[PATCH] saner handling of auto_acct_off() and DQUOT_OFF() in umount The way we currently deal with quota and process accounting that might keep vfsmount busy at umount time is inherently broken; we try to turn them off just in case (not quite correctly, at that) and a) pray umount doesn't fail (otherwise they'll stay turned off) b) pray nobody doesn anything funny just as we turn quota off Moreover, LSM provides hooks for doing the same sort of broken logics. The proper way to deal with that is to introduce the second kind of reference to vfsmount. Semantics: - when the last normal reference is dropped, all special ones are converted to normal ones and if there had been any, cleanup is done. - normal reference can be cloned into a special one - special reference can be converted to normal one; that's a no-op if we'd already passed the point of no return (i.e. mntput() had converted special references to normal and started cleanup). The way it works: e.g. starting process accounting converts the vfsmount reference pinned by the opened file into special one and turns it back to normal when it gets shut down; acct_auto_close() is done when no normal references are left. That way it does *not* obstruct umount(2) and it silently gets turned off when the last normal reference to vfsmount is gone. Which is exactly what we want... The same should be done by LSM module that holds some internal references to vfsmount and wants to shut them down on umount - it should make them special and security_sb_umount_close() will be called exactly when the last normal reference to vfsmount is gone. quota handling is even simpler - we don't use normal file IO anymore, so there's no need to hold vfsmounts at all. DQUOT_OFF() is done from deactivate_super(), where it really belongs. Signed-off-by: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-11-07 22:13:39 +00:00
} else {
struct bsd_acct_struct *acct;
acct = task_active_pid_ns(current)->bacct;
if (acct == NULL)
return 0;
spin_lock(&acct_lock);
acct_file_reopen(acct, NULL, NULL);
spin_unlock(&acct_lock);
}
[PATCH] saner handling of auto_acct_off() and DQUOT_OFF() in umount The way we currently deal with quota and process accounting that might keep vfsmount busy at umount time is inherently broken; we try to turn them off just in case (not quite correctly, at that) and a) pray umount doesn't fail (otherwise they'll stay turned off) b) pray nobody doesn anything funny just as we turn quota off Moreover, LSM provides hooks for doing the same sort of broken logics. The proper way to deal with that is to introduce the second kind of reference to vfsmount. Semantics: - when the last normal reference is dropped, all special ones are converted to normal ones and if there had been any, cleanup is done. - normal reference can be cloned into a special one - special reference can be converted to normal one; that's a no-op if we'd already passed the point of no return (i.e. mntput() had converted special references to normal and started cleanup). The way it works: e.g. starting process accounting converts the vfsmount reference pinned by the opened file into special one and turns it back to normal when it gets shut down; acct_auto_close() is done when no normal references are left. That way it does *not* obstruct umount(2) and it silently gets turned off when the last normal reference to vfsmount is gone. Which is exactly what we want... The same should be done by LSM module that holds some internal references to vfsmount and wants to shut them down on umount - it should make them special and security_sb_umount_close() will be called exactly when the last normal reference to vfsmount is gone. quota handling is even simpler - we don't use normal file IO anymore, so there's no need to hold vfsmounts at all. DQUOT_OFF() is done from deactivate_super(), where it really belongs. Signed-off-by: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-11-07 22:13:39 +00:00
return error;
}
[PATCH] saner handling of auto_acct_off() and DQUOT_OFF() in umount The way we currently deal with quota and process accounting that might keep vfsmount busy at umount time is inherently broken; we try to turn them off just in case (not quite correctly, at that) and a) pray umount doesn't fail (otherwise they'll stay turned off) b) pray nobody doesn anything funny just as we turn quota off Moreover, LSM provides hooks for doing the same sort of broken logics. The proper way to deal with that is to introduce the second kind of reference to vfsmount. Semantics: - when the last normal reference is dropped, all special ones are converted to normal ones and if there had been any, cleanup is done. - normal reference can be cloned into a special one - special reference can be converted to normal one; that's a no-op if we'd already passed the point of no return (i.e. mntput() had converted special references to normal and started cleanup). The way it works: e.g. starting process accounting converts the vfsmount reference pinned by the opened file into special one and turns it back to normal when it gets shut down; acct_auto_close() is done when no normal references are left. That way it does *not* obstruct umount(2) and it silently gets turned off when the last normal reference to vfsmount is gone. Which is exactly what we want... The same should be done by LSM module that holds some internal references to vfsmount and wants to shut them down on umount - it should make them special and security_sb_umount_close() will be called exactly when the last normal reference to vfsmount is gone. quota handling is even simpler - we don't use normal file IO anymore, so there's no need to hold vfsmounts at all. DQUOT_OFF() is done from deactivate_super(), where it really belongs. Signed-off-by: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-11-07 22:13:39 +00:00
/**
* acct_auto_close - turn off a filesystem's accounting if it is on
* @m: vfsmount being shut down
*
* If the accounting is turned on for a file in the subtree pointed to
* to by m, turn accounting off. Done when m is about to die.
*/
void acct_auto_close_mnt(struct vfsmount *m)
{
struct bsd_acct_struct *acct;
spin_lock(&acct_lock);
restart:
list_for_each_entry(acct, &acct_list, list)
if (acct->file && acct->file->f_path.mnt == m) {
acct_file_reopen(acct, NULL, NULL);
goto restart;
}
spin_unlock(&acct_lock);
}
/**
* acct_auto_close - turn off a filesystem's accounting if it is on
* @sb: super block for the filesystem
*
* If the accounting is turned on for a file in the filesystem pointed
* to by sb, turn accounting off.
*/
void acct_auto_close(struct super_block *sb)
{
struct bsd_acct_struct *acct;
spin_lock(&acct_lock);
restart:
list_for_each_entry(acct, &acct_list, list)
if (acct->file && acct->file->f_path.dentry->d_sb == sb) {
acct_file_reopen(acct, NULL, NULL);
goto restart;
}
spin_unlock(&acct_lock);
}
void acct_exit_ns(struct pid_namespace *ns)
{
struct bsd_acct_struct *acct = ns->bacct;
if (acct == NULL)
return;
spin_lock(&acct_lock);
if (acct->file != NULL)
acct_file_reopen(acct, NULL, NULL);
spin_unlock(&acct_lock);
kfree(acct);
}
/*
* encode an unsigned long into a comp_t
*
* This routine has been adopted from the encode_comp_t() function in
* the kern_acct.c file of the FreeBSD operating system. The encoding
* is a 13-bit fraction with a 3-bit (base 8) exponent.
*/
#define MANTSIZE 13 /* 13 bit mantissa. */
#define EXPSIZE 3 /* Base 8 (3 bit) exponent. */
#define MAXFRACT ((1 << MANTSIZE) - 1) /* Maximum fractional value. */
static comp_t encode_comp_t(unsigned long value)
{
int exp, rnd;
exp = rnd = 0;
while (value > MAXFRACT) {
rnd = value & (1 << (EXPSIZE - 1)); /* Round up? */
value >>= EXPSIZE; /* Base 8 exponent == 3 bit shift. */
exp++;
}
/*
* If we need to round up, do it (and handle overflow correctly).
*/
if (rnd && (++value > MAXFRACT)) {
value >>= EXPSIZE;
exp++;
}
/*
* Clean it up and polish it off.
*/
exp <<= MANTSIZE; /* Shift the exponent into place */
exp += value; /* and add on the mantissa. */
return exp;
}
#if ACCT_VERSION==1 || ACCT_VERSION==2
/*
* encode an u64 into a comp2_t (24 bits)
*
* Format: 5 bit base 2 exponent, 20 bits mantissa.
* The leading bit of the mantissa is not stored, but implied for
* non-zero exponents.
* Largest encodable value is 50 bits.
*/
#define MANTSIZE2 20 /* 20 bit mantissa. */
#define EXPSIZE2 5 /* 5 bit base 2 exponent. */
#define MAXFRACT2 ((1ul << MANTSIZE2) - 1) /* Maximum fractional value. */
#define MAXEXP2 ((1 <<EXPSIZE2) - 1) /* Maximum exponent. */
static comp2_t encode_comp2_t(u64 value)
{
int exp, rnd;
exp = (value > (MAXFRACT2>>1));
rnd = 0;
while (value > MAXFRACT2) {
rnd = value & 1;
value >>= 1;
exp++;
}
/*
* If we need to round up, do it (and handle overflow correctly).
*/
if (rnd && (++value > MAXFRACT2)) {
value >>= 1;
exp++;
}
if (exp > MAXEXP2) {
/* Overflow. Return largest representable number instead. */
return (1ul << (MANTSIZE2+EXPSIZE2-1)) - 1;
} else {
return (value & (MAXFRACT2>>1)) | (exp << (MANTSIZE2-1));
}
}
#endif
#if ACCT_VERSION==3
/*
* encode an u64 into a 32 bit IEEE float
*/
static u32 encode_float(u64 value)
{
unsigned exp = 190;
unsigned u;
if (value==0) return 0;
while ((s64)value > 0){
value <<= 1;
exp--;
}
u = (u32)(value >> 40) & 0x7fffffu;
return u | (exp << 23);
}
#endif
/*
* Write an accounting entry for an exiting process
*
* The acct_process() call is the workhorse of the process
* accounting system. The struct acct is built here and then written
* into the accounting file. This function should only be called from
* do_exit() or when switching to a different output file.
*/
/*
* do_acct_process does all actual work. Caller holds the reference to file.
*/
static void do_acct_process(struct bsd_acct_struct *acct,
struct pid_namespace *ns, struct file *file)
{
struct pacct_struct *pacct = &current->signal->pacct;
acct_t ac;
mm_segment_t fs;
unsigned long flim;
u64 elapsed;
u64 run_time;
struct timespec uptime;
[PATCH] tty: ->signal->tty locking Fix the locking of signal->tty. Use ->sighand->siglock to protect ->signal->tty; this lock is already used by most other members of ->signal/->sighand. And unless we are 'current' or the tasklist_lock is held we need ->siglock to access ->signal anyway. (NOTE: sys_unshare() is broken wrt ->sighand locking rules) Note that tty_mutex is held over tty destruction, so while holding tty_mutex any tty pointer remains valid. Otherwise the lifetime of ttys are governed by their open file handles. This leaves some holes for tty access from signal->tty (or any other non file related tty access). It solves the tty SLAB scribbles we were seeing. (NOTE: the change from group_send_sig_info to __group_send_sig_info needs to be examined by someone familiar with the security framework, I think it is safe given the SEND_SIG_PRIV from other __group_send_sig_info invocations) [schwidefsky@de.ibm.com: 3270 fix] [akpm@osdl.org: various post-viro fixes] Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Acked-by: Alan Cox <alan@redhat.com> Cc: Oleg Nesterov <oleg@tv-sign.ru> Cc: Prarit Bhargava <prarit@redhat.com> Cc: Chris Wright <chrisw@sous-sol.org> Cc: Roland McGrath <roland@redhat.com> Cc: Stephen Smalley <sds@tycho.nsa.gov> Cc: James Morris <jmorris@namei.org> Cc: "David S. Miller" <davem@davemloft.net> Cc: Jeff Dike <jdike@addtoit.com> Cc: Martin Schwidefsky <schwidefsky@de.ibm.com> Cc: Jan Kara <jack@ucw.cz> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-08 10:36:04 +00:00
struct tty_struct *tty;
const struct cred *orig_cred;
/* Perform file operations on behalf of whoever enabled accounting */
orig_cred = override_creds(file->f_cred);
/*
* First check to see if there is enough free_space to continue
* the process accounting system.
*/
if (!check_free_space(acct, file))
goto out;
/*
* Fill the accounting struct with the needed info as recorded
* by the different kernel functions.
*/
memset(&ac, 0, sizeof(acct_t));
ac.ac_version = ACCT_VERSION | ACCT_BYTEORDER;
strlcpy(ac.ac_comm, current->comm, sizeof(ac.ac_comm));
/* calculate run_time in nsec*/
do_posix_clock_monotonic_gettime(&uptime);
run_time = (u64)uptime.tv_sec*NSEC_PER_SEC + uptime.tv_nsec;
[PATCH] Fix pacct bug in multithreading case. I noticed a bug on the process accounting facility. In multi-threading process, some data would be recorded incorrectly when the group_leader dies earlier than one or more threads. The attached patch fixes this problem. See below. 'bugacct' is a test program that create a worker thread after 4 seconds sleeping, then the group_leader dies soon. The worker thread consume CPU/Memory for 6 seconds, then exit. We can estimate 10 seconds as etime and 6 seconds as stime + utime. This is a sample program which the group_leader dies earlier than other threads. The results of same binary execution on different kernel are below. -- accounted records -------------------- | btime | utime | stime | etime | minflt | majflt | comm | original | 13:16:40 | 0.00 | 0.00 | 6.10 | 171 | 0 | bugacct | patched | 13:20:21 | 5.83 | 0.18 | 10.03 | 32776 | 0 | bugacct | (*) bugacct allocates 128MB memory, thus 128MB / 4KB = 32768 of minflt is appropriate. -- Test results in original kernel ------ $ date; time -p ./bugacct Tue Mar 28 13:16:36 JST 2006 <- But pacct said btime is 13:16:40 real 10.11 <- But pacct said etime is 6.10 user 5.96 <- But pacct said utime is 0.00 sys 0.14 <- But pacct said stime is 0.00 $ -- Test results in patched kernel ------- $ date; time -p ./bugacct Tue Mar 28 13:20:21 JST 2006 real 10.04 user 5.83 sys 0.19 $ In the original 2.6.16 kernel, pacct records btime, utime, stime, etime and minflt incorrectly. In my opinion, this problem is caused by an assumption that group_leader dies last. The following section calculates process running time for etime and btime. But it means running time of the thread that dies last, not process. The start_time of the first thread in the process (group_leader) should be reduced from uptime to calculate etime and btime correctly. ---- do_acct_process() in kernel/acct.c: /* calculate run_time in nsec*/ do_posix_clock_monotonic_gettime(&uptime); run_time = (u64)uptime.tv_sec*NSEC_PER_SEC + uptime.tv_nsec; run_time -= (u64)current->start_time.tv_sec*NSEC_PER_SEC + current->start_time.tv_nsec; ---- The following section calculates stime and utime of the process. But it might count the utime and stime of the group_leader duplicatly and ignore the utime and stime of the thread dies last, when one or more threads remain after group_leader dead. The ac_utime should be calculated as the sum of the signal->utime and utime of the thread dies last. The ac_stime should be done also. ---- do_acct_process() in kernel/acct.c: jiffies = cputime_to_jiffies(cputime_add(current->group_leader->utime, current->signal->utime)); ac.ac_utime = encode_comp_t(jiffies_to_AHZ(jiffies)); jiffies = cputime_to_jiffies(cputime_add(current->group_leader->stime, current->signal->stime)); ac.ac_stime = encode_comp_t(jiffies_to_AHZ(jiffies)); ---- The part of the minflt/majflt calculation has same problem. This patch solves those problems, I think. Cc: Ingo Molnar <mingo@elte.hu> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-03-31 10:30:45 +00:00
run_time -= (u64)current->group_leader->start_time.tv_sec * NSEC_PER_SEC
+ current->group_leader->start_time.tv_nsec;
/* convert nsec -> AHZ */
elapsed = nsec_to_AHZ(run_time);
#if ACCT_VERSION==3
ac.ac_etime = encode_float(elapsed);
#else
ac.ac_etime = encode_comp_t(elapsed < (unsigned long) -1l ?
(unsigned long) elapsed : (unsigned long) -1l);
#endif
#if ACCT_VERSION==1 || ACCT_VERSION==2
{
/* new enlarged etime field */
comp2_t etime = encode_comp2_t(elapsed);
ac.ac_etime_hi = etime >> 16;
ac.ac_etime_lo = (u16) etime;
}
#endif
do_div(elapsed, AHZ);
ac.ac_btime = get_seconds() - elapsed;
/* we really need to bite the bullet and change layout */
ac.ac_uid = from_kuid_munged(file->f_cred->user_ns, orig_cred->uid);
ac.ac_gid = from_kgid_munged(file->f_cred->user_ns, orig_cred->gid);
#if ACCT_VERSION==2
ac.ac_ahz = AHZ;
#endif
#if ACCT_VERSION==1 || ACCT_VERSION==2
/* backward-compatible 16 bit fields */
ac.ac_uid16 = ac.ac_uid;
ac.ac_gid16 = ac.ac_gid;
#endif
#if ACCT_VERSION==3
ac.ac_pid = task_tgid_nr_ns(current, ns);
rcu_read_lock();
ac.ac_ppid = task_tgid_nr_ns(rcu_dereference(current->real_parent), ns);
rcu_read_unlock();
#endif
spin_lock_irq(&current->sighand->siglock);
tty = current->signal->tty; /* Safe as we hold the siglock */
ac.ac_tty = tty ? old_encode_dev(tty_devnum(tty)) : 0;
ac.ac_utime = encode_comp_t(jiffies_to_AHZ(cputime_to_jiffies(pacct->ac_utime)));
ac.ac_stime = encode_comp_t(jiffies_to_AHZ(cputime_to_jiffies(pacct->ac_stime)));
ac.ac_flag = pacct->ac_flag;
ac.ac_mem = encode_comp_t(pacct->ac_mem);
ac.ac_minflt = encode_comp_t(pacct->ac_minflt);
ac.ac_majflt = encode_comp_t(pacct->ac_majflt);
ac.ac_exitcode = pacct->ac_exitcode;
spin_unlock_irq(&current->sighand->siglock);
ac.ac_io = encode_comp_t(0 /* current->io_usage */); /* %% */
ac.ac_rw = encode_comp_t(ac.ac_io / 1024);
ac.ac_swaps = encode_comp_t(0);
/*
* Get freeze protection. If the fs is frozen, just skip the write
* as we could deadlock the system otherwise.
*/
if (!file_start_write_trylock(file))
goto out;
/*
* Kernel segment override to datasegment and write it
* to the accounting file.
*/
fs = get_fs();
set_fs(KERNEL_DS);
/*
* Accounting records are not subject to resource limits.
*/
flim = current->signal->rlim[RLIMIT_FSIZE].rlim_cur;
current->signal->rlim[RLIMIT_FSIZE].rlim_cur = RLIM_INFINITY;
file->f_op->write(file, (char *)&ac,
sizeof(acct_t), &file->f_pos);
current->signal->rlim[RLIMIT_FSIZE].rlim_cur = flim;
set_fs(fs);
file_end_write(file);
out:
revert_creds(orig_cred);
}
/**
* acct_collect - collect accounting information into pacct_struct
* @exitcode: task exit code
* @group_dead: not 0, if this thread is the last one in the process.
*/
void acct_collect(long exitcode, int group_dead)
{
struct pacct_struct *pacct = &current->signal->pacct;
cputime_t utime, stime;
unsigned long vsize = 0;
if (group_dead && current->mm) {
struct vm_area_struct *vma;
down_read(&current->mm->mmap_sem);
vma = current->mm->mmap;
while (vma) {
vsize += vma->vm_end - vma->vm_start;
vma = vma->vm_next;
}
up_read(&current->mm->mmap_sem);
}
spin_lock_irq(&current->sighand->siglock);
if (group_dead)
pacct->ac_mem = vsize / 1024;
if (thread_group_leader(current)) {
pacct->ac_exitcode = exitcode;
if (current->flags & PF_FORKNOEXEC)
pacct->ac_flag |= AFORK;
}
if (current->flags & PF_SUPERPRIV)
pacct->ac_flag |= ASU;
if (current->flags & PF_DUMPCORE)
pacct->ac_flag |= ACORE;
if (current->flags & PF_SIGNALED)
pacct->ac_flag |= AXSIG;
task_cputime(current, &utime, &stime);
pacct->ac_utime += utime;
pacct->ac_stime += stime;
pacct->ac_minflt += current->min_flt;
pacct->ac_majflt += current->maj_flt;
spin_unlock_irq(&current->sighand->siglock);
}
static void acct_process_in_ns(struct pid_namespace *ns)
{
struct file *file = NULL;
struct bsd_acct_struct *acct;
acct = ns->bacct;
/*
* accelerate the common fastpath:
*/
if (!acct || !acct->file)
return;
spin_lock(&acct_lock);
file = acct->file;
if (unlikely(!file)) {
spin_unlock(&acct_lock);
return;
}
get_file(file);
spin_unlock(&acct_lock);
do_acct_process(acct, ns, file);
fput(file);
}
/**
* acct_process - now just a wrapper around acct_process_in_ns,
* which in turn is a wrapper around do_acct_process.
*
* handles process accounting for an exiting task
*/
void acct_process(void)
{
struct pid_namespace *ns;
/*
* This loop is safe lockless, since current is still
* alive and holds its namespace, which in turn holds
* its parent.
*/
for (ns = task_active_pid_ns(current); ns != NULL; ns = ns->parent)
acct_process_in_ns(ns);
}