* a struct file opened for write. Fixed. 2/6/2000, AV.
*/
-#include <linux/config.h>
#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/jiffies.h>
#include <linux/times.h>
#include <linux/syscalls.h>
+#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
/*
* External references and all of the globals.
*/
-static void do_acct_process(long, struct file *);
+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.
*/
-struct acct_glbs {
- spinlock_t lock;
+struct bsd_acct_struct {
volatile int active;
volatile int needcheck;
struct file *file;
+ struct pid_namespace *ns;
struct timer_list timer;
+ struct list_head list;
};
-static struct acct_glbs acct_globals __cacheline_aligned = {SPIN_LOCK_UNLOCKED};
+static DEFINE_SPINLOCK(acct_lock);
+static LIST_HEAD(acct_list);
/*
* Called whenever the timer says to check the free space.
*/
-static void acct_timeout(unsigned long unused)
+static void acct_timeout(unsigned long x)
{
- acct_globals.needcheck = 1;
+ struct bsd_acct_struct *acct = (struct bsd_acct_struct *)x;
+ acct->needcheck = 1;
}
/*
* Check the amount of free space and suspend/resume accordingly.
*/
-static int check_free_space(struct file *file)
+static int check_free_space(struct bsd_acct_struct *acct, struct file *file)
{
struct kstatfs sbuf;
int res;
sector_t resume;
sector_t suspend;
- spin_lock(&acct_globals.lock);
- res = acct_globals.active;
- if (!file || !acct_globals.needcheck)
+ spin_lock(&acct_lock);
+ res = acct->active;
+ if (!file || !acct->needcheck)
goto out;
- spin_unlock(&acct_globals.lock);
+ spin_unlock(&acct_lock);
/* May block */
- if (vfs_statfs(file->f_dentry->d_inode->i_sb, &sbuf))
+ if (vfs_statfs(file->f_path.dentry, &sbuf))
return res;
suspend = sbuf.f_blocks * SUSPEND;
resume = sbuf.f_blocks * RESUME;
act = 0;
/*
- * If some joker switched acct_globals.file under us we'ld better be
+ * If some joker switched acct->file under us we'ld better be
* silent and _not_ touch anything.
*/
- spin_lock(&acct_globals.lock);
- if (file != acct_globals.file) {
+ spin_lock(&acct_lock);
+ if (file != acct->file) {
if (act)
res = act>0;
goto out;
}
- if (acct_globals.active) {
+ if (acct->active) {
if (act < 0) {
- acct_globals.active = 0;
+ acct->active = 0;
printk(KERN_INFO "Process accounting paused\n");
}
} else {
if (act > 0) {
- acct_globals.active = 1;
+ acct->active = 1;
printk(KERN_INFO "Process accounting resumed\n");
}
}
- del_timer(&acct_globals.timer);
- acct_globals.needcheck = 0;
- acct_globals.timer.expires = jiffies + ACCT_TIMEOUT*HZ;
- add_timer(&acct_globals.timer);
- res = acct_globals.active;
+ del_timer(&acct->timer);
+ acct->needcheck = 0;
+ acct->timer.expires = jiffies + ACCT_TIMEOUT*HZ;
+ add_timer(&acct->timer);
+ res = acct->active;
out:
- spin_unlock(&acct_globals.lock);
+ spin_unlock(&acct_lock);
return res;
}
/*
- * Close the old accouting file (if currently open) and then replace
+ * Close the old accounting file (if currently open) and then replace
* it with file (if non-NULL).
*
- * NOTE: acct_globals.lock MUST be held on entry and exit.
+ * NOTE: acct_lock MUST be held on entry and exit.
*/
-static void acct_file_reopen(struct file *file)
+static void acct_file_reopen(struct bsd_acct_struct *acct, struct file *file,
+ struct pid_namespace *ns)
{
struct file *old_acct = NULL;
-
- if (acct_globals.file) {
- old_acct = acct_globals.file;
- del_timer(&acct_globals.timer);
- acct_globals.active = 0;
- acct_globals.needcheck = 0;
- acct_globals.file = NULL;
+ struct pid_namespace *old_ns = NULL;
+
+ if (acct->file) {
+ old_acct = acct->file;
+ old_ns = acct->ns;
+ del_timer(&acct->timer);
+ acct->active = 0;
+ acct->needcheck = 0;
+ acct->file = NULL;
+ acct->ns = NULL;
+ list_del(&acct->list);
}
if (file) {
- acct_globals.file = file;
- acct_globals.needcheck = 0;
- acct_globals.active = 1;
+ acct->file = file;
+ acct->ns = ns;
+ acct->needcheck = 0;
+ acct->active = 1;
+ list_add(&acct->list, &acct_list);
/* It's been deleted if it was used before so this is safe */
- init_timer(&acct_globals.timer);
- acct_globals.timer.function = acct_timeout;
- acct_globals.timer.expires = jiffies + ACCT_TIMEOUT*HZ;
- add_timer(&acct_globals.timer);
+ setup_timer(&acct->timer, acct_timeout, (unsigned long)acct);
+ acct->timer.expires = jiffies + ACCT_TIMEOUT*HZ;
+ add_timer(&acct->timer);
}
if (old_acct) {
- spin_unlock(&acct_globals.lock);
- do_acct_process(0, 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_globals.lock);
+ spin_lock(&acct_lock);
}
}
-/*
- * 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.
- */
-asmlinkage long sys_acct(const char __user *name)
+static int acct_on(char *name)
{
- struct file *file = NULL;
- char *tmp;
+ struct file *file;
+ struct vfsmount *mnt;
int error;
+ struct pid_namespace *ns;
+ struct bsd_acct_struct *acct = NULL;
- if (!capable(CAP_SYS_PACCT))
- return -EPERM;
+ /* Difference from BSD - they don't do O_APPEND */
+ file = filp_open(name, O_WRONLY|O_APPEND|O_LARGEFILE, 0);
+ if (IS_ERR(file))
+ return PTR_ERR(file);
- if (name) {
- tmp = getname(name);
- if (IS_ERR(tmp)) {
- return (PTR_ERR(tmp));
- }
- /* Difference from BSD - they don't do O_APPEND */
- file = filp_open(tmp, O_WRONLY|O_APPEND, 0);
- putname(tmp);
- if (IS_ERR(file)) {
- return (PTR_ERR(file));
- }
- if (!S_ISREG(file->f_dentry->d_inode->i_mode)) {
- filp_close(file, NULL);
- return (-EACCES);
- }
+ if (!S_ISREG(file->f_path.dentry->d_inode->i_mode)) {
+ filp_close(file, NULL);
+ return -EACCES;
+ }
+
+ if (!file->f_op->write) {
+ filp_close(file, NULL);
+ return -EIO;
+ }
- if (!file->f_op->write) {
+ 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 (-EIO);
+ return -ENOMEM;
}
}
error = security_acct(file);
if (error) {
- if (file)
- filp_close(file, NULL);
+ kfree(acct);
+ filp_close(file, NULL);
return error;
}
- spin_lock(&acct_globals.lock);
- acct_file_reopen(file);
- spin_unlock(&acct_globals.lock);
+ 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);
+
+ mntput(mnt); /* it's pinned, now give up active reference */
+ kfree(acct);
- return (0);
+ return 0;
}
-/*
- * If the accouting is turned on for a file in the filesystem pointed
- * to by sb, turn accouting off.
+/**
+ * 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;
+
+ if (!capable(CAP_SYS_PACCT))
+ return -EPERM;
+
+ if (name) {
+ char *tmp = getname(name);
+ if (IS_ERR(tmp))
+ return (PTR_ERR(tmp));
+ error = acct_on(tmp);
+ putname(tmp);
+ } else {
+ struct bsd_acct_struct *acct;
+
+ acct = task_active_pid_ns(current)->bacct;
+ if (acct == NULL)
+ return 0;
+
+ error = security_acct(NULL);
+ if (!error) {
+ spin_lock(&acct_lock);
+ acct_file_reopen(acct, NULL, NULL);
+ spin_unlock(&acct_lock);
+ }
+ }
+ return error;
+}
+
+/**
+ * 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)
{
- spin_lock(&acct_globals.lock);
- if (acct_globals.file &&
- acct_globals.file->f_dentry->d_inode->i_sb == sb) {
- acct_file_reopen((struct file *)NULL);
+ 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->mnt_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;
+
+ spin_lock(&acct_lock);
+ acct = ns->bacct;
+ if (acct != NULL) {
+ if (acct->file != NULL)
+ acct_file_reopen(acct, NULL, NULL);
+
+ kfree(acct);
}
- spin_unlock(&acct_globals.lock);
+ spin_unlock(&acct_lock);
}
/*
}
/*
- * If we need to round up, do it (and handle overflow correctly).
- */
+ * 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.
- */
+ * Clean it up and polish it off.
+ */
exp <<= MANTSIZE; /* Shift the exponent into place */
exp += value; /* and add on the mantissa. */
return exp;
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));
- }
+ 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
* 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().
+ * 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(long exitcode, struct file *file)
+static void do_acct_process(struct bsd_acct_struct *acct,
+ struct pid_namespace *ns, struct file *file)
{
+ struct pacct_struct *pacct = ¤t->signal->pacct;
acct_t ac;
mm_segment_t fs;
- unsigned long vsize;
unsigned long flim;
u64 elapsed;
u64 run_time;
struct timespec uptime;
+ 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(file))
- return;
+ if (!check_free_space(acct, file))
+ goto out;
/*
* Fill the accounting struct with the needed info as recorded
/* 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;
+ 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
}
#endif
do_div(elapsed, AHZ);
- ac.ac_btime = xtime.tv_sec - elapsed;
- ac.ac_utime = encode_comp_t(jiffies_to_AHZ(
- current->signal->utime +
- current->group_leader->utime));
- ac.ac_stime = encode_comp_t(jiffies_to_AHZ(
- current->signal->stime +
- current->group_leader->stime));
+ ac.ac_btime = get_seconds() - elapsed;
/* we really need to bite the bullet and change layout */
- ac.ac_uid = current->uid;
- ac.ac_gid = current->gid;
+ ac.ac_uid = orig_cred->uid;
+ ac.ac_gid = 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 = current->uid;
- ac.ac_gid16 = current->gid;
+ ac.ac_uid16 = ac.ac_uid;
+ ac.ac_gid16 = ac.ac_gid;
#endif
#if ACCT_VERSION==3
- ac.ac_pid = current->tgid;
- ac.ac_ppid = current->parent->tgid;
+ 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
- read_lock(&tasklist_lock); /* pin current->signal */
- ac.ac_tty = current->signal->tty ?
- old_encode_dev(tty_devnum(current->signal->tty)) : 0;
- read_unlock(&tasklist_lock);
-
- ac.ac_flag = 0;
- if (current->flags & PF_FORKNOEXEC)
- ac.ac_flag |= AFORK;
- if (current->flags & PF_SUPERPRIV)
- ac.ac_flag |= ASU;
- if (current->flags & PF_DUMPCORE)
- ac.ac_flag |= ACORE;
- if (current->flags & PF_SIGNALED)
- ac.ac_flag |= AXSIG;
-
- vsize = 0;
- if (current->mm) {
- struct vm_area_struct *vma;
- down_read(¤t->mm->mmap_sem);
- vma = current->mm->mmap;
- while (vma) {
- vsize += vma->vm_end - vma->vm_start;
- vma = vma->vm_next;
- }
- up_read(¤t->mm->mmap_sem);
- }
- vsize = vsize / 1024;
- ac.ac_mem = encode_comp_t(vsize);
+ spin_lock_irq(¤t->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(¤t->sighand->siglock);
ac.ac_io = encode_comp_t(0 /* current->io_usage */); /* %% */
ac.ac_rw = encode_comp_t(ac.ac_io / 1024);
- ac.ac_minflt = encode_comp_t(current->signal->min_flt +
- current->group_leader->min_flt);
- ac.ac_majflt = encode_comp_t(current->signal->maj_flt +
- current->group_leader->maj_flt);
ac.ac_swaps = encode_comp_t(0);
- ac.ac_exitcode = exitcode;
/*
- * Kernel segment override to datasegment and write it
- * to the accounting file.
- */
+ * 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.
- */
+ * 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);
+out:
+ revert_creds(orig_cred);
}
-/*
- * acct_process - now just a wrapper around do_acct_process
+/**
+ * acct_init_pacct - initialize a new pacct_struct
+ * @pacct: per-process accounting info struct to initialize
+ */
+void acct_init_pacct(struct pacct_struct *pacct)
+{
+ memset(pacct, 0, sizeof(struct pacct_struct));
+ pacct->ac_utime = pacct->ac_stime = cputime_zero;
+}
+
+/**
+ * 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_process(long exitcode)
+void acct_collect(long exitcode, int group_dead)
+{
+ struct pacct_struct *pacct = ¤t->signal->pacct;
+ unsigned long vsize = 0;
+
+ if (group_dead && current->mm) {
+ struct vm_area_struct *vma;
+ down_read(¤t->mm->mmap_sem);
+ vma = current->mm->mmap;
+ while (vma) {
+ vsize += vma->vm_end - vma->vm_start;
+ vma = vma->vm_next;
+ }
+ up_read(¤t->mm->mmap_sem);
+ }
+
+ spin_lock_irq(¤t->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;
+ pacct->ac_utime = cputime_add(pacct->ac_utime, current->utime);
+ pacct->ac_stime = cputime_add(pacct->ac_stime, current->stime);
+ pacct->ac_minflt += current->min_flt;
+ pacct->ac_majflt += current->maj_flt;
+ spin_unlock_irq(¤t->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_globals.file)
+ if (!acct || !acct->file)
return;
- spin_lock(&acct_globals.lock);
- file = acct_globals.file;
+ spin_lock(&acct_lock);
+ file = acct->file;
if (unlikely(!file)) {
- spin_unlock(&acct_globals.lock);
+ spin_unlock(&acct_lock);
return;
}
get_file(file);
- spin_unlock(&acct_globals.lock);
+ spin_unlock(&acct_lock);
- do_acct_process(exitcode, file);
+ do_acct_process(acct, ns, file);
fput(file);
}
-
-/*
- * acct_update_integrals
- * - update mm integral fields in task_struct
+/**
+ * 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_update_integrals(struct task_struct *tsk)
+void acct_process(void)
{
- if (likely(tsk->mm)) {
- long delta = tsk->stime - tsk->acct_stimexpd;
-
- if (delta == 0)
- return;
- tsk->acct_stimexpd = tsk->stime;
- tsk->acct_rss_mem1 += delta * get_mm_counter(tsk->mm, rss);
- tsk->acct_vm_mem1 += delta * tsk->mm->total_vm;
- }
-}
+ struct pid_namespace *ns;
-/*
- * acct_clear_integrals
- * - clear the mm integral fields in task_struct
- */
-void acct_clear_integrals(struct task_struct *tsk)
-{
- if (tsk) {
- tsk->acct_stimexpd = 0;
- tsk->acct_rss_mem1 = 0;
- tsk->acct_vm_mem1 = 0;
- }
+ /*
+ * 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);
}
git://ftp.safe.ca / safe / jmp / linux-2.6 / blobdiff