Free Electrons

Embedded Linux Experts

  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
/*
 *  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>
#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.
 */
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)
{
	struct file *file;
	struct vfsmount *mnt;
	struct pid_namespace *ns;
	struct bsd_acct_struct *acct = NULL;

	/* Difference from BSD - they don't do O_APPEND */
	file = file_open_name(pathname, O_WRONLY|O_APPEND|O_LARGEFILE, 0);
	if (IS_ERR(file))
		return PTR_ERR(file);

	if (!S_ISREG(file_inode(file)->i_mode)) {
		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);

	mntput(mnt); /* it's pinned, now give up active reference */
	kfree(acct);

	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);
		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;

		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)
{
	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;
	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;
	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);
}