/*- * Copyright (c) 1994 Christopher G. Demetriou * Copyright (c) 1982, 1986, 1989, 1993 * The Regents of the University of California. All rights reserved. * (c) UNIX System Laboratories, Inc. * All or some portions of this file are derived from material licensed * to the University of California by American Telephone and Telegraph * Co. or Unix System Laboratories, Inc. and are reproduced herein with * the permission of UNIX System Laboratories, Inc. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * 3. Neither the name of the University nor the names of its contributors * may be used to endorse or promote products derived from this software * without specific prior written permission. * * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF * SUCH DAMAGE. * * @(#)kern_acct.c 8.1 (Berkeley) 6/14/93 * $FreeBSD: src/sys/kern/kern_acct.c,v 1.23.2.1 2002/07/24 18:33:55 johan Exp $ */ #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include static struct lock acct_lock = LOCK_INITIALIZER("acct_lock", 0, 0); /* * The routines implemented in this file are described in: * Leffler, et al.: The Design and Implementation of the 4.3BSD * UNIX Operating System (Addison Welley, 1989) * on pages 62-63. * * Arguably, to simplify accounting operations, this mechanism should * be replaced by one in which an accounting log file (similar to /dev/klog) * is read by a user process, etc. However, that has its own problems. */ /* * Internal accounting functions. * The former's operation is described in Leffler, et al., and the latter * was provided by UCB with the 4.4BSD-Lite release */ static comp_t encode_comp_t (u_long, u_long); static void acctwatch (void *); /* * Accounting callout handle used for periodic scheduling of * acctwatch. */ static struct callout acctwatch_handle; /* * Accounting vnode pointer, and saved vnode pointer. */ static struct vnode *acctp; static struct vnode *savacctp; /* * Values associated with enabling and disabling accounting */ static int acctsuspend = 2; /* stop accounting when < 2% free space left */ SYSCTL_INT(_kern, OID_AUTO, acct_suspend, CTLFLAG_RW, &acctsuspend, 0, "Percentage of free disk space below which accounting will be suspended"); static int acctresume = 4; /* resume when free space risen to > 4% */ SYSCTL_INT(_kern, OID_AUTO, acct_resume, CTLFLAG_RW, &acctresume, 0, "Minimum percentage of free disk space needed to resume accounting"); static int acctchkfreq = 15; /* frequency (in seconds) to check space */ SYSCTL_INT(_kern, OID_AUTO, acct_chkfreq, CTLFLAG_RW, &acctchkfreq, 0, "Frequency (in seconds) of free disk space checking"); static void acct_init(void *arg __unused) { callout_init_lk(&acctwatch_handle, &acct_lock); } SYSINIT(acct, SI_SUB_DRIVERS, SI_ORDER_ANY, acct_init, NULL); /* * Accounting system call. Written based on the specification and * previous implementation done by Mark Tinguely. * * acct(char *path) * * MPALMOSTSAFE */ int sys_acct(struct sysmsg *sysmsg, const struct acct_args *uap) { struct nlookupdata nd; struct vnode *vp; int error; error = caps_priv_check_self(SYSCAP_NOACCT); if (error) return (error); lockmgr(&acct_lock, LK_EXCLUSIVE); /* * If accounting is to be started to a file, open that file for * appending and make sure it's a 'normal'. */ if (uap->path != NULL) { error = nlookup_init(&nd, uap->path, UIO_USERSPACE, NLC_LOCKVP); if (error == 0) error = vn_open(&nd, NULL, FWRITE | O_APPEND, 0); if (error == 0 && nd.nl_open_vp->v_type != VREG) error = EACCES; if (error) { nlookup_done(&nd); goto done; } vp = nd.nl_open_vp; nd.nl_open_vp = NULL; nlookup_done(&nd); vn_unlock(vp); } else { vp = NULL; } /* * If accounting was previously enabled, kill the old space-watcher, * close the file. */ if (acctp != NULLVP || savacctp != NULLVP) { callout_stop(&acctwatch_handle); error = vn_close((acctp != NULLVP ? acctp : savacctp), FWRITE | O_APPEND, NULL); acctp = savacctp = NULLVP; } /* * If no new file opened then leave. We never did an nlookup so * don't try cleaning it up. */ if (uap->path == NULL) goto done; /* * Save the new accounting file vnode, and schedule the new * free space watcher. */ acctp = vp; acctwatch(NULL); done: lockmgr(&acct_lock, LK_RELEASE); return (error); } /* * Write out process accounting information, on process exit. * Data to be written out is specified in Leffler, et al. * and are enumerated below. (They're also noted in the system * "acct.h" header file.) */ int acct_process(struct proc *p) { struct acct acct; struct rusage *r; struct rusage ru; struct timeval tmp; struct rlimit rlim; int t; int error; struct vnode *vp; /* * If accounting isn't enabled, don't bother. Lock acct_lock * make sure. */ if (acctp == NULLVP) return 0; lockmgr(&acct_lock, LK_SHARED); vp = acctp; if (vp == NULLVP) { lockmgr(&acct_lock, LK_RELEASE); return 0; } /* * Get process accounting information. */ /* (1) The name of the command that ran */ bcopy(p->p_comm, acct.ac_comm, sizeof acct.ac_comm); /* (2) The amount of user and system time that was used */ calcru_proc(p, &ru); acct.ac_utime = encode_comp_t(ru.ru_utime.tv_sec, ru.ru_utime.tv_usec); acct.ac_stime = encode_comp_t(ru.ru_stime.tv_sec, ru.ru_stime.tv_usec); /* (3) The elapsed time the commmand ran (and its starting time) */ acct.ac_btime = p->p_start.tv_sec; microtime(&tmp); timevalsub(&tmp, &p->p_start); acct.ac_etime = encode_comp_t(tmp.tv_sec, tmp.tv_usec); /* (4) The average amount of memory used */ r = &p->p_ru; tmp = ru.ru_utime; timevaladd(&tmp, &ru.ru_stime); t = tmp.tv_sec * hz + tmp.tv_usec / ustick; if (t) acct.ac_mem = (r->ru_ixrss + r->ru_idrss + r->ru_isrss) / t; else acct.ac_mem = 0; /* (5) The number of disk I/O operations done */ acct.ac_io = encode_comp_t(r->ru_inblock + r->ru_oublock, 0); /* (6) The UID and GID of the process */ acct.ac_uid = p->p_ucred->cr_ruid; acct.ac_gid = p->p_ucred->cr_rgid; /* (7) The terminal from which the process was started */ if ((p->p_flags & P_CONTROLT) && p->p_pgrp->pg_session->s_ttyp) { acct.ac_tty = devid_from_dev(p->p_pgrp->pg_session->s_ttyp->t_dev); } else { acct.ac_tty = NOUDEV; } /* (8) The boolean flags that tell how the process terminated, etc. */ acct.ac_flag = p->p_acflag; /* * Eliminate any file size rlimit. */ rlim.rlim_cur = RLIM_INFINITY; rlim.rlim_max = RLIM_INFINITY; plimit_modify(p, RLIMIT_FSIZE, &rlim); /* * Write the accounting information to the file. */ error = vn_rdwr(UIO_WRITE, vp, (caddr_t)&acct, sizeof (acct), (off_t)0, UIO_SYSSPACE, IO_APPEND|IO_UNIT, p->p_ucred, NULL); lockmgr(&acct_lock, LK_RELEASE); return error; } /* * Encode_comp_t converts from ticks in seconds and microseconds * to ticks in 1/AHZ seconds. The encoding is described in * Leffler, et al., on page 63. */ #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(u_long s, u_long us) { int exp, rnd; exp = 0; rnd = 0; s *= AHZ; s += us / (1000000 / AHZ); /* Maximize precision. */ while (s > MAXFRACT) { rnd = s & (1 << (EXPSIZE - 1)); /* Round up? */ s >>= EXPSIZE; /* Base 8 exponent == 3 bit shift. */ exp++; } /* If we need to round up, do it (and handle overflow correctly). */ if (rnd && (++s > MAXFRACT)) { s >>= EXPSIZE; exp++; } /* Clean it up and polish it off. */ exp <<= MANTSIZE; /* Shift the exponent into place */ exp += s; /* and add on the mantissa. */ return (exp); } /* * Periodically check the file system to see if accounting * should be turned on or off. Beware the case where the vnode * has been vgone()'d out from underneath us, e.g. when the file * system containing the accounting file has been forcibly unmounted. */ /* ARGSUSED */ static void acctwatch(void *a) { struct statfs sb; if (savacctp != NULLVP) { if (savacctp->v_type == VBAD) { (void) vn_close(savacctp, FWRITE | O_APPEND, NULL); savacctp = NULLVP; return; } (void)VFS_STATFS(savacctp->v_mount, &sb, proc0.p_ucred); if (sb.f_bavail > acctresume * sb.f_blocks / 100) { acctp = savacctp; savacctp = NULLVP; log(LOG_NOTICE, "Accounting resumed\n"); } } else { if (acctp == NULLVP) return; if (acctp->v_type == VBAD) { (void) vn_close(acctp, FWRITE | O_APPEND, NULL); acctp = NULLVP; return; } (void)VFS_STATFS(acctp->v_mount, &sb, proc0.p_ucred); if (sb.f_bavail <= acctsuspend * sb.f_blocks / 100) { savacctp = acctp; acctp = NULLVP; log(LOG_NOTICE, "Accounting suspended\n"); } } callout_reset(&acctwatch_handle, acctchkfreq * hz, acctwatch, NULL); }