Index: head/usr.bin/top/top.1 =================================================================== --- head/usr.bin/top/top.1 (revision 335041) +++ head/usr.bin/top/top.1 (revision 335042) @@ -1,439 +1,440 @@ .\" $FreeBSD$ .Dd June 9, 2018 .Dt TOP 1 .Os .Sh NAME .Nm top .Nd display and update information about the top cpu processes .Sh SYNOPSIS .Nm .Op Fl CHIPSTabijnpqtuvxz .Op Fl J Ar jail .Op Fl U Ar uid .Op Fl d Ar count .Op Fl m Ar cpu|io .Op Fl s Ar time .Op Fl o Ar field .Op Fl p Ar pid .Op Ar count .Sh DESCRIPTION .Nm displays the top processes on the system and periodically updates this information. If standard output is an intelligent terminal (see below) then as many processes as will fit on the terminal screen are displayed by default. Otherwise, a good number of them are shown (around 20). Raw cpu percentage is used to rank the processes. If .Ar number is given, then the top .Ar number processes will be displayed instead of the default. .Pp .Nm makes a distinction between terminals that support advanced capabilities and those that do not. This distinction affects the choice of defaults for certain options. In the remainder of this document, an \*(lqintelligent\*(rq terminal is one that supports cursor addressing, clear screen, and clear to end of line. Conversely, a \*(lqdumb\*(rq terminal is one that does not support such features. If the output of .Nm is redirected to a file, it acts as if it were being run on a dumb terminal. .Bl -tag -width indent -compact .It Fl C Toggle CPU display mode. By default top displays the weighted CPU percentage in the WCPU column (this is the same value that .Xr ps 1 displays as CPU). Each time .It Fl C flag is passed it toggles between \*(lqraw cpu\*(rq mode and \*(lqweighted cpu\*(rq mode, showing the \*(lqCPU\*(rq or the \*(lqWCPU\*(rq column respectively. .It Fl S Show system processes in the display. Normally, system processes such as the pager and the swapper are not shown. This option makes them visible. .It Fl a Display command names derived from the argv[] vector, rather than real executable name. It it useful when you want to watch applications, that puts their status information there. If the real name differs from argv[0], it will be displayed in parenthesis. .It Fl b Use \*(lqbatch\*(rq mode. In this mode, all input from the terminal is ignored. Interrupt characters (such as ^C and ^\e) still have an effect. This is the default on a dumb terminal, or when the output is not a terminal. .It Fl H Display each thread for a multithreaded process individually. By default a single summary line is displayed for each process. .It Fl i Use \*(lqinteractive\*(rq mode. In this mode, any input is immediately read for processing. See the section on \*(lqInteractive Mode\*(rq for an explanation of which keys perform what functions. After the command is processed, the screen will immediately be updated, even if the command was not understood. This mode is the default when standard output is an intelligent terminal. .It Fl I Do not display idle processes. By default, top displays both active and idle processes. .It Fl j Display the .Xr jail 8 ID. .It Fl T Toggle displaying thread ID (tid) instead of process id (pid). .It Fl t Do not display the .Nm process itself. .It Fl display Display either 'cpu' or 'io' statistics. Default is 'cpu'. .It Fl n Use \*(lqnon-interactive\*(rq mode. This is identical to \*(lqbatch\*(rq mode. .It Fl P Display per-cpu CPU usage statistics. .It Fl q Renice .Nm to -20 so that it will run faster. This can be used when the system is being very sluggish to improve the possibility of discovering the problem. This option can only be used by root. .It Fl u Do not map uid numbers to usernames. Normally, .Nm will read as much of the file \*(lq/etc/passwd\*(rq as is necessary to map all the user id numbers it encounters into login names. This option disables all that, while possibly decreasing execution time. The uid numbers are displayed instead of the names. .It Fl v Write version number information to stderr then exit immediately. .It Fl w Display approximate swap usage for each process. .It Fl z Do not display the system idle process. .It Fl d Ar count Show only .Ar count displays, then exit. A display is considered to be one update of the screen. The default is 1 for dumb terminals. Note that for .Ar count = 1 no information is available about the percentage of time spent by the CPU in every state. .It Fl s Ar time Set the delay between screen updates to .Ar time seconds. The default delay between updates is 1 second. .It Fl o Ar field Sort the process display area on the specified field. The field name is the name of the column as seen in the output, but in lower case: \*(lqcpu\*(lq, \*(rqsize\*(lq, \*(rqres\*(lq, \*(rqtime\*(lq, \*(rqpri\*(lq, \*(rqthreads\*(lq, \*(lqtotal\*(lq, \*(rqread\*(lq, \*(rqwrite\*(lq, \*(rqfault\*(lq, \*(rqvcsw\*(lq, \*(rqivcsw\*(lq, \*(lqjid\*(lq, \*(rqswap\*(lq or \*(rqpid\*(lq. .It Fl p Ar pid Show only the process .Ar pid . .It Fl J Ar jail Show only those processes owned by .Ar jail . This may be either the .Ar jid or .Ar name of the jail. Use 0 to limit to host processes. Using this option implies .Fl j . .Pp .It Fl U Ar username Show only those processes owned by .Ar username . This option currently only accepts usernames and will not understand uid numbers. .El .Pp Both .Ar count and .Ar number fields can be specified as \*(lqinfinite\*(rq, indicating that they can stretch as far as possible. This is accomplished by using any proper prefix of the keywords \*(lqinfinity\*(rq, \*(lqmaximum\*(rq, or \*(lqall\*(rq. Boolean flags are toggles. A second specification of any of these options will negate the first. .Sh "INTERACTIVE MODE" When .Nm is running in \*(lqinteractive mode\*(rq, it reads commands from the terminal and acts upon them accordingly. In this mode, the terminal is put in \*(lqCBREAK\*(rq, so that a character will be processed as soon as it is typed. Almost always, a key will be pressed when .Nm is between displays; that is, while it is waiting for .Ar time seconds to elapse. If this is the case, the command will be processed and the display will be updated immediately thereafter (reflecting any changes that the command may have specified). This happens even if the command was incorrect. If a key is pressed while .Nm is in the middle of updating the display, it will finish the update and then process the command. Some commands require additional information, and the user will be prompted accordingly. While typing this information in, the user's erase and kill keys (as set up by the command .Xr stty 1 ) are recognized, and a newline terminates the input. .Pp These commands are currently recognized (^L refers to control-L): .Bl -tag -width indent .It ^L Redraw the screen. .It h Display a summary of the commands (help screen). Version information is included in this display. .It q Quit .Nm .It d Change the number of displays to show (prompt for new number). Remember that the next display counts as one, so typing .It d1 will make .Nm show one final display and then immediately exit. .It m Toggle the display between 'cpu' and 'io' modes. .It n or # Change the number of processes to display (prompt for new number). .It s Change the number of seconds to delay between displays (prompt for new number). .It S Toggle the display of system processes. .It a Toggle the display of process titles. .It k Send a signal (\*(lqkill\*(rq by default) to a list of processes. This acts similarly to the command .Xr kill 1 . .It r Change the priority (the \*(lqnice\*(rq) of a list of processes. This acts similarly to .Xr renice 8 . .It u Display only processes owned by a specific set of usernames (prompt for username). If the username specified is simply \*(lq+\*(rq or \*(lq-\*(rq, then processes belonging to all users will be displayed. Usernames can be added to and removed from the set by prepending them with \*(lq+\*(rq and \*(lq-\*(rq, respectively. .It o Change the order in which the display is sorted. The sort key names include \*(lqcpu\*(rq, \*(lqres\*(rq, \*(lqsize\*(rq, \*(lqtime\*(rq. The default is cpu. .It p Display a specific process (prompt for pid). If the pid specified is simply \*(lq+\*(rq, then show all processes. .It e Display a list of system errors (if any) generated by the last command. .It B H Toggle the display of threads. .It i or I Toggle the display of idle processes. .It j Toggle the display of .Xr jail 8 ID. .It J Display only processes owned by a specific jail (prompt for jail). If the jail specified is simply \*(lq+\*(rq, then processes belonging to all jails and the host will be displayed. This will also enable the display of JID. .It P Toggle the display of per-CPU statistics. .It T Toggle display of TID and PID .It t Toggle the display of the .Nm process. .It w Toggle the display of swap usage. .It z Toggle the display of the system idle process. .El .Sh "THE DISPLAY" The top few lines of the display show general information about the state of the system, including the last process id assigned to a process (on most systems), the three load averages, the current time, the number of existing processes, the number of processes in each state (sleeping, running, starting, zombies, and stopped), and a percentage of time spent in each of the processor states (user, nice, system, and idle). It also includes information about physical and virtual memory allocation. .Pp The remainder of the screen displays information about individual processes. This display is similar in spirit to .Xr ps 1 but it is not exactly the same. PID is the process id, JID, when displayed, is the .Xr jail 8 ID corresponding to the process, USERNAME is the name of the process's owner (if .Fl u is specified, a UID column will be substituted for USERNAME), PRI is the current priority of the process, NICE is the .Xr nice 1 amount, SIZE is the total size of the process (text, data, and stack), RES is the current amount of resident memory, SWAP is the approximate amount of swap, if enabled (SIZE, RES and SWAP are given in kilobytes), STATE is the current state (one of \*(lqSTART\*(rq, \*(lqRUN\*(rq (shown as \*(lqCPUn\*(rq on SMP systems), \*(lqSLEEP\*(rq, \*(lqSTOP\*(rq, \*(lqZOMB\*(rq, \*(lqWAIT\*(rq, \*(lqLOCK\*(rq or the event on which the process waits), C is the processor number on which the process is executing (visible only on SMP systems), TIME is the number of system and user cpu seconds that the process has used, WCPU, when displayed, is the weighted cpu percentage (this is the same value that .Xr ps 1 displays as CPU), CPU is the raw percentage and is the field that is sorted to determine the order of the processes, and COMMAND is the name of the command that the process is currently running (if the process is swapped out, this column is marked \*(lq\*(rq). .Pp If a process is in the \*(lqSLEEP\*(rq or \*(lqLOCK\*(rq state, the state column will report the name of the event or lock on which the process is waiting. Lock names are prefixed with an asterisk \*(lq*\*(rq while sleep events are not. .Sh AUTHORS +.An Eitan Adler Aq Mt eadler@FreeBSD.org .An William LeFebvre, EECS Department, Northwestern University .Sh BUGS The command name for swapped processes should be tracked down, but this would make the program run slower. .Pp As with .Xr ps 1 , things can change while .Nm is collecting information for an update. The picture it gives is only a close approximation to reality. .Sh ENVIRONMENT .Bl -tag -width TOP -compact .It Pa TOP Default set of arguments to .Nm . .El .Sh SEE ALSO .Xr kill 1 , .Xr ps 1 , .Xr stty 1 , .Xr getrusage 2 , .Xr humanize_number 3 , .Xr mem 4 , .Xr renice 8 .Sh DESCRIPTION OF MEMORY Mem: 61M Active, 86M Inact, 368K Laundry, 22G Wired, 102G Free ARC: 15G Total, 9303M MFU, 6155M MRU, 1464K Anon, 98M Header, 35M Other 15G Compressed, 27G Uncompressed, 1.75:1 Ratio, 174M Overhead Swap: 4096M Total, 532M Free, 13% Inuse, 80K In, 104K Out .Sh Physical Memory Stats .Bl -tag -width indent .It Active: number of bytes active .It Inact: number of clean bytes inactive .It Laundry: number of dirty bytes queued for laundering .It Wired: number of bytes wired down, including IO-level cached file data pages .It Buf: number of bytes used for IO-level disk caching .It Free: number of bytes free .It ZFS ARC Stats These stats are only displayed when the ARC is in use. .It Total: number of wired bytes used for the ZFS ARC .It MRU: number of ARC bytes holding most recently used data .It MFU: number of ARC bytes holding most frequently used data .It Anon: number of ARC bytes holding in flight data .It Header: number of ARC bytes holding headers .It Other: miscellaneous ARC bytes .It Compressed: bytes of memory used by ARC caches .It Uncompressed: bytes of data stored in ARC caches before compression .It Ratio: compression ratio of data cached in the ARC .It Swap Stats .It Total: total available swap usage .It Free: total free swap usage .It Inuse: swap usage .It In: bytes paged in from swap devices (last interval) .It Out: bytes paged out to swap devices (last interval) .El Index: head/usr.bin/top/utils.c =================================================================== --- head/usr.bin/top/utils.c (revision 335041) +++ head/usr.bin/top/utils.c (revision 335042) @@ -1,334 +1,335 @@ /* * This program may be freely redistributed, * but this entire comment MUST remain intact. * + * Copyright (c) 2018, Eitan Adler * Copyright (c) 1984, 1989, William LeFebvre, Rice University * Copyright (c) 1989, 1990, 1992, William LeFebvre, Northwestern University * * $FreeBSD$ */ /* * This file contains various handy utilities used by top. */ #include "top.h" #include "utils.h" #include #include #include #include #include #include #include #include #include #include int atoiwi(const char *str) { size_t len; len = strlen(str); if (len != 0) { if (strncmp(str, "infinity", len) == 0 || strncmp(str, "all", len) == 0 || strncmp(str, "maximum", len) == 0) { return(Infinity); } else if (str[0] == '-') { return(Invalid); } else { return((int)strtol(str, NULL, 10)); } } return(0); } /* * itoa - convert integer (decimal) to ascii string for positive numbers * only (we don't bother with negative numbers since we know we * don't use them). */ /* * How do we know that 16 will suffice? * Because the biggest number that we will * ever convert will be 2^32-1, which is 10 * digits. */ _Static_assert(sizeof(int) <= 4, "buffer too small for this sized int"); char * itoa(unsigned int val) { static char buffer[16]; /* result is built here */ /* 16 is sufficient since the largest number we will ever convert will be 2^32-1, which is 10 digits. */ sprintf(buffer, "%u", val); return (buffer); } /* * itoa7(val) - like itoa, except the number is right justified in a 7 * character field. This code is a duplication of itoa instead of * a front end to a more general routine for efficiency. */ char * itoa7(int val) { static char buffer[16]; /* result is built here */ /* 16 is sufficient since the largest number we will ever convert will be 2^32-1, which is 10 digits. */ sprintf(buffer, "%6u", val); return (buffer); } /* * digits(val) - return number of decimal digits in val. Only works for * non-negative numbers. */ int __pure2 digits(int val) { int cnt = 0; if (val == 0) { return 1; } while (val > 0) { cnt++; val /= 10; } return(cnt); } /* * string_index(string, array) - find string in array and return index */ int string_index(const char *string, const char * const *array) { size_t i = 0; while (*array != NULL) { if (strcmp(string, *array) == 0) { return(i); } array++; i++; } return(-1); } /* * argparse(line, cntp) - parse arguments in string "line", separating them * out into an argv-like array, and setting *cntp to the number of * arguments encountered. This is a simple parser that doesn't understand * squat about quotes. */ const char * const * argparse(char *line, int *cntp) { const char **ap; static const char *argv[1024] = {0}; *cntp = 1; ap = &argv[1]; while ((*ap = strsep(&line, " ")) != NULL) { if (**ap != '\0') { (*cntp)++; if (*cntp >= (int)nitems(argv)) { break; } ap++; } } return (argv); } /* * percentages(cnt, out, new, old, diffs) - calculate percentage change * between array "old" and "new", putting the percentages i "out". * "cnt" is size of each array and "diffs" is used for scratch space. * The array "old" is updated on each call. * The routine assumes modulo arithmetic. This function is especially * useful on for calculating cpu state percentages. */ long percentages(int cnt, int *out, long *new, long *old, long *diffs) { int i; long change; long total_change; long *dp; long half_total; /* initialization */ total_change = 0; dp = diffs; /* calculate changes for each state and the overall change */ for (i = 0; i < cnt; i++) { if ((change = *new - *old) < 0) { /* this only happens when the counter wraps */ change = (int) ((unsigned long)*new-(unsigned long)*old); } total_change += (*dp++ = change); *old++ = *new++; } /* avoid divide by zero potential */ if (total_change == 0) { total_change = 1; } /* calculate percentages based on overall change, rounding up */ half_total = total_change / 2l; /* Do not divide by 0. Causes Floating point exception */ if(total_change) { for (i = 0; i < cnt; i++) { *out++ = (int)((*diffs++ * 1000 + half_total) / total_change); } } /* return the total in case the caller wants to use it */ return(total_change); } /* format_time(seconds) - format number of seconds into a suitable * display that will fit within 6 characters. Note that this * routine builds its string in a static area. If it needs * to be called more than once without overwriting previous data, * then we will need to adopt a technique similar to the * one used for format_k. */ /* Explanation: We want to keep the output within 6 characters. For low values we use the format mm:ss. For values that exceed 999:59, we switch to a format that displays hours and fractions: hhh.tH. For values that exceed 999.9, we use hhhh.t and drop the "H" designator. For values that exceed 9999.9, we use "???". */ const char * format_time(long seconds) { static char result[10]; /* sanity protection */ if (seconds < 0 || seconds > (99999l * 360l)) { strcpy(result, " ???"); } else if (seconds >= (1000l * 60l)) { /* alternate (slow) method displaying hours and tenths */ sprintf(result, "%5.1fH", (double)seconds / (double)(60l * 60l)); /* It is possible that the sprintf took more than 6 characters. If so, then the "H" appears as result[6]. If not, then there is a \0 in result[6]. Either way, it is safe to step on. */ result[6] = '\0'; } else { /* standard method produces MMM:SS */ sprintf(result, "%3ld:%02ld", seconds / 60l, seconds % 60l); } return(result); } /* * format_k(amt) - format a kilobyte memory value, returning a string * suitable for display. Returns a pointer to a static * area that changes each call. "amt" is converted to a * string with a trailing "K". If "amt" is 10000 or greater, * then it is formatted as megabytes (rounded) with a * trailing "M". */ /* * Compromise time. We need to return a string, but we don't want the * caller to have to worry about freeing a dynamically allocated string. * Unfortunately, we can't just return a pointer to a static area as one * of the common uses of this function is in a large call to sprintf where * it might get invoked several times. Our compromise is to maintain an * array of strings and cycle thru them with each invocation. We make the * array large enough to handle the above mentioned case. The constant * NUM_STRINGS defines the number of strings in this array: we can tolerate * up to NUM_STRINGS calls before we start overwriting old information. * Keeping NUM_STRINGS a power of two will allow an intelligent optimizer * to convert the modulo operation into something quicker. What a hack! */ #define NUM_STRINGS 8 char * format_k(int64_t amt) { static char retarray[NUM_STRINGS][16]; static int index = 0; char *ret; ret = retarray[index]; index = (index + 1) % NUM_STRINGS; humanize_number(ret, 6, amt * 1024, "", HN_AUTOSCALE, HN_NOSPACE); return (ret); } int find_pid(pid_t pid) { kvm_t *kd = NULL; struct kinfo_proc *pbase = NULL; int nproc; int ret = 0; kd = kvm_open(NULL, _PATH_DEVNULL, NULL, O_RDONLY, NULL); if (kd == NULL) { fprintf(stderr, "top: kvm_open() failed.\n"); quit(TOP_EX_SYS_ERROR); } pbase = kvm_getprocs(kd, KERN_PROC_PID, pid, &nproc); if (pbase == NULL) { goto done; } if ((nproc == 1) && (pbase->ki_pid == pid)) { ret = 1; } done: kvm_close(kd); return ret; }