diff options
Diffstat (limited to 'src/haiku.c')
-rw-r--r-- | src/haiku.c | 286 |
1 files changed, 286 insertions, 0 deletions
diff --git a/src/haiku.c b/src/haiku.c new file mode 100644 index 00000000000..1c786012555 --- /dev/null +++ b/src/haiku.c @@ -0,0 +1,286 @@ +/* Haiku subroutines that are general to the Haiku operating system. + Copyright (C) 2021-2022 Free Software Foundation, Inc. + +This file is part of GNU Emacs. + +GNU Emacs is free software: you can redistribute it and/or modify +it under the terms of the GNU General Public License as published by +the Free Software Foundation, either version 3 of the License, or (at +your option) any later version. + +GNU Emacs is distributed in the hope that it will be useful, +but WITHOUT ANY WARRANTY; without even the implied warranty of +MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +GNU General Public License for more details. + +You should have received a copy of the GNU General Public License +along with GNU Emacs. If not, see <https://www.gnu.org/licenses/>. */ + +#include <config.h> + +#include "lisp.h" +#include "process.h" +#include "coding.h" + +#include <kernel/OS.h> + +#include <pwd.h> +#include <stdlib.h> + +Lisp_Object +list_system_processes (void) +{ + team_info info; + int32 cookie = 0; + Lisp_Object lval = Qnil; + + while (get_next_team_info (&cookie, &info) == B_OK) + lval = Fcons (make_fixnum (info.team), lval); + + return lval; +} + +Lisp_Object +system_process_attributes (Lisp_Object pid) +{ + CHECK_FIXNUM (pid); + + team_info info; + Lisp_Object lval = Qnil; + thread_info inf; + area_info area; + team_id id = (team_id) XFIXNUM (pid); + struct passwd *g; + size_t mem = 0; + + if (get_team_info (id, &info) != B_OK) + return Qnil; + + bigtime_t everything = 0, vsample = 0; + bigtime_t cpu_eaten = 0, esample = 0; + + lval = Fcons (Fcons (Qeuid, make_fixnum (info.uid)), lval); + lval = Fcons (Fcons (Qegid, make_fixnum (info.gid)), lval); + lval = Fcons (Fcons (Qthcount, make_fixnum (info.thread_count)), lval); + lval = Fcons (Fcons (Qcomm, build_string_from_utf8 (info.args)), lval); + + g = getpwuid (info.uid); + + if (g && g->pw_name) + lval = Fcons (Fcons (Quser, build_string (g->pw_name)), lval); + + /* FIXME: Calculating this makes Emacs show up as using 100% CPU! */ + + for (int32 team_cookie = 0; + get_next_team_info (&team_cookie, &info) == B_OK;) + for (int32 thread_cookie = 0; + get_next_thread_info (info.team, &thread_cookie, &inf) == B_OK;) + { + if (inf.team == id && strncmp (inf.name, "idle thread ", 12)) + cpu_eaten += inf.user_time + inf.kernel_time; + everything += inf.user_time + inf.kernel_time; + } + + sleep (0.05); + + for (int32 team_cookie = 0; + get_next_team_info (&team_cookie, &info) == B_OK;) + for (int32 thread_cookie = 0; + get_next_thread_info (info.team, &thread_cookie, &inf) == B_OK;) + { + if (inf.team == id && strncmp (inf.name, "idle thread ", 12)) + esample += inf.user_time + inf.kernel_time; + vsample += inf.user_time + inf.kernel_time; + } + + cpu_eaten = esample - cpu_eaten; + everything = vsample - everything; + + if (everything) + lval = Fcons (Fcons (Qpcpu, make_float (((double) (cpu_eaten) / + (double) (everything)) * 100)), + lval); + else + lval = Fcons (Fcons (Qpcpu, make_float (0.0)), lval); + + for (ssize_t area_cookie = 0; + get_next_area_info (id, &area_cookie, &area) == B_OK;) + mem += area.ram_size; + + system_info sinfo; + get_system_info (&sinfo); + int64 max = (int64) sinfo.max_pages * B_PAGE_SIZE; + + lval = Fcons (Fcons (Qpmem, make_float (((double) mem / + (double) max) * 100)), + lval); + lval = Fcons (Fcons (Qrss, make_fixnum (mem / 1024)), lval); + + return lval; +} + + +/* Borrowed from w32 implementation. */ + +struct load_sample +{ + time_t sample_time; + bigtime_t idle; + bigtime_t kernel; + bigtime_t user; +}; + +/* We maintain 1-sec samples for the last 16 minutes in a circular buffer. */ +static struct load_sample samples[16*60]; +static int first_idx = -1, last_idx = -1; +static int max_idx = ARRAYELTS (samples); +static unsigned num_of_processors = 0; + +static int +buf_next (int from) +{ + int next_idx = from + 1; + + if (next_idx >= max_idx) + next_idx = 0; + + return next_idx; +} + +static int +buf_prev (int from) +{ + int prev_idx = from - 1; + + if (prev_idx < 0) + prev_idx = max_idx - 1; + + return prev_idx; +} + +static double +getavg (int which) +{ + double retval = -1.0; + double tdiff; + int idx; + double span = (which == 0 ? 1.0 : (which == 1 ? 5.0 : 15.0)) * 60; + time_t now = samples[last_idx].sample_time; + + if (first_idx != last_idx) + { + for (idx = buf_prev (last_idx); ; idx = buf_prev (idx)) + { + tdiff = difftime (now, samples[idx].sample_time); + if (tdiff >= span - 2 * DBL_EPSILON * now) + { + long double sys = + (samples[last_idx].kernel + samples[last_idx].user) - + (samples[idx].kernel + samples[idx].user); + long double idl = samples[last_idx].idle - samples[idx].idle; + + retval = (idl / (sys + idl)) * num_of_processors; + break; + } + if (idx == first_idx) + break; + } + } + + return retval; +} + +static void +sample_sys_load (bigtime_t *idle, bigtime_t *system, bigtime_t *user) +{ + bigtime_t i = 0, s = 0, u = 0; + team_info info; + thread_info inf; + + for (int32 team_cookie = 0; + get_next_team_info (&team_cookie, &info) == B_OK;) + for (int32 thread_cookie = 0; + get_next_thread_info (info.team, &thread_cookie, &inf) == B_OK;) + { + if (!strncmp (inf.name, "idle thread ", 12)) + i += inf.user_time + inf.kernel_time; + else + s += inf.kernel_time, u += inf.user_time; + } + + *idle = i; + *system = s; + *user = u; +} + +int +getloadavg (double loadavg[], int nelem) +{ + int elem; + bigtime_t idle, kernel, user; + time_t now = time (NULL); + + if (num_of_processors <= 0) + { + system_info i; + if (get_system_info (&i) == B_OK) + num_of_processors = i.cpu_count; + } + + /* If system time jumped back for some reason, delete all samples + whose time is later than the current wall-clock time. This + prevents load average figures from becoming frozen for prolonged + periods of time, when system time is reset backwards. */ + if (last_idx >= 0) + { + while (difftime (now, samples[last_idx].sample_time) < -1.0) + { + if (last_idx == first_idx) + { + first_idx = last_idx = -1; + break; + } + last_idx = buf_prev (last_idx); + } + } + + /* Store another sample. We ignore samples that are less than 1 sec + apart. */ + if (last_idx < 0 + || (difftime (now, samples[last_idx].sample_time) + >= 1.0 - 2 * DBL_EPSILON * now)) + { + sample_sys_load (&idle, &kernel, &user); + last_idx = buf_next (last_idx); + samples[last_idx].sample_time = now; + samples[last_idx].idle = idle; + samples[last_idx].kernel = kernel; + samples[last_idx].user = user; + /* If the buffer has more that 15 min worth of samples, discard + the old ones. */ + if (first_idx == -1) + first_idx = last_idx; + while (first_idx != last_idx + && (difftime (now, samples[first_idx].sample_time) + >= 15.0 * 60 + 2 * DBL_EPSILON * now)) + first_idx = buf_next (first_idx); + } + + for (elem = 0; elem < nelem; elem++) + { + double avg = getavg (elem); + + if (avg < 0) + break; + loadavg[elem] = avg; + } + + /* Always return at least one element, otherwise load-average + returns nil, and Lisp programs might decide we cannot measure + system load. For example, jit-lock-stealth-load's defcustom + might decide that feature is "unsupported". */ + if (elem == 0) + loadavg[elem++] = 0.09; /* < display-time-load-average-threshold */ + + return elem; +} |