procfs_linux.c revision 1.34 1 /* $NetBSD: procfs_linux.c,v 1.34 2007/04/01 03:16:44 christos Exp $ */
2
3 /*
4 * Copyright (c) 2001 Wasabi Systems, Inc.
5 * All rights reserved.
6 *
7 * Written by Frank van der Linden for Wasabi Systems, Inc.
8 *
9 * Redistribution and use in source and binary forms, with or without
10 * modification, are permitted provided that the following conditions
11 * are met:
12 * 1. Redistributions of source code must retain the above copyright
13 * notice, this list of conditions and the following disclaimer.
14 * 2. Redistributions in binary form must reproduce the above copyright
15 * notice, this list of conditions and the following disclaimer in the
16 * documentation and/or other materials provided with the distribution.
17 * 3. All advertising materials mentioning features or use of this software
18 * must display the following acknowledgement:
19 * This product includes software developed for the NetBSD Project by
20 * Wasabi Systems, Inc.
21 * 4. The name of Wasabi Systems, Inc. may not be used to endorse
22 * or promote products derived from this software without specific prior
23 * written permission.
24 *
25 * THIS SOFTWARE IS PROVIDED BY WASABI SYSTEMS, INC. ``AS IS'' AND
26 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED
27 * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
28 * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL WASABI SYSTEMS, INC
29 * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
30 * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
31 * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
32 * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
33 * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
34 * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
35 * POSSIBILITY OF SUCH DAMAGE.
36 */
37
38 #include <sys/cdefs.h>
39 __KERNEL_RCSID(0, "$NetBSD: procfs_linux.c,v 1.34 2007/04/01 03:16:44 christos Exp $");
40
41 #include <sys/param.h>
42 #include <sys/systm.h>
43 #include <sys/time.h>
44 #include <sys/kernel.h>
45 #include <sys/proc.h>
46 #include <sys/vnode.h>
47 #include <sys/exec.h>
48 #include <sys/resource.h>
49 #include <sys/resourcevar.h>
50 #include <sys/signal.h>
51 #include <sys/signalvar.h>
52 #include <sys/tty.h>
53 #include <sys/malloc.h>
54 #include <sys/mount.h>
55 #include <sys/conf.h>
56
57 #include <miscfs/procfs/procfs.h>
58 #include <compat/linux/common/linux_exec.h>
59
60 #include <uvm/uvm_extern.h>
61 #include <uvm/uvm.h>
62
63 extern struct devsw_conv *devsw_conv;
64 extern int max_devsw_convs;
65
66 #define PGTOB(p) ((unsigned long)(p) << PAGE_SHIFT)
67 #define PGTOKB(p) ((unsigned long)(p) << (PAGE_SHIFT - 10))
68
69 #define LBFSZ (8 * 1024)
70
71 /*
72 * Linux compatible /proc/meminfo. Only active when the -o linux
73 * mountflag is used.
74 */
75 int
76 procfs_domeminfo(struct lwp *curl, struct proc *p,
77 struct pfsnode *pfs, struct uio *uio)
78 {
79 char *bf;
80 int len;
81 int error = 0;
82
83 bf = malloc(LBFSZ, M_TEMP, M_WAITOK);
84
85 len = snprintf(bf, LBFSZ,
86 " total: used: free: shared: buffers: cached:\n"
87 "Mem: %8lu %8lu %8lu %8lu %8lu %8lu\n"
88 "Swap: %8lu %8lu %8lu\n"
89 "MemTotal: %8lu kB\n"
90 "MemFree: %8lu kB\n"
91 "MemShared: %8lu kB\n"
92 "Buffers: %8lu kB\n"
93 "Cached: %8lu kB\n"
94 "SwapTotal: %8lu kB\n"
95 "SwapFree: %8lu kB\n",
96 PGTOB(uvmexp.npages),
97 PGTOB(uvmexp.npages - uvmexp.free),
98 PGTOB(uvmexp.free),
99 0L,
100 PGTOB(uvmexp.filepages),
101 PGTOB(uvmexp.anonpages + uvmexp.filepages + uvmexp.execpages),
102 PGTOB(uvmexp.swpages),
103 PGTOB(uvmexp.swpginuse),
104 PGTOB(uvmexp.swpages - uvmexp.swpginuse),
105 PGTOKB(uvmexp.npages),
106 PGTOKB(uvmexp.free),
107 0L,
108 PGTOKB(uvmexp.filepages),
109 PGTOKB(uvmexp.anonpages + uvmexp.filepages + uvmexp.execpages),
110 PGTOKB(uvmexp.swpages),
111 PGTOKB(uvmexp.swpages - uvmexp.swpginuse));
112
113 if (len == 0)
114 goto out;
115
116 error = uiomove_frombuf(bf, len, uio);
117 out:
118 free(bf, M_TEMP);
119 return error;
120 }
121
122 /*
123 * Linux compatible /proc/devices. Only active when the -o linux
124 * mountflag is used.
125 */
126 int
127 procfs_dodevices(struct lwp *curl, struct proc *p,
128 struct pfsnode *pfs, struct uio *uio)
129 {
130 char *bf;
131 int offset = 0;
132 int i, error = ENAMETOOLONG;
133
134 /* XXX elad - may need filtering. */
135
136 bf = malloc(LBFSZ, M_TEMP, M_WAITOK);
137
138 offset += snprintf(&bf[offset], LBFSZ - offset, "Character devices:\n");
139 if (offset >= LBFSZ)
140 goto out;
141
142 for (i = 0; i < max_devsw_convs; i++) {
143 if ((devsw_conv[i].d_name == NULL) ||
144 (devsw_conv[i].d_cmajor == -1))
145 continue;
146
147 offset += snprintf(&bf[offset], LBFSZ - offset,
148 "%3d %s\n", devsw_conv[i].d_cmajor, devsw_conv[i].d_name);
149 if (offset >= LBFSZ)
150 goto out;
151 }
152
153 offset += snprintf(&bf[offset], LBFSZ - offset, "\nBlock devices:\n");
154 if (offset >= LBFSZ)
155 goto out;
156
157 for (i = 0; i < max_devsw_convs; i++) {
158 if ((devsw_conv[i].d_name == NULL) ||
159 (devsw_conv[i].d_bmajor == -1))
160 continue;
161
162 offset += snprintf(&bf[offset], LBFSZ - offset,
163 "%3d %s\n", devsw_conv[i].d_bmajor, devsw_conv[i].d_name);
164 if (offset >= LBFSZ)
165 goto out;
166 }
167
168 error = uiomove_frombuf(bf, offset, uio);
169 out:
170 free(bf, M_TEMP);
171 return error;
172 }
173
174 /*
175 * Linux compatible /proc/<pid>/stat. Only active when the -o linux
176 * mountflag is used.
177 */
178 int
179 procfs_do_pid_stat(struct lwp *curl, struct lwp *l,
180 struct pfsnode *pfs, struct uio *uio)
181 {
182 char *bf;
183 struct proc *p = l->l_proc;
184 int len;
185 struct tty *tty = p->p_session->s_ttyp;
186 struct rusage *ru = &p->p_stats->p_ru;
187 struct rusage *cru = &p->p_stats->p_cru;
188 struct vmspace *vm;
189 struct vm_map *map;
190 struct vm_map_entry *entry;
191 unsigned long stext = 0, etext = 0, sstack = 0;
192 struct timeval rt;
193 int error = 0;
194
195 bf = malloc(LBFSZ, M_TEMP, M_WAITOK);
196
197 proc_vmspace_getref(p, &vm);
198 map = &vm->vm_map;
199 vm_map_lock_read(map);
200
201 for (entry = map->header.next; entry != &map->header;
202 entry = entry->next) {
203 if (UVM_ET_ISSUBMAP(entry))
204 continue;
205 /* assume text is the first entry */
206 if (stext == etext) {
207 stext = entry->start;
208 etext = entry->end;
209 break;
210 }
211 }
212 #ifdef LINUX_USRSTACK
213 if (strcmp(p->p_emul->e_name, "linux") == 0 &&
214 LINUX_USRSTACK < USRSTACK)
215 sstack = (unsigned long) LINUX_USRSTACK;
216 else
217 #endif
218 sstack = (unsigned long) USRSTACK;
219
220 /*
221 * jdk 1.6 compares low <= addr && addr < high
222 * if we put addr == high, then the test fails
223 * so eat one page.
224 */
225 sstack -= PAGE_SIZE;
226
227 vm_map_unlock_read(map);
228 uvmspace_free(vm);
229
230 mutex_enter(&proclist_lock);
231 mutex_enter(&p->p_mutex);
232 mutex_enter(&p->p_smutex);
233
234 calcru(p, NULL, NULL, NULL, &rt);
235
236 len = snprintf(bf, LBFSZ,
237 "%d (%s) %c %d %d %d %d %d "
238 "%u "
239 "%lu %lu %lu %lu %lu %lu %lu %lu "
240 "%d %d %d "
241 "%lu %lu %lu %lu %" PRIu64 " "
242 "%lu %lu %lu "
243 "%u %u "
244 "%u %u %u %u "
245 "%lu %lu %lu %d %d\n",
246
247 p->p_pid,
248 p->p_comm,
249 "0IR3SZD"[(p->p_stat > 6) ? 0 : (int)p->p_stat],
250 (p->p_pptr != NULL) ? p->p_pptr->p_pid : 0,
251
252 p->p_pgid,
253 p->p_session->s_sid,
254 tty ? tty->t_dev : 0,
255 (tty && tty->t_pgrp) ? tty->t_pgrp->pg_id : 0,
256
257 p->p_flag,
258
259 ru->ru_minflt,
260 cru->ru_minflt,
261 ru->ru_majflt,
262 cru->ru_majflt,
263 ru->ru_utime.tv_sec,
264 ru->ru_stime.tv_sec,
265 cru->ru_utime.tv_sec,
266 cru->ru_stime.tv_sec,
267
268 p->p_nice, /* XXX: priority */
269 p->p_nice,
270 0,
271
272 rt.tv_sec,
273 p->p_stats->p_start.tv_sec,
274 ru->ru_ixrss + ru->ru_idrss + ru->ru_isrss,
275 ru->ru_maxrss,
276 p->p_rlimit[RLIMIT_RSS].rlim_cur,
277
278 stext, /* start code */
279 etext, /* end code */
280 sstack, /* mm start stack */
281 0, /* XXX: pc */
282 0, /* XXX: sp */
283 p->p_sigpend.sp_set.__bits[0], /* XXX: pending */
284 0, /* XXX: held */
285 p->p_sigctx.ps_sigignore.__bits[0], /* ignored */
286 p->p_sigctx.ps_sigcatch.__bits[0], /* caught */
287
288 (unsigned long)(intptr_t)l->l_wchan,
289 ru->ru_nvcsw,
290 ru->ru_nivcsw,
291 p->p_exitsig,
292 0); /* XXX: processor */
293
294 mutex_exit(&p->p_smutex);
295 mutex_exit(&p->p_mutex);
296 mutex_exit(&proclist_lock);
297
298 if (len == 0)
299 goto out;
300
301 error = uiomove_frombuf(bf, len, uio);
302 out:
303 free(bf, M_TEMP);
304 return error;
305 }
306
307 int
308 procfs_docpuinfo(struct lwp *curl, struct proc *p,
309 struct pfsnode *pfs, struct uio *uio)
310 {
311 int len = LBFSZ;
312 char *bf = malloc(len, M_TEMP, M_WAITOK);
313 int error;
314
315 if (procfs_getcpuinfstr(bf, &len) < 0) {
316 error = ENOSPC;
317 goto done;
318 }
319
320 if (len == 0) {
321 error = 0;
322 goto done;
323 }
324
325 error = uiomove_frombuf(bf, len, uio);
326 done:
327 free(bf, M_TEMP);
328 return error;
329 }
330
331 int
332 procfs_douptime(struct lwp *curl, struct proc *p,
333 struct pfsnode *pfs, struct uio *uio)
334 {
335 char *bf;
336 int len;
337 struct timeval runtime;
338 u_int64_t idle;
339 int error = 0;
340
341 bf = malloc(LBFSZ, M_TEMP, M_WAITOK);
342
343 timersub(&curcpu()->ci_schedstate.spc_runtime, &boottime, &runtime);
344 idle = curcpu()->ci_schedstate.spc_cp_time[CP_IDLE];
345 len = snprintf(bf, LBFSZ,
346 "%lu.%02lu %" PRIu64 ".%02" PRIu64 "\n",
347 runtime.tv_sec, runtime.tv_usec / 10000,
348 idle / hz, (((idle % hz) * 100) / hz) % 100);
349
350 if (len == 0)
351 goto out;
352
353 error = uiomove_frombuf(bf, len, uio);
354 out:
355 free(bf, M_TEMP);
356 return error;
357 }
358
359 int
360 procfs_domounts(struct lwp *curl, struct proc *p,
361 struct pfsnode *pfs, struct uio *uio)
362 {
363 char *bf, *mtab = NULL;
364 const char *fsname;
365 size_t len, mtabsz = 0;
366 struct mount *mp, *nmp;
367 struct statvfs *sfs;
368 int error = 0;
369
370 /* XXX elad - may need filtering. */
371
372 bf = malloc(LBFSZ, M_TEMP, M_WAITOK);
373 simple_lock(&mountlist_slock);
374 for (mp = CIRCLEQ_FIRST(&mountlist); mp != (void *)&mountlist;
375 mp = nmp) {
376 if (vfs_busy(mp, LK_NOWAIT, &mountlist_slock)) {
377 nmp = CIRCLEQ_NEXT(mp, mnt_list);
378 continue;
379 }
380
381 sfs = &mp->mnt_stat;
382
383 /* Linux uses different names for some filesystems */
384 fsname = sfs->f_fstypename;
385 if (strcmp(fsname, "procfs") == 0)
386 fsname = "proc";
387 else if (strcmp(fsname, "ext2fs") == 0)
388 fsname = "ext2";
389
390 len = snprintf(bf, LBFSZ, "%s %s %s %s%s%s%s%s%s 0 0\n",
391 sfs->f_mntfromname,
392 sfs->f_mntonname,
393 fsname,
394 (mp->mnt_flag & MNT_RDONLY) ? "ro" : "rw",
395 (mp->mnt_flag & MNT_NOSUID) ? ",nosuid" : "",
396 (mp->mnt_flag & MNT_NOEXEC) ? ",noexec" : "",
397 (mp->mnt_flag & MNT_NODEV) ? ",nodev" : "",
398 (mp->mnt_flag & MNT_SYNCHRONOUS) ? ",sync" : "",
399 (mp->mnt_flag & MNT_NOATIME) ? ",noatime" : ""
400 );
401
402 mtab = realloc(mtab, mtabsz + len, M_TEMP, M_WAITOK);
403 memcpy(mtab + mtabsz, bf, len);
404 mtabsz += len;
405
406 simple_lock(&mountlist_slock);
407 nmp = CIRCLEQ_NEXT(mp, mnt_list);
408 vfs_unbusy(mp);
409 }
410 simple_unlock(&mountlist_slock);
411 free(bf, M_TEMP);
412
413 if (mtabsz > 0) {
414 error = uiomove_frombuf(mtab, mtabsz, uio);
415 free(mtab, M_TEMP);
416 }
417
418 return error;
419 }
420