debugger.cpp revision 18a206c81d9743481e364384affd43306911283d 
1/*
2 * Copyright (C) 2008 The Android Open Source Project
3 * All rights reserved.
4 *
5 * Redistribution and use in source and binary forms, with or without
6 * modification, are permitted provided that the following conditions
7 * are met:
8 *  * Redistributions of source code must retain the above copyright
9 *    notice, this list of conditions and the following disclaimer.
10 *  * Redistributions in binary form must reproduce the above copyright
11 *    notice, this list of conditions and the following disclaimer in
12 *    the documentation and/or other materials provided with the
13 *    distribution.
14 *
15 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
16 * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
17 * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS
18 * FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE
19 * COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT,
20 * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING,
21 * BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS
22 * OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED
23 * AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
24 * OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT
25 * OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
26 * SUCH DAMAGE.
27 */
28
29#include "linker.h"
30#include "linker_format.h"
31
32#include <stdio.h>
33#include <stdlib.h>
34#include <unistd.h>
35#include <stdbool.h>
36#include <signal.h>
37#include <sys/prctl.h>
38#include <errno.h>
39#include <sys/socket.h>
40#include <sys/un.h>
41
42#include <private/logd.h>
43
44extern "C" int tgkill(int tgid, int tid, int sig);
45
46#define DEBUGGER_SOCKET_NAME "android:debuggerd"
47
48enum debugger_action_t {
49    // dump a crash
50    DEBUGGER_ACTION_CRASH,
51    // dump a tombstone file
52    DEBUGGER_ACTION_DUMP_TOMBSTONE,
53    // dump a backtrace only back to the socket
54    DEBUGGER_ACTION_DUMP_BACKTRACE,
55};
56
57/* message sent over the socket */
58struct debugger_msg_t {
59    debugger_action_t action;
60    pid_t tid;
61};
62
63// see man(2) prctl, specifically the section about PR_GET_NAME
64#define MAX_TASK_NAME_LEN (16)
65
66static int socket_abstract_client(const char* name, int type) {
67    sockaddr_un addr;
68
69    // Test with length +1 for the *initial* '\0'.
70    size_t namelen = strlen(name);
71    if ((namelen + 1) > sizeof(addr.sun_path)) {
72        errno = EINVAL;
73        return -1;
74    }
75
76    /* This is used for abstract socket namespace, we need
77     * an initial '\0' at the start of the Unix socket path.
78     *
79     * Note: The path in this case is *not* supposed to be
80     * '\0'-terminated. ("man 7 unix" for the gory details.)
81     */
82    memset(&addr, 0, sizeof(addr));
83    addr.sun_family = AF_LOCAL;
84    addr.sun_path[0] = 0;
85    memcpy(addr.sun_path + 1, name, namelen);
86
87    socklen_t alen = namelen + offsetof(sockaddr_un, sun_path) + 1;
88
89    int s = socket(AF_LOCAL, type, 0);
90    if (s == -1) {
91        return -1;
92    }
93
94    int err = TEMP_FAILURE_RETRY(connect(s, (sockaddr*) &addr, alen));
95    if (err == -1) {
96        close(s);
97        s = -1;
98    }
99
100    return s;
101}
102
103/*
104 * Writes a summary of the signal to the log file.  We do this so that, if
105 * for some reason we're not able to contact debuggerd, there is still some
106 * indication of the failure in the log.
107 *
108 * We could be here as a result of native heap corruption, or while a
109 * mutex is being held, so we don't want to use any libc functions that
110 * could allocate memory or hold a lock.
111 */
112static void logSignalSummary(int signum, const siginfo_t* info) {
113    const char* signame;
114    switch (signum) {
115        case SIGILL:    signame = "SIGILL";     break;
116        case SIGABRT:   signame = "SIGABRT";    break;
117        case SIGBUS:    signame = "SIGBUS";     break;
118        case SIGFPE:    signame = "SIGFPE";     break;
119        case SIGSEGV:   signame = "SIGSEGV";    break;
120#if defined(SIGSTKFLT)
121        case SIGSTKFLT: signame = "SIGSTKFLT";  break;
122#endif
123        case SIGPIPE:   signame = "SIGPIPE";    break;
124        default:        signame = "???";        break;
125    }
126
127    char threadname[MAX_TASK_NAME_LEN + 1]; // one more for termination
128    if (prctl(PR_GET_NAME, (unsigned long)threadname, 0, 0, 0) != 0) {
129        strcpy(threadname, "<name unknown>");
130    } else {
131        // short names are null terminated by prctl, but the manpage
132        // implies that 16 byte names are not.
133        threadname[MAX_TASK_NAME_LEN] = 0;
134    }
135
136    char buffer[128];
137    // "info" will be NULL if the siginfo_t information was not available.
138    if (info != NULL) {
139        format_buffer(buffer, sizeof(buffer),
140            "Fatal signal %d (%s) at 0x%08x (code=%d), thread %d (%s)",
141            signum, signame, info->si_addr, info->si_code, gettid(), threadname);
142    } else {
143        format_buffer(buffer, sizeof(buffer),
144            "Fatal signal %d (%s), thread %d (%s)",
145            signum, signame, gettid(), threadname);
146    }
147
148    __libc_android_log_write(ANDROID_LOG_FATAL, "libc", buffer);
149}
150
151/*
152 * Returns true if the handler for signal "signum" has SA_SIGINFO set.
153 */
154static bool haveSiginfo(int signum) {
155    struct sigaction oldact, newact;
156
157    memset(&newact, 0, sizeof(newact));
158    newact.sa_handler = SIG_DFL;
159    newact.sa_flags = SA_RESTART;
160    sigemptyset(&newact.sa_mask);
161
162    if (sigaction(signum, &newact, &oldact) < 0) {
163        __libc_android_log_write(ANDROID_LOG_FATAL, "libc",
164            "Failed testing for SA_SIGINFO");
165        return 0;
166    }
167    bool ret = (oldact.sa_flags & SA_SIGINFO) != 0;
168
169    if (sigaction(signum, &oldact, NULL) < 0) {
170        __libc_android_log_write(ANDROID_LOG_FATAL, "libc",
171            "Restore failed in test for SA_SIGINFO");
172    }
173    return ret;
174}
175
176/*
177 * Catches fatal signals so we can ask debuggerd to ptrace us before
178 * we crash.
179 */
180void debugger_signal_handler(int n, siginfo_t* info, void*) {
181    char msgbuf[128];
182
183    /*
184     * It's possible somebody cleared the SA_SIGINFO flag, which would mean
185     * our "info" arg holds an undefined value.
186     */
187    if (!haveSiginfo(n)) {
188        info = NULL;
189    }
190
191    logSignalSummary(n, info);
192
193    pid_t tid = gettid();
194    int s = socket_abstract_client(DEBUGGER_SOCKET_NAME, SOCK_STREAM);
195
196    if (s >= 0) {
197        /* debugger knows our pid from the credentials on the
198         * local socket but we need to tell it our tid.  It
199         * is paranoid and will verify that we are giving a tid
200         * that's actually in our process
201         */
202        int  ret;
203        debugger_msg_t msg;
204        msg.action = DEBUGGER_ACTION_CRASH;
205        msg.tid = tid;
206        ret = TEMP_FAILURE_RETRY(write(s, &msg, sizeof(msg)));
207        if (ret == sizeof(msg)) {
208            /* if the write failed, there is no point to read on
209             * the file descriptor. */
210            ret = TEMP_FAILURE_RETRY(read(s, &tid, 1));
211            int saved_errno = errno;
212            notify_gdb_of_libraries();
213            errno = saved_errno;
214        }
215
216        if (ret < 0) {
217            /* read or write failed -- broken connection? */
218            format_buffer(msgbuf, sizeof(msgbuf),
219                "Failed while talking to debuggerd: %s", strerror(errno));
220            __libc_android_log_write(ANDROID_LOG_FATAL, "libc", msgbuf);
221        }
222
223        close(s);
224    } else {
225        /* socket failed; maybe process ran out of fds */
226        format_buffer(msgbuf, sizeof(msgbuf),
227            "Unable to open connection to debuggerd: %s", strerror(errno));
228        __libc_android_log_write(ANDROID_LOG_FATAL, "libc", msgbuf);
229    }
230
231    /* remove our net so we fault for real when we return */
232    signal(n, SIG_DFL);
233
234    /*
235     * These signals are not re-thrown when we resume.  This means that
236     * crashing due to (say) SIGPIPE doesn't work the way you'd expect it
237     * to.  We work around this by throwing them manually.  We don't want
238     * to do this for *all* signals because it'll screw up the address for
239     * faults like SIGSEGV.
240     */
241    switch (n) {
242        case SIGABRT:
243        case SIGFPE:
244        case SIGPIPE:
245#ifdef SIGSTKFLT
246        case SIGSTKFLT:
247#endif
248            (void) tgkill(getpid(), gettid(), n);
249            break;
250        default:    // SIGILL, SIGBUS, SIGSEGV
251            break;
252    }
253}
254
255void debugger_init() {
256    struct sigaction act;
257    memset(&act, 0, sizeof(act));
258    act.sa_sigaction = debugger_signal_handler;
259    act.sa_flags = SA_RESTART | SA_SIGINFO;
260    sigemptyset(&act.sa_mask);
261
262    sigaction(SIGILL, &act, NULL);
263    sigaction(SIGABRT, &act, NULL);
264    sigaction(SIGBUS, &act, NULL);
265    sigaction(SIGFPE, &act, NULL);
266    sigaction(SIGSEGV, &act, NULL);
267#if defined(SIGSTKFLT)
268    sigaction(SIGSTKFLT, &act, NULL);
269#endif
270    sigaction(SIGPIPE, &act, NULL);
271}
272