debugger.cpp revision d14dc3b87fbf80553f1cafa453816b7f11366627 
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
31#include <stdio.h>
32#include <stdlib.h>
33#include <unistd.h>
34#include <signal.h>
35#include <sys/prctl.h>
36#include <errno.h>
37#include <sys/socket.h>
38#include <sys/un.h>
39
40extern "C" int tgkill(int tgid, int tid, int sig);
41
42#define DEBUGGER_SOCKET_NAME "android:debuggerd"
43
44enum debugger_action_t {
45    // dump a crash
46    DEBUGGER_ACTION_CRASH,
47    // dump a tombstone file
48    DEBUGGER_ACTION_DUMP_TOMBSTONE,
49    // dump a backtrace only back to the socket
50    DEBUGGER_ACTION_DUMP_BACKTRACE,
51};
52
53/* message sent over the socket */
54struct debugger_msg_t {
55  // version 1 included:
56  debugger_action_t action;
57  pid_t tid;
58
59  // version 2 added:
60  uintptr_t abort_msg_address;
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* signal_name;
114    switch (signum) {
115        case SIGILL:    signal_name = "SIGILL";     break;
116        case SIGABRT:   signal_name = "SIGABRT";    break;
117        case SIGBUS:    signal_name = "SIGBUS";     break;
118        case SIGFPE:    signal_name = "SIGFPE";     break;
119        case SIGSEGV:   signal_name = "SIGSEGV";    break;
120#if defined(SIGSTKFLT)
121        case SIGSTKFLT: signal_name = "SIGSTKFLT";  break;
122#endif
123        case SIGPIPE:   signal_name = "SIGPIPE";    break;
124        default:        signal_name = "???";        break;
125    }
126
127    char thread_name[MAX_TASK_NAME_LEN + 1]; // one more for termination
128    if (prctl(PR_GET_NAME, (unsigned long)thread_name, 0, 0, 0) != 0) {
129        strcpy(thread_name, "<name unknown>");
130    } else {
131        // short names are null terminated by prctl, but the man page
132        // implies that 16 byte names are not.
133        thread_name[MAX_TASK_NAME_LEN] = 0;
134    }
135
136    // "info" will be NULL if the siginfo_t information was not available.
137    if (info != NULL) {
138        __libc_format_log(ANDROID_LOG_FATAL, "libc",
139                          "Fatal signal %d (%s) at 0x%08x (code=%d), thread %d (%s)",
140                          signum, signal_name, reinterpret_cast<uintptr_t>(info->si_addr),
141                          info->si_code, gettid(), thread_name);
142    } else {
143        __libc_format_log(ANDROID_LOG_FATAL, "libc",
144                          "Fatal signal %d (%s), thread %d (%s)",
145                          signum, signal_name, gettid(), thread_name);
146    }
147}
148
149/*
150 * Returns true if the handler for signal "signum" has SA_SIGINFO set.
151 */
152static bool haveSiginfo(int signum) {
153    struct sigaction oldact, newact;
154
155    memset(&newact, 0, sizeof(newact));
156    newact.sa_handler = SIG_DFL;
157    newact.sa_flags = SA_RESTART;
158    sigemptyset(&newact.sa_mask);
159
160    if (sigaction(signum, &newact, &oldact) < 0) {
161      __libc_format_log(ANDROID_LOG_WARN, "libc", "Failed testing for SA_SIGINFO: %s",
162                        strerror(errno));
163      return false;
164    }
165    bool ret = (oldact.sa_flags & SA_SIGINFO) != 0;
166
167    if (sigaction(signum, &oldact, NULL) == -1) {
168      __libc_format_log(ANDROID_LOG_WARN, "libc", "Restore failed in test for SA_SIGINFO: %s",
169                        strerror(errno));
170    }
171    return ret;
172}
173
174/*
175 * Catches fatal signals so we can ask debuggerd to ptrace us before
176 * we crash.
177 */
178void debuggerd_signal_handler(int n, siginfo_t* info, void*) {
179    /*
180     * It's possible somebody cleared the SA_SIGINFO flag, which would mean
181     * our "info" arg holds an undefined value.
182     */
183    if (!haveSiginfo(n)) {
184        info = NULL;
185    }
186
187    logSignalSummary(n, info);
188
189    pid_t tid = gettid();
190    int s = socket_abstract_client(DEBUGGER_SOCKET_NAME, SOCK_STREAM);
191
192    if (s >= 0) {
193        // debuggerd knows our pid from the credentials on the
194        // local socket but we need to tell it the tid of the crashing thread.
195        // debuggerd will be paranoid and verify that we sent a tid
196        // that's actually in our process.
197        debugger_msg_t msg;
198        msg.action = DEBUGGER_ACTION_CRASH;
199        msg.tid = tid;
200        msg.abort_msg_address = reinterpret_cast<uintptr_t>(gAbortMessage);
201        int ret = TEMP_FAILURE_RETRY(write(s, &msg, sizeof(msg)));
202        if (ret == sizeof(msg)) {
203            // if the write failed, there is no point trying to read a response.
204            ret = TEMP_FAILURE_RETRY(read(s, &tid, 1));
205            int saved_errno = errno;
206            notify_gdb_of_libraries();
207            errno = saved_errno;
208        }
209
210        if (ret < 0) {
211            /* read or write failed -- broken connection? */
212            __libc_format_log(ANDROID_LOG_FATAL, "libc", "Failed while talking to debuggerd: %s",
213                              strerror(errno));
214        }
215
216        close(s);
217    } else {
218        /* socket failed; maybe process ran out of fds */
219        __libc_format_log(ANDROID_LOG_FATAL, "libc", "Unable to open connection to debuggerd: %s",
220                          strerror(errno));
221    }
222
223    /* remove our net so we fault for real when we return */
224    signal(n, SIG_DFL);
225
226    /*
227     * These signals are not re-thrown when we resume.  This means that
228     * crashing due to (say) SIGPIPE doesn't work the way you'd expect it
229     * to.  We work around this by throwing them manually.  We don't want
230     * to do this for *all* signals because it'll screw up the address for
231     * faults like SIGSEGV.
232     */
233    switch (n) {
234        case SIGABRT:
235        case SIGFPE:
236        case SIGPIPE:
237#ifdef SIGSTKFLT
238        case SIGSTKFLT:
239#endif
240            (void) tgkill(getpid(), gettid(), n);
241            break;
242        default:    // SIGILL, SIGBUS, SIGSEGV
243            break;
244    }
245}
246
247void debuggerd_init() {
248    struct sigaction act;
249    memset(&act, 0, sizeof(act));
250    act.sa_sigaction = debuggerd_signal_handler;
251    act.sa_flags = SA_RESTART | SA_SIGINFO;
252    sigemptyset(&act.sa_mask);
253
254    sigaction(SIGILL, &act, NULL);
255    sigaction(SIGABRT, &act, NULL);
256    sigaction(SIGBUS, &act, NULL);
257    sigaction(SIGFPE, &act, NULL);
258    sigaction(SIGSEGV, &act, NULL);
259#if defined(SIGSTKFLT)
260    sigaction(SIGSTKFLT, &act, NULL);
261#endif
262    sigaction(SIGPIPE, &act, NULL);
263}
264