debugger.cpp revision e365f9d6543bc6607864ef61257505239dde15d1 
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 <errno.h>
32#include <signal.h>
33#include <stdio.h>
34#include <stdlib.h>
35#include <sys/mman.h>
36#include <sys/prctl.h>
37#include <sys/socket.h>
38#include <sys/un.h>
39#include <unistd.h>
40
41extern "C" int tgkill(int tgid, int tid, int sig);
42
43#define DEBUGGER_SOCKET_NAME "android:debuggerd"
44
45enum debugger_action_t {
46    // dump a crash
47    DEBUGGER_ACTION_CRASH,
48    // dump a tombstone file
49    DEBUGGER_ACTION_DUMP_TOMBSTONE,
50    // dump a backtrace only back to the socket
51    DEBUGGER_ACTION_DUMP_BACKTRACE,
52};
53
54/* message sent over the socket */
55struct debugger_msg_t {
56  // version 1 included:
57  debugger_action_t action;
58  pid_t tid;
59
60  // version 2 added:
61  uintptr_t abort_msg_address;
62};
63
64// see man(2) prctl, specifically the section about PR_GET_NAME
65#define MAX_TASK_NAME_LEN (16)
66
67static int socket_abstract_client(const char* name, int type) {
68    sockaddr_un addr;
69
70    // Test with length +1 for the *initial* '\0'.
71    size_t namelen = strlen(name);
72    if ((namelen + 1) > sizeof(addr.sun_path)) {
73        errno = EINVAL;
74        return -1;
75    }
76
77    /* This is used for abstract socket namespace, we need
78     * an initial '\0' at the start of the Unix socket path.
79     *
80     * Note: The path in this case is *not* supposed to be
81     * '\0'-terminated. ("man 7 unix" for the gory details.)
82     */
83    memset(&addr, 0, sizeof(addr));
84    addr.sun_family = AF_LOCAL;
85    addr.sun_path[0] = 0;
86    memcpy(addr.sun_path + 1, name, namelen);
87
88    socklen_t alen = namelen + offsetof(sockaddr_un, sun_path) + 1;
89
90    int s = socket(AF_LOCAL, type, 0);
91    if (s == -1) {
92        return -1;
93    }
94
95    int err = TEMP_FAILURE_RETRY(connect(s, reinterpret_cast<sockaddr*>(&addr), alen));
96    if (err == -1) {
97        close(s);
98        s = -1;
99    }
100
101    return s;
102}
103
104/*
105 * Writes a summary of the signal to the log file.  We do this so that, if
106 * for some reason we're not able to contact debuggerd, there is still some
107 * indication of the failure in the log.
108 *
109 * We could be here as a result of native heap corruption, or while a
110 * mutex is being held, so we don't want to use any libc functions that
111 * could allocate memory or hold a lock.
112 */
113static void log_signal_summary(int signum, const siginfo_t* info) {
114    const char* signal_name;
115    switch (signum) {
116        case SIGILL:    signal_name = "SIGILL";     break;
117        case SIGABRT:   signal_name = "SIGABRT";    break;
118        case SIGBUS:    signal_name = "SIGBUS";     break;
119        case SIGFPE:    signal_name = "SIGFPE";     break;
120        case SIGSEGV:   signal_name = "SIGSEGV";    break;
121#if defined(SIGSTKFLT)
122        case SIGSTKFLT: signal_name = "SIGSTKFLT";  break;
123#endif
124        case SIGPIPE:   signal_name = "SIGPIPE";    break;
125        default:        signal_name = "???";        break;
126    }
127
128    char thread_name[MAX_TASK_NAME_LEN + 1]; // one more for termination
129    if (prctl(PR_GET_NAME, (unsigned long)thread_name, 0, 0, 0) != 0) {
130        strcpy(thread_name, "<name unknown>");
131    } else {
132        // short names are null terminated by prctl, but the man page
133        // implies that 16 byte names are not.
134        thread_name[MAX_TASK_NAME_LEN] = 0;
135    }
136
137    // "info" will be NULL if the siginfo_t information was not available.
138    if (info != NULL) {
139        __libc_format_log(ANDROID_LOG_FATAL, "libc",
140                          "Fatal signal %d (%s) at %p (code=%d), thread %d (%s)",
141                          signum, signal_name, info->si_addr, info->si_code,
142                          gettid(), thread_name);
143    } else {
144        __libc_format_log(ANDROID_LOG_FATAL, "libc",
145                          "Fatal signal %d (%s), thread %d (%s)",
146                          signum, signal_name, gettid(), thread_name);
147    }
148}
149
150/*
151 * Returns true if the handler for signal "signum" has SA_SIGINFO set.
152 */
153static bool have_siginfo(int signum) {
154    struct sigaction old_action, new_action;
155
156    memset(&new_action, 0, sizeof(new_action));
157    new_action.sa_handler = SIG_DFL;
158    new_action.sa_flags = SA_RESTART;
159    sigemptyset(&new_action.sa_mask);
160
161    if (sigaction(signum, &new_action, &old_action) < 0) {
162      __libc_format_log(ANDROID_LOG_WARN, "libc", "Failed testing for SA_SIGINFO: %s",
163                        strerror(errno));
164      return false;
165    }
166    bool result = (old_action.sa_flags & SA_SIGINFO) != 0;
167
168    if (sigaction(signum, &old_action, NULL) == -1) {
169      __libc_format_log(ANDROID_LOG_WARN, "libc", "Restore failed in test for SA_SIGINFO: %s",
170                        strerror(errno));
171    }
172    return result;
173}
174
175/*
176 * Catches fatal signals so we can ask debuggerd to ptrace us before
177 * we crash.
178 */
179void debuggerd_signal_handler(int n, siginfo_t* info, void*) {
180    /*
181     * It's possible somebody cleared the SA_SIGINFO flag, which would mean
182     * our "info" arg holds an undefined value.
183     */
184    if (!have_siginfo(n)) {
185        info = NULL;
186    }
187
188    log_signal_summary(n, info);
189
190    pid_t tid = gettid();
191    int s = socket_abstract_client(DEBUGGER_SOCKET_NAME, SOCK_STREAM);
192
193    if (s >= 0) {
194        // debuggerd knows our pid from the credentials on the
195        // local socket but we need to tell it the tid of the crashing thread.
196        // debuggerd will be paranoid and verify that we sent a tid
197        // that's actually in our process.
198        debugger_msg_t msg;
199        msg.action = DEBUGGER_ACTION_CRASH;
200        msg.tid = tid;
201        msg.abort_msg_address = reinterpret_cast<uintptr_t>(gAbortMessage);
202        int ret = TEMP_FAILURE_RETRY(write(s, &msg, sizeof(msg)));
203        if (ret == sizeof(msg)) {
204            // if the write failed, there is no point trying to read a response.
205            ret = TEMP_FAILURE_RETRY(read(s, &tid, 1));
206            int saved_errno = errno;
207            notify_gdb_of_libraries();
208            errno = saved_errno;
209        }
210
211        if (ret < 0) {
212            /* read or write failed -- broken connection? */
213            __libc_format_log(ANDROID_LOG_FATAL, "libc", "Failed while talking to debuggerd: %s",
214                              strerror(errno));
215        }
216
217        close(s);
218    } else {
219        /* socket failed; maybe process ran out of fds */
220        __libc_format_log(ANDROID_LOG_FATAL, "libc", "Unable to open connection to debuggerd: %s",
221                          strerror(errno));
222    }
223
224    /* remove our net so we fault for real when we return */
225    signal(n, SIG_DFL);
226
227    /*
228     * These signals are not re-thrown when we resume.  This means that
229     * crashing due to (say) SIGPIPE doesn't work the way you'd expect it
230     * to.  We work around this by throwing them manually.  We don't want
231     * to do this for *all* signals because it'll screw up the address for
232     * faults like SIGSEGV.
233     */
234    switch (n) {
235        case SIGABRT:
236        case SIGFPE:
237        case SIGPIPE:
238#if defined(SIGSTKFLT)
239        case SIGSTKFLT:
240#endif
241            (void) tgkill(getpid(), gettid(), n);
242            break;
243        default:    // SIGILL, SIGBUS, SIGSEGV
244            break;
245    }
246}
247
248void debuggerd_init() {
249    struct sigaction action;
250    memset(&action, 0, sizeof(action));
251    sigemptyset(&action.sa_mask);
252    action.sa_sigaction = debuggerd_signal_handler;
253    action.sa_flags = SA_RESTART | SA_SIGINFO;
254
255    // Use the alternate signal stack if available so we can catch stack overflows.
256    action.sa_flags |= SA_ONSTACK;
257
258    sigaction(SIGABRT, &action, NULL);
259    sigaction(SIGBUS, &action, NULL);
260    sigaction(SIGFPE, &action, NULL);
261    sigaction(SIGILL, &action, NULL);
262    sigaction(SIGPIPE, &action, NULL);
263    sigaction(SIGSEGV, &action, NULL);
264#if defined(SIGSTKFLT)
265    sigaction(SIGSTKFLT, &action, NULL);
266#endif
267    sigaction(SIGTRAP, &action, NULL);
268}
269