debugger.cpp revision 8f2a5a0b40fc82126c691d5c30131d908772aab7 
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    debugger_action_t action;
56    pid_t tid;
57};
58
59// see man(2) prctl, specifically the section about PR_GET_NAME
60#define MAX_TASK_NAME_LEN (16)
61
62static int socket_abstract_client(const char* name, int type) {
63    sockaddr_un addr;
64
65    // Test with length +1 for the *initial* '\0'.
66    size_t namelen = strlen(name);
67    if ((namelen + 1) > sizeof(addr.sun_path)) {
68        errno = EINVAL;
69        return -1;
70    }
71
72    /* This is used for abstract socket namespace, we need
73     * an initial '\0' at the start of the Unix socket path.
74     *
75     * Note: The path in this case is *not* supposed to be
76     * '\0'-terminated. ("man 7 unix" for the gory details.)
77     */
78    memset(&addr, 0, sizeof(addr));
79    addr.sun_family = AF_LOCAL;
80    addr.sun_path[0] = 0;
81    memcpy(addr.sun_path + 1, name, namelen);
82
83    socklen_t alen = namelen + offsetof(sockaddr_un, sun_path) + 1;
84
85    int s = socket(AF_LOCAL, type, 0);
86    if (s == -1) {
87        return -1;
88    }
89
90    int err = TEMP_FAILURE_RETRY(connect(s, (sockaddr*) &addr, alen));
91    if (err == -1) {
92        close(s);
93        s = -1;
94    }
95
96    return s;
97}
98
99/*
100 * Writes a summary of the signal to the log file.  We do this so that, if
101 * for some reason we're not able to contact debuggerd, there is still some
102 * indication of the failure in the log.
103 *
104 * We could be here as a result of native heap corruption, or while a
105 * mutex is being held, so we don't want to use any libc functions that
106 * could allocate memory or hold a lock.
107 */
108static void logSignalSummary(int signum, const siginfo_t* info) {
109    const char* signal_name;
110    switch (signum) {
111        case SIGILL:    signal_name = "SIGILL";     break;
112        case SIGABRT:   signal_name = "SIGABRT";    break;
113        case SIGBUS:    signal_name = "SIGBUS";     break;
114        case SIGFPE:    signal_name = "SIGFPE";     break;
115        case SIGSEGV:   signal_name = "SIGSEGV";    break;
116#if defined(SIGSTKFLT)
117        case SIGSTKFLT: signal_name = "SIGSTKFLT";  break;
118#endif
119        case SIGPIPE:   signal_name = "SIGPIPE";    break;
120        default:        signal_name = "???";        break;
121    }
122
123    char thread_name[MAX_TASK_NAME_LEN + 1]; // one more for termination
124    if (prctl(PR_GET_NAME, (unsigned long)thread_name, 0, 0, 0) != 0) {
125        strcpy(thread_name, "<name unknown>");
126    } else {
127        // short names are null terminated by prctl, but the man page
128        // implies that 16 byte names are not.
129        thread_name[MAX_TASK_NAME_LEN] = 0;
130    }
131
132    // "info" will be NULL if the siginfo_t information was not available.
133    if (info != NULL) {
134        __libc_format_log(ANDROID_LOG_FATAL, "libc",
135                          "Fatal signal %d (%s) at 0x%08x (code=%d), thread %d (%s)",
136                          signum, signal_name, reinterpret_cast<uintptr_t>(info->si_addr),
137                          info->si_code, gettid(), thread_name);
138    } else {
139        __libc_format_log(ANDROID_LOG_FATAL, "libc",
140                          "Fatal signal %d (%s), thread %d (%s)",
141                          signum, signal_name, gettid(), thread_name);
142    }
143}
144
145/*
146 * Returns true if the handler for signal "signum" has SA_SIGINFO set.
147 */
148static bool haveSiginfo(int signum) {
149    struct sigaction oldact, newact;
150
151    memset(&newact, 0, sizeof(newact));
152    newact.sa_handler = SIG_DFL;
153    newact.sa_flags = SA_RESTART;
154    sigemptyset(&newact.sa_mask);
155
156    if (sigaction(signum, &newact, &oldact) < 0) {
157      __libc_format_log(ANDROID_LOG_FATAL, "libc", "Failed testing for SA_SIGINFO: %s",
158                        strerror(errno));
159      return 0;
160    }
161    bool ret = (oldact.sa_flags & SA_SIGINFO) != 0;
162
163    if (sigaction(signum, &oldact, NULL) == -1) {
164      __libc_format_log(ANDROID_LOG_FATAL, "libc", "Restore failed in test for SA_SIGINFO: %s",
165                        strerror(errno));
166    }
167    return ret;
168}
169
170/*
171 * Catches fatal signals so we can ask debuggerd to ptrace us before
172 * we crash.
173 */
174void debuggerd_signal_handler(int n, siginfo_t* info, void*) {
175    /*
176     * It's possible somebody cleared the SA_SIGINFO flag, which would mean
177     * our "info" arg holds an undefined value.
178     */
179    if (!haveSiginfo(n)) {
180        info = NULL;
181    }
182
183    logSignalSummary(n, info);
184
185    pid_t tid = gettid();
186    int s = socket_abstract_client(DEBUGGER_SOCKET_NAME, SOCK_STREAM);
187
188    if (s >= 0) {
189        /* debugger knows our pid from the credentials on the
190         * local socket but we need to tell it our tid.  It
191         * is paranoid and will verify that we are giving a tid
192         * that's actually in our process
193         */
194        int  ret;
195        debugger_msg_t msg;
196        msg.action = DEBUGGER_ACTION_CRASH;
197        msg.tid = tid;
198        ret = TEMP_FAILURE_RETRY(write(s, &msg, sizeof(msg)));
199        if (ret == sizeof(msg)) {
200            /* if the write failed, there is no point to read on
201             * the file descriptor. */
202            ret = TEMP_FAILURE_RETRY(read(s, &tid, 1));
203            int saved_errno = errno;
204            notify_gdb_of_libraries();
205            errno = saved_errno;
206        }
207
208        if (ret < 0) {
209            /* read or write failed -- broken connection? */
210            __libc_format_log(ANDROID_LOG_FATAL, "libc", "Failed while talking to debuggerd: %s",
211                              strerror(errno));
212        }
213
214        close(s);
215    } else {
216        /* socket failed; maybe process ran out of fds */
217        __libc_format_log(ANDROID_LOG_FATAL, "libc", "Unable to open connection to debuggerd: %s",
218                          strerror(errno));
219    }
220
221    /* remove our net so we fault for real when we return */
222    signal(n, SIG_DFL);
223
224    /*
225     * These signals are not re-thrown when we resume.  This means that
226     * crashing due to (say) SIGPIPE doesn't work the way you'd expect it
227     * to.  We work around this by throwing them manually.  We don't want
228     * to do this for *all* signals because it'll screw up the address for
229     * faults like SIGSEGV.
230     */
231    switch (n) {
232        case SIGABRT:
233        case SIGFPE:
234        case SIGPIPE:
235#ifdef SIGSTKFLT
236        case SIGSTKFLT:
237#endif
238            (void) tgkill(getpid(), gettid(), n);
239            break;
240        default:    // SIGILL, SIGBUS, SIGSEGV
241            break;
242    }
243}
244
245void debuggerd_init() {
246    struct sigaction act;
247    memset(&act, 0, sizeof(act));
248    act.sa_sigaction = debuggerd_signal_handler;
249    act.sa_flags = SA_RESTART | SA_SIGINFO;
250    sigemptyset(&act.sa_mask);
251
252    sigaction(SIGILL, &act, NULL);
253    sigaction(SIGABRT, &act, NULL);
254    sigaction(SIGBUS, &act, NULL);
255    sigaction(SIGFPE, &act, NULL);
256    sigaction(SIGSEGV, &act, NULL);
257#if defined(SIGSTKFLT)
258    sigaction(SIGSTKFLT, &act, NULL);
259#endif
260    sigaction(SIGPIPE, &act, NULL);
261}
262