debugger.cpp revision 6b8e321e610c0683062c0e8d8deabb64565c7403 
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
40#include <private/debug_format.h>
41#include <private/logd.h>
42
43extern "C" int tgkill(int tgid, int tid, int sig);
44
45#define DEBUGGER_SOCKET_NAME "android:debuggerd"
46
47enum debugger_action_t {
48    // dump a crash
49    DEBUGGER_ACTION_CRASH,
50    // dump a tombstone file
51    DEBUGGER_ACTION_DUMP_TOMBSTONE,
52    // dump a backtrace only back to the socket
53    DEBUGGER_ACTION_DUMP_BACKTRACE,
54};
55
56/* message sent over the socket */
57struct debugger_msg_t {
58    debugger_action_t action;
59    pid_t tid;
60};
61
62// see man(2) prctl, specifically the section about PR_GET_NAME
63#define MAX_TASK_NAME_LEN (16)
64
65static int socket_abstract_client(const char* name, int type) {
66    sockaddr_un addr;
67
68    // Test with length +1 for the *initial* '\0'.
69    size_t namelen = strlen(name);
70    if ((namelen + 1) > sizeof(addr.sun_path)) {
71        errno = EINVAL;
72        return -1;
73    }
74
75    /* This is used for abstract socket namespace, we need
76     * an initial '\0' at the start of the Unix socket path.
77     *
78     * Note: The path in this case is *not* supposed to be
79     * '\0'-terminated. ("man 7 unix" for the gory details.)
80     */
81    memset(&addr, 0, sizeof(addr));
82    addr.sun_family = AF_LOCAL;
83    addr.sun_path[0] = 0;
84    memcpy(addr.sun_path + 1, name, namelen);
85
86    socklen_t alen = namelen + offsetof(sockaddr_un, sun_path) + 1;
87
88    int s = socket(AF_LOCAL, type, 0);
89    if (s == -1) {
90        return -1;
91    }
92
93    int err = TEMP_FAILURE_RETRY(connect(s, (sockaddr*) &addr, alen));
94    if (err == -1) {
95        close(s);
96        s = -1;
97    }
98
99    return s;
100}
101
102/*
103 * Writes a summary of the signal to the log file.  We do this so that, if
104 * for some reason we're not able to contact debuggerd, there is still some
105 * indication of the failure in the log.
106 *
107 * We could be here as a result of native heap corruption, or while a
108 * mutex is being held, so we don't want to use any libc functions that
109 * could allocate memory or hold a lock.
110 */
111static void logSignalSummary(int signum, const siginfo_t* info) {
112    const char* signal_name;
113    switch (signum) {
114        case SIGILL:    signal_name = "SIGILL";     break;
115        case SIGABRT:   signal_name = "SIGABRT";    break;
116        case SIGBUS:    signal_name = "SIGBUS";     break;
117        case SIGFPE:    signal_name = "SIGFPE";     break;
118        case SIGSEGV:   signal_name = "SIGSEGV";    break;
119#if defined(SIGSTKFLT)
120        case SIGSTKFLT: signal_name = "SIGSTKFLT";  break;
121#endif
122        case SIGPIPE:   signal_name = "SIGPIPE";    break;
123        default:        signal_name = "???";        break;
124    }
125
126    char thread_name[MAX_TASK_NAME_LEN + 1]; // one more for termination
127    if (prctl(PR_GET_NAME, (unsigned long)thread_name, 0, 0, 0) != 0) {
128        strcpy(thread_name, "<name unknown>");
129    } else {
130        // short names are null terminated by prctl, but the man page
131        // implies that 16 byte names are not.
132        thread_name[MAX_TASK_NAME_LEN] = 0;
133    }
134
135    // "info" will be NULL if the siginfo_t information was not available.
136    if (info != NULL) {
137        __libc_format_log(ANDROID_LOG_FATAL, "libc",
138                          "Fatal signal %d (%s) at 0x%08x (code=%d), thread %d (%s)",
139                          signum, signal_name, reinterpret_cast<uintptr_t>(info->si_addr),
140                          info->si_code, gettid(), thread_name);
141    } else {
142        __libc_format_log(ANDROID_LOG_FATAL, "libc",
143                          "Fatal signal %d (%s), thread %d (%s)",
144                          signum, signal_name, gettid(), thread_name);
145    }
146}
147
148/*
149 * Returns true if the handler for signal "signum" has SA_SIGINFO set.
150 */
151static bool haveSiginfo(int signum) {
152    struct sigaction oldact, newact;
153
154    memset(&newact, 0, sizeof(newact));
155    newact.sa_handler = SIG_DFL;
156    newact.sa_flags = SA_RESTART;
157    sigemptyset(&newact.sa_mask);
158
159    if (sigaction(signum, &newact, &oldact) < 0) {
160        __libc_android_log_write(ANDROID_LOG_FATAL, "libc",
161            "Failed testing for SA_SIGINFO");
162        return 0;
163    }
164    bool ret = (oldact.sa_flags & SA_SIGINFO) != 0;
165
166    if (sigaction(signum, &oldact, NULL) < 0) {
167        __libc_android_log_write(ANDROID_LOG_FATAL, "libc",
168            "Restore failed in test for SA_SIGINFO");
169    }
170    return ret;
171}
172
173/*
174 * Catches fatal signals so we can ask debuggerd to ptrace us before
175 * we crash.
176 */
177void debugger_signal_handler(int n, siginfo_t* info, void*) {
178    /*
179     * It's possible somebody cleared the SA_SIGINFO flag, which would mean
180     * our "info" arg holds an undefined value.
181     */
182    if (!haveSiginfo(n)) {
183        info = NULL;
184    }
185
186    logSignalSummary(n, info);
187
188    pid_t tid = gettid();
189    int s = socket_abstract_client(DEBUGGER_SOCKET_NAME, SOCK_STREAM);
190
191    if (s >= 0) {
192        /* debugger knows our pid from the credentials on the
193         * local socket but we need to tell it our tid.  It
194         * is paranoid and will verify that we are giving a tid
195         * that's actually in our process
196         */
197        int  ret;
198        debugger_msg_t msg;
199        msg.action = DEBUGGER_ACTION_CRASH;
200        msg.tid = tid;
201        ret = TEMP_FAILURE_RETRY(write(s, &msg, sizeof(msg)));
202        if (ret == sizeof(msg)) {
203            /* if the write failed, there is no point to read on
204             * the file descriptor. */
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#ifdef 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 debugger_init() {
249    struct sigaction act;
250    memset(&act, 0, sizeof(act));
251    act.sa_sigaction = debugger_signal_handler;
252    act.sa_flags = SA_RESTART | SA_SIGINFO;
253    sigemptyset(&act.sa_mask);
254
255    sigaction(SIGILL, &act, NULL);
256    sigaction(SIGABRT, &act, NULL);
257    sigaction(SIGBUS, &act, NULL);
258    sigaction(SIGFPE, &act, NULL);
259    sigaction(SIGSEGV, &act, NULL);
260#if defined(SIGSTKFLT)
261    sigaction(SIGSTKFLT, &act, NULL);
262#endif
263    sigaction(SIGPIPE, &act, NULL);
264}
265