strace/unwind.c

322 lines
7.3 KiB
C
Raw Normal View History

/*
* Copyright (c) 2013 Luca Clementi <luca.clementi@gmail.com>
* Copyright (c) 2013-2018 The strace developers.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* 3. The name of the author may not be used to endorse or promote products
* derived from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
* IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
* OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
* IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
* INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
* NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
* DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
* THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
* THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
#include "defs.h"
#include "unwind.h"
unwind: demangle symbol names Implement demangling of C++ symbol names in stack trace using cplus_demangle function from GNU libiberty library. This is an example demangled stack trace output: fstat(5, {st_mode=S_IFREG|0664, st_size=0, ...}) = 0 > /usr/lib64/libc-2.25.so(__fxstat64+0x12) [0xffd62] > /usr/lib64/libc-2.25.so(_IO_file_doallocate+0x5f) [0x717ff] > /usr/lib64/libc-2.25.so(_IO_doallocbuf+0x79) [0x81699] > /usr/lib64/libc-2.25.so(_IO_file_overflow+0x198) [0x807b8] > /usr/lib64/libc-2.25.so(_IO_file_xsputn+0xbd) [0x7ed5d] > /usr/lib64/libc-2.25.so(fwrite_unlocked+0x60) [0x7d800] > /usr/lib64/libleveldb.so.1.18(leveldb::EnvWrapper::StartThread+0x3b6) [0x48656] > /usr/lib64/libleveldb.so.1.18(leveldb::log::Writer::EmitPhysicalRecord+0x89) [0x28bc9] > /usr/lib64/libleveldb.so.1.18(leveldb::log::Writer::AddRecord+0x9e) [0x28d9e] > /usr/lib64/libleveldb.so.1.18(leveldb::DBImpl::Write+0x208) [0x1ce18] > /usr/lib64/libleveldb.so.1.18(leveldb::DB::Put+0x59) [0x192b9] > /usr/lib64/libleveldb.so.1.18(leveldb::DBImpl::Put+0x1d) [0x1931d] > /home/yamato/var/leveldb/doc/a.out(main+0x120) [0x1107] > /usr/lib64/libc-2.25.so(__libc_start_main+0xea) [0x2088a] > /home/yamato/var/leveldb/doc/a.out(_start+0x2a) [0xf3a] * Makefile.am [USE_DEMANGLE] (strace_CPPFLAGS, strace_LDFLAGS, libiberty_LDADD): Append libiberty_CPPFLAGS, strace_LDFLAGS, and libiberty_LIBS, respectively. * configure.ac: Add --with-libiberty option. Check cplus_demangle support in libiberty. * unwind.c [USE_DEMANGLE]: Include <demangle.h>. (print_stack_frame) [USE_DEMANGLE]: Use cplus_demangle. Signed-off-by: Masatake YAMATO <yamato@redhat.com>
2017-12-18 13:24:55 +09:00
#ifdef USE_DEMANGLE
# if defined HAVE_DEMANGLE_H
# include <demangle.h>
# elif defined HAVE_LIBIBERTY_DEMANGLE_H
# include <libiberty/demangle.h>
# endif
unwind: demangle symbol names Implement demangling of C++ symbol names in stack trace using cplus_demangle function from GNU libiberty library. This is an example demangled stack trace output: fstat(5, {st_mode=S_IFREG|0664, st_size=0, ...}) = 0 > /usr/lib64/libc-2.25.so(__fxstat64+0x12) [0xffd62] > /usr/lib64/libc-2.25.so(_IO_file_doallocate+0x5f) [0x717ff] > /usr/lib64/libc-2.25.so(_IO_doallocbuf+0x79) [0x81699] > /usr/lib64/libc-2.25.so(_IO_file_overflow+0x198) [0x807b8] > /usr/lib64/libc-2.25.so(_IO_file_xsputn+0xbd) [0x7ed5d] > /usr/lib64/libc-2.25.so(fwrite_unlocked+0x60) [0x7d800] > /usr/lib64/libleveldb.so.1.18(leveldb::EnvWrapper::StartThread+0x3b6) [0x48656] > /usr/lib64/libleveldb.so.1.18(leveldb::log::Writer::EmitPhysicalRecord+0x89) [0x28bc9] > /usr/lib64/libleveldb.so.1.18(leveldb::log::Writer::AddRecord+0x9e) [0x28d9e] > /usr/lib64/libleveldb.so.1.18(leveldb::DBImpl::Write+0x208) [0x1ce18] > /usr/lib64/libleveldb.so.1.18(leveldb::DB::Put+0x59) [0x192b9] > /usr/lib64/libleveldb.so.1.18(leveldb::DBImpl::Put+0x1d) [0x1931d] > /home/yamato/var/leveldb/doc/a.out(main+0x120) [0x1107] > /usr/lib64/libc-2.25.so(__libc_start_main+0xea) [0x2088a] > /home/yamato/var/leveldb/doc/a.out(_start+0x2a) [0xf3a] * Makefile.am [USE_DEMANGLE] (strace_CPPFLAGS, strace_LDFLAGS, libiberty_LDADD): Append libiberty_CPPFLAGS, strace_LDFLAGS, and libiberty_LIBS, respectively. * configure.ac: Add --with-libiberty option. Check cplus_demangle support in libiberty. * unwind.c [USE_DEMANGLE]: Include <demangle.h>. (print_stack_frame) [USE_DEMANGLE]: Use cplus_demangle. Signed-off-by: Masatake YAMATO <yamato@redhat.com>
2017-12-18 13:24:55 +09:00
#endif
unwind: introduce queue_t for capturing stacktrace This is the second step for splitting capturing from printing. New `queue' field is added to tcb. Captured stacktrace is stored here. The field is initialized/finalized at unwind_tcb_init/unwind_tcb_fin. New API function unwind_capture_stacktrace is added. This function captures the currest stack using stracktrace_walker and records it in tcb. It's printing is delayed to the next call of unwind_print_stacktrace. unwind_print_stacktrace is extended. Now it checks queue field of the given tcb at the start of function. If the function finds a captured stack trace, the latter is printed using stracktrace_walker. Currently unwind_capture_stacktrace invocations are added directly to handlers of mmap, munmap, mprotect, and execve. Here is the difference of output with/without patch: (without patch) execve("./test-fork", ["./test-fork"], [/* 56 vars */]) = 0 > /usr/lib64/ld-2.18.so(check_one_fd.part.0+0x82) [0x11f0] (with patch) execve("./test-fork", ["./test-fork"], [/* 54 vars */]) = 0 > /usr/lib64/libc-2.18.so(execve+0x7) [0xbcd27] > /home/yamato/var/strace/strace(exec_or_die+0x10c) [0x26ac] > /home/yamato/var/strace/strace(startup_child+0x346) [0x134f6] > /home/yamato/var/strace/strace(init+0x89f) [0x13dff] > /home/yamato/var/strace/strace(main+0xa) [0x26ca] > /usr/lib64/libc-2.18.so(__libc_start_main+0xf5) [0x21d65] > /home/yamato/var/strace/strace(_start+0x29) [0x2799] In older version output lines of captured elements were built when printing. In this version they are built when capturing the stack. As result, unneeded dynamic memory allocations are avoided. Suggested by Luca Clementi. In older version the combination of snprintf and realloc were used. In this version they are replaced with asprintf. Suggested by Dmitry Levin. Signed-off-by: Masatake YAMATO <yamato@redhat.com>
2014-04-16 15:33:06 +09:00
/*
* Type used in stacktrace capturing
*/
struct call_t {
struct call_t *next;
char *output_line;
unwind: introduce queue_t for capturing stacktrace This is the second step for splitting capturing from printing. New `queue' field is added to tcb. Captured stacktrace is stored here. The field is initialized/finalized at unwind_tcb_init/unwind_tcb_fin. New API function unwind_capture_stacktrace is added. This function captures the currest stack using stracktrace_walker and records it in tcb. It's printing is delayed to the next call of unwind_print_stacktrace. unwind_print_stacktrace is extended. Now it checks queue field of the given tcb at the start of function. If the function finds a captured stack trace, the latter is printed using stracktrace_walker. Currently unwind_capture_stacktrace invocations are added directly to handlers of mmap, munmap, mprotect, and execve. Here is the difference of output with/without patch: (without patch) execve("./test-fork", ["./test-fork"], [/* 56 vars */]) = 0 > /usr/lib64/ld-2.18.so(check_one_fd.part.0+0x82) [0x11f0] (with patch) execve("./test-fork", ["./test-fork"], [/* 54 vars */]) = 0 > /usr/lib64/libc-2.18.so(execve+0x7) [0xbcd27] > /home/yamato/var/strace/strace(exec_or_die+0x10c) [0x26ac] > /home/yamato/var/strace/strace(startup_child+0x346) [0x134f6] > /home/yamato/var/strace/strace(init+0x89f) [0x13dff] > /home/yamato/var/strace/strace(main+0xa) [0x26ca] > /usr/lib64/libc-2.18.so(__libc_start_main+0xf5) [0x21d65] > /home/yamato/var/strace/strace(_start+0x29) [0x2799] In older version output lines of captured elements were built when printing. In this version they are built when capturing the stack. As result, unneeded dynamic memory allocations are avoided. Suggested by Luca Clementi. In older version the combination of snprintf and realloc were used. In this version they are replaced with asprintf. Suggested by Dmitry Levin. Signed-off-by: Masatake YAMATO <yamato@redhat.com>
2014-04-16 15:33:06 +09:00
};
struct unwind_queue_t {
struct call_t *tail;
struct call_t *head;
unwind: introduce queue_t for capturing stacktrace This is the second step for splitting capturing from printing. New `queue' field is added to tcb. Captured stacktrace is stored here. The field is initialized/finalized at unwind_tcb_init/unwind_tcb_fin. New API function unwind_capture_stacktrace is added. This function captures the currest stack using stracktrace_walker and records it in tcb. It's printing is delayed to the next call of unwind_print_stacktrace. unwind_print_stacktrace is extended. Now it checks queue field of the given tcb at the start of function. If the function finds a captured stack trace, the latter is printed using stracktrace_walker. Currently unwind_capture_stacktrace invocations are added directly to handlers of mmap, munmap, mprotect, and execve. Here is the difference of output with/without patch: (without patch) execve("./test-fork", ["./test-fork"], [/* 56 vars */]) = 0 > /usr/lib64/ld-2.18.so(check_one_fd.part.0+0x82) [0x11f0] (with patch) execve("./test-fork", ["./test-fork"], [/* 54 vars */]) = 0 > /usr/lib64/libc-2.18.so(execve+0x7) [0xbcd27] > /home/yamato/var/strace/strace(exec_or_die+0x10c) [0x26ac] > /home/yamato/var/strace/strace(startup_child+0x346) [0x134f6] > /home/yamato/var/strace/strace(init+0x89f) [0x13dff] > /home/yamato/var/strace/strace(main+0xa) [0x26ca] > /usr/lib64/libc-2.18.so(__libc_start_main+0xf5) [0x21d65] > /home/yamato/var/strace/strace(_start+0x29) [0x2799] In older version output lines of captured elements were built when printing. In this version they are built when capturing the stack. As result, unneeded dynamic memory allocations are avoided. Suggested by Luca Clementi. In older version the combination of snprintf and realloc were used. In this version they are replaced with asprintf. Suggested by Dmitry Levin. Signed-off-by: Masatake YAMATO <yamato@redhat.com>
2014-04-16 15:33:06 +09:00
};
static void queue_print(struct unwind_queue_t *queue);
static const char asprintf_error_str[] = "???";
void
unwind_init(void)
{
if (unwinder.init)
unwinder.init();
}
void
unwind_tcb_init(struct tcb *tcp)
{
if (tcp->unwind_queue)
return;
tcp->unwind_queue = xmalloc(sizeof(*tcp->unwind_queue));
tcp->unwind_queue->head = NULL;
tcp->unwind_queue->tail = NULL;
tcp->unwind_ctx = unwinder.tcb_init(tcp);
}
void
unwind_tcb_fin(struct tcb *tcp)
{
if (!tcp->unwind_queue)
return;
queue_print(tcp->unwind_queue);
free(tcp->unwind_queue);
tcp->unwind_queue = NULL;
unwind: introduce queue_t for capturing stacktrace This is the second step for splitting capturing from printing. New `queue' field is added to tcb. Captured stacktrace is stored here. The field is initialized/finalized at unwind_tcb_init/unwind_tcb_fin. New API function unwind_capture_stacktrace is added. This function captures the currest stack using stracktrace_walker and records it in tcb. It's printing is delayed to the next call of unwind_print_stacktrace. unwind_print_stacktrace is extended. Now it checks queue field of the given tcb at the start of function. If the function finds a captured stack trace, the latter is printed using stracktrace_walker. Currently unwind_capture_stacktrace invocations are added directly to handlers of mmap, munmap, mprotect, and execve. Here is the difference of output with/without patch: (without patch) execve("./test-fork", ["./test-fork"], [/* 56 vars */]) = 0 > /usr/lib64/ld-2.18.so(check_one_fd.part.0+0x82) [0x11f0] (with patch) execve("./test-fork", ["./test-fork"], [/* 54 vars */]) = 0 > /usr/lib64/libc-2.18.so(execve+0x7) [0xbcd27] > /home/yamato/var/strace/strace(exec_or_die+0x10c) [0x26ac] > /home/yamato/var/strace/strace(startup_child+0x346) [0x134f6] > /home/yamato/var/strace/strace(init+0x89f) [0x13dff] > /home/yamato/var/strace/strace(main+0xa) [0x26ca] > /usr/lib64/libc-2.18.so(__libc_start_main+0xf5) [0x21d65] > /home/yamato/var/strace/strace(_start+0x29) [0x2799] In older version output lines of captured elements were built when printing. In this version they are built when capturing the stack. As result, unneeded dynamic memory allocations are avoided. Suggested by Luca Clementi. In older version the combination of snprintf and realloc were used. In this version they are replaced with asprintf. Suggested by Dmitry Levin. Signed-off-by: Masatake YAMATO <yamato@redhat.com>
2014-04-16 15:33:06 +09:00
unwinder.tcb_fin(tcp);
tcp->unwind_ctx = NULL;
}
/*
unwind: introduce queue_t for capturing stacktrace This is the second step for splitting capturing from printing. New `queue' field is added to tcb. Captured stacktrace is stored here. The field is initialized/finalized at unwind_tcb_init/unwind_tcb_fin. New API function unwind_capture_stacktrace is added. This function captures the currest stack using stracktrace_walker and records it in tcb. It's printing is delayed to the next call of unwind_print_stacktrace. unwind_print_stacktrace is extended. Now it checks queue field of the given tcb at the start of function. If the function finds a captured stack trace, the latter is printed using stracktrace_walker. Currently unwind_capture_stacktrace invocations are added directly to handlers of mmap, munmap, mprotect, and execve. Here is the difference of output with/without patch: (without patch) execve("./test-fork", ["./test-fork"], [/* 56 vars */]) = 0 > /usr/lib64/ld-2.18.so(check_one_fd.part.0+0x82) [0x11f0] (with patch) execve("./test-fork", ["./test-fork"], [/* 54 vars */]) = 0 > /usr/lib64/libc-2.18.so(execve+0x7) [0xbcd27] > /home/yamato/var/strace/strace(exec_or_die+0x10c) [0x26ac] > /home/yamato/var/strace/strace(startup_child+0x346) [0x134f6] > /home/yamato/var/strace/strace(init+0x89f) [0x13dff] > /home/yamato/var/strace/strace(main+0xa) [0x26ca] > /usr/lib64/libc-2.18.so(__libc_start_main+0xf5) [0x21d65] > /home/yamato/var/strace/strace(_start+0x29) [0x2799] In older version output lines of captured elements were built when printing. In this version they are built when capturing the stack. As result, unneeded dynamic memory allocations are avoided. Suggested by Luca Clementi. In older version the combination of snprintf and realloc were used. In this version they are replaced with asprintf. Suggested by Dmitry Levin. Signed-off-by: Masatake YAMATO <yamato@redhat.com>
2014-04-16 15:33:06 +09:00
* printing an entry in stack to stream or buffer
*/
/*
* we want to keep the format used by backtrace_symbols from the glibc
*
* ./a.out() [0x40063d]
* ./a.out() [0x4006bb]
* ./a.out() [0x4006c6]
* /lib64/libc.so.6(__libc_start_main+0xed) [0x7fa2f8a5976d]
* ./a.out() [0x400569]
*/
#define STACK_ENTRY_SYMBOL_FMT(SYM) \
" > %s(%s+0x%lx) [0x%lx]\n", \
binary_filename, \
(SYM), \
(unsigned long) function_offset, \
true_offset
#define STACK_ENTRY_NOSYMBOL_FMT \
" > %s() [0x%lx]\n", \
binary_filename, true_offset
#define STACK_ENTRY_BUG_FMT \
" > BUG IN %s\n"
#define STACK_ENTRY_ERROR_WITH_OFFSET_FMT \
" > %s [0x%lx]\n", error, true_offset
#define STACK_ENTRY_ERROR_FMT \
" > %s\n", error
static void
print_call_cb(void *dummy,
const char *binary_filename,
const char *symbol_name,
unwind_function_offset_t function_offset,
unsigned long true_offset)
{
if (symbol_name && (symbol_name[0] != '\0')) {
#ifdef USE_DEMANGLE
char *demangled_name =
cplus_demangle(symbol_name,
DMGL_AUTO | DMGL_PARAMS);
#endif
tprintf(STACK_ENTRY_SYMBOL_FMT(
#ifdef USE_DEMANGLE
demangled_name ? demangled_name :
#endif
symbol_name));
#ifdef USE_DEMANGLE
free(demangled_name);
#endif
}
else if (binary_filename)
tprintf(STACK_ENTRY_NOSYMBOL_FMT);
else
tprintf(STACK_ENTRY_BUG_FMT, __func__);
line_ended();
}
static void
print_error_cb(void *dummy,
const char *error,
unsigned long true_offset)
{
if (true_offset)
tprintf(STACK_ENTRY_ERROR_WITH_OFFSET_FMT);
else
tprintf(STACK_ENTRY_ERROR_FMT);
line_ended();
}
unwind: introduce queue_t for capturing stacktrace This is the second step for splitting capturing from printing. New `queue' field is added to tcb. Captured stacktrace is stored here. The field is initialized/finalized at unwind_tcb_init/unwind_tcb_fin. New API function unwind_capture_stacktrace is added. This function captures the currest stack using stracktrace_walker and records it in tcb. It's printing is delayed to the next call of unwind_print_stacktrace. unwind_print_stacktrace is extended. Now it checks queue field of the given tcb at the start of function. If the function finds a captured stack trace, the latter is printed using stracktrace_walker. Currently unwind_capture_stacktrace invocations are added directly to handlers of mmap, munmap, mprotect, and execve. Here is the difference of output with/without patch: (without patch) execve("./test-fork", ["./test-fork"], [/* 56 vars */]) = 0 > /usr/lib64/ld-2.18.so(check_one_fd.part.0+0x82) [0x11f0] (with patch) execve("./test-fork", ["./test-fork"], [/* 54 vars */]) = 0 > /usr/lib64/libc-2.18.so(execve+0x7) [0xbcd27] > /home/yamato/var/strace/strace(exec_or_die+0x10c) [0x26ac] > /home/yamato/var/strace/strace(startup_child+0x346) [0x134f6] > /home/yamato/var/strace/strace(init+0x89f) [0x13dff] > /home/yamato/var/strace/strace(main+0xa) [0x26ca] > /usr/lib64/libc-2.18.so(__libc_start_main+0xf5) [0x21d65] > /home/yamato/var/strace/strace(_start+0x29) [0x2799] In older version output lines of captured elements were built when printing. In this version they are built when capturing the stack. As result, unneeded dynamic memory allocations are avoided. Suggested by Luca Clementi. In older version the combination of snprintf and realloc were used. In this version they are replaced with asprintf. Suggested by Dmitry Levin. Signed-off-by: Masatake YAMATO <yamato@redhat.com>
2014-04-16 15:33:06 +09:00
static char *
sprint_call_or_error(const char *binary_filename,
const char *symbol_name,
unwind_function_offset_t function_offset,
unwind: introduce queue_t for capturing stacktrace This is the second step for splitting capturing from printing. New `queue' field is added to tcb. Captured stacktrace is stored here. The field is initialized/finalized at unwind_tcb_init/unwind_tcb_fin. New API function unwind_capture_stacktrace is added. This function captures the currest stack using stracktrace_walker and records it in tcb. It's printing is delayed to the next call of unwind_print_stacktrace. unwind_print_stacktrace is extended. Now it checks queue field of the given tcb at the start of function. If the function finds a captured stack trace, the latter is printed using stracktrace_walker. Currently unwind_capture_stacktrace invocations are added directly to handlers of mmap, munmap, mprotect, and execve. Here is the difference of output with/without patch: (without patch) execve("./test-fork", ["./test-fork"], [/* 56 vars */]) = 0 > /usr/lib64/ld-2.18.so(check_one_fd.part.0+0x82) [0x11f0] (with patch) execve("./test-fork", ["./test-fork"], [/* 54 vars */]) = 0 > /usr/lib64/libc-2.18.so(execve+0x7) [0xbcd27] > /home/yamato/var/strace/strace(exec_or_die+0x10c) [0x26ac] > /home/yamato/var/strace/strace(startup_child+0x346) [0x134f6] > /home/yamato/var/strace/strace(init+0x89f) [0x13dff] > /home/yamato/var/strace/strace(main+0xa) [0x26ca] > /usr/lib64/libc-2.18.so(__libc_start_main+0xf5) [0x21d65] > /home/yamato/var/strace/strace(_start+0x29) [0x2799] In older version output lines of captured elements were built when printing. In this version they are built when capturing the stack. As result, unneeded dynamic memory allocations are avoided. Suggested by Luca Clementi. In older version the combination of snprintf and realloc were used. In this version they are replaced with asprintf. Suggested by Dmitry Levin. Signed-off-by: Masatake YAMATO <yamato@redhat.com>
2014-04-16 15:33:06 +09:00
unsigned long true_offset,
const char *error)
{
char *output_line = NULL;
int n;
if (symbol_name) {
#ifdef USE_DEMANGLE
char *demangled_name =
cplus_demangle(symbol_name,
DMGL_AUTO | DMGL_PARAMS);
#endif
n = asprintf(&output_line,
STACK_ENTRY_SYMBOL_FMT(
#ifdef USE_DEMANGLE
demangled_name ? demangled_name :
#endif
symbol_name));
#ifdef USE_DEMANGLE
free(demangled_name);
#endif
}
else if (binary_filename)
n = asprintf(&output_line, STACK_ENTRY_NOSYMBOL_FMT);
else if (error)
n = true_offset
? asprintf(&output_line, STACK_ENTRY_ERROR_WITH_OFFSET_FMT)
: asprintf(&output_line, STACK_ENTRY_ERROR_FMT);
else
n = asprintf(&output_line, STACK_ENTRY_BUG_FMT, __func__);
if (n < 0) {
perror_func_msg("asprintf");
output_line = (char *) asprintf_error_str;
}
return output_line;
unwind: introduce queue_t for capturing stacktrace This is the second step for splitting capturing from printing. New `queue' field is added to tcb. Captured stacktrace is stored here. The field is initialized/finalized at unwind_tcb_init/unwind_tcb_fin. New API function unwind_capture_stacktrace is added. This function captures the currest stack using stracktrace_walker and records it in tcb. It's printing is delayed to the next call of unwind_print_stacktrace. unwind_print_stacktrace is extended. Now it checks queue field of the given tcb at the start of function. If the function finds a captured stack trace, the latter is printed using stracktrace_walker. Currently unwind_capture_stacktrace invocations are added directly to handlers of mmap, munmap, mprotect, and execve. Here is the difference of output with/without patch: (without patch) execve("./test-fork", ["./test-fork"], [/* 56 vars */]) = 0 > /usr/lib64/ld-2.18.so(check_one_fd.part.0+0x82) [0x11f0] (with patch) execve("./test-fork", ["./test-fork"], [/* 54 vars */]) = 0 > /usr/lib64/libc-2.18.so(execve+0x7) [0xbcd27] > /home/yamato/var/strace/strace(exec_or_die+0x10c) [0x26ac] > /home/yamato/var/strace/strace(startup_child+0x346) [0x134f6] > /home/yamato/var/strace/strace(init+0x89f) [0x13dff] > /home/yamato/var/strace/strace(main+0xa) [0x26ca] > /usr/lib64/libc-2.18.so(__libc_start_main+0xf5) [0x21d65] > /home/yamato/var/strace/strace(_start+0x29) [0x2799] In older version output lines of captured elements were built when printing. In this version they are built when capturing the stack. As result, unneeded dynamic memory allocations are avoided. Suggested by Luca Clementi. In older version the combination of snprintf and realloc were used. In this version they are replaced with asprintf. Suggested by Dmitry Levin. Signed-off-by: Masatake YAMATO <yamato@redhat.com>
2014-04-16 15:33:06 +09:00
}
/*
* queue manipulators
*/
static void
queue_put(struct unwind_queue_t *queue,
const char *binary_filename,
const char *symbol_name,
unwind_function_offset_t function_offset,
unwind: introduce queue_t for capturing stacktrace This is the second step for splitting capturing from printing. New `queue' field is added to tcb. Captured stacktrace is stored here. The field is initialized/finalized at unwind_tcb_init/unwind_tcb_fin. New API function unwind_capture_stacktrace is added. This function captures the currest stack using stracktrace_walker and records it in tcb. It's printing is delayed to the next call of unwind_print_stacktrace. unwind_print_stacktrace is extended. Now it checks queue field of the given tcb at the start of function. If the function finds a captured stack trace, the latter is printed using stracktrace_walker. Currently unwind_capture_stacktrace invocations are added directly to handlers of mmap, munmap, mprotect, and execve. Here is the difference of output with/without patch: (without patch) execve("./test-fork", ["./test-fork"], [/* 56 vars */]) = 0 > /usr/lib64/ld-2.18.so(check_one_fd.part.0+0x82) [0x11f0] (with patch) execve("./test-fork", ["./test-fork"], [/* 54 vars */]) = 0 > /usr/lib64/libc-2.18.so(execve+0x7) [0xbcd27] > /home/yamato/var/strace/strace(exec_or_die+0x10c) [0x26ac] > /home/yamato/var/strace/strace(startup_child+0x346) [0x134f6] > /home/yamato/var/strace/strace(init+0x89f) [0x13dff] > /home/yamato/var/strace/strace(main+0xa) [0x26ca] > /usr/lib64/libc-2.18.so(__libc_start_main+0xf5) [0x21d65] > /home/yamato/var/strace/strace(_start+0x29) [0x2799] In older version output lines of captured elements were built when printing. In this version they are built when capturing the stack. As result, unneeded dynamic memory allocations are avoided. Suggested by Luca Clementi. In older version the combination of snprintf and realloc were used. In this version they are replaced with asprintf. Suggested by Dmitry Levin. Signed-off-by: Masatake YAMATO <yamato@redhat.com>
2014-04-16 15:33:06 +09:00
unsigned long true_offset,
const char *error)
{
struct call_t *call;
call = xmalloc(sizeof(*call));
unwind: introduce queue_t for capturing stacktrace This is the second step for splitting capturing from printing. New `queue' field is added to tcb. Captured stacktrace is stored here. The field is initialized/finalized at unwind_tcb_init/unwind_tcb_fin. New API function unwind_capture_stacktrace is added. This function captures the currest stack using stracktrace_walker and records it in tcb. It's printing is delayed to the next call of unwind_print_stacktrace. unwind_print_stacktrace is extended. Now it checks queue field of the given tcb at the start of function. If the function finds a captured stack trace, the latter is printed using stracktrace_walker. Currently unwind_capture_stacktrace invocations are added directly to handlers of mmap, munmap, mprotect, and execve. Here is the difference of output with/without patch: (without patch) execve("./test-fork", ["./test-fork"], [/* 56 vars */]) = 0 > /usr/lib64/ld-2.18.so(check_one_fd.part.0+0x82) [0x11f0] (with patch) execve("./test-fork", ["./test-fork"], [/* 54 vars */]) = 0 > /usr/lib64/libc-2.18.so(execve+0x7) [0xbcd27] > /home/yamato/var/strace/strace(exec_or_die+0x10c) [0x26ac] > /home/yamato/var/strace/strace(startup_child+0x346) [0x134f6] > /home/yamato/var/strace/strace(init+0x89f) [0x13dff] > /home/yamato/var/strace/strace(main+0xa) [0x26ca] > /usr/lib64/libc-2.18.so(__libc_start_main+0xf5) [0x21d65] > /home/yamato/var/strace/strace(_start+0x29) [0x2799] In older version output lines of captured elements were built when printing. In this version they are built when capturing the stack. As result, unneeded dynamic memory allocations are avoided. Suggested by Luca Clementi. In older version the combination of snprintf and realloc were used. In this version they are replaced with asprintf. Suggested by Dmitry Levin. Signed-off-by: Masatake YAMATO <yamato@redhat.com>
2014-04-16 15:33:06 +09:00
call->output_line = sprint_call_or_error(binary_filename,
symbol_name,
function_offset,
unwind: introduce queue_t for capturing stacktrace This is the second step for splitting capturing from printing. New `queue' field is added to tcb. Captured stacktrace is stored here. The field is initialized/finalized at unwind_tcb_init/unwind_tcb_fin. New API function unwind_capture_stacktrace is added. This function captures the currest stack using stracktrace_walker and records it in tcb. It's printing is delayed to the next call of unwind_print_stacktrace. unwind_print_stacktrace is extended. Now it checks queue field of the given tcb at the start of function. If the function finds a captured stack trace, the latter is printed using stracktrace_walker. Currently unwind_capture_stacktrace invocations are added directly to handlers of mmap, munmap, mprotect, and execve. Here is the difference of output with/without patch: (without patch) execve("./test-fork", ["./test-fork"], [/* 56 vars */]) = 0 > /usr/lib64/ld-2.18.so(check_one_fd.part.0+0x82) [0x11f0] (with patch) execve("./test-fork", ["./test-fork"], [/* 54 vars */]) = 0 > /usr/lib64/libc-2.18.so(execve+0x7) [0xbcd27] > /home/yamato/var/strace/strace(exec_or_die+0x10c) [0x26ac] > /home/yamato/var/strace/strace(startup_child+0x346) [0x134f6] > /home/yamato/var/strace/strace(init+0x89f) [0x13dff] > /home/yamato/var/strace/strace(main+0xa) [0x26ca] > /usr/lib64/libc-2.18.so(__libc_start_main+0xf5) [0x21d65] > /home/yamato/var/strace/strace(_start+0x29) [0x2799] In older version output lines of captured elements were built when printing. In this version they are built when capturing the stack. As result, unneeded dynamic memory allocations are avoided. Suggested by Luca Clementi. In older version the combination of snprintf and realloc were used. In this version they are replaced with asprintf. Suggested by Dmitry Levin. Signed-off-by: Masatake YAMATO <yamato@redhat.com>
2014-04-16 15:33:06 +09:00
true_offset,
error);
call->next = NULL;
if (!queue->head) {
queue->head = call;
queue->tail = call;
} else {
queue->tail->next = call;
queue->tail = call;
}
}
static void
queue_put_call(void *queue,
const char *binary_filename,
const char *symbol_name,
unwind_function_offset_t function_offset,
unwind: introduce queue_t for capturing stacktrace This is the second step for splitting capturing from printing. New `queue' field is added to tcb. Captured stacktrace is stored here. The field is initialized/finalized at unwind_tcb_init/unwind_tcb_fin. New API function unwind_capture_stacktrace is added. This function captures the currest stack using stracktrace_walker and records it in tcb. It's printing is delayed to the next call of unwind_print_stacktrace. unwind_print_stacktrace is extended. Now it checks queue field of the given tcb at the start of function. If the function finds a captured stack trace, the latter is printed using stracktrace_walker. Currently unwind_capture_stacktrace invocations are added directly to handlers of mmap, munmap, mprotect, and execve. Here is the difference of output with/without patch: (without patch) execve("./test-fork", ["./test-fork"], [/* 56 vars */]) = 0 > /usr/lib64/ld-2.18.so(check_one_fd.part.0+0x82) [0x11f0] (with patch) execve("./test-fork", ["./test-fork"], [/* 54 vars */]) = 0 > /usr/lib64/libc-2.18.so(execve+0x7) [0xbcd27] > /home/yamato/var/strace/strace(exec_or_die+0x10c) [0x26ac] > /home/yamato/var/strace/strace(startup_child+0x346) [0x134f6] > /home/yamato/var/strace/strace(init+0x89f) [0x13dff] > /home/yamato/var/strace/strace(main+0xa) [0x26ca] > /usr/lib64/libc-2.18.so(__libc_start_main+0xf5) [0x21d65] > /home/yamato/var/strace/strace(_start+0x29) [0x2799] In older version output lines of captured elements were built when printing. In this version they are built when capturing the stack. As result, unneeded dynamic memory allocations are avoided. Suggested by Luca Clementi. In older version the combination of snprintf and realloc were used. In this version they are replaced with asprintf. Suggested by Dmitry Levin. Signed-off-by: Masatake YAMATO <yamato@redhat.com>
2014-04-16 15:33:06 +09:00
unsigned long true_offset)
{
queue_put(queue,
binary_filename,
symbol_name,
function_offset,
unwind: introduce queue_t for capturing stacktrace This is the second step for splitting capturing from printing. New `queue' field is added to tcb. Captured stacktrace is stored here. The field is initialized/finalized at unwind_tcb_init/unwind_tcb_fin. New API function unwind_capture_stacktrace is added. This function captures the currest stack using stracktrace_walker and records it in tcb. It's printing is delayed to the next call of unwind_print_stacktrace. unwind_print_stacktrace is extended. Now it checks queue field of the given tcb at the start of function. If the function finds a captured stack trace, the latter is printed using stracktrace_walker. Currently unwind_capture_stacktrace invocations are added directly to handlers of mmap, munmap, mprotect, and execve. Here is the difference of output with/without patch: (without patch) execve("./test-fork", ["./test-fork"], [/* 56 vars */]) = 0 > /usr/lib64/ld-2.18.so(check_one_fd.part.0+0x82) [0x11f0] (with patch) execve("./test-fork", ["./test-fork"], [/* 54 vars */]) = 0 > /usr/lib64/libc-2.18.so(execve+0x7) [0xbcd27] > /home/yamato/var/strace/strace(exec_or_die+0x10c) [0x26ac] > /home/yamato/var/strace/strace(startup_child+0x346) [0x134f6] > /home/yamato/var/strace/strace(init+0x89f) [0x13dff] > /home/yamato/var/strace/strace(main+0xa) [0x26ca] > /usr/lib64/libc-2.18.so(__libc_start_main+0xf5) [0x21d65] > /home/yamato/var/strace/strace(_start+0x29) [0x2799] In older version output lines of captured elements were built when printing. In this version they are built when capturing the stack. As result, unneeded dynamic memory allocations are avoided. Suggested by Luca Clementi. In older version the combination of snprintf and realloc were used. In this version they are replaced with asprintf. Suggested by Dmitry Levin. Signed-off-by: Masatake YAMATO <yamato@redhat.com>
2014-04-16 15:33:06 +09:00
true_offset,
NULL);
}
static void
queue_put_error(void *queue,
const char *error,
unsigned long ip)
unwind: introduce queue_t for capturing stacktrace This is the second step for splitting capturing from printing. New `queue' field is added to tcb. Captured stacktrace is stored here. The field is initialized/finalized at unwind_tcb_init/unwind_tcb_fin. New API function unwind_capture_stacktrace is added. This function captures the currest stack using stracktrace_walker and records it in tcb. It's printing is delayed to the next call of unwind_print_stacktrace. unwind_print_stacktrace is extended. Now it checks queue field of the given tcb at the start of function. If the function finds a captured stack trace, the latter is printed using stracktrace_walker. Currently unwind_capture_stacktrace invocations are added directly to handlers of mmap, munmap, mprotect, and execve. Here is the difference of output with/without patch: (without patch) execve("./test-fork", ["./test-fork"], [/* 56 vars */]) = 0 > /usr/lib64/ld-2.18.so(check_one_fd.part.0+0x82) [0x11f0] (with patch) execve("./test-fork", ["./test-fork"], [/* 54 vars */]) = 0 > /usr/lib64/libc-2.18.so(execve+0x7) [0xbcd27] > /home/yamato/var/strace/strace(exec_or_die+0x10c) [0x26ac] > /home/yamato/var/strace/strace(startup_child+0x346) [0x134f6] > /home/yamato/var/strace/strace(init+0x89f) [0x13dff] > /home/yamato/var/strace/strace(main+0xa) [0x26ca] > /usr/lib64/libc-2.18.so(__libc_start_main+0xf5) [0x21d65] > /home/yamato/var/strace/strace(_start+0x29) [0x2799] In older version output lines of captured elements were built when printing. In this version they are built when capturing the stack. As result, unneeded dynamic memory allocations are avoided. Suggested by Luca Clementi. In older version the combination of snprintf and realloc were used. In this version they are replaced with asprintf. Suggested by Dmitry Levin. Signed-off-by: Masatake YAMATO <yamato@redhat.com>
2014-04-16 15:33:06 +09:00
{
queue_put(queue, NULL, NULL, 0, ip, error);
}
static void
queue_print(struct unwind_queue_t *queue)
unwind: introduce queue_t for capturing stacktrace This is the second step for splitting capturing from printing. New `queue' field is added to tcb. Captured stacktrace is stored here. The field is initialized/finalized at unwind_tcb_init/unwind_tcb_fin. New API function unwind_capture_stacktrace is added. This function captures the currest stack using stracktrace_walker and records it in tcb. It's printing is delayed to the next call of unwind_print_stacktrace. unwind_print_stacktrace is extended. Now it checks queue field of the given tcb at the start of function. If the function finds a captured stack trace, the latter is printed using stracktrace_walker. Currently unwind_capture_stacktrace invocations are added directly to handlers of mmap, munmap, mprotect, and execve. Here is the difference of output with/without patch: (without patch) execve("./test-fork", ["./test-fork"], [/* 56 vars */]) = 0 > /usr/lib64/ld-2.18.so(check_one_fd.part.0+0x82) [0x11f0] (with patch) execve("./test-fork", ["./test-fork"], [/* 54 vars */]) = 0 > /usr/lib64/libc-2.18.so(execve+0x7) [0xbcd27] > /home/yamato/var/strace/strace(exec_or_die+0x10c) [0x26ac] > /home/yamato/var/strace/strace(startup_child+0x346) [0x134f6] > /home/yamato/var/strace/strace(init+0x89f) [0x13dff] > /home/yamato/var/strace/strace(main+0xa) [0x26ca] > /usr/lib64/libc-2.18.so(__libc_start_main+0xf5) [0x21d65] > /home/yamato/var/strace/strace(_start+0x29) [0x2799] In older version output lines of captured elements were built when printing. In this version they are built when capturing the stack. As result, unneeded dynamic memory allocations are avoided. Suggested by Luca Clementi. In older version the combination of snprintf and realloc were used. In this version they are replaced with asprintf. Suggested by Dmitry Levin. Signed-off-by: Masatake YAMATO <yamato@redhat.com>
2014-04-16 15:33:06 +09:00
{
struct call_t *call, *tmp;
queue->tail = NULL;
call = queue->head;
queue->head = NULL;
while (call) {
tmp = call;
call = call->next;
tprints(tmp->output_line);
line_ended();
if (tmp->output_line != asprintf_error_str)
free(tmp->output_line);
unwind: introduce queue_t for capturing stacktrace This is the second step for splitting capturing from printing. New `queue' field is added to tcb. Captured stacktrace is stored here. The field is initialized/finalized at unwind_tcb_init/unwind_tcb_fin. New API function unwind_capture_stacktrace is added. This function captures the currest stack using stracktrace_walker and records it in tcb. It's printing is delayed to the next call of unwind_print_stacktrace. unwind_print_stacktrace is extended. Now it checks queue field of the given tcb at the start of function. If the function finds a captured stack trace, the latter is printed using stracktrace_walker. Currently unwind_capture_stacktrace invocations are added directly to handlers of mmap, munmap, mprotect, and execve. Here is the difference of output with/without patch: (without patch) execve("./test-fork", ["./test-fork"], [/* 56 vars */]) = 0 > /usr/lib64/ld-2.18.so(check_one_fd.part.0+0x82) [0x11f0] (with patch) execve("./test-fork", ["./test-fork"], [/* 54 vars */]) = 0 > /usr/lib64/libc-2.18.so(execve+0x7) [0xbcd27] > /home/yamato/var/strace/strace(exec_or_die+0x10c) [0x26ac] > /home/yamato/var/strace/strace(startup_child+0x346) [0x134f6] > /home/yamato/var/strace/strace(init+0x89f) [0x13dff] > /home/yamato/var/strace/strace(main+0xa) [0x26ca] > /usr/lib64/libc-2.18.so(__libc_start_main+0xf5) [0x21d65] > /home/yamato/var/strace/strace(_start+0x29) [0x2799] In older version output lines of captured elements were built when printing. In this version they are built when capturing the stack. As result, unneeded dynamic memory allocations are avoided. Suggested by Luca Clementi. In older version the combination of snprintf and realloc were used. In this version they are replaced with asprintf. Suggested by Dmitry Levin. Signed-off-by: Masatake YAMATO <yamato@redhat.com>
2014-04-16 15:33:06 +09:00
tmp->output_line = NULL;
tmp->next = NULL;
free(tmp);
}
}
/*
* printing stack
*/
void
unwind_tcb_print(struct tcb *tcp)
{
#if SUPPORTED_PERSONALITIES > 1
if (tcp->currpers != DEFAULT_PERSONALITY) {
/* disable stack trace */
return;
}
#endif
if (tcp->unwind_queue->head) {
debug_func_msg("head: tcp=%p, queue=%p",
tcp, tcp->unwind_queue->head);
queue_print(tcp->unwind_queue);
unwind-libdw: use the mmap_notify subsystem The unwind subsystem uses the mmap_cache subsystem even it uses unwind-libdw as backend. unwind-libdw doesn't need the full set of the mmap_cache subsystem; libdw has a feature for caching a memory mapping. This commit does three things. (1) Make the unwind subsystem not use the mmap_cache subsystem. The unwind subsystem never concern the memory mapping of the target. It becomes a thin layer. (2) Make unwind-libunwind use the mmap_cache subsystem directly. (3) Make unwind-libdw use the mmap_notify subsystem to know when it should call dwfl_linux_proc_report/dwfl_report_end for updating the cache. Here is a subsystem structure that this patch introduces: +-------------------------------------+ | unwind subsys | +------------------+------------------+ | unwind-libunwind | unwind-libdw | +------------------+------------------+ | mmap_cache | | +------------------+ | | mmap_notify | +-------------------------------------+ | syscall | +-------------------------------------+ mmap/munmap/mprotect/brk... * unwind.c: Don't include "mmap_cache.h". (unwind_init): Don't call mmap_cache_enable. (unwind_tcb_print, unwind_tcb_capture): Don't call mmap_cache related functions, just invoke unwinder.tcb_walk. * unwind.h (struct unwind_unwinder_t): Remove tcb_flush_cache field. * unwind-libdw.c: Include "mmap_notify.h" instead of "mmap_cache.h". (struct ctx): Add last_proc_updating field to record the generation of memory mapping that is cached by dwfl_linux_proc_report and dwfl_report_end. (mapping_generation): A variable counting how many times the memory mapping of targets has been changed. (updating_mapping_generation): New utility function for updating mapping_generation. (init): New function for registering updating_mapping_generation in the mmap_notify subsystem as a callback function. (tcb_init): Initialize ctx::last_proc_updating. (tcb_flush_cache): Rename to flush_cache_maybe. Rebuild the cache data only if the data is stale. (tcb_walk): Call flush_cache_maybe for avoiding referring staled cache data. (unwinder): Set init function, remove tcb_flush_cache field. * unwind-libunwind.c (init): Enable the mmap_cache subsystem. (tcb_walk): Call mmap_cache_rebuild_if_invalid and unw_flush_cache for updating the cache of the memory mapping before walking the stack. (tcb_walk): Rename to walk. (unwinder): Remove tcb_flush_cache field. Signed-off-by: Masatake YAMATO <yamato@redhat.com>
2018-04-30 06:45:40 +09:00
} else
unwinder.tcb_walk(tcp, print_call_cb, print_error_cb, NULL);
unwind: introduce queue_t for capturing stacktrace This is the second step for splitting capturing from printing. New `queue' field is added to tcb. Captured stacktrace is stored here. The field is initialized/finalized at unwind_tcb_init/unwind_tcb_fin. New API function unwind_capture_stacktrace is added. This function captures the currest stack using stracktrace_walker and records it in tcb. It's printing is delayed to the next call of unwind_print_stacktrace. unwind_print_stacktrace is extended. Now it checks queue field of the given tcb at the start of function. If the function finds a captured stack trace, the latter is printed using stracktrace_walker. Currently unwind_capture_stacktrace invocations are added directly to handlers of mmap, munmap, mprotect, and execve. Here is the difference of output with/without patch: (without patch) execve("./test-fork", ["./test-fork"], [/* 56 vars */]) = 0 > /usr/lib64/ld-2.18.so(check_one_fd.part.0+0x82) [0x11f0] (with patch) execve("./test-fork", ["./test-fork"], [/* 54 vars */]) = 0 > /usr/lib64/libc-2.18.so(execve+0x7) [0xbcd27] > /home/yamato/var/strace/strace(exec_or_die+0x10c) [0x26ac] > /home/yamato/var/strace/strace(startup_child+0x346) [0x134f6] > /home/yamato/var/strace/strace(init+0x89f) [0x13dff] > /home/yamato/var/strace/strace(main+0xa) [0x26ca] > /usr/lib64/libc-2.18.so(__libc_start_main+0xf5) [0x21d65] > /home/yamato/var/strace/strace(_start+0x29) [0x2799] In older version output lines of captured elements were built when printing. In this version they are built when capturing the stack. As result, unneeded dynamic memory allocations are avoided. Suggested by Luca Clementi. In older version the combination of snprintf and realloc were used. In this version they are replaced with asprintf. Suggested by Dmitry Levin. Signed-off-by: Masatake YAMATO <yamato@redhat.com>
2014-04-16 15:33:06 +09:00
}
/*
* capturing stack
*/
void
unwind_tcb_capture(struct tcb *tcp)
unwind: introduce queue_t for capturing stacktrace This is the second step for splitting capturing from printing. New `queue' field is added to tcb. Captured stacktrace is stored here. The field is initialized/finalized at unwind_tcb_init/unwind_tcb_fin. New API function unwind_capture_stacktrace is added. This function captures the currest stack using stracktrace_walker and records it in tcb. It's printing is delayed to the next call of unwind_print_stacktrace. unwind_print_stacktrace is extended. Now it checks queue field of the given tcb at the start of function. If the function finds a captured stack trace, the latter is printed using stracktrace_walker. Currently unwind_capture_stacktrace invocations are added directly to handlers of mmap, munmap, mprotect, and execve. Here is the difference of output with/without patch: (without patch) execve("./test-fork", ["./test-fork"], [/* 56 vars */]) = 0 > /usr/lib64/ld-2.18.so(check_one_fd.part.0+0x82) [0x11f0] (with patch) execve("./test-fork", ["./test-fork"], [/* 54 vars */]) = 0 > /usr/lib64/libc-2.18.so(execve+0x7) [0xbcd27] > /home/yamato/var/strace/strace(exec_or_die+0x10c) [0x26ac] > /home/yamato/var/strace/strace(startup_child+0x346) [0x134f6] > /home/yamato/var/strace/strace(init+0x89f) [0x13dff] > /home/yamato/var/strace/strace(main+0xa) [0x26ca] > /usr/lib64/libc-2.18.so(__libc_start_main+0xf5) [0x21d65] > /home/yamato/var/strace/strace(_start+0x29) [0x2799] In older version output lines of captured elements were built when printing. In this version they are built when capturing the stack. As result, unneeded dynamic memory allocations are avoided. Suggested by Luca Clementi. In older version the combination of snprintf and realloc were used. In this version they are replaced with asprintf. Suggested by Dmitry Levin. Signed-off-by: Masatake YAMATO <yamato@redhat.com>
2014-04-16 15:33:06 +09:00
{
#if SUPPORTED_PERSONALITIES > 1
if (tcp->currpers != DEFAULT_PERSONALITY) {
/* disable stack trace */
return;
}
#endif
if (tcp->unwind_queue->head)
unwind: introduce queue_t for capturing stacktrace This is the second step for splitting capturing from printing. New `queue' field is added to tcb. Captured stacktrace is stored here. The field is initialized/finalized at unwind_tcb_init/unwind_tcb_fin. New API function unwind_capture_stacktrace is added. This function captures the currest stack using stracktrace_walker and records it in tcb. It's printing is delayed to the next call of unwind_print_stacktrace. unwind_print_stacktrace is extended. Now it checks queue field of the given tcb at the start of function. If the function finds a captured stack trace, the latter is printed using stracktrace_walker. Currently unwind_capture_stacktrace invocations are added directly to handlers of mmap, munmap, mprotect, and execve. Here is the difference of output with/without patch: (without patch) execve("./test-fork", ["./test-fork"], [/* 56 vars */]) = 0 > /usr/lib64/ld-2.18.so(check_one_fd.part.0+0x82) [0x11f0] (with patch) execve("./test-fork", ["./test-fork"], [/* 54 vars */]) = 0 > /usr/lib64/libc-2.18.so(execve+0x7) [0xbcd27] > /home/yamato/var/strace/strace(exec_or_die+0x10c) [0x26ac] > /home/yamato/var/strace/strace(startup_child+0x346) [0x134f6] > /home/yamato/var/strace/strace(init+0x89f) [0x13dff] > /home/yamato/var/strace/strace(main+0xa) [0x26ca] > /usr/lib64/libc-2.18.so(__libc_start_main+0xf5) [0x21d65] > /home/yamato/var/strace/strace(_start+0x29) [0x2799] In older version output lines of captured elements were built when printing. In this version they are built when capturing the stack. As result, unneeded dynamic memory allocations are avoided. Suggested by Luca Clementi. In older version the combination of snprintf and realloc were used. In this version they are replaced with asprintf. Suggested by Dmitry Levin. Signed-off-by: Masatake YAMATO <yamato@redhat.com>
2014-04-16 15:33:06 +09:00
error_msg_and_die("bug: unprinted entries in queue");
unwind-libdw: use the mmap_notify subsystem The unwind subsystem uses the mmap_cache subsystem even it uses unwind-libdw as backend. unwind-libdw doesn't need the full set of the mmap_cache subsystem; libdw has a feature for caching a memory mapping. This commit does three things. (1) Make the unwind subsystem not use the mmap_cache subsystem. The unwind subsystem never concern the memory mapping of the target. It becomes a thin layer. (2) Make unwind-libunwind use the mmap_cache subsystem directly. (3) Make unwind-libdw use the mmap_notify subsystem to know when it should call dwfl_linux_proc_report/dwfl_report_end for updating the cache. Here is a subsystem structure that this patch introduces: +-------------------------------------+ | unwind subsys | +------------------+------------------+ | unwind-libunwind | unwind-libdw | +------------------+------------------+ | mmap_cache | | +------------------+ | | mmap_notify | +-------------------------------------+ | syscall | +-------------------------------------+ mmap/munmap/mprotect/brk... * unwind.c: Don't include "mmap_cache.h". (unwind_init): Don't call mmap_cache_enable. (unwind_tcb_print, unwind_tcb_capture): Don't call mmap_cache related functions, just invoke unwinder.tcb_walk. * unwind.h (struct unwind_unwinder_t): Remove tcb_flush_cache field. * unwind-libdw.c: Include "mmap_notify.h" instead of "mmap_cache.h". (struct ctx): Add last_proc_updating field to record the generation of memory mapping that is cached by dwfl_linux_proc_report and dwfl_report_end. (mapping_generation): A variable counting how many times the memory mapping of targets has been changed. (updating_mapping_generation): New utility function for updating mapping_generation. (init): New function for registering updating_mapping_generation in the mmap_notify subsystem as a callback function. (tcb_init): Initialize ctx::last_proc_updating. (tcb_flush_cache): Rename to flush_cache_maybe. Rebuild the cache data only if the data is stale. (tcb_walk): Call flush_cache_maybe for avoiding referring staled cache data. (unwinder): Set init function, remove tcb_flush_cache field. * unwind-libunwind.c (init): Enable the mmap_cache subsystem. (tcb_walk): Call mmap_cache_rebuild_if_invalid and unw_flush_cache for updating the cache of the memory mapping before walking the stack. (tcb_walk): Rename to walk. (unwinder): Remove tcb_flush_cache field. Signed-off-by: Masatake YAMATO <yamato@redhat.com>
2018-04-30 06:45:40 +09:00
else {
debug_func_msg("walk: tcp=%p, queue=%p",
tcp, tcp->unwind_queue->head);
unwind-libdw: use the mmap_notify subsystem The unwind subsystem uses the mmap_cache subsystem even it uses unwind-libdw as backend. unwind-libdw doesn't need the full set of the mmap_cache subsystem; libdw has a feature for caching a memory mapping. This commit does three things. (1) Make the unwind subsystem not use the mmap_cache subsystem. The unwind subsystem never concern the memory mapping of the target. It becomes a thin layer. (2) Make unwind-libunwind use the mmap_cache subsystem directly. (3) Make unwind-libdw use the mmap_notify subsystem to know when it should call dwfl_linux_proc_report/dwfl_report_end for updating the cache. Here is a subsystem structure that this patch introduces: +-------------------------------------+ | unwind subsys | +------------------+------------------+ | unwind-libunwind | unwind-libdw | +------------------+------------------+ | mmap_cache | | +------------------+ | | mmap_notify | +-------------------------------------+ | syscall | +-------------------------------------+ mmap/munmap/mprotect/brk... * unwind.c: Don't include "mmap_cache.h". (unwind_init): Don't call mmap_cache_enable. (unwind_tcb_print, unwind_tcb_capture): Don't call mmap_cache related functions, just invoke unwinder.tcb_walk. * unwind.h (struct unwind_unwinder_t): Remove tcb_flush_cache field. * unwind-libdw.c: Include "mmap_notify.h" instead of "mmap_cache.h". (struct ctx): Add last_proc_updating field to record the generation of memory mapping that is cached by dwfl_linux_proc_report and dwfl_report_end. (mapping_generation): A variable counting how many times the memory mapping of targets has been changed. (updating_mapping_generation): New utility function for updating mapping_generation. (init): New function for registering updating_mapping_generation in the mmap_notify subsystem as a callback function. (tcb_init): Initialize ctx::last_proc_updating. (tcb_flush_cache): Rename to flush_cache_maybe. Rebuild the cache data only if the data is stale. (tcb_walk): Call flush_cache_maybe for avoiding referring staled cache data. (unwinder): Set init function, remove tcb_flush_cache field. * unwind-libunwind.c (init): Enable the mmap_cache subsystem. (tcb_walk): Call mmap_cache_rebuild_if_invalid and unw_flush_cache for updating the cache of the memory mapping before walking the stack. (tcb_walk): Rename to walk. (unwinder): Remove tcb_flush_cache field. Signed-off-by: Masatake YAMATO <yamato@redhat.com>
2018-04-30 06:45:40 +09:00
unwinder.tcb_walk(tcp, queue_put_call, queue_put_error,
tcp->unwind_queue);
}
}