48 Commits

Author SHA1 Message Date
5bada1246d Merge 10.5 into 10.6 2023-04-11 16:15:19 +03:00
3261a78ea1 Merge branch '10.4' into 10.5 2023-04-03 09:34:26 +02:00
3b64244070 Handle meaningless addr2line results and increase timeout
MariaDB server prints the stack information if a crash happens.

It traverses the stack frames in function `print_with_addr_resolve`.
For *EACH* frame, it tries to parse the file name and line number of the
frame using `addr2line`, or prints `backtrace_symbols_fd` if `addr2line`
fails.

1. Logic in `addr_resolve` function uses addr2line to get the file name
   and line numbers. It has a timeout of 500ms to wait for the response
   from addr2line. However, that's not enough on small instances
   especially if the debug information is in a separate file or
   compressed.

   Increase the timeout to 5 seconds to support some edge cases, as
   experiments showed addr2line may take 2-3 seconds on some frames.

2. While parsing a frame inside of a shared library using `addr2line`,
   the file name and line numbers could be `??`, empty or `0` if the
   debug info is not loaded.
   It's easy to reproduce when glibc-debuginfo is not installed.

   Instead of printing a meaningless frame like:

       :0(__GI___poll)[0x1505e9197639]
       ...
       ??:0(__libc_start_main)[0x7ffff6c8913a]

   We want to print the frame information using `backtrace_symbols_fd`,
   with the shared library name and a hexadecimal offset.
   Stacktrace example on a real instance with this commit:

       /lib64/libc.so.6(__poll+0x49)[0x145cbf71a639]
       ...
       /lib64/libc.so.6(__libc_start_main+0xea)[0x7f4d0034d13a]

   `addr_resolve` has considered the case of meaningless combination of
   file name and line number returned by `addr2line`. e.g. `??:?`
   However, conditions like `:0` and `??:0` are not handled. So now the
   function will rollback to `backtrace_symbols_fd` in above cases.

All new code of the whole pull request, including one or several files
that are either new files or modified ones, are contributed under the
BSD-new license. I am contributing on behalf of my employer Amazon Web
Services, Inc.
2023-04-01 10:07:42 +03:00
2ac1edb1c3 Merge 10.5 into 10.6 2022-11-08 17:37:22 +02:00
a732d5e2ba Merge 10.4 into 10.5 2022-11-08 17:01:28 +02:00
92be8d2048 MDEV-29951 server hang in crash handler
When trying to output stacktrace, and addr2line is not installed, the
child process forked by start_addr2line_fork() will fail to do exec(),
and finish with exit(1).

There is a problem with exit() though - it runs exit handlers,
and for the forked copy of crashing process, it is a bad idea.

In 10.5+ code for example, exit handlers include
tpool::task_group static destructors, and it will hang infinitely
waiting for completion of the outstanding tasks.

The fix is to use _exit() instead, which skips the execution of exit
handlers
2022-11-05 18:36:43 +01:00
18a0f0c178 Fix AIX compulation (break addr resolution)
Revert c92c161585, casting to void*
cannot fool gcc. The AIX address resolution was broken anyway, so
just avoid the problem and make it compile well.
2022-10-27 11:00:54 +11:00
c92c161585 Fixed compiler warning on AIX 2022-10-19 01:02:29 +03:00
88fb89acb7 Fixes some compiler issues on AIX ( 2022-02-08 14:32:28 +02:00
df02de68f3 Fixed my_addr_resolve (cherry picked from 10.6)
When a server is compiled with -fPIE, my_addr_resolve needs to
subtract the info.dli_fbase from symbol addresses in memory for
addr2line to recognize them.  When a server is compiled without -fPIE,
my_addr_resolve should not do it.  Unfortunately not all compilers
define __PIE__ when -fPIE was used (e.g. older gcc doesn't), so we
have to resort to run-time detection.
2022-02-08 14:32:28 +02:00
a93c514595 Fixed my_addr_resolve
When a server is compiled with -fPIE, my_addr_resolve needs to subtract the info.dli_fbase from symbol addresses in memory for addr2line to recognize them.
When a server is compiled without -fPIE, my_addr_resolve should not do it.
Unfortunately not all compilers define __PIE__ when -fPIE was used
(e.g. older gcc doesn't), so we have to resort to run-time detection.
2021-05-19 22:27:27 +02:00
02e7bff882 Merge commit '10.4' into 10.5 2021-01-06 10:53:00 +01:00
384f107ae5 MDEV-21646: postfix - my_addr_resolve: static Dl_info info
Encountered the linker failure on Debug build in 10.4:

[53/585] Linking CXX executable unittest/sql/mf_iocache-t
FAILED: unittest/sql/mf_iocache-t
: && /usr/bin/c++  -pie -fPIC -fstack-protector --param=ssp-buffer-size=4 -fPIC -g -DENABLED_DEBUG_SYNC -ggdb3 -DSAFE_MUTEX -DSAFEMALLOC -DTRASH_FREED_MEMORY -Wall -Wextra -Wno-format-truncation -Wno-init-self -Wno-nonnull-compare -Wno-unused-parameter -Woverloaded-virtual -Wnon-virtual-dtor -Wvla -Wwrite-strings -Werror  -Wl,-z,relro,-z,now unittest/sql/CMakeFiles/mf_iocache-t.dir/mf_iocache-t.cc.o unittest/sql/CMakeFiles/mf_iocache-t.dir/__/__/sql/mf_iocache_encr.cc.o  -o unittest/sql/mf_iocache-t  -lpthread  mysys/libmysys.a  unittest/mytap/libmytap.a  mysys_ssl/libmysys_ssl.a  mysys/libmysys.a  dbug/libdbug.a  mysys/libmysys.a  dbug/libdbug.a  -lz  -lm  strings/libstrings.a  -lpthread  -lssl  -lcrypto  -ldl && :
/usr/bin/ld: mysys/libmysys.a(my_addr_resolve.c.o):/home/dan/repos/mariadb-server-10.4/mysys/my_addr_resolve.c:173: multiple definition of `info'; unittest/sql/CMakeFiles/mf_iocache-t.dir/mf_iocache-t.cc.o:/home/dan/repos/mariadb-server-10.4/unittest/sql/mf_iocache-t.cc:99: first defined here

We make Dl_info static as in MDEV-21646 moving it out of the function
was the main goal and having it scope limited by static doesn't affect
the function.
2020-12-15 14:47:16 +11:00
8e1e2856f2 Merge branch '10.4' into 10.5 2020-11-01 14:26:15 +01:00
65e26bc1ba Merge branch '10.1' into 10.2 2020-10-28 10:56:38 +01:00
8584349108 MDEV-14945 possible buffer overflow in stack resolver
According to https://stackoverflow.com/questions/22827510/how-to-avoid-bad-fd-set-buffer-overflow-crash
it seems that using select instead of poll can cause additional memory
allocations. As we are in a crashed state, we must prevent allocating
any memory (if possible). Thus, switch select call to poll.

Also move some bigger datastructures to global space. The code is not
run in a multithreaded context so best we don't use up stack space
if it's not needed.
2020-10-26 21:02:13 +02:00
0afccbf7b8 restore stack traces that were broken by ebfe8c4e0e 2020-03-16 01:13:01 +01:00
fad47df995 Merge branch '10.4' into 10.5 2020-03-11 17:52:49 +01:00
75d286c2cc Merge branch '10.1' into 10.2 2020-03-06 15:42:45 +01:00
345aaca14c Merge branch '5.5' into 10.1 2020-03-06 11:06:48 +01:00
be77fa914c MDEV-21646: Failure to compile my_addr_resolve.c with binutils-2.34
Emulate older API
2020-02-12 15:39:36 +01:00
826f9d4f7e Merge 10.4 into 10.5 2019-05-23 10:32:21 +03:00
ebfe8c4e0e Make it trivial to get stack traces from external programs.
To get a stacktrace in a program like aria_chk, one only have to do:

#include <my_stacktrace.h>

call my_init_stacktrace(1) in main().
2019-05-23 02:28:23 +03:00
26a14ee130 Merge 10.1 into 10.2 2019-05-13 17:54:04 +03:00
cb248f8806 Merge branch '5.5' into 10.1 2019-05-11 22:19:05 +03:00
5543b75550 Update FSF Address
* Update wrong zip-code
2019-05-11 21:29:06 +03:00
45e6d0aebf Merge branch '10.1' into 10.2 2018-04-10 17:43:18 +03:00
4c89cff558 Merge branch '10.0' into 10.1 2018-04-07 17:11:22 +03:00
df6197c8b9 compiler warning
warning: format '%p' expects argument of type 'void *', but argument 4 has type 'long int'
2018-04-03 16:19:10 +02:00
e92cc09765 MDEV-15345 Compilation fails to build my_addr_resolve.c
fix the compilation error.
no support for plugins yet.
2018-02-22 16:46:11 +01:00
4771ae4b22 Merge branch 'github/10.1' into 10.2 2018-02-06 14:50:50 +01:00
29031fe391 my_addr_resolve fixes
* use a longer timeout, 10ms is not always enough
* stop waiting as long as the data is read, do not unconditionally
  wait for timeout every time
2018-02-02 11:42:07 +01:00
d4df7bc9b1 Merge branch 'github/10.0' into 10.1 2018-02-02 10:09:44 +01:00
26e5f9dda1 MDEV-14229: Stack trace is not resolved for shared objects
Resolving a stacktrace including functions in dynamic libraries requires
us to look inside the libraries for the symbols. Addr2line needs to be
started with the correct binary for each address on the stack. To do this,
figure out which library it is using dladdr, then if the addr2line
binary was started with a different binary, fork it again with the
correct one.

We only have one addr2line process running at any point during the
stacktrace resolving step. The maximum number of forks for addr2line should
generally be around 6.

One for server stacktrace code, one for plugin code, one when going back
into server code, one for pthread library, one for libc, one for the
_start function in the server. More can come up if plugin calls server
function which goes back to a plugin, etc.
2018-01-19 16:11:34 +02:00
e336fc07bd fix stack traces when linking with libbfd
also, return different values for different errors in my_addr_resolve()
2016-12-12 20:27:20 +01:00
7d39b28093 [MDEV-9468]: Client hangs in my_addr_resolve
Account for timeout of select returning a 0 error code.
2016-01-26 21:42:37 +02:00
275f7d7d82 Remove warning in my_addr_resolve
ISO C90 does not allow mixed declarations and code.
2016-01-17 22:33:36 +02:00
b2bd10d4d6 [MDEV-9427] Server does not build on OpenSUSE 42.1
Changed code to comply to C90 standard.
2016-01-17 22:26:22 +02:00
6d3ffd2e3a Fixed a crash during stacktrace printing if addr2line failed to start.
In order to get all the input from addr2line we must read in a loop,
until the response is complete. Also, in case that the response is
malformed, we must not end up reading invalid memory.
2016-01-17 14:20:09 +02:00
1993780284 fix build on sol10-64 2015-10-15 12:23:04 +02:00
c78f594bbc MDEV-6479 stack traces in 10.1
Take into account that PIE binaries are loaded at some offset, so
addresses cannot be directly resolved with addr2line. Find this
offset and subtract it before resolving an address.
2015-02-28 19:48:22 +01:00
ca083a764f my_addr_resolve: don't resolve unknown addresses to ??:0(??), but return an error instead
(better to have an address in the output than ??:0)
2013-12-12 18:14:14 +01:00
20e706689d mysql-5.5.22 merge
mysql-test/suite/innodb/t/group_commit_crash.test:
  remove autoincrement to avoid rbr being used for insert ... select
mysql-test/suite/innodb/t/group_commit_crash_no_optimize_thread.test:
  remove autoincrement to avoid rbr being used for insert ... select
mysys/my_addr_resolve.c:
  a pointer to a buffer is returned to the caller -> the buffer cannot be on the stack
mysys/stacktrace.c:
  my_vsnprintf() is ok here, in 5.5
2012-03-28 01:04:46 +02:00
d92009231d a couple of minor post-5.5-merge cleanups
include/mysql_com.h:
  remove "shutdown levels" that aren't shutdown levels from mysql_enum_shutdown_level
mysys/my_addr_resolve.c:
  my_snprintf in 5.5 (but not in 5.3) supports %p
sql/item_func.cc:
  use a method (that exists only in 5.5) instead of directly accessing a member
sql/item_subselect.cc:
  use a method (that exists only in 5.5) instead of directly accessing a member
sql/opt_subselect.cc:
  use a method (that exists only in 5.5) instead of directly accessing a member
sql/sql_select.cc:
  use a method (that exists only in 5.5) instead of directly accessing a member
2012-03-26 12:33:49 +02:00
335de5db18 Merge mariadb 5.3->mariadb 5.5 2012-03-24 18:21:22 +01:00
00f2e3139f alternative method of resolving addresses for safemalloc and crash handler.
don't link with libbfd, exec addr2line, if it's available at run time
2012-03-13 13:34:24 +01:00
db8f698c8a new configure option: NOT_FOR_DISTRIBUTION
fix safemalloc to compile w/o libbfd.

CMakeLists.txt:
  NOT_FOR_DISTRIBUTION option
cmake/readline.cmake:
  simplify libedit/readline detection.
  never use bundled libedit.
  use system readline v6 only if NOT_FOR_DISTRIBUTION=1
configure.cmake:
  use libbfd only if NOT_FOR_DISTRIBUTION=1
include/my_stacktrace.h:
  link with libbfd even w/o safemalloc.
2011-12-14 10:53:32 +01:00
37c81d81b3 another backtrace resolver that prints source file name and line number 2011-12-11 09:00:12 +01:00