This is the mail archive of the
systemtap@sourceware.org
mailing list for the systemtap project.
[Bug uprobes/18171] setjmp/longjmp clashing with u[ret]probes, causing SIGSEGV or SIGILL
- From: "mcermak at redhat dot com" <sourceware-bugzilla at sourceware dot org>
- To: systemtap at sourceware dot org
- Date: Mon, 30 Mar 2015 14:16:56 +0000
- Subject: [Bug uprobes/18171] setjmp/longjmp clashing with u[ret]probes, causing SIGSEGV or SIGILL
- Auto-submitted: auto-generated
- References: <bug-18171-6586 at http dot sourceware dot org/bugzilla/>
https://sourceware.org/bugzilla/show_bug.cgi?id=18171
--- Comment #1 from Martin Cermak <mcermak at redhat dot com> ---
Reproduced using perf. Problem is the order of commandline switches produced by
stap2perf: Flipping -a and -x seems to help:
=======
# perf probe -x /usr/share/systemtap/testsuite/bz5274 -a
bz5274_funca_return=funca%return
# perf probe -x /usr/share/systemtap/testsuite/bz5274 -a
bz5274_funcb_return=funcb%return
...
=======
After this I have:
=======
# cat //sys/kernel/debug/tracing/uprobe_events
r:probe_bz5274/bz5274_funca_return
/usr/share/systemtap/testsuite/bz5274:0x00000000000006e6
r:probe_bz5274/bz5274_funcb_return
/usr/share/systemtap/testsuite/bz5274:0x00000000000006a0
p:probe_bz5274/bz5274_funcc
/usr/share/systemtap/testsuite/bz5274:0x000000000000065a
p:probe_bz5274/bz5274_funcd
/usr/share/systemtap/testsuite/bz5274:0x0000000000000616
r:probe_bz5274/bz5274_main_return
/usr/share/systemtap/testsuite/bz5274:0x0000000000000725
=======
Now using -e 'probe_bz5274:*' instead of -e 'probe:*' I'm getting:
=======
# /root/bin/perf record -e 'probe_bz5274:*' -aR ./bz5274
In bz5274.c: funcb :20 : i=1. Calling funcc
In bz5274.c: funcc :13 : i=1. Calling funcd
In bz5274.c: funcd :7 : i=1. Calling longjmp
In bz5274.c: funcb :20 : i=2. Calling funcc
In bz5274.c: funcc :13 : i=2. Calling funcd
In bz5274.c: funcd :7 : i=2. Calling longjmp
In bz5274.c: funcb :20 : i=3. Calling funcc
In bz5274.c: funcc :13 : i=3. Calling funcd
In bz5274.c: funcd :7 : i=3. Calling longjmp
In bz5274.c: funcb :20 : i=4. Calling funcc
In bz5274.c: funcc :13 : i=4. Calling funcd
In bz5274.c: funcd :7 : i=4. Calling longjmp
[ perf record: Woken up 1 times to write data ]
[ perf record: Captured and wrote 0.166 MB perf.data (10 samples) ]
Segmentation fault (core dumped)
[root@dhcp-1-102 testsuite]#
=======
But now I'm getting two corefiles:
=======
[root@dhcp-1-102 testsuite]# ls core*
core.2097 core.2098
[root@dhcp-1-102 testsuite]# file core.2200
core.2200: ELF 64-bit LSB core file x86-64, version 1 (SYSV), too many program
headers (163)
[root@dhcp-1-102 testsuite]# file core.2201
core.2201: ELF 64-bit LSB core file x86-64, version 1 (SYSV), SVR4-style, from
'./bz5274'
[root@dhcp-1-102 testsuite]#
[root@dhcp-1-102 testsuite]# gdb -q ./bz5274 core.2097
Reading symbols from ./bz5274...done.
warning: core file may not match specified executable file.
---------^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^---------------------------------------------------
[New LWP 2097]
Core was generated by `/root/bin/perf record -e probe_bz5274:* -aR ./bz5274'.
Program terminated with signal SIGSEGV, Segmentation fault.
#0 0x0000003e38e0ff99 in ?? ()
(gdb) info registers
rax 0x0 0
rbx 0x0 0
rcx 0xffffffffffffffff -1
rdx 0xb 11
rsi 0x831 2097
rdi 0x831 2097
rbp 0x3e389b76a0 0x3e389b76a0
rsp 0x7fff705200f8 0x7fff705200f8
r8 0x7fff70520030 140735077810224
r9 0x7fff7051fe40 140735077809728
r10 0xa5 165
r11 0x202 514
r12 0x1 1
r13 0x3e389b8e80 267237691008
r14 0x8365d0 8611280
r15 0x0 0
rip 0x3e38e0ff99 0x3e38e0ff99
eflags 0x202 [ IF ]
cs 0x33 51
ss 0x2b 43
ds 0x0 0
es 0x0 0
fs 0x0 0
gs 0x0 0
(gdb) q
[root@dhcp-1-102 testsuite]#
[root@dhcp-1-102 testsuite]# gdb -q ./bz5274 core.2098
Reading symbols from ./bz5274...done.
[New LWP 2098]
Core was generated by `./bz5274'.
Program terminated with signal SIGSEGV, Segmentation fault.
#0 0x0000000000400724 in funca (s=0xffff95e800400876 <error: Cannot access
memory at address 0xffff95e800400876>, env=0xbfc68948ffffff30) at
systemtap.base/bz5274.c:34
34 }
(gdb) info registers
rax 0x4 4
rbx 0x0 0
rcx 0x7fffffe4 2147483620
rdx 0x400702 4196098
rsi 0x4 4
rdi 0x7fff359cc380 140734092854144
rbp 0x41ff894156415741 0x41ff894156415741
rsp 0x400768 0x400768 <__libc_csu_init+8>
r8 0x7fff359cc340 140734092854080
r9 0x7fff359cc360 140734092854112
r10 0x0 0
r11 0x246 582
r12 0x400520 4195616
r13 0x7fff359cc530 140734092854576
r14 0x0 0
r15 0x0 0
rip 0x400724 0x400724 <funca+62>
eflags 0x10246 [ PF ZF IF RF ]
cs 0x33 51
ss 0x2b 43
ds 0x0 0
es 0x0 0
fs 0x0 0
gs 0x0 0
(gdb)
=======
So I assume only the second corefile is relevant.
The $rsp corruption seems to happen here too.
--
You are receiving this mail because:
You are the assignee for the bug.