Strotmann.de

Thursday Mar 10, 2011

Debugging a shortlived MacOS application

Yesterday I had to debug a MacOS X commandline program that segfaulted immediatly after starting. This program is 'rdnssd' (Recursive DNS Servers discovery Daemon, http://rdnssd.linkfanel.net/). 'rdnssd' implements the client part of RFC 5006 - IPv6 Router Advertisement Option for DNS Configuration. This function lets an IPv6 router send out DNS server IP address information as part of the Router Advertisment messages, helping client finding a DNS server without the need of DHCP or local configuration.

The 'rdnssd' source compiled without issues, but when started, it stopped immediatly with a segfault:

# sudo /usr/local/sbin/rdnssd -f
[1]    1243 segmentation fault  sudo /usr/local/sbin/rdnssd -f
On Linux I'm used to strace (and Solaris has truss) to trace the sysemcalls of a process, hoping that would point me to the error.

MacOS X 10.5/10.6 has several very useful tools build on top of DTrace. Apple Technical Note 2124 lists a good part of these tools.

A run of the process using dtrace only showed that the process died after forking, with no additional information. errinfo also didn't show any suspicious syscalls failing. But the use of opensnoop (traces the files opened by a process) showed that acutally MacOS X is writing a crash dump for the failing process with additional information:

# opensnoop | grep rdnssd       
    0   1446 rdnssd         3 /dev/urandom         
    0   1446 rdnssd         3 /dev/dtracehelper    
    0   1446 rdnssd         3 /dev/urandom         
    0   1446 rdnssd         3 /dev/urandom         
    0   1446 rdnssd         3 /usr/share/locale/en_US.UTF-8/LC_CTYPE 
    0   1448 rdnssd        -1 /etc/sysinfo.conf    
    0   1447 taskgated      4 /usr/local/sbin/rdnssd 
    0   1446 rdnssd         3 /usr/local/var/run/rdnssd.pid 
    0   1449 ReportCrash    5 /usr/local/sbin/rdnssd 
    0   1449 ReportCrash    5 /Developer/source/ipv6/ndisc6-1.0.1/rdnssd/rdnssd.o 
    0   1449 ReportCrash    5 /Developer/source/ipv6/ndisc6-1.0.1/rdnssd/icmp.o 
    0   1449 ReportCrash    5 /Developer/source/ipv6/ndisc6-1.0.1/rdnssd/rdnssd.o 
    0   1449 ReportCrash    5 /Developer/source/ipv6/ndisc6-1.0.1/rdnssd 
    0   1449 ReportCrash    5 /Developer/source/ipv6/ndisc6-1.0.1/rdnssd/icmp.o 
    0   1449 ReportCrash   -1 /Developer/source/ipv6/ndisc6-1.0.1/rdnssd/../compat/libcompat.a(ppoll.o) 
    0   1449 ReportCrash    5 /Library/Logs/DiagnosticReports/rdnssd_2011-03-10-190618_localhost.crash 
    0   1449 ReportCrash   -1 /Library/Logs/DiagnosticReports/rdnssd_2011-03-10-190618_localhost.crash 
    0   1449 ReportCrash    6 /Library/Logs/DiagnosticReports/rdnssd_2011-03-10-190618-1_localhost.crash 
    0   1449 ReportCrash    5 /Library/Application Support/CrashReporter/rdnssd_1472A851-7F63-5DAA-BB39-7AD15590893E.plist 
    0   1449 ReportCrash    5 /Library/Application Support/CrashReporter/rdnssd_1472A851-7F63-5DAA-BB39-7AD15590893E.plist 
    0   1449 ReportCrash    5 /Library/Application Support/CrashReporter/rdnssd_1472A851-7F63-5DAA-BB39-7AD15590893E.plist 
    0   1449 ReportCrash    5 /Library/Application Support/CrashReporter/rdnssd_1472A851-7F63-5DAA-BB39-7AD15590893E.plist 
   89   1450 mdworker       6 /usr/local/var/run/rdnssd.pid 
   89   1450 mdworker       6 /usr/local/var/run/rdnssd.pid 
    0     34 mds           57 /Library/Logs/DiagnosticReports/rdnssd_2011-03-10-190618-1_localhost.crash 
   89     34 mds           -1 /Library/Logs/DiagnosticReports/rdnssd_2011-03-10-190618-1_localhost.crash 
    0     34 mds           57 /Library/Logs/DiagnosticReports/rdnssd_2011-03-10-190618_localhost.crash 
   89     34 mds           -1 /Library/Logs/DiagnosticReports/rdnssd_2011-03-10-190618_localhost.crash 
   89   1450 mdworker      -1 /Library/Application Support/CrashReporter/rdnssd_1472A851-7F63-5DAA-BB39-7AD15590893E.plist 
    0     34 mds            7 /Library/Application Support/CrashReporter/rdnssd_1472A851-7F63-5DAA-BB39-7AD15590893E.plist 
   89     34 mds           -1 /Library/Application Support/CrashReporter/rdnssd_1472A851-7F63-5DAA-BB39-7AD15590893E.plist 

A look into the crash report file revealed the information I was looking for:

 more /Library/Logs/DiagnosticReports/rdnssd_2011-03-10-190618_localhost.crash 
Process:         rdnssd [1446]
Path:            /usr/local/sbin/rdnssd
Identifier:      rdnssd
Version:         ??? (???)
Code Type:       X86-64 (Native)
Parent Process:  zsh [1161]

Date/Time:       2011-03-10 19:06:17.998 +0100
OS Version:      Mac OS X 10.6.6 (10J567)
Report Version:  6

Exception Type:  EXC_BAD_ACCESS (SIGSEGV)
Exception Codes: KERN_INVALID_ADDRESS at 0x0000000000000008
Crashed Thread:  0  Dispatch queue: com.apple.main-thread

Thread 0 Crashed:  Dispatch queue: com.apple.main-thread
0   rdnssd                              0x00000001000023d2 ppoll + 66
1   rdnssd                              0x0000000100001fc2 main + 2418 (rdnssd.c:377)
2   rdnssd                              0x00000001000010a4 start + 52

Thread 0 crashed with X86 Thread State (64-bit):
  rax: 0x0000000000000000  rbx: 0x00000000ffffffff  rcx: 0x00007fff80b1ea02  rdx: 0x0000000000000000
  rdi: 0x0000000000000003  rsi: 0x00007fff5fbffa4c  rbp: 0x00007fff5fbff910  rsp: 0x00007fff5fbff8d0
   r8: 0x00007fff70125f0c   r9: 0x0000000000000000  r10: 0x00007fff5fbffa4c  r11: 0x0000000000000202
  r12: 0x0000000000000000  r13: 0x0000000000000001  r14: 0x00007fff5fbffa30  r15: 0x00007fff5fbff8dc
  rip: 0x00000001000023d2  rfl: 0x0000000000010246  cr2: 0x0000000000000008

Binary Images:
       0x100000000 -        0x100002ff7 +rdnssd ??? (???)  /usr/local/sbin/rdnssd
    0x7fff5fc00000 -     0x7fff5fc3bdef  dyld 132.1 (???)  /usr/lib/dyld
    0x7fff80abc000 -     0x7fff80c7dfff  libSystem.B.dylib 125.2.1 (compatibility 1.0.0) <71E6D4C9-F945-6EC2-998C-D61AD590DAB6> /usr/lib/libSystem.B.dylib
    0x7fff82f2c000 -     0x7fff82f30ff7  libmathCommon.A.dylib 315.0.0 (compatibility 1.0.0) <95718673-FEEE-B6ED-B127-BCDBDB60D4E5> /usr/lib/system/libmathCommon.A.dylib
    0x7fffffe00000 -     0x7fffffe01fff  libSystem.B.dylib ??? (???) <71E6D4C9-F945-6EC2-998C-D61AD590DAB6> /usr/lib/libSystem.B.dylib

The culprit is in the sourcefile rdnssd.c, line 377, in the function call 'ppoll'. The reason was a bad memory access (EXC_BAD_ACCESS). This should be all information needed to find and fix the issue.

Stay tuned...

Comments:

Post a Comment:
  • HTML Syntax: Allowed

Calendar

Feeds

Search

Links

Navigation

Info