I have to debug a huge Windows server application which we think has handle leaks.
What I'm doing:
Run WinDbg as administrator
Attach to process
!htrace -enable
!htrace -snapshot
g
Run test scenario
Ctrl-break
!htrace -diff
At this point my WinDbg show *Busy* status forever and print nothing.
What I'm doing wrong?
Related
I have a Win32 console application (built from Visual Studio as Win32 console project) which does some log file (.txt) processing. I have a separate perl program (legacy program) which now needs to start this Win32 console application and then stop when done.
Perl program starts an instance of Win32 console app using Win32::Process APIs. It can kill the console app when done by either "kill -x pid" or Win32:Process:Kill. The problem is console app needs to know if its being killed/terminated so that it can flush log handling. The console app has already registered a handler via SetConsoleCtrlHandler API but doesn't get called when killed from perl program by say kill -2/3 pid.
What do I change in perl program or in Win32 console app so that it can know when its being terminated?
Thanks!
Signal handling in Windows is a little quirky if you're used to Unix. I have done a lot of investigation into this, and wrote up my findings here (starting at line 261).
Short answer: Windows processes can set $SIG{INT}, $SIG{QUIT}, or $SIG{BREAK}. All other signal handlers are ignored. Signal them from you separate app with the builtin kill:
kill 'INT', $the_win32_logger_pid;
kill 'QUIT', $the_win32_logger_pid;
kill 'BREAK', $the_win32_logger_pid;
I have to run cmd / c from a program, run the start command xx.exe, and I capture the result (there xx.exe?). until everything is right, however, remains open the console with the error popup. how can I close the console with the error?
Usually win32 applications will close the command prompt after execution. If this isn't the case with what you're trying to run, you could:
Run it from Windows "Run" option (Windows button+R) than your program name and path in prompt.
Run it from a batch file, like so:
runMe.bat:
START "" "C:\windows\notepad.exe"
EXIT`
Than just run runMe.bat from wherever. Notice the 'exit' command that closes the command prompt after execution.
Read more about batch files, the start command, and this issue here, and there.
Good luck!
I have windbg registered as postmortem debugge (via -I command-line switch) and all works fine for non-elevated applications. But if elevated applications crashes, windbg starts and displays "Could not attach to process. Access is denied" error. Is it possible to somehow configure windbg so it will work as postmortem debugger for elevated apps?
Since you mentioned that it's an x86 app on an x64 system, you need to set these regkeys in Wow6432Node:
Windows Registry Editor Version 5.00
[HKEY_LOCAL_MACHINE\SOFTWARE\Wow6432Node\Microsoft\Windows NT\CurrentVersion\AeDebug]
"UserDebuggerHotKey"=dword:00000000
"Debugger"="\"C:\\debuggers\\windbg.exe\" -p %ld -e %ld -g"
"Auto"="1"
Note that these are both string values (REG_SZ) and I'm assuming that the windbg.exe is installed in C:\debuggers. Change that as appropriate.
I have a running Perl process that’s stuck, I’d like to poke inside with a debugger to see what’s wrong. I can’t restart the process. Can I attach the debugger to the running process? I know I can do gdb -p, but gdb does not help me. I’ve tried Enbugger, but failed:
$ perl -e 'while (1) {}'&
[1] 86836
$ gdb -p 86836
…
Attaching to process 86836.
Reading symbols for shared libraries . done
Reading symbols for shared libraries ............................. done
Reading symbols for shared libraries + done
0x000000010c1694c6 in Perl_pp_stub ()
(gdb) call (void*)Perl_eval_pv("require Enbugger;Enbugger->stop;",0)
perl(86836) malloc: *** error for object 0x3: pointer being realloc'd was not allocated
*** set a breakpoint in malloc_error_break to debug
Program received signal SIGABRT, Aborted.
0x00007fff8269d82a in __kill ()
The program being debugged was signaled while in a function called from GDB.
GDB remains in the frame where the signal was received.
To change this behavior use "set unwindonsignal on"
Evaluation of the expression containing the function (Perl_eval_pv) will be abandoned.
(gdb)
Am I doing it wrong? Are there other options?
P.S. If you think you could benefit from a debugger attached to a running process yourself, you can insert a debugger back door triggered by SIGUSR1:
use Enbugger::OnError 'USR1';
Then you can simply kill -USR1 pid and your process will jump into the debugger.
First, please use a DEBUGGING perl, if you want to inspect it with gdb.
Please define "stuck". Busy or non-busy waiting (high or low CPU), eating memory or not?
With while 1 it is busy waiting. I usually get busy waiting (endless cycles) on HV corruption in Perl_hfree_next_entry() since 5.15. Non-busy waiting is usually waiting on a blocking IO read.
I get the correct:
`0x00007fba15ab35c1 in Perl_runops_debug () at dump.c:2266`
`2266 } while ((PL_op = PL_op->op_ppaddr(aTHX)));`
and can inspect everything, much more than with a simple perl debugger. With a non-threaded perl you have to type less.
`(gdb) p Perl_op_dump(PL_op)`
and so on.
If you have to do with perl: Inside the pp_stub function it is not a good idea to enter the Enbugger runloop, you should be in the main runloop in dump.c. Set a breakpoint to the shown line.
"error for object 0x3" on eval sound like internal corruption in the context, so you should look at the cx and stack pointers. Probably because you started it in a bad context.
I've never used gdb, but maybe you could get something useful out of strace?
strace -f -s512 -p <PID>
http://metacpan.org/pod/App::Stacktrace
“perl-stacktrace prints Perl stack traces of Perl threads for a given Perl process. For each Perl frame, the full file name and line number are printed.”
I'm using Eclipse CDT with Zylin embedded debug (native) and arm-elf-gdb 6.6 on OS X 10.5.8.
I've got an OCD debugger connecting to a board with an ARM processor.
Eclipse is able to stop at the first breakpoint and gather variables from memory, but any subsequent breakpoints or single-steps fail. I can set breakpoints ok when using commandline gdb, but setting breakpoints at the exact same location gives me these errors in Eclipse:
Warning:
Cannot insert breakpoint 2.
Error accessing memory address 0x3f6: Unknown error: -1.
And my OCD connection dies with "Error: unexpected error -308"
I'm using the same OCD script in both cases.
The only differences I see are (1) Eclipse uses Zylin and the mi protocol to talk to gdb, and (2) my gdb script is invoked differently in this way:
When running from gdb, I use a script file that ends with these two lines:
load filename.out
symbol-file filename.out
In Eclipse, I copy the contents of this script file into "Debug Configurations | Commands" (I've tried both the "'Initialize' Commands" and "'Run' Commands" boxes. I replace the last two lines with load, and specify filename.out in the "Debug Configurations | Main | C/C++ Application" box.
I'd appreciate any suggestions about how to get Zylin+Eclipse to behave.
Thanks.
I'm using the same solution on a LPC2103.
Go to Eclipse, in Zylin Debug Configurations > Embedded Debug (Native):
Main tab: C/C++ Application = you_elf_file.elf
Debugger tab: GDB debugger = arm-elf-gdb; GDB command set = Standard; Protocol = mi2
Commands:
'Initialize' commands:
target remote localhost:3333
monitor arm7_9 dcc_downloads enable
monitor arm7_9 fast_memory_access enable
monitor arm7_9 dbgrq enable
monitor reset halt
monitor wait_halt
monitor sleep 500
load
'Run' commands:
continue
I'm using parallel port Wiggler and OpenOCD [Open On-Chip Debugger 0.4.0] with the cfgs that came with the installation:
openocd -f interface/parport.cfg -f target/lpc2103.cfg
It works most of the time for me.