test

Sun Studio 12: Debugging a Program With dbx

Chapter 6 Setting Breakpoints and Traces

When an event occurs, dbx allows you to stop a process, execute arbitrary commands, or print information. The simplest example of an event is a breakpoint. Examples of other events are faults, signals, system calls, calls to dlopen(), and data changes.

A trace displays information about an event in your program, such as a change in the value of a variable. Although a trace’s behavior is different from that of a breakpoint, traces and breakpoints share similar event handlers (see Event Handlers).

This chapter describes how to set, clear, and list breakpoints and traces. For complete information on the event specifications you can use in setting breakpoints and traces, see Setting Event Specifications.

The chapter is organized into the following sections:

Setting Breakpoints

In dbx, you can use three commands to set breakpoints:

The stop, when, and trace commands all take as an argument an event specification, which describes the event on which the breakpoint is based. Event specifications are discussed in detail in Setting Event Specifications.

To set machine-level breakpoints, use the stopi, wheni, and tracei commands (see Chapter 18, Debugging at the Machine-Instruction Level).

Note –

When debugging an application that is a mixture of JavaTM code and C JNI (Java Native Interface) code or C++ JNI code, you may want to set breakpoints in code that has not yet been loaded. For information on setting breakpoints on such code, see Setting Breakpoints on Code That Has Not Yet Been Loaded by the JVM Software .

Setting a stop Breakpoint at a Line of Source Code

You can set a breakpoint at a line number, using the stop at command, where n is a source code line number and filename is an optional program file name qualifier. 


(dbx) stop at filename: n

For example


(dbx) stop at main.cc:3
:

If the line specified is not an executable line of source code, dbx sets the breakpoint at the next executable line. If there is no executable line, dbx issues an error.

You can determine the line at which you wish to stop by using the file command to set the current file and the list command to list the function in which you wish to stop. Then use the stop at command to set the breakpoint on the source line:


(dbx) file t.c
(dbx) list main
10    main(int argc, char *argv[])
11    {
12        char *msg = "hello world\n";
13        printit(msg);
14    }
(dbx) stop at 13

For more information on specifying at an location event, see at [filename:]line_number.

Setting a stop Breakpoint in a Function

You can set a breakpoint in a function, using the stop in command: 


(dbx) stop in function

An In Function breakpoint suspends program execution at the beginning of the first source line in a procedure or function.

dbx should be able to determine which variable or function you are referring to except when:

Consider the following set of declarations:


int foo(double);
int foo(int);
int bar();
class x {
   int bar();
};

When you stop at a non-member function, you can type:


stop in foo(int)

to set a breakpoint at the global foo(int).

To set a breakpoint at the member function you can use the command:


stop in x::bar()

If you type:


stop in foo

dbx cannot determine whether you mean the global function foo(int) or the global function foo(double) and may be forced to display an overloaded menu for clarification.

If you type:


stop in `bar

dbx cannot determine whether you mean the global function bar() or the member function bar() and displays an overload menu.

For more information on specifying an in function event, see in function.

Setting Multiple Breaks in C++ Programs

You can check for problems related to calls to members of different classes, calls to any members of a given class, or calls to overloaded top-level functions. You can use a keyword—inmember, inclass, infunction, or inobject—with a stop, when, or trace command to set multiple breaks in C++ code.

Setting Breakpoints in Member Functions of Different Classes

To set a breakpoint in each of the object-specific variants of a particular member function (same member function name, different classes), use stop inmember.

For example, if the function draw is defined in several different classes, then to place a breakpoint in each function, type:


(dbx) stop inmember draw

For more information on specifying an inmember or inmethod event, see inmember function inmethod function.

Setting Breakpoints in Member Functions of the Same Class

To set a breakpoint in all member functions of a specific class, use the stop inclass command.

By default, breakpoints are inserted only in the class member functions defined in the class, not those that it might inherit from base classes. To insert breakpoints in the functions that inherit from the base classes also, specify the -recurse option.

To set a breakpoint in all member functions defined in the class shape, type:


(dbx) stop inclass shape

To set a breakpoint in all member functions defined in the class shape, and also in functions that inherit from the class, type:


(dbx) stop inclass shape -recurse

For more information on specifying an inclass event, see inclass classname [-recurse | -norecurse] and stop Command.

Due to the large number of breakpoints that may be inserted by stop inclass and other breakpoint selections, you should be sure to set the dbx environment variable step_events to on to speed up the step and next commands (see Efficiency Considerations).

Setting Multiple Breakpoints in Nonmember Functions

To set multiple breakpoints in nonmember functions with overloaded names (same name, different type or number of arguments), use the stop infunction command.

For example, if a C++ program has defined two versions of a function named sort()(one that passes an int type argument and the other a float) then, to place a breakpoint in both functions, type:


(dbx) stop infunction sort [command;]

For more information on specifying an infunction event, see infunction function.

Setting Breakpoints in Objects

Set an In Object breakpoint to check the operations applied to a specific object instance.

By default, an In Object breakpoint suspends program execution in all nonstatic member functions of the object’s class, including inherited ones, when called from the object. To set a breakpoint to suspend program execution in only nonstatic member functions defined in the object’s class and not inherited classes, specify the -norecurse option.

To set a breakpoint in all nonstatic member functions defined in the base class of object foo, and in all nonstatic member functions defined in inherited classes of object foo, type:


(dbx) stop inobject &foo

To set a breakpoint in all nonstatic member functions defined in the class of object foo, but not those defined in inherited classes of object foo, type:


(dbx) stop inobject &foo -norecurse

For more information on specifying an inobject event, see inobject object-expression [-recurse | -norecurse] and stop Command

Setting Data Change Breakpoints

You can use data change breakpoints in dbx to note when the value of a variable or expression has changed.

Stopping Execution When an Address Is Accessed

To stop execution when a memory address has been accessed, type:


(dbx) stop access mode address-expression [, byte-size-expression]

mode specifies how the memory was accessed. It can be composed of one or all of the letters:

r

The memory at the specified address has been read.

w

The memory has been written to.

x

The memory has been executed.

mode can also contain either of the following:

a

Stops the process after the access (default).

b

Stops the process before the access.

In both cases the program counter will point at the offending instruction. The “before” and “after” refer to the side effect.

address-expression is any expression that can be evaluated to produce an address. If you give a symbolic expression, the size of the region to be watched is automatically deduced; you can override it by specifying byte-size-expression. You can also use nonsymbolic, typeless address expressions; in which case, the size is mandatory.

In the following example, execution will stop execution after the memory address 0x4762 has been read


(dbx) stop access r 0x4762
:

In this example, execution will stop before the variable speed has be written to:


(dbx) stop access wb &speed

Keep these points in mind when using the stop access command:

For more information on specifying an access event, see access mode address-expression [, byte-size-expression] and stop Command.

Stopping Execution When Variables Change

To stop program execution if the value of a specified variable has changed, type:


(dbx) stop change variable

Keep these points in mind when using the stop change command:

For more information on specifying a change event, see change variable and stop Command.

dbx implements stop change by causing automatic single stepping together with a check on the value at each step. Stepping skips over library calls if the library was not compiled with the -g option. So, if control flows in the following manner, dbx does not trace the nested user_routine2 because tracing skips the library call and the nested call to user_routine2.


   user_routine calls
      library_routine, which calls
        user_routine2, which changes variable

The change in the value of variable appears to have occurred after the return from the library call, not in the middle of user_routine2.

dbx cannot set a breakpoint for a change in a block local variable—a variable nested in {}. If you try to set a breakpoint or trace in a block local “nested” variable, dbx issues an error informing you that it cannot perform this operation.

Note –

It is faster to watch data changes using the access event than the change event. Instead of automatically single-stepping the program, the access event uses a page protection scheme that is much faster.

Stopping Execution on a Condition

To stop program execution if a conditional statement evaluates to true, type:


(dbx) stop cond condition

The program stops executing when the condition occurs.

Keep these points in mind when using the stop cond command:

For more information on specifying a condition event, see cond condition-expression and stop Command.

Setting Filters on Breakpoints

In dbx, most of the event management commands also support an optional event filter modifier. The simplest filter instructs dbx to test for a condition after the program arrives at a breakpoint or trace handler, or after a watch condition occurs.

If this filter condition evaluates to true (non 0), the event command applies and program execution stops at the breakpoint. If the condition evaluates to false (0), dbx continues program execution as if the event had never happened.

To set a breakpoint that includes a filter at a line or in a function, add an optional- if condition modifier statement to the end of a stop or trace command.

The condition can be any valid expression, including function calls, returning Boolean or integer in the language current at the time the command is entered.

With a location-based breakpoint like in or at, the scope is that of the breakpoint location. Otherwise, the scope of the condition is the scope at the time of entry, not at the time of the event. You might have to use the backquote operator (see Backquote Operator) to specify the scope precisely.

For example, these two filters are not the same:


stop in foo -if a>5
stop cond a>5

The former breaks at foo and tests the condition. The latter automatically single steps and tests for the condition.

Using the Return Value of a Function Call as a Filter

You can use a function call as a breakpoint filter. In this example, if the value in the string str is abcde, then execution stops in function foo():


(dbx) stop in foo -if !strcmp(“abcde”,str)

Using Variable Scope as a Filter

Variable scope can be used in setting a breakpoint filter. In this example, the current scope is in function foo() and local is a local variable defined in main():


(dbx) stop access w &main`local -if pr(main`local) -in main

Using a Filter With a Conditional Event

New users sometimes confuse setting a conditional event command (a watch-type command) with using filters. Conceptually, “watching” creates a precondition that must be checked before each line of code executes (within the scope of the watch). But even a breakpoint command with a conditional trigger can also have a filter attached to it.

Consider this example:


(dbx) stop access w &speed -if speed==fast_enough

This command instructs dbx to monitor the variable, speed; if the variable speed is written to (the “watch” part), then the -if filter goes into effect. dbx checks whether the new value of speed is equal to fast_enough. If it is not, the program continues, “ignoring” the stop command.

In dbx syntax, the filter is represented in the form of an [-if condition] statement at the end of the command.


stop in function [-if condition]

If you set a breakpoint with a filter that contains function calls in a multithreaded program, dbx stops execution of all threads when it hits the breakpoint, and then evaluates the condition. If the condition is met and the function is called, dbx resumes all threads for the duration of the call.

For example, you might set the following breakpoint in a multithreaded application where many threads call lookup():


(dbx) stop in lookup -if strcmp(name, “troublesome”) == 0

dbx stops when thread t@1 calls lookup(), evaluates the condition, and calls strcmp() resuming all threads. If dbx hits the breakpoint in another thread during the function call, it issues a warning such as one of the following:


event infinite loop causes missed events in the following handlers:
...
 

Event reentrancy
first event BPT(VID 6m TID 6, PC echo+0x8)
second event BPT*VID 10, TID 10, PC echo+0x8)
the following handlers will miss events:
...

In such a case, if you can ascertain that the function called in the conditional expression will not grab a mutex, you can use the -resumeone event specification modifier to force dbx to resume only the first thread in which it hit the breakpoint. For example, you might set the following breakpoint:


(dbx) stop in lookup -resumeone -if strcmp(name, “troublesome”) == 0

The -resumeone modifier does not prevent problems in all cases. For example, it would not help if:

For detailed information on event modifiers, see Event Specification Modifiers.

Tracing Execution

Tracing collects information about what is happening in your program and displays it. If a program arrives at a breakpoint created with a trace command, the program halts and an event-specific trace information line is emitted, then the program continues.

A trace displays each line of source code as it is about to be executed. In all but the simplest programs, this trace produces volumes of output.

A more useful trace applies a filter to display information about events in your program. For example, you can trace each call to a function, every member function of a given name, every function in a class, or each exit from a function. You can also trace changes to a variable.

Setting a Trace

Set a trace by typing a trace command at the command line. The basic syntax of the trace command is:


trace event-specification [ modifier ]

For the complete syntax of the trace command, see trace Command.

The information a trace provides depends on the type of event associated with it (see Setting Event Specifications).

Controlling the Speed of a Trace

Often trace output goes by too quickly. The dbx environment variable trace_speed lets you control the delay after each trace is printed. The default delay is 0.5 seconds.

To set the interval in seconds between execution of each line of code during a trace, type:


dbxenv trace_speed number

Directing Trace Output to a File

You can direct the output of a trace to a file using the -file filename option. For example, the following command direct trace output to the file trace1:


(dbx) trace -file trace1

To revert trace output to standard output use - for filename. Trace output is always appended to filename. It is flushed whenever dbx prompts and when the application has exited. The filename is always re-opened on a new run or resumption after an attach.

Setting a when Breakpoint at a Line

A when breakpoint command accepts other dbx commands such as list, letting you write your own version of trace.


(dbx) when at 123 {list $lineno;}

The when command operates with an implied cont command. In the example above, after listing the source code at the current line, the program continues executing. If you included a stop command after the list command, the program would not continue executing.

For the complete syntax of the when command, see when Command. For detailed information on event modifiers, see Event Specification Modifiers.

Setting a Breakpoint in a Shared Library

dbx provides full debugging support for code that uses the programmatic interface to the run-time linker: code that calls dlopen(), dlclose() and their associated functions. The run-time linker binds and unbinds shared libraries during program execution. Debugging support for dlopen() and dlclose() lets you step into a function or set a breakpoint in functions in a dynamically shared library just as you can in a library linked when the program is started.

However, there are exceptions. dbx is unable to place breakpoints in loadobjects that have not been loaded (by, for example, using dlopen():

You can put the names of such loadobjects on the preload list with the loadobject command (see loadobject Command).

dbx does not forget about a loadobject that was loaded using dlopen(). For example, a breakpoint set in a freshly loaded loadobject remains until the next run command, or even if the loadobject is unloaded with dlclose() and then subsequently loaded with dlopen() again.

Listing and Clearing Breakpoints

Often, you set more than one breakpoint or trace handler during a debugging session. dbx supports commands for listing and clearing them.

Listing Breakpoints and Traces

To display a list of all active breakpoints, use the status command to display ID numbers in parentheses, which can then be used by other commands.

dbx reports multiple breakpoints set with the inmember, inclass, and infunction keywords as a single set of breakpoints with one status ID number.

Deleting Specific Breakpoints Using Handler ID Numbers

When you list breakpoints using the status command, dbx displays the ID number assigned to each breakpoint when it was created. Using the delete command, you can remove breakpoints by ID number, or use the keyword all to remove all breakpoints currently set anywhere in the program.

To delete breakpoints by ID number (in this case 3 and 5), type:


(dbx) delete 3 5

To delete all breakpoints set in the program currently loaded in dbx, type:


(dbx) delete all

For more information, see delete Command.

Enabling and Disabling Breakpoints

Each event management command (stop, trace, when) that you use to set a breakpoint creates an event handler (see Event Handlers). Each of these commands returns a number known as the handler ID (hid). You can use the handler ID as an argument to the handler command (see handler Command) to enable or disable the breakpoint.

Efficiency Considerations

Various events have different degrees of overhead in respect to the execution time of the program being debugged. Some events, like the simplest breakpoints, have practically no overhead. Events based on a single breakpoint have minimal overhead.

Multiple breakpoints such as inclass, that might result in hundreds of breakpoints, have an overhead only during creation time. dbx uses permanent breakpoints; the breakpoints are retained in the process at all times and are not taken out on every stoppage and put in on every cont command.

Note –

In the case of the step command and next command, by default all breakpoints are taken out before the process is resumed and reinserted once the step completes. If you are using many breakpoints or multiple breakpoints on prolific classes, the speed of the step command and next command slows down considerably. Use the dbx step_events environment variable to control whether breakpoints are taken out and reinserted after each step command or next command.

The slowest events are those that utilize automatic single stepping. This might be explicit and obvious as in the trace step command, which single steps through every source line. Other events, like the stop change expression or trace cond variable not only single step automatically but also have to evaluate an expression or a variable at each step.

These events are very slow, but you can often overcome the slowness by bounding the event with a function using the -in modifier. For example:


trace next -in mumble
stop change clobbered_variable -in lookup

Do not use trace -in main because the trace is effective in the functions called by main as well. Do use it in the cases where you suspect that the lookup() function is clobbering your variable.