\documentclass{article} \usepackage[fancyhdr,pdf]{latex2man} \input{common.tex} \begin{document} \begin{Name}{3}{libunwind}{David Mosberger-Tang}{Programming Library}{Introduction to libunwind} libunwind -- a (mostly) platform-independent unwind API \end{Name} \section{Synopsis} \File{\#include $<$libunwind.h$>$}\\ \noindent \Type{int} \Func{unw\_getcontext}(\Type{unw\_context\_t~*});\\ \noindent \Type{int} \Func{unw\_init\_local}(\Type{unw\_cursor\_t~*}, \Type{unw\_context\_t~*});\\ \noindent \Type{int} \Func{unw\_init\_remote}(\Type{unw\_cursor\_t~*}, \Type{unw\_addr\_space\_t}, \Type{void~*});\\ \noindent \Type{int} \Func{unw\_step}(\Type{unw\_cursor\_t~*});\\ \noindent \Type{int} \Func{unw\_get\_reg}(\Type{unw\_cursor\_t~*}, \Type{int}, \Type{unw\_word\_t~*});\\ \noindent \Type{int} \Func{unw\_get\_fpreg}(\Type{unw\_cursor\_t~*}, \Type{int}, \Type{unw\_fpreg\_t~*});\\ \noindent \Type{int} \Func{unw\_set\_reg}(\Type{unw\_cursor\_t~*}, \Type{int}, \Type{unw\_word\_t});\\ \noindent \Type{int} \Func{unw\_get\_fpreg}(\Type{unw\_cursor\_t~*}, \Type{int}, \Type{unw\_fpreg\_t});\\ \noindent \Type{int} \Func{unw\_resume}(\Type{unw\_cursor\_t~*});\\ \noindent \Type{unw\_addr\_space\_t} \Var{local\_addr\_space};\\ \noindent \Type{unw\_addr\_space\_t} \Func{unw\_create\_addr\_space}(\Type{unw\_accessors\_t}, \Type{int});\\ \noindent \Type{void} \Func{unw\_destroy\_addr\_space}(\Type{unw\_addr\_space\_t});\\ \noindent \Type{unw\_accessors\_t} \Func{unw\_get\_accessors}(\Type{unw\_addr\_space\_t});\\ \noindent \Type{void} \Func{unw\_flush\_cache}(\Type{unw\_addr\_space\_t}, \Type{unw\_word\_t}, \Type{unw\_word\_t});\\ \noindent \Type{int} \Func{unw\_set\_caching\_policy}(\Type{unw\_addr\_space\_t}, \Type{unw\_caching\_policy\_t});\\ \noindent \Type{const char *}\Func{unw\_regname}(\Type{int});\\ \noindent \Type{int} \Func{unw\_get\_proc\_info}(\Type{unw\_cursor\_t~*}, \Type{unw\_proc\_info\_t~*});\\ \noindent \Type{int} \Func{unw\_get\_save\_loc}(\Type{unw\_cursor\_t~*}, \Type{int}, \Type{unw\_save\_loc\_t~*});\\ \noindent \Type{int} \Func{unw\_is\_signal\_frame}(\Type{unw\_cursor\_t~*});\\ \noindent \Type{int} \Func{unw\_get\_proc\_name}(\Type{unw\_cursor\_t~*}, \Type{char~*}, \Type{size\_t});\\ \noindent \Type{void} \Func{\_U\_dyn\_register}(\Type{unw\_dyn\_info\_t~*});\\ \noindent \Type{void} \Func{\_U\_dyn\_cancel}(\Type{unw\_dyn\_info\_t~*});\\ \section{Local Unwinding} \Prog{Libunwind} is very easy to use when unwinding a stack from within a running program. This is called \emph{local} unwinding. Say you want to unwind the stack while executing in some function \Func{F}(). In this function, you would call \Func{unw\_getcontext}() to get a snapshot of the CPU registers (machine-state). Then you initialize an \emph{unwind~cursor} based on this snapshot. This is done with a call to \Func{unw\_init\_local}(). The cursor now points to the current frame, that is, the stack frame that corresponds to the current activation of function \Func{F}(). The unwind cursor can then be moved ``up'' (towards earlier stack frames) by calling \Func{unw\_step}(). By repeatedly calling this routine, you can uncover the entire call-chain that led to the activation of function \Func{F}(). A positive return value from \Func{unw\_step}() indicates that there are more frames in the chain, zero indicates that the end of the chain has been reached, and any negative value indicates that some sort of error has occurred. While it is not possible to directly move the unwind cursor in the ``down'' direction (towards newer stack frames), this effect can be achieved by making copyies of an unwind cursor. For example, a program that sometimes has to move ``down'' by one stack frame could maintain two cursor variables: ``\Var{curr}'' and ``\Var{prev}''. The former would be used as the current cursor and \Var{prev} would be maintained as the ``previous frame'' cursor by copying the contents of \Var{curr} to \Var{prev} right before calling \Func{unw\_step}(). With this approach, the program could move one step ``down'' simply by copying back \Var{prev} to \Var{curr} whenever that is necessary. In the mosts extreme case, a program could maintain a separate cursor for each call frame and that way it could move up and down the call frame chain at will. Given an unwind cursor, it is possible to read and write the CPU registers that were preserved for the current stack frame identified by the cursor. \Prog{Libunwind} provides several routines for this purpose: \Func{unw\_get\_reg}() reads an integer (general) register, \Func{unw\_get\_fpreg}() reads a floating-point register, \Func{unw\_set\_reg}() writes an integer register, and \Func{unw\_set\_fpreg}() writes a floating-point register. Note that, by definition, only the \emph{preserved} machine state can be accessed during an unwind operation. Normally, this state consists of the \emph{callee-saved} (``preserved'') registers. However, in some special circumstances (e.g., in a signal handler trampoline), even the \emph{caller-saved} (``scratch'') registers are preserved in the stack frame and, in those cases, \Prog{libunwind} will grant access to them as well. The exact set of registers that can be accessed via the cursor depends, of course, on the platform. However, there are two registers that can be read on all platforms: the instruction pointer (IP), sometimes also known as the ``program counter'', and the stack pointer (SP). In \Prog{libunwind}, these registers are identified by the macros \Const{UNW\_REG\_IP} and \Const{UNW\_REG\_SP}, respectively. Besides just moving the unwind cursor and reading/writing saved registers, \Prog{libunwind} also provides the ability to resume execution at an arbitrary stack frame. As you might guess, this is useful for implementing non-local gotos and the exception handling needed by some high-level languages such as Java. Resuming execution with a particular stack frame simply requires calling \Func{unw\_resume}() and passing the cursor identifying the target frame as the only argument. Normally, \Prog{libunwind} supports both local and remote unwinding (the latter will be explained in the next section). However, if you tell libunwind that your program only needs local unwinding, then a special implementation can be selected which may run much faster than the generic implementation which supports both kinds of unwinding. To select this optimized version, simply define the macro \Const{UNW\_LOCAL\_ONLY} before including the headerfile \File{$<$libunwind.h$>$}. If is perfectly OK for a single program to employ both local-only and generic unwinding. That is, whether or not \Const{UNW\_LOCAL\_ONLY} is defined is a choice that each source-file (compilation-unit) can make on its own. Independent of the setting(s) of \Const{UNW\_LOCAL\_ONLY}, you'll always link the same library into the program (normally \Opt{-l}\File{unwind}). If we put all of the above together, here is how we could use \Prog{libunwind} write function \Func{show\_backtrace}() which prints a classic stack trace: \begin{verbatim} #define UNW_LOCAL_ONLY #include void show_backtrace (void) { unw_cursor_t cursor; unw_context_t uc; unw_word_t ip, sp; unw_getcontext(&uc); unw_init_local(&cursor, &uc); while (unw_step(&cursor) > 0) { unw_get_reg(&cursor, UNW_REG_IP, &ip); unw_get_reg(&cursor, UNW_REG_SP, &sp); printf ("ip = %lx, sp = %lx\n", (long) ip, (long) sp); } } \end{verbatim} \section{Remote Unwinding} \Prog{Libunwind} can also be used to unwind a stack in a ``remote'' process. Here, ``remote'' may mean another process on the same machine or even a process on a completely different machine from the one that is running \Prog{libunwind}. Remote unwinding is typically used by debuggers and instruction-set simulators, for example. Before you can unwind a remote process, you need to create a new address-space object for that process. This is achieved with the \Func{unw\_create\_addr\_space} routine. The routine takes two arguments: a pointer to a set of \emph{accessor} routines and an integer that specifies the byte-order of the target process. The accessor routines provide \Func{libunwind} with the means to communicate with the remote process. In particular, there are callbacks to read and write the process's memory, its registers, and to access unwind information which may be needed by \Func{libunwind}. With the address space created, unwinding can be initiated by a call to \Func{unw\_init\_remote}(). This routine is very similar to \Func{unw\_init\_local}(), except that it takes an address-space object and an opaque pointer as arguments. The routine uses these arguments to fetch the initial machine state. \Prog{Libunwind} never uses the opaque pointer on its own, but instead justs passes it on to the accessor (callback) routines. Typically, this pointer is used to select, e.g., the thread within a process that is to be unwound. Once a cursor has been initialized with \Func{unw\_init\_remote}(), unwinding works exactly like in the local case. That is, you can use \Func{unw\_step} to move ``up'' in the call-chain, read and write registers, or resume execution at a particular stack frame by calling \Func{unw\_resume}. \section{Cross-platform and Multi-platform Unwinding} \Prog{Libunwind} has been designed to enable unwinding across platforms (architectures). Indeed, a single program can use \Prog{libunwind} to unwind an arbitrary number of target platforms, all at the same time! We call the machine that is running \Prog{libunwind} the \emph{host} and the machine that is running the process being unwound the \emph{target}. If the host and the target platform are the same, we call it \emph{native} unwinding. If they differ, we call it \emph{cross-platform} unwinding. The principle behind supporting native, cross-platform, and multi-platform unwinding are very simple: for native unwinding, a program includes \File{$<$libunwind.h$>$} and uses the linker switch \Opt{-l}\File{unwind}. For cross-platform unwinding, a program includes \File{$<$libunwind-}\Var{PLAT}\File{.h$>$} and uses the linker switch \Opt{-l}\File{unwind-}\Var{PLAT}, where \Var{PLAT} is the name of the target platform (e.g., \File{ia64} for IA-64, \File{hppa-elf} for ELF-based HP PA-RISC, or \File{x86} for 80386). Multi-platform unwinding works exactly like cross-platform unwinding, the only limitation is that a single source file (compilation unit) can include at most one \Prog{libunwind} header file. In other words, the platform-specific support for each supported target needs to be isolated in separate source files---a limitation that shouldn't be an issue in practice. Note that, by definition, local unwinding is possible only for the native case. Attempting to call, e.g., \Func{unw\_local\_init}() when targeting a cross-platform will result in a link-time error (unresolved references). \section{Thread- and Signal-Safety} All \Prog{libunwind} routines are thread-safe. What this means is that multiple threads may use \Prog{libunwind} simulatenously. However, any given cursor may be accessed by only one thread at any given time. To ensure thread-safety, some \Prog{libunwind} routines may have to use locking. Such routines \emph{must~not} be called from signal handlers (directly or indirectly) and are therefore \emph{not} signal-safe. The manual page for each \Prog{libunwind} routine identifies whether or not it is signal-safe, but as a general rule, any routine that may be needed for \emph{local} unwinding is signal-safe (e.g., \Func{unw\_step}() for local unwinding is signal-safe). For remote-unwinding, \emph{none} of the \Prog{libunwind} routines are guaranteed to be signal-safe. \section{Unwinding Through Dynamically Generated Code} \Func{Libunwind} provides the routines \Func{\_U\_dyn\_register}() and \Func{\_U\_dyn\_cancel} to register/cancel the information required to unwind through code that has been generated at runtime (e.g., by a just-in-time (JIT) compiler). It is important to register the information for \emph{all} dynamically generated code because otherwise, a debugger may not be able to function properly or high-level language exception handling may not work as expected. The interface for registering and canceling dynamic unwind info has been designed for maximum efficiency, so as to minimize the performance impact on JIT-compilers. In particular, both routines are guaranteed to execute in ``constant time'' (O(1)) and the data-structure encapsulating the dynamic unwind info has been designed to facilitate sharing, such that similar procedures can share much of the underlying information. \section{Caching of Unwind Info} To speed up execution, \Prog{libunwind} may aggressively cache the information it needs to perform unwinding. If a process changes during its lifetime, this creates a risk of \Prog{libunwind} using stale data. For example, this would happen if \Prog{libunwind} were to cache information about a shared library which later on gets unloaded (e.g., via \Cmd{dlclose}{3}). To prevent the risk of using stale data, \Prog{libunwind} provides two facilities: first, it is possible to flush the cached information associated with a specific address range in the target process (or the entire address space, if desired). This functionality is provided by \Func{unw\_flush\_cache}(). The second facility is provided by \Func{unw\_set\_caching\_policy}(), which lets a program select the exact caching policy in use for a given address-space object. In particular, by selecting the policy \Const{UNW\_CACHE\_NONE}, it is possible to turn off caching completely, therefore eliminating the risk of stale data alltogether (at the cost of slower execution). By default, caching is enabled for local unwinding only. \section{Files} \begin{Description} \item[\File{libunwind.h}] Headerfile to include for native (same platform) unwinding. \item[\File{libunwind-}\Var{PLAT}\File{.h}] Headerfile to include when unwind target runs on platform \Var{PLAT}. For example, to unwind an IA-64 program, the header file \File{libunwind-ia64.h} should be included. \item[\Opt{-l}\File{unwind}] Linker-switch to add when building a program that does native (same platform) unwinding. \item[\Opt{-l}\File{unwind-}\Var{PLAT}] Linker-switch to add when building a program that unwinds a program on platform \Var{PLAT}. For example, to (cross-)unwind an IA-64 program, the linker switch \File{-lunwind-ia64} should be added. Note: multiple such switches may need to be specified for programs that can unwind programss on multiple platforms. \end{Description} \section{See Also} libunwind-ia64(3), libunwind-hppa(3), libunwind-x86(3), unw\_init\_local(3), unw\_init\_remote(3), etc. \section{Author} \noindent David Mosberger-Tang\\ Hewlett-Packard Labs\\ Palo-Alto, CA 94304\\ Email: \Email{davidm@hpl.hp.com}\\ WWW: \URL{http://www.hpl.hp.com/research/linux/libunwind/}. \LatexManEnd \end{document}