@node Memory, Character Handling, Error Reporting, Top @chapter Virtual Memory Allocation And Paging @c %MENU% Allocating virtual memory and controlling paging @cindex memory allocation @cindex storage allocation This chapter describes how processes manage and use memory in a system that uses @theglibc{}. @Theglibc{} has several functions for dynamically allocating virtual memory in various ways. They vary in generality and in efficiency. The library also provides functions for controlling paging and allocation of real memory. @menu * Memory Concepts:: An introduction to concepts and terminology. * Memory Allocation:: Allocating storage for your program data * Resizing the Data Segment:: @code{brk}, @code{sbrk} * Locking Pages:: Preventing page faults @end menu Memory mapped I/O is not discussed in this chapter. @xref{Memory-mapped I/O}. @node Memory Concepts @section Process Memory Concepts One of the most basic resources a process has available to it is memory. There are a lot of different ways systems organize memory, but in a typical one, each process has one linear virtual address space, with addresses running from zero to some huge maximum. It need not be contiguous; i.e., not all of these addresses actually can be used to store data. The virtual memory is divided into pages (4 kilobytes is typical). Backing each page of virtual memory is a page of real memory (called a @dfn{frame}) or some secondary storage, usually disk space. The disk space might be swap space or just some ordinary disk file. Actually, a page of all zeroes sometimes has nothing at all backing it -- there's just a flag saying it is all zeroes. @cindex page frame @cindex frame, real memory @cindex swap space @cindex page, virtual memory The same frame of real memory or backing store can back multiple virtual pages belonging to multiple processes. This is normally the case, for example, with virtual memory occupied by @glibcadj{} code. The same real memory frame containing the @code{printf} function backs a virtual memory page in each of the existing processes that has a @code{printf} call in its program. In order for a program to access any part of a virtual page, the page must at that moment be backed by (``connected to'') a real frame. But because there is usually a lot more virtual memory than real memory, the pages must move back and forth between real memory and backing store regularly, coming into real memory when a process needs to access them and then retreating to backing store when not needed anymore. This movement is called @dfn{paging}. When a program attempts to access a page which is not at that moment backed by real memory, this is known as a @dfn{page fault}. When a page fault occurs, the kernel suspends the process, places the page into a real page frame (this is called ``paging in'' or ``faulting in''), then resumes the process so that from the process' point of view, the page was in real memory all along. In fact, to the process, all pages always seem to be in real memory. Except for one thing: the elapsed execution time of an instruction that would normally be a few nanoseconds is suddenly much, much, longer (because the kernel normally has to do I/O to complete the page-in). For programs sensitive to that, the functions described in @ref{Locking Pages} can control it. @cindex page fault @cindex paging Within each virtual address space, a process has to keep track of what is at which addresses, and that process is called memory allocation. Allocation usually brings to mind meting out scarce resources, but in the case of virtual memory, that's not a major goal, because there is generally much more of it than anyone needs. Memory allocation within a process is mainly just a matter of making sure that the same byte of memory isn't used to store two different things. Processes allocate memory in two major ways: by exec and programmatically. Actually, forking is a third way, but it's not very interesting. @xref{Creating a Process}. Exec is the operation of creating a virtual address space for a process, loading its basic program into it, and executing the program. It is done by the ``exec'' family of functions (e.g. @code{execl}). The operation takes a program file (an executable), it allocates space to load all the data in the executable, loads it, and transfers control to it. That data is most notably the instructions of the program (the @dfn{text}), but also literals and constants in the program and even some variables: C variables with the static storage class (@pxref{Memory Allocation and C}). @cindex executable @cindex literals @cindex constants Once that program begins to execute, it uses programmatic allocation to gain additional memory. In a C program with @theglibc{}, there are two kinds of programmatic allocation: automatic and dynamic. @xref{Memory Allocation and C}. Memory-mapped I/O is another form of dynamic virtual memory allocation. Mapping memory to a file means declaring that the contents of certain range of a process' addresses shall be identical to the contents of a specified regular file. The system makes the virtual memory initially contain the contents of the file, and if you modify the memory, the system writes the same modification to the file. Note that due to the magic of virtual memory and page faults, there is no reason for the system to do I/O to read the file, or allocate real memory for its contents, until the program accesses the virtual memory. @xref{Memory-mapped I/O}. @cindex memory mapped I/O @cindex memory mapped file @cindex files, accessing Just as it programmatically allocates memory, the program can programmatically deallocate (@dfn{free}) it. You can't free the memory that was allocated by exec. When the program exits or execs, you might say that all its memory gets freed, but since in both cases the address space ceases to exist, the point is really moot. @xref{Program Termination}. @cindex execing a program @cindex freeing memory @cindex exiting a program A process' virtual address space is divided into segments. A segment is a contiguous range of virtual addresses. Three important segments are: @itemize @bullet @item The @dfn{text segment} contains a program's instructions and literals and static constants. It is allocated by exec and stays the same size for the life of the virtual address space. @item The @dfn{data segment} is working storage for the program. It can be preallocated and preloaded by exec and the process can extend or shrink it by calling functions as described in @xref{Resizing the Data Segment}. Its lower end is fixed. @item The @dfn{stack segment} contains a program stack. It grows as the stack grows, but doesn't shrink when the stack shrinks. @end itemize @node Memory Allocation @section Allocating Storage For Program Data This section covers how ordinary programs manage storage for their data, including the famous @code{malloc} function and some fancier facilities special @theglibc{} and GNU Compiler. @menu * Memory Allocation and C:: How to get different kinds of allocation in C. * Unconstrained Allocation:: The @code{malloc} facility allows fully general dynamic allocation. * Allocation Debugging:: Finding memory leaks and not freed memory. * Obstacks:: Obstacks are less general than malloc but more efficient and convenient. * Variable Size Automatic:: Allocation of variable-sized blocks of automatic storage that are freed when the calling function returns. @end menu @node Memory Allocation and C @subsection Memory Allocation in C Programs The C language supports two kinds of memory allocation through the variables in C programs: @itemize @bullet @item @dfn{Static allocation} is what happens when you declare a static or global variable. Each static or global variable defines one block of space, of a fixed size. The space is allocated once, when your program is started (part of the exec operation), and is never freed. @cindex static memory allocation @cindex static storage class @item @dfn{Automatic allocation} happens when you declare an automatic variable, such as a function argument or a local variable. The space for an automatic variable is allocated when the compound statement containing the declaration is entered, and is freed when that compound statement is exited. @cindex automatic memory allocation @cindex automatic storage class In GNU C, the size of the automatic storage can be an expression that varies. In other C implementations, it must be a constant. @end itemize A third important kind of memory allocation, @dfn{dynamic allocation}, is not supported by C variables but is available via @glibcadj{} functions. @cindex dynamic memory allocation @subsubsection Dynamic Memory Allocation @cindex dynamic memory allocation @dfn{Dynamic memory allocation} is a technique in which programs determine as they are running where to store some information. You need dynamic allocation when the amount of memory you need, or how long you continue to need it, depends on factors that are not known before the program runs. For example, you may need a block to store a line read from an input file; since there is no limit to how long a line can be, you must allocate the memory dynamically and make it dynamically larger as you read more of the line. Or, you may need a block for each record or each definition in the input data; since you can't know in advance how many there will be, you must allocate a new block for each record or definition as you read it. When you use dynamic allocation, the allocation of a block of memory is an action that the program requests explicitly. You call a function or macro when you want to allocate space, and specify the size with an argument. If you want to free the space, you do so by calling another function or macro. You can do these things whenever you want, as often as you want. Dynamic allocation is not supported by C variables; there is no storage class ``dynamic'', and there can never be a C variable whose value is stored in dynamically allocated space. The only way to get dynamically allocated memory is via a system call (which is generally via a @glibcadj{} function call), and the only way to refer to dynamically allocated space is through a pointer. Because it is less convenient, and because the actual process of dynamic allocation requires more computation time, programmers generally use dynamic allocation only when neither static nor automatic allocation will serve. For example, if you want to allocate dynamically some space to hold a @code{struct foobar}, you cannot declare a variable of type @code{struct foobar} whose contents are the dynamically allocated space. But you can declare a variable of pointer type @code{struct foobar *} and assign it the address of the space. Then you can use the operators @samp{*} and @samp{->} on this pointer variable to refer to the contents of the space: @smallexample @{ struct foobar *ptr = (struct foobar *) malloc (sizeof (struct foobar)); ptr->name = x; ptr->next = current_foobar; current_foobar = ptr; @} @end smallexample @node Unconstrained Allocation @subsection Unconstrained Allocation @cindex unconstrained memory allocation @cindex @code{malloc} function @cindex heap, dynamic allocation from The most general dynamic allocation facility is @code{malloc}. It allows you to allocate blocks of memory of any size at any time, make them bigger or smaller at any time, and free the blocks individually at any time (or never). @menu * Basic Allocation:: Simple use of @code{malloc}. * Malloc Examples:: Examples of @code{malloc}. @code{xmalloc}. * Freeing after Malloc:: Use @code{free} to free a block you got with @code{malloc}. * Changing Block Size:: Use @code{realloc} to make a block bigger or smaller. * Allocating Cleared Space:: Use @code{calloc} to allocate a block and clear it. * Efficiency and Malloc:: Efficiency considerations in use of these functions. * Aligned Memory Blocks:: Allocating specially aligned memory. * Malloc Tunable Parameters:: Use @code{mallopt} to adjust allocation parameters. * Heap Consistency Checking:: Automatic checking for errors. * Hooks for Malloc:: You can use these hooks for debugging programs that use @code{malloc}. * Statistics of Malloc:: Getting information about how much memory your program is using. * Summary of Malloc:: Summary of @code{malloc} and related functions. @end menu @node Basic Allocation @subsubsection Basic Memory Allocation @cindex allocation of memory with @code{malloc} To allocate a block of memory, call @code{malloc}. The prototype for this function is in @file{stdlib.h}. @pindex stdlib.h @comment malloc.h stdlib.h @comment ISO @deftypefun {void *} malloc (size_t @var{size}) @safety{@prelim{}@mtsafe{}@asunsafe{@asulock{}}@acunsafe{@aculock{} @acsfd{} @acsmem{}}} @c Malloc hooks and __morecore pointers, as well as such parameters as @c max_n_mmaps and max_mmapped_mem, are accessed without guards, so they @c could pose a thread safety issue; in order to not declare malloc @c MT-unsafe, it's modifying the hooks and parameters while multiple @c threads are active that is regarded as unsafe. An arena's next field @c is initialized and never changed again, except for main_arena's, @c that's protected by list_lock; next_free is only modified while @c list_lock is held too. All other data members of an arena, as well @c as the metadata of the memory areas assigned to it, are only modified @c while holding the arena's mutex (fastbin pointers use catomic ops @c because they may be modified by free without taking the arena's @c lock). Some reassurance was needed for fastbins, for it wasn't clear @c how they were initialized. It turns out they are always @c zero-initialized: main_arena's, for being static data, and other @c arena's, for being just-mmapped memory. @c Leaking file descriptors and memory in case of cancellation is @c unavoidable without disabling cancellation, but the lock situation is @c a bit more complicated: we don't have fallback arenas for malloc to @c be safe to call from within signal handlers. Error-checking mutexes @c or trylock could enable us to try and use alternate arenas, even with @c -DPER_THREAD (enabled by default), but supporting interruption @c (cancellation or signal handling) while holding the arena list mutex @c would require more work; maybe blocking signals and disabling async @c cancellation while manipulating the arena lists? @c __libc_malloc @asulock @aculock @acsfd @acsmem @c force_reg ok @c *malloc_hook unguarded @c arena_lock @asulock @aculock @acsfd @acsmem @c mutex_lock @asulock @aculock @c arena_get2 @asulock @aculock @acsfd @acsmem @c get_free_list @asulock @aculock @c mutex_lock (list_lock) dup @asulock @aculock @c mutex_unlock (list_lock) dup @aculock @c mutex_lock (arena lock) dup @asulock @aculock [returns locked] @c __get_nprocs ext ok @acsfd @c NARENAS_FROM_NCORES ok @c catomic_compare_and_exchange_bool_acq ok @c _int_new_arena ok @asulock @aculock @acsmem @c new_heap ok @acsmem @c mmap ok @acsmem @c munmap ok @acsmem @c mprotect ok @c chunk2mem ok @c set_head ok @c tsd_setspecific dup ok @c mutex_init ok @c mutex_lock (just-created mutex) ok, returns locked @c mutex_lock (list_lock) dup @asulock @aculock @c atomic_write_barrier ok @c mutex_unlock (list_lock) @aculock @c catomic_decrement ok @c reused_arena @asulock @aculock @c reads&writes next_to_use and iterates over arena next without guards @c those are harmless as long as we don't drop arenas from the @c NEXT list, and we never do; when a thread terminates, @c arena_thread_freeres prepends the arena to the free_list @c NEXT_FREE list, but NEXT is never modified, so it's safe! @c mutex_trylock (arena lock) @asulock @aculock @c mutex_lock (arena lock) dup @asulock @aculock @c tsd_setspecific dup ok @c _int_malloc @acsfd @acsmem @c checked_request2size ok @c REQUEST_OUT_OF_RANGE ok @c request2size ok @c get_max_fast ok @c fastbin_index ok @c fastbin ok @c catomic_compare_and_exhange_val_acq ok @c malloc_printerr dup @mtsenv @c if we get to it, we're toast already, undefined behavior must have @c been invoked before @c libc_message @mtsenv [no leaks with cancellation disabled] @c FATAL_PREPARE ok @c pthread_setcancelstate disable ok @c libc_secure_getenv @mtsenv @c getenv @mtsenv @c open_not_cancel_2 dup @acsfd @c strchrnul ok @c WRITEV_FOR_FATAL ok @c writev ok @c mmap ok @acsmem @c munmap ok @acsmem @c BEFORE_ABORT @acsfd @c backtrace ok @c write_not_cancel dup ok @c backtrace_symbols_fd @aculock @c open_not_cancel_2 dup @acsfd @c read_not_cancel dup ok @c close_not_cancel_no_status dup @acsfd @c abort ok @c itoa_word ok @c abort ok @c check_remalloced_chunk ok/disabled @c chunk2mem dup ok @c alloc_perturb ok @c in_smallbin_range ok @c smallbin_index ok @c bin_at ok @c last ok @c malloc_consolidate ok @c get_max_fast dup ok @c clear_fastchunks ok @c unsorted_chunks dup ok @c fastbin dup ok @c atomic_exchange_acq ok @c check_inuse_chunk dup ok/disabled @c chunk_at_offset dup ok @c chunksize dup ok @c inuse_bit_at_offset dup ok @c unlink dup ok @c clear_inuse_bit_at_offset dup ok @c in_smallbin_range dup ok @c set_head dup ok @c malloc_init_state ok @c bin_at dup ok @c set_noncontiguous dup ok @c set_max_fast dup ok @c initial_top ok @c unsorted_chunks dup ok @c check_malloc_state ok/disabled @c set_inuse_bit_at_offset ok @c check_malloced_chunk ok/disabled @c largebin_index ok @c have_fastchunks ok @c unsorted_chunks ok @c bin_at ok @c chunksize ok @c chunk_at_offset ok @c set_head ok @c set_foot ok @c mark_bin ok @c idx2bit ok @c first ok @c unlink ok @c malloc_printerr dup ok @c in_smallbin_range dup ok @c idx2block ok @c idx2bit dup ok @c next_bin ok @c sysmalloc @acsfd @acsmem @c MMAP @acsmem @c set_head dup ok @c check_chunk ok/disabled @c chunk2mem dup ok @c chunksize dup ok @c chunk_at_offset dup ok @c heap_for_ptr ok @c grow_heap ok @c mprotect ok @c set_head dup ok @c new_heap @acsmem @c MMAP dup @acsmem @c munmap @acsmem @c top ok @c set_foot dup ok @c contiguous ok @c MORECORE ok @c *__morecore ok unguarded @c __default_morecore @c sbrk ok @c force_reg dup ok @c *__after_morecore_hook unguarded @c set_noncontiguous ok @c malloc_printerr dup ok @c _int_free (have_lock) @acsfd @acsmem [@asulock @aculock] @c chunksize dup ok @c mutex_unlock dup @aculock/!have_lock @c malloc_printerr dup ok @c check_inuse_chunk ok/disabled @c chunk_at_offset dup ok @c mutex_lock dup @asulock @aculock/@have_lock @c chunk2mem dup ok @c free_perturb ok @c set_fastchunks ok @c catomic_and ok @c fastbin_index dup ok @c fastbin dup ok @c catomic_compare_and_exchange_val_rel ok @c chunk_is_mmapped ok @c contiguous dup ok @c prev_inuse ok @c unlink dup ok @c inuse_bit_at_offset dup ok @c clear_inuse_bit_at_offset ok @c unsorted_chunks dup ok @c in_smallbin_range dup ok @c set_head dup ok @c set_foot dup ok @c check_free_chunk ok/disabled @c check_chunk dup ok/disabled @c have_fastchunks dup ok @c malloc_consolidate dup ok @c systrim ok @c MORECORE dup ok @c *__after_morecore_hook dup unguarded @c set_head dup ok @c check_malloc_state ok/disabled @c top dup ok @c heap_for_ptr dup ok @c heap_trim @acsfd @acsmem @c top dup ok @c chunk_at_offset dup ok @c prev_chunk ok @c chunksize dup ok @c prev_inuse dup ok @c delete_heap @acsmem @c munmap dup @acsmem @c unlink dup ok @c set_head dup ok @c shrink_heap @acsfd @c check_may_shrink_heap @acsfd @c open_not_cancel_2 @acsfd @c read_not_cancel ok @c close_not_cancel_no_status @acsfd @c MMAP dup ok @c madvise ok @c munmap_chunk @acsmem @c chunksize dup ok @c chunk_is_mmapped dup ok @c chunk2mem dup ok @c malloc_printerr dup ok @c munmap dup @acsmem @c check_malloc_state ok/disabled @c arena_get_retry @asulock @aculock @acsfd @acsmem @c mutex_unlock dup @aculock @c mutex_lock dup @asulock @aculock @c arena_get2 dup @asulock @aculock @acsfd @acsmem @c mutex_unlock @aculock @c mem2chunk ok @c chunk_is_mmapped ok @c arena_for_chunk ok @c chunk_non_main_arena ok @c heap_for_ptr ok This function returns a pointer to a newly allocated block @var{size} bytes long, or a null pointer if the block could not be allocated. @end deftypefun The contents of the block are undefined; you must initialize it yourself (or use @code{calloc} instead; @pxref{Allocating Cleared Space}). Normally you would cast the value as a pointer to the kind of object that you want to store in the block. Here we show an example of doing so, and of initializing the space with zeros using the library function @code{memset} (@pxref{Copying Strings and Arrays}): @smallexample struct foo *ptr; @dots{} ptr = (struct foo *) malloc (sizeof (struct foo)); if (ptr == 0) abort (); memset (ptr, 0, sizeof (struct foo)); @end smallexample You can store the result of @code{malloc} into any pointer variable without a cast, because @w{ISO C} automatically converts the type @code{void *} to another type of pointer when necessary. But the cast is necessary in contexts other than assignment operators or if you might want your code to run in traditional C. Remember that when allocating space for a string, the argument to @code{malloc} must be one plus the length of the string. This is because a string is terminated with a null character that doesn't count in the ``length'' of the string but does need space. For example: @smallexample char *ptr; @dots{} ptr = (char *) malloc (length + 1); @end smallexample @noindent @xref{Representation of Strings}, for more information about this. @node Malloc Examples @subsubsection Examples of @code{malloc} If no more space is available, @code{malloc} returns a null pointer. You should check the value of @emph{every} call to @code{malloc}. It is useful to write a subroutine that calls @code{malloc} and reports an error if the value is a null pointer, returning only if the value is nonzero. This function is conventionally called @code{xmalloc}. Here it is: @smallexample void * xmalloc (size_t size) @{ void *value = malloc (size); if (value == 0) fatal ("virtual memory exhausted"); return value; @} @end smallexample Here is a real example of using @code{malloc} (by way of @code{xmalloc}). The function @code{savestring} will copy a sequence of characters into a newly allocated null-terminated string: @smallexample @group char * savestring (const char *ptr, size_t len) @{ char *value = (char *) xmalloc (len + 1); value[len] = '\0'; return (char *) memcpy (value, ptr, len); @} @end group @end smallexample The block that @code{malloc} gives you is guaranteed to be aligned so that it can hold any type of data. On @gnusystems{}, the address is always a multiple of eight on 32-bit systems, and a multiple of 16 on 64-bit systems. Only rarely is any higher boundary (such as a page boundary) necessary; for those cases, use @code{aligned_alloc} or @code{posix_memalign} (@pxref{Aligned Memory Blocks}). Note that the memory located after the end of the block is likely to be in use for something else; perhaps a block already allocated by another call to @code{malloc}. If you attempt to treat the block as longer than you asked for it to be, you are liable to destroy the data that @code{malloc} uses to keep track of its blocks, or you may destroy the contents of another block. If you have already allocated a block and discover you want it to be bigger, use @code{realloc} (@pxref{Changing Block Size}). @node Freeing after Malloc @subsubsection Freeing Memory Allocated with @code{malloc} @cindex freeing memory allocated with @code{malloc} @cindex heap, freeing memory from When you no longer need a block that you got with @code{malloc}, use the function @code{free} to make the block available to be allocated again. The prototype for this function is in @file{stdlib.h}. @pindex stdlib.h @comment malloc.h stdlib.h @comment ISO @deftypefun void free (void *@var{ptr}) @safety{@prelim{}@mtsafe{}@asunsafe{@asulock{}}@acunsafe{@aculock{} @acsfd{} @acsmem{}}} @c __libc_free @asulock @aculock @acsfd @acsmem @c releasing memory into fastbins modifies the arena without taking @c its mutex, but catomic operations ensure safety. If two (or more) @c threads are running malloc and have their own arenas locked when @c each gets a signal whose handler free()s large (non-fastbin-able) @c blocks from each other's arena, we deadlock; this is a more general @c case of @asulock. @c *__free_hook unguarded @c mem2chunk ok @c chunk_is_mmapped ok, chunk bits not modified after allocation @c chunksize ok @c munmap_chunk dup @acsmem @c arena_for_chunk dup ok @c _int_free (!have_lock) dup @asulock @aculock @acsfd @acsmem The @code{free} function deallocates the block of memory pointed at by @var{ptr}. @end deftypefun @comment stdlib.h @comment Sun @deftypefun void cfree (void *@var{ptr}) @safety{@prelim{}@mtsafe{}@asunsafe{@asulock{}}@acunsafe{@aculock{} @acsfd{} @acsmem{}}} @c alias to free This function does the same thing as @code{free}. It's provided for backward compatibility with SunOS; you should use @code{free} instead. @end deftypefun Freeing a block alters the contents of the block. @strong{Do not expect to find any data (such as a pointer to the next block in a chain of blocks) in the block after freeing it.} Copy whatever you need out of the block before freeing it! Here is an example of the proper way to free all the blocks in a chain, and the strings that they point to: @smallexample struct chain @{ struct chain *next; char *name; @} void free_chain (struct chain *chain) @{ while (chain != 0) @{ struct chain *next = chain->next; free (chain->name); free (chain); chain = next; @} @} @end smallexample Occasionally, @code{free} can actually return memory to the operating system and make the process smaller. Usually, all it can do is allow a later call to @code{malloc} to reuse the space. In the meantime, the space remains in your program as part of a free-list used internally by @code{malloc}. There is no point in freeing blocks at the end of a program, because all of the program's space is given back to the system when the process terminates. @node Changing Block Size @subsubsection Changing the Size of a Block @cindex changing the size of a block (@code{malloc}) Often you do not know for certain how big a block you will ultimately need at the time you must begin to use the block. For example, the block might be a buffer that you use to hold a line being read from a file; no matter how long you make the buffer initially, you may encounter a line that is longer. You can make the block longer by calling @code{realloc}. This function is declared in @file{stdlib.h}. @pindex stdlib.h @comment malloc.h stdlib.h @comment ISO @deftypefun {void *} realloc (void *@var{ptr}, size_t @var{newsize}) @safety{@prelim{}@mtsafe{}@asunsafe{@asulock{}}@acunsafe{@aculock{} @acsfd{} @acsmem{}}} @c It may call the implementations of malloc and free, so all of their @c issues arise, plus the realloc hook, also accessed without guards. @c __libc_realloc @asulock @aculock @acsfd @acsmem @c *__realloc_hook unguarded @c __libc_free dup @asulock @aculock @acsfd @acsmem @c __libc_malloc dup @asulock @aculock @acsfd @acsmem @c mem2chunk dup ok @c chunksize dup ok @c malloc_printerr dup ok @c checked_request2size dup ok @c chunk_is_mmapped dup ok @c mremap_chunk @c chunksize dup ok @c __mremap ok @c set_head dup ok @c MALLOC_COPY ok @c memcpy ok @c munmap_chunk dup @acsmem @c arena_for_chunk dup ok @c mutex_lock (arena mutex) dup @asulock @aculock @c _int_realloc @acsfd @acsmem @c malloc_printerr dup ok @c check_inuse_chunk dup ok/disabled @c chunk_at_offset dup ok @c chunksize dup ok @c set_head_size dup ok @c chunk_at_offset dup ok @c set_head dup ok @c chunk2mem dup ok @c inuse dup ok @c unlink dup ok @c _int_malloc dup @acsfd @acsmem @c mem2chunk dup ok @c MALLOC_COPY dup ok @c _int_free (have_lock) dup @acsfd @acsmem @c set_inuse_bit_at_offset dup ok @c set_head dup ok @c mutex_unlock (arena mutex) dup @aculock @c _int_free (!have_lock) dup @asulock @aculock @acsfd @acsmem The @code{realloc} function changes the size of the block whose address is @var{ptr} to be @var{newsize}. Since the space after the end of the block may be in use, @code{realloc} may find it necessary to copy the block to a new address where more free space is available. The value of @code{realloc} is the new address of the block. If the block needs to be moved, @code{realloc} copies the old contents. If you pass a null pointer for @var{ptr}, @code{realloc} behaves just like @samp{malloc (@var{newsize})}. This can be convenient, but beware that older implementations (before @w{ISO C}) may not support this behavior, and will probably crash when @code{realloc} is passed a null pointer. @end deftypefun Like @code{malloc}, @code{realloc} may return a null pointer if no memory space is available to make the block bigger. When this happens, the original block is untouched; it has not been modified or relocated. In most cases it makes no difference what happens to the original block when @code{realloc} fails, because the application program cannot continue when it is out of memory, and the only thing to do is to give a fatal error message. Often it is convenient to write and use a subroutine, conventionally called @code{xrealloc}, that takes care of the error message as @code{xmalloc} does for @code{malloc}: @smallexample void * xrealloc (void *ptr, size_t size) @{ void *value = realloc (ptr, size); if (value == 0) fatal ("Virtual memory exhausted"); return value; @} @end smallexample You can also use @code{realloc} to make a block smaller. The reason you would do this is to avoid tying up a lot of memory space when only a little is needed. @comment The following is no longer true with the new malloc. @comment But it seems wise to keep the warning for other implementations. In several allocation implementations, making a block smaller sometimes necessitates copying it, so it can fail if no other space is available. If the new size you specify is the same as the old size, @code{realloc} is guaranteed to change nothing and return the same address that you gave. @node Allocating Cleared Space @subsubsection Allocating Cleared Space The function @code{calloc} allocates memory and clears it to zero. It is declared in @file{stdlib.h}. @pindex stdlib.h @comment malloc.h stdlib.h @comment ISO @deftypefun {void *} calloc (size_t @var{count}, size_t @var{eltsize}) @safety{@prelim{}@mtsafe{}@asunsafe{@asulock{}}@acunsafe{@aculock{} @acsfd{} @acsmem{}}} @c Same caveats as malloc. @c __libc_calloc @asulock @aculock @acsfd @acsmem @c *__malloc_hook dup unguarded @c memset dup ok @c arena_get @asulock @aculock @acsfd @acsmem @c arena_lock dup @asulock @aculock @acsfd @acsmem @c top dup ok @c chunksize dup ok @c heap_for_ptr dup ok @c _int_malloc dup @acsfd @acsmem @c arena_get_retry dup @asulock @aculock @acsfd @acsmem @c mutex_unlock dup @aculock @c mem2chunk dup ok @c chunk_is_mmapped dup ok @c MALLOC_ZERO ok @c memset dup ok This function allocates a block long enough to contain a vector of @var{count} elements, each of size @var{eltsize}. Its contents are cleared to zero before @code{calloc} returns. @end deftypefun You could define @code{calloc} as follows: @smallexample void * calloc (size_t count, size_t eltsize) @{ size_t size = count * eltsize; void *value = malloc (size); if (value != 0) memset (value, 0, size); return value; @} @end smallexample But in general, it is not guaranteed that @code{calloc} calls @code{malloc} internally. Therefore, if an application provides its own @code{malloc}/@code{realloc}/@code{free} outside the C library, it should always define @code{calloc}, too. @node Efficiency and Malloc @subsubsection Efficiency Considerations for @code{malloc} @cindex efficiency and @code{malloc} @ignore @c No longer true, see below instead. To make the best use of @code{malloc}, it helps to know that the GNU version of @code{malloc} always dispenses small amounts of memory in blocks whose sizes are powers of two. It keeps separate pools for each power of two. This holds for sizes up to a page size. Therefore, if you are free to choose the size of a small block in order to make @code{malloc} more efficient, make it a power of two. @c !!! xref getpagesize Once a page is split up for a particular block size, it can't be reused for another size unless all the blocks in it are freed. In many programs, this is unlikely to happen. Thus, you can sometimes make a program use memory more efficiently by using blocks of the same size for many different purposes. When you ask for memory blocks of a page or larger, @code{malloc} uses a different strategy; it rounds the size up to a multiple of a page, and it can coalesce and split blocks as needed. The reason for the two strategies is that it is important to allocate and free small blocks as fast as possible, but speed is less important for a large block since the program normally spends a fair amount of time using it. Also, large blocks are normally fewer in number. Therefore, for large blocks, it makes sense to use a method which takes more time to minimize the wasted space. @end ignore As opposed to other versions, the @code{malloc} in @theglibc{} does not round up block sizes to powers of two, neither for large nor for small sizes. Neighboring chunks can be coalesced on a @code{free} no matter what their size is. This makes the implementation suitable for all kinds of allocation patterns without generally incurring high memory waste through fragmentation. Very large blocks (much larger than a page) are allocated with @code{mmap} (anonymous or via @code{/dev/zero}) by this implementation. This has the great advantage that these chunks are returned to the system immediately when they are freed. Therefore, it cannot happen that a large chunk becomes ``locked'' in between smaller ones and even after calling @code{free} wastes memory. The size threshold for @code{mmap} to be used can be adjusted with @code{mallopt}. The use of @code{mmap} can also be disabled completely. @node Aligned Memory Blocks @subsubsection Allocating Aligned Memory Blocks @cindex page boundary @cindex alignment (with @code{malloc}) @pindex stdlib.h The address of a block returned by @code{malloc} or @code{realloc} in @gnusystems{} is always a multiple of eight (or sixteen on 64-bit systems). If you need a block whose address is a multiple of a higher power of two than that, use @code{aligned_alloc} or @code{posix_memalign}. @code{aligned_alloc} and @code{posix_memalign} are declared in @file{stdlib.h}. @comment stdlib.h @deftypefun {void *} aligned_alloc (size_t @var{alignment}, size_t @var{size}) @safety{@prelim{}@mtsafe{}@asunsafe{@asulock{}}@acunsafe{@aculock{} @acsfd{} @acsmem{}}} @c Alias to memalign. The @code{aligned_alloc} function allocates a block of @var{size} bytes whose address is a multiple of @var{alignment}. The @var{alignment} must be a power of two and @var{size} must be a multiple of @var{alignment}. The @code{aligned_alloc} function returns a null pointer on error and sets @code{errno} to one of the following values: @table @code @item ENOMEM There was insufficient memory available to satisfy the request. @item EINVAL @var{alignment} is not a power of two. This function was introduced in @w{ISO C11} and hence may have better portability to modern non-POSIX systems than @code{posix_memalign}. @end table @end deftypefun @comment malloc.h @comment BSD @deftypefun {void *} memalign (size_t @var{boundary}, size_t @var{size}) @safety{@prelim{}@mtsafe{}@asunsafe{@asulock{}}@acunsafe{@aculock{} @acsfd{} @acsmem{}}} @c Same issues as malloc. The padding bytes are safely freed in @c _int_memalign, with the arena still locked. @c __libc_memalign @asulock @aculock @acsfd @acsmem @c *__memalign_hook dup unguarded @c __libc_malloc dup @asulock @aculock @acsfd @acsmem @c arena_get dup @asulock @aculock @acsfd @acsmem @c _int_memalign @acsfd @acsmem @c _int_malloc dup @acsfd @acsmem @c checked_request2size dup ok @c mem2chunk dup ok @c chunksize dup ok @c chunk_is_mmapped dup ok @c set_head dup ok @c chunk2mem dup ok @c set_inuse_bit_at_offset dup ok @c set_head_size dup ok @c _int_free (have_lock) dup @acsfd @acsmem @c chunk_at_offset dup ok @c check_inuse_chunk dup ok @c arena_get_retry dup @asulock @aculock @acsfd @acsmem @c mutex_unlock dup @aculock The @code{memalign} function allocates a block of @var{size} bytes whose address is a multiple of @var{boundary}. The @var{boundary} must be a power of two! The function @code{memalign} works by allocating a somewhat larger block, and then returning an address within the block that is on the specified boundary. The @code{memalign} function returns a null pointer on error and sets @code{errno} to one of the following values: @table @code @item ENOMEM There was insufficient memory available to satisfy the request. @item EINVAL @var{alignment} is not a power of two. @end table The @code{memalign} function is obsolete and @code{aligned_alloc} or @code{posix_memalign} should be used instead. @end deftypefun @comment stdlib.h @comment POSIX @deftypefun int posix_memalign (void **@var{memptr}, size_t @var{alignment}, size_t @var{size}) @safety{@prelim{}@mtsafe{}@asunsafe{@asulock{}}@acunsafe{@aculock{} @acsfd{} @acsmem{}}} @c Calls memalign unless the requirements are not met (powerof2 macro is @c safe given an automatic variable as an argument) or there's a @c memalign hook (accessed unguarded, but safely). The @code{posix_memalign} function is similar to the @code{memalign} function in that it returns a buffer of @var{size} bytes aligned to a multiple of @var{alignment}. But it adds one requirement to the parameter @var{alignment}: the value must be a power of two multiple of @code{sizeof (void *)}. If the function succeeds in allocation memory a pointer to the allocated memory is returned in @code{*@var{memptr}} and the return value is zero. Otherwise the function returns an error value indicating the problem. The possible error values returned are: @table @code @item ENOMEM There was insufficient memory available to satisfy the request. @item EINVAL @var{alignment} is not a power of two multiple of @code{sizeof (void *)}. @end table This function was introduced in POSIX 1003.1d. Although this function is superseded by @code{aligned_alloc}, it is more portable to older POSIX systems that do not support @w{ISO C11}. @end deftypefun @comment malloc.h stdlib.h @comment BSD @deftypefun {void *} valloc (size_t @var{size}) @safety{@prelim{}@mtunsafe{@mtuinit{}}@asunsafe{@asuinit{} @asulock{}}@acunsafe{@acuinit{} @aculock{} @acsfd{} @acsmem{}}} @c __libc_valloc @mtuinit @asuinit @asulock @aculock @acsfd @acsmem @c ptmalloc_init (once) @mtsenv @asulock @aculock @acsfd @acsmem @c _dl_addr @asucorrupt? @aculock @c __rtld_lock_lock_recursive (dl_load_lock) @asucorrupt? @aculock @c _dl_find_dso_for_object ok, iterates over dl_ns and its _ns_loaded objs @c the ok above assumes no partial updates on dl_ns and _ns_loaded @c that could confuse a _dl_addr call in a signal handler @c _dl_addr_inside_object ok @c determine_info ok @c __rtld_lock_unlock_recursive (dl_load_lock) @aculock @c thread_atfork @asulock @aculock @acsfd @acsmem @c __register_atfork @asulock @aculock @acsfd @acsmem @c lll_lock (__fork_lock) @asulock @aculock @c fork_handler_alloc @asulock @aculock @acsfd @acsmem @c calloc dup @asulock @aculock @acsfd @acsmem @c __linkin_atfork ok @c catomic_compare_and_exchange_bool_acq ok @c lll_unlock (__fork_lock) @aculock @c *_environ @mtsenv @c next_env_entry ok @c strcspn dup ok @c __libc_mallopt dup @mtasuconst:mallopt [setting mp_] @c __malloc_check_init @mtasuconst:malloc_hooks [setting hooks] @c *__malloc_initialize_hook unguarded, ok @c *__memalign_hook dup ok, unguarded @c arena_get dup @asulock @aculock @acsfd @acsmem @c _int_valloc @acsfd @acsmem @c malloc_consolidate dup ok @c _int_memalign dup @acsfd @acsmem @c arena_get_retry dup @asulock @aculock @acsfd @acsmem @c _int_memalign dup @acsfd @acsmem @c mutex_unlock dup @aculock Using @code{valloc} is like using @code{memalign} and passing the page size as the value of the second argument. It is implemented like this: @smallexample void * valloc (size_t size) @{ return memalign (getpagesize (), size); @} @end smallexample @ref{Query Memory Parameters} for more information about the memory subsystem. The @code{valloc} function is obsolete and @code{aligned_alloc} or @code{posix_memalign} should be used instead. @end deftypefun @node Malloc Tunable Parameters @subsubsection Malloc Tunable Parameters You can adjust some parameters for dynamic memory allocation with the @code{mallopt} function. This function is the general SVID/XPG interface, defined in @file{malloc.h}. @pindex malloc.h @deftypefun int mallopt (int @var{param}, int @var{value}) @safety{@prelim{}@mtunsafe{@mtuinit{} @mtasuconst{:mallopt}}@asunsafe{@asuinit{} @asulock{}}@acunsafe{@acuinit{} @aculock{}}} @c __libc_mallopt @mtuinit @mtasuconst:mallopt @asuinit @asulock @aculock @c ptmalloc_init (once) dup @mtsenv @asulock @aculock @acsfd @acsmem @c mutex_lock (main_arena->mutex) @asulock @aculock @c malloc_consolidate dup ok @c set_max_fast ok @c mutex_unlock dup @aculock When calling @code{mallopt}, the @var{param} argument specifies the parameter to be set, and @var{value} the new value to be set. Possible choices for @var{param}, as defined in @file{malloc.h}, are: @table @code @comment TODO: @item M_ARENA_MAX @comment - Document ARENA_MAX env var. @comment TODO: @item M_ARENA_TEST @comment - Document ARENA_TEST env var. @comment TODO: @item M_CHECK_ACTION @item M_MMAP_MAX The maximum number of chunks to allocate with @code{mmap}. Setting this to zero disables all use of @code{mmap}. @item M_MMAP_THRESHOLD All chunks larger than this value are allocated outside the normal heap, using the @code{mmap} system call. This way it is guaranteed that the memory for these chunks can be returned to the system on @code{free}. Note that requests smaller than this threshold might still be allocated via @code{mmap}. @comment TODO: @item M_MXFAST @item M_PERTURB If non-zero, memory blocks are filled with values depending on some low order bits of this parameter when they are allocated (except when allocated by @code{calloc}) and freed. This can be used to debug the use of uninitialized or freed heap memory. Note that this option does not guarantee that the freed block will have any specific values. It only guarantees that the content the block had before it was freed will be overwritten. @item M_TOP_PAD This parameter determines the amount of extra memory to obtain from the system when a call to @code{sbrk} is required. It also specifies the number of bytes to retain when shrinking the heap by calling @code{sbrk} with a negative argument. This provides the necessary hysteresis in heap size such that excessive amounts of system calls can be avoided. @item M_TRIM_THRESHOLD This is the minimum size (in bytes) of the top-most, releasable chunk that will cause @code{sbrk} to be called with a negative argument in order to return memory to the system. @end table @end deftypefun @node Heap Consistency Checking @subsubsection Heap Consistency Checking @cindex heap consistency checking @cindex consistency checking, of heap You can ask @code{malloc} to check the consistency of dynamic memory by using the @code{mcheck} function. This function is a GNU extension, declared in @file{mcheck.h}. @pindex mcheck.h @comment mcheck.h @comment GNU @deftypefun int mcheck (void (*@var{abortfn}) (enum mcheck_status @var{status})) @safety{@prelim{}@mtunsafe{@mtasurace{:mcheck} @mtasuconst{:malloc_hooks}}@asunsafe{@asucorrupt{}}@acunsafe{@acucorrupt{}}} @c The hooks must be set up before malloc is first used, which sort of @c implies @mtuinit/@asuinit but since the function is a no-op if malloc @c was already used, that doesn't pose any safety issues. The actual @c problem is with the hooks, designed for single-threaded @c fully-synchronous operation: they manage an unguarded linked list of @c allocated blocks, and get temporarily overwritten before calling the @c allocation functions recursively while holding the old hooks. There @c are no guards for thread safety, and inconsistent hooks may be found @c within signal handlers or left behind in case of cancellation. Calling @code{mcheck} tells @code{malloc} to perform occasional consistency checks. These will catch things such as writing past the end of a block that was allocated with @code{malloc}. The @var{abortfn} argument is the function to call when an inconsistency is found. If you supply a null pointer, then @code{mcheck} uses a default function which prints a message and calls @code{abort} (@pxref{Aborting a Program}). The function you supply is called with one argument, which says what sort of inconsistency was detected; its type is described below. It is too late to begin allocation checking once you have allocated anything with @code{malloc}. So @code{mcheck} does nothing in that case. The function returns @code{-1} if you call it too late, and @code{0} otherwise (when it is successful). The easiest way to arrange to call @code{mcheck} early enough is to use the option @samp{-lmcheck} when you link your program; then you don't need to modify your program source at all. Alternatively you might use a debugger to insert a call to @code{mcheck} whenever the program is started, for example these gdb commands will automatically call @code{mcheck} whenever the program starts: @smallexample (gdb) break main Breakpoint 1, main (argc=2, argv=0xbffff964) at whatever.c:10 (gdb) command 1 Type commands for when breakpoint 1 is hit, one per line. End with a line saying just "end". >call mcheck(0) >continue >end (gdb) @dots{} @end smallexample This will however only work if no initialization function of any object involved calls any of the @code{malloc} functions since @code{mcheck} must be called before the first such function. @end deftypefun @deftypefun {enum mcheck_status} mprobe (void *@var{pointer}) @safety{@prelim{}@mtunsafe{@mtasurace{:mcheck} @mtasuconst{:malloc_hooks}}@asunsafe{@asucorrupt{}}@acunsafe{@acucorrupt{}}} @c The linked list of headers may be modified concurrently by other @c threads, and it may find a partial update if called from a signal @c handler. It's mostly read only, so cancelling it might be safe, but @c it will modify global state that, if cancellation hits at just the @c right spot, may be left behind inconsistent. This path is only taken @c if checkhdr finds an inconsistency. If the inconsistency could only @c occur because of earlier undefined behavior, that wouldn't be an @c additional safety issue problem, but because of the other concurrency @c issues in the mcheck hooks, the apparent inconsistency could be the @c result of mcheck's own internal data race. So, AC-Unsafe it is. The @code{mprobe} function lets you explicitly check for inconsistencies in a particular allocated block. You must have already called @code{mcheck} at the beginning of the program, to do its occasional checks; calling @code{mprobe} requests an additional consistency check to be done at the time of the call. The argument @var{pointer} must be a pointer returned by @code{malloc} or @code{realloc}. @code{mprobe} returns a value that says what inconsistency, if any, was found. The values are described below. @end deftypefun @deftp {Data Type} {enum mcheck_status} This enumerated type describes what kind of inconsistency was detected in an allocated block, if any. Here are the possible values: @table @code @item MCHECK_DISABLED @code{mcheck} was not called before the first allocation. No consistency checking can be done. @item MCHECK_OK No inconsistency detected. @item MCHECK_HEAD The data immediately before the block was modified. This commonly happens when an array index or pointer is decremented too far. @item MCHECK_TAIL The data immediately after the block was modified. This commonly happens when an array index or pointer is incremented too far. @item MCHECK_FREE The block was already freed. @end table @end deftp Another possibility to check for and guard against bugs in the use of @code{malloc}, @code{realloc} and @code{free} is to set the environment variable @code{MALLOC_CHECK_}. When @code{MALLOC_CHECK_} is set, a special (less efficient) implementation is used which is designed to be tolerant against simple errors, such as double calls of @code{free} with the same argument, or overruns of a single byte (off-by-one bugs). Not all such errors can be protected against, however, and memory leaks can result. If @code{MALLOC_CHECK_} is set to @code{0}, any detected heap corruption is silently ignored; if set to @code{1}, a diagnostic is printed on @code{stderr}; if set to @code{2}, @code{abort} is called immediately. This can be useful because otherwise a crash may happen much later, and the true cause for the problem is then very hard to track down. There is one problem with @code{MALLOC_CHECK_}: in SUID or SGID binaries it could possibly be exploited since diverging from the normal programs behavior it now writes something to the standard error descriptor. Therefore the use of @code{MALLOC_CHECK_} is disabled by default for SUID and SGID binaries. It can be enabled again by the system administrator by adding a file @file{/etc/suid-debug} (the content is not important it could be empty). So, what's the difference between using @code{MALLOC_CHECK_} and linking with @samp{-lmcheck}? @code{MALLOC_CHECK_} is orthogonal with respect to @samp{-lmcheck}. @samp{-lmcheck} has been added for backward compatibility. Both @code{MALLOC_CHECK_} and @samp{-lmcheck} should uncover the same bugs - but using @code{MALLOC_CHECK_} you don't need to recompile your application. @node Hooks for Malloc @subsubsection Memory Allocation Hooks @cindex allocation hooks, for @code{malloc} @Theglibc{} lets you modify the behavior of @code{malloc}, @code{realloc}, and @code{free} by specifying appropriate hook functions. You can use these hooks to help you debug programs that use dynamic memory allocation, for example. The hook variables are declared in @file{malloc.h}. @pindex malloc.h @comment malloc.h @comment GNU @defvar __malloc_hook The value of this variable is a pointer to the function that @code{malloc} uses whenever it is called. You should define this function to look like @code{malloc}; that is, like: @smallexample void *@var{function} (size_t @var{size}, const void *@var{caller}) @end smallexample The value of @var{caller} is the return address found on the stack when the @code{malloc} function was called. This value allows you to trace the memory consumption of the program. @end defvar @comment malloc.h @comment GNU @defvar __realloc_hook The value of this variable is a pointer to function that @code{realloc} uses whenever it is called. You should define this function to look like @code{realloc}; that is, like: @smallexample void *@var{function} (void *@var{ptr}, size_t @var{size}, const void *@var{caller}) @end smallexample The value of @var{caller} is the return address found on the stack when the @code{realloc} function was called. This value allows you to trace the memory consumption of the program. @end defvar @comment malloc.h @comment GNU @defvar __free_hook The value of this variable is a pointer to function that @code{free} uses whenever it is called. You should define this function to look like @code{free}; that is, like: @smallexample void @var{function} (void *@var{ptr}, const void *@var{caller}) @end smallexample The value of @var{caller} is the return address found on the stack when the @code{free} function was called. This value allows you to trace the memory consumption of the program. @end defvar @comment malloc.h @comment GNU @defvar __memalign_hook The value of this variable is a pointer to function that @code{aligned_alloc}, @code{memalign}, @code{posix_memalign} and @code{valloc} use whenever they are called. You should define this function to look like @code{aligned_alloc}; that is, like: @smallexample void *@var{function} (size_t @var{alignment}, size_t @var{size}, const void *@var{caller}) @end smallexample The value of @var{caller} is the return address found on the stack when the @code{aligned_alloc}, @code{memalign}, @code{posix_memalign} or @code{valloc} functions are called. This value allows you to trace the memory consumption of the program. @end defvar You must make sure that the function you install as a hook for one of these functions does not call that function recursively without restoring the old value of the hook first! Otherwise, your program will get stuck in an infinite recursion. Before calling the function recursively, one should make sure to restore all the hooks to their previous value. When coming back from the recursive call, all the hooks should be resaved since a hook might modify itself. @comment malloc.h @comment GNU @defvar __malloc_initialize_hook The value of this variable is a pointer to a function that is called once when the malloc implementation is initialized. This is a weak variable, so it can be overridden in the application with a definition like the following: @smallexample void (*@var{__malloc_initialize_hook}) (void) = my_init_hook; @end smallexample @end defvar An issue to look out for is the time at which the malloc hook functions can be safely installed. If the hook functions call the malloc-related functions recursively, it is necessary that malloc has already properly initialized itself at the time when @code{__malloc_hook} etc. is assigned to. On the other hand, if the hook functions provide a complete malloc implementation of their own, it is vital that the hooks are assigned to @emph{before} the very first @code{malloc} call has completed, because otherwise a chunk obtained from the ordinary, un-hooked malloc may later be handed to @code{__free_hook}, for example. In both cases, the problem can be solved by setting up the hooks from within a user-defined function pointed to by @code{__malloc_initialize_hook}---then the hooks will be set up safely at the right time. Here is an example showing how to use @code{__malloc_hook} and @code{__free_hook} properly. It installs a function that prints out information every time @code{malloc} or @code{free} is called. We just assume here that @code{realloc} and @code{memalign} are not used in our program. @smallexample /* Prototypes for __malloc_hook, __free_hook */ #include /* Prototypes for our hooks. */ static void my_init_hook (void); static void *my_malloc_hook (size_t, const void *); static void my_free_hook (void*, const void *); /* Override initializing hook from the C library. */ void (*__malloc_initialize_hook) (void) = my_init_hook; static void my_init_hook (void) @{ old_malloc_hook = __malloc_hook; old_free_hook = __free_hook; __malloc_hook = my_malloc_hook; __free_hook = my_free_hook; @} static void * my_malloc_hook (size_t size, const void *caller) @{ void *result; /* Restore all old hooks */ __malloc_hook = old_malloc_hook; __free_hook = old_free_hook; /* Call recursively */ result = malloc (size); /* Save underlying hooks */ old_malloc_hook = __malloc_hook; old_free_hook = __free_hook; /* @r{@code{printf} might call @code{malloc}, so protect it too.} */ printf ("malloc (%u) returns %p\n", (unsigned int) size, result); /* Restore our own hooks */ __malloc_hook = my_malloc_hook; __free_hook = my_free_hook; return result; @} static void my_free_hook (void *ptr, const void *caller) @{ /* Restore all old hooks */ __malloc_hook = old_malloc_hook; __free_hook = old_free_hook; /* Call recursively */ free (ptr); /* Save underlying hooks */ old_malloc_hook = __malloc_hook; old_free_hook = __free_hook; /* @r{@code{printf} might call @code{free}, so protect it too.} */ printf ("freed pointer %p\n", ptr); /* Restore our own hooks */ __malloc_hook = my_malloc_hook; __free_hook = my_free_hook; @} main () @{ @dots{} @} @end smallexample The @code{mcheck} function (@pxref{Heap Consistency Checking}) works by installing such hooks. @c __morecore, __after_morecore_hook are undocumented @c It's not clear whether to document them. @node Statistics of Malloc @subsubsection Statistics for Memory Allocation with @code{malloc} @cindex allocation statistics You can get information about dynamic memory allocation by calling the @code{mallinfo} function. This function and its associated data type are declared in @file{malloc.h}; they are an extension of the standard SVID/XPG version. @pindex malloc.h @comment malloc.h @comment GNU @deftp {Data Type} {struct mallinfo} This structure type is used to return information about the dynamic memory allocator. It contains the following members: @table @code @item int arena This is the total size of memory allocated with @code{sbrk} by @code{malloc}, in bytes. @item int ordblks This is the number of chunks not in use. (The memory allocator internally gets chunks of memory from the operating system, and then carves them up to satisfy individual @code{malloc} requests; see @ref{Efficiency and Malloc}.) @item int smblks This field is unused. @item int hblks This is the total number of chunks allocated with @code{mmap}. @item int hblkhd This is the total size of memory allocated with @code{mmap}, in bytes. @item int usmblks This field is unused. @item int fsmblks This field is unused. @item int uordblks This is the total size of memory occupied by chunks handed out by @code{malloc}. @item int fordblks This is the total size of memory occupied by free (not in use) chunks. @item int keepcost This is the size of the top-most releasable chunk that normally borders the end of the heap (i.e., the high end of the virtual address space's data segment). @end table @end deftp @comment malloc.h @comment SVID @deftypefun {struct mallinfo} mallinfo (void) @safety{@prelim{}@mtunsafe{@mtuinit{} @mtasuconst{:mallopt}}@asunsafe{@asuinit{} @asulock{}}@acunsafe{@acuinit{} @aculock{}}} @c Accessing mp_.n_mmaps and mp_.max_mmapped_mem, modified with atomics @c but non-atomically elsewhere, may get us inconsistent results. We @c mark the statistics as unsafe, rather than the fast-path functions @c that collect the possibly inconsistent data. @c __libc_mallinfo @mtuinit @mtasuconst:mallopt @asuinit @asulock @aculock @c ptmalloc_init (once) dup @mtsenv @asulock @aculock @acsfd @acsmem @c mutex_lock dup @asulock @aculock @c int_mallinfo @mtasuconst:mallopt [mp_ access on main_arena] @c malloc_consolidate dup ok @c check_malloc_state dup ok/disabled @c chunksize dup ok @c fastbin dupo ok @c bin_at dup ok @c last dup ok @c mutex_unlock @aculock This function returns information about the current dynamic memory usage in a structure of type @code{struct mallinfo}. @end deftypefun @node Summary of Malloc @subsubsection Summary of @code{malloc}-Related Functions Here is a summary of the functions that work with @code{malloc}: @table @code @item void *malloc (size_t @var{size}) Allocate a block of @var{size} bytes. @xref{Basic Allocation}. @item void free (void *@var{addr}) Free a block previously allocated by @code{malloc}. @xref{Freeing after Malloc}. @item void *realloc (void *@var{addr}, size_t @var{size}) Make a block previously allocated by @code{malloc} larger or smaller, possibly by copying it to a new location. @xref{Changing Block Size}. @item void *calloc (size_t @var{count}, size_t @var{eltsize}) Allocate a block of @var{count} * @var{eltsize} bytes using @code{malloc}, and set its contents to zero. @xref{Allocating Cleared Space}. @item void *valloc (size_t @var{size}) Allocate a block of @var{size} bytes, starting on a page boundary. @xref{Aligned Memory Blocks}. @item void *aligned_alloc (size_t @var{size}, size_t @var{alignment}) Allocate a block of @var{size} bytes, starting on an address that is a multiple of @var{alignment}. @xref{Aligned Memory Blocks}. @item int posix_memalign (void **@var{memptr}, size_t @var{alignment}, size_t @var{size}) Allocate a block of @var{size} bytes, starting on an address that is a multiple of @var{alignment}. @xref{Aligned Memory Blocks}. @item void *memalign (size_t @var{size}, size_t @var{boundary}) Allocate a block of @var{size} bytes, starting on an address that is a multiple of @var{boundary}. @xref{Aligned Memory Blocks}. @item int mallopt (int @var{param}, int @var{value}) Adjust a tunable parameter. @xref{Malloc Tunable Parameters}. @item int mcheck (void (*@var{abortfn}) (void)) Tell @code{malloc} to perform occasional consistency checks on dynamically allocated memory, and to call @var{abortfn} when an inconsistency is found. @xref{Heap Consistency Checking}. @item void *(*__malloc_hook) (size_t @var{size}, const void *@var{caller}) A pointer to a function that @code{malloc} uses whenever it is called. @item void *(*__realloc_hook) (void *@var{ptr}, size_t @var{size}, const void *@var{caller}) A pointer to a function that @code{realloc} uses whenever it is called. @item void (*__free_hook) (void *@var{ptr}, const void *@var{caller}) A pointer to a function that @code{free} uses whenever it is called. @item void (*__memalign_hook) (size_t @var{size}, size_t @var{alignment}, const void *@var{caller}) A pointer to a function that @code{aligned_alloc}, @code{memalign}, @code{posix_memalign} and @code{valloc} use whenever they are called. @item struct mallinfo mallinfo (void) Return information about the current dynamic memory usage. @xref{Statistics of Malloc}. @end table @node Allocation Debugging @subsection Allocation Debugging @cindex allocation debugging @cindex malloc debugger A complicated task when programming with languages which do not use garbage collected dynamic memory allocation is to find memory leaks. Long running programs must assure that dynamically allocated objects are freed at the end of their lifetime. If this does not happen the system runs out of memory, sooner or later. The @code{malloc} implementation in @theglibc{} provides some simple means to detect such leaks and obtain some information to find the location. To do this the application must be started in a special mode which is enabled by an environment variable. There are no speed penalties for the program if the debugging mode is not enabled. @menu * Tracing malloc:: How to install the tracing functionality. * Using the Memory Debugger:: Example programs excerpts. * Tips for the Memory Debugger:: Some more or less clever ideas. * Interpreting the traces:: What do all these lines mean? @end menu @node Tracing malloc @subsubsection How to install the tracing functionality @comment mcheck.h @comment GNU @deftypefun void mtrace (void) @safety{@prelim{}@mtunsafe{@mtsenv{} @mtasurace{:mtrace} @mtasuconst{:malloc_hooks} @mtuinit{}}@asunsafe{@asuinit{} @ascuheap{} @asucorrupt{} @asulock{}}@acunsafe{@acuinit{} @acucorrupt{} @aculock{} @acsfd{} @acsmem{}}} @c Like the mcheck hooks, these are not designed with thread safety in @c mind, because the hook pointers are temporarily modified without @c regard to other threads, signals or cancellation. @c mtrace @mtuinit @mtasurace:mtrace @mtsenv @asuinit @ascuheap @asucorrupt @acuinit @acucorrupt @aculock @acsfd @acsmem @c __libc_secure_getenv dup @mtsenv @c malloc dup @ascuheap @acsmem @c fopen dup @ascuheap @asulock @aculock @acsmem @acsfd @c fcntl dup ok @c setvbuf dup @aculock @c fprintf dup (on newly-created stream) @aculock @c __cxa_atexit (once) dup @asulock @aculock @acsmem @c free dup @ascuheap @acsmem When the @code{mtrace} function is called it looks for an environment variable named @code{MALLOC_TRACE}. This variable is supposed to contain a valid file name. The user must have write access. If the file already exists it is truncated. If the environment variable is not set or it does not name a valid file which can be opened for writing nothing is done. The behavior of @code{malloc} etc. is not changed. For obvious reasons this also happens if the application is installed with the SUID or SGID bit set. If the named file is successfully opened, @code{mtrace} installs special handlers for the functions @code{malloc}, @code{realloc}, and @code{free} (@pxref{Hooks for Malloc}). From then on, all uses of these functions are traced and protocolled into the file. There is now of course a speed penalty for all calls to the traced functions so tracing should not be enabled during normal use. This function is a GNU extension and generally not available on other systems. The prototype can be found in @file{mcheck.h}. @end deftypefun @comment mcheck.h @comment GNU @deftypefun void muntrace (void) @safety{@prelim{}@mtunsafe{@mtasurace{:mtrace} @mtasuconst{:malloc_hooks} @mtslocale{}}@asunsafe{@asucorrupt{} @ascuheap{}}@acunsafe{@acucorrupt{} @acsmem{} @aculock{} @acsfd{}}} @c muntrace @mtasurace:mtrace @mtslocale @asucorrupt @ascuheap @acucorrupt @acsmem @aculock @acsfd @c fprintf (fputs) dup @mtslocale @asucorrupt @ascuheap @acsmem @aculock @acucorrupt @c fclose dup @ascuheap @asulock @aculock @acsmem @acsfd The @code{muntrace} function can be called after @code{mtrace} was used to enable tracing the @code{malloc} calls. If no (successful) call of @code{mtrace} was made @code{muntrace} does nothing. Otherwise it deinstalls the handlers for @code{malloc}, @code{realloc}, and @code{free} and then closes the protocol file. No calls are protocolled anymore and the program runs again at full speed. This function is a GNU extension and generally not available on other systems. The prototype can be found in @file{mcheck.h}. @end deftypefun @node Using the Memory Debugger @subsubsection Example program excerpts Even though the tracing functionality does not influence the runtime behavior of the program it is not a good idea to call @code{mtrace} in all programs. Just imagine that you debug a program using @code{mtrace} and all other programs used in the debugging session also trace their @code{malloc} calls. The output file would be the same for all programs and thus is unusable. Therefore one should call @code{mtrace} only if compiled for debugging. A program could therefore start like this: @example #include int main (int argc, char *argv[]) @{ #ifdef DEBUGGING mtrace (); #endif @dots{} @} @end example This is all what is needed if you want to trace the calls during the whole runtime of the program. Alternatively you can stop the tracing at any time with a call to @code{muntrace}. It is even possible to restart the tracing again with a new call to @code{mtrace}. But this can cause unreliable results since there may be calls of the functions which are not called. Please note that not only the application uses the traced functions, also libraries (including the C library itself) use these functions. This last point is also why it is no good idea to call @code{muntrace} before the program terminated. The libraries are informed about the termination of the program only after the program returns from @code{main} or calls @code{exit} and so cannot free the memory they use before this time. So the best thing one can do is to call @code{mtrace} as the very first function in the program and never call @code{muntrace}. So the program traces almost all uses of the @code{malloc} functions (except those calls which are executed by constructors of the program or used libraries). @node Tips for the Memory Debugger @subsubsection Some more or less clever ideas You know the situation. The program is prepared for debugging and in all debugging sessions it runs well. But once it is started without debugging the error shows up. A typical example is a memory leak that becomes visible only when we turn off the debugging. If you foresee such situations you can still win. Simply use something equivalent to the following little program: @example #include #include static void enable (int sig) @{ mtrace (); signal (SIGUSR1, enable); @} static void disable (int sig) @{ muntrace (); signal (SIGUSR2, disable); @} int main (int argc, char *argv[]) @{ @dots{} signal (SIGUSR1, enable); signal (SIGUSR2, disable); @dots{} @} @end example I.e., the user can start the memory debugger any time s/he wants if the program was started with @code{MALLOC_TRACE} set in the environment. The output will of course not show the allocations which happened before the first signal but if there is a memory leak this will show up nevertheless. @node Interpreting the traces @subsubsection Interpreting the traces If you take a look at the output it will look similar to this: @example = Start @ [0x8048209] - 0x8064cc8 @ [0x8048209] - 0x8064ce0 @ [0x8048209] - 0x8064cf8 @ [0x80481eb] + 0x8064c48 0x14 @ [0x80481eb] + 0x8064c60 0x14 @ [0x80481eb] + 0x8064c78 0x14 @ [0x80481eb] + 0x8064c90 0x14 = End @end example What this all means is not really important since the trace file is not meant to be read by a human. Therefore no attention is given to readability. Instead there is a program which comes with @theglibc{} which interprets the traces and outputs a summary in an user-friendly way. The program is called @code{mtrace} (it is in fact a Perl script) and it takes one or two arguments. In any case the name of the file with the trace output must be specified. If an optional argument precedes the name of the trace file this must be the name of the program which generated the trace. @example drepper$ mtrace tst-mtrace log No memory leaks. @end example In this case the program @code{tst-mtrace} was run and it produced a trace file @file{log}. The message printed by @code{mtrace} shows there are no problems with the code, all allocated memory was freed afterwards. If we call @code{mtrace} on the example trace given above we would get a different outout: @example drepper$ mtrace errlog - 0x08064cc8 Free 2 was never alloc'd 0x8048209 - 0x08064ce0 Free 3 was never alloc'd 0x8048209 - 0x08064cf8 Free 4 was never alloc'd 0x8048209 Memory not freed: ----------------- Address Size Caller 0x08064c48 0x14 at 0x80481eb 0x08064c60 0x14 at 0x80481eb 0x08064c78 0x14 at 0x80481eb 0x08064c90 0x14 at 0x80481eb @end example We have called @code{mtrace} with only one argument and so the script has no chance to find out what is meant with the addresses given in the trace. We can do better: @example drepper$ mtrace tst errlog - 0x08064cc8 Free 2 was never alloc'd /home/drepper/tst.c:39 - 0x08064ce0 Free 3 was never alloc'd /home/drepper/tst.c:39 - 0x08064cf8 Free 4 was never alloc'd /home/drepper/tst.c:39 Memory not freed: ----------------- Address Size Caller 0x08064c48 0x14 at /home/drepper/tst.c:33 0x08064c60 0x14 at /home/drepper/tst.c:33 0x08064c78 0x14 at /home/drepper/tst.c:33 0x08064c90 0x14 at /home/drepper/tst.c:33 @end example Suddenly the output makes much more sense and the user can see immediately where the function calls causing the trouble can be found. Interpreting this output is not complicated. There are at most two different situations being detected. First, @code{free} was called for pointers which were never returned by one of the allocation functions. This is usually a very bad problem and what this looks like is shown in the first three lines of the output. Situations like this are quite rare and if they appear they show up very drastically: the program normally crashes. The other situation which is much harder to detect are memory leaks. As you can see in the output the @code{mtrace} function collects all this information and so can say that the program calls an allocation function from line 33 in the source file @file{/home/drepper/tst-mtrace.c} four times without freeing this memory before the program terminates. Whether this is a real problem remains to be investigated. @node Obstacks @subsection Obstacks @cindex obstacks An @dfn{obstack} is a pool of memory containing a stack of objects. You can create any number of separate obstacks, and then allocate objects in specified obstacks. Within each obstack, the last object allocated must always be the first one freed, but distinct obstacks are independent of each other. Aside from this one constraint of order of freeing, obstacks are totally general: an obstack can contain any number of objects of any size. They are implemented with macros, so allocation is usually very fast as long as the objects are usually small. And the only space overhead per object is the padding needed to start each object on a suitable boundary. @menu * Creating Obstacks:: How to declare an obstack in your program. * Preparing for Obstacks:: Preparations needed before you can use obstacks. * Allocation in an Obstack:: Allocating objects in an obstack. * Freeing Obstack Objects:: Freeing objects in an obstack. * Obstack Functions:: The obstack functions are both functions and macros. * Growing Objects:: Making an object bigger by stages. * Extra Fast Growing:: Extra-high-efficiency (though more complicated) growing objects. * Status of an Obstack:: Inquiries about the status of an obstack. * Obstacks Data Alignment:: Controlling alignment of objects in obstacks. * Obstack Chunks:: How obstacks obtain and release chunks; efficiency considerations. * Summary of Obstacks:: @end menu @node Creating Obstacks @subsubsection Creating Obstacks The utilities for manipulating obstacks are declared in the header file @file{obstack.h}. @pindex obstack.h @comment obstack.h @comment GNU @deftp {Data Type} {struct obstack} An obstack is represented by a data structure of type @code{struct obstack}. This structure has a small fixed size; it records the status of the obstack and how to find the space in which objects are allocated. It does not contain any of the objects themselves. You should not try to access the contents of the structure directly; use only the functions described in this chapter. @end deftp You can declare variables of type @code{struct obstack} and use them as obstacks, or you can allocate obstacks dynamically like any other kind of object. Dynamic allocation of obstacks allows your program to have a variable number of different stacks. (You can even allocate an obstack structure in another obstack, but this is rarely useful.) All the functions that work with obstacks require you to specify which obstack to use. You do this with a pointer of type @code{struct obstack *}. In the following, we often say ``an obstack'' when strictly speaking the object at hand is such a pointer. The objects in the obstack are packed into large blocks called @dfn{chunks}. The @code{struct obstack} structure points to a chain of the chunks currently in use. The obstack library obtains a new chunk whenever you allocate an object that won't fit in the previous chunk. Since the obstack library manages chunks automatically, you don't need to pay much attention to them, but you do need to supply a function which the obstack library should use to get a chunk. Usually you supply a function which uses @code{malloc} directly or indirectly. You must also supply a function to free a chunk. These matters are described in the following section. @node Preparing for Obstacks @subsubsection Preparing for Using Obstacks Each source file in which you plan to use the obstack functions must include the header file @file{obstack.h}, like this: @smallexample #include @end smallexample @findex obstack_chunk_alloc @findex obstack_chunk_free Also, if the source file uses the macro @code{obstack_init}, it must declare or define two functions or macros that will be called by the obstack library. One, @code{obstack_chunk_alloc}, is used to allocate the chunks of memory into which objects are packed. The other, @code{obstack_chunk_free}, is used to return chunks when the objects in them are freed. These macros should appear before any use of obstacks in the source file. Usually these are defined to use @code{malloc} via the intermediary @code{xmalloc} (@pxref{Unconstrained Allocation}). This is done with the following pair of macro definitions: @smallexample #define obstack_chunk_alloc xmalloc #define obstack_chunk_free free @end smallexample @noindent Though the memory you get using obstacks really comes from @code{malloc}, using obstacks is faster because @code{malloc} is called less often, for larger blocks of memory. @xref{Obstack Chunks}, for full details. At run time, before the program can use a @code{struct obstack} object as an obstack, it must initialize the obstack by calling @code{obstack_init}. @comment obstack.h @comment GNU @deftypefun int obstack_init (struct obstack *@var{obstack-ptr}) @safety{@prelim{}@mtsafe{@mtsrace{:obstack-ptr}}@assafe{}@acsafe{@acsmem{}}} @c obstack_init @mtsrace:obstack-ptr @acsmem @c _obstack_begin @acsmem @c chunkfun = obstack_chunk_alloc (suggested malloc) @c freefun = obstack_chunk_free (suggested free) @c *chunkfun @acsmem @c obstack_chunk_alloc user-supplied @c *obstack_alloc_failed_handler user-supplied @c -> print_and_abort (default) @c @c print_and_abort @c _ dup @ascuintl @c fxprintf dup @asucorrupt @aculock @acucorrupt @c exit @acucorrupt? Initialize obstack @var{obstack-ptr} for allocation of objects. This function calls the obstack's @code{obstack_chunk_alloc} function. If allocation of memory fails, the function pointed to by @code{obstack_alloc_failed_handler} is called. The @code{obstack_init} function always returns 1 (Compatibility notice: Former versions of obstack returned 0 if allocation failed). @end deftypefun Here are two examples of how to allocate the space for an obstack and initialize it. First, an obstack that is a static variable: @smallexample static struct obstack myobstack; @dots{} obstack_init (&myobstack); @end smallexample @noindent Second, an obstack that is itself dynamically allocated: @smallexample struct obstack *myobstack_ptr = (struct obstack *) xmalloc (sizeof (struct obstack)); obstack_init (myobstack_ptr); @end smallexample @comment obstack.h @comment GNU @defvar obstack_alloc_failed_handler The value of this variable is a pointer to a function that @code{obstack} uses when @code{obstack_chunk_alloc} fails to allocate memory. The default action is to print a message and abort. You should supply a function that either calls @code{exit} (@pxref{Program Termination}) or @code{longjmp} (@pxref{Non-Local Exits}) and doesn't return. @smallexample void my_obstack_alloc_failed (void) @dots{} obstack_alloc_failed_handler = &my_obstack_alloc_failed; @end smallexample @end defvar @node Allocation in an Obstack @subsubsection Allocation in an Obstack @cindex allocation (obstacks) The most direct way to allocate an object in an obstack is with @code{obstack_alloc}, which is invoked almost like @code{malloc}. @comment obstack.h @comment GNU @deftypefun {void *} obstack_alloc (struct obstack *@var{obstack-ptr}, int @var{size}) @safety{@prelim{}@mtsafe{@mtsrace{:obstack-ptr}}@assafe{}@acunsafe{@acucorrupt{} @acsmem{}}} @c obstack_alloc @mtsrace:obstack-ptr @acucorrupt @acsmem @c obstack_blank dup @mtsrace:obstack-ptr @acucorrupt @acsmem @c obstack_finish dup @mtsrace:obstack-ptr @acucorrupt This allocates an uninitialized block of @var{size} bytes in an obstack and returns its address. Here @var{obstack-ptr} specifies which obstack to allocate the block in; it is the address of the @code{struct obstack} object which represents the obstack. Each obstack function or macro requires you to specify an @var{obstack-ptr} as the first argument. This function calls the obstack's @code{obstack_chunk_alloc} function if it needs to allocate a new chunk of memory; it calls @code{obstack_alloc_failed_handler} if allocation of memory by @code{obstack_chunk_alloc} failed. @end deftypefun For example, here is a function that allocates a copy of a string @var{str} in a specific obstack, which is in the variable @code{string_obstack}: @smallexample struct obstack string_obstack; char * copystring (char *string) @{ size_t len = strlen (string) + 1; char *s = (char *) obstack_alloc (&string_obstack, len); memcpy (s, string, len); return s; @} @end smallexample To allocate a block with specified contents, use the function @code{obstack_copy}, declared like this: @comment obstack.h @comment GNU @deftypefun {void *} obstack_copy (struct obstack *@var{obstack-ptr}, void *@var{address}, int @var{size}) @safety{@prelim{}@mtsafe{@mtsrace{:obstack-ptr}}@assafe{}@acunsafe{@acucorrupt{} @acsmem{}}} @c obstack_copy @mtsrace:obstack-ptr @acucorrupt @acsmem @c obstack_grow dup @mtsrace:obstack-ptr @acucorrupt @acsmem @c obstack_finish dup @mtsrace:obstack-ptr @acucorrupt This allocates a block and initializes it by copying @var{size} bytes of data starting at @var{address}. It calls @code{obstack_alloc_failed_handler} if allocation of memory by @code{obstack_chunk_alloc} failed. @end deftypefun @comment obstack.h @comment GNU @deftypefun {void *} obstack_copy0 (struct obstack *@var{obstack-ptr}, void *@var{address}, int @var{size}) @safety{@prelim{}@mtsafe{@mtsrace{:obstack-ptr}}@assafe{}@acunsafe{@acucorrupt{} @acsmem{}}} @c obstack_copy0 @mtsrace:obstack-ptr @acucorrupt @acsmem @c obstack_grow0 dup @mtsrace:obstack-ptr @acucorrupt @acsmem @c obstack_finish dup @mtsrace:obstack-ptr @acucorrupt Like @code{obstack_copy}, but appends an extra byte containing a null character. This extra byte is not counted in the argument @var{size}. @end deftypefun The @code{obstack_copy0} function is convenient for copying a sequence of characters into an obstack as a null-terminated string. Here is an example of its use: @smallexample char * obstack_savestring (char *addr, int size) @{ return obstack_copy0 (&myobstack, addr, size); @} @end smallexample @noindent Contrast this with the previous example of @code{savestring} using @code{malloc} (@pxref{Basic Allocation}). @node Freeing Obstack Objects @subsubsection Freeing Objects in an Obstack @cindex freeing (obstacks) To free an object allocated in an obstack, use the function @code{obstack_free}. Since the obstack is a stack of objects, freeing one object automatically frees all other objects allocated more recently in the same obstack. @comment obstack.h @comment GNU @deftypefun void obstack_free (struct obstack *@var{obstack-ptr}, void *@var{object}) @safety{@prelim{}@mtsafe{@mtsrace{:obstack-ptr}}@assafe{}@acunsafe{@acucorrupt{}}} @c obstack_free @mtsrace:obstack-ptr @acucorrupt @c (obstack_free) @mtsrace:obstack-ptr @acucorrupt @c *freefun dup user-supplied If @var{object} is a null pointer, everything allocated in the obstack is freed. Otherwise, @var{object} must be the address of an object allocated in the obstack. Then @var{object} is freed, along with everything allocated in @var{obstack} since @var{object}. @end deftypefun Note that if @var{object} is a null pointer, the result is an uninitialized obstack. To free all memory in an obstack but leave it valid for further allocation, call @code{obstack_free} with the address of the first object allocated on the obstack: @smallexample obstack_free (obstack_ptr, first_object_allocated_ptr); @end smallexample Recall that the objects in an obstack are grouped into chunks. When all the objects in a chunk become free, the obstack library automatically frees the chunk (@pxref{Preparing for Obstacks}). Then other obstacks, or non-obstack allocation, can reuse the space of the chunk. @node Obstack Functions @subsubsection Obstack Functions and Macros @cindex macros The interfaces for using obstacks may be defined either as functions or as macros, depending on the compiler. The obstack facility works with all C compilers, including both @w{ISO C} and traditional C, but there are precautions you must take if you plan to use compilers other than GNU C. If you are using an old-fashioned @w{non-ISO C} compiler, all the obstack ``functions'' are actually defined only as macros. You can call these macros like functions, but you cannot use them in any other way (for example, you cannot take their address). Calling the macros requires a special precaution: namely, the first operand (the obstack pointer) may not contain any side effects, because it may be computed more than once. For example, if you write this: @smallexample obstack_alloc (get_obstack (), 4); @end smallexample @noindent you will find that @code{get_obstack} may be called several times. If you use @code{*obstack_list_ptr++} as the obstack pointer argument, you will get very strange results since the incrementation may occur several times. In @w{ISO C}, each function has both a macro definition and a function definition. The function definition is used if you take the address of the function without calling it. An ordinary call uses the macro definition by default, but you can request the function definition instead by writing the function name in parentheses, as shown here: @smallexample char *x; void *(*funcp) (); /* @r{Use the macro}. */ x = (char *) obstack_alloc (obptr, size); /* @r{Call the function}. */ x = (char *) (obstack_alloc) (obptr, size); /* @r{Take the address of the function}. */ funcp = obstack_alloc; @end smallexample @noindent This is the same situation that exists in @w{ISO C} for the standard library functions. @xref{Macro Definitions}. @strong{Warning:} When you do use the macros, you must observe the precaution of avoiding side effects in the first operand, even in @w{ISO C}. If you use the GNU C compiler, this precaution is not necessary, because various language extensions in GNU C permit defining the macros so as to compute each argument only once. @node Growing Objects @subsubsection Growing Objects @cindex growing objects (in obstacks) @cindex changing the size of a block (obstacks) Because memory in obstack chunks is used sequentially, it is possible to build up an object step by step, adding one or more bytes at a time to the end of the object. With this technique, you do not need to know how much data you will put in the object until you come to the end of it. We call this the technique of @dfn{growing objects}. The special functions for adding data to the growing object are described in this section. You don't need to do anything special when you start to grow an object. Using one of the functions to add data to the object automatically starts it. However, it is necessary to say explicitly when the object is finished. This is done with the function @code{obstack_finish}. The actual address of the object thus built up is not known until the object is finished. Until then, it always remains possible that you will add so much data that the object must be copied into a new chunk. While the obstack is in use for a growing object, you cannot use it for ordinary allocation of another object. If you try to do so, the space already added to the growing object will become part of the other object. @comment obstack.h @comment GNU @deftypefun void obstack_blank (struct obstack *@var{obstack-ptr}, int @var{size}) @safety{@prelim{}@mtsafe{@mtsrace{:obstack-ptr}}@assafe{}@acunsafe{@acucorrupt{} @acsmem{}}} @c obstack_blank @mtsrace:obstack-ptr @acucorrupt @acsmem @c _obstack_newchunk @mtsrace:obstack-ptr @acucorrupt @acsmem @c *chunkfun dup @acsmem @c *obstack_alloc_failed_handler dup user-supplied @c *freefun @c obstack_blank_fast dup @mtsrace:obstack-ptr The most basic function for adding to a growing object is @code{obstack_blank}, which adds space without initializing it. @end deftypefun @comment obstack.h @comment GNU @deftypefun void obstack_grow (struct obstack *@var{obstack-ptr}, void *@var{data}, int @var{size}) @safety{@prelim{}@mtsafe{@mtsrace{:obstack-ptr}}@assafe{}@acunsafe{@acucorrupt{} @acsmem{}}} @c obstack_grow @mtsrace:obstack-ptr @acucorrupt @acsmem @c _obstack_newchunk dup @mtsrace:obstack-ptr @acucorrupt @acsmem @c memcpy ok To add a block of initialized space, use @code{obstack_grow}, which is the growing-object analogue of @code{obstack_copy}. It adds @var{size} bytes of data to the growing object, copying the contents from @var{data}. @end deftypefun @comment obstack.h @comment GNU @deftypefun void obstack_grow0 (struct obstack *@var{obstack-ptr}, void *@var{data}, int @var{size}) @safety{@prelim{}@mtsafe{@mtsrace{:obstack-ptr}}@assafe{}@acunsafe{@acucorrupt{} @acsmem{}}} @c obstack_grow0 @mtsrace:obstack-ptr @acucorrupt @acsmem @c (no sequence point between storing NUL and incrementing next_free) @c (multiple changes to next_free => @acucorrupt) @c _obstack_newchunk dup @mtsrace:obstack-ptr @acucorrupt @acsmem @c memcpy ok This is the growing-object analogue of @code{obstack_copy0}. It adds @var{size} bytes copied from @var{data}, followed by an additional null character. @end deftypefun @comment obstack.h @comment GNU @deftypefun void obstack_1grow (struct obstack *@var{obstack-ptr}, char @var{c}) @safety{@prelim{}@mtsafe{@mtsrace{:obstack-ptr}}@assafe{}@acunsafe{@acucorrupt{} @acsmem{}}} @c obstack_1grow @mtsrace:obstack-ptr @acucorrupt @acsmem @c _obstack_newchunk dup @mtsrace:obstack-ptr @acucorrupt @acsmem @c obstack_1grow_fast dup @mtsrace:obstack-ptr @acucorrupt @acsmem To add one character at a time, use the function @code{obstack_1grow}. It adds a single byte containing @var{c} to the growing object. @end deftypefun @comment obstack.h @comment GNU @deftypefun void obstack_ptr_grow (struct obstack *@var{obstack-ptr}, void *@var{data}) @safety{@prelim{}@mtsafe{@mtsrace{:obstack-ptr}}@assafe{}@acunsafe{@acucorrupt{} @acsmem{}}} @c obstack_ptr_grow @mtsrace:obstack-ptr @acucorrupt @acsmem @c _obstack_newchunk dup @mtsrace:obstack-ptr @acucorrupt @acsmem @c obstack_ptr_grow_fast dup @mtsrace:obstack-ptr Adding the value of a pointer one can use the function @code{obstack_ptr_grow}. It adds @code{sizeof (void *)} bytes containing the value of @var{data}. @end deftypefun @comment obstack.h @comment GNU @deftypefun void obstack_int_grow (struct obstack *@var{obstack-ptr}, int @var{data}) @safety{@prelim{}@mtsafe{@mtsrace{:obstack-ptr}}@assafe{}@acunsafe{@acucorrupt{} @acsmem{}}} @c obstack_int_grow @mtsrace:obstack-ptr @acucorrupt @acsmem @c _obstack_newchunk dup @mtsrace:obstack-ptr @acucorrupt @acsmem @c obstack_int_grow_fast dup @mtsrace:obstack-ptr A single value of type @code{int} can be added by using the @code{obstack_int_grow} function. It adds @code{sizeof (int)} bytes to the growing object and initializes them with the value of @var{data}. @end deftypefun @comment obstack.h @comment GNU @deftypefun {void *} obstack_finish (struct obstack *@var{obstack-ptr}) @safety{@prelim{}@mtsafe{@mtsrace{:obstack-ptr}}@assafe{}@acunsafe{@acucorrupt{}}} @c obstack_finish @mtsrace:obstack-ptr @acucorrupt When you are finished growing the object, use the function @code{obstack_finish} to close it off and return its final address. Once you have finished the object, the obstack is available for ordinary allocation or for growing another object. This function can return a null pointer under the same conditions as @code{obstack_alloc} (@pxref{Allocation in an Obstack}). @end deftypefun When you build an object by growing it, you will probably need to know afterward how long it became. You need not keep track of this as you grow the object, because you can find out the length from the obstack just before finishing the object with the function @code{obstack_object_size}, declared as follows: @comment obstack.h @comment GNU @deftypefun int obstack_object_size (struct obstack *@var{obstack-ptr}) @safety{@prelim{}@mtsafe{@mtsrace{:obstack-ptr}}@assafe{}@acsafe{}} This function returns the current size of the growing object, in bytes. Remember to call this function @emph{before} finishing the object. After it is finished, @code{obstack_object_size} will return zero. @end deftypefun If you have started growing an object and wish to cancel it, you should finish it and then free it, like this: @smallexample obstack_free (obstack_ptr, obstack_finish (obstack_ptr)); @end smallexample @noindent This has no effect if no object was growing. @cindex shrinking objects You can use @code{obstack_blank} with a negative size argument to make the current object smaller. Just don't try to shrink it beyond zero length---there's no telling what will happen if you do that. @node Extra Fast Growing @subsubsection Extra Fast Growing Objects @cindex efficiency and obstacks The usual functions for growing objects incur overhead for checking whether there is room for the new growth in the current chunk. If you are frequently constructing objects in small steps of growth, this overhead can be significant. You can reduce the overhead by using special ``fast growth'' functions that grow the object without checking. In order to have a robust program, you must do the checking yourself. If you do this checking in the simplest way each time you are about to add data to the object, you have not saved anything, because that is what the ordinary growth functions do. But if you can arrange to check less often, or check more efficiently, then you make the program faster. The function @code{obstack_room} returns the amount of room available in the current chunk. It is declared as follows: @comment obstack.h @comment GNU @deftypefun int obstack_room (struct obstack *@var{obstack-ptr}) @safety{@prelim{}@mtsafe{@mtsrace{:obstack-ptr}}@assafe{}@acsafe{}} This returns the number of bytes that can be added safely to the current growing object (or to an object about to be started) in obstack @var{obstack} using the fast growth functions. @end deftypefun While you know there is room, you can use these fast growth functions for adding data to a growing object: @comment obstack.h @comment GNU @deftypefun void obstack_1grow_fast (struct obstack *@var{obstack-ptr}, char @var{c}) @safety{@prelim{}@mtsafe{@mtsrace{:obstack-ptr}}@assafe{}@acunsafe{@acucorrupt{} @acsmem{}}} @c obstack_1grow_fast @mtsrace:obstack-ptr @acucorrupt @acsmem @c (no sequence point between copying c and incrementing next_free) The function @code{obstack_1grow_fast} adds one byte containing the character @var{c} to the growing object in obstack @var{obstack-ptr}. @end deftypefun @comment obstack.h @comment GNU @deftypefun void obstack_ptr_grow_fast (struct obstack *@var{obstack-ptr}, void *@var{data}) @safety{@prelim{}@mtsafe{@mtsrace{:obstack-ptr}}@assafe{}@acsafe{}} @c obstack_ptr_grow_fast @mtsrace:obstack-ptr The function @code{obstack_ptr_grow_fast} adds @code{sizeof (void *)} bytes containing the value of @var{data} to the growing object in obstack @var{obstack-ptr}. @end deftypefun @comment obstack.h @comment GNU @deftypefun void obstack_int_grow_fast (struct obstack *@var{obstack-ptr}, int @var{data}) @safety{@prelim{}@mtsafe{@mtsrace{:obstack-ptr}}@assafe{}@acsafe{}} @c obstack_int_grow_fast @mtsrace:obstack-ptr The function @code{obstack_int_grow_fast} adds @code{sizeof (int)} bytes containing the value of @var{data} to the growing object in obstack @var{obstack-ptr}. @end deftypefun @comment obstack.h @comment GNU @deftypefun void obstack_blank_fast (struct obstack *@var{obstack-ptr}, int @var{size}) @safety{@prelim{}@mtsafe{@mtsrace{:obstack-ptr}}@assafe{}@acsafe{}} @c obstack_blank_fast @mtsrace:obstack-ptr The function @code{obstack_blank_fast} adds @var{size} bytes to the growing object in obstack @var{obstack-ptr} without initializing them. @end deftypefun When you check for space using @code{obstack_room} and there is not enough room for what you want to add, the fast growth functions are not safe. In this case, simply use the corresponding ordinary growth function instead. Very soon this will copy the object to a new chunk; then there will be lots of room available again. So, each time you use an ordinary growth function, check afterward for sufficient space using @code{obstack_room}. Once the object is copied to a new chunk, there will be plenty of space again, so the program will start using the fast growth functions again. Here is an example: @smallexample @group void add_string (struct obstack *obstack, const char *ptr, int len) @{ while (len > 0) @{ int room = obstack_room (obstack); if (room == 0) @{ /* @r{Not enough room. Add one character slowly,} @r{which may copy to a new chunk and make room.} */ obstack_1grow (obstack, *ptr++); len--; @} else @{ if (room > len) room = len; /* @r{Add fast as much as we have room for.} */ len -= room; while (room-- > 0) obstack_1grow_fast (obstack, *ptr++); @} @} @} @end group @end smallexample @node Status of an Obstack @subsubsection Status of an Obstack @cindex obstack status @cindex status of obstack Here are functions that provide information on the current status of allocation in an obstack. You can use them to learn about an object while still growing it. @comment obstack.h @comment GNU @deftypefun {void *} obstack_base (struct obstack *@var{obstack-ptr}) @safety{@prelim{}@mtsafe{}@asunsafe{@asucorrupt{}}@acsafe{}} This function returns the tentative address of the beginning of the currently growing object in @var{obstack-ptr}. If you finish the object immediately, it will have that address. If you make it larger first, it may outgrow the current chunk---then its address will change! If no object is growing, this value says where the next object you allocate will start (once again assuming it fits in the current chunk). @end deftypefun @comment obstack.h @comment GNU @deftypefun {void *} obstack_next_free (struct obstack *@var{obstack-ptr}) @safety{@prelim{}@mtsafe{}@asunsafe{@asucorrupt{}}@acsafe{}} This function returns the address of the first free byte in the current chunk of obstack @var{obstack-ptr}. This is the end of the currently growing object. If no object is growing, @code{obstack_next_free} returns the same value as @code{obstack_base}. @end deftypefun @comment obstack.h @comment GNU @deftypefun int obstack_object_size (struct obstack *@var{obstack-ptr}) @c dup @safety{@prelim{}@mtsafe{@mtsrace{:obstack-ptr}}@assafe{}@acsafe{}} This function returns the size in bytes of the currently growing object. This is equivalent to @smallexample obstack_next_free (@var{obstack-ptr}) - obstack_base (@var{obstack-ptr}) @end smallexample @end deftypefun @node Obstacks Data Alignment @subsubsection Alignment of Data in Obstacks @cindex alignment (in obstacks) Each obstack has an @dfn{alignment boundary}; each object allocated in the obstack automatically starts on an address that is a multiple of the specified boundary. By default, this boundary is aligned so that the object can hold any type of data. To access an obstack's alignment boundary, use the macro @code{obstack_alignment_mask}, whose function prototype looks like this: @comment obstack.h @comment GNU @deftypefn Macro int obstack_alignment_mask (struct obstack *@var{obstack-ptr}) @safety{@prelim{}@mtsafe{}@assafe{}@acsafe{}} The value is a bit mask; a bit that is 1 indicates that the corresponding bit in the address of an object should be 0. The mask value should be one less than a power of 2; the effect is that all object addresses are multiples of that power of 2. The default value of the mask is a value that allows aligned objects to hold any type of data: for example, if its value is 3, any type of data can be stored at locations whose addresses are multiples of 4. A mask value of 0 means an object can start on any multiple of 1 (that is, no alignment is required). The expansion of the macro @code{obstack_alignment_mask} is an lvalue, so you can alter the mask by assignment. For example, this statement: @smallexample obstack_alignment_mask (obstack_ptr) = 0; @end smallexample @noindent has the effect of turning off alignment processing in the specified obstack. @end deftypefn Note that a change in alignment mask does not take effect until @emph{after} the next time an object is allocated or finished in the obstack. If you are not growing an object, you can make the new alignment mask take effect immediately by calling @code{obstack_finish}. This will finish a zero-length object and then do proper alignment for the next object. @node Obstack Chunks @subsubsection Obstack Chunks @cindex efficiency of chunks @cindex chunks Obstacks work by allocating space for themselves in large chunks, and then parceling out space in the chunks to satisfy your requests. Chunks are normally 4096 bytes long unless you specify a different chunk size. The chunk size includes 8 bytes of overhead that are not actually used for storing objects. Regardless of the specified size, longer chunks will be allocated when necessary for long objects. The obstack library allocates chunks by calling the function @code{obstack_chunk_alloc}, which you must define. When a chunk is no longer needed because you have freed all the objects in it, the obstack library frees the chunk by calling @code{obstack_chunk_free}, which you must also define. These two must be defined (as macros) or declared (as functions) in each source file that uses @code{obstack_init} (@pxref{Creating Obstacks}). Most often they are defined as macros like this: @smallexample #define obstack_chunk_alloc malloc #define obstack_chunk_free free @end smallexample Note that these are simple macros (no arguments). Macro definitions with arguments will not work! It is necessary that @code{obstack_chunk_alloc} or @code{obstack_chunk_free}, alone, expand into a function name if it is not itself a function name. If you allocate chunks with @code{malloc}, the chunk size should be a power of 2. The default chunk size, 4096, was chosen because it is long enough to satisfy many typical requests on the obstack yet short enough not to waste too much memory in the portion of the last chunk not yet used. @comment obstack.h @comment GNU @deftypefn Macro int obstack_chunk_size (struct obstack *@var{obstack-ptr}) @safety{@prelim{}@mtsafe{}@assafe{}@acsafe{}} This returns the chunk size of the given obstack. @end deftypefn Since this macro expands to an lvalue, you can specify a new chunk size by assigning it a new value. Doing so does not affect the chunks already allocated, but will change the size of chunks allocated for that particular obstack in the future. It is unlikely to be useful to make the chunk size smaller, but making it larger might improve efficiency if you are allocating many objects whose size is comparable to the chunk size. Here is how to do so cleanly: @smallexample if (obstack_chunk_size (obstack_ptr) < @var{new-chunk-size}) obstack_chunk_size (obstack_ptr) = @var{new-chunk-size}; @end smallexample @node Summary of Obstacks @subsubsection Summary of Obstack Functions Here is a summary of all the functions associated with obstacks. Each takes the address of an obstack (@code{struct obstack *}) as its first argument. @table @code @item void obstack_init (struct obstack *@var{obstack-ptr}) Initialize use of an obstack. @xref{Creating Obstacks}. @item void *obstack_alloc (struct obstack *@var{obstack-ptr}, int @var{size}) Allocate an object of @var{size} uninitialized bytes. @xref{Allocation in an Obstack}. @item void *obstack_copy (struct obstack *@var{obstack-ptr}, void *@var{address}, int @var{size}) Allocate an object of @var{size} bytes, with contents copied from @var{address}. @xref{Allocation in an Obstack}. @item void *obstack_copy0 (struct obstack *@var{obstack-ptr}, void *@var{address}, int @var{size}) Allocate an object of @var{size}+1 bytes, with @var{size} of them copied from @var{address}, followed by a null character at the end. @xref{Allocation in an Obstack}. @item void obstack_free (struct obstack *@var{obstack-ptr}, void *@var{object}) Free @var{object} (and everything allocated in the specified obstack more recently than @var{object}). @xref{Freeing Obstack Objects}. @item void obstack_blank (struct obstack *@var{obstack-ptr}, int @var{size}) Add @var{size} uninitialized bytes to a growing object. @xref{Growing Objects}. @item void obstack_grow (struct obstack *@var{obstack-ptr}, void *@var{address}, int @var{size}) Add @var{size} bytes, copied from @var{address}, to a growing object. @xref{Growing Objects}. @item void obstack_grow0 (struct obstack *@var{obstack-ptr}, void *@var{address}, int @var{size}) Add @var{size} bytes, copied from @var{address}, to a growing object, and then add another byte containing a null character. @xref{Growing Objects}. @item void obstack_1grow (struct obstack *@var{obstack-ptr}, char @var{data-char}) Add one byte containing @var{data-char} to a growing object. @xref{Growing Objects}. @item void *obstack_finish (struct obstack *@var{obstack-ptr}) Finalize the object that is growing and return its permanent address. @xref{Growing Objects}. @item int obstack_object_size (struct obstack *@var{obstack-ptr}) Get the current size of the currently growing object. @xref{Growing Objects}. @item void obstack_blank_fast (struct obstack *@var{obstack-ptr}, int @var{size}) Add @var{size} uninitialized bytes to a growing object without checking that there is enough room. @xref{Extra Fast Growing}. @item void obstack_1grow_fast (struct obstack *@var{obstack-ptr}, char @var{data-char}) Add one byte containing @var{data-char} to a growing object without checking that there is enough room. @xref{Extra Fast Growing}. @item int obstack_room (struct obstack *@var{obstack-ptr}) Get the amount of room now available for growing the current object. @xref{Extra Fast Growing}. @item int obstack_alignment_mask (struct obstack *@var{obstack-ptr}) The mask used for aligning the beginning of an object. This is an lvalue. @xref{Obstacks Data Alignment}. @item int obstack_chunk_size (struct obstack *@var{obstack-ptr}) The size for allocating chunks. This is an lvalue. @xref{Obstack Chunks}. @item void *obstack_base (struct obstack *@var{obstack-ptr}) Tentative starting address of the currently growing object. @xref{Status of an Obstack}. @item void *obstack_next_free (struct obstack *@var{obstack-ptr}) Address just after the end of the currently growing object. @xref{Status of an Obstack}. @end table @node Variable Size Automatic @subsection Automatic Storage with Variable Size @cindex automatic freeing @cindex @code{alloca} function @cindex automatic storage with variable size The function @code{alloca} supports a kind of half-dynamic allocation in which blocks are allocated dynamically but freed automatically. Allocating a block with @code{alloca} is an explicit action; you can allocate as many blocks as you wish, and compute the size at run time. But all the blocks are freed when you exit the function that @code{alloca} was called from, just as if they were automatic variables declared in that function. There is no way to free the space explicitly. The prototype for @code{alloca} is in @file{stdlib.h}. This function is a BSD extension. @pindex stdlib.h @comment stdlib.h @comment GNU, BSD @deftypefun {void *} alloca (size_t @var{size}) @safety{@prelim{}@mtsafe{}@assafe{}@acsafe{}} The return value of @code{alloca} is the address of a block of @var{size} bytes of memory, allocated in the stack frame of the calling function. @end deftypefun Do not use @code{alloca} inside the arguments of a function call---you will get unpredictable results, because the stack space for the @code{alloca} would appear on the stack in the middle of the space for the function arguments. An example of what to avoid is @code{foo (x, alloca (4), y)}. @c This might get fixed in future versions of GCC, but that won't make @c it safe with compilers generally. @menu * Alloca Example:: Example of using @code{alloca}. * Advantages of Alloca:: Reasons to use @code{alloca}. * Disadvantages of Alloca:: Reasons to avoid @code{alloca}. * GNU C Variable-Size Arrays:: Only in GNU C, here is an alternative method of allocating dynamically and freeing automatically. @end menu @node Alloca Example @subsubsection @code{alloca} Example As an example of the use of @code{alloca}, here is a function that opens a file name made from concatenating two argument strings, and returns a file descriptor or minus one signifying failure: @smallexample int open2 (char *str1, char *str2, int flags, int mode) @{ char *name = (char *) alloca (strlen (str1) + strlen (str2) + 1); stpcpy (stpcpy (name, str1), str2); return open (name, flags, mode); @} @end smallexample @noindent Here is how you would get the same results with @code{malloc} and @code{free}: @smallexample int open2 (char *str1, char *str2, int flags, int mode) @{ char *name = (char *) malloc (strlen (str1) + strlen (str2) + 1); int desc; if (name == 0) fatal ("virtual memory exceeded"); stpcpy (stpcpy (name, str1), str2); desc = open (name, flags, mode); free (name); return desc; @} @end smallexample As you can see, it is simpler with @code{alloca}. But @code{alloca} has other, more important advantages, and some disadvantages. @node Advantages of Alloca @subsubsection Advantages of @code{alloca} Here are the reasons why @code{alloca} may be preferable to @code{malloc}: @itemize @bullet @item Using @code{alloca} wastes very little space and is very fast. (It is open-coded by the GNU C compiler.) @item Since @code{alloca} does not have separate pools for different sizes of block, space used for any size block can be reused for any other size. @code{alloca} does not cause memory fragmentation. @item @cindex longjmp Nonlocal exits done with @code{longjmp} (@pxref{Non-Local Exits}) automatically free the space allocated with @code{alloca} when they exit through the function that called @code{alloca}. This is the most important reason to use @code{alloca}. To illustrate this, suppose you have a function @code{open_or_report_error} which returns a descriptor, like @code{open}, if it succeeds, but does not return to its caller if it fails. If the file cannot be opened, it prints an error message and jumps out to the command level of your program using @code{longjmp}. Let's change @code{open2} (@pxref{Alloca Example}) to use this subroutine:@refill @smallexample int open2 (char *str1, char *str2, int flags, int mode) @{ char *name = (char *) alloca (strlen (str1) + strlen (str2) + 1); stpcpy (stpcpy (name, str1), str2); return open_or_report_error (name, flags, mode); @} @end smallexample @noindent Because of the way @code{alloca} works, the memory it allocates is freed even when an error occurs, with no special effort required. By contrast, the previous definition of @code{open2} (which uses @code{malloc} and @code{free}) would develop a memory leak if it were changed in this way. Even if you are willing to make more changes to fix it, there is no easy way to do so. @end itemize @node Disadvantages of Alloca @subsubsection Disadvantages of @code{alloca} @cindex @code{alloca} disadvantages @cindex disadvantages of @code{alloca} These are the disadvantages of @code{alloca} in comparison with @code{malloc}: @itemize @bullet @item If you try to allocate more memory than the machine can provide, you don't get a clean error message. Instead you get a fatal signal like the one you would get from an infinite recursion; probably a segmentation violation (@pxref{Program Error Signals}). @item Some @nongnusystems{} fail to support @code{alloca}, so it is less portable. However, a slower emulation of @code{alloca} written in C is available for use on systems with this deficiency. @end itemize @node GNU C Variable-Size Arrays @subsubsection GNU C Variable-Size Arrays @cindex variable-sized arrays In GNU C, you can replace most uses of @code{alloca} with an array of variable size. Here is how @code{open2} would look then: @smallexample int open2 (char *str1, char *str2, int flags, int mode) @{ char name[strlen (str1) + strlen (str2) + 1]; stpcpy (stpcpy (name, str1), str2); return open (name, flags, mode); @} @end smallexample But @code{alloca} is not always equivalent to a variable-sized array, for several reasons: @itemize @bullet @item A variable size array's space is freed at the end of the scope of the name of the array. The space allocated with @code{alloca} remains until the end of the function. @item It is possible to use @code{alloca} within a loop, allocating an additional block on each iteration. This is impossible with variable-sized arrays. @end itemize @strong{NB:} If you mix use of @code{alloca} and variable-sized arrays within one function, exiting a scope in which a variable-sized array was declared frees all blocks allocated with @code{alloca} during the execution of that scope. @node Resizing the Data Segment @section Resizing the Data Segment The symbols in this section are declared in @file{unistd.h}. You will not normally use the functions in this section, because the functions described in @ref{Memory Allocation} are easier to use. Those are interfaces to a @glibcadj{} memory allocator that uses the functions below itself. The functions below are simple interfaces to system calls. @comment unistd.h @comment BSD @deftypefun int brk (void *@var{addr}) @safety{@prelim{}@mtsafe{}@assafe{}@acsafe{}} @code{brk} sets the high end of the calling process' data segment to @var{addr}. The address of the end of a segment is defined to be the address of the last byte in the segment plus 1. The function has no effect if @var{addr} is lower than the low end of the data segment. (This is considered success, by the way). The function fails if it would cause the data segment to overlap another segment or exceed the process' data storage limit (@pxref{Limits on Resources}). The function is named for a common historical case where data storage and the stack are in the same segment. Data storage allocation grows upward from the bottom of the segment while the stack grows downward toward it from the top of the segment and the curtain between them is called the @dfn{break}. The return value is zero on success. On failure, the return value is @code{-1} and @code{errno} is set accordingly. The following @code{errno} values are specific to this function: @table @code @item ENOMEM The request would cause the data segment to overlap another segment or exceed the process' data storage limit. @end table @c The Brk system call in Linux (as opposed to the GNU C Library function) @c is considerably different. It always returns the new end of the data @c segment, whether it succeeds or fails. The GNU C library Brk determines @c it's a failure if and only if the system call returns an address less @c than the address requested. @end deftypefun @comment unistd.h @comment BSD @deftypefun void *sbrk (ptrdiff_t @var{delta}) @safety{@prelim{}@mtsafe{}@assafe{}@acsafe{}} This function is the same as @code{brk} except that you specify the new end of the data segment as an offset @var{delta} from the current end and on success the return value is the address of the resulting end of the data segment instead of zero. This means you can use @samp{sbrk(0)} to find out what the current end of the data segment is. @end deftypefun @node Locking Pages @section Locking Pages @cindex locking pages @cindex memory lock @cindex paging You can tell the system to associate a particular virtual memory page with a real page frame and keep it that way --- i.e., cause the page to be paged in if it isn't already and mark it so it will never be paged out and consequently will never cause a page fault. This is called @dfn{locking} a page. The functions in this chapter lock and unlock the calling process' pages. @menu * Why Lock Pages:: Reasons to read this section. * Locked Memory Details:: Everything you need to know locked memory * Page Lock Functions:: Here's how to do it. @end menu @node Why Lock Pages @subsection Why Lock Pages Because page faults cause paged out pages to be paged in transparently, a process rarely needs to be concerned about locking pages. However, there are two reasons people sometimes are: @itemize @bullet @item Speed. A page fault is transparent only insofar as the process is not sensitive to how long it takes to do a simple memory access. Time-critical processes, especially realtime processes, may not be able to wait or may not be able to tolerate variance in execution speed. @cindex realtime processing @cindex speed of execution A process that needs to lock pages for this reason probably also needs priority among other processes for use of the CPU. @xref{Priority}. In some cases, the programmer knows better than the system's demand paging allocator which pages should remain in real memory to optimize system performance. In this case, locking pages can help. @item Privacy. If you keep secrets in virtual memory and that virtual memory gets paged out, that increases the chance that the secrets will get out. If a password gets written out to disk swap space, for example, it might still be there long after virtual and real memory have been wiped clean. @end itemize Be aware that when you lock a page, that's one fewer page frame that can be used to back other virtual memory (by the same or other processes), which can mean more page faults, which means the system runs more slowly. In fact, if you lock enough memory, some programs may not be able to run at all for lack of real memory. @node Locked Memory Details @subsection Locked Memory Details A memory lock is associated with a virtual page, not a real frame. The paging rule is: If a frame backs at least one locked page, don't page it out. Memory locks do not stack. I.e., you can't lock a particular page twice so that it has to be unlocked twice before it is truly unlocked. It is either locked or it isn't. A memory lock persists until the process that owns the memory explicitly unlocks it. (But process termination and exec cause the virtual memory to cease to exist, which you might say means it isn't locked any more). Memory locks are not inherited by child processes. (But note that on a modern Unix system, immediately after a fork, the parent's and the child's virtual address space are backed by the same real page frames, so the child enjoys the parent's locks). @xref{Creating a Process}. Because of its ability to impact other processes, only the superuser can lock a page. Any process can unlock its own page. The system sets limits on the amount of memory a process can have locked and the amount of real memory it can have dedicated to it. @xref{Limits on Resources}. In Linux, locked pages aren't as locked as you might think. Two virtual pages that are not shared memory can nonetheless be backed by the same real frame. The kernel does this in the name of efficiency when it knows both virtual pages contain identical data, and does it even if one or both of the virtual pages are locked. But when a process modifies one of those pages, the kernel must get it a separate frame and fill it with the page's data. This is known as a @dfn{copy-on-write page fault}. It takes a small amount of time and in a pathological case, getting that frame may require I/O. @cindex copy-on-write page fault @cindex page fault, copy-on-write To make sure this doesn't happen to your program, don't just lock the pages. Write to them as well, unless you know you won't write to them ever. And to make sure you have pre-allocated frames for your stack, enter a scope that declares a C automatic variable larger than the maximum stack size you will need, set it to something, then return from its scope. @node Page Lock Functions @subsection Functions To Lock And Unlock Pages The symbols in this section are declared in @file{sys/mman.h}. These functions are defined by POSIX.1b, but their availability depends on your kernel. If your kernel doesn't allow these functions, they exist but always fail. They @emph{are} available with a Linux kernel. @strong{Portability Note:} POSIX.1b requires that when the @code{mlock} and @code{munlock} functions are available, the file @file{unistd.h} define the macro @code{_POSIX_MEMLOCK_RANGE} and the file @code{limits.h} define the macro @code{PAGESIZE} to be the size of a memory page in bytes. It requires that when the @code{mlockall} and @code{munlockall} functions are available, the @file{unistd.h} file define the macro @code{_POSIX_MEMLOCK}. @Theglibc{} conforms to this requirement. @comment sys/mman.h @comment POSIX.1b @deftypefun int mlock (const void *@var{addr}, size_t @var{len}) @safety{@prelim{}@mtsafe{}@assafe{}@acsafe{}} @code{mlock} locks a range of the calling process' virtual pages. The range of memory starts at address @var{addr} and is @var{len} bytes long. Actually, since you must lock whole pages, it is the range of pages that include any part of the specified range. When the function returns successfully, each of those pages is backed by (connected to) a real frame (is resident) and is marked to stay that way. This means the function may cause page-ins and have to wait for them. When the function fails, it does not affect the lock status of any pages. The return value is zero if the function succeeds. Otherwise, it is @code{-1} and @code{errno} is set accordingly. @code{errno} values specific to this function are: @table @code @item ENOMEM @itemize @bullet @item At least some of the specified address range does not exist in the calling process' virtual address space. @item The locking would cause the process to exceed its locked page limit. @end itemize @item EPERM The calling process is not superuser. @item EINVAL @var{len} is not positive. @item ENOSYS The kernel does not provide @code{mlock} capability. @end table You can lock @emph{all} a process' memory with @code{mlockall}. You unlock memory with @code{munlock} or @code{munlockall}. To avoid all page faults in a C program, you have to use @code{mlockall}, because some of the memory a program uses is hidden from the C code, e.g. the stack and automatic variables, and you wouldn't know what address to tell @code{mlock}. @end deftypefun @comment sys/mman.h @comment POSIX.1b @deftypefun int munlock (const void *@var{addr}, size_t @var{len}) @safety{@prelim{}@mtsafe{}@assafe{}@acsafe{}} @code{munlock} unlocks a range of the calling process' virtual pages. @code{munlock} is the inverse of @code{mlock} and functions completely analogously to @code{mlock}, except that there is no @code{EPERM} failure. @end deftypefun @comment sys/mman.h @comment POSIX.1b @deftypefun int mlockall (int @var{flags}) @safety{@prelim{}@mtsafe{}@assafe{}@acsafe{}} @code{mlockall} locks all the pages in a process' virtual memory address space, and/or any that are added to it in the future. This includes the pages of the code, data and stack segment, as well as shared libraries, user space kernel data, shared memory, and memory mapped files. @var{flags} is a string of single bit flags represented by the following macros. They tell @code{mlockall} which of its functions you want. All other bits must be zero. @table @code @item MCL_CURRENT Lock all pages which currently exist in the calling process' virtual address space. @item MCL_FUTURE Set a mode such that any pages added to the process' virtual address space in the future will be locked from birth. This mode does not affect future address spaces owned by the same process so exec, which replaces a process' address space, wipes out @code{MCL_FUTURE}. @xref{Executing a File}. @end table When the function returns successfully, and you specified @code{MCL_CURRENT}, all of the process' pages are backed by (connected to) real frames (they are resident) and are marked to stay that way. This means the function may cause page-ins and have to wait for them. When the process is in @code{MCL_FUTURE} mode because it successfully executed this function and specified @code{MCL_CURRENT}, any system call by the process that requires space be added to its virtual address space fails with @code{errno} = @code{ENOMEM} if locking the additional space would cause the process to exceed its locked page limit. In the case that the address space addition that can't be accommodated is stack expansion, the stack expansion fails and the kernel sends a @code{SIGSEGV} signal to the process. When the function fails, it does not affect the lock status of any pages or the future locking mode. The return value is zero if the function succeeds. Otherwise, it is @code{-1} and @code{errno} is set accordingly. @code{errno} values specific to this function are: @table @code @item ENOMEM @itemize @bullet @item At least some of the specified address range does not exist in the calling process' virtual address space. @item The locking would cause the process to exceed its locked page limit. @end itemize @item EPERM The calling process is not superuser. @item EINVAL Undefined bits in @var{flags} are not zero. @item ENOSYS The kernel does not provide @code{mlockall} capability. @end table You can lock just specific pages with @code{mlock}. You unlock pages with @code{munlockall} and @code{munlock}. @end deftypefun @comment sys/mman.h @comment POSIX.1b @deftypefun int munlockall (void) @safety{@prelim{}@mtsafe{}@assafe{}@acsafe{}} @code{munlockall} unlocks every page in the calling process' virtual address space and turn off @code{MCL_FUTURE} future locking mode. The return value is zero if the function succeeds. Otherwise, it is @code{-1} and @code{errno} is set accordingly. The only way this function can fail is for generic reasons that all functions and system calls can fail, so there are no specific @code{errno} values. @end deftypefun @ignore @c This was never actually implemented. -zw @node Relocating Allocator @section Relocating Allocator @cindex relocating memory allocator Any system of dynamic memory allocation has overhead: the amount of space it uses is more than the amount the program asks for. The @dfn{relocating memory allocator} achieves very low overhead by moving blocks in memory as necessary, on its own initiative. @c @menu @c * Relocator Concepts:: How to understand relocating allocation. @c * Using Relocator:: Functions for relocating allocation. @c @end menu @node Relocator Concepts @subsection Concepts of Relocating Allocation @ifinfo The @dfn{relocating memory allocator} achieves very low overhead by moving blocks in memory as necessary, on its own initiative. @end ifinfo When you allocate a block with @code{malloc}, the address of the block never changes unless you use @code{realloc} to change its size. Thus, you can safely store the address in various places, temporarily or permanently, as you like. This is not safe when you use the relocating memory allocator, because any and all relocatable blocks can move whenever you allocate memory in any fashion. Even calling @code{malloc} or @code{realloc} can move the relocatable blocks. @cindex handle For each relocatable block, you must make a @dfn{handle}---a pointer object in memory, designated to store the address of that block. The relocating allocator knows where each block's handle is, and updates the address stored there whenever it moves the block, so that the handle always points to the block. Each time you access the contents of the block, you should fetch its address anew from the handle. To call any of the relocating allocator functions from a signal handler is almost certainly incorrect, because the signal could happen at any time and relocate all the blocks. The only way to make this safe is to block the signal around any access to the contents of any relocatable block---not a convenient mode of operation. @xref{Nonreentrancy}. @node Using Relocator @subsection Allocating and Freeing Relocatable Blocks @pindex malloc.h In the descriptions below, @var{handleptr} designates the address of the handle. All the functions are declared in @file{malloc.h}; all are GNU extensions. @comment malloc.h @comment GNU @c @deftypefun {void *} r_alloc (void **@var{handleptr}, size_t @var{size}) This function allocates a relocatable block of size @var{size}. It stores the block's address in @code{*@var{handleptr}} and returns a non-null pointer to indicate success. If @code{r_alloc} can't get the space needed, it stores a null pointer in @code{*@var{handleptr}}, and returns a null pointer. @end deftypefun @comment malloc.h @comment GNU @c @deftypefun void r_alloc_free (void **@var{handleptr}) This function is the way to free a relocatable block. It frees the block that @code{*@var{handleptr}} points to, and stores a null pointer in @code{*@var{handleptr}} to show it doesn't point to an allocated block any more. @end deftypefun @comment malloc.h @comment GNU @c @deftypefun {void *} r_re_alloc (void **@var{handleptr}, size_t @var{size}) The function @code{r_re_alloc} adjusts the size of the block that @code{*@var{handleptr}} points to, making it @var{size} bytes long. It stores the address of the resized block in @code{*@var{handleptr}} and returns a non-null pointer to indicate success. If enough memory is not available, this function returns a null pointer and does not modify @code{*@var{handleptr}}. @end deftypefun @end ignore @ignore @comment No longer available... @comment @node Memory Warnings @comment @section Memory Usage Warnings @comment @cindex memory usage warnings @comment @cindex warnings of memory almost full @pindex malloc.c You can ask for warnings as the program approaches running out of memory space, by calling @code{memory_warnings}. This tells @code{malloc} to check memory usage every time it asks for more memory from the operating system. This is a GNU extension declared in @file{malloc.h}. @comment malloc.h @comment GNU @comment @deftypefun void memory_warnings (void *@var{start}, void (*@var{warn-func}) (const char *)) Call this function to request warnings for nearing exhaustion of virtual memory. The argument @var{start} says where data space begins, in memory. The allocator compares this against the last address used and against the limit of data space, to determine the fraction of available memory in use. If you supply zero for @var{start}, then a default value is used which is right in most circumstances. For @var{warn-func}, supply a function that @code{malloc} can call to warn you. It is called with a string (a warning message) as argument. Normally it ought to display the string for the user to read. @end deftypefun The warnings come when memory becomes 75% full, when it becomes 85% full, and when it becomes 95% full. Above 95% you get another warning each time memory usage increases. @end ignore