mremap

Standard C library (libc, -lc)

#define _GNU_SOURCE         /* See feature_test_macros(7) */
#include <sys/mman.h>

void *mremap(size_t old_size;
             void old_address[old_size], size_t old_size,
             size_t new_size, int flags, ...  /* void *new_address */);

mremap() expands (or shrinks) an existing memory mapping, potentially moving it at the same time (controlled by the flags argument and the available virtual address space).

Mappings can also simply be moved (without any resizing) by specifying equal old_size and new_size and using the MREMAP_FIXED flag (see below). Since Linux 6.17, while old_address must be mapped, old_size may span multiple mappings including unmapped areas between them when performing a simple move. The MREMAP_DONTUNMAP flag may also be specified.

Similarly, if the operation performs a shrink, that is, if old_size is greater than new_size, then old_size may also span multiple mappings, which do not have to be adjacent to one another. If this shrink is performed in-place, that is, neither MREMAP_FIXED, nor MREMAP_DONTUNMAP are specified, new_size may also span multiple VMAs. However, if the range is moved, then new_size must span only a single mapping.

If the operation is neither a MREMAP_FIXED move nor a shrink, then old_size must span only a single mapping.

old_address is the old address of the first virtual memory block that you want to expand, shrink, and/or move. Note that old_address has to be page aligned. old_size is the size of the range containing virtual memory blocks to be manipulated. new_size is the requested size of the virtual memory blocks after the resize. An optional fifth argument, new_address, may be provided; see the description of MREMAP_FIXED below.

If the value of old_size is zero, and old_address refers to a shareable mapping (see the description of MAP_SHARED in mmap(2)), then mremap() will create a new mapping of the same pages. new_size will be the size of the new mapping and the location of the new mapping may be specified with new_address; see the description of MREMAP_FIXED below. If a new mapping is requested via this method, then the MREMAP_MAYMOVE flag must also be specified.

The flags bit-mask argument may be 0, or include the following flags:

MREMAP_MAYMOVE

By default, if there is not sufficient space to expand a mapping at its current location, then mremap() fails. If this flag is specified, then the kernel is permitted to relocate the mapping to a new virtual address, if necessary. If the mapping is relocated, then absolute pointers into the old mapping location become invalid (offsets relative to the starting address of the mapping should be employed).

MREMAP_FIXED (since Linux 2.3.31)

This flag serves a similar purpose to the MAP_FIXED flag of mmap(2). If this flag is specified, then mremap() accepts a fifth argument, void *new_address, which specifies a page-aligned address to which the mapping must be moved. Any previous mapping at the address range specified by new_address and new_size is unmapped.

If MREMAP_FIXED is specified, then MREMAP_MAYMOVE must also be specified.

Since Linux 6.17, if old_size is equal to new_size and MREMAP_FIXED is specified, then old_size may span beyond the mapping in which old_address resides. In this case, gaps between mappings in the original range are maintained in the new range. The whole operation is performed atomically unless an error arises, in which case the operation may be partially completed, that is, some mappings may be moved and others not.

Moving multiple mappings is not permitted if any of those mappings have either been registered with userfaultfd(2), or map drivers that specify their own custom address mapping logic.

MREMAP_DONTUNMAP (since Linux 5.7)

This flag, which must be used in conjunction with MREMAP_MAYMOVE, remaps mappings to a new address but does not unmap them from their original address.

The MREMAP_DONTUNMAP flag can be used only with mappings that are not VM_DONTEXPAND or VM_PFNMAP. Before Linux 5.13, the MREMAP_DONTUNMAP flag could be used only with private anonymous mappings (see the description of MAP_PRIVATE and MAP_ANONYMOUS in mmap(2)).

After completion, any access to the range specified by old_address and old_size will result in a page fault. The page fault will be handled by a userfaultfd(2) handler if the address is in a range previously registered with userfaultfd(2). Otherwise, the kernel allocates a zero-filled page to handle the fault.

The MREMAP_DONTUNMAP flag may be used to atomically move a mapping while leaving the source mapped. See NOTES for some possible applications of MREMAP_DONTUNMAP.

If the memory segments specified by old_address and old_size are locked (using mlock(2) or similar), then this lock is maintained when the segments are resized and/or relocated. As a consequence, the amount of memory locked by the process may change.

On success mremap() returns a pointer to the new virtual memory area. On error, the value MAP_FAILED (that is, (void *) -1) is returned, and errno is set to indicate the error.

EAGAIN

The caller tried to expand a memory segment that is locked, but this was not possible without exceeding the RLIMIT_MEMLOCK resource limit.

EFAULT

Some address in the range old_address to old_address+old_size is an invalid virtual memory address for this process. You can also get EFAULT even if there exist mappings that cover the whole address space requested, but those mappings are of different types, and the mremap() operation being performed does not support this.

EINVAL

An invalid argument was given. Possible causes are:

ENOMEM

Not enough memory was available to complete the operation. Possible causes are:

Linux.

Prior to glibc 2.4, glibc did not expose the definition of MREMAP_FIXED, and the prototype for mremap() did not allow for the new_address argument.

mremap() changes the mapping between virtual addresses and memory pages. This can be used to implement a very efficient realloc(3).

In Linux, memory is divided into pages. A process has (one or) several linear virtual memory segments. Each virtual memory segment has one or more mappings to real memory pages (in the page table). Each virtual memory segment has its own protection (access rights), which may cause a segmentation violation (SIGSEGV) if the memory is accessed incorrectly (e.g., writing to a read-only segment). Accessing virtual memory outside of the segments will also cause a segmentation violation.

If mremap() is used to move or expand an area locked with mlock(2) or equivalent, the mremap() call will make a best effort to populate the new area but will not fail with ENOMEM if the area cannot be populated.

Before Linux 4.14, if old_size was zero and the mapping referred to by old_address was a private mapping (see the description of MAP_PRIVATE in mmap(2)), mremap() created a new private mapping unrelated to the original mapping. This behavior was unintended and probably unexpected in user-space applications (since the intention of mremap() is to create a new mapping based on the original mapping). Since Linux 4.14, mremap() fails with the error EINVAL in this scenario.

brk(2), getpagesize(2), getrlimit(2), mlock(2), mmap(2), sbrk(2), malloc(3), realloc(3)

Your favorite text book on operating systems for more information on paged memory (e.g., Modern Operating Systems by Andrew S. Tanenbaum, Inside Linux by Randolph Bentson, The Design of the UNIX Operating System by Maurice J. Bach)