VirtualAlloc2 function (memoryapi.h)
Reserves, commits, or changes the state of a region of memory within the virtual address space of a specified process (allocated memory is initialized to zero).
Syntax
PVOID VirtualAlloc2(
[in, optional] HANDLE Process,
[in, optional] PVOID BaseAddress,
[in] SIZE_T Size,
[in] ULONG AllocationType,
[in] ULONG PageProtection,
[in, out, optional] MEM_EXTENDED_PARAMETER *ExtendedParameters,
[in] ULONG ParameterCount
);
Parameters
[in, optional] Process
The handle to a process. The function allocates memory within the virtual address space of this process.
The handle must have the PROCESS_VM_OPERATION access right. For more information, see Process Security and Access Rights.
If Process is NULL, the function allocates memory for the calling process.
[in, optional] BaseAddress
The pointer that specifies a desired starting address for the region of pages that you want to allocate.
If BaseAddress is NULL, the function determines where to allocate the region.
If BaseAddress is not NULL, then any provided MEM_ADDRESS_REQUIREMENTS structure must consist of all zeroes, and the base address must be a multiple of the system allocation granularity. To determine the allocation granularity, use the GetSystemInfo function.
If this address is within an enclave that you have not initialized by calling InitializeEnclave, VirtualAlloc2 allocates a page of zeros for the enclave at that address. The page must be previously uncommitted, and will not be measured with the EEXTEND instruction of the Intel Software Guard Extensions programming model.
If the address in within an enclave that you initialized, then the allocation operation fails with the ERROR_INVALID_ADDRESS error. That is true for enclaves that do not support dynamic memory management (i.e. SGX1). SGX2 enclaves will permit allocation, and the page must be accepted by the enclave after it has been allocated.
[in] Size
The size of the region of memory to allocate, in bytes.
The size must always be a multiple of the page size.
If BaseAddress is not NULL, the function allocates all pages that contain one or more bytes in the range from BaseAddress to BaseAddress+Size. This means, for example, that a 2-byte range that straddles a page boundary causes the function to allocate both pages.
[in] AllocationType
The type of memory allocation. This parameter must contain one of the following values.
| Value | Meaning |
|---|---|
|
Allocates memory charges (from the overall size of memory and the paging files on disk) for the specified
reserved memory pages. The function also guarantees that when the caller later initially accesses the memory,
the contents will be zero. Actual physical pages are not allocated unless/until the virtual addresses are
actually accessed.
To reserve and commit pages in one step, call
VirtualAlloc2 with
Attempting to commit a specific address range by specifying MEM_COMMIT without MEM_RESERVE and a non-NULL BaseAddress fails unless the entire range has already been reserved. The resulting error code is ERROR_INVALID_ADDRESS. An attempt to commit a page that is already committed does not cause the function to fail. This means that you can commit pages without first determining the current commitment state of each page. If BaseAddress specifies an address within an enclave, AllocationType must be MEM_COMMIT. |
|
Reserves a range of the process's virtual address space without allocating any actual physical storage in
memory or in the paging file on disk.
You commit reserved pages by calling
VirtualAlloc2 again with
MEM_COMMIT. To reserve and commit pages in one step, call
VirtualAlloc2 with
Other memory allocation functions, such as malloc and LocalAlloc, cannot use reserved memory until it has been released. |
|
Replaces a placeholder with a normal private allocation. Only data/pf-backed section views are supported (no images, physical memory, etc.). When you replace a placeholder, BaseAddress and Size must exactly match those of the placeholder,
and any provided MEM_ADDRESS_REQUIREMENTS structure must consist of all zeroes.
After you replace a placeholder with a private allocation, to free that allocation back to a placeholder, see the dwFreeType parameter of VirtualFree and VirtualFreeEx. A placeholder is a type of reserved memory region. |
|
To create a placeholder, call
VirtualAlloc2 with
MEM_RESERVE | MEM_RESERVE_PLACEHOLDER and PageProtection set to PAGE_NOACCESS. To free/split/coalesce a placeholder, see the dwFreeType parameter of VirtualFree and VirtualFreeEx.
A placeholder is a type of reserved memory region. |
|
Indicates that data in the memory range specified by BaseAddress and
Size is no longer of interest. The pages should not be read from or written to
the paging file. However, the memory block will be used again later, so it should not be decommitted. This
value cannot be used with any other value.
Using this value does not guarantee that the range operated on with MEM_RESET will contain zeros. If you want the range to contain zeros, decommit the memory and then recommit it. When you use MEM_RESET, the VirtualAlloc2 function ignores the value of fProtect. However, you must still set fProtect to a valid protection value, such as PAGE_NOACCESS. VirtualAlloc2 returns an error if you use MEM_RESET and the range of memory is mapped to a file. A shared view is only acceptable if it is mapped to a paging file. |
|
MEM_RESET_UNDO should only be called on an address range to which
MEM_RESET was successfully applied earlier. It indicates that the data in the
specified memory range specified by BaseAddress and Size
is of interest to the caller and attempts to reverse the effects of MEM_RESET. If
the function succeeds, that means all data in the specified address range is intact. If the function fails,
at least some of the data in the address range has been replaced with zeroes.
This value cannot be used with any other value. If MEM_RESET_UNDO is called on an address range which was not MEM_RESET earlier, the behavior is undefined. When you specify MEM_RESET, the VirtualAlloc2 function ignores the value of PageProtection. However, you must still set PageProtection to a valid protection value, such as PAGE_NOACCESS. Windows Server 2008 R2, Windows 7, Windows Server 2008, Windows Vista, Windows Server 2003 and Windows XP: The MEM_RESET_UNDO flag is not supported until Windows 8 and Windows Server 2012. |
This parameter can also specify the following values as indicated.
| Value | Meaning |
|---|---|
|
Allocates memory using large page support.
The size and alignment must be a multiple of the large-page minimum. To obtain this value, use the GetLargePageMinimum function. If you specify this value, you must also specify MEM_RESERVE and MEM_COMMIT. |
|
Reserves an address range that can be used to map
Address Windowing Extensions (AWE)
pages.
This value must be used with MEM_RESERVE and no other values. |
|
Allocates memory at the highest possible address. This can be slower than regular allocations, especially when there are many allocations. |
[in] PageProtection
The memory protection for the region of pages to be allocated. If the pages are being committed, you can specify any one of the memory protection constants.
If BaseAddress specifies an address within an enclave, PageProtection cannot be any of the following values:
- PAGE_NOACCESS
- PAGE_GUARD
- PAGE_NOCACHE
- PAGE_WRITECOMBINE
When allocating dynamic memory for an enclave, the PageProtection parameter must be PAGE_READWRITE or PAGE_EXECUTE_READWRITE.
[in, out, optional] ExtendedParameters
An optional pointer to one or more extended parameters of type MEM_EXTENDED_PARAMETER. Each of those extended parameter values can itself have a Type field of either MemExtendedParameterAddressRequirements or MemExtendedParameterNumaNode. If no MemExtendedParameterNumaNode extended parameter is provided, then the behavior is the same as for the VirtualAlloc/MapViewOfFile functions (that is, the preferred NUMA node for the physical pages is determined based on the ideal processor of the thread that first accesses the memory).
[in] ParameterCount
The number of extended parameters pointed to by ExtendedParameters.
Return value
If the function succeeds, the return value is the base address of the allocated region of pages.
If the function fails, the return value is NULL. To get extended error information, call GetLastError.
Remarks
This function lets you specify:
- a range of virtual address space and a power-of-2 alignment restriction for new allocations
- an arbitrary number of extended parameters
- a preferred NUMA node for the physical memory as an extended parameter (see the ExtendedParameters parameter)
- a placeholder operation (specifically, replacement).
This API provides specialized techniques for managing virtual memory in support of high-performance games and server applications. For example, placeholders allow a reserved memory range to be explicitly partitioned, overlaid, and re-mapped; this can be used to implement arbitrarily extendable regions or virtual memory ring buffers. VirtualAlloc2 also allows for allocating memory with a specific memory-alignment.
Each page has an associated page state. The VirtualAlloc2 function can perform the following operations:
- Commit a region of reserved pages
- Reserve a region of free pages
- Simultaneously reserve and commit a region of free pages
VirtualAlloc2 can commit pages that are already committed, but cannot reserve pages that are already reserved. This means you can commit a range of pages, regardless of whether they have already been committed, and the function will not fail. In general however, only a minimal range of mostly uncommitted pages should be specified, because committing a large number of pages that are already committed can cause the VirtualAlloc2 call to take much longer.
You can use VirtualAlloc2 to reserve a block of pages and then make additional calls to VirtualAlloc2 to commit individual pages from the reserved block. This enables a process to reserve a range of its virtual address space without consuming physical storage until it is needed.
If the lpAddress parameter is not NULL, the function uses the lpAddress and dwSize parameters to compute the region of pages to be allocated. The current state of the entire range of pages must be compatible with the type of allocation specified by the flAllocationType parameter. Otherwise, the function fails and none of the pages is allocated. This compatibility requirement does not preclude committing an already committed page; see the preceding list.
To execute dynamically generated code, use VirtualAlloc2 to allocate memory, and the VirtualProtectEx function to grant PAGE_EXECUTE access.
The VirtualAlloc2 function can be used to reserve an Address Windowing Extensions (AWE) region of memory within the virtual address space of a specified process. This region of memory can then be used to map physical pages into and out of virtual memory as required by the application. The MEM_PHYSICAL and MEM_RESERVE values must be set in the AllocationType parameter. The MEM_COMMIT value must not be set. The page protection must be set to PAGE_READWRITE.
The VirtualFreeEx function can decommit a committed page, releasing the page's storage, or it can simultaneously decommit and release a committed page. It can also release a reserved page, making it a free page.
When creating a region that will be executable, the calling program bears responsibility for ensuring cache coherency via an appropriate call to FlushInstructionCache once the code has been set in place. Otherwise attempts to execute code out of the newly executable region may produce unpredictable results.
Examples
Scenario 1. Create a circular buffer by mapping two adjacent views of the same shared memory section.
#include <windows.h>
#include <stdio.h>
#include <stdlib.h>
//
// This function creates a ring buffer by allocating a pagefile-backed section
// and mapping two views of that section next to each other. This way if the
// last record in the buffer wraps it can still be accessed in a linear fashion
// using its base VA.
//
void*
CreateRingBuffer (
unsigned int bufferSize,
_Outptr_ void** secondaryView
)
{
BOOL result;
HANDLE section = nullptr;
SYSTEM_INFO sysInfo;
void* ringBuffer = nullptr;
void* placeholder1 = nullptr;
void* placeholder2 = nullptr;
void* view1 = nullptr;
void* view2 = nullptr;
GetSystemInfo (&sysInfo);
if ((bufferSize % sysInfo.dwAllocationGranularity) != 0) {
return nullptr;
}
//
// Reserve a placeholder region where the buffer will be mapped.
//
placeholder1 = (PCHAR) VirtualAlloc2 (
nullptr,
nullptr,
2 * bufferSize,
MEM_RESERVE | MEM_RESERVE_PLACEHOLDER,
PAGE_NOACCESS,
nullptr, 0
);
if (placeholder1 == nullptr) {
printf ("VirtualAlloc2 failed, error %#x\n", GetLastError());
goto Exit;
}
//
// Split the placeholder region into two regions of equal size.
//
result = VirtualFree (
placeholder1,
bufferSize,
MEM_RELEASE | MEM_PRESERVE_PLACEHOLDER
);
if (result == FALSE) {
printf ("VirtualFreeEx failed, error %#x\n", GetLastError());
goto Exit;
}
placeholder2 = (void*) ((ULONG_PTR) placeholder1 + bufferSize);
//
// Create a pagefile-backed section for the buffer.
//
section = CreateFileMapping (
INVALID_HANDLE_VALUE,
nullptr,
PAGE_READWRITE,
0,
bufferSize, nullptr
);
if (section == nullptr) {
printf ("CreateFileMapping failed, error %#x\n", GetLastError());
goto Exit;
}
//
// Map the section into the first placeholder region.
//
view1 = MapViewOfFile3 (
section,
nullptr,
placeholder1,
0,
bufferSize,
MEM_REPLACE_PLACEHOLDER,
PAGE_READWRITE,
nullptr, 0
);
if (view1 == nullptr) {
printf ("MapViewOfFile3 failed, error %#x\n", GetLastError());
goto Exit;
}
//
// Ownership transferred, don’t free this now.
//
placeholder1 = nullptr;
//
// Map the section into the second placeholder region.
//
view2 = MapViewOfFile3 (
section,
nullptr,
placeholder2,
0,
bufferSize,
MEM_REPLACE_PLACEHOLDER,
PAGE_READWRITE,
nullptr, 0
);
if (view2 == nullptr) {
printf ("MapViewOfFile3 failed, error %#x\n", GetLastError());
goto Exit;
}
//
// Success, return both mapped views to the caller.
//
ringBuffer = view1;
*secondaryView = view2;
placeholder2 = nullptr;
view1 = nullptr;
view2 = nullptr;
Exit:
if (section != nullptr) {
CloseHandle (section);
}
if (placeholder1 != nullptr) {
VirtualFree (placeholder1, 0, MEM_RELEASE);
}
if (placeholder2 != nullptr) {
VirtualFree (placeholder2, 0, MEM_RELEASE);
}
if (view1 != nullptr) {
UnmapViewOfFileEx (view1, 0);
}
if (view2 != nullptr) {
UnmapViewOfFileEx (view2, 0);
}
return ringBuffer;
}
int __cdecl wmain()
{
char* ringBuffer;
void* secondaryView;
unsigned int bufferSize = 0x10000;
ringBuffer = (char*) CreateRingBuffer (bufferSize, &secondaryView);
if (ringBuffer == nullptr) {
printf ("CreateRingBuffer failed\n");
return 0;
}
//
// Make sure the buffer wraps properly.
//
ringBuffer[0] = 'a';
if (ringBuffer[bufferSize] == 'a') {
printf ("The buffer wraps as expected\n");
}
UnmapViewOfFile (ringBuffer);
UnmapViewOfFile (secondaryView);
}
Scenario 2. Specify a preferred NUMA node when allocating memory.
void*
AllocateWithPreferredNode (size_t size, unsigned int numaNode)
{
MEM_EXTENDED_PARAMETER param = {0};
param.Type = MemExtendedParameterNumaNode;
param.ULong = numaNode;
return VirtualAlloc2 (
nullptr, nullptr,
size,
MEM_RESERVE | MEM_COMMIT,
PAGE_READWRITE,
¶m, 1);
}
Scenario 3. Allocate memory in a specific virtual address range (below 4GB, in this example) and with specific alignment.
void*
AllocateAlignedBelow2GB (size_t size, size_t alignment)
{
MEM_ADDRESS_REQUIREMENTS addressReqs = {0};
MEM_EXTENDED_PARAMETER param = {0};
addressReqs.Alignment = alignment;
addressReqs.HighestEndingAddress = (PVOID)(ULONG_PTR) 0x7fffffff;
param.Type = MemExtendedParameterAddressRequirements;
param.Pointer = &addressReqs;
return VirtualAlloc2 (
nullptr, nullptr,
size,
MEM_RESERVE | MEM_COMMIT,
PAGE_READWRITE,
¶m, 1);
}
Requirements
| Requirement | Value |
|---|---|
| Minimum supported client | Windows 10 [desktop apps only] |
| Minimum supported server | Windows Server 2016 [desktop apps only] |
| Target Platform | Windows |
| Header | memoryapi.h (include Windows.h) |
| Library | onecore.lib |
| DLL | Kernel32.dll |
See also
Feedback
Coming soon: Throughout 2024 we will be phasing out GitHub Issues as the feedback mechanism for content and replacing it with a new feedback system. For more information see: https://aka.ms/ContentUserFeedback.
Submit and view feedback for