reactos/ntoskrnl/mm/mminit.c

304 lines
9.6 KiB
C
Raw Normal View History

/*
* PROJECT: ReactOS Kernel
* LICENSE: GPL - See COPYING in the top level directory
* FILE: ntoskrnl/mm/mminit.c
* PURPOSE: Memory Manager Initialization
* PROGRAMMERS:
*/
/* INCLUDES ******************************************************************/
#include <ntoskrnl.h>
#define NDEBUG
#include <debug.h>
#define MODULE_INVOLVED_IN_ARM3
#include "ARM3/miarm.h"
/* GLOBALS *******************************************************************/
BOOLEAN Mm64BitPhysicalAddress = FALSE;
ULONG MmReadClusterSize;
//
// 0 | 1 is on/off paging, 2 is undocumented
//
UCHAR MmDisablePagingExecutive = 1; // Forced to off
PMMPTE MmSharedUserDataPte;
PMMSUPPORT MmKernelAddressSpace;
[NEWCC] A reintegration checkpoint for the NewCC branch, brought to you by Team NewCC. Differences with current ReactOS trunk: * A new memory area type, MEMORY_AREA_CACHE, is added, which represents a mapped region of a file. In NEWCC mode, user sections are MEMORY_AREA_CACHE type as well, and obey new semantics. In non-NEWCC mode, they aren't used. * A way of claiming a page entry for a specific thread's work is added. Placing the special SWAPENTRY value MM_WAIT_ENTRY in a page table, or in a section page table should indicate that memory management code is intended to wait for another thread to make some status change before checking the state of the page entry again. In code that uses this convention, a return value of STATUS_SUCCESS + 1 is used to indicate that the caller should use the MiWaitForPageEvent macro to wait until somebody has change the state of a wait entry before checking again. This is a lighter weight mechanism than PAGEOPs. * A way of asking the caller to perform some blocking operation without locks held is provided. This replaces some spaghettified code in which locks are repeatedly taken and broken by code that performs various blocking operations. Using this mechanism, it is possible to do a small amount of non-blocking work, fill in a request, then return STATUS_MORE_PROCESSING_REQUIRED to request that locks be dropped and the blocking operation be carried out. A MM_REQUIRED_RESOURCES structure is provided to consumers of this contract to use to accumulate state across many blocking operations. Several functions wrapping blocking operations are provided in ntoskrnl/cache/reqtools.c. * Image section pages are no longer direct mapped. This is done to simplify consolidation of ownership of pages under the data section system. At a later time, it may be possible to make data pages directly available to image sections for the same file. This is likely the only direct performance impact this code makes on non-NEWCC mode. RMAPs: * A new type of RMAP entry is introduced, distinguished by RMAP_IS_SEGMENT(Address) of the rmap entry. This kind of entry contains a pointer to a section page table node in the Process pointer, which in turn links back to the MM_SECTION_SEGMENT it belongs to. Therefore, a page belonging only to a segment (that is, a segment page that isn't mapped) can exist and be evicted using the normal page eviction mechanism in balance.c. Each of the rmap function has been modified to deal with segment rmaps. * The low 8 bits of the Address field in a segment rmap denote the entry number in the generic table node pointed to by Process that points to the page the rmap belongs to. By combining them, you can determine the file offset the page belongs to. * In NEWCC mode, MmSharePageEntry/UnsharePageEntry are not used, and instead the page reference count is used to keep track of the number of mappings of a page, allowing the last reference expiring to allow the page to be recycled without much intervention. These are still used in non-NEWCC mode. One change has been made, the count fields have been narrowed by 1 bit to make room for a dirty bit in SSE entries, needed when a page is present but unmapped. Section page tables: * The section page tables are now implemented using RtlGenericTables. This enables a fairly compact representation of section page tables without having the existence of a section object imply 4k of fake PDEs. In addition, each node in the generic table has a wide file offset that is a multiple of 256 pages, or 1 megabyte total. Besides needing wide file offsets, the only other visible change caused by the switch to generic tables for section page tables is the need to lock the section segment before interacting with the section page table. Eviction: * Page eviction in cache sections is accomplished by MmpPageOutPhysicalAddress. In the case of a shared page, it tries to remove all mappings of the indicated page. If this process fails at any point, the page will simply be drawn back into the target address spaces. After succeeding at this, if TRUE has been accumulated into the page's dirty bit in the section page table, it is written back, and then permanently removed. NewCC mode: * NEWCC mode is introduced, which rewrites the file cache to a set of cache stripes actively mapped, along with unmapped section data. * NewCC is more authentic in its interpretation of the external interface to the windows cache than the current cache manager, implementing each of the cache manager functions according to the documented interface with no preconceived ideas about how anything should be implemented internally. Cache stripes are implemented on top of section objects, using the same memory manager paths, and therefore economizing code and complexity. This replaces a rather complicated system in which pages can be owned by the cache manager and the memory manager simultaneously and they must cooperate in a fairly sophisticated way to manage them. Since they're quite interdependent in the current code, modifying either is very difficult. In NEWCC, they have a clear division of labor and thus can be worked on independently. * Several third party filesystems that use the kernel Cc interface work properly using NEWCC, including matt wu's ext3 driver. * In contrast with code that tries to make CcInitializeCacheMap and CcUninitializeCacheMap into a pair that supports reference counting, NEWCC lazily initializes the shared and private cache maps as needed and uses the presence of a PrivateCacheMap on at least one file pointing to the SharedCacheMap as an indication that the FILE_OBJECT reference in the SharedCacheMap should still be held. When the last PrivateCacheMap is discarded, that's the appropriate time to tear down caching for a specific file, as the SharedCacheMap data is allowed to be saved and reused. We honor this by making the SharedCacheMap into a depot for keeping track of the PrivateCacheMap objects associated with views of a file. svn path=/trunk/; revision=55833
2012-02-23 12:03:06 +00:00
extern KEVENT MmWaitPageEvent;
extern FAST_MUTEX MiGlobalPageOperation;
extern LIST_ENTRY MiSegmentList;
extern NTSTATUS MiRosTrimCache(ULONG Target, ULONG Priority, PULONG NrFreed);
/* PRIVATE FUNCTIONS *********************************************************/
//
// Helper function to create initial memory areas.
// The created area is always read/write.
//
CODE_SEG("INIT")
VOID
NTAPI
MiCreateArm3StaticMemoryArea(PVOID BaseAddress, SIZE_T Size, BOOLEAN Executable)
{
const ULONG Protection = Executable ? PAGE_EXECUTE_READWRITE : PAGE_READWRITE;
PVOID pBaseAddress = BaseAddress;
PMEMORY_AREA MArea;
NTSTATUS Status;
Status = MmCreateMemoryArea(MmGetKernelAddressSpace(),
MEMORY_AREA_OWNED_BY_ARM3 | MEMORY_AREA_STATIC,
&pBaseAddress,
Size,
Protection,
&MArea,
0,
PAGE_SIZE);
ASSERT(Status == STATUS_SUCCESS);
// TODO: Perhaps it would be prudent to bugcheck here, not only assert?
}
CODE_SEG("INIT")
VOID
NTAPI
MiInitSystemMemoryAreas(VOID)
{
//
// Create all the static memory areas.
//
MmLockAddressSpace(MmGetKernelAddressSpace());
#ifdef _M_AMD64
// Reserved range FFFF800000000000 - FFFFF68000000000
MiCreateArm3StaticMemoryArea((PVOID)MI_REAL_SYSTEM_RANGE_START, PTE_BASE - MI_REAL_SYSTEM_RANGE_START, FALSE);
#endif /* _M_AMD64 */
// The loader mappings. The only Executable area.
MiCreateArm3StaticMemoryArea((PVOID)KSEG0_BASE, MmBootImageSize, TRUE);
// The PTE base
MiCreateArm3StaticMemoryArea((PVOID)PTE_BASE, PTE_TOP - PTE_BASE + 1, FALSE);
// Hyperspace
MiCreateArm3StaticMemoryArea((PVOID)HYPER_SPACE, HYPER_SPACE_END - HYPER_SPACE + 1, FALSE);
// Protect the PFN database
MiCreateArm3StaticMemoryArea(MmPfnDatabase, (MxPfnAllocation << PAGE_SHIFT), FALSE);
// ReactOS requires a memory area to keep the initial NP area off-bounds
MiCreateArm3StaticMemoryArea(MmNonPagedPoolStart, MmSizeOfNonPagedPoolInBytes, FALSE);
// System PTE space
MiCreateArm3StaticMemoryArea(MmNonPagedSystemStart, (MmNumberOfSystemPtes + 1) * PAGE_SIZE, FALSE);
// Nonpaged pool expansion space
MiCreateArm3StaticMemoryArea(MmNonPagedPoolExpansionStart, (ULONG_PTR)MmNonPagedPoolEnd - (ULONG_PTR)MmNonPagedPoolExpansionStart, FALSE);
// System view space
MiCreateArm3StaticMemoryArea(MiSystemViewStart, MmSystemViewSize, FALSE);
// Session space
MiCreateArm3StaticMemoryArea(MmSessionBase, (ULONG_PTR)MiSessionSpaceEnd - (ULONG_PTR)MmSessionBase, FALSE);
// Paged pool
MiCreateArm3StaticMemoryArea(MmPagedPoolStart, MmSizeOfPagedPoolInBytes, FALSE);
// Debugger mapping
MiCreateArm3StaticMemoryArea(MI_DEBUG_MAPPING, PAGE_SIZE, FALSE);
#if defined(_X86_)
// Reserved HAL area (includes KUSER_SHARED_DATA and KPCR)
MiCreateArm3StaticMemoryArea((PVOID)MM_HAL_VA_START, MM_HAL_VA_END - MM_HAL_VA_START + 1, FALSE);
#else /* _X86_ */
#ifndef _M_AMD64
// KPCR, one page per CPU. Only for 32-bit kernel.
MiCreateArm3StaticMemoryArea(PCR, PAGE_SIZE * KeNumberProcessors, FALSE);
#endif /* _M_AMD64 */
- Implement support for reading and writing physical memory for KD. The implementation uses a reserved mapping page to map the target physical address to. On x86 this page is located at virtual address 0xFFBFF000, and the PTE for this page is the last PTE of the nonpaged pool's PDE. Other architectures may need to reserve the PTE elsewhere. - The physical memory support relies on several Mm variables and structures to be properly set up. Add a new flag, MiDbgReadyForPhysical, and set it when the debugger support can handle physical memory requests. - Protect this page with a Memory Area to make the old Mm keep its dirty hands off it. - Does not support I/O space or cache flags yet. - Add generic KeInvalidateTlbEntry to invalidate a single TLB entry for a given address instead of flushing the whole TLB. Used by the debugger physical memory support as invalidating the whole TLB for every map and unmap of its debug PTE would incur significant overhead for large copies. Replace direct usage of __invlpg() with this in x86 code too. - Fix incorrect cache flag check and set in KdpRead/WritePhysicalmemory for write combined requests. The debugger's Uncached flag was checked instead of the Write Combined flag, and the debuggers Write Combine number (0x3) was set instead of Mm's flag (0x20). - Fix implementation of MmIsAddressValid (at least for x86; other architectures will need more checks). Just check the Address' PDE and PTE valid bits instead of using Memory Areas. - Add missing ASSERTs to ensure the Memory Areas for paged pool, the PCR page, and the Shared User Data page are created. - Add missing Memory Area for the 2 pages HAL currently uses for its own mappings on x86 -- previously, those pages could have been allocated by other parts of the OS, which would have resulted in serious corruptions. svn path=/trunk/; revision=43960
2009-11-04 22:40:18 +00:00
// KUSER_SHARED_DATA
MiCreateArm3StaticMemoryArea((PVOID)KI_USER_SHARED_DATA, PAGE_SIZE, FALSE);
#endif /* _X86_ */
MmUnlockAddressSpace(MmGetKernelAddressSpace());
}
CODE_SEG("INIT")
VOID
NTAPI
MiDbgDumpAddressSpace(VOID)
{
//
// Print the memory layout
//
DPRINT1(" 0x%p - 0x%p\t%s\n",
KSEG0_BASE,
(ULONG_PTR)KSEG0_BASE + MmBootImageSize,
"Boot Loaded Image");
DPRINT1(" 0x%p - 0x%p\t%s\n",
MmPfnDatabase,
(ULONG_PTR)MmPfnDatabase + (MxPfnAllocation << PAGE_SHIFT),
"PFN Database");
DPRINT1(" 0x%p - 0x%p\t%s\n",
MmNonPagedPoolStart,
(ULONG_PTR)MmNonPagedPoolStart + MmSizeOfNonPagedPoolInBytes,
"ARM3 Non Paged Pool");
DPRINT1(" 0x%p - 0x%p\t%s\n",
MiSystemViewStart,
(ULONG_PTR)MiSystemViewStart + MmSystemViewSize,
"System View Space");
DPRINT1(" 0x%p - 0x%p\t%s\n",
MmSessionBase,
MiSessionSpaceEnd,
"Session Space");
DPRINT1(" 0x%p - 0x%p\t%s\n",
PTE_BASE, PTE_TOP,
"Page Tables");
DPRINT1(" 0x%p - 0x%p\t%s\n",
PDE_BASE, PDE_TOP,
"Page Directories");
DPRINT1(" 0x%p - 0x%p\t%s\n",
HYPER_SPACE, HYPER_SPACE_END,
"Hyperspace");
DPRINT1(" 0x%p - 0x%p\t%s\n",
MmSystemCacheStart, MmSystemCacheEnd,
"System Cache");
DPRINT1(" 0x%p - 0x%p\t%s\n",
MmPagedPoolStart,
(ULONG_PTR)MmPagedPoolStart + MmSizeOfPagedPoolInBytes,
"ARM3 Paged Pool");
DPRINT1(" 0x%p - 0x%p\t%s\n",
MmNonPagedSystemStart, MmNonPagedPoolExpansionStart,
"System PTE Space");
DPRINT1(" 0x%p - 0x%p\t%s\n",
MmNonPagedPoolExpansionStart, MmNonPagedPoolEnd,
"Non Paged Pool Expansion PTE Space");
}
CODE_SEG("INIT")
NTSTATUS
NTAPI
MmInitBsmThread(VOID)
{
NTSTATUS Status;
OBJECT_ATTRIBUTES ObjectAttributes;
HANDLE ThreadHandle;
/* Create the thread */
InitializeObjectAttributes(&ObjectAttributes, NULL, 0, NULL, NULL);
Status = PsCreateSystemThread(&ThreadHandle,
THREAD_ALL_ACCESS,
&ObjectAttributes,
NULL,
NULL,
KeBalanceSetManager,
NULL);
/* Close the handle and return status */
ZwClose(ThreadHandle);
return Status;
}
CODE_SEG("INIT")
BOOLEAN
NTAPI
MmInitSystem(IN ULONG Phase,
IN PLOADER_PARAMETER_BLOCK LoaderBlock)
{
extern MMPTE ValidKernelPte;
PMMPTE PointerPte;
MMPTE TempPte = ValidKernelPte;
PFN_NUMBER PageFrameNumber;
PLIST_ENTRY ListEntry;
PLDR_DATA_TABLE_ENTRY DataTableEntry;
/* Initialize the kernel address space */
ASSERT(Phase == 1);
[NEWCC] A reintegration checkpoint for the NewCC branch, brought to you by Team NewCC. Differences with current ReactOS trunk: * A new memory area type, MEMORY_AREA_CACHE, is added, which represents a mapped region of a file. In NEWCC mode, user sections are MEMORY_AREA_CACHE type as well, and obey new semantics. In non-NEWCC mode, they aren't used. * A way of claiming a page entry for a specific thread's work is added. Placing the special SWAPENTRY value MM_WAIT_ENTRY in a page table, or in a section page table should indicate that memory management code is intended to wait for another thread to make some status change before checking the state of the page entry again. In code that uses this convention, a return value of STATUS_SUCCESS + 1 is used to indicate that the caller should use the MiWaitForPageEvent macro to wait until somebody has change the state of a wait entry before checking again. This is a lighter weight mechanism than PAGEOPs. * A way of asking the caller to perform some blocking operation without locks held is provided. This replaces some spaghettified code in which locks are repeatedly taken and broken by code that performs various blocking operations. Using this mechanism, it is possible to do a small amount of non-blocking work, fill in a request, then return STATUS_MORE_PROCESSING_REQUIRED to request that locks be dropped and the blocking operation be carried out. A MM_REQUIRED_RESOURCES structure is provided to consumers of this contract to use to accumulate state across many blocking operations. Several functions wrapping blocking operations are provided in ntoskrnl/cache/reqtools.c. * Image section pages are no longer direct mapped. This is done to simplify consolidation of ownership of pages under the data section system. At a later time, it may be possible to make data pages directly available to image sections for the same file. This is likely the only direct performance impact this code makes on non-NEWCC mode. RMAPs: * A new type of RMAP entry is introduced, distinguished by RMAP_IS_SEGMENT(Address) of the rmap entry. This kind of entry contains a pointer to a section page table node in the Process pointer, which in turn links back to the MM_SECTION_SEGMENT it belongs to. Therefore, a page belonging only to a segment (that is, a segment page that isn't mapped) can exist and be evicted using the normal page eviction mechanism in balance.c. Each of the rmap function has been modified to deal with segment rmaps. * The low 8 bits of the Address field in a segment rmap denote the entry number in the generic table node pointed to by Process that points to the page the rmap belongs to. By combining them, you can determine the file offset the page belongs to. * In NEWCC mode, MmSharePageEntry/UnsharePageEntry are not used, and instead the page reference count is used to keep track of the number of mappings of a page, allowing the last reference expiring to allow the page to be recycled without much intervention. These are still used in non-NEWCC mode. One change has been made, the count fields have been narrowed by 1 bit to make room for a dirty bit in SSE entries, needed when a page is present but unmapped. Section page tables: * The section page tables are now implemented using RtlGenericTables. This enables a fairly compact representation of section page tables without having the existence of a section object imply 4k of fake PDEs. In addition, each node in the generic table has a wide file offset that is a multiple of 256 pages, or 1 megabyte total. Besides needing wide file offsets, the only other visible change caused by the switch to generic tables for section page tables is the need to lock the section segment before interacting with the section page table. Eviction: * Page eviction in cache sections is accomplished by MmpPageOutPhysicalAddress. In the case of a shared page, it tries to remove all mappings of the indicated page. If this process fails at any point, the page will simply be drawn back into the target address spaces. After succeeding at this, if TRUE has been accumulated into the page's dirty bit in the section page table, it is written back, and then permanently removed. NewCC mode: * NEWCC mode is introduced, which rewrites the file cache to a set of cache stripes actively mapped, along with unmapped section data. * NewCC is more authentic in its interpretation of the external interface to the windows cache than the current cache manager, implementing each of the cache manager functions according to the documented interface with no preconceived ideas about how anything should be implemented internally. Cache stripes are implemented on top of section objects, using the same memory manager paths, and therefore economizing code and complexity. This replaces a rather complicated system in which pages can be owned by the cache manager and the memory manager simultaneously and they must cooperate in a fairly sophisticated way to manage them. Since they're quite interdependent in the current code, modifying either is very difficult. In NEWCC, they have a clear division of labor and thus can be worked on independently. * Several third party filesystems that use the kernel Cc interface work properly using NEWCC, including matt wu's ext3 driver. * In contrast with code that tries to make CcInitializeCacheMap and CcUninitializeCacheMap into a pair that supports reference counting, NEWCC lazily initializes the shared and private cache maps as needed and uses the presence of a PrivateCacheMap on at least one file pointing to the SharedCacheMap as an indication that the FILE_OBJECT reference in the SharedCacheMap should still be held. When the last PrivateCacheMap is discarded, that's the appropriate time to tear down caching for a specific file, as the SharedCacheMap data is allowed to be saved and reused. We honor this by making the SharedCacheMap into a depot for keeping track of the PrivateCacheMap objects associated with views of a file. svn path=/trunk/; revision=55833
2012-02-23 12:03:06 +00:00
#ifdef NEWCC
[NEWCC] A reintegration checkpoint for the NewCC branch, brought to you by Team NewCC. Differences with current ReactOS trunk: * A new memory area type, MEMORY_AREA_CACHE, is added, which represents a mapped region of a file. In NEWCC mode, user sections are MEMORY_AREA_CACHE type as well, and obey new semantics. In non-NEWCC mode, they aren't used. * A way of claiming a page entry for a specific thread's work is added. Placing the special SWAPENTRY value MM_WAIT_ENTRY in a page table, or in a section page table should indicate that memory management code is intended to wait for another thread to make some status change before checking the state of the page entry again. In code that uses this convention, a return value of STATUS_SUCCESS + 1 is used to indicate that the caller should use the MiWaitForPageEvent macro to wait until somebody has change the state of a wait entry before checking again. This is a lighter weight mechanism than PAGEOPs. * A way of asking the caller to perform some blocking operation without locks held is provided. This replaces some spaghettified code in which locks are repeatedly taken and broken by code that performs various blocking operations. Using this mechanism, it is possible to do a small amount of non-blocking work, fill in a request, then return STATUS_MORE_PROCESSING_REQUIRED to request that locks be dropped and the blocking operation be carried out. A MM_REQUIRED_RESOURCES structure is provided to consumers of this contract to use to accumulate state across many blocking operations. Several functions wrapping blocking operations are provided in ntoskrnl/cache/reqtools.c. * Image section pages are no longer direct mapped. This is done to simplify consolidation of ownership of pages under the data section system. At a later time, it may be possible to make data pages directly available to image sections for the same file. This is likely the only direct performance impact this code makes on non-NEWCC mode. RMAPs: * A new type of RMAP entry is introduced, distinguished by RMAP_IS_SEGMENT(Address) of the rmap entry. This kind of entry contains a pointer to a section page table node in the Process pointer, which in turn links back to the MM_SECTION_SEGMENT it belongs to. Therefore, a page belonging only to a segment (that is, a segment page that isn't mapped) can exist and be evicted using the normal page eviction mechanism in balance.c. Each of the rmap function has been modified to deal with segment rmaps. * The low 8 bits of the Address field in a segment rmap denote the entry number in the generic table node pointed to by Process that points to the page the rmap belongs to. By combining them, you can determine the file offset the page belongs to. * In NEWCC mode, MmSharePageEntry/UnsharePageEntry are not used, and instead the page reference count is used to keep track of the number of mappings of a page, allowing the last reference expiring to allow the page to be recycled without much intervention. These are still used in non-NEWCC mode. One change has been made, the count fields have been narrowed by 1 bit to make room for a dirty bit in SSE entries, needed when a page is present but unmapped. Section page tables: * The section page tables are now implemented using RtlGenericTables. This enables a fairly compact representation of section page tables without having the existence of a section object imply 4k of fake PDEs. In addition, each node in the generic table has a wide file offset that is a multiple of 256 pages, or 1 megabyte total. Besides needing wide file offsets, the only other visible change caused by the switch to generic tables for section page tables is the need to lock the section segment before interacting with the section page table. Eviction: * Page eviction in cache sections is accomplished by MmpPageOutPhysicalAddress. In the case of a shared page, it tries to remove all mappings of the indicated page. If this process fails at any point, the page will simply be drawn back into the target address spaces. After succeeding at this, if TRUE has been accumulated into the page's dirty bit in the section page table, it is written back, and then permanently removed. NewCC mode: * NEWCC mode is introduced, which rewrites the file cache to a set of cache stripes actively mapped, along with unmapped section data. * NewCC is more authentic in its interpretation of the external interface to the windows cache than the current cache manager, implementing each of the cache manager functions according to the documented interface with no preconceived ideas about how anything should be implemented internally. Cache stripes are implemented on top of section objects, using the same memory manager paths, and therefore economizing code and complexity. This replaces a rather complicated system in which pages can be owned by the cache manager and the memory manager simultaneously and they must cooperate in a fairly sophisticated way to manage them. Since they're quite interdependent in the current code, modifying either is very difficult. In NEWCC, they have a clear division of labor and thus can be worked on independently. * Several third party filesystems that use the kernel Cc interface work properly using NEWCC, including matt wu's ext3 driver. * In contrast with code that tries to make CcInitializeCacheMap and CcUninitializeCacheMap into a pair that supports reference counting, NEWCC lazily initializes the shared and private cache maps as needed and uses the presence of a PrivateCacheMap on at least one file pointing to the SharedCacheMap as an indication that the FILE_OBJECT reference in the SharedCacheMap should still be held. When the last PrivateCacheMap is discarded, that's the appropriate time to tear down caching for a specific file, as the SharedCacheMap data is allowed to be saved and reused. We honor this by making the SharedCacheMap into a depot for keeping track of the PrivateCacheMap objects associated with views of a file. svn path=/trunk/; revision=55833
2012-02-23 12:03:06 +00:00
InitializeListHead(&MiSegmentList);
ExInitializeFastMutex(&MiGlobalPageOperation);
KeInitializeEvent(&MmWaitPageEvent, SynchronizationEvent, FALSE);
// Until we're fully demand paged, we can do things the old way through
// the balance manager
// CcInitView will override this...
[NEWCC] A reintegration checkpoint for the NewCC branch, brought to you by Team NewCC. Differences with current ReactOS trunk: * A new memory area type, MEMORY_AREA_CACHE, is added, which represents a mapped region of a file. In NEWCC mode, user sections are MEMORY_AREA_CACHE type as well, and obey new semantics. In non-NEWCC mode, they aren't used. * A way of claiming a page entry for a specific thread's work is added. Placing the special SWAPENTRY value MM_WAIT_ENTRY in a page table, or in a section page table should indicate that memory management code is intended to wait for another thread to make some status change before checking the state of the page entry again. In code that uses this convention, a return value of STATUS_SUCCESS + 1 is used to indicate that the caller should use the MiWaitForPageEvent macro to wait until somebody has change the state of a wait entry before checking again. This is a lighter weight mechanism than PAGEOPs. * A way of asking the caller to perform some blocking operation without locks held is provided. This replaces some spaghettified code in which locks are repeatedly taken and broken by code that performs various blocking operations. Using this mechanism, it is possible to do a small amount of non-blocking work, fill in a request, then return STATUS_MORE_PROCESSING_REQUIRED to request that locks be dropped and the blocking operation be carried out. A MM_REQUIRED_RESOURCES structure is provided to consumers of this contract to use to accumulate state across many blocking operations. Several functions wrapping blocking operations are provided in ntoskrnl/cache/reqtools.c. * Image section pages are no longer direct mapped. This is done to simplify consolidation of ownership of pages under the data section system. At a later time, it may be possible to make data pages directly available to image sections for the same file. This is likely the only direct performance impact this code makes on non-NEWCC mode. RMAPs: * A new type of RMAP entry is introduced, distinguished by RMAP_IS_SEGMENT(Address) of the rmap entry. This kind of entry contains a pointer to a section page table node in the Process pointer, which in turn links back to the MM_SECTION_SEGMENT it belongs to. Therefore, a page belonging only to a segment (that is, a segment page that isn't mapped) can exist and be evicted using the normal page eviction mechanism in balance.c. Each of the rmap function has been modified to deal with segment rmaps. * The low 8 bits of the Address field in a segment rmap denote the entry number in the generic table node pointed to by Process that points to the page the rmap belongs to. By combining them, you can determine the file offset the page belongs to. * In NEWCC mode, MmSharePageEntry/UnsharePageEntry are not used, and instead the page reference count is used to keep track of the number of mappings of a page, allowing the last reference expiring to allow the page to be recycled without much intervention. These are still used in non-NEWCC mode. One change has been made, the count fields have been narrowed by 1 bit to make room for a dirty bit in SSE entries, needed when a page is present but unmapped. Section page tables: * The section page tables are now implemented using RtlGenericTables. This enables a fairly compact representation of section page tables without having the existence of a section object imply 4k of fake PDEs. In addition, each node in the generic table has a wide file offset that is a multiple of 256 pages, or 1 megabyte total. Besides needing wide file offsets, the only other visible change caused by the switch to generic tables for section page tables is the need to lock the section segment before interacting with the section page table. Eviction: * Page eviction in cache sections is accomplished by MmpPageOutPhysicalAddress. In the case of a shared page, it tries to remove all mappings of the indicated page. If this process fails at any point, the page will simply be drawn back into the target address spaces. After succeeding at this, if TRUE has been accumulated into the page's dirty bit in the section page table, it is written back, and then permanently removed. NewCC mode: * NEWCC mode is introduced, which rewrites the file cache to a set of cache stripes actively mapped, along with unmapped section data. * NewCC is more authentic in its interpretation of the external interface to the windows cache than the current cache manager, implementing each of the cache manager functions according to the documented interface with no preconceived ideas about how anything should be implemented internally. Cache stripes are implemented on top of section objects, using the same memory manager paths, and therefore economizing code and complexity. This replaces a rather complicated system in which pages can be owned by the cache manager and the memory manager simultaneously and they must cooperate in a fairly sophisticated way to manage them. Since they're quite interdependent in the current code, modifying either is very difficult. In NEWCC, they have a clear division of labor and thus can be worked on independently. * Several third party filesystems that use the kernel Cc interface work properly using NEWCC, including matt wu's ext3 driver. * In contrast with code that tries to make CcInitializeCacheMap and CcUninitializeCacheMap into a pair that supports reference counting, NEWCC lazily initializes the shared and private cache maps as needed and uses the presence of a PrivateCacheMap on at least one file pointing to the SharedCacheMap as an indication that the FILE_OBJECT reference in the SharedCacheMap should still be held. When the last PrivateCacheMap is discarded, that's the appropriate time to tear down caching for a specific file, as the SharedCacheMap data is allowed to be saved and reused. We honor this by making the SharedCacheMap into a depot for keeping track of the PrivateCacheMap objects associated with views of a file. svn path=/trunk/; revision=55833
2012-02-23 12:03:06 +00:00
MmInitializeMemoryConsumer(MC_CACHE, MiRosTrimCache);
#else
KeInitializeEvent(&MmWaitPageEvent, SynchronizationEvent, FALSE);
#endif
[NEWCC] A reintegration checkpoint for the NewCC branch, brought to you by Team NewCC. Differences with current ReactOS trunk: * A new memory area type, MEMORY_AREA_CACHE, is added, which represents a mapped region of a file. In NEWCC mode, user sections are MEMORY_AREA_CACHE type as well, and obey new semantics. In non-NEWCC mode, they aren't used. * A way of claiming a page entry for a specific thread's work is added. Placing the special SWAPENTRY value MM_WAIT_ENTRY in a page table, or in a section page table should indicate that memory management code is intended to wait for another thread to make some status change before checking the state of the page entry again. In code that uses this convention, a return value of STATUS_SUCCESS + 1 is used to indicate that the caller should use the MiWaitForPageEvent macro to wait until somebody has change the state of a wait entry before checking again. This is a lighter weight mechanism than PAGEOPs. * A way of asking the caller to perform some blocking operation without locks held is provided. This replaces some spaghettified code in which locks are repeatedly taken and broken by code that performs various blocking operations. Using this mechanism, it is possible to do a small amount of non-blocking work, fill in a request, then return STATUS_MORE_PROCESSING_REQUIRED to request that locks be dropped and the blocking operation be carried out. A MM_REQUIRED_RESOURCES structure is provided to consumers of this contract to use to accumulate state across many blocking operations. Several functions wrapping blocking operations are provided in ntoskrnl/cache/reqtools.c. * Image section pages are no longer direct mapped. This is done to simplify consolidation of ownership of pages under the data section system. At a later time, it may be possible to make data pages directly available to image sections for the same file. This is likely the only direct performance impact this code makes on non-NEWCC mode. RMAPs: * A new type of RMAP entry is introduced, distinguished by RMAP_IS_SEGMENT(Address) of the rmap entry. This kind of entry contains a pointer to a section page table node in the Process pointer, which in turn links back to the MM_SECTION_SEGMENT it belongs to. Therefore, a page belonging only to a segment (that is, a segment page that isn't mapped) can exist and be evicted using the normal page eviction mechanism in balance.c. Each of the rmap function has been modified to deal with segment rmaps. * The low 8 bits of the Address field in a segment rmap denote the entry number in the generic table node pointed to by Process that points to the page the rmap belongs to. By combining them, you can determine the file offset the page belongs to. * In NEWCC mode, MmSharePageEntry/UnsharePageEntry are not used, and instead the page reference count is used to keep track of the number of mappings of a page, allowing the last reference expiring to allow the page to be recycled without much intervention. These are still used in non-NEWCC mode. One change has been made, the count fields have been narrowed by 1 bit to make room for a dirty bit in SSE entries, needed when a page is present but unmapped. Section page tables: * The section page tables are now implemented using RtlGenericTables. This enables a fairly compact representation of section page tables without having the existence of a section object imply 4k of fake PDEs. In addition, each node in the generic table has a wide file offset that is a multiple of 256 pages, or 1 megabyte total. Besides needing wide file offsets, the only other visible change caused by the switch to generic tables for section page tables is the need to lock the section segment before interacting with the section page table. Eviction: * Page eviction in cache sections is accomplished by MmpPageOutPhysicalAddress. In the case of a shared page, it tries to remove all mappings of the indicated page. If this process fails at any point, the page will simply be drawn back into the target address spaces. After succeeding at this, if TRUE has been accumulated into the page's dirty bit in the section page table, it is written back, and then permanently removed. NewCC mode: * NEWCC mode is introduced, which rewrites the file cache to a set of cache stripes actively mapped, along with unmapped section data. * NewCC is more authentic in its interpretation of the external interface to the windows cache than the current cache manager, implementing each of the cache manager functions according to the documented interface with no preconceived ideas about how anything should be implemented internally. Cache stripes are implemented on top of section objects, using the same memory manager paths, and therefore economizing code and complexity. This replaces a rather complicated system in which pages can be owned by the cache manager and the memory manager simultaneously and they must cooperate in a fairly sophisticated way to manage them. Since they're quite interdependent in the current code, modifying either is very difficult. In NEWCC, they have a clear division of labor and thus can be worked on independently. * Several third party filesystems that use the kernel Cc interface work properly using NEWCC, including matt wu's ext3 driver. * In contrast with code that tries to make CcInitializeCacheMap and CcUninitializeCacheMap into a pair that supports reference counting, NEWCC lazily initializes the shared and private cache maps as needed and uses the presence of a PrivateCacheMap on at least one file pointing to the SharedCacheMap as an indication that the FILE_OBJECT reference in the SharedCacheMap should still be held. When the last PrivateCacheMap is discarded, that's the appropriate time to tear down caching for a specific file, as the SharedCacheMap data is allowed to be saved and reused. We honor this by making the SharedCacheMap into a depot for keeping track of the PrivateCacheMap objects associated with views of a file. svn path=/trunk/; revision=55833
2012-02-23 12:03:06 +00:00
MmKernelAddressSpace = &PsIdleProcess->Vm;
/* Intialize system memory areas */
MiInitSystemMemoryAreas();
/* Dump the address space */
MiDbgDumpAddressSpace();
MmInitGlobalKernelPageDirectory();
MmInitializeMemoryConsumer(MC_USER, MmTrimUserMemory);
MmInitializeRmapList();
MmInitSectionImplementation();
MmInitPagingFile();
//
// Create a PTE to double-map the shared data section. We allocate it
// from paged pool so that we can't fault when trying to touch the PTE
// itself (to map it), since paged pool addresses will already be mapped
// by the fault handler.
//
MmSharedUserDataPte = ExAllocatePoolWithTag(PagedPool,
sizeof(MMPTE),
TAG_MM);
if (!MmSharedUserDataPte) return FALSE;
//
// Now get the PTE for shared data, and read the PFN that holds it
//
PointerPte = MiAddressToPte((PVOID)KI_USER_SHARED_DATA);
ASSERT(PointerPte->u.Hard.Valid == 1);
PageFrameNumber = PFN_FROM_PTE(PointerPte);
/* Build the PTE and write it */
MI_MAKE_HARDWARE_PTE_KERNEL(&TempPte,
PointerPte,
MM_READONLY,
PageFrameNumber);
*MmSharedUserDataPte = TempPte;
/* Initialize session working set support */
MiInitializeSessionWsSupport();
/* Setup session IDs */
MiInitializeSessionIds();
/* Setup the memory threshold events */
if (!MiInitializeMemoryEvents()) return FALSE;
/*
* Unmap low memory
*/
MiInitBalancerThread();
/* Initialize the balance set manager */
MmInitBsmThread();
/* Loop the boot loaded images (under lock) */
ExAcquireResourceExclusiveLite(&PsLoadedModuleResource, TRUE);
for (ListEntry = PsLoadedModuleList.Flink;
ListEntry != &PsLoadedModuleList;
ListEntry = ListEntry->Flink)
{
/* Get the data table entry */
DataTableEntry = CONTAINING_RECORD(ListEntry, LDR_DATA_TABLE_ENTRY, InLoadOrderLinks);
/* Set up the image protection */
MiWriteProtectSystemImage(DataTableEntry->DllBase);
}
ExReleaseResourceLite(&PsLoadedModuleResource);
return TRUE;
}