memobj-r0drv-nt.cpp revision c5d2523548cc57504b829f53f1362b848a84542c
/* $Id$ */
/** @file
* innotek Portable Runtime - Ring-0 Memory Objects, NT.
*/
/*
* Copyright (C) 2006-2007 innotek GmbH
*
* This file is part of VirtualBox Open Source Edition (OSE), as
* available from http://www.virtualbox.org. This file is free software;
* you can redistribute it and/or modify it under the terms of the GNU
* General Public License as published by the Free Software Foundation,
* in version 2 as it comes in the "COPYING" file of the VirtualBox OSE
* distribution. VirtualBox OSE is distributed in the hope that it will
* be useful, but WITHOUT ANY WARRANTY of any kind.
*/
/*******************************************************************************
* Header Files *
*******************************************************************************/
#include "the-nt-kernel.h"
#include <iprt/memobj.h>
#include <iprt/alloc.h>
#include <iprt/assert.h>
#include <iprt/log.h>
#include <iprt/param.h>
#include <iprt/string.h>
#include <iprt/process.h>
#include "internal/memobj.h"
/*******************************************************************************
* Defined Constants And Macros *
*******************************************************************************/
/** Maximum number of bytes we try to lock down in one go.
* This is supposed to have a limit right below 256MB, but this appears
* to actually be much lower. The values here have been determined experimentally.
*/
#ifdef RT_ARCH_X86
# define MAX_LOCK_MEM_SIZE (32*1024*1024) /* 32MB */
#endif
#ifdef RT_ARCH_AMD64
# define MAX_LOCK_MEM_SIZE (24*1024*1024) /* 24MB */
#endif
/*******************************************************************************
* Structures and Typedefs *
*******************************************************************************/
/**
* The NT version of the memory object structure.
*/
typedef struct RTR0MEMOBJNT
{
/** The core structure. */
RTR0MEMOBJINTERNAL Core;
/** The number of PMDLs (memory descriptor lists) in the array. */
unsigned cMdls;
/** Array of MDL pointers. (variable size) */
PMDL apMdls[1];
} RTR0MEMOBJNT, *PRTR0MEMOBJNT;
int rtR0MemObjNativeFree(RTR0MEMOBJ pMem)
{
PRTR0MEMOBJNT pMemNt = (PRTR0MEMOBJNT)pMem;
/*
* Release any memory that we've allocated or locked.
*/
switch (pMemNt->Core.enmType)
{
case RTR0MEMOBJTYPE_LOW:
case RTR0MEMOBJTYPE_PAGE:
break;
case RTR0MEMOBJTYPE_CONT:
break;
case RTR0MEMOBJTYPE_LOCK:
for (unsigned i = 0; i < pMemNt->cMdls; i++)
MmUnlockPages(pMemNt->apMdl[i]);
break;
case RTR0MEMOBJTYPE_PHYS:
Assert(!pMemNt->Core.u.Phys.fAllocated);
break;
case RTR0MEMOBJTYPE_RES_VIRT:
AssertMsgFailed(("RTR0MEMOBJTYPE_RES_VIRT\n"));
return VERR_INTERNAL_ERROR;
break;
case RTR0MEMOBJTYPE_MAPPING:
/* nothing to do here. */
break;
default:
AssertMsgFailed(("enmType=%d\n", pMemNt->Core.enmType));
return VERR_INTERNAL_ERROR;
}
/*
* Free any MDLs.
*/
for (unsigned i = 0; i < pMemNt->cMdls; i++)
{
MmUnlockPages(pMemNt->apMdl[i]);
IoFreeMdl(pMemNt->u.locked.papMdl[i]);
}
return VINF_SUCCESS;
}
int rtR0MemObjNativeAllocPage(PPRTR0MEMOBJINTERNAL ppMem, size_t cb, bool fExecutable)
{
/*
* Try allocate the memory and create it's IOMemoryDescriptor first.
*/
int rc = VERR_NO_PAGE_MEMORY;
AssertCompile(sizeof(IOPhysicalAddress) == 4);
void *pv = IOMallocAligned(cb, PAGE_SIZE);
if (pv)
{
IOMemoryDescriptor *pMemDesc = IOMemoryDescriptor::withAddress((vm_address_t)pv, cb, kIODirectionInOut, kernel_task);
if (pMemDesc)
{
/*
* Create the IPRT memory object.
*/
PRTR0MEMOBJNT pMemNt = (PRTR0MEMOBJNT)rtR0MemObjNew(sizeof(*pMemNt), RTR0MEMOBJTYPE_PAGE, pv, cb);
if (pMemNt)
{
pMemNt->pMemDesc = pMemDesc;
*ppMem = &pMemNt->Core;
return VINF_SUCCESS;
}
rc = VERR_NO_MEMORY;
pMemDesc->release();
}
else
rc = VERR_MEMOBJ_INIT_FAILED;
IOFreeAligned(pv, cb);
}
return rc;
}
int rtR0MemObjNativeAllocLow(PPRTR0MEMOBJINTERNAL ppMem, size_t cb, bool fExecutable)
{
#if 1
/*
* Allocating 128KB for the low page pool can bit a bit exhausting on the kernel,
* it frequnetly causes the entire box to lock up on startup.
*
* So, try allocate the memory using IOMallocAligned first and if we get any high
* physical memory we'll release it and fall back on IOMAllocContiguous.
*/
int rc = VERR_NO_PAGE_MEMORY;
AssertCompile(sizeof(IOPhysicalAddress) == 4);
void *pv = IOMallocAligned(cb, PAGE_SIZE);
if (pv)
{
IOMemoryDescriptor *pMemDesc = IOMemoryDescriptor::withAddress((vm_address_t)pv, cb, kIODirectionInOut, kernel_task);
if (pMemDesc)
{
/*
* Check if it's all below 4GB.
*/
for (IOByteCount off = 0; off < cb; off += PAGE_SIZE)
{
addr64_t Addr = pMemDesc->getPhysicalSegment64(off, NULL);
if (Addr > (uint32_t)(_4G - PAGE_SIZE))
{
/* Ok, we failed, fall back on contiguous allocation. */
pMemDesc->release();
IOFreeAligned(pv, cb);
return rtR0MemObjNativeAllocCont(ppMem, cb, fExecutable);
}
}
/*
* Create the IPRT memory object.
*/
PRTR0MEMOBJNT pMemNt = (PRTR0MEMOBJNT)rtR0MemObjNew(sizeof(*pMemNt), RTR0MEMOBJTYPE_LOW, pv, cb);
if (pMemNt)
{
pMemNt->pMemDesc = pMemDesc;
*ppMem = &pMemNt->Core;
return VINF_SUCCESS;
}
rc = VERR_NO_MEMORY;
pMemDesc->release();
}
else
rc = VERR_MEMOBJ_INIT_FAILED;
IOFreeAligned(pv, cb);
}
return rc;
#else
/*
* IOMallocContiguous is the most suitable API.
*/
return rtR0MemObjNativeAllocCont(ppMem, cb, fExecutable);
#endif
}
int rtR0MemObjNativeAllocCont(PPRTR0MEMOBJINTERNAL ppMem, size_t cb, bool fExecutable)
{
/*
* Try allocate the memory and create it's IOMemoryDescriptor first.
*/
int rc = VERR_NO_CONT_MEMORY;
AssertCompile(sizeof(IOPhysicalAddress) == 4);
void *pv = IOMallocContiguous(cb, PAGE_SIZE, NULL);
if (pv)
{
IOMemoryDescriptor *pMemDesc = IOMemoryDescriptor::withAddress((vm_address_t)pv, cb, kIODirectionInOut, kernel_task);
if (pMemDesc)
{
/* a bit of useful paranoia. */
addr64_t PhysAddr = pMemDesc->getPhysicalSegment64(0, NULL);
Assert(PhysAddr == pMemDesc->getPhysicalAddress());
if ( PhysAddr > 0
&& PhysAddr <= _4G
&& PhysAddr + cb <= _4G)
{
/*
* Create the IPRT memory object.
*/
PRTR0MEMOBJNT pMemNt = (PRTR0MEMOBJNT)rtR0MemObjNew(sizeof(*pMemNt), RTR0MEMOBJTYPE_CONT, pv, cb);
if (pMemNt)
{
pMemNt->Core.u.Cont.Phys = PhysAddr;
pMemNt->pMemDesc = pMemDesc;
*ppMem = &pMemNt->Core;
return VINF_SUCCESS;
}
rc = VERR_NO_MEMORY;
}
else
{
AssertMsgFailed(("PhysAddr=%llx\n", (unsigned long long)PhysAddr));
rc = VERR_INTERNAL_ERROR;
}
pMemDesc->release();
}
else
rc = VERR_MEMOBJ_INIT_FAILED;
IOFreeContiguous(pv, cb);
}
return rc;
}
int rtR0MemObjNativeAllocPhys(PPRTR0MEMOBJINTERNAL ppMem, size_t cb, RTHCPHYS PhysHighest)
{
#if 0 /* turned out IOMallocPhysical isn't exported yet. sigh. */
/*
* Try allocate the memory and create it's IOMemoryDescriptor first.
* Note that IOMallocPhysical is not working correctly (it's ignoring the mask).
*/
/* first calc the mask (in the hope that it'll be used) */
IOPhysicalAddress PhysMask = ~(IOPhysicalAddress)PAGE_OFFSET_MASK;
if (PhysHighest != NIL_RTHCPHYS)
{
PhysMask = ~(IOPhysicalAddress)0;
while (PhysMask > PhysHighest)
PhysMask >>= 1;
AssertReturn(PhysMask + 1 < cb, VERR_INVALID_PARAMETER);
PhysMask &= ~(IOPhysicalAddress)PAGE_OFFSET_MASK;
}
/* try allocate physical memory. */
int rc = VERR_NO_PHYS_MEMORY;
mach_vm_address_t PhysAddr64 = IOMallocPhysical(cb, PhysMask);
if (PhysAddr64)
{
IOPhysicalAddress PhysAddr = PhysAddr64;
if ( PhysAddr == PhysAddr64
&& PhysAddr < PhysHighest
&& PhysAddr + cb <= PhysHighest)
{
/* create a descriptor. */
IOMemoryDescriptor *pMemDesc = IOMemoryDescriptor::withPhysicalAddress(PhysAddr, cb, kIODirectionInOut);
if (pMemDesc)
{
Assert(PhysAddr == pMemDesc->getPhysicalAddress());
/*
* Create the IPRT memory object.
*/
PRTR0MEMOBJNT pMemNt = (PRTR0MEMOBJNT)rtR0MemObjNew(sizeof(*pMemNt), RTR0MEMOBJTYPE_PHYS, NULL, cb);
if (pMemNt)
{
pMemNt->Core.u.Phys.PhysBase = PhysAddr;
pMemNt->Core.u.Phys.fAllocated = true;
pMemNt->pMemDesc = pMemDesc;
*ppMem = &pMemNt->Core;
return VINF_SUCCESS;
}
rc = VERR_NO_MEMORY;
pMemDesc->release();
}
else
rc = VERR_MEMOBJ_INIT_FAILED;
}
else
{
AssertMsgFailed(("PhysAddr=%#llx PhysAddr64=%#llx PhysHigest=%#llx\n", (unsigned long long)PhysAddr,
(unsigned long long)PhysAddr64, (unsigned long long)PhysHighest));
rc = VERR_INTERNAL_ERROR;
}
IOFreePhysical(PhysAddr64, cb);
}
/*
* Just in case IOMallocContiguous doesn't work right, we can try fall back
* on a contiguous allcation.
*/
if (rc == VERR_INTERNAL_ERROR || rc == VERR_NO_PHYS_MEMORY)
{
int rc2 = rtR0MemObjNativeAllocCont(ppMem, cb, false);
if (RT_SUCCESS(rc2))
rc = rc2;
}
return rc;
#else
return rtR0MemObjNativeAllocCont(ppMem, cb, false);
#endif
}
int rtR0MemObjNativeAllocPhysNC(PPRTR0MEMOBJINTERNAL ppMem, size_t cb, RTHCPHYS PhysHighest)
{
/** @todo rtR0MemObjNativeAllocPhys / darwin. */
return rtR0MemObjNativeAllocPhys(ppMem, cb, PhysHighest);
}
int rtR0MemObjNativeEnterPhys(PPRTR0MEMOBJINTERNAL ppMem, RTHCPHYS Phys, size_t cb)
{
/*
* Validate the address range and create a descriptor for it.
*/
int rc = VERR_ADDRESS_TOO_BIG;
IOPhysicalAddress PhysAddr = Phys;
if (PhysAddr == Phys)
{
IOMemoryDescriptor *pMemDesc = IOMemoryDescriptor::withPhysicalAddress(PhysAddr, cb, kIODirectionInOut);
if (pMemDesc)
{
Assert(PhysAddr == pMemDesc->getPhysicalAddress());
/*
* Create the IPRT memory object.
*/
PRTR0MEMOBJNT pMemNt = (PRTR0MEMOBJNT)rtR0MemObjNew(sizeof(*pMemNt), RTR0MEMOBJTYPE_PHYS, NULL, cb);
if (pMemNt)
{
pMemNt->Core.u.Phys.PhysBase = PhysAddr;
pMemNt->Core.u.Phys.fAllocated = false;
pMemNt->pMemDesc = pMemDesc;
*ppMem = &pMemNt->Core;
return VINF_SUCCESS;
}
rc = VERR_NO_MEMORY;
pMemDesc->release();
}
}
else
AssertMsgFailed(("%#llx\n", (unsigned long long)Phys));
return rc;
}
/**
* Internal worker for locking down pages.
*
* @return IPRT status code.
*
* @param ppMem Where to store the memory object pointer.
* @param pv First page.
* @param cb Number of bytes.
* @param Task The task \a pv and \a cb refers to.
*/
static int rtR0MemObjNativeLock(PPRTR0MEMOBJINTERNAL ppMem, void *pv, size_t cb, task_t Task)
{
#ifdef USE_VM_MAP_WIRE
vm_map_t Map = get_task_map(Task);
Assert(Map);
/*
* First try lock the memory.
*/
int rc = VERR_LOCK_FAILED;
kern_return_t kr = vm_map_wire(get_task_map(Task),
(vm_map_offset_t)pv,
(vm_map_offset_t)pv + cb,
VM_PROT_DEFAULT,
0 /* not user */);
if (kr == KERN_SUCCESS)
{
/*
* Create the IPRT memory object.
*/
PRTR0MEMOBJNT pMemNt = (PRTR0MEMOBJNT)rtR0MemObjNew(sizeof(*pMemNt), RTR0MEMOBJTYPE_LOCK, pv, cb);
if (pMemNt)
{
pMemNt->Core.u.Lock.R0Process = (RTR0PROCESS)Task;
*ppMem = &pMemNt->Core;
return VINF_SUCCESS;
}
kr = vm_map_unwire(get_task_map(Task), (vm_map_offset_t)pv, (vm_map_offset_t)pv + cb, 0 /* not user */);
Assert(kr == KERN_SUCCESS);
rc = VERR_NO_MEMORY;
}
#else
/*
* Create a descriptor and try lock it (prepare).
*/
int rc = VERR_MEMOBJ_INIT_FAILED;
IOMemoryDescriptor *pMemDesc = IOMemoryDescriptor::withAddress((vm_address_t)pv, cb, kIODirectionInOut, Task);
if (pMemDesc)
{
IOReturn IORet = pMemDesc->prepare(kIODirectionInOut);
if (IORet == kIOReturnSuccess)
{
/*
* Create the IPRT memory object.
*/
PRTR0MEMOBJNT pMemNt = (PRTR0MEMOBJNT)rtR0MemObjNew(sizeof(*pMemNt), RTR0MEMOBJTYPE_LOCK, pv, cb);
if (pMemNt)
{
pMemNt->Core.u.Lock.R0Process = (RTR0PROCESS)Task;
pMemNt->pMemDesc = pMemDesc;
*ppMem = &pMemNt->Core;
return VINF_SUCCESS;
}
pMemDesc->complete();
rc = VERR_NO_MEMORY;
}
else
rc = VERR_LOCK_FAILED;
pMemDesc->release();
}
#endif
return rc;
}
int rtR0MemObjNativeLockUser(PPRTR0MEMOBJINTERNAL ppMem, void *pv, size_t cb, RTR0PROCESS R0Process)
{
return rtR0MemObjNativeLock(ppMem, pv, cb, (task_t)R0Process);
}
int rtR0MemObjNativeLockKernel(PPRTR0MEMOBJINTERNAL ppMem, void *pv, size_t cb)
{
return rtR0MemObjNativeLock(ppMem, pv, cb, kernel_task);
}
int rtR0MemObjNativeReserveKernel(PPRTR0MEMOBJINTERNAL ppMem, void *pvFixed, size_t cb, size_t uAlignment)
{
return VERR_NOT_IMPLEMENTED;
}
int rtR0MemObjNativeReserveUser(PPRTR0MEMOBJINTERNAL ppMem, RTR3PTR R3PtrFixed, size_t cb, size_t uAlignment, RTR0PROCESS R0Process)
{
return VERR_NOT_IMPLEMENTED;
}
int rtR0MemObjNativeMapKernel(PPRTR0MEMOBJINTERNAL ppMem, RTR0MEMOBJ pMemToMap, void *pvFixed, size_t uAlignment, unsigned fProt)
{
/*
* Must have a memory descriptor.
*/
int rc = VERR_INVALID_PARAMETER;
PRTR0MEMOBJNT pMemToMapDarwin = (PRTR0MEMOBJNT)pMemToMap;
if (pMemToMapDarwin->pMemDesc)
{
IOMemoryMap *pMemMap = pMemToMapDarwin->pMemDesc->map(kernel_task, kIOMapAnywhere,
kIOMapAnywhere | kIOMapDefaultCache);
if (pMemMap)
{
IOVirtualAddress VirtAddr = pMemMap->getVirtualAddress();
void *pv = (void *)(uintptr_t)VirtAddr;
if ((uintptr_t)pv == VirtAddr)
{
/*
* Create the IPRT memory object.
*/
PRTR0MEMOBJNT pMemNt = (PRTR0MEMOBJNT)rtR0MemObjNew(sizeof(*pMemNt), RTR0MEMOBJTYPE_MAPPING,
pv, pMemToMapDarwin->Core.cb);
if (pMemNt)
{
pMemNt->Core.u.Mapping.R0Process = NIL_RTR0PROCESS;
pMemNt->pMemMap = pMemMap;
*ppMem = &pMemNt->Core;
return VINF_SUCCESS;
}
rc = VERR_NO_MEMORY;
}
else
rc = VERR_ADDRESS_TOO_BIG;
pMemMap->release();
}
else
rc = VERR_MAP_FAILED;
}
return rc;
}
int rtR0MemObjNativeMapUser(PPRTR0MEMOBJINTERNAL ppMem, RTR0MEMOBJ pMemToMap, RTR3PTR R3PtrFixed, size_t uAlignment, unsigned fProt, RTR0PROCESS R0Process)
{
/*
* Must have a memory descriptor.
*/
int rc = VERR_INVALID_PARAMETER;
PRTR0MEMOBJNT pMemToMapDarwin = (PRTR0MEMOBJNT)pMemToMap;
if (pMemToMapDarwin->pMemDesc)
{
IOMemoryMap *pMemMap = pMemToMapDarwin->pMemDesc->map((task_t)R0Process, kIOMapAnywhere,
kIOMapAnywhere | kIOMapDefaultCache);
if (pMemMap)
{
IOVirtualAddress VirtAddr = pMemMap->getVirtualAddress();
void *pv = (void *)(uintptr_t)VirtAddr;
if ((uintptr_t)pv == VirtAddr)
{
/*
* Create the IPRT memory object.
*/
PRTR0MEMOBJNT pMemNt = (PRTR0MEMOBJNT)rtR0MemObjNew(sizeof(*pMemNt), RTR0MEMOBJTYPE_MAPPING,
pv, pMemToMapDarwin->Core.cb);
if (pMemNt)
{
pMemNt->Core.u.Mapping.R0Process = R0Process;
pMemNt->pMemMap = pMemMap;
*ppMem = &pMemNt->Core;
return VINF_SUCCESS;
}
rc = VERR_NO_MEMORY;
}
else
rc = VERR_ADDRESS_TOO_BIG;
pMemMap->release();
}
else
rc = VERR_MAP_FAILED;
}
return rc;
}
RTHCPHYS rtR0MemObjNativeGetPagePhysAddr(PRTR0MEMOBJINTERNAL pMem, unsigned iPage)
{
RTHCPHYS PhysAddr;
PRTR0MEMOBJNT pMemNt = (PRTR0MEMOBJNT)pMem;
#ifdef USE_VM_MAP_WIRE
/*
* Locked memory doesn't have a memory descriptor and
* needs to be handled differently.
*/
if (pMemNt->Core.enmType == RTR0MEMOBJTYPE_LOCK)
{
ppnum_t PgNo;
if (pMemNt->Core.u.Lock.R0Process == NIL_RTR0PROCESS)
PgNo = pmap_find_phys(kernel_pmap, (uintptr_t)pMemNt->Core.pv + iPage * PAGE_SIZE);
else
{
/*
* From what I can tell, Apple seems to have locked up the all the
* available interfaces that could help us obtain the pmap_t of a task
* or vm_map_t.
* So, we'll have to figure out where in the vm_map_t structure it is
* and read it our selves. ASSUMING that kernel_pmap is pointed to by
* kernel_map->pmap, we scan kernel_map to locate the structure offset.
* Not nice, but it will hopefully do the job in a reliable manner...
*
* (get_task_pmap, get_map_pmap or vm_map_pmap is what we really need btw.)
*/
static int s_offPmap = -1;
if (RT_UNLIKELY(s_offPmap == -1))
{
pmap_t const *p = (pmap_t *)kernel_map;
pmap_t const * const pEnd = p + 64;
for (; p < pEnd; p++)
if (*p == kernel_pmap)
{
s_offPmap = (uintptr_t)p - (uintptr_t)kernel_map;
break;
}
AssertReturn(s_offPmap >= 0, NIL_RTHCPHYS);
}
pmap_t Pmap = *(pmap_t *)((uintptr_t)get_task_map((task_t)pMemNt->Core.u.Lock.R0Process) + s_offPmap);
PgNo = pmap_find_phys(Pmap, (uintptr_t)pMemNt->Core.pv + iPage * PAGE_SIZE);
}
AssertReturn(PgNo, NIL_RTHCPHYS);
PhysAddr = (RTHCPHYS)PgNo << PAGE_SHIFT;
Assert((PhysAddr >> PAGE_SHIFT) == PgNo);
}
else
#endif /* USE_VM_MAP_WIRE */
{
/*
* Get the memory descriptor.
*/
IOMemoryDescriptor *pMemDesc = pMemNt->pMemDesc;
if (!pMemDesc)
pMemDesc = pMemNt->pMemMap->getMemoryDescriptor();
AssertReturn(pMemDesc, NIL_RTHCPHYS);
/*
* If we've got a memory descriptor, use getPhysicalSegment64().
*/
addr64_t Addr = pMemDesc->getPhysicalSegment64(iPage * PAGE_SIZE, NULL);
AssertMsgReturn(Addr, ("iPage=%u\n", iPage), NIL_RTHCPHYS);
PhysAddr = Addr;
AssertMsgReturn(PhysAddr == Addr, ("PhysAddr=%VHp Addr=%RX64\n", PhysAddr, (uint64_t)Addr), NIL_RTHCPHYS);
}
return PhysAddr;
}