/* * Copyright (C) 2002-2021 The DOSBox Team * * This program 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; either version 2 of the License, or * (at your option) any later version. * * This program is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * * You should have received a copy of the GNU General Public License along * with this program; if not, write to the Free Software Foundation, Inc., * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA. */ #ifndef DOSBOX_MEM_H #define DOSBOX_MEM_H #include "dosbox.h" #include "byteorder.h" #define MEM_PAGESIZE (4096U) /* HostPt and ConstHostPt is for holding linear addresses within this emulator i.e. a normal pointer. * * PhysPt is for 32-bit physical memory addresses within the emulation environment. * PhysPt64 is for 64-bit physical memory addresses for code and device/memory emulation that supports addresses above 4GB. * LinearPt is a 32-bit linear memory address from the point of view of the CPU execution context, meaning the linear address * of the code prior to translation through the page tables to physical addresses. * RealPt is a 32-bit value that holds segment in the upper 16 bits, offset in the lower 16 bits. * * Please do not mix these types up in the code, even if they happen to have the same underlying data types */ typedef uint8_t const * ConstHostPt; /* host (virtual) memory address aka ptr */ typedef uint8_t * HostPt; /* host (virtual) memory address aka ptr */ typedef uint32_t PhysPt; /* guest physical memory pointer */ typedef uint32_t LinearPt; /* guest linear memory address */ typedef uint32_t RealPt; /* guest real-mode memory address (16:16 -> seg:offset) */ typedef uint16_t SegmentVal; /* guest segment value */ typedef uint32_t PageNum; /* page frame number */ typedef uint64_t PhysPt64; /* guest physical memory pointer */ typedef int32_t MemHandle; extern HostPt MemBase; extern size_t MemSize; HostPt GetMemBase(void); bool MEM_A20_Enabled(void); void MEM_A20_Enable(bool enabled); /* Memory management / EMS mapping */ Bitu MEM_FreeTotal(void); //Free 4 kb pages Bitu MEM_FreeLargest(void); //Largest free 4 kb pages block Bitu MEM_TotalPages(void); //Total amount of 4 kb pages Bitu MEM_TotalPagesAt4GB(void); //Total amount of 4 kb pages starting at 4GB Bitu MEM_ConventionalPages(void); Bitu MEM_AllocatedPages(MemHandle handle); // amount of allocated pages of handle MemHandle MEM_AllocatePages(Bitu pages,bool sequence); MemHandle MEM_AllocatePages_A20_friendly(Bitu pages,bool sequence); MemHandle MEM_GetNextFreePage(void); void MEM_ReleasePages(MemHandle handle); bool MEM_ReAllocatePages(MemHandle & handle,Bitu pages,bool sequence); MemHandle MEM_NextHandle(MemHandle handle); MemHandle MEM_NextHandleAt(MemHandle handle,Bitu where); uint32_t MEM_HardwareAllocate(const char *name,uint32_t sz); static constexpr bool build_memlimit_32bit(void) { return sizeof(void*) < 8; } /* The following six functions are used everywhere in the end so these should be changed for Working on big or little endian machines */ static INLINE uint8_t host_readb(ConstHostPt const off) { return *off; } static INLINE void host_writeb(HostPt const off,const uint8_t val) { *off = val; } // use __builtin_bswap* for gcc >= 4.3 #if defined(WORDS_BIGENDIAN) && defined(__GNUC__) && \ (__GNUC__ > 4 || (__GNUC__ == 4 && __GNUC_MINOR__ >= 3)) static INLINE uint16_t host_readw(ConstHostPt off) { return __builtin_bswap16(*(uint16_t *)off); } static INLINE uint32_t host_readd(ConstHostPt off) { return __builtin_bswap32(*(uint32_t *)off); } static INLINE void host_writew(HostPt const off, const uint16_t val) { *(uint16_t *)off = __builtin_bswap16(val); } static INLINE void host_writed(HostPt const off, const uint32_t val) { *(uint32_t *)off = __builtin_bswap32(val); } static INLINE void host_writeq(HostPt const off, const uint64_t val) { *(uint64_t *)off = __builtin_bswap64(val); } #elif !defined(C_UNALIGNED_MEMORY) /* !defined(C_UNALIGNED_MEMORY) meaning: we're probably being compiled for a processor that doesn't like unaligned WORD access, on such processors typecasting memory as uint16_t and higher can cause a fault if the address is not aligned to that datatype when we read/write through it. */ static INLINE uint16_t host_readw(ConstHostPt const off) { return (uint16_t)host_readb(off) + ((uint16_t)host_readb(off+(uintptr_t)1U) << (uint16_t)8U); } static INLINE uint32_t host_readd(ConstHostPt const off) { return (uint32_t)host_readw(off) + ((uint32_t)host_readw(off+(uintptr_t)2U) << (uint32_t)16U); } static INLINE uint64_t host_readq(ConstHostPt const off) { return (uint64_t)host_readd(off) + ((uint64_t)host_readd(off+(uintptr_t)4U) << (uint64_t)32U); } static INLINE void host_writew(HostPt const off,const uint16_t val) { host_writeb(off, (uint8_t)(val)); host_writeb(off+1U,(uint8_t)(val >> (uint16_t)8U)); } static INLINE void host_writed(HostPt const off,const uint32_t val) { host_writew(off, (uint16_t)(val)); host_writew(off+2U,(uint16_t)(val >> (uint32_t)16U)); } static INLINE void host_writeq(HostPt const off,const uint64_t val) { host_writed(off, (uint32_t)(val)); host_writed(off+4U,(uint32_t)(val >> (uint64_t)32U)); } #else static INLINE uint16_t host_readw(ConstHostPt const off) { return le16toh((*(const uint16_t *)off)); // BSD endian.h } static INLINE uint32_t host_readd(ConstHostPt const off) { return le32toh((*(const uint32_t *)off)); // BSD endian.h } static INLINE uint64_t host_readq(ConstHostPt const off) { return le64toh((*(const uint64_t *)off)); // BSD endian.h } static INLINE void host_writew(HostPt const off,const uint16_t val) { *(uint16_t *)(off) = htole16(val); } static INLINE void host_writed(HostPt const off,const uint32_t val) { *(uint32_t *)(off) = htole32(val); } static INLINE void host_writeq(HostPt const off,const uint64_t val) { *(uint64_t *)(off) = htole64(val); } #endif static INLINE void var_write(uint8_t * const var, const uint8_t val) { host_writeb(var, val); } static INLINE void var_write(uint16_t * const var, const uint16_t val) { host_writew((HostPt)var, val); } static INLINE void var_write(uint32_t * const var, const uint32_t val) { host_writed((HostPt)var, val); } static INLINE void var_write(uint64_t * const var, const uint64_t val) { host_writeq((HostPt)var, val); } static INLINE uint16_t var_read(uint16_t * var) { return host_readw((ConstHostPt)var); } static INLINE uint32_t var_read(uint32_t * var) { return host_readd((ConstHostPt)var); } /* The Following six functions are slower but they recognize the paged memory system */ uint8_t mem_readb(const LinearPt address); uint16_t mem_readw(const LinearPt address); uint32_t mem_readd(const LinearPt address); void mem_writeb(const LinearPt address,const uint8_t val); void mem_writew(const LinearPt address,const uint16_t val); void mem_writed(const LinearPt address,const uint32_t val); void phys_writes(PhysPt addr, const char* string, Bitu length); /* WARNING: These will cause a segfault or out of bounds access IF * addr is beyond the end of memory */ /* 2024/12/22: Looking across the DOSBox-X codebase, these functions * aren't used TOO often, and where they are used, some * code has memory range checks anyway. So it doesn't hurt * performance very much if at all to just put the range * check here instead, in order not to segfault if somehow * the address given is beyond end of system memory. --J.C. * * There is one more important detail here. These functions * take only a 32-bit physical address. Which means if more * than 32 address bits are enabled on the CPU and the OS * has PSE/PAE page tables enabled, these functions will not * be able to reach above 4GB. Given how memory will be * segmented between the 'below 4GB' and 'above 4GB' regions, * if emulating 4GB or more, that is perfectly fine. S3 XGA * and ISA DMA emulation will never reach above 4GB anyway. * * The way the range check is done is ideal for performance, * yet may fail to work correctly if MemSize is very close * to zero, low enough that subtraction would cause it to * wrap back around to the largest possible value. The code, * when MemBase is a valid pointer, will never set MemSize * that small. */ /* NTS: Technically these phys_ functions are misnamed. They don't read/write * all physically addressable memory and MMIO attached to the CPU. These * functions can only address system RAM. So perhaps they should be named * physmem_read/physmem_write instead and a physmmio_read/physmmio_write * function should be added that addresses the "physical" memory addresses * accessible to the CPU. That hackery in the debugger to dump by physical * memory addresse could be a useful guide on how to do that. --J.C. */ static INLINE void phys_writeb(const PhysPt addr,const uint8_t val) { if (addr < MemSize) host_writeb(MemBase+addr,val); } static INLINE void phys_writew(const PhysPt addr,const uint16_t val) { if (addr < (MemSize-1u)) host_writew(MemBase+addr,val); } static INLINE void phys_writed(const PhysPt addr,const uint32_t val) { if (addr < (MemSize-3u)) host_writed(MemBase+addr,val); } static INLINE uint8_t phys_readb(const PhysPt addr) { if (addr < MemSize) return host_readb(MemBase+addr); else return 0xFF; } static INLINE uint16_t phys_readw(const PhysPt addr) { if (addr < (MemSize-1u)) return host_readw(MemBase+addr); else return 0xFFFFu; } static INLINE uint32_t phys_readd(const PhysPt addr) { if (addr < (MemSize-3u)) return host_readd(MemBase+addr); else return 0xFFFFFFFFu; } /* These don't check for alignment, better be sure it's correct */ void MEM_BlockWrite(LinearPt pt, const void *data, size_t size); void MEM_BlockRead(LinearPt pt,void * data,Bitu size); void MEM_BlockWrite32(LinearPt pt,void * data,Bitu size); void MEM_BlockRead32(LinearPt pt,void * data,Bitu size); void MEM_BlockCopy(LinearPt dest,LinearPt src,Bitu size); void MEM_StrCopy(LinearPt pt,char * data,Bitu size); void mem_memcpy(LinearPt dest,LinearPt src,Bitu size); Bitu mem_strlen(LinearPt pt); void mem_strcpy(LinearPt dest,LinearPt src); /* The following functions are all shortcuts to the above functions using physical addressing */ static inline constexpr LinearPt PhysMake(const uint16_t seg,const uint16_t off) { return ((LinearPt)seg << 4U) + (LinearPt)off; } static inline constexpr uint16_t RealSeg(const RealPt pt) { return (uint16_t)(pt >> 16U); } static inline constexpr uint16_t RealOff(const RealPt pt) { return (uint16_t)(pt & 0xffffu); } static inline constexpr LinearPt Real2Phys(const RealPt pt) { return (((LinearPt)RealSeg(pt) << 4U) + (LinearPt)RealOff(pt)); } static inline constexpr RealPt RealMake(const uint16_t seg,const uint16_t off) { return (((RealPt)seg << 16U) + (RealPt)off); } /* convert physical address to 4:16 real pointer (example: 0xABCDE -> 0xA000:0xBCDE) */ static inline constexpr RealPt PhysToReal416(const LinearPt phys) { return RealMake((uint16_t)((phys >> 4U) & 0xF000U),(uint16_t)(phys & 0xFFFFU)); } static inline constexpr LinearPt RealVecAddress(const uint8_t vec) { return ((unsigned int)vec << 2U); } static INLINE uint8_t real_readb(const uint16_t seg,const uint16_t off) { return mem_readb(PhysMake(seg,off)); } static INLINE uint16_t real_readw(const uint16_t seg,const uint16_t off) { return mem_readw(PhysMake(seg,off)); } static INLINE uint32_t real_readd(const uint16_t seg,const uint16_t off) { return mem_readd(PhysMake(seg,off)); } static INLINE void real_writeb(const uint16_t seg,const uint16_t off,const uint8_t val) { mem_writeb(PhysMake(seg,off),val); } static INLINE void real_writew(const uint16_t seg,const uint16_t off,const uint16_t val) { mem_writew(PhysMake(seg,off),val); } static INLINE void real_writed(const uint16_t seg,const uint16_t off,const uint32_t val) { mem_writed(PhysMake(seg,off),val); } static INLINE RealPt RealGetVec(const uint8_t vec) { return mem_readd(RealVecAddress(vec)); } static INLINE void RealSetVec(const uint8_t vec,const RealPt pt) { mem_writed(RealVecAddress(vec),pt); } static INLINE void RealSetVec(const uint8_t vec,const RealPt pt,RealPt &old) { const LinearPt addr = RealVecAddress(vec); old = mem_readd(addr); mem_writed(addr,pt); } uint8_t physdev_readb(const PhysPt64 addr); uint16_t physdev_readw(const PhysPt64 addr); uint32_t physdev_readd(const PhysPt64 addr); void physdev_writeb(const PhysPt64 addr,const uint8_t val); void physdev_writew(const PhysPt64 addr,const uint16_t val); void physdev_writed(const PhysPt64 addr,const uint32_t val); uint32_t MEM_get_address_bits(); uint32_t MEM_get_address_bits4GB(); void MEM_ResetPageHandler_Unmapped(Bitu phys_page, Bitu pages); #endif