Files
metadesk/code/base/memory.h
T
2025-02-08 22:36:11 -05:00

981 lines
27 KiB
C

#ifdef INTELLISENSE_DIRECTIVES
# pragma once
# include "context_cracking.h"
# include "linkage.h"
# include "macros.h"
# include "platform.h"
# include "base_types.h"
#endif
////////////////////////////////
//~ rjf: Units
#ifndef KILOBTYES
#define KILOBYTES( x ) ( ( x ) * ( S64 )( 1024 ) )
#endif
#ifndef MEGABYTES
#define MEGABYTES( x ) ( MD_KILOBYTES( x ) * ( S64 )( 1024 ) )
#endif
#ifndef GIGABYTES
#define GIGABYTES( x ) ( MD_MEGABYTES( x ) * ( S64 )( 1024 ) )
#endif
#ifndef TERABYTES
#define TERABYTES( x ) ( MD_GIGABYTES( x ) * ( S64 )( 1024 ) )
#endif
#ifndef KB
#define KB(n) (((U64)(n)) << 10)
#endif
#ifndef MB
#define MB(n) (((U64)(n)) << 20)
#endif
#ifndef GB
#define GB(n) (((U64)(n)) << 30)
#endif
#ifndef TB
#define TB(n) (((U64)(n)) << 40)
#endif
#ifndef thosuand
#define thousand(n) ((n) * 1000)
#endif
#ifndef million
#define million(n) ((n) * 1000000)
#endif
#ifndef billion
#define billion(n) ((n) * 1000000000)
#endif
////////////////////////////////
//~ rjf: Type -> Alignment
#ifndef align_of
# if COMPILER_MSVC
# define align_of(T) __alignof(T)
# elif COMPILER_CLANG
# define align_of(T) __alignof(T)
# elif COMPILER_GCC
# define align_of(T) __alignof__(T)
# else
# error AlignOf not defined for this compiler.
# endif
#endif
////////////////////////////////
//~ rjf: Member Offsets
#ifndef membeMD_DYN_LINKr
#define member(T, m) ( ((T*) 0)->m )
#endif
#ifndef offset_of
#define offset_of(T, m) int_from_ptr(& member(T, m))
#endif
#ifndef member_from_offset
#define member_from_offset(T, ptr, off) (T) ((((U8 *) ptr) + (off)))
#endif
#ifndef cast_from_member
#define cast_from_member(T, m, ptr) (T*) (((U8*)ptr) - offset_of(T, m))
#endif
////////////////////////////////
//~ rjf: Memory Operation Macros
// TODO(Ed): Review usage of memmove here...(I guess wanting to avoid overlap faults..)
#ifndef memory_copy
# if USE_VENDOR_MEMORY_OPS
# define memory_copy(dst, src, size) memmove((dst), (src), (size))
# else
# define memory_copy(dst, src, size) mem_move((dst), (src), (size))
#endif
#endif
#ifndef memory_set
# if USE_VENDOR_MEMORY_OPS
# define memory_set(dst, byte, size) memset((dst), (byte), (size))
# else
# define memory_set(dst, byte, size) mem_set((dst), (byte), (size))
# endif
#endif
#ifndef memory_compare
#define memory_compare(a, b, size) memcmp((a), (b), (size))
#endif
#ifndef memory_str_len
#define memory_str_len(ptr) cstr_len(ptr)
#endif
#ifndef memory_copy_struct
#define memory_copy_struct(d, s) memory_copy((d), (s), sizeof( *(d)))
#endif
#ifndef memory_copy_array
#define memory_copy_array(d, s) memory_copy((d), (s), sizeof( d))
#endif
#ifndef memory_copy_type
#define memory_copy_type(d, s, c) memory_copy((d), (s), sizeof( *(d)) * (c))
#endif
#ifndef memory_zero
#define memory_zero(s,z) memory_set((s), 0, (z))
#endif
#ifndef memory_zero_struct
#define memory_zero_struct(s) memory_zero((s), sizeof( *(s)))
#endif
#ifndef memory_zero_array
#define memory_zero_array(a) memory_zero((a), sizeof(a))
#endif
#ifndef memory_zero_type
#define memory_zero_type(m, c) memory_zero((m), sizeof( *(m)) * (c))
#endif
#ifndef memory_match
#define memory_match(a, b, z) (memory_compare((a), (b), (z)) == 0)
#endif
#ifndef memory_match_struct
#define memory_match_struct(a, b) memory_match((a), (b), sizeof(*(a)))
#endif
#ifndef memory_match_array
#define memory_match_array(a, b) memory_match((a), (b), sizeof(a))
#endif
#ifndef memory_read
#define memory_read(T, p, e) ( ((p) + sizeof(T) <= (e)) ? ( *(T*)(p)) : (0) )
#endif
#ifndef memory_consume
#define memory_consume(T, p, e) ( ((p) + sizeof(T) <= (e)) ? ((p) += sizeof(T), *(T*)((p) - sizeof(T))) : ((p) = (e),0) )
#endif
////////////////////////////////
//~ rjf: Memory Functions
inline B32
memory_is_zero(void* ptr, U64 size)
{
B32 result = 1;
// break down size
U64 extra = (size & 0x7);
U64 count8 = (size >> 3);
// check with 8-byte stride
U64* p64 = (U64*)ptr;
if (result)
{
for (U64 i = 0; i < count8; i += 1, p64 += 1) {
if (*p64 != 0){
result = 0;
goto done;
}
}
}
// check extra
if (result)
{
U8* p8 = (U8*)p64;
for (U64 i = 0; i < extra; i += 1, p8 += 1) {
if (*p8 != 0) {
result = 0;
goto done;
}
}
}
done:;
return(result);
}
inline
void* mem_move( void* destination, void const* source, SSIZE byte_count )
{
if ( destination == NULL )
{
return NULL;
}
U8* dest_ptr = rcast( U8*, destination);
U8 const* src_ptr = rcast( U8 const*, source);
if ( dest_ptr == src_ptr )
return dest_ptr;
if ( src_ptr + byte_count <= dest_ptr || dest_ptr + byte_count <= src_ptr ) // NOTE: Non-overlapping
return memory_copy( dest_ptr, src_ptr, byte_count );
if ( dest_ptr < src_ptr )
{
if ( to_uptr(src_ptr) % size_of( SSIZE ) == to_uptr(dest_ptr) % size_of( SSIZE ) )
{
while ( pcast( UPTR, dest_ptr) % size_of( SSIZE ) )
{
if ( ! byte_count-- )
return destination;
*dest_ptr++ = *src_ptr++;
}
while ( byte_count >= size_of( SSIZE ) )
{
* rcast(SSIZE*, dest_ptr) = * rcast(SSIZE const*, src_ptr);
byte_count -= size_of( SSIZE );
dest_ptr += size_of( SSIZE );
src_ptr += size_of( SSIZE );
}
}
for ( ; byte_count; byte_count-- )
*dest_ptr++ = *src_ptr++;
}
else
{
if ( ( to_uptr(src_ptr) % size_of( SSIZE ) ) == ( to_uptr(dest_ptr) % size_of( SSIZE ) ) )
{
while ( to_uptr( dest_ptr + byte_count ) % size_of( SSIZE ) )
{
if ( ! byte_count-- )
return destination;
dest_ptr[ byte_count ] = src_ptr[ byte_count ];
}
while ( byte_count >= size_of( SSIZE ) )
{
byte_count -= size_of( SSIZE );
* rcast(SSIZE*, dest_ptr + byte_count ) = * rcast( SSIZE const*, src_ptr + byte_count );
}
}
while ( byte_count )
byte_count--, dest_ptr[ byte_count ] = src_ptr[ byte_count ];
}
return destination;
}
inline
void* mem_set( void* destination, U8 fill_byte, SSIZE byte_count )
{
if ( destination == NULL )
{
return NULL;
}
SSIZE align_offset;
U8* dest_ptr = rcast( U8*, destination);
U32 fill_word = ( ( U32 )-1 ) / 255 * fill_byte;
if ( byte_count == 0 )
return destination;
dest_ptr[ 0 ] = dest_ptr[ byte_count - 1 ] = fill_byte;
if ( byte_count < 3 )
return destination;
dest_ptr[ 1 ] = dest_ptr[ byte_count - 2 ] = fill_byte;
dest_ptr[ 2 ] = dest_ptr[ byte_count - 3 ] = fill_byte;
if ( byte_count < 7 )
return destination;
dest_ptr[ 3 ] = dest_ptr[ byte_count - 4 ] = fill_byte;
if ( byte_count < 9 )
return destination;
align_offset = -to_sptr( dest_ptr ) & 3;
dest_ptr += align_offset;
byte_count -= align_offset;
byte_count &= -4;
* rcast( U32*, ( dest_ptr + 0 ) ) = fill_word;
* rcast( U32*, ( dest_ptr + byte_count - 4 ) ) = fill_word;
if ( byte_count < 9 )
return destination;
* rcast( U32*, dest_ptr + 4 ) = fill_word;
* rcast( U32*, dest_ptr + 8 ) = fill_word;
* rcast( U32*, dest_ptr + byte_count - 12 ) = fill_word;
* rcast( U32*, dest_ptr + byte_count - 8 ) = fill_word;
if ( byte_count < 25 )
return destination;
* rcast( U32*, dest_ptr + 12 ) = fill_word;
* rcast( U32*, dest_ptr + 16 ) = fill_word;
* rcast( U32*, dest_ptr + 20 ) = fill_word;
* rcast( U32*, dest_ptr + 24 ) = fill_word;
* rcast( U32*, dest_ptr + byte_count - 28 ) = fill_word;
* rcast( U32*, dest_ptr + byte_count - 24 ) = fill_word;
* rcast( U32*, dest_ptr + byte_count - 20 ) = fill_word;
* rcast( U32*, dest_ptr + byte_count - 16 ) = fill_word;
align_offset = 24 + to_uptr( dest_ptr ) & 4;
dest_ptr += align_offset;
byte_count -= align_offset;
{
U64 fill_doubleword = ( scast( U64, fill_word) << 32 ) | fill_word;
while ( byte_count > 31 )
{
* rcast( U64*, dest_ptr + 0 ) = fill_doubleword;
* rcast( U64*, dest_ptr + 8 ) = fill_doubleword;
* rcast( U64*, dest_ptr + 16 ) = fill_doubleword;
* rcast( U64*, dest_ptr + 24 ) = fill_doubleword;
byte_count -= 32;
dest_ptr += 32;
}
}
return destination;
}
////////////////////////////////
//~ rjf: Atomic Operations
#ifndef ins_atomic_u64_eval
# if OS_WINDOWS
# if ARCH_X64
# define ins_atomic_u64_eval(x) InterlockedAdd64((volatile __int64 *)(x), 0)
# define ins_atomic_u64_inc_eval(x) InterlockedIncrement64((volatile __int64 *)(x))
# define ins_atomic_u64_dec_eval(x) InterlockedDecrement64((volatile __int64 *)(x))
# define ins_atomic_u64_eval_assign(x,c) InterlockedExchange64((volatile __int64 *)(x), (c))
# define ins_atomic_u64_add_eval(x,c) InterlockedAdd64((volatile __int64 *)(x), c)
# define ins_atomic_u64_eval_cond_assign(x,k,c) InterlockedCompareExchange64((volatile __int64 *)(x), (k), (c))
# define ins_atomic_u32_eval(x,c) InterlockedAdd((volatile LONG *)(x), 0)
# define ins_atomic_u32_eval_assign(x,c) InterlockedExchange((volatile LONG *)(x), (c))
# define ins_atomic_u32_eval_cond_assign(x,k,c) InterlockedCompareExchange((volatile LONG *)(x), (k), (c))
# define ins_atomic_ptr_eval_assign(x,c) (void*) ins_atomic_u64_eval_assign((volatile __int64 *)(x), (__int64)(c))
# else
# error Atomic intrinsics not defined for this operating system / architecture combination.
# endif
# elif OS_LINUX
# if ARCH_X64
# define ins_atomic_u64_inc_eval(x) __sync_fetch_and_add((volatile U64 *)(x), 1)
# else
# error Atomic intrinsics not defined for this operating system / architecture combination.
# endif
# else
# error Atomic intrinsics not defined for this operating system.
# endif
#endif
////////////////////////////////
//~ rjf: Linked List Building Macros
//- rjf: linked list macro helpers
#ifndef check_nil
#define check_nil(nil, p) ((p) == 0 || (p) == nil)
#endif
#ifndef set_nil
#define set_nil(nil, p) ((p) = nil)
#endif
//- rjf: doubly-linked-lists
#ifndef MD_LINKED_LIST_PURE_MACRO
#define MD_LINKED_LIST_PURE_MACRO 0
#endif
#ifndef dll_insert_npz
// TODO(Ed): Review...
inline void
dll__insert_npz(
void* nil,
void** f, void** f_prev,
void** l, void** l_next,
void* p, void** p_next, void** p_next_prev,
void* n, void** n_prev, void** n_next
)
{
if (*f == nil) {
*f = n;
*l = n;
*n_prev = nil;
*n_next = nil;
}
else
{
if (p == nil) {
*n_next = *f;
*f_prev = n;
*f = n;
*n_prev = nil;
}
else
{
if (p == *l) {
*l_next = n;
*n_prev = *l;
*l = n;
*n_next = nil;
}
else
{
if ( ! (! (p == nil) && *p_next == nil)) {
*p_next_prev = n;
}
*n_next = *p_next;
*p_next = n;
*n_prev = p;
}
}
}
}
#if ! MD_LINKED_LIST_PURE_MACRO
// insert next-previous with nil
#define dll_insert_npz(nil, f, l, p, n, next, prev) dll__insert_npz(nil, &f, &f->prev, &l, &l->next, p, &p->next, &p->next->prev, n, &n->prev, &n->next)
#else
// insert next-previous with nil
#define dll_insert_npz(nil, f, l, p, n, next, prev) \
( \
check_nil(nil, f) ? ( \
(f) = (l) = (n), \
set_nil(nil, (n)->next), \
set_nil(nil, (n)->prev) \
) \
: ( \
check_nil(nil, p) ? ( \
(n)->next = (f), \
(f)->prev = (n), \
(f) = (n), \
set_nil(nil,(n)->prev) \
) \
: ((p) == (l)) ? ( \
(l)->next = (n), \
(n)->prev = (l), \
(l) = (n), \
set_nil(nil, (n)->next) \
) \
: ( \
( \
( ! check_nil(nil, p) && check_nil(nil, (p)->next) ) ? \
(0) \
: ( (p)->next->prev = (n) ) \
), \
((n)->next = (p)->next), \
((p)->next = (n)), \
((n)->prev = (p)) \
) \
) \
)
// ! MD_LINKED_LIST_PURE_MACRO
#endif
// dll_insert_npz
#endif
#ifndef dll_push_back_npz
// push-back next-previous with nil
#define dll_push_back_npz(nil, f, l, n, next, prev) dll_insert_npz(nil, f, l, l, n, next, prev)
#endif
#ifndef dll_push_front_npz
// push-fornt next-previous with nil
#define dll_push_front_npz(nil, f, l, n, next, prev) dll_insert_npz(nil, l, f, f, n, prev, next)
#endif
#ifndef dll_remove_npz
inline void
dll__remove_npz(
void* nil,
void** f,
void** l, void* l_prev,
void* n, void* n_next, void** n_next_prev,
void* n_prev, void** n_prev_next
)
{
if (n == *f) {
*f = n_next;
}
if (n == *l) {
*l = l_prev;
}
if (n_prev != nil) {
*n_prev_next = n_next;
}
if (n_next != nil) {
*n_next_prev = n_prev;
}
}
#if ! MD_LINKED_LIST_PURE_MACRO
// remove next-previous with nil
#define dll_remove_npz(nil, f, l, n, next, prev) dll__remove_npz(nil, &f, &l, l->prev, n, n->next, &n->next->prev, n->prev, &n->prev->next)
#else
// remove next-previous with nil
#define dll_remove_npz(nil, f, l, n, next, prev) \
( \
( \
(n) == (f) ? \
(f) = (n)->next \
: (0) \
), \
( \
(n) == (l) ? \
(l) = (l)->prev \
: (0) \
), \
( \
check_nil(nil,(n)->prev) ? \
(0) \
: ((n)->prev->next = (n)->next) \
), \
( \
check_nil(nil,(n)->next) ? \
(0) \
: ((n)->next->prev = (n)->prev) \
) \
)
// ! MD_LINKED_LIST_PURE_MACRO
#endif
// dll_remove_npz
#endif
//- rjf: singly-linked, doubly-headed lists (queues)
#ifndef sll_queue_push_nz
inline void
sll__queue_push_nz(
void* nil,
void** f,
void** l, void** l_next,
void* n, void** n_next
)
{
if (*f == nil) {
*f = n;
*n_next = nil;
}
else {
*l_next = n;
*l = n;
n_next = nil;
}
}
// queue-push next with nil
#if ! MD_LINKED_LIST_PURE_MACRO
#define sll_queue_push_nz(nil, f, l, n, next) sll__queue_push_nz(nil, &f, &l, &l->next, n, &n->next)
#else
#define sll_queue_push_nz(nil, f, l, n, next) \
( \
check_nil(nil, f) ? ( \
(f) = (l) = (n), \
set_nil(nil, (n)->next) \
) \
: ( \
(l)->next=(n), \
(l) = (n), \
set_nil(nil,(n)->next) \
) \
)
// ! MD_LINKED_LIST_PURE_MACRO
#endif
// sll_queue_push_nz
#endif
#ifndef sll_queue_push_front_nz
inline void
sll__queue_push_front_nz(void* nil, void** f, void** l, void* n, void** n_next) {
if (*f == nil) {
*f = n;
*l = n;
}
else {
*n_next = f;
*f = n;
}
}
// queue-push-front next with nil
#if ! MD_LINKED_LIST_PURE_MACRO
#define sll_queue_push_front_nz(nil, f, l, n, next) sll__queue_push_front_nz(nil, &f, &l, n, &n->next)
#else
#define sll_queue_push_front_nz(nil, f, l, n, next) \
( \
check_nil(nil, f) ? ( \
(f) = (l) = (n), \
set_nil(nil,(n)->next) \
) \
: ( \
(n)->next = (f), \
(f) = (n) \
) \
)
// ! MD_LINKED_LIST_PURE_MACRO
#endif
#endif
#ifndef sll_queue_pop_nz
inline void
sll__queue_pop_nz(void* nil, void** f, void* f_next, void** l)
{
if (*f == *l) {
*f = nil;
*l = nil;
}
else {
*f = f_next;
}
}
// queue-pop next with nil
#if ! MD_LINKED_LIST_PURE_MACRO
#define sll_queue_pop_nz(nil, f, l, next) sll__queue_pop_nz(nil, &f, f->next, &l)
#else
#define sll_queue_pop_nz(nil, f, l, next) \
( \
(f) == (l) ? ( \
set_nil(nil,f), \
set_nil(nil,l) \
) \
: ( \
(f)=(f)->next \
) \
)
// ! MD_LINKED_LIST_PURE_MACRO
#endif
// sll_queue_pop_nz
#endif
//- rjf: singly-linked, singly-headed lists (stacks)
#ifndef sll_stack_push_n
#define sll_stack_push_n(f,n,next) ( (n)->next = (f), (f) = (n) )
#endif
#ifndef sll_stack_pop_n
#define sll_stack_pop_n(f,next) ( (f) = (f)->next )
#endif
//- rjf: doubly-linked-list helpers
#ifndef dll_insert_np
#define dll_insert_np(f, l, p, n, next, prev) dll_insert_npz (0, f, l, p, n, next, prev)
#endif
#ifndef dll_push_back_np
#define dll_push_back_np(f, l, n, next, prev) dll_push_back_npz (0, f, l, n, next, prev)
#endif
#ifndef dll_push_front_np
#define dll_push_front_np(f, l, n, next, prev) dll_push_front_npz(0, f, l, n, next, prev)
#endif
#ifndef dll_remove_np
#define dll_remove_np(f, l, n, next, prev) dll_remove_npz (0, f, l, n, next, prev)
#endif
#ifndef dll_insert
#define dll_insert(f, l, p, n) dll_insert_npz (0, f, l, p, n, next, prev)
#endif
#ifndef dll_push_back
#define dll_push_back(f, l, n) dll_push_back_npz (0, f, l, n, next, prev)
#endif
#ifndef dll_push_front
#define dll_push_front(f, l, n) dll_push_front_npz(0, f, l, n, next, prev)
#endif
#ifndef dll_remove
#define dll_remove(f, l, n) dll_remove_npz (0, f, l, n, next, prev)
#endif
//- rjf: singly-linked, doubly-headed list helpers
#ifndef sll_queue_push_n
#define sll_queue_push_n(f, l, n, next) sll_queue_push_nz (0, f, l, n, next)
#endif
#ifndef sll_queue_push_front_n
#define sll_queue_push_front_n(f, l, n, next) sll_queue_push_front_nz(0, f, l, n, next)
#endif
#ifndef sll_queue_pop_n
#define sll_queue_pop_n(f, l, next) sll_queue_pop_nzs (0, f, l, next)
#endif
#ifndef sll_queue_push
#define sll_queue_push(f, l, n) sll_queue_push_nz (0, f, l, n, next)
#endif
#ifndef sll_queue_push_front
#define sll_queue_push_front(f, l ,n) sll_queue_push_front_nz(0, f, l, n, next)
#endif
#ifndef sll_queue_pop
#define sll_queue_pop(f, l) sll_queue_pop_nz (0, f, l, next)
#endif
//- rjf: singly-linked, singly-headed list helpers
#ifndef sll_stack_push
#define sll_stack_push(f, n) sll_stack_push_n(f, n, next)
#endif
#ifndef sll_stack_pop
#define sll_stack_pop(f) sll_stack_pop_n (f, next)
#endif
////////////////////////////////
//~ rjf: Address Sanitizer Markup
#ifndef NO_ASAN
# if COMPILER_MSVC
# if defined(__SANITIZE_ADDRESS__)
# define ASAN_ENABLED 1
# define NO_ASAN __declspec(no_sanitize_address)
# else
# define NO_ASAN
# endif
# elif COMPILER_CLANG
# if defined(__has_feature)
# if __has_feature(address_sanitizer) || defined(__SANITIZE_ADDRESS__)
# define ASAN_ENABLED 1
# endif
# endif
# define NO_ASAN __attribute__((no_sanitize("address")))
# else
# define NO_ASAN
# endif
#endif
#ifndef asan_poison_memory_region
# if MD_ASAN_ENABLED
# pragma comment(lib, "clang_rt.asan-x86_64.lib")
MD_C_API void __asan_poison_memory_region(void const volatile *addr, size_t size);
MD_C_API void __asan_unpoison_memory_region(void const volatile *addr, size_t size);
# define asan_poison_memory_region(addr, size) __asan_poison_memory_region((addr), (size))
# define asan_unpoison_memory_region(addr, size) __asan_unpoison_memory_region((addr), (size))
# else
# define asan_poison_memory_region(addr, size) ((void)(addr), (void)(size))
# define asan_unpoison_memory_region(addr, size) ((void)(addr), (void)(size))
# endif
#endif
////////////////////////////////
//~ rjf: Misc. Helper Macros
#ifndef array_count
#define array_count(a) (sizeof(a) / sizeof((a)[0]))
#endif
#ifndef ceil_integer_div
#define ceil_integer_div(a,b) (((a) + (b) - 1) / (b))
#endif
#ifndef swap
#define swap(T, a, b) do { T t__ = a; a = b; b = t__; } while(0)
#endif
#ifndef int_from_ptr
# if ARCH_64BIT
# define int_from_ptr(ptr) ((U64)(ptr))
# elif ARCH_32BIT
# define int_from_ptr(ptr) ((U32)(ptr))
# else
# error Missing pointer-to-integer cast for this architecture.
# endif
#endif
#ifndef ptr_from_int
#define ptr_from_int(i) (void*)((U8*)0 + (i))
#endif
#ifndef compose_64bit
#define compose_64bit(a,b) ((((U64)a) << 32) | ((U64)b));
#endif
#ifndef align_pow2
#define align_pow2(x,b) (((x) + (b) - 1) & ( ~((b) - 1)))
#endif
#ifndef align_down_pow2
#define align_down_pow2(x,b) ((x) & (~((b) - 1)))
#endif
#ifndef align_pad_pow2
#define align_pad_pow2(x,b) ((0-(x)) & ((b) - 1))
#endif
#ifndef is_pow2
#define is_pow2(x) ((x) != 0 && ((x ) & ((x) - 1)) == 0)
#endif
#ifndef is_pow2_or_zero
#define is_pow2_or_zero(x) ((((x) - 1) & (x)) == 0)
#endif
#ifndef extract_bit
#define extract_bit(word, idx) (((word) >> (idx)) & 1)
#endif
#ifndef zero_struct
# if LANG_CPP
# define zero_struct {}
# else
# define zero_struct {0}
# endif
#endif
#ifndef this_function_name
# if COMPILER_MSVC && COMPILER_MSVC_YEAR < 2015
# define this_function_name "unknown"
# else
# define this_function_name __func__
# endif
#endif
#ifndef read_only
# if COMPILER_MSVC || (COMPILER_CLANG && OS_WINDOWS)
# pragma section(".rdata$", read)
# define read_only __declspec(allocate(".rdata$"))
# elif (COMPILER_CLANG && OS_LINUX)
# define read_only __attribute__((section(".rodata")))
# else
// NOTE(rjf): I don't know of a useful way to do this in GCC land.
// __attribute__((section(".rodata"))) looked promising, but it introduces a
// strange warning about malformed section attributes, and it doesn't look
// like writing to that section reliably produces access violations, strangely
// enough. (It does on Clang)
# define read_only
# endif
#endif
#ifndef local_persist
#define local_persist static
#endif
#if COMPILER_MSVC
# define thread_static __declspec(thread)
#elif COMPILER_CLANG || COMPILER_GCC
# define thread_static __thread
#endif
#if COMPILER_CPP
// Already Defined
#elif COMPILER_C && __STDC_VERSION__ >= 201112L
# define thread_local _Thread_local
#elif COMPILER_MSVC
# define thread_local __declspec(thread)
#elif COMPILER_CLANG
# define thread_local __thread
#else
# error "No thread local support"
#endif
////////////////////////////////
//~ rjf: Safe Casts
inline U16
safe_cast_u16(U32 x) {
assert_always(x <= MAX_U16);
U16 result = (U16)x;
return result;
}
inline U32
safe_cast_u32(U64 x) {
assert_always(x <= MAX_U32);
U32 result = (U32)x;
return result;
}
inline S32
safe_cast_s32(S64 x) {
assert_always(x <= MAX_S32);
S32 result = (S32)x;
return result;
}
////////////////////////////////
//~ rjf: Large Base Type Functions
inline U128 u128_zero (void) { U128 v = {0}; return v; }
inline U128 u128_make (U64 v0, U64 v1) { U128 v = {v0, v1}; return v; }
inline B32 u128_match(U128 a, U128 b) { return memory_match_struct(&a, &b); }
////////////////////////////////
//~ rjf: Bit Patterns
inline U32 u32_from_u64_saturate(U64 x) { U32 x32 = (x > MAX_U32) ? MAX_U32 : (U32)x; return(x32); }
inline U64
u64_up_to_pow2(U64 x) {
if (x == 0) {
x = 1;
}
else {
x -= 1;
x |= (x >> 1);
x |= (x >> 2);
x |= (x >> 4);
x |= (x >> 8);
x |= (x >> 16);
x |= (x >> 32);
x += 1;
}
return(x);
}
inline S32
extend_sign32(U32 x, U32 size) {
U32 high_bit = size * 8;
U32 shift = 32 - high_bit;
S32 result = ((S32)x << shift) >> shift;
return result;
}
inline S64
extend_sign64(U64 x, U64 size) {
U64 high_bit = size * 8;
U64 shift = 64 - high_bit;
S64 result = ((S64)x << shift) >> shift;
return result;
}
inline F32 inf32 (void) { union { U32 u; F32 f; } x; x.u = EXPONENT32; return(x.f); }
inline F32 neg_inf32(void) { union { U32 u; F32 f; } x; x.u = SIGN32 | EXPONENT32; return(x.f); }
inline U16
bswap_u16(U16 x)
{
U16 result = (((x & 0xFF00) >> 8) |
((x & 0x00FF) << 8));
return result;
}
inline U32
bswap_u32(U32 x)
{
U32 result = (((x & 0xFF000000) >> 24) |
((x & 0x00FF0000) >> 8) |
((x & 0x0000FF00) << 8) |
((x & 0x000000FF) << 24));
return result;
}
inline U64
bswap_u64(U64 x)
{
// TODO(nick): naive bswap, replace with something that is faster like an intrinsic
U64 result = (((x & 0xFF00000000000000ULL) >> 56) |
((x & 0x00FF000000000000ULL) >> 40) |
((x & 0x0000FF0000000000ULL) >> 24) |
((x & 0x000000FF00000000ULL) >> 8) |
((x & 0x00000000FF000000ULL) << 8) |
((x & 0x0000000000FF0000ULL) << 24) |
((x & 0x000000000000FF00ULL) << 40) |
((x & 0x00000000000000FFULL) << 56));
return result;
}
#if ARCH_LITTLE_ENDIAN
# define from_be_u16(x) bswap_u16(x)
# define from_be_u32(x) bswap_u32(x)
# define from_be_u64(x) bswap_u64(x)
#else
# define from_be_u16(x) (x)
# define from_be_u32(x) (x)
# define from_be_u64(x) (x)
#endif
#if COMPILER_MSVC || (COMPILER_CLANG && OS_WINDOWS)
inline U64 count_bits_set16(U16 val) { return __popcnt16(val); }
inline U64 count_bits_set32(U32 val) { return __popcnt (val); }
inline U64 count_bits_set64(U64 val) { return __popcnt64(val); }
inline U64 ctz32(U32 mask) { unsigned long idx; _BitScanForward (&idx, mask); return idx; }
inline U64 ctz64(U64 mask) { unsigned long idx; _BitScanForward64(&idx, mask); return idx; }
inline U64 clz32(U32 mask) { unsigned long idx; _BitScanReverse (&idx, mask); return 31 - idx; }
inline U64 clz64(U64 mask) { unsigned long idx; _BitScanReverse64(&idx, mask); return 63 - idx; }
#elif COMPILER_CLANG || COMPILER_GCC
inline U64 count_bits_set16(U16 val) { NotImplemented; return 0; }
inline U64 count_bits_set32(U32 val) { NotImplemented; return 0; }
inline U64 count_bits_set64(U64 val) { NotImplemented; return 0; }
inline U64 ctz32(U32 val) { NotImplemented; return 0; }
inline U64 ctz64(U32 val) { NotImplemented; return 0; }
inline U64 clz32(U32 val) { NotImplemented; return 0; }
inline U64 clz64(U64 val) { NotImplemented; return 0; }
#else
# error "Bit intrinsic functions not defined for this compiler."
#endif