The idle thread is now actually run when the system is idle because it truly goes idle. The idle thread is made power efficient by using the hlt instruction rather than a busy loop. The new futex(2) system call is used to implement fast user-space mutexes, condition variables, and semaphores. The same backend and design is used as kutexes for truly sleeping kernel mutexes and condition variables. The new exit_thread(2) flag EXIT_THREAD_FUTEX_WAKE wakes a futex. Sleeping on clocks in the kernel now uses timers for true sleep. The interrupt worker thread now truly sleeps when idle. Kernel threads are now named. This is a compatible ABI change.
255 lines
7.5 KiB
C++
255 lines
7.5 KiB
C++
/*
|
|
* Copyright (c) 2012, 2014, 2021 Jonas 'Sortie' Termansen.
|
|
*
|
|
* Permission to use, copy, modify, and distribute this software for any
|
|
* purpose with or without fee is hereby granted, provided that the above
|
|
* copyright notice and this permission notice appear in all copies.
|
|
*
|
|
* THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
|
|
* WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
|
|
* MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
|
|
* ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
|
|
* WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
|
|
* ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
|
|
* OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
|
|
*
|
|
* copy.cpp
|
|
* The context for io operations: who made it, how should data be copied, etc.
|
|
*/
|
|
|
|
#include <assert.h>
|
|
#include <errno.h>
|
|
#include <stdint.h>
|
|
#include <string.h>
|
|
|
|
#include <sortix/mman.h>
|
|
|
|
#include <sortix/kernel/copy.h>
|
|
#include <sortix/kernel/kernel.h>
|
|
#include <sortix/kernel/kthread.h>
|
|
#include <sortix/kernel/process.h>
|
|
#include <sortix/kernel/segment.h>
|
|
#include <sortix/kernel/string.h>
|
|
|
|
// NOTE: The copy-from-user-space functions are specially made such that they
|
|
// guarantee the relevant memory exist and are unchanged during the copy
|
|
// operation. They do not protect against another thread concurrently
|
|
// modifying the memory, but the functions cope in that case and just
|
|
// transfer whatever memory they see in that case, they don't malfunction.
|
|
|
|
// TODO: We could check the page tables for extra safely.
|
|
|
|
namespace Sortix {
|
|
|
|
static bool IsInProcessAddressSpace(Process* process)
|
|
{
|
|
addr_t current_address_space;
|
|
#if defined(__i386__)
|
|
asm ( "mov %%cr3, %0" : "=r"(current_address_space) );
|
|
#elif defined(__x86_64__)
|
|
asm ( "mov %%cr3, %0" : "=r"(current_address_space) );
|
|
#else
|
|
#warning "You should set current_address_space for safety"
|
|
current_address_space = process->addrspace;
|
|
#endif
|
|
return current_address_space == process->addrspace;
|
|
}
|
|
|
|
static struct segment* FindSegment(Process* process, uintptr_t addr)
|
|
{
|
|
for ( size_t i = 0; i < process->segments_used; i++ )
|
|
{
|
|
struct segment* segment = &process->segments[i];
|
|
if ( addr < segment->addr )
|
|
continue;
|
|
if ( segment->addr + segment->size <= addr )
|
|
continue;
|
|
return segment;
|
|
}
|
|
return NULL;
|
|
}
|
|
|
|
bool CopyToUser(void* userdst_ptr, const void* ksrc_ptr, size_t count)
|
|
{
|
|
uintptr_t userdst = (uintptr_t) userdst_ptr;
|
|
uintptr_t ksrc = (uintptr_t) ksrc_ptr;
|
|
bool result = true;
|
|
Process* process = CurrentProcess();
|
|
assert(IsInProcessAddressSpace(process));
|
|
kthread_mutex_lock(&process->segment_lock);
|
|
while ( count )
|
|
{
|
|
struct segment* segment = FindSegment(process, userdst);
|
|
if ( !segment || !(segment->prot & PROT_WRITE) )
|
|
{
|
|
errno = EFAULT;
|
|
result = false;
|
|
break;
|
|
}
|
|
size_t amount = count;
|
|
size_t segment_available = segment->addr + segment->size - userdst;
|
|
if ( segment_available < amount )
|
|
amount = segment_available;
|
|
memcpy((void*) userdst, (const void*) ksrc, amount);
|
|
userdst += amount;
|
|
ksrc += amount;
|
|
count -= amount;
|
|
}
|
|
kthread_mutex_unlock(&process->segment_lock);
|
|
return result;
|
|
}
|
|
|
|
bool CopyFromUser(void* kdst_ptr, const void* usersrc_ptr, size_t count)
|
|
{
|
|
uintptr_t kdst = (uintptr_t) kdst_ptr;
|
|
uintptr_t usersrc = (uintptr_t) usersrc_ptr;
|
|
bool result = true;
|
|
Process* process = CurrentProcess();
|
|
assert(IsInProcessAddressSpace(process));
|
|
kthread_mutex_lock(&process->segment_lock);
|
|
while ( count )
|
|
{
|
|
struct segment* segment = FindSegment(process, usersrc);
|
|
if ( !segment || !(segment->prot & PROT_READ) )
|
|
{
|
|
errno = EFAULT;
|
|
result = false;
|
|
break;
|
|
}
|
|
size_t amount = count;
|
|
size_t segment_available = segment->addr + segment->size - usersrc;
|
|
if ( segment_available < amount )
|
|
amount = segment_available;
|
|
memcpy((void*) kdst, (const void*) usersrc, amount);
|
|
kdst += amount;
|
|
usersrc += amount;
|
|
count -= amount;
|
|
}
|
|
kthread_mutex_unlock(&process->segment_lock);
|
|
return result;
|
|
}
|
|
|
|
bool ReadAtomicFromUser(int* kdst_ptr, const int* usersrc_ptr)
|
|
{
|
|
uintptr_t usersrc = (uintptr_t) usersrc_ptr;
|
|
if ( usersrc & (sizeof(int) - 1) )
|
|
return errno = EINVAL, false;
|
|
Process* process = CurrentProcess();
|
|
assert(IsInProcessAddressSpace(process));
|
|
ScopedLock lock(&process->segment_lock);
|
|
struct segment* segment = FindSegment(process, usersrc);
|
|
if ( !segment || !(segment->prot & PROT_READ) )
|
|
return errno = EFAULT, false;
|
|
size_t segment_available = segment->addr + segment->size - usersrc;
|
|
if ( segment_available < sizeof(int) )
|
|
return errno = EFAULT, false;
|
|
*kdst_ptr = __atomic_load_n(usersrc_ptr, __ATOMIC_SEQ_CST);
|
|
return true;
|
|
}
|
|
|
|
bool CopyToKernel(void* kdst, const void* ksrc, size_t count)
|
|
{
|
|
memcpy(kdst, ksrc, count);
|
|
return true;
|
|
}
|
|
|
|
bool CopyFromKernel(void* kdst, const void* ksrc, size_t count)
|
|
{
|
|
memcpy(kdst, ksrc, count);
|
|
return true;
|
|
}
|
|
|
|
bool ZeroKernel(void* kdst, size_t count)
|
|
{
|
|
// TODO: We could check the page tables for extra safely.
|
|
memset(kdst, 0, count);
|
|
return true;
|
|
}
|
|
|
|
bool ZeroUser(void* userdst_ptr, size_t count)
|
|
{
|
|
uintptr_t userdst = (uintptr_t) userdst_ptr;
|
|
bool result = true;
|
|
Process* process = CurrentProcess();
|
|
assert(IsInProcessAddressSpace(process));
|
|
kthread_mutex_lock(&process->segment_lock);
|
|
while ( count )
|
|
{
|
|
struct segment* segment = FindSegment(process, userdst);
|
|
if ( !segment || !(segment->prot & PROT_WRITE) )
|
|
{
|
|
errno = EFAULT;
|
|
result = false;
|
|
break;
|
|
}
|
|
size_t amount = count;
|
|
size_t segment_available = segment->addr + segment->size - userdst;
|
|
if ( segment_available < amount )
|
|
amount = segment_available;
|
|
memset((void*) userdst, 0, amount);
|
|
userdst += amount;
|
|
count -= amount;
|
|
}
|
|
kthread_mutex_unlock(&process->segment_lock);
|
|
return result;
|
|
}
|
|
|
|
// NOTE: No overflow can happen here because the user can't make an infinitely
|
|
// long string spanning the entire address space because the user can't
|
|
// control the entire address space.
|
|
char* GetStringFromUser(const char* usersrc_str)
|
|
{
|
|
uintptr_t usersrc = (uintptr_t) usersrc_str;
|
|
size_t result_length = 0;
|
|
Process* process = CurrentProcess();
|
|
assert(IsInProcessAddressSpace(process));
|
|
|
|
kthread_mutex_lock(&process->segment_lock);
|
|
bool done = false;
|
|
while ( !done )
|
|
{
|
|
uintptr_t current_at = usersrc + result_length;
|
|
struct segment* segment = FindSegment(process, current_at);
|
|
if ( !segment || !(segment->prot & PROT_READ) )
|
|
{
|
|
kthread_mutex_unlock(&process->segment_lock);
|
|
return errno = EFAULT, (char*) NULL;
|
|
}
|
|
size_t segment_available = segment->addr + segment->size - current_at;
|
|
volatile const char* str = (volatile const char*) current_at;
|
|
size_t length = 0;
|
|
for ( ; length < segment_available; length++ )
|
|
{
|
|
char c = str[length];
|
|
if ( c == '\0' )
|
|
{
|
|
done = true;
|
|
break;
|
|
}
|
|
}
|
|
result_length += length;
|
|
}
|
|
|
|
char* result = new char[result_length + 1];
|
|
if ( !result )
|
|
{
|
|
kthread_mutex_unlock(&process->segment_lock);
|
|
return (char*) NULL;
|
|
}
|
|
|
|
memcpy(result, (const char*) usersrc, result_length);
|
|
result[result_length] = '\0';
|
|
|
|
// We have transferred a bunch of bytes from user-space and appended a zero
|
|
// byte. This is a string. If no concurrent threads were modifying the
|
|
// memory, this is the intended string. If the memory was modified, we got
|
|
// potential garbage followed by a NUL byte. This is a string, but probably
|
|
// not what was intended. If the garbage itself had a premature unexpected
|
|
// NUL byte, that's okay, the garbage string just got truncated.
|
|
|
|
kthread_mutex_unlock(&process->segment_lock);
|
|
return result;
|
|
}
|
|
|
|
} // namespace Sortix
|