eduOS/arch/x86/mm/page.c
2015-01-12 22:35:08 +01:00

562 lines
16 KiB
C

/*
* Copyright (c) 2010, Stefan Lankes, RWTH Aachen University
* 2014, Steffen Vogel, RWTH Aachen University
* All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions are met:
* * Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* * Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* * Neither the name of the University nor the names of its contributors
* may be used to endorse or promote products derived from this
* software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
* WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
* DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE FOR ANY
* DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
* (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
* LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
* ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
* SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
/**
* This is a 32/64 bit portable paging implementation for the x86 architecture
* using self-referenced page tables i.
* See http://www.noteblok.net/2014/06/14/bachelor/ for a detailed description.
*
* @author Steffen Vogel <steffen.vogel@rwth-aachen.de>
*/
#include <eduos/stdio.h>
#include <eduos/memory.h>
#include <eduos/errno.h>
#include <eduos/string.h>
#include <eduos/spinlock.h>
#include <asm/irq.h>
#include <asm/page.h>
#include <asm/multiboot.h>
/* Note that linker symbols are not variables, they have no memory
* allocated for maintaining a value, rather their address is their value. */
extern const void kernel_start;
//extern const void kernel_end;
/// This page is reserved for copying
#define PAGE_TMP (PAGE_FLOOR((size_t) &kernel_start) - PAGE_SIZE)
/** Lock for kernel space page tables */
static spinlock_t kslock = SPINLOCK_INIT;
/** A self-reference enables direct access to all page tables */
static size_t* self[PAGE_LEVELS] = {
(size_t *) 0xFFC00000,
(size_t *) 0xFFFFF000
};
/** An other self-reference for page_map_copy() */
static size_t * other[PAGE_LEVELS] = {
(size_t *) 0xFF800000,
(size_t *) 0xFFFFE000
};
/// Mapping of self referenced page map (at the end of the VAS)
// TODO: find a more generic initialization
static size_t* const current_map = (size_t*) (1 * 0xFFFFF000);
/** @brief Get the base address of the child table
*
* @param entry The parent entry
* @return The child entry
*/
static inline size_t* get_child_entry(size_t *entry)
{
size_t child = (size_t) entry;
child <<= PAGE_MAP_BITS;
return (size_t*) CANONICAL(child);
}
/** @brief Get the base address of the parent entry
*
* @param entry The child entry
* @return The parent entry
*/
static inline size_t* get_parent_entry(size_t *entry)
{
ssize_t parent = (size_t) entry;
parent >>= PAGE_MAP_BITS;
parent |= (size_t) self[0];
parent &= ~(sizeof(size_t) - 1); // align to page_entry_t
return (size_t*) CANONICAL(parent);
}
/** @brief Get the corresponding page map entry to a given virtual address
*
* Please note: this implementation requires that the tables are mapped
* at the end of VAS!
*/
static inline size_t* virt_to_entry(ssize_t addr, int level)
{
addr >>= PAGE_MAP_BITS;
addr |= (size_t) self[0]; //TODO: PAGE_MAP_PGT;
addr >>= level * PAGE_MAP_BITS;
addr &= ~(sizeof(size_t) - 1); // align to page_entry_t
return (size_t*) CANONICAL(addr);
}
/** @brief Get the corresponding virtual address to a page map entry */
static inline size_t entry_to_virt(size_t* entry, int level)
{
size_t addr = (size_t) entry;
addr <<= (level+1) * PAGE_MAP_BITS;
return CANONICAL(addr);
}
/** @brief Update page table bits (PG_*) by using arch independent flags (MAP_*) */
static inline size_t page_bits(int flags)
{
#ifdef CONFIG_X86_32
size_t bits = PG_PRESENT | PG_RW | PG_GLOBAL;
#elif defined(CONFIG_X86_64)
size_t bits = PG_PRESENT | PG_RW | PG_XD | PG_GLOBAL;
#endif
if (flags & MAP_NO_ACCESS) bits &= ~PG_PRESENT;
if (flags & MAP_READ_ONLY) bits &= ~PG_RW;
#ifdef CONFIG_X86_64
if (flags & MAP_CODE) bits &= ~PG_XD;
#endif
if (flags & MAP_USER_SPACE) bits &= ~PG_GLOBAL;
if (flags & MAP_USER_SPACE) bits |= PG_USER;
if (flags & MAP_WT) bits |= PG_PWT;
if (flags & MAP_NO_CACHE) bits |= PG_PCD;
return bits;
}
size_t virt_to_phys(size_t addr)
{
size_t vpn = addr >> PAGE_BITS; // virtual page number
size_t entry = self[0][vpn]; // page table entry
size_t off = addr & ~PAGE_MASK; // offset within page
size_t phy = entry & PAGE_MASK; // physical page frame number
return phy | off;
}
int page_map(size_t viraddr, size_t phyaddr, size_t npages, size_t bits)
{
int lvl, ret = -ENOMEM;
long vpn = viraddr >> PAGE_BITS;
long first[PAGE_LEVELS], last[PAGE_LEVELS];
/* Calculate index boundaries for page map traversal */
for (lvl=0; lvl<PAGE_LEVELS; lvl++) {
first[lvl] = (vpn ) >> (lvl * PAGE_MAP_BITS);
last[lvl] = (vpn+npages-1) >> (lvl * PAGE_MAP_BITS);
}
/** @todo: might not be sufficient! */
if (bits & PG_USER)
spinlock_irqsave_lock(&current_task->page_lock);
else
spinlock_lock(&kslock);
/* Start iterating through the entries
* beginning at the root table (PGD or PML4) */
for (lvl=PAGE_LEVELS-1; lvl>=0; lvl--) {
for (vpn=first[lvl]; vpn<=last[lvl]; vpn++) {
if (lvl) { /* PML4, PDPT, PGD */
if (!(self[lvl][vpn] & PG_PRESENT)) {
/* There's no table available which covers the region.
* Therefore we need to create a new empty table. */
size_t phyaddr = get_pages(1);
if (BUILTIN_EXPECT(!phyaddr, 0))
goto out;
if (bits & PG_USER)
atomic_int32_inc(&current_task->user_usage);
/* Reference the new table within its parent */
self[lvl][vpn] = phyaddr | bits | PG_PRESENT;
/* Fill new table with zeros */
memset(&self[lvl-1][vpn<<PAGE_MAP_BITS], 0, PAGE_SIZE);
}
}
else { /* PGT */
if (self[lvl][vpn] & PG_PRESENT)
/* There's already a page mapped at this address.
* We have to flush a single TLB entry. */
tlb_flush_one_page(vpn << PAGE_BITS);
self[lvl][vpn] = phyaddr | bits | PG_PRESENT;
phyaddr += PAGE_SIZE;
}
}
}
ret = 0;
out:
if (bits & PG_USER)
spinlock_irqsave_unlock(&current_task->page_lock);
else
spinlock_unlock(&kslock);
return ret;
}
/** Tables are freed by page_map_drop() */
int page_unmap(size_t viraddr, size_t npages)
{
/* We aquire both locks for kernel and task tables
* as we dont know to which the region belongs. */
spinlock_irqsave_lock(&current_task->page_lock);
spinlock_lock(&kslock);
/* Start iterating through the entries.
* Only the PGT entries are removed. Tables remain allocated. */
size_t vpn, start = viraddr>>PAGE_BITS;
for (vpn=start; vpn<start+npages; vpn++)
self[0][vpn] = 0;
spinlock_irqsave_unlock(&current_task->page_lock);
spinlock_unlock(&kslock);
/* This can't fail because we don't make checks here */
return 0;
}
int page_map_drop()
{
void traverse(int lvl, long vpn) {
long stop;
for (stop=vpn+PAGE_MAP_ENTRIES; vpn<stop; vpn++) {
if ((self[lvl][vpn] & PG_PRESENT) && (self[lvl][vpn] & PG_USER)) {
/* Post-order traversal */
if (lvl)
traverse(lvl-1, vpn<<PAGE_MAP_BITS);
put_pages(self[lvl][vpn] & PAGE_MASK, 1);
atomic_int32_dec(&current_task->user_usage);
}
}
}
spinlock_irqsave_lock(&current_task->page_lock);
traverse(PAGE_LEVELS-1, 0);
spinlock_irqsave_unlock(&current_task->page_lock);
/* This can't fail because we don't make checks here */
return 0;
}
int page_map_copy(task_t *dest)
{
int traverse(int lvl, long vpn) {
long stop;
for (stop=vpn+PAGE_MAP_ENTRIES; vpn<stop; vpn++) {
if (self[lvl][vpn] & PG_PRESENT) {
if (self[lvl][vpn] & PG_USER) {
size_t phyaddr = get_pages(1);
if (BUILTIN_EXPECT(!phyaddr, 0))
return -ENOMEM;
atomic_int32_inc(&dest->user_usage);
other[lvl][vpn] = phyaddr | (self[lvl][vpn] & ~PAGE_MASK);
if (lvl) /* PML4, PDPT, PGD */
traverse(lvl-1, vpn<<PAGE_MAP_BITS); /* Pre-order traversal */
else { /* PGT */
page_map(PAGE_TMP, phyaddr, 1, PG_RW);
memcpy((void*) PAGE_TMP, (void*) (vpn<<PAGE_BITS), PAGE_SIZE);
}
}
else if (self[lvl][vpn] & PG_SELF)
other[lvl][vpn] = 0;
else
other[lvl][vpn] = self[lvl][vpn];
}
else
other[lvl][vpn] = 0;
}
return 0;
}
spinlock_irqsave_lock(&current_task->page_lock);
self[PAGE_LEVELS-1][PAGE_MAP_ENTRIES-2] = dest->page_map | PG_PRESENT | PG_SELF | PG_RW;
int ret = traverse(PAGE_LEVELS-1, 0);
other[PAGE_LEVELS-1][PAGE_MAP_ENTRIES-1] = dest->page_map | PG_PRESENT | PG_SELF | PG_RW;
self [PAGE_LEVELS-1][PAGE_MAP_ENTRIES-2] = 0;
spinlock_irqsave_unlock(&current_task->page_lock);
/* Flush TLB entries of 'other' self-reference */
flush_tlb();
return ret;
}
void page_fault_handler(struct state *s)
{
size_t viraddr = read_cr2();
kprintf("Page Fault Exception (%d) at cs:ip = %#x:%#lx, task = %u, addr = %#lx, error = %#x [ %s %s %s %s %s ]\n",
s->int_no, s->cs, s->eip, current_task->id, viraddr, s->error,
(s->error & 0x4) ? "user" : "supervisor",
(s->error & 0x10) ? "instruction" : "data",
(s->error & 0x2) ? "write" : ((s->error & 0x10) ? "fetch" : "read"),
(s->error & 0x1) ? "protection" : "not present",
(s->error & 0x8) ? "reserved bit" : "\b");
while(1) HALT;
}
size_t map_region(size_t viraddr, size_t phyaddr, uint32_t npages, uint32_t flags)
{
task_t* task = current_task;
size_t* first[PAGE_LEVELS];
size_t* last[PAGE_LEVELS];
// TODO: this behaviour should be deprecated
if (!viraddr) {
int vma_flags = VMA_HEAP;
if (flags & MAP_USER_SPACE)
vma_flags |= VMA_USER;
viraddr = vma_alloc(npages * PAGE_SIZE, vma_flags);
}
size_t bits = page_bits(flags);
size_t start = viraddr;
size_t end = start + npages * PAGE_SIZE;
int traverse(int level, size_t* entry) {
size_t* stop = entry + PAGE_MAP_ENTRIES;
for (; entry != stop; entry++) {
if (entry < last[level] && entry >= first[level]) {
if (level) { // PGD, PDPT, PML4..
if (*entry & PG_PRESENT) {
if ((flags & MAP_USER_SPACE) && !(*entry & PG_USER)) {
/* We are altering entries which cover
* the kernel. So before changing them we need to
* make a private copy for the task */
size_t phyaddr = get_page();
if (BUILTIN_EXPECT(!phyaddr, 0))
return -ENOMEM;
atomic_int32_inc(&task->user_usage);
copy_page(phyaddr, *entry & PAGE_MASK);
*entry = phyaddr | (*entry & ~PAGE_MASK);
*entry &= ~PG_GLOBAL;
*entry |= PG_USER;
/* We just need to flush the table itself.
* TLB entries for the kernel remain valid
* because we've not changed them. */
tlb_flush_one_page(entry_to_virt(entry, 0));
}
}
else {
/* Theres no page map table available
* which covers the region. Therefore we will create a
* new table. */
size_t phyaddr = get_page();
if (BUILTIN_EXPECT(!phyaddr, 0))
return -ENOMEM;
if (flags & MAP_USER_SPACE)
atomic_int32_inc(&task->user_usage);
*entry = phyaddr | bits;
memset(get_child_entry(entry), 0x00, PAGE_SIZE); // fill with zeros
}
// do "pre-order" traversal if no hugepage
if (!(*entry & PG_PSE)) {
int ret = traverse(level-1, get_child_entry(entry));
if (ret < 0)
return ret;
}
}
else { // PGT
if ((*entry & PG_PRESENT) && !(flags & MAP_REMAP))
return -EINVAL;
*entry = phyaddr | bits;
if (flags & MAP_USER_SPACE)
atomic_int32_inc(&task->user_usage);
if (flags & MAP_REMAP)
tlb_flush_one_page(entry_to_virt(entry, level));
phyaddr += PAGE_SIZE;
}
}
}
return 0;
}
kprintf("map_region: map %u pages from %#lx to %#lx with flags: %#x\n", npages, viraddr, phyaddr, flags); // TODO: remove
if (BUILTIN_EXPECT(!task || !task->page_map, 0))
return 0;
// calc page tree boundaries
int i;
for (i=0; i<PAGE_LEVELS; i++) {
first[i] = virt_to_entry(start, i);
last[i] = virt_to_entry(end - 1, i) + 1; // exclusive
}
// lock tables
if (start < KERNEL_SPACE)
spinlock_lock(&kslock);
if (end >= KERNEL_SPACE)
spinlock_irqsave_lock(&task->page_lock);
int ret = traverse(PAGE_LEVELS-1, current_map);
// unlock tables
if (start < KERNEL_SPACE)
spinlock_unlock(&kslock);
if (end >= KERNEL_SPACE)
spinlock_irqsave_unlock(&task->page_lock);
return (ret) ? 0 : viraddr;
}
int unmap_region(size_t viraddr, uint32_t npages)
{
task_t* task = current_task;
size_t* first[PAGE_LEVELS];
size_t* last[PAGE_LEVELS];
size_t start = viraddr;
size_t end = start + npages * PAGE_SIZE;
kprintf("unmap_region: unmap %u pages from %#lx\n", npages, viraddr); // TODO: remove
/** @return number of page table entries which a present */
int traverse(int level, size_t* entry) {
int used = 0;
size_t* stop = entry + PAGE_MAP_ENTRIES;
for (; entry != stop; entry++) {
if (entry < last[level] && entry >= first[level]) {
if (level) { // PGD, PDPT, PML4
if ((*entry & PG_PRESENT) && !(*entry & PG_PSE)) {
// do "post-order" traversal if table is present and no hugepage
if (traverse(level-1, get_child_entry(entry)))
used++;
else { // child table is empty => delete it
*entry &= ~PG_PRESENT;
tlb_flush_one_page(entry_to_virt(entry, 0));
if (*entry & PG_USER) {
if (put_page(*entry & PAGE_MASK))
atomic_int32_dec(&task->user_usage);
}
}
}
}
else { // PGT
*entry &= ~PG_PRESENT;
tlb_flush_one_page(entry_to_virt(entry, level));
if (*entry & PG_USER)
atomic_int32_dec(&task->user_usage);
}
}
else {
if (*entry & PG_PRESENT)
used++;
}
}
return used;
}
if (BUILTIN_EXPECT(!task || !task->page_map, 0))
return 0;
// calc page tree boundaries
int i;
for (i=0; i<PAGE_LEVELS; i++) {
first[i] = virt_to_entry(start, i);
last[i] = virt_to_entry(end - 1, i) + 1; // exclusive
}
// lock tables
if (start < KERNEL_SPACE)
spinlock_lock(&kslock);
if (end >= KERNEL_SPACE)
spinlock_irqsave_lock(&task->page_lock);
traverse(PAGE_LEVELS-1, current_map);
// unlock tables
if (start < KERNEL_SPACE)
spinlock_unlock(&kslock);
if (end > KERNEL_SPACE)
spinlock_irqsave_unlock(&task->page_lock);
return 0;
}
int page_init(void)
{
size_t addr, npages;
int i;
/* Replace default pagefault handler */
irq_uninstall_handler(14);
irq_install_handler(14, page_fault_handler);
/* Map multiboot information and modules */
if (mb_info) {
addr = (size_t) mb_info & PAGE_MASK;
npages = PAGE_FLOOR(sizeof(*mb_info)) >> PAGE_BITS;
page_map(addr, addr, npages, PG_GLOBAL);
if (mb_info->flags & MULTIBOOT_INFO_MODS) {
addr = mb_info->mods_addr;
npages = PAGE_FLOOR(mb_info->mods_count*sizeof(multiboot_module_t)) >> PAGE_BITS;
page_map(addr, addr, npages, PG_GLOBAL);
multiboot_module_t* mmodule = (multiboot_module_t*) ((size_t) mb_info->mods_addr);
for(i=0; i<mb_info->mods_count; i++) {
addr = mmodule[i].mod_start;
npages = PAGE_FLOOR(mmodule[i].mod_end - mmodule[i].mod_start) >> PAGE_BITS;
page_map(addr, addr, npages, PG_GLOBAL);
}
}
}
/* Flush TLB to adopt changes above */
//flush_tlb();
return 0;
}