// Set up a two-level page table:
// kern_pgdir is its linear (virtual) address of the root
// Then turn on paging. Then effectively turn off segmentation.
// (i.e., the segment base addrs are set to zero).
//
// This function only sets up the kernel part of the address space
// (ie. addresses >= UTOP). The user part of the address space
// will be setup later.
//
// From UTOP to ULIM, the user is allowed to read but not write.
// Above ULIM the user cannot read or write.
void
mem_init(void)
{
uint32_t cr0;
size_t n;
// Ensure user & kernel struct Pages agree.
static_assert(sizeof(struct Page) == sizeof(struct UserPage));
// Find out how much memory the machine has (npages & npages_basemem).
i386_detect_memory();
// Remove this line when you're ready to test this function.
//panic("mem_init: This function is not finished\n");
//////////////////////////////////////////////////////////////////////
// create initial page directory.
kern_pgdir = (pde_t *) boot_alloc(PGSIZE);
memset(kern_pgdir, 0, PGSIZE);
//////////////////////////////////////////////////////////////////////
// Recursively insert PD in itself as a page table, to form
// a virtual page table at virtual address UVPT.
// (For now, you don't have understand the greater purpose of the
// following line.)
// Permissions: kernel R, user R
kern_pgdir[PDX(UVPT)] = PADDR(kern_pgdir) | PTE_U | PTE_P;
//////////////////////////////////////////////////////////////////////
// Allocate an array of npages 'struct Page's and store it in 'pages'.
// The kernel uses this array to keep track of physical pages: for
// each physical page, there is a corresponding struct Page in this
// array. 'npages' is the number of physical pages in memory.
pages = (Page *) boot_alloc(npages * sizeof(struct Page));
//////////////////////////////////////////////////////////////////////
// Make 'envs' point to an array of size 'NENV' of 'struct Env'.
// LAB 3: Your code here.
envs = (Env *) boot_alloc(NENV * sizeof(struct Env));
//////////////////////////////////////////////////////////////////////
// Now that we've allocated the initial kernel data structures, we set
// up the list of free physical pages. Once we've done so, all further
// memory management will go through the page_* functions. In
// particular, we can now map memory using page_map_segment
// or page_insert
page_init();
check_page_free_list(true);
check_page_alloc();
check_page();
//////////////////////////////////////////////////////////////////////
// Now we set up virtual memory
//////////////////////////////////////////////////////////////////////
// Use the physical memory that 'entry_stack' refers to as the kernel
// stack. The kernel stack grows down from virtual address KSTACKTOP.
// We consider the entire range from [KSTACKTOP-PTSIZE, KSTACKTOP)
// to be the kernel stack, but break this into two pieces:
// * [KSTACKTOP-KSTKSIZE, KSTACKTOP) -- backed by physical memory
// * [KSTACKTOP-PTSIZE, KSTACKTOP-KSTKSIZE) -- not backed; so if
// the kernel overflows its stack, it will fault rather than
// overwrite memory. Known as a "guard page".
// Permissions: kernel RW, user NONE
Page *pp = pa2page((physaddr_t)entry_stack-KERNBASE);
for (uintptr_t ptr = KSTACKTOP-KSTKSIZE; ptr < KSTACKTOP; ptr += PGSIZE) {
if (page_insert(kern_pgdir, pp, ptr, PTE_W | PTE_P) < 0)
panic("Couldn't create page table entries for stack.\n");
pp++;
}
//////////////////////////////////////////////////////////////////////
// Map all of physical memory at KERNBASE.
// Ie. the VA range [KERNBASE, 2^32) should map to
// the PA range [0, 2^32 - KERNBASE)
// We might not have 2^32 - KERNBASE bytes of physical memory, but
// we just set up the mapping anyway.
// Permissions: kernel RW, user NONE
page_map_segment(kern_pgdir, KERNBASE, 0xFFFFFFFF-KERNBASE, 0x0, PTE_W | PTE_P);
//print_page_table(kern_pgdir, false, false);
//////////////////////////////////////////////////////////////////////
// Map the 'envs' array read-only by the user at linear address UENVS.
// Permissions: kernel R, user R
// (That's the UENVS version; 'envs' itself is kernel RW, user NONE.)
// LAB 3: Your code here.
page_map_segment(kern_pgdir, (uintptr_t) UENVS, ROUNDUP(NENV*sizeof(struct Env), PGSIZE), PADDR(envs), PTE_U | PTE_P);
//////////////////////////////////////////////////////////////////////
// Map 'pages' read-only by the user at linear address UPAGES.
// Permissions: kernel R, user R
// (That's the UPAGES version; 'pages' itself is kernel RW, user NONE.)
// LAB 3: Your code here.
page_map_segment(kern_pgdir, UPAGES, ROUNDUP(npages*sizeof(struct Page), PGSIZE), PADDR(pages), PTE_U | PTE_P);
// Check that the initial page directory has been set up correctly.
check_kern_pgdir();
// Switch from the minimal entry page directory to the full kern_pgdir
// page table we just created. Our instruction pointer should be
// somewhere between KERNBASE and KERNBASE+4MB right now, which is
// mapped the same way by both page tables.
//
// If the machine reboots at this point, you've probably set up your
// kern_pgdir wrong.
lcr3(PADDR(kern_pgdir));
// entry.S set the really important flags in cr0 (including enabling
// paging). Here we configure the rest of the flags we need.
cr0 = rcr0();
cr0 |= CR0_PE|CR0_PG|CR0_AM|CR0_WP|CR0_NE|CR0_MP;
cr0 &= ~(CR0_TS|CR0_EM);
lcr0(cr0);
// Some more checks, only possible after kern_pgdir is installed.
check_page_installed_pgdir();
}
// do_pgfault - interrupt handler to process the page fault execption
int
do_pgfault(struct mm_struct *mm, uint32_t error_code, uintptr_t addr) {
if (mm == NULL) {
assert(current != NULL);
panic("page fault in kernel thread: pid = %d, %d %08x.\n",
current->pid, error_code, addr);
}
bool need_unlock = 1;
if (!try_lock_mm(mm)) {
if (current != NULL && mm->locked_by == current->pid) {
need_unlock = 0;
}
else {
lock_mm(mm);
}
}
int ret = -E_INVAL;
struct vma_struct *vma = find_vma(mm, addr);
if (vma == NULL || vma->vm_start > addr) {
goto failed;
}
if (vma->vm_flags & VM_STACK) {
if (addr < vma->vm_start + PGSIZE) {
goto failed;
}
}
switch (error_code & 3) {
default:
/* default is 3: write, present */
case 2: /* write, not present */
if (!(vma->vm_flags & VM_WRITE)) {
goto failed;
}
break;
case 1: /* read, present */
goto failed;
case 0: /* read, not present */
if (!(vma->vm_flags & (VM_READ | VM_EXEC))) {
goto failed;
}
}
uint32_t perm = PTE_U;
if (vma->vm_flags & VM_WRITE) {
perm |= PTE_W;
}
addr = ROUNDDOWN(addr, PGSIZE);
ret = -E_NO_MEM;
pte_t *ptep;
if ((ptep = get_pte(mm->pgdir, addr, 1)) == NULL) {
goto failed;
}
if (*ptep == 0) {
if (!(vma->vm_flags & VM_SHARE)) {
if (pgdir_alloc_page(mm->pgdir, addr, perm) == NULL) {
goto failed;
}
}
else {
lock_shmem(vma->shmem);
uintptr_t shmem_addr = addr - vma->vm_start + vma->shmem_off;
pte_t *sh_ptep = shmem_get_entry(vma->shmem, shmem_addr, 1);
if (sh_ptep == NULL || *sh_ptep == 0) {
unlock_shmem(vma->shmem);
goto failed;
}
unlock_shmem(vma->shmem);
if (*sh_ptep & PTE_P) {
page_insert(mm->pgdir, pa2page(*sh_ptep), addr, perm);
}
else {
swap_duplicate(*ptep);
*ptep = *sh_ptep;
}
}
}
else {
struct Page *page, *newpage = NULL;
bool cow = ((vma->vm_flags & (VM_SHARE | VM_WRITE)) == VM_WRITE), may_copy = 1;
assert(!(*ptep & PTE_P) || ((error_code & 2) && !(*ptep & PTE_W) && cow));
if (cow) {
newpage = alloc_page();
}
if (*ptep & PTE_P) {
page = pte2page(*ptep);
}
else {
if ((ret = swap_in_page(*ptep, &page)) != 0) {
if (newpage != NULL) {
free_page(newpage);
}
goto failed;
}
if (!(error_code & 2) && cow) {
perm &= ~PTE_W;
may_copy = 0;
}
}
if (cow && may_copy) {
if (page_ref(page) + swap_page_count(page) > 1) {
if (newpage == NULL) {
goto failed;
}
memcpy(page2kva(newpage), page2kva(page), PGSIZE);
page = newpage, newpage = NULL;
}
}
page_insert(mm->pgdir, page, addr, perm);
if (newpage != NULL) {
free_page(newpage);
}
}
ret = 0;
failed:
if (need_unlock) {
unlock_mm(mm);
}
return ret;
}
void*
frame_evict (void* uaddr)
{
/* 1. Choose a frame to evict, using your page replacement algorithm.
The "accessed" and "dirty" bits in the page table, described below,
will come in handy. */
struct frame_table_entry *fte = NULL;
switch (PAGE_EVICTION_ALGORITHM)
{
/* First in first out */
case PAGE_EVICTION_FIFO:
fte = frame_evict_choose_fifo ();
break;
/* Second chance */
case PAGE_EVICTION_SECONDCHANCE:
fte = frame_evict_choose_secondchance ();
break;
default:
PANIC ("Invalid eviction algorithm choice.");
}
ASSERT (fte != NULL);
/* 2. Remove references to the frame from any page table that refers to it.
Unless you have implemented sharing, only a single page should refer to
a frame at any given time. */
pagedir_clear_page (fte->owner->pagedir, pg_round_down (fte->uaddr));
/* 3. If necessary, write the page to the file system or to swap.
The evicted frame may then be used to store a different page. */
struct page *p_evict =
page_lookup (fte->owner->pages, pg_round_down (fte->uaddr));
if (p_evict == NULL)
PANIC ("Failed to get supp page for existing page.");
/* Page to be evicted is in swap */
if (p_evict->page_location_option == FILESYS)
{
if (p_evict->writable)
{
file_write_at (p_evict->file, fte->kaddr, p_evict->page_read_bytes,
p_evict->ofs);
}
}
else if (p_evict->page_location_option == ALLZERO)
{
// All zero, so can just be overwritten
}
else
{
// From stack
int index = swap_to_disk (pg_round_down (fte->uaddr));
/* Creates a supp page and insert it into pages. */
struct page *p = page_create ();
if (p == NULL)
PANIC ("Failed to get supp page for swap slot.");
p->addr = fte->uaddr;
p->page_location_option = SWAPSLOT;
p->swap_index = index;
page_insert (fte->owner->pages, &p->hash_elem);
}
/* Replace virtual address with new virtual address */
fte->owner = thread_current ();
fte->uaddr = uaddr;
/* Reinsert the frame table entry into the frame table */
lock_acquire (&frame_table_lock);
list_remove (&fte->elem);
list_push_front (&frame_table, &fte->elem);
lock_release (&frame_table_lock);
return fte->kaddr;
}
/* do_pgfault - interrupt handler to process the page fault execption
* @mm : the control struct for a set of vma using the same PDT
* @error_code : the error code recorded in trapframe->tf_err which is setted by x86 hardware
* @addr : the addr which causes a memory access exception, (the contents of the CR2 register)
*
* CALL GRAPH: trap--> trap_dispatch-->pgfault_handler-->do_pgfault
* The processor provides ucore's do_pgfault function with two items of information to aid in diagnosing
* the exception and recovering from it.
* (1) The contents of the CR2 register. The processor loads the CR2 register with the
* 32-bit linear address that generated the exception. The do_pgfault fun can
* use this address to locate the corresponding page directory and page-table
* entries.
* (2) An error code on the kernel stack. The error code for a page fault has a format different from
* that for other exceptions. The error code tells the exception handler three things:
* -- The P flag (bit 0) indicates whether the exception was due to a not-present page (0)
* or to either an access rights violation or the use of a reserved bit (1).
* -- The W/R flag (bit 1) indicates whether the memory access that caused the exception
* was a read (0) or write (1).
* -- The U/S flag (bit 2) indicates whether the processor was executing at user mode (1)
* or supervisor mode (0) at the time of the exception.
*/
int
do_pgfault(struct mm_struct *mm, uint32_t error_code, uintptr_t addr) {
int ret = -E_INVAL;
//try to find a vma which include addr
struct vma_struct *vma = find_vma(mm, addr);
pgfault_num++;
//If the addr is in the range of a mm's vma?
if (vma == NULL || vma->vm_start > addr) {
cprintf("not valid addr %x, and can not find it in vma\n", addr);
goto failed;
}
//check the error_code
switch (error_code & 3) {
default:
/* error code flag : default is 3 ( W/R=1, P=1): write, present */
case 2: /* error code flag : (W/R=1, P=0): write, not present */
if (!(vma->vm_flags & VM_WRITE)) {
cprintf("do_pgfault failed: error code flag = write AND not present, but the addr's vma cannot write\n");
goto failed;
}
break;
case 1: /* error code flag : (W/R=0, P=1): read, present */
cprintf("do_pgfault failed: error code flag = read AND present\n");
goto failed;
case 0: /* error code flag : (W/R=0, P=0): read, not present */
if (!(vma->vm_flags & (VM_READ | VM_EXEC))) {
cprintf("do_pgfault failed: error code flag = read AND not present, but the addr's vma cannot read or exec\n");
goto failed;
}
}
/* IF (write an existed addr ) OR
* (write an non_existed addr && addr is writable) OR
* (read an non_existed addr && addr is readable)
* THEN
* continue process
*/
uint32_t perm = PTE_U;
if (vma->vm_flags & VM_WRITE) {
perm |= PTE_W;
}
addr = ROUNDDOWN(addr, PGSIZE);
ret = -E_NO_MEM;
pte_t *ptep=NULL;
/*LAB3 EXERCISE 1: 2013012213
* Maybe you want help comment, BELOW comments can help you finish the code
*
* Some Useful MACROs and DEFINEs, you can use them in below implementation.
* MACROs or Functions:
* get_pte : get an pte and return the kernel virtual address of this pte for la
* if the PT contians this pte didn't exist, alloc a page for PT (notice the 3th parameter '1')
* pgdir_alloc_page : call alloc_page & page_insert functions to allocate a page size memory & setup
* an addr map pa<--->la with linear address la and the PDT pgdir
* DEFINES:
* VM_WRITE : If vma->vm_flags & VM_WRITE == 1/0, then the vma is writable/non writable
* PTE_W 0x002 // page table/directory entry flags bit : Writeable
* PTE_U 0x004 // page table/directory entry flags bit : User can access
* VARIABLES:
* mm->pgdir : the PDT of these vma
*
*/
/*LAB3 EXERCISE 1: 2013012213*/
if ((ptep = get_pte(mm->pgdir, addr, 1)) == NULL) {
cprintf("get_pte in do_pgfault failed\n");
goto failed;
}
if (*ptep == 0) {
if (pgdir_alloc_page(mm->pgdir, addr, perm) == NULL) {
cprintf("pgdir_alloc_page in do_pgfault failed\n");
goto failed;
}
}
else {
/*LAB3 EXERCISE 2: 2013012213
* Now we think this pte is a swap entry, we should load data from disk to a page with phy addr,
* and map the phy addr with logical addr, trigger swap manager to record the access situation of this page.
*
* Some Useful MACROs and DEFINEs, you can use them in below implementation.
* MACROs or Functions:
* swap_in(mm, addr, &page) : alloc a memory page, then according to the swap entry in PTE for addr,
* find the addr of disk page, read the content of disk page into this memroy page
* page_insert ： build the map of phy addr of an Page with the linear addr la
* swap_map_swappable ： set the page swappable
*/
/*
* LAB5 CHALLENGE ( the implmentation Copy on Write)
There are 2 situlations when code comes here.
1) *ptep & PTE_P == 1, it means one process try to write a readonly page.
If the vma includes this addr is writable, then we can set the page writable by rewrite the *ptep.
This method could be used to implement the Copy on Write (COW) thchnology(a fast fork process method).
2) *ptep & PTE_P == 0 & but *ptep!=0, it means this pte is a swap entry.
We should add the LAB3's results here.
*/
struct Page *page = NULL;
if (*ptep & PTE_P)
panic("error write a non-writable pte");
else {
if(swap_init_ok) {
if ((ret = swap_in(mm, addr, &page)) != 0) {
cprintf("swap_in in do_pgfault failed\n");
goto failed;
}
}
else {
cprintf("no swap_init_ok but ptep is %x, failed\n", *ptep);
goto failed;
}
}
page_insert(mm->pgdir, page, addr, perm);
swap_map_swappable(mm, addr, page, 1);
page->pra_vaddr = addr;
}
ret = 0;
failed:
return ret;
}
// check page_insert, page_remove, &c
static void
check_page(void)
{
struct PageInfo *pp, *pp0, *pp1, *pp2;
struct PageInfo *fl;
pte_t *ptep, *ptep1;
void *va;
uintptr_t mm1, mm2;
int i;
extern pde_t entry_pgdir[];
// should be able to allocate three pages
pp0 = pp1 = pp2 = 0;
assert((pp0 = page_alloc(0)));
assert((pp1 = page_alloc(0)));
assert((pp2 = page_alloc(0)));
assert(pp0);
assert(pp1 && pp1 != pp0);
assert(pp2 && pp2 != pp1 && pp2 != pp0);
// temporarily steal the rest of the free pages
fl = page_free_list;
page_free_list = 0;
// should be no free memory
assert(!page_alloc(0));
// there is no page allocated at address 0
assert(page_lookup(kern_pgdir, (void *) 0x0, &ptep) == NULL);
// there is no free memory, so we can't allocate a page table
assert(page_insert(kern_pgdir, pp1, 0x0, PTE_W) < 0);
// free pp0 and try again: pp0 should be used for page table
page_free(pp0);
assert(page_insert(kern_pgdir, pp1, 0x0, PTE_W) == 0);
assert(PTE_ADDR(kern_pgdir[0]) == page2pa(pp0));
assert(check_va2pa(kern_pgdir, 0x0) == page2pa(pp1));
assert(pp1->pp_ref == 1);
assert(pp0->pp_ref == 1);
// should be able to map pp2 at PGSIZE because pp0 is already allocated for page table
assert(page_insert(kern_pgdir, pp2, (void*) PGSIZE, PTE_W) == 0);
assert(check_va2pa(kern_pgdir, PGSIZE) == page2pa(pp2));
assert(pp2->pp_ref == 1);
// should be no free memory
assert(!page_alloc(0));
// should be able to map pp2 at PGSIZE because it's already there
assert(page_insert(kern_pgdir, pp2, (void*) PGSIZE, PTE_W) == 0);
assert(check_va2pa(kern_pgdir, PGSIZE) == page2pa(pp2));
assert(pp2->pp_ref == 1);
// pp2 should NOT be on the free list
// could happen in ref counts are handled sloppily in page_insert
//cprintf("p2: %p, free_list %p, p2 ref: %d", pp2, page_free_list, (int)pp2->pp_ref);
assert(!page_alloc(0));
// check that pgdir_walk returns a pointer to the pte
// 从这里也可以推测出pgdir_walk的功能（因为page table entry的歧义:
// 是pointer to page table, 还是pointer of entry in page table)
// 给定虚拟地址va, kern_pgdir[PDX(va)]是va二级页表, page table的物理地址。
// 再KADDR一下，就成了page table的虚拟地址，即ptep
// ptep + PTX(va)即va在page table中的表项的位置，的虚拟地址
// 这就是pgdir_walk需要返回的。
ptep = (pte_t *) KADDR(PTE_ADDR(kern_pgdir[PDX(PGSIZE)]));
assert(pgdir_walk(kern_pgdir, (void*)PGSIZE, 0) == ptep+PTX(PGSIZE));
// should be able to change permissions too.
assert(page_insert(kern_pgdir, pp2, (void*) PGSIZE, PTE_W|PTE_U) == 0);
assert(check_va2pa(kern_pgdir, PGSIZE) == page2pa(pp2));
assert(pp2->pp_ref == 1);
assert(*pgdir_walk(kern_pgdir, (void*) PGSIZE, 0) & PTE_U);
assert(kern_pgdir[0] & PTE_U);
// should be able to remap with fewer permissions
assert(page_insert(kern_pgdir, pp2, (void*) PGSIZE, PTE_W) == 0);
assert(*pgdir_walk(kern_pgdir, (void*) PGSIZE, 0) & PTE_W);
assert(!(*pgdir_walk(kern_pgdir, (void*) PGSIZE, 0) & PTE_U));
// should not be able to map at PTSIZE because need free page for page table
assert(page_insert(kern_pgdir, pp0, (void*) PTSIZE, PTE_W) < 0);
// insert pp1 at PGSIZE (replacing pp2)
assert(page_insert(kern_pgdir, pp1, (void*) PGSIZE, PTE_W) == 0);
assert(!(*pgdir_walk(kern_pgdir, (void*) PGSIZE, 0) & PTE_U));
// should have pp1 at both 0 and PGSIZE, pp2 nowhere, ...
assert(check_va2pa(kern_pgdir, 0) == page2pa(pp1));
assert(check_va2pa(kern_pgdir, PGSIZE) == page2pa(pp1));
// ... and ref counts should reflect this
assert(pp1->pp_ref == 2);
assert(pp2->pp_ref == 0);
// pp2 should be returned by page_alloc
assert((pp = page_alloc(0)) && pp == pp2);
// unmapping pp1 at 0 should keep pp1 at PGSIZE
page_remove(kern_pgdir, 0x0);
assert(check_va2pa(kern_pgdir, 0x0) == ~0);
assert(check_va2pa(kern_pgdir, PGSIZE) == page2pa(pp1));
assert(pp1->pp_ref == 1);
assert(pp2->pp_ref == 0);
// unmapping pp1 at PGSIZE should free it
page_remove(kern_pgdir, (void*) PGSIZE);
assert(check_va2pa(kern_pgdir, 0x0) == ~0);
assert(check_va2pa(kern_pgdir, PGSIZE) == ~0);
assert(pp1->pp_ref == 0);
assert(pp2->pp_ref == 0);
// so it should be returned by page_alloc
assert((pp = page_alloc(0)) && pp == pp1);
// should be no free memory
assert(!page_alloc(0));
// forcibly take pp0 back
assert(PTE_ADDR(kern_pgdir[0]) == page2pa(pp0));
kern_pgdir[0] = 0;
assert(pp0->pp_ref == 1);
pp0->pp_ref = 0;
// check pointer arithmetic in pgdir_walk
page_free(pp0);
va = (void*)(PGSIZE * NPDENTRIES + PGSIZE);
ptep = pgdir_walk(kern_pgdir, va, 1);
ptep1 = (pte_t *) KADDR(PTE_ADDR(kern_pgdir[PDX(va)]));
assert(ptep == ptep1 + PTX(va));
kern_pgdir[PDX(va)] = 0;
pp0->pp_ref = 0;
// check that new page tables get cleared
memset(page2kva(pp0), 0xFF, PGSIZE);
page_free(pp0);
pgdir_walk(kern_pgdir, 0x0, 1);
ptep = (pte_t *) page2kva(pp0);
for(i=0; i<NPTENTRIES; i++)
assert((ptep[i] & PTE_P) == 0);
kern_pgdir[0] = 0;
pp0->pp_ref = 0;
// give free list back
page_free_list = fl;
// free the pages we took
page_free(pp0);
page_free(pp1);
page_free(pp2);
// test mmio_map_region
mm1 = (uintptr_t) mmio_map_region(0, 4097);
mm2 = (uintptr_t) mmio_map_region(0, 4096);
// check that they're in the right region
assert(mm1 >= MMIOBASE && mm1 + 8096 < MMIOLIM);
assert(mm2 >= MMIOBASE && mm2 + 8096 < MMIOLIM);
// check that they're page-aligned
assert(mm1 % PGSIZE == 0 && mm2 % PGSIZE == 0);
// check that they don't overlap
assert(mm1 + 8096 <= mm2);
// check page mappings
assert(check_va2pa(kern_pgdir, mm1) == 0);
assert(check_va2pa(kern_pgdir, mm1+PGSIZE) == PGSIZE);
assert(check_va2pa(kern_pgdir, mm2) == 0);
assert(check_va2pa(kern_pgdir, mm2+PGSIZE) == ~0);
// check permissions
assert(*pgdir_walk(kern_pgdir, (void*) mm1, 0) & (PTE_W|PTE_PWT|PTE_PCD));
assert(!(*pgdir_walk(kern_pgdir, (void*) mm1, 0) & PTE_U));
// clear the mappings
*pgdir_walk(kern_pgdir, (void*) mm1, 0) = 0;
*pgdir_walk(kern_pgdir, (void*) mm1 + PGSIZE, 0) = 0;
*pgdir_walk(kern_pgdir, (void*) mm2, 0) = 0;
cprintf("check_page() succeeded!\n");
}
int do_pgfault(struct mm_struct *mm, machine_word_t error_code, uintptr_t addr)
{
if (mm == NULL) {
assert(current != NULL);
/* Chen Yuheng
* give handler a chance to deal with it
*/
kprintf
("page fault in kernel thread: pid = %d, name = %s, %d %08x.\n",
current->pid, current->name, error_code, addr);
return -E_KILLED;
}
bool need_unlock = 1;
if (!try_lock_mm(mm)) {
if (current != NULL && mm->locked_by == current->pid) {
need_unlock = 0;
} else {
lock_mm(mm);
}
}
int ret = -E_INVAL;
struct vma_struct *vma = find_vma(mm, addr);
if (vma == NULL || vma->vm_start > addr) {
goto failed;
}
if (vma->vm_flags & VM_STACK) {
if (addr < vma->vm_start + PGSIZE) {
goto failed;
}
}
//kprintf("@ %x %08x\n", vma->vm_flags, vma->vm_start);
//assert((vma->vm_flags & VM_IO)==0);
if (vma->vm_flags & VM_IO) {
ret = -E_INVAL;
goto failed;
}
switch (error_code & 3) {
default:
/* default is 3: write, present */
case 2: /* write, not present */
if (!(vma->vm_flags & VM_WRITE)) {
goto failed;
}
break;
case 1: /* read, present */
goto failed;
case 0: /* read, not present */
if (!(vma->vm_flags & (VM_READ | VM_EXEC))) {
goto failed;
}
}
pte_perm_t perm, nperm;
#ifdef ARCH_ARM
/* ARM9 software emulated PTE_xxx */
perm = PTE_P | PTE_U;
if (vma->vm_flags & VM_WRITE) {
perm |= PTE_W;
}
#else
ptep_unmap(&perm);
ptep_set_u_read(&perm);
if (vma->vm_flags & VM_WRITE) {
ptep_set_u_write(&perm);
}
#endif
addr = ROUNDDOWN(addr, PGSIZE);
ret = -E_NO_MEM;
pte_t *ptep;
if ((ptep = get_pte(mm->pgdir, addr, 1)) == NULL) {
goto failed;
}
if (ptep_invalid(ptep)) {
#ifdef UCONFIG_BIONIC_LIBC
if (vma->mfile.file != NULL) {
struct file *file = vma->mfile.file;
off_t old_pos = file->pos, new_pos =
vma->mfile.offset + addr - vma->vm_start;
#ifdef SHARE_MAPPED_FILE
struct mapped_addr *maddr =
find_maddr(file, new_pos, NULL);
if (maddr == NULL) {
#endif // SHARE_MAPPED_FILE
struct Page *page;
if ((page = alloc_page()) == NULL) {
assert(false);
goto failed;
}
nperm = perm;
#ifdef ARCH_ARM
/* ARM9 software emulated PTE_xxx */
nperm &= ~PTE_W;
#else
ptep_unset_s_write(&nperm);
#endif
page_insert_pte(mm->pgdir, page, ptep, addr,
nperm);
if ((ret =
filestruct_setpos(file, new_pos)) != 0) {
assert(false);
goto failed;
}
filestruct_read(file, page2kva(page), PGSIZE);
if ((ret =
filestruct_setpos(file, old_pos)) != 0) {
assert(false);
goto failed;
}
#ifdef SHARE_MAPPED_FILE
if ((maddr = (struct mapped_addr *)
kmalloc(sizeof(struct mapped_addr))) !=
NULL) {
maddr->page = page;
maddr->offset = new_pos;
page->maddr = maddr;
list_add(&
(file->node->mapped_addr_list),
&(maddr->list));
} else {
assert(false);
}
} else {
nperm = perm;
#ifdef ARCH_ARM
/* ARM9 software emulated PTE_xxx */
nperm &= ~PTE_W;
#else
ptep_unset_s_write(&nperm);
#endif
page_insert_pte(mm->pgdir, maddr->page, ptep,
addr, nperm);
}
#endif //SHARE_MAPPED_FILE
} else
#endif //UCONFIG_BIONIC_LIBC
if (!(vma->vm_flags & VM_SHARE)) {
if (pgdir_alloc_page(mm->pgdir, addr, perm) == NULL) {
goto failed;
}
#ifdef UCONFIG_BIONIC_LIBC
if (vma->vm_flags & VM_ANONYMOUS) {
memset((void *)addr, 0, PGSIZE);
}
#endif //UCONFIG_BIONIC_LIBC
} else { //shared mem
lock_shmem(vma->shmem);
uintptr_t shmem_addr =
addr - vma->vm_start + vma->shmem_off;
pte_t *sh_ptep =
shmem_get_entry(vma->shmem, shmem_addr, 1);
if (sh_ptep == NULL || ptep_invalid(sh_ptep)) {
unlock_shmem(vma->shmem);
goto failed;
}
unlock_shmem(vma->shmem);
if (ptep_present(sh_ptep)) {
page_insert(mm->pgdir, pa2page(*sh_ptep), addr,
perm);
} else {
#ifdef UCONFIG_SWAP
swap_duplicate(*ptep);
ptep_copy(ptep, sh_ptep);
#else
panic("NO SWAP\n");
#endif
}
}
} else { //a present page, handle copy-on-write (cow)
struct Page *page, *newpage = NULL;
bool cow =
((vma->vm_flags & (VM_SHARE | VM_WRITE)) == VM_WRITE),
may_copy = 1;
#if 1
if (!(!ptep_present(ptep)
|| ((error_code & 2) && !ptep_u_write(ptep) && cow))) {
//assert(PADDR(mm->pgdir) == rcr3());
kprintf("%p %p %d %d %x\n", *ptep, addr, error_code,
cow, vma->vm_flags);
assert(0);
}
#endif
if (cow) {
newpage = alloc_page();
}
if (ptep_present(ptep)) {
page = pte2page(*ptep);
} else {
#ifdef UCONFIG_SWAP
if ((ret = swap_in_page(*ptep, &page)) != 0) {
if (newpage != NULL) {
free_page(newpage);
}
goto failed;
}
#else
assert(0);
#endif
if (!(error_code & 2) && cow) {
#ifdef ARCH_ARM
//#warning ARM9 software emulated PTE_xxx
perm &= ~PTE_W;
#else
ptep_unset_s_write(&perm);
#endif
may_copy = 0;
}
}
if (cow && may_copy) {
#ifdef UCONFIG_SWAP
if (page_ref(page) + swap_page_count(page) > 1) {
#else
if (page_ref(page) > 1) {
#endif
if (newpage == NULL) {
goto failed;
}
memcpy(page2kva(newpage), page2kva(page),
PGSIZE);
//kprintf("COW!\n");
page = newpage, newpage = NULL;
}
}
#ifdef UCONFIG_BIONIC_LIBC
else if (vma->mfile.file != NULL) {
#ifdef UCONFIG_SWAP
assert(page_reg(page) + swap_page_count(page) == 1);
#else
assert(page_ref(page) == 1);
#endif
#ifdef SHARE_MAPPED_FILE
off_t offset = vma->mfile.offset + addr - vma->vm_start;
struct mapped_addr *maddr =
find_maddr(vma->mfile.file, offset, page);
if (maddr != NULL) {
list_del(&(maddr->list));
kfree(maddr);
page->maddr = NULL;
assert(find_maddr(vma->mfile.file, offset, page)
== NULL);
} else {
}
#endif //SHARE_MAPPED_FILE
}
#endif //UCONFIG_BIONIC_LIBC
else {
}
page_insert(mm->pgdir, page, addr, perm);
if (newpage != NULL) {
free_page(newpage);
}
}
ret = 0;
failed:
if (need_unlock) {
unlock_mm(mm);
}
return ret;
}
/* ucore use copy-on-write when forking a new process,
* thus copy_range only copy pdt/pte and set their permission to
* READONLY, a write will be handled in pgfault
*/
int
copy_range(pgd_t *to, pgd_t *from, uintptr_t start, uintptr_t end, bool share) {
assert(start % PGSIZE == 0 && end % PGSIZE == 0);
assert(USER_ACCESS(start, end));
do {
pte_t *ptep = get_pte(from, start, 0), *nptep;
if (ptep == NULL) {
if (get_pud(from, start, 0) == NULL) {
start = ROUNDDOWN(start + PUSIZE, PUSIZE);
}
else if (get_pmd(from, start, 0) == NULL) {
start = ROUNDDOWN(start + PMSIZE, PMSIZE);
}
else {
start = ROUNDDOWN(start + PTSIZE, PTSIZE);
}
continue ;
}
if (*ptep != 0) {
if ((nptep = get_pte(to, start, 1)) == NULL) {
return -E_NO_MEM;
}
int ret;
//kprintf("%08x %08x %08x\n", nptep, *nptep, start);
assert(*ptep != 0 && *nptep == 0);
#ifdef ARCH_ARM
//TODO add code to handle swap
if (ptep_present(ptep)){
//no body should be able to write this page
//before a W-pgfault
pte_perm_t perm = PTE_P;
if(ptep_u_read(ptep))
perm |= PTE_U;
if(!share){
//Original page should be set to readonly!
//because Copy-on-write may happen
//after the current proccess modifies its page
ptep_set_perm(ptep, perm);
}else{
if(ptep_u_write(ptep)){
perm |= PTE_W;
}
}
struct Page *page = pte2page(*ptep);
ret = page_insert(to, page, start, perm);
}
#else /* ARCH_ARM */
if (ptep_present(ptep)) {
pte_perm_t perm = ptep_get_perm(ptep, PTE_USER);
struct Page *page = pte2page(*ptep);
if (!share && ptep_s_write(ptep)) {
ptep_unset_s_write(&perm);
pte_perm_t perm_with_swap_stat = ptep_get_perm(ptep, PTE_SWAP);
ptep_set_perm(&perm_with_swap_stat, perm);
page_insert(from, page, start, perm_with_swap_stat);
}
ret = page_insert(to, page, start, perm);
assert(ret == 0);
}
#endif /* ARCH_ARM */
else {
#ifdef CONFIG_NO_SWAP
assert(0);
#endif
swap_entry_t entry;
ptep_copy(&entry, ptep);
swap_duplicate(entry);
ptep_copy(nptep, &entry);
}
}
start += PGSIZE;
} while (start != 0 && start < end);
#ifdef ARCH_ARM
/* we have modified the PTE of the original
* process, so invalidate TLB */
tlb_invalidate_all();
#endif
return 0;
}
//
// Set up the initial program binary, stack, and processor flags
// for a user process.
// This function is ONLY called during kernel initialization,
// before running the first user-mode environment.
//
// This function loads all loadable segments from the ELF binary image
// into the environment's user memory, starting at the appropriate
// virtual addresses indicated in the ELF program header.
// At the same time it clears to zero any portions of these segments
// that are marked in the program header as being mapped
// but not actually present in the ELF file - i.e., the program's bss section.
//
// All this is very similar to what our boot loader does, except the boot
// loader also needs to read the code from disk. Take a look at
// boot/main.c to get ideas.
//
// Finally, this function maps one page for the program's initial stack.
//
// load_icode panics if it encounters problems.
// - How might load_icode fail? What might be wrong with the given input?
//
static void
load_icode(struct Env *e, uint8_t *binary, size_t size)
{
// Hints:
// Load each program segment into virtual memory
// at the address specified in the ELF section header.
// You should only load segments with ph->p_type == ELF_PROG_LOAD.
// Each segment's virtual address can be found in ph->p_va
// and its size in memory can be found in ph->p_memsz.
// The ph->p_filesz bytes from the ELF binary, starting at
// 'binary + ph->p_offset', should be copied to virtual address
// ph->p_va. Any remaining memory bytes should be cleared to zero.
// (The ELF header should have ph->p_filesz <= ph->p_memsz.)
// Use functions from the previous lab to allocate and map pages.
//
// All page protection bits should be user read/write for now.
// ELF segments are not necessarily page-aligned, but you can
// assume for this function that no two segments will touch
// the same virtual page.
//
// You may find a function like segment_alloc useful.
//
// Loading the segments is much simpler if you can move data
// directly into the virtual addresses stored in the ELF binary.
// So which page directory should be in force during
// this function?
//
// Hint:
// You must also do something with the program's entry point,
// to make sure that the environment starts executing there.
// What? (See env_run() and env_pop_tf() below.)
// LAB 3: Your code here.
struct Elf *env_elf;
struct Proghdr *ph;
struct Page *pg;
int i;
unsigned int old_cr3;
env_elf = (struct Elf *)binary;
old_cr3 = rcr3();
lcr3(PADDR(e->env_pgdir));
if( env_elf->e_magic != ELF_MAGIC)
return;
ph = (struct Proghdr*)((unsigned int)env_elf + env_elf->e_phoff);
for(i=0; i < env_elf->e_phnum;i++){
if(ph->p_type == ELF_PROG_LOAD){
segment_alloc(e,(void *)ph->p_va, ph->p_memsz);
memset((void *)ph->p_va, 0, ph->p_memsz);
memmove((void *)ph->p_va, (void *)((unsigned int)env_elf + ph->p_offset), ph->p_filesz);
}
ph++;
}
e->env_tf.tf_eip = env_elf->e_entry;
// Now map one page for the program's initial stack
// at virtual address USTACKTOP - PGSIZE.
// LAB 3: Your code here.
if(page_alloc(&pg) != 0){
cprintf("load_icode page_alloc fail!!\n");
return;
}
page_insert(e->env_pgdir, pg, (void *)(USTACKTOP - PGSIZE), PTE_U | PTE_W);
lcr3(old_cr3);
}
// Try to send 'value' to the target env 'envid'.
// If srcva < UTOP, then also send page currently mapped at 'srcva',
// so that receiver gets a duplicate mapping of the same page.
//
// The send fails with a return value of -E_IPC_NOT_RECV if the
// target is not blocked, waiting for an IPC.
//
// The send also can fail for the other reasons listed below.
//
// Otherwise, the send succeeds, and the target's ipc fields are
// updated as follows:
// env_ipc_recving is set to 0 to block future sends;
// env_ipc_from is set to the sending envid;
// env_ipc_value is set to the 'value' parameter;
// env_ipc_perm is set to 'perm' if a page was transferred, 0 otherwise.
// The target environment is marked runnable again, returning 0
// from the paused sys_ipc_recv system call. (Hint: does the
// sys_ipc_recv function ever actually return?)
//
// If the sender wants to send a page but the receiver isn't asking for one,
// then no page mapping is transferred, but no error occurs.
// The ipc only happens when no errors occur.
//
// Returns 0 on success, < 0 on error.
// Errors are:
// -E_BAD_ENV if environment envid doesn't currently exist.
// (No need to check permissions.)
// -E_IPC_NOT_RECV if envid is not currently blocked in sys_ipc_recv,
// or another environment managed to send first.
// -E_INVAL if srcva < UTOP but srcva is not page-aligned.
// -E_INVAL if srcva < UTOP and perm is inappropriate
// (see sys_page_alloc).
// -E_INVAL if srcva < UTOP but srcva is not mapped in the caller's
// address space.
// -E_INVAL if (perm & PTE_W), but srcva is read-only in the
// current environment's address space.
// -E_NO_MEM if there's not enough memory to map srcva in envid's
// address space.
static int
sys_ipc_try_send(envid_t envid, uint32_t value, void *srcva, unsigned perm)
{
// LAB 4: Your code here.
/* lj */
struct Env *denv = NULL;
struct Page *pg = NULL;
pte_t *pte = NULL;
int ret = 0;
//cprintf("%x try_send to %x [%d]\n", curenv->env_id, envid, value);
if((ret = envid2env(envid, &denv, 0)) < 0) {
return -E_BAD_ENV;
}
else if(!denv->env_ipc_recving) {
//cprintf("00\n");
return -E_IPC_NOT_RECV;
}
if((void *)-1 != srcva) {
if((size_t)srcva > UTOP) {
return -E_INVAL;
}
else if((size_t)srcva & (PGSIZE-1)) {
return -E_INVAL;
}
else if((PTE_U | PTE_P) != (perm & (PTE_U | PTE_P))) {
return -E_INVAL;
}
else if(perm & ~(PTE_U | PTE_P | PTE_W | PTE_AVAIL)) {
return -E_INVAL;
}
else if(NULL == (pg = page_lookup(curenv->env_pgdir, srcva, &pte))) {
return -E_INVAL;
}
else if((perm & PTE_W) != (*pte & PTE_W)) {
return -E_INVAL;
}
}
denv->env_ipc_from = curenv->env_id;
denv->env_ipc_value = value;
if((void *)-1 != srcva) {
if((void *)-1 == denv->env_ipc_dstva) {
denv->env_ipc_perm = 0;
}
else if((ret = page_insert(denv->env_pgdir, pg, denv->env_ipc_dstva, perm)) < 0) {
}
else {
denv->env_ipc_perm = perm;
}
}
/* After a env sent ipc to denv,
* set denv runnable whether it is successfully sent
* */
denv->env_ipc_recving = 0;
denv->env_tf.tf_regs.reg_eax = ret;
denv->env_status = ENV_RUNNABLE;
//panic("sys_ipc_try_send not implemented");
//cprintf("ret %d\n",ret);
//cprintf("%x try_send to %x [%d] awake %x\n", curenv->env_id, envid, value, denv->env_id);
return ret;
}
// Try to send 'value' to the target env 'envid'.
// If srcva < UTOP, then also send page currently mapped at 'srcva',
// so that receiver gets a duplicate mapping of the same page.
//
// The send fails with a return value of -E_IPC_NOT_RECV if the
// target is not blocked, waiting for an IPC.
//
// The send also can fail for the other reasons listed below.
//
// Otherwise, the send succeeds, and the target's ipc fields are
// updated as follows:
// env_ipc_recving is set to 0 to block future sends;
// env_ipc_from is set to the sending envid;
// env_ipc_value is set to the 'value' parameter;
// env_ipc_perm is set to 'perm' if a page was transferred, 0 otherwise.
// The target environment is marked runnable again, returning 0
// from the paused sys_ipc_recv system call. (Hint: does the
// sys_ipc_recv function ever actually return?)
//
// If the sender wants to send a page but the receiver isn't asking for one,
// then no page mapping is transferred, but no error occurs.
// The ipc only happens when no errors occur.
//
// Returns 0 on success, < 0 on error.
// Errors are:
// -E_BAD_ENV if environment envid doesn't currently exist.
// (No need to check permissions.)
// -E_IPC_NOT_RECV if envid is not currently blocked in sys_ipc_recv,
// or another environment managed to send first.
// -E_INVAL if srcva < UTOP but srcva is not page-aligned.
// -E_INVAL if srcva < UTOP and perm is inappropriate
// (see sys_page_alloc).
// -E_INVAL if srcva < UTOP but srcva is not mapped in the caller's
// address space.
// -E_INVAL if (perm & PTE_W), but srcva is read-only in the
// current environment's address space.
// -E_NO_MEM if there's not enough memory to map srcva in envid's
// address space.
static int
sys_ipc_try_send(envid_t envid, uint32_t value, void *srcva, unsigned perm)
{
// LAB 4: Your code here.
struct Env *e;
int ret = envid2env(envid, &e, 0);
if(ret)
return ret;
if(!e->env_ipc_recving)
return -E_IPC_NOT_RECV;
if(srcva < (void*)UTOP)
{
pte_t *pte;
struct PageInfo *pg = page_lookup(curenv->env_pgdir, srcva, &pte);
if (!pg) return -E_INVAL;
//if ((*pte & perm) != perm) return -E_INVAL;
if ((perm & PTE_W) && !(*pte & PTE_W)) return -E_INVAL;
if (srcva != ROUNDDOWN(srcva, PGSIZE)) return -E_INVAL;
if (e->env_ipc_dstva < (void*)UTOP)
{
ret = page_insert(e->env_pgdir, pg, e->env_ipc_dstva, perm);
if (ret) return ret;
e->env_ipc_perm = perm;
}
}
e->env_ipc_recving = 0;
e->env_ipc_from = curenv->env_id;
e->env_ipc_value = value;
e->env_status = ENV_RUNNABLE;
e->env_tf.tf_regs.reg_eax = 0;
return 0;
/*
struct Env *e;
struct PageInfo *page;
pte_t *pte;
if (envid2env(envid, &e, 0) != 0)
return -E_BAD_ENV;
if (e->env_ipc_recving == 0)
return -E_IPC_NOT_RECV;
e->env_ipc_recving = 0;
e->env_ipc_from = curenv->env_id;
e->env_ipc_value = value;
if ((uint32_t) srcva < UTOP) {
if ((uint32_t) srcva % PGSIZE != 0)
return -E_INVAL;
if ((perm & PTE_U) == 0
|| (perm & PTE_P) == 0
|| (perm & ~PTE_SYSCALL) != 0)
return -E_INVAL;
if ((page = page_lookup(curenv->env_pgdir, srcva, &pte)) == NULL)
return -E_INVAL;
if ((perm & PTE_W) != 0 && (*pte & PTE_W) == 0)
return -E_INVAL;
if ((page_insert(e->env_pgdir, page, e->env_ipc_dstva, perm))!=0) {
return -E_NO_MEM;
}
cprintf("sys_ipc_try_send: from 0x%x\n", e->env_ipc_from);
e->env_ipc_perm = perm;
} else {
e->env_ipc_perm = 0;
//cprintf("sys_ipc_try_send: perm=0 from 0x%x\n", e->env_ipc_from);
}
e->env_status = ENV_RUNNABLE;
return 0;
*/
// panic("sys_ipc_try_send not implemented");
}
//
// Set up the initial program binary, stack, and processor flags
// for a user process.
//
// This function loads all loadable segments from the ELF binary image
// into the environment's user memory, starting at the appropriate
// virtual addresses indicated in the ELF program header.
// It also clears to zero any portions of these segments
// that are marked in the program header as being mapped
// but not actually present in the ELF file -- i.e., the program's bss section.
//
// Finally, this function maps one page for the program's initial stack.
//
// load_elf panics if it encounters problems.
// - How might load_elf fail? What might be wrong with the given input?
//
static void
load_elf(struct Env *e, uint8_t *binary, size_t size)
{
struct Elf *elf = (struct Elf *) binary;
// Load each program segment into environment 'e's virtual memory
// at the address specified in the ELF section header.
// Only load segments with ph->p_type == ELF_PROG_LOAD.
// Each segment's virtual address can be found in ph->p_va
// and its size in memory can be found in ph->p_memsz.
// The ph->p_filesz bytes from the ELF binary, starting at
// 'binary + ph->p_offset', should be copied to virtual address
// ph->p_va. Any remaining memory bytes should be cleared to zero.
// (The ELF header should have ph->p_filesz <= ph->p_memsz.)
// Use functions from the previous lab to allocate and map pages.
//
// All page protection bits should be user read/write for now.
// ELF segments are not necessarily page-aligned, but you can
// assume for this function that no two segments will touch
// the same virtual page.
//
// You may find a function like segment_alloc useful.
//
// Loading the segments is much simpler if you can move data
// directly into the virtual addresses stored in the ELF binary.
// So which page directory should be in force during
// this function?
//
// All this is very similar to what our boot loader does, except the
// boot loader reads the code from disk and doesn't check whether
// segments are loadable. Take a look at boot/main.c to get ideas.
//
// You must also store the program's entry point somewhere,
// to make sure that the environment starts executing at that point.
// See env_run() and env_iret() below.
// LAB 3: Your code here.
lcr3(PADDR(e->env_pgdir));
if (elf->e_magic != ELF_MAGIC)
panic("Invalid Elf Magic");
struct Proghdr *ph = (struct Proghdr *) ((uint8_t *) elf + elf->e_phoff);
int ph_num = elf->e_phnum;
// iterate over all program headers
for (; --ph_num >= 0; ph++)
if (ph->p_type == ELF_PROG_LOAD)
{
segment_alloc(e, ph->p_va, ph->p_memsz);
// copy data from binary to address space
memmove((void *)ph->p_va, binary + ph->p_offset, ph->p_filesz);
}
// set entry point for new env
e->env_tf.tf_eip = elf->e_entry;
e->env_tf.tf_esp = USTACKTOP;
// Now map one page for the program's initial stack
// at virtual address USTACKTOP - PGSIZE.
// (What should the permissions be?)
struct Page *p;
if ((p = page_alloc()) == NULL || page_insert(e->env_pgdir, p, USTACKTOP-PGSIZE, PTE_U|PTE_W|PTE_P))
panic("segment_alloc: Can't allocate page");
memset(page2kva(p), 0, PGSIZE);
}
// Block until a value is ready. Record that you want to receive
// using the env_ipc_recving and env_ipc_dstva fields of struct Env,
// mark yourself not runnable, and then give up the CPU.
//
// If 'dstva' is < UTOP, then you are willing to receive a page of data.
// 'dstva' is the virtual address at which the sent page should be mapped.
//
// This function only returns on error, but the system call will eventually
// return 0 on success.
// Return < 0 on error. Errors are:
// -E_INVAL if dstva < UTOP but dstva is not page-aligned.
static int
sys_ipc_recv(void *dstva)
{
// LAB 4: Your code here.
if((uint32_t)dstva < UTOP)
{
if((uint32_t)dstva % PGSIZE != 0)
{
return -E_INVAL;
}
curenv->env_ipc_dstva = dstva;
}
else
{
curenv->env_ipc_dstva = (void*)0xFFFFFFFF;
}
// LAB 4 CHALLENGE: Check if another env has queued a send to us
// If they do, receive it and return, short circuiting on the first
// received
// i keeps track of the last index received from
static int i = 0;
static int recvs = RECV_LIMIT;
// k is the counting index,
int k = i;
// j is the stop index, when j == k we end
int j;
// Loop over all envs circularly starting with the last one we received
// from
// Every RECV_LIMIT receives from the same env, start one after the
// last one we received from.
// This should eliminate livelock when one env is repeatedly queueing
// messages to this env faster than we receive, but it still provides
// good performance when we only want to receive from a small number of
// envs
int shift_j = 1;
for(j = k-1; k != j ; k = (k+1)%NENV)
{
// We need to shift j up 1 to that envs[i] is checked
// This solves the problem of finding the end of the ring
if(shift_j)
{
j = (j+1) % NENV;
shift_j = 0;
}
struct Env *env = &envs[k];
if(env->env_status == ENV_NOT_RUNNABLE &&
env->env_ipc_send_to == curenv->env_id)
{
// This environment has a send waiting for us!
curenv->env_ipc_value = env->env_ipc_send_value;
curenv->env_ipc_from = env->env_id;
// Map the page iff they sent one AND we want one
if((uint32_t)env->env_ipc_send_srcva < UTOP &&
(uint32_t)curenv->env_ipc_dstva < UTOP)
{
int r;
struct Page *page = page_lookup(env->env_pgdir,
env->env_ipc_send_srcva, NULL);
r = page_insert(curenv->env_pgdir, page, curenv->env_ipc_dstva,
env->env_ipc_send_perm);
if(r < 0)
{
// Error mapping, we need to make the send call receive the
// error and ignore it here as if it never happened
env->env_ipc_send_to = 0;
env->env_tf.tf_regs.reg_eax = r;
env->env_status = ENV_RUNNABLE;
continue;
}
curenv->env_ipc_perm = env->env_ipc_send_perm;
}
else
{
// If we do not map, clear perm
curenv->env_ipc_perm = 0;
}
// We have received it, check that we do not need to increment i
if(k == i)
{
if (--recvs == 0)
{
// If we have received RECV_LIMIT messages in a row from the same
// env, start checking one past that env next time to prevent
// livelock
recvs = RECV_LIMIT;
k++;
}
}
else
{
recvs = RECV_LIMIT;
}
i = k;
// Queued send has been received, ready the other env and return
env->env_ipc_send_to = 0;
env->env_tf.tf_regs.reg_eax = 0;
env->env_status = ENV_RUNNABLE;
return 0;
}
}
curenv->env_ipc_recving = 1;
curenv->env_status = ENV_NOT_RUNNABLE;
sched_yield();
return 0;
}
// Try to send 'value' to the target env 'envid'.
// If srcva < UTOP, then also send page currently mapped at 'srcva',
// so that receiver gets a duplicate mapping of the same page.
//
// The send fails with a return value of -E_IPC_NOT_RECV if the
// target is not blocked, waiting for an IPC.
//
// The send also can fail for the other reasons listed below.
//
// Otherwise, the send succeeds, and the target's ipc fields are
// updated as follows:
// env_ipc_recving is set to 0 to block future sends;
// env_ipc_from is set to the sending envid;
// env_ipc_value is set to the 'value' parameter;
// env_ipc_perm is set to 'perm' if a page was transferred, 0 otherwise.
// The target environment is marked runnable again, returning 0
// from the paused sys_ipc_recv system call. (Hint: does the
// sys_ipc_recv function ever actually return?)
//
// If the sender wants to send a page but the receiver isn't asking for one,
// then no page mapping is transferred, but no error occurs.
// The ipc only happens when no errors occur.
//
// Returns 0 on success, < 0 on error.
// Errors are:
// -E_BAD_ENV if environment envid doesn't currently exist.
// (No need to check permissions.)
// -E_IPC_NOT_RECV if envid is not currently blocked in sys_ipc_recv,
// or another environment managed to send first.
// -E_INVAL if srcva < UTOP but srcva is not page-aligned.
// -E_INVAL if srcva < UTOP and perm is inappropriate
// (see sys_page_alloc).
// -E_INVAL if srcva < UTOP but srcva is not mapped in the caller's
// address space.
// -E_INVAL if (perm & PTE_W), but srcva is read-only in the
// current environment's address space.
// -E_NO_MEM if there's not enough memory to map srcva in envid's
// address space.
static int
sys_ipc_try_send(envid_t envid, uint32_t value, void *srcva, unsigned perm)
{
// LAB 4: Your code here.
struct Env *env;
int r;
uint32_t srcint = (uint32_t)srcva;
struct Page *srcpage;
// Make sure env exists
if ((r = envid2env(envid, &env, 0) < 0))
return r;
int map_page = 0; // By default, we do not map a page
// First check that our input makes sense
if (srcint < UTOP)
{
// Check if srcva < UTOP but not page-aligned
if(srcint % PGSIZE != 0)
{
return -E_INVAL;
}
// Check that permissions are appropriate
if (!( (perm & PTE_P) &&
(perm & PTE_U) &&
!(perm & ~(PTE_P | PTE_U | PTE_AVAIL | PTE_W))))
{
return -E_INVAL;
}
// Make sure that srcva is mapped in current env
pte_t *pte;
if ((srcpage = page_lookup(curenv->env_pgdir, srcva, &pte)) == NULL)
{
return -E_INVAL;
}
// Ensure that source and perm are either both RO or W
if ((perm & PTE_W) != (*pte & PTE_W))
{
return -E_INVAL;
}
// Page mapping checks out, we are mapping a page if
// they want it
map_page = 1;
}
// Make sure env receiving
if (!env->env_ipc_recving)
{
// CHALLENGE: If the env is not receiving, set up
// send fields in our env and sleep for response.
// Note: The receiver now has the responsibility to
// map its own page based off of what we have entered in
// our struct (the va is guaranteed to be valid if it is < UTOP)
// It must also set our EAX to the error code for the mapping.
curenv->env_ipc_send_to = envid;
curenv->env_ipc_send_value = value;
curenv->env_ipc_send_srcva = srcva;
curenv->env_ipc_send_perm = perm;
// Sleep and yield
curenv->env_status = ENV_NOT_RUNNABLE;
sched_yield();
}
// The env is receiving, set their fields
// Only map the page if our checks passed AND the destination VA
// is < UTOP
if(map_page && (uint32_t) env->env_ipc_dstva < UTOP)
{
// Map the page
r = page_insert(env->env_pgdir, srcpage, env->env_ipc_dstva, perm);
if(r < 0)
return r;
// Update perm
env->env_ipc_perm = perm;
}
// If we made it here, there were no errors, update fields
env->env_ipc_recving = 0;
env->env_ipc_from = curenv->env_id;
env->env_ipc_value = value;
env->env_status = ENV_RUNNABLE;
env->env_tf.tf_regs.reg_eax = 0;
return 0;
}
/** @brief Function deals with the page fault
*
* This is an amazing function.
*
* @param void
* @return void
*/
void page_fault_handler(void){
int fault_addr = get_cr2();
mutex_lock(&cur_task->pcb_mutex);
uint32_t align_addr = fault_addr & PGALIGN_MASK;
uint32_t *ptep = NULL;
Page *phy_page = NULL;
phy_page = page_lookup(cur_task->task_pgdir, align_addr, &ptep);
uint32_t pte = *ptep;
/** Catch COW page fualt */
if((ptep != NULL) &&(pte & PTE_P) && (pte & PTE_COW))
{
mutex_lock(&mtx_m.frame_mutex);
if(phy_page->pp_ref == 1)
{
*ptep = (pte | PTE_W) &(~PTE_COW);
mutex_unlock(&mtx_m.frame_mutex);
}
else{
mutex_unlock(&mtx_m.frame_mutex);
if(pte & PTE_RWMARK){
lprintf("ERROR: Cannot access TXT or ro_data area!");
mutex_unlock(&cur_task->pcb_mutex);
sys_vanish();
return;
}
Page *new_page = page_alloc();
if(new_page == NULL){
lprintf("ERROR: No pages for COW in page_fault_handler");
mutex_unlock(&cur_task->pcb_mutex);
sys_vanish();
return;
}
uint32_t *temp_page = smemalign(PAGE_SIZE, PAGE_SIZE);
if (temp_page == NULL){
lprintf("ERROR: No memory for temp_page in page_fault_handler!");
mutex_unlock(&cur_task->pcb_mutex);
sys_vanish();
return;
}
/* Copy the physical page to a temporary page in kernel space */
memcpy((void*)temp_page, (void*)align_addr, PAGE_SIZE);
if(page_insert(cur_task->task_pgdir, new_page, align_addr,
PTE_P | PTE_U | PTE_W) < 0)
{
lprintf("ERROR: No memory for COW in page_fault_handler");
mutex_unlock(&cur_task->pcb_mutex);
sys_vanish();
return;
}
/* Copy the content to the new mapped physical page */
memcpy((void*)align_addr, (void*)temp_page, PAGE_SIZE);
/* Free the temp physical page */
sfree(temp_page, PAGE_SIZE);
mutex_lock(&mtx_m.frame_mutex);
phy_page->pp_ref--;
mutex_unlock(&mtx_m.frame_mutex);
}
}
/** Catch the ZFOD */
else if ((ptep != NULL) &&(pte & PTE_P) && (pte & PTE_ZFOD))
{
Page *pg = page_alloc();
if(pg == NULL){
lprintf("ERROR: No pages for ZFOD in page_fault_handler");
mutex_unlock(&cur_task->pcb_mutex);
sys_vanish();
return;
}
uint32_t perm = PTE_P | PTE_U | PTE_W;
if(page_insert(cur_task->task_pgdir, pg, align_addr, perm) < 0)
{
lprintf("ERROR: No memory for ZFOD in page_fault_handler");
mutex_unlock(&cur_task->pcb_mutex);
sys_vanish();
return;
}
bzero((void*)align_addr, PAGE_SIZE);
}
/* Check if installed swexn handler can fix this up */
else if(cur_thread->swexn_eip != NULL)
{
mutex_unlock(&cur_task->pcb_mutex);
swexn_handler(HAS_ERROR_CODE, SWEXN_CAUSE_PAGEFAULT);
return;
}
else{
mutex_unlock(&cur_task->pcb_mutex);
sys_vanish();
return;
}
mutex_unlock(&cur_task->pcb_mutex);
return;
}
// check page_insert, page_remove, &c
static void
check_page(void)
{
struct Page *pp, *pp0, *pp1, *pp2;
struct Page *fl;
pte_t *ptep, *ptep1;
uintptr_t va;
int i;
// should be able to allocate three pages
pp0 = pp1 = pp2 = 0;
assert((pp0 = page_alloc()));
assert((pp1 = page_alloc()));
assert((pp2 = page_alloc()));
assert(pp0);
assert(pp1 && pp1 != pp0);
assert(pp2 && pp2 != pp1 && pp2 != pp0);
// temporarily steal the rest of the free pages
fl = page_free_list;
page_free_list = 0;
// should be no free memory
assert(!page_alloc());
// there is no page allocated at address 0
assert(page_lookup(kern_pgdir, 0x0, &ptep) == NULL);
// there is no free memory, so we can't allocate a page table
assert(page_insert(kern_pgdir, pp1, 0x0, PTE_W) < 0);
// free pp0 and try again: pp0 should be used for page table
page_free(pp0);
assert(page_insert(kern_pgdir, pp1, 0x0, PTE_W) == 0);
assert(PTE_ADDR(kern_pgdir[0]) == page2pa(pp0));
assert(check_va2pa(kern_pgdir, 0x0) == page2pa(pp1));
assert(pp1->pp_ref == 1);
assert(pp0->pp_ref == 1);
// should be able to map pp2 at PGSIZE because pp0 is already allocated for page table
assert(page_insert(kern_pgdir, pp2, PGSIZE, PTE_W) == 0);
assert(check_va2pa(kern_pgdir, PGSIZE) == page2pa(pp2));
assert(pp2->pp_ref == 1);
// should be no free memory
assert(!page_alloc());
// should be able to map pp2 at PGSIZE because it's already there
assert(page_insert(kern_pgdir, pp2, PGSIZE, PTE_W) == 0);
assert(check_va2pa(kern_pgdir, PGSIZE) == page2pa(pp2));
assert(pp2->pp_ref == 1);
// pp2 should NOT be on the free list
// could happen in ref counts are handled sloppily in page_insert
assert(!page_alloc());
// check that pgdir_walk returns a pointer to the pte
ptep = (pte_t *) KADDR(PTE_ADDR(kern_pgdir[PDX(PGSIZE)]));
assert(pgdir_walk(kern_pgdir, PGSIZE, false) == ptep+PTX(PGSIZE));
// should be able to change permissions too.
assert(page_insert(kern_pgdir, pp2, PGSIZE, PTE_W|PTE_U) == 0);
assert(check_va2pa(kern_pgdir, PGSIZE) == page2pa(pp2));
assert(pp2->pp_ref == 1);
assert(*pgdir_walk(kern_pgdir, PGSIZE, false) & PTE_U);
assert(kern_pgdir[0] & PTE_U);
// should not be able to map at PTSIZE because need free page for page table
assert(page_insert(kern_pgdir, pp0, PTSIZE, PTE_W) < 0);
// insert pp1 at PGSIZE (replacing pp2)
assert(page_insert(kern_pgdir, pp1, PGSIZE, PTE_W) == 0);
assert(!(*pgdir_walk(kern_pgdir, PGSIZE, false) & PTE_U));
// should have pp1 at both 0 and PGSIZE, pp2 nowhere, ...
assert(check_va2pa(kern_pgdir, 0) == page2pa(pp1));
assert(check_va2pa(kern_pgdir, PGSIZE) == page2pa(pp1));
// ... and ref counts should reflect this
assert(pp1->pp_ref == 2);
assert(pp2->pp_ref == 0);
// pp2 should be returned by page_alloc
assert((pp = page_alloc()) && pp == pp2);
// unmapping pp1 at 0 should keep pp1 at PGSIZE
page_remove(kern_pgdir, 0x0);
assert(check_va2pa(kern_pgdir, 0x0) == (physaddr_t) ~0);
assert(check_va2pa(kern_pgdir, PGSIZE) == page2pa(pp1));
assert(pp1->pp_ref == 1);
assert(pp2->pp_ref == 0);
// unmapping pp1 at PGSIZE should free it
page_remove(kern_pgdir, PGSIZE);
assert(check_va2pa(kern_pgdir, 0x0) == (physaddr_t) ~0);
assert(check_va2pa(kern_pgdir, PGSIZE) == (physaddr_t) ~0);
assert(pp1->pp_ref == 0);
assert(pp2->pp_ref == 0);
// so it should be returned by page_alloc
assert((pp = page_alloc()) && pp == pp1);
// should be no free memory
assert(!page_alloc());
// forcibly take pp0 back
assert(PTE_ADDR(kern_pgdir[0]) == page2pa(pp0));
kern_pgdir[0] = 0;
assert(pp0->pp_ref == 1);
pp0->pp_ref = 0;
// check pointer arithmetic in pgdir_walk
page_free(pp0);
va = PGSIZE * NPDENTRIES + PGSIZE;
ptep = pgdir_walk(kern_pgdir, va, true);
ptep1 = (pte_t *) KADDR(PTE_ADDR(kern_pgdir[PDX(va)]));
assert(ptep == ptep1 + PTX(va));
kern_pgdir[PDX(va)] = 0;
pp0->pp_ref = 0;
// check that new page tables get cleared
memset(page2kva(pp0), 0xFF, PGSIZE);
page_free(pp0);
pgdir_walk(kern_pgdir, 0x0, true);
ptep = (pte_t *) page2kva(pp0);
for(i=0; i<NPTENTRIES; i++)
assert((ptep[i] & PTE_P) == 0);
kern_pgdir[0] = 0;
pp0->pp_ref = 0;
// give free list back
page_free_list = fl;
// free the pages we took
page_free(pp0);
page_free(pp1);
page_free(pp2);
cprintf("check_page() succeeded!\n");
}
// check_swap - check the correctness of swap & page replacement algorithm
static void
check_swap(void) {
size_t nr_free_pages_store = nr_free_pages();
size_t slab_allocated_store = slab_allocated();
size_t offset;
for (offset = 2; offset < max_swap_offset; offset ++) {
mem_map[offset] = 1;
}
struct mm_struct *mm = mm_create();
assert(mm != NULL);
extern struct mm_struct *check_mm_struct;
assert(check_mm_struct == NULL);
check_mm_struct = mm;
pgd_t *pgdir = mm->pgdir = boot_pgdir;
assert(pgdir[0] == 0);
struct vma_struct *vma = vma_create(0, PTSIZE, VM_WRITE | VM_READ);
assert(vma != NULL);
insert_vma_struct(mm, vma);
struct Page *rp0 = alloc_page(), *rp1 = alloc_page();
assert(rp0 != NULL && rp1 != NULL);
uint32_t perm = PTE_U | PTE_W;
int ret = page_insert(pgdir, rp1, 0, perm);
assert(ret == 0 && page_ref(rp1) == 1);
page_ref_inc(rp1);
ret = page_insert(pgdir, rp0, 0, perm);
assert(ret == 0 && page_ref(rp1) == 1 && page_ref(rp0) == 1);
// check try_alloc_swap_entry
swap_entry_t entry = try_alloc_swap_entry();
assert(swap_offset(entry) == 1);
mem_map[1] = 1;
assert(try_alloc_swap_entry() == 0);
// set rp1, Swap, Active, add to hash_list, active_list
swap_page_add(rp1, entry);
swap_active_list_add(rp1);
assert(PageSwap(rp1));
mem_map[1] = 0;
entry = try_alloc_swap_entry();
assert(swap_offset(entry) == 1);
assert(!PageSwap(rp1));
// check swap_remove_entry
assert(swap_hash_find(entry) == NULL);
mem_map[1] = 2;
swap_remove_entry(entry);
assert(mem_map[1] == 1);
swap_page_add(rp1, entry);
swap_inactive_list_add(rp1);
swap_remove_entry(entry);
assert(PageSwap(rp1));
assert(rp1->index == entry && mem_map[1] == 0);
// check page_launder, move page from inactive_list to active_list
assert(page_ref(rp1) == 1);
assert(nr_active_pages == 0 && nr_inactive_pages == 1);
assert(list_next(&(inactive_list.swap_list)) == &(rp1->swap_link));
page_launder();
assert(nr_active_pages == 1 && nr_inactive_pages == 0);
assert(PageSwap(rp1) && PageActive(rp1));
entry = try_alloc_swap_entry();
assert(swap_offset(entry) == 1);
assert(!PageSwap(rp1) && nr_active_pages == 0);
assert(list_empty(&(active_list.swap_list)));
// set rp1 inactive again
assert(page_ref(rp1) == 1);
swap_page_add(rp1, 0);
assert(PageSwap(rp1) && swap_offset(rp1->index) == 1);
swap_inactive_list_add(rp1);
mem_map[1] = 1;
assert(nr_inactive_pages == 1);
page_ref_dec(rp1);
size_t count = nr_free_pages();
swap_remove_entry(entry);
assert(nr_inactive_pages == 0 && nr_free_pages() == count + 1);
// check swap_out_mm
pte_t *ptep0 = get_pte(pgdir, 0, 0), *ptep1;
assert(ptep0 != NULL && pte2page(*ptep0) == rp0);
ret = swap_out_mm(mm, 0);
assert(ret == 0);
ret = swap_out_mm(mm, 10);
assert(ret == 1 && mm->swap_address == PGSIZE);
ret = swap_out_mm(mm, 10);
assert(ret == 0 && *ptep0 == entry && mem_map[1] == 1);
assert(PageDirty(rp0) && PageActive(rp0) && page_ref(rp0) == 0);
assert(nr_active_pages == 1 && list_next(&(active_list.swap_list)) == &(rp0->swap_link));
// check refill_inactive_scan()
refill_inactive_scan();
assert(!PageActive(rp0) && page_ref(rp0) == 0);
assert(nr_inactive_pages == 1 && list_next(&(inactive_list.swap_list)) == &(rp0->swap_link));
page_ref_inc(rp0);
page_launder();
assert(PageActive(rp0) && page_ref(rp0) == 1);
assert(nr_active_pages == 1 && list_next(&(active_list.swap_list)) == &(rp0->swap_link));
page_ref_dec(rp0);
refill_inactive_scan();
assert(!PageActive(rp0));
// save data in rp0
int i;
for (i = 0; i < PGSIZE; i ++) {
((char *)page2kva(rp0))[i] = (char)i;
}
page_launder();
assert(nr_inactive_pages == 0 && list_empty(&(inactive_list.swap_list)));
assert(mem_map[1] == 1);
rp1 = alloc_page();
assert(rp1 != NULL);
ret = swapfs_read(entry, rp1);
assert(ret == 0);
for (i = 0; i < PGSIZE; i ++) {
assert(((char *)page2kva(rp1))[i] == (char)i);
}
// page fault now
*(char *)0 = 0xEF;
rp0 = pte2page(*ptep0);
assert(page_ref(rp0) == 1);
assert(PageSwap(rp0) && PageActive(rp0));
entry = try_alloc_swap_entry();
assert(swap_offset(entry) == 1 && mem_map[1] == SWAP_UNUSED);
assert(!PageSwap(rp0) && nr_active_pages == 0 && nr_inactive_pages == 0);
// clear accessed flag
assert(rp0 == pte2page(*ptep0));
assert(!PageSwap(rp0));
ret = swap_out_mm(mm, 10);
assert(ret == 0);
assert(!PageSwap(rp0) && (*ptep0 & PTE_P));
// change page table
ret = swap_out_mm(mm, 10);
assert(ret == 1);
assert(*ptep0 == entry && page_ref(rp0) == 0 && mem_map[1] == 1);
count = nr_free_pages();
refill_inactive_scan();
page_launder();
assert(count + 1 == nr_free_pages());
ret = swapfs_read(entry, rp1);
assert(ret == 0 && *(char *)(page2kva(rp1)) == (char)0xEF);
free_page(rp1);
// duplictate *ptep0
ptep1 = get_pte(pgdir, PGSIZE, 0);
assert(ptep1 != NULL && *ptep1 == 0);
swap_duplicate(*ptep0);
*ptep1 = *ptep0;
// page fault again
// update for copy on write
*(char *)1 = 0x88;
*(char *)(PGSIZE) = 0x8F;
*(char *)(PGSIZE + 1) = 0xFF;
assert(pte2page(*ptep0) != pte2page(*ptep1));
assert(*(char *)0 == (char)0xEF);
assert(*(char *)1 == (char)0x88);
assert(*(char *)(PGSIZE) == (char)0x8F);
assert(*(char *)(PGSIZE + 1) == (char)0xFF);
rp0 = pte2page(*ptep0);
rp1 = pte2page(*ptep1);
assert(!PageSwap(rp0) && PageSwap(rp1) && PageActive(rp1));
entry = try_alloc_swap_entry();
assert(!PageSwap(rp0) && !PageSwap(rp1));
assert(swap_offset(entry) == 1 && mem_map[1] == SWAP_UNUSED);
assert(list_empty(&(active_list.swap_list)));
assert(list_empty(&(inactive_list.swap_list)));
page_insert(pgdir, rp0, PGSIZE, perm | PTE_A);
// check swap_out_mm
*(char *)0 = *(char *)PGSIZE = 0xEE;
mm->swap_address = PGSIZE * 2;
ret = swap_out_mm(mm, 2);
assert(ret == 0);
assert((*ptep0 & PTE_P) && !(*ptep0 & PTE_A));
assert((*ptep1 & PTE_P) && !(*ptep1 & PTE_A));
ret = swap_out_mm(mm, 2);
assert(ret == 2);
assert(mem_map[1] == 2 && page_ref(rp0) == 0);
refill_inactive_scan();
page_launder();
assert(mem_map[1] == 2 && swap_hash_find(entry) == NULL);
// check copy entry
swap_remove_entry(entry);
*ptep1 = 0;
assert(mem_map[1] == 1);
swap_entry_t store;
ret = swap_copy_entry(entry, &store);
assert(ret == -E_NO_MEM);
mem_map[2] = SWAP_UNUSED;
ret = swap_copy_entry(entry, &store);
assert(ret == 0 && swap_offset(store) == 2 && mem_map[2] == 0);
mem_map[2] = 1;
*ptep1 = store;
assert(*(char *)PGSIZE == (char)0xEE && *(char *)(PGSIZE + 1)== (char)0x88);
*(char *)PGSIZE = 1, *(char *)(PGSIZE + 1) = 2;
assert(*(char *)0 == (char)0xEE && *(char *)1 == (char)0x88);
ret = swap_in_page(entry, &rp0);
assert(ret == 0);
ret = swap_in_page(store, &rp1);
assert(ret == 0);
assert(rp1 != rp0);
// free memory
swap_list_del(rp0), swap_list_del(rp1);
swap_page_del(rp0), swap_page_del(rp1);
assert(page_ref(rp0) == 1 && page_ref(rp1) == 1);
assert(nr_active_pages == 0 && list_empty(&(active_list.swap_list)));
assert(nr_inactive_pages == 0 && list_empty(&(inactive_list.swap_list)));
for (i = 0; i < HASH_LIST_SIZE; i ++) {
assert(list_empty(hash_list + i));
}
page_remove(pgdir, 0);
page_remove(pgdir, PGSIZE);
free_page(pa2page(PMD_ADDR(*get_pmd(pgdir, 0, 0))));
free_page(pa2page(PUD_ADDR(*get_pud(pgdir, 0, 0))));
free_page(pa2page(PGD_ADDR(*get_pgd(pgdir, 0, 0))));
pgdir[0] = 0;
mm->pgdir = NULL;
mm_destroy(mm);
check_mm_struct = NULL;
assert(nr_active_pages == 0 && nr_inactive_pages == 0);
for (offset = 0; offset < max_swap_offset; offset ++) {
mem_map[offset] = SWAP_UNUSED;
}
assert(nr_free_pages_store == nr_free_pages());
assert(slab_allocated_store == slab_allocated());
cprintf("check_swap() succeeded.\n");
}
// Map the page of memory at 'srcva' in srcenvid's address space
// at 'dstva' in dstenvid's address space with permission 'perm'.
// Perm has the same restrictions as in sys_page_alloc, except
// that it also must not grant write access to a read-only
// page.
//
// Return 0 on success, < 0 on error. Errors are:
// -E_BAD_ENV if srcenvid and/or dstenvid doesn't currently exist,
// or the caller doesn't have permission to change one of them.
// -E_INVAL if srcva >= UTOP or srcva is not page-aligned,
// or dstva >= UTOP or dstva is not page-aligned.
// -E_INVAL is srcva is not mapped in srcenvid's address space.
// -E_INVAL if perm is inappropriate (see sys_page_alloc).
// -E_INVAL if (perm & PTE_W), but srcva is read-only in srcenvid's
// address space.
// -E_NO_MEM if there's no memory to allocate any necessary page tables.
static int
sys_page_map(envid_t srcenvid, void *srcva,
envid_t dstenvid, void *dstva, int perm)
{
// Hint: This function is a wrapper around page_lookup() and
// page_insert() from kern/pmap.c.
// Again, most of the new code you write should be to check the
// parameters for correctness.
// Use the third argument to page_lookup() to
// check the current permissions on the page.
// LAB 4: Your code here.
struct Env* srcenv;
struct Env*dstenv;
struct Page * pp;
pte_t *srcpte;
pte_t *dstpte;
/*
if(dstenvid==0)
dstenvid=curenv->env_id;
if(srcenvid==0)
*/
if(envid2env(srcenvid, &srcenv,1) < 0)
return E_BAD_ENV;
if(envid2env(dstenvid, &dstenv,1) < 0)
return E_BAD_ENV;
if((uint32_t)srcva > UTOP || ((uint32_t)srcva &(PGSIZE-1)) != 0 )
return -E_INVAL;
if((uint32_t)dstva > UTOP || ((uint32_t)dstva &(PGSIZE-1)) != 0 )
return -E_INVAL;
pp=page_lookup(srcenv->env_pgdir,srcva,0);
if(!pp)
return -E_INVAL;
srcpte = pgdir_walk(srcenv->env_pgdir,srcva,0);
if(!srcpte || !(*srcpte &PTE_P))
return -E_INVAL;
if((perm & (PTE_P | PTE_U)) != (PTE_P | PTE_U))
return -E_INVAL;
if( !(*srcpte | PTE_W) && (perm & PTE_W) )
return -E_INVAL;
/* dstpte = pgdir_walk(dstenv->env_pgdir,dstpte,1);
if(!dstpte)
return -E_NO_MEM;
*dstpte =PTE_ADDR(*srcpte) | perm;
*/
if(page_insert(dstenv->env_pgdir, pp, dstva, perm) < 0)
return -E_NO_MEM;
return 0;
//panic("sys_page_map not implemented");
}
//
// Set up the initial program binary, stack, and processor flags
// for a user process.
// This function is ONLY called during kernel initialization,
// before running the first user-mode environment.
//
// This function loads all loadable segments from the ELF binary image
// into the environment's user memory, starting at the appropriate
// virtual addresses indicated in the ELF program header.
// At the same time it clears to zero any portions of these segments
// that are marked in the program header as being mapped
// but not actually present in the ELF file - i.e., the program's bss section.
//
// All this is very similar to what our boot loader does, except the boot
// loader also needs to read the code from disk. Take a look at
// boot/main.c to get ideas.
//
// Finally, this function maps one page for the program's initial stack.
//
// load_icode panics if it encounters problems.
// - How might load_icode fail? What might be wrong with the given input?
//
static void
load_icode(struct Env *e, uint8_t *binary, size_t size)
{
// Hints:
// Load each program segment into virtual memory
// at the address specified in the ELF section header.
// You should only load segments with ph->p_type == ELF_PROG_LOAD.
// Each segment's virtual address can be found in ph->p_va
// and its size in memory can be found in ph->p_memsz.
// The ph->p_filesz bytes from the ELF binary, starting at
// 'binary + ph->p_offset', should be copied to virtual address
// ph->p_va. Any remaining memory bytes should be cleared to zero.
// (The ELF header should have ph->p_filesz <= ph->p_memsz.)
// Use functions from the previous lab to allocate and map pages.
//
// All page protection bits should be user read/write for now.
// ELF segments are not necessarily page-aligned, but you can
// assume for this function that no two segments will touch
// the same virtual page.
//
// You may find a function like segment_alloc useful.
//
// Loading the segments is much simpler if you can move data
// directly into the virtual addresses stored in the ELF binary.
// So which page directory should be in force during
// this function?
//
// Hint:
// You must also do something with the program's entry point,
// to make sure that the environment starts executing there.
// What? (See env_run() and env_pop_tf() below.)
// LAB 3: Your code here.
//cprintf("Begin to load icode\n");
struct Proghdr *ph, *eph;
ph = (struct Proghdr *) (binary + ((struct Elf *)binary)->e_phoff);
eph = ph + ((struct Elf *)binary)->e_phnum;
lcr3(e->env_cr3);
while (ph < eph)
{
if(ph->p_type == ELF_PROG_LOAD)
{
segment_alloc(e, (void*)ph->p_va,ph->p_memsz);
memcpy((void *)ph->p_va, binary + ph->p_offset, ph->p_filesz);
memset((void *)(ph->p_va + ph->p_filesz), 0x0, ph->p_memsz - ph->p_filesz);
}
ph++;
}
//lcr3(boot_cr3);
//cprintf("segment copy success\n");
// Now map one page for the program's initial stack
// at virtual address USTACKTOP - PGSIZE.
// LAB 3: Your code here.
struct Page * user_stack;
if(page_alloc(&user_stack) == -E_NO_MEM)
panic("No memory to alloc for user stack");
page_insert(e->env_pgdir, user_stack, (void *)(USTACKTOP - PGSIZE),PTE_W|PTE_U|PTE_P);
e->env_tf.tf_eip = ((struct Elf*)binary)->e_entry;
}
// Try to send 'value' to the target env 'envid'.
// If srcva < UTOP, then also send page currently mapped at 'srcva',
// so that receiver gets a duplicate mapping of the same page.
//
// The send fails with a return value of -E_IPC_NOT_RECV if the
// target is not blocked, waiting for an IPC.
//
// The send also can fail for the other reasons listed below.
//
// Otherwise, the send succeeds, and the target's ipc fields are
// updated as follows:
// env_ipc_recving is set to 0 to block future sends;
// env_ipc_from is set to the sending envid;
// env_ipc_value is set to the 'value' parameter;
// env_ipc_perm is set to 'perm' if a page was transferred, 0 otherwise.
// The target environment is marked runnable again, returning 0
// from the paused sys_ipc_recv system call. (Hint: does the
// sys_ipc_recv function ever actually return?)
//
// If the sender wants to send a page but the receiver isn't asking for one,
// then no page mapping is transferred, but no error occurs.
// The ipc only happens when no errors occur.
//
// Returns 0 on success, < 0 on error.
// Errors are:
// -E_BAD_ENV if environment envid doesn't currently exist.
// (No need to check permissions.)
// -E_IPC_NOT_RECV if envid is not currently blocked in sys_ipc_recv,
// or another environment managed to send first.
// -E_INVAL if srcva < UTOP but srcva is not page-aligned.
// -E_INVAL if srcva < UTOP and perm is inappropriate
// (see sys_page_alloc).
// -E_INVAL if srcva < UTOP but srcva is not mapped in the caller's
// address space.
// -E_INVAL if (perm & PTE_W), but srcva is read-only in the
// current environment's address space.
// -E_NO_MEM if there's not enough memory to map srcva in envid's
// address space.
static int
sys_ipc_try_send(envid_t envid, uint32_t value, void *srcva, unsigned perm)
{
// LAB 4: Your code here.
struct Env *rcv;
pte_t *pte;
struct Page *pp;
envid_t jdos_client = 0;
struct Env *e;
int i, r;
if (curenv->env_alien &&
((curenv->env_hosteid & 0xfff00000) ==
(envid & 0xfff00000))) {
goto djos_send;
}
// Is receiver valid?
if (envid2env(envid, &rcv, 0) < 0) {
return -E_BAD_ENV;
}
if (rcv->env_status == ENV_SUSPENDED) {
return -E_IPC_NOT_RECV;
}
if (rcv->env_status == ENV_LEASED) { // is leased?
djos_send:
for (i = 0; i < NENV; i++) {
if (envs[i].env_type == ENV_TYPE_JDOSC) {
jdos_client = envs[i].env_id;
break;
}
}
// jdos client running?
if (!jdos_client) return -E_BAD_ENV;
if ((r = envid2env(jdos_client, &e, 0)) < 0) return r;
// Mark suspended and try to send ipc
curenv->env_status = ENV_SUSPENDED;
sys_page_alloc(curenv->env_id, (void *) IPCSND,
PTE_U|PTE_P|PTE_W);
*((envid_t *) IPCSND) = envid;
*((uint32_t *)(IPCSND + sizeof(envid_t))) = value;
*((unsigned *)(IPCSND + sizeof(envid_t) +
sizeof(uint32_t))) = perm;
//can't write to page
r = sys_ipc_try_send(jdos_client, CLIENT_SEND_IPC,
(void *) IPCSND, PTE_U|PTE_P);
sys_page_unmap(curenv->env_id, (void *) IPCSND);
// Failed to send ipc, back to running!
if (r < 0) {
cprintf("sys_send_ipc: failed to send ipc %d\n", r);
curenv->env_status = ENV_RUNNABLE;
return r;
}
}
else {
// Is receiver waiting?
if (!rcv->env_ipc_recving) {
return -E_IPC_NOT_RECV;
}
// Try mapping page from sender to receiver (if receiver
// wants it, and sender wants to send it)
// NOTE: Can't use sys_map_page as it checks for env perms
if ((uint32_t) rcv->env_ipc_dstva < UTOP &&
(uint32_t) srcva < UTOP) {
if (!(pp = page_lookup(curenv->env_pgdir,
srcva, &pte)))
return -E_INVAL;
if ((perm & PTE_W) && !(*pte & PTE_W))
return -E_INVAL;
if (page_insert(rcv->env_pgdir, pp,
rcv->env_ipc_dstva, perm) < 0)
return -E_NO_MEM;
}
// Set fields which mark receiver as not waiting
rcv->env_ipc_recving = 0;
rcv->env_ipc_dstva = (void *) UTOP; // invalid dstva
// Set received data fields of receiver
rcv->env_ipc_value = value;
rcv->env_ipc_from = curenv->env_id;
rcv->env_ipc_perm = perm;
// Mark receiver as RUNNABLE
rcv->env_status = ENV_RUNNABLE;
}
return 0;
}
//
// Set up the initial program binary, stack, and processor flags
// for a user process.
// This function is ONLY called during kernel initialization,
// before running the first user-mode environment.
//
// This function loads all loadable segments from the ELF binary image
// into the environment's user memory, starting at the appropriate
// virtual addresses indicated in the ELF program header.
// At the same time it clears to zero any portions of these segments
// that are marked in the program header as being mapped
// but not actually present in the ELF file - i.e., the program's bss section.
//
// All this is very similar to what our boot loader does, except the boot
// loader also needs to read the code from disk. Take a look at
// boot/main.c to get ideas.
//
// Finally, this function maps one page for the program's initial stack.
//
// load_icode panics if it encounters problems.
// - How might load_icode fail? What might be wrong with the given input?
//
//
static void
load_icode(struct Env *e, uint8_t *binary, size_t size)
{
// Hints:
// Load each program segment into virtual memory
// at the address specified in the ELF section header.
// You should only load segments with ph->p_type == ELF_PROG_LOAD.
// Each segment's virtual address can be found in ph->p_va
// and its size in memory can be found in ph->p_memsz.
// The ph->p_filesz bytes from the ELF binary, starting at
// 'binary + ph->p_offset', should be copied to virtual address
// ph->p_va. Any remaining memory bytes should be cleared to zero.
// (The ELF header should have ph->p_filesz <= ph->p_memsz.) /* p_filesz <= ph->memsz ,sunus*/
// Use functions from the previous lab to allocate and map pages.
//
// All page protection bits should be user read/write for now.
// ELF segments are not necessarily page-aligned, but you can
// assume for this function that no two segments will touch
// the same virtual page.
//
// You may find a function like segment_alloc useful.
//
// Loading the segments is much simpler if you can move data
// directly into the virtual addresses stored in the ELF binary.
// So which page directory should be in force during
// this function?
//
// You must also do something with the program's entry point,
// to make sure that the environment starts executing there.
// What? (See env_run() and env_pop_tf() below.)
// LAB 3: Your code here.
// Now map one page for the program's initial stack
// at virtual address USTACKTOP - PGSIZE.
// LAB 3: Your code here.
// You must also do something with the program's entry point, /* TO BE CODED!*/ /* DONE, LINE 328,338 and 339 */
// to make sure that the environment starts executing there. /* DEC 10,2010 */
// What? (See env_run() and env_pop_tf() below.) /* sunus */
// LAB 3: Your code here.
//DEC 09,2010 sunus
struct Proghdr *ph,*eph;
struct Elf *env_elf;
struct Page *pstack;
env_elf = (struct Elf *)binary;
assert(env_elf->e_magic == ELF_MAGIC);
ph = (struct Proghdr *)((uint8_t *)binary + env_elf->e_phoff);
eph = ph + env_elf->e_phnum;
lcr3(e->env_cr3); // we will use env_cr3 for a little while :D
for( ; ph < eph ; ph++)
{
if(ph->p_type == ELF_PROG_LOAD)
{
segment_alloc(e, (void *)ph->p_va,ph->p_memsz);
memmove((void *)ph->p_va, (void *)(binary + ph->p_offset), ph->p_filesz);
memset(((void *)ph->p_va + ph->p_filesz), 0, (ph->p_memsz - ph->p_filesz)); // .bss matters
}
}
lcr3(boot_cr3); // restore boot_cr3
e->env_tf.tf_eip = env_elf->e_entry;
// Now map one page for the program's initial stack
// at virtual address USTACKTOP - PGSIZE.
// LAB 3: Your code here.
//DEC 10,2010,sunus
assert(page_alloc(&pstack) == 0);
assert(page_insert(e->env_pgdir, pstack,(void *)(USTACKTOP - PGSIZE), PTE_U|PTE_W) == 0);
return ;
}
/* do_pgfault - interrupt handler to process the page fault execption
* @mm : the control struct for a set of vma using the same PDT
* @error_code : the error code recorded in trapframe->tf_err which is setted by x86 hardware
* @addr : the addr which causes a memory access exception, (the contents of the CR2 register)
*
* CALL GRAPH: trap--> trap_dispatch-->pgfault_handler-->do_pgfault
* The processor provides ucore's do_pgfault function with two items of information to aid in diagnosing
* the exception and recovering from it.
* (1) The contents of the CR2 register. The processor loads the CR2 register with the
* 32-bit linear address that generated the exception. The do_pgfault fun can
* use this address to locate the corresponding page directory and page-table
* entries.
* (2) An error code on the kernel stack. The error code for a page fault has a format different from
* that for other exceptions. The error code tells the exception handler three things:
* -- The P flag (bit 0) indicates whether the exception was due to a not-present page (0)
* or to either an access rights violation or the use of a reserved bit (1).
* -- The W/R flag (bit 1) indicates whether the memory access that caused the exception
* was a read (0) or write (1).
* -- The U/S flag (bit 2) indicates whether the processor was executing at user mode (1)
* or supervisor mode (0) at the time of the exception.
*/
int
do_pgfault(struct mm_struct *mm, uint32_t error_code, uintptr_t addr) {
int ret = -E_INVAL;
//try to find a vma which include addr
struct vma_struct *vma = find_vma(mm, addr);
pgfault_num++;
//If the addr is in the range of a mm's vma?
if (vma == NULL || vma->vm_start > addr) {
cprintf("not valid addr %x, and can not find it in vma\n", addr);
goto failed;
}
//check the error_code
switch (error_code & 3) {
default:
/* error code flag : default is 3 ( W/R=1, P=1): write, present */
case 2: /* error code flag : (W/R=1, P=0): write, not present */
if (!(vma->vm_flags & VM_WRITE)) {
cprintf("do_pgfault failed: error code flag = write AND not present, but the addr's vma cannot write\n");
goto failed;
}
break;
case 1: /* error code flag : (W/R=0, P=1): read, present */
cprintf("do_pgfault failed: error code flag = read AND present\n");
goto failed;
case 0: /* error code flag : (W/R=0, P=0): read, not present */
if (!(vma->vm_flags & (VM_READ | VM_EXEC))) {
cprintf("do_pgfault failed: error code flag = read AND not present, but the addr's vma cannot read or exec\n");
goto failed;
}
}
/* IF (write an existed addr ) OR
* (write an non_existed addr && addr is writable) OR
* (read an non_existed addr && addr is readable)
* THEN
* continue process
*/
uint32_t perm = PTE_U;
if (vma->vm_flags & VM_WRITE) {
perm |= PTE_W;
}
addr = ROUNDDOWN(addr, PGSIZE);
ret = -E_NO_MEM;
pte_t *ptep=NULL;
/*LAB3 EXERCISE 1: YOUR CODE
* Maybe you want help comment, BELOW comments can help you finish the code
*
* Some Useful MACROs and DEFINEs, you can use them in below implementation.
* MACROs or Functions:
* get_pte : get an pte and return the kernel virtual address of this pte for la
* if the PT contians this pte didn't exist, alloc a page for PT (notice the 3th parameter '1')
* pgdir_alloc_page : call alloc_page & page_insert functions to allocate a page size memory & setup
* an addr map pa<--->la with linear address la and the PDT pgdir
* DEFINES:
* VM_WRITE : If vma->vm_flags & VM_WRITE == 1/0, then the vma is writable/non writable
* PTE_W 0x002 // page table/directory entry flags bit : Writeable
* PTE_U 0x004 // page table/directory entry flags bit : User can access
* VARIABLES:
* mm->pgdir : the PDT of these vma
*
*/
#if 0
/*LAB3 EXERCISE 1: 2013010617*/
ptep = ??? //(1) try to find a pte, if pte's PT(Page Table) isn't existed, then create a PT.
if (*ptep == 0) {
//(2) if the phy addr isn't exist, then alloc a page & map the phy addr with logical addr
}
else {
/*LAB3 EXERCISE 2: 2013010617
* Now we think this pte is a swap entry, we should load data from disk to a page with phy addr,
* and map the phy addr with logical addr, trigger swap manager to record the access situation of this page.
*
* Some Useful MACROs and DEFINEs, you can use them in below implementation.
* MACROs or Functions:
* swap_in(mm, addr, &page) : alloc a memory page, then according to the swap entry in PTE for addr,
* find the addr of disk page, read the content of disk page into this memroy page
* page_insert ： build the map of phy addr of an Page with the linear addr la
* swap_map_swappable ： set the page swappable
*/
if(swap_init_ok) {
struct Page *page=NULL;
//(1）According to the mm AND addr, try to load the content of right disk page
// into the memory which page managed.
//(2) According to the mm, addr AND page, setup the map of phy addr <---> logical addr
//(3) make the page swappable.
}
else {
cprintf("no swap_init_ok but ptep is %x, failed\n",*ptep);
goto failed;
}
}
#endif
// try to find a pte, if pte's PT(Page Table) isn't existed, then create a PT.
// (notice the 3th parameter '1')
if ((ptep = get_pte(mm->pgdir, addr, 1)) == NULL) {
cprintf("get_pte in do_pgfault failed\n");
goto failed;
}
if (*ptep == 0) { // if the phy addr isn't exist, then alloc a page & map the phy addr with logical addr
if (pgdir_alloc_page(mm->pgdir, addr, perm) == NULL) {
cprintf("pgdir_alloc_page in do_pgfault failed\n");
goto failed;
}
}
else { // if this pte is a swap entry, then load data from disk to a page with phy addr
// and call page_insert to map the phy addr with logical addr
if(swap_init_ok) {
struct Page *page=NULL;
if ((ret = swap_in(mm, addr, &page)) != 0) {
cprintf("swap_in in do_pgfault failed\n");
goto failed;
}
page_insert(mm->pgdir, page, addr, perm);
swap_map_swappable(mm, addr, page, 1);
page->pra_vaddr = addr;
}
else {
cprintf("no swap_init_ok but ptep is %x, failed\n",*ptep);
goto failed;
}
}
ret = 0;
failed:
return ret;
}
// Map the page of memory at 'srcva' in srcenvid's address space
// at 'dstva' in dstenvid's address space with permission 'perm'.
// Perm has the same restrictions as in sys_page_alloc, except
// that it also must not grant write access to a read-only
// page.
//
// Return 0 on success, < 0 on error. Errors are:
// -E_BAD_ENV if srcenvid and/or dstenvid doesn't currently exist,
// or the caller doesn't have permission to change one of them.
// -E_INVAL if srcva >= UTOP or srcva is not page-aligned,
// or dstva >= UTOP or dstva is not page-aligned.
// -E_INVAL is srcva is not mapped in srcenvid's address space.
// -E_INVAL if perm is inappropriate (see sys_page_alloc).
// -E_INVAL if (perm & PTE_W), but srcva is read-only in srcenvid's
// address space.
// -E_NO_MEM if there's no memory to allocate any necessary page tables.
static int
sys_page_map(envid_t srcenvid, void *srcva,
envid_t dstenvid, void *dstva, int perm)
{
// Hint: This function is a wrapper around page_lookup() and
// page_insert() from kern/pmap.c.
// Again, most of the new code you write should be to check the
// parameters for correctness.
// Use the third argument to page_lookup() to
// check the current permissions on the page.
// LAB 4: Your code here.
cprintf("DEBUG sys_page_map() called\n");
struct Env *se, *de;
int err = envid2env(srcenvid, &se, 1);
if (err) {
cprintf("9 err %d", err);
return err; //bad environment
}
err = envid2env(dstenvid, &de, 1);
if (err) {
cprintf("10 err %d", err);
return err; //bad environment
}
// -E_INVAL if srcva >= UTOP or srcva is not page-aligned,
// or dstva >= UTOP or dstva is not page-aligned.
if (srcva>=(void*)UTOP || dstva>=(void*)UTOP || ROUNDDOWN(srcva,PGSIZE)!=srcva || ROUNDDOWN(dstva,PGSIZE)!=dstva) {
cprintf("11 err");
return -E_INVAL;
}
// -E_INVAL is srcva is not mapped in srcenvid's address space.
pte_t *pte;
struct PageInfo *pag = page_lookup(se->env_pgdir, srcva, &pte);
if (!pag) {
cprintf("12 err");
return -E_INVAL;
}
// -E_INVAL if perm is inappropriate (see sys_page_alloc).
int flag = PTE_U | PTE_P;
if ((perm & flag) != flag) {
cprintf("13 err");
return -E_INVAL;
}
// -E_INVAL if (perm & PTE_W), but srcva is read-only in srcenvid's
// address space.
if (((*pte&PTE_W) == 0) && (perm&PTE_W)) {
cprintf("14 err");
return -E_INVAL;
}
// -E_NO_MEM if there's no memory to allocate any necessary page tables.
err = page_insert(de->env_pgdir, pag, dstva, perm);
cprintf("15 err %d", err);
return err;
//panic("sys_page_map not implemented");
}