2007-07-26 17:41:02 +00:00
|
|
|
/*P:700 The pagetable code, on the other hand, still shows the scars of
|
|
|
|
* previous encounters. It's functional, and as neat as it can be in the
|
|
|
|
* circumstances, but be wary, for these things are subtle and break easily.
|
|
|
|
* The Guest provides a virtual to physical mapping, but we can neither trust
|
|
|
|
* it nor use it: we verify and convert it here to point the hardware to the
|
|
|
|
* actual Guest pages when running the Guest. :*/
|
|
|
|
|
|
|
|
/* Copyright (C) Rusty Russell IBM Corporation 2006.
|
2007-07-19 08:49:23 +00:00
|
|
|
* GPL v2 and any later version */
|
|
|
|
#include <linux/mm.h>
|
|
|
|
#include <linux/types.h>
|
|
|
|
#include <linux/spinlock.h>
|
|
|
|
#include <linux/random.h>
|
|
|
|
#include <linux/percpu.h>
|
|
|
|
#include <asm/tlbflush.h>
|
2007-10-22 01:03:36 +00:00
|
|
|
#include <asm/uaccess.h>
|
2007-07-19 08:49:23 +00:00
|
|
|
#include "lg.h"
|
|
|
|
|
2007-07-26 17:41:05 +00:00
|
|
|
/*M:008 We hold reference to pages, which prevents them from being swapped.
|
|
|
|
* It'd be nice to have a callback in the "struct mm_struct" when Linux wants
|
|
|
|
* to swap out. If we had this, and a shrinker callback to trim PTE pages, we
|
|
|
|
* could probably consider launching Guests as non-root. :*/
|
|
|
|
|
2007-07-26 17:41:04 +00:00
|
|
|
/*H:300
|
|
|
|
* The Page Table Code
|
|
|
|
*
|
|
|
|
* We use two-level page tables for the Guest. If you're not entirely
|
|
|
|
* comfortable with virtual addresses, physical addresses and page tables then
|
2007-10-25 05:02:50 +00:00
|
|
|
* I recommend you review arch/x86/lguest/boot.c's "Page Table Handling" (with
|
|
|
|
* diagrams!).
|
2007-07-26 17:41:04 +00:00
|
|
|
*
|
|
|
|
* The Guest keeps page tables, but we maintain the actual ones here: these are
|
|
|
|
* called "shadow" page tables. Which is a very Guest-centric name: these are
|
|
|
|
* the real page tables the CPU uses, although we keep them up to date to
|
|
|
|
* reflect the Guest's. (See what I mean about weird naming? Since when do
|
|
|
|
* shadows reflect anything?)
|
|
|
|
*
|
|
|
|
* Anyway, this is the most complicated part of the Host code. There are seven
|
|
|
|
* parts to this:
|
2007-10-25 05:02:50 +00:00
|
|
|
* (i) Looking up a page table entry when the Guest faults,
|
|
|
|
* (ii) Making sure the Guest stack is mapped,
|
|
|
|
* (iii) Setting up a page table entry when the Guest tells us one has changed,
|
2007-07-26 17:41:04 +00:00
|
|
|
* (iv) Switching page tables,
|
2007-10-25 05:02:50 +00:00
|
|
|
* (v) Flushing (throwing away) page tables,
|
2007-07-26 17:41:04 +00:00
|
|
|
* (vi) Mapping the Switcher when the Guest is about to run,
|
|
|
|
* (vii) Setting up the page tables initially.
|
|
|
|
:*/
|
|
|
|
|
|
|
|
|
|
|
|
/* 1024 entries in a page table page maps 1024 pages: 4MB. The Switcher is
|
|
|
|
* conveniently placed at the top 4MB, so it uses a separate, complete PTE
|
|
|
|
* page. */
|
2007-10-22 01:03:33 +00:00
|
|
|
#define SWITCHER_PGD_INDEX (PTRS_PER_PGD - 1)
|
2007-07-19 08:49:23 +00:00
|
|
|
|
2007-07-26 17:41:04 +00:00
|
|
|
/* We actually need a separate PTE page for each CPU. Remember that after the
|
|
|
|
* Switcher code itself comes two pages for each CPU, and we don't want this
|
|
|
|
* CPU's guest to see the pages of any other CPU. */
|
2007-10-22 01:03:33 +00:00
|
|
|
static DEFINE_PER_CPU(pte_t *, switcher_pte_pages);
|
2007-07-19 08:49:23 +00:00
|
|
|
#define switcher_pte_page(cpu) per_cpu(switcher_pte_pages, cpu)
|
|
|
|
|
2007-10-25 05:02:50 +00:00
|
|
|
/*H:320 The page table code is curly enough to need helper functions to keep it
|
|
|
|
* clear and clean.
|
2007-07-26 17:41:04 +00:00
|
|
|
*
|
2007-10-22 01:03:33 +00:00
|
|
|
* There are two functions which return pointers to the shadow (aka "real")
|
2007-07-26 17:41:04 +00:00
|
|
|
* page tables.
|
|
|
|
*
|
|
|
|
* spgd_addr() takes the virtual address and returns a pointer to the top-level
|
2007-10-25 05:02:50 +00:00
|
|
|
* page directory entry (PGD) for that address. Since we keep track of several
|
|
|
|
* page tables, the "i" argument tells us which one we're interested in (it's
|
2007-07-26 17:41:04 +00:00
|
|
|
* usually the current one). */
|
2008-01-17 21:19:42 +00:00
|
|
|
static pgd_t *spgd_addr(struct lg_cpu *cpu, u32 i, unsigned long vaddr)
|
2007-07-19 08:49:23 +00:00
|
|
|
{
|
2007-10-22 01:03:33 +00:00
|
|
|
unsigned int index = pgd_index(vaddr);
|
2007-07-19 08:49:23 +00:00
|
|
|
|
2007-07-26 17:41:04 +00:00
|
|
|
/* We kill any Guest trying to touch the Switcher addresses. */
|
2007-07-19 08:49:23 +00:00
|
|
|
if (index >= SWITCHER_PGD_INDEX) {
|
2008-01-17 21:19:42 +00:00
|
|
|
kill_guest(cpu, "attempt to access switcher pages");
|
2007-07-19 08:49:23 +00:00
|
|
|
index = 0;
|
|
|
|
}
|
2007-07-26 17:41:04 +00:00
|
|
|
/* Return a pointer index'th pgd entry for the i'th page table. */
|
2008-01-17 21:19:42 +00:00
|
|
|
return &cpu->lg->pgdirs[i].pgdir[index];
|
2007-07-19 08:49:23 +00:00
|
|
|
}
|
|
|
|
|
2007-10-25 05:02:50 +00:00
|
|
|
/* This routine then takes the page directory entry returned above, which
|
|
|
|
* contains the address of the page table entry (PTE) page. It then returns a
|
|
|
|
* pointer to the PTE entry for the given address. */
|
2008-01-17 21:09:49 +00:00
|
|
|
static pte_t *spte_addr(pgd_t spgd, unsigned long vaddr)
|
2007-07-19 08:49:23 +00:00
|
|
|
{
|
2007-10-22 01:03:33 +00:00
|
|
|
pte_t *page = __va(pgd_pfn(spgd) << PAGE_SHIFT);
|
2007-07-26 17:41:04 +00:00
|
|
|
/* You should never call this if the PGD entry wasn't valid */
|
2007-10-22 01:03:33 +00:00
|
|
|
BUG_ON(!(pgd_flags(spgd) & _PAGE_PRESENT));
|
|
|
|
return &page[(vaddr >> PAGE_SHIFT) % PTRS_PER_PTE];
|
2007-07-19 08:49:23 +00:00
|
|
|
}
|
|
|
|
|
2007-07-26 17:41:04 +00:00
|
|
|
/* These two functions just like the above two, except they access the Guest
|
|
|
|
* page tables. Hence they return a Guest address. */
|
2008-01-07 13:05:37 +00:00
|
|
|
static unsigned long gpgd_addr(struct lg_cpu *cpu, unsigned long vaddr)
|
2007-07-19 08:49:23 +00:00
|
|
|
{
|
2007-10-22 01:03:33 +00:00
|
|
|
unsigned int index = vaddr >> (PGDIR_SHIFT);
|
2008-01-07 13:05:37 +00:00
|
|
|
return cpu->lg->pgdirs[cpu->cpu_pgd].gpgdir + index * sizeof(pgd_t);
|
2007-07-19 08:49:23 +00:00
|
|
|
}
|
|
|
|
|
2008-01-17 21:18:08 +00:00
|
|
|
static unsigned long gpte_addr(pgd_t gpgd, unsigned long vaddr)
|
2007-07-19 08:49:23 +00:00
|
|
|
{
|
2007-10-22 01:03:33 +00:00
|
|
|
unsigned long gpage = pgd_pfn(gpgd) << PAGE_SHIFT;
|
|
|
|
BUG_ON(!(pgd_flags(gpgd) & _PAGE_PRESENT));
|
|
|
|
return gpage + ((vaddr>>PAGE_SHIFT) % PTRS_PER_PTE) * sizeof(pte_t);
|
2007-07-19 08:49:23 +00:00
|
|
|
}
|
|
|
|
|
2007-07-26 17:41:04 +00:00
|
|
|
/*H:350 This routine takes a page number given by the Guest and converts it to
|
|
|
|
* an actual, physical page number. It can fail for several reasons: the
|
|
|
|
* virtual address might not be mapped by the Launcher, the write flag is set
|
|
|
|
* and the page is read-only, or the write flag was set and the page was
|
|
|
|
* shared so had to be copied, but we ran out of memory.
|
|
|
|
*
|
|
|
|
* This holds a reference to the page, so release_pte() is careful to
|
|
|
|
* put that back. */
|
2007-07-19 08:49:23 +00:00
|
|
|
static unsigned long get_pfn(unsigned long virtpfn, int write)
|
|
|
|
{
|
|
|
|
struct page *page;
|
2007-07-26 17:41:04 +00:00
|
|
|
/* This value indicates failure. */
|
2007-07-19 08:49:23 +00:00
|
|
|
unsigned long ret = -1UL;
|
|
|
|
|
2007-07-26 17:41:04 +00:00
|
|
|
/* get_user_pages() is a complex interface: it gets the "struct
|
|
|
|
* vm_area_struct" and "struct page" assocated with a range of pages.
|
|
|
|
* It also needs the task's mmap_sem held, and is not very quick.
|
|
|
|
* It returns the number of pages it got. */
|
2007-07-19 08:49:23 +00:00
|
|
|
down_read(¤t->mm->mmap_sem);
|
|
|
|
if (get_user_pages(current, current->mm, virtpfn << PAGE_SHIFT,
|
|
|
|
1, write, 1, &page, NULL) == 1)
|
|
|
|
ret = page_to_pfn(page);
|
|
|
|
up_read(¤t->mm->mmap_sem);
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
2007-07-26 17:41:04 +00:00
|
|
|
/*H:340 Converting a Guest page table entry to a shadow (ie. real) page table
|
|
|
|
* entry can be a little tricky. The flags are (almost) the same, but the
|
|
|
|
* Guest PTE contains a virtual page number: the CPU needs the real page
|
|
|
|
* number. */
|
2008-01-17 21:19:42 +00:00
|
|
|
static pte_t gpte_to_spte(struct lg_cpu *cpu, pte_t gpte, int write)
|
2007-07-19 08:49:23 +00:00
|
|
|
{
|
2007-10-22 01:03:33 +00:00
|
|
|
unsigned long pfn, base, flags;
|
2007-07-19 08:49:23 +00:00
|
|
|
|
2007-07-26 17:41:04 +00:00
|
|
|
/* The Guest sets the global flag, because it thinks that it is using
|
|
|
|
* PGE. We only told it to use PGE so it would tell us whether it was
|
|
|
|
* flushing a kernel mapping or a userspace mapping. We don't actually
|
|
|
|
* use the global bit, so throw it away. */
|
2007-10-22 01:03:33 +00:00
|
|
|
flags = (pte_flags(gpte) & ~_PAGE_GLOBAL);
|
2007-07-26 17:41:04 +00:00
|
|
|
|
2007-10-22 01:03:26 +00:00
|
|
|
/* The Guest's pages are offset inside the Launcher. */
|
2008-01-17 21:19:42 +00:00
|
|
|
base = (unsigned long)cpu->lg->mem_base / PAGE_SIZE;
|
2007-10-22 01:03:26 +00:00
|
|
|
|
2007-07-26 17:41:04 +00:00
|
|
|
/* We need a temporary "unsigned long" variable to hold the answer from
|
|
|
|
* get_pfn(), because it returns 0xFFFFFFFF on failure, which wouldn't
|
|
|
|
* fit in spte.pfn. get_pfn() finds the real physical number of the
|
|
|
|
* page, given the virtual number. */
|
2007-10-22 01:03:33 +00:00
|
|
|
pfn = get_pfn(base + pte_pfn(gpte), write);
|
2007-07-19 08:49:23 +00:00
|
|
|
if (pfn == -1UL) {
|
2008-01-17 21:19:42 +00:00
|
|
|
kill_guest(cpu, "failed to get page %lu", pte_pfn(gpte));
|
2007-07-26 17:41:04 +00:00
|
|
|
/* When we destroy the Guest, we'll go through the shadow page
|
|
|
|
* tables and release_pte() them. Make sure we don't think
|
|
|
|
* this one is valid! */
|
2007-10-22 01:03:33 +00:00
|
|
|
flags = 0;
|
2007-07-19 08:49:23 +00:00
|
|
|
}
|
2007-10-22 01:03:33 +00:00
|
|
|
/* Now we assemble our shadow PTE from the page number and flags. */
|
|
|
|
return pfn_pte(pfn, __pgprot(flags));
|
2007-07-19 08:49:23 +00:00
|
|
|
}
|
|
|
|
|
2007-07-26 17:41:04 +00:00
|
|
|
/*H:460 And to complete the chain, release_pte() looks like this: */
|
2007-10-22 01:03:33 +00:00
|
|
|
static void release_pte(pte_t pte)
|
2007-07-19 08:49:23 +00:00
|
|
|
{
|
2007-07-26 17:41:04 +00:00
|
|
|
/* Remember that get_user_pages() took a reference to the page, in
|
|
|
|
* get_pfn()? We have to put it back now. */
|
2007-10-22 01:03:33 +00:00
|
|
|
if (pte_flags(pte) & _PAGE_PRESENT)
|
|
|
|
put_page(pfn_to_page(pte_pfn(pte)));
|
2007-07-19 08:49:23 +00:00
|
|
|
}
|
2007-07-26 17:41:04 +00:00
|
|
|
/*:*/
|
2007-07-19 08:49:23 +00:00
|
|
|
|
2008-01-17 21:19:42 +00:00
|
|
|
static void check_gpte(struct lg_cpu *cpu, pte_t gpte)
|
2007-07-19 08:49:23 +00:00
|
|
|
{
|
2007-10-22 01:03:33 +00:00
|
|
|
if ((pte_flags(gpte) & (_PAGE_PWT|_PAGE_PSE))
|
2008-01-17 21:19:42 +00:00
|
|
|
|| pte_pfn(gpte) >= cpu->lg->pfn_limit)
|
|
|
|
kill_guest(cpu, "bad page table entry");
|
2007-07-19 08:49:23 +00:00
|
|
|
}
|
|
|
|
|
2008-01-17 21:19:42 +00:00
|
|
|
static void check_gpgd(struct lg_cpu *cpu, pgd_t gpgd)
|
2007-07-19 08:49:23 +00:00
|
|
|
{
|
2008-01-17 21:19:42 +00:00
|
|
|
if ((pgd_flags(gpgd) & ~_PAGE_TABLE) ||
|
|
|
|
(pgd_pfn(gpgd) >= cpu->lg->pfn_limit))
|
|
|
|
kill_guest(cpu, "bad page directory entry");
|
2007-07-19 08:49:23 +00:00
|
|
|
}
|
|
|
|
|
2007-07-26 17:41:04 +00:00
|
|
|
/*H:330
|
2007-10-25 05:02:50 +00:00
|
|
|
* (i) Looking up a page table entry when the Guest faults.
|
2007-07-26 17:41:04 +00:00
|
|
|
*
|
|
|
|
* We saw this call in run_guest(): when we see a page fault in the Guest, we
|
|
|
|
* come here. That's because we only set up the shadow page tables lazily as
|
|
|
|
* they're needed, so we get page faults all the time and quietly fix them up
|
|
|
|
* and return to the Guest without it knowing.
|
|
|
|
*
|
|
|
|
* If we fixed up the fault (ie. we mapped the address), this routine returns
|
2007-10-25 05:02:50 +00:00
|
|
|
* true. Otherwise, it was a real fault and we need to tell the Guest. */
|
2008-01-07 13:05:37 +00:00
|
|
|
int demand_page(struct lg_cpu *cpu, unsigned long vaddr, int errcode)
|
2007-07-19 08:49:23 +00:00
|
|
|
{
|
2007-10-22 01:03:33 +00:00
|
|
|
pgd_t gpgd;
|
|
|
|
pgd_t *spgd;
|
2007-07-19 08:49:23 +00:00
|
|
|
unsigned long gpte_ptr;
|
2007-10-22 01:03:33 +00:00
|
|
|
pte_t gpte;
|
|
|
|
pte_t *spte;
|
2007-07-19 08:49:23 +00:00
|
|
|
|
2007-07-26 17:41:04 +00:00
|
|
|
/* First step: get the top-level Guest page table entry. */
|
2008-01-17 21:19:42 +00:00
|
|
|
gpgd = lgread(cpu, gpgd_addr(cpu, vaddr), pgd_t);
|
2007-07-26 17:41:04 +00:00
|
|
|
/* Toplevel not present? We can't map it in. */
|
2007-10-22 01:03:33 +00:00
|
|
|
if (!(pgd_flags(gpgd) & _PAGE_PRESENT))
|
2007-07-19 08:49:23 +00:00
|
|
|
return 0;
|
|
|
|
|
2007-07-26 17:41:04 +00:00
|
|
|
/* Now look at the matching shadow entry. */
|
2008-01-17 21:19:42 +00:00
|
|
|
spgd = spgd_addr(cpu, cpu->cpu_pgd, vaddr);
|
2007-10-22 01:03:33 +00:00
|
|
|
if (!(pgd_flags(*spgd) & _PAGE_PRESENT)) {
|
2007-07-26 17:41:04 +00:00
|
|
|
/* No shadow entry: allocate a new shadow PTE page. */
|
2007-07-19 08:49:23 +00:00
|
|
|
unsigned long ptepage = get_zeroed_page(GFP_KERNEL);
|
2007-07-26 17:41:04 +00:00
|
|
|
/* This is not really the Guest's fault, but killing it is
|
|
|
|
* simple for this corner case. */
|
2007-07-19 08:49:23 +00:00
|
|
|
if (!ptepage) {
|
2008-01-17 21:19:42 +00:00
|
|
|
kill_guest(cpu, "out of memory allocating pte page");
|
2007-07-19 08:49:23 +00:00
|
|
|
return 0;
|
|
|
|
}
|
2007-07-26 17:41:04 +00:00
|
|
|
/* We check that the Guest pgd is OK. */
|
2008-01-17 21:19:42 +00:00
|
|
|
check_gpgd(cpu, gpgd);
|
2007-07-26 17:41:04 +00:00
|
|
|
/* And we copy the flags to the shadow PGD entry. The page
|
|
|
|
* number in the shadow PGD is the page we just allocated. */
|
2007-10-22 01:03:33 +00:00
|
|
|
*spgd = __pgd(__pa(ptepage) | pgd_flags(gpgd));
|
2007-07-19 08:49:23 +00:00
|
|
|
}
|
|
|
|
|
2007-07-26 17:41:04 +00:00
|
|
|
/* OK, now we look at the lower level in the Guest page table: keep its
|
|
|
|
* address, because we might update it later. */
|
2008-01-17 21:18:08 +00:00
|
|
|
gpte_ptr = gpte_addr(gpgd, vaddr);
|
2008-01-17 21:19:42 +00:00
|
|
|
gpte = lgread(cpu, gpte_ptr, pte_t);
|
2007-07-19 08:49:23 +00:00
|
|
|
|
2007-07-26 17:41:04 +00:00
|
|
|
/* If this page isn't in the Guest page tables, we can't page it in. */
|
2007-10-22 01:03:33 +00:00
|
|
|
if (!(pte_flags(gpte) & _PAGE_PRESENT))
|
2007-07-19 08:49:23 +00:00
|
|
|
return 0;
|
|
|
|
|
2007-07-26 17:41:04 +00:00
|
|
|
/* Check they're not trying to write to a page the Guest wants
|
|
|
|
* read-only (bit 2 of errcode == write). */
|
2007-10-22 01:03:33 +00:00
|
|
|
if ((errcode & 2) && !(pte_flags(gpte) & _PAGE_RW))
|
2007-07-19 08:49:23 +00:00
|
|
|
return 0;
|
|
|
|
|
2007-10-25 05:02:50 +00:00
|
|
|
/* User access to a kernel-only page? (bit 3 == user access) */
|
2007-10-22 01:03:33 +00:00
|
|
|
if ((errcode & 4) && !(pte_flags(gpte) & _PAGE_USER))
|
2007-07-19 08:49:23 +00:00
|
|
|
return 0;
|
|
|
|
|
2007-07-26 17:41:04 +00:00
|
|
|
/* Check that the Guest PTE flags are OK, and the page number is below
|
|
|
|
* the pfn_limit (ie. not mapping the Launcher binary). */
|
2008-01-17 21:19:42 +00:00
|
|
|
check_gpte(cpu, gpte);
|
2007-10-25 05:02:50 +00:00
|
|
|
|
2007-07-26 17:41:04 +00:00
|
|
|
/* Add the _PAGE_ACCESSED and (for a write) _PAGE_DIRTY flag */
|
2007-10-22 01:03:33 +00:00
|
|
|
gpte = pte_mkyoung(gpte);
|
2007-07-19 08:49:23 +00:00
|
|
|
if (errcode & 2)
|
2007-10-22 01:03:33 +00:00
|
|
|
gpte = pte_mkdirty(gpte);
|
2007-07-19 08:49:23 +00:00
|
|
|
|
2007-07-26 17:41:04 +00:00
|
|
|
/* Get the pointer to the shadow PTE entry we're going to set. */
|
2008-01-17 21:09:49 +00:00
|
|
|
spte = spte_addr(*spgd, vaddr);
|
2007-07-26 17:41:04 +00:00
|
|
|
/* If there was a valid shadow PTE entry here before, we release it.
|
|
|
|
* This can happen with a write to a previously read-only entry. */
|
2007-07-19 08:49:23 +00:00
|
|
|
release_pte(*spte);
|
|
|
|
|
2007-07-26 17:41:04 +00:00
|
|
|
/* If this is a write, we insist that the Guest page is writable (the
|
|
|
|
* final arg to gpte_to_spte()). */
|
2007-10-22 01:03:33 +00:00
|
|
|
if (pte_dirty(gpte))
|
2008-01-17 21:19:42 +00:00
|
|
|
*spte = gpte_to_spte(cpu, gpte, 1);
|
2007-10-22 01:03:33 +00:00
|
|
|
else
|
2007-07-26 17:41:04 +00:00
|
|
|
/* If this is a read, don't set the "writable" bit in the page
|
|
|
|
* table entry, even if the Guest says it's writable. That way
|
2007-10-25 05:02:50 +00:00
|
|
|
* we will come back here when a write does actually occur, so
|
|
|
|
* we can update the Guest's _PAGE_DIRTY flag. */
|
2008-01-17 21:19:42 +00:00
|
|
|
*spte = gpte_to_spte(cpu, pte_wrprotect(gpte), 0);
|
2007-07-19 08:49:23 +00:00
|
|
|
|
2007-07-26 17:41:04 +00:00
|
|
|
/* Finally, we write the Guest PTE entry back: we've set the
|
|
|
|
* _PAGE_ACCESSED and maybe the _PAGE_DIRTY flags. */
|
2008-01-17 21:19:42 +00:00
|
|
|
lgwrite(cpu, gpte_ptr, pte_t, gpte);
|
2007-07-26 17:41:04 +00:00
|
|
|
|
2007-10-25 05:02:50 +00:00
|
|
|
/* The fault is fixed, the page table is populated, the mapping
|
|
|
|
* manipulated, the result returned and the code complete. A small
|
|
|
|
* delay and a trace of alliteration are the only indications the Guest
|
|
|
|
* has that a page fault occurred at all. */
|
2007-07-19 08:49:23 +00:00
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
|
2007-10-25 05:02:50 +00:00
|
|
|
/*H:360
|
|
|
|
* (ii) Making sure the Guest stack is mapped.
|
2007-07-26 17:41:04 +00:00
|
|
|
*
|
2007-10-25 05:02:50 +00:00
|
|
|
* Remember that direct traps into the Guest need a mapped Guest kernel stack.
|
|
|
|
* pin_stack_pages() calls us here: we could simply call demand_page(), but as
|
|
|
|
* we've seen that logic is quite long, and usually the stack pages are already
|
|
|
|
* mapped, so it's overkill.
|
2007-07-26 17:41:04 +00:00
|
|
|
*
|
|
|
|
* This is a quick version which answers the question: is this virtual address
|
|
|
|
* mapped by the shadow page tables, and is it writable? */
|
2008-01-07 13:05:37 +00:00
|
|
|
static int page_writable(struct lg_cpu *cpu, unsigned long vaddr)
|
2007-07-19 08:49:23 +00:00
|
|
|
{
|
2007-10-22 01:03:33 +00:00
|
|
|
pgd_t *spgd;
|
2007-07-19 08:49:23 +00:00
|
|
|
unsigned long flags;
|
|
|
|
|
2007-10-25 05:02:50 +00:00
|
|
|
/* Look at the current top level entry: is it present? */
|
2008-01-17 21:19:42 +00:00
|
|
|
spgd = spgd_addr(cpu, cpu->cpu_pgd, vaddr);
|
2007-10-22 01:03:33 +00:00
|
|
|
if (!(pgd_flags(*spgd) & _PAGE_PRESENT))
|
2007-07-19 08:49:23 +00:00
|
|
|
return 0;
|
|
|
|
|
2007-07-26 17:41:04 +00:00
|
|
|
/* Check the flags on the pte entry itself: it must be present and
|
|
|
|
* writable. */
|
2008-01-17 21:09:49 +00:00
|
|
|
flags = pte_flags(*(spte_addr(*spgd, vaddr)));
|
2007-10-22 01:03:33 +00:00
|
|
|
|
2007-07-19 08:49:23 +00:00
|
|
|
return (flags & (_PAGE_PRESENT|_PAGE_RW)) == (_PAGE_PRESENT|_PAGE_RW);
|
|
|
|
}
|
|
|
|
|
2007-07-26 17:41:04 +00:00
|
|
|
/* So, when pin_stack_pages() asks us to pin a page, we check if it's already
|
|
|
|
* in the page tables, and if not, we call demand_page() with error code 2
|
|
|
|
* (meaning "write"). */
|
2008-01-07 13:05:37 +00:00
|
|
|
void pin_page(struct lg_cpu *cpu, unsigned long vaddr)
|
2007-07-19 08:49:23 +00:00
|
|
|
{
|
2008-01-07 13:05:37 +00:00
|
|
|
if (!page_writable(cpu, vaddr) && !demand_page(cpu, vaddr, 2))
|
2008-01-17 21:19:42 +00:00
|
|
|
kill_guest(cpu, "bad stack page %#lx", vaddr);
|
2007-07-19 08:49:23 +00:00
|
|
|
}
|
|
|
|
|
2007-07-26 17:41:04 +00:00
|
|
|
/*H:450 If we chase down the release_pgd() code, it looks like this: */
|
2007-10-22 01:03:33 +00:00
|
|
|
static void release_pgd(struct lguest *lg, pgd_t *spgd)
|
2007-07-19 08:49:23 +00:00
|
|
|
{
|
2007-07-26 17:41:04 +00:00
|
|
|
/* If the entry's not present, there's nothing to release. */
|
2007-10-22 01:03:33 +00:00
|
|
|
if (pgd_flags(*spgd) & _PAGE_PRESENT) {
|
2007-07-19 08:49:23 +00:00
|
|
|
unsigned int i;
|
2007-07-26 17:41:04 +00:00
|
|
|
/* Converting the pfn to find the actual PTE page is easy: turn
|
|
|
|
* the page number into a physical address, then convert to a
|
|
|
|
* virtual address (easy for kernel pages like this one). */
|
2007-10-22 01:03:33 +00:00
|
|
|
pte_t *ptepage = __va(pgd_pfn(*spgd) << PAGE_SHIFT);
|
2007-07-26 17:41:04 +00:00
|
|
|
/* For each entry in the page, we might need to release it. */
|
2007-10-22 01:03:33 +00:00
|
|
|
for (i = 0; i < PTRS_PER_PTE; i++)
|
2007-07-19 08:49:23 +00:00
|
|
|
release_pte(ptepage[i]);
|
2007-07-26 17:41:04 +00:00
|
|
|
/* Now we can free the page of PTEs */
|
2007-07-19 08:49:23 +00:00
|
|
|
free_page((long)ptepage);
|
2007-10-25 05:02:50 +00:00
|
|
|
/* And zero out the PGD entry so we never release it twice. */
|
2007-10-22 01:03:33 +00:00
|
|
|
*spgd = __pgd(0);
|
2007-07-19 08:49:23 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2007-10-25 05:02:50 +00:00
|
|
|
/*H:445 We saw flush_user_mappings() twice: once from the flush_user_mappings()
|
|
|
|
* hypercall and once in new_pgdir() when we re-used a top-level pgdir page.
|
|
|
|
* It simply releases every PTE page from 0 up to the Guest's kernel address. */
|
2007-07-19 08:49:23 +00:00
|
|
|
static void flush_user_mappings(struct lguest *lg, int idx)
|
|
|
|
{
|
|
|
|
unsigned int i;
|
2007-07-26 17:41:04 +00:00
|
|
|
/* Release every pgd entry up to the kernel's address. */
|
2007-10-22 01:03:36 +00:00
|
|
|
for (i = 0; i < pgd_index(lg->kernel_address); i++)
|
2007-07-19 08:49:23 +00:00
|
|
|
release_pgd(lg, lg->pgdirs[idx].pgdir + i);
|
|
|
|
}
|
|
|
|
|
2007-10-25 05:02:50 +00:00
|
|
|
/*H:440 (v) Flushing (throwing away) page tables,
|
|
|
|
*
|
|
|
|
* The Guest has a hypercall to throw away the page tables: it's used when a
|
|
|
|
* large number of mappings have been changed. */
|
2008-01-07 13:05:37 +00:00
|
|
|
void guest_pagetable_flush_user(struct lg_cpu *cpu)
|
2007-07-19 08:49:23 +00:00
|
|
|
{
|
2007-07-26 17:41:04 +00:00
|
|
|
/* Drop the userspace part of the current page table. */
|
2008-01-07 13:05:37 +00:00
|
|
|
flush_user_mappings(cpu->lg, cpu->cpu_pgd);
|
2007-07-19 08:49:23 +00:00
|
|
|
}
|
2007-07-26 17:41:04 +00:00
|
|
|
/*:*/
|
2007-07-19 08:49:23 +00:00
|
|
|
|
2007-10-22 01:03:36 +00:00
|
|
|
/* We walk down the guest page tables to get a guest-physical address */
|
2008-01-07 13:05:37 +00:00
|
|
|
unsigned long guest_pa(struct lg_cpu *cpu, unsigned long vaddr)
|
2007-10-22 01:03:36 +00:00
|
|
|
{
|
|
|
|
pgd_t gpgd;
|
|
|
|
pte_t gpte;
|
|
|
|
|
|
|
|
/* First step: get the top-level Guest page table entry. */
|
2008-01-17 21:19:42 +00:00
|
|
|
gpgd = lgread(cpu, gpgd_addr(cpu, vaddr), pgd_t);
|
2007-10-22 01:03:36 +00:00
|
|
|
/* Toplevel not present? We can't map it in. */
|
|
|
|
if (!(pgd_flags(gpgd) & _PAGE_PRESENT))
|
2008-01-17 21:19:42 +00:00
|
|
|
kill_guest(cpu, "Bad address %#lx", vaddr);
|
2007-10-22 01:03:36 +00:00
|
|
|
|
2008-01-17 21:19:42 +00:00
|
|
|
gpte = lgread(cpu, gpte_addr(gpgd, vaddr), pte_t);
|
2007-10-22 01:03:36 +00:00
|
|
|
if (!(pte_flags(gpte) & _PAGE_PRESENT))
|
2008-01-17 21:19:42 +00:00
|
|
|
kill_guest(cpu, "Bad address %#lx", vaddr);
|
2007-10-22 01:03:36 +00:00
|
|
|
|
|
|
|
return pte_pfn(gpte) * PAGE_SIZE | (vaddr & ~PAGE_MASK);
|
|
|
|
}
|
|
|
|
|
2007-07-26 17:41:04 +00:00
|
|
|
/* We keep several page tables. This is a simple routine to find the page
|
|
|
|
* table (if any) corresponding to this top-level address the Guest has given
|
|
|
|
* us. */
|
2007-07-19 08:49:23 +00:00
|
|
|
static unsigned int find_pgdir(struct lguest *lg, unsigned long pgtable)
|
|
|
|
{
|
|
|
|
unsigned int i;
|
|
|
|
for (i = 0; i < ARRAY_SIZE(lg->pgdirs); i++)
|
2007-10-22 01:03:34 +00:00
|
|
|
if (lg->pgdirs[i].gpgdir == pgtable)
|
2007-07-19 08:49:23 +00:00
|
|
|
break;
|
|
|
|
return i;
|
|
|
|
}
|
|
|
|
|
2007-07-26 17:41:04 +00:00
|
|
|
/*H:435 And this is us, creating the new page directory. If we really do
|
|
|
|
* allocate a new one (and so the kernel parts are not there), we set
|
|
|
|
* blank_pgdir. */
|
2008-01-07 13:05:37 +00:00
|
|
|
static unsigned int new_pgdir(struct lg_cpu *cpu,
|
2007-10-22 01:03:34 +00:00
|
|
|
unsigned long gpgdir,
|
2007-07-19 08:49:23 +00:00
|
|
|
int *blank_pgdir)
|
|
|
|
{
|
|
|
|
unsigned int next;
|
|
|
|
|
2007-07-26 17:41:04 +00:00
|
|
|
/* We pick one entry at random to throw out. Choosing the Least
|
|
|
|
* Recently Used might be better, but this is easy. */
|
2008-01-17 21:19:42 +00:00
|
|
|
next = random32() % ARRAY_SIZE(cpu->lg->pgdirs);
|
2007-07-26 17:41:04 +00:00
|
|
|
/* If it's never been allocated at all before, try now. */
|
2008-01-17 21:19:42 +00:00
|
|
|
if (!cpu->lg->pgdirs[next].pgdir) {
|
|
|
|
cpu->lg->pgdirs[next].pgdir =
|
|
|
|
(pgd_t *)get_zeroed_page(GFP_KERNEL);
|
2007-07-26 17:41:04 +00:00
|
|
|
/* If the allocation fails, just keep using the one we have */
|
2008-01-17 21:19:42 +00:00
|
|
|
if (!cpu->lg->pgdirs[next].pgdir)
|
2008-01-07 13:05:37 +00:00
|
|
|
next = cpu->cpu_pgd;
|
2007-07-19 08:49:23 +00:00
|
|
|
else
|
2007-07-26 17:41:04 +00:00
|
|
|
/* This is a blank page, so there are no kernel
|
|
|
|
* mappings: caller must map the stack! */
|
2007-07-19 08:49:23 +00:00
|
|
|
*blank_pgdir = 1;
|
|
|
|
}
|
2007-07-26 17:41:04 +00:00
|
|
|
/* Record which Guest toplevel this shadows. */
|
2008-01-17 21:19:42 +00:00
|
|
|
cpu->lg->pgdirs[next].gpgdir = gpgdir;
|
2007-07-19 08:49:23 +00:00
|
|
|
/* Release all the non-kernel mappings. */
|
2008-01-17 21:19:42 +00:00
|
|
|
flush_user_mappings(cpu->lg, next);
|
2007-07-19 08:49:23 +00:00
|
|
|
|
|
|
|
return next;
|
|
|
|
}
|
|
|
|
|
2007-07-26 17:41:04 +00:00
|
|
|
/*H:430 (iv) Switching page tables
|
|
|
|
*
|
2007-10-25 05:02:50 +00:00
|
|
|
* Now we've seen all the page table setting and manipulation, let's see what
|
|
|
|
* what happens when the Guest changes page tables (ie. changes the top-level
|
|
|
|
* pgdir). This occurs on almost every context switch. */
|
2008-01-07 13:05:35 +00:00
|
|
|
void guest_new_pagetable(struct lg_cpu *cpu, unsigned long pgtable)
|
2007-07-19 08:49:23 +00:00
|
|
|
{
|
|
|
|
int newpgdir, repin = 0;
|
|
|
|
|
2007-07-26 17:41:04 +00:00
|
|
|
/* Look to see if we have this one already. */
|
2008-01-17 21:19:42 +00:00
|
|
|
newpgdir = find_pgdir(cpu->lg, pgtable);
|
2007-07-26 17:41:04 +00:00
|
|
|
/* If not, we allocate or mug an existing one: if it's a fresh one,
|
|
|
|
* repin gets set to 1. */
|
2008-01-17 21:19:42 +00:00
|
|
|
if (newpgdir == ARRAY_SIZE(cpu->lg->pgdirs))
|
2008-01-07 13:05:37 +00:00
|
|
|
newpgdir = new_pgdir(cpu, pgtable, &repin);
|
2007-07-26 17:41:04 +00:00
|
|
|
/* Change the current pgd index to the new one. */
|
2008-01-07 13:05:37 +00:00
|
|
|
cpu->cpu_pgd = newpgdir;
|
2007-07-26 17:41:04 +00:00
|
|
|
/* If it was completely blank, we map in the Guest kernel stack */
|
2007-07-19 08:49:23 +00:00
|
|
|
if (repin)
|
2008-01-07 13:05:35 +00:00
|
|
|
pin_stack_pages(cpu);
|
2007-07-19 08:49:23 +00:00
|
|
|
}
|
|
|
|
|
2007-07-26 17:41:04 +00:00
|
|
|
/*H:470 Finally, a routine which throws away everything: all PGD entries in all
|
2007-10-25 05:02:50 +00:00
|
|
|
* the shadow page tables, including the Guest's kernel mappings. This is used
|
|
|
|
* when we destroy the Guest. */
|
2007-07-19 08:49:23 +00:00
|
|
|
static void release_all_pagetables(struct lguest *lg)
|
|
|
|
{
|
|
|
|
unsigned int i, j;
|
|
|
|
|
2007-07-26 17:41:04 +00:00
|
|
|
/* Every shadow pagetable this Guest has */
|
2007-07-19 08:49:23 +00:00
|
|
|
for (i = 0; i < ARRAY_SIZE(lg->pgdirs); i++)
|
|
|
|
if (lg->pgdirs[i].pgdir)
|
2007-07-26 17:41:04 +00:00
|
|
|
/* Every PGD entry except the Switcher at the top */
|
2007-07-19 08:49:23 +00:00
|
|
|
for (j = 0; j < SWITCHER_PGD_INDEX; j++)
|
|
|
|
release_pgd(lg, lg->pgdirs[i].pgdir + j);
|
|
|
|
}
|
|
|
|
|
2007-07-26 17:41:04 +00:00
|
|
|
/* We also throw away everything when a Guest tells us it's changed a kernel
|
|
|
|
* mapping. Since kernel mappings are in every page table, it's easiest to
|
2007-10-25 05:02:50 +00:00
|
|
|
* throw them all away. This traps the Guest in amber for a while as
|
|
|
|
* everything faults back in, but it's rare. */
|
2008-01-07 13:05:35 +00:00
|
|
|
void guest_pagetable_clear_all(struct lg_cpu *cpu)
|
2007-07-19 08:49:23 +00:00
|
|
|
{
|
2008-01-07 13:05:35 +00:00
|
|
|
release_all_pagetables(cpu->lg);
|
2007-07-26 17:41:04 +00:00
|
|
|
/* We need the Guest kernel stack mapped again. */
|
2008-01-07 13:05:35 +00:00
|
|
|
pin_stack_pages(cpu);
|
2007-07-19 08:49:23 +00:00
|
|
|
}
|
2007-10-25 05:02:50 +00:00
|
|
|
/*:*/
|
|
|
|
/*M:009 Since we throw away all mappings when a kernel mapping changes, our
|
|
|
|
* performance sucks for guests using highmem. In fact, a guest with
|
|
|
|
* PAGE_OFFSET 0xc0000000 (the default) and more than about 700MB of RAM is
|
|
|
|
* usually slower than a Guest with less memory.
|
|
|
|
*
|
|
|
|
* This, of course, cannot be fixed. It would take some kind of... well, I
|
|
|
|
* don't know, but the term "puissant code-fu" comes to mind. :*/
|
2007-07-19 08:49:23 +00:00
|
|
|
|
2007-07-26 17:41:04 +00:00
|
|
|
/*H:420 This is the routine which actually sets the page table entry for then
|
|
|
|
* "idx"'th shadow page table.
|
|
|
|
*
|
|
|
|
* Normally, we can just throw out the old entry and replace it with 0: if they
|
|
|
|
* use it demand_page() will put the new entry in. We need to do this anyway:
|
|
|
|
* The Guest expects _PAGE_ACCESSED to be set on its PTE the first time a page
|
|
|
|
* is read from, and _PAGE_DIRTY when it's written to.
|
|
|
|
*
|
|
|
|
* But Avi Kivity pointed out that most Operating Systems (Linux included) set
|
|
|
|
* these bits on PTEs immediately anyway. This is done to save the CPU from
|
|
|
|
* having to update them, but it helps us the same way: if they set
|
|
|
|
* _PAGE_ACCESSED then we can put a read-only PTE entry in immediately, and if
|
|
|
|
* they set _PAGE_DIRTY then we can put a writable PTE entry in immediately.
|
|
|
|
*/
|
2008-01-17 21:19:42 +00:00
|
|
|
static void do_set_pte(struct lg_cpu *cpu, int idx,
|
2007-10-22 01:03:33 +00:00
|
|
|
unsigned long vaddr, pte_t gpte)
|
2007-07-19 08:49:23 +00:00
|
|
|
{
|
2007-10-25 05:02:50 +00:00
|
|
|
/* Look up the matching shadow page directory entry. */
|
2008-01-17 21:19:42 +00:00
|
|
|
pgd_t *spgd = spgd_addr(cpu, idx, vaddr);
|
2007-07-26 17:41:04 +00:00
|
|
|
|
|
|
|
/* If the top level isn't present, there's no entry to update. */
|
2007-10-22 01:03:33 +00:00
|
|
|
if (pgd_flags(*spgd) & _PAGE_PRESENT) {
|
2007-07-26 17:41:04 +00:00
|
|
|
/* Otherwise, we start by releasing the existing entry. */
|
2008-01-17 21:09:49 +00:00
|
|
|
pte_t *spte = spte_addr(*spgd, vaddr);
|
2007-07-19 08:49:23 +00:00
|
|
|
release_pte(*spte);
|
2007-07-26 17:41:04 +00:00
|
|
|
|
|
|
|
/* If they're setting this entry as dirty or accessed, we might
|
|
|
|
* as well put that entry they've given us in now. This shaves
|
|
|
|
* 10% off a copy-on-write micro-benchmark. */
|
2007-10-22 01:03:33 +00:00
|
|
|
if (pte_flags(gpte) & (_PAGE_DIRTY | _PAGE_ACCESSED)) {
|
2008-01-17 21:19:42 +00:00
|
|
|
check_gpte(cpu, gpte);
|
|
|
|
*spte = gpte_to_spte(cpu, gpte,
|
2007-10-22 01:03:33 +00:00
|
|
|
pte_flags(gpte) & _PAGE_DIRTY);
|
2007-07-19 08:49:23 +00:00
|
|
|
} else
|
2007-10-25 05:02:50 +00:00
|
|
|
/* Otherwise kill it and we can demand_page() it in
|
|
|
|
* later. */
|
2007-10-22 01:03:33 +00:00
|
|
|
*spte = __pte(0);
|
2007-07-19 08:49:23 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2007-07-26 17:41:04 +00:00
|
|
|
/*H:410 Updating a PTE entry is a little trickier.
|
|
|
|
*
|
|
|
|
* We keep track of several different page tables (the Guest uses one for each
|
|
|
|
* process, so it makes sense to cache at least a few). Each of these have
|
|
|
|
* identical kernel parts: ie. every mapping above PAGE_OFFSET is the same for
|
|
|
|
* all processes. So when the page table above that address changes, we update
|
|
|
|
* all the page tables, not just the current one. This is rare.
|
|
|
|
*
|
|
|
|
* The benefit is that when we have to track a new page table, we can copy keep
|
|
|
|
* all the kernel mappings. This speeds up context switch immensely. */
|
2008-01-17 21:19:42 +00:00
|
|
|
void guest_set_pte(struct lg_cpu *cpu,
|
2007-10-22 01:03:34 +00:00
|
|
|
unsigned long gpgdir, unsigned long vaddr, pte_t gpte)
|
2007-07-19 08:49:23 +00:00
|
|
|
{
|
2007-07-26 17:41:04 +00:00
|
|
|
/* Kernel mappings must be changed on all top levels. Slow, but
|
|
|
|
* doesn't happen often. */
|
2008-01-17 21:19:42 +00:00
|
|
|
if (vaddr >= cpu->lg->kernel_address) {
|
2007-07-19 08:49:23 +00:00
|
|
|
unsigned int i;
|
2008-01-17 21:19:42 +00:00
|
|
|
for (i = 0; i < ARRAY_SIZE(cpu->lg->pgdirs); i++)
|
|
|
|
if (cpu->lg->pgdirs[i].pgdir)
|
|
|
|
do_set_pte(cpu, i, vaddr, gpte);
|
2007-07-19 08:49:23 +00:00
|
|
|
} else {
|
2007-07-26 17:41:04 +00:00
|
|
|
/* Is this page table one we have a shadow for? */
|
2008-01-17 21:19:42 +00:00
|
|
|
int pgdir = find_pgdir(cpu->lg, gpgdir);
|
|
|
|
if (pgdir != ARRAY_SIZE(cpu->lg->pgdirs))
|
2007-07-26 17:41:04 +00:00
|
|
|
/* If so, do the update. */
|
2008-01-17 21:19:42 +00:00
|
|
|
do_set_pte(cpu, pgdir, vaddr, gpte);
|
2007-07-19 08:49:23 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2007-07-26 17:41:04 +00:00
|
|
|
/*H:400
|
2007-10-25 05:02:50 +00:00
|
|
|
* (iii) Setting up a page table entry when the Guest tells us one has changed.
|
2007-07-26 17:41:04 +00:00
|
|
|
*
|
|
|
|
* Just like we did in interrupts_and_traps.c, it makes sense for us to deal
|
|
|
|
* with the other side of page tables while we're here: what happens when the
|
|
|
|
* Guest asks for a page table to be updated?
|
|
|
|
*
|
|
|
|
* We already saw that demand_page() will fill in the shadow page tables when
|
|
|
|
* needed, so we can simply remove shadow page table entries whenever the Guest
|
|
|
|
* tells us they've changed. When the Guest tries to use the new entry it will
|
|
|
|
* fault and demand_page() will fix it up.
|
|
|
|
*
|
|
|
|
* So with that in mind here's our code to to update a (top-level) PGD entry:
|
|
|
|
*/
|
2007-10-22 01:03:34 +00:00
|
|
|
void guest_set_pmd(struct lguest *lg, unsigned long gpgdir, u32 idx)
|
2007-07-19 08:49:23 +00:00
|
|
|
{
|
|
|
|
int pgdir;
|
|
|
|
|
2007-07-26 17:41:04 +00:00
|
|
|
/* The kernel seems to try to initialize this early on: we ignore its
|
|
|
|
* attempts to map over the Switcher. */
|
2007-07-19 08:49:23 +00:00
|
|
|
if (idx >= SWITCHER_PGD_INDEX)
|
|
|
|
return;
|
|
|
|
|
2007-07-26 17:41:04 +00:00
|
|
|
/* If they're talking about a page table we have a shadow for... */
|
2007-10-22 01:03:34 +00:00
|
|
|
pgdir = find_pgdir(lg, gpgdir);
|
2007-07-19 08:49:23 +00:00
|
|
|
if (pgdir < ARRAY_SIZE(lg->pgdirs))
|
2007-07-26 17:41:04 +00:00
|
|
|
/* ... throw it away. */
|
2007-07-19 08:49:23 +00:00
|
|
|
release_pgd(lg, lg->pgdirs[pgdir].pgdir + idx);
|
|
|
|
}
|
|
|
|
|
2007-07-26 17:41:04 +00:00
|
|
|
/*H:500 (vii) Setting up the page tables initially.
|
|
|
|
*
|
|
|
|
* When a Guest is first created, the Launcher tells us where the toplevel of
|
|
|
|
* its first page table is. We set some things up here: */
|
2007-07-19 08:49:23 +00:00
|
|
|
int init_guest_pagetable(struct lguest *lg, unsigned long pgtable)
|
|
|
|
{
|
2007-07-26 17:41:04 +00:00
|
|
|
/* We start on the first shadow page table, and give it a blank PGD
|
|
|
|
* page. */
|
2008-01-07 13:05:37 +00:00
|
|
|
lg->pgdirs[0].gpgdir = pgtable;
|
|
|
|
lg->pgdirs[0].pgdir = (pgd_t *)get_zeroed_page(GFP_KERNEL);
|
|
|
|
if (!lg->pgdirs[0].pgdir)
|
2007-07-19 08:49:23 +00:00
|
|
|
return -ENOMEM;
|
2008-01-07 13:05:37 +00:00
|
|
|
lg->cpus[0].cpu_pgd = 0;
|
2007-07-19 08:49:23 +00:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2007-10-22 01:03:36 +00:00
|
|
|
/* When the Guest calls LHCALL_LGUEST_INIT we do more setup. */
|
2008-01-17 21:19:42 +00:00
|
|
|
void page_table_guest_data_init(struct lg_cpu *cpu)
|
2007-10-22 01:03:36 +00:00
|
|
|
{
|
|
|
|
/* We get the kernel address: above this is all kernel memory. */
|
2008-01-17 21:19:42 +00:00
|
|
|
if (get_user(cpu->lg->kernel_address,
|
|
|
|
&cpu->lg->lguest_data->kernel_address)
|
2007-10-22 01:03:36 +00:00
|
|
|
/* We tell the Guest that it can't use the top 4MB of virtual
|
|
|
|
* addresses used by the Switcher. */
|
2008-01-17 21:19:42 +00:00
|
|
|
|| put_user(4U*1024*1024, &cpu->lg->lguest_data->reserve_mem)
|
|
|
|
|| put_user(cpu->lg->pgdirs[0].gpgdir, &cpu->lg->lguest_data->pgdir))
|
|
|
|
kill_guest(cpu, "bad guest page %p", cpu->lg->lguest_data);
|
2007-10-22 01:03:36 +00:00
|
|
|
|
|
|
|
/* In flush_user_mappings() we loop from 0 to
|
|
|
|
* "pgd_index(lg->kernel_address)". This assumes it won't hit the
|
|
|
|
* Switcher mappings, so check that now. */
|
2008-01-17 21:19:42 +00:00
|
|
|
if (pgd_index(cpu->lg->kernel_address) >= SWITCHER_PGD_INDEX)
|
|
|
|
kill_guest(cpu, "bad kernel address %#lx",
|
|
|
|
cpu->lg->kernel_address);
|
2007-10-22 01:03:36 +00:00
|
|
|
}
|
|
|
|
|
2007-07-26 17:41:04 +00:00
|
|
|
/* When a Guest dies, our cleanup is fairly simple. */
|
2007-07-19 08:49:23 +00:00
|
|
|
void free_guest_pagetable(struct lguest *lg)
|
|
|
|
{
|
|
|
|
unsigned int i;
|
|
|
|
|
2007-07-26 17:41:04 +00:00
|
|
|
/* Throw away all page table pages. */
|
2007-07-19 08:49:23 +00:00
|
|
|
release_all_pagetables(lg);
|
2007-07-26 17:41:04 +00:00
|
|
|
/* Now free the top levels: free_page() can handle 0 just fine. */
|
2007-07-19 08:49:23 +00:00
|
|
|
for (i = 0; i < ARRAY_SIZE(lg->pgdirs); i++)
|
|
|
|
free_page((long)lg->pgdirs[i].pgdir);
|
|
|
|
}
|
|
|
|
|
2007-07-26 17:41:04 +00:00
|
|
|
/*H:480 (vi) Mapping the Switcher when the Guest is about to run.
|
|
|
|
*
|
2007-10-25 05:02:50 +00:00
|
|
|
* The Switcher and the two pages for this CPU need to be visible in the
|
2007-07-26 17:41:04 +00:00
|
|
|
* Guest (and not the pages for other CPUs). We have the appropriate PTE pages
|
2007-10-25 05:02:50 +00:00
|
|
|
* for each CPU already set up, we just need to hook them in now we know which
|
|
|
|
* Guest is about to run on this CPU. */
|
2008-01-07 13:05:30 +00:00
|
|
|
void map_switcher_in_guest(struct lg_cpu *cpu, struct lguest_pages *pages)
|
2007-07-19 08:49:23 +00:00
|
|
|
{
|
2007-10-22 01:03:33 +00:00
|
|
|
pte_t *switcher_pte_page = __get_cpu_var(switcher_pte_pages);
|
|
|
|
pgd_t switcher_pgd;
|
|
|
|
pte_t regs_pte;
|
2008-01-07 13:05:32 +00:00
|
|
|
unsigned long pfn;
|
2007-07-19 08:49:23 +00:00
|
|
|
|
2007-07-26 17:41:04 +00:00
|
|
|
/* Make the last PGD entry for this Guest point to the Switcher's PTE
|
|
|
|
* page for this CPU (with appropriate flags). */
|
2008-01-19 01:59:08 +00:00
|
|
|
switcher_pgd = __pgd(__pa(switcher_pte_page) | __PAGE_KERNEL);
|
2007-10-22 01:03:33 +00:00
|
|
|
|
2008-01-07 13:05:37 +00:00
|
|
|
cpu->lg->pgdirs[cpu->cpu_pgd].pgdir[SWITCHER_PGD_INDEX] = switcher_pgd;
|
2007-07-19 08:49:23 +00:00
|
|
|
|
2007-07-26 17:41:04 +00:00
|
|
|
/* We also change the Switcher PTE page. When we're running the Guest,
|
|
|
|
* we want the Guest's "regs" page to appear where the first Switcher
|
|
|
|
* page for this CPU is. This is an optimization: when the Switcher
|
|
|
|
* saves the Guest registers, it saves them into the first page of this
|
|
|
|
* CPU's "struct lguest_pages": if we make sure the Guest's register
|
|
|
|
* page is already mapped there, we don't have to copy them out
|
|
|
|
* again. */
|
2008-01-07 13:05:32 +00:00
|
|
|
pfn = __pa(cpu->regs_page) >> PAGE_SHIFT;
|
2008-01-19 01:59:08 +00:00
|
|
|
regs_pte = pfn_pte(pfn, __pgprot(__PAGE_KERNEL));
|
2007-10-22 01:03:33 +00:00
|
|
|
switcher_pte_page[(unsigned long)pages/PAGE_SIZE%PTRS_PER_PTE] = regs_pte;
|
2007-07-19 08:49:23 +00:00
|
|
|
}
|
2007-07-26 17:41:04 +00:00
|
|
|
/*:*/
|
2007-07-19 08:49:23 +00:00
|
|
|
|
|
|
|
static void free_switcher_pte_pages(void)
|
|
|
|
{
|
|
|
|
unsigned int i;
|
|
|
|
|
|
|
|
for_each_possible_cpu(i)
|
|
|
|
free_page((long)switcher_pte_page(i));
|
|
|
|
}
|
|
|
|
|
2007-07-26 17:41:04 +00:00
|
|
|
/*H:520 Setting up the Switcher PTE page for given CPU is fairly easy, given
|
|
|
|
* the CPU number and the "struct page"s for the Switcher code itself.
|
|
|
|
*
|
|
|
|
* Currently the Switcher is less than a page long, so "pages" is always 1. */
|
2007-07-19 08:49:23 +00:00
|
|
|
static __init void populate_switcher_pte_page(unsigned int cpu,
|
|
|
|
struct page *switcher_page[],
|
|
|
|
unsigned int pages)
|
|
|
|
{
|
|
|
|
unsigned int i;
|
2007-10-22 01:03:33 +00:00
|
|
|
pte_t *pte = switcher_pte_page(cpu);
|
2007-07-19 08:49:23 +00:00
|
|
|
|
2007-07-26 17:41:04 +00:00
|
|
|
/* The first entries are easy: they map the Switcher code. */
|
2007-07-19 08:49:23 +00:00
|
|
|
for (i = 0; i < pages; i++) {
|
2007-10-22 01:03:33 +00:00
|
|
|
pte[i] = mk_pte(switcher_page[i],
|
|
|
|
__pgprot(_PAGE_PRESENT|_PAGE_ACCESSED));
|
2007-07-19 08:49:23 +00:00
|
|
|
}
|
|
|
|
|
2007-07-26 17:41:04 +00:00
|
|
|
/* The only other thing we map is this CPU's pair of pages. */
|
2007-07-19 08:49:23 +00:00
|
|
|
i = pages + cpu*2;
|
|
|
|
|
2007-07-26 17:41:04 +00:00
|
|
|
/* First page (Guest registers) is writable from the Guest */
|
2007-10-22 01:03:33 +00:00
|
|
|
pte[i] = pfn_pte(page_to_pfn(switcher_page[i]),
|
|
|
|
__pgprot(_PAGE_PRESENT|_PAGE_ACCESSED|_PAGE_RW));
|
|
|
|
|
2007-07-26 17:41:04 +00:00
|
|
|
/* The second page contains the "struct lguest_ro_state", and is
|
|
|
|
* read-only. */
|
2007-10-22 01:03:33 +00:00
|
|
|
pte[i+1] = pfn_pte(page_to_pfn(switcher_page[i+1]),
|
|
|
|
__pgprot(_PAGE_PRESENT|_PAGE_ACCESSED));
|
2007-07-19 08:49:23 +00:00
|
|
|
}
|
|
|
|
|
2007-10-25 05:02:50 +00:00
|
|
|
/* We've made it through the page table code. Perhaps our tired brains are
|
|
|
|
* still processing the details, or perhaps we're simply glad it's over.
|
|
|
|
*
|
|
|
|
* If nothing else, note that all this complexity in juggling shadow page
|
|
|
|
* tables in sync with the Guest's page tables is for one reason: for most
|
|
|
|
* Guests this page table dance determines how bad performance will be. This
|
|
|
|
* is why Xen uses exotic direct Guest pagetable manipulation, and why both
|
|
|
|
* Intel and AMD have implemented shadow page table support directly into
|
|
|
|
* hardware.
|
|
|
|
*
|
|
|
|
* There is just one file remaining in the Host. */
|
|
|
|
|
2007-07-26 17:41:04 +00:00
|
|
|
/*H:510 At boot or module load time, init_pagetables() allocates and populates
|
|
|
|
* the Switcher PTE page for each CPU. */
|
2007-07-19 08:49:23 +00:00
|
|
|
__init int init_pagetables(struct page **switcher_page, unsigned int pages)
|
|
|
|
{
|
|
|
|
unsigned int i;
|
|
|
|
|
|
|
|
for_each_possible_cpu(i) {
|
2007-10-22 01:03:33 +00:00
|
|
|
switcher_pte_page(i) = (pte_t *)get_zeroed_page(GFP_KERNEL);
|
2007-07-19 08:49:23 +00:00
|
|
|
if (!switcher_pte_page(i)) {
|
|
|
|
free_switcher_pte_pages();
|
|
|
|
return -ENOMEM;
|
|
|
|
}
|
|
|
|
populate_switcher_pte_page(i, switcher_page, pages);
|
|
|
|
}
|
|
|
|
return 0;
|
|
|
|
}
|
2007-07-26 17:41:04 +00:00
|
|
|
/*:*/
|
2007-07-19 08:49:23 +00:00
|
|
|
|
2007-07-26 17:41:04 +00:00
|
|
|
/* Cleaning up simply involves freeing the PTE page for each CPU. */
|
2007-07-19 08:49:23 +00:00
|
|
|
void free_pagetables(void)
|
|
|
|
{
|
|
|
|
free_switcher_pte_pages();
|
|
|
|
}
|