/*P:400 This contains run_guest() which actually calls into the Host<->Guest
* Switcher and analyzes the return, such as determining if the Guest wants the
- * Host to do something. This file also contains useful helper routines, and a
- * couple of non-obvious setup and teardown pieces which were implemented after
- * days of debugging pain. :*/
+ * Host to do something. This file also contains useful helper routines. :*/
#include <linux/module.h>
#include <linux/stringify.h>
#include <linux/stddef.h>
* easy.
*/
- /* We allocate an array of "struct page"s. map_vm_area() wants the
- * pages in this form, rather than just an array of pointers. */
+ /* We allocate an array of struct page pointers. map_vm_area() wants
+ * this, rather than just an array of pages. */
switcher_page = kmalloc(sizeof(switcher_page[0])*TOTAL_SWITCHER_PAGES,
GFP_KERNEL);
if (!switcher_page) {
switcher_page[i] = virt_to_page(addr);
}
+ /* First we check that the Switcher won't overlap the fixmap area at
+ * the top of memory. It's currently nowhere near, but it could have
+ * very strange effects if it ever happened. */
+ if (SWITCHER_ADDR + (TOTAL_SWITCHER_PAGES+1)*PAGE_SIZE > FIXADDR_START){
+ err = -ENOMEM;
+ printk("lguest: mapping switcher would thwack fixmap\n");
+ goto free_pages;
+ }
+
/* Now we reserve the "virtual memory area" we want: 0xFFC00000
* (SWITCHER_ADDR). We might not get it in theory, but in practice
- * it's worked so far. */
+ * it's worked so far. The end address needs +1 because __get_vm_area
+ * allocates an extra guard page, so we need space for that. */
switcher_vma = __get_vm_area(TOTAL_SWITCHER_PAGES * PAGE_SIZE,
- VM_ALLOC, SWITCHER_ADDR, VMALLOC_END);
+ VM_ALLOC, SWITCHER_ADDR, SWITCHER_ADDR
+ + (TOTAL_SWITCHER_PAGES+1) * PAGE_SIZE);
if (!switcher_vma) {
err = -ENOMEM;
printk("lguest: could not map switcher pages high\n");
/* Now we just need to free the pages we copied the switcher into */
for (i = 0; i < TOTAL_SWITCHER_PAGES; i++)
__free_pages(switcher_page[i], 0);
+ kfree(switcher_page);
}
-/*L:305
+/*H:032
* Dealing With Guest Memory.
*
+ * Before we go too much further into the Host, we need to grok the routines
+ * we use to deal with Guest memory.
+ *
* When the Guest gives us (what it thinks is) a physical address, we can use
* the normal copy_from_user() & copy_to_user() on the corresponding place in
* the memory region allocated by the Launcher.
return (addr+len) / PAGE_SIZE < lg->pfn_limit && (addr+len >= addr);
}
-/* This is a convenient routine to get a 32-bit value from the Guest (a very
- * common operation). Here we can see how useful the kill_lguest() routine we
- * met in the Launcher can be: we return a random value (0) instead of needing
- * to return an error. */
-u32 lgread_u32(struct lguest *lg, unsigned long addr)
+/* This routine copies memory from the Guest. Here we can see how useful the
+ * kill_lguest() routine we met in the Launcher can be: we return a random
+ * value (all zeroes) instead of needing to return an error. */
+void __lgread(struct lg_cpu *cpu, void *b, unsigned long addr, unsigned bytes)
{
- u32 val = 0;
-
- /* Don't let them access lguest binary. */
- if (!lguest_address_ok(lg, addr, sizeof(val))
- || get_user(val, (u32 *)(lg->mem_base + addr)) != 0)
- kill_guest(lg, "bad read address %#lx: pfn_limit=%u membase=%p", addr, lg->pfn_limit, lg->mem_base);
- return val;
-}
-
-/* Same thing for writing a value. */
-void lgwrite_u32(struct lguest *lg, unsigned long addr, u32 val)
-{
- if (!lguest_address_ok(lg, addr, sizeof(val))
- || put_user(val, (u32 *)(lg->mem_base + addr)) != 0)
- kill_guest(lg, "bad write address %#lx", addr);
-}
-
-/* This routine is more generic, and copies a range of Guest bytes into a
- * buffer. If the copy_from_user() fails, we fill the buffer with zeroes, so
- * the caller doesn't end up using uninitialized kernel memory. */
-void lgread(struct lguest *lg, void *b, unsigned long addr, unsigned bytes)
-{
- if (!lguest_address_ok(lg, addr, bytes)
- || copy_from_user(b, lg->mem_base + addr, bytes) != 0) {
+ if (!lguest_address_ok(cpu->lg, addr, bytes)
+ || copy_from_user(b, cpu->lg->mem_base + addr, bytes) != 0) {
/* copy_from_user should do this, but as we rely on it... */
memset(b, 0, bytes);
- kill_guest(lg, "bad read address %#lx len %u", addr, bytes);
+ kill_guest(cpu, "bad read address %#lx len %u", addr, bytes);
}
}
-/* Similarly, our generic routine to copy into a range of Guest bytes. */
-void lgwrite(struct lguest *lg, unsigned long addr, const void *b,
- unsigned bytes)
+/* This is the write (copy into Guest) version. */
+void __lgwrite(struct lg_cpu *cpu, unsigned long addr, const void *b,
+ unsigned bytes)
{
- if (!lguest_address_ok(lg, addr, bytes)
- || copy_to_user(lg->mem_base + addr, b, bytes) != 0)
- kill_guest(lg, "bad write address %#lx len %u", addr, bytes);
+ if (!lguest_address_ok(cpu->lg, addr, bytes)
+ || copy_to_user(cpu->lg->mem_base + addr, b, bytes) != 0)
+ kill_guest(cpu, "bad write address %#lx len %u", addr, bytes);
}
-/* (end of memory access helper routines) :*/
+/*:*/
/*H:030 Let's jump straight to the the main loop which runs the Guest.
* Remember, this is called by the Launcher reading /dev/lguest, and we keep
* going around and around until something interesting happens. */
-int run_guest(struct lguest *lg, unsigned long __user *user)
+int run_guest(struct lg_cpu *cpu, unsigned long __user *user)
{
/* We stop running once the Guest is dead. */
- while (!lg->dead) {
- /* First we run any hypercalls the Guest wants done: either in
- * the hypercall ring in "struct lguest_data", or directly by
- * using int 31 (LGUEST_TRAP_ENTRY). */
- do_hypercalls(lg);
- /* It's possible the Guest did a SEND_DMA hypercall to the
+ while (!cpu->lg->dead) {
+ /* First we run any hypercalls the Guest wants done. */
+ if (cpu->hcall)
+ do_hypercalls(cpu);
+
+ /* It's possible the Guest did a NOTIFY hypercall to the
* Launcher, in which case we return from the read() now. */
- if (lg->dma_is_pending) {
- if (put_user(lg->pending_dma, user) ||
- put_user(lg->pending_key, user+1))
+ if (cpu->pending_notify) {
+ if (put_user(cpu->pending_notify, user))
return -EFAULT;
- return sizeof(unsigned long)*2;
+ return sizeof(cpu->pending_notify);
}
/* Check for signals */
return -ERESTARTSYS;
/* If Waker set break_out, return to Launcher. */
- if (lg->break_out)
+ if (cpu->break_out)
return -EAGAIN;
- /* Check if there are any interrupts which can be delivered
- * now: if so, this sets up the hander to be executed when we
- * next run the Guest. */
- maybe_do_interrupt(lg);
+ /* Check if there are any interrupts which can be delivered now:
+ * if so, this sets up the hander to be executed when we next
+ * run the Guest. */
+ maybe_do_interrupt(cpu);
/* All long-lived kernel loops need to check with this horrible
* thing called the freezer. If the Host is trying to suspend,
/* Just make absolutely sure the Guest is still alive. One of
* those hypercalls could have been fatal, for example. */
- if (lg->dead)
+ if (cpu->lg->dead)
break;
/* If the Guest asked to be stopped, we sleep. The Guest's
- * clock timer or LHCALL_BREAK from the Waker will wake us. */
- if (lg->halted) {
+ * clock timer or LHREQ_BREAK from the Waker will wake us. */
+ if (cpu->halted) {
set_current_state(TASK_INTERRUPTIBLE);
schedule();
continue;
local_irq_disable();
/* Actually run the Guest until something happens. */
- lguest_arch_run_guest(lg);
+ lguest_arch_run_guest(cpu);
/* Now we're ready to be interrupted or moved to other CPUs */
local_irq_enable();
/* Now we deal with whatever happened to the Guest. */
- lguest_arch_handle_trap(lg);
+ lguest_arch_handle_trap(cpu);
}
+ /* Special case: Guest is 'dead' but wants a reboot. */
+ if (cpu->lg->dead == ERR_PTR(-ERESTART))
+ return -ERESTART;
+
/* The Guest is dead => "No such file or directory" */
return -ENOENT;
}
/* Lguest can't run under Xen, VMI or itself. It does Tricky Stuff. */
if (paravirt_enabled()) {
- printk("lguest is afraid of %s\n", pv_info.name);
+ printk("lguest is afraid of being a guest\n");
return -EPERM;
}
/* First we put the Switcher up in very high virtual memory. */
err = map_switcher();
if (err)
- return err;
+ goto out;
/* Now we set up the pagetable implementation for the Guests. */
err = init_pagetables(switcher_page, SHARED_SWITCHER_PAGES);
- if (err) {
- unmap_switcher();
- return err;
- }
+ if (err)
+ goto unmap;
- /* The I/O subsystem needs some things initialized. */
- lguest_io_init();
+ /* We might need to reserve an interrupt vector. */
+ err = init_interrupts();
+ if (err)
+ goto free_pgtables;
/* /dev/lguest needs to be registered. */
err = lguest_device_init();
- if (err) {
- free_pagetables();
- unmap_switcher();
- return err;
- }
+ if (err)
+ goto free_interrupts;
/* Finally we do some architecture-specific setup. */
lguest_arch_host_init();
/* All good! */
return 0;
+
+free_interrupts:
+ free_interrupts();
+free_pgtables:
+ free_pagetables();
+unmap:
+ unmap_switcher();
+out:
+ return err;
}
/* Cleaning up is just the same code, backwards. With a little French. */
static void __exit fini(void)
{
lguest_device_remove();
+ free_interrupts();
free_pagetables();
unmap_switcher();