summaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authorRusty Russell <rusty@rustcorp.com.au>2009-07-30 16:03:45 -0600
committerRusty Russell <rusty@rustcorp.com.au>2009-07-30 16:03:46 +0930
commita91d74a3c4de8115295ee87350c13a329164aaaf (patch)
tree02c862fccc9abedf7fc354061e69c4b5fbcce06d
parent2e04ef76916d1e29a077ea9d0f2003c8fd86724d (diff)
lguest: update commentry
Every so often, after code shuffles, I need to go through and unbitrot the Lguest Journey (see drivers/lguest/README). Since we now use RCU in a simple form in one place I took the opportunity to expand that explanation. Signed-off-by: Rusty Russell <rusty@rustcorp.com.au> Cc: Ingo Molnar <mingo@redhat.com> Cc: Paul McKenney <paulmck@linux.vnet.ibm.com>
-rw-r--r--Documentation/lguest/lguest.c184
-rw-r--r--arch/x86/include/asm/lguest_hcall.h8
-rw-r--r--arch/x86/lguest/boot.c99
-rw-r--r--arch/x86/lguest/i386_head.S2
-rw-r--r--drivers/lguest/core.c7
-rw-r--r--drivers/lguest/hypercalls.c6
-rw-r--r--drivers/lguest/lguest_device.c11
-rw-r--r--drivers/lguest/lguest_user.c100
-rw-r--r--drivers/lguest/page_tables.c84
-rw-r--r--drivers/lguest/x86/core.c2
-rw-r--r--drivers/lguest/x86/switcher_32.S6
11 files changed, 398 insertions, 111 deletions
diff --git a/Documentation/lguest/lguest.c b/Documentation/lguest/lguest.c
index aa66a52b73e9..45163651b519 100644
--- a/Documentation/lguest/lguest.c
+++ b/Documentation/lguest/lguest.c
@@ -49,7 +49,7 @@
#include "linux/virtio_ring.h"
#include "asm/bootparam.h"
/*L:110
- * We can ignore the 39 include files we need for this program, but I do want
+ * We can ignore the 42 include files we need for this program, but I do want
* to draw attention to the use of kernel-style types.
*
* As Linus said, "C is a Spartan language, and so should your naming be." I
@@ -305,6 +305,11 @@ static void *map_zeroed_pages(unsigned int num)
PROT_READ|PROT_WRITE|PROT_EXEC, MAP_PRIVATE, fd, 0);
if (addr == MAP_FAILED)
err(1, "Mmaping %u pages of /dev/zero", num);
+
+ /*
+ * One neat mmap feature is that you can close the fd, and it
+ * stays mapped.
+ */
close(fd);
return addr;
@@ -557,7 +562,7 @@ static void tell_kernel(unsigned long start)
}
/*:*/
-/*
+/*L:200
* Device Handling.
*
* When the Guest gives us a buffer, it sends an array of addresses and sizes.
@@ -608,7 +613,10 @@ static unsigned next_desc(struct vring_desc *desc,
return next;
}
-/* This actually sends the interrupt for this virtqueue */
+/*
+ * This actually sends the interrupt for this virtqueue, if we've used a
+ * buffer.
+ */
static void trigger_irq(struct virtqueue *vq)
{
unsigned long buf[] = { LHREQ_IRQ, vq->config.irq };
@@ -629,12 +637,12 @@ static void trigger_irq(struct virtqueue *vq)
}
/*
- * This looks in the virtqueue and for the first available buffer, and converts
+ * This looks in the virtqueue for the first available buffer, and converts
* it to an iovec for convenient access. Since descriptors consist of some
* number of output then some number of input descriptors, it's actually two
* iovecs, but we pack them into one and note how many of each there were.
*
- * This function returns the descriptor number found.
+ * This function waits if necessary, and returns the descriptor number found.
*/
static unsigned wait_for_vq_desc(struct virtqueue *vq,
struct iovec iov[],
@@ -644,10 +652,14 @@ static unsigned wait_for_vq_desc(struct virtqueue *vq,
struct vring_desc *desc;
u16 last_avail = lg_last_avail(vq);
+ /* There's nothing available? */
while (last_avail == vq->vring.avail->idx) {
u64 event;
- /* OK, tell Guest about progress up to now. */
+ /*
+ * Since we're about to sleep, now is a good time to tell the
+ * Guest about what we've used up to now.
+ */
trigger_irq(vq);
/* OK, now we need to know about added descriptors. */
@@ -734,8 +746,9 @@ static unsigned wait_for_vq_desc(struct virtqueue *vq,
}
/*
- * After we've used one of their buffers, we tell them about it. We'll then
- * want to send them an interrupt, using trigger_irq().
+ * After we've used one of their buffers, we tell the Guest about it. Sometime
+ * later we'll want to send them an interrupt using trigger_irq(); note that
+ * wait_for_vq_desc() does that for us if it has to wait.
*/
static void add_used(struct virtqueue *vq, unsigned int head, int len)
{
@@ -782,12 +795,12 @@ static void console_input(struct virtqueue *vq)
struct console_abort *abort = vq->dev->priv;
struct iovec iov[vq->vring.num];
- /* Make sure there's a descriptor waiting. */
+ /* Make sure there's a descriptor available. */
head = wait_for_vq_desc(vq, iov, &out_num, &in_num);
if (out_num)
errx(1, "Output buffers in console in queue?");
- /* Read it in. */
+ /* Read into it. This is where we usually wait. */
len = readv(STDIN_FILENO, iov, in_num);
if (len <= 0) {
/* Ran out of input? */
@@ -800,6 +813,7 @@ static void console_input(struct virtqueue *vq)
pause();
}
+ /* Tell the Guest we used a buffer. */
add_used_and_trigger(vq, head, len);
/*
@@ -834,15 +848,23 @@ static void console_output(struct virtqueue *vq)
unsigned int head, out, in;
struct iovec iov[vq->vring.num];
+ /* We usually wait in here, for the Guest to give us something. */
head = wait_for_vq_desc(vq, iov, &out, &in);
if (in)
errx(1, "Input buffers in console output queue?");
+
+ /* writev can return a partial write, so we loop here. */
while (!iov_empty(iov, out)) {
int len = writev(STDOUT_FILENO, iov, out);
if (len <= 0)
err(1, "Write to stdout gave %i", len);
iov_consume(iov, out, len);
}
+
+ /*
+ * We're finished with that buffer: if we're going to sleep,
+ * wait_for_vq_desc() will prod the Guest with an interrupt.
+ */
add_used(vq, head, 0);
}
@@ -862,15 +884,30 @@ static void net_output(struct virtqueue *vq)
unsigned int head, out, in;
struct iovec iov[vq->vring.num];
+ /* We usually wait in here for the Guest to give us a packet. */
head = wait_for_vq_desc(vq, iov, &out, &in);
if (in)
errx(1, "Input buffers in net output queue?");
+ /*
+ * Send the whole thing through to /dev/net/tun. It expects the exact
+ * same format: what a coincidence!
+ */
if (writev(net_info->tunfd, iov, out) < 0)
errx(1, "Write to tun failed?");
+
+ /*
+ * Done with that one; wait_for_vq_desc() will send the interrupt if
+ * all packets are processed.
+ */
add_used(vq, head, 0);
}
-/* Will reading from this file descriptor block? */
+/*
+ * Handling network input is a bit trickier, because I've tried to optimize it.
+ *
+ * First we have a helper routine which tells is if from this file descriptor
+ * (ie. the /dev/net/tun device) will block:
+ */
static bool will_block(int fd)
{
fd_set fdset;
@@ -880,7 +917,11 @@ static bool will_block(int fd)
return select(fd+1, &fdset, NULL, NULL, &zero) != 1;
}
-/* This handles packets coming in from the tun device to our Guest. */
+/*
+ * This handles packets coming in from the tun device to our Guest. Like all
+ * service routines, it gets called again as soon as it returns, so you don't
+ * see a while(1) loop here.
+ */
static void net_input(struct virtqueue *vq)
{
int len;
@@ -888,21 +929,38 @@ static void net_input(struct virtqueue *vq)
struct iovec iov[vq->vring.num];
struct net_info *net_info = vq->dev->priv;
+ /*
+ * Get a descriptor to write an incoming packet into. This will also
+ * send an interrupt if they're out of descriptors.
+ */
head = wait_for_vq_desc(vq, iov, &out, &in);
if (out)
errx(1, "Output buffers in net input queue?");
- /* Deliver interrupt now, since we're about to sleep. */
+ /*
+ * If it looks like we'll block reading from the tun device, send them
+ * an interrupt.
+ */
if (vq->pending_used && will_block(net_info->tunfd))
trigger_irq(vq);
+ /*
+ * Read in the packet. This is where we normally wait (when there's no
+ * incoming network traffic).
+ */
len = readv(net_info->tunfd, iov, in);
if (len <= 0)
err(1, "Failed to read from tun.");
+
+ /*
+ * Mark that packet buffer as used, but don't interrupt here. We want
+ * to wait until we've done as much work as we can.
+ */
add_used(vq, head, len);
}
+/*:*/
-/* This is the helper to create threads. */
+/* This is the helper to create threads: run the service routine in a loop. */
static int do_thread(void *_vq)
{
struct virtqueue *vq = _vq;
@@ -950,11 +1008,14 @@ static void reset_device(struct device *dev)
signal(SIGCHLD, (void *)kill_launcher);
}
+/*L:216
+ * This actually creates the thread which services the virtqueue for a device.
+ */
static void create_thread(struct virtqueue *vq)
{
/*
- * Create stack for thread and run it. Since the stack grows upwards,
- * we point the stack pointer to the end of this region.
+ * Create stack for thread. Since the stack grows upwards, we point
+ * the stack pointer to the end of this region.
*/
char *stack = malloc(32768);
unsigned long args[] = { LHREQ_EVENTFD,
@@ -966,17 +1027,22 @@ static void create_thread(struct virtqueue *vq)
err(1, "Creating eventfd");
args[2] = vq->eventfd;
- /* Attach an eventfd to this virtqueue: it will go off
- * when the Guest does an LHCALL_NOTIFY for this vq. */
+ /*
+ * Attach an eventfd to this virtqueue: it will go off when the Guest
+ * does an LHCALL_NOTIFY for this vq.
+ */
if (write(lguest_fd, &args, sizeof(args)) != 0)
err(1, "Attaching eventfd");
- /* CLONE_VM: because it has to access the Guest memory, and
- * SIGCHLD so we get a signal if it dies. */
+ /*
+ * CLONE_VM: because it has to access the Guest memory, and SIGCHLD so
+ * we get a signal if it dies.
+ */
vq->thread = clone(do_thread, stack + 32768, CLONE_VM | SIGCHLD, vq);
if (vq->thread == (pid_t)-1)
err(1, "Creating clone");
- /* We close our local copy, now the child has it. */
+
+ /* We close our local copy now the child has it. */
close(vq->eventfd);
}
@@ -1028,7 +1094,10 @@ static void update_device_status(struct device *dev)
}
}
-/* This is the generic routine we call when the Guest uses LHCALL_NOTIFY. */
+/*L:215
+ * This is the generic routine we call when the Guest uses LHCALL_NOTIFY. In
+ * particular, it's used to notify us of device status changes during boot.
+ */
static void handle_output(unsigned long addr)
{
struct device *i;
@@ -1037,18 +1106,32 @@ static void handle_output(unsigned long addr)
for (i = devices.dev; i; i = i->next) {
struct virtqueue *vq;
- /* Notifications to device descriptors update device status. */
+ /*
+ * Notifications to device descriptors mean they updated the
+ * device status.
+ */
if (from_guest_phys(addr) == i->desc) {
update_device_status(i);
return;
}
- /* Devices *can* be used before status is set to DRIVER_OK. */
+ /*
+ * Devices *can* be used before status is set to DRIVER_OK.
+ * The original plan was that they would never do this: they
+ * would always finish setting up their status bits before
+ * actually touching the virtqueues. In practice, we allowed
+ * them to, and they do (eg. the disk probes for partition
+ * tables as part of initialization).
+ *
+ * If we see this, we start the device: once it's running, we
+ * expect the device to catch all the notifications.
+ */
for (vq = i->vq; vq; vq = vq->next) {
if (addr != vq->config.pfn*getpagesize())
continue;
if (i->running)
errx(1, "Notification on running %s", i->name);
+ /* This just calls create_thread() for each virtqueue */
start_device(i);
return;
}
@@ -1132,6 +1215,11 @@ static void add_virtqueue(struct device *dev, unsigned int num_descs,
vq->next = NULL;
vq->last_avail_idx = 0;
vq->dev = dev;
+
+ /*
+ * This is the routine the service thread will run, and its Process ID
+ * once it's running.
+ */
vq->service = service;
vq->thread = (pid_t)-1;
@@ -1202,7 +1290,8 @@ static void set_config(struct device *dev, unsigned len, const void *conf)
/*
* This routine does all the creation and setup of a new device, including
- * calling new_dev_desc() to allocate the descriptor and device memory.
+ * calling new_dev_desc() to allocate the descriptor and device memory. We
+ * don't actually start the service threads until later.
*
* See what I mean about userspace being boring?
*/
@@ -1478,19 +1567,7 @@ static void setup_tun_net(char *arg)
verbose("device %u: tun %s: %s\n",
devices.device_num, tapif, arg);
}
-
-/*
- * Our block (disk) device should be really simple: the Guest asks for a block
- * number and we read or write that position in the file. Unfortunately, that
- * was amazingly slow: the Guest waits until the read is finished before
- * running anything else, even if it could have been doing useful work.
- *
- * We could use async I/O, except it's reputed to suck so hard that characters
- * actually go missing from your code when you try to use it.
- *
- * So this was one reason why lguest now does all virtqueue servicing in
- * separate threads: it's more efficient and more like a real device.
- */
+/*:*/
/* This hangs off device->priv. */
struct vblk_info
@@ -1512,8 +1589,16 @@ struct vblk_info
/*L:210
* The Disk
*
- * Remember that the block device is handled by a separate I/O thread. We head
- * straight into the core of that thread here:
+ * The disk only has one virtqueue, so it only has one thread. It is really
+ * simple: the Guest asks for a block number and we read or write that position
+ * in the file.
+ *
+ * Before we serviced each virtqueue in a separate thread, that was unacceptably
+ * slow: the Guest waits until the read is finished before running anything
+ * else, even if it could have been doing useful work.
+ *
+ * We could have used async I/O, except it's reputed to suck so hard that
+ * characters actually go missing from your code when you try to use it.
*/
static void blk_request(struct virtqueue *vq)
{
@@ -1525,7 +1610,10 @@ static void blk_request(struct virtqueue *vq)
struct iovec iov[vq->vring.num];
off64_t off;
- /* Get the next request. */
+ /*
+ * Get the next request, where we normally wait. It triggers the
+ * interrupt to acknowledge previously serviced requests (if any).
+ */
head = wait_for_vq_desc(vq, iov, &out_num, &in_num);
/*
@@ -1539,6 +1627,10 @@ static void blk_request(struct virtqueue *vq)
out = convert(&iov[0], struct virtio_blk_outhdr);
in = convert(&iov[out_num+in_num-1], u8);
+ /*
+ * For historical reasons, block operations are expressed in 512 byte
+ * "sectors".
+ */
off = out->sector * 512;
/*
@@ -1614,6 +1706,7 @@ static void blk_request(struct virtqueue *vq)
if (out->type & VIRTIO_BLK_T_BARRIER)
fdatasync(vblk->fd);
+ /* Finished that request. */
add_used(vq, head, wlen);
}
@@ -1682,9 +1775,8 @@ static void rng_input(struct virtqueue *vq)
errx(1, "Output buffers in rng?");
/*
- * This is why we convert to iovecs: the readv() call uses them, and so
- * it reads straight into the Guest's buffer. We loop to make sure we
- * fill it.
+ * Just like the console write, we loop to cover the whole iovec.
+ * In this case, short reads actually happen quite a bit.
*/
while (!iov_empty(iov, in_num)) {
len = readv(rng_info->rfd, iov, in_num);
@@ -1818,7 +1910,9 @@ int main(int argc, char *argv[])
devices.lastdev = NULL;
devices.next_irq = 1;
+ /* We're CPU 0. In fact, that's the only CPU possible right now. */
cpu_id = 0;
+
/*
* We need to know how much memory so we can set up the device
* descriptor and memory pages for the devices as we parse the command
@@ -1926,7 +2020,7 @@ int main(int argc, char *argv[])
*/
tell_kernel(start);
- /* Ensure that we terminate if a child dies. */
+ /* Ensure that we terminate if a device-servicing child dies. */
signal(SIGCHLD, kill_launcher);
/* If we exit via err(), this kills all the threads, restores tty. */
diff --git a/arch/x86/include/asm/lguest_hcall.h b/arch/x86/include/asm/lguest_hcall.h
index cceb73e12e50..ba0eed8aa1a6 100644
--- a/arch/x86/include/asm/lguest_hcall.h
+++ b/arch/x86/include/asm/lguest_hcall.h
@@ -35,10 +35,10 @@
* operations? There are two ways: the direct way is to make a "hypercall",
* to make requests of the Host Itself.
*
- * We use the KVM hypercall mechanism. Seventeen hypercalls are
- * available: the hypercall number is put in the %eax register, and the
- * arguments (when required) are placed in %ebx, %ecx, %edx and %esi.
- * If a return value makes sense, it's returned in %eax.
+ * We use the KVM hypercall mechanism, though completely different hypercall
+ * numbers. Seventeen hypercalls are available: the hypercall number is put in
+ * the %eax register, and the arguments (when required) are placed in %ebx,
+ * %ecx, %edx and %esi. If a return value makes sense, it's returned in %eax.
*
* Grossly invalid calls result in Sudden Death at the hands of the vengeful
* Host, rather than returning failure. This reflects Winston Churchill's
diff --git a/arch/x86/lguest/boot.c b/arch/x86/lguest/boot.c
index 025c04d18f2b..d677fa9ca650 100644
--- a/arch/x86/lguest/boot.c
+++ b/arch/x86/lguest/boot.c
@@ -154,6 +154,7 @@ static void lazy_hcall1(unsigned long call,
async_hcall(call, arg1, 0, 0, 0);
}
+/* You can imagine what lazy_hcall2, 3 and 4 look like. :*/
static void lazy_hcall2(unsigned long call,
unsigned long arg1,
unsigned long arg2)
@@ -189,8 +190,10 @@ static void lazy_hcall4(unsigned long call,
}
#endif
-/* When lazy mode is turned off reset the per-cpu lazy mode variable and then
- * issue the do-nothing hypercall to flush any stored calls. */
+/*G:036
+ * When lazy mode is turned off reset the per-cpu lazy mode variable and then
+ * issue the do-nothing hypercall to flush any stored calls.
+:*/
static void lguest_leave_lazy_mmu_mode(void)
{
kvm_hypercall0(LHCALL_FLUSH_ASYNC);
@@ -250,13 +253,11 @@ extern void lg_irq_enable(void);
extern void lg_restore_fl(unsigned long flags);
/*M:003
- * Note that we don't check for outstanding interrupts when we re-enable them
- * (or when we unmask an interrupt). This seems to work for the moment, since
- * interrupts are rare and we'll just get the interrupt on the next timer tick,
- * but now we can run with CONFIG_NO_HZ, we should revisit this. One way would
- * be to put the "irq_enabled" field in a page by itself, and have the Host
- * write-protect it when an interrupt comes in when irqs are disabled. There
- * will then be a page fault as soon as interrupts are re-enabled.
+ * We could be more efficient in our checking of outstanding interrupts, rather
+ * than using a branch. One way would be to put the "irq_enabled" field in a
+ * page by itself, and have the Host write-protect it when an interrupt comes
+ * in when irqs are disabled. There will then be a page fault as soon as
+ * interrupts are re-enabled.
*
* A better method is to implement soft interrupt disable generally for x86:
* instead of disabling interrupts, we set a flag. If an interrupt does come
@@ -568,7 +569,7 @@ static void lguest_write_cr4(unsigned long val)
* cr3 ---> +---------+
* | --------->+---------+
* | | | PADDR1 |
- * Top-level | | PADDR2 |
+ * Mid-level | | PADDR2 |
* (PMD) page | | |
* | | Lower-level |
* | | (PTE) page |
@@ -588,23 +589,62 @@ static void lguest_write_cr4(unsigned long val)
* Index into top Index into second Offset within page
* page directory page pagetable page
*
- * The kernel spends a lot of time changing both the top-level page directory
- * and lower-level pagetable pages. The Guest doesn't know physical addresses,
- * so while it maintains these page tables exactly like normal, it also needs
- * to keep the Host informed whenever it makes a change: the Host will create
- * the real page tables based on the Guests'.
+ * Now, unfortunately, this isn't the whole story: Intel added Physical Address
+ * Extension (PAE) to allow 32 bit systems to use 64GB of memory (ie. 36 bits).
+ * These are held in 64-bit page table entries, so we can now only fit 512
+ * entries in a page, and the neat three-level tree breaks down.
+ *
+ * The result is a four level page table:
+ *
+ * cr3 --> [ 4 Upper ]
+ * [ Level ]
+ * [ Entries ]
+ * [(PUD Page)]---> +---------+
+ * | --------->+---------+
+ * | | | PADDR1 |
+ * Mid-level | | PADDR2 |
+ * (PMD) page | | |
+ * | | Lower-level |
+ * | | (PTE) page |
+ * | | | |
+ * .... ....
+ *
+ *
+ * And the virtual address is decoded as:
+ *
+ * 1 1 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
+ * |<-2->|<--- 9 bits ---->|<---- 9 bits --->|<------ 12 bits ------>|
+ * Index into Index into mid Index into lower Offset within page
+ * top entries directory page pagetable page
+ *
+ * It's too hard to switch between these two formats at runtime, so Linux only
+ * supports one or the other depending on whether CONFIG_X86_PAE is set. Many
+ * distributions turn it on, and not just for people with silly amounts of
+ * memory: the larger PTE entries allow room for the NX bit, which lets the
+ * kernel disable execution of pages and increase security.
+ *
+ * This was a problem for lguest, which couldn't run on these distributions;
+ * then Matias Zabaljauregui figured it all out and implemented it, and only a
+ * handful of puppies were crushed in the process!
+ *
+ * Back to our point: the kernel spends a lot of time changing both the
+ * top-level page directory and lower-level pagetable pages. The Guest doesn't
+ * know physical addresses, so while it maintains these page tables exactly
+ * like normal, it also needs to keep the Host informed whenever it makes a
+ * change: the Host will create the real page tables based on the Guests'.
*/
/*
- * The Guest calls this to set a second-level entry (pte), ie. to map a page
- * into a process' address space. We set the entry then tell the Host the
- * toplevel and address this corresponds to. The Guest uses one pagetable per
- * process, so we need to tell the Host which one we're changing (mm->pgd).
+ * The Guest calls this after it has set a second-level entry (pte), ie. to map
+ * a page into a process' address space. Wetell the Host the toplevel and
+ * address this corresponds to. The Guest uses one pagetable per process, so
+ * we need to tell the Host which one we're changing (mm->pgd).
*/
static void lguest_pte_update(struct mm_struct *mm, unsigned long addr,
pte_t *ptep)
{
#ifdef CONFIG_X86_PAE
+ /* PAE needs to hand a 64 bit page table entry, so it uses two args. */
lazy_hcall4(LHCALL_SET_PTE, __pa(mm->pgd), addr,
ptep->pte_low, ptep->pte_high);
#else
@@ -612,6 +652,7 @@ static void lguest_pte_update(struct mm_struct *mm, unsigned long addr,
#endif
}
+/* This is the "set and update" combo-meal-deal version. */
static void lguest_set_pte_at(struct mm_struct *mm, unsigned long addr,
pte_t *ptep, pte_t pteval)
{
@@ -672,6 +713,11 @@ static void lguest_set_pte(pte_t *ptep, pte_t pteval)
}
#ifdef CONFIG_X86_PAE
+/*
+ * With 64-bit PTE values, we need to be careful setting them: if we set 32
+ * bits at a time, the hardware could see a weird half-set entry. These
+ * versions ensure we update all 64 bits at once.
+ */
static void lguest_set_pte_atomic(pte_t *ptep, pte_t pte)
{
native_set_pte_atomic(ptep, pte);
@@ -679,13 +725,14 @@ static void lguest_set_pte_atomic(pte_t *ptep, pte_t pte)
lazy_hcall1(LHCALL_FLUSH_TLB, 1);
}
-void lguest_pte_clear(struct mm_struct *mm, unsigned long addr, pte_t *ptep)
+static void lguest_pte_clear(struct mm_struct *mm, unsigned long addr,
+ pte_t *ptep)
{
native_pte_clear(mm, addr, ptep);
lguest_pte_update(mm, addr, ptep);
}
-void lguest_pmd_clear(pmd_t *pmdp)
+static void lguest_pmd_clear(pmd_t *pmdp)
{
lguest_set_pmd(pmdp, __pmd(0));
}
@@ -784,6 +831,14 @@ static void __init lguest_init_IRQ(void)
irq_ctx_init(smp_processor_id());
}
+/*
+ * With CONFIG_SPARSE_IRQ, interrupt descriptors are allocated as-needed, so
+ * rather than set them in lguest_init_IRQ we are called here every time an
+ * lguest device needs an interrupt.
+ *
+ * FIXME: irq_to_desc_alloc_node() can fail due to lack of memory, we should
+ * pass that up!
+ */
void lguest_setup_irq(unsigned int irq)
{
irq_to_desc_alloc_node(irq, 0);
@@ -1298,7 +1353,7 @@ __init void lguest_init(void)
*/
switch_to_new_gdt(0);
- /* As described in head_32.S, we map the first 128M of memory. */
+ /* We actually boot with all memory mapped, but let's say 128MB. */
max_pfn_mapped = (128*1024*1024) >> PAGE_SHIFT;
/*
diff --git a/arch/x86/lguest/i386_head.S b/arch/x86/lguest/i386_head.S
index db6aa95eb054..27eac0faee48 100644
--- a/arch/x86/lguest/i386_head.S
+++ b/arch/x86/lguest/i386_head.S
@@ -102,6 +102,7 @@ send_interrupts:
* create one manually here.
*/
.byte 0x0f,0x01,0xc1 /* KVM_HYPERCALL */
+ /* Put eax back the way we found it. */
popl %eax
ret
@@ -125,6 +126,7 @@ ENTRY(lg_restore_fl)
jnz send_interrupts
/* Again, the normal path has used no extra registers. Clever, huh? */
ret
+/*:*/
/* These demark the EIP range where host should never deliver interrupts. */
.global lguest_noirq_start
diff --git a/drivers/lguest/core.c b/drivers/lguest/core.c
index cd058bc903ff..1e2cb846b3c9 100644
--- a/drivers/lguest/core.c
+++ b/drivers/lguest/core.c
@@ -217,10 +217,15 @@ int run_guest(struct lg_cpu *cpu, unsigned long __user *user)
/*
* It's possible the Guest did a NOTIFY hypercall to the
- * Launcher, in which case we return from the read() now.
+ * Launcher.
*/
if (cpu->pending_notify) {
+ /*
+ * Does it just needs to write to a registered
+ * eventfd (ie. the appropriate virtqueue thread)?
+ */
if (!send_notify_to_eventfd(cpu)) {
+ /* OK, we tell the main Laucher. */
if (put_user(cpu->pending_notify, user))
return -EFAULT;
return sizeof(cpu->pending_notify);
diff --git a/drivers/lguest/hypercalls.c b/drivers/lguest/hypercalls.c
index 787ab4bc09f0..83511eb0923d 100644
--- a/drivers/lguest/hypercalls.c
+++ b/drivers/lguest/hypercalls.c
@@ -59,7 +59,7 @@ static void do_hcall(struct lg_cpu *cpu, struct hcall_args *args)
case LHCALL_SHUTDOWN: {
char msg[128];
/*
- * Shutdown is such a trivial hypercall that we do it in four
+ * Shutdown is such a trivial hypercall that we do it in five
* lines right here.
*
* If the lgread fails, it will call kill_guest() itself; the
@@ -245,6 +245,10 @@ static void initialize(struct lg_cpu *cpu)
* device), the Guest will still see the old page. In practice, this never
* happens: why would the Guest read a page which it has never written to? But
* a similar scenario might one day bite us, so it's worth mentioning.
+ *
+ * Note that if we used a shared anonymous mapping in the Launcher instead of
+ * mapping /dev/zero private, we wouldn't worry about cop-on-write. And we
+ * need that to switch the Launcher to processes (away from threads) anyway.
:*/
/*H:100
diff --git a/drivers/lguest/lguest_device.c b/drivers/lguest/lguest_device.c
index cc000e79c3d1..1401c1ace1ec 100644
--- a/drivers/lguest/lguest_device.c
+++ b/drivers/lguest/lguest_device.c
@@ -236,7 +236,7 @@ static void lg_notify(struct virtqueue *vq)
extern void lguest_setup_irq(unsigned int irq);
/*
- * This routine finds the first virtqueue described in the configuration of
+ * This routine finds the Nth virtqueue described in the configuration of
* this device and sets it up.
*
* This is kind of an ugly duckling. It'd be nicer to have a standard
@@ -244,9 +244,6 @@ extern void lguest_setup_irq(unsigned int irq);
* everyone wants to do it differently. The KVM coders want the Guest to
* allocate its own pages and tell the Host where they are, but for lguest it's
* simpler for the Host to simply tell us where the pages are.
- *
- * So we provide drivers with a "find the Nth virtqueue and set it up"
- * function.
*/
static struct virtqueue *lg_find_vq(struct virtio_device *vdev,
unsigned index,
@@ -422,7 +419,11 @@ static void add_lguest_device(struct lguest_device_desc *d,
/* This devices' parent is the lguest/ dir. */
ldev->vdev.dev.parent = lguest_root;
- /* We have a unique device index thanks to the dev_index counter. */
+ /*
+ * The device type comes straight from the descriptor. There's also a
+ * device vendor field in the virtio_device struct, which we leave as
+ * 0.
+ */
ldev->vdev.id.device = d->type;
/*
* We have a simple set of routines for querying the device's
diff --git a/drivers/lguest/lguest_user.c b/drivers/lguest/lguest_user.c
index 7e92017103dc..b4d3f7ca554f 100644
--- a/drivers/lguest/lguest_user.c
+++ b/drivers/lguest/lguest_user.c
@@ -1,9 +1,8 @@
-/*P:200
- * This contains all the /dev/lguest code, whereby the userspace launcher
+/*P:200 This contains all the /dev/lguest code, whereby the userspace launcher
* controls and communicates with the Guest. For example, the first write will
- * tell us the Guest's memory layout, pagetable, entry point and kernel address
- * offset. A read will run the Guest until something happens, such as a signal
- * or the Guest doing a NOTIFY out to the Launcher.
+ * tell us the Guest's memory layout and entry point. A read will run the
+ * Guest until something happens, such as a signal or the Guest doing a NOTIFY
+ * out to the Launcher.
:*/
#include <linux/uaccess.h>
#include <linux/miscdevice.h>
@@ -13,14 +12,41 @@
#include <linux/file.h>
#include "lg.h"
+/*L:056
+ * Before we move on, let's jump ahead and look at what the kernel does when
+ * it needs to look up the eventfds. That will complete our picture of how we
+ * use RCU.
+ *
+ * The notification value is in cpu->pending_notify: we return true if it went
+ * to an eventfd.
+ */
bool send_notify_to_eventfd(struct lg_cpu *cpu)
{
unsigned int i;
struct lg_eventfd_map *map;
- /* lg->eventfds is RCU-protected */
+ /*
+ * This "rcu_read_lock()" helps track when someone is still looking at
+ * the (RCU-using) eventfds array. It's not actually a lock at all;
+ * indeed it's a noop in many configurations. (You didn't expect me to
+ * explain all the RCU secrets here, did you?)
+ */
rcu_read_lock();
+ /*
+ * rcu_dereference is the counter-side of rcu_assign_pointer(); it
+ * makes sure we don't access the memory pointed to by
+ * cpu->lg->eventfds before cpu->lg->eventfds is set. Sounds crazy,
+ * but Alpha allows this! Paul McKenney points out that a really
+ * aggressive compiler could have the same effect:
+ * http://lists.ozlabs.org/pipermail/lguest/2009-July/001560.html
+ *
+ * So play safe, use rcu_dereference to get the rcu-protected pointer:
+ */
map = rcu_dereference(cpu->lg->eventfds);
+ /*
+ * Simple array search: even if they add an eventfd while we do this,
+ * we'll continue to use the old array and just won't see the new one.
+ */
for (i = 0; i < map->num; i++) {
if (map->map[i].addr == cpu->pending_notify) {
eventfd_signal(map->map[i].event, 1);
@@ -28,14 +54,43 @@ bool send_notify_to_eventfd(struct lg_cpu *cpu)
break;
}
}
+ /* We're done with the rcu-protected variable cpu->lg->eventfds. */
rcu_read_unlock();
+
+ /* If we cleared the notification, it's because we found a match. */
return cpu->pending_notify == 0;
}
+/*L:055
+ * One of the more tricksy tricks in the Linux Kernel is a technique called
+ * Read Copy Update. Since one point of lguest is to teach lguest journeyers
+ * about kernel coding, I use it here. (In case you're curious, other purposes
+ * include learning about virtualization and instilling a deep appreciation for
+ * simplicity and puppies).
+ *
+ * We keep a simple array which maps LHCALL_NOTIFY values to eventfds, but we
+ * add new eventfds without ever blocking readers from accessing the array.
+ * The current Launcher only does this during boot, so that never happens. But
+ * Read Copy Update is cool, and adding a lock risks damaging even more puppies
+ * than this code does.
+ *
+ * We allocate a brand new one-larger array, copy the old one and add our new
+ * element. Then we make the lg eventfd pointer point to the new array.
+ * That's the easy part: now we need to free the old one, but we need to make
+ * sure no slow CPU somewhere is still looking at it. That's what
+ * synchronize_rcu does for us: waits until every CPU has indicated that it has
+ * moved on to know it's no longer using the old one.
+ *
+ * If that's unclear, see http://en.wikipedia.org/wiki/Read-copy-update.
+ */
static int add_eventfd(struct lguest *lg, unsigned long addr, int fd)
{
struct lg_eventfd_map *new, *old = lg->eventfds;
+ /*
+ * We don't allow notifications on value 0 anyway (pending_notify of
+ * 0 means "nothing pending").
+ */
if (!addr)
return -EINVAL;
@@ -62,12 +117,20 @@ static int add_eventfd(struct lguest *lg, unsigned long addr, int fd)
}
new->num++;
- /* Now put new one in place. */
+ /*
+ * Now put new one in place: rcu_assign_pointer() is a fancy way of
+ * doing "lg->eventfds = new", but it uses memory barriers to make
+ * absolutely sure that the contents of "new" written above is nailed
+ * down before we actually do the assignment.
+ *
+ * We have to think about these kinds of things when we're operating on
+ * live data without locks.
+ */
rcu_assign_pointer(lg->eventfds, new);
/*
* We're not in a big hurry. Wait until noone's looking at old
- * version, then delete it.
+ * version, then free it.
*/
synchronize_rcu();
kfree(old);
@@ -75,6 +138,14 @@ static int add_eventfd(struct lguest *lg, unsigned long addr, int fd)
return 0;
}
+/*L:052
+ * Receiving notifications from the Guest is usually done by attaching a
+ * particular LHCALL_NOTIFY value to an event filedescriptor. The eventfd will
+ * become readable when the Guest does an LHCALL_NOTIFY with that value.
+ *
+ * This is really convenient for processing each virtqueue in a separate
+ * thread.
+ */
static int attach_eventfd(struct lguest *lg, const unsigned long __user *input)
{
unsigned long addr, fd;
@@ -86,6 +157,11 @@ static int attach_eventfd(struct lguest *lg, const unsigned long __user *input)
if (get_user(fd, input) != 0)
return -EFAULT;
+ /*
+ * Just make sure two callers don't add eventfds at once. We really
+ * only need to lock against callers adding to the same Guest, so using
+ * the Big Lguest Lock is overkill. But this is setup, not a fast path.
+ */
mutex_lock(&lguest_lock);
err = add_eventfd(lg, addr, fd);
mutex_unlock(&lguest_lock);
@@ -106,6 +182,10 @@ static int user_send_irq(struct lg_cpu *cpu, const unsigned long __user *input)
if (irq >= LGUEST_IRQS)
return -EINVAL;
+ /*
+ * Next time the Guest runs, the core code will see if it can deliver
+ * this interrupt.
+ */
set_interrupt(cpu, irq);
return 0;
}
@@ -307,10 +387,10 @@ unlock:
* The first operation the Launcher does must be a write. All writes
* start with an unsigned long number: for the first write this must be
* LHREQ_INITIALIZE to set up the Guest. After that the Launcher can use
- * writes of other values to send interrupts.
+ * writes of other values to send interrupts or set up receipt of notifications.
*
* Note that we overload the "offset" in the /dev/lguest file to indicate what
- * CPU number we're dealing with. Currently this is always 0, since we only
+ * CPU number we're dealing with. Currently this is always 0 since we only
* support uniprocessor Guests, but you can see the beginnings of SMP support
* here.
*/
diff --git a/drivers/lguest/page_tables.c b/drivers/lguest/page_tables.c
index 3da902e4b4cb..a8d0aee3bc0e 100644
--- a/drivers/lguest/page_tables.c
+++ b/drivers/lguest/page_tables.c
@@ -29,10 +29,10 @@
/*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
- * I recommend you review arch/x86/lguest/boot.c's "Page Table Handling" (with
- * diagrams!).
+ * We use two-level page tables for the Guest, or three-level with PAE. If
+ * you're not entirely comfortable with virtual addresses, physical addresses
+ * and page tables then I recommend you review arch/x86/lguest/boot.c's "Page
+ * Table Handling" (with diagrams!).
*
* 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
@@ -52,9 +52,8 @@
:*/
/*
- * 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.
+ * The Switcher uses the complete top PTE page. That's 1024 PTE entries (4MB)
+ * or 512 PTE entries with PAE (2MB).
*/
#define SWITCHER_PGD_INDEX (PTRS_PER_PGD - 1)
@@ -81,7 +80,8 @@ static DEFINE_PER_CPU(pte_t *, switcher_pte_pages);
/*H:320
* The page table code is curly enough to need helper functions to keep it
- * clear and clean.
+ * clear and clean. The kernel itself provides many of them; one advantage
+ * of insisting that the Guest and Host use the same CONFIG_PAE setting.
*
* There are two functions which return pointers to the shadow (aka "real")
* page tables.
@@ -155,7 +155,7 @@ static pte_t *spte_addr(struct lg_cpu *cpu, pgd_t spgd, unsigned long vaddr)
}
/*
- * These two functions just like the above two, except they access the Guest
+ * These functions are just like the above two, except they access the Guest
* page tables. Hence they return a Guest address.
*/
static unsigned long gpgd_addr(struct lg_cpu *cpu, unsigned long vaddr)
@@ -165,6 +165,7 @@ static unsigned long gpgd_addr(struct lg_cpu *cpu, unsigned long vaddr)
}
#ifdef CONFIG_X86_PAE
+/* Follow the PGD to the PMD. */
static unsigned long gpmd_addr(pgd_t gpgd, unsigned long vaddr)
{
unsigned long gpage = pgd_pfn(gpgd) << PAGE_SHIFT;
@@ -172,6 +173,7 @@ static unsigned long gpmd_addr(pgd_t gpgd, unsigned long vaddr)
return gpage + pmd_index(vaddr) * sizeof(pmd_t);
}
+/* Follow the PMD to the PTE. */
static unsigned long gpte_addr(struct lg_cpu *cpu,
pmd_t gpmd, unsigned long vaddr)
{
@@ -181,6 +183,7 @@ static unsigned long gpte_addr(struct lg_cpu *cpu,
return gpage + pte_index(vaddr) * sizeof(pte_t);
}
#else
+/* Follow the PGD to the PTE (no mid-level for !PAE). */
static unsigned long gpte_addr(struct lg_cpu *cpu,
pgd_t gpgd, unsigned long vaddr)
{
@@ -314,6 +317,7 @@ bool demand_page(struct lg_cpu *cpu, unsigned long vaddr, int errcode)
pte_t gpte;
pte_t *spte;
+ /* Mid level for PAE. */
#ifdef CONFIG_X86_PAE
pmd_t *spmd;
pmd_t gpmd;
@@ -391,6 +395,8 @@ bool demand_page(struct lg_cpu *cpu, unsigned long vaddr, int errcode)
*/
gpte_ptr = gpte_addr(cpu, gpgd, vaddr);
#endif
+
+ /* Read the actual PTE value. */
gpte = lgread(cpu, gpte_ptr, pte_t);
/* If this page isn't in the Guest page tables, we can't page it in. */
@@ -507,6 +513,7 @@ void pin_page(struct lg_cpu *cpu, unsigned long vaddr)
if (!page_writable(cpu, vaddr) && !demand_page(cpu, vaddr, 2))
kill_guest(cpu, "bad stack page %#lx", vaddr);
}
+/*:*/
#ifdef CONFIG_X86_PAE
static void release_pmd(pmd_t *spmd)
@@ -543,7 +550,11 @@ static void release_pgd(pgd_t *spgd)
}
#else /* !CONFIG_X86_PAE */
-/*H:450 If we chase down the release_pgd() code, it looks like this: */
+/*H:450
+ * If we chase down the release_pgd() code, the non-PAE version looks like
+ * this. The PAE version is almost identical, but instead of calling
+ * release_pte it calls release_pmd(), which looks much like this.
+ */
static void release_pgd(pgd_t *spgd)
{
/* If the entry's not present, there's nothing to release. */
@@ -898,17 +909,21 @@ void guest_set_pgd(struct lguest *lg, unsigned long gpgdir, u32 idx)
/* ... throw it away. */
release_pgd(lg->pgdirs[pgdir].pgdir + idx);
}
+
#ifdef CONFIG_X86_PAE
+/* For setting a mid-level, we just throw everything away. It's easy. */
void guest_set_pmd(struct lguest *lg, unsigned long pmdp, u32 idx)
{
guest_pagetable_clear_all(&lg->cpus[0]);
}
#endif
-/*
- * Once we know how much memory we have we can construct simple identity (which
+/*H:505
+ * To get through boot, we construct simple identity page mappings (which
* set virtual == physical) and linear mappings which will get the Guest far
- * enough into the boot to create its own.
+ * enough into the boot to create its own. The linear mapping means we
+ * simplify the Guest boot, but it makes assumptions about their PAGE_OFFSET,
+ * as you'll see.
*
* We lay them out of the way, just below the initrd (which is why we need to
* know its size here).
@@ -944,6 +959,10 @@ static unsigned long setup_pagetables(struct lguest *lg,
linear = (void *)pgdir - linear_pages * PAGE_SIZE;
#ifdef CONFIG_X86_PAE
+ /*
+ * And the single mid page goes below that. We only use one, but
+ * that's enough to map 1G, which definitely gets us through boot.
+ */
pmds = (void *)linear - PAGE_SIZE;
#endif
/*
@@ -957,13 +976,14 @@ static unsigned long setup_pagetables(struct lguest *lg,
return -EFAULT;
}
+#ifdef CONFIG_X86_PAE
/*
- * The top level points to the linear page table pages above.
- * We setup the identity and linear mappings here.
+ * Make the Guest PMD entries point to the corresponding place in the
+ * linear mapping (up to one page worth of PMD).
*/
-#ifdef CONFIG_X86_PAE
for (i = j = 0; i < mapped_pages && j < PTRS_PER_PMD;
i += PTRS_PER_PTE, j++) {
+ /* FIXME: native_set_pmd is overkill here. */
native_set_pmd(&pmd, __pmd(((unsigned long)(linear + i)
- mem_base) | _PAGE_PRESENT | _PAGE_RW | _PAGE_USER));
@@ -971,18 +991,36 @@ static unsigned long setup_pagetables(struct lguest *lg,
return -EFAULT;
}
+ /* One PGD entry, pointing to that PMD page. */
set_pgd(&pgd, __pgd(((u32)pmds - mem_base) | _PAGE_PRESENT));
+ /* Copy it in as the first PGD entry (ie. addresses 0-1G). */
if (copy_to_user(&pgdir[0], &pgd, sizeof(pgd)) != 0)
return -EFAULT;
+ /*
+ * And the third PGD entry (ie. addresses 3G-4G).
+ *
+ * FIXME: This assumes that PAGE_OFFSET for the Guest is 0xC0000000.
+ */
if (copy_to_user(&pgdir[3], &pgd, sizeof(pgd)) != 0)
return -EFAULT;
#else
+ /*
+ * The top level points to the linear page table pages above.
+ * We setup the identity and linear mappings here.
+ */
phys_linear = (unsigned long)linear - mem_base;
for (i = 0; i < mapped_pages; i += PTRS_PER_PTE) {
pgd_t pgd;
+ /*
+ * Create a PGD entry which points to the right part of the
+ * linear PTE pages.
+ */
pgd = __pgd((phys_linear + i * sizeof(pte_t)) |
(_PAGE_PRESENT | _PAGE_RW | _PAGE_USER));
+ /*
+ * Copy it into the PGD page at 0 and PAGE_OFFSET.
+ */
if (copy_to_user(&pgdir[i / PTRS_PER_PTE], &pgd, sizeof(pgd))
|| copy_to_user(&pgdir[pgd_index(PAGE_OFFSET)
+ i / PTRS_PER_PTE],
@@ -992,8 +1030,8 @@ static unsigned long setup_pagetables(struct lguest *lg,
#endif
/*
- * We return the top level (guest-physical) address: remember where
- * this is.
+ * We return the top level (guest-physical) address: we remember where
+ * this is to write it into lguest_data when the Guest initializes.
*/
return (unsigned long)pgdir - mem_base;
}
@@ -1031,7 +1069,9 @@ int init_guest_pagetable(struct lguest *lg)
lg->pgdirs[0].pgdir = (pgd_t *)get_zeroed_page(GFP_KERNEL);
if (!lg->pgdirs[0].pgdir)
return -ENOMEM;
+
#ifdef CONFIG_X86_PAE
+ /* For PAE, we also create the initial mid-level. */
pgd = lg->pgdirs[0].pgdir;
pmd_table = (pmd_t *) get_zeroed_page(GFP_KERNEL);
if (!pmd_table)
@@ -1040,11 +1080,13 @@ int init_guest_pagetable(struct lguest *lg)
set_pgd(pgd + SWITCHER_PGD_INDEX,
__pgd(__pa(pmd_table) | _PAGE_PRESENT));
#endif
+
+ /* This is the current page table. */
lg->cpus[0].cpu_pgd = 0;
return 0;
}
-/* When the Guest calls LHCALL_LGUEST_INIT we do more setup. */
+/*H:508 When the Guest calls LHCALL_LGUEST_INIT we do more setup. */
void page_table_guest_data_init(struct lg_cpu *cpu)
{
/* We get the kernel address: above this is all kernel memory. */
@@ -1105,12 +1147,16 @@ void map_switcher_in_guest(struct lg_cpu *cpu, struct lguest_pages *pages)
pmd_t switcher_pmd;
pmd_t *pmd_table;
+ /* FIXME: native_set_pmd is overkill here. */
native_set_pmd(&switcher_pmd, pfn_pmd(__pa(switcher_pte_page) >>
PAGE_SHIFT, PAGE_KERNEL_EXEC));
+ /* Figure out where the pmd page is, by reading the PGD, and converting
+ * it to a virtual address. */
pmd_table = __va(pgd_pfn(cpu->lg->
pgdirs[cpu->cpu_pgd].pgdir[SWITCHER_PGD_INDEX])
<< PAGE_SHIFT);
+ /* Now write it into the shadow page table. */
native_set_pmd(&pmd_table[SWITCHER_PMD_INDEX], switcher_pmd);
#else
pgd_t switcher_pgd;
diff --git a/drivers/lguest/x86/core.c b/drivers/lguest/x86/core.c
index 96f7d88ec7f8..6ae388849a3b 100644
--- a/drivers/lguest/x86/core.c
+++ b/drivers/lguest/x86/core.c
@@ -187,7 +187,7 @@ static void run_guest_once(struct lg_cpu *cpu, struct lguest_pages *pages)
* also simplify copy_in_guest_info(). Note that we'd still need to restore
* things when we exit to Launcher userspace, but that's fairly easy.
*
- * We could also try using this hooks for PGE, but that might be too expensive.
+ * We could also try using these hooks for PGE, but that might be too expensive.
*
* The hooks were designed for KVM, but we can also put them to good use.
:*/
diff --git a/drivers/lguest/x86/switcher_32.S b/drivers/lguest/x86/switcher_32.S
index 6dec09793836..40634b0db9f7 100644
--- a/drivers/lguest/x86/switcher_32.S
+++ b/drivers/lguest/x86/switcher_32.S
@@ -1,7 +1,7 @@
/*P:900
- * This is the Switcher: code which sits at 0xFFC00000 astride both the
- * Host and Guest to do the low-level Guest<->Host switch. It is as simple as
- * it can be made, but it's naturally very specific to x86.
+ * This is the Switcher: code which sits at 0xFFC00000 (or 0xFFE00000) astride
+ * both the Host and Guest to do the low-level Guest<->Host switch. It is as
+ * simple as it can be made, but it's naturally very specific to x86.
*
* You have now completed Preparation. If this has whet your appetite; if you
* are feeling invigorated and refreshed then the next, more challenging stage