/*P:100 This is the Launcher code, a simple program which lays out the
- * "physical" memory for the new Guest by mapping the kernel image and the
- * virtual devices, then reads repeatedly from /dev/lguest to run the Guest.
-:*/
+ * "physical" memory for the new Guest by mapping the kernel image and
+ * the virtual devices, then opens /dev/lguest to tell the kernel
+ * about the Guest and control it. :*/
#define _LARGEFILE64_SOURCE
#define _GNU_SOURCE
#include <stdio.h>
#include <zlib.h>
#include <assert.h>
#include <sched.h>
-/*L:110 We can ignore the 30 include files we need for this program, but I do
+#include <limits.h>
+#include <stddef.h>
+#include <signal.h>
+#include "linux/lguest_launcher.h"
+#include "linux/virtio_config.h"
+#include "linux/virtio_net.h"
+#include "linux/virtio_blk.h"
+#include "linux/virtio_console.h"
+#include "linux/virtio_rng.h"
+#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 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
- * like these abbreviations and the header we need uses them, so we define them
- * here.
- */
+ * like these abbreviations, so we define them here. Note that u64 is always
+ * unsigned long long, which works on all Linux systems: this means that we can
+ * use %llu in printf for any u64. */
typedef unsigned long long u64;
typedef uint32_t u32;
typedef uint16_t u16;
typedef uint8_t u8;
-#include "linux/lguest_launcher.h"
-#include "linux/pci_ids.h"
-#include "linux/virtio_config.h"
-#include "linux/virtio_net.h"
-#include "linux/virtio_blk.h"
-#include "linux/virtio_console.h"
-#include "linux/virtio_ring.h"
-#include "asm-x86/e820.h"
/*:*/
#define PAGE_PRESENT 0x7 /* Present, RW, Execute */
#endif
/* We can have up to 256 pages for devices. */
#define DEVICE_PAGES 256
-/* This fits nicely in a single 4096-byte page. */
-#define VIRTQUEUE_NUM 127
+/* This will occupy 3 pages: it must be a power of 2. */
+#define VIRTQUEUE_NUM 256
/*L:120 verbose is both a global flag and a macro. The C preprocessor allows
* this, and although I wouldn't recommend it, it works quite nicely here. */
do { if (verbose) printf(args); } while(0)
/*:*/
-/* The pipe to send commands to the waker process */
-static int waker_fd;
+/* File descriptors for the Waker. */
+struct {
+ int pipe[2];
+ int lguest_fd;
+} waker_fds;
+
/* The pointer to the start of guest memory. */
static void *guest_base;
/* The maximum guest physical address allowed, and maximum possible. */
static unsigned long guest_limit, guest_max;
+/* The pipe for signal hander to write to. */
+static int timeoutpipe[2];
+static unsigned int timeout_usec = 500;
+
+/* a per-cpu variable indicating whose vcpu is currently running */
+static unsigned int __thread cpu_id;
/* This is our list of devices. */
struct device_list
/* The descriptor page for the devices. */
u8 *descpage;
- /* The tail of the last descriptor. */
- unsigned int desc_used;
-
/* A single linked list of devices. */
struct device *dev;
- /* ... And an end pointer so we can easily append new devices */
- struct device **lastdev;
+ /* And a pointer to the last device for easy append and also for
+ * configuration appending. */
+ struct device *lastdev;
};
/* The list of Guest devices, based on command line arguments. */
/* Any queues attached to this device */
struct virtqueue *vq;
+ /* Handle status being finalized (ie. feature bits stable). */
+ void (*ready)(struct device *me);
+
/* Device-specific data. */
void *priv;
};
/* Last available index we saw. */
u16 last_avail_idx;
- /* The routine to call when the Guest pings us. */
- void (*handle_output)(int fd, struct virtqueue *me);
+ /* The routine to call when the Guest pings us, or timeout. */
+ void (*handle_output)(int fd, struct virtqueue *me, bool timeout);
+
+ /* Outstanding buffers */
+ unsigned int inflight;
+
+ /* Is this blocked awaiting a timer? */
+ bool blocked;
};
+/* Remember the arguments to the program so we can "reboot" */
+static char **main_args;
+
/* Since guest is UP and we don't run at the same time, we don't need barriers.
* But I include them in the code in case others copy it. */
#define wmb()
return iov->iov_base;
}
+/* Wrapper for the last available index. Makes it easier to change. */
+#define lg_last_avail(vq) ((vq)->last_avail_idx)
+
/* The virtio configuration space is defined to be little-endian. x86 is
* little-endian too, but it's nice to be explicit so we have these helpers. */
#define cpu_to_le16(v16) (v16)
#define cpu_to_le64(v64) (v64)
#define le16_to_cpu(v16) (v16)
#define le32_to_cpu(v32) (v32)
-#define le64_to_cpu(v32) (v64)
+#define le64_to_cpu(v64) (v64)
+
+/* Is this iovec empty? */
+static bool iov_empty(const struct iovec iov[], unsigned int num_iov)
+{
+ unsigned int i;
+
+ for (i = 0; i < num_iov; i++)
+ if (iov[i].iov_len)
+ return false;
+ return true;
+}
+
+/* Take len bytes from the front of this iovec. */
+static void iov_consume(struct iovec iov[], unsigned num_iov, unsigned len)
+{
+ unsigned int i;
+
+ for (i = 0; i < num_iov; i++) {
+ unsigned int used;
+
+ used = iov[i].iov_len < len ? iov[i].iov_len : len;
+ iov[i].iov_base += used;
+ iov[i].iov_len -= used;
+ len -= used;
+ }
+ assert(len == 0);
+}
+
+/* The device virtqueue descriptors are followed by feature bitmasks. */
+static u8 *get_feature_bits(struct device *dev)
+{
+ return (u8 *)(dev->desc + 1)
+ + dev->desc->num_vq * sizeof(struct lguest_vqconfig);
+}
/*L:100 The Launcher code itself takes us out into userspace, that scary place
* where pointers run wild and free! Unfortunately, like most userspace
PROT_READ|PROT_WRITE|PROT_EXEC, MAP_PRIVATE, fd, 0);
if (addr == MAP_FAILED)
err(1, "Mmaping %u pages of /dev/zero", num);
+ close(fd);
return addr;
}
err(1, "Reading program headers");
/* Try all the headers: there are usually only three. A read-only one,
- * a read-write one, and a "note" section which isn't loadable. */
+ * a read-write one, and a "note" section which we don't load. */
for (i = 0; i < ehdr->e_phnum; i++) {
/* If this isn't a loadable segment, we ignore it */
if (phdr[i].p_type != PT_LOAD)
return ehdr->e_entry;
}
-/*L:160 Unfortunately the entire ELF image isn't compressed: the segments
- * which need loading are extracted and compressed raw. This denies us the
- * information we need to make a fully-general loader. */
-static unsigned long unpack_bzimage(int fd)
+/*L:150 A bzImage, unlike an ELF file, is not meant to be loaded. You're
+ * supposed to jump into it and it will unpack itself. We used to have to
+ * perform some hairy magic because the unpacking code scared me.
+ *
+ * Fortunately, Jeremy Fitzhardinge convinced me it wasn't that hard and wrote
+ * a small patch to jump over the tricky bits in the Guest, so now we just read
+ * the funky header so we know where in the file to load, and away we go! */
+static unsigned long load_bzimage(int fd)
{
- gzFile f;
- int ret, len = 0;
- /* A bzImage always gets loaded at physical address 1M. This is
- * actually configurable as CONFIG_PHYSICAL_START, but as the comment
- * there says, "Don't change this unless you know what you are doing".
- * Indeed. */
- void *img = from_guest_phys(0x100000);
+ struct boot_params boot;
+ int r;
+ /* Modern bzImages get loaded at 1M. */
+ void *p = from_guest_phys(0x100000);
- /* gzdopen takes our file descriptor (carefully placed at the start of
- * the GZIP header we found) and returns a gzFile. */
- f = gzdopen(fd, "rb");
- /* We read it into memory in 64k chunks until we hit the end. */
- while ((ret = gzread(f, img + len, 65536)) > 0)
- len += ret;
- if (ret < 0)
- err(1, "reading image from bzImage");
+ /* Go back to the start of the file and read the header. It should be
+ * a Linux boot header (see Documentation/x86/i386/boot.txt) */
+ lseek(fd, 0, SEEK_SET);
+ read(fd, &boot, sizeof(boot));
- verbose("Unpacked size %i addr %p\n", len, img);
+ /* Inside the setup_hdr, we expect the magic "HdrS" */
+ if (memcmp(&boot.hdr.header, "HdrS", 4) != 0)
+ errx(1, "This doesn't look like a bzImage to me");
- /* The entry point for a bzImage is always the first byte */
- return (unsigned long)img;
-}
+ /* Skip over the extra sectors of the header. */
+ lseek(fd, (boot.hdr.setup_sects+1) * 512, SEEK_SET);
-/*L:150 A bzImage, unlike an ELF file, is not meant to be loaded. You're
- * supposed to jump into it and it will unpack itself. We can't do that
- * because the Guest can't run the unpacking code, and adding features to
- * lguest kills puppies, so we don't want to.
- *
- * The bzImage is formed by putting the decompressing code in front of the
- * compressed kernel code. So we can simple scan through it looking for the
- * first "gzip" header, and start decompressing from there. */
-static unsigned long load_bzimage(int fd)
-{
- unsigned char c;
- int state = 0;
+ /* Now read everything into memory. in nice big chunks. */
+ while ((r = read(fd, p, 65536)) > 0)
+ p += r;
- /* GZIP header is 0x1F 0x8B <method> <flags>... <compressed-by>. */
- while (read(fd, &c, 1) == 1) {
- switch (state) {
- case 0:
- if (c == 0x1F)
- state++;
- break;
- case 1:
- if (c == 0x8B)
- state++;
- else
- state = 0;
- break;
- case 2 ... 8:
- state++;
- break;
- case 9:
- /* Seek back to the start of the gzip header. */
- lseek(fd, -10, SEEK_CUR);
- /* One final check: "compressed under UNIX". */
- if (c != 0x03)
- state = -1;
- else
- return unpack_bzimage(fd);
- }
- }
- errx(1, "Could not find kernel in bzImage");
+ /* Finally, code32_start tells us where to enter the kernel. */
+ return boot.hdr.code32_start;
}
/*L:140 Loading the kernel is easy when it's a "vmlinux", but most kernels
- * come wrapped up in the self-decompressing "bzImage" format. With some funky
- * coding, we can load those, too. */
+ * come wrapped up in the self-decompressing "bzImage" format. With a little
+ * work, we can load those, too. */
static unsigned long load_kernel(int fd)
{
Elf32_Ehdr hdr;
if (memcmp(hdr.e_ident, ELFMAG, SELFMAG) == 0)
return map_elf(fd, &hdr);
- /* Otherwise we assume it's a bzImage, and try to unpack it */
+ /* Otherwise we assume it's a bzImage, and try to load it. */
return load_bzimage(fd);
}
/* We return the initrd size. */
return len;
}
-
-/* Once we know how much memory we have, we can construct simple linear page
- * tables which set virtual == physical which will get the Guest far enough
- * into the boot to create its own.
- *
- * We lay them out of the way, just below the initrd (which is why we need to
- * know its size). */
-static unsigned long setup_pagetables(unsigned long mem,
- unsigned long initrd_size)
-{
- unsigned long *pgdir, *linear;
- unsigned int mapped_pages, i, linear_pages;
- unsigned int ptes_per_page = getpagesize()/sizeof(void *);
-
- mapped_pages = mem/getpagesize();
-
- /* Each PTE page can map ptes_per_page pages: how many do we need? */
- linear_pages = (mapped_pages + ptes_per_page-1)/ptes_per_page;
-
- /* We put the toplevel page directory page at the top of memory. */
- pgdir = from_guest_phys(mem) - initrd_size - getpagesize();
-
- /* Now we use the next linear_pages pages as pte pages */
- linear = (void *)pgdir - linear_pages*getpagesize();
-
- /* Linear mapping is easy: put every page's address into the mapping in
- * order. PAGE_PRESENT contains the flags Present, Writable and
- * Executable. */
- for (i = 0; i < mapped_pages; i++)
- linear[i] = ((i * getpagesize()) | PAGE_PRESENT);
-
- /* The top level points to the linear page table pages above. */
- for (i = 0; i < mapped_pages; i += ptes_per_page) {
- pgdir[i/ptes_per_page]
- = ((to_guest_phys(linear) + i*sizeof(void *))
- | PAGE_PRESENT);
- }
-
- verbose("Linear mapping of %u pages in %u pte pages at %#lx\n",
- mapped_pages, linear_pages, to_guest_phys(linear));
-
- /* We return the top level (guest-physical) address: the kernel needs
- * to know where it is. */
- return to_guest_phys(pgdir);
-}
+/*:*/
/* Simple routine to roll all the commandline arguments together with spaces
* between them. */
unsigned int i, len = 0;
for (i = 0; args[i]; i++) {
+ if (i) {
+ strcat(dst+len, " ");
+ len++;
+ }
strcpy(dst+len, args[i]);
- strcat(dst+len, " ");
- len += strlen(args[i]) + 1;
+ len += strlen(args[i]);
}
/* In case it's empty. */
dst[len] = '\0';
}
-/* This is where we actually tell the kernel to initialize the Guest. We saw
- * the arguments it expects when we looked at initialize() in lguest_user.c:
- * the base of guest "physical" memory, the top physical page to allow, the
- * top level pagetable and the entry point for the Guest. */
-static int tell_kernel(unsigned long pgdir, unsigned long start)
+/*L:185 This is where we actually tell the kernel to initialize the Guest. We
+ * saw the arguments it expects when we looked at initialize() in lguest_user.c:
+ * the base of Guest "physical" memory, the top physical page to allow and the
+ * entry point for the Guest. */
+static int tell_kernel(unsigned long start)
{
unsigned long args[] = { LHREQ_INITIALIZE,
(unsigned long)guest_base,
- guest_limit / getpagesize(), pgdir, start };
+ guest_limit / getpagesize(), start };
int fd;
verbose("Guest: %p - %p (%#lx)\n",
/*L:200
* The Waker.
*
- * With a console and network devices, we can have lots of input which we need
- * to process. We could try to tell the kernel what file descriptors to watch,
- * but handing a file descriptor mask through to the kernel is fairly icky.
+ * With console, block and network devices, we can have lots of input which we
+ * need to process. We could try to tell the kernel what file descriptors to
+ * watch, but handing a file descriptor mask through to the kernel is fairly
+ * icky.
*
- * Instead, we fork off a process which watches the file descriptors and writes
- * the LHREQ_BREAK command to the /dev/lguest filedescriptor to tell the Host
- * loop to stop running the Guest. This causes it to return from the
+ * Instead, we clone off a thread which watches the file descriptors and writes
+ * the LHREQ_BREAK command to the /dev/lguest file descriptor to tell the Host
+ * stop running the Guest. This causes the Launcher to return from the
* /dev/lguest read with -EAGAIN, where it will write to /dev/lguest to reset
* the LHREQ_BREAK and wake us up again.
*
* This, of course, is merely a different *kind* of icky.
+ *
+ * Given my well-known antipathy to threads, I'd prefer to use processes. But
+ * it's easier to share Guest memory with threads, and trivial to share the
+ * devices.infds as the Launcher changes it.
*/
-static void wake_parent(int pipefd, int lguest_fd)
+static int waker(void *unused)
{
- /* Add the pipe from the Launcher to the fdset in the device_list, so
- * we watch it, too. */
- add_device_fd(pipefd);
+ /* Close the write end of the pipe: only the Launcher has it open. */
+ close(waker_fds.pipe[1]);
for (;;) {
fd_set rfds = devices.infds;
unsigned long args[] = { LHREQ_BREAK, 1 };
+ unsigned int maxfd = devices.max_infd;
+
+ /* We also listen to the pipe from the Launcher. */
+ FD_SET(waker_fds.pipe[0], &rfds);
+ if (waker_fds.pipe[0] > maxfd)
+ maxfd = waker_fds.pipe[0];
/* Wait until input is ready from one of the devices. */
- select(devices.max_infd+1, &rfds, NULL, NULL, NULL);
- /* Is it a message from the Launcher? */
- if (FD_ISSET(pipefd, &rfds)) {
- int fd;
- /* If read() returns 0, it means the Launcher has
- * exited. We silently follow. */
- if (read(pipefd, &fd, sizeof(fd)) == 0)
- exit(0);
- /* Otherwise it's telling us to change what file
- * descriptors we're to listen to. */
- if (fd >= 0)
- FD_SET(fd, &devices.infds);
- else
- FD_CLR(-fd - 1, &devices.infds);
- } else /* Send LHREQ_BREAK command. */
- write(lguest_fd, args, sizeof(args));
+ select(maxfd+1, &rfds, NULL, NULL, NULL);
+
+ /* Message from Launcher? */
+ if (FD_ISSET(waker_fds.pipe[0], &rfds)) {
+ char c;
+ /* If this fails, then assume Launcher has exited.
+ * Don't do anything on exit: we're just a thread! */
+ if (read(waker_fds.pipe[0], &c, 1) != 1)
+ _exit(0);
+ continue;
+ }
+
+ /* Send LHREQ_BREAK command to snap the Launcher out of it. */
+ pwrite(waker_fds.lguest_fd, args, sizeof(args), cpu_id);
}
+ return 0;
}
/* This routine just sets up a pipe to the Waker process. */
-static int setup_waker(int lguest_fd)
+static void setup_waker(int lguest_fd)
{
- int pipefd[2], child;
+ /* This pipe is closed when Launcher dies, telling Waker. */
+ if (pipe(waker_fds.pipe) != 0)
+ err(1, "Creating pipe for Waker");
- /* We create a pipe to talk to the waker, and also so it knows when the
- * Launcher dies (and closes pipe). */
- pipe(pipefd);
- child = fork();
- if (child == -1)
- err(1, "forking");
-
- if (child == 0) {
- /* Close the "writing" end of our copy of the pipe */
- close(pipefd[1]);
- wake_parent(pipefd[0], lguest_fd);
- }
- /* Close the reading end of our copy of the pipe. */
- close(pipefd[0]);
+ /* Waker also needs to know the lguest fd */
+ waker_fds.lguest_fd = lguest_fd;
- /* Here is the fd used to talk to the waker. */
- return pipefd[1];
+ if (clone(waker, malloc(4096) + 4096, CLONE_VM | SIGCHLD, NULL) == -1)
+ err(1, "Creating Waker");
}
-/*L:210
+/*
* Device Handling.
*
- * When the Guest sends DMA to us, it sends us an array of addresses and sizes.
+ * When the Guest gives us a buffer, it sends an array of addresses and sizes.
* We need to make sure it's not trying to reach into the Launcher itself, so
- * we have a convenient routine which check it and exits with an error message
+ * we have a convenient routine which checks it and exits with an error message
* if something funny is going on:
*/
static void *_check_pointer(unsigned long addr, unsigned int size,
/* A macro which transparently hands the line number to the real function. */
#define check_pointer(addr,size) _check_pointer(addr, size, __LINE__)
-/* This function returns the next descriptor in the chain, or vq->vring.num. */
+/* Each buffer in the virtqueues is actually a chain of descriptors. This
+ * function returns the next descriptor in the chain, or vq->vring.num if we're
+ * at the end. */
static unsigned next_desc(struct virtqueue *vq, unsigned int i)
{
unsigned int next;
unsigned int *out_num, unsigned int *in_num)
{
unsigned int i, head;
+ u16 last_avail;
/* Check it isn't doing very strange things with descriptor numbers. */
- if ((u16)(vq->vring.avail->idx - vq->last_avail_idx) > vq->vring.num)
+ last_avail = lg_last_avail(vq);
+ if ((u16)(vq->vring.avail->idx - last_avail) > vq->vring.num)
errx(1, "Guest moved used index from %u to %u",
- vq->last_avail_idx, vq->vring.avail->idx);
+ last_avail, vq->vring.avail->idx);
/* If there's nothing new since last we looked, return invalid. */
- if (vq->vring.avail->idx == vq->last_avail_idx)
+ if (vq->vring.avail->idx == last_avail)
return vq->vring.num;
/* Grab the next descriptor number they're advertising, and increment
* the index we've seen. */
- head = vq->vring.avail->ring[vq->last_avail_idx++ % vq->vring.num];
+ head = vq->vring.avail->ring[last_avail % vq->vring.num];
+ lg_last_avail(vq)++;
/* If their number is silly, that's a fatal mistake. */
if (head >= vq->vring.num)
errx(1, "Looped descriptor");
} while ((i = next_desc(vq, i)) != vq->vring.num);
+ vq->inflight++;
return head;
}
-/* Once we've used one of their buffers, we tell them about it. We'll then
+/* 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(). */
static void add_used(struct virtqueue *vq, unsigned int head, int len)
{
struct vring_used_elem *used;
- /* Get a pointer to the next entry in the used ring. */
+ /* The virtqueue contains a ring of used buffers. Get a pointer to the
+ * next entry in that used ring. */
used = &vq->vring.used->ring[vq->vring.used->idx % vq->vring.num];
used->id = head;
used->len = len;
/* Make sure buffer is written before we update index. */
wmb();
vq->vring.used->idx++;
+ vq->inflight--;
}
/* This actually sends the interrupt for this virtqueue */
{
unsigned long buf[] = { LHREQ_IRQ, vq->config.irq };
- if (vq->vring.avail->flags & VRING_AVAIL_F_NO_INTERRUPT)
+ /* If they don't want an interrupt, don't send one, unless empty. */
+ if ((vq->vring.avail->flags & VRING_AVAIL_F_NO_INTERRUPT)
+ && vq->inflight)
return;
/* Send the Guest an interrupt tell them we used something up. */
trigger_irq(fd, vq);
}
-/* Here is the input terminal setting we save, and the routine to restore them
- * on exit so the user can see what they type next. */
+/*
+ * The Console
+ *
+ * Here is the input terminal setting we save, and the routine to restore them
+ * on exit so the user gets their terminal back. */
static struct termios orig_term;
static void restore_term(void)
{
unsigned long args[] = { LHREQ_BREAK, 0 };
/* Close the fd so Waker will know it has to
* exit. */
- close(waker_fd);
- /* Just in case waker is blocked in BREAK, send
+ close(waker_fds.pipe[1]);
+ /* Just in case Waker is blocked in BREAK, send
* unbreak now. */
write(fd, args, sizeof(args));
exit(2);
/* Handling output for console is simple: we just get all the output buffers
* and write them to stdout. */
-static void handle_console_output(int fd, struct virtqueue *vq)
+static void handle_console_output(int fd, struct virtqueue *vq, bool timeout)
{
unsigned int head, out, in;
int len;
}
}
-/* Handling output for network is also simple: we get all the output buffers
+/* This is called when we no longer want to hear about Guest changes to a
+ * virtqueue. This is more efficient in high-traffic cases, but it means we
+ * have to set a timer to check if any more changes have occurred. */
+static void block_vq(struct virtqueue *vq)
+{
+ struct itimerval itm;
+
+ vq->vring.used->flags |= VRING_USED_F_NO_NOTIFY;
+ vq->blocked = true;
+
+ itm.it_interval.tv_sec = 0;
+ itm.it_interval.tv_usec = 0;
+ itm.it_value.tv_sec = 0;
+ itm.it_value.tv_usec = timeout_usec;
+
+ setitimer(ITIMER_REAL, &itm, NULL);
+}
+
+/*
+ * The Network
+ *
+ * Handling output for network is also simple: we get all the output buffers
* and write them (ignoring the first element) to this device's file descriptor
- * (stdout). */
-static void handle_net_output(int fd, struct virtqueue *vq)
+ * (/dev/net/tun).
+ */
+static void handle_net_output(int fd, struct virtqueue *vq, bool timeout)
{
- unsigned int head, out, in;
+ unsigned int head, out, in, num = 0;
int len;
struct iovec iov[vq->vring.num];
+ static int last_timeout_num;
/* Keep getting output buffers from the Guest until we run out. */
while ((head = get_vq_desc(vq, iov, &out, &in)) != vq->vring.num) {
if (in)
errx(1, "Input buffers in output queue?");
- /* Check header, but otherwise ignore it (we said we supported
- * no features). */
- (void)convert(&iov[0], struct virtio_net_hdr);
- len = writev(vq->dev->fd, iov+1, out-1);
+ len = writev(vq->dev->fd, iov, out);
+ if (len < 0)
+ err(1, "Writing network packet to tun");
add_used_and_trigger(fd, vq, head, len);
+ num++;
+ }
+
+ /* Block further kicks and set up a timer if we saw anything. */
+ if (!timeout && num)
+ block_vq(vq);
+
+ /* We never quite know how long should we wait before we check the
+ * queue again for more packets. We start at 500 microseconds, and if
+ * we get fewer packets than last time, we assume we made the timeout
+ * too small and increase it by 10 microseconds. Otherwise, we drop it
+ * by one microsecond every time. It seems to work well enough. */
+ if (timeout) {
+ if (num < last_timeout_num)
+ timeout_usec += 10;
+ else if (timeout_usec > 1)
+ timeout_usec--;
+ last_timeout_num = num;
}
}
unsigned int head, in_num, out_num;
int len;
struct iovec iov[dev->vq->vring.num];
- struct virtio_net_hdr *hdr;
/* First we need a network buffer from the Guests's recv virtqueue. */
head = get_vq_desc(dev->vq, iov, &out_num, &in_num);
* early, the Guest won't be ready yet. Wait until the device
* status says it's ready. */
/* FIXME: Actually want DRIVER_ACTIVE here. */
- if (dev->desc->status & VIRTIO_CONFIG_S_DRIVER_OK)
- warn("network: no dma buffer!");
+
+ /* Now tell it we want to know if new things appear. */
+ dev->vq->vring.used->flags &= ~VRING_USED_F_NO_NOTIFY;
+ wmb();
+
/* We'll turn this back on if input buffers are registered. */
return false;
} else if (out_num)
errx(1, "Output buffers in network recv queue?");
- /* First element is the header: we set it to 0 (no features). */
- hdr = convert(&iov[0], struct virtio_net_hdr);
- hdr->flags = 0;
- hdr->gso_type = VIRTIO_NET_HDR_GSO_NONE;
-
/* Read the packet from the device directly into the Guest's buffer. */
- len = readv(dev->fd, iov+1, in_num-1);
+ len = readv(dev->fd, iov, in_num);
if (len <= 0)
err(1, "reading network");
/* Tell the Guest about the new packet. */
- add_used_and_trigger(fd, dev->vq, head, sizeof(*hdr) + len);
+ add_used_and_trigger(fd, dev->vq, head, len);
verbose("tun input packet len %i [%02x %02x] (%s)\n", len,
((u8 *)iov[1].iov_base)[0], ((u8 *)iov[1].iov_base)[1],
return true;
}
-/* This callback ensures we try again, in case we stopped console or net
+/*L:215 This is the callback attached to the network and console input
+ * virtqueues: it ensures we try again, in case we stopped console or net
* delivery because Guest didn't have any buffers. */
-static void enable_fd(int fd, struct virtqueue *vq)
+static void enable_fd(int fd, struct virtqueue *vq, bool timeout)
{
add_device_fd(vq->dev->fd);
- /* Tell waker to listen to it again */
- write(waker_fd, &vq->dev->fd, sizeof(vq->dev->fd));
+ /* Snap the Waker out of its select loop. */
+ write(waker_fds.pipe[1], "", 1);
+}
+
+static void net_enable_fd(int fd, struct virtqueue *vq, bool timeout)
+{
+ /* We don't need to know again when Guest refills receive buffer. */
+ vq->vring.used->flags |= VRING_USED_F_NO_NOTIFY;
+ enable_fd(fd, vq, timeout);
+}
+
+/* When the Guest tells us they updated the status field, we handle it. */
+static void update_device_status(struct device *dev)
+{
+ struct virtqueue *vq;
+
+ /* This is a reset. */
+ if (dev->desc->status == 0) {
+ verbose("Resetting device %s\n", dev->name);
+
+ /* Clear any features they've acked. */
+ memset(get_feature_bits(dev) + dev->desc->feature_len, 0,
+ dev->desc->feature_len);
+
+ /* Zero out the virtqueues. */
+ for (vq = dev->vq; vq; vq = vq->next) {
+ memset(vq->vring.desc, 0,
+ vring_size(vq->config.num, LGUEST_VRING_ALIGN));
+ lg_last_avail(vq) = 0;
+ }
+ } else if (dev->desc->status & VIRTIO_CONFIG_S_FAILED) {
+ warnx("Device %s configuration FAILED", dev->name);
+ } else if (dev->desc->status & VIRTIO_CONFIG_S_DRIVER_OK) {
+ unsigned int i;
+
+ verbose("Device %s OK: offered", dev->name);
+ for (i = 0; i < dev->desc->feature_len; i++)
+ verbose(" %02x", get_feature_bits(dev)[i]);
+ verbose(", accepted");
+ for (i = 0; i < dev->desc->feature_len; i++)
+ verbose(" %02x", get_feature_bits(dev)
+ [dev->desc->feature_len+i]);
+
+ if (dev->ready)
+ dev->ready(dev);
+ }
}
/* This is the generic routine we call when the Guest uses LHCALL_NOTIFY. */
struct device *i;
struct virtqueue *vq;
- /* Check each virtqueue. */
+ /* Check each device and virtqueue. */
for (i = devices.dev; i; i = i->next) {
+ /* Notifications to device descriptors update device status. */
+ if (from_guest_phys(addr) == i->desc) {
+ update_device_status(i);
+ return;
+ }
+
+ /* Notifications to virtqueues mean output has occurred. */
for (vq = i->vq; vq; vq = vq->next) {
- if (vq->config.pfn == addr/getpagesize()
- && vq->handle_output) {
- verbose("Output to %s\n", vq->dev->name);
- vq->handle_output(fd, vq);
+ if (vq->config.pfn != addr/getpagesize())
+ continue;
+
+ /* Guest should acknowledge (and set features!) before
+ * using the device. */
+ if (i->desc->status == 0) {
+ warnx("%s gave early output", i->name);
return;
}
+
+ if (strcmp(vq->dev->name, "console") != 0)
+ verbose("Output to %s\n", vq->dev->name);
+ if (vq->handle_output)
+ vq->handle_output(fd, vq, false);
+ return;
}
}
strnlen(from_guest_phys(addr), guest_limit - addr));
}
-/* This is called when the waker wakes us up: check for incoming file
+static void handle_timeout(int fd)
+{
+ char buf[32];
+ struct device *i;
+ struct virtqueue *vq;
+
+ /* Clear the pipe */
+ read(timeoutpipe[0], buf, sizeof(buf));
+
+ /* Check each device and virtqueue: flush blocked ones. */
+ for (i = devices.dev; i; i = i->next) {
+ for (vq = i->vq; vq; vq = vq->next) {
+ if (!vq->blocked)
+ continue;
+
+ vq->vring.used->flags &= ~VRING_USED_F_NO_NOTIFY;
+ vq->blocked = false;
+ if (vq->handle_output)
+ vq->handle_output(fd, vq, true);
+ }
+ }
+}
+
+/* This is called when the Waker wakes us up: check for incoming file
* descriptors. */
static void handle_input(int fd)
{
for (;;) {
struct device *i;
fd_set fds = devices.infds;
+ int num;
+ num = select(devices.max_infd+1, &fds, NULL, NULL, &poll);
+ /* Could get interrupted */
+ if (num < 0)
+ continue;
/* If nothing is ready, we're done. */
- if (select(devices.max_infd+1, &fds, NULL, NULL, &poll) == 0)
+ if (num == 0)
break;
- /* Otherwise, call the device(s) which have readable
- * file descriptors and a method of handling them. */
+ /* Otherwise, call the device(s) which have readable file
+ * descriptors and a method of handling them. */
for (i = devices.dev; i; i = i->next) {
if (i->handle_input && FD_ISSET(i->fd, &fds)) {
- int dev_fd;
if (i->handle_input(fd, i))
continue;
* should no longer service it. Networking and
* console do this when there's no input
* buffers to deliver into. Console also uses
- * it when it discovers that stdin is
- * closed. */
+ * it when it discovers that stdin is closed. */
FD_CLR(i->fd, &devices.infds);
- /* Tell waker to ignore it too, by sending a
- * negative fd number (-1, since 0 is a valid
- * FD number). */
- dev_fd = -i->fd - 1;
- write(waker_fd, &dev_fd, sizeof(dev_fd));
}
}
+
+ /* Is this the timeout fd? */
+ if (FD_ISSET(timeoutpipe[0], &fds))
+ handle_timeout(fd);
}
}
*
* All devices need a descriptor so the Guest knows it exists, and a "struct
* device" so the Launcher can keep track of it. We have common helper
- * routines to allocate them.
- *
- * This routine allocates a new "struct lguest_device_desc" from descriptor
- * table just above the Guest's normal memory. It returns a pointer to that
- * descriptor. */
-static struct lguest_device_desc *new_dev_desc(u16 type)
-{
- struct lguest_device_desc *d;
-
- /* We only have one page for all the descriptors. */
- if (devices.desc_used + sizeof(*d) > getpagesize())
- errx(1, "Too many devices");
-
- /* We don't need to set config_len or status: page is 0 already. */
- d = (void *)devices.descpage + devices.desc_used;
- d->type = type;
- devices.desc_used += sizeof(*d);
+ * routines to allocate and manage them.
+ */
- return d;
+/* The layout of the device page is a "struct lguest_device_desc" followed by a
+ * number of virtqueue descriptors, then two sets of feature bits, then an
+ * array of configuration bytes. This routine returns the configuration
+ * pointer. */
+static u8 *device_config(const struct device *dev)
+{
+ return (void *)(dev->desc + 1)
+ + dev->desc->num_vq * sizeof(struct lguest_vqconfig)
+ + dev->desc->feature_len * 2;
}
-/* Each device descriptor is followed by some configuration information.
- * The first byte is a "status" byte for the Guest to report what's happening.
- * After that are fields: u8 type, u8 len, [... len bytes...].
- *
- * This routine adds a new field to an existing device's descriptor. It only
- * works for the last device, but that's OK because that's how we use it. */
-static void add_desc_field(struct device *dev, u8 type, u8 len, const void *c)
+/* This routine allocates a new "struct lguest_device_desc" from descriptor
+ * table page just above the Guest's normal memory. It returns a pointer to
+ * that descriptor. */
+static struct lguest_device_desc *new_dev_desc(u16 type)
{
- /* This is the last descriptor, right? */
- assert(devices.descpage + devices.desc_used
- == (u8 *)(dev->desc + 1) + dev->desc->config_len);
+ struct lguest_device_desc d = { .type = type };
+ void *p;
- /* We only have one page of device descriptions. */
- if (devices.desc_used + 2 + len > getpagesize())
- errx(1, "Too many devices");
+ /* Figure out where the next device config is, based on the last one. */
+ if (devices.lastdev)
+ p = device_config(devices.lastdev)
+ + devices.lastdev->desc->config_len;
+ else
+ p = devices.descpage;
- /* Copy in the new config header: type then length. */
- devices.descpage[devices.desc_used++] = type;
- devices.descpage[devices.desc_used++] = len;
- memcpy(devices.descpage + devices.desc_used, c, len);
- devices.desc_used += len;
+ /* We only have one page for all the descriptors. */
+ if (p + sizeof(d) > (void *)devices.descpage + getpagesize())
+ errx(1, "Too many devices");
- /* Update the device descriptor length: two byte head then data. */
- dev->desc->config_len += 2 + len;
+ /* p might not be aligned, so we memcpy in. */
+ return memcpy(p, &d, sizeof(d));
}
-/* This routine adds a virtqueue to a device. We specify how many descriptors
- * the virtqueue is to have. */
+/* Each device descriptor is followed by the description of its virtqueues. We
+ * specify how many descriptors the virtqueue is to have. */
static void add_virtqueue(struct device *dev, unsigned int num_descs,
- void (*handle_output)(int fd, struct virtqueue *me))
+ void (*handle_output)(int, struct virtqueue *, bool))
{
unsigned int pages;
struct virtqueue **i, *vq = malloc(sizeof(*vq));
void *p;
- /* First we need some pages for this virtqueue. */
- pages = (vring_size(num_descs) + getpagesize() - 1) / getpagesize();
+ /* First we need some memory for this virtqueue. */
+ pages = (vring_size(num_descs, LGUEST_VRING_ALIGN) + getpagesize() - 1)
+ / getpagesize();
p = get_pages(pages);
+ /* Initialize the virtqueue */
+ vq->next = NULL;
+ vq->last_avail_idx = 0;
+ vq->dev = dev;
+ vq->inflight = 0;
+ vq->blocked = false;
+
/* Initialize the configuration. */
vq->config.num = num_descs;
vq->config.irq = devices.next_irq++;
vq->config.pfn = to_guest_phys(p) / getpagesize();
/* Initialize the vring. */
- vring_init(&vq->vring, num_descs, p);
+ vring_init(&vq->vring, num_descs, p, LGUEST_VRING_ALIGN);
- /* Add the configuration information to this device's descriptor. */
- add_desc_field(dev, VIRTIO_CONFIG_F_VIRTQUEUE,
- sizeof(vq->config), &vq->config);
+ /* Append virtqueue to this device's descriptor. We use
+ * device_config() to get the end of the device's current virtqueues;
+ * we check that we haven't added any config or feature information
+ * yet, otherwise we'd be overwriting them. */
+ assert(dev->desc->config_len == 0 && dev->desc->feature_len == 0);
+ memcpy(device_config(dev), &vq->config, sizeof(vq->config));
+ dev->desc->num_vq++;
+
+ verbose("Virtqueue page %#lx\n", to_guest_phys(p));
/* Add to tail of list, so dev->vq is first vq, dev->vq->next is
* second. */
for (i = &dev->vq; *i; i = &(*i)->next);
*i = vq;
- /* Link virtqueue back to device. */
- vq->dev = dev;
-
- /* Set up handler. */
+ /* Set the routine to call when the Guest does something to this
+ * virtqueue. */
vq->handle_output = handle_output;
+
+ /* As an optimization, set the advisory "Don't Notify Me" flag if we
+ * don't have a handler */
if (!handle_output)
vq->vring.used->flags = VRING_USED_F_NO_NOTIFY;
}
+/* The first half of the feature bitmask is for us to advertise features. The
+ * second half is for the Guest to accept features. */
+static void add_feature(struct device *dev, unsigned bit)
+{
+ u8 *features = get_feature_bits(dev);
+
+ /* We can't extend the feature bits once we've added config bytes */
+ if (dev->desc->feature_len <= bit / CHAR_BIT) {
+ assert(dev->desc->config_len == 0);
+ dev->desc->feature_len = (bit / CHAR_BIT) + 1;
+ }
+
+ features[bit / CHAR_BIT] |= (1 << (bit % CHAR_BIT));
+}
+
+/* This routine sets the configuration fields for an existing device's
+ * descriptor. It only works for the last device, but that's OK because that's
+ * how we use it. */
+static void set_config(struct device *dev, unsigned len, const void *conf)
+{
+ /* Check we haven't overflowed our single page. */
+ if (device_config(dev) + len > devices.descpage + getpagesize())
+ errx(1, "Too many devices");
+
+ /* Copy in the config information, and store the length. */
+ memcpy(device_config(dev), conf, len);
+ dev->desc->config_len = len;
+}
+
/* This routine does all the creation and setup of a new device, including
- * caling new_dev_desc() to allocate the descriptor and device memory. */
+ * calling new_dev_desc() to allocate the descriptor and device memory.
+ *
+ * See what I mean about userspace being boring? */
static struct device *new_device(const char *name, u16 type, int fd,
bool (*handle_input)(int, struct device *))
{
struct device *dev = malloc(sizeof(*dev));
- /* Append to device list. Prepending to a single-linked list is
- * easier, but the user expects the devices to be arranged on the bus
- * in command-line order. The first network device on the command line
- * is eth0, the first block device /dev/lgba, etc. */
- *devices.lastdev = dev;
- dev->next = NULL;
- devices.lastdev = &dev->next;
-
/* Now we populate the fields one at a time. */
dev->fd = fd;
/* If we have an input handler for this file descriptor, then we add it
dev->desc = new_dev_desc(type);
dev->handle_input = handle_input;
dev->name = name;
+ dev->vq = NULL;
+ dev->ready = NULL;
+
+ /* Append to device list. Prepending to a single-linked list is
+ * easier, but the user expects the devices to be arranged on the bus
+ * in command-line order. The first network device on the command line
+ * is eth0, the first block device /dev/vda, etc. */
+ if (devices.lastdev)
+ devices.lastdev->next = dev;
+ else
+ devices.dev = dev;
+ devices.lastdev = dev;
+
return dev;
}
/* The console needs two virtqueues: the input then the output. When
* they put something the input queue, we make sure we're listening to
* stdin. When they put something in the output queue, we write it to
- * stdout. */
+ * stdout. */
add_virtqueue(dev, VIRTQUEUE_NUM, enable_fd);
add_virtqueue(dev, VIRTQUEUE_NUM, handle_console_output);
}
/*:*/
+static void timeout_alarm(int sig)
+{
+ write(timeoutpipe[1], "", 1);
+}
+
+static void setup_timeout(void)
+{
+ if (pipe(timeoutpipe) != 0)
+ err(1, "Creating timeout pipe");
+
+ if (fcntl(timeoutpipe[1], F_SETFL,
+ fcntl(timeoutpipe[1], F_GETFL) | O_NONBLOCK) != 0)
+ err(1, "Making timeout pipe nonblocking");
+
+ add_device_fd(timeoutpipe[0]);
+ signal(SIGALRM, timeout_alarm);
+}
+
/*M:010 Inter-guest networking is an interesting area. Simplest is to have a
* --sharenet=<name> option which opens or creates a named pipe. This can be
* used to send packets to another guest in a 1:1 manner.
static u32 str2ip(const char *ipaddr)
{
- unsigned int byte[4];
+ unsigned int b[4];
- sscanf(ipaddr, "%u.%u.%u.%u", &byte[0], &byte[1], &byte[2], &byte[3]);
- return (byte[0] << 24) | (byte[1] << 16) | (byte[2] << 8) | byte[3];
+ if (sscanf(ipaddr, "%u.%u.%u.%u", &b[0], &b[1], &b[2], &b[3]) != 4)
+ errx(1, "Failed to parse IP address '%s'", ipaddr);
+ return (b[0] << 24) | (b[1] << 16) | (b[2] << 8) | b[3];
+}
+
+static void str2mac(const char *macaddr, unsigned char mac[6])
+{
+ unsigned int m[6];
+ if (sscanf(macaddr, "%02x:%02x:%02x:%02x:%02x:%02x",
+ &m[0], &m[1], &m[2], &m[3], &m[4], &m[5]) != 6)
+ errx(1, "Failed to parse mac address '%s'", macaddr);
+ mac[0] = m[0];
+ mac[1] = m[1];
+ mac[2] = m[2];
+ mac[3] = m[3];
+ mac[4] = m[4];
+ mac[5] = m[5];
}
/* This code is "adapted" from libbridge: it attaches the Host end of the
errx(1, "interface %s does not exist!", if_name);
strncpy(ifr.ifr_name, br_name, IFNAMSIZ);
+ ifr.ifr_name[IFNAMSIZ-1] = '\0';
ifr.ifr_ifindex = ifidx;
if (ioctl(fd, SIOCBRADDIF, &ifr) < 0)
err(1, "can't add %s to bridge %s", if_name, br_name);
/* This sets up the Host end of the network device with an IP address, brings
* it up so packets will flow, the copies the MAC address into the hwaddr
* pointer. */
-static void configure_device(int fd, const char *devname, u32 ipaddr,
- unsigned char hwaddr[6])
+static void configure_device(int fd, const char *tapif, u32 ipaddr)
{
struct ifreq ifr;
struct sockaddr_in *sin = (struct sockaddr_in *)&ifr.ifr_addr;
- /* Don't read these incantations. Just cut & paste them like I did! */
memset(&ifr, 0, sizeof(ifr));
- strcpy(ifr.ifr_name, devname);
+ strcpy(ifr.ifr_name, tapif);
+
+ /* Don't read these incantations. Just cut & paste them like I did! */
sin->sin_family = AF_INET;
sin->sin_addr.s_addr = htonl(ipaddr);
if (ioctl(fd, SIOCSIFADDR, &ifr) != 0)
- err(1, "Setting %s interface address", devname);
+ err(1, "Setting %s interface address", tapif);
ifr.ifr_flags = IFF_UP;
if (ioctl(fd, SIOCSIFFLAGS, &ifr) != 0)
- err(1, "Bringing interface %s up", devname);
-
- /* SIOC stands for Socket I/O Control. G means Get (vs S for Set
- * above). IF means Interface, and HWADDR is hardware address.
- * Simple! */
- if (ioctl(fd, SIOCGIFHWADDR, &ifr) != 0)
- err(1, "getting hw address for %s", devname);
- memcpy(hwaddr, ifr.ifr_hwaddr.sa_data, 6);
+ err(1, "Bringing interface %s up", tapif);
}
-/*L:195 Our network is a Host<->Guest network. This can either use bridging or
- * routing, but the principle is the same: it uses the "tun" device to inject
- * packets into the Host as if they came in from a normal network card. We
- * just shunt packets between the Guest and the tun device. */
-static void setup_tun_net(const char *arg)
+static int get_tun_device(char tapif[IFNAMSIZ])
{
- struct device *dev;
struct ifreq ifr;
- int netfd, ipfd;
- u32 ip;
- const char *br_name = NULL;
- u8 hwaddr[6];
+ int netfd;
+
+ /* Start with this zeroed. Messy but sure. */
+ memset(&ifr, 0, sizeof(ifr));
/* We open the /dev/net/tun device and tell it we want a tap device. A
* tap device is like a tun device, only somehow different. To tell
* the truth, I completely blundered my way through this code, but it
* works now! */
netfd = open_or_die("/dev/net/tun", O_RDWR);
- memset(&ifr, 0, sizeof(ifr));
- ifr.ifr_flags = IFF_TAP | IFF_NO_PI;
+ ifr.ifr_flags = IFF_TAP | IFF_NO_PI | IFF_VNET_HDR;
strcpy(ifr.ifr_name, "tap%d");
if (ioctl(netfd, TUNSETIFF, &ifr) != 0)
err(1, "configuring /dev/net/tun");
+
+ if (ioctl(netfd, TUNSETOFFLOAD,
+ TUN_F_CSUM|TUN_F_TSO4|TUN_F_TSO6|TUN_F_TSO_ECN) != 0)
+ err(1, "Could not set features for tun device");
+
/* We don't need checksums calculated for packets coming in this
* device: trust us! */
ioctl(netfd, TUNSETNOCSUM, 1);
+ memcpy(tapif, ifr.ifr_name, IFNAMSIZ);
+ return netfd;
+}
+
+/*L:195 Our network is a Host<->Guest network. This can either use bridging or
+ * routing, but the principle is the same: it uses the "tun" device to inject
+ * packets into the Host as if they came in from a normal network card. We
+ * just shunt packets between the Guest and the tun device. */
+static void setup_tun_net(char *arg)
+{
+ struct device *dev;
+ int netfd, ipfd;
+ u32 ip = INADDR_ANY;
+ bool bridging = false;
+ char tapif[IFNAMSIZ], *p;
+ struct virtio_net_config conf;
+
+ netfd = get_tun_device(tapif);
+
/* First we create a new network device. */
dev = new_device("net", VIRTIO_ID_NET, netfd, handle_tun_input);
/* Network devices need a receive and a send queue, just like
* console. */
- add_virtqueue(dev, VIRTQUEUE_NUM, enable_fd);
+ add_virtqueue(dev, VIRTQUEUE_NUM, net_enable_fd);
add_virtqueue(dev, VIRTQUEUE_NUM, handle_net_output);
/* We need a socket to perform the magic network ioctls to bring up the
/* If the command line was --tunnet=bridge:<name> do bridging. */
if (!strncmp(BRIDGE_PFX, arg, strlen(BRIDGE_PFX))) {
- ip = INADDR_ANY;
- br_name = arg + strlen(BRIDGE_PFX);
- add_to_bridge(ipfd, ifr.ifr_name, br_name);
- } else /* It is an IP address to set up the device with */
- ip = str2ip(arg);
+ arg += strlen(BRIDGE_PFX);
+ bridging = true;
+ }
- /* Set up the tun device, and get the mac address for the interface. */
- configure_device(ipfd, ifr.ifr_name, ip, hwaddr);
+ /* A mac address may follow the bridge name or IP address */
+ p = strchr(arg, ':');
+ if (p) {
+ str2mac(p+1, conf.mac);
+ add_feature(dev, VIRTIO_NET_F_MAC);
+ *p = '\0';
+ }
- /* Tell Guest what MAC address to use. */
- add_desc_field(dev, VIRTIO_CONFIG_NET_MAC_F, sizeof(hwaddr), hwaddr);
+ /* arg is now either an IP address or a bridge name */
+ if (bridging)
+ add_to_bridge(ipfd, tapif, arg);
+ else
+ ip = str2ip(arg);
- /* We don't seed the socket any more; setup is done. */
+ /* Set up the tun device. */
+ configure_device(ipfd, tapif, ip);
+
+ add_feature(dev, VIRTIO_F_NOTIFY_ON_EMPTY);
+ /* Expect Guest to handle everything except UFO */
+ add_feature(dev, VIRTIO_NET_F_CSUM);
+ add_feature(dev, VIRTIO_NET_F_GUEST_CSUM);
+ add_feature(dev, VIRTIO_NET_F_GUEST_TSO4);
+ add_feature(dev, VIRTIO_NET_F_GUEST_TSO6);
+ add_feature(dev, VIRTIO_NET_F_GUEST_ECN);
+ add_feature(dev, VIRTIO_NET_F_HOST_TSO4);
+ add_feature(dev, VIRTIO_NET_F_HOST_TSO6);
+ add_feature(dev, VIRTIO_NET_F_HOST_ECN);
+ set_config(dev, sizeof(conf), &conf);
+
+ /* We don't need the socket any more; setup is done. */
close(ipfd);
- verbose("device %u: tun net %u.%u.%u.%u\n",
- devices.device_num++,
- (u8)(ip>>24),(u8)(ip>>16),(u8)(ip>>8),(u8)ip);
- if (br_name)
- verbose("attached to bridge: %s\n", br_name);
-}
+ devices.device_num++;
+ if (bridging)
+ verbose("device %u: tun %s attached to bridge: %s\n",
+ devices.device_num, tapif, arg);
+ else
+ verbose("device %u: tun %s: %s\n",
+ devices.device_num, tapif, arg);
+}
-/*
- * Block device.
+/* 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.
*
- * Serving a block device is really easy: the Guest asks for a block number and
- * we read or write that position in the file.
- *
- * Unfortunately, this is amazingly slow: the Guest waits until the read is
- * finished before running anything else, even if it could be 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.
+ * 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 we farm the I/O out to thread, and communicate with it via a pipe. */
-/* This hangs off device->priv, with the data. */
+/* This hangs off device->priv. */
struct vblk_info
{
/* The size of the file. */
int done_fd;
};
-/* This is the core of the I/O thread. It returns true if it did something. */
+/*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:
+ */
static bool service_io(struct device *dev)
{
struct vblk_info *vblk = dev->priv;
unsigned int head, out_num, in_num, wlen;
int ret;
- struct virtio_blk_inhdr *in;
+ u8 *in;
struct virtio_blk_outhdr *out;
struct iovec iov[dev->vq->vring.num];
off64_t off;
+ /* See if there's a request waiting. If not, nothing to do. */
head = get_vq_desc(dev->vq, iov, &out_num, &in_num);
if (head == dev->vq->vring.num)
return false;
+ /* Every block request should contain at least one output buffer
+ * (detailing the location on disk and the type of request) and one
+ * input buffer (to hold the result). */
if (out_num == 0 || in_num == 0)
errx(1, "Bad virtblk cmd %u out=%u in=%u",
head, out_num, in_num);
out = convert(&iov[0], struct virtio_blk_outhdr);
- in = convert(&iov[out_num+in_num-1], struct virtio_blk_inhdr);
+ in = convert(&iov[out_num+in_num-1], u8);
off = out->sector * 512;
- /* This is how we implement barriers. Pretty poor, no? */
+ /* The block device implements "barriers", where the Guest indicates
+ * that it wants all previous writes to occur before this write. We
+ * don't have a way of asking our kernel to do a barrier, so we just
+ * synchronize all the data in the file. Pretty poor, no? */
if (out->type & VIRTIO_BLK_T_BARRIER)
fdatasync(vblk->fd);
+ /* In general the virtio block driver is allowed to try SCSI commands.
+ * It'd be nice if we supported eject, for example, but we don't. */
if (out->type & VIRTIO_BLK_T_SCSI_CMD) {
fprintf(stderr, "Scsi commands unsupported\n");
- in->status = VIRTIO_BLK_S_UNSUPP;
- wlen = sizeof(in);
+ *in = VIRTIO_BLK_S_UNSUPP;
+ wlen = sizeof(*in);
} else if (out->type & VIRTIO_BLK_T_OUT) {
/* Write */
/* Die, bad Guest, die. */
errx(1, "Write past end %llu+%u", off, ret);
}
- wlen = sizeof(in);
- in->status = (ret >= 0 ? VIRTIO_BLK_S_OK : VIRTIO_BLK_S_IOERR);
+ wlen = sizeof(*in);
+ *in = (ret >= 0 ? VIRTIO_BLK_S_OK : VIRTIO_BLK_S_IOERR);
} else {
/* Read */
ret = readv(vblk->fd, iov+1, in_num-1);
verbose("READ from sector %llu: %i\n", out->sector, ret);
if (ret >= 0) {
- wlen = sizeof(in) + ret;
- in->status = VIRTIO_BLK_S_OK;
+ wlen = sizeof(*in) + ret;
+ *in = VIRTIO_BLK_S_OK;
} else {
- wlen = sizeof(in);
- in->status = VIRTIO_BLK_S_IOERR;
+ wlen = sizeof(*in);
+ *in = VIRTIO_BLK_S_IOERR;
}
}
+ /* OK, so we noted that it was pretty poor to use an fdatasync as a
+ * barrier. But Christoph Hellwig points out that we need a sync
+ * *afterwards* as well: "Barriers specify no reordering to the front
+ * or the back." And Jens Axboe confirmed it, so here we are: */
+ if (out->type & VIRTIO_BLK_T_BARRIER)
+ fdatasync(vblk->fd);
+
/* We can't trigger an IRQ, because we're not the Launcher. It does
* that when we tell it we're done. */
add_used(dev->vq, head, wlen);
/* When this read fails, it means Launcher died, so we follow. */
while (read(vblk->workpipe[0], &c, 1) == 1) {
- /* We acknowledge each request immediately, to reduce latency,
+ /* We acknowledge each request immediately to reduce latency,
* rather than waiting until we've done them all. I haven't
- * measured to see if it makes any difference. */
+ * measured to see if it makes any difference.
+ *
+ * That would be an interesting test, wouldn't it? You could
+ * also try having more than one I/O thread. */
while (service_io(dev))
write(vblk->done_fd, &c, 1);
}
return 0;
}
-/* When the thread says some I/O is done, we interrupt the Guest. */
+/* Now we've seen the I/O thread, we return to the Launcher to see what happens
+ * when that thread tells us it's completed some I/O. */
static bool handle_io_finish(int fd, struct device *dev)
{
char c;
- /* If child died, presumably it printed message. */
+ /* If the I/O thread died, presumably it printed the error, so we
+ * simply exit. */
if (read(dev->fd, &c, 1) != 1)
exit(1);
return true;
}
-/* When the Guest submits some I/O, we wake the I/O thread. */
-static void handle_virtblk_output(int fd, struct virtqueue *vq)
+/* When the Guest submits some I/O, we just need to wake the I/O thread. */
+static void handle_virtblk_output(int fd, struct virtqueue *vq, bool timeout)
{
struct vblk_info *vblk = vq->dev->priv;
char c = 0;
exit(1);
}
-/* This creates a virtual block device. */
+/*L:198 This actually sets up a virtual block device. */
static void setup_block_file(const char *filename)
{
int p[2];
struct device *dev;
struct vblk_info *vblk;
void *stack;
- u64 cap;
- unsigned int val;
+ struct virtio_blk_config conf;
/* This is the pipe the I/O thread will use to tell us I/O is done. */
pipe(p);
/* The device responds to return from I/O thread. */
dev = new_device("block", VIRTIO_ID_BLOCK, p[0], handle_io_finish);
- /* The device has a virtqueue. */
+ /* The device has one virtqueue, where the Guest places requests. */
add_virtqueue(dev, VIRTQUEUE_NUM, handle_virtblk_output);
/* Allocate the room for our own bookkeeping */
vblk->fd = open_or_die(filename, O_RDWR|O_LARGEFILE);
vblk->len = lseek64(vblk->fd, 0, SEEK_END);
+ /* We support barriers. */
+ add_feature(dev, VIRTIO_BLK_F_BARRIER);
+
/* Tell Guest how many sectors this device has. */
- cap = cpu_to_le64(vblk->len / 512);
- add_desc_field(dev, VIRTIO_CONFIG_BLK_F_CAPACITY, sizeof(cap), &cap);
+ conf.capacity = cpu_to_le64(vblk->len / 512);
/* Tell Guest not to put in too many descriptors at once: two are used
* for the in and out elements. */
- val = cpu_to_le32(VIRTQUEUE_NUM - 2);
- add_desc_field(dev, VIRTIO_CONFIG_BLK_F_SEG_MAX, sizeof(val), &val);
+ add_feature(dev, VIRTIO_BLK_F_SEG_MAX);
+ conf.seg_max = cpu_to_le32(VIRTQUEUE_NUM - 2);
+
+ set_config(dev, sizeof(conf), &conf);
/* The I/O thread writes to this end of the pipe when done. */
vblk->done_fd = p[1];
- /* This is how we tell the I/O thread about more work. */
+ /* This is the second pipe, which is how we tell the I/O thread about
+ * more work. */
pipe(vblk->workpipe);
- /* Create stack for thread and run it */
+ /* Create stack for thread and run it. Since stack grows upwards, we
+ * point the stack pointer to the end of this region. */
stack = malloc(32768);
- if (clone(io_thread, stack + 32768, CLONE_VM, dev) == -1)
+ /* SIGCHLD - We dont "wait" for our cloned thread, so prevent it from
+ * becoming a zombie. */
+ if (clone(io_thread, stack + 32768, CLONE_VM | SIGCHLD, dev) == -1)
err(1, "Creating clone");
/* We don't need to keep the I/O thread's end of the pipes open. */
close(vblk->workpipe[0]);
verbose("device %u: virtblock %llu sectors\n",
- devices.device_num, cap);
+ devices.device_num, le64_to_cpu(conf.capacity));
+}
+
+/* Our random number generator device reads from /dev/random into the Guest's
+ * input buffers. The usual case is that the Guest doesn't want random numbers
+ * and so has no buffers although /dev/random is still readable, whereas
+ * console is the reverse.
+ *
+ * The same logic applies, however. */
+static bool handle_rng_input(int fd, struct device *dev)
+{
+ int len;
+ unsigned int head, in_num, out_num, totlen = 0;
+ struct iovec iov[dev->vq->vring.num];
+
+ /* First we need a buffer from the Guests's virtqueue. */
+ head = get_vq_desc(dev->vq, iov, &out_num, &in_num);
+
+ /* If they're not ready for input, stop listening to this file
+ * descriptor. We'll start again once they add an input buffer. */
+ if (head == dev->vq->vring.num)
+ return false;
+
+ if (out_num)
+ 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. */
+ while (!iov_empty(iov, in_num)) {
+ len = readv(dev->fd, iov, in_num);
+ if (len <= 0)
+ err(1, "Read from /dev/random gave %i", len);
+ iov_consume(iov, in_num, len);
+ totlen += len;
+ }
+
+ /* Tell the Guest about the new input. */
+ add_used_and_trigger(fd, dev->vq, head, totlen);
+
+ /* Everything went OK! */
+ return true;
+}
+
+/* And this creates a "hardware" random number device for the Guest. */
+static void setup_rng(void)
+{
+ struct device *dev;
+ int fd;
+
+ fd = open_or_die("/dev/random", O_RDONLY);
+
+ /* The device responds to return from I/O thread. */
+ dev = new_device("rng", VIRTIO_ID_RNG, fd, handle_rng_input);
+
+ /* The device has one virtqueue, where the Guest places inbufs. */
+ add_virtqueue(dev, VIRTQUEUE_NUM, enable_fd);
+
+ verbose("device %u: rng\n", devices.device_num++);
}
/* That's the end of device setup. */
-/*L:220 Finally we reach the core of the Launcher, which runs the Guest, serves
+/*L:230 Reboot is pretty easy: clean up and exec() the Launcher afresh. */
+static void __attribute__((noreturn)) restart_guest(void)
+{
+ unsigned int i;
+
+ /* Since we don't track all open fds, we simply close everything beyond
+ * stderr. */
+ for (i = 3; i < FD_SETSIZE; i++)
+ close(i);
+
+ /* The exec automatically gets rid of the I/O and Waker threads. */
+ execv(main_args[0], main_args);
+ err(1, "Could not exec %s", main_args[0]);
+}
+
+/*L:220 Finally we reach the core of the Launcher which runs the Guest, serves
* its input and output, and finally, lays it to rest. */
static void __attribute__((noreturn)) run_guest(int lguest_fd)
{
int readval;
/* We read from the /dev/lguest device to run the Guest. */
- readval = read(lguest_fd, ¬ify_addr, sizeof(notify_addr));
+ readval = pread(lguest_fd, ¬ify_addr,
+ sizeof(notify_addr), cpu_id);
/* One unsigned long means the Guest did HCALL_NOTIFY */
if (readval == sizeof(notify_addr)) {
/* ENOENT means the Guest died. Reading tells us why. */
} else if (errno == ENOENT) {
char reason[1024] = { 0 };
- read(lguest_fd, reason, sizeof(reason)-1);
+ pread(lguest_fd, reason, sizeof(reason)-1, cpu_id);
errx(1, "%s", reason);
- /* EAGAIN means the waker wanted us to look at some input.
+ /* ERESTART means that we need to reboot the guest */
+ } else if (errno == ERESTART) {
+ restart_guest();
+ /* EAGAIN means a signal (timeout).
* Anything else means a bug or incompatible change. */
} else if (errno != EAGAIN)
err(1, "Running guest failed");
- /* Service input, then unset the BREAK which releases
- * the Waker. */
+ /* Only service input on thread for CPU 0. */
+ if (cpu_id != 0)
+ continue;
+
+ /* Service input, then unset the BREAK to release the Waker. */
handle_input(lguest_fd);
- if (write(lguest_fd, args, sizeof(args)) < 0)
+ if (pwrite(lguest_fd, args, sizeof(args), cpu_id) < 0)
err(1, "Resetting break");
}
}
-/*
- * This is the end of the Launcher.
+/*L:240
+ * This is the end of the Launcher. The good news: we are over halfway
+ * through! The bad news: the most fiendish part of the code still lies ahead
+ * of us.
*
- * But wait! We've seen I/O from the Launcher, and we've seen I/O from the
- * Drivers. If we were to see the Host kernel I/O code, our understanding
- * would be complete... :*/
+ * Are you ready? Take a deep breath and join me in the core of the Host, in
+ * "make Host".
+ :*/
static struct option opts[] = {
{ "verbose", 0, NULL, 'v' },
{ "tunnet", 1, NULL, 't' },
{ "block", 1, NULL, 'b' },
+ { "rng", 0, NULL, 'r' },
{ "initrd", 1, NULL, 'i' },
{ NULL },
};
static void usage(void)
{
errx(1, "Usage: lguest [--verbose] "
- "[--tunnet=(<ipaddr>|bridge:<bridgename>)\n"
+ "[--tunnet=(<ipaddr>:<macaddr>|bridge:<bridgename>:<macaddr>)\n"
"|--block=<filename>|--initrd=<filename>]...\n"
"<mem-in-mb> vmlinux [args...]");
}
{
/* Memory, top-level pagetable, code startpoint and size of the
* (optional) initrd. */
- unsigned long mem = 0, pgdir, start, initrd_size = 0;
- /* A temporary and the /dev/lguest file descriptor. */
+ unsigned long mem = 0, start, initrd_size = 0;
+ /* Two temporaries and the /dev/lguest file descriptor. */
int i, c, lguest_fd;
/* The boot information for the Guest. */
- void *boot;
+ struct boot_params *boot;
/* If they specify an initrd file to load. */
const char *initrd_name = NULL;
+ /* Save the args: we "reboot" by execing ourselves again. */
+ main_args = argv;
+ /* We don't "wait" for the children, so prevent them from becoming
+ * zombies. */
+ signal(SIGCHLD, SIG_IGN);
+
/* First we initialize the device list. Since console and network
* device receive input from a file descriptor, we keep an fdset
* (infds) and the maximum fd number (max_infd) with the head of the
- * list. We also keep a pointer to the last device, for easy appending
- * to the list. Finally, we keep the next interrupt number to hand out
- * (1: remember that 0 is used by the timer). */
+ * list. We also keep a pointer to the last device. Finally, we keep
+ * the next interrupt number to use for devices (1: remember that 0 is
+ * used by the timer). */
FD_ZERO(&devices.infds);
devices.max_infd = -1;
- devices.lastdev = &devices.dev;
+ devices.lastdev = NULL;
devices.next_irq = 1;
+ 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
* line. So we quickly look through the arguments to find the amount
case 'b':
setup_block_file(optarg);
break;
+ case 'r':
+ setup_rng();
+ break;
case 'i':
initrd_name = optarg;
break;
/* We always have a console device */
setup_console();
+ /* We can timeout waiting for Guest network transmit. */
+ setup_timeout();
+
/* Now we load the kernel */
start = load_kernel(open_or_die(argv[optind+1], O_RDONLY));
initrd_size = load_initrd(initrd_name, mem);
/* These are the location in the Linux boot header where the
* start and size of the initrd are expected to be found. */
- *(unsigned long *)(boot+0x218) = mem - initrd_size;
- *(unsigned long *)(boot+0x21c) = initrd_size;
+ boot->hdr.ramdisk_image = mem - initrd_size;
+ boot->hdr.ramdisk_size = initrd_size;
/* The bootloader type 0xFF means "unknown"; that's OK. */
- *(unsigned char *)(boot+0x210) = 0xFF;
+ boot->hdr.type_of_loader = 0xFF;
}
- /* Set up the initial linear pagetables, starting below the initrd. */
- pgdir = setup_pagetables(mem, initrd_size);
-
/* The Linux boot header contains an "E820" memory map: ours is a
* simple, single region. */
- *(char*)(boot+E820NR) = 1;
- *((struct e820entry *)(boot+E820MAP))
- = ((struct e820entry) { 0, mem, E820_RAM });
+ boot->e820_entries = 1;
+ boot->e820_map[0] = ((struct e820entry) { 0, mem, E820_RAM });
/* The boot header contains a command line pointer: we put the command
- * line after the boot header (at address 4096) */
- *(u32 *)(boot + 0x228) = 4096;
- concat(boot + 4096, argv+optind+2);
+ * line after the boot header. */
+ boot->hdr.cmd_line_ptr = to_guest_phys(boot + 1);
+ /* We use a simple helper to copy the arguments separated by spaces. */
+ concat((char *)(boot + 1), argv+optind+2);
/* Boot protocol version: 2.07 supports the fields for lguest. */
- *(u16 *)(boot + 0x206) = 0x207;
+ boot->hdr.version = 0x207;
/* The hardware_subarch value of "1" tells the Guest it's an lguest. */
- *(u32 *)(boot + 0x23c) = 1;
+ boot->hdr.hardware_subarch = 1;
- /* Set bit 6 of the loadflags (aka. KEEP_SEGMENTS) so the entry path
- * does not try to reload segment registers. */
- *(u8 *)(boot + 0x211) |= (1 << 6);
+ /* Tell the entry path not to try to reload segment registers. */
+ boot->hdr.loadflags |= KEEP_SEGMENTS;
/* We tell the kernel to initialize the Guest: this returns the open
* /dev/lguest file descriptor. */
- lguest_fd = tell_kernel(pgdir, start);
+ lguest_fd = tell_kernel(start);
- /* We fork off a child process, which wakes the Launcher whenever one
- * of the input file descriptors needs attention. Otherwise we would
- * run the Guest until it tries to output something. */
- waker_fd = setup_waker(lguest_fd);
+ /* We clone off a thread, which wakes the Launcher whenever one of the
+ * input file descriptors needs attention. We call this the Waker, and
+ * we'll cover it in a moment. */
+ setup_waker(lguest_fd);
/* Finally, run the Guest. This doesn't return. */
run_guest(lguest_fd);