/*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.
- *
- * The only trick: the Makefile links it at a high address so it will be clear
- * of the guest memory region. It means that each Guest cannot have more than
- * about 2.5G of memory on a normally configured Host. :*/
+ * "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 <termios.h>
#include <getopt.h>
#include <zlib.h>
-/*L:110 We can ignore the 28 include files we need for this program, but I do
+#include <assert.h>
+#include <sched.h>
+#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 "asm-x86/e820.h"
/*:*/
#define PAGE_PRESENT 0x7 /* Present, RW, Execute */
#ifndef SIOCBRADDIF
#define SIOCBRADDIF 0x89a2 /* add interface to bridge */
#endif
+/* We can have up to 256 pages for devices. */
+#define DEVICE_PAGES 256
+/* 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;
-/* The top of guest physical memory. */
-static u32 top;
+/* 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
fd_set infds;
int max_infd;
+ /* Counter to assign interrupt numbers. */
+ unsigned int next_irq;
+
+ /* Counter to print out convenient device numbers. */
+ unsigned int device_num;
+
/* The descriptor page for the devices. */
- struct lguest_device_desc *descs;
+ u8 *descpage;
/* 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. */
+static struct device_list devices;
+
/* The device structure describes a single device. */
struct device
{
/* The linked-list pointer. */
struct device *next;
- /* The descriptor for this device, as mapped into the Guest. */
+
+ /* The this device's descriptor, as mapped into the Guest. */
struct lguest_device_desc *desc;
- /* The memory page(s) of this device, if any. Also mapped in Guest. */
- void *mem;
+
+ /* The name of this device, for --verbose. */
+ const char *name;
/* If handle_input is set, it wants to be called when this file
* descriptor is ready. */
int fd;
bool (*handle_input)(int fd, struct device *me);
- /* If handle_output is set, it wants to be called when the Guest sends
- * DMA to this key. */
- unsigned long watch_key;
- u32 (*handle_output)(int fd, const struct iovec *iov,
- unsigned int num, struct device *me);
+ /* 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;
};
+/* The virtqueue structure describes a queue attached to a device. */
+struct virtqueue
+{
+ struct virtqueue *next;
+
+ /* Which device owns me. */
+ struct device *dev;
+
+ /* The configuration for this queue. */
+ struct lguest_vqconfig config;
+
+ /* The actual ring of buffers. */
+ struct vring vring;
+
+ /* Last available index we saw. */
+ u16 last_avail_idx;
+
+ /* 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()
+
+/* Convert an iovec element to the given type.
+ *
+ * This is a fairly ugly trick: we need to know the size of the type and
+ * alignment requirement to check the pointer is kosher. It's also nice to
+ * have the name of the type in case we report failure.
+ *
+ * Typing those three things all the time is cumbersome and error prone, so we
+ * have a macro which sets them all up and passes to the real function. */
+#define convert(iov, type) \
+ ((type *)_convert((iov), sizeof(type), __alignof__(type), #type))
+
+static void *_convert(struct iovec *iov, size_t size, size_t align,
+ const char *name)
+{
+ if (iov->iov_len != size)
+ errx(1, "Bad iovec size %zu for %s", iov->iov_len, name);
+ if ((unsigned long)iov->iov_base % align != 0)
+ errx(1, "Bad alignment %p for %s", iov->iov_base, name);
+ 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_le32(v32) (v32)
+#define cpu_to_le64(v64) (v64)
+#define le16_to_cpu(v16) (v16)
+#define le32_to_cpu(v32) (v32)
+#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
+ * programs, it's quite boring (which is why everyone likes to hack on the
+ * kernel!). Perhaps if you make up an Lguest Drinking Game at this point, it
+ * will get you through this section. Or, maybe not.
+ *
+ * The Launcher sets up a big chunk of memory to be the Guest's "physical"
+ * memory and stores it in "guest_base". In other words, Guest physical ==
+ * Launcher virtual with an offset.
+ *
+ * This can be tough to get your head around, but usually it just means that we
+ * use these trivial conversion functions when the Guest gives us it's
+ * "physical" addresses: */
+static void *from_guest_phys(unsigned long addr)
+{
+ return guest_base + addr;
+}
+
+static unsigned long to_guest_phys(const void *addr)
+{
+ return (addr - guest_base);
+}
+
/*L:130
* Loading the Kernel.
*
return fd;
}
-/* map_zeroed_pages() takes a (page-aligned) address and a number of pages. */
-static void *map_zeroed_pages(unsigned long addr, unsigned int num)
+/* map_zeroed_pages() takes a number of pages. */
+static void *map_zeroed_pages(unsigned int num)
{
- /* We cache the /dev/zero file-descriptor so we only open it once. */
- static int fd = -1;
-
- if (fd == -1)
- fd = open_or_die("/dev/zero", O_RDONLY);
+ int fd = open_or_die("/dev/zero", O_RDONLY);
+ void *addr;
/* We use a private mapping (ie. if we write to the page, it will be
- * copied), and obviously we insist that it be mapped where we ask. */
- if (mmap((void *)addr, getpagesize() * num,
- PROT_READ|PROT_WRITE|PROT_EXEC, MAP_FIXED|MAP_PRIVATE, fd, 0)
- != (void *)addr)
- err(1, "Mmaping %u pages of /dev/zero @%p", num, (void *)addr);
-
- /* Returning the address is just a courtesy: can simplify callers. */
- return (void *)addr;
+ * copied). */
+ addr = mmap(NULL, getpagesize() * num,
+ 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;
}
-/* To find out where to start we look for the magic Guest string, which marks
- * the code we see in lguest_asm.S. This is a hack which we are currently
- * plotting to replace with the normal Linux entry point. */
-static unsigned long entry_point(void *start, void *end,
- unsigned long page_offset)
+/* Get some more pages for a device. */
+static void *get_pages(unsigned int num)
{
- void *p;
-
- /* The scan gives us the physical starting address. We want the
- * virtual address in this case, and fortunately, we already figured
- * out the physical-virtual difference and passed it here in
- * "page_offset". */
- for (p = start; p < end; p++)
- if (memcmp(p, "GenuineLguest", strlen("GenuineLguest")) == 0)
- return (long)p + strlen("GenuineLguest") + page_offset;
+ void *addr = from_guest_phys(guest_limit);
- errx(1, "Is this image a genuine lguest?");
+ guest_limit += num * getpagesize();
+ if (guest_limit > guest_max)
+ errx(1, "Not enough memory for devices");
+ return addr;
}
/* This routine is used to load the kernel or initrd. It tries mmap, but if
* by all modern binaries on Linux including the kernel.
*
* The ELF headers give *two* addresses: a physical address, and a virtual
- * address. The Guest kernel expects to be placed in memory at the physical
- * address, and the page tables set up so it will correspond to that virtual
- * address. We return the difference between the virtual and physical
- * addresses in the "page_offset" pointer.
+ * address. We use the physical address; the Guest will map itself to the
+ * virtual address.
*
* We return the starting address. */
-static unsigned long map_elf(int elf_fd, const Elf32_Ehdr *ehdr,
- unsigned long *page_offset)
+static unsigned long map_elf(int elf_fd, const Elf32_Ehdr *ehdr)
{
Elf32_Phdr phdr[ehdr->e_phnum];
unsigned int i;
- unsigned long start = -1UL, end = 0;
/* Sanity checks on the main ELF header: an x86 executable with a
* reasonable number of correctly-sized program headers. */
if (read(elf_fd, phdr, sizeof(phdr)) != sizeof(phdr))
err(1, "Reading program headers");
- /* We don't know page_offset yet. */
- *page_offset = 0;
-
/* 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)
verbose("Section %i: size %i addr %p\n",
i, phdr[i].p_memsz, (void *)phdr[i].p_paddr);
- /* We expect a simple linear address space: every segment must
- * have the same difference between virtual (p_vaddr) and
- * physical (p_paddr) address. */
- if (!*page_offset)
- *page_offset = phdr[i].p_vaddr - phdr[i].p_paddr;
- else if (*page_offset != phdr[i].p_vaddr - phdr[i].p_paddr)
- errx(1, "Page offset of section %i different", i);
-
- /* We track the first and last address we mapped, so we can
- * tell entry_point() where to scan. */
- if (phdr[i].p_paddr < start)
- start = phdr[i].p_paddr;
- if (phdr[i].p_paddr + phdr[i].p_filesz > end)
- end = phdr[i].p_paddr + phdr[i].p_filesz;
-
/* We map this section of the file at its physical address. */
- map_at(elf_fd, (void *)phdr[i].p_paddr,
+ map_at(elf_fd, from_guest_phys(phdr[i].p_paddr),
phdr[i].p_offset, phdr[i].p_filesz);
}
- return entry_point((void *)start, (void *)end, *page_offset);
+ /* The entry point is given in the ELF header. */
+ return ehdr->e_entry;
}
-/*L:170 Prepare to be SHOCKED and AMAZED. And possibly a trifle nauseated.
- *
- * We know that CONFIG_PAGE_OFFSET sets what virtual address the kernel expects
- * to be. We don't know what that option was, but we can figure it out
- * approximately by looking at the addresses in the code. I chose the common
- * case of reading a memory location into the %eax register:
- *
- * movl <some-address>, %eax
- *
- * This gets encoded as five bytes: "0xA1 <4-byte-address>". For example,
- * "0xA1 0x18 0x60 0x47 0xC0" reads the address 0xC0476018 into %eax.
- *
- * In this example can guess that the kernel was compiled with
- * CONFIG_PAGE_OFFSET set to 0xC0000000 (it's always a round number). If the
- * kernel were larger than 16MB, we might see 0xC1 addresses show up, but our
- * kernel isn't that bloated yet.
- *
- * Unfortunately, x86 has variable-length instructions, so finding this
- * particular instruction properly involves writing a disassembler. Instead,
- * we rely on statistics. We look for "0xA1" and tally the different bytes
- * which occur 4 bytes later (the "0xC0" in our example above). When one of
- * those bytes appears three times, we can be reasonably confident that it
- * forms the start of CONFIG_PAGE_OFFSET.
+/*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.
*
- * This is amazingly reliable. */
-static unsigned long intuit_page_offset(unsigned char *img, unsigned long len)
+ * 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)
{
- unsigned int i, possibilities[256] = { 0 };
+ struct boot_params boot;
+ int r;
+ /* Modern bzImages get loaded at 1M. */
+ void *p = from_guest_phys(0x100000);
- for (i = 0; i + 4 < len; i++) {
- /* mov 0xXXXXXXXX,%eax */
- if (img[i] == 0xA1 && ++possibilities[img[i+4]] > 3)
- return (unsigned long)img[i+4] << 24;
- }
- errx(1, "could not determine page offset");
-}
-
-/*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, unsigned long *page_offset)
-{
- 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 = (void *)0x100000;
+ /* 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));
- /* 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");
+ /* 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");
- verbose("Unpacked size %i addr %p\n", len, img);
+ /* Skip over the extra sectors of the header. */
+ lseek(fd, (boot.hdr.setup_sects+1) * 512, SEEK_SET);
- /* Without the ELF header, we can't tell virtual-physical gap. This is
- * CONFIG_PAGE_OFFSET, and people do actually change it. Fortunately,
- * I have a clever way of figuring it out from the code itself. */
- *page_offset = intuit_page_offset(img, len);
+ /* Now read everything into memory. in nice big chunks. */
+ while ((r = read(fd, p, 65536)) > 0)
+ p += r;
- return entry_point(img, img + len, *page_offset);
-}
-
-/*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 long *page_offset)
-{
- unsigned char c;
- int state = 0;
-
- /* 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, page_offset);
- }
- }
- 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. */
-static unsigned long load_kernel(int fd, unsigned long *page_offset)
+ * 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 it's an ELF file, it starts with "\177ELF" */
if (memcmp(hdr.e_ident, ELFMAG, SELFMAG) == 0)
- return map_elf(fd, &hdr, page_offset);
+ return map_elf(fd, &hdr);
- /* Otherwise we assume it's a bzImage, and try to unpack it */
- return load_bzimage(fd, page_offset);
+ /* Otherwise we assume it's a bzImage, and try to load it. */
+ return load_bzimage(fd);
}
/* This is a trivial little helper to align pages. Andi Kleen hated it because
/* We map the initrd at the top of memory, but mmap wants it to be
* page-aligned, so we round the size up for that. */
len = page_align(st.st_size);
- map_at(ifd, (void *)mem - len, 0, st.st_size);
+ map_at(ifd, from_guest_phys(mem - len), 0, st.st_size);
/* Once a file is mapped, you can close the file descriptor. It's a
* little odd, but quite useful. */
close(ifd);
/* We return the initrd size. */
return len;
}
-
-/* Once we know how much memory we have, and the address the Guest kernel
- * expects, we can construct simple linear page tables 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 page_offset)
-{
- u32 *pgdir, *linear;
- unsigned int mapped_pages, i, linear_pages;
- unsigned int ptes_per_page = getpagesize()/sizeof(u32);
-
- /* Ideally we map all physical memory starting at page_offset.
- * However, if page_offset is 0xC0000000 we can only map 1G of physical
- * (0xC0000000 + 1G overflows). */
- if (mem <= -page_offset)
- mapped_pages = mem/getpagesize();
- else
- mapped_pages = -page_offset/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 = (void *)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. The
- * entry representing page_offset points to the first one, and they
- * continue from there. */
- for (i = 0; i < mapped_pages; i += ptes_per_page) {
- pgdir[(i + page_offset/getpagesize())/ptes_per_page]
- = (((u32)linear + i*sizeof(u32)) | PAGE_PRESENT);
- }
-
- verbose("Linear mapping of %u pages in %u pte pages at %p\n",
- mapped_pages, linear_pages, linear);
-
- /* We return the top level (guest-physical) address: the kernel needs
- * to know where it is. */
- return (unsigned long)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 top physical page to allow, the top level pagetable, the entry point and
- * the page_offset constant for the Guest. */
-static int tell_kernel(u32 pgdir, u32 start, u32 page_offset)
+/*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)
{
- u32 args[] = { LHREQ_INITIALIZE,
- top/getpagesize(), pgdir, start, page_offset };
+ unsigned long args[] = { LHREQ_INITIALIZE,
+ (unsigned long)guest_base,
+ guest_limit / getpagesize(), start };
int fd;
+ verbose("Guest: %p - %p (%#lx)\n",
+ guest_base, guest_base + guest_limit, guest_limit);
fd = open_or_die("/dev/lguest", O_RDWR);
if (write(fd, args, sizeof(args)) < 0)
err(1, "Writing to /dev/lguest");
}
/*:*/
-static void set_fd(int fd, struct device_list *devices)
+static void add_device_fd(int fd)
{
- FD_SET(fd, &devices->infds);
- if (fd > devices->max_infd)
- devices->max_infd = fd;
+ FD_SET(fd, &devices.infds);
+ if (fd > devices.max_infd)
+ devices.max_infd = fd;
}
/*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, struct device_list *devices)
+static int waker(void *unused)
{
- /* Add the pipe from the Launcher to the fdset in the device_list, so
- * we watch it, too. */
- set_fd(pipefd, devices);
+ /* Close the write end of the pipe: only the Launcher has it open. */
+ close(waker_fds.pipe[1]);
for (;;) {
- fd_set rfds = devices->infds;
- u32 args[] = { LHREQ_BREAK, 1 };
+ 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 ignorefd;
- /* If read() returns 0, it means the Launcher has
- * exited. We silently follow. */
- if (read(pipefd, &ignorefd, sizeof(ignorefd)) == 0)
- exit(0);
- /* Otherwise it's telling us there's a problem with one
- * of the devices, and we should ignore that file
- * descriptor from now on. */
- FD_CLR(ignorefd, &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, struct device_list *device_list)
+static void setup_waker(int lguest_fd)
{
- int pipefd[2], child;
-
- /* 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");
+ /* This pipe is closed when Launcher dies, telling Waker. */
+ if (pipe(waker_fds.pipe) != 0)
+ err(1, "Creating pipe for Waker");
- if (child == 0) {
- /* Close the "writing" end of our copy of the pipe */
- close(pipefd[1]);
- wake_parent(pipefd[0], lguest_fd, device_list);
- }
- /* 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,
{
/* We have to separately check addr and addr+size, because size could
* be huge and addr + size might wrap around. */
- if (addr >= top || addr + size >= top)
- errx(1, "%s:%i: Invalid address %li", __FILE__, line, addr);
+ if (addr >= guest_limit || addr + size >= guest_limit)
+ errx(1, "%s:%i: Invalid address %#lx", __FILE__, line, addr);
/* We return a pointer for the caller's convenience, now we know it's
* safe to use. */
- return (void *)addr;
+ return from_guest_phys(addr);
}
/* A macro which transparently hands the line number to the real function. */
#define check_pointer(addr,size) _check_pointer(addr, size, __LINE__)
-/* The Guest has given us the address of a "struct lguest_dma". We check it's
- * OK and convert it to an iovec (which is a simple array of ptr/size
- * pairs). */
-static u32 *dma2iov(unsigned long dma, struct iovec iov[], unsigned *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 i;
- struct lguest_dma *udma;
-
- /* First we make sure that the array memory itself is valid. */
- udma = check_pointer(dma, sizeof(*udma));
- /* Now we check each element */
- for (i = 0; i < LGUEST_MAX_DMA_SECTIONS; i++) {
- /* A zero length ends the array. */
- if (!udma->len[i])
- break;
+ unsigned int next;
- iov[i].iov_base = check_pointer(udma->addr[i], udma->len[i]);
- iov[i].iov_len = udma->len[i];
- }
- *num = i;
+ /* If this descriptor says it doesn't chain, we're done. */
+ if (!(vq->vring.desc[i].flags & VRING_DESC_F_NEXT))
+ return vq->vring.num;
+
+ /* Check they're not leading us off end of descriptors. */
+ next = vq->vring.desc[i].next;
+ /* Make sure compiler knows to grab that: we don't want it changing! */
+ wmb();
- /* We return the pointer to where the caller should write the amount of
- * the buffer used. */
- return &udma->used_len;
+ if (next >= vq->vring.num)
+ errx(1, "Desc next is %u", next);
+
+ return next;
}
-/* This routine gets a DMA buffer from the Guest for a given key, and converts
- * it to an iovec array. It returns the interrupt the Guest wants when we're
- * finished, and a pointer to the "used_len" field to fill in. */
-static u32 *get_dma_buffer(int fd, void *key,
- struct iovec iov[], unsigned int *num, u32 *irq)
+/* This looks in the virtqueue and 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, or vq->vring.num (which
+ * is never a valid descriptor number) if none was found. */
+static unsigned get_vq_desc(struct virtqueue *vq,
+ struct iovec iov[],
+ unsigned int *out_num, unsigned int *in_num)
{
- u32 buf[] = { LHREQ_GETDMA, (u32)key };
- unsigned long udma;
- u32 *res;
+ unsigned int i, head;
+ u16 last_avail;
+
+ /* Check it isn't doing very strange things with descriptor numbers. */
+ 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",
+ last_avail, vq->vring.avail->idx);
+
+ /* If there's nothing new since last we looked, return invalid. */
+ 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[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, "Guest says index %u is available", head);
+
+ /* When we start there are none of either input nor output. */
+ *out_num = *in_num = 0;
+
+ i = head;
+ do {
+ /* Grab the first descriptor, and check it's OK. */
+ iov[*out_num + *in_num].iov_len = vq->vring.desc[i].len;
+ iov[*out_num + *in_num].iov_base
+ = check_pointer(vq->vring.desc[i].addr,
+ vq->vring.desc[i].len);
+ /* If this is an input descriptor, increment that count. */
+ if (vq->vring.desc[i].flags & VRING_DESC_F_WRITE)
+ (*in_num)++;
+ else {
+ /* If it's an output descriptor, they're all supposed
+ * to come before any input descriptors. */
+ if (*in_num)
+ errx(1, "Descriptor has out after in");
+ (*out_num)++;
+ }
- /* Ask the kernel for a DMA buffer corresponding to this key. */
- udma = write(fd, buf, sizeof(buf));
- /* They haven't registered any, or they're all used? */
- if (udma == (unsigned long)-1)
- return NULL;
+ /* If we've got too many, that implies a descriptor loop. */
+ if (*out_num + *in_num > vq->vring.num)
+ errx(1, "Looped descriptor");
+ } while ((i = next_desc(vq, i)) != vq->vring.num);
- /* Convert it into our iovec array */
- res = dma2iov(udma, iov, num);
- /* The kernel stashes irq in ->used_len to get it out to us. */
- *irq = *res;
- /* Return a pointer to ((struct lguest_dma *)udma)->used_len. */
- return res;
+ vq->inflight++;
+ return head;
}
-/* This is a convenient routine to send the Guest an interrupt. */
-static void trigger_irq(int fd, u32 irq)
+/* 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)
{
- u32 buf[] = { LHREQ_IRQ, irq };
+ struct vring_used_elem *used;
+
+ /* 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 */
+static void trigger_irq(int fd, struct virtqueue *vq)
+{
+ unsigned long buf[] = { LHREQ_IRQ, vq->config.irq };
+
+ /* 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. */
if (write(fd, buf, sizeof(buf)) != 0)
- err(1, "Triggering irq %i", irq);
+ err(1, "Triggering irq %i", vq->config.irq);
}
-/* This simply sets up an iovec array where we can put data to be discarded.
- * This happens when the Guest doesn't want or can't handle the input: we have
- * to get rid of it somewhere, and if we bury it in the ceiling space it will
- * start to smell after a week. */
-static void discard_iovec(struct iovec *iov, unsigned int *num)
+/* And here's the combo meal deal. Supersize me! */
+static void add_used_and_trigger(int fd, struct virtqueue *vq,
+ unsigned int head, int len)
{
- static char discard_buf[1024];
- *num = 1;
- iov->iov_base = discard_buf;
- iov->iov_len = sizeof(discard_buf);
+ add_used(vq, head, len);
+ 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)
{
/* This is the routine which handles console input (ie. stdin). */
static bool handle_console_input(int fd, struct device *dev)
{
- u32 irq = 0, *lenp;
int len;
- unsigned int num;
- struct iovec iov[LGUEST_MAX_DMA_SECTIONS];
+ unsigned int head, in_num, out_num;
+ struct iovec iov[dev->vq->vring.num];
struct console_abort *abort = dev->priv;
- /* First we get the console buffer from the Guest. The key is dev->mem
- * which was set to 0 in setup_console(). */
- lenp = get_dma_buffer(fd, dev->mem, iov, &num, &irq);
- if (!lenp) {
- /* If it's not ready for input, warn and set up to discard. */
- warn("console: no dma buffer!");
- discard_iovec(iov, &num);
- }
+ /* First we need a console buffer from the Guests's input 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 console in queue?");
/* This is why we convert to iovecs: the readv() call uses them, and so
* it reads straight into the Guest's buffer. */
- len = readv(dev->fd, iov, num);
+ len = readv(dev->fd, iov, in_num);
if (len <= 0) {
/* This implies that the console is closed, is /dev/null, or
- * something went terribly wrong. We still go through the rest
- * of the logic, though, especially the exit handling below. */
+ * something went terribly wrong. */
warnx("Failed to get console input, ignoring console.");
- len = 0;
+ /* Put the input terminal back. */
+ restore_term();
+ /* Remove callback from input vq, so it doesn't restart us. */
+ dev->vq->handle_output = NULL;
+ /* Stop listening to this fd: don't call us again. */
+ return false;
}
- /* If we read the data into the Guest, fill in the length and send the
- * interrupt. */
- if (lenp) {
- *lenp = len;
- trigger_irq(fd, irq);
- }
+ /* Tell the Guest about the new input. */
+ add_used_and_trigger(fd, dev->vq, head, len);
/* Three ^C within one second? Exit.
*
struct timeval now;
gettimeofday(&now, NULL);
if (now.tv_sec <= abort->start.tv_sec+1) {
- u32 args[] = { LHREQ_BREAK, 0 };
+ 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);
/* Any other key resets the abort counter. */
abort->count = 0;
- /* Now, if we didn't read anything, put the input terminal back and
- * return failure (meaning, don't call us again). */
- if (!len) {
- restore_term();
- return false;
- }
/* Everything went OK! */
return true;
}
-/* Handling console output is much simpler than input. */
-static u32 handle_console_output(int fd, const struct iovec *iov,
- unsigned num, struct device*dev)
+/* 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, bool timeout)
{
- /* Whatever the Guest sends, write it to standard output. Return the
- * number of bytes written. */
- return writev(STDOUT_FILENO, iov, num);
+ unsigned int head, out, in;
+ int len;
+ struct iovec iov[vq->vring.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?");
+ len = writev(STDOUT_FILENO, iov, out);
+ add_used_and_trigger(fd, vq, head, len);
+ }
}
-/* Guest->Host network output is also pretty easy. */
-static u32 handle_tun_output(int fd, const struct iovec *iov,
- unsigned num, struct device *dev)
+/* 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)
{
- /* We put a flag in the "priv" pointer of the network device, and set
- * it as soon as we see output. We'll see why in handle_tun_input() */
- *(bool *)dev->priv = true;
- /* Whatever packet the Guest sent us, write it out to the tun
- * device. */
- return writev(dev->fd, iov, num);
+ 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);
}
-/* This matches the peer_key() in lguest_net.c. The key for any given slot
- * is the address of the network device's page plus 4 * the slot number. */
-static unsigned long peer_offset(unsigned int peernum)
+/*
+ * 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
+ * (/dev/net/tun).
+ */
+static void handle_net_output(int fd, struct virtqueue *vq, bool timeout)
{
- return 4 * peernum;
+ 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?");
+ 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;
+ }
}
-/* This is where we handle a packet coming in from the tun device */
+/* This is where we handle a packet coming in from the tun device to our
+ * Guest. */
static bool handle_tun_input(int fd, struct device *dev)
{
- u32 irq = 0, *lenp;
+ unsigned int head, in_num, out_num;
int len;
- unsigned num;
- struct iovec iov[LGUEST_MAX_DMA_SECTIONS];
+ struct iovec iov[dev->vq->vring.num];
- /* First we get a buffer the Guest has bound to its key. */
- lenp = get_dma_buffer(fd, dev->mem+peer_offset(NET_PEERNUM), iov, &num,
- &irq);
- if (!lenp) {
+ /* First we need a network buffer from the Guests's recv virtqueue. */
+ head = get_vq_desc(dev->vq, iov, &out_num, &in_num);
+ if (head == dev->vq->vring.num) {
/* Now, it's expected that if we try to send a packet too
- * early, the Guest won't be ready yet. This is why we set a
- * flag when the Guest sends its first packet. If it's sent a
- * packet we assume it should be ready to receive them.
- *
- * Actually, this is what the status bits in the descriptor are
- * for: we should *use* them. FIXME! */
- if (*(bool *)dev->priv)
- warn("network: no dma buffer!");
- discard_iovec(iov, &num);
- }
+ * early, the Guest won't be ready yet. Wait until the device
+ * status says it's ready. */
+ /* FIXME: Actually want DRIVER_ACTIVE here. */
+
+ /* 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?");
/* Read the packet from the device directly into the Guest's buffer. */
- len = readv(dev->fd, iov, num);
+ len = readv(dev->fd, iov, in_num);
if (len <= 0)
err(1, "reading network");
- /* Write the used_len, and trigger the interrupt for the Guest */
- if (lenp) {
- *lenp = len;
- trigger_irq(fd, irq);
- }
+ /* Tell the Guest about the new packet. */
+ add_used_and_trigger(fd, dev->vq, head, len);
+
verbose("tun input packet len %i [%02x %02x] (%s)\n", len,
- ((u8 *)iov[0].iov_base)[0], ((u8 *)iov[0].iov_base)[1],
- lenp ? "sent" : "discarded");
+ ((u8 *)iov[1].iov_base)[0], ((u8 *)iov[1].iov_base)[1],
+ head != dev->vq->vring.num ? "sent" : "discarded");
+
/* All good. */
return true;
}
-/* The last device handling routine is block output: the Guest has sent a DMA
- * to the block device. It will have placed the command it wants in the
- * "struct lguest_block_page". */
-static u32 handle_block_output(int fd, const struct iovec *iov,
- unsigned num, struct device *dev)
-{
- struct lguest_block_page *p = dev->mem;
- u32 irq, *lenp;
- unsigned int len, reply_num;
- struct iovec reply[LGUEST_MAX_DMA_SECTIONS];
- off64_t device_len, off = (off64_t)p->sector * 512;
-
- /* First we extract the device length from the dev->priv pointer. */
- device_len = *(off64_t *)dev->priv;
-
- /* We first check that the read or write is within the length of the
- * block file. */
- if (off >= device_len)
- errx(1, "Bad offset %llu vs %llu", off, device_len);
- /* Move to the right location in the block file. This shouldn't fail,
- * but best to check. */
- if (lseek64(dev->fd, off, SEEK_SET) != off)
- err(1, "Bad seek to sector %i", p->sector);
-
- verbose("Block: %s at offset %llu\n", p->type ? "WRITE" : "READ", off);
-
- /* They were supposed to bind a reply buffer at key equal to the start
- * of the block device memory. We need this to tell them when the
- * request is finished. */
- lenp = get_dma_buffer(fd, dev->mem, reply, &reply_num, &irq);
- if (!lenp)
- err(1, "Block request didn't give us a dma buffer");
-
- if (p->type) {
- /* A write request. The DMA they sent contained the data, so
- * write it out. */
- len = writev(dev->fd, iov, num);
- /* Grr... Now we know how long the "struct lguest_dma" they
- * sent was, we make sure they didn't try to write over the end
- * of the block file (possibly extending it). */
- if (off + len > device_len) {
- /* Trim it back to the correct length */
- ftruncate64(dev->fd, device_len);
- /* Die, bad Guest, die. */
- errx(1, "Write past end %llu+%u", off, len);
- }
- /* The reply length is 0: we just send back an empty DMA to
- * interrupt them and tell them the write is finished. */
- *lenp = 0;
- } else {
- /* A read request. They sent an empty DMA to start the
- * request, and we put the read contents into the reply
- * buffer. */
- len = readv(dev->fd, reply, reply_num);
- *lenp = len;
- }
+/*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, bool timeout)
+{
+ add_device_fd(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);
+}
- /* The result is 1 (done), 2 if there was an error (short read or
- * write). */
- p->result = 1 + (p->bytes != len);
- /* Now tell them we've used their reply buffer. */
- trigger_irq(fd, irq);
+/* When the Guest tells us they updated the status field, we handle it. */
+static void update_device_status(struct device *dev)
+{
+ struct virtqueue *vq;
- /* We're supposed to return the number of bytes of the output buffer we
- * used. But the block device uses the "result" field instead, so we
- * don't bother. */
- return 0;
+ /* 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 sends some DMA out. */
-static void handle_output(int fd, unsigned long dma, unsigned long key,
- struct device_list *devices)
+/* This is the generic routine we call when the Guest uses LHCALL_NOTIFY. */
+static void handle_output(int fd, unsigned long addr)
{
struct device *i;
- u32 *lenp;
- struct iovec iov[LGUEST_MAX_DMA_SECTIONS];
- unsigned num = 0;
-
- /* Convert the "struct lguest_dma" they're sending to a "struct
- * iovec". */
- lenp = dma2iov(dma, iov, &num);
-
- /* Check each device: if they expect output to this key, tell them to
- * handle it. */
- for (i = devices->dev; i; i = i->next) {
- if (i->handle_output && key == i->watch_key) {
- /* We write the result straight into the used_len field
- * for them. */
- *lenp = i->handle_output(fd, iov, num, i);
+ struct virtqueue *vq;
+
+ /* 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())
+ 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;
}
}
- /* This can happen: the kernel sends any SEND_DMA which doesn't match
- * another Guest to us. It could be that another Guest just left a
- * network, for example. But it's unusual. */
- warnx("Pending dma %p, key %p", (void *)dma, (void *)key);
+ /* Early console write is done using notify on a nul-terminated string
+ * in Guest memory. */
+ if (addr >= guest_limit)
+ errx(1, "Bad NOTIFY %#lx", addr);
+
+ write(STDOUT_FILENO, from_guest_phys(addr),
+ strnlen(from_guest_phys(addr), guest_limit - addr));
+}
+
+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
+/* This is called when the Waker wakes us up: check for incoming file
* descriptors. */
-static void handle_input(int fd, struct device_list *devices)
+static void handle_input(int fd)
{
/* select() wants a zeroed timeval to mean "don't wait". */
struct timeval poll = { .tv_sec = 0, .tv_usec = 0 };
for (;;) {
struct device *i;
- fd_set fds = devices->infds;
+ 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. */
- for (i = devices->dev; i; i = i->next) {
+ /* 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)) {
+ if (i->handle_input(fd, i))
+ continue;
+
/* If handle_input() returns false, it means we
- * should no longer service it.
- * handle_console_input() does this. */
- if (!i->handle_input(fd, i)) {
- /* Clear it from the set of input file
- * descriptors kept at the head of the
- * device list. */
- FD_CLR(i->fd, &devices->infds);
- /* Tell waker to ignore it too... */
- write(waker_fd, &i->fd, sizeof(i->fd));
- }
+ * 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. */
+ FD_CLR(i->fd, &devices.infds);
}
}
+
+ /* 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 in the devices array just above the Guest's normal memory. */
-static struct lguest_device_desc *
-new_dev_desc(struct lguest_device_desc *descs,
- u16 type, u16 features, u16 num_pages)
+ * routines to allocate and manage them.
+ */
+
+/* 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)
{
- unsigned int i;
+ return (void *)(dev->desc + 1)
+ + dev->desc->num_vq * sizeof(struct lguest_vqconfig)
+ + dev->desc->feature_len * 2;
+}
- for (i = 0; i < LGUEST_MAX_DEVICES; i++) {
- if (!descs[i].type) {
- descs[i].type = type;
- descs[i].features = features;
- descs[i].num_pages = num_pages;
- /* If they said the device needs memory, we allocate
- * that now, bumping up the top of Guest memory. */
- if (num_pages) {
- map_zeroed_pages(top, num_pages);
- descs[i].pfn = top/getpagesize();
- top += num_pages*getpagesize();
- }
- return &descs[i];
- }
+/* 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)
+{
+ struct lguest_device_desc d = { .type = type };
+ void *p;
+
+ /* 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;
+
+ /* We only have one page for all the descriptors. */
+ if (p + sizeof(d) > (void *)devices.descpage + getpagesize())
+ errx(1, "Too many devices");
+
+ /* p might not be aligned, so we memcpy in. */
+ return memcpy(p, &d, sizeof(d));
+}
+
+/* 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, struct virtqueue *, bool))
+{
+ unsigned int pages;
+ struct virtqueue **i, *vq = malloc(sizeof(*vq));
+ void *p;
+
+ /* 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, LGUEST_VRING_ALIGN);
+
+ /* 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;
+
+ /* 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;
}
- errx(1, "too many devices");
-}
-
-/* This monster routine does all the creation and setup of a new device,
- * including caling new_dev_desc() to allocate the descriptor and device
- * memory. */
-static struct device *new_device(struct device_list *devices,
- u16 type, u16 num_pages, u16 features,
- int fd,
- bool (*handle_input)(int, struct device *),
- unsigned long watch_off,
- u32 (*handle_output)(int,
- const struct iovec *,
- unsigned,
- struct device *))
+
+ 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)
{
- struct device *dev = malloc(sizeof(*dev));
+ /* Check we haven't overflowed our single page. */
+ if (device_config(dev) + len > devices.descpage + getpagesize())
+ errx(1, "Too many devices");
- /* 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;
+ /* 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
+ * 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));
/* 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
* to the device_list's fdset and maxfd. */
if (handle_input)
- set_fd(dev->fd, devices);
- dev->desc = new_dev_desc(devices->descs, type, features, num_pages);
- dev->mem = (void *)(dev->desc->pfn * getpagesize());
+ add_device_fd(dev->fd);
+ dev->desc = new_dev_desc(type);
dev->handle_input = handle_input;
- dev->watch_key = (unsigned long)dev->mem + watch_off;
- dev->handle_output = handle_output;
+ 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;
}
/* Our first setup routine is the console. It's a fairly simple device, but
* UNIX tty handling makes it uglier than it could be. */
-static void setup_console(struct device_list *devices)
+static void setup_console(void)
{
struct device *dev;
atexit(restore_term);
}
- /* We don't currently require any memory for the console, so we ask for
- * 0 pages. */
- dev = new_device(devices, LGUEST_DEVICE_T_CONSOLE, 0, 0,
- STDIN_FILENO, handle_console_input,
- LGUEST_CONSOLE_DMA_KEY, handle_console_output);
+ dev = new_device("console", VIRTIO_ID_CONSOLE,
+ STDIN_FILENO, handle_console_input);
/* We store the console state in dev->priv, and initialize it. */
dev->priv = malloc(sizeof(struct console_abort));
((struct console_abort *)dev->priv)->count = 0;
- verbose("device %p: console\n",
- (void *)(dev->desc->pfn * getpagesize()));
-}
-/* Setting up a block file is also fairly straightforward. */
-static void setup_block_file(const char *filename, struct device_list *devices)
-{
- int fd;
- struct device *dev;
- off64_t *device_len;
- struct lguest_block_page *p;
-
- /* We open with O_LARGEFILE because otherwise we get stuck at 2G. We
- * open with O_DIRECT because otherwise our benchmarks go much too
- * fast. */
- fd = open_or_die(filename, O_RDWR|O_LARGEFILE|O_DIRECT);
-
- /* We want one page, and have no input handler (the block file never
- * has anything interesting to say to us). Our timing will be quite
- * random, so it should be a reasonable randomness source. */
- dev = new_device(devices, LGUEST_DEVICE_T_BLOCK, 1,
- LGUEST_DEVICE_F_RANDOMNESS,
- fd, NULL, 0, handle_block_output);
-
- /* We store the device size in the private area */
- device_len = dev->priv = malloc(sizeof(*device_len));
- /* This is the safe way of establishing the size of our device: it
- * might be a normal file or an actual block device like /dev/hdb. */
- *device_len = lseek64(fd, 0, SEEK_END);
-
- /* The device memory is a "struct lguest_block_page". It's zeroed
- * already, we just need to put in the device size. Block devices
- * think in sectors (ie. 512 byte chunks), so we translate here. */
- p = dev->mem;
- p->num_sectors = *device_len/512;
- verbose("device %p: block %i sectors\n",
- (void *)(dev->desc->pfn * getpagesize()), p->num_sectors);
+ /* 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. */
+ add_virtqueue(dev, VIRTQUEUE_NUM, enable_fd);
+ add_virtqueue(dev, VIRTQUEUE_NUM, handle_console_output);
+
+ verbose("device %u: console\n", devices.device_num++);
}
+/*:*/
-/*
- * Network Devices.
- *
- * Setting up network devices is quite a pain, because we have three types.
- * First, we have the inter-Guest network. This is a file which is mapped into
- * the address space of the Guests who are on the network. Because it is a
- * shared mapping, the same page underlies all the devices, and they can send
- * DMA to each other.
- *
- * Remember from our network driver, the Guest is told what slot in the page it
- * is to use. We use exclusive fnctl locks to reserve a slot. If another
- * Guest is using a slot, the lock will fail and we try another. Because fnctl
- * locks are cleaned up automatically when we die, this cleverly means that our
- * reservation on the slot will vanish if we crash. */
-static unsigned int find_slot(int netfd, const char *filename)
-{
- struct flock fl;
-
- fl.l_type = F_WRLCK;
- fl.l_whence = SEEK_SET;
- fl.l_len = 1;
- /* Try a 1 byte lock in each possible position number */
- for (fl.l_start = 0;
- fl.l_start < getpagesize()/sizeof(struct lguest_net);
- fl.l_start++) {
- /* If we succeed, return the slot number. */
- if (fcntl(netfd, F_SETLK, &fl) == 0)
- return fl.l_start;
- }
- errx(1, "No free slots in network file %s", filename);
+static void timeout_alarm(int sig)
+{
+ write(timeoutpipe[1], "", 1);
}
-/* This function sets up the network file */
-static void setup_net_file(const char *filename,
- struct device_list *devices)
+static void setup_timeout(void)
{
- int netfd;
- struct device *dev;
+ if (pipe(timeoutpipe) != 0)
+ err(1, "Creating timeout pipe");
- /* We don't use open_or_die() here: for friendliness we create the file
- * if it doesn't already exist. */
- netfd = open(filename, O_RDWR, 0);
- if (netfd < 0) {
- if (errno == ENOENT) {
- netfd = open(filename, O_RDWR|O_CREAT, 0600);
- if (netfd >= 0) {
- /* If we succeeded, initialize the file with a
- * blank page. */
- char page[getpagesize()];
- memset(page, 0, sizeof(page));
- write(netfd, page, sizeof(page));
- }
- }
- if (netfd < 0)
- err(1, "cannot open net file '%s'", filename);
- }
+ if (fcntl(timeoutpipe[1], F_SETFL,
+ fcntl(timeoutpipe[1], F_GETFL) | O_NONBLOCK) != 0)
+ err(1, "Making timeout pipe nonblocking");
- /* We need 1 page, and the features indicate the slot to use and that
- * no checksum is needed. We never touch this device again; it's
- * between the Guests on the network, so we don't register input or
- * output handlers. */
- dev = new_device(devices, LGUEST_DEVICE_T_NET, 1,
- find_slot(netfd, filename)|LGUEST_NET_F_NOCSUM,
- -1, NULL, 0, NULL);
-
- /* Map the shared file. */
- if (mmap(dev->mem, getpagesize(), PROT_READ|PROT_WRITE,
- MAP_FIXED|MAP_SHARED, netfd, 0) != dev->mem)
- err(1, "could not mmap '%s'", filename);
- verbose("device %p: shared net %s, peer %i\n",
- (void *)(dev->desc->pfn * getpagesize()), filename,
- dev->desc->features & ~LGUEST_NET_F_NOCSUM);
+ 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.
+ *
+ * More sopisticated is to use one of the tools developed for project like UML
+ * to do networking.
+ *
+ * Faster is to do virtio bonding in kernel. Doing this 1:1 would be
+ * completely generic ("here's my vring, attach to your vring") and would work
+ * for any traffic. Of course, namespace and permissions issues need to be
+ * dealt with. A more sophisticated "multi-channel" virtio_net.c could hide
+ * multiple inter-guest channels behind one interface, although it would
+ * require some manner of hotplugging new virtio channels.
+ *
+ * Finally, we could implement a virtio network switch in the kernel. :*/
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 (in practice, the Host's slot in the network device's memory). */
-static void configure_device(int fd, const char *devname, u32 ipaddr,
- unsigned char hwaddr[6])
+ * pointer. */
+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 The other kind of network is a Host<->Guest network. This can either
- * use briding 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, struct device_list *devices)
+static int get_tun_device(char tapif[IFNAMSIZ])
{
- struct device *dev;
struct ifreq ifr;
- int netfd, ipfd;
- u32 ip;
- const char *br_name = NULL;
+ 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);
- /* We create the net device with 1 page, using the features field of
- * the descriptor to tell the Guest it is in slot 1 (NET_PEERNUM), and
- * that the device has fairly random timing. We do *not* specify
- * LGUEST_NET_F_NOCSUM: these packets can reach the real world.
- *
- * We will put our MAC address is slot 0 for the Guest to see, so
- * it will send packets to us using the key "peer_offset(0)": */
- dev = new_device(devices, LGUEST_DEVICE_T_NET, 1,
- NET_PEERNUM|LGUEST_DEVICE_F_RANDOMNESS, netfd,
- handle_tun_input, peer_offset(0), handle_tun_output);
+ memcpy(tapif, ifr.ifr_name, IFNAMSIZ);
+ return netfd;
+}
- /* We keep a flag which says whether we've seen packets come out from
- * this network device. */
- dev->priv = malloc(sizeof(bool));
- *(bool *)dev->priv = false;
+/*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, 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
* tap interface, connect to the bridge etc. Any socket will do! */
/* 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;
+ }
- /* We are peer 0, ie. first slot, so we hand dev->mem to this routine
- * to write the MAC address at the start of the device memory. */
- configure_device(ipfd, ifr.ifr_name, ip, dev->mem);
+ /* 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';
+ }
- /* Set "promisc" bit: we want every single packet if we're going to
- * bridge to other machines (and otherwise it doesn't matter). */
- *((u8 *)dev->mem) |= 0x1;
+ /* arg is now either an IP address or a bridge name */
+ if (bridging)
+ add_to_bridge(ipfd, tapif, arg);
+ else
+ ip = str2ip(arg);
+ /* 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 %p: tun net %u.%u.%u.%u\n",
- (void *)(dev->desc->pfn * getpagesize()),
- (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);
+}
+
+/* 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 we farm the I/O out to thread, and communicate with it via a pipe. */
+
+/* This hangs off device->priv. */
+struct vblk_info
+{
+ /* The size of the file. */
+ off64_t len;
+
+ /* The file descriptor for the file. */
+ int fd;
+
+ /* IO thread listens on this file descriptor [0]. */
+ int workpipe[2];
+
+ /* IO thread writes to this file descriptor to mark it done, then
+ * Launcher triggers interrupt to Guest. */
+ int done_fd;
+};
+
+/*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;
+ 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], u8);
+ off = out->sector * 512;
+
+ /* 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 = VIRTIO_BLK_S_UNSUPP;
+ wlen = sizeof(*in);
+ } else if (out->type & VIRTIO_BLK_T_OUT) {
+ /* Write */
+
+ /* Move to the right location in the block file. This can fail
+ * if they try to write past end. */
+ if (lseek64(vblk->fd, off, SEEK_SET) != off)
+ err(1, "Bad seek to sector %llu", out->sector);
+
+ ret = writev(vblk->fd, iov+1, out_num-1);
+ verbose("WRITE to sector %llu: %i\n", out->sector, ret);
+
+ /* Grr... Now we know how long the descriptor they sent was, we
+ * make sure they didn't try to write over the end of the block
+ * file (possibly extending it). */
+ if (ret > 0 && off + ret > vblk->len) {
+ /* Trim it back to the correct length */
+ ftruncate64(vblk->fd, vblk->len);
+ /* Die, bad Guest, die. */
+ errx(1, "Write past end %llu+%u", off, ret);
+ }
+ wlen = sizeof(*in);
+ *in = (ret >= 0 ? VIRTIO_BLK_S_OK : VIRTIO_BLK_S_IOERR);
+ } else {
+ /* Read */
+
+ /* Move to the right location in the block file. This can fail
+ * if they try to read past end. */
+ if (lseek64(vblk->fd, off, SEEK_SET) != off)
+ err(1, "Bad seek to sector %llu", out->sector);
+
+ 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 = VIRTIO_BLK_S_OK;
+ } else {
+ 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);
+ return true;
+}
+
+/* This is the thread which actually services the I/O. */
+static int io_thread(void *_dev)
+{
+ struct device *dev = _dev;
+ struct vblk_info *vblk = dev->priv;
+ char c;
+
+ /* Close other side of workpipe so we get 0 read when main dies. */
+ close(vblk->workpipe[1]);
+ /* Close the other side of the done_fd pipe. */
+ close(dev->fd);
+
+ /* 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,
+ * rather than waiting until we've done them all. I haven't
+ * 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;
+}
+
+/* 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 the I/O thread died, presumably it printed the error, so we
+ * simply exit. */
+ if (read(dev->fd, &c, 1) != 1)
+ exit(1);
+
+ /* It did some work, so trigger the irq. */
+ trigger_irq(fd, dev->vq);
+ return true;
+}
+
+/* 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;
+
+ /* Wake up I/O thread and tell it to go to work! */
+ if (write(vblk->workpipe[1], &c, 1) != 1)
+ /* Presumably it indicated why it died. */
+ exit(1);
+}
+
+/*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;
+ 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 one virtqueue, where the Guest places requests. */
+ add_virtqueue(dev, VIRTQUEUE_NUM, handle_virtblk_output);
+
+ /* Allocate the room for our own bookkeeping */
+ vblk = dev->priv = malloc(sizeof(*vblk));
+
+ /* First we open the file and store the length. */
+ 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. */
+ 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. */
+ 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 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. Since stack grows upwards, we
+ * point the stack pointer to the end of this region. */
+ stack = malloc(32768);
+ /* 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->done_fd);
+ close(vblk->workpipe[0]);
+
+ verbose("device %u: virtblock %llu sectors\n",
+ 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, struct device_list *device_list)
+static void __attribute__((noreturn)) run_guest(int lguest_fd)
{
for (;;) {
- u32 args[] = { LHREQ_BREAK, 0 };
- unsigned long arr[2];
+ unsigned long args[] = { LHREQ_BREAK, 0 };
+ unsigned long notify_addr;
int readval;
/* We read from the /dev/lguest device to run the Guest. */
- readval = read(lguest_fd, arr, sizeof(arr));
+ readval = pread(lguest_fd, ¬ify_addr,
+ sizeof(notify_addr), cpu_id);
- /* The read can only really return sizeof(arr) (the Guest did a
- * SEND_DMA to us), or an error. */
-
- /* For a successful read, arr[0] is the address of the "struct
- * lguest_dma", and arr[1] is the key the Guest sent to. */
- if (readval == sizeof(arr)) {
- handle_output(lguest_fd, arr[0], arr[1], device_list);
+ /* One unsigned long means the Guest did HCALL_NOTIFY */
+ if (readval == sizeof(notify_addr)) {
+ verbose("Notify on address %#lx\n", notify_addr);
+ handle_output(lguest_fd, notify_addr);
continue;
/* 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. */
- handle_input(lguest_fd, device_list);
- if (write(lguest_fd, args, sizeof(args)) < 0)
+ /* 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 (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' },
- { "sharenet", 1, NULL, 's' },
{ "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] "
- "[--sharenet=<filename>|--tunnet=(<ipaddr>|bridge:<bridgename>)\n"
+ "[--tunnet=(<ipaddr>:<macaddr>|bridge:<bridgename>:<macaddr>)\n"
"|--block=<filename>|--initrd=<filename>]...\n"
"<mem-in-mb> vmlinux [args...]");
}
-/*L:100 The Launcher code itself takes us out into userspace, that scary place
- * where pointers run wild and free! Unfortunately, like most userspace
- * programs, it's quite boring (which is why everyone like to hack on the
- * kernel!). Perhaps if you make up an Lguest Drinking Game at this point, it
- * will get you through this section. Or, maybe not.
- *
- * The Launcher binary sits up high, usually starting at address 0xB8000000.
- * Everything below this is the "physical" memory for the Guest. For example,
- * if the Guest were to write a "1" at physical address 0, we would see a "1"
- * in the Launcher at "(int *)0". Guest physical == Launcher virtual.
- *
- * This can be tough to get your head around, but usually it just means that we
- * don't need to do any conversion when the Guest gives us it's "physical"
- * addresses.
- */
+/*L:105 The main routine is where the real work begins: */
int main(int argc, char *argv[])
{
- /* Memory, top-level pagetable, code startpoint, PAGE_OFFSET and size
- * of the (optional) initrd. */
- unsigned long mem = 0, pgdir, start, page_offset, initrd_size = 0;
- /* A temporary and the /dev/lguest file descriptor. */
+ /* Memory, top-level pagetable, code startpoint and size of the
+ * (optional) initrd. */
+ unsigned long mem = 0, start, initrd_size = 0;
+ /* Two temporaries and the /dev/lguest file descriptor. */
int i, c, lguest_fd;
- /* The list of Guest devices, based on command line arguments. */
- struct device_list device_list;
- /* The boot information for the Guest: at guest-physical address 0. */
- void *boot = (void *)0;
+ /* The boot information for the Guest. */
+ 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. */
- device_list.max_infd = -1;
- device_list.dev = NULL;
- device_list.lastdev = &device_list.dev;
- FD_ZERO(&device_list.infds);
-
+ * 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 = 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
* of memory now. */
for (i = 1; i < argc; i++) {
if (argv[i][0] != '-') {
- mem = top = atoi(argv[i]) * 1024 * 1024;
- device_list.descs = map_zeroed_pages(top, 1);
- top += getpagesize();
+ mem = atoi(argv[i]) * 1024 * 1024;
+ /* We start by mapping anonymous pages over all of
+ * guest-physical memory range. This fills it with 0,
+ * and ensures that the Guest won't be killed when it
+ * tries to access it. */
+ guest_base = map_zeroed_pages(mem / getpagesize()
+ + DEVICE_PAGES);
+ guest_limit = mem;
+ guest_max = mem + DEVICE_PAGES*getpagesize();
+ devices.descpage = get_pages(1);
break;
}
}
case 'v':
verbose = true;
break;
- case 's':
- setup_net_file(optarg, &device_list);
- break;
case 't':
- setup_tun_net(optarg, &device_list);
+ setup_tun_net(optarg);
break;
case 'b':
- setup_block_file(optarg, &device_list);
+ setup_block_file(optarg);
+ break;
+ case 'r':
+ setup_rng();
break;
case 'i':
initrd_name = optarg;
if (optind + 2 > argc)
usage();
+ verbose("Guest base is at %p\n", guest_base);
+
/* We always have a console device */
- setup_console(&device_list);
+ setup_console();
- /* We start by mapping anonymous pages over all of guest-physical
- * memory range. This fills it with 0, and ensures that the Guest
- * won't be killed when it tries to access it. */
- map_zeroed_pages(0, mem / getpagesize());
+ /* 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),
- &page_offset);
+ start = load_kernel(open_or_die(argv[optind+1], O_RDONLY));
+
+ /* Boot information is stashed at physical address 0 */
+ boot = from_guest_phys(0);
/* Map the initrd image if requested (at top of physical memory) */
if (initrd_name) {
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, page_offset);
-
/* 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) */
- *(void **)(boot + 0x228) = boot + 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. */
+ boot->hdr.version = 0x207;
+
+ /* The hardware_subarch value of "1" tells the Guest it's an lguest. */
+ boot->hdr.hardware_subarch = 1;
- /* The guest type value of "1" tells the Guest it's under lguest. */
- *(int *)(boot + 0x23c) = 1;
+ /* 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, page_offset);
+ 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, &device_list);
+ /* 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, &device_list);
+ run_guest(lguest_fd);
}
/*:*/