path: root/usr.sbin/vmctl/vmctl.c

/*	$OpenBSD: vmctl.c,v 1.58 2018/09/16 02:43:11 millert Exp $	*/

/*
 * Copyright (c) 2014 Mike Larkin <mlarkin@openbsd.org>
 *
 * Permission to use, copy, modify, and distribute this software for any
 * purpose with or without fee is hereby granted, provided that the above
 * copyright notice and this permission notice appear in all copies.
 *
 * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
 * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
 * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
 * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
 * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
 * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
 * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
 */

#include <sys/queue.h>
#include <sys/uio.h>
#include <sys/stat.h>
#include <sys/socket.h>
#include <sys/un.h>

#include <machine/vmmvar.h>

#include <ctype.h>
#include <err.h>
#include <errno.h>
#include <fcntl.h>
#include <imsg.h>
#include <limits.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <unistd.h>
#include <util.h>
#include <pwd.h>
#include <grp.h>

#include "vmd.h"
#include "vmctl.h"
#include "atomicio.h"

extern char *__progname;
uint32_t info_id;
char info_name[VMM_MAX_NAME_LEN];
int info_console;

/*
 * vm_start
 *
 * Request vmd to start the VM defined by the supplied parameters
 *
 * Parameters:
 *  start_id: optional ID of the VM
 *  name: optional name of the VM
 *  memsize: memory size (MB) of the VM to create
 *  nnics: number of vionet network interfaces to create
 *  nics: switch names of the network interfaces to create
 *  ndisks: number of disk images
 *  disks: disk image file names
 *  kernel: kernel image to load
 *  iso: iso image file
 *  instance: create instance from vm
 *
 * Return:
 *  0 if the request to start the VM was sent successfully.
 *  ENOMEM if a memory allocation failure occurred.
 */
int
vm_start(uint32_t start_id, const char *name, int memsize, int nnics,
    char **nics, int ndisks, char **disks, int *disktypes, char *kernel,
    char *iso, char *instance)
{
	struct vmop_create_params *vmc;
	struct vm_create_params *vcp;
	unsigned int flags = 0;
	int i;
	const char *s;

	if (memsize)
		flags |= VMOP_CREATE_MEMORY;
	if (nnics)
		flags |= VMOP_CREATE_NETWORK;
	if (ndisks)
		flags |= VMOP_CREATE_DISK;
	if (kernel)
		flags |= VMOP_CREATE_KERNEL;
	if (iso)
		flags |= VMOP_CREATE_CDROM;
	if (instance)
		flags |= VMOP_CREATE_INSTANCE;
	else if (flags != 0) {
		if (memsize < 1)
			memsize = VM_DEFAULT_MEMORY;
		if (ndisks > VMM_MAX_DISKS_PER_VM)
			errx(1, "too many disks");
		else if (ndisks == 0)
			warnx("starting without disks");
		if (kernel == NULL && ndisks == 0 && !iso)
			errx(1, "no kernel or disk/cdrom specified");
		if (nnics == -1)
			nnics = 0;
		if (nnics > VMM_MAX_NICS_PER_VM)
			errx(1, "too many network interfaces");
		if (nnics == 0)
			warnx("starting without network interfaces");
	}

	if ((vmc = calloc(1, sizeof(struct vmop_create_params))) == NULL)
		return (ENOMEM);

	vmc->vmc_flags = flags;

	/* vcp includes configuration that is shared with the kernel */
	vcp = &vmc->vmc_params;

	/*
	 * XXX: vmd(8) fills in the actual memory ranges. vmctl(8)
	 * just passes in the actual memory size in MB here.
	 */
	vcp->vcp_nmemranges = 1;
	vcp->vcp_memranges[0].vmr_size = memsize;

	vcp->vcp_ncpus = 1;
	vcp->vcp_ndisks = ndisks;
	vcp->vcp_nnics = nnics;
	vcp->vcp_id = start_id;

	for (i = 0 ; i < ndisks; i++) {
		if (strlcpy(vcp->vcp_disks[i], disks[i],
		    sizeof(vcp->vcp_disks[i])) >=
		    sizeof(vcp->vcp_disks[i]))
			errx(1, "disk path too long");
		vmc->vmc_disktypes[i] = disktypes[i];
	}
	for (i = 0 ; i < nnics; i++) {
		vmc->vmc_ifflags[i] = VMIFF_UP;

		if (strcmp(".", nics[i]) == 0) {
			/* Add a "local" interface */
			(void)strlcpy(vmc->vmc_ifswitch[i], "",
			    sizeof(vmc->vmc_ifswitch[i]));
			vmc->vmc_ifflags[i] |= VMIFF_LOCAL;
		} else {
			/* Add an interface to a switch */
			if (strlcpy(vmc->vmc_ifswitch[i], nics[i],
			    sizeof(vmc->vmc_ifswitch[i])) >=
			    sizeof(vmc->vmc_ifswitch[i]))
				errx(1, "interface name too long");
		}
	}
	if (name != NULL) {
		/*
		 * Allow VMs names with alphanumeric characters, dot, hyphen
		 * and underscore. But disallow dot, hyphen and underscore at
		 * the start.
		 */
		if (*name == '-' || *name == '.' || *name == '_')
			errx(1, "invalid VM name");

		for (s = name; *s != '\0'; ++s) {
			if (!(isalnum(*s) || *s == '.' || *s == '-' ||
			    *s == '_'))
				errx(1, "invalid VM name");
		}

		if (strlcpy(vcp->vcp_name, name,
		    sizeof(vcp->vcp_name)) >= sizeof(vcp->vcp_name))
			errx(1, "vm name too long");
	}
	if (kernel != NULL)
		if (strlcpy(vcp->vcp_kernel, kernel,
		    sizeof(vcp->vcp_kernel)) >= sizeof(vcp->vcp_kernel))
			errx(1, "kernel name too long");
	if (iso != NULL)
		if (strlcpy(vcp->vcp_cdrom, iso,
		    sizeof(vcp->vcp_cdrom)) >= sizeof(vcp->vcp_cdrom))
			errx(1, "cdrom name too long");
	if (instance != NULL)
		if (strlcpy(vmc->vmc_instance, instance,
		    sizeof(vmc->vmc_instance)) >= sizeof(vmc->vmc_instance))
			errx(1, "instance vm name too long");

	imsg_compose(ibuf, IMSG_VMDOP_START_VM_REQUEST, 0, 0, -1,
	    vmc, sizeof(struct vmop_create_params));

	free(vcp);
	return (0);
}

/*
 * vm_start_complete
 *
 * Callback function invoked when we are expecting an
 * IMSG_VMDOP_START_VMM_RESPONSE message indicating the completion of
 * a start vm operation.
 *
 * Parameters:
 *  imsg : response imsg received from vmd
 *  ret  : return value
 *  autoconnect : open the console after startup
 *
 * Return:
 *  Always 1 to indicate we have processed the return message (even if it
 *  was an incorrect/failure message)
 *
 *  The function also sets 'ret' to the error code as follows:
 *   0     : Message successfully processed
 *   EINVAL: Invalid or unexpected response from vmd
 *   EIO   : vm_start command failed
 *   ENOENT: a specified component of the VM could not be found (disk image,
 *    BIOS firmware image, etc)
 */
int
vm_start_complete(struct imsg *imsg, int *ret, int autoconnect)
{
	struct vmop_result *vmr;
	int res;

	if (imsg->hdr.type == IMSG_VMDOP_START_VM_RESPONSE) {
		vmr = (struct vmop_result *)imsg->data;
		res = vmr->vmr_result;
		if (res) {
			switch (res) {
			case VMD_BIOS_MISSING:
				warnx("vmm bios firmware file not found.");
				*ret = ENOENT;
				break;
			case VMD_DISK_MISSING:
				warnx("could not open disk image(s)");
				*ret = ENOENT;
				break;
			case VMD_DISK_INVALID:
				warnx("specified disk image(s) are "
				    "not regular files");
				*ret = ENOENT;
				break;
			case VMD_CDROM_MISSING:
				warnx("could not find specified iso image");
				*ret = ENOENT;
				break;
			case VMD_CDROM_INVALID:
				warnx("specified iso image is not a regular "
				    "file");
				*ret = ENOENT;
				break;
			default:
				errno = res;
				warn("start vm command failed");
				*ret = EIO;
			}
		} else if (autoconnect) {
			/* does not return */
			ctl_openconsole(vmr->vmr_ttyname);
		} else {
			warnx("started vm %d successfully, tty %s",
			    vmr->vmr_id, vmr->vmr_ttyname);
			*ret = 0;
		}
	} else {
		warnx("unexpected response received from vmd");
		*ret = EINVAL;
	}

	return (1);
}

void
send_vm(uint32_t id, const char *name)
{
	struct vmop_id vid;
	int fds[2], readn, writen;
	long pagesz;
	char *buf;

	pagesz = getpagesize();
	buf = malloc(pagesz);
	if (buf == NULL)
		errx(1, "%s: memory allocation failure", __func__);

	memset(&vid, 0, sizeof(vid));
	vid.vid_id = id;
	if (name != NULL)
		strlcpy(vid.vid_name, name, sizeof(vid.vid_name));
	if (socketpair(AF_UNIX, SOCK_STREAM, PF_UNSPEC, fds) == -1) {
		warnx("%s: socketpair creation failed", __func__);
	} else {
		imsg_compose(ibuf, IMSG_VMDOP_SEND_VM_REQUEST, 0, 0, fds[0],
				&vid, sizeof(vid));
		imsg_flush(ibuf);
		while (1) {
			readn = atomicio(read, fds[1], buf, pagesz);
			if (!readn)
				break;
			writen = atomicio(vwrite, STDOUT_FILENO, buf,
					readn);
			if (writen != readn)
				break;
		}
		if (vid.vid_id)
			warnx("sent vm %d successfully", vid.vid_id);
		else
			warnx("sent vm %s successfully", vid.vid_name);
	}

	free(buf);
}

void
vm_receive(uint32_t id, const char *name)
{
	struct vmop_id vid;
	int fds[2], readn, writen;
	long pagesz;
	char *buf;

	pagesz = getpagesize();
	buf = malloc(pagesz);
	if (buf == NULL)
		errx(1, "%s: memory allocation failure", __func__);

	memset(&vid, 0, sizeof(vid));
	if (name != NULL)
		strlcpy(vid.vid_name, name, sizeof(vid.vid_name));
	if (socketpair(AF_UNIX, SOCK_STREAM, PF_UNSPEC, fds) == -1) {
		warnx("%s: socketpair creation failed", __func__);
	} else {
		imsg_compose(ibuf, IMSG_VMDOP_RECEIVE_VM_REQUEST, 0, 0, fds[0],
		    &vid, sizeof(vid));
		imsg_flush(ibuf);
		while (1) {
			readn = atomicio(read, STDIN_FILENO, buf, pagesz);
			if (!readn) {
				close(fds[1]);
				break;
			}
			writen = atomicio(vwrite, fds[1], buf, readn);
			if (writen != readn)
				break;
		}
	}

	free(buf);
}

void
pause_vm(uint32_t pause_id, const char *name)
{
	struct vmop_id vid;

	memset(&vid, 0, sizeof(vid));
	vid.vid_id = pause_id;
	if (name != NULL)
		(void)strlcpy(vid.vid_name, name, sizeof(vid.vid_name));

	imsg_compose(ibuf, IMSG_VMDOP_PAUSE_VM, 0, 0, -1,
	    &vid, sizeof(vid));
}

int
pause_vm_complete(struct imsg *imsg, int *ret)
{
	struct vmop_result *vmr;
	int res;

	if (imsg->hdr.type == IMSG_VMDOP_PAUSE_VM_RESPONSE) {
		vmr = (struct vmop_result *)imsg->data;
		res = vmr->vmr_result;
		if (res) {
			errno = res;
			warn("pause vm command failed");
			*ret = EIO;
		} else {
			warnx("paused vm %d successfully", vmr->vmr_id);
			*ret = 0;
		}
	} else {
		warnx("unexpected response received from vmd");
		*ret = EINVAL;
	}

	return (1);
}

void
unpause_vm(uint32_t pause_id, const char *name)
{
	struct vmop_id vid;

	memset(&vid, 0, sizeof(vid));
	vid.vid_id = pause_id;
	if (name != NULL)
		(void)strlcpy(vid.vid_name, name, sizeof(vid.vid_name));

	imsg_compose(ibuf, IMSG_VMDOP_UNPAUSE_VM, 0, 0, -1,
	    &vid, sizeof(vid));
}

int
unpause_vm_complete(struct imsg *imsg, int *ret)
{
	struct vmop_result *vmr;
	int res;

	if (imsg->hdr.type == IMSG_VMDOP_UNPAUSE_VM_RESPONSE) {
		vmr = (struct vmop_result *)imsg->data;
		res = vmr->vmr_result;
		if (res) {
			errno = res;
			warn("unpause vm command failed");
			*ret = EIO;
		} else {
			warnx("unpaused vm %d successfully", vmr->vmr_id);
			*ret = 0;
		}
	} else {
		warnx("unexpected response received from vmd");
		*ret = EINVAL;
	}

	return (1);
}

/*
 * terminate_vm
 *
 * Request vmd to stop the VM indicated
 *
 * Parameters:
 *  terminate_id: ID of the vm to be terminated
 *  name: optional name of the VM to be terminated
 *  flags: VMOP_FORCE or VMOP_WAIT flags
 */
void
terminate_vm(uint32_t terminate_id, const char *name, unsigned int flags)
{
	struct vmop_id vid;

	memset(&vid, 0, sizeof(vid));
	vid.vid_id = terminate_id;
	if (name != NULL)
		(void)strlcpy(vid.vid_name, name, sizeof(vid.vid_name));

	vid.vid_flags = flags & (VMOP_FORCE|VMOP_WAIT);

	imsg_compose(ibuf, IMSG_VMDOP_TERMINATE_VM_REQUEST,
	    0, 0, -1, &vid, sizeof(vid));
}

/*
 * terminate_vm_complete
 *
 * Callback function invoked when we are expecting an
 * IMSG_VMDOP_TERMINATE_VMM_RESPONSE message indicating the completion of
 * a terminate vm operation.
 *
 * Parameters:
 *  imsg : response imsg received from vmd
 *  ret  : return value
 *  flags: VMOP_FORCE or VMOP_WAIT flags
 *
 * Return:
 *  Always 1 to indicate we have processed the return message (even if it
 *  was an incorrect/failure message)
 *
 *  The function also sets 'ret' to the error code as follows:
 *   0     : Message successfully processed
 *   EINVAL: Invalid or unexpected response from vmd
 *   EIO   : terminate_vm command failed
 */
int
terminate_vm_complete(struct imsg *imsg, int *ret, unsigned int flags)
{
	struct vmop_result *vmr;
	int res;

	if (imsg->hdr.type == IMSG_VMDOP_TERMINATE_VM_RESPONSE) {
		vmr = (struct vmop_result *)imsg->data;
		res = vmr->vmr_result;
		if (res) {
			switch (res) {
			case VMD_VM_STOP_INVALID:
				warnx("cannot stop vm that is not running");
				*ret = EINVAL;
				break;
			case ENOENT:
				warnx("vm not found");
				*ret = EIO;
				break;
			default:
				errno = res;
				warn("terminate vm command failed");
				*ret = EIO;
			}
		} else if (flags & VMOP_WAIT) {
			warnx("terminated vm %d", vmr->vmr_id);
		} else if (flags & VMOP_FORCE) {
			warnx("requested to terminate vm %d", vmr->vmr_id);
		} else {
			warnx("requested to shutdown vm %d", vmr->vmr_id);
			*ret = 0;
		}
	} else {
		warnx("unexpected response received from vmd");
		*ret = EINVAL;
	}
	errno = *ret;

	return (1);
}

/*
 * get_info_vm
 *
 * Return the list of all running VMs or find a specific VM by ID or name.
 *
 * Parameters:
 *  id: optional ID of a VM to list
 *  name: optional name of a VM to list
 *  console: if true, open the console of the selected VM (by name or ID)
 *
 * Request a list of running VMs from vmd
 */
void
get_info_vm(uint32_t id, const char *name, int console)
{
	info_id = id;
	if (name != NULL)
		(void)strlcpy(info_name, name, sizeof(info_name));
	info_console = console;
	imsg_compose(ibuf, IMSG_VMDOP_GET_INFO_VM_REQUEST, 0, 0, -1, NULL, 0);
}

/*
 * check_info_id
 *
 * Check if requested name or ID of a VM matches specified arguments
 *
 * Parameters:
 *  name: name of the VM
 *  id: ID of the VM
 */
int
check_info_id(const char *name, uint32_t id)
{
	if (info_id == 0 && *info_name == '\0')
		return (-1);
	if (info_id != 0 && info_id == id)
		return (1);
	if (*info_name != '\0' && name && strcmp(info_name, name) == 0)
		return (1);
	return (0);
}

/*
 * add_info
 *
 * Callback function invoked when we are expecting an
 * IMSG_VMDOP_GET_INFO_VM_DATA message indicating the receipt of additional
 * "list vm" data, or an IMSG_VMDOP_GET_INFO_VM_END_DATA message indicating
 * that no additional "list vm" data will be forthcoming.
 *
 * Parameters:
 *  imsg : response imsg received from vmd
 *  ret  : return value
 *
 * Return:
 *  0     : the returned data was successfully added to the "list vm" data.
 *          The caller can expect more data.
 *  1     : IMSG_VMDOP_GET_INFO_VM_END_DATA received (caller should not call
 *          add_info again), or an error occurred adding the returned data
 *          to the "list vm" data. The caller should check the value of
 *          'ret' to determine which case occurred.
 *
 * This function does not return if a VM is found and info_console is set.
 *
 *  The function also sets 'ret' to the error code as follows:
 *   0     : Message successfully processed
 *   EINVAL: Invalid or unexpected response from vmd
 *   ENOMEM: memory allocation failure
 */
int
add_info(struct imsg *imsg, int *ret)
{
	static size_t ct = 0;
	static struct vmop_info_result *vir = NULL;

	if (imsg->hdr.type == IMSG_VMDOP_GET_INFO_VM_DATA) {
		vir = reallocarray(vir, ct + 1,
		    sizeof(struct vmop_info_result));
		if (vir == NULL) {
			*ret = ENOMEM;
			return (1);
		}
		memcpy(&vir[ct], imsg->data, sizeof(struct vmop_info_result));
		ct++;
		*ret = 0;
		return (0);
	} else if (imsg->hdr.type == IMSG_VMDOP_GET_INFO_VM_END_DATA) {
		if (info_console)
			vm_console(vir, ct);
		else
			print_vm_info(vir, ct);
		free(vir);
		*ret = 0;
		return (1);
	} else {
		*ret = EINVAL;
		return (1);
	}
}

/*
 * print_vm_info
 *
 * Prints the vm information returned from vmd in 'list' to stdout.
 *
 * Parameters
 *  list: the vm information (consolidated) returned from vmd via imsg
 *  ct  : the size (number of elements in 'list') of the result
 */
void
print_vm_info(struct vmop_info_result *list, size_t ct)
{
	struct vm_info_result *vir;
	struct vmop_info_result *vmi;
	size_t i, j;
	char *vcpu_state, *tty;
	char curmem[FMT_SCALED_STRSIZE];
	char maxmem[FMT_SCALED_STRSIZE];
	char user[16], group[16];
	const char *name;

	printf("%5s %5s %5s %7s %7s %7s %12s %s\n", "ID", "PID", "VCPUS",
	    "MAXMEM", "CURMEM", "TTY", "OWNER", "NAME");

	for (i = 0; i < ct; i++) {
		vmi = &list[i];
		vir = &vmi->vir_info;
		if (check_info_id(vir->vir_name, vir->vir_id)) {
			/* get user name */
			name = user_from_uid(vmi->vir_uid, 1);
			if (name == NULL)
				(void)snprintf(user, sizeof(user),
				    "%d", vmi->vir_uid);
			else
				(void)strlcpy(user, name, sizeof(user));
			/* get group name */
			if (vmi->vir_gid != -1) {
				if (vmi->vir_uid == 0)
					*user = '\0';
				name = group_from_gid(vmi->vir_gid, 1);
				if (name == NULL)
					(void)snprintf(group, sizeof(group),
					    ":%lld", vmi->vir_gid);
				else
					(void)snprintf(group, sizeof(group),
					    ":%s", name);
				(void)strlcat(user, group, sizeof(user));
			}

			(void)strlcpy(curmem, "-", sizeof(curmem));
			(void)strlcpy(maxmem, "-", sizeof(maxmem));

			(void)fmt_scaled(vir->vir_memory_size * 1024 * 1024,
			    maxmem);

			if (vir->vir_creator_pid != 0 && vir->vir_id != 0) {
				if (*vmi->vir_ttyname == '\0')
					tty = "-";
				/* get tty - skip /dev/ path */
				else if ((tty = strrchr(vmi->vir_ttyname,
				    '/')) == NULL || ++tty == '\0')
					tty = list[i].vir_ttyname;

				(void)fmt_scaled(vir->vir_used_size, curmem);

				/* running vm */
				printf("%5u %5u %5zd %7s %7s %7s %12s %s\n",
				    vir->vir_id, vir->vir_creator_pid,
				    vir->vir_ncpus, maxmem, curmem,
				    tty, user, vir->vir_name);
			} else {
				/* disabled vm */
				printf("%5u %5s %5zd %7s %7s %7s %12s %s\n",
				    vir->vir_id, "-",
				    vir->vir_ncpus, maxmem, curmem,
				    "-", user, vir->vir_name);
			}
		}
		if (check_info_id(vir->vir_name, vir->vir_id) > 0) {
			for (j = 0; j < vir->vir_ncpus; j++) {
				if (vir->vir_vcpu_state[j] ==
				    VCPU_STATE_STOPPED)
					vcpu_state = "STOPPED";
				else if (vir->vir_vcpu_state[j] ==
				    VCPU_STATE_RUNNING)
					vcpu_state = "RUNNING";
				else
					vcpu_state = "UNKNOWN";

				printf(" VCPU: %2zd STATE: %s\n",
				    j, vcpu_state);
			}
		}
	}
}

/*
 * vm_console
 *
 * Connects to the vm console returned from vmd in 'list'.
 *
 * Parameters
 *  list: the vm information (consolidated) returned from vmd via imsg
 *  ct  : the size (number of elements in 'list') of the result
 */
__dead void
vm_console(struct vmop_info_result *list, size_t ct)
{
	struct vmop_info_result *vir;
	size_t i;

	for (i = 0; i < ct; i++) {
		vir = &list[i];
		if ((check_info_id(vir->vir_info.vir_name,
		    vir->vir_info.vir_id) > 0) &&
			(vir->vir_ttyname[0] != '\0')) {
			/* does not return */
			ctl_openconsole(vir->vir_ttyname);
		}
	}

	errx(1, "console not found");
}

/*
 * create_raw_imagefile
 *
 * Create an empty imagefile with the specified path and size.
 *
 * Parameters:
 *  imgfile_path: path to the image file to create
 *  imgsize     : size of the image file to create (in MB)
 *
 * Return:
 *  EEXIST: The requested image file already exists
 *  0     : Image file successfully created
 *  Exxxx : Various other Exxxx errno codes due to other I/O errors
 */
int
create_raw_imagefile(const char *imgfile_path, long imgsize)
{
	int fd, ret;

	/* Refuse to overwrite an existing image */
	fd = open(imgfile_path, O_RDWR | O_CREAT | O_TRUNC | O_EXCL,
	    S_IRUSR | S_IWUSR);
	if (fd == -1)
		return (errno);

	/* Extend to desired size */
	if (ftruncate(fd, (off_t)imgsize * 1024 * 1024) == -1) {
		ret = errno;
		close(fd);
		unlink(imgfile_path);
		return (ret);
	}

	ret = close(fd);
	return (ret);
}

/*
 * create_imagefile
 *
 * Create an empty qcow2 imagefile with the specified path and size.
 *
 * Parameters:
 *  imgfile_path: path to the image file to create
 *  imgsize     : size of the image file to create (in MB)
 *
 * Return:
 *  EEXIST: The requested image file already exists
 *  0     : Image file successfully created
 *  Exxxx : Various other Exxxx errno codes due to other I/O errors
 */
#define ALIGN(sz, align) \
	((sz + align - 1) & ~(align - 1))
int
create_qc2_imagefile(const char *imgfile_path, long imgsize)
{
	struct qcheader {
		char magic[4];
		uint32_t version;
		uint64_t backingoff;
		uint32_t backingsz;
		uint32_t clustershift;
		uint64_t disksz;
		uint32_t cryptmethod;
		uint32_t l1sz;
		uint64_t l1off;
		uint64_t refoff;
		uint32_t refsz;
		uint32_t snapcount;
		uint64_t snapsz;
		/* v3 additions */
		uint64_t incompatfeatures;
		uint64_t compatfeatures;
		uint64_t autoclearfeatures;
		uint32_t reforder;
		uint32_t headersz;
	} __packed hdr;
	int fd, ret;
	uint64_t l1sz, refsz, disksz, initsz, clustersz;
	uint64_t l1off, refoff, v, i;
	uint16_t refs;

	disksz = 1024*1024*imgsize;
	clustersz = (1<<16);
	l1off = ALIGN(sizeof hdr, clustersz);
	l1sz = disksz / (clustersz*clustersz/8);
	if (l1sz == 0)
		l1sz = 1;

	refoff = ALIGN(l1off + 8*l1sz, clustersz);
	refsz = disksz / (clustersz*clustersz*clustersz/2);
	if (refsz == 0)
		refsz = 1;

	initsz = ALIGN(refoff + refsz*clustersz, clustersz);

	memcpy(hdr.magic, "QFI\xfb", 4);
	hdr.version		= htobe32(3);
	hdr.backingoff		= htobe64(0);
	hdr.backingsz		= htobe32(0);
	hdr.clustershift	= htobe32(16);
	hdr.disksz		= htobe64(disksz);
	hdr.cryptmethod		= htobe32(0);
	hdr.l1sz		= htobe32(l1sz);
	hdr.l1off		= htobe64(l1off);
	hdr.refoff		= htobe64(refoff);
	hdr.refsz		= htobe32(refsz);
	hdr.snapcount		= htobe32(0);
	hdr.snapsz		= htobe64(0);
	hdr.incompatfeatures	= htobe64(0);
	hdr.compatfeatures	= htobe64(0);
	hdr.autoclearfeatures	= htobe64(0);
	hdr.reforder		= htobe32(4);
	hdr.headersz		= htobe32(sizeof hdr);

	/* Refuse to overwrite an existing image */
	fd = open(imgfile_path, O_RDWR | O_CREAT | O_TRUNC | O_EXCL,
	    S_IRUSR | S_IWUSR);
	if (fd == -1)
		return (errno);

	/* Write out the header */
	if (write(fd, &hdr, sizeof hdr) != sizeof hdr)
		goto error;

	/* Extend to desired size, and add one refcount cluster */
	if (ftruncate(fd, (off_t)initsz + clustersz) == -1)
		goto error;

	/* 
	 * Paranoia: if our disk image takes more than one cluster
	 * to refcount the initial image, fail.
	 */
	if (initsz/clustersz > clustersz/2) {
		errno = ERANGE;
		goto error;
	}

	/* Add a refcount block, and refcount ourselves. */
	v = htobe64(initsz);
	if (pwrite(fd, &v, 8, refoff) != 8)
		goto error;
	for (i = 0; i < initsz/clustersz + 1; i++) {
		refs = htobe16(1);
		if (pwrite(fd, &refs, 2, initsz + 2*i) != 2)
			goto error;
	}

	ret = close(fd);
	return (ret);
error:
	ret = errno;
	close(fd);
	unlink(imgfile_path);
	return (errno);
}