Index: head/usr.sbin/bhyveload/bhyveload.8 =================================================================== --- head/usr.sbin/bhyveload/bhyveload.8 (revision 296101) +++ head/usr.sbin/bhyveload/bhyveload.8 (revision 296102) @@ -1,173 +1,180 @@ .\" .\" Copyright (c) 2012 NetApp Inc .\" All rights reserved. .\" .\" Redistribution and use in source and binary forms, with or without .\" modification, are permitted provided that the following conditions .\" are met: .\" 1. Redistributions of source code must retain the above copyright .\" notice, this list of conditions and the following disclaimer. .\" 2. Redistributions in binary form must reproduce the above copyright .\" notice, this list of conditions and the following disclaimer in the .\" documentation and/or other materials provided with the distribution. .\" .\" THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND .\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE .\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE .\" ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE .\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL .\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS .\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) .\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT .\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY .\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF .\" SUCH DAMAGE. .\" .\" $FreeBSD$ .\" -.Dd October 7, 2015 +.Dd February 26, 2016 .Dt BHYVELOAD 8 .Os .Sh NAME .Nm bhyveload .Nd load a .Fx guest inside a bhyve virtual machine .Sh SYNOPSIS .Nm +.Op Fl C .Op Fl S .Op Fl c Ar cons-dev .Op Fl d Ar disk-path .Op Fl e Ar name=value .Op Fl h Ar host-path .Op Fl l Ar os-loader .Op Fl m Ar mem-size .Ar vmname .Sh DESCRIPTION .Nm is used to load a .Fx guest inside a .Xr bhyve 4 virtual machine. .Pp .Nm is based on .Xr loader 8 and will present an interface identical to the .Fx loader on the user's terminal. This behavior can be changed by specifying a different OS loader. .Pp The virtual machine is identified as .Ar vmname and will be created if it does not already exist. .Sh OPTIONS The following options are available: .Bl -tag -width indent .It Fl c Ar cons-dev .Ar cons-dev is a .Xr tty 4 device to use for .Nm terminal I/O. .Pp The text string "stdio" is also accepted and selects the use of unbuffered standard I/O. This is the default value. .It Fl d Ar disk-path The .Ar disk-path is the pathname of the guest's boot disk image. .It Fl e Ar name=value Set the .Fx loader environment variable .Ar name to .Ar value . .Pp The option may be used more than once to set more than one environment variable. .It Fl h Ar host-path The .Ar host-path is the directory at the top of the guest's boot filesystem. .It Fl l Ar os-loader Specify a different OS loader. By default .Nm will use .Pa /boot/userboot.so , which presents a standard .Fx loader. .It Fl m Ar mem-size Xo .Sm off .Op Cm K | k | M | m | G | g | T | t .Xc .Sm on .Ar mem-size is the amount of memory allocated to the guest. .Pp The .Ar mem-size argument may be suffixed with one of .Cm K , .Cm M , .Cm G or .Cm T (either upper or lower case) to indicate a multiple of Kilobytes, Megabytes, Gigabytes or Terabytes respectively. .Pp The default value of .Ar mem-size is 256M. +.It Fl C +Include guest memory in the core file when +.Nm +dumps core. +This is intended for debugging an OS loader as it allows inspection of +the guest memory. .It Fl S Wire guest memory. .El .Sh EXAMPLES To create a virtual machine named .Ar freebsd-vm that boots off the ISO image .Pa /freebsd/release.iso and has 1GB memory allocated to it: .Pp .Dl "bhyveload -m 1G -d /freebsd/release.iso freebsd-vm" .Pp To create a virtual machine named .Ar test-vm with 256MB of memory allocated, the guest root filesystem under the host directory .Pa /user/images/test and terminal I/O sent to the .Xr nmdm 4 device .Pa /dev/nmdm1B .Pp .Dl "bhyveload -m 256MB -h /usr/images/test -c /dev/nmdm1B test-vm" .Sh SEE ALSO .Xr bhyve 4 , .Xr nmdm 4 , .Xr vmm 4 , .Xr bhyve 8 , .Xr loader 8 .Sh HISTORY .Nm first appeared in .Fx 10.0 , and was developed at NetApp Inc. .Sh AUTHORS .Nm was developed by .An -nosplit .An Neel Natu Aq Mt neel@FreeBSD.org at NetApp Inc with a lot of help from .An Doug Rabson Aq Mt dfr@FreeBSD.org . .Sh BUGS .Nm can only load .Fx as a guest. Index: head/usr.sbin/bhyveload/bhyveload.c =================================================================== --- head/usr.sbin/bhyveload/bhyveload.c (revision 296101) +++ head/usr.sbin/bhyveload/bhyveload.c (revision 296102) @@ -1,791 +1,794 @@ /*- * Copyright (c) 2011 NetApp, Inc. * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * * THIS SOFTWARE IS PROVIDED BY NETAPP, INC ``AS IS'' AND * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE * ARE DISCLAIMED. IN NO EVENT SHALL NETAPP, INC OR CONTRIBUTORS BE LIABLE * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF * SUCH DAMAGE. * * $FreeBSD$ */ /*- * Copyright (c) 2011 Google, Inc. * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF * SUCH DAMAGE. * * $FreeBSD$ */ #include __FBSDID("$FreeBSD$"); #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include "userboot.h" #define MB (1024 * 1024UL) #define GB (1024 * 1024 * 1024UL) #define BSP 0 #define NDISKS 32 static char *host_base; static struct termios term, oldterm; static int disk_fd[NDISKS]; static int ndisks; static int consin_fd, consout_fd; static char *vmname, *progname; static struct vmctx *ctx; static uint64_t gdtbase, cr3, rsp; static void cb_exit(void *arg, int v); /* * Console i/o callbacks */ static void cb_putc(void *arg, int ch) { char c = ch; (void) write(consout_fd, &c, 1); } static int cb_getc(void *arg) { char c; if (read(consin_fd, &c, 1) == 1) return (c); return (-1); } static int cb_poll(void *arg) { int n; if (ioctl(consin_fd, FIONREAD, &n) >= 0) return (n > 0); return (0); } /* * Host filesystem i/o callbacks */ struct cb_file { int cf_isdir; size_t cf_size; struct stat cf_stat; union { int fd; DIR *dir; } cf_u; }; static int cb_open(void *arg, const char *filename, void **hp) { struct stat st; struct cb_file *cf; char path[PATH_MAX]; if (!host_base) return (ENOENT); strlcpy(path, host_base, PATH_MAX); if (path[strlen(path) - 1] == '/') path[strlen(path) - 1] = 0; strlcat(path, filename, PATH_MAX); cf = malloc(sizeof(struct cb_file)); if (stat(path, &cf->cf_stat) < 0) { free(cf); return (errno); } cf->cf_size = st.st_size; if (S_ISDIR(cf->cf_stat.st_mode)) { cf->cf_isdir = 1; cf->cf_u.dir = opendir(path); if (!cf->cf_u.dir) goto out; *hp = cf; return (0); } if (S_ISREG(cf->cf_stat.st_mode)) { cf->cf_isdir = 0; cf->cf_u.fd = open(path, O_RDONLY); if (cf->cf_u.fd < 0) goto out; *hp = cf; return (0); } out: free(cf); return (EINVAL); } static int cb_close(void *arg, void *h) { struct cb_file *cf = h; if (cf->cf_isdir) closedir(cf->cf_u.dir); else close(cf->cf_u.fd); free(cf); return (0); } static int cb_isdir(void *arg, void *h) { struct cb_file *cf = h; return (cf->cf_isdir); } static int cb_read(void *arg, void *h, void *buf, size_t size, size_t *resid) { struct cb_file *cf = h; ssize_t sz; if (cf->cf_isdir) return (EINVAL); sz = read(cf->cf_u.fd, buf, size); if (sz < 0) return (EINVAL); *resid = size - sz; return (0); } static int cb_readdir(void *arg, void *h, uint32_t *fileno_return, uint8_t *type_return, size_t *namelen_return, char *name) { struct cb_file *cf = h; struct dirent *dp; if (!cf->cf_isdir) return (EINVAL); dp = readdir(cf->cf_u.dir); if (!dp) return (ENOENT); /* * Note: d_namlen is in the range 0..255 and therefore less * than PATH_MAX so we don't need to test before copying. */ *fileno_return = dp->d_fileno; *type_return = dp->d_type; *namelen_return = dp->d_namlen; memcpy(name, dp->d_name, dp->d_namlen); name[dp->d_namlen] = 0; return (0); } static int cb_seek(void *arg, void *h, uint64_t offset, int whence) { struct cb_file *cf = h; if (cf->cf_isdir) return (EINVAL); if (lseek(cf->cf_u.fd, offset, whence) < 0) return (errno); return (0); } static int cb_stat(void *arg, void *h, int *mode, int *uid, int *gid, uint64_t *size) { struct cb_file *cf = h; *mode = cf->cf_stat.st_mode; *uid = cf->cf_stat.st_uid; *gid = cf->cf_stat.st_gid; *size = cf->cf_stat.st_size; return (0); } /* * Disk image i/o callbacks */ static int cb_diskread(void *arg, int unit, uint64_t from, void *to, size_t size, size_t *resid) { ssize_t n; if (unit < 0 || unit >= ndisks ) return (EIO); n = pread(disk_fd[unit], to, size, from); if (n < 0) return (errno); *resid = size - n; return (0); } static int cb_diskioctl(void *arg, int unit, u_long cmd, void *data) { struct stat sb; if (unit < 0 || unit >= ndisks) return (EBADF); switch (cmd) { case DIOCGSECTORSIZE: *(u_int *)data = 512; break; case DIOCGMEDIASIZE: if (fstat(disk_fd[unit], &sb) == 0) *(off_t *)data = sb.st_size; else return (ENOTTY); break; default: return (ENOTTY); } return (0); } /* * Guest virtual machine i/o callbacks */ static int cb_copyin(void *arg, const void *from, uint64_t to, size_t size) { char *ptr; to &= 0x7fffffff; ptr = vm_map_gpa(ctx, to, size); if (ptr == NULL) return (EFAULT); memcpy(ptr, from, size); return (0); } static int cb_copyout(void *arg, uint64_t from, void *to, size_t size) { char *ptr; from &= 0x7fffffff; ptr = vm_map_gpa(ctx, from, size); if (ptr == NULL) return (EFAULT); memcpy(to, ptr, size); return (0); } static void cb_setreg(void *arg, int r, uint64_t v) { int error; enum vm_reg_name vmreg; vmreg = VM_REG_LAST; switch (r) { case 4: vmreg = VM_REG_GUEST_RSP; rsp = v; break; default: break; } if (vmreg == VM_REG_LAST) { printf("test_setreg(%d): not implemented\n", r); cb_exit(NULL, USERBOOT_EXIT_QUIT); } error = vm_set_register(ctx, BSP, vmreg, v); if (error) { perror("vm_set_register"); cb_exit(NULL, USERBOOT_EXIT_QUIT); } } static void cb_setmsr(void *arg, int r, uint64_t v) { int error; enum vm_reg_name vmreg; vmreg = VM_REG_LAST; switch (r) { case MSR_EFER: vmreg = VM_REG_GUEST_EFER; break; default: break; } if (vmreg == VM_REG_LAST) { printf("test_setmsr(%d): not implemented\n", r); cb_exit(NULL, USERBOOT_EXIT_QUIT); } error = vm_set_register(ctx, BSP, vmreg, v); if (error) { perror("vm_set_msr"); cb_exit(NULL, USERBOOT_EXIT_QUIT); } } static void cb_setcr(void *arg, int r, uint64_t v) { int error; enum vm_reg_name vmreg; vmreg = VM_REG_LAST; switch (r) { case 0: vmreg = VM_REG_GUEST_CR0; break; case 3: vmreg = VM_REG_GUEST_CR3; cr3 = v; break; case 4: vmreg = VM_REG_GUEST_CR4; break; default: break; } if (vmreg == VM_REG_LAST) { printf("test_setcr(%d): not implemented\n", r); cb_exit(NULL, USERBOOT_EXIT_QUIT); } error = vm_set_register(ctx, BSP, vmreg, v); if (error) { perror("vm_set_cr"); cb_exit(NULL, USERBOOT_EXIT_QUIT); } } static void cb_setgdt(void *arg, uint64_t base, size_t size) { int error; error = vm_set_desc(ctx, BSP, VM_REG_GUEST_GDTR, base, size - 1, 0); if (error != 0) { perror("vm_set_desc(gdt)"); cb_exit(NULL, USERBOOT_EXIT_QUIT); } gdtbase = base; } static void cb_exec(void *arg, uint64_t rip) { int error; if (cr3 == 0) error = vm_setup_freebsd_registers_i386(ctx, BSP, rip, gdtbase, rsp); else error = vm_setup_freebsd_registers(ctx, BSP, rip, cr3, gdtbase, rsp); if (error) { perror("vm_setup_freebsd_registers"); cb_exit(NULL, USERBOOT_EXIT_QUIT); } cb_exit(NULL, 0); } /* * Misc */ static void cb_delay(void *arg, int usec) { usleep(usec); } static void cb_exit(void *arg, int v) { tcsetattr(consout_fd, TCSAFLUSH, &oldterm); exit(v); } static void cb_getmem(void *arg, uint64_t *ret_lowmem, uint64_t *ret_highmem) { *ret_lowmem = vm_get_lowmem_size(ctx); *ret_highmem = vm_get_highmem_size(ctx); } struct env { const char *str; /* name=value */ SLIST_ENTRY(env) next; }; static SLIST_HEAD(envhead, env) envhead; static void addenv(const char *str) { struct env *env; env = malloc(sizeof(struct env)); env->str = str; SLIST_INSERT_HEAD(&envhead, env, next); } static const char * cb_getenv(void *arg, int num) { int i; struct env *env; i = 0; SLIST_FOREACH(env, &envhead, next) { if (i == num) return (env->str); i++; } return (NULL); } static int cb_vm_set_register(void *arg, int vcpu, int reg, uint64_t val) { return (vm_set_register(ctx, vcpu, reg, val)); } static int cb_vm_set_desc(void *arg, int vcpu, int reg, uint64_t base, u_int limit, u_int access) { return (vm_set_desc(ctx, vcpu, reg, base, limit, access)); } static struct loader_callbacks cb = { .getc = cb_getc, .putc = cb_putc, .poll = cb_poll, .open = cb_open, .close = cb_close, .isdir = cb_isdir, .read = cb_read, .readdir = cb_readdir, .seek = cb_seek, .stat = cb_stat, .diskread = cb_diskread, .diskioctl = cb_diskioctl, .copyin = cb_copyin, .copyout = cb_copyout, .setreg = cb_setreg, .setmsr = cb_setmsr, .setcr = cb_setcr, .setgdt = cb_setgdt, .exec = cb_exec, .delay = cb_delay, .exit = cb_exit, .getmem = cb_getmem, .getenv = cb_getenv, /* Version 4 additions */ .vm_set_register = cb_vm_set_register, .vm_set_desc = cb_vm_set_desc, }; static int altcons_open(char *path) { struct stat sb; int err; int fd; /* * Allow stdio to be passed in so that the same string * can be used for the bhyveload console and bhyve com-port * parameters */ if (!strcmp(path, "stdio")) return (0); err = stat(path, &sb); if (err == 0) { if (!S_ISCHR(sb.st_mode)) err = ENOTSUP; else { fd = open(path, O_RDWR | O_NONBLOCK); if (fd < 0) err = errno; else consin_fd = consout_fd = fd; } } return (err); } static int disk_open(char *path) { int err, fd; if (ndisks >= NDISKS) return (ERANGE); err = 0; fd = open(path, O_RDONLY); if (fd > 0) { disk_fd[ndisks] = fd; ndisks++; } else err = errno; return (err); } static void usage(void) { fprintf(stderr, "usage: %s [-S][-c ] [-d ] [-e ]\n" " %*s [-h ] [-m mem-size] \n", progname, (int)strlen(progname), ""); exit(1); } int main(int argc, char** argv) { char *loader; void *h; void (*func)(struct loader_callbacks *, void *, int, int); uint64_t mem_size; int opt, error, need_reinit, memflags; progname = basename(argv[0]); loader = NULL; memflags = 0; mem_size = 256 * MB; consin_fd = STDIN_FILENO; consout_fd = STDOUT_FILENO; - while ((opt = getopt(argc, argv, "Sc:d:e:h:l:m:")) != -1) { + while ((opt = getopt(argc, argv, "CSc:d:e:h:l:m:")) != -1) { switch (opt) { case 'c': error = altcons_open(optarg); if (error != 0) errx(EX_USAGE, "Could not open '%s'", optarg); break; case 'd': error = disk_open(optarg); if (error != 0) errx(EX_USAGE, "Could not open '%s'", optarg); break; case 'e': addenv(optarg); break; case 'h': host_base = optarg; break; case 'l': if (loader != NULL) errx(EX_USAGE, "-l can only be given once"); loader = strdup(optarg); if (loader == NULL) err(EX_OSERR, "malloc"); break; case 'm': error = vm_parse_memsize(optarg, &mem_size); if (error != 0) errx(EX_USAGE, "Invalid memsize '%s'", optarg); + break; + case 'C': + memflags |= VM_MEM_F_INCORE; break; case 'S': memflags |= VM_MEM_F_WIRED; break; case '?': usage(); } } argc -= optind; argv += optind; if (argc != 1) usage(); vmname = argv[0]; need_reinit = 0; error = vm_create(vmname); if (error) { if (errno != EEXIST) { perror("vm_create"); exit(1); } need_reinit = 1; } ctx = vm_open(vmname); if (ctx == NULL) { perror("vm_open"); exit(1); } if (need_reinit) { error = vm_reinit(ctx); if (error) { perror("vm_reinit"); exit(1); } } vm_set_memflags(ctx, memflags); error = vm_setup_memory(ctx, mem_size, VM_MMAP_ALL); if (error) { perror("vm_setup_memory"); exit(1); } if (loader == NULL) { loader = strdup("/boot/userboot.so"); if (loader == NULL) err(EX_OSERR, "malloc"); } h = dlopen(loader, RTLD_LOCAL); if (!h) { printf("%s\n", dlerror()); free(loader); return (1); } func = dlsym(h, "loader_main"); if (!func) { printf("%s\n", dlerror()); free(loader); return (1); } tcgetattr(consout_fd, &term); oldterm = term; cfmakeraw(&term); term.c_cflag |= CLOCAL; tcsetattr(consout_fd, TCSAFLUSH, &term); addenv("smbios.bios.vendor=BHYVE"); addenv("boot_serial=1"); func(&cb, NULL, USERBOOT_VERSION_4, ndisks); free(loader); return (0); }