xen_internal.c 65.0 KB
Newer Older
1 2 3
/*
 * xen_internal.c: direct access to Xen hypervisor level
 *
D
Daniel Veillard 已提交
4
 * Copyright (C) 2005, 2006 Red Hat, Inc.
5 6 7 8 9 10
 *
 * See COPYING.LIB for the License of this software
 *
 * Daniel Veillard <veillard@redhat.com>
 */

11 12
#ifdef WITH_XEN

13 14
#include <stdio.h>
#include <string.h>
15
/* required for uint8_t, uint32_t, etc ... */
16 17 18 19 20 21 22
#include <stdint.h>
#include <sys/types.h>
#include <sys/stat.h>
#include <unistd.h>
#include <fcntl.h>
#include <sys/mman.h>
#include <sys/ioctl.h>
23
#include <limits.h>
24
#include <stdint.h>
25 26 27
#include <regex.h>
#include <errno.h>
#include <sys/utsname.h>
28 29

/* required for dom0_getdomaininfo_t */
30
#include <xen/dom0_ops.h>
31
#include <xen/version.h>
32
#include <xen/xen.h>
33
#include <xen/linux/privcmd.h>
34

35 36 37
/* required for shutdown flags */
#include <xen/sched.h>

38 39
#include "xml.h"

40 41
/* #define DEBUG */
/*
42
 * so far there is 2 versions of the structures usable for doing
43 44 45 46
 * hypervisor calls.
 */
/* the old one */
typedef struct v0_hypercall_struct {
47 48
    unsigned long op;
    unsigned long arg[5];
49 50 51 52 53 54 55 56 57 58
} v0_hypercall_t;
#define XEN_V0_IOCTL_HYPERCALL_CMD \
        _IOC(_IOC_NONE, 'P', 0, sizeof(v0_hypercall_t))

/* the new one */
typedef struct v1_hypercall_struct
{
    uint64_t op;
    uint64_t arg[5];
} v1_hypercall_t;
59 60
#define XEN_V1_IOCTL_HYPERCALL_CMD                  \
    _IOC(_IOC_NONE, 'P', 0, sizeof(v1_hypercall_t))
61

62 63 64 65 66 67 68 69
typedef v1_hypercall_t hypercall_t;

#ifndef __HYPERVISOR_sysctl
#define __HYPERVISOR_sysctl 35
#endif
#ifndef __HYPERVISOR_domctl
#define __HYPERVISOR_domctl 36
#endif
70

71 72
static int xen_ioctl_hypercall_cmd = 0;
static int initialized = 0;
73
static int in_init = 0;
74
static int hv_version = 0;
75 76 77
static int hypervisor_version = 2;
static int sys_interface_version = -1;
static int dom_interface_version = -1;
78
static int kb_per_pages = 0;
79

80 81 82 83 84 85 86 87
/* Regular expressions used by xenHypervisorGetCapabilities, and
 * compiled once by xenHypervisorInit.  Note that these are POSIX.2
 * extended regular expressions (regex(7)).
 */
static const char *flags_hvm_re = "^flags[[:blank:]]+:.* (vmx|svm)[[:space:]]";
static regex_t flags_hvm_rec;
static const char *flags_pae_re = "^flags[[:blank:]]+:.* pae[[:space:]]";
static regex_t flags_pae_rec;
88
static const char *xen_cap_re = "(xen|hvm)-[[:digit:]]+\\.[[:digit:]]+-(x86_32|x86_64|ia64|powerpc64)(p|be)?";
89 90
static regex_t xen_cap_rec;

91 92 93 94 95 96 97 98 99 100 101 102 103 104 105
/*
 * The content of the structures for a getdomaininfolist system hypercall
 */
#ifndef DOMFLAGS_DYING
#define DOMFLAGS_DYING     (1<<0) /* Domain is scheduled to die.             */
#define DOMFLAGS_SHUTDOWN  (1<<2) /* The guest OS has shut down.             */
#define DOMFLAGS_PAUSED    (1<<3) /* Currently paused by control software.   */
#define DOMFLAGS_BLOCKED   (1<<4) /* Currently blocked pending an event.     */
#define DOMFLAGS_RUNNING   (1<<5) /* Domain is currently running.            */
#define DOMFLAGS_CPUMASK      255 /* CPU to which this domain is bound.      */
#define DOMFLAGS_CPUSHIFT       8
#define DOMFLAGS_SHUTDOWNMASK 255 /* DOMFLAGS_SHUTDOWN guest-supplied code.  */
#define DOMFLAGS_SHUTDOWNSHIFT 16
#endif

106 107 108 109 110 111 112 113 114 115 116
/*
 * These flags explain why a system is in the state of "shutdown".  Normally,
 * They are defined in xen/sched.h
 */
#ifndef SHUTDOWN_poweroff
#define SHUTDOWN_poweroff   0  /* Domain exited normally. Clean up and kill. */
#define SHUTDOWN_reboot     1  /* Clean up, kill, and then restart.          */
#define SHUTDOWN_suspend    2  /* Clean up, save suspend info, kill.         */
#define SHUTDOWN_crash      3  /* Tell controller we've crashed.             */
#endif

117 118 119 120 121 122 123 124 125
#define XEN_V0_OP_GETDOMAININFOLIST	38
#define XEN_V1_OP_GETDOMAININFOLIST	38
#define XEN_V2_OP_GETDOMAININFOLIST	6

struct xen_v0_getdomaininfo {
    domid_t  domain;	/* the domain number */
    uint32_t flags;	/* falgs, see before */
    uint64_t tot_pages;	/* total number of pages used */
    uint64_t max_pages;	/* maximum number of pages allowed */
126
    unsigned long shared_info_frame; /* MFN of shared_info struct */
127 128 129 130 131 132 133 134
    uint64_t cpu_time;  /* CPU time used */
    uint32_t nr_online_vcpus;  /* Number of VCPUs currently online. */
    uint32_t max_vcpu_id; /* Maximum VCPUID in use by this domain. */
    uint32_t ssidref;
    xen_domain_handle_t handle;
};
typedef struct xen_v0_getdomaininfo xen_v0_getdomaininfo;

135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180
struct xen_v2_getdomaininfo {
    domid_t  domain;	/* the domain number */
    uint32_t flags;	/* falgs, see before */
    uint64_t tot_pages;	/* total number of pages used */
    uint64_t max_pages;	/* maximum number of pages allowed */
    uint64_t shared_info_frame; /* MFN of shared_info struct */
    uint64_t cpu_time;  /* CPU time used */
    uint32_t nr_online_vcpus;  /* Number of VCPUs currently online. */
    uint32_t max_vcpu_id; /* Maximum VCPUID in use by this domain. */
    uint32_t ssidref;
    xen_domain_handle_t handle;
};
typedef struct xen_v2_getdomaininfo xen_v2_getdomaininfo;

union xen_getdomaininfo {
    struct xen_v0_getdomaininfo v0;
    struct xen_v2_getdomaininfo v2;
};
typedef union xen_getdomaininfo xen_getdomaininfo;

union xen_getdomaininfolist {
    struct xen_v0_getdomaininfo *v0;
    struct xen_v2_getdomaininfo *v2;
};
typedef union xen_getdomaininfolist xen_getdomaininfolist;

#define XEN_GETDOMAININFOLIST_ALLOC(domlist, size)                      \
    (hypervisor_version < 2 ?                                           \
     ((domlist.v0 = malloc(sizeof(xen_v0_getdomaininfo)*(size))) != NULL) : \
     ((domlist.v2 = malloc(sizeof(xen_v2_getdomaininfo)*(size))) != NULL))

#define XEN_GETDOMAININFOLIST_FREE(domlist)     \
    (hypervisor_version < 2 ?                   \
     free(domlist.v0) :                         \
     free(domlist.v2))

#define XEN_GETDOMAININFOLIST_CLEAR(domlist, size)                  \
    (hypervisor_version < 2 ?                                       \
     memset(domlist.v0, 0, sizeof(xen_v0_getdomaininfo) * size) :   \
     memset(domlist.v2, 0, sizeof(xen_v2_getdomaininfo) * size))

#define XEN_GETDOMAININFOLIST_DOMAIN(domlist, n)    \
    (hypervisor_version < 2 ?                       \
     domlist.v0[n].domain :                         \
     domlist.v2[n].domain)

181 182 183 184
#define XEN_GETDOMAININFOLIST_DATA(domlist)     \
    (hypervisor_version < 2 ?                   \
     (void*)(domlist->v0) :                     \
     (void*)(domlist->v2))
185

186 187 188 189
#define XEN_GETDOMAININFO_SIZE                  \
    (hypervisor_version < 2 ?                   \
     sizeof(xen_v0_getdomaininfo) :             \
     sizeof(xen_v2_getdomaininfo))
190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210

#define XEN_GETDOMAININFO_CLEAR(dominfo)                        \
    (hypervisor_version < 2 ?                                   \
     memset(&(dominfo.v0), 0, sizeof(xen_v0_getdomaininfo)) :   \
     memset(&(dominfo.v2), 0, sizeof(xen_v2_getdomaininfo)))

#define XEN_GETDOMAININFO_DOMAIN(dominfo)       \
    (hypervisor_version < 2 ?                   \
     dominfo.v0.domain :                        \
     dominfo.v2.domain)

#define XEN_GETDOMAININFO_CPUTIME(dominfo)      \
    (hypervisor_version < 2 ?                   \
     dominfo.v0.cpu_time :                      \
     dominfo.v2.cpu_time)

#define XEN_GETDOMAININFO_CPUCOUNT(dominfo)     \
    (hypervisor_version < 2 ?                   \
     dominfo.v0.nr_online_vcpus :               \
     dominfo.v2.nr_online_vcpus)

211 212 213 214 215
#define XEN_GETDOMAININFO_MAXCPUID(dominfo)  \
    (hypervisor_version < 2 ?                   \
     dominfo.v0.max_vcpu_id :                   \
     dominfo.v2.max_vcpu_id)

216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233
#define XEN_GETDOMAININFO_FLAGS(dominfo)        \
    (hypervisor_version < 2 ?                   \
     dominfo.v0.flags :                         \
     dominfo.v2.flags)

#define XEN_GETDOMAININFO_TOT_PAGES(dominfo)    \
    (hypervisor_version < 2 ?                   \
     dominfo.v0.tot_pages :                     \
     dominfo.v2.tot_pages)

#define XEN_GETDOMAININFO_MAX_PAGES(dominfo)    \
    (hypervisor_version < 2 ?                   \
     dominfo.v0.max_pages :                     \
     dominfo.v2.max_pages)



struct xen_v0_getdomaininfolistop {
234 235 236 237 238
    domid_t   first_domain;
    uint32_t  max_domains;
    struct xen_v0_getdomaininfo *buffer;
    uint32_t  num_domains;
};
239 240 241 242 243 244 245 246 247 248 249 250
typedef struct xen_v0_getdomaininfolistop xen_v0_getdomaininfolistop;


struct xen_v2_getdomaininfolistop {
    domid_t   first_domain;
    uint32_t  max_domains;
    struct xen_v2_getdomaininfo *buffer;
    uint32_t  num_domains;
};
typedef struct xen_v2_getdomaininfolistop xen_v2_getdomaininfolistop;


251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342

struct xen_v0_domainop {
    domid_t   domain;
};
typedef struct xen_v0_domainop xen_v0_domainop;

/*
 * The informations for a destroydomain system hypercall
 */
#define XEN_V0_OP_DESTROYDOMAIN	9
#define XEN_V1_OP_DESTROYDOMAIN	9
#define XEN_V2_OP_DESTROYDOMAIN	2

/*
 * The informations for a pausedomain system hypercall
 */
#define XEN_V0_OP_PAUSEDOMAIN	10
#define XEN_V1_OP_PAUSEDOMAIN	10
#define XEN_V2_OP_PAUSEDOMAIN	3

/*
 * The informations for an unpausedomain system hypercall
 */
#define XEN_V0_OP_UNPAUSEDOMAIN	11
#define XEN_V1_OP_UNPAUSEDOMAIN	11
#define XEN_V2_OP_UNPAUSEDOMAIN	4

/*
 * The informations for an setmaxmem system hypercall
 */
#define XEN_V0_OP_SETMAXMEM	28
#define XEN_V1_OP_SETMAXMEM	28
#define XEN_V2_OP_SETMAXMEM	14

struct xen_v0_setmaxmem {
    domid_t	domain;
    uint64_t	maxmem;
};
typedef struct xen_v0_setmaxmem xen_v0_setmaxmem;
typedef struct xen_v0_setmaxmem xen_v1_setmaxmem;

struct xen_v2_setmaxmem {
    uint64_t	maxmem;
};
typedef struct xen_v2_setmaxmem xen_v2_setmaxmem;

/*
 * The informations for an setmaxvcpu system hypercall
 */
#define XEN_V0_OP_SETMAXVCPU	41
#define XEN_V1_OP_SETMAXVCPU	41
#define XEN_V2_OP_SETMAXVCPU	15

struct xen_v0_setmaxvcpu {
    domid_t	domain;
    uint32_t	maxvcpu;
};
typedef struct xen_v0_setmaxvcpu xen_v0_setmaxvcpu;
typedef struct xen_v0_setmaxvcpu xen_v1_setmaxvcpu;

struct xen_v2_setmaxvcpu {
    uint32_t	maxvcpu;
};
typedef struct xen_v2_setmaxvcpu xen_v2_setmaxvcpu;

/*
 * The informations for an setvcpumap system hypercall
 * Note that between 1 and 2 the limitation to 64 physical CPU was lifted
 * hence the difference in structures
 */
#define XEN_V0_OP_SETVCPUMAP	20
#define XEN_V1_OP_SETVCPUMAP	20
#define XEN_V2_OP_SETVCPUMAP	9

struct xen_v0_setvcpumap {
    domid_t	domain;
    uint32_t	vcpu;
    cpumap_t    cpumap;
};
typedef struct xen_v0_setvcpumap xen_v0_setvcpumap;
typedef struct xen_v0_setvcpumap xen_v1_setvcpumap;

struct xen_v2_cpumap {
    uint8_t    *bitmap;
    uint32_t    nr_cpus;
};
struct xen_v2_setvcpumap {
    uint32_t	vcpu;
    struct xen_v2_cpumap cpumap;
};
typedef struct xen_v2_setvcpumap xen_v2_setvcpumap;

343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378
/*
 * The informations for an vcpuinfo system hypercall
 */
#define XEN_V0_OP_GETVCPUINFO   43
#define XEN_V1_OP_GETVCPUINFO	43
#define XEN_V2_OP_GETVCPUINFO   14

struct xen_v0_vcpuinfo {
    domid_t	domain;		/* owner's domain */
    uint32_t	vcpu;		/* the vcpu number */
    uint8_t	online;		/* seen as on line */
    uint8_t	blocked;	/* blocked on event */
    uint8_t	running;	/* scheduled on CPU */
    uint64_t    cpu_time;	/* nanosecond of CPU used */
    uint32_t	cpu;		/* current mapping */
    cpumap_t	cpumap;		/* deprecated in V2 */
};
typedef struct xen_v0_vcpuinfo xen_v0_vcpuinfo;
typedef struct xen_v0_vcpuinfo xen_v1_vcpuinfo;

struct xen_v2_vcpuinfo {
    uint32_t	vcpu;		/* the vcpu number */
    uint8_t	online;		/* seen as on line */
    uint8_t	blocked;	/* blocked on event */
    uint8_t	running;	/* scheduled on CPU */
    uint64_t    cpu_time;	/* nanosecond of CPU used */
    uint32_t	cpu;		/* current mapping */
};
typedef struct xen_v2_vcpuinfo xen_v2_vcpuinfo;

/*
 * from V2 the pinning of a vcpu is read with a separate call
 */
#define XEN_V2_OP_GETVCPUMAP	25
typedef struct xen_v2_setvcpumap xen_v2_getvcpumap;

379 380 381 382 383 384 385 386 387
/*
 * The hypercall operation structures also have changed on
 * changeset 86d26e6ec89b
 */
/* the old structure */
struct xen_op_v0 {
    uint32_t cmd;
    uint32_t interface_version;
    union {
388 389 390 391 392 393 394
        xen_v0_getdomaininfolistop getdomaininfolist;
        xen_v0_domainop          domain;
        xen_v0_setmaxmem         setmaxmem;
        xen_v0_setmaxvcpu        setmaxvcpu;
        xen_v0_setvcpumap        setvcpumap;
        xen_v0_vcpuinfo          getvcpuinfo;
        uint8_t padding[128];
395 396 397 398 399 400 401 402 403 404
    } u;
};
typedef struct xen_op_v0 xen_op_v0;
typedef struct xen_op_v0 xen_op_v1;

/* the new structure for systems operations */
struct xen_op_v2_sys {
    uint32_t cmd;
    uint32_t interface_version;
    union {
405 406
        xen_v2_getdomaininfolistop getdomaininfolist;
        uint8_t padding[128];
407 408 409 410 411 412 413 414 415 416
    } u;
};
typedef struct xen_op_v2_sys xen_op_v2_sys;

/* the new structure for domains operation */
struct xen_op_v2_dom {
    uint32_t cmd;
    uint32_t interface_version;
    domid_t  domain;
    union {
417 418 419 420 421 422
        xen_v2_setmaxmem         setmaxmem;
        xen_v2_setmaxvcpu        setmaxvcpu;
        xen_v2_setvcpumap        setvcpumap;
        xen_v2_vcpuinfo          getvcpuinfo;
        xen_v2_getvcpumap        getvcpumap;
        uint8_t padding[128];
423 424 425
    } u;
};
typedef struct xen_op_v2_dom xen_op_v2_dom;
426 427

#include "internal.h"
428
#include "driver.h"
429 430 431 432
#include "xen_internal.h"

#define XEN_HYPERVISOR_SOCKET "/proc/xen/privcmd"

433
#ifndef PROXY
434
static const char * xenHypervisorGetType(virConnectPtr conn);
435
static unsigned long xenHypervisorGetMaxMemory(virDomainPtr domain);
436
#endif
437

438
#ifndef PROXY
439
static virDriver xenHypervisorDriver = {
440
    VIR_DRV_XEN_HYPERVISOR,
441
    "Xen",
442 443 444
    (DOM0_INTERFACE_VERSION >> 24) * 1000000 +
    ((DOM0_INTERFACE_VERSION >> 16) & 0xFF) * 1000 +
    (DOM0_INTERFACE_VERSION & 0xFFFF),
445 446
    xenHypervisorOpen, /* open */
    xenHypervisorClose, /* close */
447
    xenHypervisorGetType, /* type */
448
    xenHypervisorGetVersion, /* version */
449
    xenHypervisorNumOfMaxVcpus, /* getMaxVcpus */
450
    NULL, /* nodeGetInfo */
451
    xenHypervisorGetCapabilities, /* getCapabilities */
452 453
    xenHypervisorListDomains, /* listDomains */
    xenHypervisorNumOfDomains, /* numOfDomains */
454 455 456 457 458 459 460
    NULL, /* domainCreateLinux */
    NULL, /* domainLookupByID */
    NULL, /* domainLookupByUUID */
    NULL, /* domainLookupByName */
    xenHypervisorPauseDomain, /* domainSuspend */
    xenHypervisorResumeDomain, /* domainResume */
    NULL, /* domainShutdown */
461
    NULL, /* domainReboot */
462 463
    xenHypervisorDestroyDomain, /* domainDestroy */
    NULL, /* domainGetOSType */
464
    xenHypervisorGetMaxMemory, /* domainGetMaxMemory */
465
    xenHypervisorSetMaxMemory, /* domainSetMaxMemory */
466
    NULL, /* domainSetMemory */
467 468
    xenHypervisorGetDomainInfo, /* domainGetInfo */
    NULL, /* domainSave */
469
    NULL, /* domainRestore */
D
Daniel Veillard 已提交
470
    NULL, /* domainCoreDump */
471 472
    xenHypervisorSetVcpus, /* domainSetVcpus */
    xenHypervisorPinVcpu, /* domainPinVcpu */
473
    xenHypervisorGetVcpus, /* domainGetVcpus */
474
    xenHypervisorGetVcpuMax, /* domainGetMaxVcpus */
475
    NULL, /* domainDumpXML */
476 477 478 479 480
    NULL, /* listDefinedDomains */
    NULL, /* numOfDefinedDomains */
    NULL, /* domainCreate */
    NULL, /* domainDefineXML */
    NULL, /* domainUndefine */
481 482
    NULL, /* domainAttachDevice */
    NULL, /* domainDetachDevice */
483 484
    NULL, /* domainGetAutostart */
    NULL, /* domainSetAutostart */
485
};
486
#endif /* !PROXY */
487

488 489 490 491
/**
 * virXenError:
 * @error: the error number
 * @info: extra information string
492
 * @value: extra information number
493 494 495 496
 *
 * Handle an error at the xend daemon interface
 */
static void
497 498
virXenError(virErrorNumber error, const char *info, int value)
{
499
    const char *errmsg;
500

501
    if ((error == VIR_ERR_OK) || (in_init != 0))
502 503 504
        return;

    errmsg = __virErrorMsg(error, info);
505
    __virRaiseError(NULL, NULL, NULL, VIR_FROM_XEN, error, VIR_ERR_ERROR,
506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530
                    errmsg, info, NULL, value, 0, errmsg, info);
}

/**
 * virXenPerror:
 * @conn: the connection (if available)
 * @msg: name of system call or file (as in perror(3))
 *
 * Raise error from a failed system call, using errno as the source.
 */
static void
virXenPerror (virConnectPtr conn, const char *msg)
{
    char *msg_s;

    msg_s = malloc (strlen (msg) + 10);
    if (msg_s) {
        strcpy (msg_s, msg);
        strcat (msg_s, ": %s");
    }

    __virRaiseError (conn, NULL, NULL,
                     VIR_FROM_XEN, VIR_ERR_SYSTEM_ERROR, VIR_ERR_ERROR,
                     msg, NULL, NULL, errno, 0,
                     msg_s ? msg_s : msg, strerror (errno));
531 532
}

533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712
/**
 * xenHypervisorDoV0Op:
 * @handle: the handle to the Xen hypervisor
 * @op: pointer to the hyperviros operation structure
 *
 * Do an hypervisor operation though the old interface,
 * this leads to an hypervisor call through ioctl.
 *
 * Returns 0 in case of success and -1 in case of error.
 */
static int
xenHypervisorDoV0Op(int handle, xen_op_v0 * op)
{
    int ret;
    v0_hypercall_t hc;

    memset(&hc, 0, sizeof(hc));
    op->interface_version = hv_version << 8;
    hc.op = __HYPERVISOR_dom0_op;
    hc.arg[0] = (unsigned long) op;

    if (mlock(op, sizeof(dom0_op_t)) < 0) {
        virXenError(VIR_ERR_XEN_CALL, " locking", sizeof(*op));
        return (-1);
    }

    ret = ioctl(handle, xen_ioctl_hypercall_cmd, (unsigned long) &hc);
    if (ret < 0) {
        virXenError(VIR_ERR_XEN_CALL, " ioctl ", xen_ioctl_hypercall_cmd);
    }

    if (munlock(op, sizeof(dom0_op_t)) < 0) {
        virXenError(VIR_ERR_XEN_CALL, " releasing", sizeof(*op));
        ret = -1;
    }

    if (ret < 0)
        return (-1);

    return (0);
}
/**
 * xenHypervisorDoV1Op:
 * @handle: the handle to the Xen hypervisor
 * @op: pointer to the hyperviros operation structure
 *
 * Do an hypervisor v1 operation, this leads to an hypervisor call through
 * ioctl.
 *
 * Returns 0 in case of success and -1 in case of error.
 */
static int
xenHypervisorDoV1Op(int handle, xen_op_v1* op)
{
    int ret;
    hypercall_t hc;

    memset(&hc, 0, sizeof(hc));
    op->interface_version = DOM0_INTERFACE_VERSION;
    hc.op = __HYPERVISOR_dom0_op;
    hc.arg[0] = (unsigned long) op;

    if (mlock(op, sizeof(dom0_op_t)) < 0) {
        virXenError(VIR_ERR_XEN_CALL, " locking", sizeof(*op));
        return (-1);
    }

    ret = ioctl(handle, xen_ioctl_hypercall_cmd, (unsigned long) &hc);
    if (ret < 0) {
        virXenError(VIR_ERR_XEN_CALL, " ioctl ", xen_ioctl_hypercall_cmd);
    }

    if (munlock(op, sizeof(dom0_op_t)) < 0) {
        virXenError(VIR_ERR_XEN_CALL, " releasing", sizeof(*op));
        ret = -1;
    }

    if (ret < 0)
        return (-1);

    return (0);
}

/**
 * xenHypervisorDoV2Sys:
 * @handle: the handle to the Xen hypervisor
 * @op: pointer to the hypervisor operation structure
 *
 * Do an hypervisor v2 stsyem operation, this leads to an hypervisor
 * call through ioctl.
 *
 * Returns 0 in case of success and -1 in case of error.
 */
static int
xenHypervisorDoV2Sys(int handle, xen_op_v2_sys* op)
{
    int ret;
    hypercall_t hc;

    memset(&hc, 0, sizeof(hc));
    op->interface_version = sys_interface_version;
    hc.op = __HYPERVISOR_sysctl;
    hc.arg[0] = (unsigned long) op;

    if (mlock(op, sizeof(dom0_op_t)) < 0) {
        virXenError(VIR_ERR_XEN_CALL, " locking", sizeof(*op));
        return (-1);
    }

    ret = ioctl(handle, xen_ioctl_hypercall_cmd, (unsigned long) &hc);
    if (ret < 0) {
        virXenError(VIR_ERR_XEN_CALL, " ioctl ", xen_ioctl_hypercall_cmd);
    }

    if (munlock(op, sizeof(dom0_op_t)) < 0) {
        virXenError(VIR_ERR_XEN_CALL, " releasing", sizeof(*op));
        ret = -1;
    }

    if (ret < 0)
        return (-1);

    return (0);
}

/**
 * xenHypervisorDoV2Dom:
 * @handle: the handle to the Xen hypervisor
 * @op: pointer to the hypervisor domain operation structure
 *
 * Do an hypervisor v2 domain operation, this leads to an hypervisor
 * call through ioctl.
 *
 * Returns 0 in case of success and -1 in case of error.
 */
static int
xenHypervisorDoV2Dom(int handle, xen_op_v2_dom* op)
{
    int ret;
    hypercall_t hc;

    memset(&hc, 0, sizeof(hc));
    op->interface_version = dom_interface_version;
    hc.op = __HYPERVISOR_domctl;
    hc.arg[0] = (unsigned long) op;

    if (mlock(op, sizeof(dom0_op_t)) < 0) {
        virXenError(VIR_ERR_XEN_CALL, " locking", sizeof(*op));
        return (-1);
    }

    ret = ioctl(handle, xen_ioctl_hypercall_cmd, (unsigned long) &hc);
    if (ret < 0) {
        virXenError(VIR_ERR_XEN_CALL, " ioctl ", xen_ioctl_hypercall_cmd);
    }

    if (munlock(op, sizeof(dom0_op_t)) < 0) {
        virXenError(VIR_ERR_XEN_CALL, " releasing", sizeof(*op));
        ret = -1;
    }

    if (ret < 0)
        return (-1);

    return (0);
}

/**
 * virXen_getdomaininfolist:
 * @handle: the hypervisor handle
 * @first_domain: first domain in the range
 * @maxids: maximum number of domains to list
 * @dominfos: output structures
 *
 * Do a low level hypercall to list existing domains informations
 *
 * Returns the number of domains or -1 in case of failure
 */
static int
virXen_getdomaininfolist(int handle, int first_domain, int maxids,
713
                         xen_getdomaininfolist *dominfos)
714 715 716
{
    int ret = -1;

717 718
    if (mlock(XEN_GETDOMAININFOLIST_DATA(dominfos),
              XEN_GETDOMAININFO_SIZE * maxids) < 0) {
719
        virXenError(VIR_ERR_XEN_CALL, " locking",
720
                    XEN_GETDOMAININFO_SIZE * maxids);
721 722 723 724 725 726
        return (-1);
    }
    if (hypervisor_version > 1) {
        xen_op_v2_sys op;

        memset(&op, 0, sizeof(op));
727 728 729 730 731 732 733 734
        op.cmd = XEN_V2_OP_GETDOMAININFOLIST;
        op.u.getdomaininfolist.first_domain = (domid_t) first_domain;
        op.u.getdomaininfolist.max_domains = maxids;
        op.u.getdomaininfolist.buffer = dominfos->v2;
        op.u.getdomaininfolist.num_domains = maxids;
        ret = xenHypervisorDoV2Sys(handle, &op);
        if (ret == 0)
            ret = op.u.getdomaininfolist.num_domains;
735 736 737 738
    } else if (hypervisor_version == 1) {
        xen_op_v1 op;

        memset(&op, 0, sizeof(op));
739 740 741 742 743 744 745 746
        op.cmd = XEN_V1_OP_GETDOMAININFOLIST;
        op.u.getdomaininfolist.first_domain = (domid_t) first_domain;
        op.u.getdomaininfolist.max_domains = maxids;
        op.u.getdomaininfolist.buffer = dominfos->v0;
        op.u.getdomaininfolist.num_domains = maxids;
        ret = xenHypervisorDoV1Op(handle, &op);
        if (ret == 0)
            ret = op.u.getdomaininfolist.num_domains;
747 748 749 750
    } else if (hypervisor_version == 0) {
        xen_op_v0 op;

        memset(&op, 0, sizeof(op));
751 752 753 754 755 756 757 758
        op.cmd = XEN_V0_OP_GETDOMAININFOLIST;
        op.u.getdomaininfolist.first_domain = (domid_t) first_domain;
        op.u.getdomaininfolist.max_domains = maxids;
        op.u.getdomaininfolist.buffer = dominfos->v0;
        op.u.getdomaininfolist.num_domains = maxids;
        ret = xenHypervisorDoV0Op(handle, &op);
        if (ret == 0)
            ret = op.u.getdomaininfolist.num_domains;
759
    }
760 761
    if (munlock(XEN_GETDOMAININFOLIST_DATA(dominfos),
                XEN_GETDOMAININFO_SIZE * maxids) < 0) {
762
        virXenError(VIR_ERR_XEN_CALL, " release",
763
                    XEN_GETDOMAININFO_SIZE * maxids);
764 765 766 767 768
        ret = -1;
    }
    return(ret);
}

769 770 771 772 773 774 775 776 777 778 779 780 781 782 783
static int
virXen_getdomaininfo(int handle, int first_domain,
                     xen_getdomaininfo *dominfo) {
    xen_getdomaininfolist dominfos;

    if (hypervisor_version < 2) {
        dominfos.v0 = &(dominfo->v0);
    } else {
        dominfos.v2 = &(dominfo->v2);
    }

    return virXen_getdomaininfolist(handle, first_domain, 1, &dominfos);
}


784
#ifndef PROXY
785 786 787 788 789 790 791 792 793 794
/**
 * virXen_pausedomain:
 * @handle: the hypervisor handle
 * @id: the domain id
 *
 * Do a low level hypercall to pause the domain
 *
 * Returns 0 or -1 in case of failure
 */
static int
795
virXen_pausedomain(int handle, int id)
796 797 798 799 800 801 802
{
    int ret = -1;

    if (hypervisor_version > 1) {
        xen_op_v2_dom op;

        memset(&op, 0, sizeof(op));
803 804 805
        op.cmd = XEN_V2_OP_PAUSEDOMAIN;
        op.domain = (domid_t) id;
        ret = xenHypervisorDoV2Dom(handle, &op);
806 807 808 809
    } else if (hypervisor_version == 1) {
        xen_op_v1 op;

        memset(&op, 0, sizeof(op));
810 811 812
        op.cmd = XEN_V1_OP_PAUSEDOMAIN;
        op.u.domain.domain = (domid_t) id;
        ret = xenHypervisorDoV1Op(handle, &op);
813 814 815 816
    } else if (hypervisor_version == 0) {
        xen_op_v0 op;

        memset(&op, 0, sizeof(op));
817 818 819
        op.cmd = XEN_V0_OP_PAUSEDOMAIN;
        op.u.domain.domain = (domid_t) id;
        ret = xenHypervisorDoV0Op(handle, &op);
820 821 822 823 824 825 826 827 828 829 830 831 832 833
    }
    return(ret);
}

/**
 * virXen_unpausedomain:
 * @handle: the hypervisor handle
 * @id: the domain id
 *
 * Do a low level hypercall to unpause the domain
 *
 * Returns 0 or -1 in case of failure
 */
static int
834
virXen_unpausedomain(int handle, int id)
835 836 837 838 839 840 841
{
    int ret = -1;

    if (hypervisor_version > 1) {
        xen_op_v2_dom op;

        memset(&op, 0, sizeof(op));
842 843 844
        op.cmd = XEN_V2_OP_UNPAUSEDOMAIN;
        op.domain = (domid_t) id;
        ret = xenHypervisorDoV2Dom(handle, &op);
845 846 847 848
    } else if (hypervisor_version == 1) {
        xen_op_v1 op;

        memset(&op, 0, sizeof(op));
849 850 851
        op.cmd = XEN_V1_OP_UNPAUSEDOMAIN;
        op.u.domain.domain = (domid_t) id;
        ret = xenHypervisorDoV1Op(handle, &op);
852 853 854 855
    } else if (hypervisor_version == 0) {
        xen_op_v0 op;

        memset(&op, 0, sizeof(op));
856 857 858
        op.cmd = XEN_V0_OP_UNPAUSEDOMAIN;
        op.u.domain.domain = (domid_t) id;
        ret = xenHypervisorDoV0Op(handle, &op);
859 860 861 862 863 864 865 866 867 868 869 870 871 872
    }
    return(ret);
}

/**
 * virXen_destroydomain:
 * @handle: the hypervisor handle
 * @id: the domain id
 *
 * Do a low level hypercall to destroy the domain
 *
 * Returns 0 or -1 in case of failure
 */
static int
873
virXen_destroydomain(int handle, int id)
874 875 876 877 878 879 880
{
    int ret = -1;

    if (hypervisor_version > 1) {
        xen_op_v2_dom op;

        memset(&op, 0, sizeof(op));
881 882 883
        op.cmd = XEN_V2_OP_DESTROYDOMAIN;
        op.domain = (domid_t) id;
        ret = xenHypervisorDoV2Dom(handle, &op);
884 885 886 887
    } else if (hypervisor_version == 1) {
        xen_op_v1 op;

        memset(&op, 0, sizeof(op));
888 889 890
        op.cmd = XEN_V1_OP_DESTROYDOMAIN;
        op.u.domain.domain = (domid_t) id;
        ret = xenHypervisorDoV1Op(handle, &op);
891 892 893 894
    } else if (hypervisor_version == 0) {
        xen_op_v0 op;

        memset(&op, 0, sizeof(op));
895 896 897
        op.cmd = XEN_V0_OP_DESTROYDOMAIN;
        op.u.domain.domain = (domid_t) id;
        ret = xenHypervisorDoV0Op(handle, &op);
898 899 900 901 902 903 904 905 906 907 908 909 910 911 912
    }
    return(ret);
}

/**
 * virXen_setmaxmem:
 * @handle: the hypervisor handle
 * @id: the domain id
 * @memory: the amount of memory in kilobytes
 *
 * Do a low level hypercall to change the max memory amount
 *
 * Returns 0 or -1 in case of failure
 */
static int
913
virXen_setmaxmem(int handle, int id, unsigned long memory)
914 915 916 917 918 919 920
{
    int ret = -1;

    if (hypervisor_version > 1) {
        xen_op_v2_dom op;

        memset(&op, 0, sizeof(op));
921 922 923 924
        op.cmd = XEN_V2_OP_SETMAXMEM;
        op.domain = (domid_t) id;
        op.u.setmaxmem.maxmem = memory;
        ret = xenHypervisorDoV2Dom(handle, &op);
925 926 927 928
    } else if (hypervisor_version == 1) {
        xen_op_v1 op;

        memset(&op, 0, sizeof(op));
929 930 931 932
        op.cmd = XEN_V1_OP_SETMAXMEM;
        op.u.setmaxmem.domain = (domid_t) id;
        op.u.setmaxmem.maxmem = memory;
        ret = xenHypervisorDoV1Op(handle, &op);
933
    } else if (hypervisor_version == 0) {
934
        xen_op_v0 op;
935 936

        memset(&op, 0, sizeof(op));
937 938 939 940
        op.cmd = XEN_V0_OP_SETMAXMEM;
        op.u.setmaxmem.domain = (domid_t) id;
        op.u.setmaxmem.maxmem = memory;
        ret = xenHypervisorDoV0Op(handle, &op);
941 942 943 944 945 946 947 948 949 950 951 952 953 954 955
    }
    return(ret);
}

/**
 * virXen_setmaxvcpus:
 * @handle: the hypervisor handle
 * @id: the domain id
 * @vcpus: the numbers of vcpus
 *
 * Do a low level hypercall to change the max vcpus amount
 *
 * Returns 0 or -1 in case of failure
 */
static int
956
virXen_setmaxvcpus(int handle, int id, unsigned int vcpus)
957 958 959 960 961 962 963
{
    int ret = -1;

    if (hypervisor_version > 1) {
        xen_op_v2_dom op;

        memset(&op, 0, sizeof(op));
964 965 966 967
        op.cmd = XEN_V2_OP_SETMAXVCPU;
        op.domain = (domid_t) id;
        op.u.setmaxvcpu.maxvcpu = vcpus;
        ret = xenHypervisorDoV2Dom(handle, &op);
968 969 970 971
    } else if (hypervisor_version == 1) {
        xen_op_v1 op;

        memset(&op, 0, sizeof(op));
972 973 974 975
        op.cmd = XEN_V1_OP_SETMAXVCPU;
        op.u.setmaxvcpu.domain = (domid_t) id;
        op.u.setmaxvcpu.maxvcpu = vcpus;
        ret = xenHypervisorDoV1Op(handle, &op);
976
    } else if (hypervisor_version == 0) {
977
        xen_op_v0 op;
978 979

        memset(&op, 0, sizeof(op));
980 981 982 983
        op.cmd = XEN_V0_OP_SETMAXVCPU;
        op.u.setmaxvcpu.domain = (domid_t) id;
        op.u.setmaxvcpu.maxvcpu = vcpus;
        ret = xenHypervisorDoV0Op(handle, &op);
984 985 986 987 988 989 990 991 992 993
    }
    return(ret);
}

/**
 * virXen_setvcpumap:
 * @handle: the hypervisor handle
 * @id: the domain id
 * @vcpu: the vcpu to map
 * @cpumap: the bitmap for this vcpu
994
 * @maplen: the size of the bitmap in bytes
995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008
 *
 * Do a low level hypercall to change the pinning for vcpu
 *
 * Returns 0 or -1 in case of failure
 */
static int
virXen_setvcpumap(int handle, int id, unsigned int vcpu,
                  unsigned char * cpumap, int maplen)
{
    int ret = -1;

    if (hypervisor_version > 1) {
        xen_op_v2_dom op;

1009 1010 1011 1012
        if (mlock(cpumap, maplen) < 0) {
            virXenError(VIR_ERR_XEN_CALL, " locking", maplen);
            return (-1);
        }
1013
        memset(&op, 0, sizeof(op));
1014 1015 1016 1017 1018 1019 1020 1021 1022 1023
        op.cmd = XEN_V2_OP_SETVCPUMAP;
        op.domain = (domid_t) id;
        op.u.setvcpumap.vcpu = vcpu;
        op.u.setvcpumap.cpumap.bitmap = cpumap;
        op.u.setvcpumap.cpumap.nr_cpus = maplen * 8;
        ret = xenHypervisorDoV2Dom(handle, &op);
        if (munlock(cpumap, maplen) < 0) {
            virXenError(VIR_ERR_XEN_CALL, " release", maplen);
            ret = -1;
        }
1024
    } else {
1025 1026 1027
        cpumap_t xen_cpumap; /* limited to 64 CPUs in old hypervisors */
        uint64_t *pm = &xen_cpumap;
        int j;
1028

1029 1030
        if ((maplen > (int)sizeof(cpumap_t)) || (sizeof(cpumap_t) & 7))
            return (-1);
1031

1032 1033 1034
        memset(pm, 0, sizeof(cpumap_t));
        for (j = 0; j < maplen; j++)
            *(pm + (j / 8)) |= cpumap[j] << (8 * (j & 7));
1035 1036

        if (hypervisor_version == 1) {
1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054
            xen_op_v1 op;

            memset(&op, 0, sizeof(op));
            op.cmd = XEN_V1_OP_SETVCPUMAP;
            op.u.setvcpumap.domain = (domid_t) id;
            op.u.setvcpumap.vcpu = vcpu;
            op.u.setvcpumap.cpumap = xen_cpumap;
            ret = xenHypervisorDoV1Op(handle, &op);
        } else if (hypervisor_version == 0) {
            xen_op_v0 op;

            memset(&op, 0, sizeof(op));
            op.cmd = XEN_V0_OP_SETVCPUMAP;
            op.u.setvcpumap.domain = (domid_t) id;
            op.u.setvcpumap.vcpu = vcpu;
            op.u.setvcpumap.cpumap = xen_cpumap;
            ret = xenHypervisorDoV0Op(handle, &op);
        }
1055 1056 1057
    }
    return(ret);
}
1058 1059
#endif /* !PROXY*/

1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073
/**
 * virXen_getvcpusinfo:
 * @handle: the hypervisor handle
 * @id: the domain id
 * @vcpu: the vcpu to map
 * @cpumap: the bitmap for this vcpu
 * @maplen: the size of the bitmap in bytes
 *
 * Do a low level hypercall to change the pinning for vcpu
 *
 * Returns 0 or -1 in case of failure
 */
static int
virXen_getvcpusinfo(int handle, int id, unsigned int vcpu, virVcpuInfoPtr ipt,
1074
                    unsigned char *cpumap, int maplen)
1075 1076 1077 1078 1079 1080 1081
{
    int ret = -1;

    if (hypervisor_version > 1) {
        xen_op_v2_dom op;

        memset(&op, 0, sizeof(op));
1082 1083
        op.cmd = XEN_V2_OP_GETVCPUINFO;
        op.domain = (domid_t) id;
1084
        op.u.getvcpuinfo.vcpu = (uint16_t) vcpu;
1085 1086 1087
        ret = xenHypervisorDoV2Dom(handle, &op);
        if (ret < 0)
            return(-1);
1088
        ipt->number = vcpu;
1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112
        if (op.u.getvcpuinfo.online) {
            if (op.u.getvcpuinfo.running) ipt->state = VIR_VCPU_RUNNING;
            if (op.u.getvcpuinfo.blocked) ipt->state = VIR_VCPU_BLOCKED;
        }
        else ipt->state = VIR_VCPU_OFFLINE;
        ipt->cpuTime = op.u.getvcpuinfo.cpu_time;
        ipt->cpu = op.u.getvcpuinfo.online ? (int)op.u.getvcpuinfo.cpu : -1;
        if ((cpumap != NULL) && (maplen > 0)) {
            if (mlock(cpumap, maplen) < 0) {
                virXenError(VIR_ERR_XEN_CALL, " locking", maplen);
                return (-1);
            }
            memset(&op, 0, sizeof(op));
            op.cmd = XEN_V2_OP_GETVCPUMAP;
            op.domain = (domid_t) id;
            op.u.setvcpumap.vcpu = vcpu;
            op.u.setvcpumap.cpumap.bitmap = cpumap;
            op.u.setvcpumap.cpumap.nr_cpus = maplen * 8;
            ret = xenHypervisorDoV2Dom(handle, &op);
            if (munlock(cpumap, maplen) < 0) {
                virXenError(VIR_ERR_XEN_CALL, " release", maplen);
                ret = -1;
            }
        }
1113
    } else {
1114 1115 1116 1117 1118
        int mapl = maplen;
        int cpu;

        if (maplen > (int)sizeof(cpumap_t))
            mapl = (int)sizeof(cpumap_t);
1119 1120

        if (hypervisor_version == 1) {
1121 1122 1123 1124 1125 1126 1127 1128 1129
            xen_op_v1 op;

            memset(&op, 0, sizeof(op));
            op.cmd = XEN_V1_OP_GETVCPUINFO;
            op.u.getvcpuinfo.domain = (domid_t) id;
            op.u.getvcpuinfo.vcpu = vcpu;
            ret = xenHypervisorDoV1Op(handle, &op);
            if (ret < 0)
                return(-1);
1130
            ipt->number = vcpu;
1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153
            if (op.u.getvcpuinfo.online) {
                if (op.u.getvcpuinfo.running) ipt->state = VIR_VCPU_RUNNING;
                if (op.u.getvcpuinfo.blocked) ipt->state = VIR_VCPU_BLOCKED;
            }
            else ipt->state = VIR_VCPU_OFFLINE;
            ipt->cpuTime = op.u.getvcpuinfo.cpu_time;
            ipt->cpu = op.u.getvcpuinfo.online ? (int)op.u.getvcpuinfo.cpu : -1;
            if ((cpumap != NULL) && (maplen > 0)) {
                for (cpu = 0; cpu < (mapl * 8); cpu++) {
                    if (op.u.getvcpuinfo.cpumap & ((uint64_t)1<<cpu))
                        VIR_USE_CPU(cpumap, cpu);
                }
            }
        } else if (hypervisor_version == 0) {
            xen_op_v1 op;

            memset(&op, 0, sizeof(op));
            op.cmd = XEN_V0_OP_GETVCPUINFO;
            op.u.getvcpuinfo.domain = (domid_t) id;
            op.u.getvcpuinfo.vcpu = vcpu;
            ret = xenHypervisorDoV0Op(handle, &op);
            if (ret < 0)
                return(-1);
1154
            ipt->number = vcpu;
1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168
            if (op.u.getvcpuinfo.online) {
                if (op.u.getvcpuinfo.running) ipt->state = VIR_VCPU_RUNNING;
                if (op.u.getvcpuinfo.blocked) ipt->state = VIR_VCPU_BLOCKED;
            }
            else ipt->state = VIR_VCPU_OFFLINE;
            ipt->cpuTime = op.u.getvcpuinfo.cpu_time;
            ipt->cpu = op.u.getvcpuinfo.online ? (int)op.u.getvcpuinfo.cpu : -1;
            if ((cpumap != NULL) && (maplen > 0)) {
                for (cpu = 0; cpu < (mapl * 8); cpu++) {
                    if (op.u.getvcpuinfo.cpumap & ((uint64_t)1<<cpu))
                        VIR_USE_CPU(cpumap, cpu);
                }
            }
        }
1169 1170 1171
    }
    return(ret);
}
1172

1173 1174 1175 1176 1177 1178
/**
 * xenHypervisorInit:
 *
 * Initialize the hypervisor layer. Try to detect the kind of interface
 * used i.e. pre or post changeset 10277
 */
1179
static int
D
Daniel P. Berrange 已提交
1180
xenHypervisorInit(void)
1181
{
1182
    int fd, ret, cmd, errcode;
1183
    hypercall_t hc;
1184
    v0_hypercall_t v0_hc;
1185
    xen_getdomaininfo info;
1186
    virVcpuInfoPtr ipt;
1187 1188

    if (initialized) {
1189
        if (hypervisor_version == -1)
1190 1191
            return(-1);
        return(0);
1192 1193
    }
    initialized = 1;
1194
    in_init = 1;
1195

1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231
    /* Compile regular expressions used by xenHypervisorGetCapabilities.
     * Note that errors here are really internal errors since these
     * regexps should never fail to compile.
     */
    errcode = regcomp (&flags_hvm_rec, flags_hvm_re, REG_EXTENDED);
    if (errcode != 0) {
        char error[100];
        regerror (errcode, &flags_hvm_rec, error, sizeof error);
        regfree (&flags_hvm_rec);
        virXenError (VIR_ERR_INTERNAL_ERROR, error, 0);
        in_init = 0;
        return -1;
    }
    errcode = regcomp (&flags_pae_rec, flags_pae_re, REG_EXTENDED);
    if (errcode != 0) {
        char error[100];
        regerror (errcode, &flags_pae_rec, error, sizeof error);
        regfree (&flags_pae_rec);
        regfree (&flags_hvm_rec);
        virXenError (VIR_ERR_INTERNAL_ERROR, error, 0);
        in_init = 0;
        return -1;
    }
    errcode = regcomp (&xen_cap_rec, xen_cap_re, REG_EXTENDED);
    if (errcode != 0) {
        char error[100];
        regerror (errcode, &xen_cap_rec, error, sizeof error);
        regfree (&xen_cap_rec);
        regfree (&flags_pae_rec);
        regfree (&flags_hvm_rec);
        virXenError (VIR_ERR_INTERNAL_ERROR, error, 0);
        in_init = 0;
        return -1;
    }

    /* Xen hypervisor version detection begins. */
1232 1233
    ret = open(XEN_HYPERVISOR_SOCKET, O_RDWR);
    if (ret < 0) {
1234
        hypervisor_version = -1;
1235 1236 1237 1238
        return (-1);
    }
    fd = ret;

1239 1240 1241 1242
    /*
     * The size of the hypervisor call block changed July 2006
     * this detect if we are using the new or old hypercall_t structure
     */
1243 1244 1245 1246 1247 1248 1249 1250
    hc.op = __HYPERVISOR_xen_version;
    hc.arg[0] = (unsigned long) XENVER_version;
    hc.arg[1] = 0;

    cmd = IOCTL_PRIVCMD_HYPERCALL;
    ret = ioctl(fd, cmd, (unsigned long) &hc);

    if ((ret != -1) && (ret != 0)) {
D
Daniel Veillard 已提交
1251
#ifdef DEBUG
1252
        fprintf(stderr, "Using new hypervisor call: %X\n", ret);
D
Daniel Veillard 已提交
1253
#endif
1254 1255 1256
        hv_version = ret;
        xen_ioctl_hypercall_cmd = cmd;
        goto detect_v2;
1257
    }
1258

1259 1260 1261 1262 1263 1264 1265 1266
    /*
     * check if the old hypercall are actually working
     */
    v0_hc.op = __HYPERVISOR_xen_version;
    v0_hc.arg[0] = (unsigned long) XENVER_version;
    v0_hc.arg[1] = 0;
    cmd = _IOC(_IOC_NONE, 'P', 0, sizeof(v0_hypercall_t));
    ret = ioctl(fd, cmd, (unsigned long) &v0_hc);
1267
    if ((ret != -1) && (ret != 0)) {
D
Daniel Veillard 已提交
1268
#ifdef DEBUG
1269
        fprintf(stderr, "Using old hypervisor call: %X\n", ret);
D
Daniel Veillard 已提交
1270
#endif
1271 1272
        hv_version = ret;
        xen_ioctl_hypercall_cmd = cmd;
1273
        hypervisor_version = 0;
1274
        goto done;
1275 1276
    }

1277 1278 1279 1280 1281
    /*
     * we faild to make any hypercall
     */

    hypervisor_version = -1;
1282 1283
    virXenError(VIR_ERR_XEN_CALL, " ioctl ", IOCTL_PRIVCMD_HYPERCALL);
    close(fd);
1284 1285 1286
    in_init = 0;
    return(-1);

1287
 detect_v2:
1288 1289 1290 1291 1292 1293
    /*
     * The hypercalls were refactored into 3 different section in August 2006
     * Try to detect if we are running a version post 3.0.2 with the new ones
     * or the old ones
     */
    hypervisor_version = 2;
1294 1295 1296 1297 1298 1299 1300 1301 1302

    ipt = malloc(sizeof(virVcpuInfo));
    if (ipt == NULL){
#ifdef DEBUG
        fprintf(stderr, "Memory allocation failed at xenHypervisorIniti()\n");
#endif
        return(-1);
    }
    /* Currently consider RHEL5.0 Fedora7 and xen-unstable */
1303
    sys_interface_version = 2; /* XEN_SYSCTL_INTERFACE_VERSION */
1304
    if (virXen_getdomaininfo(fd, 0, &info) == 1) {
1305 1306 1307
        /* RHEL 5.0 */
        dom_interface_version = 3; /* XEN_DOMCTL_INTERFACE_VERSION */
        if (virXen_getvcpusinfo(fd, 0, 0, ipt, NULL, 0) == 0){
D
Daniel Veillard 已提交
1308
#ifdef DEBUG
1309
            fprintf(stderr, "Using hypervisor call v2, sys ver2 dom ver3\n");
D
Daniel Veillard 已提交
1310
#endif
1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332
            goto done;
        }
        /* Fedora 7 */
        dom_interface_version = 4; /* XEN_DOMCTL_INTERFACE_VERSION */
        if (virXen_getvcpusinfo(fd, 0, 0, ipt, NULL, 0) == 0){
#ifdef DEBUG
            fprintf(stderr, "Using hypervisor call v2, sys ver2 dom ver4\n");
#endif
            goto done;
        }
    }

    sys_interface_version = 3; /* XEN_SYSCTL_INTERFACE_VERSION */
    if (virXen_getdomaininfo(fd, 0, &info) == 1) {
        /* xen-unstable */
        dom_interface_version = 5; /* XEN_DOMCTL_INTERFACE_VERSION */
        if (virXen_getvcpusinfo(fd, 0, 0, ipt, NULL, 0) == 0){
#ifdef DEBUG
            fprintf(stderr, "Using hypervisor call v2, sys ver3 dom ver5\n");
#endif
            goto done;
        }
1333
    }
1334

1335 1336
    hypervisor_version = 1;
    sys_interface_version = -1;
1337
    if (virXen_getdomaininfo(fd, 0, &info) == 1) {
D
Daniel Veillard 已提交
1338
#ifdef DEBUG
1339
        fprintf(stderr, "Using hypervisor call v1\n");
D
Daniel Veillard 已提交
1340
#endif
1341
        goto done;
1342 1343 1344 1345 1346 1347 1348 1349 1350 1351
    }

    /*
     * we faild to make the getdomaininfolist hypercall
     */

    hypervisor_version = -1;
    virXenError(VIR_ERR_XEN_CALL, " ioctl ", IOCTL_PRIVCMD_HYPERCALL);
    close(fd);
    in_init = 0;
1352 1353
    return(-1);

1354
 done:
1355
    close(fd);
1356
    in_init = 0;
1357 1358 1359
    return(0);
}

1360
#ifndef PROXY
1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372
/**
 * xenHypervisorRegister:
 *
 * Registers the xenHypervisor driver
 */
void xenHypervisorRegister(void)
{
    if (initialized == 0)
        xenHypervisorInit();

    virRegisterDriver(&xenHypervisorDriver);
}
1373
#endif /* !PROXY */
1374

1375 1376
/**
 * xenHypervisorOpen:
1377 1378 1379
 * @conn: pointer to the connection block
 * @name: URL for the target, NULL for local
 * @flags: combination of virDrvOpenFlag(s)
1380 1381 1382
 *
 * Connects to the Xen hypervisor.
 *
1383
 * Returns 0 or -1 in case of error.
1384
 */
1385
int
1386
xenHypervisorOpen(virConnectPtr conn, const char *name, int flags)
1387
{
1388 1389
    int ret;

1390 1391 1392
    if (initialized == 0)
        xenHypervisorInit();

1393
    if ((name != NULL) && (strcasecmp(name, "xen")))
1394 1395 1396 1397
        return(-1);

    conn->handle = -1;

1398
    ret = open(XEN_HYPERVISOR_SOCKET, O_RDWR);
1399
    if (ret < 0) {
1400
        if (!(flags & VIR_DRV_OPEN_QUIET))
1401 1402
            virXenError(VIR_ERR_NO_XEN, XEN_HYPERVISOR_SOCKET, 0);
        return (-1);
1403
    }
1404
    conn->handle = ret;
1405

1406
    return(0);
1407 1408 1409 1410
}

/**
 * xenHypervisorClose:
1411
 * @conn: pointer to the connection block
1412 1413 1414 1415 1416
 *
 * Close the connection to the Xen hypervisor.
 *
 * Returns 0 in case of success or -1 in case of error.
 */
1417
int
1418
xenHypervisorClose(virConnectPtr conn)
1419
{
1420 1421
    int ret;

1422
    if ((conn == NULL) || (conn->handle < 0))
1423
        return (-1);
1424

1425
    ret = close(conn->handle);
1426
    if (ret < 0)
1427 1428
        return (-1);
    return (0);
1429 1430 1431
}


1432
#ifndef PROXY
1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451
/**
 * xenHypervisorGetType:
 * @conn: pointer to the Xen Hypervisor block
 *
 * Get the version level of the Hypervisor running.
 *
 * Returns -1 in case of error, 0 otherwise. if the version can't be
 *    extracted by lack of capacities returns 0 and @hvVer is 0, otherwise
 *    @hvVer value is major * 1,000,000 + minor * 1,000 + release
 */
static const char *
xenHypervisorGetType(virConnectPtr conn)
{
    if (!VIR_IS_CONNECT(conn)) {
        virXenError(VIR_ERR_INVALID_CONN, __FUNCTION__, 0);
        return (NULL);
    }
    return("Xen");
}
1452
#endif
1453

1454 1455
/**
 * xenHypervisorGetVersion:
1456 1457
 * @conn: pointer to the connection block
 * @hvVer: where to store the version
1458 1459 1460
 *
 * Call the hypervisor to extracts his own internal API version
 *
1461
 * Returns 0 in case of success, -1 in case of error
1462
 */
1463 1464
int
xenHypervisorGetVersion(virConnectPtr conn, unsigned long *hvVer)
1465
{
1466 1467
    if ((conn == NULL) || (conn->handle < 0) || (hvVer == NULL))
        return (-1);
1468
    *hvVer = (hv_version >> 16) * 1000000 + (hv_version & 0xFFFF) * 1000;
1469
    return(0);
1470 1471
}

1472 1473 1474
/**
 * xenHypervisorGetCapabilities:
 * @conn: pointer to the connection block
1475 1476
 * @cpuinfo: file handle containing /proc/cpuinfo data, or NULL
 * @capabilities: file handle containing /sys/hypervisor/properties/capabilities data, or NULL
1477 1478 1479 1480
 *
 * Return the capabilities of this hypervisor.
 */
char *
1481 1482 1483
xenHypervisorMakeCapabilitiesXML(virConnectPtr conn ATTRIBUTE_UNUSED,
                                 const char *hostmachine,
                                 FILE *cpuinfo, FILE *capabilities)
1484 1485
{
    char line[1024], *str, *token;
1486
    regmatch_t subs[4];
1487
    char *saveptr = NULL;
1488 1489 1490 1491
    int i, r;

    char hvm_type[4] = ""; /* "vmx" or "svm" (or "" if not in CPU). */
    int host_pae = 0;
1492
    struct guest_arch {
1493 1494 1495 1496
        const char *model;
        int bits;
        int hvm;
        int pae;
1497
        int nonpae;
1498
        int ia64_be;
1499
    } guest_archs[32];
1500 1501 1502 1503 1504
    int nr_guest_archs = 0;

    virBufferPtr xml;
    char *xml_str;

1505 1506
    memset(guest_archs, 0, sizeof(guest_archs));

1507 1508 1509 1510
    /* /proc/cpuinfo: flags: Intel calls HVM "vmx", AMD calls it "svm".
     * It's not clear if this will work on IA64, let alone other
     * architectures and non-Linux. (XXX)
     */
1511 1512 1513 1514 1515 1516 1517 1518 1519 1520
    if (cpuinfo) {
        while (fgets (line, sizeof line, cpuinfo)) {
            if (regexec (&flags_hvm_rec, line, sizeof(subs)/sizeof(regmatch_t), subs, 0) == 0
                && subs[0].rm_so != -1) {
                strncpy (hvm_type,
                         &line[subs[1].rm_so], subs[1].rm_eo-subs[1].rm_so+1);
                hvm_type[subs[1].rm_eo-subs[1].rm_so] = '\0';
            } else if (regexec (&flags_pae_rec, line, 0, NULL, 0) == 0)
                host_pae = 1;
        }
1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547
    }

    /* Most of the useful info is in /sys/hypervisor/properties/capabilities
     * which is documented in the code in xen-unstable.hg/xen/arch/.../setup.c.
     *
     * It is a space-separated list of supported guest architectures.
     *
     * For x86:
     *    TYP-VER-ARCH[p]
     *    ^   ^   ^    ^
     *    |   |   |    +-- PAE supported
     *    |   |   +------- x86_32 or x86_64
     *    |   +----------- the version of Xen, eg. "3.0"
     *    +--------------- "xen" or "hvm" for para or full virt respectively
     *
     * For PPC this file appears to be always empty (?)
     *
     * For IA64:
     *    TYP-VER-ARCH[be]
     *    ^   ^   ^    ^
     *    |   |   |    +-- Big-endian supported
     *    |   |   +------- always "ia64"
     *    |   +----------- the version of Xen, eg. "3.0"
     *    +--------------- "xen" or "hvm" for para or full virt respectively
     */

    /* Expecting one line in this file - ignore any more. */
1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586
    if (fgets (line, sizeof line, capabilities)) {
        /* Split the line into tokens.  strtok_r is OK here because we "own"
         * this buffer.  Parse out the features from each token.
         */
        for (str = line, nr_guest_archs = 0;
             nr_guest_archs < sizeof guest_archs / sizeof guest_archs[0]
                 && (token = strtok_r (str, " ", &saveptr)) != NULL;
             str = NULL) {

            if (regexec (&xen_cap_rec, token, sizeof subs / sizeof subs[0],
                         subs, 0) == 0) {
                int hvm = strncmp (&token[subs[1].rm_so], "hvm", 3) == 0;
                const char *model;
                int bits, pae = 0, nonpae = 0, ia64_be = 0;
                if (strncmp (&token[subs[2].rm_so], "x86_32", 6) == 0) {
                    model = "i686";
                    bits = 32;
                    if (strncmp (&token[subs[3].rm_so], "p", 1) == 0)
                        pae = 1;
                    else
                        nonpae = 1;
                }
                else if (strncmp (&token[subs[2].rm_so], "x86_64", 6) == 0) {
                    model = "x86_64";
                    bits = 64;
                }
                else if (strncmp (&token[subs[2].rm_so], "ia64", 4) == 0) {
                    model = "ia64";
                    bits = 64;
                    if (strncmp (&token[subs[3].rm_so], "be", 2) == 0)
                        ia64_be = 1;
                }
                else if (strncmp (&token[subs[2].rm_so], "powerpc64", 4) == 0) {
                    model = "ppc64";
                    bits = 64;
                } else {
                    /* XXX surely no other Xen archs exist  */
                    continue;
                }
1587

1588 1589 1590 1591 1592 1593 1594
                /* Search for existing matching (model,hvm) tuple */
                for (i = 0 ; i < nr_guest_archs ; i++) {
                    if (!strcmp(guest_archs[i].model, model) &&
                        guest_archs[i].hvm == hvm) {
                        break;
                    }
                }
1595

1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616
                /* Too many arch flavours - highly unlikely ! */
                if (i >= sizeof(guest_archs)/sizeof(guest_archs[0]))
                    continue;
                /* Didn't find a match, so create a new one */
                if (i == nr_guest_archs)
                    nr_guest_archs++;

                guest_archs[i].model = model;
                guest_archs[i].bits = bits;
                guest_archs[i].hvm = hvm;

                /* Careful not to overwrite a previous positive
                   setting with a negative one here - some archs
                   can do both pae & non-pae, but Xen reports
                   separately capabilities so we're merging archs */
                if (pae)
                    guest_archs[i].pae = pae;
                if (nonpae)
                    guest_archs[i].nonpae = nonpae;
                if (ia64_be)
                    guest_archs[i].ia64_be = ia64_be;
1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630
            }
        }
    }

    /* Construct the final XML. */
    xml = virBufferNew (1024);
    if (!xml) return NULL;
    r = virBufferVSprintf (xml,
                           "\
<capabilities>\n\
  <host>\n\
    <cpu>\n\
      <arch>%s</arch>\n\
      <features>\n",
1631 1632
                           hostmachine);
    if (r == -1) goto vir_buffer_failed;
1633 1634 1635 1636 1637 1638

    if (strcmp (hvm_type, "") != 0) {
        r = virBufferVSprintf (xml,
                               "\
        <%s/>\n",
                               hvm_type);
1639
        if (r == -1) goto vir_buffer_failed;
1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 1656 1657 1658 1659 1660
    }
    if (host_pae) {
        r = virBufferAdd (xml, "\
        <pae/>\n", -1);
        if (r == -1) goto vir_buffer_failed;
    }
    r = virBufferAdd (xml,
                      "\
      </features>\n\
    </cpu>\n\
  </host>\n", -1);
    if (r == -1) goto vir_buffer_failed;

    for (i = 0; i < nr_guest_archs; ++i) {
        r = virBufferVSprintf (xml,
                               "\
\n\
  <guest>\n\
    <os_type>%s</os_type>\n\
    <arch name=\"%s\">\n\
      <wordsize>%d</wordsize>\n\
1661
      <domain type=\"xen\"></domain>\n",
1662 1663
                               guest_archs[i].hvm ? "hvm" : "xen",
                               guest_archs[i].model,
1664
                               guest_archs[i].bits);
1665 1666
        if (r == -1) goto vir_buffer_failed;
        if (guest_archs[i].hvm) {
1667
            r = virBufferVSprintf (xml,
1668
                              "\
1669
      <emulator>/usr/lib%s/xen/bin/qemu-dm</emulator>\n\
1670 1671
      <machine>pc</machine>\n\
      <machine>isapc</machine>\n\
1672 1673
      <loader>/usr/lib/xen/boot/hvmloader</loader>\n",
                                   guest_archs[i].bits == 64 ? "64" : "");
1674 1675 1676 1677 1678
            if (r == -1) goto vir_buffer_failed;
        }
        r = virBufferAdd (xml,
                          "\
    </arch>\n\
1679
    <features>\n", -1);
1680 1681 1682 1683
        if (r == -1) goto vir_buffer_failed;
        if (guest_archs[i].pae) {
            r = virBufferAdd (xml,
                              "\
1684 1685 1686 1687 1688 1689
      <pae/>\n", -1);
            if (r == -1) goto vir_buffer_failed;
        }
        if (guest_archs[i].nonpae) {
            r = virBufferAdd (xml,
                              "\
1690
      <nonpae/>\n", -1);
1691 1692 1693 1694 1695
            if (r == -1) goto vir_buffer_failed;
        }
        if (guest_archs[i].ia64_be) {
            r = virBufferAdd (xml,
                              "\
1696
      <ia64_be/>\n", -1);
1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 1716
            if (r == -1) goto vir_buffer_failed;
        }
        r = virBufferAdd (xml,
                          "\
    </features>\n\
  </guest>\n", -1);
        if (r == -1) goto vir_buffer_failed;
    }
    r = virBufferAdd (xml,
                      "\
</capabilities>\n", -1);
    if (r == -1) goto vir_buffer_failed;
    xml_str = strdup (xml->content);
    if (!xml_str) {
        virXenError(VIR_ERR_NO_MEMORY, "strdup", 0);
        goto vir_buffer_failed;
    }
    virBufferFree (xml);

    return xml_str;
1717 1718 1719 1720 1721 1722 1723 1724 1725 1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741 1742 1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756 1757 1758 1759 1760 1761 1762 1763

 vir_buffer_failed:
    virBufferFree (xml);
    return NULL;
}

/**
 * xenHypervisorGetCapabilities:
 * @conn: pointer to the connection block
 *
 * Return the capabilities of this hypervisor.
 */
char *
xenHypervisorGetCapabilities (virConnectPtr conn)
{
    char *xml;
    FILE *cpuinfo, *capabilities;
    struct utsname utsname;

    /* Really, this never fails - look at the man-page. */
    uname (&utsname);

    cpuinfo = fopen ("/proc/cpuinfo", "r");
    if (cpuinfo == NULL) {
        if (errno != ENOENT) {
            virXenPerror (conn, "/proc/cpuinfo");
            return NULL;
        }
    }

    capabilities = fopen ("/sys/hypervisor/properties/capabilities", "r");
    if (capabilities == NULL) {
        if (errno != ENOENT) {
            fclose(cpuinfo);
            virXenPerror (conn, "/sys/hypervisor/properties/capabilities");
            return NULL;
        }
    }

    xml = xenHypervisorMakeCapabilitiesXML(conn, utsname.machine, cpuinfo, capabilities);

    if (cpuinfo)
        fclose(cpuinfo);
    if (capabilities)
        fclose(capabilities);

    return xml;
1764 1765
}

1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776
/**
 * xenHypervisorNumOfDomains:
 * @conn: pointer to the connection block
 *
 * Provides the number of active domains.
 *
 * Returns the number of domain found or -1 in case of error
 */
int
xenHypervisorNumOfDomains(virConnectPtr conn)
{
1777
    xen_getdomaininfolist dominfos;
1778 1779 1780 1781 1782 1783 1784
    int ret, nbids;
    static int last_maxids = 2;
    int maxids = last_maxids;

    if ((conn == NULL) || (conn->handle < 0))
        return (-1);

1785 1786
 retry:
    if (!(XEN_GETDOMAININFOLIST_ALLOC(dominfos, maxids))) {
1787
        virXenError(VIR_ERR_NO_MEMORY, _("allocating %d domain info"),
1788 1789
                    maxids);
        return(-1);
1790 1791
    }

1792 1793 1794
    XEN_GETDOMAININFOLIST_CLEAR(dominfos, maxids);

    ret = virXen_getdomaininfolist(conn->handle, 0, maxids, &dominfos);
1795

1796
    XEN_GETDOMAININFOLIST_FREE(dominfos);
1797 1798 1799 1800

    if (ret < 0)
        return (-1);

1801
    nbids = ret;
1802 1803 1804
    if (nbids == maxids) {
        last_maxids *= 2;
        maxids *= 2;
1805
        goto retry;
1806 1807 1808 1809 1810 1811 1812 1813 1814 1815 1816 1817 1818 1819 1820 1821 1822 1823 1824
    }
    if ((nbids < 0) || (nbids > maxids))
        return(-1);
    return(nbids);
}

/**
 * xenHypervisorListDomains:
 * @conn: pointer to the connection block
 * @ids: array to collect the list of IDs of active domains
 * @maxids: size of @ids
 *
 * Collect the list of active domains, and store their ID in @maxids
 *
 * Returns the number of domain found or -1 in case of error
 */
int
xenHypervisorListDomains(virConnectPtr conn, int *ids, int maxids)
{
1825
    xen_getdomaininfolist dominfos;
1826 1827 1828 1829 1830 1831
    int ret, nbids, i;

    if ((conn == NULL) || (conn->handle < 0) ||
        (ids == NULL) || (maxids < 1))
        return (-1);

1832
    if (!(XEN_GETDOMAININFOLIST_ALLOC(dominfos, maxids))) {
1833
        virXenError(VIR_ERR_NO_MEMORY, "allocating %d domain info",
1834 1835
                    maxids);
        return(-1);
1836
    }
1837 1838

    XEN_GETDOMAININFOLIST_CLEAR(dominfos, maxids);
1839 1840
    memset(ids, 0, maxids * sizeof(int));

1841
    ret = virXen_getdomaininfolist(conn->handle, 0, maxids, &dominfos);
1842 1843

    if (ret < 0) {
1844
        XEN_GETDOMAININFOLIST_FREE(dominfos);
1845 1846 1847
        return (-1);
    }

1848
    nbids = ret;
1849
    if ((nbids < 0) || (nbids > maxids)) {
1850
        XEN_GETDOMAININFOLIST_FREE(dominfos);
1851 1852 1853 1854
        return(-1);
    }

    for (i = 0;i < nbids;i++) {
1855
        ids[i] = XEN_GETDOMAININFOLIST_DOMAIN(dominfos, i);
1856 1857
    }

1858
    XEN_GETDOMAININFOLIST_FREE(dominfos);
1859 1860 1861
    return (nbids);
}

1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875
/**
 * xenHypervisorNumOfMaxVcpus:
 *
 * Returns the maximum of CPU defined by Xen.
 */
int
xenHypervisorNumOfMaxVcpus(virConnectPtr conn)
{
    if ((conn == NULL) || (conn->handle < 0))
        return (-1);

    return MAX_VIRT_CPUS;
}

1876
/**
1877 1878 1879
 * xenHypervisorGetDomMaxMemory:
 * @conn: connection data
 * @id: domain id
1880
 *
1881
 * Retrieve the maximum amount of physical memory allocated to a
1882
 * domain.
1883 1884 1885
 *
 * Returns the memory size in kilobytes or 0 in case of error.
 */
1886 1887
unsigned long
xenHypervisorGetDomMaxMemory(virConnectPtr conn, int id)
1888
{
1889
    xen_getdomaininfo dominfo;
1890 1891
    int ret;

1892
    if ((conn == NULL) || (conn->handle < 0))
1893 1894
        return (0);

1895 1896 1897 1898 1899 1900
    if (kb_per_pages == 0) {
        kb_per_pages = sysconf(_SC_PAGESIZE) / 1024;
	if (kb_per_pages <= 0) 
	    kb_per_pages = 4;
    }

1901
    XEN_GETDOMAININFO_CLEAR(dominfo);
1902

1903
    ret = virXen_getdomaininfo(conn->handle, id, &dominfo);
1904

1905
    if ((ret < 0) || (XEN_GETDOMAININFO_DOMAIN(dominfo) != id))
1906 1907
        return (0);

1908
    return((unsigned long) XEN_GETDOMAININFO_MAX_PAGES(dominfo) * kb_per_pages);
1909 1910
}

1911
#ifndef PROXY
1912 1913 1914
/**
 * xenHypervisorGetMaxMemory:
 * @domain: a domain object or NULL
1915
 *
1916 1917 1918 1919 1920 1921 1922 1923 1924 1925
 * Retrieve the maximum amount of physical memory allocated to a
 * domain. If domain is NULL, then this get the amount of memory reserved
 * to Domain0 i.e. the domain where the application runs.
 *
 * Returns the memory size in kilobytes or 0 in case of error.
 */
static unsigned long
xenHypervisorGetMaxMemory(virDomainPtr domain)
{
    if ((domain == NULL) || (domain->conn == NULL) ||
1926
        (domain->conn->handle < 0) || (domain->id < 0))
1927 1928
        return (0);

1929
    return(xenHypervisorGetDomMaxMemory(domain->conn, domain->id));
1930
}
1931
#endif
1932

1933
/**
1934 1935 1936
 * xenHypervisorGetDomInfo:
 * @conn: connection data
 * @id: the domain ID
1937
 * @info: the place where information should be stored
1938
 *
1939
 * Do an hypervisor call to get the related set of domain information.
1940 1941 1942 1943
 *
 * Returns 0 in case of success, -1 in case of error.
 */
int
1944
xenHypervisorGetDomInfo(virConnectPtr conn, int id, virDomainInfoPtr info)
1945
{
1946
    xen_getdomaininfo dominfo;
1947
    int ret;
1948
    uint32_t domain_flags, domain_state, domain_shutdown_cause;
1949 1950 1951 1952 1953 1954

    if (kb_per_pages == 0) {
        kb_per_pages = sysconf(_SC_PAGESIZE) / 1024;
	if (kb_per_pages <= 0) 
	    kb_per_pages = 4;
    }
1955

1956
    if ((conn == NULL) || (conn->handle < 0) || (info == NULL))
1957
        return (-1);
1958

1959
    memset(info, 0, sizeof(virDomainInfo));
1960
    XEN_GETDOMAININFO_CLEAR(dominfo);
1961

1962
    ret = virXen_getdomaininfo(conn->handle, id, &dominfo);
1963

1964
    if ((ret < 0) || (XEN_GETDOMAININFO_DOMAIN(dominfo) != id))
1965
        return (-1);
1966

1967 1968 1969
    domain_flags = XEN_GETDOMAININFO_FLAGS(dominfo);
    domain_state = domain_flags & 0xFF;
    switch (domain_state) {
1970 1971 1972 1973
	case DOMFLAGS_DYING:
	    info->state = VIR_DOMAIN_SHUTDOWN;
	    break;
	case DOMFLAGS_SHUTDOWN:
1974 1975 1976 1977 1978 1979 1980 1981 1982
            /* The domain is shutdown.  Determine the cause. */
            domain_shutdown_cause = domain_flags >> DOMFLAGS_SHUTDOWNSHIFT;
            switch (domain_shutdown_cause) {
                case SHUTDOWN_crash:
                    info->state = VIR_DOMAIN_CRASHED;
                    break;
                default:
                    info->state = VIR_DOMAIN_SHUTOFF;
            }
1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001
	    break;
	case DOMFLAGS_PAUSED:
	    info->state = VIR_DOMAIN_PAUSED;
	    break;
	case DOMFLAGS_BLOCKED:
	    info->state = VIR_DOMAIN_BLOCKED;
	    break;
	case DOMFLAGS_RUNNING:
	    info->state = VIR_DOMAIN_RUNNING;
	    break;
	default:
	    info->state = VIR_DOMAIN_NONE;
    }

    /*
     * the API brings back the cpu time in nanoseconds,
     * convert to microseconds, same thing convert to
     * kilobytes from page counts
     */
2002
    info->cpuTime = XEN_GETDOMAININFO_CPUTIME(dominfo);
2003
    info->memory = XEN_GETDOMAININFO_TOT_PAGES(dominfo) * kb_per_pages;
2004 2005 2006
    info->maxMem = XEN_GETDOMAININFO_MAX_PAGES(dominfo);
    if(info->maxMem != UINT_MAX)
        info->maxMem *= kb_per_pages;
2007
    info->nrVirtCpu = XEN_GETDOMAININFO_CPUCOUNT(dominfo);
2008
    return (0);
2009 2010
}

2011 2012 2013
/**
 * xenHypervisorGetDomainInfo:
 * @domain: pointer to the domain block
2014
 * @info: the place where information should be stored
2015
 *
2016
 * Do an hypervisor call to get the related set of domain information.
2017 2018 2019 2020 2021 2022 2023 2024
 *
 * Returns 0 in case of success, -1 in case of error.
 */
int
xenHypervisorGetDomainInfo(virDomainPtr domain, virDomainInfoPtr info)
{
    if ((domain == NULL) || (domain->conn == NULL) ||
        (domain->conn->handle < 0) || (info == NULL) ||
2025
        (domain->id < 0))
2026
        return (-1);
2027

2028
    return(xenHypervisorGetDomInfo(domain->conn, domain->id, info));
2029 2030 2031

}

2032
#ifndef PROXY
2033 2034
/**
 * xenHypervisorPauseDomain:
2035
 * @domain: pointer to the domain block
2036 2037 2038 2039 2040 2041
 *
 * Do an hypervisor call to pause the given domain
 *
 * Returns 0 in case of success, -1 in case of error.
 */
int
2042
xenHypervisorPauseDomain(virDomainPtr domain)
2043
{
2044 2045
    int ret;

2046
    if ((domain == NULL) || (domain->conn == NULL) ||
2047
        (domain->conn->handle < 0) || (domain->id < 0))
2048 2049
        return (-1);

2050
    ret = virXen_pausedomain(domain->conn->handle, domain->id);
2051
    if (ret < 0)
2052 2053
        return (-1);
    return (0);
2054 2055 2056 2057
}

/**
 * xenHypervisorResumeDomain:
2058
 * @domain: pointer to the domain block
2059 2060 2061 2062 2063 2064
 *
 * Do an hypervisor call to resume the given domain
 *
 * Returns 0 in case of success, -1 in case of error.
 */
int
2065
xenHypervisorResumeDomain(virDomainPtr domain)
2066
{
2067 2068
    int ret;

2069
    if ((domain == NULL) || (domain->conn == NULL) ||
2070
        (domain->conn->handle < 0) || (domain->id < 0))
2071 2072
        return (-1);

2073
    ret = virXen_unpausedomain(domain->conn->handle, domain->id);
2074
    if (ret < 0)
2075 2076
        return (-1);
    return (0);
2077 2078 2079 2080
}

/**
 * xenHypervisorDestroyDomain:
2081
 * @domain: pointer to the domain block
2082 2083 2084 2085 2086 2087
 *
 * Do an hypervisor call to destroy the given domain
 *
 * Returns 0 in case of success, -1 in case of error.
 */
int
2088
xenHypervisorDestroyDomain(virDomainPtr domain)
2089
{
2090 2091
    int ret;

2092
    if ((domain == NULL) || (domain->conn == NULL) ||
2093
        (domain->conn->handle < 0) || (domain->id < 0))
2094 2095
        return (-1);

2096
    ret = virXen_destroydomain(domain->conn->handle, domain->id);
2097
    if (ret < 0)
2098 2099
        return (-1);
    return (0);
2100 2101
}

2102 2103
/**
 * xenHypervisorSetMaxMemory:
2104
 * @domain: pointer to the domain block
2105 2106 2107 2108 2109 2110 2111
 * @memory: the max memory size in kilobytes.
 *
 * Do an hypervisor call to change the maximum amount of memory used
 *
 * Returns 0 in case of success, -1 in case of error.
 */
int
2112
xenHypervisorSetMaxMemory(virDomainPtr domain, unsigned long memory)
2113
{
2114 2115
    int ret;

2116
    if ((domain == NULL) || (domain->conn == NULL) ||
2117
        (domain->conn->handle < 0) || (domain->id < 0))
2118 2119
        return (-1);

2120
    ret = virXen_setmaxmem(domain->conn->handle, domain->id, memory);
2121
    if (ret < 0)
2122 2123
        return (-1);
    return (0);
2124
}
2125
#endif /* PROXY */
2126

2127
#ifndef PROXY
2128 2129 2130 2131 2132 2133 2134 2135 2136 2137 2138 2139 2140
/**
 * xenHypervisorSetVcpus:
 * @domain: pointer to domain object
 * @nvcpus: the new number of virtual CPUs for this domain
 *
 * Dynamically change the number of virtual CPUs used by the domain.
 *
 * Returns 0 in case of success, -1 in case of failure.
 */

int
xenHypervisorSetVcpus(virDomainPtr domain, unsigned int nvcpus)
{
2141
    int ret;
2142

2143
    if ((domain == NULL) || (domain->conn == NULL) ||
2144
        (domain->conn->handle < 0) || (domain->id < 0) ||
2145
        (nvcpus < 1))
2146
        return (-1);
2147

2148
    ret = virXen_setmaxvcpus(domain->conn->handle, domain->id, nvcpus);
2149
    if (ret < 0)
2150
        return (-1);
2151
    return (0);
2152 2153 2154 2155 2156 2157 2158 2159
}

/**
 * xenHypervisorPinVcpu:
 * @domain: pointer to domain object
 * @vcpu: virtual CPU number
 * @cpumap: pointer to a bit map of real CPUs (in 8-bit bytes)
 * @maplen: length of cpumap in bytes
2160
 *
2161 2162 2163 2164 2165 2166 2167 2168 2169
 * Dynamically change the real CPUs which can be allocated to a virtual CPU.
 *
 * Returns 0 in case of success, -1 in case of failure.
 */

int
xenHypervisorPinVcpu(virDomainPtr domain, unsigned int vcpu,
                     unsigned char *cpumap, int maplen)
{
2170
    int ret;
2171

2172
    if ((domain == NULL) || (domain->conn == NULL) ||
2173
        (domain->conn->handle < 0) || (domain->id < 0) ||
2174
        (cpumap == NULL) || (maplen < 1))
2175 2176
        return (-1);

2177
    ret = virXen_setvcpumap(domain->conn->handle, domain->id, vcpu,
2178 2179
                            cpumap, maplen);
    if (ret < 0)
2180
        return (-1);
2181
    return (0);
2182
}
2183
#endif
2184 2185 2186 2187 2188 2189 2190 2191 2192 2193 2194 2195 2196 2197

/**
 * virDomainGetVcpus:
 * @domain: pointer to domain object, or NULL for Domain0
 * @info: pointer to an array of virVcpuInfo structures (OUT)
 * @maxinfo: number of structures in info array
 * @cpumaps: pointer to an bit map of real CPUs for all vcpus of this domain (in 8-bit bytes) (OUT)
 *	If cpumaps is NULL, then no cupmap information is returned by the API.
 *	It's assumed there is <maxinfo> cpumap in cpumaps array.
 *	The memory allocated to cpumaps must be (maxinfo * maplen) bytes
 *	(ie: calloc(maxinfo, maplen)).
 *	One cpumap inside cpumaps has the format described in virDomainPinVcpu() API.
 * @maplen: number of bytes in one cpumap, from 1 up to size of CPU map in
 *	underlying virtualization system (Xen...).
2198
 *
2199 2200 2201 2202 2203
 * Extract information about virtual CPUs of domain, store it in info array
 * and also in cpumaps if this pointer is'nt NULL.
 *
 * Returns the number of info filled in case of success, -1 in case of failure.
 */
2204
#ifndef PROXY
2205 2206
int
xenHypervisorGetVcpus(virDomainPtr domain, virVcpuInfoPtr info, int maxinfo,
2207
                      unsigned char *cpumaps, int maplen)
2208
{
2209
    xen_getdomaininfo dominfo;
2210 2211
    int ret;

2212
    virVcpuInfoPtr ipt;
2213
    int nbinfo, i;
2214

2215
    if ((domain == NULL) || (domain->conn == NULL) ||
2216
        (domain->conn->handle < 0) || (domain->id < 0) ||
2217 2218
        (info == NULL) || (maxinfo < 1) ||
        (sizeof(cpumap_t) & 7))
2219
        return (-1);
2220
    if ((cpumaps != NULL) && (maplen < 1))
2221
        return -1;
2222 2223

    /* first get the number of virtual CPUs in this domain */
2224
    XEN_GETDOMAININFO_CLEAR(dominfo);
2225
    ret = virXen_getdomaininfo(domain->conn->handle, domain->id,
2226
                               &dominfo);
2227

2228
    if ((ret < 0) || (XEN_GETDOMAININFO_DOMAIN(dominfo) != domain->id))
2229
        return (-1);
2230
    nbinfo = XEN_GETDOMAININFO_CPUCOUNT(dominfo) + 1;
2231 2232 2233
    if (nbinfo > maxinfo) nbinfo = maxinfo;

    if (cpumaps != NULL)
2234
        memset(cpumaps, 0, maxinfo * maplen);
2235

2236 2237
    for (i = 0, ipt = info; i < nbinfo; i++, ipt++) {
        if ((cpumaps != NULL) && (i < maxinfo)) {
2238
            ret = virXen_getvcpusinfo(domain->conn->handle, domain->id, i,
2239 2240 2241 2242 2243 2244
                                      ipt,
                                      (unsigned char *)VIR_GET_CPUMAP(cpumaps, maplen, i),
                                      maplen);
            if (ret < 0)
                return(-1);
        } else {
2245
            ret = virXen_getvcpusinfo(domain->conn->handle, domain->id, i,
2246 2247 2248 2249
                                      ipt, NULL, 0);
            if (ret < 0)
                return(-1);
        }
2250 2251 2252
    }
    return nbinfo;
}
2253 2254
#endif

2255 2256 2257 2258 2259 2260 2261 2262 2263 2264 2265 2266 2267 2268 2269 2270 2271 2272 2273 2274 2275 2276 2277 2278 2279 2280 2281 2282 2283 2284 2285 2286 2287 2288 2289
/**
 * xenHypervisorGetVcpuMax:
 *
 *  Returns the maximum number of virtual CPUs supported for
 *  the guest VM. If the guest is inactive, this is the maximum
 *  of CPU defined by Xen. If the guest is running this reflect
 *  the maximum number of virtual CPUs the guest was booted with.
 */
int
xenHypervisorGetVcpuMax(virDomainPtr domain)
{
    xen_getdomaininfo dominfo;
    int ret;
    int maxcpu;

    if ((domain == NULL) || (domain->conn == NULL) ||
        (domain->conn->handle < 0))
        return (-1);

    /* inactive domain */
    if (domain->id < 0) {
        maxcpu = MAX_VIRT_CPUS;
    } else {
        XEN_GETDOMAININFO_CLEAR(dominfo);
        ret = virXen_getdomaininfo(domain->conn->handle, domain->id,
                                   &dominfo);

        if ((ret < 0) || (XEN_GETDOMAININFO_DOMAIN(dominfo) != domain->id))
            return (-1);
        maxcpu = XEN_GETDOMAININFO_MAXCPUID(dominfo) + 1;
    }

    return maxcpu;
}

2290
#endif /* WITH_XEN */
2291 2292 2293 2294 2295
/*
 * vim: set tabstop=4:
 * vim: set shiftwidth=4:
 * vim: set expandtab:
 */
2296 2297 2298 2299 2300 2301 2302 2303
/*
 * Local variables:
 *  indent-tabs-mode: nil
 *  c-indent-level: 4
 *  c-basic-offset: 4
 *  tab-width: 4
 * End:
 */