xen_internal.c 97.0 KB
Newer Older
1 2 3
/*
 * xen_internal.c: direct access to Xen hypervisor level
 *
4
 * Copyright (C) 2005, 2006, 2007, 2008 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
#include <config.h>
14

15 16
#include <stdio.h>
#include <string.h>
17
/* required for uint8_t, uint32_t, etc ... */
18 19 20 21 22 23 24
#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>
25
#include <limits.h>
26
#include <stdint.h>
27 28 29
#include <regex.h>
#include <errno.h>
#include <sys/utsname.h>
30

31
#include "xs_internal.h"
32
#include "stats_linux.h"
33
#include "xend_internal.h"
34 35

/* required for dom0_getdomaininfo_t */
36
#include <xen/dom0_ops.h>
37
#include <xen/version.h>
38
#ifdef HAVE_XEN_LINUX_PRIVCMD_H
39
#include <xen/linux/privcmd.h>
40 41 42 43 44
#else
#ifdef HAVE_XEN_SYS_PRIVCMD_H
#include <xen/sys/privcmd.h>
#endif
#endif
45

46 47 48
/* required for shutdown flags */
#include <xen/sched.h>

49
#include "buf.h"
50
#include "capabilities.h"
51

52 53 54
#define DEBUG(fmt,...) VIR_DEBUG(__FILE__, fmt, __VA_ARGS__)
#define DEBUG0(msg) VIR_DEBUG(__FILE__, "%s", msg)

55
/*
56
 * so far there is 2 versions of the structures usable for doing
57 58 59 60
 * hypervisor calls.
 */
/* the old one */
typedef struct v0_hypercall_struct {
61 62
    unsigned long op;
    unsigned long arg[5];
63
} v0_hypercall_t;
64 65

#ifdef __linux__
66 67 68 69 70 71 72 73
#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;
74 75
#define XEN_V1_IOCTL_HYPERCALL_CMD                  \
    _IOC(_IOC_NONE, 'P', 0, sizeof(v1_hypercall_t))
76
typedef v1_hypercall_t hypercall_t;
77 78 79 80 81
#elif define(__sun__)
typedef privcmd_hypercall_t hypercall_t;
#else
#error "unsupported platform"
#endif
82 83 84 85 86 87 88

#ifndef __HYPERVISOR_sysctl
#define __HYPERVISOR_sysctl 35
#endif
#ifndef __HYPERVISOR_domctl
#define __HYPERVISOR_domctl 36
#endif
89

90 91
static int xen_ioctl_hypercall_cmd = 0;
static int initialized = 0;
92
static int in_init = 0;
93
static int hv_version = 0;
94 95 96
static int hypervisor_version = 2;
static int sys_interface_version = -1;
static int dom_interface_version = -1;
97
static int kb_per_pages = 0;
98

99 100 101 102 103 104 105 106
/* 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;
107
static const char *xen_cap_re = "(xen|hvm)-[[:digit:]]+\\.[[:digit:]]+-(x86_32|x86_64|ia64|powerpc64)(p|be)?";
108 109
static regex_t xen_cap_rec;

110 111 112 113 114
/*
 * The content of the structures for a getdomaininfolist system hypercall
 */
#ifndef DOMFLAGS_DYING
#define DOMFLAGS_DYING     (1<<0) /* Domain is scheduled to die.             */
115
#define DOMFLAGS_HVM       (1<<1) /* Domain is HVM                           */
116 117 118 119 120 121 122 123 124 125
#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

126 127 128 129 130 131 132 133 134 135 136
/*
 * 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

137 138 139 140 141 142
#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 */
R
Richard W.M. Jones 已提交
143
    uint32_t flags;	/* flags, see before */
144 145
    uint64_t tot_pages;	/* total number of pages used */
    uint64_t max_pages;	/* maximum number of pages allowed */
146
    unsigned long shared_info_frame; /* MFN of shared_info struct */
147 148 149 150 151 152 153 154
    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;

155 156
struct xen_v2_getdomaininfo {
    domid_t  domain;	/* the domain number */
R
Richard W.M. Jones 已提交
157
    uint32_t flags;	/* flags, see before */
158 159 160 161 162 163 164 165 166 167 168
    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;

169 170 171 172 173 174 175

/* As of Hypervisor Call v2,  DomCtl v5 we are now 8-byte aligned
   even on 32-bit archs when dealing with uint64_t */
#define ALIGN_64 __attribute__((aligned(8)))

struct xen_v2d5_getdomaininfo {
    domid_t  domain;	/* the domain number */
R
Richard W.M. Jones 已提交
176
    uint32_t flags;	/* flags, see before */
177 178 179 180 181 182 183 184 185 186 187
    uint64_t tot_pages ALIGN_64;	/* total number of pages used */
    uint64_t max_pages ALIGN_64;	/* maximum number of pages allowed */
    uint64_t shared_info_frame ALIGN_64; /* MFN of shared_info struct */
    uint64_t cpu_time ALIGN_64;  /* 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_v2d5_getdomaininfo xen_v2d5_getdomaininfo;

188 189 190
union xen_getdomaininfo {
    struct xen_v0_getdomaininfo v0;
    struct xen_v2_getdomaininfo v2;
191
    struct xen_v2d5_getdomaininfo v2d5;
192 193 194 195 196 197
};
typedef union xen_getdomaininfo xen_getdomaininfo;

union xen_getdomaininfolist {
    struct xen_v0_getdomaininfo *v0;
    struct xen_v2_getdomaininfo *v2;
198
    struct xen_v2d5_getdomaininfo *v2d5;
199 200 201
};
typedef union xen_getdomaininfolist xen_getdomaininfolist;

202 203 204 205 206 207 208 209 210 211 212 213

struct xen_v2_getschedulerid {
    uint32_t sched_id; /* Get Scheduler ID from Xen */
};
typedef struct xen_v2_getschedulerid xen_v2_getschedulerid;


union xen_getschedulerid {
    struct xen_v2_getschedulerid *v2;
};
typedef union xen_getschedulerid xen_getschedulerid;

214 215 216 217 218 219 220 221 222
struct xen_v2s4_availheap {
    uint32_t min_bitwidth;  /* Smallest address width (zero if don't care). */
    uint32_t max_bitwidth;  /* Largest address width (zero if don't care). */
    int32_t  node;          /* NUMA node (-1 for sum across all nodes). */
    uint64_t avail_bytes;   /* Bytes available in the specified region. */
};

typedef struct xen_v2s4_availheap  xen_v2s4_availheap;

223

224 225
#define XEN_GETDOMAININFOLIST_ALLOC(domlist, size)                      \
    (hypervisor_version < 2 ?                                           \
226
     ((domlist.v0 = malloc(sizeof(*domlist.v0)*(size))) != NULL) :      \
227
     (dom_interface_version < 5 ?                                       \
228 229
      ((domlist.v2 = malloc(sizeof(*domlist.v2)*(size))) != NULL) :     \
      ((domlist.v2d5 = malloc(sizeof(*domlist.v2d5)*(size))) != NULL)))
230 231 232 233 234 235 236 237

#define XEN_GETDOMAININFOLIST_FREE(domlist)        \
    (hypervisor_version < 2 ?                      \
     free(domlist.v0) :                            \
     (dom_interface_version < 5 ?                  \
      free(domlist.v2) :                           \
      free(domlist.v2d5)))

238 239 240 241 242 243
#define XEN_GETDOMAININFOLIST_CLEAR(domlist, size)            \
    (hypervisor_version < 2 ?                                 \
     memset(domlist.v0, 0, sizeof(*domlist.v0) * size) :      \
     (dom_interface_version < 5 ?                             \
      memset(domlist.v2, 0, sizeof(*domlist.v2) * size) :     \
      memset(domlist.v2d5, 0, sizeof(*domlist.v2d5) * size)))
244 245 246 247

#define XEN_GETDOMAININFOLIST_DOMAIN(domlist, n)    \
    (hypervisor_version < 2 ?                       \
     domlist.v0[n].domain :                         \
248 249 250 251
     (dom_interface_version < 5 ?                   \
      domlist.v2[n].domain :                        \
      domlist.v2d5[n].domain))

252 253 254 255 256 257 258
#define XEN_GETDOMAININFOLIST_UUID(domlist, n)      \
    (hypervisor_version < 2 ?                       \
     domlist.v0[n].handle :                         \
     (dom_interface_version < 5 ?                   \
      domlist.v2[n].handle :                        \
      domlist.v2d5[n].handle))

259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278
#define XEN_GETDOMAININFOLIST_DATA(domlist)        \
    (hypervisor_version < 2 ?                      \
     (void*)(domlist->v0) :                        \
     (dom_interface_version < 5 ?                  \
      (void*)(domlist->v2) :                       \
      (void*)(domlist->v2d5)))

#define XEN_GETDOMAININFO_SIZE                     \
    (hypervisor_version < 2 ?                      \
     sizeof(xen_v0_getdomaininfo) :                \
     (dom_interface_version < 5 ?                  \
      sizeof(xen_v2_getdomaininfo) :               \
      sizeof(xen_v2d5_getdomaininfo)))

#define XEN_GETDOMAININFO_CLEAR(dominfo)                           \
    (hypervisor_version < 2 ?                                      \
     memset(&(dominfo.v0), 0, sizeof(xen_v0_getdomaininfo)) :      \
     (dom_interface_version < 5 ?                                  \
      memset(&(dominfo.v2), 0, sizeof(xen_v2_getdomaininfo)) :     \
      memset(&(dominfo.v2d5), 0, sizeof(xen_v2d5_getdomaininfo))))
279 280 281 282

#define XEN_GETDOMAININFO_DOMAIN(dominfo)       \
    (hypervisor_version < 2 ?                   \
     dominfo.v0.domain :                        \
283 284 285
     (dom_interface_version < 5 ?               \
      dominfo.v2.domain :                       \
      dominfo.v2d5.domain))
286 287 288 289

#define XEN_GETDOMAININFO_CPUTIME(dominfo)      \
    (hypervisor_version < 2 ?                   \
     dominfo.v0.cpu_time :                      \
290 291 292
     (dom_interface_version < 5 ?               \
      dominfo.v2.cpu_time :                     \
      dominfo.v2d5.cpu_time))
293 294 295 296

#define XEN_GETDOMAININFO_CPUCOUNT(dominfo)     \
    (hypervisor_version < 2 ?                   \
     dominfo.v0.nr_online_vcpus :               \
297 298 299
     (dom_interface_version < 5 ?               \
      dominfo.v2.nr_online_vcpus :              \
      dominfo.v2d5.nr_online_vcpus))
300

301 302 303
#define XEN_GETDOMAININFO_MAXCPUID(dominfo)  \
    (hypervisor_version < 2 ?                   \
     dominfo.v0.max_vcpu_id :                   \
304 305 306
     (dom_interface_version < 5 ?               \
      dominfo.v2.max_vcpu_id :                  \
      dominfo.v2d5.max_vcpu_id))
307

308 309 310
#define XEN_GETDOMAININFO_FLAGS(dominfo)        \
    (hypervisor_version < 2 ?                   \
     dominfo.v0.flags :                         \
311 312 313
     (dom_interface_version < 5 ?               \
      dominfo.v2.flags :                        \
      dominfo.v2d5.flags))
314 315 316 317

#define XEN_GETDOMAININFO_TOT_PAGES(dominfo)    \
    (hypervisor_version < 2 ?                   \
     dominfo.v0.tot_pages :                     \
318 319 320
     (dom_interface_version < 5 ?               \
      dominfo.v2.tot_pages :                    \
      dominfo.v2d5.tot_pages))
321 322 323 324

#define XEN_GETDOMAININFO_MAX_PAGES(dominfo)    \
    (hypervisor_version < 2 ?                   \
     dominfo.v0.max_pages :                     \
325 326 327
     (dom_interface_version < 5 ?               \
      dominfo.v2.max_pages :                    \
      dominfo.v2d5.max_pages))
328

329 330 331 332 333 334 335
#define XEN_GETDOMAININFO_UUID(dominfo)         \
    (hypervisor_version < 2 ?                   \
     dominfo.v0.handle :                        \
     (dom_interface_version < 5 ?               \
      dominfo.v2.handle :                       \
      dominfo.v2d5.handle))

336

337 338 339 340
static int
lock_pages(void *addr, size_t len)
{
#ifdef __linux__
341
        return (mlock(addr, len));
342
#elif define(__sun)
343
        return (0);
344 345 346 347 348 349 350
#endif
}

static int
unlock_pages(void *addr, size_t len)
{
#ifdef __linux__
351
        return (munlock(addr, len));
352
#elif define(__sun)
353
        return (0);
354 355 356
#endif
}

357 358

struct xen_v0_getdomaininfolistop {
359 360 361 362 363
    domid_t   first_domain;
    uint32_t  max_domains;
    struct xen_v0_getdomaininfo *buffer;
    uint32_t  num_domains;
};
364 365 366 367 368 369 370 371 372 373 374
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;

375 376 377 378
/* As of HV version 2, sysctl version 3 the *buffer pointer is 64-bit aligned */
struct xen_v2s3_getdomaininfolistop {
    domid_t   first_domain;
    uint32_t  max_domains;
379 380 381 382 383 384
#ifdef __BIG_ENDIAN__
    struct {
        int __pad[(sizeof (long long) - sizeof (struct xen_v2d5_getdomaininfo *)) / sizeof (int)];
        struct xen_v2d5_getdomaininfo *v;
    } buffer;
#else
385 386 387 388
    union {
        struct xen_v2d5_getdomaininfo *v;
        uint64_t pad ALIGN_64;
    } buffer;
389
#endif
390 391 392 393
    uint32_t  num_domains;
};
typedef struct xen_v2s3_getdomaininfolistop xen_v2s3_getdomaininfolistop;

394

395 396 397 398 399 400 401

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

/*
402
 * The information for a destroydomain system hypercall
403 404 405 406 407 408
 */
#define XEN_V0_OP_DESTROYDOMAIN	9
#define XEN_V1_OP_DESTROYDOMAIN	9
#define XEN_V2_OP_DESTROYDOMAIN	2

/*
409
 * The information for a pausedomain system hypercall
410 411 412 413 414 415
 */
#define XEN_V0_OP_PAUSEDOMAIN	10
#define XEN_V1_OP_PAUSEDOMAIN	10
#define XEN_V2_OP_PAUSEDOMAIN	3

/*
416
 * The information for an unpausedomain system hypercall
417 418 419 420 421 422
 */
#define XEN_V0_OP_UNPAUSEDOMAIN	11
#define XEN_V1_OP_UNPAUSEDOMAIN	11
#define XEN_V2_OP_UNPAUSEDOMAIN	4

/*
423
 * The information for an setmaxmem system hypercall
424 425 426
 */
#define XEN_V0_OP_SETMAXMEM	28
#define XEN_V1_OP_SETMAXMEM	28
427
#define XEN_V2_OP_SETMAXMEM	11
428 429 430 431 432 433 434 435 436 437 438 439 440

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;

441 442 443 444 445
struct xen_v2d5_setmaxmem {
    uint64_t	maxmem ALIGN_64;
};
typedef struct xen_v2d5_setmaxmem xen_v2d5_setmaxmem;

446
/*
447
 * The information for an setmaxvcpu system hypercall
448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465
 */
#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;

/*
466
 * The information for an setvcpumap system hypercall
467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491
 * 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;

492 493
/* HV version 2, Dom version 5 requires 64-bit alignment */
struct xen_v2d5_cpumap {
494 495 496 497 498 499
#ifdef __BIG_ENDIAN__
    struct {
        int __pad[(sizeof (long long) - sizeof (uint8_t *)) / sizeof (int)];
        uint8_t *v;
    } bitmap;
#else
500 501 502 503
    union {
        uint8_t    *v;
        uint64_t   pad ALIGN_64;
    } bitmap;
504
#endif
505 506 507 508 509 510 511 512
    uint32_t    nr_cpus;
};
struct xen_v2d5_setvcpumap {
    uint32_t	vcpu;
    struct xen_v2d5_cpumap cpumap;
};
typedef struct xen_v2d5_setvcpumap xen_v2d5_setvcpumap;

513
/*
514
 * The information for an vcpuinfo system hypercall
515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542
 */
#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;

543 544 545 546 547 548 549 550 551 552
struct xen_v2d5_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 ALIGN_64; /* nanosecond of CPU used */
    uint32_t	cpu;		/* current mapping */
};
typedef struct xen_v2d5_vcpuinfo xen_v2d5_vcpuinfo;

553 554 555 556 557
/*
 * 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;
558
typedef struct xen_v2d5_setvcpumap xen_v2d5_getvcpumap;
559

560 561 562 563 564
/*
 * from V2 we get the scheduler information
 */
#define XEN_V2_OP_GETSCHEDULERID	4

565 566 567 568 569
/*
 * from V2 we get the available heap information
 */
#define XEN_V2_OP_GETAVAILHEAP  	9

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
/*
 * from V2 we get the scheduler parameter
 */
#define XEN_V2_OP_SCHEDULER		16
/* Scheduler types. */
#define XEN_SCHEDULER_SEDF       4
#define XEN_SCHEDULER_CREDIT     5
/* get/set scheduler parameters */
#define XEN_DOMCTL_SCHEDOP_putinfo 0
#define XEN_DOMCTL_SCHEDOP_getinfo 1

struct xen_v2_setschedinfo {
    uint32_t sched_id;
    uint32_t cmd;
    union {
        struct xen_domctl_sched_sedf {
            uint64_t period ALIGN_64;
            uint64_t slice  ALIGN_64;
            uint64_t latency ALIGN_64;
            uint32_t extratime;
            uint32_t weight;
        } sedf;
        struct xen_domctl_sched_credit {
            uint16_t weight;
            uint16_t cap;
        } credit;
    } u;
};
typedef struct xen_v2_setschedinfo xen_v2_setschedinfo;
typedef struct xen_v2_setschedinfo xen_v2_getschedinfo;


602 603 604 605 606 607 608 609 610
/*
 * 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 {
611 612 613 614 615 616 617
        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];
618 619 620 621 622 623 624 625 626 627
    } 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 {
628 629
        xen_v2_getdomaininfolistop   getdomaininfolist;
        xen_v2s3_getdomaininfolistop getdomaininfolists3;
630
        xen_v2_getschedulerid        getschedulerid;
631
        xen_v2s4_availheap           availheap;
632
        uint8_t padding[128];
633 634 635 636 637 638 639 640 641 642
    } 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 {
643
        xen_v2_setmaxmem         setmaxmem;
644
        xen_v2d5_setmaxmem       setmaxmemd5;
645 646
        xen_v2_setmaxvcpu        setmaxvcpu;
        xen_v2_setvcpumap        setvcpumap;
647
        xen_v2d5_setvcpumap      setvcpumapd5;
648
        xen_v2_vcpuinfo          getvcpuinfo;
649
        xen_v2d5_vcpuinfo        getvcpuinfod5;
650
        xen_v2_getvcpumap        getvcpumap;
651
        xen_v2d5_getvcpumap      getvcpumapd5;
652 653
        xen_v2_setschedinfo      setschedinfo;
        xen_v2_getschedinfo      getschedinfo;
654
        uint8_t padding[128];
655 656 657
    } u;
};
typedef struct xen_op_v2_dom xen_op_v2_dom;
658 659

#include "internal.h"
660
#include "driver.h"
661
#include "xen_unified.h"
662 663
#include "xen_internal.h"

664 665 666 667 668 669 670 671 672 673 674
#ifdef __linux__
#define XEN_HYPERVISOR_SOCKET	"/proc/xen/privcmd"
#define HYPERVISOR_CAPABILITIES	"/sys/hypervisor/properties/capabilities"
#define CPUINFO			"/proc/cpuinfo"
#elif define(__sun__)
#define XEN_HYPERVISOR_SOCKET	"/dev/xen/privcmd"
#define HYPERVISOR_CAPABILITIES	""
#define CPUINFO			"/dev/cpu/self/cpuid"
#else
#error "unsupported platform"
#endif
675

676
#ifndef PROXY
677
static unsigned long xenHypervisorGetMaxMemory(virDomainPtr domain);
678
#endif
679

680
#ifndef PROXY
681
struct xenUnifiedDriver xenHypervisorDriver = {
682 683 684
    xenHypervisorOpen, /* open */
    xenHypervisorClose, /* close */
    xenHypervisorGetVersion, /* version */
685 686
    NULL, /* hostname */
    NULL, /* URI */
687
    NULL, /* nodeGetInfo */
688
    xenHypervisorGetCapabilities, /* getCapabilities */
689 690
    xenHypervisorListDomains, /* listDomains */
    xenHypervisorNumOfDomains, /* numOfDomains */
691 692 693 694
    NULL, /* domainCreateLinux */
    xenHypervisorPauseDomain, /* domainSuspend */
    xenHypervisorResumeDomain, /* domainResume */
    NULL, /* domainShutdown */
695
    NULL, /* domainReboot */
696
    xenHypervisorDestroyDomain, /* domainDestroy */
697
    xenHypervisorDomainGetOSType, /* domainGetOSType */
698
    xenHypervisorGetMaxMemory, /* domainGetMaxMemory */
699
    xenHypervisorSetMaxMemory, /* domainSetMaxMemory */
700
    NULL, /* domainSetMemory */
701 702
    xenHypervisorGetDomainInfo, /* domainGetInfo */
    NULL, /* domainSave */
703
    NULL, /* domainRestore */
D
Daniel Veillard 已提交
704
    NULL, /* domainCoreDump */
705 706
    xenHypervisorSetVcpus, /* domainSetVcpus */
    xenHypervisorPinVcpu, /* domainPinVcpu */
707
    xenHypervisorGetVcpus, /* domainGetVcpus */
708
    xenHypervisorGetVcpuMax, /* domainGetMaxVcpus */
709 710 711 712 713
    NULL, /* listDefinedDomains */
    NULL, /* numOfDefinedDomains */
    NULL, /* domainCreate */
    NULL, /* domainDefineXML */
    NULL, /* domainUndefine */
714 715
    NULL, /* domainAttachDevice */
    NULL, /* domainDetachDevice */
716 717
    NULL, /* domainGetAutostart */
    NULL, /* domainSetAutostart */
718 719 720
    xenHypervisorGetSchedulerType, /* domainGetSchedulerType */
    xenHypervisorGetSchedulerParameters, /* domainGetSchedulerParameters */
    xenHypervisorSetSchedulerParameters, /* domainSetSchedulerParameters */
721
};
722
#endif /* !PROXY */
723

724 725
/**
 * virXenError:
726
 * @conn: connection, if known
727 728
 * @error: the error number
 * @info: extra information string
729
 * @value: extra information number
730 731 732 733
 *
 * Handle an error at the xend daemon interface
 */
static void
734 735
virXenError(virConnectPtr conn,
            virErrorNumber error, const char *info, int value)
736
{
737
    const char *errmsg;
738

739
    if ((error == VIR_ERR_OK) || (in_init != 0))
740 741 742
        return;

    errmsg = __virErrorMsg(error, info);
743
    __virRaiseError(conn, NULL, NULL, VIR_FROM_XEN, error, VIR_ERR_ERROR,
744
                    errmsg, info, NULL, value, 0, errmsg, info, value);
745 746
}

747 748
#ifndef PROXY

749 750
/**
 * virXenErrorFunc:
751
 * @conn: connection, if known
752 753 754 755 756 757 758 759
 * @error: the error number
 * @func: the function failing
 * @info: extra information string
 * @value: extra information number
 *
 * Handle an error at the xend daemon interface
 */
static void
760 761
virXenErrorFunc(virConnectPtr conn,
                virErrorNumber error, const char *func, const char *info,
762 763
                int value)
{
D
Daniel Veillard 已提交
764
    char fullinfo[1000];
765 766 767 768 769 770 771 772 773
    const char *errmsg;

    if ((error == VIR_ERR_OK) || (in_init != 0))
        return;


    errmsg = __virErrorMsg(error, info);
    if (func != NULL) {
        snprintf(fullinfo, 999, "%s: %s", func, info);
774 775 776 777
        fullinfo[999] = 0;
        __virRaiseError(conn, NULL, NULL, VIR_FROM_XEN, error, VIR_ERR_ERROR,
                        errmsg, fullinfo, NULL, value, 0, errmsg, fullinfo,
                        value);
778
    } else {
779 780 781
        __virRaiseError(conn, NULL, NULL, VIR_FROM_XEN, error, VIR_ERR_ERROR,
                        errmsg, info, NULL, value, 0, errmsg, info,
                        value);
782 783 784
    }
}

785 786
#endif /* PROXY */

787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808
/**
 * 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));
809 810
}

811 812 813
/**
 * xenHypervisorDoV0Op:
 * @handle: the handle to the Xen hypervisor
R
Richard W.M. Jones 已提交
814
 * @op: pointer to the hypervisor operation structure
815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831
 *
 * 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;

832
    if (lock_pages(op, sizeof(dom0_op_t)) < 0) {
833
        virXenError(NULL, VIR_ERR_XEN_CALL, " locking", sizeof(*op));
834 835 836 837 838
        return (-1);
    }

    ret = ioctl(handle, xen_ioctl_hypercall_cmd, (unsigned long) &hc);
    if (ret < 0) {
839
        virXenError(NULL, VIR_ERR_XEN_CALL, " ioctl ", xen_ioctl_hypercall_cmd);
840 841
    }

842
    if (unlock_pages(op, sizeof(dom0_op_t)) < 0) {
843
        virXenError(NULL, VIR_ERR_XEN_CALL, " releasing", sizeof(*op));
844 845 846 847 848 849 850 851 852 853 854
        ret = -1;
    }

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

    return (0);
}
/**
 * xenHypervisorDoV1Op:
 * @handle: the handle to the Xen hypervisor
R
Richard W.M. Jones 已提交
855
 * @op: pointer to the hypervisor operation structure
856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872
 *
 * 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;

873
    if (lock_pages(op, sizeof(dom0_op_t)) < 0) {
874
        virXenError(NULL, VIR_ERR_XEN_CALL, " locking", sizeof(*op));
875 876 877 878 879
        return (-1);
    }

    ret = ioctl(handle, xen_ioctl_hypercall_cmd, (unsigned long) &hc);
    if (ret < 0) {
880
        virXenError(NULL, VIR_ERR_XEN_CALL, " ioctl ", xen_ioctl_hypercall_cmd);
881 882
    }

883
    if (unlock_pages(op, sizeof(dom0_op_t)) < 0) {
884
        virXenError(NULL, VIR_ERR_XEN_CALL, " releasing", sizeof(*op));
885 886 887 888 889 890 891 892 893 894 895 896 897 898
        ret = -1;
    }

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

    return (0);
}

/**
 * xenHypervisorDoV2Sys:
 * @handle: the handle to the Xen hypervisor
 * @op: pointer to the hypervisor operation structure
 *
R
Richard W.M. Jones 已提交
899
 * Do an hypervisor v2 system operation, this leads to an hypervisor
900 901 902 903 904 905 906 907 908 909 910 911 912 913 914
 * 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;

915
    if (lock_pages(op, sizeof(dom0_op_t)) < 0) {
916
        virXenError(NULL, VIR_ERR_XEN_CALL, " locking", sizeof(*op));
917 918 919 920 921
        return (-1);
    }

    ret = ioctl(handle, xen_ioctl_hypercall_cmd, (unsigned long) &hc);
    if (ret < 0) {
922
        virXenError(NULL, VIR_ERR_XEN_CALL, " sys ioctl ", xen_ioctl_hypercall_cmd);
923 924
    }

925
    if (unlock_pages(op, sizeof(dom0_op_t)) < 0) {
926
        virXenError(NULL, VIR_ERR_XEN_CALL, " releasing", sizeof(*op));
927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956
        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;

957
    if (lock_pages(op, sizeof(dom0_op_t)) < 0) {
958
        virXenError(NULL, VIR_ERR_XEN_CALL, " locking", sizeof(*op));
959 960 961 962 963
        return (-1);
    }

    ret = ioctl(handle, xen_ioctl_hypercall_cmd, (unsigned long) &hc);
    if (ret < 0) {
964
        virXenError(NULL, VIR_ERR_XEN_CALL, " ioctl ", xen_ioctl_hypercall_cmd);
965 966
    }

967
    if (unlock_pages(op, sizeof(dom0_op_t)) < 0) {
968
        virXenError(NULL, VIR_ERR_XEN_CALL, " releasing", sizeof(*op));
969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984
        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
 *
985
 * Do a low level hypercall to list existing domains information
986 987 988 989 990
 *
 * Returns the number of domains or -1 in case of failure
 */
static int
virXen_getdomaininfolist(int handle, int first_domain, int maxids,
991
                         xen_getdomaininfolist *dominfos)
992 993 994
{
    int ret = -1;

995
    if (lock_pages(XEN_GETDOMAININFOLIST_DATA(dominfos),
996
              XEN_GETDOMAININFO_SIZE * maxids) < 0) {
997
        virXenError(NULL, VIR_ERR_XEN_CALL, " locking",
998
                    XEN_GETDOMAININFO_SIZE * maxids);
999 1000 1001 1002 1003 1004
        return (-1);
    }
    if (hypervisor_version > 1) {
        xen_op_v2_sys op;

        memset(&op, 0, sizeof(op));
1005
        op.cmd = XEN_V2_OP_GETDOMAININFOLIST;
1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017

        if (sys_interface_version < 3) {
            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;
        } else {
            op.u.getdomaininfolists3.first_domain = (domid_t) first_domain;
            op.u.getdomaininfolists3.max_domains = maxids;
            op.u.getdomaininfolists3.buffer.v = dominfos->v2d5;
            op.u.getdomaininfolists3.num_domains = maxids;
        }
1018
        ret = xenHypervisorDoV2Sys(handle, &op);
1019 1020 1021 1022 1023 1024 1025

        if (ret == 0) {
            if (sys_interface_version < 3)
                ret = op.u.getdomaininfolist.num_domains;
            else
                ret = op.u.getdomaininfolists3.num_domains;
        }
1026 1027 1028 1029
    } else if (hypervisor_version == 1) {
        xen_op_v1 op;

        memset(&op, 0, sizeof(op));
1030 1031 1032 1033 1034 1035 1036 1037
        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;
1038 1039 1040 1041
    } else if (hypervisor_version == 0) {
        xen_op_v0 op;

        memset(&op, 0, sizeof(op));
1042 1043 1044 1045 1046 1047 1048 1049
        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;
1050
    }
1051
    if (unlock_pages(XEN_GETDOMAININFOLIST_DATA(dominfos),
1052
                XEN_GETDOMAININFO_SIZE * maxids) < 0) {
1053
        virXenError(NULL, VIR_ERR_XEN_CALL, " release",
1054
                    XEN_GETDOMAININFO_SIZE * maxids);
1055 1056 1057 1058 1059
        ret = -1;
    }
    return(ret);
}

1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074
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);
}


1075
#ifndef PROXY
1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091
/**
 * xenHypervisorGetSchedulerType:
 * @domain: pointer to the Xen Hypervisor block
 * @nparams:give a number of scheduler parameters.
 *
 * Do a low level hypercall to get scheduler type
 *
 * Returns scheduler name or NULL in case of failure
 */
char *
xenHypervisorGetSchedulerType(virDomainPtr domain, int *nparams)
{
    char *schedulertype = NULL;
    xenUnifiedPrivatePtr priv;

    if ((domain == NULL) || (domain->conn == NULL)) {
1092
        virXenErrorFunc(NULL, VIR_ERR_INTERNAL_ERROR, __FUNCTION__,
1093
                        "domain or conn is NULL", 0);
1094 1095 1096 1097 1098
        return NULL;
    }

    priv = (xenUnifiedPrivatePtr) domain->conn->privateData;
    if (priv->handle < 0 || domain->id < 0) {
1099
        virXenErrorFunc(domain->conn, VIR_ERR_INTERNAL_ERROR, __FUNCTION__,
1100
                        "priv->handle or domain->id invalid", 0);
1101 1102 1103 1104 1105 1106
        return NULL;
    }

    /*
     * Support only dom_interface_version >=5
     * (Xen3.1.0 or later)
1107
     * TODO: check on Xen 3.0.3
1108 1109
     */
    if (dom_interface_version < 5) {
1110
        virXenErrorFunc(domain->conn, VIR_ERR_NO_XEN, __FUNCTION__,
1111
                        "unsupported in dom interface < 5", 0);
1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122
        return NULL;
    }

    if (hypervisor_version > 1) {
        xen_op_v2_sys op;
        int ret;

        memset(&op, 0, sizeof(op));
        op.cmd = XEN_V2_OP_GETSCHEDULERID;
        ret = xenHypervisorDoV2Sys(priv->handle, &op);
        if (ret < 0)
1123
            return(NULL);
1124 1125

        switch (op.u.getschedulerid.sched_id){
1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137
            case XEN_SCHEDULER_SEDF:
                schedulertype = strdup("sedf");
                if (nparams)
                    *nparams = 6;
                break;
            case XEN_SCHEDULER_CREDIT:
                schedulertype = strdup("credit");
                if (nparams)
                    *nparams = 2;
                break;
            default:
                break;
1138 1139 1140 1141 1142 1143
        }
    }

    return schedulertype;
}

1144 1145 1146
static const char *str_weight = "weight";
static const char *str_cap = "cap";

1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161
/**
 * xenHypervisorGetSchedulerParameters:
 * @domain: pointer to the Xen Hypervisor block
 * @params: pointer to scheduler parameters.
 *     This memory area should be allocated before calling.
 * @nparams:this parameter should be same as
 *     a given number of scheduler parameters.
 *     from xenHypervisorGetSchedulerType().
 *
 * Do a low level hypercall to get scheduler parameters
 *
 * Returns 0 or -1 in case of failure
 */
int
xenHypervisorGetSchedulerParameters(virDomainPtr domain,
1162
                                 virSchedParameterPtr params, int *nparams)
1163 1164 1165 1166
{
    xenUnifiedPrivatePtr priv;

    if ((domain == NULL) || (domain->conn == NULL)) {
1167
        virXenErrorFunc(NULL, VIR_ERR_INTERNAL_ERROR, __FUNCTION__,
1168
                        "domain or conn is NULL", 0);
1169 1170 1171 1172 1173
        return -1;
    }

    priv = (xenUnifiedPrivatePtr) domain->conn->privateData;
    if (priv->handle < 0 || domain->id < 0) {
1174
        virXenErrorFunc(domain->conn, VIR_ERR_INTERNAL_ERROR, __FUNCTION__,
1175
                        "priv->handle or domain->id invalid", 0);
1176 1177 1178 1179 1180 1181 1182 1183 1184
        return -1;
    }

    /*
     * Support only dom_interface_version >=5
     * (Xen3.1.0 or later)
     * TODO: check on Xen 3.0.3
     */
    if (dom_interface_version < 5) {
1185
        virXenErrorFunc(domain->conn, VIR_ERR_NO_XEN, __FUNCTION__,
1186
                        "unsupported in dom interface < 5", 0);
1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198
        return -1;
    }

    if (hypervisor_version > 1) {
        xen_op_v2_sys op_sys;
        xen_op_v2_dom op_dom;
        int ret;

        memset(&op_sys, 0, sizeof(op_sys));
        op_sys.cmd = XEN_V2_OP_GETSCHEDULERID;
        ret = xenHypervisorDoV2Sys(priv->handle, &op_sys);
        if (ret < 0)
1199
            return -1;
1200 1201

        switch (op_sys.u.getschedulerid.sched_id){
1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218
            case XEN_SCHEDULER_SEDF:
                /* TODO: Implement for Xen/SEDF */
                TODO
                return(-1);
            case XEN_SCHEDULER_CREDIT:
                if (*nparams < 2)
                    return(-1);
                memset(&op_dom, 0, sizeof(op_dom));
                op_dom.cmd = XEN_V2_OP_SCHEDULER;
                op_dom.domain = (domid_t) domain->id;
                op_dom.u.getschedinfo.sched_id = XEN_SCHEDULER_CREDIT;
                op_dom.u.getschedinfo.cmd = XEN_DOMCTL_SCHEDOP_getinfo;
                ret = xenHypervisorDoV2Dom(priv->handle, &op_dom);
                if (ret < 0)
                    return(-1);

                strncpy (params[0].field, str_weight, VIR_DOMAIN_SCHED_FIELD_LENGTH);
1219
        params[0].field[VIR_DOMAIN_SCHED_FIELD_LENGTH-1] = '\0';
1220 1221
                params[0].type = VIR_DOMAIN_SCHED_FIELD_UINT;
                params[0].value.ui = op_dom.u.getschedinfo.u.credit.weight;
1222

1223
                strncpy (params[1].field, str_cap, VIR_DOMAIN_SCHED_FIELD_LENGTH);
1224
        params[1].field[VIR_DOMAIN_SCHED_FIELD_LENGTH-1] = '\0';
1225 1226
                params[1].type = VIR_DOMAIN_SCHED_FIELD_UINT;
                params[1].value.ui = op_dom.u.getschedinfo.u.credit.cap;
1227

1228 1229 1230
                *nparams = 2;
                break;
            default:
1231
        virXenErrorFunc(domain->conn, VIR_ERR_INVALID_ARG, __FUNCTION__,
1232 1233
                        "Unknown scheduler", op_sys.u.getschedulerid.sched_id);
                return -1;
1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250
        }
    }

    return 0;
}

/**
 * xenHypervisorSetSchedulerParameters:
 * @domain: pointer to the Xen Hypervisor block
 * @nparams:give a number of scheduler setting parameters .
 *
 * Do a low level hypercall to set scheduler parameters
 *
 * Returns 0 or -1 in case of failure
 */
int
xenHypervisorSetSchedulerParameters(virDomainPtr domain,
1251
                                 virSchedParameterPtr params, int nparams)
1252 1253
{
    int i;
1254
    unsigned int val;
1255
    xenUnifiedPrivatePtr priv;
1256
    char buf[256];
1257 1258

    if ((domain == NULL) || (domain->conn == NULL)) {
1259
        virXenErrorFunc (NULL, VIR_ERR_INTERNAL_ERROR, __FUNCTION__,
1260
                         "domain or conn is NULL", 0);
1261 1262 1263 1264
        return -1;
    }

    if ((nparams == 0) || (params == NULL)) {
1265
        virXenErrorFunc (domain->conn, VIR_ERR_INVALID_ARG, __FUNCTION__,
1266 1267
                         "Noparameters given", 0);
        return(-1);
1268 1269 1270 1271
    }

    priv = (xenUnifiedPrivatePtr) domain->conn->privateData;
    if (priv->handle < 0 || domain->id < 0) {
1272
        virXenErrorFunc (domain->conn, VIR_ERR_INTERNAL_ERROR, __FUNCTION__,
1273
                         "priv->handle or domain->id invalid", 0);
1274 1275 1276 1277 1278 1279 1280 1281 1282
        return -1;
    }

    /*
     * Support only dom_interface_version >=5
     * (Xen3.1.0 or later)
     * TODO: check on Xen 3.0.3
     */
    if (dom_interface_version < 5) {
1283
        virXenErrorFunc(domain->conn, VIR_ERR_NO_XEN, __FUNCTION__,
1284
                        "unsupported in dom interface < 5", 0);
1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301
        return -1;
    }

    if (hypervisor_version > 1) {
        xen_op_v2_sys op_sys;
        xen_op_v2_dom op_dom;
        int ret;

        memset(&op_sys, 0, sizeof(op_sys));
        op_sys.cmd = XEN_V2_OP_GETSCHEDULERID;
        ret = xenHypervisorDoV2Sys(priv->handle, &op_sys);
        if (ret == -1) return -1;

        switch (op_sys.u.getschedulerid.sched_id){
        case XEN_SCHEDULER_SEDF:
            /* TODO: Implement for Xen/SEDF */
            TODO
1302
            return(-1);
1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313
        case XEN_SCHEDULER_CREDIT: {
            int weight_set = 0;
            int cap_set = 0;

            memset(&op_dom, 0, sizeof(op_dom));
            op_dom.cmd = XEN_V2_OP_SCHEDULER;
            op_dom.domain = (domid_t) domain->id;
            op_dom.u.getschedinfo.sched_id = XEN_SCHEDULER_CREDIT;
            op_dom.u.getschedinfo.cmd = XEN_DOMCTL_SCHEDOP_putinfo;

            /*
1314 1315
             * credit scheduler parameters
             * following values do not change the parameters
1316 1317 1318 1319 1320
             */
            op_dom.u.getschedinfo.u.credit.weight = 0;
            op_dom.u.getschedinfo.u.credit.cap    = (uint16_t)~0U;

            for (i = 0; i < nparams; i++) {
1321
                memset(&buf, 0, sizeof(buf));
1322
                if (STREQ (params[i].field, str_weight) &&
1323
                    params[i].type == VIR_DOMAIN_SCHED_FIELD_UINT) {
1324 1325
                    val = params[i].value.ui;
                    if ((val < 1) || (val > USHRT_MAX)) {
1326
                        snprintf(buf, sizeof(buf), _("Credit scheduler weight parameter (%d) is out of range (1-65535)"), val);
1327
                        virXenErrorFunc (domain->conn, VIR_ERR_INVALID_ARG, __FUNCTION__, buf, val);
1328 1329
                        return(-1);
                    }
1330
                    op_dom.u.getschedinfo.u.credit.weight = val;
1331 1332
                    weight_set = 1;
                } else if (STREQ (params[i].field, str_cap) &&
1333
                    params[i].type == VIR_DOMAIN_SCHED_FIELD_UINT) {
1334 1335
                    val = params[i].value.ui;
                    if (val > USHRT_MAX) {
1336
                        snprintf(buf, sizeof(buf), _("Credit scheduler cap parameter (%d) is out of range (0-65535)"), val);
1337
                        virXenErrorFunc (domain->conn, VIR_ERR_INVALID_ARG, __FUNCTION__, buf, val);
1338 1339
                        return(-1);
                    }
1340
                    op_dom.u.getschedinfo.u.credit.cap = val;
1341 1342
                    cap_set = 1;
                } else {
1343 1344
                    virXenErrorFunc (domain->conn, VIR_ERR_INVALID_ARG, __FUNCTION__,
                                     "Credit scheduler accepts 'cap' and 'weight' integer parameters",
1345 1346 1347
                                     0);
                    return(-1);
                }
1348 1349 1350 1351
            }

            ret = xenHypervisorDoV2Dom(priv->handle, &op_dom);
            if (ret < 0)
1352
                return -1;
1353
            break;
1354
        }
1355
        default:
1356
            virXenErrorFunc(domain->conn, VIR_ERR_INVALID_ARG, __FUNCTION__,
1357 1358 1359 1360
                        "Unknown scheduler", op_sys.u.getschedulerid.sched_id);
            return -1;
        }
    }
1361

1362 1363 1364
    return 0;
}

1365 1366 1367 1368 1369 1370

int
xenHypervisorDomainBlockStats (virDomainPtr dom,
                               const char *path,
                               struct _virDomainBlockStats *stats)
{
1371 1372
#ifdef __linux__
    xenUnifiedPrivatePtr priv;
1373

1374 1375 1376
    priv = (xenUnifiedPrivatePtr) dom->conn->privateData;
    return xenLinuxDomainBlockStats (priv, dom, path, stats);
#else
1377
    virXenErrorFunc (dom->conn, VIR_ERR_NO_SUPPORT, __FUNCTION__,
1378 1379 1380 1381
                     "block statistics not supported on this platform",
                     dom->id);
    return -1;
#endif
1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395
}

/* Paths have the form vif<domid>.<n> (this interface checks that
 * <domid> is the real domain ID and returns an error if not).
 *
 * In future we may allow you to query bridge stats (virbrX or
 * xenbrX), but that will probably be through a separate
 * virNetwork interface, as yet not decided.
 */
int
xenHypervisorDomainInterfaceStats (virDomainPtr dom,
                                   const char *path,
                                   struct _virDomainInterfaceStats *stats)
{
1396
#ifdef __linux__
1397 1398
    int rqdomid, device;

1399 1400 1401
    /* Verify that the vif requested is one belonging to the current
     * domain.
     */
1402
    if (sscanf (path, "vif%d.%d", &rqdomid, &device) != 2) {
1403
        virXenErrorFunc (dom->conn, VIR_ERR_INVALID_ARG, __FUNCTION__,
1404 1405 1406 1407
                         "invalid path, should be vif<domid>.<n>.", 0);
        return -1;
    }
    if (rqdomid != dom->id) {
1408
        virXenErrorFunc (dom->conn, VIR_ERR_INVALID_ARG, __FUNCTION__,
1409 1410 1411 1412
                         "invalid path, vif<domid> should match this domain ID", 0);
        return -1;
    }

1413 1414
    return linuxDomainInterfaceStats (dom->conn, path, stats);
#else
1415
    virXenErrorFunc (dom->conn, VIR_ERR_NO_SUPPORT, __FUNCTION__,
1416 1417
                     "/proc/net/dev: Interface not found", 0);
    return -1;
1418
#endif
1419 1420
}

1421 1422 1423 1424 1425 1426 1427 1428 1429 1430
/**
 * 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
1431
virXen_pausedomain(int handle, int id)
1432 1433 1434 1435 1436 1437 1438
{
    int ret = -1;

    if (hypervisor_version > 1) {
        xen_op_v2_dom op;

        memset(&op, 0, sizeof(op));
1439 1440 1441
        op.cmd = XEN_V2_OP_PAUSEDOMAIN;
        op.domain = (domid_t) id;
        ret = xenHypervisorDoV2Dom(handle, &op);
1442 1443 1444 1445
    } else if (hypervisor_version == 1) {
        xen_op_v1 op;

        memset(&op, 0, sizeof(op));
1446 1447 1448
        op.cmd = XEN_V1_OP_PAUSEDOMAIN;
        op.u.domain.domain = (domid_t) id;
        ret = xenHypervisorDoV1Op(handle, &op);
1449 1450 1451 1452
    } else if (hypervisor_version == 0) {
        xen_op_v0 op;

        memset(&op, 0, sizeof(op));
1453 1454 1455
        op.cmd = XEN_V0_OP_PAUSEDOMAIN;
        op.u.domain.domain = (domid_t) id;
        ret = xenHypervisorDoV0Op(handle, &op);
1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469
    }
    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
1470
virXen_unpausedomain(int handle, int id)
1471 1472 1473 1474 1475 1476 1477
{
    int ret = -1;

    if (hypervisor_version > 1) {
        xen_op_v2_dom op;

        memset(&op, 0, sizeof(op));
1478 1479 1480
        op.cmd = XEN_V2_OP_UNPAUSEDOMAIN;
        op.domain = (domid_t) id;
        ret = xenHypervisorDoV2Dom(handle, &op);
1481 1482 1483 1484
    } else if (hypervisor_version == 1) {
        xen_op_v1 op;

        memset(&op, 0, sizeof(op));
1485 1486 1487
        op.cmd = XEN_V1_OP_UNPAUSEDOMAIN;
        op.u.domain.domain = (domid_t) id;
        ret = xenHypervisorDoV1Op(handle, &op);
1488 1489 1490 1491
    } else if (hypervisor_version == 0) {
        xen_op_v0 op;

        memset(&op, 0, sizeof(op));
1492 1493 1494
        op.cmd = XEN_V0_OP_UNPAUSEDOMAIN;
        op.u.domain.domain = (domid_t) id;
        ret = xenHypervisorDoV0Op(handle, &op);
1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508
    }
    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
1509
virXen_destroydomain(int handle, int id)
1510 1511 1512 1513 1514 1515 1516
{
    int ret = -1;

    if (hypervisor_version > 1) {
        xen_op_v2_dom op;

        memset(&op, 0, sizeof(op));
1517 1518 1519
        op.cmd = XEN_V2_OP_DESTROYDOMAIN;
        op.domain = (domid_t) id;
        ret = xenHypervisorDoV2Dom(handle, &op);
1520 1521 1522 1523
    } else if (hypervisor_version == 1) {
        xen_op_v1 op;

        memset(&op, 0, sizeof(op));
1524 1525 1526
        op.cmd = XEN_V1_OP_DESTROYDOMAIN;
        op.u.domain.domain = (domid_t) id;
        ret = xenHypervisorDoV1Op(handle, &op);
1527 1528 1529 1530
    } else if (hypervisor_version == 0) {
        xen_op_v0 op;

        memset(&op, 0, sizeof(op));
1531 1532 1533
        op.cmd = XEN_V0_OP_DESTROYDOMAIN;
        op.u.domain.domain = (domid_t) id;
        ret = xenHypervisorDoV0Op(handle, &op);
1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548
    }
    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
1549
virXen_setmaxmem(int handle, int id, unsigned long memory)
1550 1551 1552 1553 1554 1555 1556
{
    int ret = -1;

    if (hypervisor_version > 1) {
        xen_op_v2_dom op;

        memset(&op, 0, sizeof(op));
1557 1558
        op.cmd = XEN_V2_OP_SETMAXMEM;
        op.domain = (domid_t) id;
1559 1560 1561 1562
        if (dom_interface_version < 5)
            op.u.setmaxmem.maxmem = memory;
        else
            op.u.setmaxmemd5.maxmem = memory;
1563
        ret = xenHypervisorDoV2Dom(handle, &op);
1564 1565 1566 1567
    } else if (hypervisor_version == 1) {
        xen_op_v1 op;

        memset(&op, 0, sizeof(op));
1568 1569 1570 1571
        op.cmd = XEN_V1_OP_SETMAXMEM;
        op.u.setmaxmem.domain = (domid_t) id;
        op.u.setmaxmem.maxmem = memory;
        ret = xenHypervisorDoV1Op(handle, &op);
1572
    } else if (hypervisor_version == 0) {
1573
        xen_op_v0 op;
1574 1575

        memset(&op, 0, sizeof(op));
1576 1577 1578 1579
        op.cmd = XEN_V0_OP_SETMAXMEM;
        op.u.setmaxmem.domain = (domid_t) id;
        op.u.setmaxmem.maxmem = memory;
        ret = xenHypervisorDoV0Op(handle, &op);
1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594
    }
    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
1595
virXen_setmaxvcpus(int handle, int id, unsigned int vcpus)
1596 1597 1598 1599 1600 1601 1602
{
    int ret = -1;

    if (hypervisor_version > 1) {
        xen_op_v2_dom op;

        memset(&op, 0, sizeof(op));
1603 1604 1605 1606
        op.cmd = XEN_V2_OP_SETMAXVCPU;
        op.domain = (domid_t) id;
        op.u.setmaxvcpu.maxvcpu = vcpus;
        ret = xenHypervisorDoV2Dom(handle, &op);
1607 1608 1609 1610
    } else if (hypervisor_version == 1) {
        xen_op_v1 op;

        memset(&op, 0, sizeof(op));
1611 1612 1613 1614
        op.cmd = XEN_V1_OP_SETMAXVCPU;
        op.u.setmaxvcpu.domain = (domid_t) id;
        op.u.setmaxvcpu.maxvcpu = vcpus;
        ret = xenHypervisorDoV1Op(handle, &op);
1615
    } else if (hypervisor_version == 0) {
1616
        xen_op_v0 op;
1617 1618

        memset(&op, 0, sizeof(op));
1619 1620 1621 1622
        op.cmd = XEN_V0_OP_SETMAXVCPU;
        op.u.setmaxvcpu.domain = (domid_t) id;
        op.u.setmaxvcpu.maxvcpu = vcpus;
        ret = xenHypervisorDoV0Op(handle, &op);
1623 1624 1625 1626 1627 1628 1629 1630 1631 1632
    }
    return(ret);
}

/**
 * virXen_setvcpumap:
 * @handle: the hypervisor handle
 * @id: the domain id
 * @vcpu: the vcpu to map
 * @cpumap: the bitmap for this vcpu
1633
 * @maplen: the size of the bitmap in bytes
1634 1635 1636 1637 1638 1639 1640 1641 1642 1643
 *
 * 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;
1644 1645 1646
    unsigned char *new = NULL;
    unsigned char *bitmap = NULL;
    uint32_t nr_cpus;
1647 1648 1649 1650

    if (hypervisor_version > 1) {
        xen_op_v2_dom op;

1651
        if (lock_pages(cpumap, maplen) < 0) {
1652
            virXenError(NULL, VIR_ERR_XEN_CALL, " locking", maplen);
1653 1654
            return (-1);
        }
1655
        memset(&op, 0, sizeof(op));
1656 1657
        op.cmd = XEN_V2_OP_SETVCPUMAP;
        op.domain = (domid_t) id;
1658 1659 1660 1661 1662 1663

        /* The allocated memory to cpumap must be 'sizeof(uint64_t)' byte *
         * for Xen, and also nr_cpus must be 'sizeof(uint64_t) * 8'       */
        if (maplen < 8) {
            new = calloc(1, sizeof(uint64_t));
            if (!new) {
1664
                virXenErrorFunc(NULL, VIR_ERR_NO_MEMORY, __FUNCTION__,
1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675
                                "allocating private data", 0);
                return (-1);
            }
            memcpy(new, cpumap, maplen);
            bitmap = new;
            nr_cpus = sizeof(uint64_t) * 8;
        } else {
            bitmap = cpumap;
            nr_cpus = maplen * 8;
        }

1676 1677
        if (dom_interface_version < 5) {
            op.u.setvcpumap.vcpu = vcpu;
1678 1679
            op.u.setvcpumap.cpumap.bitmap = bitmap;
            op.u.setvcpumap.cpumap.nr_cpus = nr_cpus;
1680 1681
        } else {
            op.u.setvcpumapd5.vcpu = vcpu;
1682 1683
            op.u.setvcpumapd5.cpumap.bitmap.v = bitmap;
            op.u.setvcpumapd5.cpumap.nr_cpus = nr_cpus;
1684
        }
1685
        ret = xenHypervisorDoV2Dom(handle, &op);
1686
        free(new);
1687

1688
        if (unlock_pages(cpumap, maplen) < 0) {
1689
            virXenError(NULL, VIR_ERR_XEN_CALL, " release", maplen);
1690 1691
            ret = -1;
        }
1692
    } else {
1693 1694 1695
        cpumap_t xen_cpumap; /* limited to 64 CPUs in old hypervisors */
        uint64_t *pm = &xen_cpumap;
        int j;
1696

1697 1698
        if ((maplen > (int)sizeof(cpumap_t)) || (sizeof(cpumap_t) & 7))
            return (-1);
1699

1700 1701 1702
        memset(pm, 0, sizeof(cpumap_t));
        for (j = 0; j < maplen; j++)
            *(pm + (j / 8)) |= cpumap[j] << (8 * (j & 7));
1703 1704

        if (hypervisor_version == 1) {
1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718 1719 1720 1721 1722
            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);
        }
1723 1724 1725
    }
    return(ret);
}
1726 1727
#endif /* !PROXY*/

1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741
/**
 * 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,
1742
                    unsigned char *cpumap, int maplen)
1743 1744 1745 1746 1747 1748 1749
{
    int ret = -1;

    if (hypervisor_version > 1) {
        xen_op_v2_dom op;

        memset(&op, 0, sizeof(op));
1750 1751
        op.cmd = XEN_V2_OP_GETVCPUINFO;
        op.domain = (domid_t) id;
1752 1753 1754 1755
        if (dom_interface_version < 5)
            op.u.getvcpuinfo.vcpu = (uint16_t) vcpu;
        else
            op.u.getvcpuinfod5.vcpu = (uint16_t) vcpu;
1756
        ret = xenHypervisorDoV2Dom(handle, &op);
1757

1758 1759
        if (ret < 0)
            return(-1);
1760
        ipt->number = vcpu;
1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782
        if (dom_interface_version < 5) {
            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;
        } else {
            if (op.u.getvcpuinfod5.online) {
                if (op.u.getvcpuinfod5.running)
                    ipt->state = VIR_VCPU_RUNNING;
                if (op.u.getvcpuinfod5.blocked)
                    ipt->state = VIR_VCPU_BLOCKED;
            } else
                ipt->state = VIR_VCPU_OFFLINE;

            ipt->cpuTime = op.u.getvcpuinfod5.cpu_time;
            ipt->cpu = op.u.getvcpuinfod5.online ? (int)op.u.getvcpuinfod5.cpu : -1;
1783 1784
        }
        if ((cpumap != NULL) && (maplen > 0)) {
1785
            if (lock_pages(cpumap, maplen) < 0) {
1786
                virXenError(NULL, VIR_ERR_XEN_CALL, " locking", maplen);
1787 1788
                return (-1);
            }
1789
            memset(cpumap, 0, maplen);
1790 1791 1792
            memset(&op, 0, sizeof(op));
            op.cmd = XEN_V2_OP_GETVCPUMAP;
            op.domain = (domid_t) id;
1793 1794 1795 1796 1797 1798 1799 1800 1801
            if (dom_interface_version < 5) {
                op.u.getvcpumap.vcpu = vcpu;
                op.u.getvcpumap.cpumap.bitmap = cpumap;
                op.u.getvcpumap.cpumap.nr_cpus = maplen * 8;
            } else {
                op.u.getvcpumapd5.vcpu = vcpu;
                op.u.getvcpumapd5.cpumap.bitmap.v = cpumap;
                op.u.getvcpumapd5.cpumap.nr_cpus = maplen * 8;
            }
1802
            ret = xenHypervisorDoV2Dom(handle, &op);
1803
            if (unlock_pages(cpumap, maplen) < 0) {
1804
                virXenError(NULL, VIR_ERR_XEN_CALL, " release", maplen);
1805 1806 1807
                ret = -1;
            }
        }
1808
    } else {
1809 1810 1811 1812 1813
        int mapl = maplen;
        int cpu;

        if (maplen > (int)sizeof(cpumap_t))
            mapl = (int)sizeof(cpumap_t);
1814 1815

        if (hypervisor_version == 1) {
1816 1817 1818 1819 1820 1821 1822 1823 1824
            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);
1825
            ipt->number = vcpu;
1826 1827 1828 1829 1830 1831 1832 1833 1834 1835 1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 1846 1847 1848
            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);
1849
            ipt->number = vcpu;
1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863
            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);
                }
            }
        }
1864 1865 1866
    }
    return(ret);
}
1867

1868 1869 1870 1871 1872 1873
/**
 * xenHypervisorInit:
 *
 * Initialize the hypervisor layer. Try to detect the kind of interface
 * used i.e. pre or post changeset 10277
 */
1874
int
D
Daniel P. Berrange 已提交
1875
xenHypervisorInit(void)
1876
{
1877
    int fd, ret, cmd, errcode;
1878
    hypercall_t hc;
1879
    v0_hypercall_t v0_hc;
1880
    xen_getdomaininfo info;
D
Daniel Veillard 已提交
1881
    virVcpuInfoPtr ipt = NULL;
1882 1883

    if (initialized) {
1884
        if (hypervisor_version == -1)
1885
            return (-1);
1886
        return(0);
1887 1888
    }
    initialized = 1;
1889
    in_init = 1;
1890

1891 1892 1893 1894 1895 1896 1897 1898 1899
    /* 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);
1900
        virXenError (NULL, VIR_ERR_INTERNAL_ERROR, error, 0);
1901 1902 1903 1904 1905 1906 1907 1908 1909
        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);
1910
        virXenError (NULL, VIR_ERR_INTERNAL_ERROR, error, 0);
1911 1912 1913 1914 1915 1916 1917 1918 1919 1920
        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);
1921
        virXenError (NULL, VIR_ERR_INTERNAL_ERROR, error, 0);
1922 1923 1924 1925 1926
        in_init = 0;
        return -1;
    }

    /* Xen hypervisor version detection begins. */
1927 1928
    ret = open(XEN_HYPERVISOR_SOCKET, O_RDWR);
    if (ret < 0) {
1929
        hypervisor_version = -1;
1930
        return(-1);
1931 1932 1933
    }
    fd = ret;

1934 1935 1936 1937
    /*
     * The size of the hypervisor call block changed July 2006
     * this detect if we are using the new or old hypercall_t structure
     */
1938 1939 1940 1941 1942 1943 1944 1945
    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)) {
1946
        DEBUG("Using new hypervisor call: %X\n", ret);
1947 1948 1949
        hv_version = ret;
        xen_ioctl_hypercall_cmd = cmd;
        goto detect_v2;
1950
    }
1951

1952
#ifndef __sun__
1953 1954 1955 1956 1957 1958 1959 1960
    /*
     * 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);
1961
    if ((ret != -1) && (ret != 0)) {
1962
        DEBUG("Using old hypervisor call: %X\n", ret);
1963 1964
        hv_version = ret;
        xen_ioctl_hypercall_cmd = cmd;
1965
        hypervisor_version = 0;
1966
        goto done;
1967
    }
1968
#endif
1969

1970
    /*
R
Richard W.M. Jones 已提交
1971
     * we failed to make any hypercall
1972 1973 1974
     */

    hypervisor_version = -1;
1975
    virXenError(NULL, VIR_ERR_XEN_CALL, " ioctl ", IOCTL_PRIVCMD_HYPERCALL);
1976
    close(fd);
1977 1978 1979
    in_init = 0;
    return(-1);

1980
 detect_v2:
1981 1982 1983 1984 1985 1986
    /*
     * 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;
1987

1988
    ipt = malloc(sizeof(*ipt));
1989
    if (ipt == NULL){
1990
        virXenError(NULL, VIR_ERR_NO_MEMORY, __FUNCTION__, 0);
1991 1992
        return(-1);
    }
1993
    /* Currently consider RHEL5.0 Fedora7, xen-3.1, and xen-unstable */
1994
    sys_interface_version = 2; /* XEN_SYSCTL_INTERFACE_VERSION */
1995
    if (virXen_getdomaininfo(fd, 0, &info) == 1) {
1996 1997 1998
        /* RHEL 5.0 */
        dom_interface_version = 3; /* XEN_DOMCTL_INTERFACE_VERSION */
        if (virXen_getvcpusinfo(fd, 0, 0, ipt, NULL, 0) == 0){
1999
            DEBUG0("Using hypervisor call v2, sys ver2 dom ver3\n");
2000 2001 2002 2003 2004
            goto done;
        }
        /* Fedora 7 */
        dom_interface_version = 4; /* XEN_DOMCTL_INTERFACE_VERSION */
        if (virXen_getvcpusinfo(fd, 0, 0, ipt, NULL, 0) == 0){
2005
            DEBUG0("Using hypervisor call v2, sys ver2 dom ver4\n");
2006 2007 2008 2009 2010 2011
            goto done;
        }
    }

    sys_interface_version = 3; /* XEN_SYSCTL_INTERFACE_VERSION */
    if (virXen_getdomaininfo(fd, 0, &info) == 1) {
2012
        /* xen-3.1 */
2013 2014
        dom_interface_version = 5; /* XEN_DOMCTL_INTERFACE_VERSION */
        if (virXen_getvcpusinfo(fd, 0, 0, ipt, NULL, 0) == 0){
2015
            DEBUG0("Using hypervisor call v2, sys ver3 dom ver5\n");
2016 2017
            goto done;
        }
2018
    }
2019

2020 2021
    sys_interface_version = 4; /* XEN_SYSCTL_INTERFACE_VERSION */
    if (virXen_getdomaininfo(fd, 0, &info) == 1) {
2022
        /* Fedora 8 */
2023 2024
        dom_interface_version = 5; /* XEN_DOMCTL_INTERFACE_VERSION */
        if (virXen_getvcpusinfo(fd, 0, 0, ipt, NULL, 0) == 0){
2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035
            DEBUG0("Using hypervisor call v2, sys ver4 dom ver5\n");
            goto done;
        }
    }

    sys_interface_version = 6; /* XEN_SYSCTL_INTERFACE_VERSION */
    if (virXen_getdomaininfo(fd, 0, &info) == 1) {
        /* Xen 3.2, Fedora 9 */
        dom_interface_version = 5; /* XEN_DOMCTL_INTERFACE_VERSION */
        if (virXen_getvcpusinfo(fd, 0, 0, ipt, NULL, 0) == 0){
            DEBUG0("Using hypervisor call v2, sys ver6 dom ver5\n");
2036 2037 2038 2039
            goto done;
        }
    }

2040 2041
    hypervisor_version = 1;
    sys_interface_version = -1;
2042
    if (virXen_getdomaininfo(fd, 0, &info) == 1) {
2043
        DEBUG0("Using hypervisor call v1\n");
2044
        goto done;
2045 2046 2047
    }

    /*
R
Richard W.M. Jones 已提交
2048
     * we failed to make the getdomaininfolist hypercall
2049 2050
     */

2051
    DEBUG0("Failed to find any Xen hypervisor method\n");
2052
    hypervisor_version = -1;
2053
    virXenError(NULL, VIR_ERR_XEN_CALL, " ioctl ", IOCTL_PRIVCMD_HYPERCALL);
2054 2055
    close(fd);
    in_init = 0;
2056
    free(ipt);
2057 2058
    return(-1);

2059
 done:
2060
    close(fd);
2061
    in_init = 0;
2062
    free(ipt);
2063 2064 2065
    return(0);
}

2066 2067
/**
 * xenHypervisorOpen:
2068 2069 2070
 * @conn: pointer to the connection block
 * @name: URL for the target, NULL for local
 * @flags: combination of virDrvOpenFlag(s)
2071 2072 2073
 *
 * Connects to the Xen hypervisor.
 *
2074
 * Returns 0 or -1 in case of error.
2075
 */
2076
int
2077
xenHypervisorOpen(virConnectPtr conn,
2078
                  xmlURIPtr uri ATTRIBUTE_UNUSED,
2079
                  virConnectAuthPtr auth ATTRIBUTE_UNUSED,
2080
                  int flags ATTRIBUTE_UNUSED)
2081
{
2082
    int ret;
2083
    xenUnifiedPrivatePtr priv = (xenUnifiedPrivatePtr) conn->privateData;
2084

2085
    if (initialized == 0)
2086 2087
        if (xenHypervisorInit() == -1)
            return -1;
2088

2089
    priv->handle = -1;
2090

2091
    ret = open(XEN_HYPERVISOR_SOCKET, O_RDWR);
2092
    if (ret < 0) {
2093
        virXenError(conn, VIR_ERR_NO_XEN, XEN_HYPERVISOR_SOCKET, 0);
2094
        return (-1);
2095
    }
2096 2097

    priv->handle = ret;
2098

2099
    return(0);
2100 2101 2102 2103
}

/**
 * xenHypervisorClose:
2104
 * @conn: pointer to the connection block
2105 2106 2107 2108 2109
 *
 * Close the connection to the Xen hypervisor.
 *
 * Returns 0 in case of success or -1 in case of error.
 */
2110
int
2111
xenHypervisorClose(virConnectPtr conn)
2112
{
2113
    int ret;
2114
    xenUnifiedPrivatePtr priv;
2115

2116
    if (conn == NULL)
2117
        return (-1);
2118

2119 2120 2121 2122 2123 2124
    priv = (xenUnifiedPrivatePtr) conn->privateData;

    if (priv->handle < 0)
        return -1;

    ret = close(priv->handle);
2125
    if (ret < 0)
2126
        return (-1);
2127

2128
    return (0);
2129 2130 2131
}


2132 2133
/**
 * xenHypervisorGetVersion:
2134 2135
 * @conn: pointer to the connection block
 * @hvVer: where to store the version
2136 2137 2138
 *
 * Call the hypervisor to extracts his own internal API version
 *
2139
 * Returns 0 in case of success, -1 in case of error
2140
 */
2141 2142
int
xenHypervisorGetVersion(virConnectPtr conn, unsigned long *hvVer)
2143
{
2144 2145 2146 2147 2148 2149
    xenUnifiedPrivatePtr priv;

    if (conn == NULL)
        return -1;
    priv = (xenUnifiedPrivatePtr) conn->privateData;
    if (priv->handle < 0 || hvVer == NULL)
2150
        return (-1);
2151
    *hvVer = (hv_version >> 16) * 1000000 + (hv_version & 0xFFFF) * 1000;
2152
    return(0);
2153 2154
}

2155 2156 2157 2158 2159 2160 2161 2162 2163 2164 2165 2166 2167 2168 2169 2170 2171 2172 2173 2174 2175 2176 2177 2178 2179 2180 2181 2182 2183 2184 2185 2186 2187 2188 2189 2190 2191 2192 2193 2194 2195 2196 2197 2198 2199 2200 2201 2202 2203 2204 2205 2206 2207 2208 2209 2210 2211 2212 2213 2214 2215 2216 2217 2218 2219 2220 2221 2222 2223 2224 2225 2226 2227 2228 2229 2230 2231 2232 2233 2234 2235 2236 2237 2238 2239 2240 2241 2242 2243 2244 2245 2246 2247 2248 2249 2250 2251 2252 2253 2254 2255 2256 2257 2258 2259 2260 2261 2262 2263 2264 2265 2266 2267 2268 2269
struct guest_arch {
    const char *model;
    int bits;
    int hvm;
    int pae;
    int nonpae;
    int ia64_be;
};


static virCapsPtr
xenHypervisorBuildCapabilities(virConnectPtr conn,
                               const char *hostmachine,
                               int host_pae,
                               char *hvm_type,
                               struct guest_arch *guest_archs,
                               int nr_guest_archs) {
    virCapsPtr caps;
    int i;
    int hv_major = hv_version >> 16;
    int hv_minor = hv_version & 0xFFFF;

    if ((caps = virCapabilitiesNew(hostmachine, 1, 1)) == NULL)
        goto no_memory;
    if (hvm_type && STRNEQ(hvm_type, "") &&
        virCapabilitiesAddHostFeature(caps, hvm_type) < 0)
        goto no_memory;
    if (host_pae &&
        virCapabilitiesAddHostFeature(caps, "pae") < 0)
        goto no_memory;


    if (virCapabilitiesAddHostMigrateTransport(caps,
                                               "xenmigr") < 0)
        goto no_memory;


    if (sys_interface_version >= 4) {
        if (xenDaemonNodeGetTopology(conn, caps) != 0) {
            virCapabilitiesFree(caps);
            return NULL;
        }
    }

    for (i = 0; i < nr_guest_archs; ++i) {
        virCapsGuestPtr guest;
        const char const *machines[] = { guest_archs[i].hvm ? "xenfv" : "xenpv" };

        if ((guest = virCapabilitiesAddGuest(caps,
                                             guest_archs[i].hvm ? "hvm" : "xen",
                                             guest_archs[i].model,
                                             guest_archs[i].bits,
                                             (STREQ(hostmachine, "x86_64") ?
                                              "/usr/lib64/xen/bin/qemu-dm" :
                                              "/usr/lib/xen/bin/qemu-dm"),
                                             (guest_archs[i].hvm ?
                                              "/usr/lib/xen/boot/hvmloader" :
                                              NULL),
                                             1,
                                             machines)) == NULL)
            goto no_memory;

        if (virCapabilitiesAddGuestDomain(guest,
                                          "xen",
                                          NULL,
                                          NULL,
                                          0,
                                          NULL) == NULL)
            goto no_memory;

        if (guest_archs[i].pae &&
            virCapabilitiesAddGuestFeature(guest,
                                           "pae",
                                           1,
                                           0) == NULL)
            goto no_memory;

        if (guest_archs[i].nonpae &&
            virCapabilitiesAddGuestFeature(guest,
                                           "nonpae",
                                           1,
                                           0) == NULL)
            goto no_memory;

        if (guest_archs[i].ia64_be &&
            virCapabilitiesAddGuestFeature(guest,
                                           "ia64_be",
                                           1,
                                           0) == NULL)
            goto no_memory;

        if (guest_archs[i].hvm) {
            if (virCapabilitiesAddGuestFeature(guest,
                                               "acpi",
                                               1, 1) == NULL)
                goto no_memory;

            // In Xen 3.1.0, APIC is always on and can't be toggled
            if (virCapabilitiesAddGuestFeature(guest,
                                               "apic",
                                               1,
                                               (hv_major > 3 &&
                                                hv_minor > 0 ?
                                                0 : 1)) == NULL)
                goto no_memory;
        }
    }

    return caps;

 no_memory:
    virCapabilitiesFree(caps);
    return NULL;
}

2270 2271 2272
/**
 * xenHypervisorGetCapabilities:
 * @conn: pointer to the connection block
2273 2274
 * @cpuinfo: file handle containing /proc/cpuinfo data, or NULL
 * @capabilities: file handle containing /sys/hypervisor/properties/capabilities data, or NULL
2275 2276 2277 2278
 *
 * Return the capabilities of this hypervisor.
 */
char *
2279
xenHypervisorMakeCapabilitiesXML(virConnectPtr conn,
2280 2281
                                 const char *hostmachine,
                                 FILE *cpuinfo, FILE *capabilities)
2282 2283
{
    char line[1024], *str, *token;
2284
    regmatch_t subs[4];
2285
    char *saveptr = NULL;
2286
    int i;
2287 2288 2289

    char hvm_type[4] = ""; /* "vmx" or "svm" (or "" if not in CPU). */
    int host_pae = 0;
2290
    struct guest_arch guest_archs[32];
2291 2292
    int nr_guest_archs = 0;

2293
    char *xml;
2294

2295 2296

    virCapsPtr caps = NULL;
2297

2298 2299
    memset(guest_archs, 0, sizeof(guest_archs));

2300 2301 2302 2303
    /* /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)
     */
2304 2305 2306 2307 2308 2309 2310 2311 2312 2313
    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;
        }
2314 2315 2316 2317 2318 2319 2320 2321 2322 2323 2324 2325 2326 2327 2328 2329 2330 2331 2332 2333 2334 2335 2336 2337 2338 2339 2340
    }

    /* 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. */
2341
    if ((capabilities) && (fgets (line, sizeof line, capabilities))) {
2342 2343 2344 2345 2346 2347 2348 2349 2350 2351
        /* 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) {
2352
                int hvm = STRPREFIX(&token[subs[1].rm_so], "hvm");
2353 2354
                const char *model;
                int bits, pae = 0, nonpae = 0, ia64_be = 0;
2355 2356

                if (STRPREFIX(&token[subs[2].rm_so], "x86_32")) {
2357 2358
                    model = "i686";
                    bits = 32;
2359 2360
                    if (subs[3].rm_so != -1 &&
                        STRPREFIX(&token[subs[3].rm_so], "p"))
2361 2362 2363 2364
                        pae = 1;
                    else
                        nonpae = 1;
                }
2365
                else if (STRPREFIX(&token[subs[2].rm_so], "x86_64")) {
2366 2367 2368
                    model = "x86_64";
                    bits = 64;
                }
2369
                else if (STRPREFIX(&token[subs[2].rm_so], "ia64")) {
2370 2371
                    model = "ia64";
                    bits = 64;
2372 2373
                    if (subs[3].rm_so != -1 &&
                        STRPREFIX(&token[subs[3].rm_so], "be"))
2374 2375
                        ia64_be = 1;
                }
2376
                else if (STRPREFIX(&token[subs[2].rm_so], "powerpc64")) {
2377 2378 2379 2380 2381 2382
                    model = "ppc64";
                    bits = 64;
                } else {
                    /* XXX surely no other Xen archs exist  */
                    continue;
                }
2383

2384 2385
                /* Search for existing matching (model,hvm) tuple */
                for (i = 0 ; i < nr_guest_archs ; i++) {
2386
                    if (STREQ(guest_archs[i].model, model) &&
2387 2388 2389 2390
                        guest_archs[i].hvm == hvm) {
                        break;
                    }
                }
2391

2392 2393 2394 2395 2396 2397 2398 2399 2400 2401 2402 2403 2404 2405 2406 2407 2408 2409 2410 2411 2412
                /* 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;
2413 2414 2415 2416
            }
        }
    }

2417 2418 2419 2420 2421 2422 2423 2424 2425
    if ((caps = xenHypervisorBuildCapabilities(conn,
                                               hostmachine,
                                               host_pae,
                                               hvm_type,
                                               guest_archs,
                                               nr_guest_archs)) == NULL)
        goto no_memory;
    if ((xml = virCapabilitiesFormatXML(caps)) == NULL)
        goto no_memory;
2426

2427 2428
    virCapabilitiesFree(caps);
    return xml;
2429

2430
 no_memory:
2431
    virXenError(conn, VIR_ERR_NO_MEMORY, __FUNCTION__, 0);
2432
    virCapabilitiesFree(caps);
2433 2434 2435 2436 2437 2438 2439 2440 2441 2442 2443 2444 2445 2446 2447 2448 2449 2450 2451 2452 2453 2454 2455 2456 2457 2458 2459 2460 2461 2462 2463 2464 2465 2466 2467 2468 2469 2470 2471 2472 2473 2474 2475 2476
    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;
2477 2478
}

2479 2480 2481 2482 2483 2484 2485 2486 2487 2488 2489
/**
 * 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)
{
2490
    xen_getdomaininfolist dominfos;
2491 2492 2493
    int ret, nbids;
    static int last_maxids = 2;
    int maxids = last_maxids;
2494
    xenUnifiedPrivatePtr priv;
2495

2496 2497 2498 2499
    if (conn == NULL)
        return -1;
    priv = (xenUnifiedPrivatePtr) conn->privateData;
    if (priv->handle < 0)
2500 2501
        return (-1);

2502 2503
 retry:
    if (!(XEN_GETDOMAININFOLIST_ALLOC(dominfos, maxids))) {
2504
        virXenError(conn, VIR_ERR_NO_MEMORY, _("allocating %d domain info"),
2505 2506
                    maxids);
        return(-1);
2507 2508
    }

2509 2510
    XEN_GETDOMAININFOLIST_CLEAR(dominfos, maxids);

2511
    ret = virXen_getdomaininfolist(priv->handle, 0, maxids, &dominfos);
2512

2513
    XEN_GETDOMAININFOLIST_FREE(dominfos);
2514 2515 2516 2517

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

2518
    nbids = ret;
2519 2520 2521 2522 2523
    /* Can't possibly have more than 65,000 concurrent guests
     * so limit how many times we try, to avoid increasing
     * without bound & thus allocating all of system memory !
     * XXX I'll regret this comment in a few years time ;-)
     */
2524
    if (nbids == maxids) {
2525 2526 2527 2528 2529 2530
        if (maxids < 65000) {
            last_maxids *= 2;
            maxids *= 2;
            goto retry;
        }
        nbids = -1;
2531 2532 2533 2534 2535 2536 2537 2538 2539 2540 2541 2542 2543 2544 2545 2546 2547 2548 2549
    }
    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)
{
2550
    xen_getdomaininfolist dominfos;
2551
    int ret, nbids, i;
2552 2553 2554 2555
    xenUnifiedPrivatePtr priv;

    if (conn == NULL)
        return -1;
2556

2557 2558
    priv = (xenUnifiedPrivatePtr) conn->privateData;
    if (priv->handle < 0 ||
2559 2560 2561
        (ids == NULL) || (maxids < 1))
        return (-1);

2562
    if (!(XEN_GETDOMAININFOLIST_ALLOC(dominfos, maxids))) {
2563
        virXenError(conn, VIR_ERR_NO_MEMORY, "allocating %d domain info",
2564 2565
                    maxids);
        return(-1);
2566
    }
2567 2568

    XEN_GETDOMAININFOLIST_CLEAR(dominfos, maxids);
2569 2570
    memset(ids, 0, maxids * sizeof(int));

2571
    ret = virXen_getdomaininfolist(priv->handle, 0, maxids, &dominfos);
2572 2573

    if (ret < 0) {
2574
        XEN_GETDOMAININFOLIST_FREE(dominfos);
2575 2576 2577
        return (-1);
    }

2578
    nbids = ret;
2579
    if ((nbids < 0) || (nbids > maxids)) {
2580
        XEN_GETDOMAININFOLIST_FREE(dominfos);
2581 2582 2583 2584
        return(-1);
    }

    for (i = 0;i < nbids;i++) {
2585
        ids[i] = XEN_GETDOMAININFOLIST_DOMAIN(dominfos, i);
2586 2587
    }

2588
    XEN_GETDOMAININFOLIST_FREE(dominfos);
2589 2590 2591
    return (nbids);
}

2592 2593 2594 2595 2596 2597 2598 2599 2600 2601 2602 2603 2604 2605 2606 2607 2608 2609 2610 2611 2612 2613 2614 2615 2616 2617 2618 2619 2620 2621 2622 2623 2624 2625 2626 2627 2628 2629 2630 2631 2632 2633 2634 2635 2636 2637 2638 2639 2640 2641 2642 2643 2644 2645 2646 2647 2648 2649 2650 2651 2652 2653 2654 2655 2656 2657 2658 2659 2660 2661 2662 2663 2664 2665 2666 2667 2668 2669 2670

#ifndef PROXY
char *
xenHypervisorDomainGetOSType (virDomainPtr dom)
{
    xenUnifiedPrivatePtr priv;
    xen_getdomaininfo dominfo;

    priv = (xenUnifiedPrivatePtr) dom->conn->privateData;
    if (priv->handle < 0)
        return (NULL);

    /* HV's earlier than 3.1.0 don't include the HVM flags in guests status*/
    if (hypervisor_version < 2 ||
        dom_interface_version < 4)
        return (NULL);

    XEN_GETDOMAININFO_CLEAR(dominfo);

    if (virXen_getdomaininfo(priv->handle, dom->id, &dominfo) < 0)
        return (NULL);

    if (XEN_GETDOMAININFO_DOMAIN(dominfo) != dom->id)
        return (NULL);

    if (XEN_GETDOMAININFO_FLAGS(dominfo) & DOMFLAGS_HVM)
        return strdup("hvm");
    return strdup("linux");
}

virDomainPtr
xenHypervisorLookupDomainByID(virConnectPtr conn,
                              int id)
{
    xenUnifiedPrivatePtr priv;
    xen_getdomaininfo dominfo;
    virDomainPtr ret;
    char *name;

    priv = (xenUnifiedPrivatePtr) conn->privateData;
    if (priv->handle < 0)
        return (NULL);

    XEN_GETDOMAININFO_CLEAR(dominfo);

    if (virXen_getdomaininfo(priv->handle, id, &dominfo) < 0)
        return (NULL);

    if (XEN_GETDOMAININFO_DOMAIN(dominfo) != id)
        return (NULL);


    if (!(name = xenStoreDomainGetName(conn, id)))
        return (NULL);

    ret = virGetDomain(conn, name, XEN_GETDOMAININFO_UUID(dominfo));
    if (ret)
        ret->id = id;
    free(name);
    return ret;
}


virDomainPtr
xenHypervisorLookupDomainByUUID(virConnectPtr conn,
                                const unsigned char *uuid)
{
    xen_getdomaininfolist dominfos;
    xenUnifiedPrivatePtr priv;
    virDomainPtr ret;
    char *name;
    int maxids = 100, nids, i, id;

    priv = (xenUnifiedPrivatePtr) conn->privateData;
    if (priv->handle < 0)
        return (NULL);

 retry:
    if (!(XEN_GETDOMAININFOLIST_ALLOC(dominfos, maxids))) {
2671
        virXenError(conn, VIR_ERR_NO_MEMORY, "allocating %d domain info",
2672 2673 2674 2675 2676 2677 2678 2679 2680 2681 2682 2683 2684 2685 2686 2687 2688 2689 2690 2691 2692 2693 2694 2695 2696 2697 2698 2699 2700 2701 2702 2703 2704 2705 2706 2707 2708 2709 2710 2711 2712 2713 2714 2715 2716 2717 2718 2719 2720 2721
                    maxids);
        return(NULL);
    }

    XEN_GETDOMAININFOLIST_CLEAR(dominfos, maxids);

    nids = virXen_getdomaininfolist(priv->handle, 0, maxids, &dominfos);

    if (nids < 0) {
        XEN_GETDOMAININFOLIST_FREE(dominfos);
        return (NULL);
    }

    /* Can't possibly have more than 65,000 concurrent guests
     * so limit how many times we try, to avoid increasing
     * without bound & thus allocating all of system memory !
     * XXX I'll regret this comment in a few years time ;-)
     */
    if (nids == maxids) {
        XEN_GETDOMAININFOLIST_FREE(dominfos);
        if (maxids < 65000) {
            maxids *= 2;
            goto retry;
        }
        return (NULL);
    }

    id = -1;
    for (i = 0 ; i < nids ; i++) {
        if (memcmp(XEN_GETDOMAININFOLIST_UUID(dominfos, i), uuid, VIR_UUID_BUFLEN) == 0) {
            id = XEN_GETDOMAININFOLIST_DOMAIN(dominfos, i);
            break;
        }
    }
    XEN_GETDOMAININFOLIST_FREE(dominfos);

    if (id == -1)
        return (NULL);

    if (!(name = xenStoreDomainGetName(conn, id)))
        return (NULL);

    ret = virGetDomain(conn, name, uuid);
    if (ret)
        ret->id = id;
    free(name);
    return ret;
}
#endif

2722
/**
2723
 * xenHypervisorGetMaxVcpus:
2724 2725 2726 2727
 *
 * Returns the maximum of CPU defined by Xen.
 */
int
2728 2729
xenHypervisorGetMaxVcpus(virConnectPtr conn,
                         const char *type ATTRIBUTE_UNUSED)
2730
{
2731 2732 2733 2734 2735 2736
    xenUnifiedPrivatePtr priv;

    if (conn == NULL)
        return -1;
    priv = (xenUnifiedPrivatePtr) conn->privateData;
    if (priv->handle < 0)
2737 2738 2739 2740 2741
        return (-1);

    return MAX_VIRT_CPUS;
}

2742
/**
2743 2744 2745
 * xenHypervisorGetDomMaxMemory:
 * @conn: connection data
 * @id: domain id
2746
 *
2747
 * Retrieve the maximum amount of physical memory allocated to a
2748
 * domain.
2749 2750 2751
 *
 * Returns the memory size in kilobytes or 0 in case of error.
 */
2752 2753
unsigned long
xenHypervisorGetDomMaxMemory(virConnectPtr conn, int id)
2754
{
2755
    xenUnifiedPrivatePtr priv;
2756
    xen_getdomaininfo dominfo;
2757 2758
    int ret;

2759 2760 2761 2762 2763 2764
    if (conn == NULL)
        return 0;

    priv = (xenUnifiedPrivatePtr) conn->privateData;
    if (priv->handle < 0)
        return 0;
2765

2766 2767
    if (kb_per_pages == 0) {
        kb_per_pages = sysconf(_SC_PAGESIZE) / 1024;
2768 2769
        if (kb_per_pages <= 0)
            kb_per_pages = 4;
2770 2771
    }

2772
    XEN_GETDOMAININFO_CLEAR(dominfo);
2773

2774
    ret = virXen_getdomaininfo(priv->handle, id, &dominfo);
2775

2776
    if ((ret < 0) || (XEN_GETDOMAININFO_DOMAIN(dominfo) != id))
2777 2778
        return (0);

2779
    return((unsigned long) XEN_GETDOMAININFO_MAX_PAGES(dominfo) * kb_per_pages);
2780 2781
}

2782
#ifndef PROXY
2783 2784 2785
/**
 * xenHypervisorGetMaxMemory:
 * @domain: a domain object or NULL
2786
 *
2787 2788 2789 2790 2791 2792 2793 2794 2795
 * 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)
{
2796 2797 2798 2799 2800 2801 2802
    xenUnifiedPrivatePtr priv;

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

    priv = (xenUnifiedPrivatePtr) domain->conn->privateData;
    if (priv->handle < 0 || domain->id < 0)
2803 2804
        return (0);

2805
    return(xenHypervisorGetDomMaxMemory(domain->conn, domain->id));
2806
}
2807
#endif
2808

2809
/**
2810 2811 2812
 * xenHypervisorGetDomInfo:
 * @conn: connection data
 * @id: the domain ID
2813
 * @info: the place where information should be stored
2814
 *
2815
 * Do an hypervisor call to get the related set of domain information.
2816 2817 2818 2819
 *
 * Returns 0 in case of success, -1 in case of error.
 */
int
2820
xenHypervisorGetDomInfo(virConnectPtr conn, int id, virDomainInfoPtr info)
2821
{
2822
    xenUnifiedPrivatePtr priv;
2823
    xen_getdomaininfo dominfo;
2824
    int ret;
2825
    uint32_t domain_flags, domain_state, domain_shutdown_cause;
2826 2827 2828

    if (kb_per_pages == 0) {
        kb_per_pages = sysconf(_SC_PAGESIZE) / 1024;
2829 2830
        if (kb_per_pages <= 0)
            kb_per_pages = 4;
2831
    }
2832

2833 2834 2835 2836 2837
    if (conn == NULL)
        return -1;

    priv = (xenUnifiedPrivatePtr) conn->privateData;
    if (priv->handle < 0 || info == NULL)
2838
        return (-1);
2839

2840
    memset(info, 0, sizeof(virDomainInfo));
2841
    XEN_GETDOMAININFO_CLEAR(dominfo);
2842

2843
    ret = virXen_getdomaininfo(priv->handle, id, &dominfo);
2844

2845
    if ((ret < 0) || (XEN_GETDOMAININFO_DOMAIN(dominfo) != id))
2846
        return (-1);
2847

2848
    domain_flags = XEN_GETDOMAININFO_FLAGS(dominfo);
2849 2850
    domain_flags &= ~DOMFLAGS_HVM; /* Mask out HVM flags */
    domain_state = domain_flags & 0xFF; /* Mask out high bits */
2851
    switch (domain_state) {
2852 2853 2854 2855
        case DOMFLAGS_DYING:
            info->state = VIR_DOMAIN_SHUTDOWN;
            break;
        case DOMFLAGS_SHUTDOWN:
2856 2857 2858 2859 2860 2861 2862 2863 2864
            /* 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;
            }
2865 2866 2867 2868 2869 2870 2871 2872 2873 2874 2875 2876
            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_NOSTATE;
2877 2878 2879 2880 2881 2882 2883
    }

    /*
     * the API brings back the cpu time in nanoseconds,
     * convert to microseconds, same thing convert to
     * kilobytes from page counts
     */
2884
    info->cpuTime = XEN_GETDOMAININFO_CPUTIME(dominfo);
2885
    info->memory = XEN_GETDOMAININFO_TOT_PAGES(dominfo) * kb_per_pages;
2886 2887 2888
    info->maxMem = XEN_GETDOMAININFO_MAX_PAGES(dominfo);
    if(info->maxMem != UINT_MAX)
        info->maxMem *= kb_per_pages;
2889
    info->nrVirtCpu = XEN_GETDOMAININFO_CPUCOUNT(dominfo);
2890
    return (0);
2891 2892
}

2893 2894 2895
/**
 * xenHypervisorGetDomainInfo:
 * @domain: pointer to the domain block
2896
 * @info: the place where information should be stored
2897
 *
2898
 * Do an hypervisor call to get the related set of domain information.
2899 2900 2901 2902 2903 2904
 *
 * Returns 0 in case of success, -1 in case of error.
 */
int
xenHypervisorGetDomainInfo(virDomainPtr domain, virDomainInfoPtr info)
{
2905 2906 2907 2908 2909 2910 2911
    xenUnifiedPrivatePtr priv;

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

    priv = (xenUnifiedPrivatePtr) domain->conn->privateData;
    if (priv->handle < 0 || info == NULL ||
2912
        (domain->id < 0))
2913
        return (-1);
2914

2915
    return(xenHypervisorGetDomInfo(domain->conn, domain->id, info));
2916 2917 2918

}

2919
#ifndef PROXY
2920 2921 2922 2923
/**
 * xenHypervisorNodeGetCellsFreeMemory:
 * @conn: pointer to the hypervisor connection
 * @freeMems: pointer to the array of unsigned long long
2924 2925
 * @startCell: index of first cell to return freeMems info on.
 * @maxCells: Maximum number of cells for which freeMems information can
2926 2927 2928 2929
 *            be returned.
 *
 * This call returns the amount of free memory in one or more NUMA cells.
 * The @freeMems array must be allocated by the caller and will be filled
2930 2931 2932
 * with the amount of free memory in kilobytes for each cell requested,
 * starting with startCell (in freeMems[0]), up to either
 * (startCell + maxCells), or the number of additional cells in the node,
2933 2934 2935 2936 2937 2938 2939 2940 2941 2942 2943
 * whichever is smaller.
 *
 * Returns the number of entries filled in freeMems, or -1 in case of error.
 */
int
xenHypervisorNodeGetCellsFreeMemory(virConnectPtr conn, unsigned long long *freeMems,
                                    int startCell, int maxCells)
{
    xen_op_v2_sys op_sys;
    int i, j, ret;
    xenUnifiedPrivatePtr priv;
2944
    int nbNodeCells;
2945

2946
    if (conn == NULL) {
2947
        virXenErrorFunc (conn, VIR_ERR_INVALID_ARG, __FUNCTION__,
2948 2949 2950
                        "invalid argument", 0);
        return -1;
    }
2951

2952 2953
    nbNodeCells = xenNbCells(conn);
    if (nbNodeCells < 0) {
2954
        virXenErrorFunc (conn, VIR_ERR_XEN_CALL, __FUNCTION__,
2955 2956
                         "cannot determine actual number of cells",0);
        return(-1);
2957 2958
    }

2959
    if ((maxCells < 1) || (startCell >= nbNodeCells)) {
2960
        virXenErrorFunc (conn, VIR_ERR_INVALID_ARG, __FUNCTION__,
2961 2962 2963
                        "invalid argument", 0);
        return -1;
    }
2964

2965 2966 2967 2968
    /*
     * Support only sys_interface_version >=4
     */
    if (sys_interface_version < 4) {
2969
        virXenErrorFunc (conn, VIR_ERR_XEN_CALL, __FUNCTION__,
2970 2971 2972 2973 2974 2975
                        "unsupported in sys interface < 4", 0);
        return -1;
    }

    priv = (xenUnifiedPrivatePtr) conn->privateData;
    if (priv->handle < 0) {
2976
        virXenErrorFunc (conn, VIR_ERR_INTERNAL_ERROR, __FUNCTION__,
2977 2978 2979 2980 2981 2982 2983 2984 2985 2986 2987 2988 2989 2990 2991 2992 2993 2994 2995
                        "priv->handle invalid", 0);
        return -1;
    }

    memset(&op_sys, 0, sizeof(op_sys));
    op_sys.cmd = XEN_V2_OP_GETAVAILHEAP;

    for (i = startCell, j = 0;(i < nbNodeCells) && (j < maxCells);i++,j++) {
        op_sys.u.availheap.node = i;
        ret = xenHypervisorDoV2Sys(priv->handle, &op_sys);
        if (ret < 0) {
            return(-1);
        }
        freeMems[j] = op_sys.u.availheap.avail_bytes;
    }
    return (j);
}


2996 2997
/**
 * xenHypervisorPauseDomain:
2998
 * @domain: pointer to the domain block
2999 3000 3001 3002 3003 3004
 *
 * Do an hypervisor call to pause the given domain
 *
 * Returns 0 in case of success, -1 in case of error.
 */
int
3005
xenHypervisorPauseDomain(virDomainPtr domain)
3006
{
3007
    int ret;
3008
    xenUnifiedPrivatePtr priv;
3009

3010 3011 3012 3013 3014
    if ((domain == NULL) || (domain->conn == NULL))
        return -1;

    priv = (xenUnifiedPrivatePtr) domain->conn->privateData;
    if (priv->handle < 0 || domain->id < 0)
3015 3016
        return (-1);

3017
    ret = virXen_pausedomain(priv->handle, domain->id);
3018
    if (ret < 0)
3019 3020
        return (-1);
    return (0);
3021 3022 3023 3024
}

/**
 * xenHypervisorResumeDomain:
3025
 * @domain: pointer to the domain block
3026 3027 3028 3029 3030 3031
 *
 * Do an hypervisor call to resume the given domain
 *
 * Returns 0 in case of success, -1 in case of error.
 */
int
3032
xenHypervisorResumeDomain(virDomainPtr domain)
3033
{
3034
    int ret;
3035 3036 3037 3038
    xenUnifiedPrivatePtr priv;

    if ((domain == NULL) || (domain->conn == NULL))
        return -1;
3039

3040 3041
    priv = (xenUnifiedPrivatePtr) domain->conn->privateData;
    if (priv->handle < 0 || domain->id < 0)
3042 3043
        return (-1);

3044
    ret = virXen_unpausedomain(priv->handle, domain->id);
3045
    if (ret < 0)
3046 3047
        return (-1);
    return (0);
3048 3049 3050 3051
}

/**
 * xenHypervisorDestroyDomain:
3052
 * @domain: pointer to the domain block
3053 3054 3055 3056 3057 3058
 *
 * Do an hypervisor call to destroy the given domain
 *
 * Returns 0 in case of success, -1 in case of error.
 */
int
3059
xenHypervisorDestroyDomain(virDomainPtr domain)
3060
{
3061
    int ret;
3062 3063 3064 3065
    xenUnifiedPrivatePtr priv;

    if (domain == NULL || domain->conn == NULL)
        return -1;
3066

3067 3068
    priv = (xenUnifiedPrivatePtr) domain->conn->privateData;
    if (priv->handle < 0 || domain->id < 0)
3069 3070
        return (-1);

3071
    ret = virXen_destroydomain(priv->handle, domain->id);
3072
    if (ret < 0)
3073 3074
        return (-1);
    return (0);
3075 3076
}

3077 3078
/**
 * xenHypervisorSetMaxMemory:
3079
 * @domain: pointer to the domain block
3080 3081 3082 3083 3084 3085 3086
 * @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
3087
xenHypervisorSetMaxMemory(virDomainPtr domain, unsigned long memory)
3088
{
3089
    int ret;
3090
    xenUnifiedPrivatePtr priv;
3091

3092 3093 3094 3095 3096
    if (domain == NULL || domain->conn == NULL)
        return -1;

    priv = (xenUnifiedPrivatePtr) domain->conn->privateData;
    if (priv->handle < 0 || domain->id < 0)
3097 3098
        return (-1);

3099
    ret = virXen_setmaxmem(priv->handle, domain->id, memory);
3100
    if (ret < 0)
3101 3102
        return (-1);
    return (0);
3103
}
3104
#endif /* PROXY */
3105

3106
#ifndef PROXY
3107 3108 3109 3110 3111 3112 3113 3114 3115 3116 3117 3118 3119
/**
 * 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)
{
3120
    int ret;
3121 3122 3123 3124
    xenUnifiedPrivatePtr priv;

    if (domain == NULL || domain->conn == NULL)
        return -1;
3125

3126 3127
    priv = (xenUnifiedPrivatePtr) domain->conn->privateData;
    if (priv->handle < 0 || domain->id < 0 || nvcpus < 1)
3128
        return (-1);
3129

3130
    ret = virXen_setmaxvcpus(priv->handle, domain->id, nvcpus);
3131
    if (ret < 0)
3132
        return (-1);
3133
    return (0);
3134 3135 3136 3137 3138 3139 3140 3141
}

/**
 * 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
3142
 *
3143 3144 3145 3146 3147 3148 3149 3150 3151
 * 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)
{
3152
    int ret;
3153
    xenUnifiedPrivatePtr priv;
3154

3155 3156 3157 3158 3159
    if (domain == NULL || domain->conn == NULL)
        return -1;

    priv = (xenUnifiedPrivatePtr) domain->conn->privateData;
    if (priv->handle < 0 || (domain->id < 0) ||
3160
        (cpumap == NULL) || (maplen < 1))
3161 3162
        return (-1);

3163
    ret = virXen_setvcpumap(priv->handle, domain->id, vcpu,
3164 3165
                            cpumap, maplen);
    if (ret < 0)
3166
        return (-1);
3167
    return (0);
3168
}
3169
#endif
3170 3171 3172 3173 3174 3175 3176

/**
 * 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)
D
Daniel Veillard 已提交
3177
 *	If cpumaps is NULL, then no cpumap information is returned by the API.
3178 3179 3180 3181 3182 3183
 *	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...).
3184
 *
3185
 * Extract information about virtual CPUs of domain, store it in info array
R
Richard W.M. Jones 已提交
3186
 * and also in cpumaps if this pointer isn't NULL.
3187 3188 3189
 *
 * Returns the number of info filled in case of success, -1 in case of failure.
 */
3190
#ifndef PROXY
3191 3192
int
xenHypervisorGetVcpus(virDomainPtr domain, virVcpuInfoPtr info, int maxinfo,
3193
                      unsigned char *cpumaps, int maplen)
3194
{
3195
    xen_getdomaininfo dominfo;
3196
    int ret;
3197
    xenUnifiedPrivatePtr priv;
3198
    virVcpuInfoPtr ipt;
3199
    int nbinfo, i;
3200

3201 3202 3203 3204 3205
    if (domain == NULL || domain->conn == NULL)
        return -1;

    priv = (xenUnifiedPrivatePtr) domain->conn->privateData;
    if (priv->handle < 0 || (domain->id < 0) ||
3206 3207
        (info == NULL) || (maxinfo < 1) ||
        (sizeof(cpumap_t) & 7))
3208
        return (-1);
3209
    if ((cpumaps != NULL) && (maplen < 1))
3210
        return -1;
3211 3212

    /* first get the number of virtual CPUs in this domain */
3213
    XEN_GETDOMAININFO_CLEAR(dominfo);
3214
    ret = virXen_getdomaininfo(priv->handle, domain->id,
3215
                               &dominfo);
3216

3217
    if ((ret < 0) || (XEN_GETDOMAININFO_DOMAIN(dominfo) != domain->id))
3218
        return (-1);
3219
    nbinfo = XEN_GETDOMAININFO_CPUCOUNT(dominfo) + 1;
3220 3221 3222
    if (nbinfo > maxinfo) nbinfo = maxinfo;

    if (cpumaps != NULL)
3223
        memset(cpumaps, 0, maxinfo * maplen);
3224

3225 3226
    for (i = 0, ipt = info; i < nbinfo; i++, ipt++) {
        if ((cpumaps != NULL) && (i < maxinfo)) {
3227
            ret = virXen_getvcpusinfo(priv->handle, domain->id, i,
3228 3229 3230 3231 3232 3233
                                      ipt,
                                      (unsigned char *)VIR_GET_CPUMAP(cpumaps, maplen, i),
                                      maplen);
            if (ret < 0)
                return(-1);
        } else {
3234
            ret = virXen_getvcpusinfo(priv->handle, domain->id, i,
3235 3236 3237 3238
                                      ipt, NULL, 0);
            if (ret < 0)
                return(-1);
        }
3239 3240 3241
    }
    return nbinfo;
}
3242
#endif /* PROXY */
3243

3244 3245 3246 3247 3248 3249 3250 3251 3252 3253 3254 3255 3256 3257
/**
 * 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;
3258 3259 3260 3261
    xenUnifiedPrivatePtr priv;

    if (domain == NULL || domain->conn == NULL)
        return -1;
3262

3263 3264
    priv = (xenUnifiedPrivatePtr) domain->conn->privateData;
    if (priv->handle < 0)
3265 3266 3267 3268 3269 3270 3271
        return (-1);

    /* inactive domain */
    if (domain->id < 0) {
        maxcpu = MAX_VIRT_CPUS;
    } else {
        XEN_GETDOMAININFO_CLEAR(dominfo);
3272
        ret = virXen_getdomaininfo(priv->handle, domain->id,
3273 3274 3275 3276 3277 3278 3279 3280 3281 3282
                                   &dominfo);

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

    return maxcpu;
}

3283
#endif /* WITH_XEN */