xen_internal.c 94.4 KB
Newer Older
1 2 3
/*
 * xen_internal.c: direct access to Xen hypervisor level
 *
D
Daniel Veillard 已提交
4
 * Copyright (C) 2005, 2006 Red Hat, Inc.
5 6 7 8 9 10
 *
 * See COPYING.LIB for the License of this software
 *
 * Daniel Veillard <veillard@redhat.com>
 */

11 12
#ifdef WITH_XEN

13 14
#include "config.h"

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

51 52
/* #define DEBUG */
/*
53
 * so far there is 2 versions of the structures usable for doing
54 55 56 57
 * hypervisor calls.
 */
/* the old one */
typedef struct v0_hypercall_struct {
58 59
    unsigned long op;
    unsigned long arg[5];
60
} v0_hypercall_t;
61 62

#ifdef __linux__
63 64 65 66 67 68 69 70
#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;
71 72
#define XEN_V1_IOCTL_HYPERCALL_CMD                  \
    _IOC(_IOC_NONE, 'P', 0, sizeof(v1_hypercall_t))
73
typedef v1_hypercall_t hypercall_t;
74 75 76 77 78
#elif define(__sun__)
typedef privcmd_hypercall_t hypercall_t;
#else
#error "unsupported platform"
#endif
79 80 81 82 83 84 85

#ifndef __HYPERVISOR_sysctl
#define __HYPERVISOR_sysctl 35
#endif
#ifndef __HYPERVISOR_domctl
#define __HYPERVISOR_domctl 36
#endif
86

87 88
static int xen_ioctl_hypercall_cmd = 0;
static int initialized = 0;
89
static int in_init = 0;
90
static int hv_version = 0;
91 92 93
static int hypervisor_version = 2;
static int sys_interface_version = -1;
static int dom_interface_version = -1;
94
static int kb_per_pages = 0;
95

96 97 98 99 100 101 102 103
/* 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;
104
static const char *xen_cap_re = "(xen|hvm)-[[:digit:]]+\\.[[:digit:]]+-(x86_32|x86_64|ia64|powerpc64)(p|be)?";
105 106
static regex_t xen_cap_rec;

107 108 109 110 111
/*
 * The content of the structures for a getdomaininfolist system hypercall
 */
#ifndef DOMFLAGS_DYING
#define DOMFLAGS_DYING     (1<<0) /* Domain is scheduled to die.             */
112
#define DOMFLAGS_HVM       (1<<1) /* Domain is HVM                           */
113 114 115 116 117 118 119 120 121 122
#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

123 124 125 126 127 128 129 130 131 132 133
/*
 * 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

134 135 136 137 138 139 140 141 142
#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 */
    uint32_t flags;	/* falgs, see before */
    uint64_t tot_pages;	/* total number of pages used */
    uint64_t max_pages;	/* maximum number of pages allowed */
143
    unsigned long shared_info_frame; /* MFN of shared_info struct */
144 145 146 147 148 149 150 151
    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;

152 153 154 155 156 157 158 159 160 161 162 163 164 165
struct xen_v2_getdomaininfo {
    domid_t  domain;	/* the domain number */
    uint32_t flags;	/* falgs, see before */
    uint64_t tot_pages;	/* total number of pages used */
    uint64_t max_pages;	/* maximum number of pages allowed */
    uint64_t shared_info_frame; /* MFN of shared_info struct */
    uint64_t cpu_time;  /* CPU time used */
    uint32_t nr_online_vcpus;  /* Number of VCPUs currently online. */
    uint32_t max_vcpu_id; /* Maximum VCPUID in use by this domain. */
    uint32_t ssidref;
    xen_domain_handle_t handle;
};
typedef struct xen_v2_getdomaininfo xen_v2_getdomaininfo;

166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184

/* 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 */
    uint32_t flags;	/* falgs, see before */
    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;

185 186 187
union xen_getdomaininfo {
    struct xen_v0_getdomaininfo v0;
    struct xen_v2_getdomaininfo v2;
188
    struct xen_v2d5_getdomaininfo v2d5;
189 190 191 192 193 194
};
typedef union xen_getdomaininfo xen_getdomaininfo;

union xen_getdomaininfolist {
    struct xen_v0_getdomaininfo *v0;
    struct xen_v2_getdomaininfo *v2;
195
    struct xen_v2d5_getdomaininfo *v2d5;
196 197 198
};
typedef union xen_getdomaininfolist xen_getdomaininfolist;

199 200 201 202 203 204 205 206 207 208 209 210

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;

211 212 213 214 215 216 217 218 219
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;

220

221 222 223
#define XEN_GETDOMAININFOLIST_ALLOC(domlist, size)                      \
    (hypervisor_version < 2 ?                                           \
     ((domlist.v0 = malloc(sizeof(xen_v0_getdomaininfo)*(size))) != NULL) : \
224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240
     (dom_interface_version < 5 ?                                       \
      ((domlist.v2 = malloc(sizeof(xen_v2_getdomaininfo)*(size))) != NULL) : \
      ((domlist.v2d5 = malloc(sizeof(xen_v2d5_getdomaininfo)*(size))) != NULL)))

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

#define XEN_GETDOMAININFOLIST_CLEAR(domlist, size)                     \
    (hypervisor_version < 2 ?                                          \
     memset(domlist.v0, 0, sizeof(xen_v0_getdomaininfo) * size) :      \
     (dom_interface_version < 5 ?                                      \
      memset(domlist.v2, 0, sizeof(xen_v2_getdomaininfo) * size) :     \
      memset(domlist.v2d5, 0, sizeof(xen_v2d5_getdomaininfo) * size)))
241 242 243 244

#define XEN_GETDOMAININFOLIST_DOMAIN(domlist, n)    \
    (hypervisor_version < 2 ?                       \
     domlist.v0[n].domain :                         \
245 246 247 248
     (dom_interface_version < 5 ?                   \
      domlist.v2[n].domain :                        \
      domlist.v2d5[n].domain))

249 250 251 252 253 254 255
#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))

256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275
#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))))
276 277 278 279

#define XEN_GETDOMAININFO_DOMAIN(dominfo)       \
    (hypervisor_version < 2 ?                   \
     dominfo.v0.domain :                        \
280 281 282
     (dom_interface_version < 5 ?               \
      dominfo.v2.domain :                       \
      dominfo.v2d5.domain))
283 284 285 286

#define XEN_GETDOMAININFO_CPUTIME(dominfo)      \
    (hypervisor_version < 2 ?                   \
     dominfo.v0.cpu_time :                      \
287 288 289
     (dom_interface_version < 5 ?               \
      dominfo.v2.cpu_time :                     \
      dominfo.v2d5.cpu_time))
290 291 292 293

#define XEN_GETDOMAININFO_CPUCOUNT(dominfo)     \
    (hypervisor_version < 2 ?                   \
     dominfo.v0.nr_online_vcpus :               \
294 295 296
     (dom_interface_version < 5 ?               \
      dominfo.v2.nr_online_vcpus :              \
      dominfo.v2d5.nr_online_vcpus))
297

298 299 300
#define XEN_GETDOMAININFO_MAXCPUID(dominfo)  \
    (hypervisor_version < 2 ?                   \
     dominfo.v0.max_vcpu_id :                   \
301 302 303
     (dom_interface_version < 5 ?               \
      dominfo.v2.max_vcpu_id :                  \
      dominfo.v2d5.max_vcpu_id))
304

305 306 307
#define XEN_GETDOMAININFO_FLAGS(dominfo)        \
    (hypervisor_version < 2 ?                   \
     dominfo.v0.flags :                         \
308 309 310
     (dom_interface_version < 5 ?               \
      dominfo.v2.flags :                        \
      dominfo.v2d5.flags))
311 312 313 314

#define XEN_GETDOMAININFO_TOT_PAGES(dominfo)    \
    (hypervisor_version < 2 ?                   \
     dominfo.v0.tot_pages :                     \
315 316 317
     (dom_interface_version < 5 ?               \
      dominfo.v2.tot_pages :                    \
      dominfo.v2d5.tot_pages))
318 319 320 321

#define XEN_GETDOMAININFO_MAX_PAGES(dominfo)    \
    (hypervisor_version < 2 ?                   \
     dominfo.v0.max_pages :                     \
322 323 324
     (dom_interface_version < 5 ?               \
      dominfo.v2.max_pages :                    \
      dominfo.v2d5.max_pages))
325

326 327 328 329 330 331 332
#define XEN_GETDOMAININFO_UUID(dominfo)         \
    (hypervisor_version < 2 ?                   \
     dominfo.v0.handle :                        \
     (dom_interface_version < 5 ?               \
      dominfo.v2.handle :                       \
      dominfo.v2d5.handle))

333

334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353
static int
lock_pages(void *addr, size_t len)
{
#ifdef __linux__
	return (mlock(addr, len));
#elif define(__sun)
	return (0);
#endif
}

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

354 355

struct xen_v0_getdomaininfolistop {
356 357 358 359 360
    domid_t   first_domain;
    uint32_t  max_domains;
    struct xen_v0_getdomaininfo *buffer;
    uint32_t  num_domains;
};
361 362 363 364 365 366 367 368 369 370 371
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;

372 373 374 375
/* 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;
376 377 378 379 380 381
#ifdef __BIG_ENDIAN__
    struct {
        int __pad[(sizeof (long long) - sizeof (struct xen_v2d5_getdomaininfo *)) / sizeof (int)];
        struct xen_v2d5_getdomaininfo *v;
    } buffer;
#else
382 383 384 385
    union {
        struct xen_v2d5_getdomaininfo *v;
        uint64_t pad ALIGN_64;
    } buffer;
386
#endif
387 388 389 390
    uint32_t  num_domains;
};
typedef struct xen_v2s3_getdomaininfolistop xen_v2s3_getdomaininfolistop;

391

392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423

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

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

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

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

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

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;

438 439 440 441 442
struct xen_v2d5_setmaxmem {
    uint64_t	maxmem ALIGN_64;
};
typedef struct xen_v2d5_setmaxmem xen_v2d5_setmaxmem;

443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488
/*
 * The informations for an setmaxvcpu system hypercall
 */
#define XEN_V0_OP_SETMAXVCPU	41
#define XEN_V1_OP_SETMAXVCPU	41
#define XEN_V2_OP_SETMAXVCPU	15

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

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

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

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

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

489 490
/* HV version 2, Dom version 5 requires 64-bit alignment */
struct xen_v2d5_cpumap {
491 492 493 494 495 496
#ifdef __BIG_ENDIAN__
    struct {
        int __pad[(sizeof (long long) - sizeof (uint8_t *)) / sizeof (int)];
        uint8_t *v;
    } bitmap;
#else
497 498 499 500
    union {
        uint8_t    *v;
        uint64_t   pad ALIGN_64;
    } bitmap;
501
#endif
502 503 504 505 506 507 508 509
    uint32_t    nr_cpus;
};
struct xen_v2d5_setvcpumap {
    uint32_t	vcpu;
    struct xen_v2d5_cpumap cpumap;
};
typedef struct xen_v2d5_setvcpumap xen_v2d5_setvcpumap;

510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539
/*
 * The informations for an vcpuinfo system hypercall
 */
#define XEN_V0_OP_GETVCPUINFO   43
#define XEN_V1_OP_GETVCPUINFO	43
#define XEN_V2_OP_GETVCPUINFO   14

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

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

540 541 542 543 544 545 546 547 548 549
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;

550 551 552 553 554
/*
 * 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;
555
typedef struct xen_v2d5_setvcpumap xen_v2d5_getvcpumap;
556

557 558 559 560 561
/*
 * from V2 we get the scheduler information
 */
#define XEN_V2_OP_GETSCHEDULERID	4

562 563 564 565 566
/*
 * from V2 we get the available heap information
 */
#define XEN_V2_OP_GETAVAILHEAP  	9

567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598
/*
 * 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;


599 600 601 602 603 604 605 606 607
/*
 * 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 {
608 609 610 611 612 613 614
        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];
615 616 617 618 619 620 621 622 623 624
    } 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 {
625 626
        xen_v2_getdomaininfolistop   getdomaininfolist;
        xen_v2s3_getdomaininfolistop getdomaininfolists3;
627
        xen_v2_getschedulerid        getschedulerid;
628
        xen_v2s4_availheap           availheap;
629
        uint8_t padding[128];
630 631 632 633 634 635 636 637 638 639
    } 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 {
640
        xen_v2_setmaxmem         setmaxmem;
641
        xen_v2d5_setmaxmem       setmaxmemd5;
642 643
        xen_v2_setmaxvcpu        setmaxvcpu;
        xen_v2_setvcpumap        setvcpumap;
644
        xen_v2d5_setvcpumap      setvcpumapd5;
645
        xen_v2_vcpuinfo          getvcpuinfo;
646
        xen_v2d5_vcpuinfo        getvcpuinfod5;
647
        xen_v2_getvcpumap        getvcpumap;
648
        xen_v2d5_getvcpumap      getvcpumapd5;
649 650
        xen_v2_setschedinfo      setschedinfo;
        xen_v2_getschedinfo      getschedinfo;
651
        uint8_t padding[128];
652 653 654
    } u;
};
typedef struct xen_op_v2_dom xen_op_v2_dom;
655 656

#include "internal.h"
657
#include "driver.h"
658
#include "xen_unified.h"
659 660
#include "xen_internal.h"

661 662 663 664 665 666 667 668 669 670 671
#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
672

673
#ifndef PROXY
674
static unsigned long xenHypervisorGetMaxMemory(virDomainPtr domain);
675
#endif
676

677
#ifndef PROXY
678
struct xenUnifiedDriver xenHypervisorDriver = {
679 680 681
    xenHypervisorOpen, /* open */
    xenHypervisorClose, /* close */
    xenHypervisorGetVersion, /* version */
682 683
    NULL, /* hostname */
    NULL, /* URI */
684
    NULL, /* nodeGetInfo */
685
    xenHypervisorGetCapabilities, /* getCapabilities */
686 687
    xenHypervisorListDomains, /* listDomains */
    xenHypervisorNumOfDomains, /* numOfDomains */
688 689 690 691
    NULL, /* domainCreateLinux */
    xenHypervisorPauseDomain, /* domainSuspend */
    xenHypervisorResumeDomain, /* domainResume */
    NULL, /* domainShutdown */
692
    NULL, /* domainReboot */
693
    xenHypervisorDestroyDomain, /* domainDestroy */
694
    xenHypervisorDomainGetOSType, /* domainGetOSType */
695
    xenHypervisorGetMaxMemory, /* domainGetMaxMemory */
696
    xenHypervisorSetMaxMemory, /* domainSetMaxMemory */
697
    NULL, /* domainSetMemory */
698 699
    xenHypervisorGetDomainInfo, /* domainGetInfo */
    NULL, /* domainSave */
700
    NULL, /* domainRestore */
D
Daniel Veillard 已提交
701
    NULL, /* domainCoreDump */
702 703
    xenHypervisorSetVcpus, /* domainSetVcpus */
    xenHypervisorPinVcpu, /* domainPinVcpu */
704
    xenHypervisorGetVcpus, /* domainGetVcpus */
705
    xenHypervisorGetVcpuMax, /* domainGetMaxVcpus */
706 707 708 709 710
    NULL, /* listDefinedDomains */
    NULL, /* numOfDefinedDomains */
    NULL, /* domainCreate */
    NULL, /* domainDefineXML */
    NULL, /* domainUndefine */
711 712
    NULL, /* domainAttachDevice */
    NULL, /* domainDetachDevice */
713 714
    NULL, /* domainGetAutostart */
    NULL, /* domainSetAutostart */
715 716 717
    xenHypervisorGetSchedulerType, /* domainGetSchedulerType */
    xenHypervisorGetSchedulerParameters, /* domainGetSchedulerParameters */
    xenHypervisorSetSchedulerParameters, /* domainSetSchedulerParameters */
718
};
719
#endif /* !PROXY */
720

721 722
/**
 * virXenError:
723
 * @conn: connection, if known
724 725
 * @error: the error number
 * @info: extra information string
726
 * @value: extra information number
727 728 729 730
 *
 * Handle an error at the xend daemon interface
 */
static void
731 732
virXenError(virConnectPtr conn,
            virErrorNumber error, const char *info, int value)
733
{
734
    const char *errmsg;
735

736
    if ((error == VIR_ERR_OK) || (in_init != 0))
737 738 739
        return;

    errmsg = __virErrorMsg(error, info);
740
    __virRaiseError(conn, NULL, NULL, VIR_FROM_XEN, error, VIR_ERR_ERROR,
741
                    errmsg, info, NULL, value, 0, errmsg, info, value);
742 743
}

744 745
#ifndef PROXY

746 747
/**
 * virXenErrorFunc:
748
 * @conn: connection, if known
749 750 751 752 753 754 755 756
 * @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
757 758
virXenErrorFunc(virConnectPtr conn,
                virErrorNumber error, const char *func, const char *info,
759 760
                int value)
{
D
Daniel Veillard 已提交
761
    char fullinfo[1000];
762 763 764 765 766 767 768 769 770 771
    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);
	fullinfo[999] = 0;
772
	__virRaiseError(conn, NULL, NULL, VIR_FROM_XEN, error, VIR_ERR_ERROR,
773 774
			errmsg, fullinfo, NULL, value, 0, errmsg, fullinfo,
			value);
775
    } else {
776
	__virRaiseError(conn, NULL, NULL, VIR_FROM_XEN, error, VIR_ERR_ERROR,
777 778
			errmsg, info, NULL, value, 0, errmsg, info,
			value);
779 780 781
    }
}

782 783
#endif /* PROXY */

784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805
/**
 * 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));
806 807
}

808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828
/**
 * xenHypervisorDoV0Op:
 * @handle: the handle to the Xen hypervisor
 * @op: pointer to the hyperviros operation structure
 *
 * Do an hypervisor operation though the old interface,
 * this leads to an hypervisor call through ioctl.
 *
 * Returns 0 in case of success and -1 in case of error.
 */
static int
xenHypervisorDoV0Op(int handle, xen_op_v0 * op)
{
    int ret;
    v0_hypercall_t hc;

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

829
    if (lock_pages(op, sizeof(dom0_op_t)) < 0) {
830
        virXenError(NULL, VIR_ERR_XEN_CALL, " locking", sizeof(*op));
831 832 833 834 835
        return (-1);
    }

    ret = ioctl(handle, xen_ioctl_hypercall_cmd, (unsigned long) &hc);
    if (ret < 0) {
836
        virXenError(NULL, VIR_ERR_XEN_CALL, " ioctl ", xen_ioctl_hypercall_cmd);
837 838
    }

839
    if (unlock_pages(op, sizeof(dom0_op_t)) < 0) {
840
        virXenError(NULL, VIR_ERR_XEN_CALL, " releasing", sizeof(*op));
841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869
        ret = -1;
    }

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

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

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

870
    if (lock_pages(op, sizeof(dom0_op_t)) < 0) {
871
        virXenError(NULL, VIR_ERR_XEN_CALL, " locking", sizeof(*op));
872 873 874 875 876
        return (-1);
    }

    ret = ioctl(handle, xen_ioctl_hypercall_cmd, (unsigned long) &hc);
    if (ret < 0) {
877
        virXenError(NULL, VIR_ERR_XEN_CALL, " ioctl ", xen_ioctl_hypercall_cmd);
878 879
    }

880
    if (unlock_pages(op, sizeof(dom0_op_t)) < 0) {
881
        virXenError(NULL, VIR_ERR_XEN_CALL, " releasing", sizeof(*op));
882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911
        ret = -1;
    }

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

    return (0);
}

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

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

912
    if (lock_pages(op, sizeof(dom0_op_t)) < 0) {
913
        virXenError(NULL, VIR_ERR_XEN_CALL, " locking", sizeof(*op));
914 915 916 917 918
        return (-1);
    }

    ret = ioctl(handle, xen_ioctl_hypercall_cmd, (unsigned long) &hc);
    if (ret < 0) {
919
        virXenError(NULL, VIR_ERR_XEN_CALL, " sys ioctl ", xen_ioctl_hypercall_cmd);
920 921
    }

922
    if (unlock_pages(op, sizeof(dom0_op_t)) < 0) {
923
        virXenError(NULL, VIR_ERR_XEN_CALL, " releasing", sizeof(*op));
924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953
        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;

954
    if (lock_pages(op, sizeof(dom0_op_t)) < 0) {
955
        virXenError(NULL, VIR_ERR_XEN_CALL, " locking", sizeof(*op));
956 957 958 959 960
        return (-1);
    }

    ret = ioctl(handle, xen_ioctl_hypercall_cmd, (unsigned long) &hc);
    if (ret < 0) {
961
        virXenError(NULL, VIR_ERR_XEN_CALL, " ioctl ", xen_ioctl_hypercall_cmd);
962 963
    }

964
    if (unlock_pages(op, sizeof(dom0_op_t)) < 0) {
965
        virXenError(NULL, VIR_ERR_XEN_CALL, " releasing", sizeof(*op));
966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987
        ret = -1;
    }

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

    return (0);
}

/**
 * virXen_getdomaininfolist:
 * @handle: the hypervisor handle
 * @first_domain: first domain in the range
 * @maxids: maximum number of domains to list
 * @dominfos: output structures
 *
 * Do a low level hypercall to list existing domains informations
 *
 * Returns the number of domains or -1 in case of failure
 */
static int
virXen_getdomaininfolist(int handle, int first_domain, int maxids,
988
                         xen_getdomaininfolist *dominfos)
989 990 991
{
    int ret = -1;

992
    if (lock_pages(XEN_GETDOMAININFOLIST_DATA(dominfos),
993
              XEN_GETDOMAININFO_SIZE * maxids) < 0) {
994
        virXenError(NULL, VIR_ERR_XEN_CALL, " locking",
995
                    XEN_GETDOMAININFO_SIZE * maxids);
996 997 998 999 1000 1001
        return (-1);
    }
    if (hypervisor_version > 1) {
        xen_op_v2_sys op;

        memset(&op, 0, sizeof(op));
1002
        op.cmd = XEN_V2_OP_GETDOMAININFOLIST;
1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014

        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;
        }
1015
        ret = xenHypervisorDoV2Sys(handle, &op);
1016 1017 1018 1019 1020 1021 1022

        if (ret == 0) {
            if (sys_interface_version < 3)
                ret = op.u.getdomaininfolist.num_domains;
            else
                ret = op.u.getdomaininfolists3.num_domains;
        }
1023 1024 1025 1026
    } else if (hypervisor_version == 1) {
        xen_op_v1 op;

        memset(&op, 0, sizeof(op));
1027 1028 1029 1030 1031 1032 1033 1034
        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;
1035 1036 1037 1038
    } else if (hypervisor_version == 0) {
        xen_op_v0 op;

        memset(&op, 0, sizeof(op));
1039 1040 1041 1042 1043 1044 1045 1046
        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;
1047
    }
1048
    if (unlock_pages(XEN_GETDOMAININFOLIST_DATA(dominfos),
1049
                XEN_GETDOMAININFO_SIZE * maxids) < 0) {
1050
        virXenError(NULL, VIR_ERR_XEN_CALL, " release",
1051
                    XEN_GETDOMAININFO_SIZE * maxids);
1052 1053 1054 1055 1056
        ret = -1;
    }
    return(ret);
}

1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071
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);
}


1072
#ifndef PROXY
1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088
/**
 * 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)) {
1089
        virXenErrorFunc(NULL, VIR_ERR_INTERNAL_ERROR, __FUNCTION__,
1090 1091 1092 1093 1094 1095
			"domain or conn is NULL", 0);
        return NULL;
    }

    priv = (xenUnifiedPrivatePtr) domain->conn->privateData;
    if (priv->handle < 0 || domain->id < 0) {
1096
        virXenErrorFunc(domain->conn, VIR_ERR_INTERNAL_ERROR, __FUNCTION__,
1097 1098 1099 1100 1101 1102 1103
			"priv->handle or domain->id invalid", 0);
        return NULL;
    }

    /*
     * Support only dom_interface_version >=5
     * (Xen3.1.0 or later)
1104
     * TODO: check on Xen 3.0.3
1105 1106
     */
    if (dom_interface_version < 5) {
1107
        virXenErrorFunc(domain->conn, VIR_ERR_NO_XEN, __FUNCTION__,
1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140
		        "unsupported in dom interface < 5", 0);
        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)
	    return(NULL);

        switch (op.u.getschedulerid.sched_id){
	    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;
        }
    }

    return schedulertype;
}

1141 1142 1143
static const char *str_weight = "weight";
static const char *str_cap = "cap";

1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163
/**
 * 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,
				 virSchedParameterPtr params, int *nparams)
{
    xenUnifiedPrivatePtr priv;

    if ((domain == NULL) || (domain->conn == NULL)) {
1164
        virXenErrorFunc(NULL, VIR_ERR_INTERNAL_ERROR, __FUNCTION__,
1165 1166 1167 1168 1169 1170
			"domain or conn is NULL", 0);
        return -1;
    }

    priv = (xenUnifiedPrivatePtr) domain->conn->privateData;
    if (priv->handle < 0 || domain->id < 0) {
1171
        virXenErrorFunc(domain->conn, VIR_ERR_INTERNAL_ERROR, __FUNCTION__,
1172 1173 1174 1175 1176 1177 1178 1179 1180 1181
			"priv->handle or domain->id invalid", 0);
        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) {
1182
        virXenErrorFunc(domain->conn, VIR_ERR_NO_XEN, __FUNCTION__,
1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214
			"unsupported in dom interface < 5", 0);
        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)
	    return -1;

        switch (op_sys.u.getschedulerid.sched_id){
	    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);

1215 1216
		strncpy (params[0].field, str_weight, VIR_DOMAIN_SCHED_FIELD_LENGTH);
        params[0].field[VIR_DOMAIN_SCHED_FIELD_LENGTH-1] = '\0';
1217 1218 1219
		params[0].type = VIR_DOMAIN_SCHED_FIELD_UINT;
		params[0].value.ui = op_dom.u.getschedinfo.u.credit.weight;

1220 1221
		strncpy (params[1].field, str_cap, VIR_DOMAIN_SCHED_FIELD_LENGTH);
        params[1].field[VIR_DOMAIN_SCHED_FIELD_LENGTH-1] = '\0';
1222 1223 1224 1225 1226 1227
		params[1].type = VIR_DOMAIN_SCHED_FIELD_UINT;
		params[1].value.ui = op_dom.u.getschedinfo.u.credit.cap;

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

    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,
				 virSchedParameterPtr params, int nparams)
{
    int i;
1251
    unsigned int val;
1252
    xenUnifiedPrivatePtr priv;
1253
    char buf[256];
1254 1255

    if ((domain == NULL) || (domain->conn == NULL)) {
1256
        virXenErrorFunc (NULL, VIR_ERR_INTERNAL_ERROR, __FUNCTION__,
1257 1258 1259 1260 1261
	                 "domain or conn is NULL", 0);
        return -1;
    }

    if ((nparams == 0) || (params == NULL)) {
1262
        virXenErrorFunc (domain->conn, VIR_ERR_INVALID_ARG, __FUNCTION__,
1263 1264 1265 1266 1267 1268
			 "Noparameters given", 0);
	return(-1);
    }

    priv = (xenUnifiedPrivatePtr) domain->conn->privateData;
    if (priv->handle < 0 || domain->id < 0) {
1269
        virXenErrorFunc (domain->conn, VIR_ERR_INTERNAL_ERROR, __FUNCTION__,
1270 1271 1272 1273 1274 1275 1276 1277 1278 1279
	                 "priv->handle or domain->id invalid", 0);
        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) {
1280
        virXenErrorFunc(domain->conn, VIR_ERR_NO_XEN, __FUNCTION__,
D
Daniel Veillard 已提交
1281
			"unsupported in dom interface < 5", 0);
1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310
        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
	    return(-1);
        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;

            /*
1311 1312
             * credit scheduler parameters
             * following values do not change the parameters
1313 1314 1315 1316 1317
             */
            op_dom.u.getschedinfo.u.credit.weight = 0;
            op_dom.u.getschedinfo.u.credit.cap    = (uint16_t)~0U;

            for (i = 0; i < nparams; i++) {
1318
                memset(&buf, 0, sizeof(buf));
1319
                if (STREQ (params[i].field, str_weight) &&
1320
                    params[i].type == VIR_DOMAIN_SCHED_FIELD_UINT) {
1321 1322
		    val = params[i].value.ui;
		    if ((val < 1) || (val > USHRT_MAX)) {
1323
                        snprintf(buf, sizeof(buf), _("Credit scheduler weight parameter (%d) is out of range (1-65535)"), val);
1324
                        virXenErrorFunc (domain->conn, VIR_ERR_INVALID_ARG, __FUNCTION__, buf, val);
1325 1326 1327
			return(-1);
		    }
                    op_dom.u.getschedinfo.u.credit.weight = val;
1328
		    weight_set = 1;
1329
		} else if (STREQ (params[i].field, str_cap) &&
1330
                    params[i].type == VIR_DOMAIN_SCHED_FIELD_UINT) {
1331 1332
		    val = params[i].value.ui;
		    if (val > USHRT_MAX) {
1333
                        snprintf(buf, sizeof(buf), _("Credit scheduler cap parameter (%d) is out of range (0-65535)"), val);
1334
                        virXenErrorFunc (domain->conn, VIR_ERR_INVALID_ARG, __FUNCTION__, buf, val);
1335 1336 1337
			return(-1);
		    }
                    op_dom.u.getschedinfo.u.credit.cap = val;
1338 1339
		    cap_set = 1;
	        } else {
1340 1341
                    virXenErrorFunc (domain->conn, VIR_ERR_INVALID_ARG, __FUNCTION__,
                                     "Credit scheduler accepts 'cap' and 'weight' integer parameters",
1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352
				     0);
		    return(-1);
		}
            }

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

1362 1363 1364 1365 1366 1367

int
xenHypervisorDomainBlockStats (virDomainPtr dom,
                               const char *path,
                               struct _virDomainBlockStats *stats)
{
1368 1369
#ifdef __linux__
    xenUnifiedPrivatePtr priv;
1370

1371 1372 1373
    priv = (xenUnifiedPrivatePtr) dom->conn->privateData;
    return xenLinuxDomainBlockStats (priv, dom, path, stats);
#else
1374
    virXenErrorFunc (dom->conn, VIR_ERR_NO_SUPPORT, __FUNCTION__,
1375 1376 1377 1378
                     "block statistics not supported on this platform",
                     dom->id);
    return -1;
#endif
1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392
}

/* 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)
{
1393
#ifdef __linux__
1394 1395
    int rqdomid, device;

1396 1397 1398
    /* Verify that the vif requested is one belonging to the current
     * domain.
     */
1399
    if (sscanf (path, "vif%d.%d", &rqdomid, &device) != 2) {
1400
        virXenErrorFunc (dom->conn, VIR_ERR_INVALID_ARG, __FUNCTION__,
1401 1402 1403 1404
                         "invalid path, should be vif<domid>.<n>.", 0);
        return -1;
    }
    if (rqdomid != dom->id) {
1405
        virXenErrorFunc (dom->conn, VIR_ERR_INVALID_ARG, __FUNCTION__,
1406 1407 1408 1409
                         "invalid path, vif<domid> should match this domain ID", 0);
        return -1;
    }

1410 1411
    return linuxDomainInterfaceStats (dom->conn, path, stats);
#else
1412
    virXenErrorFunc (dom->conn, VIR_ERR_NO_SUPPORT, __FUNCTION__,
1413 1414
                     "/proc/net/dev: Interface not found", 0);
    return -1;
1415
#endif
1416 1417
}

1418 1419 1420 1421 1422 1423 1424 1425 1426 1427
/**
 * 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
1428
virXen_pausedomain(int handle, int id)
1429 1430 1431 1432 1433 1434 1435
{
    int ret = -1;

    if (hypervisor_version > 1) {
        xen_op_v2_dom op;

        memset(&op, 0, sizeof(op));
1436 1437 1438
        op.cmd = XEN_V2_OP_PAUSEDOMAIN;
        op.domain = (domid_t) id;
        ret = xenHypervisorDoV2Dom(handle, &op);
1439 1440 1441 1442
    } else if (hypervisor_version == 1) {
        xen_op_v1 op;

        memset(&op, 0, sizeof(op));
1443 1444 1445
        op.cmd = XEN_V1_OP_PAUSEDOMAIN;
        op.u.domain.domain = (domid_t) id;
        ret = xenHypervisorDoV1Op(handle, &op);
1446 1447 1448 1449
    } else if (hypervisor_version == 0) {
        xen_op_v0 op;

        memset(&op, 0, sizeof(op));
1450 1451 1452
        op.cmd = XEN_V0_OP_PAUSEDOMAIN;
        op.u.domain.domain = (domid_t) id;
        ret = xenHypervisorDoV0Op(handle, &op);
1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466
    }
    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
1467
virXen_unpausedomain(int handle, int id)
1468 1469 1470 1471 1472 1473 1474
{
    int ret = -1;

    if (hypervisor_version > 1) {
        xen_op_v2_dom op;

        memset(&op, 0, sizeof(op));
1475 1476 1477
        op.cmd = XEN_V2_OP_UNPAUSEDOMAIN;
        op.domain = (domid_t) id;
        ret = xenHypervisorDoV2Dom(handle, &op);
1478 1479 1480 1481
    } else if (hypervisor_version == 1) {
        xen_op_v1 op;

        memset(&op, 0, sizeof(op));
1482 1483 1484
        op.cmd = XEN_V1_OP_UNPAUSEDOMAIN;
        op.u.domain.domain = (domid_t) id;
        ret = xenHypervisorDoV1Op(handle, &op);
1485 1486 1487 1488
    } else if (hypervisor_version == 0) {
        xen_op_v0 op;

        memset(&op, 0, sizeof(op));
1489 1490 1491
        op.cmd = XEN_V0_OP_UNPAUSEDOMAIN;
        op.u.domain.domain = (domid_t) id;
        ret = xenHypervisorDoV0Op(handle, &op);
1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505
    }
    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
1506
virXen_destroydomain(int handle, int id)
1507 1508 1509 1510 1511 1512 1513
{
    int ret = -1;

    if (hypervisor_version > 1) {
        xen_op_v2_dom op;

        memset(&op, 0, sizeof(op));
1514 1515 1516
        op.cmd = XEN_V2_OP_DESTROYDOMAIN;
        op.domain = (domid_t) id;
        ret = xenHypervisorDoV2Dom(handle, &op);
1517 1518 1519 1520
    } else if (hypervisor_version == 1) {
        xen_op_v1 op;

        memset(&op, 0, sizeof(op));
1521 1522 1523
        op.cmd = XEN_V1_OP_DESTROYDOMAIN;
        op.u.domain.domain = (domid_t) id;
        ret = xenHypervisorDoV1Op(handle, &op);
1524 1525 1526 1527
    } else if (hypervisor_version == 0) {
        xen_op_v0 op;

        memset(&op, 0, sizeof(op));
1528 1529 1530
        op.cmd = XEN_V0_OP_DESTROYDOMAIN;
        op.u.domain.domain = (domid_t) id;
        ret = xenHypervisorDoV0Op(handle, &op);
1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545
    }
    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
1546
virXen_setmaxmem(int handle, int id, unsigned long memory)
1547 1548 1549 1550 1551 1552 1553
{
    int ret = -1;

    if (hypervisor_version > 1) {
        xen_op_v2_dom op;

        memset(&op, 0, sizeof(op));
1554 1555
        op.cmd = XEN_V2_OP_SETMAXMEM;
        op.domain = (domid_t) id;
1556 1557 1558 1559
        if (dom_interface_version < 5)
            op.u.setmaxmem.maxmem = memory;
        else
            op.u.setmaxmemd5.maxmem = memory;
1560
        ret = xenHypervisorDoV2Dom(handle, &op);
1561 1562 1563 1564
    } else if (hypervisor_version == 1) {
        xen_op_v1 op;

        memset(&op, 0, sizeof(op));
1565 1566 1567 1568
        op.cmd = XEN_V1_OP_SETMAXMEM;
        op.u.setmaxmem.domain = (domid_t) id;
        op.u.setmaxmem.maxmem = memory;
        ret = xenHypervisorDoV1Op(handle, &op);
1569
    } else if (hypervisor_version == 0) {
1570
        xen_op_v0 op;
1571 1572

        memset(&op, 0, sizeof(op));
1573 1574 1575 1576
        op.cmd = XEN_V0_OP_SETMAXMEM;
        op.u.setmaxmem.domain = (domid_t) id;
        op.u.setmaxmem.maxmem = memory;
        ret = xenHypervisorDoV0Op(handle, &op);
1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591
    }
    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
1592
virXen_setmaxvcpus(int handle, int id, unsigned int vcpus)
1593 1594 1595 1596 1597 1598 1599
{
    int ret = -1;

    if (hypervisor_version > 1) {
        xen_op_v2_dom op;

        memset(&op, 0, sizeof(op));
1600 1601 1602 1603
        op.cmd = XEN_V2_OP_SETMAXVCPU;
        op.domain = (domid_t) id;
        op.u.setmaxvcpu.maxvcpu = vcpus;
        ret = xenHypervisorDoV2Dom(handle, &op);
1604 1605 1606 1607
    } else if (hypervisor_version == 1) {
        xen_op_v1 op;

        memset(&op, 0, sizeof(op));
1608 1609 1610 1611
        op.cmd = XEN_V1_OP_SETMAXVCPU;
        op.u.setmaxvcpu.domain = (domid_t) id;
        op.u.setmaxvcpu.maxvcpu = vcpus;
        ret = xenHypervisorDoV1Op(handle, &op);
1612
    } else if (hypervisor_version == 0) {
1613
        xen_op_v0 op;
1614 1615

        memset(&op, 0, sizeof(op));
1616 1617 1618 1619
        op.cmd = XEN_V0_OP_SETMAXVCPU;
        op.u.setmaxvcpu.domain = (domid_t) id;
        op.u.setmaxvcpu.maxvcpu = vcpus;
        ret = xenHypervisorDoV0Op(handle, &op);
1620 1621 1622 1623 1624 1625 1626 1627 1628 1629
    }
    return(ret);
}

/**
 * virXen_setvcpumap:
 * @handle: the hypervisor handle
 * @id: the domain id
 * @vcpu: the vcpu to map
 * @cpumap: the bitmap for this vcpu
1630
 * @maplen: the size of the bitmap in bytes
1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644
 *
 * Do a low level hypercall to change the pinning for vcpu
 *
 * Returns 0 or -1 in case of failure
 */
static int
virXen_setvcpumap(int handle, int id, unsigned int vcpu,
                  unsigned char * cpumap, int maplen)
{
    int ret = -1;

    if (hypervisor_version > 1) {
        xen_op_v2_dom op;

1645
        if (lock_pages(cpumap, maplen) < 0) {
1646
            virXenError(NULL, VIR_ERR_XEN_CALL, " locking", maplen);
1647 1648
            return (-1);
        }
1649
        memset(&op, 0, sizeof(op));
1650 1651
        op.cmd = XEN_V2_OP_SETVCPUMAP;
        op.domain = (domid_t) id;
1652 1653 1654 1655 1656 1657 1658 1659 1660
        if (dom_interface_version < 5) {
            op.u.setvcpumap.vcpu = vcpu;
            op.u.setvcpumap.cpumap.bitmap = cpumap;
            op.u.setvcpumap.cpumap.nr_cpus = maplen * 8;
        } else {
            op.u.setvcpumapd5.vcpu = vcpu;
            op.u.setvcpumapd5.cpumap.bitmap.v = cpumap;
            op.u.setvcpumapd5.cpumap.nr_cpus = maplen * 8;
        }
1661
        ret = xenHypervisorDoV2Dom(handle, &op);
1662

1663
        if (unlock_pages(cpumap, maplen) < 0) {
1664
            virXenError(NULL, VIR_ERR_XEN_CALL, " release", maplen);
1665 1666
            ret = -1;
        }
1667
    } else {
1668 1669 1670
        cpumap_t xen_cpumap; /* limited to 64 CPUs in old hypervisors */
        uint64_t *pm = &xen_cpumap;
        int j;
1671

1672 1673
        if ((maplen > (int)sizeof(cpumap_t)) || (sizeof(cpumap_t) & 7))
            return (-1);
1674

1675 1676 1677
        memset(pm, 0, sizeof(cpumap_t));
        for (j = 0; j < maplen; j++)
            *(pm + (j / 8)) |= cpumap[j] << (8 * (j & 7));
1678 1679

        if (hypervisor_version == 1) {
1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697
            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);
        }
1698 1699 1700
    }
    return(ret);
}
1701 1702
#endif /* !PROXY*/

1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 1716
/**
 * 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,
1717
                    unsigned char *cpumap, int maplen)
1718 1719 1720 1721 1722 1723 1724
{
    int ret = -1;

    if (hypervisor_version > 1) {
        xen_op_v2_dom op;

        memset(&op, 0, sizeof(op));
1725 1726
        op.cmd = XEN_V2_OP_GETVCPUINFO;
        op.domain = (domid_t) id;
1727 1728 1729 1730
        if (dom_interface_version < 5)
            op.u.getvcpuinfo.vcpu = (uint16_t) vcpu;
        else
            op.u.getvcpuinfod5.vcpu = (uint16_t) vcpu;
1731
        ret = xenHypervisorDoV2Dom(handle, &op);
1732

1733 1734
        if (ret < 0)
            return(-1);
1735
        ipt->number = vcpu;
1736 1737 1738 1739 1740 1741 1742 1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756 1757
        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;
1758 1759
        }
        if ((cpumap != NULL) && (maplen > 0)) {
1760
            if (lock_pages(cpumap, maplen) < 0) {
1761
                virXenError(NULL, VIR_ERR_XEN_CALL, " locking", maplen);
1762 1763
                return (-1);
            }
1764
            memset(cpumap, 0, maplen);
1765 1766 1767
            memset(&op, 0, sizeof(op));
            op.cmd = XEN_V2_OP_GETVCPUMAP;
            op.domain = (domid_t) id;
1768 1769 1770 1771 1772 1773 1774 1775 1776
            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;
            }
1777
            ret = xenHypervisorDoV2Dom(handle, &op);
1778
            if (unlock_pages(cpumap, maplen) < 0) {
1779
                virXenError(NULL, VIR_ERR_XEN_CALL, " release", maplen);
1780 1781 1782
                ret = -1;
            }
        }
1783
    } else {
1784 1785 1786 1787 1788
        int mapl = maplen;
        int cpu;

        if (maplen > (int)sizeof(cpumap_t))
            mapl = (int)sizeof(cpumap_t);
1789 1790

        if (hypervisor_version == 1) {
1791 1792 1793 1794 1795 1796 1797 1798 1799
            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);
1800
            ipt->number = vcpu;
1801 1802 1803 1804 1805 1806 1807 1808 1809 1810 1811 1812 1813 1814 1815 1816 1817 1818 1819 1820 1821 1822 1823
            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);
1824
            ipt->number = vcpu;
1825 1826 1827 1828 1829 1830 1831 1832 1833 1834 1835 1836 1837 1838
            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);
                }
            }
        }
1839 1840 1841
    }
    return(ret);
}
1842

1843 1844 1845 1846 1847 1848
/**
 * xenHypervisorInit:
 *
 * Initialize the hypervisor layer. Try to detect the kind of interface
 * used i.e. pre or post changeset 10277
 */
1849
int
D
Daniel P. Berrange 已提交
1850
xenHypervisorInit(void)
1851
{
1852
    int fd, ret, cmd, errcode;
1853
    hypercall_t hc;
1854
    v0_hypercall_t v0_hc;
1855
    xen_getdomaininfo info;
D
Daniel Veillard 已提交
1856
    virVcpuInfoPtr ipt = NULL;
1857 1858

    if (initialized) {
1859
        if (hypervisor_version == -1)
1860
            return (-1);
1861
        return(0);
1862 1863
    }
    initialized = 1;
1864
    in_init = 1;
1865

1866 1867 1868 1869 1870 1871 1872 1873 1874
    /* 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);
1875
        virXenError (NULL, VIR_ERR_INTERNAL_ERROR, error, 0);
1876 1877 1878 1879 1880 1881 1882 1883 1884
        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);
1885
        virXenError (NULL, VIR_ERR_INTERNAL_ERROR, error, 0);
1886 1887 1888 1889 1890 1891 1892 1893 1894 1895
        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);
1896
        virXenError (NULL, VIR_ERR_INTERNAL_ERROR, error, 0);
1897 1898 1899 1900 1901
        in_init = 0;
        return -1;
    }

    /* Xen hypervisor version detection begins. */
1902 1903
    ret = open(XEN_HYPERVISOR_SOCKET, O_RDWR);
    if (ret < 0) {
1904
        hypervisor_version = -1;
1905
        return(-1);
1906 1907 1908
    }
    fd = ret;

1909 1910 1911 1912
    /*
     * The size of the hypervisor call block changed July 2006
     * this detect if we are using the new or old hypercall_t structure
     */
1913 1914 1915 1916 1917 1918 1919 1920
    hc.op = __HYPERVISOR_xen_version;
    hc.arg[0] = (unsigned long) XENVER_version;
    hc.arg[1] = 0;

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

    if ((ret != -1) && (ret != 0)) {
D
Daniel Veillard 已提交
1921
#ifdef DEBUG
1922
        fprintf(stderr, "Using new hypervisor call: %X\n", ret);
D
Daniel Veillard 已提交
1923
#endif
1924 1925 1926
        hv_version = ret;
        xen_ioctl_hypercall_cmd = cmd;
        goto detect_v2;
1927
    }
1928

1929
#ifndef __sun__
1930 1931 1932 1933 1934 1935 1936 1937
    /*
     * 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);
1938
    if ((ret != -1) && (ret != 0)) {
D
Daniel Veillard 已提交
1939
#ifdef DEBUG
1940
        fprintf(stderr, "Using old hypervisor call: %X\n", ret);
D
Daniel Veillard 已提交
1941
#endif
1942 1943
        hv_version = ret;
        xen_ioctl_hypercall_cmd = cmd;
1944
        hypervisor_version = 0;
1945
        goto done;
1946
    }
1947
#endif
1948

1949 1950 1951 1952 1953
    /*
     * we faild to make any hypercall
     */

    hypervisor_version = -1;
1954
    virXenError(NULL, VIR_ERR_XEN_CALL, " ioctl ", IOCTL_PRIVCMD_HYPERCALL);
1955
    close(fd);
1956 1957 1958
    in_init = 0;
    return(-1);

1959
 detect_v2:
1960 1961 1962 1963 1964 1965
    /*
     * 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;
1966 1967 1968 1969

    ipt = malloc(sizeof(virVcpuInfo));
    if (ipt == NULL){
#ifdef DEBUG
1970
        fprintf(stderr, "Memory allocation failed at xenHypervisorInit()\n");
1971 1972 1973
#endif
        return(-1);
    }
1974
    /* Currently consider RHEL5.0 Fedora7, xen-3.1, and xen-unstable */
1975
    sys_interface_version = 2; /* XEN_SYSCTL_INTERFACE_VERSION */
1976
    if (virXen_getdomaininfo(fd, 0, &info) == 1) {
1977 1978 1979
        /* RHEL 5.0 */
        dom_interface_version = 3; /* XEN_DOMCTL_INTERFACE_VERSION */
        if (virXen_getvcpusinfo(fd, 0, 0, ipt, NULL, 0) == 0){
D
Daniel Veillard 已提交
1980
#ifdef DEBUG
1981
            fprintf(stderr, "Using hypervisor call v2, sys ver2 dom ver3\n");
D
Daniel Veillard 已提交
1982
#endif
1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996
            goto done;
        }
        /* Fedora 7 */
        dom_interface_version = 4; /* XEN_DOMCTL_INTERFACE_VERSION */
        if (virXen_getvcpusinfo(fd, 0, 0, ipt, NULL, 0) == 0){
#ifdef DEBUG
            fprintf(stderr, "Using hypervisor call v2, sys ver2 dom ver4\n");
#endif
            goto done;
        }
    }

    sys_interface_version = 3; /* XEN_SYSCTL_INTERFACE_VERSION */
    if (virXen_getdomaininfo(fd, 0, &info) == 1) {
1997
        /* xen-3.1 */
1998 1999 2000 2001 2002 2003 2004
        dom_interface_version = 5; /* XEN_DOMCTL_INTERFACE_VERSION */
        if (virXen_getvcpusinfo(fd, 0, 0, ipt, NULL, 0) == 0){
#ifdef DEBUG
            fprintf(stderr, "Using hypervisor call v2, sys ver3 dom ver5\n");
#endif
            goto done;
        }
2005
    }
2006

2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018
    sys_interface_version = 4; /* XEN_SYSCTL_INTERFACE_VERSION */
    if (virXen_getdomaininfo(fd, 0, &info) == 1) {
        /* xen-unstable */
        dom_interface_version = 5; /* XEN_DOMCTL_INTERFACE_VERSION */
        if (virXen_getvcpusinfo(fd, 0, 0, ipt, NULL, 0) == 0){
#ifdef DEBUG
            fprintf(stderr, "Using hypervisor call v2, sys ver4 dom ver5\n");
#endif
            goto done;
        }
    }

2019 2020
    hypervisor_version = 1;
    sys_interface_version = -1;
2021
    if (virXen_getdomaininfo(fd, 0, &info) == 1) {
D
Daniel Veillard 已提交
2022
#ifdef DEBUG
2023
        fprintf(stderr, "Using hypervisor call v1\n");
D
Daniel Veillard 已提交
2024
#endif
2025
        goto done;
2026 2027 2028 2029 2030 2031 2032
    }

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

    hypervisor_version = -1;
2033
    virXenError(NULL, VIR_ERR_XEN_CALL, " ioctl ", IOCTL_PRIVCMD_HYPERCALL);
2034 2035
    close(fd);
    in_init = 0;
2036 2037
    if (ipt)
        free(ipt);
2038 2039
    return(-1);

2040
 done:
2041
    close(fd);
2042
    in_init = 0;
2043 2044
    if (ipt)
        free(ipt);
2045 2046 2047
    return(0);
}

2048 2049
/**
 * xenHypervisorOpen:
2050 2051 2052
 * @conn: pointer to the connection block
 * @name: URL for the target, NULL for local
 * @flags: combination of virDrvOpenFlag(s)
2053 2054 2055
 *
 * Connects to the Xen hypervisor.
 *
2056
 * Returns 0 or -1 in case of error.
2057
 */
2058
int
2059
xenHypervisorOpen(virConnectPtr conn,
2060
                  xmlURIPtr uri ATTRIBUTE_UNUSED,
2061
                  virConnectAuthPtr auth ATTRIBUTE_UNUSED,
2062
                  int flags ATTRIBUTE_UNUSED)
2063
{
2064
    int ret;
2065
    xenUnifiedPrivatePtr priv = (xenUnifiedPrivatePtr) conn->privateData;
2066

2067
    if (initialized == 0)
2068 2069
        if (xenHypervisorInit() == -1)
            return -1;
2070

2071
    priv->handle = -1;
2072

2073
    ret = open(XEN_HYPERVISOR_SOCKET, O_RDWR);
2074
    if (ret < 0) {
2075
        virXenError(conn, VIR_ERR_NO_XEN, XEN_HYPERVISOR_SOCKET, 0);
2076
        return (-1);
2077
    }
2078 2079

    priv->handle = ret;
2080

2081
    return(0);
2082 2083 2084 2085
}

/**
 * xenHypervisorClose:
2086
 * @conn: pointer to the connection block
2087 2088 2089 2090 2091
 *
 * Close the connection to the Xen hypervisor.
 *
 * Returns 0 in case of success or -1 in case of error.
 */
2092
int
2093
xenHypervisorClose(virConnectPtr conn)
2094
{
2095
    int ret;
2096
    xenUnifiedPrivatePtr priv;
2097

2098
    if (conn == NULL)
2099
        return (-1);
2100

2101 2102 2103 2104 2105 2106
    priv = (xenUnifiedPrivatePtr) conn->privateData;

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

    ret = close(priv->handle);
2107
    if (ret < 0)
2108
        return (-1);
2109

2110
    return (0);
2111 2112 2113
}


2114 2115
/**
 * xenHypervisorGetVersion:
2116 2117
 * @conn: pointer to the connection block
 * @hvVer: where to store the version
2118 2119 2120
 *
 * Call the hypervisor to extracts his own internal API version
 *
2121
 * Returns 0 in case of success, -1 in case of error
2122
 */
2123 2124
int
xenHypervisorGetVersion(virConnectPtr conn, unsigned long *hvVer)
2125
{
2126 2127 2128 2129 2130 2131
    xenUnifiedPrivatePtr priv;

    if (conn == NULL)
        return -1;
    priv = (xenUnifiedPrivatePtr) conn->privateData;
    if (priv->handle < 0 || hvVer == NULL)
2132
        return (-1);
2133
    *hvVer = (hv_version >> 16) * 1000000 + (hv_version & 0xFFFF) * 1000;
2134
    return(0);
2135 2136
}

2137 2138 2139
/**
 * xenHypervisorGetCapabilities:
 * @conn: pointer to the connection block
2140 2141
 * @cpuinfo: file handle containing /proc/cpuinfo data, or NULL
 * @capabilities: file handle containing /sys/hypervisor/properties/capabilities data, or NULL
2142 2143 2144 2145
 *
 * Return the capabilities of this hypervisor.
 */
char *
2146
xenHypervisorMakeCapabilitiesXML(virConnectPtr conn,
2147 2148
                                 const char *hostmachine,
                                 FILE *cpuinfo, FILE *capabilities)
2149 2150
{
    char line[1024], *str, *token;
2151
    regmatch_t subs[4];
2152
    char *saveptr = NULL;
2153
    int i, r, topology;
2154 2155 2156

    char hvm_type[4] = ""; /* "vmx" or "svm" (or "" if not in CPU). */
    int host_pae = 0;
2157
    struct guest_arch {
2158 2159 2160 2161
        const char *model;
        int bits;
        int hvm;
        int pae;
2162
        int nonpae;
2163
        int ia64_be;
2164
    } guest_archs[32];
2165 2166 2167 2168 2169
    int nr_guest_archs = 0;

    virBufferPtr xml;
    char *xml_str;

2170 2171 2172
    int hv_major = hv_version >> 16;
    int hv_minor = hv_version & 0xFFFF;

2173 2174
    memset(guest_archs, 0, sizeof(guest_archs));

2175 2176 2177 2178
    /* /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)
     */
2179 2180 2181 2182 2183 2184 2185 2186 2187 2188
    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;
        }
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
    }

    /* 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. */
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
    if (fgets (line, sizeof line, capabilities)) {
        /* Split the line into tokens.  strtok_r is OK here because we "own"
         * this buffer.  Parse out the features from each token.
         */
        for (str = line, nr_guest_archs = 0;
             nr_guest_archs < sizeof guest_archs / sizeof guest_archs[0]
                 && (token = strtok_r (str, " ", &saveptr)) != NULL;
             str = NULL) {

            if (regexec (&xen_cap_rec, token, sizeof subs / sizeof subs[0],
                         subs, 0) == 0) {
                int hvm = strncmp (&token[subs[1].rm_so], "hvm", 3) == 0;
                const char *model;
                int bits, pae = 0, nonpae = 0, ia64_be = 0;
                if (strncmp (&token[subs[2].rm_so], "x86_32", 6) == 0) {
                    model = "i686";
                    bits = 32;
                    if (strncmp (&token[subs[3].rm_so], "p", 1) == 0)
                        pae = 1;
                    else
                        nonpae = 1;
                }
                else if (strncmp (&token[subs[2].rm_so], "x86_64", 6) == 0) {
                    model = "x86_64";
                    bits = 64;
                }
                else if (strncmp (&token[subs[2].rm_so], "ia64", 4) == 0) {
                    model = "ia64";
                    bits = 64;
                    if (strncmp (&token[subs[3].rm_so], "be", 2) == 0)
                        ia64_be = 1;
                }
                else if (strncmp (&token[subs[2].rm_so], "powerpc64", 4) == 0) {
                    model = "ppc64";
                    bits = 64;
                } else {
                    /* XXX surely no other Xen archs exist  */
                    continue;
                }
2255

2256 2257 2258 2259 2260 2261 2262
                /* Search for existing matching (model,hvm) tuple */
                for (i = 0 ; i < nr_guest_archs ; i++) {
                    if (!strcmp(guest_archs[i].model, model) &&
                        guest_archs[i].hvm == hvm) {
                        break;
                    }
                }
2263

2264 2265 2266 2267 2268 2269 2270 2271 2272 2273 2274 2275 2276 2277 2278 2279 2280 2281 2282 2283 2284
                /* 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;
2285 2286 2287 2288 2289 2290
            }
        }
    }

    /* Construct the final XML. */
    xml = virBufferNew (1024);
2291
    if (!xml) {
2292
        virXenError(conn, VIR_ERR_NO_MEMORY, __FUNCTION__, 0);
2293 2294
        return NULL;
    }
2295 2296 2297 2298 2299 2300 2301
    r = virBufferVSprintf (xml,
                           "\
<capabilities>\n\
  <host>\n\
    <cpu>\n\
      <arch>%s</arch>\n\
      <features>\n",
2302 2303
                           hostmachine);
    if (r == -1) goto vir_buffer_failed;
2304 2305 2306 2307 2308 2309

    if (strcmp (hvm_type, "") != 0) {
        r = virBufferVSprintf (xml,
                               "\
        <%s/>\n",
                               hvm_type);
2310
        if (r == -1) goto vir_buffer_failed;
2311 2312 2313 2314 2315 2316 2317 2318 2319 2320
    }
    if (host_pae) {
        r = virBufferAdd (xml, "\
        <pae/>\n", -1);
        if (r == -1) goto vir_buffer_failed;
    }
    r = virBufferAdd (xml,
                      "\
      </features>\n\
    </cpu>\n\
2321 2322 2323 2324 2325 2326
    <migration_features>\n\
      <live/>\n\
      <uri_transports>\n\
        <uri_transport>xenmigr</uri_transport>\n\
      </uri_transports>\n\
    </migration_features>\n\
2327 2328 2329
  </host>\n", -1);
    if (r == -1) goto vir_buffer_failed;

2330 2331 2332 2333 2334 2335
    if (sys_interface_version >= 4) {
        topology = xenDaemonNodeGetTopology(conn, xml);
        if (topology != 0)
            goto topology_failed;
    }

2336 2337 2338 2339 2340 2341 2342 2343
    for (i = 0; i < nr_guest_archs; ++i) {
        r = virBufferVSprintf (xml,
                               "\
\n\
  <guest>\n\
    <os_type>%s</os_type>\n\
    <arch name=\"%s\">\n\
      <wordsize>%d</wordsize>\n\
2344
      <domain type=\"xen\"></domain>\n",
2345 2346
                               guest_archs[i].hvm ? "hvm" : "xen",
                               guest_archs[i].model,
2347
                               guest_archs[i].bits);
2348 2349
        if (r == -1) goto vir_buffer_failed;
        if (guest_archs[i].hvm) {
2350
            r = virBufferVSprintf (xml,
2351
                              "\
2352
      <emulator>/usr/lib%s/xen/bin/qemu-dm</emulator>\n\
2353 2354
      <machine>pc</machine>\n\
      <machine>isapc</machine>\n\
2355 2356
      <loader>/usr/lib/xen/boot/hvmloader</loader>\n",
                                   guest_archs[i].bits == 64 ? "64" : "");
2357 2358 2359 2360 2361
            if (r == -1) goto vir_buffer_failed;
        }
        r = virBufferAdd (xml,
                          "\
    </arch>\n\
2362
    <features>\n", -1);
2363 2364 2365 2366
        if (r == -1) goto vir_buffer_failed;
        if (guest_archs[i].pae) {
            r = virBufferAdd (xml,
                              "\
2367 2368 2369 2370
      <pae/>\n", -1);
            if (r == -1) goto vir_buffer_failed;
        }
        if (guest_archs[i].nonpae) {
2371
            r = virBufferAdd (xml, "      <nonpae/>\n", -1);
2372 2373 2374
            if (r == -1) goto vir_buffer_failed;
        }
        if (guest_archs[i].ia64_be) {
2375
            r = virBufferAdd (xml, "      <ia64_be/>\n", -1);
2376 2377
            if (r == -1) goto vir_buffer_failed;
        }
2378 2379 2380 2381 2382 2383 2384 2385 2386 2387 2388 2389 2390 2391 2392
        if (guest_archs[i].hvm) {
            r = virBufferAdd (xml, "      <acpi default='on' toggle='yes'/>\n",
	                      -1);
            if (r == -1) goto vir_buffer_failed;
            // In Xen 3.1.0, APIC is always on and can't be toggled
            if (hv_major >= 3 && hv_minor > 0) {
                r = virBufferAdd (xml,
		              "      <apic default='off' toggle='no'/>\n", -1);
            } else {
                r = virBufferAdd (xml,
		              "      <apic default='on' toggle='yes'/>\n", -1);
            }
            if (r == -1) goto vir_buffer_failed;
        }
        r = virBufferAdd (xml, "\
2393 2394 2395 2396
    </features>\n\
  </guest>\n", -1);
        if (r == -1) goto vir_buffer_failed;
    }
2397

2398 2399 2400 2401 2402
    r = virBufferAdd (xml,
                      "\
</capabilities>\n", -1);
    if (r == -1) goto vir_buffer_failed;
    xml_str = strdup (xml->content);
2403
    if (!xml_str) goto vir_buffer_failed;
2404 2405 2406
    virBufferFree (xml);

    return xml_str;
2407 2408

 vir_buffer_failed:
2409
    virXenError(conn, VIR_ERR_NO_MEMORY, __FUNCTION__, 0);
2410
 topology_failed:
2411 2412 2413 2414 2415 2416 2417 2418 2419 2420 2421 2422 2423 2424 2425 2426 2427 2428 2429 2430 2431 2432 2433 2434 2435 2436 2437 2438 2439 2440 2441 2442 2443 2444 2445 2446 2447 2448 2449 2450 2451 2452 2453 2454 2455
    virBufferFree (xml);
    return NULL;
}

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

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

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

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

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

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

    return xml;
2456 2457
}

2458 2459 2460 2461 2462 2463 2464 2465 2466 2467 2468
/**
 * 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)
{
2469
    xen_getdomaininfolist dominfos;
2470 2471 2472
    int ret, nbids;
    static int last_maxids = 2;
    int maxids = last_maxids;
2473
    xenUnifiedPrivatePtr priv;
2474

2475 2476 2477 2478
    if (conn == NULL)
        return -1;
    priv = (xenUnifiedPrivatePtr) conn->privateData;
    if (priv->handle < 0)
2479 2480
        return (-1);

2481 2482
 retry:
    if (!(XEN_GETDOMAININFOLIST_ALLOC(dominfos, maxids))) {
2483
        virXenError(conn, VIR_ERR_NO_MEMORY, _("allocating %d domain info"),
2484 2485
                    maxids);
        return(-1);
2486 2487
    }

2488 2489
    XEN_GETDOMAININFOLIST_CLEAR(dominfos, maxids);

2490
    ret = virXen_getdomaininfolist(priv->handle, 0, maxids, &dominfos);
2491

2492
    XEN_GETDOMAININFOLIST_FREE(dominfos);
2493 2494 2495 2496

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

2497
    nbids = ret;
2498 2499 2500 2501 2502
    /* 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 ;-)
     */
2503
    if (nbids == maxids) {
2504 2505 2506 2507 2508 2509
        if (maxids < 65000) {
            last_maxids *= 2;
            maxids *= 2;
            goto retry;
        }
        nbids = -1;
2510 2511 2512 2513 2514 2515 2516 2517 2518 2519 2520 2521 2522 2523 2524 2525 2526 2527 2528
    }
    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)
{
2529
    xen_getdomaininfolist dominfos;
2530
    int ret, nbids, i;
2531 2532 2533 2534
    xenUnifiedPrivatePtr priv;

    if (conn == NULL)
        return -1;
2535

2536 2537
    priv = (xenUnifiedPrivatePtr) conn->privateData;
    if (priv->handle < 0 ||
2538 2539 2540
        (ids == NULL) || (maxids < 1))
        return (-1);

2541
    if (!(XEN_GETDOMAININFOLIST_ALLOC(dominfos, maxids))) {
2542
        virXenError(conn, VIR_ERR_NO_MEMORY, "allocating %d domain info",
2543 2544
                    maxids);
        return(-1);
2545
    }
2546 2547

    XEN_GETDOMAININFOLIST_CLEAR(dominfos, maxids);
2548 2549
    memset(ids, 0, maxids * sizeof(int));

2550
    ret = virXen_getdomaininfolist(priv->handle, 0, maxids, &dominfos);
2551 2552

    if (ret < 0) {
2553
        XEN_GETDOMAININFOLIST_FREE(dominfos);
2554 2555 2556
        return (-1);
    }

2557
    nbids = ret;
2558
    if ((nbids < 0) || (nbids > maxids)) {
2559
        XEN_GETDOMAININFOLIST_FREE(dominfos);
2560 2561 2562 2563
        return(-1);
    }

    for (i = 0;i < nbids;i++) {
2564
        ids[i] = XEN_GETDOMAININFOLIST_DOMAIN(dominfos, i);
2565 2566
    }

2567
    XEN_GETDOMAININFOLIST_FREE(dominfos);
2568 2569 2570
    return (nbids);
}

2571 2572 2573 2574 2575 2576 2577 2578 2579 2580 2581 2582 2583 2584 2585 2586 2587 2588 2589 2590 2591 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

#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))) {
2650
        virXenError(conn, VIR_ERR_NO_MEMORY, "allocating %d domain info",
2651 2652 2653 2654 2655 2656 2657 2658 2659 2660 2661 2662 2663 2664 2665 2666 2667 2668 2669 2670 2671 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
                    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

2701
/**
2702
 * xenHypervisorGetMaxVcpus:
2703 2704 2705 2706
 *
 * Returns the maximum of CPU defined by Xen.
 */
int
2707 2708
xenHypervisorGetMaxVcpus(virConnectPtr conn,
                         const char *type ATTRIBUTE_UNUSED)
2709
{
2710 2711 2712 2713 2714 2715
    xenUnifiedPrivatePtr priv;

    if (conn == NULL)
        return -1;
    priv = (xenUnifiedPrivatePtr) conn->privateData;
    if (priv->handle < 0)
2716 2717 2718 2719 2720
        return (-1);

    return MAX_VIRT_CPUS;
}

2721
/**
2722 2723 2724
 * xenHypervisorGetDomMaxMemory:
 * @conn: connection data
 * @id: domain id
2725
 *
2726
 * Retrieve the maximum amount of physical memory allocated to a
2727
 * domain.
2728 2729 2730
 *
 * Returns the memory size in kilobytes or 0 in case of error.
 */
2731 2732
unsigned long
xenHypervisorGetDomMaxMemory(virConnectPtr conn, int id)
2733
{
2734
    xenUnifiedPrivatePtr priv;
2735
    xen_getdomaininfo dominfo;
2736 2737
    int ret;

2738 2739 2740 2741 2742 2743
    if (conn == NULL)
        return 0;

    priv = (xenUnifiedPrivatePtr) conn->privateData;
    if (priv->handle < 0)
        return 0;
2744

2745 2746 2747 2748 2749 2750
    if (kb_per_pages == 0) {
        kb_per_pages = sysconf(_SC_PAGESIZE) / 1024;
	if (kb_per_pages <= 0) 
	    kb_per_pages = 4;
    }

2751
    XEN_GETDOMAININFO_CLEAR(dominfo);
2752

2753
    ret = virXen_getdomaininfo(priv->handle, id, &dominfo);
2754

2755
    if ((ret < 0) || (XEN_GETDOMAININFO_DOMAIN(dominfo) != id))
2756 2757
        return (0);

2758
    return((unsigned long) XEN_GETDOMAININFO_MAX_PAGES(dominfo) * kb_per_pages);
2759 2760
}

2761
#ifndef PROXY
2762 2763 2764
/**
 * xenHypervisorGetMaxMemory:
 * @domain: a domain object or NULL
2765
 *
2766 2767 2768 2769 2770 2771 2772 2773 2774
 * 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)
{
2775 2776 2777 2778 2779 2780 2781
    xenUnifiedPrivatePtr priv;

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

    priv = (xenUnifiedPrivatePtr) domain->conn->privateData;
    if (priv->handle < 0 || domain->id < 0)
2782 2783
        return (0);

2784
    return(xenHypervisorGetDomMaxMemory(domain->conn, domain->id));
2785
}
2786
#endif
2787

2788
/**
2789 2790 2791
 * xenHypervisorGetDomInfo:
 * @conn: connection data
 * @id: the domain ID
2792
 * @info: the place where information should be stored
2793
 *
2794
 * Do an hypervisor call to get the related set of domain information.
2795 2796 2797 2798
 *
 * Returns 0 in case of success, -1 in case of error.
 */
int
2799
xenHypervisorGetDomInfo(virConnectPtr conn, int id, virDomainInfoPtr info)
2800
{
2801
    xenUnifiedPrivatePtr priv;
2802
    xen_getdomaininfo dominfo;
2803
    int ret;
2804
    uint32_t domain_flags, domain_state, domain_shutdown_cause;
2805 2806 2807 2808 2809 2810

    if (kb_per_pages == 0) {
        kb_per_pages = sysconf(_SC_PAGESIZE) / 1024;
	if (kb_per_pages <= 0) 
	    kb_per_pages = 4;
    }
2811

2812 2813 2814 2815 2816
    if (conn == NULL)
        return -1;

    priv = (xenUnifiedPrivatePtr) conn->privateData;
    if (priv->handle < 0 || info == NULL)
2817
        return (-1);
2818

2819
    memset(info, 0, sizeof(virDomainInfo));
2820
    XEN_GETDOMAININFO_CLEAR(dominfo);
2821

2822
    ret = virXen_getdomaininfo(priv->handle, id, &dominfo);
2823

2824
    if ((ret < 0) || (XEN_GETDOMAININFO_DOMAIN(dominfo) != id))
2825
        return (-1);
2826

2827
    domain_flags = XEN_GETDOMAININFO_FLAGS(dominfo);
2828 2829
    domain_flags &= ~DOMFLAGS_HVM; /* Mask out HVM flags */
    domain_state = domain_flags & 0xFF; /* Mask out high bits */
2830
    switch (domain_state) {
2831 2832 2833 2834
	case DOMFLAGS_DYING:
	    info->state = VIR_DOMAIN_SHUTDOWN;
	    break;
	case DOMFLAGS_SHUTDOWN:
2835 2836 2837 2838 2839 2840 2841 2842 2843
            /* 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;
            }
2844 2845 2846 2847 2848 2849 2850 2851 2852 2853 2854 2855 2856 2857 2858 2859 2860 2861 2862
	    break;
	case DOMFLAGS_PAUSED:
	    info->state = VIR_DOMAIN_PAUSED;
	    break;
	case DOMFLAGS_BLOCKED:
	    info->state = VIR_DOMAIN_BLOCKED;
	    break;
	case DOMFLAGS_RUNNING:
	    info->state = VIR_DOMAIN_RUNNING;
	    break;
	default:
	    info->state = VIR_DOMAIN_NONE;
    }

    /*
     * the API brings back the cpu time in nanoseconds,
     * convert to microseconds, same thing convert to
     * kilobytes from page counts
     */
2863
    info->cpuTime = XEN_GETDOMAININFO_CPUTIME(dominfo);
2864
    info->memory = XEN_GETDOMAININFO_TOT_PAGES(dominfo) * kb_per_pages;
2865 2866 2867
    info->maxMem = XEN_GETDOMAININFO_MAX_PAGES(dominfo);
    if(info->maxMem != UINT_MAX)
        info->maxMem *= kb_per_pages;
2868
    info->nrVirtCpu = XEN_GETDOMAININFO_CPUCOUNT(dominfo);
2869
    return (0);
2870 2871
}

2872 2873 2874
/**
 * xenHypervisorGetDomainInfo:
 * @domain: pointer to the domain block
2875
 * @info: the place where information should be stored
2876
 *
2877
 * Do an hypervisor call to get the related set of domain information.
2878 2879 2880 2881 2882 2883
 *
 * Returns 0 in case of success, -1 in case of error.
 */
int
xenHypervisorGetDomainInfo(virDomainPtr domain, virDomainInfoPtr info)
{
2884 2885 2886 2887 2888 2889 2890
    xenUnifiedPrivatePtr priv;

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

    priv = (xenUnifiedPrivatePtr) domain->conn->privateData;
    if (priv->handle < 0 || info == NULL ||
2891
        (domain->id < 0))
2892
        return (-1);
2893

2894
    return(xenHypervisorGetDomInfo(domain->conn, domain->id, info));
2895 2896 2897

}

2898
#ifndef PROXY
2899 2900 2901 2902 2903 2904 2905 2906 2907 2908 2909 2910 2911 2912 2913 2914 2915 2916 2917 2918 2919 2920 2921 2922
/**
 * xenHypervisorNodeGetCellsFreeMemory:
 * @conn: pointer to the hypervisor connection
 * @freeMems: pointer to the array of unsigned long long
 * @startCell: index of first cell to return freeMems info on. 
 * @maxCells: Maximum number of cells for which freeMems information can 
 *            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
 * 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, 
 * 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;
2923
    int nbNodeCells;
2924
    
2925
    if (conn == NULL) {
2926
        virXenErrorFunc (conn, VIR_ERR_INVALID_ARG, __FUNCTION__,
2927 2928 2929
                        "invalid argument", 0);
        return -1;
    }
2930

2931 2932
    nbNodeCells = xenNbCells(conn);
    if (nbNodeCells < 0) {
2933
        virXenErrorFunc (conn, VIR_ERR_XEN_CALL, __FUNCTION__,
2934 2935
			 "cannot determine actual number of cells",0);
	return(-1);
2936 2937
    }

2938
    if ((maxCells < 1) || (startCell >= nbNodeCells)) {
2939
        virXenErrorFunc (conn, VIR_ERR_INVALID_ARG, __FUNCTION__,
2940 2941 2942
                        "invalid argument", 0);
        return -1;
    }
2943

2944 2945 2946 2947
    /*
     * Support only sys_interface_version >=4
     */
    if (sys_interface_version < 4) {
2948
        virXenErrorFunc (conn, VIR_ERR_XEN_CALL, __FUNCTION__,
2949 2950 2951 2952 2953 2954
                        "unsupported in sys interface < 4", 0);
        return -1;
    }

    priv = (xenUnifiedPrivatePtr) conn->privateData;
    if (priv->handle < 0) {
2955
        virXenErrorFunc (conn, VIR_ERR_INTERNAL_ERROR, __FUNCTION__,
2956 2957 2958 2959 2960 2961 2962 2963 2964 2965 2966 2967 2968 2969 2970 2971 2972 2973 2974
                        "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);
}


2975 2976
/**
 * xenHypervisorPauseDomain:
2977
 * @domain: pointer to the domain block
2978 2979 2980 2981 2982 2983
 *
 * Do an hypervisor call to pause the given domain
 *
 * Returns 0 in case of success, -1 in case of error.
 */
int
2984
xenHypervisorPauseDomain(virDomainPtr domain)
2985
{
2986
    int ret;
2987
    xenUnifiedPrivatePtr priv;
2988

2989 2990 2991 2992 2993
    if ((domain == NULL) || (domain->conn == NULL))
        return -1;

    priv = (xenUnifiedPrivatePtr) domain->conn->privateData;
    if (priv->handle < 0 || domain->id < 0)
2994 2995
        return (-1);

2996
    ret = virXen_pausedomain(priv->handle, domain->id);
2997
    if (ret < 0)
2998 2999
        return (-1);
    return (0);
3000 3001 3002 3003
}

/**
 * xenHypervisorResumeDomain:
3004
 * @domain: pointer to the domain block
3005 3006 3007 3008 3009 3010
 *
 * Do an hypervisor call to resume the given domain
 *
 * Returns 0 in case of success, -1 in case of error.
 */
int
3011
xenHypervisorResumeDomain(virDomainPtr domain)
3012
{
3013
    int ret;
3014 3015 3016 3017
    xenUnifiedPrivatePtr priv;

    if ((domain == NULL) || (domain->conn == NULL))
        return -1;
3018

3019 3020
    priv = (xenUnifiedPrivatePtr) domain->conn->privateData;
    if (priv->handle < 0 || domain->id < 0)
3021 3022
        return (-1);

3023
    ret = virXen_unpausedomain(priv->handle, domain->id);
3024
    if (ret < 0)
3025 3026
        return (-1);
    return (0);
3027 3028 3029 3030
}

/**
 * xenHypervisorDestroyDomain:
3031
 * @domain: pointer to the domain block
3032 3033 3034 3035 3036 3037
 *
 * Do an hypervisor call to destroy the given domain
 *
 * Returns 0 in case of success, -1 in case of error.
 */
int
3038
xenHypervisorDestroyDomain(virDomainPtr domain)
3039
{
3040
    int ret;
3041 3042 3043 3044
    xenUnifiedPrivatePtr priv;

    if (domain == NULL || domain->conn == NULL)
        return -1;
3045

3046 3047
    priv = (xenUnifiedPrivatePtr) domain->conn->privateData;
    if (priv->handle < 0 || domain->id < 0)
3048 3049
        return (-1);

3050
    ret = virXen_destroydomain(priv->handle, domain->id);
3051
    if (ret < 0)
3052 3053
        return (-1);
    return (0);
3054 3055
}

3056 3057
/**
 * xenHypervisorSetMaxMemory:
3058
 * @domain: pointer to the domain block
3059 3060 3061 3062 3063 3064 3065
 * @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
3066
xenHypervisorSetMaxMemory(virDomainPtr domain, unsigned long memory)
3067
{
3068
    int ret;
3069
    xenUnifiedPrivatePtr priv;
3070

3071 3072 3073 3074 3075
    if (domain == NULL || domain->conn == NULL)
        return -1;

    priv = (xenUnifiedPrivatePtr) domain->conn->privateData;
    if (priv->handle < 0 || domain->id < 0)
3076 3077
        return (-1);

3078
    ret = virXen_setmaxmem(priv->handle, domain->id, memory);
3079
    if (ret < 0)
3080 3081
        return (-1);
    return (0);
3082
}
3083
#endif /* PROXY */
3084

3085
#ifndef PROXY
3086 3087 3088 3089 3090 3091 3092 3093 3094 3095 3096 3097 3098
/**
 * 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)
{
3099
    int ret;
3100 3101 3102 3103
    xenUnifiedPrivatePtr priv;

    if (domain == NULL || domain->conn == NULL)
        return -1;
3104

3105 3106
    priv = (xenUnifiedPrivatePtr) domain->conn->privateData;
    if (priv->handle < 0 || domain->id < 0 || nvcpus < 1)
3107
        return (-1);
3108

3109
    ret = virXen_setmaxvcpus(priv->handle, domain->id, nvcpus);
3110
    if (ret < 0)
3111
        return (-1);
3112
    return (0);
3113 3114 3115 3116 3117 3118 3119 3120
}

/**
 * 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
3121
 *
3122 3123 3124 3125 3126 3127 3128 3129 3130
 * 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)
{
3131
    int ret;
3132
    xenUnifiedPrivatePtr priv;
3133

3134 3135 3136 3137 3138
    if (domain == NULL || domain->conn == NULL)
        return -1;

    priv = (xenUnifiedPrivatePtr) domain->conn->privateData;
    if (priv->handle < 0 || (domain->id < 0) ||
3139
        (cpumap == NULL) || (maplen < 1))
3140 3141
        return (-1);

3142
    ret = virXen_setvcpumap(priv->handle, domain->id, vcpu,
3143 3144
                            cpumap, maplen);
    if (ret < 0)
3145
        return (-1);
3146
    return (0);
3147
}
3148
#endif
3149 3150 3151 3152 3153 3154 3155 3156 3157 3158 3159 3160 3161 3162

/**
 * virDomainGetVcpus:
 * @domain: pointer to domain object, or NULL for Domain0
 * @info: pointer to an array of virVcpuInfo structures (OUT)
 * @maxinfo: number of structures in info array
 * @cpumaps: pointer to an bit map of real CPUs for all vcpus of this domain (in 8-bit bytes) (OUT)
 *	If cpumaps is NULL, then no cupmap information is returned by the API.
 *	It's assumed there is <maxinfo> cpumap in cpumaps array.
 *	The memory allocated to cpumaps must be (maxinfo * maplen) bytes
 *	(ie: calloc(maxinfo, maplen)).
 *	One cpumap inside cpumaps has the format described in virDomainPinVcpu() API.
 * @maplen: number of bytes in one cpumap, from 1 up to size of CPU map in
 *	underlying virtualization system (Xen...).
3163
 *
3164 3165 3166 3167 3168
 * Extract information about virtual CPUs of domain, store it in info array
 * and also in cpumaps if this pointer is'nt NULL.
 *
 * Returns the number of info filled in case of success, -1 in case of failure.
 */
3169
#ifndef PROXY
3170 3171
int
xenHypervisorGetVcpus(virDomainPtr domain, virVcpuInfoPtr info, int maxinfo,
3172
                      unsigned char *cpumaps, int maplen)
3173
{
3174
    xen_getdomaininfo dominfo;
3175
    int ret;
3176
    xenUnifiedPrivatePtr priv;
3177
    virVcpuInfoPtr ipt;
3178
    int nbinfo, i;
3179

3180 3181 3182 3183 3184
    if (domain == NULL || domain->conn == NULL)
        return -1;

    priv = (xenUnifiedPrivatePtr) domain->conn->privateData;
    if (priv->handle < 0 || (domain->id < 0) ||
3185 3186
        (info == NULL) || (maxinfo < 1) ||
        (sizeof(cpumap_t) & 7))
3187
        return (-1);
3188
    if ((cpumaps != NULL) && (maplen < 1))
3189
        return -1;
3190 3191

    /* first get the number of virtual CPUs in this domain */
3192
    XEN_GETDOMAININFO_CLEAR(dominfo);
3193
    ret = virXen_getdomaininfo(priv->handle, domain->id,
3194
                               &dominfo);
3195

3196
    if ((ret < 0) || (XEN_GETDOMAININFO_DOMAIN(dominfo) != domain->id))
3197
        return (-1);
3198
    nbinfo = XEN_GETDOMAININFO_CPUCOUNT(dominfo) + 1;
3199 3200 3201
    if (nbinfo > maxinfo) nbinfo = maxinfo;

    if (cpumaps != NULL)
3202
        memset(cpumaps, 0, maxinfo * maplen);
3203

3204 3205
    for (i = 0, ipt = info; i < nbinfo; i++, ipt++) {
        if ((cpumaps != NULL) && (i < maxinfo)) {
3206
            ret = virXen_getvcpusinfo(priv->handle, domain->id, i,
3207 3208 3209 3210 3211 3212
                                      ipt,
                                      (unsigned char *)VIR_GET_CPUMAP(cpumaps, maplen, i),
                                      maplen);
            if (ret < 0)
                return(-1);
        } else {
3213
            ret = virXen_getvcpusinfo(priv->handle, domain->id, i,
3214 3215 3216 3217
                                      ipt, NULL, 0);
            if (ret < 0)
                return(-1);
        }
3218 3219 3220
    }
    return nbinfo;
}
3221
#endif /* PROXY */
3222

3223 3224 3225 3226 3227 3228 3229 3230 3231 3232 3233 3234 3235 3236
/**
 * 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;
3237 3238 3239 3240
    xenUnifiedPrivatePtr priv;

    if (domain == NULL || domain->conn == NULL)
        return -1;
3241

3242 3243
    priv = (xenUnifiedPrivatePtr) domain->conn->privateData;
    if (priv->handle < 0)
3244 3245 3246 3247 3248 3249 3250
        return (-1);

    /* inactive domain */
    if (domain->id < 0) {
        maxcpu = MAX_VIRT_CPUS;
    } else {
        XEN_GETDOMAININFO_CLEAR(dominfo);
3251
        ret = virXen_getdomaininfo(priv->handle, domain->id,
3252 3253 3254 3255 3256 3257 3258 3259 3260 3261
                                   &dominfo);

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

    return maxcpu;
}

3262
#endif /* WITH_XEN */
3263 3264 3265 3266 3267
/*
 * vim: set tabstop=4:
 * vim: set shiftwidth=4:
 * vim: set expandtab:
 */
3268 3269 3270 3271 3272 3273 3274 3275
/*
 * Local variables:
 *  indent-tabs-mode: nil
 *  c-indent-level: 4
 *  c-basic-offset: 4
 *  tab-width: 4
 * End:
 */