drivers/block/xen-blkback/common.h

   1 /*
   2  * This program is free software; you can redistribute it and/or
   3  * modify it under the terms of the GNU General Public License version 2
   4  * as published by the Free Software Foundation; or, when distributed
   5  * separately from the Linux kernel or incorporated into other
   6  * software packages, subject to the following license:
   7  *
   8  * Permission is hereby granted, free of charge, to any person obtaining a copy
   9  * of this source file (the "Software"), to deal in the Software without
  10  * restriction, including without limitation the rights to use, copy, modify,
  11  * merge, publish, distribute, sublicense, and/or sell copies of the Software,
  12  * and to permit persons to whom the Software is furnished to do so, subject to
  13  * the following conditions:
  14  *
  15  * The above copyright notice and this permission notice shall be included in
  16  * all copies or substantial portions of the Software.
  17  *
  18  * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
  19  * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
  20  * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
  21  * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
  22  * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
  23  * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
  24  * IN THE SOFTWARE.
  25  */
  26
  27 #ifndef __XEN_BLKIF__BACKEND__COMMON_H__
  28 #define __XEN_BLKIF__BACKEND__COMMON_H__
  29
  30 #include <linux/module.h>
  31 #include <linux/interrupt.h>
  32 #include <linux/slab.h>
  33 #include <linux/blkdev.h>
  34 #include <linux/vmalloc.h>
  35 #include <linux/wait.h>
  36 #include <linux/io.h>
  37 #include <linux/rbtree.h>
  38 #include <asm/setup.h>
  39 #include <asm/pgalloc.h>
  40 #include <asm/hypervisor.h>
  41 #include <xen/grant_table.h>
  42 #include <xen/xenbus.h>
  43 #include <xen/interface/io/ring.h>
  44 #include <xen/interface/io/blkif.h>
  45 #include <xen/interface/io/protocols.h>
  46
  47 extern unsigned int xen_blkif_max_ring_order;
  48 /*
  49  * This is the maximum number of segments that would be allowed in indirect
  50  * requests. This value will also be passed to the frontend.
  51  */
  52 #define MAX_INDIRECT_SEGMENTS 256
  53
  54 #define SEGS_PER_INDIRECT_FRAME \
  55         (PAGE_SIZE/sizeof(struct blkif_request_segment))
  56 #define MAX_INDIRECT_PAGES \
  57         ((MAX_INDIRECT_SEGMENTS + SEGS_PER_INDIRECT_FRAME - 1)/SEGS_PER_INDIRECT_FRAME)
  58 #define INDIRECT_PAGES(_segs) \
  59         ((_segs + SEGS_PER_INDIRECT_FRAME - 1)/SEGS_PER_INDIRECT_FRAME)
  60
  61 /* Not a real protocol.  Used to generate ring structs which contain
  62  * the elements common to all protocols only.  This way we get a
  63  * compiler-checkable way to use common struct elements, so we can
  64  * avoid using switch(protocol) in a number of places.  */
  65 struct blkif_common_request {
  66         char dummy;
  67 };
  68 struct blkif_common_response {
  69         char dummy;
  70 };
  71
  72 struct blkif_x86_32_request_rw {
  73         uint8_t        nr_segments;  /* number of segments                   */
  74         blkif_vdev_t   handle;       /* only for read/write requests         */
  75         uint64_t       id;           /* private guest value, echoed in resp  */
  76         blkif_sector_t sector_number;/* start sector idx on disk (r/w only)  */
  77         struct blkif_request_segment seg[BLKIF_MAX_SEGMENTS_PER_REQUEST];
  78 } __attribute__((__packed__));
  79
  80 struct blkif_x86_32_request_discard {
  81         uint8_t        flag;         /* BLKIF_DISCARD_SECURE or zero         */
  82         blkif_vdev_t   _pad1;        /* was "handle" for read/write requests */
  83         uint64_t       id;           /* private guest value, echoed in resp  */
  84         blkif_sector_t sector_number;/* start sector idx on disk (r/w only)  */
  85         uint64_t       nr_sectors;
  86 } __attribute__((__packed__));
  87
  88 struct blkif_x86_32_request_other {
  89         uint8_t        _pad1;
  90         blkif_vdev_t   _pad2;
  91         uint64_t       id;           /* private guest value, echoed in resp  */
  92 } __attribute__((__packed__));
  93
  94 struct blkif_x86_32_request_indirect {
  95         uint8_t        indirect_op;
  96         uint16_t       nr_segments;
  97         uint64_t       id;
  98         blkif_sector_t sector_number;
  99         blkif_vdev_t   handle;
 100         uint16_t       _pad1;
 101         grant_ref_t    indirect_grefs[BLKIF_MAX_INDIRECT_PAGES_PER_REQUEST];
 102         /*
 103          * The maximum number of indirect segments (and pages) that will
 104          * be used is determined by MAX_INDIRECT_SEGMENTS, this value
 105          * is also exported to the guest (via xenstore
 106          * feature-max-indirect-segments entry), so the frontend knows how
 107          * many indirect segments the backend supports.
 108          */
 109         uint64_t       _pad2;        /* make it 64 byte aligned */
 110 } __attribute__((__packed__));
 111
 112 struct blkif_x86_32_request {
 113         uint8_t        operation;    /* BLKIF_OP_???                         */
 114         union {
 115                 struct blkif_x86_32_request_rw rw;
 116                 struct blkif_x86_32_request_discard discard;
 117                 struct blkif_x86_32_request_other other;
 118                 struct blkif_x86_32_request_indirect indirect;
 119         } u;
 120 } __attribute__((__packed__));
 121
 122 /* i386 protocol version */
 123 #pragma pack(push, 4)
 124 struct blkif_x86_32_response {
 125         uint64_t        id;              /* copied from request */
 126         uint8_t         operation;       /* copied from request */
 127         int16_t         status;          /* BLKIF_RSP_???       */
 128 };
 129 #pragma pack(pop)
 130 /* x86_64 protocol version */
 131
 132 struct blkif_x86_64_request_rw {
 133         uint8_t        nr_segments;  /* number of segments                   */
 134         blkif_vdev_t   handle;       /* only for read/write requests         */
 135         uint32_t       _pad1;        /* offsetof(blkif_reqest..,u.rw.id)==8  */
 136         uint64_t       id;
 137         blkif_sector_t sector_number;/* start sector idx on disk (r/w only)  */
 138         struct blkif_request_segment seg[BLKIF_MAX_SEGMENTS_PER_REQUEST];
 139 } __attribute__((__packed__));
 140
 141 struct blkif_x86_64_request_discard {
 142         uint8_t        flag;         /* BLKIF_DISCARD_SECURE or zero         */
 143         blkif_vdev_t   _pad1;        /* was "handle" for read/write requests */
 144         uint32_t       _pad2;        /* offsetof(blkif_..,u.discard.id)==8   */
 145         uint64_t       id;
 146         blkif_sector_t sector_number;/* start sector idx on disk (r/w only)  */
 147         uint64_t       nr_sectors;
 148 } __attribute__((__packed__));
 149
 150 struct blkif_x86_64_request_other {
 151         uint8_t        _pad1;
 152         blkif_vdev_t   _pad2;
 153         uint32_t       _pad3;        /* offsetof(blkif_..,u.discard.id)==8   */
 154         uint64_t       id;           /* private guest value, echoed in resp  */
 155 } __attribute__((__packed__));
 156
 157 struct blkif_x86_64_request_indirect {
 158         uint8_t        indirect_op;
 159         uint16_t       nr_segments;
 160         uint32_t       _pad1;        /* offsetof(blkif_..,u.indirect.id)==8   */
 161         uint64_t       id;
 162         blkif_sector_t sector_number;
 163         blkif_vdev_t   handle;
 164         uint16_t       _pad2;
 165         grant_ref_t    indirect_grefs[BLKIF_MAX_INDIRECT_PAGES_PER_REQUEST];
 166         /*
 167          * The maximum number of indirect segments (and pages) that will
 168          * be used is determined by MAX_INDIRECT_SEGMENTS, this value
 169          * is also exported to the guest (via xenstore
 170          * feature-max-indirect-segments entry), so the frontend knows how
 171          * many indirect segments the backend supports.
 172          */
 173         uint32_t       _pad3;        /* make it 64 byte aligned */
 174 } __attribute__((__packed__));
 175
 176 struct blkif_x86_64_request {
 177         uint8_t        operation;    /* BLKIF_OP_???                         */
 178         union {
 179                 struct blkif_x86_64_request_rw rw;
 180                 struct blkif_x86_64_request_discard discard;
 181                 struct blkif_x86_64_request_other other;
 182                 struct blkif_x86_64_request_indirect indirect;
 183         } u;
 184 } __attribute__((__packed__));
 185
 186 struct blkif_x86_64_response {
 187         uint64_t       __attribute__((__aligned__(8))) id;
 188         uint8_t         operation;       /* copied from request */
 189         int16_t         status;          /* BLKIF_RSP_???       */
 190 };
 191
 192 DEFINE_RING_TYPES(blkif_common, struct blkif_common_request,
 193                   struct blkif_common_response);
 194 DEFINE_RING_TYPES(blkif_x86_32, struct blkif_x86_32_request,
 195                   struct blkif_x86_32_response);
 196 DEFINE_RING_TYPES(blkif_x86_64, struct blkif_x86_64_request,
 197                   struct blkif_x86_64_response);
 198
 199 union blkif_back_rings {
 200         struct blkif_back_ring        native;
 201         struct blkif_common_back_ring common;
 202         struct blkif_x86_32_back_ring x86_32;
 203         struct blkif_x86_64_back_ring x86_64;
 204 };
 205
 206 enum blkif_protocol {
 207         BLKIF_PROTOCOL_NATIVE = 1,
 208         BLKIF_PROTOCOL_X86_32 = 2,
 209         BLKIF_PROTOCOL_X86_64 = 3,
 210 };
 211
 212 /*
 213  * Default protocol if the frontend doesn't specify one.
 214  */
 215 #ifdef CONFIG_X86
 216 #  define BLKIF_PROTOCOL_DEFAULT BLKIF_PROTOCOL_X86_32
 217 #else
 218 #  define BLKIF_PROTOCOL_DEFAULT BLKIF_PROTOCOL_NATIVE
 219 #endif
 220
 221 struct xen_vbd {
 222         /* What the domain refers to this vbd as. */
 223         blkif_vdev_t            handle;
 224         /* Non-zero -> read-only */
 225         unsigned char           readonly;
 226         /* VDISK_xxx */
 227         unsigned char           type;
 228         /* phys device that this vbd maps to. */
 229         u32                     pdevice;
 230         struct block_device     *bdev;
 231         /* Cached size parameter. */
 232         sector_t                size;
 233         unsigned int            flush_support:1;
 234         unsigned int            discard_secure:1;
 235         unsigned int            feature_gnt_persistent:1;
 236         unsigned int            overflow_max_grants:1;
 237 };
 238
 239 struct backend_info;
 240
 241 /* Number of available flags */
 242 #define PERSISTENT_GNT_FLAGS_SIZE       2
 243 /* This persistent grant is currently in use */
 244 #define PERSISTENT_GNT_ACTIVE           0
 245 /*
 246  * This persistent grant has been used, this flag is set when we remove the
 247  * PERSISTENT_GNT_ACTIVE, to know that this grant has been used recently.
 248  */
 249 #define PERSISTENT_GNT_WAS_ACTIVE       1
 250
 251 /* Number of requests that we can fit in a ring */
 252 #define XEN_BLKIF_REQS_PER_PAGE         32
 253
 254 struct persistent_gnt {
 255         struct page *page;
 256         grant_ref_t gnt;
 257         grant_handle_t handle;
 258         DECLARE_BITMAP(flags, PERSISTENT_GNT_FLAGS_SIZE);
 259         struct rb_node node;
 260         struct list_head remove_node;
 261 };
 262
 263 struct xen_blkif {
 264         /* Unique identifier for this interface. */
 265         domid_t                 domid;
 266         unsigned int            handle;
 267         /* Physical parameters of the comms window. */
 268         unsigned int            irq;
 269         /* Comms information. */
 270         enum blkif_protocol     blk_protocol;
 271         union blkif_back_rings  blk_rings;
 272         void                    *blk_ring;
 273         /* The VBD attached to this interface. */
 274         struct xen_vbd          vbd;
 275         /* Back pointer to the backend_info. */
 276         struct backend_info     *be;
 277         /* Private fields. */
 278         spinlock_t              blk_ring_lock;
 279         atomic_t                refcnt;
 280
 281         wait_queue_head_t       wq;
 282         /* for barrier (drain) requests */
 283         struct completion       drain_complete;
 284         atomic_t                drain;
 285         atomic_t                inflight;
 286         /* One thread per one blkif. */
 287         struct task_struct      *xenblkd;
 288         unsigned int            waiting_reqs;
 289
 290         /* tree to store persistent grants */
 291         struct rb_root          persistent_gnts;
 292         unsigned int            persistent_gnt_c;
 293         atomic_t                persistent_gnt_in_use;
 294         unsigned long           next_lru;
 295
 296         /* used by the kworker that offload work from the persistent purge */
 297         struct list_head        persistent_purge_list;
 298         struct work_struct      persistent_purge_work;
 299
 300         /* buffer of free pages to map grant refs */
 301         spinlock_t              free_pages_lock;
 302         int                     free_pages_num;
 303         struct list_head        free_pages;
 304
 305         /* List of all 'pending_req' available */
 306         struct list_head        pending_free;
 307         /* And its spinlock. */
 308         spinlock_t              pending_free_lock;
 309         wait_queue_head_t       pending_free_wq;
 310
 311         /* statistics */
 312         unsigned long           st_print;
 313         unsigned long long                      st_rd_req;
 314         unsigned long long                      st_wr_req;
 315         unsigned long long                      st_oo_req;
 316         unsigned long long                      st_f_req;
 317         unsigned long long                      st_ds_req;
 318         unsigned long long                      st_rd_sect;
 319         unsigned long long                      st_wr_sect;
 320
 321         struct work_struct      free_work;
 322         /* Thread shutdown wait queue. */
 323         wait_queue_head_t       shutdown_wq;
 324         unsigned int nr_ring_pages;
 325 };
 326
 327 struct seg_buf {
 328         unsigned long offset;
 329         unsigned int nsec;
 330 };
 331
 332 struct grant_page {
 333         struct page             *page;
 334         struct persistent_gnt   *persistent_gnt;
 335         grant_handle_t          handle;
 336         grant_ref_t             gref;
 337 };
 338
 339 /*
 340  * Each outstanding request that we've passed to the lower device layers has a
 341  * 'pending_req' allocated to it. Each buffer_head that completes decrements
 342  * the pendcnt towards zero. When it hits zero, the specified domain has a
 343  * response queued for it, with the saved 'id' passed back.
 344  */
 345 struct pending_req {
 346         struct xen_blkif        *blkif;
 347         u64                     id;
 348         int                     nr_segs;
 349         atomic_t                pendcnt;
 350         unsigned short          operation;
 351         int                     status;
 352         struct list_head        free_list;
 353         struct grant_page       *segments[MAX_INDIRECT_SEGMENTS];
 354         /* Indirect descriptors */
 355         struct grant_page       *indirect_pages[MAX_INDIRECT_PAGES];
 356         struct seg_buf          seg[MAX_INDIRECT_SEGMENTS];
 357         struct bio              *biolist[MAX_INDIRECT_SEGMENTS];
 358         struct gnttab_unmap_grant_ref unmap[MAX_INDIRECT_SEGMENTS];
 359         struct page                   *unmap_pages[MAX_INDIRECT_SEGMENTS];
 360         struct gntab_unmap_queue_data gnttab_unmap_data;
 361 };
 362
 363
 364 #define vbd_sz(_v)      ((_v)->bdev->bd_part ? \
 365                          (_v)->bdev->bd_part->nr_sects : \
 366                           get_capacity((_v)->bdev->bd_disk))
 367
 368 #define xen_blkif_get(_b) (atomic_inc(&(_b)->refcnt))
 369 #define xen_blkif_put(_b)                               \
 370         do {                                            \
 371                 if (atomic_dec_and_test(&(_b)->refcnt)) \
 372                         schedule_work(&(_b)->free_work);\
 373         } while (0)
 374
 375 struct phys_req {
 376         unsigned short          dev;
 377         blkif_sector_t          nr_sects;
 378         struct block_device     *bdev;
 379         blkif_sector_t          sector_number;
 380 };
 381 int xen_blkif_interface_init(void);
 382
 383 int xen_blkif_xenbus_init(void);
 384
 385 irqreturn_t xen_blkif_be_int(int irq, void *dev_id);
 386 int xen_blkif_schedule(void *arg);
 387 int xen_blkif_purge_persistent(void *arg);
 388 void xen_blkbk_free_caches(struct xen_blkif *blkif);
 389
 390 int xen_blkbk_flush_diskcache(struct xenbus_transaction xbt,
 391                               struct backend_info *be, int state);
 392
 393 int xen_blkbk_barrier(struct xenbus_transaction xbt,
 394                       struct backend_info *be, int state);
 395 struct xenbus_device *xen_blkbk_xenbus(struct backend_info *be);
 396 void xen_blkbk_unmap_purged_grants(struct work_struct *work);
 397
 398 static inline void blkif_get_x86_32_req(struct blkif_request *dst,
 399                                         struct blkif_x86_32_request *src)
 400 {
 401         int i, n = BLKIF_MAX_SEGMENTS_PER_REQUEST, j;
 402         dst->operation = src->operation;
 403         switch (src->operation) {
 404         case BLKIF_OP_READ:
 405         case BLKIF_OP_WRITE:
 406         case BLKIF_OP_WRITE_BARRIER:
 407         case BLKIF_OP_FLUSH_DISKCACHE:
 408                 dst->u.rw.nr_segments = src->u.rw.nr_segments;
 409                 dst->u.rw.handle = src->u.rw.handle;
 410                 dst->u.rw.id = src->u.rw.id;
 411                 dst->u.rw.sector_number = src->u.rw.sector_number;
 412                 barrier();
 413                 if (n > dst->u.rw.nr_segments)
 414                         n = dst->u.rw.nr_segments;
 415                 for (i = 0; i < n; i++)
 416                         dst->u.rw.seg[i] = src->u.rw.seg[i];
 417                 break;
 418         case BLKIF_OP_DISCARD:
 419                 dst->u.discard.flag = src->u.discard.flag;
 420                 dst->u.discard.id = src->u.discard.id;
 421                 dst->u.discard.sector_number = src->u.discard.sector_number;
 422                 dst->u.discard.nr_sectors = src->u.discard.nr_sectors;
 423                 break;
 424         case BLKIF_OP_INDIRECT:
 425                 dst->u.indirect.indirect_op = src->u.indirect.indirect_op;
 426                 dst->u.indirect.nr_segments = src->u.indirect.nr_segments;
 427                 dst->u.indirect.handle = src->u.indirect.handle;
 428                 dst->u.indirect.id = src->u.indirect.id;
 429                 dst->u.indirect.sector_number = src->u.indirect.sector_number;
 430                 barrier();
 431                 j = min(MAX_INDIRECT_PAGES, INDIRECT_PAGES(dst->u.indirect.nr_segments));
 432                 for (i = 0; i < j; i++)
 433                         dst->u.indirect.indirect_grefs[i] =
 434                                 src->u.indirect.indirect_grefs[i];
 435                 break;
 436         default:
 437                 /*
 438                  * Don't know how to translate this op. Only get the
 439                  * ID so failure can be reported to the frontend.
 440                  */
 441                 dst->u.other.id = src->u.other.id;
 442                 break;
 443         }
 444 }
 445
 446 static inline void blkif_get_x86_64_req(struct blkif_request *dst,
 447                                         struct blkif_x86_64_request *src)
 448 {
 449         int i, n = BLKIF_MAX_SEGMENTS_PER_REQUEST, j;
 450         dst->operation = src->operation;
 451         switch (src->operation) {
 452         case BLKIF_OP_READ:
 453         case BLKIF_OP_WRITE:
 454         case BLKIF_OP_WRITE_BARRIER:
 455         case BLKIF_OP_FLUSH_DISKCACHE:
 456                 dst->u.rw.nr_segments = src->u.rw.nr_segments;
 457                 dst->u.rw.handle = src->u.rw.handle;
 458                 dst->u.rw.id = src->u.rw.id;
 459                 dst->u.rw.sector_number = src->u.rw.sector_number;
 460                 barrier();
 461                 if (n > dst->u.rw.nr_segments)
 462                         n = dst->u.rw.nr_segments;
 463                 for (i = 0; i < n; i++)
 464                         dst->u.rw.seg[i] = src->u.rw.seg[i];
 465                 break;
 466         case BLKIF_OP_DISCARD:
 467                 dst->u.discard.flag = src->u.discard.flag;
 468                 dst->u.discard.id = src->u.discard.id;
 469                 dst->u.discard.sector_number = src->u.discard.sector_number;
 470                 dst->u.discard.nr_sectors = src->u.discard.nr_sectors;
 471                 break;
 472         case BLKIF_OP_INDIRECT:
 473                 dst->u.indirect.indirect_op = src->u.indirect.indirect_op;
 474                 dst->u.indirect.nr_segments = src->u.indirect.nr_segments;
 475                 dst->u.indirect.handle = src->u.indirect.handle;
 476                 dst->u.indirect.id = src->u.indirect.id;
 477                 dst->u.indirect.sector_number = src->u.indirect.sector_number;
 478                 barrier();
 479                 j = min(MAX_INDIRECT_PAGES, INDIRECT_PAGES(dst->u.indirect.nr_segments));
 480                 for (i = 0; i < j; i++)
 481                         dst->u.indirect.indirect_grefs[i] =
 482                                 src->u.indirect.indirect_grefs[i];
 483                 break;
 484         default:
 485                 /*
 486                  * Don't know how to translate this op. Only get the
 487                  * ID so failure can be reported to the frontend.
 488                  */
 489                 dst->u.other.id = src->u.other.id;
 490                 break;
 491         }
 492 }
 493
 494 #endif /* __XEN_BLKIF__BACKEND__COMMON_H__ */