arch/ia64/sn/kernel/xpc_channel.c

   1 /*
   2  * This file is subject to the terms and conditions of the GNU General Public
   3  * License.  See the file "COPYING" in the main directory of this archive
   4  * for more details.
   5  *
   6  * Copyright (c) 2004-2006 Silicon Graphics, Inc.  All Rights Reserved.
   7  */
   8
   9
  10 /*
  11  * Cross Partition Communication (XPC) channel support.
  12  *
  13  *      This is the part of XPC that manages the channels and
  14  *      sends/receives messages across them to/from other partitions.
  15  *
  16  */
  17
  18
  19 #include <linux/kernel.h>
  20 #include <linux/init.h>
  21 #include <linux/sched.h>
  22 #include <linux/cache.h>
  23 #include <linux/interrupt.h>
  24 #include <linux/slab.h>
  25 #include <linux/mutex.h>
  26 #include <linux/completion.h>
  27 #include <asm/sn/bte.h>
  28 #include <asm/sn/sn_sal.h>
  29 #include <asm/sn/xpc.h>
  30
  31
  32 /*
  33  * Set up the initial values for the XPartition Communication channels.
  34  */
  35 static void
  36 xpc_initialize_channels(struct xpc_partition *part, partid_t partid)
  37 {
  38         int ch_number;
  39         struct xpc_channel *ch;
  40
  41
  42         for (ch_number = 0; ch_number < part->nchannels; ch_number++) {
  43                 ch = &part->channels[ch_number];
  44
  45                 ch->partid = partid;
  46                 ch->number = ch_number;
  47                 ch->flags = XPC_C_DISCONNECTED;
  48
  49                 ch->local_GP = &part->local_GPs[ch_number];
  50                 ch->local_openclose_args =
  51                                         &part->local_openclose_args[ch_number];
  52
  53                 atomic_set(&ch->kthreads_assigned, 0);
  54                 atomic_set(&ch->kthreads_idle, 0);
  55                 atomic_set(&ch->kthreads_active, 0);
  56
  57                 atomic_set(&ch->references, 0);
  58                 atomic_set(&ch->n_to_notify, 0);
  59
  60                 spin_lock_init(&ch->lock);
  61                 mutex_init(&ch->msg_to_pull_mutex);
  62                 init_completion(&ch->wdisconnect_wait);
  63
  64                 atomic_set(&ch->n_on_msg_allocate_wq, 0);
  65                 init_waitqueue_head(&ch->msg_allocate_wq);
  66                 init_waitqueue_head(&ch->idle_wq);
  67         }
  68 }
  69
  70
  71 /*
  72  * Setup the infrastructure necessary to support XPartition Communication
  73  * between the specified remote partition and the local one.
  74  */
  75 enum xpc_retval
  76 xpc_setup_infrastructure(struct xpc_partition *part)
  77 {
  78         int ret, cpuid;
  79         struct timer_list *timer;
  80         partid_t partid = XPC_PARTID(part);
  81
  82
  83         /*
  84          * Zero out MOST of the entry for this partition. Only the fields
  85          * starting with `nchannels' will be zeroed. The preceding fields must
  86          * remain `viable' across partition ups and downs, since they may be
  87          * referenced during this memset() operation.
  88          */
  89         memset(&part->nchannels, 0, sizeof(struct xpc_partition) -
  90                                 offsetof(struct xpc_partition, nchannels));
  91
  92         /*
  93          * Allocate all of the channel structures as a contiguous chunk of
  94          * memory.
  95          */
  96         part->channels = kmalloc(sizeof(struct xpc_channel) * XPC_NCHANNELS,
  97                                                                 GFP_KERNEL);
  98         if (part->channels == NULL) {
  99                 dev_err(xpc_chan, "can't get memory for channels\n");
 100                 return xpcNoMemory;
 101         }
 102         memset(part->channels, 0, sizeof(struct xpc_channel) * XPC_NCHANNELS);
 103
 104         part->nchannels = XPC_NCHANNELS;
 105
 106
 107         /* allocate all the required GET/PUT values */
 108
 109         part->local_GPs = xpc_kmalloc_cacheline_aligned(XPC_GP_SIZE,
 110                                         GFP_KERNEL, &part->local_GPs_base);
 111         if (part->local_GPs == NULL) {
 112                 kfree(part->channels);
 113                 part->channels = NULL;
 114                 dev_err(xpc_chan, "can't get memory for local get/put "
 115                         "values\n");
 116                 return xpcNoMemory;
 117         }
 118         memset(part->local_GPs, 0, XPC_GP_SIZE);
 119
 120         part->remote_GPs = xpc_kmalloc_cacheline_aligned(XPC_GP_SIZE,
 121                                         GFP_KERNEL, &part->remote_GPs_base);
 122         if (part->remote_GPs == NULL) {
 123                 kfree(part->channels);
 124                 part->channels = NULL;
 125                 kfree(part->local_GPs_base);
 126                 part->local_GPs = NULL;
 127                 dev_err(xpc_chan, "can't get memory for remote get/put "
 128                         "values\n");
 129                 return xpcNoMemory;
 130         }
 131         memset(part->remote_GPs, 0, XPC_GP_SIZE);
 132
 133
 134         /* allocate all the required open and close args */
 135
 136         part->local_openclose_args = xpc_kmalloc_cacheline_aligned(
 137                                         XPC_OPENCLOSE_ARGS_SIZE, GFP_KERNEL,
 138                                         &part->local_openclose_args_base);
 139         if (part->local_openclose_args == NULL) {
 140                 kfree(part->channels);
 141                 part->channels = NULL;
 142                 kfree(part->local_GPs_base);
 143                 part->local_GPs = NULL;
 144                 kfree(part->remote_GPs_base);
 145                 part->remote_GPs = NULL;
 146                 dev_err(xpc_chan, "can't get memory for local connect args\n");
 147                 return xpcNoMemory;
 148         }
 149         memset(part->local_openclose_args, 0, XPC_OPENCLOSE_ARGS_SIZE);
 150
 151         part->remote_openclose_args = xpc_kmalloc_cacheline_aligned(
 152                                         XPC_OPENCLOSE_ARGS_SIZE, GFP_KERNEL,
 153                                         &part->remote_openclose_args_base);
 154         if (part->remote_openclose_args == NULL) {
 155                 kfree(part->channels);
 156                 part->channels = NULL;
 157                 kfree(part->local_GPs_base);
 158                 part->local_GPs = NULL;
 159                 kfree(part->remote_GPs_base);
 160                 part->remote_GPs = NULL;
 161                 kfree(part->local_openclose_args_base);
 162                 part->local_openclose_args = NULL;
 163                 dev_err(xpc_chan, "can't get memory for remote connect args\n");
 164                 return xpcNoMemory;
 165         }
 166         memset(part->remote_openclose_args, 0, XPC_OPENCLOSE_ARGS_SIZE);
 167
 168
 169         xpc_initialize_channels(part, partid);
 170
 171         atomic_set(&part->nchannels_active, 0);
 172         atomic_set(&part->nchannels_engaged, 0);
 173
 174
 175         /* local_IPI_amo were set to 0 by an earlier memset() */
 176
 177         /* Initialize this partitions AMO_t structure */
 178         part->local_IPI_amo_va = xpc_IPI_init(partid);
 179
 180         spin_lock_init(&part->IPI_lock);
 181
 182         atomic_set(&part->channel_mgr_requests, 1);
 183         init_waitqueue_head(&part->channel_mgr_wq);
 184
 185         sprintf(part->IPI_owner, "xpc%02d", partid);
 186         ret = request_irq(SGI_XPC_NOTIFY, xpc_notify_IRQ_handler, SA_SHIRQ,
 187                                 part->IPI_owner, (void *) (u64) partid);
 188         if (ret != 0) {
 189                 kfree(part->channels);
 190                 part->channels = NULL;
 191                 kfree(part->local_GPs_base);
 192                 part->local_GPs = NULL;
 193                 kfree(part->remote_GPs_base);
 194                 part->remote_GPs = NULL;
 195                 kfree(part->local_openclose_args_base);
 196                 part->local_openclose_args = NULL;
 197                 kfree(part->remote_openclose_args_base);
 198                 part->remote_openclose_args = NULL;
 199                 dev_err(xpc_chan, "can't register NOTIFY IRQ handler, "
 200                         "errno=%d\n", -ret);
 201                 return xpcLackOfResources;
 202         }
 203
 204         /* Setup a timer to check for dropped IPIs */
 205         timer = &part->dropped_IPI_timer;
 206         init_timer(timer);
 207         timer->function = (void (*)(unsigned long)) xpc_dropped_IPI_check;
 208         timer->data = (unsigned long) part;
 209         timer->expires = jiffies + XPC_P_DROPPED_IPI_WAIT;
 210         add_timer(timer);
 211
 212         /*
 213          * With the setting of the partition setup_state to XPC_P_SETUP, we're
 214          * declaring that this partition is ready to go.
 215          */
 216         part->setup_state = XPC_P_SETUP;
 217
 218
 219         /*
 220          * Setup the per partition specific variables required by the
 221          * remote partition to establish channel connections with us.
 222          *
 223          * The setting of the magic # indicates that these per partition
 224          * specific variables are ready to be used.
 225          */
 226         xpc_vars_part[partid].GPs_pa = __pa(part->local_GPs);
 227         xpc_vars_part[partid].openclose_args_pa =
 228                                         __pa(part->local_openclose_args);
 229         xpc_vars_part[partid].IPI_amo_pa = __pa(part->local_IPI_amo_va);
 230         cpuid = raw_smp_processor_id(); /* any CPU in this partition will do */
 231         xpc_vars_part[partid].IPI_nasid = cpuid_to_nasid(cpuid);
 232         xpc_vars_part[partid].IPI_phys_cpuid = cpu_physical_id(cpuid);
 233         xpc_vars_part[partid].nchannels = part->nchannels;
 234         xpc_vars_part[partid].magic = XPC_VP_MAGIC1;
 235
 236         return xpcSuccess;
 237 }
 238
 239
 240 /*
 241  * Create a wrapper that hides the underlying mechanism for pulling a cacheline
 242  * (or multiple cachelines) from a remote partition.
 243  *
 244  * src must be a cacheline aligned physical address on the remote partition.
 245  * dst must be a cacheline aligned virtual address on this partition.
 246  * cnt must be an cacheline sized
 247  */
 248 static enum xpc_retval
 249 xpc_pull_remote_cachelines(struct xpc_partition *part, void *dst,
 250                                 const void *src, size_t cnt)
 251 {
 252         bte_result_t bte_ret;
 253
 254
 255         DBUG_ON((u64) src != L1_CACHE_ALIGN((u64) src));
 256         DBUG_ON((u64) dst != L1_CACHE_ALIGN((u64) dst));
 257         DBUG_ON(cnt != L1_CACHE_ALIGN(cnt));
 258
 259         if (part->act_state == XPC_P_DEACTIVATING) {
 260                 return part->reason;
 261         }
 262
 263         bte_ret = xp_bte_copy((u64) src, (u64) ia64_tpa((u64) dst),
 264                                 (u64) cnt, (BTE_NORMAL | BTE_WACQUIRE), NULL);
 265         if (bte_ret == BTE_SUCCESS) {
 266                 return xpcSuccess;
 267         }
 268
 269         dev_dbg(xpc_chan, "xp_bte_copy() from partition %d failed, ret=%d\n",
 270                 XPC_PARTID(part), bte_ret);
 271
 272         return xpc_map_bte_errors(bte_ret);
 273 }
 274
 275
 276 /*
 277  * Pull the remote per partititon specific variables from the specified
 278  * partition.
 279  */
 280 enum xpc_retval
 281 xpc_pull_remote_vars_part(struct xpc_partition *part)
 282 {
 283         u8 buffer[L1_CACHE_BYTES * 2];
 284         struct xpc_vars_part *pulled_entry_cacheline =
 285                         (struct xpc_vars_part *) L1_CACHE_ALIGN((u64) buffer);
 286         struct xpc_vars_part *pulled_entry;
 287         u64 remote_entry_cacheline_pa, remote_entry_pa;
 288         partid_t partid = XPC_PARTID(part);
 289         enum xpc_retval ret;
 290
 291
 292         /* pull the cacheline that contains the variables we're interested in */
 293
 294         DBUG_ON(part->remote_vars_part_pa !=
 295                                 L1_CACHE_ALIGN(part->remote_vars_part_pa));
 296         DBUG_ON(sizeof(struct xpc_vars_part) != L1_CACHE_BYTES / 2);
 297
 298         remote_entry_pa = part->remote_vars_part_pa +
 299                         sn_partition_id * sizeof(struct xpc_vars_part);
 300
 301         remote_entry_cacheline_pa = (remote_entry_pa & ~(L1_CACHE_BYTES - 1));
 302
 303         pulled_entry = (struct xpc_vars_part *) ((u64) pulled_entry_cacheline +
 304                                 (remote_entry_pa & (L1_CACHE_BYTES - 1)));
 305
 306         ret = xpc_pull_remote_cachelines(part, pulled_entry_cacheline,
 307                                         (void *) remote_entry_cacheline_pa,
 308                                         L1_CACHE_BYTES);
 309         if (ret != xpcSuccess) {
 310                 dev_dbg(xpc_chan, "failed to pull XPC vars_part from "
 311                         "partition %d, ret=%d\n", partid, ret);
 312                 return ret;
 313         }
 314
 315
 316         /* see if they've been set up yet */
 317
 318         if (pulled_entry->magic != XPC_VP_MAGIC1 &&
 319                                 pulled_entry->magic != XPC_VP_MAGIC2) {
 320
 321                 if (pulled_entry->magic != 0) {
 322                         dev_dbg(xpc_chan, "partition %d's XPC vars_part for "
 323                                 "partition %d has bad magic value (=0x%lx)\n",
 324                                 partid, sn_partition_id, pulled_entry->magic);
 325                         return xpcBadMagic;
 326                 }
 327
 328                 /* they've not been initialized yet */
 329                 return xpcRetry;
 330         }
 331
 332         if (xpc_vars_part[partid].magic == XPC_VP_MAGIC1) {
 333
 334                 /* validate the variables */
 335
 336                 if (pulled_entry->GPs_pa == 0 ||
 337                                 pulled_entry->openclose_args_pa == 0 ||
 338                                         pulled_entry->IPI_amo_pa == 0) {
 339
 340                         dev_err(xpc_chan, "partition %d's XPC vars_part for "
 341                                 "partition %d are not valid\n", partid,
 342                                 sn_partition_id);
 343                         return xpcInvalidAddress;
 344                 }
 345
 346                 /* the variables we imported look to be valid */
 347
 348                 part->remote_GPs_pa = pulled_entry->GPs_pa;
 349                 part->remote_openclose_args_pa =
 350                                         pulled_entry->openclose_args_pa;
 351                 part->remote_IPI_amo_va =
 352                                       (AMO_t *) __va(pulled_entry->IPI_amo_pa);
 353                 part->remote_IPI_nasid = pulled_entry->IPI_nasid;
 354                 part->remote_IPI_phys_cpuid = pulled_entry->IPI_phys_cpuid;
 355
 356                 if (part->nchannels > pulled_entry->nchannels) {
 357                         part->nchannels = pulled_entry->nchannels;
 358                 }
 359
 360                 /* let the other side know that we've pulled their variables */
 361
 362                 xpc_vars_part[partid].magic = XPC_VP_MAGIC2;
 363         }
 364
 365         if (pulled_entry->magic == XPC_VP_MAGIC1) {
 366                 return xpcRetry;
 367         }
 368
 369         return xpcSuccess;
 370 }
 371
 372
 373 /*
 374  * Get the IPI flags and pull the openclose args and/or remote GPs as needed.
 375  */
 376 static u64
 377 xpc_get_IPI_flags(struct xpc_partition *part)
 378 {
 379         unsigned long irq_flags;
 380         u64 IPI_amo;
 381         enum xpc_retval ret;
 382
 383
 384         /*
 385          * See if there are any IPI flags to be handled.
 386          */
 387
 388         spin_lock_irqsave(&part->IPI_lock, irq_flags);
 389         if ((IPI_amo = part->local_IPI_amo) != 0) {
 390                 part->local_IPI_amo = 0;
 391         }
 392         spin_unlock_irqrestore(&part->IPI_lock, irq_flags);
 393
 394
 395         if (XPC_ANY_OPENCLOSE_IPI_FLAGS_SET(IPI_amo)) {
 396                 ret = xpc_pull_remote_cachelines(part,
 397                                         part->remote_openclose_args,
 398                                         (void *) part->remote_openclose_args_pa,
 399                                         XPC_OPENCLOSE_ARGS_SIZE);
 400                 if (ret != xpcSuccess) {
 401                         XPC_DEACTIVATE_PARTITION(part, ret);
 402
 403                         dev_dbg(xpc_chan, "failed to pull openclose args from "
 404                                 "partition %d, ret=%d\n", XPC_PARTID(part),
 405                                 ret);
 406
 407                         /* don't bother processing IPIs anymore */
 408                         IPI_amo = 0;
 409                 }
 410         }
 411
 412         if (XPC_ANY_MSG_IPI_FLAGS_SET(IPI_amo)) {
 413                 ret = xpc_pull_remote_cachelines(part, part->remote_GPs,
 414                                                 (void *) part->remote_GPs_pa,
 415                                                 XPC_GP_SIZE);
 416                 if (ret != xpcSuccess) {
 417                         XPC_DEACTIVATE_PARTITION(part, ret);
 418
 419                         dev_dbg(xpc_chan, "failed to pull GPs from partition "
 420                                 "%d, ret=%d\n", XPC_PARTID(part), ret);
 421
 422                         /* don't bother processing IPIs anymore */
 423                         IPI_amo = 0;
 424                 }
 425         }
 426
 427         return IPI_amo;
 428 }
 429
 430
 431 /*
 432  * Allocate the local message queue and the notify queue.
 433  */
 434 static enum xpc_retval
 435 xpc_allocate_local_msgqueue(struct xpc_channel *ch)
 436 {
 437         unsigned long irq_flags;
 438         int nentries;
 439         size_t nbytes;
 440
 441
 442         // >>> may want to check for ch->flags & XPC_C_DISCONNECTING between
 443         // >>> iterations of the for-loop, bail if set?
 444
 445         // >>> should we impose a minumum #of entries? like 4 or 8?
 446         for (nentries = ch->local_nentries; nentries > 0; nentries--) {
 447
 448                 nbytes = nentries * ch->msg_size;
 449                 ch->local_msgqueue = xpc_kmalloc_cacheline_aligned(nbytes,
 450                                                 GFP_KERNEL,
 451                                                 &ch->local_msgqueue_base);
 452                 if (ch->local_msgqueue == NULL) {
 453                         continue;
 454                 }
 455                 memset(ch->local_msgqueue, 0, nbytes);
 456
 457                 nbytes = nentries * sizeof(struct xpc_notify);
 458                 ch->notify_queue = kmalloc(nbytes, GFP_KERNEL);
 459                 if (ch->notify_queue == NULL) {
 460                         kfree(ch->local_msgqueue_base);
 461                         ch->local_msgqueue = NULL;
 462                         continue;
 463                 }
 464                 memset(ch->notify_queue, 0, nbytes);
 465
 466                 spin_lock_irqsave(&ch->lock, irq_flags);
 467                 if (nentries < ch->local_nentries) {
 468                         dev_dbg(xpc_chan, "nentries=%d local_nentries=%d, "
 469                                 "partid=%d, channel=%d\n", nentries,
 470                                 ch->local_nentries, ch->partid, ch->number);
 471
 472                         ch->local_nentries = nentries;
 473                 }
 474                 spin_unlock_irqrestore(&ch->lock, irq_flags);
 475                 return xpcSuccess;
 476         }
 477
 478         dev_dbg(xpc_chan, "can't get memory for local message queue and notify "
 479                 "queue, partid=%d, channel=%d\n", ch->partid, ch->number);
 480         return xpcNoMemory;
 481 }
 482
 483
 484 /*
 485  * Allocate the cached remote message queue.
 486  */
 487 static enum xpc_retval
 488 xpc_allocate_remote_msgqueue(struct xpc_channel *ch)
 489 {
 490         unsigned long irq_flags;
 491         int nentries;
 492         size_t nbytes;
 493
 494
 495         DBUG_ON(ch->remote_nentries <= 0);
 496
 497         // >>> may want to check for ch->flags & XPC_C_DISCONNECTING between
 498         // >>> iterations of the for-loop, bail if set?
 499
 500         // >>> should we impose a minumum #of entries? like 4 or 8?
 501         for (nentries = ch->remote_nentries; nentries > 0; nentries--) {
 502
 503                 nbytes = nentries * ch->msg_size;
 504                 ch->remote_msgqueue = xpc_kmalloc_cacheline_aligned(nbytes,
 505                                                 GFP_KERNEL,
 506                                                 &ch->remote_msgqueue_base);
 507                 if (ch->remote_msgqueue == NULL) {
 508                         continue;
 509                 }
 510                 memset(ch->remote_msgqueue, 0, nbytes);
 511
 512                 spin_lock_irqsave(&ch->lock, irq_flags);
 513                 if (nentries < ch->remote_nentries) {
 514                         dev_dbg(xpc_chan, "nentries=%d remote_nentries=%d, "
 515                                 "partid=%d, channel=%d\n", nentries,
 516                                 ch->remote_nentries, ch->partid, ch->number);
 517
 518                         ch->remote_nentries = nentries;
 519                 }
 520                 spin_unlock_irqrestore(&ch->lock, irq_flags);
 521                 return xpcSuccess;
 522         }
 523
 524         dev_dbg(xpc_chan, "can't get memory for cached remote message queue, "
 525                 "partid=%d, channel=%d\n", ch->partid, ch->number);
 526         return xpcNoMemory;
 527 }
 528
 529
 530 /*
 531  * Allocate message queues and other stuff associated with a channel.
 532  *
 533  * Note: Assumes all of the channel sizes are filled in.
 534  */
 535 static enum xpc_retval
 536 xpc_allocate_msgqueues(struct xpc_channel *ch)
 537 {
 538         unsigned long irq_flags;
 539         enum xpc_retval ret;
 540
 541
 542         DBUG_ON(ch->flags & XPC_C_SETUP);
 543
 544         if ((ret = xpc_allocate_local_msgqueue(ch)) != xpcSuccess) {
 545                 return ret;
 546         }
 547
 548         if ((ret = xpc_allocate_remote_msgqueue(ch)) != xpcSuccess) {
 549                 kfree(ch->local_msgqueue_base);
 550                 ch->local_msgqueue = NULL;
 551                 kfree(ch->notify_queue);
 552                 ch->notify_queue = NULL;
 553                 return ret;
 554         }
 555
 556         spin_lock_irqsave(&ch->lock, irq_flags);
 557         ch->flags |= XPC_C_SETUP;
 558         spin_unlock_irqrestore(&ch->lock, irq_flags);
 559
 560         return xpcSuccess;
 561 }
 562
 563
 564 /*
 565  * Process a connect message from a remote partition.
 566  *
 567  * Note: xpc_process_connect() is expecting to be called with the
 568  * spin_lock_irqsave held and will leave it locked upon return.
 569  */
 570 static void
 571 xpc_process_connect(struct xpc_channel *ch, unsigned long *irq_flags)
 572 {
 573         enum xpc_retval ret;
 574
 575
 576         DBUG_ON(!spin_is_locked(&ch->lock));
 577
 578         if (!(ch->flags & XPC_C_OPENREQUEST) ||
 579                                 !(ch->flags & XPC_C_ROPENREQUEST)) {
 580                 /* nothing more to do for now */
 581                 return;
 582         }
 583         DBUG_ON(!(ch->flags & XPC_C_CONNECTING));
 584
 585         if (!(ch->flags & XPC_C_SETUP)) {
 586                 spin_unlock_irqrestore(&ch->lock, *irq_flags);
 587                 ret = xpc_allocate_msgqueues(ch);
 588                 spin_lock_irqsave(&ch->lock, *irq_flags);
 589
 590                 if (ret != xpcSuccess) {
 591                         XPC_DISCONNECT_CHANNEL(ch, ret, irq_flags);
 592                 }
 593                 if (ch->flags & (XPC_C_CONNECTED | XPC_C_DISCONNECTING)) {
 594                         return;
 595                 }
 596
 597                 DBUG_ON(!(ch->flags & XPC_C_SETUP));
 598                 DBUG_ON(ch->local_msgqueue == NULL);
 599                 DBUG_ON(ch->remote_msgqueue == NULL);
 600         }
 601
 602         if (!(ch->flags & XPC_C_OPENREPLY)) {
 603                 ch->flags |= XPC_C_OPENREPLY;
 604                 xpc_IPI_send_openreply(ch, irq_flags);
 605         }
 606
 607         if (!(ch->flags & XPC_C_ROPENREPLY)) {
 608                 return;
 609         }
 610
 611         DBUG_ON(ch->remote_msgqueue_pa == 0);
 612
 613         ch->flags = (XPC_C_CONNECTED | XPC_C_SETUP);    /* clear all else */
 614
 615         dev_info(xpc_chan, "channel %d to partition %d connected\n",
 616                 ch->number, ch->partid);
 617
 618         spin_unlock_irqrestore(&ch->lock, *irq_flags);
 619         xpc_create_kthreads(ch, 1);
 620         spin_lock_irqsave(&ch->lock, *irq_flags);
 621 }
 622
 623
 624 /*
 625  * Notify those who wanted to be notified upon delivery of their message.
 626  */
 627 static void
 628 xpc_notify_senders(struct xpc_channel *ch, enum xpc_retval reason, s64 put)
 629 {
 630         struct xpc_notify *notify;
 631         u8 notify_type;
 632         s64 get = ch->w_remote_GP.get - 1;
 633
 634
 635         while (++get < put && atomic_read(&ch->n_to_notify) > 0) {
 636
 637                 notify = &ch->notify_queue[get % ch->local_nentries];
 638
 639                 /*
 640                  * See if the notify entry indicates it was associated with
 641                  * a message who's sender wants to be notified. It is possible
 642                  * that it is, but someone else is doing or has done the
 643                  * notification.
 644                  */
 645                 notify_type = notify->type;
 646                 if (notify_type == 0 ||
 647                                 cmpxchg(&notify->type, notify_type, 0) !=
 648                                                                 notify_type) {
 649                         continue;
 650                 }
 651
 652                 DBUG_ON(notify_type != XPC_N_CALL);
 653
 654                 atomic_dec(&ch->n_to_notify);
 655
 656                 if (notify->func != NULL) {
 657                         dev_dbg(xpc_chan, "notify->func() called, notify=0x%p, "
 658                                 "msg_number=%ld, partid=%d, channel=%d\n",
 659                                 (void *) notify, get, ch->partid, ch->number);
 660
 661                         notify->func(reason, ch->partid, ch->number,
 662                                                                 notify->key);
 663
 664                         dev_dbg(xpc_chan, "notify->func() returned, "
 665                                 "notify=0x%p, msg_number=%ld, partid=%d, "
 666                                 "channel=%d\n", (void *) notify, get,
 667                                 ch->partid, ch->number);
 668                 }
 669         }
 670 }
 671
 672
 673 /*
 674  * Free up message queues and other stuff that were allocated for the specified
 675  * channel.
 676  *
 677  * Note: ch->reason and ch->reason_line are left set for debugging purposes,
 678  * they're cleared when XPC_C_DISCONNECTED is cleared.
 679  */
 680 static void
 681 xpc_free_msgqueues(struct xpc_channel *ch)
 682 {
 683         DBUG_ON(!spin_is_locked(&ch->lock));
 684         DBUG_ON(atomic_read(&ch->n_to_notify) != 0);
 685
 686         ch->remote_msgqueue_pa = 0;
 687         ch->func = NULL;
 688         ch->key = NULL;
 689         ch->msg_size = 0;
 690         ch->local_nentries = 0;
 691         ch->remote_nentries = 0;
 692         ch->kthreads_assigned_limit = 0;
 693         ch->kthreads_idle_limit = 0;
 694
 695         ch->local_GP->get = 0;
 696         ch->local_GP->put = 0;
 697         ch->remote_GP.get = 0;
 698         ch->remote_GP.put = 0;
 699         ch->w_local_GP.get = 0;
 700         ch->w_local_GP.put = 0;
 701         ch->w_remote_GP.get = 0;
 702         ch->w_remote_GP.put = 0;
 703         ch->next_msg_to_pull = 0;
 704
 705         if (ch->flags & XPC_C_SETUP) {
 706                 ch->flags &= ~XPC_C_SETUP;
 707
 708                 dev_dbg(xpc_chan, "ch->flags=0x%x, partid=%d, channel=%d\n",
 709                         ch->flags, ch->partid, ch->number);
 710
 711                 kfree(ch->local_msgqueue_base);
 712                 ch->local_msgqueue = NULL;
 713                 kfree(ch->remote_msgqueue_base);
 714                 ch->remote_msgqueue = NULL;
 715                 kfree(ch->notify_queue);
 716                 ch->notify_queue = NULL;
 717         }
 718 }
 719
 720
 721 /*
 722  * spin_lock_irqsave() is expected to be held on entry.
 723  */
 724 static void
 725 xpc_process_disconnect(struct xpc_channel *ch, unsigned long *irq_flags)
 726 {
 727         struct xpc_partition *part = &xpc_partitions[ch->partid];
 728         u32 channel_was_connected = (ch->flags & XPC_C_WASCONNECTED);
 729
 730
 731         DBUG_ON(!spin_is_locked(&ch->lock));
 732
 733         if (!(ch->flags & XPC_C_DISCONNECTING)) {
 734                 return;
 735         }
 736
 737         DBUG_ON(!(ch->flags & XPC_C_CLOSEREQUEST));
 738
 739         /* make sure all activity has settled down first */
 740
 741         if (atomic_read(&ch->references) > 0) {
 742                 return;
 743         }
 744         DBUG_ON(atomic_read(&ch->kthreads_assigned) != 0);
 745
 746         if (part->act_state == XPC_P_DEACTIVATING) {
 747                 /* can't proceed until the other side disengages from us */
 748                 if (xpc_partition_engaged(1UL << ch->partid)) {
 749                         return;
 750                 }
 751
 752         } else {
 753
 754                 /* as long as the other side is up do the full protocol */
 755
 756                 if (!(ch->flags & XPC_C_RCLOSEREQUEST)) {
 757                         return;
 758                 }
 759
 760                 if (!(ch->flags & XPC_C_CLOSEREPLY)) {
 761                         ch->flags |= XPC_C_CLOSEREPLY;
 762                         xpc_IPI_send_closereply(ch, irq_flags);
 763                 }
 764
 765                 if (!(ch->flags & XPC_C_RCLOSEREPLY)) {
 766                         return;
 767                 }
 768         }
 769
 770         /* wake those waiting for notify completion */
 771         if (atomic_read(&ch->n_to_notify) > 0) {
 772                 /* >>> we do callout while holding ch->lock */
 773                 xpc_notify_senders(ch, ch->reason, ch->w_local_GP.put);
 774         }
 775
 776         /* both sides are disconnected now */
 777
 778         if (ch->flags & XPC_C_CONNECTCALLOUT) {
 779                 spin_unlock_irqrestore(&ch->lock, *irq_flags);
 780                 xpc_disconnect_callout(ch, xpcDisconnected);
 781                 spin_lock_irqsave(&ch->lock, *irq_flags);
 782         }
 783
 784         /* it's now safe to free the channel's message queues */
 785         xpc_free_msgqueues(ch);
 786
 787         /* mark disconnected, clear all other flags except XPC_C_WDISCONNECT */
 788         ch->flags = (XPC_C_DISCONNECTED | (ch->flags & XPC_C_WDISCONNECT));
 789
 790         atomic_dec(&part->nchannels_active);
 791
 792         if (channel_was_connected) {
 793                 dev_info(xpc_chan, "channel %d to partition %d disconnected, "
 794                         "reason=%d\n", ch->number, ch->partid, ch->reason);
 795         }
 796
 797         if (ch->flags & XPC_C_WDISCONNECT) {
 798                 /* we won't lose the CPU since we're holding ch->lock */
 799                 complete(&ch->wdisconnect_wait);
 800         } else if (ch->delayed_IPI_flags) {
 801                 if (part->act_state != XPC_P_DEACTIVATING) {
 802                         /* time to take action on any delayed IPI flags */
 803                         spin_lock(&part->IPI_lock);
 804                         XPC_SET_IPI_FLAGS(part->local_IPI_amo, ch->number,
 805                                                         ch->delayed_IPI_flags);
 806                         spin_unlock(&part->IPI_lock);
 807                 }
 808                 ch->delayed_IPI_flags = 0;
 809         }
 810 }
 811
 812
 813 /*
 814  * Process a change in the channel's remote connection state.
 815  */
 816 static void
 817 xpc_process_openclose_IPI(struct xpc_partition *part, int ch_number,
 818                                 u8 IPI_flags)
 819 {
 820         unsigned long irq_flags;
 821         struct xpc_openclose_args *args =
 822                                 &part->remote_openclose_args[ch_number];
 823         struct xpc_channel *ch = &part->channels[ch_number];
 824         enum xpc_retval reason;
 825
 826
 827
 828         spin_lock_irqsave(&ch->lock, irq_flags);
 829
 830 again:
 831
 832         if ((ch->flags & XPC_C_DISCONNECTED) &&
 833                                         (ch->flags & XPC_C_WDISCONNECT)) {
 834                 /*
 835                  * Delay processing IPI flags until thread waiting disconnect
 836                  * has had a chance to see that the channel is disconnected.
 837                  */
 838                 ch->delayed_IPI_flags |= IPI_flags;
 839                 spin_unlock_irqrestore(&ch->lock, irq_flags);
 840                 return;
 841         }
 842
 843
 844         if (IPI_flags & XPC_IPI_CLOSEREQUEST) {
 845
 846                 dev_dbg(xpc_chan, "XPC_IPI_CLOSEREQUEST (reason=%d) received "
 847                         "from partid=%d, channel=%d\n", args->reason,
 848                         ch->partid, ch->number);
 849
 850                 /*
 851                  * If RCLOSEREQUEST is set, we're probably waiting for
 852                  * RCLOSEREPLY. We should find it and a ROPENREQUEST packed
 853                  * with this RCLOSEREQUEST in the IPI_flags.
 854                  */
 855
 856                 if (ch->flags & XPC_C_RCLOSEREQUEST) {
 857                         DBUG_ON(!(ch->flags & XPC_C_DISCONNECTING));
 858                         DBUG_ON(!(ch->flags & XPC_C_CLOSEREQUEST));
 859                         DBUG_ON(!(ch->flags & XPC_C_CLOSEREPLY));
 860                         DBUG_ON(ch->flags & XPC_C_RCLOSEREPLY);
 861
 862                         DBUG_ON(!(IPI_flags & XPC_IPI_CLOSEREPLY));
 863                         IPI_flags &= ~XPC_IPI_CLOSEREPLY;
 864                         ch->flags |= XPC_C_RCLOSEREPLY;
 865
 866                         /* both sides have finished disconnecting */
 867                         xpc_process_disconnect(ch, &irq_flags);
 868                         DBUG_ON(!(ch->flags & XPC_C_DISCONNECTED));
 869                         goto again;
 870                 }
 871
 872                 if (ch->flags & XPC_C_DISCONNECTED) {
 873                         if (!(IPI_flags & XPC_IPI_OPENREQUEST)) {
 874                                 if ((XPC_GET_IPI_FLAGS(part->local_IPI_amo,
 875                                          ch_number) & XPC_IPI_OPENREQUEST)) {
 876
 877                                         DBUG_ON(ch->delayed_IPI_flags != 0);
 878                                         spin_lock(&part->IPI_lock);
 879                                         XPC_SET_IPI_FLAGS(part->local_IPI_amo,
 880                                                         ch_number,
 881                                                         XPC_IPI_CLOSEREQUEST);
 882                                         spin_unlock(&part->IPI_lock);
 883                                 }
 884                                 spin_unlock_irqrestore(&ch->lock, irq_flags);
 885                                 return;
 886                         }
 887
 888                         XPC_SET_REASON(ch, 0, 0);
 889                         ch->flags &= ~XPC_C_DISCONNECTED;
 890
 891                         atomic_inc(&part->nchannels_active);
 892                         ch->flags |= (XPC_C_CONNECTING | XPC_C_ROPENREQUEST);
 893                 }
 894
 895                 IPI_flags &= ~(XPC_IPI_OPENREQUEST | XPC_IPI_OPENREPLY);
 896
 897                 /*
 898                  * The meaningful CLOSEREQUEST connection state fields are:
 899                  *      reason = reason connection is to be closed
 900                  */
 901
 902                 ch->flags |= XPC_C_RCLOSEREQUEST;
 903
 904                 if (!(ch->flags & XPC_C_DISCONNECTING)) {
 905                         reason = args->reason;
 906                         if (reason <= xpcSuccess || reason > xpcUnknownReason) {
 907                                 reason = xpcUnknownReason;
 908                         } else if (reason == xpcUnregistering) {
 909                                 reason = xpcOtherUnregistering;
 910                         }
 911
 912                         XPC_DISCONNECT_CHANNEL(ch, reason, &irq_flags);
 913
 914                         DBUG_ON(IPI_flags & XPC_IPI_CLOSEREPLY);
 915                         spin_unlock_irqrestore(&ch->lock, irq_flags);
 916                         return;
 917                 }
 918
 919                 xpc_process_disconnect(ch, &irq_flags);
 920         }
 921
 922
 923         if (IPI_flags & XPC_IPI_CLOSEREPLY) {
 924
 925                 dev_dbg(xpc_chan, "XPC_IPI_CLOSEREPLY received from partid=%d,"
 926                         " channel=%d\n", ch->partid, ch->number);
 927
 928                 if (ch->flags & XPC_C_DISCONNECTED) {
 929                         DBUG_ON(part->act_state != XPC_P_DEACTIVATING);
 930                         spin_unlock_irqrestore(&ch->lock, irq_flags);
 931                         return;
 932                 }
 933
 934                 DBUG_ON(!(ch->flags & XPC_C_CLOSEREQUEST));
 935
 936                 if (!(ch->flags & XPC_C_RCLOSEREQUEST)) {
 937                         if ((XPC_GET_IPI_FLAGS(part->local_IPI_amo, ch_number)
 938                                                 & XPC_IPI_CLOSEREQUEST)) {
 939
 940                                 DBUG_ON(ch->delayed_IPI_flags != 0);
 941                                 spin_lock(&part->IPI_lock);
 942                                 XPC_SET_IPI_FLAGS(part->local_IPI_amo,
 943                                                 ch_number, XPC_IPI_CLOSEREPLY);
 944                                 spin_unlock(&part->IPI_lock);
 945                         }
 946                         spin_unlock_irqrestore(&ch->lock, irq_flags);
 947                         return;
 948                 }
 949
 950                 ch->flags |= XPC_C_RCLOSEREPLY;
 951
 952                 if (ch->flags & XPC_C_CLOSEREPLY) {
 953                         /* both sides have finished disconnecting */
 954                         xpc_process_disconnect(ch, &irq_flags);
 955                 }
 956         }
 957
 958
 959         if (IPI_flags & XPC_IPI_OPENREQUEST) {
 960
 961                 dev_dbg(xpc_chan, "XPC_IPI_OPENREQUEST (msg_size=%d, "
 962                         "local_nentries=%d) received from partid=%d, "
 963                         "channel=%d\n", args->msg_size, args->local_nentries,
 964                         ch->partid, ch->number);
 965
 966                 if (part->act_state == XPC_P_DEACTIVATING ||
 967                                         (ch->flags & XPC_C_ROPENREQUEST)) {
 968                         spin_unlock_irqrestore(&ch->lock, irq_flags);
 969                         return;
 970                 }
 971
 972                 if (ch->flags & (XPC_C_DISCONNECTING | XPC_C_WDISCONNECT)) {
 973                         ch->delayed_IPI_flags |= XPC_IPI_OPENREQUEST;
 974                         spin_unlock_irqrestore(&ch->lock, irq_flags);
 975                         return;
 976                 }
 977                 DBUG_ON(!(ch->flags & (XPC_C_DISCONNECTED |
 978                                                         XPC_C_OPENREQUEST)));
 979                 DBUG_ON(ch->flags & (XPC_C_ROPENREQUEST | XPC_C_ROPENREPLY |
 980                                         XPC_C_OPENREPLY | XPC_C_CONNECTED));
 981
 982                 /*
 983                  * The meaningful OPENREQUEST connection state fields are:
 984                  *      msg_size = size of channel's messages in bytes
 985                  *      local_nentries = remote partition's local_nentries
 986                  */
 987                 if (args->msg_size == 0 || args->local_nentries == 0) {
 988                         /* assume OPENREQUEST was delayed by mistake */
 989                         spin_unlock_irqrestore(&ch->lock, irq_flags);
 990                         return;
 991                 }
 992
 993                 ch->flags |= (XPC_C_ROPENREQUEST | XPC_C_CONNECTING);
 994                 ch->remote_nentries = args->local_nentries;
 995
 996
 997                 if (ch->flags & XPC_C_OPENREQUEST) {
 998                         if (args->msg_size != ch->msg_size) {
 999                                 XPC_DISCONNECT_CHANNEL(ch, xpcUnequalMsgSizes,
1000                                                                 &irq_flags);
1001                                 spin_unlock_irqrestore(&ch->lock, irq_flags);
1002                                 return;
1003                         }
1004                 } else {
1005                         ch->msg_size = args->msg_size;
1006
1007                         XPC_SET_REASON(ch, 0, 0);
1008                         ch->flags &= ~XPC_C_DISCONNECTED;
1009
1010                         atomic_inc(&part->nchannels_active);
1011                 }
1012
1013                 xpc_process_connect(ch, &irq_flags);
1014         }
1015
1016
1017         if (IPI_flags & XPC_IPI_OPENREPLY) {
1018
1019                 dev_dbg(xpc_chan, "XPC_IPI_OPENREPLY (local_msgqueue_pa=0x%lx, "
1020                         "local_nentries=%d, remote_nentries=%d) received from "
1021                         "partid=%d, channel=%d\n", args->local_msgqueue_pa,
1022                         args->local_nentries, args->remote_nentries,
1023                         ch->partid, ch->number);
1024
1025                 if (ch->flags & (XPC_C_DISCONNECTING | XPC_C_DISCONNECTED)) {
1026                         spin_unlock_irqrestore(&ch->lock, irq_flags);
1027                         return;
1028                 }
1029                 if (!(ch->flags & XPC_C_OPENREQUEST)) {
1030                         XPC_DISCONNECT_CHANNEL(ch, xpcOpenCloseError,
1031                                                                 &irq_flags);
1032                         spin_unlock_irqrestore(&ch->lock, irq_flags);
1033                         return;
1034                 }
1035
1036                 DBUG_ON(!(ch->flags & XPC_C_ROPENREQUEST));
1037                 DBUG_ON(ch->flags & XPC_C_CONNECTED);
1038
1039                 /*
1040                  * The meaningful OPENREPLY connection state fields are:
1041                  *      local_msgqueue_pa = physical address of remote
1042                  *                          partition's local_msgqueue
1043                  *      local_nentries = remote partition's local_nentries
1044                  *      remote_nentries = remote partition's remote_nentries
1045                  */
1046                 DBUG_ON(args->local_msgqueue_pa == 0);
1047                 DBUG_ON(args->local_nentries == 0);
1048                 DBUG_ON(args->remote_nentries == 0);
1049
1050                 ch->flags |= XPC_C_ROPENREPLY;
1051                 ch->remote_msgqueue_pa = args->local_msgqueue_pa;
1052
1053                 if (args->local_nentries < ch->remote_nentries) {
1054                         dev_dbg(xpc_chan, "XPC_IPI_OPENREPLY: new "
1055                                 "remote_nentries=%d, old remote_nentries=%d, "
1056                                 "partid=%d, channel=%d\n",
1057                                 args->local_nentries, ch->remote_nentries,
1058                                 ch->partid, ch->number);
1059
1060                         ch->remote_nentries = args->local_nentries;
1061                 }
1062                 if (args->remote_nentries < ch->local_nentries) {
1063                         dev_dbg(xpc_chan, "XPC_IPI_OPENREPLY: new "
1064                                 "local_nentries=%d, old local_nentries=%d, "
1065                                 "partid=%d, channel=%d\n",
1066                                 args->remote_nentries, ch->local_nentries,
1067                                 ch->partid, ch->number);
1068
1069                         ch->local_nentries = args->remote_nentries;
1070                 }
1071
1072                 xpc_process_connect(ch, &irq_flags);
1073         }
1074
1075         spin_unlock_irqrestore(&ch->lock, irq_flags);
1076 }
1077
1078
1079 /*
1080  * Attempt to establish a channel connection to a remote partition.
1081  */
1082 static enum xpc_retval
1083 xpc_connect_channel(struct xpc_channel *ch)
1084 {
1085         unsigned long irq_flags;
1086         struct xpc_registration *registration = &xpc_registrations[ch->number];
1087
1088
1089         if (mutex_trylock(&registration->mutex) == 0) {
1090                 return xpcRetry;
1091         }
1092
1093         if (!XPC_CHANNEL_REGISTERED(ch->number)) {
1094                 mutex_unlock(&registration->mutex);
1095                 return xpcUnregistered;
1096         }
1097
1098         spin_lock_irqsave(&ch->lock, irq_flags);
1099
1100         DBUG_ON(ch->flags & XPC_C_CONNECTED);
1101         DBUG_ON(ch->flags & XPC_C_OPENREQUEST);
1102
1103         if (ch->flags & XPC_C_DISCONNECTING) {
1104                 spin_unlock_irqrestore(&ch->lock, irq_flags);
1105                 mutex_unlock(&registration->mutex);
1106                 return ch->reason;
1107         }
1108
1109
1110         /* add info from the channel connect registration to the channel */
1111
1112         ch->kthreads_assigned_limit = registration->assigned_limit;
1113         ch->kthreads_idle_limit = registration->idle_limit;
1114         DBUG_ON(atomic_read(&ch->kthreads_assigned) != 0);
1115         DBUG_ON(atomic_read(&ch->kthreads_idle) != 0);
1116         DBUG_ON(atomic_read(&ch->kthreads_active) != 0);
1117
1118         ch->func = registration->func;
1119         DBUG_ON(registration->func == NULL);
1120         ch->key = registration->key;
1121
1122         ch->local_nentries = registration->nentries;
1123
1124         if (ch->flags & XPC_C_ROPENREQUEST) {
1125                 if (registration->msg_size != ch->msg_size) {
1126                         /* the local and remote sides aren't the same */
1127
1128                         /*
1129                          * Because XPC_DISCONNECT_CHANNEL() can block we're
1130                          * forced to up the registration sema before we unlock
1131                          * the channel lock. But that's okay here because we're
1132                          * done with the part that required the registration
1133                          * sema. XPC_DISCONNECT_CHANNEL() requires that the
1134                          * channel lock be locked and will unlock and relock
1135                          * the channel lock as needed.
1136                          */
1137                         mutex_unlock(&registration->mutex);
1138                         XPC_DISCONNECT_CHANNEL(ch, xpcUnequalMsgSizes,
1139                                                                 &irq_flags);
1140                         spin_unlock_irqrestore(&ch->lock, irq_flags);
1141                         return xpcUnequalMsgSizes;
1142                 }
1143         } else {
1144                 ch->msg_size = registration->msg_size;
1145
1146                 XPC_SET_REASON(ch, 0, 0);
1147                 ch->flags &= ~XPC_C_DISCONNECTED;
1148
1149                 atomic_inc(&xpc_partitions[ch->partid].nchannels_active);
1150         }
1151
1152         mutex_unlock(&registration->mutex);
1153
1154
1155         /* initiate the connection */
1156
1157         ch->flags |= (XPC_C_OPENREQUEST | XPC_C_CONNECTING);
1158         xpc_IPI_send_openrequest(ch, &irq_flags);
1159
1160         xpc_process_connect(ch, &irq_flags);
1161
1162         spin_unlock_irqrestore(&ch->lock, irq_flags);
1163
1164         return xpcSuccess;
1165 }
1166
1167
1168 /*
1169  * Clear some of the msg flags in the local message queue.
1170  */
1171 static inline void
1172 xpc_clear_local_msgqueue_flags(struct xpc_channel *ch)
1173 {
1174         struct xpc_msg *msg;
1175         s64 get;
1176
1177
1178         get = ch->w_remote_GP.get;
1179         do {
1180                 msg = (struct xpc_msg *) ((u64) ch->local_msgqueue +
1181                                 (get % ch->local_nentries) * ch->msg_size);
1182                 msg->flags = 0;
1183         } while (++get < (volatile s64) ch->remote_GP.get);
1184 }
1185
1186
1187 /*
1188  * Clear some of the msg flags in the remote message queue.
1189  */
1190 static inline void
1191 xpc_clear_remote_msgqueue_flags(struct xpc_channel *ch)
1192 {
1193         struct xpc_msg *msg;
1194         s64 put;
1195
1196
1197         put = ch->w_remote_GP.put;
1198         do {
1199                 msg = (struct xpc_msg *) ((u64) ch->remote_msgqueue +
1200                                 (put % ch->remote_nentries) * ch->msg_size);
1201                 msg->flags = 0;
1202         } while (++put < (volatile s64) ch->remote_GP.put);
1203 }
1204
1205
1206 static void
1207 xpc_process_msg_IPI(struct xpc_partition *part, int ch_number)
1208 {
1209         struct xpc_channel *ch = &part->channels[ch_number];
1210         int nmsgs_sent;
1211
1212
1213         ch->remote_GP = part->remote_GPs[ch_number];
1214
1215
1216         /* See what, if anything, has changed for each connected channel */
1217
1218         xpc_msgqueue_ref(ch);
1219
1220         if (ch->w_remote_GP.get == ch->remote_GP.get &&
1221                                 ch->w_remote_GP.put == ch->remote_GP.put) {
1222                 /* nothing changed since GPs were last pulled */
1223                 xpc_msgqueue_deref(ch);
1224                 return;
1225         }
1226
1227         if (!(ch->flags & XPC_C_CONNECTED)){
1228                 xpc_msgqueue_deref(ch);
1229                 return;
1230         }
1231
1232
1233         /*
1234          * First check to see if messages recently sent by us have been
1235          * received by the other side. (The remote GET value will have
1236          * changed since we last looked at it.)
1237          */
1238
1239         if (ch->w_remote_GP.get != ch->remote_GP.get) {
1240
1241                 /*
1242                  * We need to notify any senders that want to be notified
1243                  * that their sent messages have been received by their
1244                  * intended recipients. We need to do this before updating
1245                  * w_remote_GP.get so that we don't allocate the same message
1246                  * queue entries prematurely (see xpc_allocate_msg()).
1247                  */
1248                 if (atomic_read(&ch->n_to_notify) > 0) {
1249                         /*
1250                          * Notify senders that messages sent have been
1251                          * received and delivered by the other side.
1252                          */
1253                         xpc_notify_senders(ch, xpcMsgDelivered,
1254                                                         ch->remote_GP.get);
1255                 }
1256
1257                 /*
1258                  * Clear msg->flags in previously sent messages, so that
1259                  * they're ready for xpc_allocate_msg().
1260                  */
1261                 xpc_clear_local_msgqueue_flags(ch);
1262
1263                 ch->w_remote_GP.get = ch->remote_GP.get;
1264
1265                 dev_dbg(xpc_chan, "w_remote_GP.get changed to %ld, partid=%d, "
1266                         "channel=%d\n", ch->w_remote_GP.get, ch->partid,
1267                         ch->number);
1268
1269                 /*
1270                  * If anyone was waiting for message queue entries to become
1271                  * available, wake them up.
1272                  */
1273                 if (atomic_read(&ch->n_on_msg_allocate_wq) > 0) {
1274                         wake_up(&ch->msg_allocate_wq);
1275                 }
1276         }
1277
1278
1279         /*
1280          * Now check for newly sent messages by the other side. (The remote
1281          * PUT value will have changed since we last looked at it.)
1282          */
1283
1284         if (ch->w_remote_GP.put != ch->remote_GP.put) {
1285                 /*
1286                  * Clear msg->flags in previously received messages, so that
1287                  * they're ready for xpc_get_deliverable_msg().
1288                  */
1289                 xpc_clear_remote_msgqueue_flags(ch);
1290
1291                 ch->w_remote_GP.put = ch->remote_GP.put;
1292
1293                 dev_dbg(xpc_chan, "w_remote_GP.put changed to %ld, partid=%d, "
1294                         "channel=%d\n", ch->w_remote_GP.put, ch->partid,
1295                         ch->number);
1296
1297                 nmsgs_sent = ch->w_remote_GP.put - ch->w_local_GP.get;
1298                 if (nmsgs_sent > 0) {
1299                         dev_dbg(xpc_chan, "msgs waiting to be copied and "
1300                                 "delivered=%d, partid=%d, channel=%d\n",
1301                                 nmsgs_sent, ch->partid, ch->number);
1302
1303                         if (ch->flags & XPC_C_CONNECTCALLOUT) {
1304                                 xpc_activate_kthreads(ch, nmsgs_sent);
1305                         }
1306                 }
1307         }
1308
1309         xpc_msgqueue_deref(ch);
1310 }
1311
1312
1313 void
1314 xpc_process_channel_activity(struct xpc_partition *part)
1315 {
1316         unsigned long irq_flags;
1317         u64 IPI_amo, IPI_flags;
1318         struct xpc_channel *ch;
1319         int ch_number;
1320         u32 ch_flags;
1321
1322
1323         IPI_amo = xpc_get_IPI_flags(part);
1324
1325         /*
1326          * Initiate channel connections for registered channels.
1327          *
1328          * For each connected channel that has pending messages activate idle
1329          * kthreads and/or create new kthreads as needed.
1330          */
1331
1332         for (ch_number = 0; ch_number < part->nchannels; ch_number++) {
1333                 ch = &part->channels[ch_number];
1334
1335
1336                 /*
1337                  * Process any open or close related IPI flags, and then deal
1338                  * with connecting or disconnecting the channel as required.
1339                  */
1340
1341                 IPI_flags = XPC_GET_IPI_FLAGS(IPI_amo, ch_number);
1342
1343                 if (XPC_ANY_OPENCLOSE_IPI_FLAGS_SET(IPI_flags)) {
1344                         xpc_process_openclose_IPI(part, ch_number, IPI_flags);
1345                 }
1346
1347                 ch_flags = ch->flags;   /* need an atomic snapshot of flags */
1348
1349                 if (ch_flags & XPC_C_DISCONNECTING) {
1350                         spin_lock_irqsave(&ch->lock, irq_flags);
1351                         xpc_process_disconnect(ch, &irq_flags);
1352                         spin_unlock_irqrestore(&ch->lock, irq_flags);
1353                         continue;
1354                 }
1355
1356                 if (part->act_state == XPC_P_DEACTIVATING) {
1357                         continue;
1358                 }
1359
1360                 if (!(ch_flags & XPC_C_CONNECTED)) {
1361                         if (!(ch_flags & XPC_C_OPENREQUEST)) {
1362                                 DBUG_ON(ch_flags & XPC_C_SETUP);
1363                                 (void) xpc_connect_channel(ch);
1364                         } else {
1365                                 spin_lock_irqsave(&ch->lock, irq_flags);
1366                                 xpc_process_connect(ch, &irq_flags);
1367                                 spin_unlock_irqrestore(&ch->lock, irq_flags);
1368                         }
1369                         continue;
1370                 }
1371
1372
1373                 /*
1374                  * Process any message related IPI flags, this may involve the
1375                  * activation of kthreads to deliver any pending messages sent
1376                  * from the other partition.
1377                  */
1378
1379                 if (XPC_ANY_MSG_IPI_FLAGS_SET(IPI_flags)) {
1380                         xpc_process_msg_IPI(part, ch_number);
1381                 }
1382         }
1383 }
1384
1385
1386 /*
1387  * XPC's heartbeat code calls this function to inform XPC that a partition is
1388  * going down.  XPC responds by tearing down the XPartition Communication
1389  * infrastructure used for the just downed partition.
1390  *
1391  * XPC's heartbeat code will never call this function and xpc_partition_up()
1392  * at the same time. Nor will it ever make multiple calls to either function
1393  * at the same time.
1394  */
1395 void
1396 xpc_partition_going_down(struct xpc_partition *part, enum xpc_retval reason)
1397 {
1398         unsigned long irq_flags;
1399         int ch_number;
1400         struct xpc_channel *ch;
1401
1402
1403         dev_dbg(xpc_chan, "deactivating partition %d, reason=%d\n",
1404                 XPC_PARTID(part), reason);
1405
1406         if (!xpc_part_ref(part)) {
1407                 /* infrastructure for this partition isn't currently set up */
1408                 return;
1409         }
1410
1411
1412         /* disconnect channels associated with the partition going down */
1413
1414         for (ch_number = 0; ch_number < part->nchannels; ch_number++) {
1415                 ch = &part->channels[ch_number];
1416
1417                 xpc_msgqueue_ref(ch);
1418                 spin_lock_irqsave(&ch->lock, irq_flags);
1419
1420                 XPC_DISCONNECT_CHANNEL(ch, reason, &irq_flags);
1421
1422                 spin_unlock_irqrestore(&ch->lock, irq_flags);
1423                 xpc_msgqueue_deref(ch);
1424         }
1425
1426         xpc_wakeup_channel_mgr(part);
1427
1428         xpc_part_deref(part);
1429 }
1430
1431
1432 /*
1433  * Teardown the infrastructure necessary to support XPartition Communication
1434  * between the specified remote partition and the local one.
1435  */
1436 void
1437 xpc_teardown_infrastructure(struct xpc_partition *part)
1438 {
1439         partid_t partid = XPC_PARTID(part);
1440
1441
1442         /*
1443          * We start off by making this partition inaccessible to local
1444          * processes by marking it as no longer setup. Then we make it
1445          * inaccessible to remote processes by clearing the XPC per partition
1446          * specific variable's magic # (which indicates that these variables
1447          * are no longer valid) and by ignoring all XPC notify IPIs sent to
1448          * this partition.
1449          */
1450
1451         DBUG_ON(atomic_read(&part->nchannels_engaged) != 0);
1452         DBUG_ON(atomic_read(&part->nchannels_active) != 0);
1453         DBUG_ON(part->setup_state != XPC_P_SETUP);
1454         part->setup_state = XPC_P_WTEARDOWN;
1455
1456         xpc_vars_part[partid].magic = 0;
1457
1458
1459         free_irq(SGI_XPC_NOTIFY, (void *) (u64) partid);
1460
1461
1462         /*
1463          * Before proceding with the teardown we have to wait until all
1464          * existing references cease.
1465          */
1466         wait_event(part->teardown_wq, (atomic_read(&part->references) == 0));
1467
1468
1469         /* now we can begin tearing down the infrastructure */
1470
1471         part->setup_state = XPC_P_TORNDOWN;
1472
1473         /* in case we've still got outstanding timers registered... */
1474         del_timer_sync(&part->dropped_IPI_timer);
1475
1476         kfree(part->remote_openclose_args_base);
1477         part->remote_openclose_args = NULL;
1478         kfree(part->local_openclose_args_base);
1479         part->local_openclose_args = NULL;
1480         kfree(part->remote_GPs_base);
1481         part->remote_GPs = NULL;
1482         kfree(part->local_GPs_base);
1483         part->local_GPs = NULL;
1484         kfree(part->channels);
1485         part->channels = NULL;
1486         part->local_IPI_amo_va = NULL;
1487 }
1488
1489
1490 /*
1491  * Called by XP at the time of channel connection registration to cause
1492  * XPC to establish connections to all currently active partitions.
1493  */
1494 void
1495 xpc_initiate_connect(int ch_number)
1496 {
1497         partid_t partid;
1498         struct xpc_partition *part;
1499         struct xpc_channel *ch;
1500
1501
1502         DBUG_ON(ch_number < 0 || ch_number >= XPC_NCHANNELS);
1503
1504         for (partid = 1; partid < XP_MAX_PARTITIONS; partid++) {
1505                 part = &xpc_partitions[partid];
1506
1507                 if (xpc_part_ref(part)) {
1508                         ch = &part->channels[ch_number];
1509
1510                         /*
1511                          * Initiate the establishment of a connection on the
1512                          * newly registered channel to the remote partition.
1513                          */
1514                         xpc_wakeup_channel_mgr(part);
1515                         xpc_part_deref(part);
1516                 }
1517         }
1518 }
1519
1520
1521 void
1522 xpc_connected_callout(struct xpc_channel *ch)
1523 {
1524         /* let the registerer know that a connection has been established */
1525
1526         if (ch->func != NULL) {
1527                 dev_dbg(xpc_chan, "ch->func() called, reason=xpcConnected, "
1528                         "partid=%d, channel=%d\n", ch->partid, ch->number);
1529
1530                 ch->func(xpcConnected, ch->partid, ch->number,
1531                                 (void *) (u64) ch->local_nentries, ch->key);
1532
1533                 dev_dbg(xpc_chan, "ch->func() returned, reason=xpcConnected, "
1534                         "partid=%d, channel=%d\n", ch->partid, ch->number);
1535         }
1536 }
1537
1538
1539 /*
1540  * Called by XP at the time of channel connection unregistration to cause
1541  * XPC to teardown all current connections for the specified channel.
1542  *
1543  * Before returning xpc_initiate_disconnect() will wait until all connections
1544  * on the specified channel have been closed/torndown. So the caller can be
1545  * assured that they will not be receiving any more callouts from XPC to the
1546  * function they registered via xpc_connect().
1547  *
1548  * Arguments:
1549  *
1550  *      ch_number - channel # to unregister.
1551  */
1552 void
1553 xpc_initiate_disconnect(int ch_number)
1554 {
1555         unsigned long irq_flags;
1556         partid_t partid;
1557         struct xpc_partition *part;
1558         struct xpc_channel *ch;
1559
1560
1561         DBUG_ON(ch_number < 0 || ch_number >= XPC_NCHANNELS);
1562
1563         /* initiate the channel disconnect for every active partition */
1564         for (partid = 1; partid < XP_MAX_PARTITIONS; partid++) {
1565                 part = &xpc_partitions[partid];
1566
1567                 if (xpc_part_ref(part)) {
1568                         ch = &part->channels[ch_number];
1569                         xpc_msgqueue_ref(ch);
1570
1571                         spin_lock_irqsave(&ch->lock, irq_flags);
1572
1573                         if (!(ch->flags & XPC_C_DISCONNECTED)) {
1574                                 ch->flags |= XPC_C_WDISCONNECT;
1575
1576                                 XPC_DISCONNECT_CHANNEL(ch, xpcUnregistering,
1577                                                                 &irq_flags);
1578                         }
1579
1580                         spin_unlock_irqrestore(&ch->lock, irq_flags);
1581
1582                         xpc_msgqueue_deref(ch);
1583                         xpc_part_deref(part);
1584                 }
1585         }
1586
1587         xpc_disconnect_wait(ch_number);
1588 }
1589
1590
1591 /*
1592  * To disconnect a channel, and reflect it back to all who may be waiting.
1593  *
1594  * An OPEN is not allowed until XPC_C_DISCONNECTING is cleared by
1595  * xpc_process_disconnect(), and if set, XPC_C_WDISCONNECT is cleared by
1596  * xpc_disconnect_wait().
1597  *
1598  * THE CHANNEL IS TO BE LOCKED BY THE CALLER AND WILL REMAIN LOCKED UPON RETURN.
1599  */
1600 void
1601 xpc_disconnect_channel(const int line, struct xpc_channel *ch,
1602                         enum xpc_retval reason, unsigned long *irq_flags)
1603 {
1604         u32 channel_was_connected = (ch->flags & XPC_C_CONNECTED);
1605
1606
1607         DBUG_ON(!spin_is_locked(&ch->lock));
1608
1609         if (ch->flags & (XPC_C_DISCONNECTING | XPC_C_DISCONNECTED)) {
1610                 return;
1611         }
1612         DBUG_ON(!(ch->flags & (XPC_C_CONNECTING | XPC_C_CONNECTED)));
1613
1614         dev_dbg(xpc_chan, "reason=%d, line=%d, partid=%d, channel=%d\n",
1615                 reason, line, ch->partid, ch->number);
1616
1617         XPC_SET_REASON(ch, reason, line);
1618
1619         ch->flags |= (XPC_C_CLOSEREQUEST | XPC_C_DISCONNECTING);
1620         /* some of these may not have been set */
1621         ch->flags &= ~(XPC_C_OPENREQUEST | XPC_C_OPENREPLY |
1622                         XPC_C_ROPENREQUEST | XPC_C_ROPENREPLY |
1623                         XPC_C_CONNECTING | XPC_C_CONNECTED);
1624
1625         xpc_IPI_send_closerequest(ch, irq_flags);
1626
1627         if (channel_was_connected) {
1628                 ch->flags |= XPC_C_WASCONNECTED;
1629         }
1630
1631         spin_unlock_irqrestore(&ch->lock, *irq_flags);
1632
1633         /* wake all idle kthreads so they can exit */
1634         if (atomic_read(&ch->kthreads_idle) > 0) {
1635                 wake_up_all(&ch->idle_wq);
1636         }
1637
1638         /* wake those waiting to allocate an entry from the local msg queue */
1639         if (atomic_read(&ch->n_on_msg_allocate_wq) > 0) {
1640                 wake_up(&ch->msg_allocate_wq);
1641         }
1642
1643         spin_lock_irqsave(&ch->lock, *irq_flags);
1644 }
1645
1646
1647 void
1648 xpc_disconnect_callout(struct xpc_channel *ch, enum xpc_retval reason)
1649 {
1650         /*
1651          * Let the channel's registerer know that the channel is being
1652          * disconnected. We don't want to do this if the registerer was never
1653          * informed of a connection being made.
1654          */
1655
1656         if (ch->func != NULL) {
1657                 dev_dbg(xpc_chan, "ch->func() called, reason=%d, partid=%d, "
1658                         "channel=%d\n", reason, ch->partid, ch->number);
1659
1660                 ch->func(reason, ch->partid, ch->number, NULL, ch->key);
1661
1662                 dev_dbg(xpc_chan, "ch->func() returned, reason=%d, partid=%d, "
1663                         "channel=%d\n", reason, ch->partid, ch->number);
1664         }
1665 }
1666
1667
1668 /*
1669  * Wait for a message entry to become available for the specified channel,
1670  * but don't wait any longer than 1 jiffy.
1671  */
1672 static enum xpc_retval
1673 xpc_allocate_msg_wait(struct xpc_channel *ch)
1674 {
1675         enum xpc_retval ret;
1676
1677
1678         if (ch->flags & XPC_C_DISCONNECTING) {
1679                 DBUG_ON(ch->reason == xpcInterrupted);  // >>> Is this true?
1680                 return ch->reason;
1681         }
1682
1683         atomic_inc(&ch->n_on_msg_allocate_wq);
1684         ret = interruptible_sleep_on_timeout(&ch->msg_allocate_wq, 1);
1685         atomic_dec(&ch->n_on_msg_allocate_wq);
1686
1687         if (ch->flags & XPC_C_DISCONNECTING) {
1688                 ret = ch->reason;
1689                 DBUG_ON(ch->reason == xpcInterrupted);  // >>> Is this true?
1690         } else if (ret == 0) {
1691                 ret = xpcTimeout;
1692         } else {
1693                 ret = xpcInterrupted;
1694         }
1695
1696         return ret;
1697 }
1698
1699
1700 /*
1701  * Allocate an entry for a message from the message queue associated with the
1702  * specified channel.
1703  */
1704 static enum xpc_retval
1705 xpc_allocate_msg(struct xpc_channel *ch, u32 flags,
1706                         struct xpc_msg **address_of_msg)
1707 {
1708         struct xpc_msg *msg;
1709         enum xpc_retval ret;
1710         s64 put;
1711
1712
1713         /* this reference will be dropped in xpc_send_msg() */
1714         xpc_msgqueue_ref(ch);
1715
1716         if (ch->flags & XPC_C_DISCONNECTING) {
1717                 xpc_msgqueue_deref(ch);
1718                 return ch->reason;
1719         }
1720         if (!(ch->flags & XPC_C_CONNECTED)) {
1721                 xpc_msgqueue_deref(ch);
1722                 return xpcNotConnected;
1723         }
1724
1725
1726         /*
1727          * Get the next available message entry from the local message queue.
1728          * If none are available, we'll make sure that we grab the latest
1729          * GP values.
1730          */
1731         ret = xpcTimeout;
1732
1733         while (1) {
1734
1735                 put = (volatile s64) ch->w_local_GP.put;
1736                 if (put - (volatile s64) ch->w_remote_GP.get <
1737                                                         ch->local_nentries) {
1738
1739                         /* There are available message entries. We need to try
1740                          * to secure one for ourselves. We'll do this by trying
1741                          * to increment w_local_GP.put as long as someone else
1742                          * doesn't beat us to it. If they do, we'll have to
1743                          * try again.
1744                          */
1745                         if (cmpxchg(&ch->w_local_GP.put, put, put + 1) ==
1746                                                                         put) {
1747                                 /* we got the entry referenced by put */
1748                                 break;
1749                         }
1750                         continue;       /* try again */
1751                 }
1752
1753
1754                 /*
1755                  * There aren't any available msg entries at this time.
1756                  *
1757                  * In waiting for a message entry to become available,
1758                  * we set a timeout in case the other side is not
1759                  * sending completion IPIs. This lets us fake an IPI
1760                  * that will cause the IPI handler to fetch the latest
1761                  * GP values as if an IPI was sent by the other side.
1762                  */
1763                 if (ret == xpcTimeout) {
1764                         xpc_IPI_send_local_msgrequest(ch);
1765                 }
1766
1767                 if (flags & XPC_NOWAIT) {
1768                         xpc_msgqueue_deref(ch);
1769                         return xpcNoWait;
1770                 }
1771
1772                 ret = xpc_allocate_msg_wait(ch);
1773                 if (ret != xpcInterrupted && ret != xpcTimeout) {
1774                         xpc_msgqueue_deref(ch);
1775                         return ret;
1776                 }
1777         }
1778
1779
1780         /* get the message's address and initialize it */
1781         msg = (struct xpc_msg *) ((u64) ch->local_msgqueue +
1782                                 (put % ch->local_nentries) * ch->msg_size);
1783
1784
1785         DBUG_ON(msg->flags != 0);
1786         msg->number = put;
1787
1788         dev_dbg(xpc_chan, "w_local_GP.put changed to %ld; msg=0x%p, "
1789                 "msg_number=%ld, partid=%d, channel=%d\n", put + 1,
1790                 (void *) msg, msg->number, ch->partid, ch->number);
1791
1792         *address_of_msg = msg;
1793
1794         return xpcSuccess;
1795 }
1796
1797
1798 /*
1799  * Allocate an entry for a message from the message queue associated with the
1800  * specified channel. NOTE that this routine can sleep waiting for a message
1801  * entry to become available. To not sleep, pass in the XPC_NOWAIT flag.
1802  *
1803  * Arguments:
1804  *
1805  *      partid - ID of partition to which the channel is connected.
1806  *      ch_number - channel #.
1807  *      flags - see xpc.h for valid flags.
1808  *      payload - address of the allocated payload area pointer (filled in on
1809  *                return) in which the user-defined message is constructed.
1810  */
1811 enum xpc_retval
1812 xpc_initiate_allocate(partid_t partid, int ch_number, u32 flags, void **payload)
1813 {
1814         struct xpc_partition *part = &xpc_partitions[partid];
1815         enum xpc_retval ret = xpcUnknownReason;
1816         struct xpc_msg *msg;
1817
1818
1819         DBUG_ON(partid <= 0 || partid >= XP_MAX_PARTITIONS);
1820         DBUG_ON(ch_number < 0 || ch_number >= part->nchannels);
1821
1822         *payload = NULL;
1823
1824         if (xpc_part_ref(part)) {
1825                 ret = xpc_allocate_msg(&part->channels[ch_number], flags, &msg);
1826                 xpc_part_deref(part);
1827
1828                 if (msg != NULL) {
1829                         *payload = &msg->payload;
1830                 }
1831         }
1832
1833         return ret;
1834 }
1835
1836
1837 /*
1838  * Now we actually send the messages that are ready to be sent by advancing
1839  * the local message queue's Put value and then send an IPI to the recipient
1840  * partition.
1841  */
1842 static void
1843 xpc_send_msgs(struct xpc_channel *ch, s64 initial_put)
1844 {
1845         struct xpc_msg *msg;
1846         s64 put = initial_put + 1;
1847         int send_IPI = 0;
1848
1849
1850         while (1) {
1851
1852                 while (1) {
1853                         if (put == (volatile s64) ch->w_local_GP.put) {
1854                                 break;
1855                         }
1856
1857                         msg = (struct xpc_msg *) ((u64) ch->local_msgqueue +
1858                                (put % ch->local_nentries) * ch->msg_size);
1859
1860                         if (!(msg->flags & XPC_M_READY)) {
1861                                 break;
1862                         }
1863
1864                         put++;
1865                 }
1866
1867                 if (put == initial_put) {
1868                         /* nothing's changed */
1869                         break;
1870                 }
1871
1872                 if (cmpxchg_rel(&ch->local_GP->put, initial_put, put) !=
1873                                                                 initial_put) {
1874                         /* someone else beat us to it */
1875                         DBUG_ON((volatile s64) ch->local_GP->put < initial_put);
1876                         break;
1877                 }
1878
1879                 /* we just set the new value of local_GP->put */
1880
1881                 dev_dbg(xpc_chan, "local_GP->put changed to %ld, partid=%d, "
1882                         "channel=%d\n", put, ch->partid, ch->number);
1883
1884                 send_IPI = 1;
1885
1886                 /*
1887                  * We need to ensure that the message referenced by
1888                  * local_GP->put is not XPC_M_READY or that local_GP->put
1889                  * equals w_local_GP.put, so we'll go have a look.
1890                  */
1891                 initial_put = put;
1892         }
1893
1894         if (send_IPI) {
1895                 xpc_IPI_send_msgrequest(ch);
1896         }
1897 }
1898
1899
1900 /*
1901  * Common code that does the actual sending of the message by advancing the
1902  * local message queue's Put value and sends an IPI to the partition the
1903  * message is being sent to.
1904  */
1905 static enum xpc_retval
1906 xpc_send_msg(struct xpc_channel *ch, struct xpc_msg *msg, u8 notify_type,
1907                         xpc_notify_func func, void *key)
1908 {
1909         enum xpc_retval ret = xpcSuccess;
1910         struct xpc_notify *notify = notify;
1911         s64 put, msg_number = msg->number;
1912
1913
1914         DBUG_ON(notify_type == XPC_N_CALL && func == NULL);
1915         DBUG_ON((((u64) msg - (u64) ch->local_msgqueue) / ch->msg_size) !=
1916                                         msg_number % ch->local_nentries);
1917         DBUG_ON(msg->flags & XPC_M_READY);
1918
1919         if (ch->flags & XPC_C_DISCONNECTING) {
1920                 /* drop the reference grabbed in xpc_allocate_msg() */
1921                 xpc_msgqueue_deref(ch);
1922                 return ch->reason;
1923         }
1924
1925         if (notify_type != 0) {
1926                 /*
1927                  * Tell the remote side to send an ACK interrupt when the
1928                  * message has been delivered.
1929                  */
1930                 msg->flags |= XPC_M_INTERRUPT;
1931
1932                 atomic_inc(&ch->n_to_notify);
1933
1934                 notify = &ch->notify_queue[msg_number % ch->local_nentries];
1935                 notify->func = func;
1936                 notify->key = key;
1937                 notify->type = notify_type;
1938
1939                 // >>> is a mb() needed here?
1940
1941                 if (ch->flags & XPC_C_DISCONNECTING) {
1942                         /*
1943                          * An error occurred between our last error check and
1944                          * this one. We will try to clear the type field from
1945                          * the notify entry. If we succeed then
1946                          * xpc_disconnect_channel() didn't already process
1947                          * the notify entry.
1948                          */
1949                         if (cmpxchg(&notify->type, notify_type, 0) ==
1950                                                                 notify_type) {
1951                                 atomic_dec(&ch->n_to_notify);
1952                                 ret = ch->reason;
1953                         }
1954
1955                         /* drop the reference grabbed in xpc_allocate_msg() */
1956                         xpc_msgqueue_deref(ch);
1957                         return ret;
1958                 }
1959         }
1960
1961         msg->flags |= XPC_M_READY;
1962
1963         /*
1964          * The preceding store of msg->flags must occur before the following
1965          * load of ch->local_GP->put.
1966          */
1967         mb();
1968
1969         /* see if the message is next in line to be sent, if so send it */
1970
1971         put = ch->local_GP->put;
1972         if (put == msg_number) {
1973                 xpc_send_msgs(ch, put);
1974         }
1975
1976         /* drop the reference grabbed in xpc_allocate_msg() */
1977         xpc_msgqueue_deref(ch);
1978         return ret;
1979 }
1980
1981
1982 /*
1983  * Send a message previously allocated using xpc_initiate_allocate() on the
1984  * specified channel connected to the specified partition.
1985  *
1986  * This routine will not wait for the message to be received, nor will
1987  * notification be given when it does happen. Once this routine has returned
1988  * the message entry allocated via xpc_initiate_allocate() is no longer
1989  * accessable to the caller.
1990  *
1991  * This routine, although called by users, does not call xpc_part_ref() to
1992  * ensure that the partition infrastructure is in place. It relies on the
1993  * fact that we called xpc_msgqueue_ref() in xpc_allocate_msg().
1994  *
1995  * Arguments:
1996  *
1997  *      partid - ID of partition to which the channel is connected.
1998  *      ch_number - channel # to send message on.
1999  *      payload - pointer to the payload area allocated via
2000  *                      xpc_initiate_allocate().
2001  */
2002 enum xpc_retval
2003 xpc_initiate_send(partid_t partid, int ch_number, void *payload)
2004 {
2005         struct xpc_partition *part = &xpc_partitions[partid];
2006         struct xpc_msg *msg = XPC_MSG_ADDRESS(payload);
2007         enum xpc_retval ret;
2008
2009
2010         dev_dbg(xpc_chan, "msg=0x%p, partid=%d, channel=%d\n", (void *) msg,
2011                 partid, ch_number);
2012
2013         DBUG_ON(partid <= 0 || partid >= XP_MAX_PARTITIONS);
2014         DBUG_ON(ch_number < 0 || ch_number >= part->nchannels);
2015         DBUG_ON(msg == NULL);
2016
2017         ret = xpc_send_msg(&part->channels[ch_number], msg, 0, NULL, NULL);
2018
2019         return ret;
2020 }
2021
2022
2023 /*
2024  * Send a message previously allocated using xpc_initiate_allocate on the
2025  * specified channel connected to the specified partition.
2026  *
2027  * This routine will not wait for the message to be sent. Once this routine
2028  * has returned the message entry allocated via xpc_initiate_allocate() is no
2029  * longer accessable to the caller.
2030  *
2031  * Once the remote end of the channel has received the message, the function
2032  * passed as an argument to xpc_initiate_send_notify() will be called. This
2033  * allows the sender to free up or re-use any buffers referenced by the
2034  * message, but does NOT mean the message has been processed at the remote
2035  * end by a receiver.
2036  *
2037  * If this routine returns an error, the caller's function will NOT be called.
2038  *
2039  * This routine, although called by users, does not call xpc_part_ref() to
2040  * ensure that the partition infrastructure is in place. It relies on the
2041  * fact that we called xpc_msgqueue_ref() in xpc_allocate_msg().
2042  *
2043  * Arguments:
2044  *
2045  *      partid - ID of partition to which the channel is connected.
2046  *      ch_number - channel # to send message on.
2047  *      payload - pointer to the payload area allocated via
2048  *                      xpc_initiate_allocate().
2049  *      func - function to call with asynchronous notification of message
2050  *                receipt. THIS FUNCTION MUST BE NON-BLOCKING.
2051  *      key - user-defined key to be passed to the function when it's called.
2052  */
2053 enum xpc_retval
2054 xpc_initiate_send_notify(partid_t partid, int ch_number, void *payload,
2055                                 xpc_notify_func func, void *key)
2056 {
2057         struct xpc_partition *part = &xpc_partitions[partid];
2058         struct xpc_msg *msg = XPC_MSG_ADDRESS(payload);
2059         enum xpc_retval ret;
2060
2061
2062         dev_dbg(xpc_chan, "msg=0x%p, partid=%d, channel=%d\n", (void *) msg,
2063                 partid, ch_number);
2064
2065         DBUG_ON(partid <= 0 || partid >= XP_MAX_PARTITIONS);
2066         DBUG_ON(ch_number < 0 || ch_number >= part->nchannels);
2067         DBUG_ON(msg == NULL);
2068         DBUG_ON(func == NULL);
2069
2070         ret = xpc_send_msg(&part->channels[ch_number], msg, XPC_N_CALL,
2071                                                                 func, key);
2072         return ret;
2073 }
2074
2075
2076 static struct xpc_msg *
2077 xpc_pull_remote_msg(struct xpc_channel *ch, s64 get)
2078 {
2079         struct xpc_partition *part = &xpc_partitions[ch->partid];
2080         struct xpc_msg *remote_msg, *msg;
2081         u32 msg_index, nmsgs;
2082         u64 msg_offset;
2083         enum xpc_retval ret;
2084
2085
2086         if (mutex_lock_interruptible(&ch->msg_to_pull_mutex) != 0) {
2087                 /* we were interrupted by a signal */
2088                 return NULL;
2089         }
2090
2091         while (get >= ch->next_msg_to_pull) {
2092
2093                 /* pull as many messages as are ready and able to be pulled */
2094
2095                 msg_index = ch->next_msg_to_pull % ch->remote_nentries;
2096
2097                 DBUG_ON(ch->next_msg_to_pull >=
2098                                         (volatile s64) ch->w_remote_GP.put);
2099                 nmsgs =  (volatile s64) ch->w_remote_GP.put -
2100                                                 ch->next_msg_to_pull;
2101                 if (msg_index + nmsgs > ch->remote_nentries) {
2102                         /* ignore the ones that wrap the msg queue for now */
2103                         nmsgs = ch->remote_nentries - msg_index;
2104                 }
2105
2106                 msg_offset = msg_index * ch->msg_size;
2107                 msg = (struct xpc_msg *) ((u64) ch->remote_msgqueue +
2108                                                                 msg_offset);
2109                 remote_msg = (struct xpc_msg *) (ch->remote_msgqueue_pa +
2110                                                                 msg_offset);
2111
2112                 if ((ret = xpc_pull_remote_cachelines(part, msg, remote_msg,
2113                                 nmsgs * ch->msg_size)) != xpcSuccess) {
2114
2115                         dev_dbg(xpc_chan, "failed to pull %d msgs starting with"
2116                                 " msg %ld from partition %d, channel=%d, "
2117                                 "ret=%d\n", nmsgs, ch->next_msg_to_pull,
2118                                 ch->partid, ch->number, ret);
2119
2120                         XPC_DEACTIVATE_PARTITION(part, ret);
2121
2122                         mutex_unlock(&ch->msg_to_pull_mutex);
2123                         return NULL;
2124                 }
2125
2126                 mb();   /* >>> this may not be needed, we're not sure */
2127
2128                 ch->next_msg_to_pull += nmsgs;
2129         }
2130
2131         mutex_unlock(&ch->msg_to_pull_mutex);
2132
2133         /* return the message we were looking for */
2134         msg_offset = (get % ch->remote_nentries) * ch->msg_size;
2135         msg = (struct xpc_msg *) ((u64) ch->remote_msgqueue + msg_offset);
2136
2137         return msg;
2138 }
2139
2140
2141 /*
2142  * Get a message to be delivered.
2143  */
2144 static struct xpc_msg *
2145 xpc_get_deliverable_msg(struct xpc_channel *ch)
2146 {
2147         struct xpc_msg *msg = NULL;
2148         s64 get;
2149
2150
2151         do {
2152                 if ((volatile u32) ch->flags & XPC_C_DISCONNECTING) {
2153                         break;
2154                 }
2155
2156                 get = (volatile s64) ch->w_local_GP.get;
2157                 if (get == (volatile s64) ch->w_remote_GP.put) {
2158                         break;
2159                 }
2160
2161                 /* There are messages waiting to be pulled and delivered.
2162                  * We need to try to secure one for ourselves. We'll do this
2163                  * by trying to increment w_local_GP.get and hope that no one
2164                  * else beats us to it. If they do, we'll we'll simply have
2165                  * to try again for the next one.
2166                  */
2167
2168                 if (cmpxchg(&ch->w_local_GP.get, get, get + 1) == get) {
2169                         /* we got the entry referenced by get */
2170
2171                         dev_dbg(xpc_chan, "w_local_GP.get changed to %ld, "
2172                                 "partid=%d, channel=%d\n", get + 1,
2173                                 ch->partid, ch->number);
2174
2175                         /* pull the message from the remote partition */
2176
2177                         msg = xpc_pull_remote_msg(ch, get);
2178
2179                         DBUG_ON(msg != NULL && msg->number != get);
2180                         DBUG_ON(msg != NULL && (msg->flags & XPC_M_DONE));
2181                         DBUG_ON(msg != NULL && !(msg->flags & XPC_M_READY));
2182
2183                         break;
2184                 }
2185
2186         } while (1);
2187
2188         return msg;
2189 }
2190
2191
2192 /*
2193  * Deliver a message to its intended recipient.
2194  */
2195 void
2196 xpc_deliver_msg(struct xpc_channel *ch)
2197 {
2198         struct xpc_msg *msg;
2199
2200
2201         if ((msg = xpc_get_deliverable_msg(ch)) != NULL) {
2202
2203                 /*
2204                  * This ref is taken to protect the payload itself from being
2205                  * freed before the user is finished with it, which the user
2206                  * indicates by calling xpc_initiate_received().
2207                  */
2208                 xpc_msgqueue_ref(ch);
2209
2210                 atomic_inc(&ch->kthreads_active);
2211
2212                 if (ch->func != NULL) {
2213                         dev_dbg(xpc_chan, "ch->func() called, msg=0x%p, "
2214                                 "msg_number=%ld, partid=%d, channel=%d\n",
2215                                 (void *) msg, msg->number, ch->partid,
2216                                 ch->number);
2217
2218                         /* deliver the message to its intended recipient */
2219                         ch->func(xpcMsgReceived, ch->partid, ch->number,
2220                                         &msg->payload, ch->key);
2221
2222                         dev_dbg(xpc_chan, "ch->func() returned, msg=0x%p, "
2223                                 "msg_number=%ld, partid=%d, channel=%d\n",
2224                                 (void *) msg, msg->number, ch->partid,
2225                                 ch->number);
2226                 }
2227
2228                 atomic_dec(&ch->kthreads_active);
2229         }
2230 }
2231
2232
2233 /*
2234  * Now we actually acknowledge the messages that have been delivered and ack'd
2235  * by advancing the cached remote message queue's Get value and if requested
2236  * send an IPI to the message sender's partition.
2237  */
2238 static void
2239 xpc_acknowledge_msgs(struct xpc_channel *ch, s64 initial_get, u8 msg_flags)
2240 {
2241         struct xpc_msg *msg;
2242         s64 get = initial_get + 1;
2243         int send_IPI = 0;
2244
2245
2246         while (1) {
2247
2248                 while (1) {
2249                         if (get == (volatile s64) ch->w_local_GP.get) {
2250                                 break;
2251                         }
2252
2253                         msg = (struct xpc_msg *) ((u64) ch->remote_msgqueue +
2254                                (get % ch->remote_nentries) * ch->msg_size);
2255
2256                         if (!(msg->flags & XPC_M_DONE)) {
2257                                 break;
2258                         }
2259
2260                         msg_flags |= msg->flags;
2261                         get++;
2262                 }
2263
2264                 if (get == initial_get) {
2265                         /* nothing's changed */
2266                         break;
2267                 }
2268
2269                 if (cmpxchg_rel(&ch->local_GP->get, initial_get, get) !=
2270                                                                 initial_get) {
2271                         /* someone else beat us to it */
2272                         DBUG_ON((volatile s64) ch->local_GP->get <=
2273                                                                 initial_get);
2274                         break;
2275                 }
2276
2277                 /* we just set the new value of local_GP->get */
2278
2279                 dev_dbg(xpc_chan, "local_GP->get changed to %ld, partid=%d, "
2280                         "channel=%d\n", get, ch->partid, ch->number);
2281
2282                 send_IPI = (msg_flags & XPC_M_INTERRUPT);
2283
2284                 /*
2285                  * We need to ensure that the message referenced by
2286                  * local_GP->get is not XPC_M_DONE or that local_GP->get
2287                  * equals w_local_GP.get, so we'll go have a look.
2288                  */
2289                 initial_get = get;
2290         }
2291
2292         if (send_IPI) {
2293                 xpc_IPI_send_msgrequest(ch);
2294         }
2295 }
2296
2297
2298 /*
2299  * Acknowledge receipt of a delivered message.
2300  *
2301  * If a message has XPC_M_INTERRUPT set, send an interrupt to the partition
2302  * that sent the message.
2303  *
2304  * This function, although called by users, does not call xpc_part_ref() to
2305  * ensure that the partition infrastructure is in place. It relies on the
2306  * fact that we called xpc_msgqueue_ref() in xpc_deliver_msg().
2307  *
2308  * Arguments:
2309  *
2310  *      partid - ID of partition to which the channel is connected.
2311  *      ch_number - channel # message received on.
2312  *      payload - pointer to the payload area allocated via
2313  *                      xpc_initiate_allocate().
2314  */
2315 void
2316 xpc_initiate_received(partid_t partid, int ch_number, void *payload)
2317 {
2318         struct xpc_partition *part = &xpc_partitions[partid];
2319         struct xpc_channel *ch;
2320         struct xpc_msg *msg = XPC_MSG_ADDRESS(payload);
2321         s64 get, msg_number = msg->number;
2322
2323
2324         DBUG_ON(partid <= 0 || partid >= XP_MAX_PARTITIONS);
2325         DBUG_ON(ch_number < 0 || ch_number >= part->nchannels);
2326
2327         ch = &part->channels[ch_number];
2328
2329         dev_dbg(xpc_chan, "msg=0x%p, msg_number=%ld, partid=%d, channel=%d\n",
2330                 (void *) msg, msg_number, ch->partid, ch->number);
2331
2332         DBUG_ON((((u64) msg - (u64) ch->remote_msgqueue) / ch->msg_size) !=
2333                                         msg_number % ch->remote_nentries);
2334         DBUG_ON(msg->flags & XPC_M_DONE);
2335
2336         msg->flags |= XPC_M_DONE;
2337
2338         /*
2339          * The preceding store of msg->flags must occur before the following
2340          * load of ch->local_GP->get.
2341          */
2342         mb();
2343
2344         /*
2345          * See if this message is next in line to be acknowledged as having
2346          * been delivered.
2347          */
2348         get = ch->local_GP->get;
2349         if (get == msg_number) {
2350                 xpc_acknowledge_msgs(ch, get, msg->flags);
2351         }
2352
2353         /* the call to xpc_msgqueue_ref() was done by xpc_deliver_msg()  */
2354         xpc_msgqueue_deref(ch);
2355 }
2356