relay.c 29.0 KB
Newer Older
1 2 3
/*
 * Public API and common code for kernel->userspace relay file support.
 *
4
 * See Documentation/filesystems/relay.txt for an overview.
5 6 7 8 9
 *
 * Copyright (C) 2002-2005 - Tom Zanussi (zanussi@us.ibm.com), IBM Corp
 * Copyright (C) 1999-2005 - Karim Yaghmour (karim@opersys.com)
 *
 * Moved to kernel/relay.c by Paul Mundt, 2006.
10 11
 * November 2006 - CPU hotplug support by Mathieu Desnoyers
 * 	(mathieu.desnoyers@polymtl.ca)
12 13 14 15 16 17 18 19 20 21 22
 *
 * This file is released under the GPL.
 */
#include <linux/errno.h>
#include <linux/stddef.h>
#include <linux/slab.h>
#include <linux/module.h>
#include <linux/string.h>
#include <linux/relay.h>
#include <linux/vmalloc.h>
#include <linux/mm.h>
23
#include <linux/cpu.h>
24
#include <linux/splice.h>
25 26 27 28

/* list of open channels, for cpu hotplug */
static DEFINE_MUTEX(relay_channels_mutex);
static LIST_HEAD(relay_channels);
29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82

/*
 * close() vm_op implementation for relay file mapping.
 */
static void relay_file_mmap_close(struct vm_area_struct *vma)
{
	struct rchan_buf *buf = vma->vm_private_data;
	buf->chan->cb->buf_unmapped(buf, vma->vm_file);
}

/*
 * nopage() vm_op implementation for relay file mapping.
 */
static struct page *relay_buf_nopage(struct vm_area_struct *vma,
				     unsigned long address,
				     int *type)
{
	struct page *page;
	struct rchan_buf *buf = vma->vm_private_data;
	unsigned long offset = address - vma->vm_start;

	if (address > vma->vm_end)
		return NOPAGE_SIGBUS; /* Disallow mremap */
	if (!buf)
		return NOPAGE_OOM;

	page = vmalloc_to_page(buf->start + offset);
	if (!page)
		return NOPAGE_OOM;
	get_page(page);

	if (type)
		*type = VM_FAULT_MINOR;

	return page;
}

/*
 * vm_ops for relay file mappings.
 */
static struct vm_operations_struct relay_file_mmap_ops = {
	.nopage = relay_buf_nopage,
	.close = relay_file_mmap_close,
};

/**
 *	relay_mmap_buf: - mmap channel buffer to process address space
 *	@buf: relay channel buffer
 *	@vma: vm_area_struct describing memory to be mapped
 *
 *	Returns 0 if ok, negative on error
 *
 *	Caller should already have grabbed mmap_sem.
 */
83
static int relay_mmap_buf(struct rchan_buf *buf, struct vm_area_struct *vma)
84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105
{
	unsigned long length = vma->vm_end - vma->vm_start;
	struct file *filp = vma->vm_file;

	if (!buf)
		return -EBADF;

	if (length != (unsigned long)buf->chan->alloc_size)
		return -EINVAL;

	vma->vm_ops = &relay_file_mmap_ops;
	vma->vm_private_data = buf;
	buf->chan->cb->buf_mapped(buf, filp);

	return 0;
}

/**
 *	relay_alloc_buf - allocate a channel buffer
 *	@buf: the buffer struct
 *	@size: total size of the buffer
 *
106
 *	Returns a pointer to the resulting buffer, %NULL if unsuccessful. The
107
 *	passed in size will get page aligned, if it isn't already.
108
 */
109
static void *relay_alloc_buf(struct rchan_buf *buf, size_t *size)
110 111 112 113
{
	void *mem;
	unsigned int i, j, n_pages;

114 115
	*size = PAGE_ALIGN(*size);
	n_pages = *size >> PAGE_SHIFT;
116 117 118 119 120 121 122 123 124

	buf->page_array = kcalloc(n_pages, sizeof(struct page *), GFP_KERNEL);
	if (!buf->page_array)
		return NULL;

	for (i = 0; i < n_pages; i++) {
		buf->page_array[i] = alloc_page(GFP_KERNEL);
		if (unlikely(!buf->page_array[i]))
			goto depopulate;
T
Tom Zanussi 已提交
125
		set_page_private(buf->page_array[i], (unsigned long)buf);
126 127 128 129 130
	}
	mem = vmap(buf->page_array, n_pages, VM_MAP, PAGE_KERNEL);
	if (!mem)
		goto depopulate;

131
	memset(mem, 0, *size);
132 133 134 135 136 137 138 139 140 141 142 143
	buf->page_count = n_pages;
	return mem;

depopulate:
	for (j = 0; j < i; j++)
		__free_page(buf->page_array[j]);
	kfree(buf->page_array);
	return NULL;
}

/**
 *	relay_create_buf - allocate and initialize a channel buffer
144
 *	@chan: the relay channel
145
 *
146
 *	Returns channel buffer if successful, %NULL otherwise.
147
 */
148
static struct rchan_buf *relay_create_buf(struct rchan *chan)
149
{
150
	struct rchan_buf *buf = kzalloc(sizeof(struct rchan_buf), GFP_KERNEL);
151 152 153 154 155 156 157
	if (!buf)
		return NULL;

	buf->padding = kmalloc(chan->n_subbufs * sizeof(size_t *), GFP_KERNEL);
	if (!buf->padding)
		goto free_buf;

158
	buf->start = relay_alloc_buf(buf, &chan->alloc_size);
159 160 161 162 163 164 165 166 167 168 169 170 171 172 173
	if (!buf->start)
		goto free_buf;

	buf->chan = chan;
	kref_get(&buf->chan->kref);
	return buf;

free_buf:
	kfree(buf->padding);
	kfree(buf);
	return NULL;
}

/**
 *	relay_destroy_channel - free the channel struct
174
 *	@kref: target kernel reference that contains the relay channel
175 176 177
 *
 *	Should only be called from kref_put().
 */
178
static void relay_destroy_channel(struct kref *kref)
179 180 181 182 183 184 185 186 187
{
	struct rchan *chan = container_of(kref, struct rchan, kref);
	kfree(chan);
}

/**
 *	relay_destroy_buf - destroy an rchan_buf struct and associated buffer
 *	@buf: the buffer struct
 */
188
static void relay_destroy_buf(struct rchan_buf *buf)
189 190 191 192 193 194 195 196 197 198
{
	struct rchan *chan = buf->chan;
	unsigned int i;

	if (likely(buf->start)) {
		vunmap(buf->start);
		for (i = 0; i < buf->page_count; i++)
			__free_page(buf->page_array[i]);
		kfree(buf->page_array);
	}
199
	chan->buf[buf->cpu] = NULL;
200 201 202 203 204 205 206
	kfree(buf->padding);
	kfree(buf);
	kref_put(&chan->kref, relay_destroy_channel);
}

/**
 *	relay_remove_buf - remove a channel buffer
207
 *	@kref: target kernel reference that contains the relay buffer
208 209 210 211 212
 *
 *	Removes the file from the fileystem, which also frees the
 *	rchan_buf_struct and the channel buffer.  Should only be called from
 *	kref_put().
 */
213
static void relay_remove_buf(struct kref *kref)
214 215 216 217 218 219 220 221 222 223 224 225
{
	struct rchan_buf *buf = container_of(kref, struct rchan_buf, kref);
	buf->chan->cb->remove_buf_file(buf->dentry);
	relay_destroy_buf(buf);
}

/**
 *	relay_buf_empty - boolean, is the channel buffer empty?
 *	@buf: channel buffer
 *
 *	Returns 1 if the buffer is empty, 0 otherwise.
 */
226
static int relay_buf_empty(struct rchan_buf *buf)
227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313
{
	return (buf->subbufs_produced - buf->subbufs_consumed) ? 0 : 1;
}

/**
 *	relay_buf_full - boolean, is the channel buffer full?
 *	@buf: channel buffer
 *
 *	Returns 1 if the buffer is full, 0 otherwise.
 */
int relay_buf_full(struct rchan_buf *buf)
{
	size_t ready = buf->subbufs_produced - buf->subbufs_consumed;
	return (ready >= buf->chan->n_subbufs) ? 1 : 0;
}
EXPORT_SYMBOL_GPL(relay_buf_full);

/*
 * High-level relay kernel API and associated functions.
 */

/*
 * rchan_callback implementations defining default channel behavior.  Used
 * in place of corresponding NULL values in client callback struct.
 */

/*
 * subbuf_start() default callback.  Does nothing.
 */
static int subbuf_start_default_callback (struct rchan_buf *buf,
					  void *subbuf,
					  void *prev_subbuf,
					  size_t prev_padding)
{
	if (relay_buf_full(buf))
		return 0;

	return 1;
}

/*
 * buf_mapped() default callback.  Does nothing.
 */
static void buf_mapped_default_callback(struct rchan_buf *buf,
					struct file *filp)
{
}

/*
 * buf_unmapped() default callback.  Does nothing.
 */
static void buf_unmapped_default_callback(struct rchan_buf *buf,
					  struct file *filp)
{
}

/*
 * create_buf_file_create() default callback.  Does nothing.
 */
static struct dentry *create_buf_file_default_callback(const char *filename,
						       struct dentry *parent,
						       int mode,
						       struct rchan_buf *buf,
						       int *is_global)
{
	return NULL;
}

/*
 * remove_buf_file() default callback.  Does nothing.
 */
static int remove_buf_file_default_callback(struct dentry *dentry)
{
	return -EINVAL;
}

/* relay channel default callbacks */
static struct rchan_callbacks default_channel_callbacks = {
	.subbuf_start = subbuf_start_default_callback,
	.buf_mapped = buf_mapped_default_callback,
	.buf_unmapped = buf_unmapped_default_callback,
	.create_buf_file = create_buf_file_default_callback,
	.remove_buf_file = remove_buf_file_default_callback,
};

/**
 *	wakeup_readers - wake up readers waiting on a channel
314
 *	@data: contains the channel buffer
315
 *
316
 *	This is the timer function used to defer reader waking.
317
 */
318
static void wakeup_readers(unsigned long data)
319
{
320
	struct rchan_buf *buf = (struct rchan_buf *)data;
321 322 323 324 325 326 327 328
	wake_up_interruptible(&buf->read_wait);
}

/**
 *	__relay_reset - reset a channel buffer
 *	@buf: the channel buffer
 *	@init: 1 if this is a first-time initialization
 *
329
 *	See relay_reset() for description of effect.
330
 */
A
Andrew Morton 已提交
331
static void __relay_reset(struct rchan_buf *buf, unsigned int init)
332 333 334 335 336 337
{
	size_t i;

	if (init) {
		init_waitqueue_head(&buf->read_wait);
		kref_init(&buf->kref);
338 339 340
		setup_timer(&buf->timer, wakeup_readers, (unsigned long)buf);
	} else
		del_timer_sync(&buf->timer);
341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362

	buf->subbufs_produced = 0;
	buf->subbufs_consumed = 0;
	buf->bytes_consumed = 0;
	buf->finalized = 0;
	buf->data = buf->start;
	buf->offset = 0;

	for (i = 0; i < buf->chan->n_subbufs; i++)
		buf->padding[i] = 0;

	buf->chan->cb->subbuf_start(buf, buf->data, NULL, 0);
}

/**
 *	relay_reset - reset the channel
 *	@chan: the channel
 *
 *	This has the effect of erasing all data from all channel buffers
 *	and restarting the channel in its initial state.  The buffers
 *	are not freed, so any mappings are still in effect.
 *
363
 *	NOTE. Care should be taken that the channel isn't actually
364 365 366 367 368 369 370 371 372
 *	being used by anything when this call is made.
 */
void relay_reset(struct rchan *chan)
{
	unsigned int i;

	if (!chan)
		return;

373 374 375
 	if (chan->is_global && chan->buf[0]) {
		__relay_reset(chan->buf[0], 0);
		return;
376
	}
377 378 379 380 381 382

	mutex_lock(&relay_channels_mutex);
	for_each_online_cpu(i)
		if (chan->buf[i])
			__relay_reset(chan->buf[i], 0);
	mutex_unlock(&relay_channels_mutex);
383 384 385
}
EXPORT_SYMBOL_GPL(relay_reset);

386
/*
387 388
 *	relay_open_buf - create a new relay channel buffer
 *
389
 *	used by relay_open() and CPU hotplug.
390
 */
391
static struct rchan_buf *relay_open_buf(struct rchan *chan, unsigned int cpu)
392
{
393
 	struct rchan_buf *buf = NULL;
394
	struct dentry *dentry;
395
 	char *tmpname;
396

397
 	if (chan->is_global)
398 399
		return chan->buf[0];

400 401 402 403 404
	tmpname = kzalloc(NAME_MAX + 1, GFP_KERNEL);
 	if (!tmpname)
 		goto end;
 	snprintf(tmpname, NAME_MAX, "%s%d", chan->base_filename, cpu);

405 406
	buf = relay_create_buf(chan);
	if (!buf)
407 408 409 410
 		goto free_name;

 	buf->cpu = cpu;
 	__relay_reset(buf, 1);
411 412

	/* Create file in fs */
413 414 415 416
 	dentry = chan->cb->create_buf_file(tmpname, chan->parent, S_IRUSR,
 					   buf, &chan->is_global);
 	if (!dentry)
 		goto free_buf;
417 418 419

	buf->dentry = dentry;

420 421 422 423 424 425 426 427 428
 	if(chan->is_global) {
 		chan->buf[0] = buf;
 		buf->cpu = 0;
  	}

 	goto free_name;

free_buf:
 	relay_destroy_buf(buf);
429
 	buf = NULL;
430 431 432
free_name:
 	kfree(tmpname);
end:
433 434 435 436 437 438 439 440 441 442 443
	return buf;
}

/**
 *	relay_close_buf - close a channel buffer
 *	@buf: channel buffer
 *
 *	Marks the buffer finalized and restores the default callbacks.
 *	The channel buffer and channel buffer data structure are then freed
 *	automatically when the last reference is given up.
 */
A
Andrew Morton 已提交
444
static void relay_close_buf(struct rchan_buf *buf)
445 446
{
	buf->finalized = 1;
447
	del_timer_sync(&buf->timer);
448 449 450
	kref_put(&buf->kref, relay_remove_buf);
}

A
Andrew Morton 已提交
451
static void setup_callbacks(struct rchan *chan,
452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471
				   struct rchan_callbacks *cb)
{
	if (!cb) {
		chan->cb = &default_channel_callbacks;
		return;
	}

	if (!cb->subbuf_start)
		cb->subbuf_start = subbuf_start_default_callback;
	if (!cb->buf_mapped)
		cb->buf_mapped = buf_mapped_default_callback;
	if (!cb->buf_unmapped)
		cb->buf_unmapped = buf_unmapped_default_callback;
	if (!cb->create_buf_file)
		cb->create_buf_file = create_buf_file_default_callback;
	if (!cb->remove_buf_file)
		cb->remove_buf_file = remove_buf_file_default_callback;
	chan->cb = cb;
}

472 473 474 475 476 477
/**
 * 	relay_hotcpu_callback - CPU hotplug callback
 * 	@nb: notifier block
 * 	@action: hotplug action to take
 * 	@hcpu: CPU number
 *
478
 * 	Returns the success/failure of the operation. (%NOTIFY_OK, %NOTIFY_BAD)
479 480 481 482 483 484 485 486 487 488
 */
static int __cpuinit relay_hotcpu_callback(struct notifier_block *nb,
				unsigned long action,
				void *hcpu)
{
	unsigned int hotcpu = (unsigned long)hcpu;
	struct rchan *chan;

	switch(action) {
	case CPU_UP_PREPARE:
489
	case CPU_UP_PREPARE_FROZEN:
490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505
		mutex_lock(&relay_channels_mutex);
		list_for_each_entry(chan, &relay_channels, list) {
			if (chan->buf[hotcpu])
				continue;
			chan->buf[hotcpu] = relay_open_buf(chan, hotcpu);
			if(!chan->buf[hotcpu]) {
				printk(KERN_ERR
					"relay_hotcpu_callback: cpu %d buffer "
					"creation failed\n", hotcpu);
				mutex_unlock(&relay_channels_mutex);
				return NOTIFY_BAD;
			}
		}
		mutex_unlock(&relay_channels_mutex);
		break;
	case CPU_DEAD:
506
	case CPU_DEAD_FROZEN:
507 508 509 510 511 512 513
		/* No need to flush the cpu : will be flushed upon
		 * final relay_flush() call. */
		break;
	}
	return NOTIFY_OK;
}

514 515 516
/**
 *	relay_open - create a new relay channel
 *	@base_filename: base name of files to create
517
 *	@parent: dentry of parent directory, %NULL for root directory
518 519 520
 *	@subbuf_size: size of sub-buffers
 *	@n_subbufs: number of sub-buffers
 *	@cb: client callback functions
521
 *	@private_data: user-defined data
522
 *
523
 *	Returns channel pointer if successful, %NULL otherwise.
524 525 526 527
 *
 *	Creates a channel buffer for each cpu using the sizes and
 *	attributes specified.  The created channel buffer files
 *	will be named base_filename0...base_filenameN-1.  File
528
 *	permissions will be %S_IRUSR.
529 530 531 532 533
 */
struct rchan *relay_open(const char *base_filename,
			 struct dentry *parent,
			 size_t subbuf_size,
			 size_t n_subbufs,
534 535
			 struct rchan_callbacks *cb,
			 void *private_data)
536 537 538 539 540 541 542 543 544
{
	unsigned int i;
	struct rchan *chan;
	if (!base_filename)
		return NULL;

	if (!(subbuf_size && n_subbufs))
		return NULL;

545
	chan = kzalloc(sizeof(struct rchan), GFP_KERNEL);
546 547 548 549 550 551 552
	if (!chan)
		return NULL;

	chan->version = RELAYFS_CHANNEL_VERSION;
	chan->n_subbufs = n_subbufs;
	chan->subbuf_size = subbuf_size;
	chan->alloc_size = FIX_SIZE(subbuf_size * n_subbufs);
553 554 555
	chan->parent = parent;
	chan->private_data = private_data;
	strlcpy(chan->base_filename, base_filename, NAME_MAX);
556 557 558
	setup_callbacks(chan, cb);
	kref_init(&chan->kref);

559
	mutex_lock(&relay_channels_mutex);
560
	for_each_online_cpu(i) {
561
		chan->buf[i] = relay_open_buf(chan, i);
562 563 564
		if (!chan->buf[i])
			goto free_bufs;
	}
565 566
	list_add(&chan->list, &relay_channels);
	mutex_unlock(&relay_channels_mutex);
567 568 569 570

	return chan;

free_bufs:
571
	for_each_online_cpu(i) {
572 573 574 575 576 577
		if (!chan->buf[i])
			break;
		relay_close_buf(chan->buf[i]);
	}

	kref_put(&chan->kref, relay_destroy_channel);
578
	mutex_unlock(&relay_channels_mutex);
579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605
	return NULL;
}
EXPORT_SYMBOL_GPL(relay_open);

/**
 *	relay_switch_subbuf - switch to a new sub-buffer
 *	@buf: channel buffer
 *	@length: size of current event
 *
 *	Returns either the length passed in or 0 if full.
 *
 *	Performs sub-buffer-switch tasks such as invoking callbacks,
 *	updating padding counts, waking up readers, etc.
 */
size_t relay_switch_subbuf(struct rchan_buf *buf, size_t length)
{
	void *old, *new;
	size_t old_subbuf, new_subbuf;

	if (unlikely(length > buf->chan->subbuf_size))
		goto toobig;

	if (buf->offset != buf->chan->subbuf_size + 1) {
		buf->prev_padding = buf->chan->subbuf_size - buf->offset;
		old_subbuf = buf->subbufs_produced % buf->chan->n_subbufs;
		buf->padding[old_subbuf] = buf->prev_padding;
		buf->subbufs_produced++;
606 607 608
		buf->dentry->d_inode->i_size += buf->chan->subbuf_size -
			buf->padding[old_subbuf];
		smp_mb();
609 610 611 612 613 614 615 616
		if (waitqueue_active(&buf->read_wait))
			/*
			 * Calling wake_up_interruptible() from here
			 * will deadlock if we happen to be logging
			 * from the scheduler (trying to re-grab
			 * rq->lock), so defer it.
			 */
			__mod_timer(&buf->timer, jiffies + 1);
617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650
	}

	old = buf->data;
	new_subbuf = buf->subbufs_produced % buf->chan->n_subbufs;
	new = buf->start + new_subbuf * buf->chan->subbuf_size;
	buf->offset = 0;
	if (!buf->chan->cb->subbuf_start(buf, new, old, buf->prev_padding)) {
		buf->offset = buf->chan->subbuf_size + 1;
		return 0;
	}
	buf->data = new;
	buf->padding[new_subbuf] = 0;

	if (unlikely(length + buf->offset > buf->chan->subbuf_size))
		goto toobig;

	return length;

toobig:
	buf->chan->last_toobig = length;
	return 0;
}
EXPORT_SYMBOL_GPL(relay_switch_subbuf);

/**
 *	relay_subbufs_consumed - update the buffer's sub-buffers-consumed count
 *	@chan: the channel
 *	@cpu: the cpu associated with the channel buffer to update
 *	@subbufs_consumed: number of sub-buffers to add to current buf's count
 *
 *	Adds to the channel buffer's consumed sub-buffer count.
 *	subbufs_consumed should be the number of sub-buffers newly consumed,
 *	not the total consumed.
 *
651
 *	NOTE. Kernel clients don't need to call this function if the channel
652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685
 *	mode is 'overwrite'.
 */
void relay_subbufs_consumed(struct rchan *chan,
			    unsigned int cpu,
			    size_t subbufs_consumed)
{
	struct rchan_buf *buf;

	if (!chan)
		return;

	if (cpu >= NR_CPUS || !chan->buf[cpu])
		return;

	buf = chan->buf[cpu];
	buf->subbufs_consumed += subbufs_consumed;
	if (buf->subbufs_consumed > buf->subbufs_produced)
		buf->subbufs_consumed = buf->subbufs_produced;
}
EXPORT_SYMBOL_GPL(relay_subbufs_consumed);

/**
 *	relay_close - close the channel
 *	@chan: the channel
 *
 *	Closes all channel buffers and frees the channel.
 */
void relay_close(struct rchan *chan)
{
	unsigned int i;

	if (!chan)
		return;

686 687 688 689 690 691 692
	mutex_lock(&relay_channels_mutex);
	if (chan->is_global && chan->buf[0])
		relay_close_buf(chan->buf[0]);
	else
		for_each_possible_cpu(i)
			if (chan->buf[i])
				relay_close_buf(chan->buf[i]);
693 694 695 696 697 698

	if (chan->last_toobig)
		printk(KERN_WARNING "relay: one or more items not logged "
		       "[item size (%Zd) > sub-buffer size (%Zd)]\n",
		       chan->last_toobig, chan->subbuf_size);

699
	list_del(&chan->list);
700
	kref_put(&chan->kref, relay_destroy_channel);
701
	mutex_unlock(&relay_channels_mutex);
702 703 704 705 706 707 708
}
EXPORT_SYMBOL_GPL(relay_close);

/**
 *	relay_flush - close the channel
 *	@chan: the channel
 *
709
 *	Flushes all channel buffers, i.e. forces buffer switch.
710 711 712 713 714 715 716 717
 */
void relay_flush(struct rchan *chan)
{
	unsigned int i;

	if (!chan)
		return;

718 719 720
	if (chan->is_global && chan->buf[0]) {
		relay_switch_subbuf(chan->buf[0], 0);
		return;
721
	}
722 723 724 725 726 727

	mutex_lock(&relay_channels_mutex);
	for_each_possible_cpu(i)
		if (chan->buf[i])
			relay_switch_subbuf(chan->buf[i], 0);
	mutex_unlock(&relay_channels_mutex);
728 729 730 731 732 733 734 735 736 737 738 739
}
EXPORT_SYMBOL_GPL(relay_flush);

/**
 *	relay_file_open - open file op for relay files
 *	@inode: the inode
 *	@filp: the file
 *
 *	Increments the channel buffer refcount.
 */
static int relay_file_open(struct inode *inode, struct file *filp)
{
740
	struct rchan_buf *buf = inode->i_private;
741 742 743 744 745 746 747 748 749 750 751
	kref_get(&buf->kref);
	filp->private_data = buf;

	return 0;
}

/**
 *	relay_file_mmap - mmap file op for relay files
 *	@filp: the file
 *	@vma: the vma describing what to map
 *
752
 *	Calls upon relay_mmap_buf() to map the file into user space.
753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799
 */
static int relay_file_mmap(struct file *filp, struct vm_area_struct *vma)
{
	struct rchan_buf *buf = filp->private_data;
	return relay_mmap_buf(buf, vma);
}

/**
 *	relay_file_poll - poll file op for relay files
 *	@filp: the file
 *	@wait: poll table
 *
 *	Poll implemention.
 */
static unsigned int relay_file_poll(struct file *filp, poll_table *wait)
{
	unsigned int mask = 0;
	struct rchan_buf *buf = filp->private_data;

	if (buf->finalized)
		return POLLERR;

	if (filp->f_mode & FMODE_READ) {
		poll_wait(filp, &buf->read_wait, wait);
		if (!relay_buf_empty(buf))
			mask |= POLLIN | POLLRDNORM;
	}

	return mask;
}

/**
 *	relay_file_release - release file op for relay files
 *	@inode: the inode
 *	@filp: the file
 *
 *	Decrements the channel refcount, as the filesystem is
 *	no longer using it.
 */
static int relay_file_release(struct inode *inode, struct file *filp)
{
	struct rchan_buf *buf = filp->private_data;
	kref_put(&buf->kref, relay_remove_buf);

	return 0;
}

800
/*
801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816
 *	relay_file_read_consume - update the consumed count for the buffer
 */
static void relay_file_read_consume(struct rchan_buf *buf,
				    size_t read_pos,
				    size_t bytes_consumed)
{
	size_t subbuf_size = buf->chan->subbuf_size;
	size_t n_subbufs = buf->chan->n_subbufs;
	size_t read_subbuf;

	if (buf->bytes_consumed + bytes_consumed > subbuf_size) {
		relay_subbufs_consumed(buf->chan, buf->cpu, 1);
		buf->bytes_consumed = 0;
	}

	buf->bytes_consumed += bytes_consumed;
M
Masami Hiramatsu 已提交
817 818 819 820
	if (!read_pos)
		read_subbuf = buf->subbufs_consumed % n_subbufs;
	else
		read_subbuf = read_pos / buf->chan->subbuf_size;
821 822 823 824 825 826 827 828 829
	if (buf->bytes_consumed + buf->padding[read_subbuf] == subbuf_size) {
		if ((read_subbuf == buf->subbufs_produced % n_subbufs) &&
		    (buf->offset == subbuf_size))
			return;
		relay_subbufs_consumed(buf->chan, buf->cpu, 1);
		buf->bytes_consumed = 0;
	}
}

830
/*
831 832 833 834 835 836
 *	relay_file_read_avail - boolean, are there unconsumed bytes available?
 */
static int relay_file_read_avail(struct rchan_buf *buf, size_t read_pos)
{
	size_t subbuf_size = buf->chan->subbuf_size;
	size_t n_subbufs = buf->chan->n_subbufs;
837 838
	size_t produced = buf->subbufs_produced;
	size_t consumed = buf->subbufs_consumed;
839

840
	relay_file_read_consume(buf, read_pos, 0);
841

842 843 844 845
	if (unlikely(buf->offset > subbuf_size)) {
		if (produced == consumed)
			return 0;
		return 1;
846 847
	}

848
	if (unlikely(produced - consumed >= n_subbufs)) {
M
Masami Hiramatsu 已提交
849
		consumed = produced - n_subbufs + 1;
850
		buf->subbufs_consumed = consumed;
M
Masami Hiramatsu 已提交
851
		buf->bytes_consumed = 0;
852 853 854 855 856 857 858 859 860
	}
	
	produced = (produced % n_subbufs) * subbuf_size + buf->offset;
	consumed = (consumed % n_subbufs) * subbuf_size + buf->bytes_consumed;

	if (consumed > produced)
		produced += n_subbufs * subbuf_size;
	
	if (consumed == produced)
861 862 863 864 865 866 867
		return 0;

	return 1;
}

/**
 *	relay_file_read_subbuf_avail - return bytes available in sub-buffer
868 869
 *	@read_pos: file read position
 *	@buf: relay channel buffer
870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894
 */
static size_t relay_file_read_subbuf_avail(size_t read_pos,
					   struct rchan_buf *buf)
{
	size_t padding, avail = 0;
	size_t read_subbuf, read_offset, write_subbuf, write_offset;
	size_t subbuf_size = buf->chan->subbuf_size;

	write_subbuf = (buf->data - buf->start) / subbuf_size;
	write_offset = buf->offset > subbuf_size ? subbuf_size : buf->offset;
	read_subbuf = read_pos / subbuf_size;
	read_offset = read_pos % subbuf_size;
	padding = buf->padding[read_subbuf];

	if (read_subbuf == write_subbuf) {
		if (read_offset + padding < write_offset)
			avail = write_offset - (read_offset + padding);
	} else
		avail = (subbuf_size - padding) - read_offset;

	return avail;
}

/**
 *	relay_file_read_start_pos - find the first available byte to read
895 896
 *	@read_pos: file read position
 *	@buf: relay channel buffer
897
 *
898
 *	If the @read_pos is in the middle of padding, return the
899 900 901 902 903 904 905 906 907
 *	position of the first actually available byte, otherwise
 *	return the original value.
 */
static size_t relay_file_read_start_pos(size_t read_pos,
					struct rchan_buf *buf)
{
	size_t read_subbuf, padding, padding_start, padding_end;
	size_t subbuf_size = buf->chan->subbuf_size;
	size_t n_subbufs = buf->chan->n_subbufs;
D
David Wilder 已提交
908
	size_t consumed = buf->subbufs_consumed % n_subbufs;
909

D
David Wilder 已提交
910 911
	if (!read_pos)
		read_pos = consumed * subbuf_size + buf->bytes_consumed;
912 913 914 915 916 917 918 919 920 921 922 923 924 925
	read_subbuf = read_pos / subbuf_size;
	padding = buf->padding[read_subbuf];
	padding_start = (read_subbuf + 1) * subbuf_size - padding;
	padding_end = (read_subbuf + 1) * subbuf_size;
	if (read_pos >= padding_start && read_pos < padding_end) {
		read_subbuf = (read_subbuf + 1) % n_subbufs;
		read_pos = read_subbuf * subbuf_size;
	}

	return read_pos;
}

/**
 *	relay_file_read_end_pos - return the new read position
926 927 928
 *	@read_pos: file read position
 *	@buf: relay channel buffer
 *	@count: number of bytes to be read
929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949
 */
static size_t relay_file_read_end_pos(struct rchan_buf *buf,
				      size_t read_pos,
				      size_t count)
{
	size_t read_subbuf, padding, end_pos;
	size_t subbuf_size = buf->chan->subbuf_size;
	size_t n_subbufs = buf->chan->n_subbufs;

	read_subbuf = read_pos / subbuf_size;
	padding = buf->padding[read_subbuf];
	if (read_pos % subbuf_size + count + padding == subbuf_size)
		end_pos = (read_subbuf + 1) * subbuf_size;
	else
		end_pos = read_pos + count;
	if (end_pos >= subbuf_size * n_subbufs)
		end_pos = 0;

	return end_pos;
}

950
/*
951
 *	subbuf_read_actor - read up to one subbuf's worth of data
952
 */
953 954 955 956 957
static int subbuf_read_actor(size_t read_start,
			     struct rchan_buf *buf,
			     size_t avail,
			     read_descriptor_t *desc,
			     read_actor_t actor)
958 959
{
	void *from;
960
	int ret = 0;
961 962

	from = buf->start + read_start;
963
	ret = avail;
A
Al Viro 已提交
964
	if (copy_to_user(desc->arg.buf, from, avail)) {
965 966
		desc->error = -EFAULT;
		ret = 0;
967
	}
968 969 970 971
	desc->arg.data += ret;
	desc->written += ret;
	desc->count -= ret;

972 973 974
	return ret;
}

975 976 977 978 979 980
typedef int (*subbuf_actor_t) (size_t read_start,
			       struct rchan_buf *buf,
			       size_t avail,
			       read_descriptor_t *desc,
			       read_actor_t actor);

981
/*
982 983
 *	relay_file_read_subbufs - read count bytes, bridging subbuf boundaries
 */
A
Andrew Morton 已提交
984 985 986 987
static ssize_t relay_file_read_subbufs(struct file *filp, loff_t *ppos,
					subbuf_actor_t subbuf_actor,
					read_actor_t actor,
					read_descriptor_t *desc)
988
{
989 990 991
	struct rchan_buf *buf = filp->private_data;
	size_t read_start, avail;
	int ret;
992

A
Al Viro 已提交
993
	if (!desc->count)
994 995
		return 0;

996
	mutex_lock(&filp->f_path.dentry->d_inode->i_mutex);
997
	do {
998 999 1000 1001 1002 1003
		if (!relay_file_read_avail(buf, *ppos))
			break;

		read_start = relay_file_read_start_pos(*ppos, buf);
		avail = relay_file_read_subbuf_avail(read_start, buf);
		if (!avail)
1004 1005
			break;

A
Al Viro 已提交
1006 1007 1008
		avail = min(desc->count, avail);
		ret = subbuf_actor(read_start, buf, avail, desc, actor);
		if (desc->error < 0)
1009 1010 1011 1012 1013 1014
			break;

		if (ret) {
			relay_file_read_consume(buf, read_start, ret);
			*ppos = relay_file_read_end_pos(buf, read_start, ret);
		}
A
Al Viro 已提交
1015
	} while (desc->count && ret);
1016
	mutex_unlock(&filp->f_path.dentry->d_inode->i_mutex);
1017

A
Al Viro 已提交
1018
	return desc->written;
1019 1020 1021 1022 1023 1024 1025
}

static ssize_t relay_file_read(struct file *filp,
			       char __user *buffer,
			       size_t count,
			       loff_t *ppos)
{
A
Al Viro 已提交
1026 1027 1028 1029 1030 1031 1032
	read_descriptor_t desc;
	desc.written = 0;
	desc.count = count;
	desc.arg.buf = buffer;
	desc.error = 0;
	return relay_file_read_subbufs(filp, ppos, subbuf_read_actor,
				       NULL, &desc);
1033 1034
}

1035 1036 1037 1038 1039 1040 1041 1042 1043 1044
static void relay_consume_bytes(struct rchan_buf *rbuf, int bytes_consumed)
{
	rbuf->bytes_consumed += bytes_consumed;

	if (rbuf->bytes_consumed >= rbuf->chan->subbuf_size) {
		relay_subbufs_consumed(rbuf->chan, rbuf->cpu, 1);
		rbuf->bytes_consumed %= rbuf->chan->subbuf_size;
	}
}

T
Tom Zanussi 已提交
1045 1046
static void relay_pipe_buf_release(struct pipe_inode_info *pipe,
				   struct pipe_buffer *buf)
1047
{
T
Tom Zanussi 已提交
1048 1049 1050
	struct rchan_buf *rbuf;

	rbuf = (struct rchan_buf *)page_private(buf->page);
1051
	relay_consume_bytes(rbuf, buf->private);
T
Tom Zanussi 已提交
1052 1053 1054 1055 1056 1057
}

static struct pipe_buf_operations relay_pipe_buf_ops = {
	.can_merge = 0,
	.map = generic_pipe_buf_map,
	.unmap = generic_pipe_buf_unmap,
1058
	.confirm = generic_pipe_buf_confirm,
T
Tom Zanussi 已提交
1059 1060 1061 1062 1063
	.release = relay_pipe_buf_release,
	.steal = generic_pipe_buf_steal,
	.get = generic_pipe_buf_get,
};

T
Tom Zanussi 已提交
1064
/*
T
Tom Zanussi 已提交
1065 1066 1067 1068 1069 1070 1071 1072 1073
 *	subbuf_splice_actor - splice up to one subbuf's worth of data
 */
static int subbuf_splice_actor(struct file *in,
			       loff_t *ppos,
			       struct pipe_inode_info *pipe,
			       size_t len,
			       unsigned int flags,
			       int *nonpad_ret)
{
1074
	unsigned int pidx, poff, total_len, subbuf_pages, ret;
T
Tom Zanussi 已提交
1075 1076
	struct rchan_buf *rbuf = in->private_data;
	unsigned int subbuf_size = rbuf->chan->subbuf_size;
1077 1078 1079
	uint64_t pos = (uint64_t) *ppos;
	uint32_t alloc_size = (uint32_t) rbuf->chan->alloc_size;
	size_t read_start = (size_t) do_div(pos, alloc_size);
T
Tom Zanussi 已提交
1080 1081 1082
	size_t read_subbuf = read_start / subbuf_size;
	size_t padding = rbuf->padding[read_subbuf];
	size_t nonpad_end = read_subbuf * subbuf_size + subbuf_size - padding;
1083 1084 1085 1086 1087 1088 1089 1090 1091
	struct page *pages[PIPE_BUFFERS];
	struct partial_page partial[PIPE_BUFFERS];
	struct splice_pipe_desc spd = {
		.pages = pages,
		.nr_pages = 0,
		.partial = partial,
		.flags = flags,
		.ops = &relay_pipe_buf_ops,
	};
T
Tom Zanussi 已提交
1092 1093 1094 1095

	if (rbuf->subbufs_produced == rbuf->subbufs_consumed)
		return 0;

1096 1097 1098 1099 1100
	/*
	 * Adjust read len, if longer than what is available
	 */
	if (len > (subbuf_size - read_start % subbuf_size))
		len = subbuf_size - read_start % subbuf_size;
T
Tom Zanussi 已提交
1101 1102 1103 1104 1105

	subbuf_pages = rbuf->chan->alloc_size >> PAGE_SHIFT;
	pidx = (read_start / PAGE_SIZE) % subbuf_pages;
	poff = read_start & ~PAGE_MASK;

1106 1107 1108
	for (total_len = 0; spd.nr_pages < subbuf_pages; spd.nr_pages++) {
		unsigned int this_len, this_end, private;
		unsigned int cur_pos = read_start + total_len;
T
Tom Zanussi 已提交
1109

1110
		if (!len)
T
Tom Zanussi 已提交
1111 1112
			break;

1113 1114
		this_len = min_t(unsigned long, len, PAGE_SIZE - poff);
		private = this_len;
T
Tom Zanussi 已提交
1115

1116 1117
		spd.pages[spd.nr_pages] = rbuf->page_array[pidx];
		spd.partial[spd.nr_pages].offset = poff;
T
Tom Zanussi 已提交
1118

1119 1120 1121 1122
		this_end = cur_pos + this_len;
		if (this_end >= nonpad_end) {
			this_len = nonpad_end - cur_pos;
			private = this_len + padding;
T
Tom Zanussi 已提交
1123
		}
1124 1125
		spd.partial[spd.nr_pages].len = this_len;
		spd.partial[spd.nr_pages].private = private;
T
Tom Zanussi 已提交
1126

1127 1128 1129 1130
		len -= this_len;
		total_len += this_len;
		poff = 0;
		pidx = (pidx + 1) % subbuf_pages;
T
Tom Zanussi 已提交
1131

1132 1133
		if (this_end >= nonpad_end) {
			spd.nr_pages++;
T
Tom Zanussi 已提交
1134 1135 1136 1137
			break;
		}
	}

1138 1139
	if (!spd.nr_pages)
		return 0;
T
Tom Zanussi 已提交
1140

1141 1142 1143
	ret = *nonpad_ret = splice_to_pipe(pipe, &spd);
	if (ret < 0 || ret < total_len)
		return ret;
T
Tom Zanussi 已提交
1144

1145 1146 1147 1148
        if (read_start + ret == nonpad_end)
                ret += padding;

        return ret;
T
Tom Zanussi 已提交
1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189
}

static ssize_t relay_file_splice_read(struct file *in,
				      loff_t *ppos,
				      struct pipe_inode_info *pipe,
				      size_t len,
				      unsigned int flags)
{
	ssize_t spliced;
	int ret;
	int nonpad_ret = 0;

	ret = 0;
	spliced = 0;

	while (len) {
		ret = subbuf_splice_actor(in, ppos, pipe, len, flags, &nonpad_ret);
		if (ret < 0)
			break;
		else if (!ret) {
			if (spliced)
				break;
			if (flags & SPLICE_F_NONBLOCK) {
				ret = -EAGAIN;
				break;
			}
		}

		*ppos += ret;
		if (ret > len)
			len = 0;
		else
			len -= ret;
		spliced += nonpad_ret;
		nonpad_ret = 0;
	}

	if (spliced)
		return spliced;

	return ret;
1190 1191
}

1192
const struct file_operations relay_file_operations = {
1193 1194 1195 1196 1197 1198
	.open		= relay_file_open,
	.poll		= relay_file_poll,
	.mmap		= relay_file_mmap,
	.read		= relay_file_read,
	.llseek		= no_llseek,
	.release	= relay_file_release,
T
Tom Zanussi 已提交
1199
	.splice_read	= relay_file_splice_read,
1200 1201
};
EXPORT_SYMBOL_GPL(relay_file_operations);
1202 1203 1204 1205 1206 1207 1208 1209 1210

static __init int relay_init(void)
{

	hotcpu_notifier(relay_hotcpu_callback, 0);
	return 0;
}

module_init(relay_init);