1// SPDX-License-Identifier: GPL-2.0-only
2/* Network filesystem write subrequest result collection, assessment
3 * and retrying.
4 *
5 * Copyright (C) 2024 Red Hat, Inc. All Rights Reserved.
6 * Written by David Howells (dhowells@redhat.com)
7 */
8
9#include <linux/export.h>
10#include <linux/fs.h>
11#include <linux/mm.h>
12#include <linux/pagemap.h>
13#include <linux/slab.h>
14#include "internal.h"
15
16/* Notes made in the collector */
17#define HIT_PENDING		0x01	/* A front op was still pending */
18#define SOME_EMPTY		0x02	/* One of more streams are empty */
19#define ALL_EMPTY		0x04	/* All streams are empty */
20#define MAYBE_DISCONTIG		0x08	/* A front op may be discontiguous (rounded to PAGE_SIZE) */
21#define NEED_REASSESS		0x10	/* Need to loop round and reassess */
22#define REASSESS_DISCONTIG	0x20	/* Reassess discontiguity if contiguity advances */
23#define MADE_PROGRESS		0x40	/* Made progress cleaning up a stream or the folio set */
24#define BUFFERED		0x80	/* The pagecache needs cleaning up */
25#define NEED_RETRY		0x100	/* A front op requests retrying */
26#define SAW_FAILURE		0x200	/* One stream or hit a permanent failure */
27
28/*
29 * Successful completion of write of a folio to the server and/or cache.  Note
30 * that we are not allowed to lock the folio here on pain of deadlocking with
31 * truncate.
32 */
33int netfs_folio_written_back(struct folio *folio)
34{
35	enum netfs_folio_trace why = netfs_folio_trace_clear;
36	struct netfs_folio *finfo;
37	struct netfs_group *group = NULL;
38	int gcount = 0;
39
40	if ((finfo = netfs_folio_info(folio))) {
41		/* Streaming writes cannot be redirtied whilst under writeback,
42		 * so discard the streaming record.
43		 */
44		folio_detach_private(folio);
45		group = finfo->netfs_group;
46		gcount++;
47		kfree(finfo);
48		why = netfs_folio_trace_clear_s;
49		goto end_wb;
50	}
51
52	if ((group = netfs_folio_group(folio))) {
53		if (group == NETFS_FOLIO_COPY_TO_CACHE) {
54			why = netfs_folio_trace_clear_cc;
55			folio_detach_private(folio);
56			goto end_wb;
57		}
58
59		/* Need to detach the group pointer if the page didn't get
60		 * redirtied.  If it has been redirtied, then it must be within
61		 * the same group.
62		 */
63		why = netfs_folio_trace_redirtied;
64		if (!folio_test_dirty(folio)) {
65			folio_detach_private(folio);
66			gcount++;
67			why = netfs_folio_trace_clear_g;
68		}
69	}
70
71end_wb:
72	trace_netfs_folio(folio, why);
73	folio_end_writeback(folio);
74	return gcount;
75}
76
77/*
78 * Get hold of a folio we have under writeback.  We don't want to get the
79 * refcount on it.
80 */
81static struct folio *netfs_writeback_lookup_folio(struct netfs_io_request *wreq, loff_t pos)
82{
83	XA_STATE(xas, &wreq->mapping->i_pages, pos / PAGE_SIZE);
84	struct folio *folio;
85
86	rcu_read_lock();
87
88	for (;;) {
89		xas_reset(&xas);
90		folio = xas_load(&xas);
91		if (xas_retry(&xas, folio))
92			continue;
93
94		if (!folio || xa_is_value(folio))
95			kdebug("R=%08x: folio %lx (%llx) not present",
96			       wreq->debug_id, xas.xa_index, pos / PAGE_SIZE);
97		BUG_ON(!folio || xa_is_value(folio));
98
99		if (folio == xas_reload(&xas))
100			break;
101	}
102
103	rcu_read_unlock();
104
105	if (WARN_ONCE(!folio_test_writeback(folio),
106		      "R=%08x: folio %lx is not under writeback\n",
107		      wreq->debug_id, folio->index)) {
108		trace_netfs_folio(folio, netfs_folio_trace_not_under_wback);
109	}
110	return folio;
111}
112
113/*
114 * Unlock any folios we've finished with.
115 */
116static void netfs_writeback_unlock_folios(struct netfs_io_request *wreq,
117					  unsigned long long collected_to,
118					  unsigned int *notes)
119{
120	for (;;) {
121		struct folio *folio;
122		struct netfs_folio *finfo;
123		unsigned long long fpos, fend;
124		size_t fsize, flen;
125
126		folio = netfs_writeback_lookup_folio(wreq, wreq->cleaned_to);
127
128		fpos = folio_pos(folio);
129		fsize = folio_size(folio);
130		finfo = netfs_folio_info(folio);
131		flen = finfo ? finfo->dirty_offset + finfo->dirty_len : fsize;
132
133		fend = min_t(unsigned long long, fpos + flen, wreq->i_size);
134
135		trace_netfs_collect_folio(wreq, folio, fend, collected_to);
136
137		if (fpos + fsize > wreq->contiguity) {
138			trace_netfs_collect_contig(wreq, fpos + fsize,
139						   netfs_contig_trace_unlock);
140			wreq->contiguity = fpos + fsize;
141		}
142
143		/* Unlock any folio we've transferred all of. */
144		if (collected_to < fend)
145			break;
146
147		wreq->nr_group_rel += netfs_folio_written_back(folio);
148		wreq->cleaned_to = fpos + fsize;
149		*notes |= MADE_PROGRESS;
150
151		if (fpos + fsize >= collected_to)
152			break;
153	}
154}
155
156/*
157 * Perform retries on the streams that need it.
158 */
159static void netfs_retry_write_stream(struct netfs_io_request *wreq,
160				     struct netfs_io_stream *stream)
161{
162	struct list_head *next;
163
164	_enter("R=%x[%x:]", wreq->debug_id, stream->stream_nr);
165
166	if (list_empty(&stream->subrequests))
167		return;
168
169	if (stream->source == NETFS_UPLOAD_TO_SERVER &&
170	    wreq->netfs_ops->retry_request)
171		wreq->netfs_ops->retry_request(wreq, stream);
172
173	if (unlikely(stream->failed))
174		return;
175
176	/* If there's no renegotiation to do, just resend each failed subreq. */
177	if (!stream->prepare_write) {
178		struct netfs_io_subrequest *subreq;
179
180		list_for_each_entry(subreq, &stream->subrequests, rreq_link) {
181			if (test_bit(NETFS_SREQ_FAILED, &subreq->flags))
182				break;
183			if (__test_and_clear_bit(NETFS_SREQ_NEED_RETRY, &subreq->flags)) {
184				__set_bit(NETFS_SREQ_RETRYING, &subreq->flags);
185				netfs_get_subrequest(subreq, netfs_sreq_trace_get_resubmit);
186				netfs_reissue_write(stream, subreq);
187			}
188		}
189		return;
190	}
191
192	next = stream->subrequests.next;
193
194	do {
195		struct netfs_io_subrequest *subreq = NULL, *from, *to, *tmp;
196		unsigned long long start, len;
197		size_t part;
198		bool boundary = false;
199
200		/* Go through the stream and find the next span of contiguous
201		 * data that we then rejig (cifs, for example, needs the wsize
202		 * renegotiating) and reissue.
203		 */
204		from = list_entry(next, struct netfs_io_subrequest, rreq_link);
205		to = from;
206		start = from->start + from->transferred;
207		len   = from->len   - from->transferred;
208
209		if (test_bit(NETFS_SREQ_FAILED, &from->flags) ||
210		    !test_bit(NETFS_SREQ_NEED_RETRY, &from->flags))
211			return;
212
213		list_for_each_continue(next, &stream->subrequests) {
214			subreq = list_entry(next, struct netfs_io_subrequest, rreq_link);
215			if (subreq->start + subreq->transferred != start + len ||
216			    test_bit(NETFS_SREQ_BOUNDARY, &subreq->flags) ||
217			    !test_bit(NETFS_SREQ_NEED_RETRY, &subreq->flags))
218				break;
219			to = subreq;
220			len += to->len;
221		}
222
223		/* Work through the sublist. */
224		subreq = from;
225		list_for_each_entry_from(subreq, &stream->subrequests, rreq_link) {
226			if (!len)
227				break;
228			/* Renegotiate max_len (wsize) */
229			trace_netfs_sreq(subreq, netfs_sreq_trace_retry);
230			__clear_bit(NETFS_SREQ_NEED_RETRY, &subreq->flags);
231			__set_bit(NETFS_SREQ_RETRYING, &subreq->flags);
232			stream->prepare_write(subreq);
233
234			part = min(len, subreq->max_len);
235			subreq->len = part;
236			subreq->start = start;
237			subreq->transferred = 0;
238			len -= part;
239			start += part;
240			if (len && subreq == to &&
241			    __test_and_clear_bit(NETFS_SREQ_BOUNDARY, &to->flags))
242				boundary = true;
243
244			netfs_get_subrequest(subreq, netfs_sreq_trace_get_resubmit);
245			netfs_reissue_write(stream, subreq);
246			if (subreq == to)
247				break;
248		}
249
250		/* If we managed to use fewer subreqs, we can discard the
251		 * excess; if we used the same number, then we're done.
252		 */
253		if (!len) {
254			if (subreq == to)
255				continue;
256			list_for_each_entry_safe_from(subreq, tmp,
257						      &stream->subrequests, rreq_link) {
258				trace_netfs_sreq(subreq, netfs_sreq_trace_discard);
259				list_del(&subreq->rreq_link);
260				netfs_put_subrequest(subreq, false, netfs_sreq_trace_put_done);
261				if (subreq == to)
262					break;
263			}
264			continue;
265		}
266
267		/* We ran out of subrequests, so we need to allocate some more
268		 * and insert them after.
269		 */
270		do {
271			subreq = netfs_alloc_subrequest(wreq);
272			subreq->source		= to->source;
273			subreq->start		= start;
274			subreq->max_len		= len;
275			subreq->max_nr_segs	= INT_MAX;
276			subreq->debug_index	= atomic_inc_return(&wreq->subreq_counter);
277			subreq->stream_nr	= to->stream_nr;
278			__set_bit(NETFS_SREQ_RETRYING, &subreq->flags);
279
280			trace_netfs_sreq_ref(wreq->debug_id, subreq->debug_index,
281					     refcount_read(&subreq->ref),
282					     netfs_sreq_trace_new);
283			netfs_get_subrequest(subreq, netfs_sreq_trace_get_resubmit);
284
285			list_add(&subreq->rreq_link, &to->rreq_link);
286			to = list_next_entry(to, rreq_link);
287			trace_netfs_sreq(subreq, netfs_sreq_trace_retry);
288
289			switch (stream->source) {
290			case NETFS_UPLOAD_TO_SERVER:
291				netfs_stat(&netfs_n_wh_upload);
292				subreq->max_len = min(len, wreq->wsize);
293				break;
294			case NETFS_WRITE_TO_CACHE:
295				netfs_stat(&netfs_n_wh_write);
296				break;
297			default:
298				WARN_ON_ONCE(1);
299			}
300
301			stream->prepare_write(subreq);
302
303			part = min(len, subreq->max_len);
304			subreq->len = subreq->transferred + part;
305			len -= part;
306			start += part;
307			if (!len && boundary) {
308				__set_bit(NETFS_SREQ_BOUNDARY, &to->flags);
309				boundary = false;
310			}
311
312			netfs_reissue_write(stream, subreq);
313			if (!len)
314				break;
315
316		} while (len);
317
318	} while (!list_is_head(next, &stream->subrequests));
319}
320
321/*
322 * Perform retries on the streams that need it.  If we're doing content
323 * encryption and the server copy changed due to a third-party write, we may
324 * need to do an RMW cycle and also rewrite the data to the cache.
325 */
326static void netfs_retry_writes(struct netfs_io_request *wreq)
327{
328	struct netfs_io_subrequest *subreq;
329	struct netfs_io_stream *stream;
330	int s;
331
332	/* Wait for all outstanding I/O to quiesce before performing retries as
333	 * we may need to renegotiate the I/O sizes.
334	 */
335	for (s = 0; s < NR_IO_STREAMS; s++) {
336		stream = &wreq->io_streams[s];
337		if (!stream->active)
338			continue;
339
340		list_for_each_entry(subreq, &stream->subrequests, rreq_link) {
341			wait_on_bit(&subreq->flags, NETFS_SREQ_IN_PROGRESS,
342				    TASK_UNINTERRUPTIBLE);
343		}
344	}
345
346	// TODO: Enc: Fetch changed partial pages
347	// TODO: Enc: Reencrypt content if needed.
348	// TODO: Enc: Wind back transferred point.
349	// TODO: Enc: Mark cache pages for retry.
350
351	for (s = 0; s < NR_IO_STREAMS; s++) {
352		stream = &wreq->io_streams[s];
353		if (stream->need_retry) {
354			stream->need_retry = false;
355			netfs_retry_write_stream(wreq, stream);
356		}
357	}
358}
359
360/*
361 * Collect and assess the results of various write subrequests.  We may need to
362 * retry some of the results - or even do an RMW cycle for content crypto.
363 *
364 * Note that we have a number of parallel, overlapping lists of subrequests,
365 * one to the server and one to the local cache for example, which may not be
366 * the same size or starting position and may not even correspond in boundary
367 * alignment.
368 */
369static void netfs_collect_write_results(struct netfs_io_request *wreq)
370{
371	struct netfs_io_subrequest *front, *remove;
372	struct netfs_io_stream *stream;
373	unsigned long long collected_to;
374	unsigned int notes;
375	int s;
376
377	_enter("%llx-%llx", wreq->start, wreq->start + wreq->len);
378	trace_netfs_collect(wreq);
379	trace_netfs_rreq(wreq, netfs_rreq_trace_collect);
380
381reassess_streams:
382	smp_rmb();
383	collected_to = ULLONG_MAX;
384	if (wreq->origin == NETFS_WRITEBACK)
385		notes = ALL_EMPTY | BUFFERED | MAYBE_DISCONTIG;
386	else if (wreq->origin == NETFS_WRITETHROUGH)
387		notes = ALL_EMPTY | BUFFERED;
388	else
389		notes = ALL_EMPTY;
390
391	/* Remove completed subrequests from the front of the streams and
392	 * advance the completion point on each stream.  We stop when we hit
393	 * something that's in progress.  The issuer thread may be adding stuff
394	 * to the tail whilst we're doing this.
395	 *
396	 * We must not, however, merge in discontiguities that span whole
397	 * folios that aren't under writeback.  This is made more complicated
398	 * by the folios in the gap being of unpredictable sizes - if they even
399	 * exist - but we don't want to look them up.
400	 */
401	for (s = 0; s < NR_IO_STREAMS; s++) {
402		loff_t rstart, rend;
403
404		stream = &wreq->io_streams[s];
405		/* Read active flag before list pointers */
406		if (!smp_load_acquire(&stream->active))
407			continue;
408
409		front = stream->front;
410		while (front) {
411			trace_netfs_collect_sreq(wreq, front);
412			//_debug("sreq [%x] %llx %zx/%zx",
413			//       front->debug_index, front->start, front->transferred, front->len);
414
415			/* Stall if there may be a discontinuity. */
416			rstart = round_down(front->start, PAGE_SIZE);
417			if (rstart > wreq->contiguity) {
418				if (wreq->contiguity > stream->collected_to) {
419					trace_netfs_collect_gap(wreq, stream,
420								wreq->contiguity, 'D');
421					stream->collected_to = wreq->contiguity;
422				}
423				notes |= REASSESS_DISCONTIG;
424				break;
425			}
426			rend = round_up(front->start + front->len, PAGE_SIZE);
427			if (rend > wreq->contiguity) {
428				trace_netfs_collect_contig(wreq, rend,
429							   netfs_contig_trace_collect);
430				wreq->contiguity = rend;
431				if (notes & REASSESS_DISCONTIG)
432					notes |= NEED_REASSESS;
433			}
434			notes &= ~MAYBE_DISCONTIG;
435
436			/* Stall if the front is still undergoing I/O. */
437			if (test_bit(NETFS_SREQ_IN_PROGRESS, &front->flags)) {
438				notes |= HIT_PENDING;
439				break;
440			}
441			smp_rmb(); /* Read counters after I-P flag. */
442
443			if (stream->failed) {
444				stream->collected_to = front->start + front->len;
445				notes |= MADE_PROGRESS | SAW_FAILURE;
446				goto cancel;
447			}
448			if (front->start + front->transferred > stream->collected_to) {
449				stream->collected_to = front->start + front->transferred;
450				stream->transferred = stream->collected_to - wreq->start;
451				notes |= MADE_PROGRESS;
452			}
453			if (test_bit(NETFS_SREQ_FAILED, &front->flags)) {
454				stream->failed = true;
455				stream->error = front->error;
456				if (stream->source == NETFS_UPLOAD_TO_SERVER)
457					mapping_set_error(wreq->mapping, front->error);
458				notes |= NEED_REASSESS | SAW_FAILURE;
459				break;
460			}
461			if (front->transferred < front->len) {
462				stream->need_retry = true;
463				notes |= NEED_RETRY | MADE_PROGRESS;
464				break;
465			}
466
467		cancel:
468			/* Remove if completely consumed. */
469			spin_lock(&wreq->lock);
470
471			remove = front;
472			list_del_init(&front->rreq_link);
473			front = list_first_entry_or_null(&stream->subrequests,
474							 struct netfs_io_subrequest, rreq_link);
475			stream->front = front;
476			if (!front) {
477				unsigned long long jump_to = atomic64_read(&wreq->issued_to);
478
479				if (stream->collected_to < jump_to) {
480					trace_netfs_collect_gap(wreq, stream, jump_to, 'A');
481					stream->collected_to = jump_to;
482				}
483			}
484
485			spin_unlock(&wreq->lock);
486			netfs_put_subrequest(remove, false,
487					     notes & SAW_FAILURE ?
488					     netfs_sreq_trace_put_cancel :
489					     netfs_sreq_trace_put_done);
490		}
491
492		if (front)
493			notes &= ~ALL_EMPTY;
494		else
495			notes |= SOME_EMPTY;
496
497		if (stream->collected_to < collected_to)
498			collected_to = stream->collected_to;
499	}
500
501	if (collected_to != ULLONG_MAX && collected_to > wreq->collected_to)
502		wreq->collected_to = collected_to;
503
504	/* If we have an empty stream, we need to jump it forward over any gap
505	 * otherwise the collection point will never advance.
506	 *
507	 * Note that the issuer always adds to the stream with the lowest
508	 * so-far submitted start, so if we see two consecutive subreqs in one
509	 * stream with nothing between then in another stream, then the second
510	 * stream has a gap that can be jumped.
511	 */
512	if (notes & SOME_EMPTY) {
513		unsigned long long jump_to = wreq->start + READ_ONCE(wreq->submitted);
514
515		for (s = 0; s < NR_IO_STREAMS; s++) {
516			stream = &wreq->io_streams[s];
517			if (stream->active &&
518			    stream->front &&
519			    stream->front->start < jump_to)
520				jump_to = stream->front->start;
521		}
522
523		for (s = 0; s < NR_IO_STREAMS; s++) {
524			stream = &wreq->io_streams[s];
525			if (stream->active &&
526			    !stream->front &&
527			    stream->collected_to < jump_to) {
528				trace_netfs_collect_gap(wreq, stream, jump_to, 'B');
529				stream->collected_to = jump_to;
530			}
531		}
532	}
533
534	for (s = 0; s < NR_IO_STREAMS; s++) {
535		stream = &wreq->io_streams[s];
536		if (stream->active)
537			trace_netfs_collect_stream(wreq, stream);
538	}
539
540	trace_netfs_collect_state(wreq, wreq->collected_to, notes);
541
542	/* Unlock any folios that we have now finished with. */
543	if (notes & BUFFERED) {
544		unsigned long long clean_to = min(wreq->collected_to, wreq->contiguity);
545
546		if (wreq->cleaned_to < clean_to)
547			netfs_writeback_unlock_folios(wreq, clean_to, &notes);
548	} else {
549		wreq->cleaned_to = wreq->collected_to;
550	}
551
552	// TODO: Discard encryption buffers
553
554	/* If all streams are discontiguous with the last folio we cleared, we
555	 * may need to skip a set of folios.
556	 */
557	if ((notes & (MAYBE_DISCONTIG | ALL_EMPTY)) == MAYBE_DISCONTIG) {
558		unsigned long long jump_to = ULLONG_MAX;
559
560		for (s = 0; s < NR_IO_STREAMS; s++) {
561			stream = &wreq->io_streams[s];
562			if (stream->active && stream->front &&
563			    stream->front->start < jump_to)
564				jump_to = stream->front->start;
565		}
566
567		trace_netfs_collect_contig(wreq, jump_to, netfs_contig_trace_jump);
568		wreq->contiguity = jump_to;
569		wreq->cleaned_to = jump_to;
570		wreq->collected_to = jump_to;
571		for (s = 0; s < NR_IO_STREAMS; s++) {
572			stream = &wreq->io_streams[s];
573			if (stream->collected_to < jump_to)
574				stream->collected_to = jump_to;
575		}
576		//cond_resched();
577		notes |= MADE_PROGRESS;
578		goto reassess_streams;
579	}
580
581	if (notes & NEED_RETRY)
582		goto need_retry;
583	if ((notes & MADE_PROGRESS) && test_bit(NETFS_RREQ_PAUSE, &wreq->flags)) {
584		trace_netfs_rreq(wreq, netfs_rreq_trace_unpause);
585		clear_bit_unlock(NETFS_RREQ_PAUSE, &wreq->flags);
586		wake_up_bit(&wreq->flags, NETFS_RREQ_PAUSE);
587	}
588
589	if (notes & NEED_REASSESS) {
590		//cond_resched();
591		goto reassess_streams;
592	}
593	if (notes & MADE_PROGRESS) {
594		//cond_resched();
595		goto reassess_streams;
596	}
597
598out:
599	netfs_put_group_many(wreq->group, wreq->nr_group_rel);
600	wreq->nr_group_rel = 0;
601	_leave(" = %x", notes);
602	return;
603
604need_retry:
605	/* Okay...  We're going to have to retry one or both streams.  Note
606	 * that any partially completed op will have had any wholly transferred
607	 * folios removed from it.
608	 */
609	_debug("retry");
610	netfs_retry_writes(wreq);
611	goto out;
612}
613
614/*
615 * Perform the collection of subrequests, folios and encryption buffers.
616 */
617void netfs_write_collection_worker(struct work_struct *work)
618{
619	struct netfs_io_request *wreq = container_of(work, struct netfs_io_request, work);
620	struct netfs_inode *ictx = netfs_inode(wreq->inode);
621	size_t transferred;
622	int s;
623
624	_enter("R=%x", wreq->debug_id);
625
626	netfs_see_request(wreq, netfs_rreq_trace_see_work);
627	if (!test_bit(NETFS_RREQ_IN_PROGRESS, &wreq->flags)) {
628		netfs_put_request(wreq, false, netfs_rreq_trace_put_work);
629		return;
630	}
631
632	netfs_collect_write_results(wreq);
633
634	/* We're done when the app thread has finished posting subreqs and all
635	 * the queues in all the streams are empty.
636	 */
637	if (!test_bit(NETFS_RREQ_ALL_QUEUED, &wreq->flags)) {
638		netfs_put_request(wreq, false, netfs_rreq_trace_put_work);
639		return;
640	}
641	smp_rmb(); /* Read ALL_QUEUED before lists. */
642
643	transferred = LONG_MAX;
644	for (s = 0; s < NR_IO_STREAMS; s++) {
645		struct netfs_io_stream *stream = &wreq->io_streams[s];
646		if (!stream->active)
647			continue;
648		if (!list_empty(&stream->subrequests)) {
649			netfs_put_request(wreq, false, netfs_rreq_trace_put_work);
650			return;
651		}
652		if (stream->transferred < transferred)
653			transferred = stream->transferred;
654	}
655
656	/* Okay, declare that all I/O is complete. */
657	wreq->transferred = transferred;
658	trace_netfs_rreq(wreq, netfs_rreq_trace_write_done);
659
660	if (wreq->io_streams[1].active &&
661	    wreq->io_streams[1].failed) {
662		/* Cache write failure doesn't prevent writeback completion
663		 * unless we're in disconnected mode.
664		 */
665		ictx->ops->invalidate_cache(wreq);
666	}
667
668	if (wreq->cleanup)
669		wreq->cleanup(wreq);
670
671	if (wreq->origin == NETFS_DIO_WRITE &&
672	    wreq->mapping->nrpages) {
673		/* mmap may have got underfoot and we may now have folios
674		 * locally covering the region we just wrote.  Attempt to
675		 * discard the folios, but leave in place any modified locally.
676		 * ->write_iter() is prevented from interfering by the DIO
677		 * counter.
678		 */
679		pgoff_t first = wreq->start >> PAGE_SHIFT;
680		pgoff_t last = (wreq->start + wreq->transferred - 1) >> PAGE_SHIFT;
681		invalidate_inode_pages2_range(wreq->mapping, first, last);
682	}
683
684	if (wreq->origin == NETFS_DIO_WRITE)
685		inode_dio_end(wreq->inode);
686
687	_debug("finished");
688	trace_netfs_rreq(wreq, netfs_rreq_trace_wake_ip);
689	clear_bit_unlock(NETFS_RREQ_IN_PROGRESS, &wreq->flags);
690	wake_up_bit(&wreq->flags, NETFS_RREQ_IN_PROGRESS);
691
692	if (wreq->iocb) {
693		size_t written = min(wreq->transferred, wreq->len);
694		wreq->iocb->ki_pos += written;
695		if (wreq->iocb->ki_complete)
696			wreq->iocb->ki_complete(
697				wreq->iocb, wreq->error ? wreq->error : written);
698		wreq->iocb = VFS_PTR_POISON;
699	}
700
701	netfs_clear_subrequests(wreq, false);
702	netfs_put_request(wreq, false, netfs_rreq_trace_put_work_complete);
703}
704
705/*
706 * Wake the collection work item.
707 */
708void netfs_wake_write_collector(struct netfs_io_request *wreq, bool was_async)
709{
710	if (!work_pending(&wreq->work)) {
711		netfs_get_request(wreq, netfs_rreq_trace_get_work);
712		if (!queue_work(system_unbound_wq, &wreq->work))
713			netfs_put_request(wreq, was_async, netfs_rreq_trace_put_work_nq);
714	}
715}
716
717/**
718 * netfs_write_subrequest_terminated - Note the termination of a write operation.
719 * @_op: The I/O request that has terminated.
720 * @transferred_or_error: The amount of data transferred or an error code.
721 * @was_async: The termination was asynchronous
722 *
723 * This tells the library that a contributory write I/O operation has
724 * terminated, one way or another, and that it should collect the results.
725 *
726 * The caller indicates in @transferred_or_error the outcome of the operation,
727 * supplying a positive value to indicate the number of bytes transferred or a
728 * negative error code.  The library will look after reissuing I/O operations
729 * as appropriate and writing downloaded data to the cache.
730 *
731 * If @was_async is true, the caller might be running in softirq or interrupt
732 * context and we can't sleep.
733 *
734 * When this is called, ownership of the subrequest is transferred back to the
735 * library, along with a ref.
736 *
737 * Note that %_op is a void* so that the function can be passed to
738 * kiocb::term_func without the need for a casting wrapper.
739 */
740void netfs_write_subrequest_terminated(void *_op, ssize_t transferred_or_error,
741				       bool was_async)
742{
743	struct netfs_io_subrequest *subreq = _op;
744	struct netfs_io_request *wreq = subreq->rreq;
745	struct netfs_io_stream *stream = &wreq->io_streams[subreq->stream_nr];
746
747	_enter("%x[%x] %zd", wreq->debug_id, subreq->debug_index, transferred_or_error);
748
749	switch (subreq->source) {
750	case NETFS_UPLOAD_TO_SERVER:
751		netfs_stat(&netfs_n_wh_upload_done);
752		break;
753	case NETFS_WRITE_TO_CACHE:
754		netfs_stat(&netfs_n_wh_write_done);
755		break;
756	case NETFS_INVALID_WRITE:
757		break;
758	default:
759		BUG();
760	}
761
762	if (IS_ERR_VALUE(transferred_or_error)) {
763		subreq->error = transferred_or_error;
764		if (subreq->error == -EAGAIN)
765			set_bit(NETFS_SREQ_NEED_RETRY, &subreq->flags);
766		else
767			set_bit(NETFS_SREQ_FAILED, &subreq->flags);
768		trace_netfs_failure(wreq, subreq, transferred_or_error, netfs_fail_write);
769
770		switch (subreq->source) {
771		case NETFS_WRITE_TO_CACHE:
772			netfs_stat(&netfs_n_wh_write_failed);
773			break;
774		case NETFS_UPLOAD_TO_SERVER:
775			netfs_stat(&netfs_n_wh_upload_failed);
776			break;
777		default:
778			break;
779		}
780		trace_netfs_rreq(wreq, netfs_rreq_trace_set_pause);
781		set_bit(NETFS_RREQ_PAUSE, &wreq->flags);
782	} else {
783		if (WARN(transferred_or_error > subreq->len - subreq->transferred,
784			 "Subreq excess write: R=%x[%x] %zd > %zu - %zu",
785			 wreq->debug_id, subreq->debug_index,
786			 transferred_or_error, subreq->len, subreq->transferred))
787			transferred_or_error = subreq->len - subreq->transferred;
788
789		subreq->error = 0;
790		subreq->transferred += transferred_or_error;
791
792		if (subreq->transferred < subreq->len)
793			set_bit(NETFS_SREQ_NEED_RETRY, &subreq->flags);
794	}
795
796	trace_netfs_sreq(subreq, netfs_sreq_trace_terminated);
797
798	clear_bit_unlock(NETFS_SREQ_IN_PROGRESS, &subreq->flags);
799	wake_up_bit(&subreq->flags, NETFS_SREQ_IN_PROGRESS);
800
801	/* If we are at the head of the queue, wake up the collector,
802	 * transferring a ref to it if we were the ones to do so.
803	 */
804	if (list_is_first(&subreq->rreq_link, &stream->subrequests))
805		netfs_wake_write_collector(wreq, was_async);
806
807	netfs_put_subrequest(subreq, was_async, netfs_sreq_trace_put_terminated);
808}
809EXPORT_SYMBOL(netfs_write_subrequest_terminated);
810