Merge patch series "netfs: Read performance improvements and "single-blob" support"

David Howells <dhowells@redhat.com> says:

This set of patches is primarily about two things: improving read
performance and supporting monolithic single-blob objects that have to be
read/written as such (e.g. AFS directory contents).  The implementation of
the two parts is interwoven as each makes the other possible.

READ PERFORMANCE
================

The read performance improvements are intended to speed up some loss of
performance detected in cifs and to a lesser extend in afs.  The problem is
that we queue too many work items during the collection of read results:
each individual subrequest is collected by its own work item, and then they
have to interact with each other when a series of subrequests don't exactly
align with the pattern of folios that are being read by the overall
request.

Whilst the processing of the pages covered by individual subrequests as
they complete potentially allows folios to be woken in parallel and with
minimum delay, it can shuffle wakeups for sequential reads out of order -
and that is the most common I/O pattern.

The final assessment and cleanup of an operation is then held up until the
last I/O completes - and for a synchronous sequential operation, this means
the bouncing around of work items just adds latency.

Two changes have been made to make this work:

 (1) All collection is now done in a single "work item" that works
     progressively through the subrequests as they complete (and also
     dispatches retries as necessary).

 (2) For readahead and AIO, this work item be done on a workqueue and can
     run in parallel with the ultimate consumer of the data; for
     synchronous direct or unbuffered reads, the collection is run in the
     application thread and not offloaded.

Functions such as smb2_readv_callback() then just tell netfslib that the
subrequest has terminated; netfslib does a minimal bit of processing on the
spot - stat counting and tracing mostly - and then queues/wakes up the
worker.  This simplifies the logic as the collector just walks sequentially
through the subrequests as they complete and walks through the folios, if
buffered, unlocking them as it goes.  It also keeps to a minimum the amount
of latency injected into the filesystem's low-level I/O handling

The way netfs supports filesystems using the deprecated PG_private_2 flag
is changed: folios are flagged and added to a write request as they
complete and that takes care of scheduling the writes to the cache.  The
originating read request can then just unlock the pages whatever happens.

SINGLE-BLOB OBJECT SUPPORT
==========================

Single-blob objects are files for which the content of the file must be
read from or written to the server in a single operation because reading
them in parts may yield inconsistent results.  AFS directories are an
example of this as there exists the possibility that the contents are
generated on the fly and would differ between reads or might change due to
third party interference.

Such objects will be written to and retrieved from the cache if one is
present, though we allow/may need to propose multiple subrequests to do so.
The important part is that read from/write to the *server* is monolithic.

Single blob reading is, for the moment, fully synchronous and does result
collection in the application thread and, also for the moment, the API is
supplied the buffer in the form of a folio_queue chain rather than using
the pagecache.

AFS CHANGES
===========

This series makes a number of changes to the kafs filesystem, primarily in
the area of directory handling:

 (1) AFS's FetchData RPC reply processing is made partially asynchronous
     which allows the netfs_io_request's outstanding operation counter to
     be removed as part of reducing the collection to a single work item.

 (2) Directory and symlink reading are plumbed through netfslib using the
     single-blob object API and are now cacheable with fscache.  This also
     allows the afs_read struct to be eliminated and netfs_io_subrequest to
     be used directly instead.

 (3) Directory and symlink content are now stored in a folio_queue buffer
     rather than in the pagecache.  This means we don't require the RCU
     read lock and xarray iteration to access it, and folios won't randomly
     disappear under us because the VM wants them back.

     There are some downsides to this, though: the storage folios are no
     longer known to the VM, drop_caches can't flush them, the folios are
     not migrateable.  The inode must also be marked dirty manually to get
     the data written to the cache in the background.

 (4) The vnode operation lock is changed from a mutex struct to a private
     lock implementation.  The problem is that the lock now needs to be
     dropped in a separate thread and mutexes don't permit that.

 (5) When a new directory or symlink is created, we now initialise it
     locally and mark it valid rather than downloading it (we know what
     it's likely to look like).

 (6) We now use the in-directory hashtable to reduce the number of entries
     we need to scan when doing a lookup.  The edit routines have to
     maintain the hash chains.

 (7) Cancellation (e.g. by signal) of an async call after the rxrpc_call
     has been set up is now offloaded to the worker thread as there will be
     a notification from rxrpc upon completion.  This avoids a double
     cleanup.

SUPPORTING CHANGES
==================

To support the above some other changes are also made:

 (1) A "rolling buffer" implementation is created to abstract out the two
     separate folio_queue chaining implementations I had (one for read and
     one for write).

 (2) Functions are provided to create/extend a buffer in a folio_queue
     chain and tear it down again.  This is used to handle AFS directories,
     but could also be used to create bounce buffers for content crypto and
     transport crypto.

 (3) The was_async argument is dropped from netfs_read_subreq_terminated().
     Instead we wake the read collection work item by either queuing it or
     waking up the app thread.

 (4) We don't need to use BH-excluding locks when communicating between the
     issuing thread and the collection thread as neither of them now run in
     BH context.

MISCELLANY
==========

Also included are a number of new tracepoints; a split of the netfslib
write collection code to put retrying into its own file (it gets more
complicated with content encryption).

There are also some minor fixes AFS included, including fixing the AFS
directory format struct layout, reducing some directory over-invalidation
and making afs_mkdir() translate EEXIST to ENOTEMPY (which is not available
on all systems the servers support).

Finally, there's a patch to try and detect entry into the folio unlock
function with no folio_queue structs in the buffer (which isn't allowed in
the cases that can get there).  This is a debugging patch, but should be
minimal overhead.

* patches from https://lore.kernel.org/r/20241216204124.3752367-1-dhowells@redhat.com: (31 commits)
  netfs: Report on NULL folioq in netfs_writeback_unlock_folios()
  afs: Add a tracepoint for afs_read_receive()
  afs: Locally initialise the contents of a new symlink on creation
  afs: Use the contained hashtable to search a directory
  afs: Make afs_mkdir() locally initialise a new directory's content
  netfs: Change the read result collector to only use one work item
  afs: Make {Y,}FS.FetchData an asynchronous operation
  afs: Fix cleanup of immediately failed async calls
  afs: Eliminate afs_read
  afs: Use netfslib for symlinks, allowing them to be cached
  afs: Use netfslib for directories
  afs: Make afs_init_request() get a key if not given a file
  netfs: Add support for caching single monolithic objects such as AFS dirs
  netfs: Add functions to build/clean a buffer in a folio_queue
  afs: Add more tracepoints to do with tracking validity
  cachefiles: Add auxiliary data trace
  cachefiles: Add some subrequest tracepoints
  netfs: Remove some extraneous directory invalidations
  afs: Fix directory format encoding struct
  afs: Fix EEXIST error returned from afs_rmdir() to be ENOTEMPTY
  ...

Link: https://lore.kernel.org/r/20241216204124.3752367-1-dhowells@redhat.com
Signed-off-by: Christian Brauner <brauner@kernel.org>

1 files changed, 226 insertions, 0 deletions

diff --git a/fs/netfs/rolling_buffer.c b/fs/netfs/rolling_buffer.c
new file mode 100644
index 000000000000..75d97af14b4a
--- /dev/null
+++ b/fs/netfs/rolling_buffer.c
@@ -0,0 +1,226 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/* Rolling buffer helpers
+ *
+ * Copyright (C) 2024 Red Hat, Inc. All Rights Reserved.
+ * Written by David Howells (dhowells@redhat.com)
+ */
+
+#include <linux/bitops.h>
+#include <linux/pagemap.h>
+#include <linux/rolling_buffer.h>
+#include <linux/slab.h>
+#include "internal.h"
+
+static atomic_t debug_ids;
+
+/**
+ * netfs_folioq_alloc - Allocate a folio_queue struct
+ * @rreq_id: Associated debugging ID for tracing purposes
+ * @gfp: Allocation constraints
+ * @trace: Trace tag to indicate the purpose of the allocation
+ *
+ * Allocate, initialise and account the folio_queue struct and log a trace line
+ * to mark the allocation.
+ */
+struct folio_queue *netfs_folioq_alloc(unsigned int rreq_id, gfp_t gfp,
+				       unsigned int /*enum netfs_folioq_trace*/ trace)
+{
+	struct folio_queue *fq;
+
+	fq = kmalloc(sizeof(*fq), gfp);
+	if (fq) {
+		netfs_stat(&netfs_n_folioq);
+		folioq_init(fq, rreq_id);
+		fq->debug_id = atomic_inc_return(&debug_ids);
+		trace_netfs_folioq(fq, trace);
+	}
+	return fq;
+}
+EXPORT_SYMBOL(netfs_folioq_alloc);
+
+/**
+ * netfs_folioq_free - Free a folio_queue struct
+ * @folioq: The object to free
+ * @trace: Trace tag to indicate which free
+ *
+ * Free and unaccount the folio_queue struct.
+ */
+void netfs_folioq_free(struct folio_queue *folioq,
+		       unsigned int /*enum netfs_trace_folioq*/ trace)
+{
+	trace_netfs_folioq(folioq, trace);
+	netfs_stat_d(&netfs_n_folioq);
+	kfree(folioq);
+}
+EXPORT_SYMBOL(netfs_folioq_free);
+
+/*
+ * Initialise a rolling buffer.  We allocate an empty folio queue struct to so
+ * that the pointers can be independently driven by the producer and the
+ * consumer.
+ */
+int rolling_buffer_init(struct rolling_buffer *roll, unsigned int rreq_id,
+			unsigned int direction)
+{
+	struct folio_queue *fq;
+
+	fq = netfs_folioq_alloc(rreq_id, GFP_NOFS, netfs_trace_folioq_rollbuf_init);
+	if (!fq)
+		return -ENOMEM;
+
+	roll->head = fq;
+	roll->tail = fq;
+	iov_iter_folio_queue(&roll->iter, direction, fq, 0, 0, 0);
+	return 0;
+}
+
+/*
+ * Add another folio_queue to a rolling buffer if there's no space left.
+ */
+int rolling_buffer_make_space(struct rolling_buffer *roll)
+{
+	struct folio_queue *fq, *head = roll->head;
+
+	if (!folioq_full(head))
+		return 0;
+
+	fq = netfs_folioq_alloc(head->rreq_id, GFP_NOFS, netfs_trace_folioq_make_space);
+	if (!fq)
+		return -ENOMEM;
+	fq->prev = head;
+
+	roll->head = fq;
+	if (folioq_full(head)) {
+		/* Make sure we don't leave the master iterator pointing to a
+		 * block that might get immediately consumed.
+		 */
+		if (roll->iter.folioq == head &&
+		    roll->iter.folioq_slot == folioq_nr_slots(head)) {
+			roll->iter.folioq = fq;
+			roll->iter.folioq_slot = 0;
+		}
+	}
+
+	/* Make sure the initialisation is stored before the next pointer.
+	 *
+	 * [!] NOTE: After we set head->next, the consumer is at liberty to
+	 * immediately delete the old head.
+	 */
+	smp_store_release(&head->next, fq);
+	return 0;
+}
+
+/*
+ * Decant the list of folios to read into a rolling buffer.
+ */
+ssize_t rolling_buffer_load_from_ra(struct rolling_buffer *roll,
+				    struct readahead_control *ractl,
+				    struct folio_batch *put_batch)
+{
+	struct folio_queue *fq;
+	struct page **vec;
+	int nr, ix, to;
+	ssize_t size = 0;
+
+	if (rolling_buffer_make_space(roll) < 0)
+		return -ENOMEM;
+
+	fq = roll->head;
+	vec = (struct page **)fq->vec.folios;
+	nr = __readahead_batch(ractl, vec + folio_batch_count(&fq->vec),
+			       folio_batch_space(&fq->vec));
+	ix = fq->vec.nr;
+	to = ix + nr;
+	fq->vec.nr = to;
+	for (; ix < to; ix++) {
+		struct folio *folio = folioq_folio(fq, ix);
+		unsigned int order = folio_order(folio);
+
+		fq->orders[ix] = order;
+		size += PAGE_SIZE << order;
+		trace_netfs_folio(folio, netfs_folio_trace_read);
+		if (!folio_batch_add(put_batch, folio))
+			folio_batch_release(put_batch);
+	}
+	WRITE_ONCE(roll->iter.count, roll->iter.count + size);
+
+	/* Store the counter after setting the slot. */
+	smp_store_release(&roll->next_head_slot, to);
+
+	for (; ix < folioq_nr_slots(fq); ix++)
+		folioq_clear(fq, ix);
+
+	return size;
+}
+
+/*
+ * Append a folio to the rolling buffer.
+ */
+ssize_t rolling_buffer_append(struct rolling_buffer *roll, struct folio *folio,
+			      unsigned int flags)
+{
+	ssize_t size = folio_size(folio);
+	int slot;
+
+	if (rolling_buffer_make_space(roll) < 0)
+		return -ENOMEM;
+
+	slot = folioq_append(roll->head, folio);
+	if (flags & ROLLBUF_MARK_1)
+		folioq_mark(roll->head, slot);
+	if (flags & ROLLBUF_MARK_2)
+		folioq_mark2(roll->head, slot);
+
+	WRITE_ONCE(roll->iter.count, roll->iter.count + size);
+
+	/* Store the counter after setting the slot. */
+	smp_store_release(&roll->next_head_slot, slot);
+	return size;
+}
+
+/*
+ * Delete a spent buffer from a rolling queue and return the next in line.  We
+ * don't return the last buffer to keep the pointers independent, but return
+ * NULL instead.
+ */
+struct folio_queue *rolling_buffer_delete_spent(struct rolling_buffer *roll)
+{
+	struct folio_queue *spent = roll->tail, *next = READ_ONCE(spent->next);
+
+	if (!next)
+		return NULL;
+	next->prev = NULL;
+	netfs_folioq_free(spent, netfs_trace_folioq_delete);
+	roll->tail = next;
+	return next;
+}
+
+/*
+ * Clear out a rolling queue.  Folios that have mark 1 set are put.
+ */
+void rolling_buffer_clear(struct rolling_buffer *roll)
+{
+	struct folio_batch fbatch;
+	struct folio_queue *p;
+
+	folio_batch_init(&fbatch);
+
+	while ((p = roll->tail)) {
+		roll->tail = p->next;
+		for (int slot = 0; slot < folioq_count(p); slot++) {
+			struct folio *folio = folioq_folio(p, slot);
+
+			if (!folio)
+				continue;
+			if (folioq_is_marked(p, slot)) {
+				trace_netfs_folio(folio, netfs_folio_trace_put);
+				if (!folio_batch_add(&fbatch, folio))
+					folio_batch_release(&fbatch);
+			}
+		}
+
+		netfs_folioq_free(p, netfs_trace_folioq_clear);
+	}
+
+	folio_batch_release(&fbatch);
+}