* [PATCH 0/4] netfs: Miscellaneous fixes
@ 2025-05-19 9:07 David Howells
2025-05-19 9:07 ` [PATCH 1/4] netfs: Fix oops in write-retry from mis-resetting the subreq iterator David Howells
` (4 more replies)
0 siblings, 5 replies; 6+ messages in thread
From: David Howells @ 2025-05-19 9:07 UTC (permalink / raw)
To: Christian Brauner
Cc: David Howells, Paulo Alcantara, netfs, linux-afs, linux-cifs,
linux-nfs, ceph-devel, v9fs, linux-fsdevel, linux-kernel
Hi Christian,
Here are some miscellaneous fixes and changes for netfslib, if you could
pull them:
(1) Fix an oops in write-retry due to mis-resetting the I/O iterator.
(2) Fix the recording of transferred bytes for short DIO reads.
(3) Fix a request's work item to not require a reference, thereby avoiding
the need to get rid of it in BH/IRQ context.
(4) Fix waiting and waking to be consistent about the waitqueue used.
The patches can also be found here:
https://git.kernel.org/pub/scm/linux/kernel/git/dhowells/linux-fs.git/log/?h=netfs-fixes
Thanks,
David
David Howells (3):
netfs: Fix oops in write-retry from mis-resetting the subreq iterator
netfs: Fix the request's work item to not require a ref
netfs: Fix wait/wake to be consistent about the waitqueue used
Paulo Alcantara (1):
netfs: Fix setting of transferred bytes with short DIO reads
fs/9p/vfs_addr.c | 2 +-
fs/afs/write.c | 8 +-
fs/cachefiles/io.c | 16 +--
fs/ceph/addr.c | 2 +-
fs/erofs/fscache.c | 6 +-
fs/netfs/buffered_read.c | 32 +++--
fs/netfs/buffered_write.c | 2 +-
fs/netfs/direct_read.c | 10 +-
fs/netfs/direct_write.c | 12 +-
fs/netfs/fscache_io.c | 10 +-
fs/netfs/internal.h | 42 +++++--
fs/netfs/misc.c | 218 +++++++++++++++++++++++++++++++++++
fs/netfs/objects.c | 47 ++++----
fs/netfs/read_collect.c | 178 ++++------------------------
fs/netfs/read_pgpriv2.c | 4 +-
fs/netfs/read_retry.c | 26 +----
fs/netfs/read_single.c | 6 +-
fs/netfs/write_collect.c | 81 +++++--------
fs/netfs/write_issue.c | 38 +++---
fs/netfs/write_retry.c | 19 ++-
fs/smb/client/cifsproto.h | 3 +-
fs/smb/client/cifssmb.c | 4 +-
fs/smb/client/file.c | 7 +-
fs/smb/client/smb2pdu.c | 4 +-
include/linux/fscache.h | 2 +-
include/linux/netfs.h | 14 +--
include/trace/events/netfs.h | 7 +-
net/9p/client.c | 6 +-
28 files changed, 427 insertions(+), 379 deletions(-)
^ permalink raw reply [flat|nested] 6+ messages in thread
* [PATCH 1/4] netfs: Fix oops in write-retry from mis-resetting the subreq iterator
2025-05-19 9:07 [PATCH 0/4] netfs: Miscellaneous fixes David Howells
@ 2025-05-19 9:07 ` David Howells
2025-05-19 9:07 ` [PATCH 2/4] netfs: Fix setting of transferred bytes with short DIO reads David Howells
` (3 subsequent siblings)
4 siblings, 0 replies; 6+ messages in thread
From: David Howells @ 2025-05-19 9:07 UTC (permalink / raw)
To: Christian Brauner
Cc: David Howells, Paulo Alcantara, netfs, linux-afs, linux-cifs,
linux-nfs, ceph-devel, v9fs, linux-fsdevel, linux-kernel,
syzbot+25b83a6f2c702075fcbc
Fix the resetting of the subrequest iterator in netfs_retry_write_stream()
to use the iterator-reset function as the iterator may have been shortened
by a previous retry. In such a case, the amount of data to be written by
the subrequest is not "subreq->len" but "subreq->len -
subreq->transferred".
Without this, KASAN may see an error in iov_iter_revert():
BUG: KASAN: slab-out-of-bounds in iov_iter_revert lib/iov_iter.c:633 [inline]
BUG: KASAN: slab-out-of-bounds in iov_iter_revert+0x443/0x5a0 lib/iov_iter.c:611
Read of size 4 at addr ffff88802912a0b8 by task kworker/u32:7/1147
CPU: 1 UID: 0 PID: 1147 Comm: kworker/u32:7 Not tainted 6.15.0-rc6-syzkaller-00052-g9f35e33144ae #0 PREEMPT(full)
Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.16.3-debian-1.16.3-2~bpo12+1 04/01/2014
Workqueue: events_unbound netfs_write_collection_worker
Call Trace:
<TASK>
__dump_stack lib/dump_stack.c:94 [inline]
dump_stack_lvl+0x116/0x1f0 lib/dump_stack.c:120
print_address_description mm/kasan/report.c:408 [inline]
print_report+0xc3/0x670 mm/kasan/report.c:521
kasan_report+0xe0/0x110 mm/kasan/report.c:634
iov_iter_revert lib/iov_iter.c:633 [inline]
iov_iter_revert+0x443/0x5a0 lib/iov_iter.c:611
netfs_retry_write_stream fs/netfs/write_retry.c:44 [inline]
netfs_retry_writes+0x166d/0x1a50 fs/netfs/write_retry.c:231
netfs_collect_write_results fs/netfs/write_collect.c:352 [inline]
netfs_write_collection_worker+0x23fd/0x3830 fs/netfs/write_collect.c:374
process_one_work+0x9cf/0x1b70 kernel/workqueue.c:3238
process_scheduled_works kernel/workqueue.c:3319 [inline]
worker_thread+0x6c8/0xf10 kernel/workqueue.c:3400
kthread+0x3c2/0x780 kernel/kthread.c:464
ret_from_fork+0x45/0x80 arch/x86/kernel/process.c:153
ret_from_fork_asm+0x1a/0x30 arch/x86/entry/entry_64.S:245
</TASK>
Fixes: cd0277ed0c18 ("netfs: Use new folio_queue data type and iterator instead of xarray iter")
Reported-by: syzbot+25b83a6f2c702075fcbc@syzkaller.appspotmail.com
Closes: https://syzkaller.appspot.com/bug?extid=25b83a6f2c702075fcbc
Signed-off-by: David Howells <dhowells@redhat.com>
cc: Paulo Alcantara <pc@manguebit.com>
cc: netfs@lists.linux.dev
cc: linux-fsdevel@vger.kernel.org
Tested-by: syzbot+25b83a6f2c702075fcbc@syzkaller.appspotmail.com
---
fs/netfs/write_retry.c | 5 +++--
1 file changed, 3 insertions(+), 2 deletions(-)
diff --git a/fs/netfs/write_retry.c b/fs/netfs/write_retry.c
index 545d33079a77..9b1ca8b0f4dd 100644
--- a/fs/netfs/write_retry.c
+++ b/fs/netfs/write_retry.c
@@ -39,9 +39,10 @@ static void netfs_retry_write_stream(struct netfs_io_request *wreq,
if (test_bit(NETFS_SREQ_FAILED, &subreq->flags))
break;
if (__test_and_clear_bit(NETFS_SREQ_NEED_RETRY, &subreq->flags)) {
- struct iov_iter source = subreq->io_iter;
+ struct iov_iter source;
- iov_iter_revert(&source, subreq->len - source.count);
+ netfs_reset_iter(subreq);
+ source = subreq->io_iter;
netfs_get_subrequest(subreq, netfs_sreq_trace_get_resubmit);
netfs_reissue_write(stream, subreq, &source);
}
^ permalink raw reply related [flat|nested] 6+ messages in thread
* [PATCH 2/4] netfs: Fix setting of transferred bytes with short DIO reads
2025-05-19 9:07 [PATCH 0/4] netfs: Miscellaneous fixes David Howells
2025-05-19 9:07 ` [PATCH 1/4] netfs: Fix oops in write-retry from mis-resetting the subreq iterator David Howells
@ 2025-05-19 9:07 ` David Howells
2025-05-19 9:07 ` [PATCH 3/4] netfs: Fix the request's work item to not require a ref David Howells
` (2 subsequent siblings)
4 siblings, 0 replies; 6+ messages in thread
From: David Howells @ 2025-05-19 9:07 UTC (permalink / raw)
To: Christian Brauner
Cc: David Howells, Paulo Alcantara, netfs, linux-afs, linux-cifs,
linux-nfs, ceph-devel, v9fs, linux-fsdevel, linux-kernel,
Nicolas Baranger
From: Paulo Alcantara <pc@manguebit.com>
A netfslib request comprises an ordered stream of subrequests that, when
doing an unbuffered/DIO read, are contiguous. The subrequests may be
performed in parallel, but may not be fully completed.
For instance, if we try and make a 256KiB DIO read from a 3-byte file with
a 64KiB rsize and 256KiB bsize, netfslib will attempt to make a read of
256KiB, broken up into four 64KiB subreads, with the expectation that the
first will be short and the subsequent three be completely devoid - but we
do all four on the basis that the file may have been changed by a third
party.
The read-collection code, however, walks through all the subreqs and
advances the notion of how much data has been read in the stream to the
start of each subreq plus its amount transferred (which are 3, 0, 0, 0 for
the example above) - which gives an amount apparently read of 3*64KiB -
which is incorrect.
Fix the collection code to cut short the calculation of the transferred
amount with the first short subrequest in an unbuffered read; everything
beyond that must be ignored as there's a hole that cannot be filled. This
applies both to shortness due to hitting the EOF and shortness due to an
error.
This is achieved by setting a flag on the request when we collect the first
short subrequest (collection is done in ascending order).
This can be tested by mounting a cifs volume with rsize=65536,bsize=262144
and doing a 256k DIO read of a very small file (e.g. 3 bytes). read()
should return 3, not >3.
This problem came in when netfs_read_collection() set rreq->transferred to
stream->transferred, even for DIO. Prior to that, netfs_rreq_assess_dio()
just went over the list and added up the subreqs till it met a short one -
but now the subreqs are discarded earlier.
Fixes: e2d46f2ec332 ("netfs: Change the read result collector to only use one work item")
Reported-by: Nicolas Baranger <nicolas.baranger@3xo.fr>
Closes: https://lore.kernel.org/all/10bec2430ed4df68bde10ed95295d093@3xo.fr/
Signed-off-by: Paulo Alcantara (Red Hat) <pc@manguebit.com>
Signed-off-by: David Howells <dhowells@redhat.com>
cc: netfs@lists.linux.dev
cc: linux-fsdevel@vger.kernel.org
---
fs/netfs/read_collect.c | 21 +++++----------------
include/linux/netfs.h | 1 +
2 files changed, 6 insertions(+), 16 deletions(-)
diff --git a/fs/netfs/read_collect.c b/fs/netfs/read_collect.c
index 23c75755ad4e..d3cf27b2697c 100644
--- a/fs/netfs/read_collect.c
+++ b/fs/netfs/read_collect.c
@@ -280,9 +280,13 @@ static void netfs_collect_read_results(struct netfs_io_request *rreq)
stream->need_retry = true;
notes |= NEED_RETRY | MADE_PROGRESS;
break;
+ } else if (test_bit(NETFS_RREQ_SHORT_TRANSFER, &rreq->flags)) {
+ notes |= MADE_PROGRESS;
} else {
if (!stream->failed)
- stream->transferred = stream->collected_to - rreq->start;
+ stream->transferred += transferred;
+ if (front->transferred < front->len)
+ set_bit(NETFS_RREQ_SHORT_TRANSFER, &rreq->flags);
notes |= MADE_PROGRESS;
}
@@ -342,23 +346,8 @@ static void netfs_collect_read_results(struct netfs_io_request *rreq)
*/
static void netfs_rreq_assess_dio(struct netfs_io_request *rreq)
{
- struct netfs_io_subrequest *subreq;
- struct netfs_io_stream *stream = &rreq->io_streams[0];
unsigned int i;
- /* Collect unbuffered reads and direct reads, adding up the transfer
- * sizes until we find the first short or failed subrequest.
- */
- list_for_each_entry(subreq, &stream->subrequests, rreq_link) {
- rreq->transferred += subreq->transferred;
-
- if (subreq->transferred < subreq->len ||
- test_bit(NETFS_SREQ_FAILED, &subreq->flags)) {
- rreq->error = subreq->error;
- break;
- }
- }
-
if (rreq->origin == NETFS_DIO_READ) {
for (i = 0; i < rreq->direct_bv_count; i++) {
flush_dcache_page(rreq->direct_bv[i].bv_page);
diff --git a/include/linux/netfs.h b/include/linux/netfs.h
index c86a11cfc4a3..497c4f4698f6 100644
--- a/include/linux/netfs.h
+++ b/include/linux/netfs.h
@@ -279,6 +279,7 @@ struct netfs_io_request {
#define NETFS_RREQ_USE_IO_ITER 12 /* Use ->io_iter rather than ->i_pages */
#define NETFS_RREQ_ALL_QUEUED 13 /* All subreqs are now queued */
#define NETFS_RREQ_RETRYING 14 /* Set if we're in the retry path */
+#define NETFS_RREQ_SHORT_TRANSFER 15 /* Set if we have a short transfer */
#define NETFS_RREQ_USE_PGPRIV2 31 /* [DEPRECATED] Use PG_private_2 to mark
* write to cache on read */
const struct netfs_request_ops *netfs_ops;
^ permalink raw reply related [flat|nested] 6+ messages in thread
* [PATCH 3/4] netfs: Fix the request's work item to not require a ref
2025-05-19 9:07 [PATCH 0/4] netfs: Miscellaneous fixes David Howells
2025-05-19 9:07 ` [PATCH 1/4] netfs: Fix oops in write-retry from mis-resetting the subreq iterator David Howells
2025-05-19 9:07 ` [PATCH 2/4] netfs: Fix setting of transferred bytes with short DIO reads David Howells
@ 2025-05-19 9:07 ` David Howells
2025-05-19 9:07 ` [PATCH 4/4] netfs: Fix wait/wake to be consistent about the waitqueue used David Howells
2025-05-20 11:34 ` [PATCH 0/4] netfs: Miscellaneous fixes Christian Brauner
4 siblings, 0 replies; 6+ messages in thread
From: David Howells @ 2025-05-19 9:07 UTC (permalink / raw)
To: Christian Brauner
Cc: David Howells, Paulo Alcantara, netfs, linux-afs, linux-cifs,
linux-nfs, ceph-devel, v9fs, linux-fsdevel, linux-kernel,
Marc Dionne, Steve French
When the netfs_io_request struct's work item is queued, it must be supplied
with a ref to the work item struct to prevent it being deallocated whilst
on the queue or whilst it is being processed. This is tricky to manage as
we have to get a ref before we try and queue it and then we may find it's
already queued and is thus already holding a ref - in which case we have to
try and get rid of the ref again.
The problem comes if we're in BH or IRQ context and need to drop the ref:
if netfs_put_request() reduces the count to 0, we have to do the cleanup -
but the cleanup may need to wait.
Fix this by adding a new work item to the request, ->cleanup_work, and
dispatching that when the refcount hits zero. That can then synchronously
cancel any outstanding work on the main work item before doing the cleanup.
Adding a new work item also deals with another problem upstream where it's
sometimes changing the work func in the put function and requeuing it -
which has occasionally in the past caused the cleanup to happen
incorrectly.
As a bonus, this allows us to get rid of the 'was_async' parameter from a
bunch of functions. This indicated whether the put function might not be
permitted to sleep.
Fixes: 3d3c95046742 ("netfs: Provide readahead and readpage netfs helpers")
Signed-off-by: David Howells <dhowells@redhat.com>
cc: Paulo Alcantara <pc@manguebit.com>
cc: Marc Dionne <marc.dionne@auristor.com>
cc: Steve French <stfrench@microsoft.com>
cc: linux-cifs@vger.kernel.org
cc: netfs@lists.linux.dev
cc: linux-fsdevel@vger.kernel.org
---
fs/9p/vfs_addr.c | 2 +-
fs/afs/write.c | 8 ++---
fs/cachefiles/io.c | 16 +++++-----
fs/ceph/addr.c | 2 +-
fs/erofs/fscache.c | 6 ++--
fs/netfs/buffered_read.c | 30 +++++++++---------
fs/netfs/direct_read.c | 6 ++--
fs/netfs/direct_write.c | 2 +-
fs/netfs/fscache_io.c | 10 +++---
fs/netfs/internal.h | 11 +++----
fs/netfs/objects.c | 47 +++++++++++++--------------
fs/netfs/read_collect.c | 44 ++++++++++++++++----------
fs/netfs/read_pgpriv2.c | 4 +--
fs/netfs/read_retry.c | 2 +-
fs/netfs/read_single.c | 6 ++--
fs/netfs/write_collect.c | 61 +++++++++++++++++-------------------
fs/netfs/write_issue.c | 16 +++++-----
fs/netfs/write_retry.c | 2 +-
fs/smb/client/cifsproto.h | 3 +-
fs/smb/client/cifssmb.c | 4 +--
fs/smb/client/file.c | 7 ++---
fs/smb/client/smb2pdu.c | 4 +--
include/linux/fscache.h | 2 +-
include/linux/netfs.h | 13 ++++----
include/trace/events/netfs.h | 7 ++---
net/9p/client.c | 6 ++--
26 files changed, 159 insertions(+), 162 deletions(-)
diff --git a/fs/9p/vfs_addr.c b/fs/9p/vfs_addr.c
index 32619d146cbc..b5a4a28e0fe7 100644
--- a/fs/9p/vfs_addr.c
+++ b/fs/9p/vfs_addr.c
@@ -59,7 +59,7 @@ static void v9fs_issue_write(struct netfs_io_subrequest *subreq)
len = p9_client_write(fid, subreq->start, &subreq->io_iter, &err);
if (len > 0)
__set_bit(NETFS_SREQ_MADE_PROGRESS, &subreq->flags);
- netfs_write_subrequest_terminated(subreq, len ?: err, false);
+ netfs_write_subrequest_terminated(subreq, len ?: err);
}
/**
diff --git a/fs/afs/write.c b/fs/afs/write.c
index 18b0a9f1615e..7df7b2f5e7b2 100644
--- a/fs/afs/write.c
+++ b/fs/afs/write.c
@@ -120,17 +120,17 @@ static void afs_issue_write_worker(struct work_struct *work)
#if 0 // Error injection
if (subreq->debug_index == 3)
- return netfs_write_subrequest_terminated(subreq, -ENOANO, false);
+ return netfs_write_subrequest_terminated(subreq, -ENOANO);
if (!subreq->retry_count) {
set_bit(NETFS_SREQ_NEED_RETRY, &subreq->flags);
- return netfs_write_subrequest_terminated(subreq, -EAGAIN, false);
+ return netfs_write_subrequest_terminated(subreq, -EAGAIN);
}
#endif
op = afs_alloc_operation(wreq->netfs_priv, vnode->volume);
if (IS_ERR(op))
- return netfs_write_subrequest_terminated(subreq, -EAGAIN, false);
+ return netfs_write_subrequest_terminated(subreq, -EAGAIN);
afs_op_set_vnode(op, 0, vnode);
op->file[0].dv_delta = 1;
@@ -166,7 +166,7 @@ static void afs_issue_write_worker(struct work_struct *work)
break;
}
- netfs_write_subrequest_terminated(subreq, ret < 0 ? ret : subreq->len, false);
+ netfs_write_subrequest_terminated(subreq, ret < 0 ? ret : subreq->len);
}
void afs_issue_write(struct netfs_io_subrequest *subreq)
diff --git a/fs/cachefiles/io.c b/fs/cachefiles/io.c
index 92058ae43488..c08e4a66ac07 100644
--- a/fs/cachefiles/io.c
+++ b/fs/cachefiles/io.c
@@ -63,7 +63,7 @@ static void cachefiles_read_complete(struct kiocb *iocb, long ret)
ret = -ESTALE;
}
- ki->term_func(ki->term_func_priv, ret, ki->was_async);
+ ki->term_func(ki->term_func_priv, ret);
}
cachefiles_put_kiocb(ki);
@@ -188,7 +188,7 @@ static int cachefiles_read(struct netfs_cache_resources *cres,
presubmission_error:
if (term_func)
- term_func(term_func_priv, ret < 0 ? ret : skipped, false);
+ term_func(term_func_priv, ret < 0 ? ret : skipped);
return ret;
}
@@ -271,7 +271,7 @@ static void cachefiles_write_complete(struct kiocb *iocb, long ret)
atomic_long_sub(ki->b_writing, &object->volume->cache->b_writing);
set_bit(FSCACHE_COOKIE_HAVE_DATA, &object->cookie->flags);
if (ki->term_func)
- ki->term_func(ki->term_func_priv, ret, ki->was_async);
+ ki->term_func(ki->term_func_priv, ret);
cachefiles_put_kiocb(ki);
}
@@ -301,7 +301,7 @@ int __cachefiles_write(struct cachefiles_object *object,
ki = kzalloc(sizeof(struct cachefiles_kiocb), GFP_KERNEL);
if (!ki) {
if (term_func)
- term_func(term_func_priv, -ENOMEM, false);
+ term_func(term_func_priv, -ENOMEM);
return -ENOMEM;
}
@@ -366,7 +366,7 @@ static int cachefiles_write(struct netfs_cache_resources *cres,
{
if (!fscache_wait_for_operation(cres, FSCACHE_WANT_WRITE)) {
if (term_func)
- term_func(term_func_priv, -ENOBUFS, false);
+ term_func(term_func_priv, -ENOBUFS);
trace_netfs_sreq(term_func_priv, netfs_sreq_trace_cache_nowrite);
return -ENOBUFS;
}
@@ -665,7 +665,7 @@ static void cachefiles_issue_write(struct netfs_io_subrequest *subreq)
pre = CACHEFILES_DIO_BLOCK_SIZE - off;
if (pre >= len) {
fscache_count_dio_misfit();
- netfs_write_subrequest_terminated(subreq, len, false);
+ netfs_write_subrequest_terminated(subreq, len);
return;
}
subreq->transferred += pre;
@@ -691,7 +691,7 @@ static void cachefiles_issue_write(struct netfs_io_subrequest *subreq)
len -= post;
if (len == 0) {
fscache_count_dio_misfit();
- netfs_write_subrequest_terminated(subreq, post, false);
+ netfs_write_subrequest_terminated(subreq, post);
return;
}
iov_iter_truncate(&subreq->io_iter, len);
@@ -703,7 +703,7 @@ static void cachefiles_issue_write(struct netfs_io_subrequest *subreq)
&start, &len, len, true);
cachefiles_end_secure(cache, saved_cred);
if (ret < 0) {
- netfs_write_subrequest_terminated(subreq, ret, false);
+ netfs_write_subrequest_terminated(subreq, ret);
return;
}
diff --git a/fs/ceph/addr.c b/fs/ceph/addr.c
index 29be367905a1..557c326561fd 100644
--- a/fs/ceph/addr.c
+++ b/fs/ceph/addr.c
@@ -539,7 +539,7 @@ static void ceph_set_page_fscache(struct page *page)
folio_start_private_2(page_folio(page)); /* [DEPRECATED] */
}
-static void ceph_fscache_write_terminated(void *priv, ssize_t error, bool was_async)
+static void ceph_fscache_write_terminated(void *priv, ssize_t error)
{
struct inode *inode = priv;
diff --git a/fs/erofs/fscache.c b/fs/erofs/fscache.c
index 9c9129bca346..34517ca9df91 100644
--- a/fs/erofs/fscache.c
+++ b/fs/erofs/fscache.c
@@ -102,8 +102,7 @@ static void erofs_fscache_req_io_put(struct erofs_fscache_io *io)
erofs_fscache_req_put(req);
}
-static void erofs_fscache_req_end_io(void *priv,
- ssize_t transferred_or_error, bool was_async)
+static void erofs_fscache_req_end_io(void *priv, ssize_t transferred_or_error)
{
struct erofs_fscache_io *io = priv;
struct erofs_fscache_rq *req = io->private;
@@ -180,8 +179,7 @@ struct erofs_fscache_bio {
struct bio_vec bvecs[BIO_MAX_VECS];
};
-static void erofs_fscache_bio_endio(void *priv,
- ssize_t transferred_or_error, bool was_async)
+static void erofs_fscache_bio_endio(void *priv, ssize_t transferred_or_error)
{
struct erofs_fscache_bio *io = priv;
diff --git a/fs/netfs/buffered_read.c b/fs/netfs/buffered_read.c
index 0d1b6d35ff3b..cb6202efc466 100644
--- a/fs/netfs/buffered_read.c
+++ b/fs/netfs/buffered_read.c
@@ -262,9 +262,9 @@ static void netfs_read_to_pagecache(struct netfs_io_request *rreq)
if (ret < 0) {
subreq->error = ret;
/* Not queued - release both refs. */
- netfs_put_subrequest(subreq, false,
+ netfs_put_subrequest(subreq,
netfs_sreq_trace_put_cancel);
- netfs_put_subrequest(subreq, false,
+ netfs_put_subrequest(subreq,
netfs_sreq_trace_put_cancel);
break;
}
@@ -297,8 +297,8 @@ static void netfs_read_to_pagecache(struct netfs_io_request *rreq)
subreq->error = ret;
trace_netfs_sreq(subreq, netfs_sreq_trace_cancel);
/* Not queued - release both refs. */
- netfs_put_subrequest(subreq, false, netfs_sreq_trace_put_cancel);
- netfs_put_subrequest(subreq, false, netfs_sreq_trace_put_cancel);
+ netfs_put_subrequest(subreq, netfs_sreq_trace_put_cancel);
+ netfs_put_subrequest(subreq, netfs_sreq_trace_put_cancel);
break;
}
size -= slice;
@@ -365,12 +365,10 @@ void netfs_readahead(struct readahead_control *ractl)
goto cleanup_free;
netfs_read_to_pagecache(rreq);
- netfs_put_request(rreq, true, netfs_rreq_trace_put_return);
- return;
+ return netfs_put_request(rreq, netfs_rreq_trace_put_return);
cleanup_free:
- netfs_put_request(rreq, false, netfs_rreq_trace_put_failed);
- return;
+ return netfs_put_request(rreq, netfs_rreq_trace_put_failed);
}
EXPORT_SYMBOL(netfs_readahead);
@@ -470,11 +468,11 @@ static int netfs_read_gaps(struct file *file, struct folio *folio)
folio_mark_uptodate(folio);
}
folio_unlock(folio);
- netfs_put_request(rreq, false, netfs_rreq_trace_put_return);
+ netfs_put_request(rreq, netfs_rreq_trace_put_return);
return ret < 0 ? ret : 0;
discard:
- netfs_put_request(rreq, false, netfs_rreq_trace_put_discard);
+ netfs_put_request(rreq, netfs_rreq_trace_put_discard);
alloc_error:
folio_unlock(folio);
return ret;
@@ -530,11 +528,11 @@ int netfs_read_folio(struct file *file, struct folio *folio)
netfs_read_to_pagecache(rreq);
ret = netfs_wait_for_read(rreq);
- netfs_put_request(rreq, false, netfs_rreq_trace_put_return);
+ netfs_put_request(rreq, netfs_rreq_trace_put_return);
return ret < 0 ? ret : 0;
discard:
- netfs_put_request(rreq, false, netfs_rreq_trace_put_discard);
+ netfs_put_request(rreq, netfs_rreq_trace_put_discard);
alloc_error:
folio_unlock(folio);
return ret;
@@ -689,7 +687,7 @@ int netfs_write_begin(struct netfs_inode *ctx,
ret = netfs_wait_for_read(rreq);
if (ret < 0)
goto error;
- netfs_put_request(rreq, false, netfs_rreq_trace_put_return);
+ netfs_put_request(rreq, netfs_rreq_trace_put_return);
have_folio:
ret = folio_wait_private_2_killable(folio);
@@ -701,7 +699,7 @@ int netfs_write_begin(struct netfs_inode *ctx,
return 0;
error_put:
- netfs_put_request(rreq, false, netfs_rreq_trace_put_failed);
+ netfs_put_request(rreq, netfs_rreq_trace_put_failed);
error:
if (folio) {
folio_unlock(folio);
@@ -752,11 +750,11 @@ int netfs_prefetch_for_write(struct file *file, struct folio *folio,
netfs_read_to_pagecache(rreq);
ret = netfs_wait_for_read(rreq);
- netfs_put_request(rreq, false, netfs_rreq_trace_put_return);
+ netfs_put_request(rreq, netfs_rreq_trace_put_return);
return ret < 0 ? ret : 0;
error_put:
- netfs_put_request(rreq, false, netfs_rreq_trace_put_discard);
+ netfs_put_request(rreq, netfs_rreq_trace_put_discard);
error:
_leave(" = %d", ret);
return ret;
diff --git a/fs/netfs/direct_read.c b/fs/netfs/direct_read.c
index 5e3f0aeb51f3..cb3c6dc0b165 100644
--- a/fs/netfs/direct_read.c
+++ b/fs/netfs/direct_read.c
@@ -85,7 +85,7 @@ static int netfs_dispatch_unbuffered_reads(struct netfs_io_request *rreq)
if (rreq->netfs_ops->prepare_read) {
ret = rreq->netfs_ops->prepare_read(subreq);
if (ret < 0) {
- netfs_put_subrequest(subreq, false, netfs_sreq_trace_put_cancel);
+ netfs_put_subrequest(subreq, netfs_sreq_trace_put_cancel);
break;
}
}
@@ -144,7 +144,7 @@ static ssize_t netfs_unbuffered_read(struct netfs_io_request *rreq, bool sync)
ret = netfs_dispatch_unbuffered_reads(rreq);
if (!rreq->submitted) {
- netfs_put_request(rreq, false, netfs_rreq_trace_put_no_submit);
+ netfs_put_request(rreq, netfs_rreq_trace_put_no_submit);
inode_dio_end(rreq->inode);
ret = 0;
goto out;
@@ -236,7 +236,7 @@ ssize_t netfs_unbuffered_read_iter_locked(struct kiocb *iocb, struct iov_iter *i
}
out:
- netfs_put_request(rreq, false, netfs_rreq_trace_put_return);
+ netfs_put_request(rreq, netfs_rreq_trace_put_return);
if (ret > 0)
orig_count -= ret;
return ret;
diff --git a/fs/netfs/direct_write.c b/fs/netfs/direct_write.c
index 42ce53cc216e..c98f1676f86d 100644
--- a/fs/netfs/direct_write.c
+++ b/fs/netfs/direct_write.c
@@ -117,7 +117,7 @@ ssize_t netfs_unbuffered_write_iter_locked(struct kiocb *iocb, struct iov_iter *
}
out:
- netfs_put_request(wreq, false, netfs_rreq_trace_put_return);
+ netfs_put_request(wreq, netfs_rreq_trace_put_return);
return ret;
}
EXPORT_SYMBOL(netfs_unbuffered_write_iter_locked);
diff --git a/fs/netfs/fscache_io.c b/fs/netfs/fscache_io.c
index b1722a82c03d..e4308457633c 100644
--- a/fs/netfs/fscache_io.c
+++ b/fs/netfs/fscache_io.c
@@ -192,8 +192,7 @@ EXPORT_SYMBOL(__fscache_clear_page_bits);
/*
* Deal with the completion of writing the data to the cache.
*/
-static void fscache_wreq_done(void *priv, ssize_t transferred_or_error,
- bool was_async)
+static void fscache_wreq_done(void *priv, ssize_t transferred_or_error)
{
struct fscache_write_request *wreq = priv;
@@ -202,8 +201,7 @@ static void fscache_wreq_done(void *priv, ssize_t transferred_or_error,
wreq->set_bits);
if (wreq->term_func)
- wreq->term_func(wreq->term_func_priv, transferred_or_error,
- was_async);
+ wreq->term_func(wreq->term_func_priv, transferred_or_error);
fscache_end_operation(&wreq->cache_resources);
kfree(wreq);
}
@@ -255,14 +253,14 @@ void __fscache_write_to_cache(struct fscache_cookie *cookie,
return;
abandon_end:
- return fscache_wreq_done(wreq, ret, false);
+ return fscache_wreq_done(wreq, ret);
abandon_free:
kfree(wreq);
abandon:
if (using_pgpriv2)
fscache_clear_page_bits(mapping, start, len, cond);
if (term_func)
- term_func(term_func_priv, ret, false);
+ term_func(term_func_priv, ret);
}
EXPORT_SYMBOL(__fscache_write_to_cache);
diff --git a/fs/netfs/internal.h b/fs/netfs/internal.h
index 1c4f953c3d68..b6500a7cda81 100644
--- a/fs/netfs/internal.h
+++ b/fs/netfs/internal.h
@@ -23,7 +23,7 @@
/*
* buffered_read.c
*/
-void netfs_cache_read_terminated(void *priv, ssize_t transferred_or_error, bool was_async);
+void netfs_cache_read_terminated(void *priv, ssize_t transferred_or_error);
int netfs_prefetch_for_write(struct file *file, struct folio *folio,
size_t offset, size_t len);
@@ -71,9 +71,8 @@ struct netfs_io_request *netfs_alloc_request(struct address_space *mapping,
loff_t start, size_t len,
enum netfs_io_origin origin);
void netfs_get_request(struct netfs_io_request *rreq, enum netfs_rreq_ref_trace what);
-void netfs_clear_subrequests(struct netfs_io_request *rreq, bool was_async);
-void netfs_put_request(struct netfs_io_request *rreq, bool was_async,
- enum netfs_rreq_ref_trace what);
+void netfs_clear_subrequests(struct netfs_io_request *rreq);
+void netfs_put_request(struct netfs_io_request *rreq, enum netfs_rreq_ref_trace what);
struct netfs_io_subrequest *netfs_alloc_subrequest(struct netfs_io_request *rreq);
static inline void netfs_see_request(struct netfs_io_request *rreq,
@@ -94,7 +93,7 @@ static inline void netfs_see_subrequest(struct netfs_io_subrequest *subreq,
*/
void netfs_read_collection_worker(struct work_struct *work);
void netfs_wake_read_collector(struct netfs_io_request *rreq);
-void netfs_cache_read_terminated(void *priv, ssize_t transferred_or_error, bool was_async);
+void netfs_cache_read_terminated(void *priv, ssize_t transferred_or_error);
ssize_t netfs_wait_for_read(struct netfs_io_request *rreq);
void netfs_wait_for_pause(struct netfs_io_request *rreq);
@@ -177,7 +176,7 @@ static inline void netfs_stat_d(atomic_t *stat)
*/
int netfs_folio_written_back(struct folio *folio);
void netfs_write_collection_worker(struct work_struct *work);
-void netfs_wake_write_collector(struct netfs_io_request *wreq, bool was_async);
+void netfs_wake_write_collector(struct netfs_io_request *wreq);
/*
* write_issue.c
diff --git a/fs/netfs/objects.c b/fs/netfs/objects.c
index dc6b41ef18b0..d3eb9ba3013a 100644
--- a/fs/netfs/objects.c
+++ b/fs/netfs/objects.c
@@ -10,6 +10,8 @@
#include <linux/delay.h>
#include "internal.h"
+static void netfs_free_request(struct work_struct *work);
+
/*
* Allocate an I/O request and initialise it.
*/
@@ -34,6 +36,7 @@ struct netfs_io_request *netfs_alloc_request(struct address_space *mapping,
}
memset(rreq, 0, kmem_cache_size(cache));
+ INIT_WORK(&rreq->cleanup_work, netfs_free_request);
rreq->start = start;
rreq->len = len;
rreq->origin = origin;
@@ -49,7 +52,7 @@ struct netfs_io_request *netfs_alloc_request(struct address_space *mapping,
INIT_LIST_HEAD(&rreq->io_streams[0].subrequests);
INIT_LIST_HEAD(&rreq->io_streams[1].subrequests);
init_waitqueue_head(&rreq->waitq);
- refcount_set(&rreq->ref, 1);
+ refcount_set(&rreq->ref, 2);
if (origin == NETFS_READAHEAD ||
origin == NETFS_READPAGE ||
@@ -63,7 +66,9 @@ struct netfs_io_request *netfs_alloc_request(struct address_space *mapping,
INIT_WORK(&rreq->work, netfs_write_collection_worker);
}
+ /* The IN_PROGRESS flag comes with a ref. */
__set_bit(NETFS_RREQ_IN_PROGRESS, &rreq->flags);
+
if (file && file->f_flags & O_NONBLOCK)
__set_bit(NETFS_RREQ_NONBLOCK, &rreq->flags);
if (rreq->netfs_ops->init_request) {
@@ -75,7 +80,7 @@ struct netfs_io_request *netfs_alloc_request(struct address_space *mapping,
}
atomic_inc(&ctx->io_count);
- trace_netfs_rreq_ref(rreq->debug_id, 1, netfs_rreq_trace_new);
+ trace_netfs_rreq_ref(rreq->debug_id, refcount_read(&rreq->ref), netfs_rreq_trace_new);
netfs_proc_add_rreq(rreq);
netfs_stat(&netfs_n_rh_rreq);
return rreq;
@@ -89,7 +94,7 @@ void netfs_get_request(struct netfs_io_request *rreq, enum netfs_rreq_ref_trace
trace_netfs_rreq_ref(rreq->debug_id, r + 1, what);
}
-void netfs_clear_subrequests(struct netfs_io_request *rreq, bool was_async)
+void netfs_clear_subrequests(struct netfs_io_request *rreq)
{
struct netfs_io_subrequest *subreq;
struct netfs_io_stream *stream;
@@ -101,8 +106,7 @@ void netfs_clear_subrequests(struct netfs_io_request *rreq, bool was_async)
subreq = list_first_entry(&stream->subrequests,
struct netfs_io_subrequest, rreq_link);
list_del(&subreq->rreq_link);
- netfs_put_subrequest(subreq, was_async,
- netfs_sreq_trace_put_clear);
+ netfs_put_subrequest(subreq, netfs_sreq_trace_put_clear);
}
}
}
@@ -118,13 +122,19 @@ static void netfs_free_request_rcu(struct rcu_head *rcu)
static void netfs_free_request(struct work_struct *work)
{
struct netfs_io_request *rreq =
- container_of(work, struct netfs_io_request, work);
+ container_of(work, struct netfs_io_request, cleanup_work);
struct netfs_inode *ictx = netfs_inode(rreq->inode);
unsigned int i;
trace_netfs_rreq(rreq, netfs_rreq_trace_free);
+
+ /* Cancel/flush the result collection worker. That does not carry a
+ * ref of its own, so we must wait for it somewhere.
+ */
+ cancel_work_sync(&rreq->work);
+
netfs_proc_del_rreq(rreq);
- netfs_clear_subrequests(rreq, false);
+ netfs_clear_subrequests(rreq);
if (rreq->netfs_ops->free_request)
rreq->netfs_ops->free_request(rreq);
if (rreq->cache_resources.ops)
@@ -145,8 +155,7 @@ static void netfs_free_request(struct work_struct *work)
call_rcu(&rreq->rcu, netfs_free_request_rcu);
}
-void netfs_put_request(struct netfs_io_request *rreq, bool was_async,
- enum netfs_rreq_ref_trace what)
+void netfs_put_request(struct netfs_io_request *rreq, enum netfs_rreq_ref_trace what)
{
unsigned int debug_id;
bool dead;
@@ -156,15 +165,8 @@ void netfs_put_request(struct netfs_io_request *rreq, bool was_async,
debug_id = rreq->debug_id;
dead = __refcount_dec_and_test(&rreq->ref, &r);
trace_netfs_rreq_ref(debug_id, r - 1, what);
- if (dead) {
- if (was_async) {
- rreq->work.func = netfs_free_request;
- if (!queue_work(system_unbound_wq, &rreq->work))
- WARN_ON(1);
- } else {
- netfs_free_request(&rreq->work);
- }
- }
+ if (dead)
+ WARN_ON(!queue_work(system_unbound_wq, &rreq->cleanup_work));
}
}
@@ -206,8 +208,7 @@ void netfs_get_subrequest(struct netfs_io_subrequest *subreq,
what);
}
-static void netfs_free_subrequest(struct netfs_io_subrequest *subreq,
- bool was_async)
+static void netfs_free_subrequest(struct netfs_io_subrequest *subreq)
{
struct netfs_io_request *rreq = subreq->rreq;
@@ -216,10 +217,10 @@ static void netfs_free_subrequest(struct netfs_io_subrequest *subreq,
rreq->netfs_ops->free_subrequest(subreq);
mempool_free(subreq, rreq->netfs_ops->subrequest_pool ?: &netfs_subrequest_pool);
netfs_stat_d(&netfs_n_rh_sreq);
- netfs_put_request(rreq, was_async, netfs_rreq_trace_put_subreq);
+ netfs_put_request(rreq, netfs_rreq_trace_put_subreq);
}
-void netfs_put_subrequest(struct netfs_io_subrequest *subreq, bool was_async,
+void netfs_put_subrequest(struct netfs_io_subrequest *subreq,
enum netfs_sreq_ref_trace what)
{
unsigned int debug_index = subreq->debug_index;
@@ -230,5 +231,5 @@ void netfs_put_subrequest(struct netfs_io_subrequest *subreq, bool was_async,
dead = __refcount_dec_and_test(&subreq->ref, &r);
trace_netfs_sreq_ref(debug_id, debug_index, r - 1, what);
if (dead)
- netfs_free_subrequest(subreq, was_async);
+ netfs_free_subrequest(subreq);
}
diff --git a/fs/netfs/read_collect.c b/fs/netfs/read_collect.c
index d3cf27b2697c..1197ebce5675 100644
--- a/fs/netfs/read_collect.c
+++ b/fs/netfs/read_collect.c
@@ -301,7 +301,7 @@ static void netfs_collect_read_results(struct netfs_io_request *rreq)
struct netfs_io_subrequest, rreq_link);
stream->front = front;
spin_unlock(&rreq->lock);
- netfs_put_subrequest(remove, false,
+ netfs_put_subrequest(remove,
notes & ABANDON_SREQ ?
netfs_sreq_trace_put_abandon :
netfs_sreq_trace_put_done);
@@ -399,7 +399,7 @@ static void netfs_rreq_assess_single(struct netfs_io_request *rreq)
* Note that we're in normal kernel thread context at this point, possibly
* running on a workqueue.
*/
-static void netfs_read_collection(struct netfs_io_request *rreq)
+static bool netfs_read_collection(struct netfs_io_request *rreq)
{
struct netfs_io_stream *stream = &rreq->io_streams[0];
@@ -409,11 +409,11 @@ static void netfs_read_collection(struct netfs_io_request *rreq)
* queue is empty.
*/
if (!test_bit(NETFS_RREQ_ALL_QUEUED, &rreq->flags))
- return;
+ return false;
smp_rmb(); /* Read ALL_QUEUED before subreq lists. */
if (!list_empty(&stream->subrequests))
- return;
+ return false;
/* Okay, declare that all I/O is complete. */
rreq->transferred = stream->transferred;
@@ -436,12 +436,14 @@ static void netfs_read_collection(struct netfs_io_request *rreq)
trace_netfs_rreq(rreq, netfs_rreq_trace_wake_ip);
clear_and_wake_up_bit(NETFS_RREQ_IN_PROGRESS, &rreq->flags);
+ /* As we cleared NETFS_RREQ_IN_PROGRESS, we acquired its ref. */
trace_netfs_rreq(rreq, netfs_rreq_trace_done);
- netfs_clear_subrequests(rreq, false);
+ netfs_clear_subrequests(rreq);
netfs_unlock_abandoned_read_pages(rreq);
if (unlikely(rreq->copy_to_cache))
netfs_pgpriv2_end_copy_to_cache(rreq);
+ return true;
}
void netfs_read_collection_worker(struct work_struct *work)
@@ -449,9 +451,13 @@ void netfs_read_collection_worker(struct work_struct *work)
struct netfs_io_request *rreq = container_of(work, struct netfs_io_request, work);
netfs_see_request(rreq, netfs_rreq_trace_see_work);
- if (test_bit(NETFS_RREQ_IN_PROGRESS, &rreq->flags))
- netfs_read_collection(rreq);
- netfs_put_request(rreq, false, netfs_rreq_trace_put_work);
+ if (test_bit(NETFS_RREQ_IN_PROGRESS, &rreq->flags)) {
+ if (netfs_read_collection(rreq))
+ /* Drop the ref from the IN_PROGRESS flag. */
+ netfs_put_request(rreq, netfs_rreq_trace_put_work_ip);
+ else
+ netfs_see_request(rreq, netfs_rreq_trace_see_work_complete);
+ }
}
/*
@@ -461,11 +467,7 @@ void netfs_wake_read_collector(struct netfs_io_request *rreq)
{
if (test_bit(NETFS_RREQ_OFFLOAD_COLLECTION, &rreq->flags) &&
!test_bit(NETFS_RREQ_RETRYING, &rreq->flags)) {
- if (!work_pending(&rreq->work)) {
- netfs_get_request(rreq, netfs_rreq_trace_get_work);
- if (!queue_work(system_unbound_wq, &rreq->work))
- netfs_put_request(rreq, true, netfs_rreq_trace_put_work_nq);
- }
+ queue_work(system_unbound_wq, &rreq->work);
} else {
trace_netfs_rreq(rreq, netfs_rreq_trace_wake_queue);
wake_up(&rreq->waitq);
@@ -580,14 +582,14 @@ void netfs_read_subreq_terminated(struct netfs_io_subrequest *subreq)
test_bit(NETFS_RREQ_RETRYING, &rreq->flags))
netfs_wake_read_collector(rreq);
- netfs_put_subrequest(subreq, true, netfs_sreq_trace_put_terminated);
+ netfs_put_subrequest(subreq, netfs_sreq_trace_put_terminated);
}
EXPORT_SYMBOL(netfs_read_subreq_terminated);
/*
* Handle termination of a read from the cache.
*/
-void netfs_cache_read_terminated(void *priv, ssize_t transferred_or_error, bool was_async)
+void netfs_cache_read_terminated(void *priv, ssize_t transferred_or_error)
{
struct netfs_io_subrequest *subreq = priv;
@@ -623,7 +625,11 @@ ssize_t netfs_wait_for_read(struct netfs_io_request *rreq)
(!test_bit(NETFS_SREQ_IN_PROGRESS, &subreq->flags) ||
test_bit(NETFS_SREQ_MADE_PROGRESS, &subreq->flags))) {
__set_current_state(TASK_RUNNING);
- netfs_read_collection(rreq);
+ if (netfs_read_collection(rreq)) {
+ /* Drop the ref from the NETFS_RREQ_IN_PROGRESS flag. */
+ netfs_put_request(rreq, netfs_rreq_trace_put_work_ip);
+ break;
+ }
continue;
}
@@ -678,7 +684,11 @@ void netfs_wait_for_pause(struct netfs_io_request *rreq)
(!test_bit(NETFS_SREQ_IN_PROGRESS, &subreq->flags) ||
test_bit(NETFS_SREQ_MADE_PROGRESS, &subreq->flags))) {
__set_current_state(TASK_RUNNING);
- netfs_read_collection(rreq);
+ if (netfs_read_collection(rreq)) {
+ /* Drop the ref from the NETFS_RREQ_IN_PROGRESS flag. */
+ netfs_put_request(rreq, netfs_rreq_trace_put_work_ip);
+ break;
+ }
continue;
}
}
diff --git a/fs/netfs/read_pgpriv2.c b/fs/netfs/read_pgpriv2.c
index cf7727060215..5bbe906a551d 100644
--- a/fs/netfs/read_pgpriv2.c
+++ b/fs/netfs/read_pgpriv2.c
@@ -116,7 +116,7 @@ static struct netfs_io_request *netfs_pgpriv2_begin_copy_to_cache(
return creq;
cancel_put:
- netfs_put_request(creq, false, netfs_rreq_trace_put_return);
+ netfs_put_request(creq, netfs_rreq_trace_put_return);
cancel:
rreq->copy_to_cache = ERR_PTR(-ENOBUFS);
clear_bit(NETFS_RREQ_FOLIO_COPY_TO_CACHE, &rreq->flags);
@@ -155,7 +155,7 @@ void netfs_pgpriv2_end_copy_to_cache(struct netfs_io_request *rreq)
smp_wmb(); /* Write lists before ALL_QUEUED. */
set_bit(NETFS_RREQ_ALL_QUEUED, &creq->flags);
- netfs_put_request(creq, false, netfs_rreq_trace_put_return);
+ netfs_put_request(creq, netfs_rreq_trace_put_return);
creq->copy_to_cache = NULL;
}
diff --git a/fs/netfs/read_retry.c b/fs/netfs/read_retry.c
index 0f294b26e08c..1378dc7fa2cc 100644
--- a/fs/netfs/read_retry.c
+++ b/fs/netfs/read_retry.c
@@ -173,7 +173,7 @@ static void netfs_retry_read_subrequests(struct netfs_io_request *rreq)
&stream->subrequests, rreq_link) {
trace_netfs_sreq(subreq, netfs_sreq_trace_superfluous);
list_del(&subreq->rreq_link);
- netfs_put_subrequest(subreq, false, netfs_sreq_trace_put_done);
+ netfs_put_subrequest(subreq, netfs_sreq_trace_put_done);
if (subreq == to)
break;
}
diff --git a/fs/netfs/read_single.c b/fs/netfs/read_single.c
index fea0ecdecc53..fa622a6cd56d 100644
--- a/fs/netfs/read_single.c
+++ b/fs/netfs/read_single.c
@@ -142,7 +142,7 @@ static int netfs_single_dispatch_read(struct netfs_io_request *rreq)
set_bit(NETFS_RREQ_ALL_QUEUED, &rreq->flags);
return ret;
cancel:
- netfs_put_subrequest(subreq, false, netfs_sreq_trace_put_cancel);
+ netfs_put_subrequest(subreq, netfs_sreq_trace_put_cancel);
return ret;
}
@@ -185,11 +185,11 @@ ssize_t netfs_read_single(struct inode *inode, struct file *file, struct iov_ite
netfs_single_dispatch_read(rreq);
ret = netfs_wait_for_read(rreq);
- netfs_put_request(rreq, true, netfs_rreq_trace_put_return);
+ netfs_put_request(rreq, netfs_rreq_trace_put_return);
return ret;
cleanup_free:
- netfs_put_request(rreq, false, netfs_rreq_trace_put_failed);
+ netfs_put_request(rreq, netfs_rreq_trace_put_failed);
return ret;
}
EXPORT_SYMBOL(netfs_read_single);
diff --git a/fs/netfs/write_collect.c b/fs/netfs/write_collect.c
index 3fca59e6475d..7241d1fd2c14 100644
--- a/fs/netfs/write_collect.c
+++ b/fs/netfs/write_collect.c
@@ -280,7 +280,7 @@ static void netfs_collect_write_results(struct netfs_io_request *wreq)
struct netfs_io_subrequest, rreq_link);
stream->front = front;
spin_unlock(&wreq->lock);
- netfs_put_subrequest(remove, false,
+ netfs_put_subrequest(remove,
notes & SAW_FAILURE ?
netfs_sreq_trace_put_cancel :
netfs_sreq_trace_put_done);
@@ -356,30 +356,21 @@ static void netfs_collect_write_results(struct netfs_io_request *wreq)
/*
* Perform the collection of subrequests, folios and encryption buffers.
*/
-void netfs_write_collection_worker(struct work_struct *work)
+static bool netfs_write_collection(struct netfs_io_request *wreq)
{
- struct netfs_io_request *wreq = container_of(work, struct netfs_io_request, work);
struct netfs_inode *ictx = netfs_inode(wreq->inode);
size_t transferred;
int s;
_enter("R=%x", wreq->debug_id);
- netfs_see_request(wreq, netfs_rreq_trace_see_work);
- if (!test_bit(NETFS_RREQ_IN_PROGRESS, &wreq->flags)) {
- netfs_put_request(wreq, false, netfs_rreq_trace_put_work);
- return;
- }
-
netfs_collect_write_results(wreq);
/* We're done when the app thread has finished posting subreqs and all
* the queues in all the streams are empty.
*/
- if (!test_bit(NETFS_RREQ_ALL_QUEUED, &wreq->flags)) {
- netfs_put_request(wreq, false, netfs_rreq_trace_put_work);
- return;
- }
+ if (!test_bit(NETFS_RREQ_ALL_QUEUED, &wreq->flags))
+ return false;
smp_rmb(); /* Read ALL_QUEUED before lists. */
transferred = LONG_MAX;
@@ -387,10 +378,8 @@ void netfs_write_collection_worker(struct work_struct *work)
struct netfs_io_stream *stream = &wreq->io_streams[s];
if (!stream->active)
continue;
- if (!list_empty(&stream->subrequests)) {
- netfs_put_request(wreq, false, netfs_rreq_trace_put_work);
- return;
- }
+ if (!list_empty(&stream->subrequests))
+ return false;
if (stream->transferred < transferred)
transferred = stream->transferred;
}
@@ -430,6 +419,7 @@ void netfs_write_collection_worker(struct work_struct *work)
_debug("finished");
trace_netfs_rreq(wreq, netfs_rreq_trace_wake_ip);
clear_and_wake_up_bit(NETFS_RREQ_IN_PROGRESS, &wreq->flags);
+ /* As we cleared NETFS_RREQ_IN_PROGRESS, we acquired its ref. */
if (wreq->iocb) {
size_t written = min(wreq->transferred, wreq->len);
@@ -440,27 +430,36 @@ void netfs_write_collection_worker(struct work_struct *work)
wreq->iocb = VFS_PTR_POISON;
}
- netfs_clear_subrequests(wreq, false);
- netfs_put_request(wreq, false, netfs_rreq_trace_put_work_complete);
+ netfs_clear_subrequests(wreq);
+ return true;
+}
+
+void netfs_write_collection_worker(struct work_struct *work)
+{
+ struct netfs_io_request *rreq = container_of(work, struct netfs_io_request, work);
+
+ netfs_see_request(rreq, netfs_rreq_trace_see_work);
+ if (test_bit(NETFS_RREQ_IN_PROGRESS, &rreq->flags)) {
+ if (netfs_write_collection(rreq))
+ /* Drop the ref from the IN_PROGRESS flag. */
+ netfs_put_request(rreq, netfs_rreq_trace_put_work_ip);
+ else
+ netfs_see_request(rreq, netfs_rreq_trace_see_work_complete);
+ }
}
/*
* Wake the collection work item.
*/
-void netfs_wake_write_collector(struct netfs_io_request *wreq, bool was_async)
+void netfs_wake_write_collector(struct netfs_io_request *wreq)
{
- if (!work_pending(&wreq->work)) {
- netfs_get_request(wreq, netfs_rreq_trace_get_work);
- if (!queue_work(system_unbound_wq, &wreq->work))
- netfs_put_request(wreq, was_async, netfs_rreq_trace_put_work_nq);
- }
+ queue_work(system_unbound_wq, &wreq->work);
}
/**
* netfs_write_subrequest_terminated - Note the termination of a write operation.
* @_op: The I/O request that has terminated.
* @transferred_or_error: The amount of data transferred or an error code.
- * @was_async: The termination was asynchronous
*
* This tells the library that a contributory write I/O operation has
* terminated, one way or another, and that it should collect the results.
@@ -470,17 +469,13 @@ void netfs_wake_write_collector(struct netfs_io_request *wreq, bool was_async)
* negative error code. The library will look after reissuing I/O operations
* as appropriate and writing downloaded data to the cache.
*
- * If @was_async is true, the caller might be running in softirq or interrupt
- * context and we can't sleep.
- *
* When this is called, ownership of the subrequest is transferred back to the
* library, along with a ref.
*
* Note that %_op is a void* so that the function can be passed to
* kiocb::term_func without the need for a casting wrapper.
*/
-void netfs_write_subrequest_terminated(void *_op, ssize_t transferred_or_error,
- bool was_async)
+void netfs_write_subrequest_terminated(void *_op, ssize_t transferred_or_error)
{
struct netfs_io_subrequest *subreq = _op;
struct netfs_io_request *wreq = subreq->rreq;
@@ -543,8 +538,8 @@ void netfs_write_subrequest_terminated(void *_op, ssize_t transferred_or_error,
* transferring a ref to it if we were the ones to do so.
*/
if (list_is_first(&subreq->rreq_link, &stream->subrequests))
- netfs_wake_write_collector(wreq, was_async);
+ netfs_wake_write_collector(wreq);
- netfs_put_subrequest(subreq, was_async, netfs_sreq_trace_put_terminated);
+ netfs_put_subrequest(subreq, netfs_sreq_trace_put_terminated);
}
EXPORT_SYMBOL(netfs_write_subrequest_terminated);
diff --git a/fs/netfs/write_issue.c b/fs/netfs/write_issue.c
index 77279fc5b5a7..8744ed3faf29 100644
--- a/fs/netfs/write_issue.c
+++ b/fs/netfs/write_issue.c
@@ -134,7 +134,7 @@ struct netfs_io_request *netfs_create_write_req(struct address_space *mapping,
return wreq;
nomem:
wreq->error = -ENOMEM;
- netfs_put_request(wreq, false, netfs_rreq_trace_put_failed);
+ netfs_put_request(wreq, netfs_rreq_trace_put_failed);
return ERR_PTR(-ENOMEM);
}
@@ -233,7 +233,7 @@ static void netfs_do_issue_write(struct netfs_io_stream *stream,
_enter("R=%x[%x],%zx", wreq->debug_id, subreq->debug_index, subreq->len);
if (test_bit(NETFS_SREQ_FAILED, &subreq->flags))
- return netfs_write_subrequest_terminated(subreq, subreq->error, false);
+ return netfs_write_subrequest_terminated(subreq, subreq->error);
trace_netfs_sreq(subreq, netfs_sreq_trace_submit);
stream->issue_write(subreq);
@@ -542,7 +542,7 @@ static void netfs_end_issue_write(struct netfs_io_request *wreq)
}
if (needs_poke)
- netfs_wake_write_collector(wreq, false);
+ netfs_wake_write_collector(wreq);
}
/*
@@ -599,8 +599,9 @@ int netfs_writepages(struct address_space *mapping,
netfs_end_issue_write(wreq);
mutex_unlock(&ictx->wb_lock);
+ netfs_wake_write_collector(wreq);
- netfs_put_request(wreq, false, netfs_rreq_trace_put_return);
+ netfs_put_request(wreq, netfs_rreq_trace_put_return);
_leave(" = %d", error);
return error;
@@ -694,7 +695,7 @@ int netfs_end_writethrough(struct netfs_io_request *wreq, struct writeback_contr
wait_on_bit(&wreq->flags, NETFS_RREQ_IN_PROGRESS, TASK_UNINTERRUPTIBLE);
ret = wreq->error;
}
- netfs_put_request(wreq, false, netfs_rreq_trace_put_return);
+ netfs_put_request(wreq, netfs_rreq_trace_put_return);
return ret;
}
@@ -885,7 +886,7 @@ int netfs_writeback_single(struct address_space *mapping,
goto couldnt_start;
}
- trace_netfs_write(wreq, netfs_write_trace_writeback);
+ trace_netfs_write(wreq, netfs_write_trace_writeback_single);
netfs_stat(&netfs_n_wh_writepages);
if (__test_and_set_bit(NETFS_RREQ_UPLOAD_TO_SERVER, &wreq->flags))
@@ -914,8 +915,9 @@ int netfs_writeback_single(struct address_space *mapping,
set_bit(NETFS_RREQ_ALL_QUEUED, &wreq->flags);
mutex_unlock(&ictx->wb_lock);
+ netfs_wake_write_collector(wreq);
- netfs_put_request(wreq, false, netfs_rreq_trace_put_return);
+ netfs_put_request(wreq, netfs_rreq_trace_put_return);
_leave(" = %d", ret);
return ret;
diff --git a/fs/netfs/write_retry.c b/fs/netfs/write_retry.c
index 9b1ca8b0f4dd..7408f6bb8e42 100644
--- a/fs/netfs/write_retry.c
+++ b/fs/netfs/write_retry.c
@@ -132,7 +132,7 @@ static void netfs_retry_write_stream(struct netfs_io_request *wreq,
&stream->subrequests, rreq_link) {
trace_netfs_sreq(subreq, netfs_sreq_trace_discard);
list_del(&subreq->rreq_link);
- netfs_put_subrequest(subreq, false, netfs_sreq_trace_put_done);
+ netfs_put_subrequest(subreq, netfs_sreq_trace_put_done);
if (subreq == to)
break;
}
diff --git a/fs/smb/client/cifsproto.h b/fs/smb/client/cifsproto.h
index ecf774a8f1ca..66093fa78aed 100644
--- a/fs/smb/client/cifsproto.h
+++ b/fs/smb/client/cifsproto.h
@@ -151,8 +151,7 @@ extern bool is_size_safe_to_change(struct cifsInodeInfo *cifsInode, __u64 eof,
bool from_readdir);
extern void cifs_update_eof(struct cifsInodeInfo *cifsi, loff_t offset,
unsigned int bytes_written);
-void cifs_write_subrequest_terminated(struct cifs_io_subrequest *wdata, ssize_t result,
- bool was_async);
+void cifs_write_subrequest_terminated(struct cifs_io_subrequest *wdata, ssize_t result);
extern struct cifsFileInfo *find_writable_file(struct cifsInodeInfo *, int);
extern int cifs_get_writable_file(struct cifsInodeInfo *cifs_inode,
int flags,
diff --git a/fs/smb/client/cifssmb.c b/fs/smb/client/cifssmb.c
index f55457b4b82e..477792c07d45 100644
--- a/fs/smb/client/cifssmb.c
+++ b/fs/smb/client/cifssmb.c
@@ -1725,7 +1725,7 @@ cifs_writev_callback(struct mid_q_entry *mid)
server->credits, server->in_flight,
0, cifs_trace_rw_credits_write_response_clear);
wdata->credits.value = 0;
- cifs_write_subrequest_terminated(wdata, result, true);
+ cifs_write_subrequest_terminated(wdata, result);
release_mid(mid);
trace_smb3_rw_credits(credits.rreq_debug_id, credits.rreq_debug_index, 0,
server->credits, server->in_flight,
@@ -1813,7 +1813,7 @@ cifs_async_writev(struct cifs_io_subrequest *wdata)
out:
if (rc) {
add_credits_and_wake_if(wdata->server, &wdata->credits, 0);
- cifs_write_subrequest_terminated(wdata, rc, false);
+ cifs_write_subrequest_terminated(wdata, rc);
}
}
diff --git a/fs/smb/client/file.c b/fs/smb/client/file.c
index 950aa4f912f5..3000c8a9d3ea 100644
--- a/fs/smb/client/file.c
+++ b/fs/smb/client/file.c
@@ -130,7 +130,7 @@ static void cifs_issue_write(struct netfs_io_subrequest *subreq)
else
trace_netfs_sreq(subreq, netfs_sreq_trace_fail);
add_credits_and_wake_if(wdata->server, &wdata->credits, 0);
- cifs_write_subrequest_terminated(wdata, rc, false);
+ cifs_write_subrequest_terminated(wdata, rc);
goto out;
}
@@ -2423,8 +2423,7 @@ int cifs_lock(struct file *file, int cmd, struct file_lock *flock)
return rc;
}
-void cifs_write_subrequest_terminated(struct cifs_io_subrequest *wdata, ssize_t result,
- bool was_async)
+void cifs_write_subrequest_terminated(struct cifs_io_subrequest *wdata, ssize_t result)
{
struct netfs_io_request *wreq = wdata->rreq;
struct netfs_inode *ictx = netfs_inode(wreq->inode);
@@ -2441,7 +2440,7 @@ void cifs_write_subrequest_terminated(struct cifs_io_subrequest *wdata, ssize_t
netfs_resize_file(ictx, wrend, true);
}
- netfs_write_subrequest_terminated(&wdata->subreq, result, was_async);
+ netfs_write_subrequest_terminated(&wdata->subreq, result);
}
struct cifsFileInfo *find_readable_file(struct cifsInodeInfo *cifs_inode,
diff --git a/fs/smb/client/smb2pdu.c b/fs/smb/client/smb2pdu.c
index 4e28632b5fd6..399185ca7cac 100644
--- a/fs/smb/client/smb2pdu.c
+++ b/fs/smb/client/smb2pdu.c
@@ -4888,7 +4888,7 @@ smb2_writev_callback(struct mid_q_entry *mid)
0, cifs_trace_rw_credits_write_response_clear);
wdata->credits.value = 0;
trace_netfs_sreq(&wdata->subreq, netfs_sreq_trace_io_progress);
- cifs_write_subrequest_terminated(wdata, result ?: written, true);
+ cifs_write_subrequest_terminated(wdata, result ?: written);
release_mid(mid);
trace_smb3_rw_credits(rreq_debug_id, subreq_debug_index, 0,
server->credits, server->in_flight,
@@ -5061,7 +5061,7 @@ smb2_async_writev(struct cifs_io_subrequest *wdata)
-(int)wdata->credits.value,
cifs_trace_rw_credits_write_response_clear);
add_credits_and_wake_if(wdata->server, &wdata->credits, 0);
- cifs_write_subrequest_terminated(wdata, rc, true);
+ cifs_write_subrequest_terminated(wdata, rc);
}
}
diff --git a/include/linux/fscache.h b/include/linux/fscache.h
index 9de27643607f..266e6c9e6f83 100644
--- a/include/linux/fscache.h
+++ b/include/linux/fscache.h
@@ -628,7 +628,7 @@ static inline void fscache_write_to_cache(struct fscache_cookie *cookie,
term_func, term_func_priv,
using_pgpriv2, caching);
else if (term_func)
- term_func(term_func_priv, -ENOBUFS, false);
+ term_func(term_func_priv, -ENOBUFS);
}
diff --git a/include/linux/netfs.h b/include/linux/netfs.h
index 497c4f4698f6..c3f230732f51 100644
--- a/include/linux/netfs.h
+++ b/include/linux/netfs.h
@@ -51,8 +51,7 @@ enum netfs_io_source {
NETFS_INVALID_WRITE,
} __mode(byte);
-typedef void (*netfs_io_terminated_t)(void *priv, ssize_t transferred_or_error,
- bool was_async);
+typedef void (*netfs_io_terminated_t)(void *priv, ssize_t transferred_or_error);
/*
* Per-inode context. This wraps the VFS inode.
@@ -223,9 +222,10 @@ enum netfs_io_origin {
*/
struct netfs_io_request {
union {
- struct work_struct work;
+ struct work_struct cleanup_work; /* Deferred cleanup work */
struct rcu_head rcu;
};
+ struct work_struct work; /* Result collector work */
struct inode *inode; /* The file being accessed */
struct address_space *mapping; /* The mapping being accessed */
struct kiocb *iocb; /* AIO completion vector */
@@ -270,7 +270,7 @@ struct netfs_io_request {
#define NETFS_RREQ_NO_UNLOCK_FOLIO 2 /* Don't unlock no_unlock_folio on completion */
#define NETFS_RREQ_DONT_UNLOCK_FOLIOS 3 /* Don't unlock the folios on completion */
#define NETFS_RREQ_FAILED 4 /* The request failed */
-#define NETFS_RREQ_IN_PROGRESS 5 /* Unlocked when the request completes */
+#define NETFS_RREQ_IN_PROGRESS 5 /* Unlocked when the request completes (has ref) */
#define NETFS_RREQ_FOLIO_COPY_TO_CACHE 6 /* Copy current folio to cache from read */
#define NETFS_RREQ_UPLOAD_TO_SERVER 8 /* Need to write to the server */
#define NETFS_RREQ_NONBLOCK 9 /* Don't block if possible (O_NONBLOCK) */
@@ -440,15 +440,14 @@ void netfs_read_subreq_terminated(struct netfs_io_subrequest *subreq);
void netfs_get_subrequest(struct netfs_io_subrequest *subreq,
enum netfs_sreq_ref_trace what);
void netfs_put_subrequest(struct netfs_io_subrequest *subreq,
- bool was_async, enum netfs_sreq_ref_trace what);
+ enum netfs_sreq_ref_trace what);
ssize_t netfs_extract_user_iter(struct iov_iter *orig, size_t orig_len,
struct iov_iter *new,
iov_iter_extraction_t extraction_flags);
size_t netfs_limit_iter(const struct iov_iter *iter, size_t start_offset,
size_t max_size, size_t max_segs);
void netfs_prepare_write_failed(struct netfs_io_subrequest *subreq);
-void netfs_write_subrequest_terminated(void *_op, ssize_t transferred_or_error,
- bool was_async);
+void netfs_write_subrequest_terminated(void *_op, ssize_t transferred_or_error);
void netfs_queue_write_request(struct netfs_io_subrequest *subreq);
int netfs_start_io_read(struct inode *inode);
diff --git a/include/trace/events/netfs.h b/include/trace/events/netfs.h
index f880835f7695..402c5e82e7b8 100644
--- a/include/trace/events/netfs.h
+++ b/include/trace/events/netfs.h
@@ -30,6 +30,7 @@
EM(netfs_write_trace_dio_write, "DIO-WRITE") \
EM(netfs_write_trace_unbuffered_write, "UNB-WRITE") \
EM(netfs_write_trace_writeback, "WRITEBACK") \
+ EM(netfs_write_trace_writeback_single, "WB-SINGLE") \
E_(netfs_write_trace_writethrough, "WRITETHRU")
#define netfs_rreq_origins \
@@ -128,17 +129,15 @@
#define netfs_rreq_ref_traces \
EM(netfs_rreq_trace_get_for_outstanding,"GET OUTSTND") \
EM(netfs_rreq_trace_get_subreq, "GET SUBREQ ") \
- EM(netfs_rreq_trace_get_work, "GET WORK ") \
EM(netfs_rreq_trace_put_complete, "PUT COMPLT ") \
EM(netfs_rreq_trace_put_discard, "PUT DISCARD") \
EM(netfs_rreq_trace_put_failed, "PUT FAILED ") \
EM(netfs_rreq_trace_put_no_submit, "PUT NO-SUBM") \
EM(netfs_rreq_trace_put_return, "PUT RETURN ") \
EM(netfs_rreq_trace_put_subreq, "PUT SUBREQ ") \
- EM(netfs_rreq_trace_put_work, "PUT WORK ") \
- EM(netfs_rreq_trace_put_work_complete, "PUT WORK CP") \
- EM(netfs_rreq_trace_put_work_nq, "PUT WORK NQ") \
+ EM(netfs_rreq_trace_put_work_ip, "PUT WORK IP ") \
EM(netfs_rreq_trace_see_work, "SEE WORK ") \
+ EM(netfs_rreq_trace_see_work_complete, "SEE WORK CP") \
E_(netfs_rreq_trace_new, "NEW ")
#define netfs_sreq_ref_traces \
diff --git a/net/9p/client.c b/net/9p/client.c
index 61461b9fa134..5c1ca57ccd28 100644
--- a/net/9p/client.c
+++ b/net/9p/client.c
@@ -1704,7 +1704,7 @@ p9_client_write_subreq(struct netfs_io_subrequest *subreq)
start, len, &subreq->io_iter);
}
if (IS_ERR(req)) {
- netfs_write_subrequest_terminated(subreq, PTR_ERR(req), false);
+ netfs_write_subrequest_terminated(subreq, PTR_ERR(req));
return;
}
@@ -1712,7 +1712,7 @@ p9_client_write_subreq(struct netfs_io_subrequest *subreq)
if (err) {
trace_9p_protocol_dump(clnt, &req->rc);
p9_req_put(clnt, req);
- netfs_write_subrequest_terminated(subreq, err, false);
+ netfs_write_subrequest_terminated(subreq, err);
return;
}
@@ -1724,7 +1724,7 @@ p9_client_write_subreq(struct netfs_io_subrequest *subreq)
p9_debug(P9_DEBUG_9P, "<<< RWRITE count %d\n", len);
p9_req_put(clnt, req);
- netfs_write_subrequest_terminated(subreq, written, false);
+ netfs_write_subrequest_terminated(subreq, written);
}
EXPORT_SYMBOL(p9_client_write_subreq);
^ permalink raw reply related [flat|nested] 6+ messages in thread
* [PATCH 4/4] netfs: Fix wait/wake to be consistent about the waitqueue used
2025-05-19 9:07 [PATCH 0/4] netfs: Miscellaneous fixes David Howells
` (2 preceding siblings ...)
2025-05-19 9:07 ` [PATCH 3/4] netfs: Fix the request's work item to not require a ref David Howells
@ 2025-05-19 9:07 ` David Howells
2025-05-20 11:34 ` [PATCH 0/4] netfs: Miscellaneous fixes Christian Brauner
4 siblings, 0 replies; 6+ messages in thread
From: David Howells @ 2025-05-19 9:07 UTC (permalink / raw)
To: Christian Brauner
Cc: David Howells, Paulo Alcantara, netfs, linux-afs, linux-cifs,
linux-nfs, ceph-devel, v9fs, linux-fsdevel, linux-kernel,
Marc Dionne, Steve French, Ihor Solodrai, Eric Van Hensbergen,
Latchesar Ionkov, Dominique Martinet, Christian Schoenebeck,
Jeff Layton
Fix further inconsistencies in the use of waitqueues
(clear_and_wake_up_bit() vs private waitqueue).
Move some of this stuff from the read and write sides into common code so
that it can be done in fewer places.
To make this work, async I/O needs to set NETFS_RREQ_OFFLOAD_COLLECTION to
indicate that a workqueue will do the collecting and places that call the
wait function need to deal with it returning the amount transferred.
Fixes: e2d46f2ec332 ("netfs: Change the read result collector to only use one work item")
Signed-off-by: David Howells <dhowells@redhat.com>
cc: Marc Dionne <marc.dionne@auristor.com>
cc: Steve French <stfrench@microsoft.com>
cc: Ihor Solodrai <ihor.solodrai@pm.me>
cc: Eric Van Hensbergen <ericvh@kernel.org>
cc: Latchesar Ionkov <lucho@ionkov.net>
cc: Dominique Martinet <asmadeus@codewreck.org>
cc: Christian Schoenebeck <linux_oss@crudebyte.com>
cc: Paulo Alcantara <pc@manguebit.com>
cc: Jeff Layton <jlayton@kernel.org>
cc: v9fs@lists.linux.dev
cc: linux-cifs@vger.kernel.org
cc: netfs@lists.linux.dev
cc: linux-fsdevel@vger.kernel.org
---
fs/netfs/buffered_read.c | 2 +-
fs/netfs/buffered_write.c | 2 +-
fs/netfs/direct_read.c | 4 +-
fs/netfs/direct_write.c | 10 +-
fs/netfs/internal.h | 33 ++++--
fs/netfs/misc.c | 218 ++++++++++++++++++++++++++++++++++++++
fs/netfs/read_collect.c | 139 +-----------------------
fs/netfs/read_retry.c | 24 +----
fs/netfs/write_collect.c | 36 ++-----
fs/netfs/write_issue.c | 28 +++--
fs/netfs/write_retry.c | 12 +--
11 files changed, 284 insertions(+), 224 deletions(-)
diff --git a/fs/netfs/buffered_read.c b/fs/netfs/buffered_read.c
index cb6202efc466..fd4619275801 100644
--- a/fs/netfs/buffered_read.c
+++ b/fs/netfs/buffered_read.c
@@ -312,7 +312,7 @@ static void netfs_read_to_pagecache(struct netfs_io_request *rreq)
if (unlikely(size > 0)) {
smp_wmb(); /* Write lists before ALL_QUEUED. */
set_bit(NETFS_RREQ_ALL_QUEUED, &rreq->flags);
- netfs_wake_read_collector(rreq);
+ netfs_wake_collector(rreq);
}
/* Defer error return as we may need to wait for outstanding I/O. */
diff --git a/fs/netfs/buffered_write.c b/fs/netfs/buffered_write.c
index b4826360a411..dbb544e183d1 100644
--- a/fs/netfs/buffered_write.c
+++ b/fs/netfs/buffered_write.c
@@ -386,7 +386,7 @@ ssize_t netfs_perform_write(struct kiocb *iocb, struct iov_iter *iter,
wbc_detach_inode(&wbc);
if (ret2 == -EIOCBQUEUED)
return ret2;
- if (ret == 0)
+ if (ret == 0 && ret2 < 0)
ret = ret2;
}
diff --git a/fs/netfs/direct_read.c b/fs/netfs/direct_read.c
index cb3c6dc0b165..a24e63d2c818 100644
--- a/fs/netfs/direct_read.c
+++ b/fs/netfs/direct_read.c
@@ -103,7 +103,7 @@ static int netfs_dispatch_unbuffered_reads(struct netfs_io_request *rreq)
rreq->netfs_ops->issue_read(subreq);
if (test_bit(NETFS_RREQ_PAUSE, &rreq->flags))
- netfs_wait_for_pause(rreq);
+ netfs_wait_for_paused_read(rreq);
if (test_bit(NETFS_RREQ_FAILED, &rreq->flags))
break;
if (test_bit(NETFS_RREQ_BLOCKED, &rreq->flags) &&
@@ -115,7 +115,7 @@ static int netfs_dispatch_unbuffered_reads(struct netfs_io_request *rreq)
if (unlikely(size > 0)) {
smp_wmb(); /* Write lists before ALL_QUEUED. */
set_bit(NETFS_RREQ_ALL_QUEUED, &rreq->flags);
- netfs_wake_read_collector(rreq);
+ netfs_wake_collector(rreq);
}
return ret;
diff --git a/fs/netfs/direct_write.c b/fs/netfs/direct_write.c
index c98f1676f86d..fa9a5bf3c6d5 100644
--- a/fs/netfs/direct_write.c
+++ b/fs/netfs/direct_write.c
@@ -87,6 +87,8 @@ ssize_t netfs_unbuffered_write_iter_locked(struct kiocb *iocb, struct iov_iter *
}
__set_bit(NETFS_RREQ_USE_IO_ITER, &wreq->flags);
+ if (async)
+ __set_bit(NETFS_RREQ_OFFLOAD_COLLECTION, &wreq->flags);
/* Copy the data into the bounce buffer and encrypt it. */
// TODO
@@ -105,13 +107,9 @@ ssize_t netfs_unbuffered_write_iter_locked(struct kiocb *iocb, struct iov_iter *
if (!async) {
trace_netfs_rreq(wreq, netfs_rreq_trace_wait_ip);
- wait_on_bit(&wreq->flags, NETFS_RREQ_IN_PROGRESS,
- TASK_UNINTERRUPTIBLE);
- ret = wreq->error;
- if (ret == 0) {
- ret = wreq->transferred;
+ ret = netfs_wait_for_write(wreq);
+ if (ret > 0)
iocb->ki_pos += ret;
- }
} else {
ret = -EIOCBQUEUED;
}
diff --git a/fs/netfs/internal.h b/fs/netfs/internal.h
index b6500a7cda81..e2ee9183392b 100644
--- a/fs/netfs/internal.h
+++ b/fs/netfs/internal.h
@@ -62,6 +62,14 @@ static inline void netfs_proc_del_rreq(struct netfs_io_request *rreq) {}
struct folio_queue *netfs_buffer_make_space(struct netfs_io_request *rreq,
enum netfs_folioq_trace trace);
void netfs_reset_iter(struct netfs_io_subrequest *subreq);
+void netfs_wake_collector(struct netfs_io_request *rreq);
+void netfs_subreq_clear_in_progress(struct netfs_io_subrequest *subreq);
+void netfs_wait_for_in_progress_stream(struct netfs_io_request *rreq,
+ struct netfs_io_stream *stream);
+ssize_t netfs_wait_for_read(struct netfs_io_request *rreq);
+ssize_t netfs_wait_for_write(struct netfs_io_request *rreq);
+void netfs_wait_for_paused_read(struct netfs_io_request *rreq);
+void netfs_wait_for_paused_write(struct netfs_io_request *rreq);
/*
* objects.c
@@ -91,11 +99,9 @@ static inline void netfs_see_subrequest(struct netfs_io_subrequest *subreq,
/*
* read_collect.c
*/
+bool netfs_read_collection(struct netfs_io_request *rreq);
void netfs_read_collection_worker(struct work_struct *work);
-void netfs_wake_read_collector(struct netfs_io_request *rreq);
void netfs_cache_read_terminated(void *priv, ssize_t transferred_or_error);
-ssize_t netfs_wait_for_read(struct netfs_io_request *rreq);
-void netfs_wait_for_pause(struct netfs_io_request *rreq);
/*
* read_pgpriv2.c
@@ -175,8 +181,8 @@ static inline void netfs_stat_d(atomic_t *stat)
* write_collect.c
*/
int netfs_folio_written_back(struct folio *folio);
+bool netfs_write_collection(struct netfs_io_request *wreq);
void netfs_write_collection_worker(struct work_struct *work);
-void netfs_wake_write_collector(struct netfs_io_request *wreq);
/*
* write_issue.c
@@ -197,8 +203,8 @@ struct netfs_io_request *netfs_begin_writethrough(struct kiocb *iocb, size_t len
int netfs_advance_writethrough(struct netfs_io_request *wreq, struct writeback_control *wbc,
struct folio *folio, size_t copied, bool to_page_end,
struct folio **writethrough_cache);
-int netfs_end_writethrough(struct netfs_io_request *wreq, struct writeback_control *wbc,
- struct folio *writethrough_cache);
+ssize_t netfs_end_writethrough(struct netfs_io_request *wreq, struct writeback_control *wbc,
+ struct folio *writethrough_cache);
int netfs_unbuffered_write(struct netfs_io_request *wreq, bool may_wait, size_t len);
/*
@@ -253,6 +259,21 @@ static inline void netfs_put_group_many(struct netfs_group *netfs_group, int nr)
netfs_group->free(netfs_group);
}
+/*
+ * Clear and wake up a NETFS_RREQ_* flag bit on a request.
+ */
+static inline void netfs_wake_rreq_flag(struct netfs_io_request *rreq,
+ unsigned int rreq_flag,
+ enum netfs_rreq_trace trace)
+{
+ if (test_bit(rreq_flag, &rreq->flags)) {
+ trace_netfs_rreq(rreq, trace);
+ clear_bit_unlock(rreq_flag, &rreq->flags);
+ smp_mb__after_atomic(); /* Set flag before task state */
+ wake_up(&rreq->waitq);
+ }
+}
+
/*
* fscache-cache.c
*/
diff --git a/fs/netfs/misc.c b/fs/netfs/misc.c
index 7099aa07737a..77e7f7c79d27 100644
--- a/fs/netfs/misc.c
+++ b/fs/netfs/misc.c
@@ -313,3 +313,221 @@ bool netfs_release_folio(struct folio *folio, gfp_t gfp)
return true;
}
EXPORT_SYMBOL(netfs_release_folio);
+
+/*
+ * Wake the collection work item.
+ */
+void netfs_wake_collector(struct netfs_io_request *rreq)
+{
+ if (test_bit(NETFS_RREQ_OFFLOAD_COLLECTION, &rreq->flags) &&
+ !test_bit(NETFS_RREQ_RETRYING, &rreq->flags)) {
+ queue_work(system_unbound_wq, &rreq->work);
+ } else {
+ trace_netfs_rreq(rreq, netfs_rreq_trace_wake_queue);
+ wake_up(&rreq->waitq);
+ }
+}
+
+/*
+ * Mark a subrequest as no longer being in progress and, if need be, wake the
+ * collector.
+ */
+void netfs_subreq_clear_in_progress(struct netfs_io_subrequest *subreq)
+{
+ struct netfs_io_request *rreq = subreq->rreq;
+ struct netfs_io_stream *stream = &rreq->io_streams[subreq->stream_nr];
+
+ clear_bit_unlock(NETFS_SREQ_IN_PROGRESS, &subreq->flags);
+ smp_mb__after_atomic(); /* Clear IN_PROGRESS before task state */
+
+ /* If we are at the head of the queue, wake up the collector. */
+ if (list_is_first(&subreq->rreq_link, &stream->subrequests) ||
+ test_bit(NETFS_RREQ_RETRYING, &rreq->flags))
+ netfs_wake_collector(rreq);
+}
+
+/*
+ * Wait for all outstanding I/O in a stream to quiesce.
+ */
+void netfs_wait_for_in_progress_stream(struct netfs_io_request *rreq,
+ struct netfs_io_stream *stream)
+{
+ struct netfs_io_subrequest *subreq;
+ DEFINE_WAIT(myself);
+
+ list_for_each_entry(subreq, &stream->subrequests, rreq_link) {
+ if (!test_bit(NETFS_SREQ_IN_PROGRESS, &subreq->flags))
+ continue;
+
+ trace_netfs_rreq(rreq, netfs_rreq_trace_wait_queue);
+ for (;;) {
+ prepare_to_wait(&rreq->waitq, &myself, TASK_UNINTERRUPTIBLE);
+
+ if (!test_bit(NETFS_SREQ_IN_PROGRESS, &subreq->flags))
+ break;
+
+ trace_netfs_sreq(subreq, netfs_sreq_trace_wait_for);
+ schedule();
+ trace_netfs_rreq(rreq, netfs_rreq_trace_woke_queue);
+ }
+ }
+
+ finish_wait(&rreq->waitq, &myself);
+}
+
+/*
+ * Perform collection in app thread if not offloaded to workqueue.
+ */
+static int netfs_collect_in_app(struct netfs_io_request *rreq,
+ bool (*collector)(struct netfs_io_request *rreq))
+{
+ bool need_collect = false, inactive = true;
+
+ for (int i = 0; i < NR_IO_STREAMS; i++) {
+ struct netfs_io_subrequest *subreq;
+ struct netfs_io_stream *stream = &rreq->io_streams[i];
+
+ if (!stream->active)
+ continue;
+ inactive = false;
+ trace_netfs_collect_stream(rreq, stream);
+ subreq = list_first_entry_or_null(&stream->subrequests,
+ struct netfs_io_subrequest,
+ rreq_link);
+ if (subreq &&
+ (!test_bit(NETFS_SREQ_IN_PROGRESS, &subreq->flags) ||
+ test_bit(NETFS_SREQ_MADE_PROGRESS, &subreq->flags))) {
+ need_collect = true;
+ break;
+ }
+ }
+
+ if (!need_collect && !inactive)
+ return 0; /* Sleep */
+
+ __set_current_state(TASK_RUNNING);
+ if (collector(rreq)) {
+ /* Drop the ref from the NETFS_RREQ_IN_PROGRESS flag. */
+ netfs_put_request(rreq, netfs_rreq_trace_put_work_ip);
+ return 1; /* Done */
+ }
+
+ if (inactive) {
+ WARN(true, "Failed to collect inactive req R=%08x\n",
+ rreq->debug_id);
+ cond_resched();
+ }
+ return 2; /* Again */
+}
+
+/*
+ * Wait for a request to complete, successfully or otherwise.
+ */
+static ssize_t netfs_wait_for_request(struct netfs_io_request *rreq,
+ bool (*collector)(struct netfs_io_request *rreq))
+{
+ DEFINE_WAIT(myself);
+ ssize_t ret;
+
+ for (;;) {
+ trace_netfs_rreq(rreq, netfs_rreq_trace_wait_queue);
+ prepare_to_wait(&rreq->waitq, &myself, TASK_UNINTERRUPTIBLE);
+
+ if (!test_bit(NETFS_RREQ_OFFLOAD_COLLECTION, &rreq->flags)) {
+ switch (netfs_collect_in_app(rreq, collector)) {
+ case 0:
+ break;
+ case 1:
+ goto all_collected;
+ case 2:
+ continue;
+ }
+ }
+
+ if (!test_bit(NETFS_RREQ_IN_PROGRESS, &rreq->flags))
+ break;
+
+ schedule();
+ trace_netfs_rreq(rreq, netfs_rreq_trace_woke_queue);
+ }
+
+all_collected:
+ finish_wait(&rreq->waitq, &myself);
+
+ ret = rreq->error;
+ if (ret == 0) {
+ ret = rreq->transferred;
+ switch (rreq->origin) {
+ case NETFS_DIO_READ:
+ case NETFS_DIO_WRITE:
+ case NETFS_READ_SINGLE:
+ case NETFS_UNBUFFERED_WRITE:
+ break;
+ default:
+ if (rreq->submitted < rreq->len) {
+ trace_netfs_failure(rreq, NULL, ret, netfs_fail_short_read);
+ ret = -EIO;
+ }
+ break;
+ }
+ }
+
+ return ret;
+}
+
+ssize_t netfs_wait_for_read(struct netfs_io_request *rreq)
+{
+ return netfs_wait_for_request(rreq, netfs_read_collection);
+}
+
+ssize_t netfs_wait_for_write(struct netfs_io_request *rreq)
+{
+ return netfs_wait_for_request(rreq, netfs_write_collection);
+}
+
+/*
+ * Wait for a paused operation to unpause or complete in some manner.
+ */
+static void netfs_wait_for_pause(struct netfs_io_request *rreq,
+ bool (*collector)(struct netfs_io_request *rreq))
+{
+ DEFINE_WAIT(myself);
+
+ trace_netfs_rreq(rreq, netfs_rreq_trace_wait_pause);
+
+ for (;;) {
+ trace_netfs_rreq(rreq, netfs_rreq_trace_wait_queue);
+ prepare_to_wait(&rreq->waitq, &myself, TASK_UNINTERRUPTIBLE);
+
+ if (!test_bit(NETFS_RREQ_OFFLOAD_COLLECTION, &rreq->flags)) {
+ switch (netfs_collect_in_app(rreq, collector)) {
+ case 0:
+ break;
+ case 1:
+ goto all_collected;
+ case 2:
+ continue;
+ }
+ }
+
+ if (!test_bit(NETFS_RREQ_IN_PROGRESS, &rreq->flags) ||
+ !test_bit(NETFS_RREQ_PAUSE, &rreq->flags))
+ break;
+
+ schedule();
+ trace_netfs_rreq(rreq, netfs_rreq_trace_woke_queue);
+ }
+
+all_collected:
+ finish_wait(&rreq->waitq, &myself);
+}
+
+void netfs_wait_for_paused_read(struct netfs_io_request *rreq)
+{
+ return netfs_wait_for_pause(rreq, netfs_read_collection);
+}
+
+void netfs_wait_for_paused_write(struct netfs_io_request *rreq)
+{
+ return netfs_wait_for_pause(rreq, netfs_write_collection);
+}
diff --git a/fs/netfs/read_collect.c b/fs/netfs/read_collect.c
index 1197ebce5675..900dd51c3b94 100644
--- a/fs/netfs/read_collect.c
+++ b/fs/netfs/read_collect.c
@@ -315,14 +315,8 @@ static void netfs_collect_read_results(struct netfs_io_request *rreq)
if (notes & NEED_RETRY)
goto need_retry;
- if ((notes & MADE_PROGRESS) && test_bit(NETFS_RREQ_PAUSE, &rreq->flags)) {
- trace_netfs_rreq(rreq, netfs_rreq_trace_unpause);
- clear_bit_unlock(NETFS_RREQ_PAUSE, &rreq->flags);
- smp_mb__after_atomic(); /* Set PAUSE before task state */
- wake_up(&rreq->waitq);
- }
-
if (notes & MADE_PROGRESS) {
+ netfs_wake_rreq_flag(rreq, NETFS_RREQ_PAUSE, netfs_rreq_trace_unpause);
//cond_resched();
goto reassess;
}
@@ -399,7 +393,7 @@ static void netfs_rreq_assess_single(struct netfs_io_request *rreq)
* Note that we're in normal kernel thread context at this point, possibly
* running on a workqueue.
*/
-static bool netfs_read_collection(struct netfs_io_request *rreq)
+bool netfs_read_collection(struct netfs_io_request *rreq)
{
struct netfs_io_stream *stream = &rreq->io_streams[0];
@@ -434,8 +428,7 @@ static bool netfs_read_collection(struct netfs_io_request *rreq)
}
task_io_account_read(rreq->transferred);
- trace_netfs_rreq(rreq, netfs_rreq_trace_wake_ip);
- clear_and_wake_up_bit(NETFS_RREQ_IN_PROGRESS, &rreq->flags);
+ netfs_wake_rreq_flag(rreq, NETFS_RREQ_IN_PROGRESS, netfs_rreq_trace_wake_ip);
/* As we cleared NETFS_RREQ_IN_PROGRESS, we acquired its ref. */
trace_netfs_rreq(rreq, netfs_rreq_trace_done);
@@ -460,20 +453,6 @@ void netfs_read_collection_worker(struct work_struct *work)
}
}
-/*
- * Wake the collection work item.
- */
-void netfs_wake_read_collector(struct netfs_io_request *rreq)
-{
- if (test_bit(NETFS_RREQ_OFFLOAD_COLLECTION, &rreq->flags) &&
- !test_bit(NETFS_RREQ_RETRYING, &rreq->flags)) {
- queue_work(system_unbound_wq, &rreq->work);
- } else {
- trace_netfs_rreq(rreq, netfs_rreq_trace_wake_queue);
- wake_up(&rreq->waitq);
- }
-}
-
/**
* netfs_read_subreq_progress - Note progress of a read operation.
* @subreq: The read request that has terminated.
@@ -502,7 +481,7 @@ void netfs_read_subreq_progress(struct netfs_io_subrequest *subreq)
list_is_first(&subreq->rreq_link, &stream->subrequests)
) {
__set_bit(NETFS_SREQ_MADE_PROGRESS, &subreq->flags);
- netfs_wake_read_collector(rreq);
+ netfs_wake_collector(rreq);
}
}
EXPORT_SYMBOL(netfs_read_subreq_progress);
@@ -526,7 +505,6 @@ EXPORT_SYMBOL(netfs_read_subreq_progress);
void netfs_read_subreq_terminated(struct netfs_io_subrequest *subreq)
{
struct netfs_io_request *rreq = subreq->rreq;
- struct netfs_io_stream *stream = &rreq->io_streams[0];
switch (subreq->source) {
case NETFS_READ_FROM_CACHE:
@@ -573,15 +551,7 @@ void netfs_read_subreq_terminated(struct netfs_io_subrequest *subreq)
}
trace_netfs_sreq(subreq, netfs_sreq_trace_terminated);
-
- clear_bit_unlock(NETFS_SREQ_IN_PROGRESS, &subreq->flags);
- smp_mb__after_atomic(); /* Clear IN_PROGRESS before task state */
-
- /* If we are at the head of the queue, wake up the collector. */
- if (list_is_first(&subreq->rreq_link, &stream->subrequests) ||
- test_bit(NETFS_RREQ_RETRYING, &rreq->flags))
- netfs_wake_read_collector(rreq);
-
+ netfs_subreq_clear_in_progress(subreq);
netfs_put_subrequest(subreq, netfs_sreq_trace_put_terminated);
}
EXPORT_SYMBOL(netfs_read_subreq_terminated);
@@ -604,102 +574,3 @@ void netfs_cache_read_terminated(void *priv, ssize_t transferred_or_error)
}
netfs_read_subreq_terminated(subreq);
}
-
-/*
- * Wait for the read operation to complete, successfully or otherwise.
- */
-ssize_t netfs_wait_for_read(struct netfs_io_request *rreq)
-{
- struct netfs_io_subrequest *subreq;
- struct netfs_io_stream *stream = &rreq->io_streams[0];
- DEFINE_WAIT(myself);
- ssize_t ret;
-
- for (;;) {
- trace_netfs_rreq(rreq, netfs_rreq_trace_wait_queue);
- prepare_to_wait(&rreq->waitq, &myself, TASK_UNINTERRUPTIBLE);
-
- subreq = list_first_entry_or_null(&stream->subrequests,
- struct netfs_io_subrequest, rreq_link);
- if (subreq &&
- (!test_bit(NETFS_SREQ_IN_PROGRESS, &subreq->flags) ||
- test_bit(NETFS_SREQ_MADE_PROGRESS, &subreq->flags))) {
- __set_current_state(TASK_RUNNING);
- if (netfs_read_collection(rreq)) {
- /* Drop the ref from the NETFS_RREQ_IN_PROGRESS flag. */
- netfs_put_request(rreq, netfs_rreq_trace_put_work_ip);
- break;
- }
- continue;
- }
-
- if (!test_bit(NETFS_RREQ_IN_PROGRESS, &rreq->flags))
- break;
-
- schedule();
- trace_netfs_rreq(rreq, netfs_rreq_trace_woke_queue);
- }
-
- finish_wait(&rreq->waitq, &myself);
-
- ret = rreq->error;
- if (ret == 0) {
- ret = rreq->transferred;
- switch (rreq->origin) {
- case NETFS_DIO_READ:
- case NETFS_READ_SINGLE:
- ret = rreq->transferred;
- break;
- default:
- if (rreq->submitted < rreq->len) {
- trace_netfs_failure(rreq, NULL, ret, netfs_fail_short_read);
- ret = -EIO;
- }
- break;
- }
- }
-
- return ret;
-}
-
-/*
- * Wait for a paused read operation to unpause or complete in some manner.
- */
-void netfs_wait_for_pause(struct netfs_io_request *rreq)
-{
- struct netfs_io_subrequest *subreq;
- struct netfs_io_stream *stream = &rreq->io_streams[0];
- DEFINE_WAIT(myself);
-
- trace_netfs_rreq(rreq, netfs_rreq_trace_wait_pause);
-
- for (;;) {
- trace_netfs_rreq(rreq, netfs_rreq_trace_wait_queue);
- prepare_to_wait(&rreq->waitq, &myself, TASK_UNINTERRUPTIBLE);
-
- if (!test_bit(NETFS_RREQ_OFFLOAD_COLLECTION, &rreq->flags)) {
- subreq = list_first_entry_or_null(&stream->subrequests,
- struct netfs_io_subrequest, rreq_link);
- if (subreq &&
- (!test_bit(NETFS_SREQ_IN_PROGRESS, &subreq->flags) ||
- test_bit(NETFS_SREQ_MADE_PROGRESS, &subreq->flags))) {
- __set_current_state(TASK_RUNNING);
- if (netfs_read_collection(rreq)) {
- /* Drop the ref from the NETFS_RREQ_IN_PROGRESS flag. */
- netfs_put_request(rreq, netfs_rreq_trace_put_work_ip);
- break;
- }
- continue;
- }
- }
-
- if (!test_bit(NETFS_RREQ_IN_PROGRESS, &rreq->flags) ||
- !test_bit(NETFS_RREQ_PAUSE, &rreq->flags))
- break;
-
- schedule();
- trace_netfs_rreq(rreq, netfs_rreq_trace_woke_queue);
- }
-
- finish_wait(&rreq->waitq, &myself);
-}
diff --git a/fs/netfs/read_retry.c b/fs/netfs/read_retry.c
index 1378dc7fa2cc..b99e84a8170a 100644
--- a/fs/netfs/read_retry.c
+++ b/fs/netfs/read_retry.c
@@ -257,35 +257,15 @@ static void netfs_retry_read_subrequests(struct netfs_io_request *rreq)
*/
void netfs_retry_reads(struct netfs_io_request *rreq)
{
- struct netfs_io_subrequest *subreq;
struct netfs_io_stream *stream = &rreq->io_streams[0];
- DEFINE_WAIT(myself);
netfs_stat(&netfs_n_rh_retry_read_req);
- set_bit(NETFS_RREQ_RETRYING, &rreq->flags);
-
/* Wait for all outstanding I/O to quiesce before performing retries as
* we may need to renegotiate the I/O sizes.
*/
- list_for_each_entry(subreq, &stream->subrequests, rreq_link) {
- if (!test_bit(NETFS_SREQ_IN_PROGRESS, &subreq->flags))
- continue;
-
- trace_netfs_rreq(rreq, netfs_rreq_trace_wait_queue);
- for (;;) {
- prepare_to_wait(&rreq->waitq, &myself, TASK_UNINTERRUPTIBLE);
-
- if (!test_bit(NETFS_SREQ_IN_PROGRESS, &subreq->flags))
- break;
-
- trace_netfs_sreq(subreq, netfs_sreq_trace_wait_for);
- schedule();
- trace_netfs_rreq(rreq, netfs_rreq_trace_woke_queue);
- }
-
- finish_wait(&rreq->waitq, &myself);
- }
+ set_bit(NETFS_RREQ_RETRYING, &rreq->flags);
+ netfs_wait_for_in_progress_stream(rreq, stream);
clear_bit(NETFS_RREQ_RETRYING, &rreq->flags);
trace_netfs_rreq(rreq, netfs_rreq_trace_resubmit);
diff --git a/fs/netfs/write_collect.c b/fs/netfs/write_collect.c
index 7241d1fd2c14..0ce7b53e7fe8 100644
--- a/fs/netfs/write_collect.c
+++ b/fs/netfs/write_collect.c
@@ -321,18 +321,14 @@ static void netfs_collect_write_results(struct netfs_io_request *wreq)
if (notes & NEED_RETRY)
goto need_retry;
- if ((notes & MADE_PROGRESS) && test_bit(NETFS_RREQ_PAUSE, &wreq->flags)) {
- trace_netfs_rreq(wreq, netfs_rreq_trace_unpause);
- clear_bit_unlock(NETFS_RREQ_PAUSE, &wreq->flags);
- smp_mb__after_atomic(); /* Set PAUSE before task state */
- wake_up(&wreq->waitq);
- }
- if (notes & NEED_REASSESS) {
+ if (notes & MADE_PROGRESS) {
+ netfs_wake_rreq_flag(wreq, NETFS_RREQ_PAUSE, netfs_rreq_trace_unpause);
//cond_resched();
goto reassess_streams;
}
- if (notes & MADE_PROGRESS) {
+
+ if (notes & NEED_REASSESS) {
//cond_resched();
goto reassess_streams;
}
@@ -356,7 +352,7 @@ static void netfs_collect_write_results(struct netfs_io_request *wreq)
/*
* Perform the collection of subrequests, folios and encryption buffers.
*/
-static bool netfs_write_collection(struct netfs_io_request *wreq)
+bool netfs_write_collection(struct netfs_io_request *wreq)
{
struct netfs_inode *ictx = netfs_inode(wreq->inode);
size_t transferred;
@@ -417,8 +413,7 @@ static bool netfs_write_collection(struct netfs_io_request *wreq)
inode_dio_end(wreq->inode);
_debug("finished");
- trace_netfs_rreq(wreq, netfs_rreq_trace_wake_ip);
- clear_and_wake_up_bit(NETFS_RREQ_IN_PROGRESS, &wreq->flags);
+ netfs_wake_rreq_flag(wreq, NETFS_RREQ_IN_PROGRESS, netfs_rreq_trace_wake_ip);
/* As we cleared NETFS_RREQ_IN_PROGRESS, we acquired its ref. */
if (wreq->iocb) {
@@ -448,14 +443,6 @@ void netfs_write_collection_worker(struct work_struct *work)
}
}
-/*
- * Wake the collection work item.
- */
-void netfs_wake_write_collector(struct netfs_io_request *wreq)
-{
- queue_work(system_unbound_wq, &wreq->work);
-}
-
/**
* netfs_write_subrequest_terminated - Note the termination of a write operation.
* @_op: The I/O request that has terminated.
@@ -479,7 +466,6 @@ void netfs_write_subrequest_terminated(void *_op, ssize_t transferred_or_error)
{
struct netfs_io_subrequest *subreq = _op;
struct netfs_io_request *wreq = subreq->rreq;
- struct netfs_io_stream *stream = &wreq->io_streams[subreq->stream_nr];
_enter("%x[%x] %zd", wreq->debug_id, subreq->debug_index, transferred_or_error);
@@ -531,15 +517,7 @@ void netfs_write_subrequest_terminated(void *_op, ssize_t transferred_or_error)
}
trace_netfs_sreq(subreq, netfs_sreq_trace_terminated);
-
- clear_and_wake_up_bit(NETFS_SREQ_IN_PROGRESS, &subreq->flags);
-
- /* If we are at the head of the queue, wake up the collector,
- * transferring a ref to it if we were the ones to do so.
- */
- if (list_is_first(&subreq->rreq_link, &stream->subrequests))
- netfs_wake_write_collector(wreq);
-
+ netfs_subreq_clear_in_progress(subreq);
netfs_put_subrequest(subreq, netfs_sreq_trace_put_terminated);
}
EXPORT_SYMBOL(netfs_write_subrequest_terminated);
diff --git a/fs/netfs/write_issue.c b/fs/netfs/write_issue.c
index 8744ed3faf29..50bee2c4130d 100644
--- a/fs/netfs/write_issue.c
+++ b/fs/netfs/write_issue.c
@@ -542,7 +542,7 @@ static void netfs_end_issue_write(struct netfs_io_request *wreq)
}
if (needs_poke)
- netfs_wake_write_collector(wreq);
+ netfs_wake_collector(wreq);
}
/*
@@ -576,6 +576,7 @@ int netfs_writepages(struct address_space *mapping,
goto couldnt_start;
}
+ __set_bit(NETFS_RREQ_OFFLOAD_COLLECTION, &wreq->flags);
trace_netfs_write(wreq, netfs_write_trace_writeback);
netfs_stat(&netfs_n_wh_writepages);
@@ -599,7 +600,7 @@ int netfs_writepages(struct address_space *mapping,
netfs_end_issue_write(wreq);
mutex_unlock(&ictx->wb_lock);
- netfs_wake_write_collector(wreq);
+ netfs_wake_collector(wreq);
netfs_put_request(wreq, netfs_rreq_trace_put_return);
_leave(" = %d", error);
@@ -674,11 +675,11 @@ int netfs_advance_writethrough(struct netfs_io_request *wreq, struct writeback_c
/*
* End a write operation used when writing through the pagecache.
*/
-int netfs_end_writethrough(struct netfs_io_request *wreq, struct writeback_control *wbc,
- struct folio *writethrough_cache)
+ssize_t netfs_end_writethrough(struct netfs_io_request *wreq, struct writeback_control *wbc,
+ struct folio *writethrough_cache)
{
struct netfs_inode *ictx = netfs_inode(wreq->inode);
- int ret;
+ ssize_t ret;
_enter("R=%x", wreq->debug_id);
@@ -689,12 +690,10 @@ int netfs_end_writethrough(struct netfs_io_request *wreq, struct writeback_contr
mutex_unlock(&ictx->wb_lock);
- if (wreq->iocb) {
+ if (wreq->iocb)
ret = -EIOCBQUEUED;
- } else {
- wait_on_bit(&wreq->flags, NETFS_RREQ_IN_PROGRESS, TASK_UNINTERRUPTIBLE);
- ret = wreq->error;
- }
+ else
+ ret = netfs_wait_for_write(wreq);
netfs_put_request(wreq, netfs_rreq_trace_put_return);
return ret;
}
@@ -723,10 +722,8 @@ int netfs_unbuffered_write(struct netfs_io_request *wreq, bool may_wait, size_t
start += part;
len -= part;
rolling_buffer_advance(&wreq->buffer, part);
- if (test_bit(NETFS_RREQ_PAUSE, &wreq->flags)) {
- trace_netfs_rreq(wreq, netfs_rreq_trace_wait_pause);
- wait_event(wreq->waitq, !test_bit(NETFS_RREQ_PAUSE, &wreq->flags));
- }
+ if (test_bit(NETFS_RREQ_PAUSE, &wreq->flags))
+ netfs_wait_for_paused_write(wreq);
if (test_bit(NETFS_RREQ_FAILED, &wreq->flags))
break;
}
@@ -886,6 +883,7 @@ int netfs_writeback_single(struct address_space *mapping,
goto couldnt_start;
}
+ __set_bit(NETFS_RREQ_OFFLOAD_COLLECTION, &wreq->flags);
trace_netfs_write(wreq, netfs_write_trace_writeback_single);
netfs_stat(&netfs_n_wh_writepages);
@@ -915,7 +913,7 @@ int netfs_writeback_single(struct address_space *mapping,
set_bit(NETFS_RREQ_ALL_QUEUED, &wreq->flags);
mutex_unlock(&ictx->wb_lock);
- netfs_wake_write_collector(wreq);
+ netfs_wake_collector(wreq);
netfs_put_request(wreq, netfs_rreq_trace_put_return);
_leave(" = %d", ret);
diff --git a/fs/netfs/write_retry.c b/fs/netfs/write_retry.c
index 7408f6bb8e42..9d1d8a8bab72 100644
--- a/fs/netfs/write_retry.c
+++ b/fs/netfs/write_retry.c
@@ -200,7 +200,6 @@ static void netfs_retry_write_stream(struct netfs_io_request *wreq,
*/
void netfs_retry_writes(struct netfs_io_request *wreq)
{
- struct netfs_io_subrequest *subreq;
struct netfs_io_stream *stream;
int s;
@@ -209,16 +208,13 @@ void netfs_retry_writes(struct netfs_io_request *wreq)
/* Wait for all outstanding I/O to quiesce before performing retries as
* we may need to renegotiate the I/O sizes.
*/
+ set_bit(NETFS_RREQ_RETRYING, &wreq->flags);
for (s = 0; s < NR_IO_STREAMS; s++) {
stream = &wreq->io_streams[s];
- if (!stream->active)
- continue;
-
- list_for_each_entry(subreq, &stream->subrequests, rreq_link) {
- wait_on_bit(&subreq->flags, NETFS_SREQ_IN_PROGRESS,
- TASK_UNINTERRUPTIBLE);
- }
+ if (stream->active)
+ netfs_wait_for_in_progress_stream(wreq, stream);
}
+ clear_bit(NETFS_RREQ_RETRYING, &wreq->flags);
// TODO: Enc: Fetch changed partial pages
// TODO: Enc: Reencrypt content if needed.
^ permalink raw reply related [flat|nested] 6+ messages in thread
* Re: [PATCH 0/4] netfs: Miscellaneous fixes
2025-05-19 9:07 [PATCH 0/4] netfs: Miscellaneous fixes David Howells
` (3 preceding siblings ...)
2025-05-19 9:07 ` [PATCH 4/4] netfs: Fix wait/wake to be consistent about the waitqueue used David Howells
@ 2025-05-20 11:34 ` Christian Brauner
4 siblings, 0 replies; 6+ messages in thread
From: Christian Brauner @ 2025-05-20 11:34 UTC (permalink / raw)
To: Christian Brauner, David Howells
Cc: Paulo Alcantara, netfs, linux-afs, linux-cifs, linux-nfs,
ceph-devel, v9fs, linux-fsdevel, linux-kernel
On Mon, 19 May 2025 10:07:00 +0100, David Howells wrote:
> Here are some miscellaneous fixes and changes for netfslib, if you could
> pull them:
>
> (1) Fix an oops in write-retry due to mis-resetting the I/O iterator.
>
> (2) Fix the recording of transferred bytes for short DIO reads.
>
> [...]
Applied to the vfs-6.16.netfs branch of the vfs/vfs.git tree.
Patches in the vfs-6.16.netfs branch should appear in linux-next soon.
Please report any outstanding bugs that were missed during review in a
new review to the original patch series allowing us to drop it.
It's encouraged to provide Acked-bys and Reviewed-bys even though the
patch has now been applied. If possible patch trailers will be updated.
Note that commit hashes shown below are subject to change due to rebase,
trailer updates or similar. If in doubt, please check the listed branch.
tree: https://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs.git
branch: vfs-6.16.netfs
[1/4] netfs: Fix oops in write-retry from mis-resetting the subreq iterator
https://git.kernel.org/vfs/vfs/c/cd084c7184ce
[2/4] netfs: Fix setting of transferred bytes with short DIO reads
https://git.kernel.org/vfs/vfs/c/2973904c9b79
[3/4] netfs: Fix the request's work item to not require a ref
https://git.kernel.org/vfs/vfs/c/537b296114cc
[4/4] netfs: Fix wait/wake to be consistent about the waitqueue used
https://git.kernel.org/vfs/vfs/c/2a6d0284a4a3
^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2025-05-20 11:34 UTC | newest]
Thread overview: 6+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2025-05-19 9:07 [PATCH 0/4] netfs: Miscellaneous fixes David Howells
2025-05-19 9:07 ` [PATCH 1/4] netfs: Fix oops in write-retry from mis-resetting the subreq iterator David Howells
2025-05-19 9:07 ` [PATCH 2/4] netfs: Fix setting of transferred bytes with short DIO reads David Howells
2025-05-19 9:07 ` [PATCH 3/4] netfs: Fix the request's work item to not require a ref David Howells
2025-05-19 9:07 ` [PATCH 4/4] netfs: Fix wait/wake to be consistent about the waitqueue used David Howells
2025-05-20 11:34 ` [PATCH 0/4] netfs: Miscellaneous fixes Christian Brauner
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).