From: Kent Overstreet <kent.overstreet@linux.dev>
To: Jens Axboe <axboe@kernel.dk>
Cc: torvalds@linux-foundation.org, linux-kernel@vger.kernel.org,
linux-fsdevel@vger.kernel.org, linux-bcachefs@vger.kernel.org,
Christoph Hellwig <hch@lst.de>,
Christian Brauner <brauner@kernel.org>
Subject: Re: [GIT PULL] bcachefs
Date: Thu, 6 Jul 2023 16:15:10 -0400 [thread overview]
Message-ID: <20230706201510.sh5ukzfsf5vdxvrf@moria.home.lan> (raw)
In-Reply-To: <2e635579-37ba-ddfc-a2ab-e6c080ab4971@kernel.dk>
On Wed, Jun 28, 2023 at 03:17:43PM -0600, Jens Axboe wrote:
> On 6/28/23 2:44?PM, Jens Axboe wrote:
> > On 6/28/23 11:52?AM, Kent Overstreet wrote:
> >> On Wed, Jun 28, 2023 at 10:57:02AM -0600, Jens Axboe wrote:
> >>> I discussed this with Christian offline. I have a patch that is pretty
> >>> simple, but it does mean that you'd wait for delayed fput flush off
> >>> umount. Which seems kind of iffy.
> >>>
> >>> I think we need to back up a bit and consider if the kill && umount
> >>> really is sane. If you kill a task that has open files, then any fput
> >>> from that task will end up being delayed. This means that the umount may
> >>> very well fail.
> >>>
> >>> It'd be handy if we could have umount wait for that to finish, but I'm
> >>> not at all confident this is a sane solution for all cases. And as
> >>> discussed, we have no way to even identify which files we'd need to
> >>> flush out of the delayed list.
> >>>
> >>> Maybe the test case just needs fixing? Christian suggested lazy/detach
> >>> umount and wait for sb release. There's an fsnotify hook for that,
> >>> fsnotify_sb_delete(). Obviously this is a bit more involved, but seems
> >>> to me that this would be the way to make it more reliable when killing
> >>> of tasks with open files are involved.
> >>
> >> No, this is a real breakage. Any time we introduce unexpected
> >> asynchrony there's the potential for breakage: case in point, there was
> >> a filesystem that made rm asynchronous, then there were scripts out
> >> there that deleted until df showed under some threshold.. whoops...
> >
> > This is nothing new - any fput done from an exiting task will end up
> > being deferred. The window may be a bit wider now or a bit different,
> > but it's the same window. If an application assumes it can kill && wait
> > on a task and be guaranteed that the files are released as soon as wait
> > returns, it is mistaken. That is NOT the case.
>
> Case in point, just changed my reproducer to use aio instead of
> io_uring. Here's the full script:
>
> #!/bin/bash
>
> DEV=/dev/nvme1n1
> MNT=/data
> ITER=0
>
> while true; do
> echo loop $ITER
> sudo mount $DEV $MNT
> fio --name=test --ioengine=aio --iodepth=2 --filename=$MNT/foo --size=1g --buffered=1 --overwrite=0 --numjobs=12 --minimal --rw=randread --output=/dev/null &
> Y=$(($RANDOM % 3))
> X=$(($RANDOM % 10))
> VAL="$Y.$X"
> sleep $VAL
> ps -e | grep fio > /dev/null 2>&1
> while [ $? -eq 0 ]; do
> killall -9 fio > /dev/null 2>&1
> echo will wait
> wait > /dev/null 2>&1
> echo done waiting
> ps -e | grep "fio " > /dev/null 2>&1
> done
> sudo umount /data
> if [ $? -ne 0 ]; then
> break
> fi
> ((ITER++))
> done
>
> and if I run that, fails on the first umount attempt in that loop:
>
> axboe@m1max-kvm ~> bash test2.sh
> loop 0
> will wait
> done waiting
> umount: /data: target is busy.
Your test fails because fio by default spawns off multiple processes,
and just calling wait does not wait for the subprocesses.
When I pass --thread to fio, your test passes.
I have a patch to avoid use of the delayed_fput list in the aio path,
but curiously it seems not to be needed - perhaps there's some other
synchronization I haven't found yet. I'm including the patch below in
case the technique is useful for io_uring:
diff --git a/fs/aio.c b/fs/aio.c
index b3e14a9fe3..00cb953efa 100644
--- a/fs/aio.c
+++ b/fs/aio.c
@@ -211,6 +211,7 @@ struct aio_kiocb {
* for cancellation */
refcount_t ki_refcnt;
+ struct task_struct *ki_task;
/*
* If the aio_resfd field of the userspace iocb is not zero,
* this is the underlying eventfd context to deliver events to.
@@ -321,7 +322,7 @@ static void put_aio_ring_file(struct kioctx *ctx)
ctx->aio_ring_file = NULL;
spin_unlock(&i_mapping->private_lock);
- fput(aio_ring_file);
+ __fput_sync(aio_ring_file);
}
}
@@ -1068,6 +1069,7 @@ static inline struct aio_kiocb *aio_get_req(struct kioctx *ctx)
INIT_LIST_HEAD(&req->ki_list);
refcount_set(&req->ki_refcnt, 2);
req->ki_eventfd = NULL;
+ req->ki_task = get_task_struct(current);
return req;
}
@@ -1104,8 +1106,9 @@ static inline void iocb_destroy(struct aio_kiocb *iocb)
if (iocb->ki_eventfd)
eventfd_ctx_put(iocb->ki_eventfd);
if (iocb->ki_filp)
- fput(iocb->ki_filp);
+ fput_for_task(iocb->ki_filp, iocb->ki_task);
percpu_ref_put(&iocb->ki_ctx->reqs);
+ put_task_struct(iocb->ki_task);
kmem_cache_free(kiocb_cachep, iocb);
}
diff --git a/fs/file_table.c b/fs/file_table.c
index 372653b926..137f87f55e 100644
--- a/fs/file_table.c
+++ b/fs/file_table.c
@@ -367,12 +367,13 @@ EXPORT_SYMBOL_GPL(flush_delayed_fput);
static DECLARE_DELAYED_WORK(delayed_fput_work, delayed_fput);
-void fput(struct file *file)
+void fput_for_task(struct file *file, struct task_struct *task)
{
if (atomic_long_dec_and_test(&file->f_count)) {
- struct task_struct *task = current;
+ if (!task && likely(!in_interrupt() && !(current->flags & PF_KTHREAD)))
+ task = current;
- if (likely(!in_interrupt() && !(task->flags & PF_KTHREAD))) {
+ if (task) {
init_task_work(&file->f_rcuhead, ____fput);
if (!task_work_add(task, &file->f_rcuhead, TWA_RESUME))
return;
@@ -388,6 +389,11 @@ void fput(struct file *file)
}
}
+void fput(struct file *file)
+{
+ fput_for_task(file, NULL);
+}
+
/*
* synchronous analog of fput(); for kernel threads that might be needed
* in some umount() (and thus can't use flush_delayed_fput() without
@@ -405,6 +411,7 @@ void __fput_sync(struct file *file)
}
}
+EXPORT_SYMBOL(fput_for_task);
EXPORT_SYMBOL(fput);
EXPORT_SYMBOL(__fput_sync);
diff --git a/include/linux/file.h b/include/linux/file.h
index 39704eae83..667a68f477 100644
--- a/include/linux/file.h
+++ b/include/linux/file.h
@@ -12,7 +12,9 @@
#include <linux/errno.h>
struct file;
+struct task_struct;
+extern void fput_for_task(struct file *, struct task_struct *);
extern void fput(struct file *);
struct file_operations;
next prev parent reply other threads:[~2023-07-06 20:15 UTC|newest]
Thread overview: 140+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-06-26 21:46 [GIT PULL] bcachefs Kent Overstreet
2023-06-26 23:11 ` Jens Axboe
2023-06-27 0:06 ` Kent Overstreet
2023-06-27 1:13 ` Jens Axboe
2023-06-27 2:05 ` Kent Overstreet
2023-06-27 2:59 ` Jens Axboe
2023-06-27 3:10 ` Kent Overstreet
2023-06-27 17:16 ` Jens Axboe
2023-06-27 20:15 ` Kent Overstreet
2023-06-27 22:05 ` Dave Chinner
2023-06-27 22:41 ` Kent Overstreet
2023-06-28 14:40 ` Jens Axboe
2023-06-28 14:48 ` Thomas Weißschuh
2023-06-28 14:58 ` Jens Axboe
2023-06-28 3:16 ` Jens Axboe
2023-06-28 4:01 ` Kent Overstreet
2023-06-28 14:58 ` Jens Axboe
2023-06-28 15:22 ` Jens Axboe
2023-06-28 17:56 ` Kent Overstreet
2023-06-28 20:45 ` Jens Axboe
2023-06-28 16:57 ` Jens Axboe
2023-06-28 17:33 ` Christian Brauner
2023-06-28 17:52 ` Kent Overstreet
2023-06-28 20:44 ` Jens Axboe
2023-06-28 21:17 ` Jens Axboe
2023-06-28 22:13 ` Kent Overstreet
2023-06-28 22:33 ` Jens Axboe
2023-06-28 22:55 ` Kent Overstreet
2023-06-28 23:14 ` Jens Axboe
2023-06-28 23:50 ` Kent Overstreet
2023-06-29 1:00 ` Dave Chinner
2023-06-29 1:33 ` Jens Axboe
2023-06-29 11:18 ` Christian Brauner
2023-06-29 14:17 ` Kent Overstreet
2023-06-29 15:31 ` Kent Overstreet
2023-06-30 9:40 ` Christian Brauner
2023-07-06 15:20 ` Kent Overstreet
2023-07-06 16:26 ` Jens Axboe
2023-07-06 16:34 ` Kent Overstreet
2023-06-29 1:29 ` Jens Axboe
2023-07-06 20:15 ` Kent Overstreet [this message]
2023-06-28 17:54 ` Kent Overstreet
2023-06-28 20:54 ` Jens Axboe
2023-06-28 22:14 ` Jens Axboe
2023-06-28 23:04 ` Kent Overstreet
2023-06-28 23:11 ` Jens Axboe
2023-06-27 2:33 ` Kent Overstreet
2023-06-27 2:59 ` Jens Axboe
2023-06-27 3:19 ` Matthew Wilcox
2023-06-27 3:22 ` Kent Overstreet
2023-06-27 3:52 ` Christoph Hellwig
2023-06-27 4:36 ` Kent Overstreet
2023-07-06 15:56 ` Kent Overstreet
2023-07-06 16:40 ` Josef Bacik
2023-07-06 17:38 ` Kent Overstreet
2023-07-06 19:17 ` Eric Sandeen
2023-07-06 19:31 ` Kent Overstreet
2023-07-06 21:19 ` Darrick J. Wong
2023-07-06 22:43 ` Kent Overstreet
2023-07-07 13:13 ` Jan Kara
2023-07-07 13:52 ` Kent Overstreet
2023-07-07 8:48 ` Christian Brauner
2023-07-07 9:18 ` Kent Overstreet
2023-07-07 16:26 ` James Bottomley
2023-07-07 16:48 ` Kent Overstreet
2023-07-07 17:04 ` James Bottomley
2023-07-07 17:26 ` Kent Overstreet
2023-07-08 3:54 ` Matthew Wilcox
2023-07-08 4:10 ` Kent Overstreet
2023-07-08 4:31 ` Kent Overstreet
2023-07-08 15:02 ` Theodore Ts'o
2023-07-08 15:23 ` Kent Overstreet
2023-07-08 16:42 ` James Bottomley
2023-07-09 1:16 ` Kent Overstreet
2023-07-07 9:35 ` Kent Overstreet
2023-07-07 2:04 ` Theodore Ts'o
2023-07-07 12:18 ` Brian Foster
2023-07-07 14:49 ` Kent Overstreet
2023-07-12 2:54 ` Kent Overstreet
2023-07-12 19:48 ` Kees Cook
2023-07-12 19:57 ` Kent Overstreet
2023-07-12 22:10 ` Darrick J. Wong
2023-07-12 23:57 ` Kent Overstreet
2023-08-09 1:27 ` Linus Torvalds
2023-08-10 15:54 ` Kent Overstreet
2023-08-10 16:40 ` Linus Torvalds
2023-08-10 18:02 ` Kent Overstreet
2023-08-10 18:09 ` Linus Torvalds
2023-08-10 17:52 ` Jan Kara
2023-08-11 2:47 ` Kent Overstreet
2023-08-11 8:10 ` Jan Kara
2023-08-11 8:13 ` Christian Brauner
2023-08-10 22:39 ` Darrick J. Wong
2023-08-10 23:47 ` Linus Torvalds
2023-08-11 2:40 ` Jens Axboe
2023-08-11 4:03 ` Kent Overstreet
2023-08-11 5:20 ` Linus Torvalds
2023-08-11 5:29 ` Kent Overstreet
2023-08-11 5:53 ` Linus Torvalds
2023-08-11 7:52 ` Christian Brauner
2023-08-11 14:31 ` Jens Axboe
2023-08-11 3:45 ` Kent Overstreet
2023-08-21 0:09 ` Dave Chinner
2023-08-10 23:07 ` Matthew Wilcox
2023-08-11 10:54 ` Christian Brauner
2023-08-11 12:58 ` Kent Overstreet
2023-08-14 7:25 ` Christian Brauner
2023-08-14 15:23 ` Kent Overstreet
2023-08-11 13:21 ` Kent Overstreet
2023-08-11 22:56 ` Darrick J. Wong
2023-08-14 7:21 ` Christian Brauner
2023-08-14 15:27 ` Kent Overstreet
2023-09-03 3:25 Kent Overstreet
2023-09-05 13:24 ` Christoph Hellwig
2023-09-06 0:00 ` Kent Overstreet
2023-09-06 0:41 ` Matthew Wilcox
2023-09-06 16:10 ` Kent Overstreet
2023-09-06 17:57 ` Darrick J. Wong
2023-09-08 9:37 ` Christoph Hellwig
2023-09-06 19:36 ` Linus Torvalds
2023-09-06 20:02 ` Linus Torvalds
2023-09-06 20:20 ` Linus Torvalds
2023-09-06 21:55 ` Arnaldo Carvalho de Melo
2023-09-06 23:13 ` David Sterba
2023-09-06 23:34 ` Linus Torvalds
2023-09-06 23:46 ` Arnaldo Carvalho de Melo
2023-09-06 23:53 ` Arnaldo Carvalho de Melo
2023-09-06 23:16 ` Linus Torvalds
2023-09-10 0:53 ` Kent Overstreet
2023-09-07 20:37 ` Kent Overstreet
2023-09-07 20:51 ` Linus Torvalds
2023-09-07 23:40 ` Kent Overstreet
2023-09-08 6:29 ` Martin Steigerwald
2023-09-08 9:11 ` Joshua Ashton
2023-09-06 22:28 ` Nathan Chancellor
2023-09-07 0:03 ` Kees Cook
2023-09-07 14:29 ` Chris Mason
2023-09-07 20:39 ` Kent Overstreet
2023-09-08 10:50 ` Brian Foster
2023-09-08 23:05 ` Dave Chinner
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20230706201510.sh5ukzfsf5vdxvrf@moria.home.lan \
--to=kent.overstreet@linux.dev \
--cc=axboe@kernel.dk \
--cc=brauner@kernel.org \
--cc=hch@lst.de \
--cc=linux-bcachefs@vger.kernel.org \
--cc=linux-fsdevel@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=torvalds@linux-foundation.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
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).