From: Christian Brauner <brauner@kernel.org>
To: NeilBrown <neilb@suse.de>
Cc: Linus Torvalds <torvalds@linux-foundation.org>,
Peter Zijlstra <peterz@infradead.org>,
Ingo Molnar <mingo@redhat.com>, Jeff Layton <jlayton@kernel.org>,
Jan Kara <jack@suse.cz>,
linux-fsdevel@vger.kernel.org
Subject: Re: [PATCH RFC v2 1/6] fs: add i_state helpers
Date: Thu, 22 Aug 2024 10:27:17 +0200 [thread overview]
Message-ID: <20240822-knipsen-bebildert-b6f94efcb429@brauner> (raw)
In-Reply-To: <172427833589.6062.8614016543522604940@noble.neil.brown.name>
On Thu, Aug 22, 2024 at 08:12:15AM GMT, NeilBrown wrote:
> On Thu, 22 Aug 2024, Christian Brauner wrote:
> > The i_state member is an unsigned long so that it can be used with the
> > wait bit infrastructure which expects unsigned long. This wastes 4 bytes
> > which we're unlikely to ever use. Switch to using the var event wait
> > mechanism using the address of the bit. Thanks to Linus for the address
> > idea.
> >
> > Signed-off-by: Christian Brauner <brauner@kernel.org>
> > ---
> > fs/inode.c | 10 ++++++++++
> > include/linux/fs.h | 16 ++++++++++++++++
> > 2 files changed, 26 insertions(+)
> >
> > diff --git a/fs/inode.c b/fs/inode.c
> > index 154f8689457f..f2a2f6351ec3 100644
> > --- a/fs/inode.c
> > +++ b/fs/inode.c
> > @@ -472,6 +472,16 @@ static void __inode_add_lru(struct inode *inode, bool rotate)
> > inode->i_state |= I_REFERENCED;
> > }
> >
> > +struct wait_queue_head *inode_bit_waitqueue(struct wait_bit_queue_entry *wqe,
> > + struct inode *inode, u32 bit)
> > +{
> > + void *bit_address;
> > +
> > + bit_address = inode_state_wait_address(inode, bit);
> > + init_wait_var_entry(wqe, bit_address, 0);
> > + return __var_waitqueue(bit_address);
> > +}
> > +
> > /*
> > * Add inode to LRU if needed (inode is unused and clean).
> > *
> > diff --git a/include/linux/fs.h b/include/linux/fs.h
> > index 23e7d46b818a..a5b036714d74 100644
> > --- a/include/linux/fs.h
> > +++ b/include/linux/fs.h
> > @@ -744,6 +744,22 @@ struct inode {
> > void *i_private; /* fs or device private pointer */
> > } __randomize_layout;
> >
> > +/*
> > + * Get bit address from inode->i_state to use with wait_var_event()
> > + * infrastructre.
> > + */
> > +#define inode_state_wait_address(inode, bit) ((char *)&(inode)->i_state + (bit))
> > +
> > +struct wait_queue_head *inode_bit_waitqueue(struct wait_bit_queue_entry *wqe,
> > + struct inode *inode, u32 bit);
> > +
> > +static inline void inode_wake_up_bit(struct inode *inode, u32 bit)
> > +{
> > + /* Ensure @bit will be seen cleared/set when caller is woken up. */
>
> The above comment is wrong. I think I once thought it was correct too
> but now I know better (I hope).
> A better comment might be
> /* Insert memory barrier as recommended by wake_up_var() */
> but even that is unnecessary as we don't need the memory barrier.
>
> A careful reading of memory-barriers.rst shows that *when the process is
> actually woken* there are sufficient barriers in wake_up_process() and
> prepare_wait_event() and the scheduler and (particularly)
> set_current_state() so that a value set before the wake_up is seen after
> the schedule().
>
> So this barrier isn't about the bit. This barrier is about the
> wait_queue. In particular it is about waitqueue_active() call at the
> start of wake_up_var(). If that test wasn't there and if instead
> wake_up_var() conditionally called __wake_up(), then there would be no
Did you mean "unconditionally called"?
> need for any barrier. A comment near wake_up_bit() makes it clear that
> the barrier is only needed when the spinlock is avoided.
>
> On a weakly ordered arch, this test can be effected *before* the write
> of the bit. If the waiter adds itself to the wait queue and then tests
> the bit before the bit is set, but after the waitqueue_active() test is
> put in effect, then the wake_up will never be sent.
>
> But .... this is all academic of this code because you don't need a
> barrier at all. The wake_up happens in a spin_locked region, and the
> wait is entirely inside the same spin_lock, except for the schedule. A
> later patch has:
> spin_unlock(); schedule(); spin_lock();
>
> So there is no room for a race. If the bit is cleared before the
> wait_var_event() equivalent, then no wakeup is needed. When the lock is
> dropped after the bit is cleared the unlock will have all the barriers
> needed for the bit to be visible.
>
> The only other place that the bit can be cleared is concurrent with the
> above schedule() while the spinlock isn't held by the waiter. In that
> case it is again clear that no barrier is needed - or that the
> spin_unlock/lock provide all the needed barriers.
>
> So a better comment would be
>
> /* no barrier needs as both waker and waiter are in spin-locked regions */
Thanks for the analysis. I was under the impression that wait_on_inode()
was called in contexts where no barrier is guaranteed and the bit isn't
checked with spin_lock() held.
next prev parent reply other threads:[~2024-08-22 8:27 UTC|newest]
Thread overview: 49+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-08-21 15:47 [PATCH RFC v2 0/6] inode: turn i_state into u32 Christian Brauner
2024-08-21 15:47 ` [PATCH RFC v2 1/6] fs: add i_state helpers Christian Brauner
2024-08-21 22:12 ` NeilBrown
2024-08-21 22:47 ` Linus Torvalds
2024-08-21 23:34 ` Dave Chinner
2024-08-23 0:08 ` NeilBrown
2024-08-22 8:27 ` Christian Brauner [this message]
2024-08-22 8:37 ` Linus Torvalds
2024-08-23 0:14 ` NeilBrown
2024-08-23 2:52 ` Linus Torvalds
2024-08-23 3:05 ` Linus Torvalds
2024-08-23 3:44 ` Linus Torvalds
2024-08-23 5:01 ` NeilBrown
2024-08-23 12:47 ` [PATCH v3 0/6] inode: turn i_state into u32 Christian Brauner
2024-08-23 12:47 ` [PATCH v3 1/6] fs: add i_state helpers Christian Brauner
2024-09-05 16:01 ` Jan Kara
2024-08-23 12:47 ` [PATCH v3 2/6] fs: reorder i_state bits Christian Brauner
2024-09-05 16:02 ` Jan Kara
2024-08-23 12:47 ` [PATCH v3 3/6] inode: port __I_SYNC to var event Christian Brauner
2024-09-05 16:02 ` Jan Kara
2024-08-23 12:47 ` [PATCH v3 4/6] inode: port __I_NEW " Christian Brauner
2024-09-06 13:30 ` Jan Kara
2024-08-23 12:47 ` [PATCH v3 5/6] inode: port __I_LRU_ISOLATING " Christian Brauner
2024-09-09 7:35 ` Jan Kara
2024-08-23 12:47 ` [PATCH v3 6/6] inode: make i_state a u32 Christian Brauner
2024-09-09 7:35 ` Jan Kara
2024-08-23 15:06 ` [PATCH v3 0/6] inode: turn i_state into u32 Josef Bacik
2024-08-21 22:28 ` [PATCH RFC v2 1/6] fs: add i_state helpers Dave Chinner
2024-08-21 22:53 ` Linus Torvalds
2024-08-21 15:47 ` [PATCH RFC v2 2/6] fs: reorder i_state bits Christian Brauner
2024-08-21 15:47 ` [PATCH RFC v2 3/6] writeback: port __I_SYNC to var event Christian Brauner
2024-08-21 15:47 ` [PATCH RFC v2 4/6] inode: port __I_NEW " Christian Brauner
2024-08-23 0:31 ` NeilBrown
2024-08-23 8:20 ` Christian Brauner
2024-08-23 11:07 ` Christian Brauner
2024-08-21 15:47 ` [PATCH RFC v2 5/6] inode: port __I_LRU_ISOLATING " Christian Brauner
2024-08-21 19:41 ` Jeff Layton
2024-08-22 8:53 ` Christian Brauner
2024-08-22 9:48 ` Mateusz Guzik
2024-08-22 11:10 ` Christian Brauner
2024-08-22 12:46 ` Mateusz Guzik
2024-08-23 0:36 ` NeilBrown
2024-08-23 2:24 ` Linus Torvalds
2024-08-23 8:29 ` Christian Brauner
2024-08-21 15:47 ` [PATCH RFC v2 6/6] inode: make i_state a u32 Christian Brauner
2024-08-21 21:03 ` Andreas Dilger
2024-08-22 8:31 ` Christian Brauner
2024-08-21 19:15 ` [PATCH RFC v2 0/6] inode: turn i_state into u32 Josef Bacik
2024-08-21 19:42 ` Jeff Layton
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=20240822-knipsen-bebildert-b6f94efcb429@brauner \
--to=brauner@kernel.org \
--cc=jack@suse.cz \
--cc=jlayton@kernel.org \
--cc=linux-fsdevel@vger.kernel.org \
--cc=mingo@redhat.com \
--cc=neilb@suse.de \
--cc=peterz@infradead.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