All of lore.kernel.org
 help / color / mirror / Atom feed
From: Linus Torvalds <torvalds@linux-foundation.org>
To: Steven Rostedt <rostedt@goodmis.org>
Cc: Masami Hiramatsu <mhiramat@kernel.org>,
	Mathieu Desnoyers <mathieu.desnoyers@efficios.com>,
	 LKML <linux-kernel@vger.kernel.org>,
	 Linux Trace Devel <linux-trace-devel@vger.kernel.org>,
	Christian Brauner <brauner@kernel.org>,
	 Ajay Kaher <ajay.kaher@broadcom.com>,
	Geert Uytterhoeven <geert@linux-m68k.org>,
	 linux-fsdevel <linux-fsdevel@vger.kernel.org>
Subject: Re: [PATCH] eventfs: Have inodes have unique inode numbers
Date: Sun, 28 Jan 2024 14:17:43 -0800	[thread overview]
Message-ID: <CAHk-=wg0bfqL9Yn-KnamLTvTpw+zbAa=og_JRPjZHgJ5m9iCZA@mail.gmail.com> (raw)
In-Reply-To: <20240128170125.7d51aa8f@rorschach.local.home>

On Sun, 28 Jan 2024 at 14:01, Steven Rostedt <rostedt@goodmis.org> wrote:
>
> Basically you are saying that when the ei is created, it should have a
> ref count of 1. If the lookup happens and does the
> atomic_inc_not_zero() it will only increment if the ref count is not 0
> (which is basically equivalent to is_freed).

Exactly.

That's what we do with almost all our data structures.

You can use the kref() infrastructure to give this a bit more
structure, and avoid doing the atomics by hand. So then "get a ref" is
literally

    kref_get(&ei->kref);

and releasing it is

    kref_put(&ei->kref, ei_release);

so you don't have the write out that "if (atomic_dec_and_test(..) kfree();".

And "ei_release()" looks just something like this:

    static void ei_release(struct kref *ref)
    {
        kfree(container_of(ref, struct eventfs_inode, kref);
    }

and that's literally all you need (ok, you need to add the 'kref' to
the eventfs_inode, and initialize it at allocation time, of course).

> And then we can have deletion of the object happen in both where the
> caller (kprobes) deletes the directory and in the final iput()
> reference (can I use iput and not the d_release()?), that it does the
> same as well.
>
> Where whatever sees the refcount of zero calls rcu_free?

Having looked more at your code, I actually don't see you even needing
rcu_free().

It's not that you don't need SRCU (which is *much* heavier than RCU) -
you don't even need the regular quick RCU at all.

As far as I can see, you do all the lookups under eventfs_mutex, so
there are actually no RCU users.

And the VFS code (that obviously uses RCU internally) has a ref to it
and just a direct pointer, so again, there's no real RCU involved.

You seem to have used SRCU as a "I don't want to do refcounts" thing.
I bet you'll notice that it clarifies things *enormously* to just use
refcounts.

                Linus

  reply	other threads:[~2024-01-28 22:18 UTC|newest]

Thread overview: 52+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-01-26 20:02 [PATCH] eventfs: Have inodes have unique inode numbers Steven Rostedt
2024-01-26 20:25 ` Linus Torvalds
2024-01-26 21:26   ` Steven Rostedt
2024-01-26 21:31     ` Linus Torvalds
2024-01-26 21:43       ` Steven Rostedt
2024-01-26 21:36     ` Linus Torvalds
2024-01-26 21:42       ` Steven Rostedt
2024-01-26 21:49       ` Linus Torvalds
2024-01-26 22:08         ` Steven Rostedt
2024-01-26 22:26           ` Linus Torvalds
2024-01-27 14:47             ` Steven Rostedt
2024-01-28 14:42               ` Steven Rostedt
2024-01-26 22:14         ` Mathieu Desnoyers
2024-01-26 22:29           ` Linus Torvalds
2024-01-26 22:41             ` Mathieu Desnoyers
2024-01-26 22:49               ` Linus Torvalds
2024-01-29 16:00                 ` Mathieu Desnoyers
2024-01-29 18:58                   ` Linus Torvalds
2024-01-26 22:34           ` Matthew Wilcox
2024-01-26 22:40             ` Mathieu Desnoyers
2024-01-26 22:48             ` Linus Torvalds
2024-01-26 23:04               ` Matthew Wilcox
2024-01-26 23:11                 ` Linus Torvalds
2024-01-26 23:17                   ` Linus Torvalds
2024-01-27  9:36                     ` Andreas Schwab
2024-01-27 21:47         ` Linus Torvalds
2024-01-28 20:15           ` Steven Rostedt
2024-01-28 20:53             ` Linus Torvalds
2024-01-28 21:08               ` Linus Torvalds
2024-01-28 22:01                 ` Steven Rostedt
2024-01-28 22:17                   ` Linus Torvalds [this message]
2024-01-28 22:26                     ` Steven Rostedt
2024-01-28 21:11               ` Steven Rostedt
2024-01-28 21:19               ` Steven Rostedt
2024-01-28 21:43                 ` Linus Torvalds
2024-01-28 22:07                   ` Linus Torvalds
2024-01-28 22:17                     ` Steven Rostedt
2024-01-28 22:25                       ` Linus Torvalds
2024-01-28 22:51           ` Steven Rostedt
2024-01-28 23:24             ` Linus Torvalds
2024-01-28 23:59               ` Steven Rostedt
2024-01-29  0:21                 ` Steven Rostedt
2024-01-29  1:00                   ` Linus Torvalds
2024-01-29  1:42                     ` Linus Torvalds
2024-01-29  2:32                       ` Steven Rostedt
2024-01-29  3:40                         ` Steven Rostedt
2024-01-29  4:01                           ` Linus Torvalds
2024-01-29  2:09                     ` Steven Rostedt
2024-01-29  6:44                       ` Amir Goldstein
2024-01-29  9:32                         ` Steven Rostedt
2024-01-27 15:26       ` David Laight
2024-01-27 20:01         ` Linus Torvalds

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='CAHk-=wg0bfqL9Yn-KnamLTvTpw+zbAa=og_JRPjZHgJ5m9iCZA@mail.gmail.com' \
    --to=torvalds@linux-foundation.org \
    --cc=ajay.kaher@broadcom.com \
    --cc=brauner@kernel.org \
    --cc=geert@linux-m68k.org \
    --cc=linux-fsdevel@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-trace-devel@vger.kernel.org \
    --cc=mathieu.desnoyers@efficios.com \
    --cc=mhiramat@kernel.org \
    --cc=rostedt@goodmis.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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.