Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Date:   Mon, 12 Jun 2023 09:45:28 +0200
From:   Christian Brauner <brauner@kernel.org>
To:     wenyang.linux@foxmail.com
Cc:     Alexander Viro <viro@zeniv.linux.org.uk>,
        Jens Axboe <axboe@kernel.dk>, Christoph Hellwig <hch@lst.de>,
        Dylan Yudaken <dylany@fb.com>,
        David Woodhouse <dwmw@amazon.co.uk>,
        Matthew Wilcox <willy@infradead.org>,
        Eric Biggers <ebiggers@google.com>,
        linux-fsdevel@vger.kernel.org, linux-kernel@vger.kernel.org
Subject: Re: [PATCH] eventfd: show flags in fdinfo
Message-ID: <20230612-atomkraftgegner-erziehen-8ee4a0c9c606@brauner>
References: <tencent_59C3AA88A8F1829226E5D3619837FC4A9E09@qq.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Disposition: inline
In-Reply-To: <tencent_59C3AA88A8F1829226E5D3619837FC4A9E09@qq.com>
Precedence: bulk

On Mon, Jun 12, 2023 at 02:59:47AM +0800, wenyang.linux@foxmail.com wrote:
> From: Wen Yang <wenyang.linux@foxmail.com>
> 
> The flags should be displayed in fdinfo, as different flags
> could affect the behavior of eventfd.
> 
> Signed-off-by: Wen Yang <wenyang.linux@foxmail.com>
> Cc: Alexander Viro <viro@zeniv.linux.org.uk>
> Cc: Jens Axboe <axboe@kernel.dk>
> Cc: Christian Brauner <brauner@kernel.org>
> Cc: Christoph Hellwig <hch@lst.de>
> Cc: Dylan Yudaken <dylany@fb.com>
> Cc: David Woodhouse <dwmw@amazon.co.uk>
> Cc: Matthew Wilcox <willy@infradead.org>
> Cc: Eric Biggers <ebiggers@google.com>
> Cc: linux-fsdevel@vger.kernel.org
> Cc: linux-kernel@vger.kernel.org
> ---
>  fs/eventfd.c | 9 +++++----
>  1 file changed, 5 insertions(+), 4 deletions(-)
> 
> diff --git a/fs/eventfd.c b/fs/eventfd.c
> index 6c06a527747f..5b5448e65f6f 100644
> --- a/fs/eventfd.c
> +++ b/fs/eventfd.c
> @@ -33,10 +33,10 @@ struct eventfd_ctx {
>  	/*
>  	 * Every time that a write(2) is performed on an eventfd, the
>  	 * value of the __u64 being written is added to "count" and a
> -	 * wakeup is performed on "wqh". A read(2) will return the "count"
> -	 * value to userspace, and will reset "count" to zero. The kernel
> -	 * side eventfd_signal() also, adds to the "count" counter and
> -	 * issue a wakeup.
> +	 * wakeup is performed on "wqh". If EFD_SEMAPHORE flag was not
> +	 * specified, a read(2) will return the "count" value to userspace,
> +	 * and will reset "count" to zero. The kernel side eventfd_signal()
> +	 * also, adds to the "count" counter and issue a wakeup.
>  	 */
>  	__u64 count;
>  	unsigned int flags;
> @@ -301,6 +301,7 @@ static void eventfd_show_fdinfo(struct seq_file *m, struct file *f)
>  		   (unsigned long long)ctx->count);
>  	spin_unlock_irq(&ctx->wqh.lock);
>  	seq_printf(m, "eventfd-id: %d\n", ctx->id);
> +	seq_printf(m, "eventfd-flags: 0%o\n", ctx->flags);

EFD_CLOEXEC and EFD_NONBLOCK are mapped to generic O_* flags and are
included in fdinfo output already:

pos:    0
flags:  02000002 -> O_CLOEXEC/EFD_CLOEXEC | O_RDWR
mnt_id: 15
ino:    12497
eventfd-count:                0
eventfd-id: 156

So the only thing you really care about is EFD_SEMAPHORE. Since this
changes the type of the eventfd I would just do either:

eventfd-semaphore: {0,1}

or - if we can reasonably expect a third type:

eventfd-type: {0, ..., n}

though I suspect since it hasn't changed since 2.6-something that
eventfd-semaphore: {0,1} is probably fine.


so you only really care about EFD_SEMAPHORE.