[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <35d71280-380c-bcda-9e82-32bd8e6cb631@loongson.cn>
Date: Tue, 15 Aug 2023 14:01:33 +0800
From: Hongchen Zhang <zhanghongchen@...ngson.cn>
To: David Howells <dhowells@...hat.com>
Cc: Alexander Viro <viro@...iv.linux.org.uk>,
Christian Brauner <brauner@...nel.org>,
Steve French <stfrench@...rosoft.com>,
Jens Axboe <axboe@...nel.dk>, David Disseldorp <ddiss@...e.de>,
Luis Chamberlain <mcgrof@...nel.org>,
Nick Alcock <nick.alcock@...cle.com>,
linux-fsdevel@...r.kernel.org, linux-kernel@...r.kernel.org,
loongson-kernel@...ts.loongnix.cn
Subject: Re: [PATCH v5 2/2] pipe: use __pipe_{lock,unlock} instead of spinlock
On 2023/8/14 下午4:47, David Howells wrote:
> Hongchen Zhang <zhanghongchen@...ngson.cn> wrote:
>
>> - spin_lock_irq(&pipe->rd_wait.lock);
>> + __pipe_lock(pipe);
>
> This mustn't sleep. post_one_notification() needs to be callable with a
> spinlock held.
Hi David,
Where is the usage scenario for post_one_notification, or what usage
scenarios will it be used in the future?
IMO, it is not advisable for a debugging tool to affect the performance
of an syscall.
>
> David
>
--
Best Regards
Hongchen Zhang
Powered by blists - more mailing lists