netdev - Re: [PATCH 1/1] net/ipv4/inet_fragment: Batch fqdir destroy works

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Wed, 9 Dec 2020 15:16:59 -0800
From:   Jakub Kicinski <kuba@...nel.org>
To:     SeongJae Park <sjpark@...zon.com>
Cc:     <davem@...emloft.net>, SeongJae Park <sjpark@...zon.de>,
        <kuznet@....inr.ac.ru>, <paulmck@...nel.org>,
        <netdev@...r.kernel.org>, <rcu@...r.kernel.org>,
        <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH 1/1] net/ipv4/inet_fragment: Batch fqdir destroy works

On Tue, 8 Dec 2020 10:45:29 +0100 SeongJae Park wrote:
> From: SeongJae Park <sjpark@...zon.de>
> 
> In 'fqdir_exit()', a work for destruction of the 'fqdir' is enqueued.
> The work function, 'fqdir_work_fn()', calls 'rcu_barrier()'.  In case of
> intensive 'fqdir_exit()' (e.g., frequent 'unshare(CLONE_NEWNET)'
> systemcalls), this increased contention could result in unacceptably
> high latency of 'rcu_barrier()'.  This commit avoids such contention by
> doing the destruction in batched manner, as similar to that of
> 'cleanup_net()'.
> 
> Signed-off-by: SeongJae Park <sjpark@...zon.de>

Looks fine to me, but you haven't CCed Florian or Eric who where the
last two people to touch this function. Please repost CCing them and
fixing the nit below, thanks!

>  static void fqdir_work_fn(struct work_struct *work)
>  {
> -	struct fqdir *fqdir = container_of(work, struct fqdir, destroy_work);
> -	struct inet_frags *f = fqdir->f;
> +	struct llist_node *kill_list;
> +	struct fqdir *fqdir;
> +	struct inet_frags *f;

nit: reorder fqdir and f to keep reverse xmas tree variable ordering.