Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Subject: Re: [PATCH] kernel/groups.c: use bsearch library function
To:     Thomas Meyer <thomas@m3y3r.de>, linux-kernel@vger.kernel.org
References: <20191007192632.29535-1-thomas@m3y3r.de>
From:   Rasmus Villemoes <linux@rasmusvillemoes.dk>
Message-ID: <60e43953-a7f9-c52e-150c-74059d1b377b@rasmusvillemoes.dk>
Date:   Wed, 9 Oct 2019 09:46:21 +0200
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:60.0) Gecko/20100101
 Thunderbird/60.8.0
MIME-Version: 1.0
In-Reply-To: <20191007192632.29535-1-thomas@m3y3r.de>
Content-Type: text/plain; charset=windows-1252
Content-Language: en-US
Content-Transfer-Encoding: 7bit
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On 07/10/2019 21.26, Thomas Meyer wrote:
> commit b7b2562f7252 ("kernel/groups.c: use sort library function")
> introduced the sort library function.
> also use the bsearch library function instead of open-coding the binary
> search.

Yes, but please note the difference between sorting the group_info and
searching it: The former is done quite rarely - the setgroups syscall is
used roughly once per login-session.

But the searching of that structure is done more or less every time a
user accesses a file not owned by that user (e.g., any time a normal
user accesses anything in /usr) - at least if I'm reading
acl_permission_check() right.

So using a callback-based interface, especially in a post-spectre world,
may have a somewhat large performance impact.

Rasmus