netdev - Re: [RFC][PATCH] rcu: Use typeof(p) instead of typeof(*p) *

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 linux-cve-announce 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]

Message-ID: <639278914.2878.1633457192964.JavaMail.zimbra@efficios.com>
Date:   Tue, 5 Oct 2021 14:06:32 -0400 (EDT)
From:   Mathieu Desnoyers <mathieu.desnoyers@...icios.com>
To:     Rasmus Villemoes <linux@...musvillemoes.dk>
Cc:     rostedt <rostedt@...dmis.org>,
        linux-kernel <linux-kernel@...r.kernel.org>,
        Linus Torvalds <torvalds@...ux-foundation.org>,
        Paul <paulmck@...ux.vnet.ibm.com>,
        Josh Triplett <josh@...htriplett.org>,
        Lai Jiangshan <jiangshanlai@...il.com>,
        "Joel Fernandes, Google" <joel@...lfernandes.org>,
        Pablo Neira Ayuso <pablo@...filter.org>,
        Jozsef Kadlecsik <kadlec@...filter.org>,
        Florian Westphal <fw@...len.de>,
        "David S. Miller" <davem@...emloft.net>,
        Hideaki YOSHIFUJI <yoshfuji@...ux-ipv6.org>,
        David Ahern <dsahern@...nel.org>,
        Jakub Kicinski <kuba@...nel.org>, rcu <rcu@...r.kernel.org>,
        netfilter-devel <netfilter-devel@...r.kernel.org>,
        coreteam <coreteam@...filter.org>,
        netdev <netdev@...r.kernel.org>
Subject: Re: [RFC][PATCH] rcu: Use typeof(p) instead of typeof(*p) *

----- On Oct 5, 2021, at 2:01 PM, Rasmus Villemoes linux@...musvillemoes.dk wrote:

> On 05/10/2021 15.47, Steven Rostedt wrote:
> 
>> That is, instead of declaring: typeof(*p) *_p; just do:
>>  typeof(p) _p;
>> 
>> Also had to update a lot of the function pointer initialization in the
>> networking code, as a function address must be passed as an argument in
>> RCU_INIT_POINTER()
> 
> I would think that one could avoid that churn by saying
> 
>  typeof((p) + 0)
> 
> instead of just "typeof(p)", to force the decay to a pointer.

If the type of @p is an integer, (p) + 0 is still valid, so it will not
prevent users from passing an integer type as argument, which is what
the current implementation prevents.

Also, AFAIU, the compiler wants to know the sizeof(p) in order to evaluate
(p + 0). Steven's goal is to hide the structure declaration, so that would
not work either.

Thanks,

Mathieu

-- 
Mathieu Desnoyers
EfficiOS Inc.
http://www.efficios.com