linux-kernel - Re: [PATCH] packet: Don't include <linux/rculist.h>

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] [day] [month] [year] [list]

Message-ID: <3ac4ee1a-b6ad-283d-6747-1b2e15fb27f3@wanadoo.fr>
Date:   Wed, 7 Dec 2022 07:29:19 +0100
From:   Christophe JAILLET <christophe.jaillet@...adoo.fr>
To:     Eric Dumazet <edumazet@...gle.com>
Cc:     "David S. Miller" <davem@...emloft.net>,
        Jakub Kicinski <kuba@...nel.org>,
        Paolo Abeni <pabeni@...hat.com>, linux-kernel@...r.kernel.org,
        kernel-janitors@...r.kernel.org, netdev@...r.kernel.org
Subject: Re: [PATCH] packet: Don't include <linux/rculist.h>

Le 05/12/2022 à 06:24, Eric Dumazet a écrit :
> On Sat, Dec 3, 2022 at 5:49 PM Christophe JAILLET
> <christophe.jaillet@...adoo.fr> wrote:
>>
>> There is no need to include <linux/rculist.h> here.
>>
>> Prefer the less invasive <linux/types.h> which is needed for 'hlist_head'.
>>
>> Signed-off-by: Christophe JAILLET <christophe.jaillet@...adoo.fr>
>> ---
>> Let see if build-bots agree with me!
>>
> 
> net/packet/af_packet.c does not explicitly include linux/rculist.h
> 
> It might be provided by include/linux/netdevice.h, but I wonder if
> this is best practice.

At least, it is not what I expect.

My goal is to avoid some unneeded includes AND the related indirect 
needed includes that are buried somewhere in the dependency hell.

I missed the one in af_packet.c

I'll repost a v2 with the fix for af_packet.c (and double-check if some 
other are also needed)

> 
>> Just declaring 'struct mutex' and 'struct hlist_head' would also be an
>> option.
> 
> I do not get it, see [1]

Just forget about it.

Requirement for:
    struct my_struct {
           struct another_struct            x;

and
    struct my_struct {
           struct another_struct            *x;
                                           ~~~
are not the same, even if 'my_struct' is not used at all...

(*ashamed *)

CJ

> 
>> It would remove the need of any include, but is more likely to break
>> something.
> 
> I do not see why you are even trying this ?
> 
>> ---
>>   include/net/netns/packet.h | 2 +-
>>   1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/include/net/netns/packet.h b/include/net/netns/packet.h
>> index aae69bb43cde..74750865df36 100644
>> --- a/include/net/netns/packet.h
>> +++ b/include/net/netns/packet.h
>> @@ -5,8 +5,8 @@
>>   #ifndef __NETNS_PACKET_H__
>>   #define __NETNS_PACKET_H__
>>
>> -#include <linux/rculist.h>
>>   #include <linux/mutex.h>
>> +#include <linux/types.h>
>>
>>   struct netns_packet {
>>          struct mutex            sklist_lock;
> 
> [1] Definition of 'struct mutex' is definitely needed here.
> 
>> --
>> 2.34.1
>>
>