netdev - Re: [PATCH net-next v2 02/11] sock: Make sk

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: <03207138-4067-d3dc-c904-e9ff4a8ed197@gmail.com>
Date:   Mon, 16 Dec 2019 18:25:55 -0800
From:   Eric Dumazet <eric.dumazet@...il.com>
To:     Mat Martineau <mathew.j.martineau@...ux.intel.com>,
        netdev@...r.kernel.org, mptcp@...ts.01.org
Cc:     Paolo Abeni <pabeni@...hat.com>
Subject: Re: [PATCH net-next v2 02/11] sock: Make sk_protocol a 16-bit value



On 12/16/19 4:24 PM, Mat Martineau wrote:
> Match the 16-bit width of skbuff->protocol. Fills an 8-bit hole so
> sizeof(struct sock) does not change.
> 
> Co-developed-by: Paolo Abeni <pabeni@...hat.com>
> Signed-off-by: Paolo Abeni <pabeni@...hat.com>
> Signed-off-by: Mat Martineau <mathew.j.martineau@...ux.intel.com>
> ---
> 
> Changes for v2:
> 
> Moved sk_pacing_shift back to a regular struct member, adjacent to the
> bitfield. gcc then packs the bitfield and the u8 in a single 32-bit word.
> 
> 
>  include/net/sock.h          | 4 ++--
>  include/trace/events/sock.h | 2 +-
>  2 files changed, 3 insertions(+), 3 deletions(-)
> 
> diff --git a/include/net/sock.h b/include/net/sock.h
> index 81dc811aad2e..0930f46c600c 100644
> --- a/include/net/sock.h
> +++ b/include/net/sock.h
> @@ -456,10 +456,10 @@ struct sock {
>  				sk_no_check_tx : 1,
>  				sk_no_check_rx : 1,
>  				sk_userlocks : 4,
> -				sk_protocol  : 8,
>  				sk_type      : 16;
> -	u16			sk_gso_max_segs;
>  	u8			sk_pacing_shift;
> +	u16			sk_protocol;
> +	u16			sk_gso_max_segs;
>  	unsigned long	        sk_lingertime;
>  	struct proto		*sk_prot_creator;
>  	rwlock_t		sk_callback_lock;


sk_type is no longer at a 16bit (u16) aligned location, which might force the compiler
to do extra operations to fetch sk->sk_type on some arches.

Note that I do not even know why sk_type is a 16bit field, we might be able to convert it to 8bit.