[Devel] Re: [PATCH] Make common helpers for seq_files that work with list_head-s

Andrew Morton akpm at linux-foundation.org
Thu May 17 10:36:55 PDT 2007


On Thu, 17 May 2007 19:19:40 +0400 Pavel Emelianov <xemul at sw.ru> wrote:

> Many places in kernel use seq_file API to iterate over a regular
> list_head. The code for such iteration is identical in all the
> places, so it's worth introducing a common helpers.
> 
> This makes code more than 300 lines smaller.
> 
> Cc-ed are people, who maintain the code altered by the patch.
> 
> Signed-off-by: Pavel Emelianov <xemul at openvz.org>
> 
>  block/genhd.c                                         |   40 +++-----
>  crypto/proc.c                                         |   17 ---
>  drivers/char/misc.c                                   |   18 ----
>  drivers/input/input.c                                 |   29 ------
>  drivers/isdn/capi/kcapi_proc.c                        |   28 ------
>  fs/afs/proc.c                                         |   81 ++----------------
>  fs/namespace.c                                        |   14 ---
>  fs/nfs/client.c                                       |   54 +-----------
>  fs/proc/proc_tty.c                                    |   15 ---
>  include/linux/seq_file.h                              |   35 +++++++
>  kernel/module.c                                       |   17 ---
>  mm/slab.c                                             |   28 +-----
>  net/atm/br2684.c                                      |   22 ----
>  net/core/sock.c                                       |   39 --------
>  net/ipv4/netfilter/nf_conntrack_l3proto_ipv4_compat.c |   27 ------
>  net/netfilter/nf_conntrack_expect.c                   |   27 ------
>  net/rxrpc/ar-proc.c                                   |   48 +---------
>  17 files changed, 116 insertions(+), 423 deletions(-)

Can't complain about the diffstat.  Please experiment with uninlining
seq_list_start(), see if that reduces overall text size.

And as Dmitry indicated, it would be less disruptive if we could have the
one core patch then a stream of per-subsystem patches, please.




More information about the Devel mailing list