lkml.org 
[lkml]   [2011]   [Feb]   [4]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
    /
    Date
    SubjectRe: [PATCH 0/5] net: sysctl: share ipv4/ipv6 sysctl tables
    From
    On Fri, Feb 4, 2011 at 6:37 AM, Lucian Adrian Grijincu
    <lucian.grijincu@gmail.com> wrote:
    > Each network device gets the same 25/24 sysctl entries for ipv4/ipv6
    > in /proc/sys/net/ipv4/conf/DEVNAME and /proc/sys/net/ipv6/conf/DEVNAME
    >
    > Unfortunately, space is wasted holding very much similar data.
    > Fortunately, with some tricks these entries can be shared between all
    > network devices.
    >
    >
    > The single entry in 'struct ctl_table' that was modified at runtime
    > for leaf ctl_table nodes and prevented sharing was 'parent'. This
    > field was first introduces for selinux and then was used to implement
    > sysctl_check_table. Selinux recently removed the need for this field:
    > * http://thread.gmane.org/gmane.linux.kernel.lsm/12623
    > * LKML-Reference: 1296519474-15714-1-git-send-email-lucian.grijincu@gmail.com
    >
    >
    > Remove the need for 'parent' in sysctl_check_table and remove the
    > 'parent' field:
    >
    >  [PATCH 1/5] sysctl: faster reimplementation of sysctl_check_table
    >  [PATCH 2/5] sysctl: remove useless ctl_table->parent field
    >
    >
    > Pave the way for sharing of ipv4/6 tables: allow data to be stored in
    > the nodes above the leafs that will be shared:
    >
    >  [PATCH 3/5] sysctl: write ctl_table->extra2 to entries created from ctl_path
    >
    >
    > Finally share the leaf sysctl tables for ipv4/ipv6:
    >
    >  [PATCH 4/5] ipv4: share sysctl net/ipv4/conf/DEVNAME/ tables
    >  [PATCH 5/5] ipv6: share sysctl net/ipv6/conf/DEVNAME/ tables

    Meh.

    First you remove ->parent, then heroically pass "struct file *"
    to sysctl handlers which duplicates all information already passed
    and brings dcache into picture.

    Binary sysctl rewrite confused you into thinking that d_name.name
    is the way, but it isn't.
    For binary sysctl(2) you wouldn't get d_name.name.
    --
    To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
    the body of a message to majordomo@vger.kernel.org
    More majordomo info at http://vger.kernel.org/majordomo-info.html
    Please read the FAQ at http://www.tux.org/lkml/

    \
     
     \ /
      Last update: 2011-02-04 11:53    [W:0.031 / U:155.684 seconds]
    ©2003-2016 Jasper Spaans. hosted at Digital OceanAdvertise on this site