netdev
[Top] [All Lists]

[PATCH 2/9] decnet: Don't use sk_protinfo + private sock slab cache

To: "David S. Miller" <davem@xxxxxxxxxxxxx>, Patrick Caulfield <patrick@xxxxxxxxxxxxxxx>
Subject: [PATCH 2/9] decnet: Don't use sk_protinfo + private sock slab cache
From: Arnaldo Carvalho de Melo <acme@xxxxxxxxxxxxxxxx>
Date: Thu, 20 Jan 2005 00:04:10 -0200
Cc: linux-decnet-user@xxxxxxxxxxxxxxxxxxxxx, Networking Team <netdev@xxxxxxxxxxx>
Organization: Conectiva S.A.
Sender: netdev-bounce@xxxxxxxxxxx
User-agent: Mozilla Thunderbird 1.0 (X11/20041220)
Hi David, Patrick,

        The log has the reason for this patch.

Regards,

- Arnaldo


===================================================================


ChangeSet@xxxxxx, 2005-01-19 23:02:32-02:00, acme@xxxxxxxxxxxxxxxxxxxxxx
  [DECNET] Don't use sk_protinfo + private sock slab cache
  
  DecNET already uses a private sock slab cache, but initializes 
sk->sk_protinfo,
  pointing to (sk + 1), this is wrong because at sk_free time we call
  sk->sk_destruct, that by default points to sock_def_destruct, that does a 
kfree
  on sk->sk_protinfo, since it was initialized at net_proto_family->create() 
time
  (dn_create), but in decnet sk_protinfo was not kmalloced, it was allocated
  piggybacked to struct sock.
  
  This doesn't causes problems because decnet sets sk->sk_destruct to a custom
  function that doesn't calls kfree(sk->sk_protinfo), but to reach a long time
  goal of killing sk_protinfo lets just make DN_SK return sk + 1.
  
  I left merging dn_scp with dn_sock for later, as the current state suits my
  needs to introduce connection_sock.
  
  Signed-off-by: Arnaldo Carvalho de Melo <acme@xxxxxxxxxxxxxxxx>
  Signed-off-by: David S. Miller <davem@xxxxxxxxxxxxx>


 include/net/dn.h       |    6 +++++-
 net/decnet/af_decnet.c |    3 +--
 2 files changed, 6 insertions(+), 3 deletions(-)


diff -Nru a/include/net/dn.h b/include/net/dn.h
--- a/include/net/dn.h  2005-01-19 23:18:06 -02:00
+++ b/include/net/dn.h  2005-01-19 23:18:06 -02:00
@@ -2,6 +2,7 @@
 #define _NET_DN_H
 
 #include <linux/dn.h>
+#include <net/sock.h>
 #include <asm/byteorder.h>
 
 typedef unsigned short dn_address;
@@ -133,7 +134,10 @@
 
 };
 
-#define DN_SK(__sk) ((struct dn_scp *)(__sk)->sk_protinfo)
+static inline struct dn_scp *DN_SK(struct sock *sk)
+{
+       return (struct dn_scp *)(sk + 1);
+}
 
 /*
  * src,dst : Source and Destination DECnet addresses
diff -Nru a/net/decnet/af_decnet.c b/net/decnet/af_decnet.c
--- a/net/decnet/af_decnet.c    2005-01-19 23:18:06 -02:00
+++ b/net/decnet/af_decnet.c    2005-01-19 23:18:06 -02:00
@@ -456,8 +456,6 @@
        if  (!sk)
                goto out;
 
-       sk->sk_protinfo = scp = (struct dn_scp *)(sk + 1);
-
        if (sock)
                sock->ops = &dn_proto_ops;
        sock_init_data(sock, sk);
@@ -471,6 +469,7 @@
        sk->sk_allocation  = gfp;
 
        /* Initialization of DECnet Session Control Port                */
+       scp = DN_SK(sk);
        scp->state      = DN_O;         /* Open                 */
        scp->numdat     = 1;            /* Next data seg to tx  */
        scp->numoth     = 1;            /* Next oth data to tx  */



<Prev in Thread] Current Thread [Next in Thread>
  • [PATCH 2/9] decnet: Don't use sk_protinfo + private sock slab cache, Arnaldo Carvalho de Melo <=