[patch 0/1][RFC] add a private field to the sock structure

[patch 0/1][RFC] add a private field to the sock structure

[dlezcano]

When a socket is created it is sometime useful to store a specific information
for this socket.
 
This information can be for examples: 
	* a creation time
	* a pid
	* a uid/gid
	* a container identifier
	* a pointer to a more specific structure
	* ...

The following patch is a proposition to add a private anonymous pointer
field to the common part of the sock structure.

-- 

[patch 1/1][RFC] add a private field to the sock structure

[dlezcano]

[-- Attachment #1: sock-private-data.patch --]
[-- Type: text/plain, Size: 2443 bytes --]

From: Daniel Lezcano <dlezcano@fr.ibm.com>

Store private information for a socket

This patch adds a field to the common socket structure. This field
is a anonymous pointer which allow to store an information about
the socket

Signed-off-by: Daniel Lezcano <dlezcano@fr.ibm.com>

---
 include/net/inet_timewait_sock.h |    1 +
 include/net/sock.h               |    3 +++
 net/ipv4/inet_timewait_sock.c    |    1 +
 3 files changed, 5 insertions(+)

Index: net-2.6.24-bf/include/net/sock.h
===================================================================
--- net-2.6.24-bf.orig/include/net/sock.h
+++ net-2.6.24-bf/include/net/sock.h
@@ -106,6 +106,7 @@
  *	@skc_refcnt: reference count
  *	@skc_hash: hash value used with various protocol lookup tables
  *	@skc_prot: protocol handlers inside a network family
+ *      @skc_private: field used to store private data
  *
  *	This is the minimal network layer representation of sockets, the header
  *	for struct sock and struct inet_timewait_sock.
@@ -120,6 +121,7 @@
 	atomic_t		skc_refcnt;
 	unsigned int		skc_hash;
 	struct proto		*skc_prot;
+	void                    *skc_private;
 };
 
 /**
@@ -196,6 +198,7 @@
 #define sk_refcnt		__sk_common.skc_refcnt
 #define sk_hash			__sk_common.skc_hash
 #define sk_prot			__sk_common.skc_prot
+#define sk_private		__sk_common.skc_private
 	unsigned char		sk_shutdown : 2,
 				sk_no_check : 2,
 				sk_userlocks : 4;
Index: net-2.6.24-bf/net/ipv4/inet_timewait_sock.c
===================================================================
--- net-2.6.24-bf.orig/net/ipv4/inet_timewait_sock.c
+++ net-2.6.24-bf/net/ipv4/inet_timewait_sock.c
@@ -108,6 +108,7 @@
 		tw->tw_hash	    = sk->sk_hash;
 		tw->tw_ipv6only	    = 0;
 		tw->tw_prot	    = sk->sk_prot_creator;
+		tw->tw_private      = sk->sk_private;
 		atomic_set(&tw->tw_refcnt, 1);
 		inet_twsk_dead_node_init(tw);
 		__module_get(tw->tw_prot->owner);
Index: net-2.6.24-bf/include/net/inet_timewait_sock.h
===================================================================
--- net-2.6.24-bf.orig/include/net/inet_timewait_sock.h
+++ net-2.6.24-bf/include/net/inet_timewait_sock.h
@@ -115,6 +115,7 @@
 #define tw_refcnt		__tw_common.skc_refcnt
 #define tw_hash			__tw_common.skc_hash
 #define tw_prot			__tw_common.skc_prot
+#define tw_private              __tw_common.skc_private
 	volatile unsigned char	tw_substate;
 	/* 3 bits hole, try to pack */
 	unsigned char		tw_rcv_wscale;

-- 

Re: [patch 1/1][RFC] add a private field to the sock structure

[Christoph Hellwig]

On Wed, Aug 29, 2007 at 06:41:15PM +0200, dlezcano@fr.ibm.com wrote:
> From: Daniel Lezcano <dlezcano@fr.ibm.com>
> 
> Store private information for a socket
> 
> This patch adds a field to the common socket structure. This field
> is a anonymous pointer which allow to store an information about
> the socket

If you have orivtae information it should have a typed pointer to it.
Please also post the patches actually using this field.

Re: [patch 0/1][RFC] add a private field to the sock structure

[David Miller]

From: dlezcano@fr.ibm.com
Date: Wed, 29 Aug 2007 18:41:14 +0200

> When a socket is created it is sometime useful to store a specific information
> for this socket.
>  
> This information can be for examples: 
> 	* a creation time
> 	* a pid
> 	* a uid/gid
> 	* a container identifier
> 	* a pointer to a more specific structure
> 	* ...
> 
> The following patch is a proposition to add a private anonymous pointer
> field to the common part of the sock structure.

We got rid of the private field a long time ago because not only
is it not needed, it tends to get abused.

Find another way to implement your feature.

Re: [patch 0/1][RFC] add a private field to the sock structure

[Arnaldo Carvalho de Melo]

Em Wed, Aug 29, 2007 at 11:23:16AM -0700, David Miller escreveu:
> From: dlezcano@fr.ibm.com
> Date: Wed, 29 Aug 2007 18:41:14 +0200
> 
> > When a socket is created it is sometime useful to store a specific information
> > for this socket.
> >  
> > This information can be for examples: 
> > 	* a creation time
> > 	* a pid
> > 	* a uid/gid
> > 	* a container identifier
> > 	* a pointer to a more specific structure
> > 	* ...
> > 
> > The following patch is a proposition to add a private anonymous pointer
> > field to the common part of the sock structure.
> 
> We got rid of the private field a long time ago because not only
> is it not needed, it tends to get abused.

Nah, it still there, sk_protinfo, its just ax25 that uses it
(nudge(Ralf)). How do we state that a struct field is deprecated and
will be removed soon(tm)?

There is another, sk_user_data, that is used only by the rpc, pppol2tp,
iscsi, ncpfs, smbfs and dlm guys, see? We need a deprecate flag for
sk_protinfo, if not people will use it! :-)

- Arnaldo

Re: [patch 0/1][RFC] add a private field to the sock structure

[Daniel Lezcano]

Arnaldo Carvalho de Melo wrote:
> Em Wed, Aug 29, 2007 at 11:23:16AM -0700, David Miller escreveu:
>> From: dlezcano@fr.ibm.com
>> Date: Wed, 29 Aug 2007 18:41:14 +0200
>>
>>> When a socket is created it is sometime useful to store a specific information
>>> for this socket.
>>>  
>>> This information can be for examples: 
>>> 	* a creation time
>>> 	* a pid
>>> 	* a uid/gid
>>> 	* a container identifier
>>> 	* a pointer to a more specific structure
>>> 	* ...
>>>
>>> The following patch is a proposition to add a private anonymous pointer
>>> field to the common part of the sock structure.
>> We got rid of the private field a long time ago because not only
>> is it not needed, it tends to get abused.
> 
> Nah, it still there, sk_protinfo, its just ax25 that uses it
> (nudge(Ralf)). How do we state that a struct field is deprecated and
> will be removed soon(tm)?
> 
> There is another, sk_user_data, that is used only by the rpc, pppol2tp,
> iscsi, ncpfs, smbfs and dlm guys, see? We need a deprecate flag for
> sk_protinfo, if not people will use it! :-)

Thank you for your feedbacks. I will play with the sk_user_data.

Is there a chance to have this field in the timewait sockets ? :)

Regards
	--Daniel

Re: [patch 0/1][RFC] add a private field to the sock structure

[David Miller]

From: Arnaldo Carvalho de Melo <acme@ghostprotocols.net>
Date: Wed, 29 Aug 2007 16:18:31 -0300

> Nah, it still there, sk_protinfo, its just ax25 that uses it
> (nudge(Ralf)). How do we state that a struct field is deprecated and
> will be removed soon(tm)?

You get rid of all the in-tree users and then just kill it :)

More seriously, I don't think we need a deprecation schedule
for a struct member, either it's not used in-tree anymore or
it isn't.  We'll go crazy with any other policy and it's
difficult enough shrinking these things :-)

Re: [patch 0/1][RFC] add a private field to the sock structure

[Arnaldo Carvalho de Melo]

Em Thu, Aug 30, 2007 at 10:40:49PM -0700, David Miller escreveu:
> From: Arnaldo Carvalho de Melo <acme@ghostprotocols.net>
> Date: Wed, 29 Aug 2007 16:18:31 -0300
> 
> > Nah, it still there, sk_protinfo, its just ax25 that uses it
> > (nudge(Ralf)). How do we state that a struct field is deprecated and
> > will be removed soon(tm)?
> 
> You get rid of all the in-tree users and then just kill it :)
> 
> More seriously, I don't think we need a deprecation schedule
> for a struct member, either it's not used in-tree anymore or
> it isn't.  We'll go crazy with any other policy and it's
> difficult enough shrinking these things :-)

OK, I think its enough to say to dlezcano that he should avoid using
sk_protinfo, as soon as we do the work on AX.25 sk_protinfo will be
removed.

- Arnaldo