From: Federico Vaga <federico.vaga@gmail.com>
To: "David S. Miller" <davem@davemloft.net>,
Yuchung Cheng <ycheng@google.com>,
"Paul E. McKenney" <paulmck@linux.vnet.ibm.com>,
David Howells <dhowells@redhat.com>,
Samuel Ortiz <sameo@linux.intel.com>,
Neil Horman <nhorman@tuxdriver.com>,
Eric Dumazet <edumazet@google.com>, Li Zefan <lizefan@huawei.com>,
Andrew Morton <akpm@linux-foundation.org>
Cc: linux-kernel@vger.kernel.org, netdev@vger.kernel.org,
Alessandro Rubini <rubini@gnudd.com>
Subject: [RFC] assign some address families for local use
Date: Thu, 16 May 2013 17:49:46 +0200 [thread overview]
Message-ID: <1483365.AWhs4mCphi@harkonnen> (raw)
[-- Attachment #1: Type: text/plain, Size: 964 bytes --]
Hello,
We are working on new protocols and we think that is useful to have some
address protocol families index assigned for local use. So we will not have
conflict every time a new protocol is included within the Linux kernel.
Doubt: index 27 and 28 are not assigned to any address family, can be
explicitly assigned for local use?
We also thought to increase AF_MAX to 64 to avoid to modify it every time.
Doubt: array like af_family_key_strings (net/core/sock.c) will have some NULL
pointer. I see that a string is specified also for index 27 and 28 even if
there is not a protocol assigned for these. Is a NULL string a problem for
these vectors? Typically is used in this way:
af_family_clock_key_strings[newsk->sk_family]
So, if I set sk_family with an unassigned index I will have a NULL pointer and
a DEBUG_LOCK_WARN_ON() from lockdep_init_map() (kernel/lockdep.c)
I attached to this email the patch that do these stuff.
--
Federico Vaga
[-- Attachment #2: 0001-include-linux-socket.h-assign-address-families-for-l.patch --]
[-- Type: text/x-patch, Size: 4304 bytes --]
>From 8ce4f2576aa8e95ea22921c31bdffd049460951d Mon Sep 17 00:00:00 2001
From: Federico Vaga <federico.vaga@gmail.com>
Date: Wed, 15 May 2013 12:32:03 +0200
Subject: [PATCH] include/linux/socket.h: assign address families for local use
The patch assigns 4 address families for local use only. This is
useful because it allows to maintain an address family outside kernel
source without conflict. It is also useful during development until a
number is officially assigned.
This is the same kind of policy applied for major number
(Documentation/devices.text)
This patch also increases the number of maximum address (protocol)
families to 64. In this way for a while nobody need to increase this
value. The cost, in terms of memory, is tiny. I made an (very)
approximate calculation about the cost of an unused address family by
following NPROTO, AF_MAX and PF_MAX usage. If I did not big errors it
should be about 70Byte on 32bit systems and 130Byte on 64bit systems for
each new address family.
I also compiled a kernel on a x86_64 machine:
Without patch
text data bss dec hex filename
10935491 1398904 1175552 13509947 ce253b vmlinux
With patch
text data bss dec hex filename
10935427 1399544 1175552 13510523 ce277b vmlinux
Signed-off-by: Federico Vaga <federico.vaga@gmail.com>
---
include/linux/socket.h | 12 +++++++++++-
net/core/sock.c | 12 +++++++++---
2 files changed, 20 insertions(+), 4 deletions(-)
diff --git a/include/linux/socket.h b/include/linux/socket.h
index 428c37a..4775d69 100644
--- a/include/linux/socket.h
+++ b/include/linux/socket.h
@@ -179,7 +179,12 @@ struct ucred {
#define AF_ALG 38 /* Algorithm sockets */
#define AF_NFC 39 /* NFC sockets */
#define AF_VSOCK 40 /* vSockets */
-#define AF_MAX 41 /* For now.. */
+#define AF_LOCAL1 41 /* Local use sockets */
+#define AF_LOCAL2 42 /* Local use sockets */
+#define AF_LOCAL3 43 /* Local use sockets */
+#define AF_LOCAL4 44 /* Local use sockets */
+/* new address families here */
+#define AF_MAX 64
/* Protocol families, same as address families. */
#define PF_UNSPEC AF_UNSPEC
@@ -223,6 +228,11 @@ struct ucred {
#define PF_ALG AF_ALG
#define PF_NFC AF_NFC
#define PF_VSOCK AF_VSOCK
+#define PF_LOCAL1 AF_LOCAL1
+#define PF_LOCAL2 AF_LOCAL2
+#define PF_LOCAL3 AF_LOCAL3
+#define PF_LOCAL4 AF_LOCAL4
+/* new protocol families here */
#define PF_MAX AF_MAX
/* Maximum queue length specifiable by listen. */
diff --git a/net/core/sock.c b/net/core/sock.c
index 6ba327d..9bf66ab 100644
--- a/net/core/sock.c
+++ b/net/core/sock.c
@@ -210,7 +210,9 @@ static const char *const af_family_key_strings[AF_MAX+1] = {
"sk_lock-AF_TIPC" , "sk_lock-AF_BLUETOOTH", "sk_lock-IUCV" ,
"sk_lock-AF_RXRPC" , "sk_lock-AF_ISDN" , "sk_lock-AF_PHONET" ,
"sk_lock-AF_IEEE802154", "sk_lock-AF_CAIF" , "sk_lock-AF_ALG" ,
- "sk_lock-AF_NFC" , "sk_lock-AF_MAX"
+ "sk_lock-AF_NFC" , "sk_lock-AF_LOCAL1" , "sk_lock-AF_LOCAL2" ,
+ "sk_lock-AF_LOCAL3", "sk_lock-AF_LOCAL4" ,
+ [AF_MAX] = "sk_lock-AF_MAX"
};
static const char *const af_family_slock_key_strings[AF_MAX+1] = {
"slock-AF_UNSPEC", "slock-AF_UNIX" , "slock-AF_INET" ,
@@ -226,7 +228,9 @@ static const char *const af_family_slock_key_strings[AF_MAX+1] = {
"slock-AF_TIPC" , "slock-AF_BLUETOOTH", "slock-AF_IUCV" ,
"slock-AF_RXRPC" , "slock-AF_ISDN" , "slock-AF_PHONET" ,
"slock-AF_IEEE802154", "slock-AF_CAIF" , "slock-AF_ALG" ,
- "slock-AF_NFC" , "slock-AF_MAX"
+ "slock-AF_NFC" , "slock-AF_LOCAL1" , "slock-AF_LOCAL2" ,
+ "slock-AF_LOCAL3", "slock-AF_LOCAL4" ,
+ [AF_MAX] = "slock-AF_MAX"
};
static const char *const af_family_clock_key_strings[AF_MAX+1] = {
"clock-AF_UNSPEC", "clock-AF_UNIX" , "clock-AF_INET" ,
@@ -242,7 +246,9 @@ static const char *const af_family_clock_key_strings[AF_MAX+1] = {
"clock-AF_TIPC" , "clock-AF_BLUETOOTH", "clock-AF_IUCV" ,
"clock-AF_RXRPC" , "clock-AF_ISDN" , "clock-AF_PHONET" ,
"clock-AF_IEEE802154", "clock-AF_CAIF" , "clock-AF_ALG" ,
- "clock-AF_NFC" , "clock-AF_MAX"
+ "clock-AF_NFC" , "clock-AF_LOCAL1" , "clock-AF_LOCAL2" ,
+ "clock-AF_LOCAL3", "clock-AF_LOCAL4" ,
+ [AF_MAX] = "clock-AF_MAX"
};
/*
--
1.8.1.4
reply other threads:[~2013-05-16 15:49 UTC|newest]
Thread overview: [no followups] expand[flat|nested] mbox.gz Atom feed
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=1483365.AWhs4mCphi@harkonnen \
--to=federico.vaga@gmail.com \
--cc=akpm@linux-foundation.org \
--cc=davem@davemloft.net \
--cc=dhowells@redhat.com \
--cc=edumazet@google.com \
--cc=linux-kernel@vger.kernel.org \
--cc=lizefan@huawei.com \
--cc=netdev@vger.kernel.org \
--cc=nhorman@tuxdriver.com \
--cc=paulmck@linux.vnet.ibm.com \
--cc=rubini@gnudd.com \
--cc=sameo@linux.intel.com \
--cc=ycheng@google.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox