All of lore.kernel.org
 help / color / mirror / Atom feed
From: "Bastien Roucariès" <rouca@debian.org>
To: "Michael Kerrisk (man-pages)" <mtk.manpages@gmail.com>
Cc: "linux-man@vger.kernel.org" <linux-man@vger.kernel.org>
Subject: Improve getsockname
Date: Tue, 17 Jan 2023 10:22:25 +0000	[thread overview]
Message-ID: <2889739.XEG2VLDMQ1@portable-bastien> (raw)


[-- Attachment #1.1: Type: text/plain, Size: 481 bytes --]

Hi,

I have a lot of student that does not use correctly getsockname in case of 
dual stack. 

May be this kind of discussion should be factorized in  sockaddr_storage (the 
historical note particularly).

i suppose the same should be done for getpeername

I think a safe programming example may be given that accept a socket as stdin 
and print information on it. Using socat it could be simple to test. maybe 
forcing ENOTSUPP if *addr > sizeof(sockadd_storage)

Regards

Bastien

[-- Attachment #1.2: 0001-Document-use-of-struct-sockaddr_storage-in-getsockna.patch --]
[-- Type: text/x-patch, Size: 2699 bytes --]

From 0afb3ad23f8ea09331f21a377e3ad19c44e4df18 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Bastien=20Roucari=C3=A8s?= <rouca@debian.org>
Date: Tue, 17 Jan 2023 10:07:43 +0000
Subject: [PATCH] Document use of struct sockaddr_storage in getsockname

Document:
- storage requierement
- future compatibility
---
 man2/getsockname.2 | 56 ++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 56 insertions(+)

diff --git a/man2/getsockname.2 b/man2/getsockname.2
index e6e8980c9..5914c9e12 100644
--- a/man2/getsockname.2
+++ b/man2/getsockname.2
@@ -39,6 +39,17 @@ The returned address is truncated if the buffer provided is too small;
 in this case,
 .I addrlen
 will return a value greater than was supplied to the call.
+.PP
+For greater portability
+.I addr
+should point to a structure of type
+.I sockaddr_storage.
+.I sockaddr_storage
+structure is large enough to hold any of the other
+.I sockaddr_*
+variants and always well aligned. On return, it should be cast to the correct
+.I sockaddr_*
+type, according to the current protocol family, given by the member ss_family.
 .SH RETURN VALUE
 On success, zero is returned.
 On error, \-1 is returned, and
@@ -80,10 +91,55 @@ For background on the
 .I socklen_t
 type, see
 .BR accept (2).
+.PP
+Security and portability wise, use of
+.I struct sockaddr_storage
+type as
+.I addr
+and
+.I addrlen
+set to
+.BI "sizeof(struct sockaddr_storage)"
+is strongly encouraged. Particularly this usage avoid bugs in dual stack IPv4+IPv6 configuration.
+.PP
+Historical use of
+.I addr
+requires one to use a structure specific to the protocol family in use,
+such as
+.I sockaddr_in
+(AF_INET or IPv4),
+.I sockaddr_in6
+(AF_INET6 or IPv6), or
+.I sockaddr_un
+(AF_UNIX)
+cast to a
+.I (struct sockaddr *).
+The purpose of the
+.I struct sockaddr *
+type
+is purely to allow casting of  domain-specific  socket  address  types  to  a
+"generic" type, so as to avoid compiler warnings about type mismatches in calls to the sockets API.
+Nevertheless,
+. I struct sockaddr *
+is too small to hold newer protocol address (for instance IPv6) and not always well aligned.
+.PP
+Even if
+.I sockaddr_storage
+type is large enough at compilation time to hold any of the
+.I sockaddr_*
+variants known by the library,
+this guarantee will not hold for future. Newer protocol may need an increase of
+.I sockaddr_storage
+size or alignement requirement, and safe program should always runtime check if returned
+.I addr
+is smaller or equal to compile time
+.BI "sizeof(struct sockaddr_storage)"
+size.
 .SH SEE ALSO
 .BR bind (2),
 .BR socket (2),
 .BR getifaddrs (3),
+.BR sockaddr_storage (3type),
 .BR ip (7),
 .BR socket (7),
 .BR unix (7)
-- 
2.39.0


[-- Attachment #2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 833 bytes --]

             reply	other threads:[~2023-01-17 10:22 UTC|newest]

Thread overview: 10+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-01-17 10:22 Bastien Roucariès [this message]
2023-01-19 12:42 ` Improve getsockname Alejandro Colomar
2023-01-19 19:44   ` Bastien Roucariès
2023-01-19 20:19     ` struct sockaddr_storage, union (was: Improve getsockname) Alejandro Colomar
2023-01-19 21:00       ` Bastien Roucariès
2023-01-19 21:19         ` Alejandro Colomar
2023-01-19 21:38           ` Bastien Roucariès
2023-01-19 23:31             ` Alejandro Colomar
2023-01-20  0:12               ` Alejandro Colomar
2023-01-20 21:11               ` Bastien Roucariès

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=2889739.XEG2VLDMQ1@portable-bastien \
    --to=rouca@debian.org \
    --cc=linux-man@vger.kernel.org \
    --cc=mtk.manpages@gmail.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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.