netdev - Re: behavior of recvmmsg() on blocking sockets

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <84621a61003241635h65687366t6b28148281d529c9@mail.gmail.com>
Date:	Wed, 24 Mar 2010 18:35:23 -0500
From:	Brandon Black <blblack@...il.com>
To:	netdev@...r.kernel.org
Cc:	linux-kernel@...r.kernel.org
Subject: Re: behavior of recvmmsg() on blocking sockets

On Wed, Mar 24, 2010 at 1:34 PM,  <drepper@...il.com> wrote:
> On Wed, Mar 24, 2010 at 11:28, Brandon Black <blblack@...il.com> wrote:
>>
>> I still think having a "block until at least one packet arrives" mode
>> for recvmmsg() makes sense though.
>
> I agree.  This is the mode I've seen people asking for.  They want the call
> to return as quickly as possible if there is data and then with as many
> messages as possible.  A MSG_WAITFORONE flag would do the trick nicely.

This patch might be woefully inadequate, as I'm not intimately
familiar with the rest of the Linux socket code, and I'm not sure what
the impact is of (a) adding that new flag, which is the first beyond
the 16-bit space, and (b) having that extra undefined flag present
during the underlying recvmsg() calls, but this patch Works For Me for
my isolated case.  Thoughts? (hoping gmail doesn't mangle this)

[blblack@xpc kernels]$ diff -u linux-2.6.33-orig/net/socket.c
linux-2.6.33/net/socket.c
--- linux-2.6.33-orig/net/socket.c	2010-02-24 12:52:17.000000000 -0600
+++ linux-2.6.33/net/socket.c	2010-03-24 18:10:37.156234986 -0500
@@ -2133,7 +2133,10 @@

 		if (err)
 			break;
-		++datagrams;
+
+                /* MSG_WAITFORONE turns on MSG_DONTWAIT after one packet */
+		if (!datagrams++ && flags & MSG_WAITFORONE)
+			flags |= MSG_DONTWAIT;

 		if (timeout) {
 			ktime_get_ts(timeout);
[blblack@xpc kernels]$ diff -u
linux-2.6.33-orig/include/linux/socket.h
linux-2.6.33/include/linux/socket.h
--- linux-2.6.33-orig/include/linux/socket.h	2010-02-24 12:52:17.000000000 -0600
+++ linux-2.6.33/include/linux/socket.h	2010-03-24 17:35:14.009266280 -0500
@@ -255,6 +255,7 @@
 #define MSG_ERRQUEUE	0x2000	/* Fetch message from error queue */
 #define MSG_NOSIGNAL	0x4000	/* Do not generate SIGPIPE */
 #define MSG_MORE	0x8000	/* Sender will send more */
+#define MSG_WAITFORONE  0x10000 /* recvmmsg(): block until 1+ packets avail */

 #define MSG_EOF         MSG_FIN
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html