[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Thu, 04 Apr 2013 03:22:07 -0700
From: ebiederm@...ssion.com (Eric W. Biederman)
To: dingtianhong <dingtianhong@...wei.com>
Cc: "David S. Miller" <davem@...emloft.net>,
Sven Joachim <svenjoac@....de>,
Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
<linux-kernel@...r.kernel.org>, <stable@...r.kernel.org>,
Eric Dumazet <edumazet@...gle.com>,
Andy Lutomirski <luto@...capital.net>,
Karel Srot <ksrot@...hat.com>, <netdev@...r.kernel.org>,
Eric Dumazet <eric.dumazet@...il.com>
Subject: Re: [PATCH 1/2] Revert "af_unix: dont send SCM_CREDENTIAL when dest socket is NULL"
dingtianhong <dingtianhong@...wei.com> writes:
> On 2013/4/4 10:13, Eric W. Biederman wrote:
>>
>> This reverts commit 14134f6584212d585b310ce95428014b653dfaf6.
>>
>> The problem that the above patch was meant to address is that af_unix
>> messages are not being coallesced because we are sending unnecesarry
>> credentials. Not sending credentials in maybe_add_creds totally
>> breaks unconnected unix domain sockets that wish to send credentails
>> to other sockets.
>>
>
> thanks for check the question and make a fix solution, but I still doubt that if unconnected unix
> domain socket wish to send credentails to oher sockets, why dont set
> SOCK_PASSCRED on sock->flags, I think the user need to decide the param
> and shouldnt send creds by default way.
The big issue is the semantics are the receiver sets SOCK_PASSCRED when
they want to receive credentials. When transmitting packets from
unconnected or unaccepted sockets we don't know if the receiver has set
SOCK_PASSCRED so when in doubt transmit. Historically we always
tranmitted credentials.
Furthermore we have a real regression in udev that breaks systems, so
this patch must be reverted.
Eric
--
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
Powered by blists - more mailing lists