From: Oliver Hartkopp <socketcan@hartkopp•net>
To: Patrick McHardy <kaber@trash•net>, Urs Thuermann <urs@isnogud•escape.de>
Cc: David Miller <davem@davemloft•net>,
Thomas Gleixner <tglx@linutronix•de>,
Oliver Hartkopp <oliver.hartkopp@volkswagen•de>,
Urs Thuermann <urs.thuermann@volkswagen•de>,
netdev@vger•kernel.org
Subject: Re: [patch 5/7] CAN: Add virtual CAN netdevice driver
Date: Sat, 02 Jun 2007 21:58:31 +0200 [thread overview]
Message-ID: <4661CBE7.7040901@hartkopp.net> (raw)
In-Reply-To: <46613DBE.3010602@hartkopp.net>
The more i was thinking about my own suggested solution the more it
turned out to be ugly for me ...
Summarizing we have two problems to solve:
1. Identify the originating sk to potentially trash my own sent messages.
2. Indicate a requested local loopback to the lower (driver) layer.
Regarding point 1.:
skb->sk is intentionally set to NULL, when ever skb_orphan() or
skb_clone() is invoked to cut the reference to the sk. Performing a
loopback this is a reasonable thing to do as also skb->destructor(skb)
is called in skb_orphan().
Indeed skb->sk is completely unused in the rx path, so we just would
have to 'preserve' skb->sk the way 'up' whenever we make use of
skb_orphan() or skb_clone().
E.g. in af_can.c the deliver() function would be changed like this:
static inline void deliver(struct sk_buff *skb, struct receiver *r)
{
+ struct sock *sk = skb->sk;
struct sk_buff *clone = skb_clone(skb, GFP_ATOMIC);
DBG("skbuff %p cloned to %p\n", skb, clone);
if (clone) {
+ clone->sk = sk;
r->func(clone, r->data);
r->matches++;
}
}
So there is a proper way to make the originating sk avaliable when the
skb reaches the receiving socket.
Regarding point 2.:
To indicate an outgoing skb to be looped back it would just be enough to
set skb->tstamp at skb creation time. This would be similar to set the
timestamp at netdev receive time. skb->tstamp is also completely unused
in the tx path. In the case of local loopback setting the timestamp at
this creation stage is not even a wrong semantic behaviour.
Assuming some clarifying comments in the source code this looks like a
simple to use and clearly arranged implementation to me. I'll make a
test implementation tomorrow to see how if it smells as good as i
expected ;-)
Best regards,
Oliver
Oliver Hartkopp wrote:
> Patrick McHardy wrote:
>
>> Oliver Hartkopp wrote:
>>
>>
>>> Patrick McHardy wrote:
>>>
>>>
>> Yes, its working, but only in certain combinations and you're breaking
>> the rules for skb->cb, making it impossible for other layers to use.
>> skb->sk is "stable" at the output path, the regular loopback device
>> orphans the skb in hard_start_xmit. So you can at least use it there.
>>
>>
>>
>>> Would therefore skb->cb left unchanged in my skb's? Or is there any flag
>>> that can be set in the skb to keep the packet scheduler's hands off?
>>>
>>>
>> No, and I don't think we want a flag to signal that something is
>> violating the rules for skb->cb, there are other users of this
>> besides qdiscs.
>>
>>
>
> Hm - regarding Patricks and Urs' last mails i just had the idea to put
> the sk-reference that's needed for this special
> CAN-only-loopback-functionality into the data section of the skb, e.g.
> by introducing a new struct can_skb_data:
>
> struct can_skb_data {
> struct can_frame cf;
> sock *txsk;
> };
>
> So instead of allocating the space of struct can_frame the alloc_skb()
> would allocate the size of struct can_skb_data.
> The needed txsk would be stable in any case and could be used like the
> currently missused skb->cb. This would also lead to a type proof(!)
> implementation.
>
> In raw_rcv() in raw.c there could be a check for the size of struct
> can_skb_data first before checking the txsk - this would also guarantee
> the backward compatibility for current CAN drivers that allocate only
> the size of struct can_frame. For me this looks like a safe and
> compatible (Kernel & CAN) solution.
>
> Any objections/comments for this approach?
>
> Best regards,
> Oliver
>
>
> -
> To unsubscribe from this list: send the line "unsubscribe netdev" in
> the body of a message to majordomo@vger•kernel.org
> More majordomo info at http://vger.kernel.org/majordomo-info.html
>
>
next prev parent reply other threads:[~2007-06-02 19:58 UTC|newest]
Thread overview: 97+ messages / expand[flat|nested] mbox.gz Atom feed top
2007-05-30 13:11 [patch 0/7] CAN: Add new PF_CAN protocol family, update Urs Thuermann
2007-05-30 13:11 ` [patch 1/7] CAN: Allocate protocol numbers for PF_CAN Urs Thuermann
2007-05-30 13:11 ` [patch 2/7] CAN: Add PF_CAN core module Urs Thuermann
2007-05-30 13:11 ` [patch 3/7] CAN: Add raw protocol Urs Thuermann
2007-05-30 13:11 ` [patch 4/7] CAN: Add broadcast manager (bcm) protocol Urs Thuermann
2007-05-30 13:11 ` [patch 5/7] CAN: Add virtual CAN netdevice driver Urs Thuermann
2007-05-30 17:13 ` Patrick McHardy
2007-05-30 18:39 ` Oliver Hartkopp
2007-05-30 19:16 ` Patrick McHardy
2007-05-30 19:48 ` Oliver Hartkopp
2007-05-30 19:58 ` Patrick McHardy
2007-05-31 5:13 ` Oliver Hartkopp
2007-05-31 20:25 ` Oliver Hartkopp
2007-06-01 14:54 ` Patrick McHardy
2007-06-02 9:51 ` Oliver Hartkopp
2007-06-02 19:58 ` Oliver Hartkopp [this message]
2007-06-03 19:16 ` Oliver Hartkopp
2007-06-04 11:53 ` Patrick McHardy
2007-06-04 12:44 ` Urs Thuermann
2007-06-06 11:39 ` Patrick McHardy
2007-06-04 12:17 ` Urs Thuermann
2007-06-04 16:32 ` Oliver Hartkopp
2007-06-04 18:26 ` Oliver Hartkopp
2007-06-02 8:09 ` Urs Thuermann
2007-06-03 18:27 ` Patrick McHardy
2007-05-30 17:28 ` Stephen Hemminger
2007-05-30 18:57 ` Oliver Hartkopp
2007-05-30 19:01 ` Stephen Hemminger
2007-05-30 19:25 ` Oliver Hartkopp
2007-05-30 13:11 ` [patch 6/7] CAN: Add maintainer entries Urs Thuermann
2007-05-30 13:11 ` [patch 7/7] CAN: Add documentation Urs Thuermann
-- strict thread matches above, loose matches on Subject: below --
2007-11-16 15:02 [PATCH 0/7] CAN: New PF_CAN protocol family for 2.6.25, update Urs Thuermann
2007-11-16 15:02 ` [PATCH 5/7] CAN: Add virtual CAN netdevice driver Urs Thuermann
2007-11-14 12:13 [PATCH 0/7] CAN: New PF_CAN protocol family for 2.6.25 Urs Thuermann
2007-11-14 12:13 ` [PATCH 5/7] CAN: Add virtual CAN netdevice driver Urs Thuermann
2007-11-14 12:51 ` Patrick McHardy
2007-11-14 13:36 ` Urs Thuermann
2007-11-14 13:45 ` Patrick McHardy
2007-10-05 10:49 [PATCH 0/7] CAN: Add new PF_CAN protocol family, try #10 Urs Thuermann
2007-10-05 10:49 ` [PATCH 5/7] CAN: Add virtual CAN netdevice driver Urs Thuermann
2007-10-02 13:10 [PATCH 0/7] CAN: Add new PF_CAN protocol family, try #9 Urs Thuermann
2007-10-02 13:10 ` [PATCH 5/7] CAN: Add virtual CAN netdevice driver Urs Thuermann
2007-10-02 14:20 ` Arnaldo Carvalho de Melo
2007-10-02 15:07 ` Oliver Hartkopp
2007-10-02 16:46 ` Arnaldo Carvalho de Melo
2007-10-02 21:02 ` Oliver Hartkopp
2007-10-02 21:43 ` Arnaldo Carvalho de Melo
2007-10-02 21:50 ` David Miller
2007-10-03 7:06 ` Oliver Hartkopp
2007-10-02 21:52 ` Stephen Hemminger
2007-10-02 22:04 ` David Miller
2007-10-03 17:47 ` Oliver Hartkopp
2007-10-04 11:52 ` Urs Thuermann
2007-09-25 12:20 [PATCH 0/7] CAN: Add new PF_CAN protocol family, try #8 Urs Thuermann
2007-09-25 12:20 ` [PATCH 5/7] CAN: Add virtual CAN netdevice driver Urs Thuermann
2007-09-25 13:26 ` YOSHIFUJI Hideaki / 吉藤英明
2007-09-25 14:00 ` Urs Thuermann
2007-09-25 14:47 ` Patrick McHardy
2007-09-25 17:55 ` Oliver Hartkopp
2007-09-25 17:53 ` Patrick McHardy
2007-09-25 19:09 ` Oliver Hartkopp
2007-09-20 18:43 [PATCH 0/7] CAN: Add new PF_CAN protocol family, try #7 Urs Thuermann
2007-09-20 18:43 ` [PATCH 5/7] CAN: Add virtual CAN netdevice driver Urs Thuermann
2007-09-27 15:54 ` Eric W. Biederman
2007-09-27 16:16 ` Eric W. Biederman
2007-09-27 19:18 ` David Miller
2007-09-28 8:48 ` Oliver Hartkopp
2007-09-28 16:52 ` Eric W. Biederman
2007-09-28 18:33 ` Oliver Hartkopp
2007-09-17 10:03 [PATCH 0/7] CAN: Add new PF_CAN protocol family, try #6 Urs Thuermann
2007-09-17 10:03 ` [PATCH 5/7] CAN: Add virtual CAN netdevice driver Urs Thuermann
2007-09-18 15:02 ` Patrick McHardy
2007-09-18 22:24 ` Urs Thuermann
2007-09-19 6:26 ` Oliver Hartkopp
2007-09-19 8:41 ` Patrick McHardy
2007-08-04 2:06 [patch 0/7] CAN: Add new PF_CAN protocol family, try #5 Urs Thuermann
2007-08-04 2:07 ` [patch 5/7] CAN: Add virtual CAN netdevice driver Urs Thuermann
2007-06-22 3:44 [patch 0/7] CAN: Add new PF_CAN protocol family, try #3 Urs Thuermann
2007-06-22 3:44 ` [patch 5/7] CAN: Add virtual CAN netdevice driver Urs Thuermann
2007-06-22 11:02 ` Patrick McHardy
2007-06-22 12:22 ` Urs Thuermann
2007-06-22 12:38 ` Patrick McHardy
2007-06-23 12:05 ` Oliver Hartkopp
2007-06-23 12:52 ` Patrick McHardy
2007-06-23 15:13 ` Oliver Hartkopp
2007-06-23 16:25 ` Patrick McHardy
2007-06-23 16:42 ` Oliver Hartkopp
2007-06-23 17:13 ` Patrick McHardy
2007-07-04 11:37 ` Urs Thuermann
2007-07-04 14:01 ` Patrick McHardy
2007-07-09 11:37 ` Urs Thuermann
2007-07-09 14:18 ` Patrick McHardy
2007-07-09 15:27 ` Oliver Hartkopp
2007-07-11 19:41 ` Oliver Hartkopp
2007-07-11 22:52 ` Patrick McHardy
2007-07-16 6:05 ` Oliver Hartkopp
2007-07-16 8:37 ` David Miller
2007-07-16 13:08 ` Patrick McHardy
2007-07-16 16:27 ` Oliver Hartkopp
2007-07-16 13:07 ` Patrick McHardy
2007-07-16 16:00 ` Oliver Hartkopp
2007-06-23 21:01 ` David Miller
2007-06-23 21:44 ` Oliver Hartkopp
2007-06-23 20:51 ` David Miller
2007-06-23 21:49 ` Oliver Hartkopp
2007-05-16 14:51 [patch 0/7] CAN: Add new PF_CAN protocol family Urs Thuermann
2007-05-16 14:51 ` [patch 5/7] CAN: Add virtual CAN netdevice driver Urs Thuermann
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=4661CBE7.7040901@hartkopp.net \
--to=socketcan@hartkopp$(echo .)net \
--cc=davem@davemloft$(echo .)net \
--cc=kaber@trash$(echo .)net \
--cc=netdev@vger$(echo .)kernel.org \
--cc=oliver.hartkopp@volkswagen$(echo .)de \
--cc=tglx@linutronix$(echo .)de \
--cc=urs.thuermann@volkswagen$(echo .)de \
--cc=urs@isnogud$(echo .)escape.de \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
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