OpenVZ Forum: Devel » [PATCH] Copy mac_len in skb

Home » Mailing lists » Devel » [PATCH] Copy mac_len in skb_clone() as well

Show: Today's Messages :: Show Polls :: Message Navigator
E-mail to friend

[PATCH] Copy mac_len in skb_clone() as well [message #11154]

Wed, 14 March 2007 13:00

Alexey Dobriyan
Messages: 195
Registered: August 2006

Senior Member

ANK says: "It is rarely used, that's wy it was not noticed.
But in the places, where it is used, it should be disaster."

Signed-off-by: Alexey Dobriyan <adobriyan@sw.ru>
---

net/core/skbuff.c | 1 +
1 file changed, 1 insertion(+)

--- a/net/core/skbuff.c
+++ b/net/core/skbuff.c
@@ -463,6 +463,7 @@ #endif
memcpy(n->cb, skb->cb, sizeof(skb->cb));
C(len);
C(data_len);
+ C(mac_len);
C(csum);
C(local_df);
n->cloned = 1;

Report message to a moderator

Re: [PATCH] Copy mac_len in skb_clone() as well [message #11195 is a reply to message #11154]

Thu, 15 March 2007 10:02

davem
Messages: 463
Registered: February 2006

Senior Member

From: Alexey Dobriyan <adobriyan@sw.ru>
Date: Wed, 14 Mar 2007 16:07:11 +0300

> ANK says: "It is rarely used, that's wy it was not noticed.
> But in the places, where it is used, it should be disaster."
>
> Signed-off-by: Alexey Dobriyan <adobriyan@sw.ru>

Applied.

What bug triggered that helped you discover this? Or is it
merely from a code audit?

Thanks.

Report message to a moderator

Re: Re: [PATCH] Copy mac_len in skb_clone() as well [message #11196 is a reply to message #11195]

Thu, 15 March 2007 10:19

dev
Messages: 1693
Registered: September 2005
Location: Moscow

Senior Member

David Miller wrote:
> From: Alexey Dobriyan <adobriyan@sw.ru>
> Date: Wed, 14 Mar 2007 16:07:11 +0300
>
>
>>ANK says: "It is rarely used, that's wy it was not noticed.
>>But in the places, where it is used, it should be disaster."
>>
>>Signed-off-by: Alexey Dobriyan <adobriyan@sw.ru>
>
>
> Applied.
>
> What bug triggered that helped you discover this? Or is it
> merely from a code audit?
Ohhh, it is a fairy-tale to tell the truth :)
We had some unexplainable problems with java application in OpenVZ kernel.
It didn't work sometimes, but worked fine (!) with CONFIG_SLAB_DEBUG.
Alexey blamed java :), but ...
Then we found that poising one of the bits in slab cache was curing it.
After that we found that the problem is related to fclone cache.
And then we found that not all the fields are initialized during cloning.
The bug was related to our own skb->field we introduced,
but we analyzed the code and found this as well.

Thanks,
Kirill

Report message to a moderator

Re: [PATCH] Copy mac_len in skb_clone() as well [message #11224 is a reply to message #11195]

Thu, 15 March 2007 16:04

Alexey Kuznetsov
Messages: 18
Registered: February 2006

Junior Member

Hello!

> What bug triggered that helped you discover this? Or is it
> merely from a code audit?

I asked the same question. :-)

openvz added some another fields to skbuff and when it was found
that they are lost while clone, he tried to figure out how all this works
and looked for another examples of this kind.

As I understand, the problem can be seen only in xfrmX_tunnel_input.
If uninitialized mac_len obtained from slab is more than current head room
it could corrupt memory.

Also, it looks like the fix is incomplete. copy_skb_header() also does not
copy this field. But it will be initialized to 0 by alloc_skb in this case
and xfrmX_tunnel_input() just will not copy mac header.

Alexey

Report message to a moderator

Re: Re: [PATCH] Copy mac_len in skb_clone() as well [message #11242 is a reply to message #11196]

Fri, 16 March 2007 01:08

davem
Messages: 463
Registered: February 2006

Senior Member

From: Kirill Korotaev <dev@sw.ru>
Date: Thu, 15 Mar 2007 13:33:12 +0300

> David Miller wrote:
> > From: Alexey Dobriyan <adobriyan@sw.ru>
> > Date: Wed, 14 Mar 2007 16:07:11 +0300
> >
> >
> >>ANK says: "It is rarely used, that's wy it was not noticed.
> >>But in the places, where it is used, it should be disaster."
> >>
> >>Signed-off-by: Alexey Dobriyan <adobriyan@sw.ru>
> >
> >
> > Applied.
> >
> > What bug triggered that helped you discover this? Or is it
> > merely from a code audit?
> Ohhh, it is a fairy-tale to tell the truth :)
> We had some unexplainable problems with java application in OpenVZ kernel.
> It didn't work sometimes, but worked fine (!) with CONFIG_SLAB_DEBUG.
> Alexey blamed java :), but ...
> Then we found that poising one of the bits in slab cache was curing it.
> After that we found that the problem is related to fclone cache.
> And then we found that not all the fields are initialized during cloning.
> The bug was related to our own skb->field we introduced,
> but we analyzed the code and found this as well.

Thanks for the detailed information.

Report message to a moderator

Previous Topic:	[PATCH 1/2] mm: move common segment checks to separate helper function (v6)
Next Topic:	Monitoring /proc/user_beancounters via Perl Script?

Goto Forum:

-=] Back to Top [=-

[ Syndicate this forum (XML) ] [

] [

]

Current Time: Fri Jul 26 23:52:49 GMT 2024

Total time taken to generate the page: 0.03203 seconds