netdev - Re: [PATCH net-next] pktgen: Add NOINIT option to leave packet data uninitialized

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: <1400250203.7973.173.camel@edumazet-glaptop2.roam.corp.google.com>
Date:	Fri, 16 May 2014 07:23:23 -0700
From:	Eric Dumazet <eric.dumazet@...il.com>
To:	Ben Greear <greearb@...delatech.com>
Cc:	Thomas Graf <tgraf@...g.ch>, davem@...emloft.net,
	netdev@...r.kernel.org
Subject: Re: [PATCH net-next] pktgen: Add NOINIT option to leave packet data
 uninitialized

On Fri, 2014-05-16 at 06:28 -0700, Ben Greear wrote:
> 
> On 05/16/2014 06:16 AM, Eric Dumazet wrote:
> > On Fri, 2014-05-16 at 12:34 +0200, Thomas Graf wrote:
> >> The memset() on the packet data is expensive and severely limiting
> >> the pps throughput for large frame sizes.
> >>
> >> Considering that pktgen requires root privileges to run it is safe
> >> to introduce an option to optionally avoid the memset() and leave
> >> the packet data uninitialized.
> >>
> >> VM test results, 2 VCPU, 2 threads, pkt_size=9K:
> >>
> >>      12.44%  -11.15%  [kernel.kallsyms]   [k] memset
> >>       4.84%   +1.58%  [kernel.kallsyms]   [k] get_page_from_freelist
> >>
> >> Signed-off-by: Thomas Graf <tgraf@...g.ch>
> >> ---
> >>   Documentation/networking/pktgen.txt |  3 +++
> >>   net/core/pktgen.c                   | 16 ++++++++++++++--
> >>   2 files changed, 17 insertions(+), 2 deletions(-)
> >
> > We do not want to leak content of memory, even if 'root' is willing to.
> >
> > Sorry, you need to do something else if you really care.
> 
> Is there currently no other way for root to grab memory contents and
> send them over the network?
> 
> I mean, it could load it's own module and do bad things if nothing else?

pktgen has already a way to send fragments which are not memset for
every packet. We share a 'zero' page for that.

So only headers need to be really built for each packet.


--
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