netdev - Re: [6.6.x REGRESSION][BISECTED] dev_snmp6: broken Ip6OutOctets accounting for forwarded IPv6 packets

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: <ZbJ5Wfx7jNfXBpAP@pc11.op.pod.cz>
Date: Thu, 25 Jan 2024 16:08:09 +0100
From: Vitezslav Samel <vitezslav@...el.cz>
To: Jakub Kicinski <kuba@...nel.org>
Cc: David Ahern <dsahern@...il.com>, heng guo <heng.guo@...driver.com>,
	netdev@...r.kernel.org
Subject: Re: [6.6.x REGRESSION][BISECTED] dev_snmp6: broken Ip6OutOctets
 accounting for forwarded IPv6 packets

On Thu, Jan 25, 2024 at 07:47:40 +0100, Vitezslav Samel wrote:
> On Wed, Jan 24, 2024 at 17:46:52 -0800, Jakub Kicinski wrote:
> > On Thu, 25 Jan 2024 08:37:11 +0800 heng guo wrote:
> > > >> Heng Guo, David, any thoughts on this? Revert?  
> > > > Revert is best; Heng Guo can revisit the math and try again.
> > > >
> > > > The patch in question basically negated IPSTATS_MIB_OUTOCTETS; I see it
> > > > shown in proc but never bumped in the datapath.  
> > > [HG]: Yes please revert it. I verified the patch on ipv4, seems I should 
> > > not touch the codes to ipv6. Sorry for it.
> > 
> > Would you mind sending a patch with a revert, explaining the situation,
> > the right Fixes tag and a link to Vitezslav's report?
> 
>   I took a look at current master and found that there is yet another
> commit since 6.6.x which touches this area: commit b4a11b2033b7 by Heng Guo
> ("net: fix IPSTATS_MIB_OUTPKGS increment in OutForwDatagrams"). It went
> in v6.7-rc1.
> 
>   I will test current master this afternoon and report back.

  Test 1: Linus' current master: IPv6 octets accounting is OK
  Test 2: 6.6.13 with b4a11b2033b7 ("net: fix IPSTATS_MIB_OUTPKGS
          increment in OutForwDatagrams") on top is also OK.

  Seems like my problem was solved in master already, but
it still exists in 6.6.y. IMHO commit b4a11b2033b7 should be
marked as for-stable-6.6.y and forwarded to GregKH. AFAIK only 6.6.y
stable tree is affected.

  But beware: I only tested my specific problem and I don't know if the
commit with fix doesn't break anything else.

	Thanks,

		Vita