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-next>] [day] [month] [year] [list]
Message-Id: <20230315111842.1589296-1-razor@blackwall.org>
Date:   Wed, 15 Mar 2023 13:18:39 +0200
From:   Nikolay Aleksandrov <razor@...ckwall.org>
To:     netdev@...r.kernel.org
Cc:     monis@...taire.com, syoshida@...hat.com, j.vosburgh@...il.com,
        andy@...yhouse.net, kuba@...nel.org, davem@...emloft.net,
        pabeni@...hat.com, edumazet@...gle.com,
        syzbot+9dfc3f3348729cc82277@...kaller.appspotmail.com,
        michal.kubiak@...el.com, jtoppins@...hat.com,
        Nikolay Aleksandrov <razor@...ckwall.org>
Subject: [PATCH net v3 0/3] bonding: properly restore flags when bond changes ether type

Hi,
A bug was reported by syzbot[1] that causes a warning and a myriad of
other potential issues if a bond, that is also a slave, fails to enslave a
non-eth device. While fixing that bug I found that we have the same
issues when such enslave passes and after that the bond changes back to
ARPHRD_ETHER (again due to ether_setup). This set fixes all issues by
extracting the ether_setup() sequence in a helper which does the right
thing about bond flags when it needs to change back to ARPHRD_ETHER. It
also adds selftests for these cases.

Patch 01 adds the new bond_ether_setup helper and fixes the issues when a
bond device changes its ether type due to successful enslave. Patch 02
fixes the issues when it changes its ether type due to an unsuccessful
enslave. Note we need two patches because the bugs were introduced by
different commits. Patch 03 adds the new selftests.

Due to the comment adjustment and squash, could you please review
patch 01 again? I've kept the other acks since there were no code
changes.

v3: squash the helper patch and the first fix, adjust the comment above
    it to be explicit about the bond device, no code changes
v2: new set, all patches are new due to new approach of fixing these bugs

Thanks,
 Nik

[1] https://syzkaller.appspot.com/bug?id=391c7b1f6522182899efba27d891f1743e8eb3ef

Nikolay Aleksandrov (3):
  bonding: restore IFF_MASTER/SLAVE flags on bond enslave ether type
    change
  bonding: restore bond's IFF_SLAVE flag if a non-eth dev enslave fails
  selftests: bonding: add tests for ether type changes

 drivers/net/bonding/bond_main.c               | 23 +++--
 .../selftests/drivers/net/bonding/Makefile    |  3 +-
 .../net/bonding/bond-eth-type-change.sh       | 85 +++++++++++++++++++
 3 files changed, 103 insertions(+), 8 deletions(-)
 create mode 100755 tools/testing/selftests/drivers/net/bonding/bond-eth-type-change.sh

-- 
2.39.1

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ