netdev - Re: [PATCH net 2/2] selftests: rtnetlink: add a bond test trying to enslave non-eth dev

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: <95af874d-626e-00d0-b418-052323e00dd1@blackwall.org>
Date:   Mon, 13 Mar 2023 15:34:45 +0200
From:   Nikolay Aleksandrov <razor@...ckwall.org>
To:     netdev@...r.kernel.org
Cc:     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
Subject: Re: [PATCH net 2/2] selftests: rtnetlink: add a bond test trying to
 enslave non-eth dev

On 13/03/2023 15:32, Nikolay Aleksandrov wrote:
> On 13/03/2023 15:28, Nikolay Aleksandrov wrote:
> [snip]
>> +kci_test_enslaved_bond_non_eth()
>> +{
>> +	local ret=0
>> +
>> +	ip link add name test-nlmon0 type nlmon
>> +	ip link add name test-bond0 type bond
>> +	ip link add name test-bond1 type bond
>> +	ip link set dev test-bond0 master test-bond1
>> +	ip link set dev test-nlmon0 master test-bond0 1>/dev/null 2>/dev/null
>> +
>> +	ip -d l sh dev test-bond0 | grep -q "SLAVE"
>> +	if [ $? -ne 0 ]; then
>> +		echo "FAIL: IFF_SLAVE flag is missing from the bond device"
>> +		check_err 1
>> +	fi
>> +	ip -d l sh dev test-bond0 | grep -q "MASTER"
>> +	if [ $? -ne 0 ]; then
>> +		echo "FAIL: IFF_MASTER flag is missing from the bond device"
>> +		check_err 1
>> +	fi
>> +
>> +	# on error we return before cleaning up as that may hang the system
> 
> I wasn't sure if this part was ok, let me know if you prefer to always attempt cleaning up
> and I'll send v2 moving the return after the cleanup attempt.
> 

Actually sorry for the noise, I'll just send v2 instead with that change.
I don't have a good argument why we shouldn't attempt a cleanup after printing
the error messages even if the system hangs.