netdev - Re: [PATCH net-next] net: vlan: Avoid using BUG() in vlan_proto

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] [day] [month] [year] [list]

Message-ID: <74995756-2a48-7e63-3806-e98b624177e0@gmail.com>
Date:   Thu, 24 Sep 2020 16:54:15 -0700
From:   Florian Fainelli <f.fainelli@...il.com>
To:     Jakub Kicinski <kuba@...nel.org>
Cc:     netdev@...r.kernel.org, olteanv@...il.com, vladimir.oltean@....com
Subject: Re: [PATCH net-next] net: vlan: Avoid using BUG() in vlan_proto_idx()



On 9/24/2020 4:46 PM, Jakub Kicinski wrote:
> On Wed, 23 Sep 2020 21:16:27 -0700 Florian Fainelli wrote:
>> While we should always make sure that we specify a valid VLAN protocol
>> to vlan_proto_idx(), killing the machine when an invalid value is
>> specified is too harsh and not helpful for debugging. All callers are
>> capable of dealing with an error returned by vlan_proto_idx() so check
>> the index value and propagate it accordingly.
>>
>> Signed-off-by: Florian Fainelli <f.fainelli@...il.com>
> 
> Perhaps it's heresy but I wonder if the error checking is worth it
> or we'd be better off WARNing and assuming normal Q tag.. unlikely
> someone is getting it wrong now given the BUG().

This showed up while working with Vladimir on another patch set where we 
were able to submit packets with an incorrect skb->vlan_proto submitted 
via the DSA stacking model. It should not happen, but arguably crashing 
the kernel was not helping either.

> 
>> diff --git a/net/8021q/vlan.c b/net/8021q/vlan.c
>> index d4bcfd8f95bf..6c08de1116c1 100644
>> --- a/net/8021q/vlan.c
>> +++ b/net/8021q/vlan.c
>> @@ -57,6 +57,9 @@ static int vlan_group_prealloc_vid(struct vlan_group *vg,
>>   	ASSERT_RTNL();
>>   
>>   	pidx  = vlan_proto_idx(vlan_proto);
>> +	if (pidx < 0)
>> +		return -EINVAL;
>> +
>>   	vidx  = vlan_id / VLAN_GROUP_ARRAY_PART_LEN;
>>   	array = vg->vlan_devices_arrays[pidx][vidx];
>>   	if (array != NULL)
>> diff --git a/net/8021q/vlan.h b/net/8021q/vlan.h
>> index bb7ec1a3915d..143e9c12dbd6 100644
>> --- a/net/8021q/vlan.h
>> +++ b/net/8021q/vlan.h
>> @@ -44,8 +44,8 @@ static inline unsigned int vlan_proto_idx(__be16 proto)
> 
> adjust return type

Ah yes, indeed.

> 
>>   	case htons(ETH_P_8021AD):
>>   		return VLAN_PROTO_8021AD;
>>   	default:
>> -		BUG();
>> -		return 0;
>> +		WARN(1, "invalid VLAN protocol: 0x%04x\n", htons(proto));
> 
> ntohs()

OK, there is also a variable name pdix instead of pidx, I will resend 
shortly, thanks!
-- 
Florian