Hi Veerendranatha, 

> On Apr 3, 2017, at 8:54 AM, Veerendranatha Reddy Vallem 
> <[email protected]> wrote:
> 
> Dear Authors,
> Can you please confirm the actions need to take while validating extended 
> TLVs for malformed extended LSA check while parsing the LSA.
> Can you please confirm which case we can discard LSA or in which case we will 
> ignore  TLV, but LSA is consider as valid and processed other TLVs.
>  
> S.No            Scenario                                                      
>                                                                         Action
> 1                     Sum of TLV lengths is more than total LSA body length   
>                                      Discard the LSA by considering malformed 
> packet, no ACK

I’d certainly agree with this. 


>  
> 2                     Sum of TLV lengths is same as total LSA  body length, 
> but                                   Ignore the current TLV and continue to 
> next TLV (or)
>                        Sub TLV length/ sum of sub TLV length in one TLV       
>                                           Discard the LSA by considering 
> malformed packet with  no ACK??
>                         exceeds TLV length     


Unless the correct TLV length is known in advance, I don’t see how you can 
parse this correctly. I’d consider it a malformed packet. 



> 3.                    Sum of TLV and sub TLV lengths are  same as total LSA 
> body length                
>                        But TLV length is not correct as per TLV type
>                       (Ex: Router Link TLV has fixed length of 16 bytes, but 
> provided
>                        8 bytes data in one  of router Link TLV and TLV length 
> also set as 8 bytes)      Discard the current TLV and continue to next TLV 
> (or)
>                                                                               
>                                                                               
>       Discard the LSA by considering malformed packet with no ACK??


I’d just log an error in this case but certainly wouldn’t change the LSA 
content. Depending on the usage, the LSA can be marked as invalid. 




> 4.                   Sum of TLV and sub TLV lengths are  same as total LSA 
> body length                
>                        TLV length is correct as per length of sub TLVs, but 
> sub TLV length is
>                        not correct as per sub TLV type
>                       (Ex: Adj SID sub TLV has at least fixed length of 8 
> bytes, but provided
>                        4 bytes data in one  of Adj SID sub TLV and sub TLV 
> length also set as 4 bytes)      Discard the current TLV and continue to next 
> TLV (or)
>                                                                               
>                                                                               
>             Discard the LSA by considering malformed packet with no ACK??


Same as #3. 

Thanks,
Acee



>  
> Regards,
> Veerendranath
>                       
> _______________________________________________
> OSPF mailing list
> [email protected] <mailto:[email protected]>
> https://www.ietf.org/mailman/listinfo/ospf 
> <https://www.ietf.org/mailman/listinfo/ospf>
_______________________________________________
OSPF mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/ospf

Reply via email to