Re: [PATCH] openvswitch: supply a dummy err_handler of gre_cisco_protocol to prevent kernel crash

[wei zhang]

AtÂ2014-04-01Â08:49:53,"JesseÂGross"Â<jesse@xxxxxxxxxx>Âwrote:
>OnÂSun,ÂMarÂ30,Â2014ÂatÂ5:12ÂAM,ÂweiÂzhangÂ<asuka.com@xxxxxxx>Âwrote:
>>ÂAtÂ2014-03-29Â06:02:25,"JesseÂGross"Â<jesse@xxxxxxxxxx>Âwrote:

>>ÂMaybeÂIÂmisunderstandÂsomething?ÂIÂthinkÂifÂweÂdiscardÂallÂpacketÂpassÂtoÂus
>>ÂwhenÂweÂuseÂgreÂvport,ÂnewÂgre_cisco_protocolÂwhichÂhasÂlowerÂpriorityÂcould
>>ÂnotÂseeÂtheÂpacketÂintendedÂtoÂit.
>
>That'sÂtrueÂbutÂinÂthisÂcaseÂitÂwouldÂalsoÂnotÂseeÂanyÂdataÂpackets,
>soÂIÂdon'tÂthinkÂthatÂsituationÂwouldÂworkÂwellÂanyways.
>
>>ÂIÂcheckedÂtheÂimplementationÂofÂtheÂipgre_err(),ÂwhichÂhasÂbeÂcalledÂbefore
>>ÂtheÂerr_handlerÂofÂgreÂvport.ÂItÂuseÂtheÂtheÂ(localÂaddress,ÂremoteÂaddress,Âkey)
>>ÂtoÂdistinguishÂtheÂpacketÂwhichÂisÂrealyÂintendedÂtoÂit,ÂalthoughÂitÂcouldÂnot
>>ÂalwaysÂgetÂtheÂkeyÂfromÂtheÂicmpÂpacket.ÂShouldÂweÂdoÂasÂtheÂsameÂasÂit?
>>ÂI'mÂnotÂsureÂthisÂisÂfeasible,ÂanyÂadviceÂisÂappreciate.
>
>OVSÂdoesÂflowÂbasedÂmatchingÂratherÂthanÂusingÂaÂstaticÂsetÂof
>configurationÂparameters,ÂsoÂeverythingÂ"matches"ÂinÂsomeÂway
>(althoughÂtheÂresultÂmightÂbeÂtoÂdrop).Â

So the flow based match could dynamically determine by the ovs daemon, we could
not find out the belonging of the packet as far as we callÂovs_dp_upcall(), isn't it?

>ThisÂgenerallyÂmeansÂthatÂOVS
>isÂtheÂreceiverÂofÂlastÂresortÂandÂnothingÂcurrentlyÂhasÂaÂlower
>priority.Â

Thanks for your kind help, this clarify my misunderstanding!

>ThatÂactuallyÂmeansÂtheÂdifferenceÂbetweenÂtheÂpatchesÂis
>somewhatÂacademicÂbutÂitÂseemsÂmoreÂrobustÂforÂtheÂlogicÂtoÂbe
>consistent.

Regards,
Wei ZhangÂ