diff mbox series

[ovs-dev,branch-2.8,2/2] datapath: Clear the L4 portion of the key for "later" fragments

Message ID 1567101350-27905-2-git-send-email-gvrose8192@gmail.com
State Changes Requested
Headers show
Series [ovs-dev,branch-2.8,1/2] datapath: Properly set L4 keys on "later" IP fragments | expand

Commit Message

Gregory Rose Aug. 29, 2019, 5:55 p.m. UTC 
From: Justin Pettit <jpettit@ovn.org>

Upstream commit:
    commit 0754b4e8cdf3eec6e4122e79af26ed9bab20f8f8
    Author: Justin Pettit <jpettit@ovn.org>
    Date:   Tue Aug 27 07:58:10 2019 -0700

    openvswitch: Clear the L4 portion of the key for "later" fragments.

    Only the first fragment in a datagram contains the L4 headers.  When the
    Open vSwitch module parses a packet, it always sets the IP protocol
    field in the key, but can only set the L4 fields on the first fragment.
    The original behavior would not clear the L4 portion of the key, so
    garbage values would be sent in the key for "later" fragments.  This
    patch clears the L4 fields in that circumstance to prevent sending those
    garbage values as part of the upcall.

    Signed-off-by: Justin Pettit <jpettit@ovn.org>
    Acked-by: Pravin B Shelar <pshelar@ovn.org>
    Signed-off-by: David S. Miller <davem@davemloft.net>

Signed-off-by: Greg Rose <gvrose8192@gmail.com>
Signed-off-by: Justin Pettit <jpettit@ovn.org>
---
 datapath/flow.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)

Comments

0-day Robot Aug. 29, 2019, 9:41 p.m. UTC | #1 
Bleep bloop.  Greetings Greg Rose, I am a robot and I have tried out your patch.
Thanks for your contribution.

I encountered some error that I wasn't expecting.  See the details below.


checkpatch:
WARNING: Unexpected sign-offs from developers who are not authors or co-authors or committers: Greg Rose <gvrose8192@gmail.com>
Lines checked: 58, Warnings: 1, Errors: 0


Please check this out.  If you feel there has been an error, please email aconole@redhat.com

Thanks,
0-day Robot
Justin Pettit Aug. 29, 2019, 10:25 p.m. UTC | #2 
> On Aug 29, 2019, at 10:55 AM, Greg Rose <gvrose8192@gmail.com> wrote:
> 
> diff --git a/datapath/flow.c b/datapath/flow.c
> index 083288f..92fc6ac 100644
> --- a/datapath/flow.c
> +++ b/datapath/flow.c
> @@ -529,6 +529,7 @@ static int key_extract_l3l4(struct sk_buff *skb, struct sw_flow_key *key)
> 		offset = nh->frag_off & htons(IP_OFFSET);
> 		if (offset) {
> 			key->ip.frag = OVS_FRAG_TYPE_LATER;
> +			memset(&key->tp, 0, sizeof(key->tp));
> 			return 0;
> 		}
> 		if (nh->frag_off & htons(IP_MF) ||
> @@ -647,8 +648,11 @@ static int key_extract_l3l4(struct sk_buff *skb, struct sw_flow_key *key)
> 			return error;
> 		}
> 
> -		if (key->ip.frag == OVS_FRAG_TYPE_LATER)
> +		if (key->ip.frag == OVS_FRAG_TYPE_LATER) {
> +			memset(&key->tp, 0, sizeof(key->tp));
> 			return 0;
> +		}
> +#ifdef HAVE_SKB_GSO_UDP

My system's kernel is too new to be built against an OVS this old, but I noticed this patch for OVS versions 2.5 through 2.9 introduce this #ifdef without a corresponding #endif.  Was this intentional?  Does it even build?

--Justin
Gregory Rose Aug. 30, 2019, 1:46 a.m. UTC | #3 
On 8/29/2019 3:25 PM, Justin Pettit wrote:
>> On Aug 29, 2019, at 10:55 AM, Greg Rose <gvrose8192@gmail.com> wrote:
>>
>> diff --git a/datapath/flow.c b/datapath/flow.c
>> index 083288f..92fc6ac 100644
>> --- a/datapath/flow.c
>> +++ b/datapath/flow.c
>> @@ -529,6 +529,7 @@ static int key_extract_l3l4(struct sk_buff *skb, struct sw_flow_key *key)
>> 		offset = nh->frag_off & htons(IP_OFFSET);
>> 		if (offset) {
>> 			key->ip.frag = OVS_FRAG_TYPE_LATER;
>> +			memset(&key->tp, 0, sizeof(key->tp));
>> 			return 0;
>> 		}
>> 		if (nh->frag_off & htons(IP_MF) ||
>> @@ -647,8 +648,11 @@ static int key_extract_l3l4(struct sk_buff *skb, struct sw_flow_key *key)
>> 			return error;
>> 		}
>>
>> -		if (key->ip.frag == OVS_FRAG_TYPE_LATER)
>> +		if (key->ip.frag == OVS_FRAG_TYPE_LATER) {
>> +			memset(&key->tp, 0, sizeof(key->tp));
>> 			return 0;
>> +		}
>> +#ifdef HAVE_SKB_GSO_UDP
> My system's kernel is too new to be built against an OVS this old, but I noticed this patch for OVS versions 2.5 through 2.9 introduce this #ifdef without a corresponding #endif.  Was this intentional?  Does it even build?
>
> --Justin
>
>
Conflict resolutions error - I'll probably have to resend branches 2.8 
through 2.5.

- Greg
Justin Pettit Aug. 30, 2019, 5:44 a.m. UTC | #4 
> On Aug 29, 2019, at 6:46 PM, Gregory Rose <gvrose8192@gmail.com> wrote:
> 
> 
> On 8/29/2019 3:25 PM, Justin Pettit wrote:
>>> On Aug 29, 2019, at 10:55 AM, Greg Rose <gvrose8192@gmail.com> wrote:
>>> 
>>> diff --git a/datapath/flow.c b/datapath/flow.c
>>> index 083288f..92fc6ac 100644
>>> --- a/datapath/flow.c
>>> +++ b/datapath/flow.c
>>> @@ -529,6 +529,7 @@ static int key_extract_l3l4(struct sk_buff *skb, struct sw_flow_key *key)
>>>        offset = nh->frag_off & htons(IP_OFFSET);
>>>        if (offset) {
>>>            key->ip.frag = OVS_FRAG_TYPE_LATER;
>>> +            memset(&key->tp, 0, sizeof(key->tp));
>>>            return 0;
>>>        }
>>>        if (nh->frag_off & htons(IP_MF) ||
>>> @@ -647,8 +648,11 @@ static int key_extract_l3l4(struct sk_buff *skb, struct sw_flow_key *key)
>>>            return error;
>>>        }
>>> 
>>> -        if (key->ip.frag == OVS_FRAG_TYPE_LATER)
>>> +        if (key->ip.frag == OVS_FRAG_TYPE_LATER) {
>>> +            memset(&key->tp, 0, sizeof(key->tp));
>>>            return 0;
>>> +        }
>>> +#ifdef HAVE_SKB_GSO_UDP
>> My system's kernel is too new to be built against an OVS this old, but I noticed this patch for OVS versions 2.5 through 2.9 introduce this #ifdef without a corresponding #endif.  Was this intentional?  Does it even build?
>> 
>> --Justin
>> 
>> 
> Conflict resolutions error - I'll probably have to resend branches 2.8 through 2.5.

Thanks.  Can you please at least compile-check them before sending out v2?

--Justin
Gregory Rose Aug. 30, 2019, 3:42 p.m. UTC | #5 
On 8/29/2019 10:44 PM, Justin Pettit wrote:
>> On Aug 29, 2019, at 6:46 PM, Gregory Rose <gvrose8192@gmail.com> wrote:
>>
>>
>> On 8/29/2019 3:25 PM, Justin Pettit wrote:
>>>> On Aug 29, 2019, at 10:55 AM, Greg Rose <gvrose8192@gmail.com> wrote:
>>>>
>>>> diff --git a/datapath/flow.c b/datapath/flow.c
>>>> index 083288f..92fc6ac 100644
>>>> --- a/datapath/flow.c
>>>> +++ b/datapath/flow.c
>>>> @@ -529,6 +529,7 @@ static int key_extract_l3l4(struct sk_buff *skb, struct sw_flow_key *key)
>>>>         offset = nh->frag_off & htons(IP_OFFSET);
>>>>         if (offset) {
>>>>             key->ip.frag = OVS_FRAG_TYPE_LATER;
>>>> +            memset(&key->tp, 0, sizeof(key->tp));
>>>>             return 0;
>>>>         }
>>>>         if (nh->frag_off & htons(IP_MF) ||
>>>> @@ -647,8 +648,11 @@ static int key_extract_l3l4(struct sk_buff *skb, struct sw_flow_key *key)
>>>>             return error;
>>>>         }
>>>>
>>>> -        if (key->ip.frag == OVS_FRAG_TYPE_LATER)
>>>> +        if (key->ip.frag == OVS_FRAG_TYPE_LATER) {
>>>> +            memset(&key->tp, 0, sizeof(key->tp));
>>>>             return 0;
>>>> +        }
>>>> +#ifdef HAVE_SKB_GSO_UDP
>>> My system's kernel is too new to be built against an OVS this old, but I noticed this patch for OVS versions 2.5 through 2.9 introduce this #ifdef without a corresponding #endif.  Was this intentional?  Does it even build?
>>>
>>> --Justin
>>>
>>>
>> Conflict resolutions error - I'll probably have to resend branches 2.8 through 2.5.
> Thanks.  Can you please at least compile-check them before sending out v2?


I'll have to wait for a travis check to finish - I can't seem to get any 
branches older than 2.9 to compile on my
own system anymore due to some Python issues.  So I'll just let travis 
do the work.

Sorry about that, trying to do too many things at once...

- Greg
diff mbox series

Patch

diff --git a/datapath/flow.c b/datapath/flow.c
index 083288f..92fc6ac 100644
--- a/datapath/flow.c
+++ b/datapath/flow.c
@@ -529,6 +529,7 @@  static int key_extract_l3l4(struct sk_buff *skb, struct sw_flow_key *key)
 		offset = nh->frag_off & htons(IP_OFFSET);
 		if (offset) {
 			key->ip.frag = OVS_FRAG_TYPE_LATER;
+			memset(&key->tp, 0, sizeof(key->tp));
 			return 0;
 		}
 		if (nh->frag_off & htons(IP_MF) ||
@@ -647,8 +648,11 @@  static int key_extract_l3l4(struct sk_buff *skb, struct sw_flow_key *key)
 			return error;
 		}
 
-		if (key->ip.frag == OVS_FRAG_TYPE_LATER)
+		if (key->ip.frag == OVS_FRAG_TYPE_LATER) {
+			memset(&key->tp, 0, sizeof(key->tp));
 			return 0;
+		}
+#ifdef HAVE_SKB_GSO_UDP
 		if (skb_shinfo(skb)->gso_type & SKB_GSO_UDP)
 			key->ip.frag = OVS_FRAG_TYPE_FIRST;