Re: [MPTCP] [GIT] Sync with net-next: conflict

Re: [MPTCP] [GIT] Sync with net-next: conflict

[Paolo Abeni]

[-- Attachment #1: Type: text/plain, Size: 979 bytes --]

On Tue, 2019-05-21 at 16:09 +0200, Matthieu Baerts wrote:
> On 21/05/2019 16:03, Paolo Abeni wrote:
> > On Tue, 2019-05-21 at 14:26 +0200, Matthieu Baerts wrote:
> > > The commit 8b2e02e94442 (tcp: clean ext on tx recycle) from Paolo was
> > > conflicting with a new one from upstream:  (tcp: do not
> > > recycle cloned skbs) from Eric Dumazet.
> > > 
> > > I simply added the new condition (&& !skb_cloned(skb)). I hope it is
> > > alright: a73376082e4c6e3c9a48fd887f759ab1e4118477
> > 
> > The conflict resolution LGTM, thanks!
> > 
> > Not sure if the following actually makes any sense, if the merge commit
> > could be squashed into "tcp: clean ext on tx recycle" that would
> > 
> Sorry, the commit I mentioned is one from the topgit tree. During the
> export, it is "squashed" correctly:
> 
> https://github.com/multipath-tcp/mptcp_net-next/commit/eddf1c63aa7f382376638ab96082dc773c30eb50
> 
> Is it fine for you?

Yep, perfect, thanks!

Paolo

Re: [MPTCP] [GIT] Sync with net-next: conflict

[Matthieu Baerts]

[-- Attachment #1: Type: text/plain, Size: 1120 bytes --]



On 21/05/2019 16:03, Paolo Abeni wrote:
> Hi,
> 
> On Tue, 2019-05-21 at 14:26 +0200, Matthieu Baerts wrote:
>>
>> The commit 8b2e02e94442 (tcp: clean ext on tx recycle) from Paolo was
>> conflicting with a new one from upstream:  (tcp: do not
>> recycle cloned skbs) from Eric Dumazet.
>>
>> I simply added the new condition (&& !skb_cloned(skb)). I hope it is
>> alright: a73376082e4c6e3c9a48fd887f759ab1e4118477
> 
> The conflict resolution LGTM, thanks!
> 
> Not sure if the following actually makes any sense, if the merge commit
> could be squashed into "tcp: clean ext on tx recycle" that would
> simplify/cleanup the git history.

Sorry, the commit I mentioned is one from the topgit tree. During the
export, it is "squashed" correctly:

https://github.com/multipath-tcp/mptcp_net-next/commit/eddf1c63aa7f382376638ab96082dc773c30eb50

Is it fine for you?

Cheers,
Matt

> Cheers,
> 
> Paolo
> 

-- 
Matthieu Baerts | R&D Engineer
matthieu.baerts(a)tessares.net
Tessares SA | Hybrid Access Solutions
www.tessares.net
1 Avenue Jean Monnet, 1348 Louvain-la-Neuve, Belgium

Re: [MPTCP] [GIT] Sync with net-next: conflict

[Paolo Abeni]

[-- Attachment #1: Type: text/plain, Size: 1339 bytes --]

Hi,

On Tue, 2019-05-21 at 14:26 +0200, Matthieu Baerts wrote:
> diff --cc include/net/sock.h
> index fd982f976925,0680fa988497..000000000000
> --- a/include/net/sock.h
> +++ b/include/net/sock.h
> @@@ -1473,8 -1473,7 +1473,12 @@@ static inline void sk_wmem_free_skb(str
>         sock_set_flag(sk, SOCK_QUEUE_SHRUNK);
>         sk->sk_wmem_queued -= skb->truesize;
>         sk_mem_uncharge(sk, skb->truesize);
> ++<<<<<<< t/tcp-clean-ext-on-tx-recycle
>  +      if (!sk->sk_tx_sb_cache) {
>  +              skb_ext_clear(skb);
> ++=======
> +       if (!sk->sk_tx_skb_cache && !skb_cloned(skb)) {
> ++>>>>>>> top-bases/t/tcp-clean-ext-on-tx-recycle
>                 skb_zcopy_clear(skb, true);
>                 sk->sk_tx_skb_cache = skb;
>                 return;
> 
> 
> The commit 8b2e02e94442 (tcp: clean ext on tx recycle) from Paolo was
> conflicting with a new one from upstream:  (tcp: do not
> recycle cloned skbs) from Eric Dumazet.
> 
> I simply added the new condition (&& !skb_cloned(skb)). I hope it is
> alright: a73376082e4c6e3c9a48fd887f759ab1e4118477

The conflict resolution LGTM, thanks!

Not sure if the following actually makes any sense, if the merge commit
could be squashed into "tcp: clean ext on tx recycle" that would
simplify/cleanup the git history.

Cheers,

Paolo

[MPTCP] [GIT] Sync with net-next: conflict

[Matthieu Baerts]

[-- Attachment #1: Type: text/plain, Size: 1303 bytes --]

Hello,

With the last sync with net-next, a conflict was found:


diff --cc include/net/sock.h
index fd982f976925,0680fa988497..000000000000
--- a/include/net/sock.h
+++ b/include/net/sock.h
@@@ -1473,8 -1473,7 +1473,12 @@@ static inline void sk_wmem_free_skb(str
        sock_set_flag(sk, SOCK_QUEUE_SHRUNK);
        sk->sk_wmem_queued -= skb->truesize;
        sk_mem_uncharge(sk, skb->truesize);
++<<<<<<< t/tcp-clean-ext-on-tx-recycle
 +      if (!sk->sk_tx_skb_cache) {
 +              skb_ext_clear(skb);
++=======
+       if (!sk->sk_tx_skb_cache && !skb_cloned(skb)) {
++>>>>>>> top-bases/t/tcp-clean-ext-on-tx-recycle
                skb_zcopy_clear(skb, true);
                sk->sk_tx_skb_cache = skb;
                return;


The commit 8b2e02e94442 (tcp: clean ext on tx recycle) from Paolo was
conflicting with a new one from upstream: 858f50174467 (tcp: do not
recycle cloned skbs) from Eric Dumazet.

I simply added the new condition (&& !skb_cloned(skb)). I hope it is
alright: a73376082e4c6e3c9a48fd887f759ab1e4118477

The new sync has just been pushed.

Cheers,
Matt
-- 
Matthieu Baerts | R&D Engineer
matthieu.baerts(a)tessares.net
Tessares SA | Hybrid Access Solutions
www.tessares.net
1 Avenue Jean Monnet, 1348 Louvain-la-Neuve, Belgium