[dpdk-stable] Release note for the stable/lts releases

[dpdk-stable] Release note for the stable/lts releases

[Mcnamara, John]

Hi,

As part of the release process for the stable/lts releases, we should update the release notes.

I suggest adding a section like the following, with the applied patches, to the existing release notes in the docs folder:

    Fixes in 16.07.01 Stable Release
    --------------------------------

    The following fixes were applied in DPDK 16.07.01 Stable Release:

    * examples: fix unusual-interpreter
    * ip_pipeline: fix lcore mapping for varying SMT 
    * kni: fix crash for KNI interface remove
    * other patches in alphabetical order
    * ...

Does that seem okay to everyone?

John

Re: [dpdk-stable] Release note for the stable/lts releases

[Christian Ehrhardt]

On Thu, Sep 29, 2016 at 6:39 PM, Mcnamara, John <john.mcnamara@intel.com>
wrote:

> Does that seem okay to everyone?


Seems ok to me at least :-)

Out of backport experience sometimes it is almost more interesting to cover
which ones where not applied and a one liner why.
So if some patch made it to qualify to be looked at, but then be rejected
it would be great to show up in a list:

    The following fixes were NOT applied in DPDK 16.07.01 Stable Release:

    * examples: fix very-very-unusual-interpreter; Reason: too unusual to
happen

Also I'd pretty much like to get a commit id of the original patch in that
list, so
    * kni: fix crash when removing interface
would be
    * [5544a453] kni: fix crash when removing interface

What about these suggestions?


-- 
Christian Ehrhardt
Software Engineer, Ubuntu Server
Canonical Ltd

Re: [dpdk-stable] Release note for the stable/lts releases

[Yuanhan Liu]

On Thu, Sep 29, 2016 at 07:29:58PM +0200, Christian Ehrhardt wrote:
> 
> On Thu, Sep 29, 2016 at 6:39 PM, Mcnamara, John <john.mcnamara@intel.com>
> wrote:
> 
>     Does that seem okay to everyone?
> 
> 
> Seems ok to me at least :-)

+1

> Out of backport experience sometimes it is almost more interesting to cover
> which ones where not applied and a one liner why.
> So if some patch made it to qualify to be looked at, but then be rejected it
> would be great to show up in a list:
> 
>     The following fixes were NOT applied in DPDK 16.07.01 Stable Release:
> 
>     * examples: fix very-very-unusual-interpreter; Reason: too unusual to
> happen

Hmm, is it an example to show rejected patches only? Or, do you actually
mean we should not apply it to stable branch? If so, I'm just wondering
why haven't you stated that while I sent out for comments before.

> 
> Also I'd pretty much like to get a commit id of the original patch in that
> list, so
>     * kni: fix crash when removing interface
> would be
>     * [5544a453] kni: fix crash when removing interface
> 
> What about these suggestions?

Does that matter? I mean, why we should put it in the release note?

OTOH, I have included the original commit info while cherry-pick it from
upstream. For example,

---
commit 6f55e57db50b1729c33a011eafc7704753cb8860
Author: Ferruh Yigit <ferruh.yigit@intel.com>
Date:   Wed Aug 31 17:48:48 2016 +0100

    app/testpmd: fix help of MTU set commmand

    [ upstream commit b2fbb98789d4e7311a415052897d98b6850861de ]

    Fixes: ae03d0d18adf ("app/testpmd: command to configure MTU")

    Signed-off-by: Ferruh Yigit <ferruh.yigit@intel.com>
    Acked-by: Pablo de Lara <pablo.de.lara.guarch@intel.com>

diff --git a/app/test-pmd/cmdline.c b/app/test-pmd/cmdline.c
index f90befc..17d238f 100644
--- a/app/test-pmd/cmdline.c
+++ b/app/test-pmd/cmdline.c
@@ -1367,7 +1367,7 @@ cmdline_parse_token_num_t cmd_config_mtu_value =
 cmdline_parse_inst_t cmd_config_mtu = {
        .f = cmd_config_mtu_parsed,
        .data = NULL,
-       .help_str = "port config mtu value",
+       .help_str = "port config mtu port_id value",
        .tokens = {
                (void *)&cmd_config_mtu_port,
                (void *)&cmd_config_mtu_keyword,


	--yliu

Re: [dpdk-stable] Release note for the stable/lts releases

[Christian Ehrhardt]

On Mon, Oct 10, 2016 at 9:52 AM, Yuanhan Liu <yuanhan.liu@linux.intel.com>
wrote:

>
> Hmm, is it an example to show rejected patches only? Or, do you actually
> mean we should not apply it to stable branch? If so, I'm just wondering
> why haven't you stated that while I sent out for comments before.
>

No objection to the patch implied - as you assumed this was only meant to
be an example.

>
> > Also I'd pretty much like to get a commit id of the original patch in
> that
> > list, so
> >     * kni: fix crash when removing interface
> > would be
> >     * [5544a453] kni: fix crash when removing interface
> >
> > What about these suggestions?
>
> Does that matter? I mean, why we should put it in the release note?
>

You are right no need in the release notes.


> OTOH, I have included the original commit info while cherry-pick it from
> upstream. For example,


I like how you included the original commit info - that provides what I
wanted to see - thanks!



-- 
Christian Ehrhardt
Software Engineer, Ubuntu Server
Canonical Ltd