[ANN] U-Boot v2020.07-rc2 released

[Heiko Schocher]

Hello Tom,

Am 13.05.2020 um 16:56 schrieb Tom Rini:
> On Wed, May 13, 2020 at 04:52:26PM +0200, Marek Vasut wrote:
>> On 5/13/20 4:49 PM, Tom Rini wrote:
>>> On Wed, May 13, 2020 at 08:13:45AM +0200, Heiko Schocher wrote:
>>>> Hello Heinrich,
>>>>
>>>> Am 13.05.2020 um 07:53 schrieb Heinrich Schuchardt:
>>>>> On 5/13/20 5:54 AM, Heiko Schocher wrote:
>>>>>> Hello Tom,
>>>>>>
>>>>>> Am 12.05.2020 um 15:28 schrieb Heiko Schocher:
>>>>>>> Hello Tom,
>>>>>>>
>>>>>>> Am 12.05.2020 um 00:28 schrieb Tom Rini:
>>>>>>>> Hey all,
>>>>>>>>
>>>>>>>> It's release day and I've tagged v2020.07-rc2.  At this point out we
>>>>>>>> should be seeing stabilization, clean-up and localized new features.
>>>>>>>>
>>>>>>>> Once again, for a changelog,
>>>>>>>> git log --merges v2020.07-rc1..v2020.07-rc2
>>>>>>>> and as always, I ask for more details in the PRs people send me so I can
>>>>>>>> put them in the merge commit.
>>>>>>>>
>>>>>>>> I'm planning to follow the every-other-week RC schedule and release on
>>>>>>>> July 6th.  I'm also thinking about opening -next again on June 8th as
>>>>>>>> that will give us a bit more time to focus on stability and regression
>>>>>>>> fixing.  Thanks all!
>>>>>>>>
>>>>>>>
>>>>>>> FYI:
>>>>>>>
>>>>>>> I have now running the following 4 boards in my daily automated tbot
>>>>>>> testsetup (build and install U-boot on the boards, and run test.py)
>>>>>>>
>>>>>>> wandboard DL (imx6)
>>>>>>> http://xeidos.ddns.net/ubtestresults/result/69
>>>>>>>
>>>>>>> socrates (mpc85xx)
>>>>>>> http://xeidos.ddns.net/ubtestresults/result/70
>>>>>>>
>>>>>>> BBB (am335x)
>>>>>>> http://xeidos.ddns.net/ubtestresults/result/71
>>>>>>>
>>>>>>> aristainetos2 (imx6) (with some patches on top, gpio fixes just posted
>>>>>>> also on this board with secure boot enabled)
>>>>>>> http://xeidos.ddns.net/ubtestresults/result/72
>>>>>>>
>>>>>>> Hope I find time to get this up and running for some of my at91 boards
>>>>>>
>>>>>> May I am nerving with this stuff .. but today morning, I see for the
>>>>>> wandboard with commit 10bca13ea6 the image size increased around ~35K
>>>>>> compared to commit 2a38d2239
>>>>>>
>>>>>> http://xeidos.ddns.net/ubtestresults/stats/wandboard_defconfig/10
>>>>>>
>>>>>> bye,
>>>>>> Heiko
>>>>>
>>>>> Hello Heiko,
>>>>>
>>>>> I am not able to reproduce your 35k increase. This is what I got:
>>>>>
>>>>> 10bca13ea6
>>>>> 668860 u-boot-with-spl.imx
>>>>> 599164 u-boot.bin
>>>>>
>>>>> 2a38d2239
>>>>> 668780 u-boot-with-spl.imx
>>>>> 599084 u-boot.bin
>>>>>
>>>>> 80 additional bytes seems to be in line with the few lines added to
>>>>> drivers/gpio/gpio-uclass.c in 47bd533e9dd0 ("gpio: emulate open drain &
>>>>> open source in dm_gpio_set_value()").
>>>>>
>>>>> I am building on Debian Buster with gcc (Debian 9.3.0-11) 9.3.0.
>>>>>
>>>>> Please, indicate the size of which file is shown in the caption of your
>>>>> graphs.
>>>>
>>>> Yes, I add this, when I am back in the office!
>>>>
>>>>> Please, check if your distribution updated your build system between the
>>>>> two builds.
>>>>
>>>> No, I use the same toolchain:
>>>>
>>>> http://xeidos.ddns.net/ubtestresults/result/75
>>>> http://xeidos.ddns.net/ubtestresults/result/69
>>>>
>>>> Ah, yes my fault! I measured the size of u-boot.img now the size of
>>>> u-boot.bin as I made some changes ...
>>>>
>>>> hs at lab-1:tbot  $ ls -al
>>>> total 1292
>>>> drwxrwxr-x. 2 hs hs   4096 Apr 17 09:29 .
>>>> drwxrwxr-x. 3 hs hs   4096 Apr 17 09:28 ..
>>>> -rw-rw-rw-. 1 hs hs  48128 May 13 04:00 SPL
>>>> -rw-rw-rw-. 1 hs hs  84476 May 13 04:00 System.map
>>>> -rw-rw-rw-. 1 hs hs 597544 May 13 04:00 u-boot.bin
>>>> -rw-rw-rw-. 1 hs hs  15553 May 13 04:00 u-boot.cfg
>>>> -rw-rw-rw-. 1 hs hs 562116 May 13 04:00 u-boot.img
>>>>
>>>> So, sorry for the noise! I add to the database now also the name
>>>> of the image.
>>>
>>> I'm glad you got this sorted out but I would _really_ encourage you to

Yes, sorry for the noise, but it is a proof of concept

I would first add now the posssibility to store more than one U-Boot image
with name and size in the DB ... so I can add this into my graph ...

I think it is very interesting to have the posssibility to see how the
image sizes change over the time

Also I can think of running tests with different toolchains...

>>> make use of the size tracking information we have had in buildman for
>>> forever to track the function size changes, when you build for commit A
>>> and B.  This is what I do for every PR, FWIW.

Yes, adding this is also now on my Todolist...

Thanks!

>> Do we now reflect on that growth in any way ?
> 
> For the past several years, yes.

And it would be good to collect this data somewhere / somehow ...
Thats why I startet with "my" server ...

bye,
Heiko
-- 
DENX Software Engineering GmbH,      Managing Director: Wolfgang Denk
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: +49-8142-66989-52   Fax: +49-8142-66989-80   Email: hs at denx.de