GSoC'18-Introduction

Christian Mauderer list at c-mauderer.de
Mon Feb 26 20:54:31 UTC 2018 

Am 26.02.2018 um 18:01 schrieb Russell Haley:
> On Mon, Feb 26, 2018 at 12:04 AM, Christian Mauderer
> <christian.mauderer at embedded-brains.de> wrote:
>> Hello Udit,
>>
>> Am 26.02.2018 um 08:30 schrieb Russell Haley:
>>> On Sun, Feb 25, 2018 at 11:08 PM, Udit agarwal <dev.madaari at gmail.com> wrote:
>>>> Hi,
>>>> So, the error turned out to be due to missing dts only, after building the
>>>> scripts again with FDT, both the I2C sample code(Logs:here),
>>>> media01(Logs:here) example worked flawlessly, and thus i was able to read
>>>> sd-card data. What could be my next possible task on libbsd?
>>
>> Most of the time, students are encouraged to fix some small bugs or
>> create some small patches during this phase of GSoC. So you might want
>> to do that.
>>
>> I think there are only few if any (small) open bugs in the RT regarding
>> libbsd. But the BBB BSP currently lacks a getentropy implementation
>> based on the hardware random number generator. That could be a nice
>> small patch and it would increase the strength of any cryptography (like
>> for example WiFi) on the BBB.
>>
>>>> Moreover, I tried building wifi sample app as well, but i have Atheros
>>>> AR9271(Firmware name:  ath9k_htc/htc_9271-1.4.0.fw) wireless card. Following
>>>> sichen's blog posts(here) i tried building support for my wireless card, but
>>>> it looks like FreeBSD itself lacks support for this( as mentioned here )
>>>> Is there any other way, apart from using FreeBSD for generating support for
>>>> my wifi card?
>>>
>>> I know you said "apart from using FreeBSD" but there is a good chance
>>> that partial support for your card is lying around in FreeBSD HEAD. If
>>> you're really keen you could go on the freebsd-arm at freebsd.org mailing
>>> list and ask. Alternatively you pop onto IRC. Adrian and the other
>>> guys on there are the ones that can help you out.
>>>
>>
>> If FreeBSD doesn't support your dongle, it's quite hard to get it
>> working. You would either have to write the driver yourself or find some
>> FreeBSD compatible one. I would say that writing a WiFi driver yourself
>> would be a project that heavily exceeds a GSoC project. For the other
>> solution (finding a driver) Russell already gave the best hints: Talk to
>> the FreeBSD developers. But I would expect that you won't have a quick
>> success there too.
>>
>> So the simplest solution would be to buy some (cheap) WiFi dongle that
>> is supported or can easily be ported.
> 
> Even drivers that work need someone to go through and validate their
> performance. The following is a link to a conversation I had about
> WiFi drivers (albeit a different radios) with marginal or no support.
> While my radio was not supported, the original complainants was.
> 
> http://freebsd.1045724.x6.nabble.com/template/NamlServlet.jtp?macro=search_page&node=3556680&query=Beaglebone+Black+%2B+FreeBSD+%2B+USB+WiFi+%3D+WAP&n=3556680

Yes of course a new driver needs someone to test it too and if some
fixes are necessary that can be a great GSoC project on it's own. @Udit:
Please note that if you want to do that as a project we should get at
least one FreeBSD programmer into the boat.

> 
> While I am unsure of what RTEMS supports, there is also an SDIO driver
> in FreeBSD that is somewhat new that nobody has properly
> tested/compared against the original sdhc driver. I was trying to get
> there but ran out of time (I was using DTrace, RTEMS will have
> different benchmark tools). The original port was on BBB. An excellent
> project would be to go and build the original sdhc driver and the new
> SDIO driver and benchmark them (In FreeBSD, this was done by using two
> different kernel configs). This driver is also in the process of being
> adopted to support SDIO WiFi chips, which are wildly popular now.

With the libbsd RTEMS basically should be able to take over new drivers
quite easily. User space tools are a little harder.

It's really great to hear that there is a SDIO support for FreeBSD in
preparation.

Russells suggestion could make a nice project too. Porting the new
driver and compare different driver versions on different systems would
give us some nice performance charts and could help finding bugs in the
new FreeBSD driver.

@Russell: It seems that you have some experience with FreeBSD. If you
have time and if you feel like it you could think about mentoring or
co-mentoring a RTEMS or rtems-libbsd project.

> 
>>> For What It's Worth,
>>> Russ
>>>
>>>> Moreover, regarding libbsd documentation, i am keeping note of every step,
>>>> and will soon update the document(libbsd.txt).
>>
>> That would be a great patch too. Thanks.
> 
> I can't overstate what a fantastic learning opportunity documentation
> support is:
> 
> - It forces one to go back and validate everything that you *think*
> you know and explain it.
> - It demonstrates that you understand the source material.
> - You learn new tools and and options in the process (RTEMS has a great toolset)
> 
> In my opinion it's far better to partially succeed but have good
> documents, then to finish something once and have none.

Unfortunately we have only few documentation for the libbsd. I agree
that a better documentation would be great. Therefore: Every patch
improving the documentation is great.

Best regards

Christian

> 
> There is my morning randomness, let me know if any of it makes sense. ;)
> 
> Russ
> 
>> Best regards
>>
>> Christian
>>
>>>>
>>>> Thanks,
>>>> Udit Agarwal
>>>>
>>>> On Sun, Feb 25, 2018 at 11:15 PM, Russell Haley <russ.haley at gmail.com>
>>>> wrote:
>>>>>
>>>>> Sorry for the top post.
>>>>>
>>>>> Dts file in freebsd tree are here. The Dts files are imported from Linux
>>>>> for compatability.
>>>>>
>>>>> https://svnweb.freebsd.org/base/head/sys/gnu/dts/arm/
>>>>>
>>>>> Russ
>>>>>
>>>>> Sent from my BlackBerry 10 smartphone on the Virgin Mobile network.
>>>>>   Original Message
>>>>> From: Christian Mauderer
>>>>> Sent: Sunday, February 25, 2018 2:32 AM
>>>>> To: Udit agarwal; Users
>>>>> Subject: Re: GSoC'18-Introduction
>>>>>
>>>>> Am 25.02.2018 um 08:41 schrieb Udit agarwal:
>>>>>> Hi,
>>>>>> I did build libbsd for beagleboneBlack and executed selectpollkqueue01
>>>>>> test successfully.(Logs:here
>>>>>> <https://gist.github.com/madaari/ae7b3334514a9aec3b063eaa3a0dcd05>)
>>>>>> However, while trying to access SD card, media01 test failed with an
>>>>>> error "fatal source: 9 (RTEMS_FATAL_SOURCE_EXCEPTION)  " (Logs:here
>>>>>> <https://gist.github.com/madaari/c6d524bc06e896359f9535d90c0a447a>)
>>>>>> I tried debugging the script, by putting printf statements in various
>>>>>> functions, but it seems like it's failing even before entering
>>>>>> test_main().
>>>>>> Also, is there a better method with which we can debug RTEMS scripts?
>>>>>>
>>>>>> Thanks,
>>>>>> Udit Agarwal
>>>>>
>>>>> Hello Udit,
>>>>>
>>>>> it's great that you can already compile and run some of the libbsd
>>>>> examples.
>>>>>
>>>>>
>>>>> Regarding your problem with the SD card: Note that the BBB BSP uses the
>>>>> flattened device tree (at least the libbsd on BBB does). Therefore you
>>>>> have to load a dtb too. My U-Boot boot lines looks like follows:
>>>>>
>>>>> fatload mmc 0 0x80800000 rtems-app.img
>>>>> fatload mmc 0 0x88000000 am335x-boneblack.dtb
>>>>> bootm 0x80800000 - 0x88000000
>>>>>
>>>>> I create a SD-Card image with a script that is based on one that is (or
>>>>> was) contained some time back in the BSP. You can have a look here:
>>>>>
>>>>>
>>>>> https://gitlab.com/c-mauderer/rtems-bbb/blob/master/build/create-sdcardimage.sh
>>>>>
>>>>> The script does a lot of things that are not necessary any more (like
>>>>> copying a U-Boot to the SD card). Basically the necessary points would
>>>>> be to create a image and an uEnv.txt (with the lines from 78 to 83). You
>>>>> can use the dtb file that is provided by a FreeBSD or a Linux for BBB.
>>>>>
>>>>>
>>>>> Regarding debugging the application: I would strongly recommend to use
>>>>> some kind of hardware JTAG debugger. Some Beagle Bone variants have an
>>>>> on-board debugger (I think mainly the older white one). For others you
>>>>> can use hardware debuggers like quite a number of adapters supported by
>>>>> OpenOCD (FlySwatter2, simple FTDI-based ones, ...), Segger J-Link (in
>>>>> your case the EDU version) or whatever you find that can debug the BBB.
>>>>>
>>>>> It may be possible to use the libdebugger as a serial debugger. But I
>>>>> never used it so I'm not sure how well you can debug through the libbsd
>>>>> with it. I'm also not sure whether it works for the BBB. Chris Johns
>>>>> added the ARM support some time back. Maybe he can give you any hints if
>>>>> you want to try libdebugger.
>>>>>
>>>>> Best Regards
>>>>>
>>>>> Christian Mauderer
>>>>>
>>>>>>
>>>>>> On Fri, Feb 23, 2018 at 11:25 PM, Christian Mauderer <list at c-mauderer.de
>>>>>> <mailto:list at c-mauderer.de>> wrote:
>>>>>>
>>>>>> Am 23.02.2018 um 05:17 schrieb Udit agarwal:
>>>>>>> Hi,
>>>>>>> Thanks for your help mentors,
>>>>>>> Since last time, I have managed to modify and execute hello world
>>>>>>> application on both the platforms, my local machine(Ubuntu 14.04) and
>>>>>>> BeagleBone Black.
>>>>>>> I have attached the screenshots and patch file as proof and  detailed
>>>>>>> logs can be found on these(1
>>>>>>> <https://gist.github.com/madaari/8aae37012d0d175b859c6471a5b13248
>>>>>> <https://gist.github.com/madaari/8aae37012d0d175b859c6471a5b13248>>
>>>>>> and 2
>>>>>>> <https://gist.github.com/madaari/a7f1d223fb0513e4890025fd1ba6184d
>>>>>> <https://gist.github.com/madaari/a7f1d223fb0513e4890025fd1ba6184d>>)
>>>>>> gists.
>>>>>>> Moreover, I went through all the mentioned projects, and found a few
>>>>>>> of
>>>>>>> them(Like extending support for PRU, integration of networking stack
>>>>>>> ,porting RTEMS on PocketBeagle etc) really interesting and doable
>>>>>>> withing the given time constraint!
>>>>>>> So, i thought of starting off by studying a bit about the networking
>>>>>>> stack, its implementation etc. Please recommend few resources from
>>>>>>> where
>>>>>>> i can explore this.
>>>>>>>
>>>>>>> Regards,
>>>>>>> Udit agarwal
>>>>>>>
>>>>>>
>>>>>> Hello Udit,
>>>>>>
>>>>>> it's great that you already have build and successfully executed the
>>>>>> example on BBB.
>>>>>>
>>>>>> Regarding the network stack. First of all: Note that there are at least
>>>>>> three network stacks in RTEMS:
>>>>>>
>>>>>> 1. The "old" one which is integrated into the RTEMS sources. This one
>>>>>> exists since quite some time in RTEMS. It is a fork of a really old
>>>>>> version of the FreeBSD network stack and it has been heavily adapted
>>>>>> during the time. I think there is a manual regarding this stack on the
>>>>>> docs.rtems.org <http://docs.rtems.org>. It has only IPv4 support.
>>>>>>
>>>>>> 2. The "new" one which lives in it's own libbsd repository. That one is
>>>>>> a fork of the latest FreeBSD development version. One of the important
>>>>>> points during the development has been that the sources should be as
>>>>>> close as possible to the upstream FreeBSD sources. This makes it simpler
>>>>>> to keep up to the FreeBSD development. This one can IPv4 and IPv6 and
>>>>>> should be basically able to support everything FreeBSD does. Beneath
>>>>>> that it also brings in USB and SD-Card support and some other subsystems
>>>>>> of FreeBSD.
>>>>>>
>>>>>> 3. I'm quite sure that I have seen some posts about the lwIP
>>>>>> (lightweight IP) stack used together with RTEMS.
>>>>>>
>>>>>>
>>>>>> Most network stack related projects on the open projects page refer to
>>>>>> the libbsd (second one in my list). So you most likely should
>>>>>> concentrate on that one.
>>>>>>
>>>>>> The big disadvantage of the libbsd is that it heavily lacks
>>>>>> documentation. The main documentation is the one big libbsd.txt in the
>>>>>> source tree (https://git.rtems.org/rtems-libbsd/tree/libbsd.txt
>>>>>> <https://git.rtems.org/rtems-libbsd/tree/libbsd.txt>) and
>>>>>> some other files (like CONTRIBUGING.md).
>>>>>>
>>>>>> If you want to work on the libbsd, I would suggest that you try to build
>>>>>> it for the Beagle Bone Black and for example access the SD card content.
>>>>>>
>>>>>> For that I would suggest to take a look at the libbsd.txt and maybe at
>>>>>> the work of last years students (for the libbsd in 2017 that is mainly
>>>>>> Sichen Zhao). Please ask any question you have on the mailing list. And
>>>>>> please write down everything you learn and add it to the libbsd.txt. It
>>>>>> would be a really great start to improve that document.
>>>>>>
>>>>>> Best regards
>>>>>>
>>>>>> Christian Mauderer
>>>>>>
>>>>>>> On Thu, Feb 22, 2018 at 12:28 AM, Christian Mauderer
>>>>>>> <list at c-mauderer.de <mailto:list at c-mauderer.de>
>>>>>>> <mailto:list at c-mauderer.de <mailto:list at c-mauderer.de>>> wrote:
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>     Am 21.02.2018 um 00:14 schrieb Joel Sherrill:
>>>>>>>     >
>>>>>>>     >
>>>>>>>     > On Tue, Feb 20, 2018 at 3:23 PM, Gedare Bloom <gedare at rtems.org
>>>>>>> <mailto:gedare at rtems.org> <mailto:gedare at rtems.org
>>>>>> <mailto:gedare at rtems.org>>
>>>>>>>     > <mailto:gedare at rtems.org <mailto:gedare at rtems.org>
>>>>>> <mailto:gedare at rtems.org <mailto:gedare at rtems.org>>>> wrote:
>>>>>>>     >
>>>>>>>     >     Hello and welcome,
>>>>>>>     >
>>>>>>>     >     For projects that intend to use a specific board, we
>>>>>> require that you
>>>>>>>     >     can demonstrate some proficiency with the board itself.
>>>>>> In this case,
>>>>>>>     >     you should be able to run RTEMS on the BBB first. Now,
>>>>>> there has been
>>>>>>>     >     significant progress made on BBB support, so hopefully
>>>>>> someone may
>>>>>>>     >     chime in with what else might remain to be done with it.
>>>>>> With a little
>>>>>>>     >     bit of digging, you should be able to find prior
>>>>>> students and mentors
>>>>>>>     >     for BBB related projects.
>>>>>>>     >
>>>>>>>     >
>>>>>>>     > My recollection is that the student working on Wifi got it
>>>>>> working
>>>>>>>     with a
>>>>>>>     > USB dongle on either the BBB or Pi. The Wifi stack still
>>>>>> needs work
>>>>>>>     > independent of any BSP and we have listed that as a project.
>>>>>>>     >
>>>>>>>     > But I thought the BBB was in pretty good shape overall. But
>>>>>> that's
>>>>>>>     just
>>>>>>>     > from memory.
>>>>>>>     >
>>>>>>>     > --joel
>>>>>>>     >
>>>>>>>
>>>>>>>     Hello,
>>>>>>>
>>>>>>>     I have been a mentor for Sichen last year. He has done some driver
>>>>>>>     support for the core RTEMS BSP of the BBB last year and he
>>>>>> ported the
>>>>>>>     necessary libbsd drivers to use a USB WiFi dongle on that
>>>>>> board. He also
>>>>>>>     put a lot of work into the encrypted WiFi support.
>>>>>>>
>>>>>>>     I think the year before Punit has worked on the BBB BSP.
>>>>>>>
>>>>>>>     Like Joel already said, the core BBB is quite well supported.
>>>>>> But there
>>>>>>>     are still some drivers like CAN, Ethernet or USB OTG (to name
>>>>>> a few)
>>>>>>>     that could be implemented. I've posted an update to the ticket
>>>>>> on BBB
>>>>>>>     (#2891) some weeks back.
>>>>>>>
>>>>>>>     There are also still two projects regarding the WiFi support
>>>>>> that could
>>>>>>>     be done. The BBB is a really good platform for that. That
>>>>>> would be the
>>>>>>>     tickets #3222 and #3223.
>>>>>>>
>>>>>>>     You can find the tickets linked on the open projects page in
>>>>>> the wiki.
>>>>>>>
>>>>>>>     Regards
>>>>>>>
>>>>>>>     Christian
>>>>>>>
>>>>>>>     >
>>>>>>>     >
>>>>>>>     >     Gedare
>>>>>>>     >
>>>>>>>     >     On Tue, Feb 20, 2018 at 1:56 PM, Udit agarwal
>>>>>> <dev.madaari at gmail.com <mailto:dev.madaari at gmail.com>
>>>>>> <mailto:dev.madaari at gmail.com <mailto:dev.madaari at gmail.com>>
>>>>>>>     >     <mailto:dev.madaari at gmail.com <mailto:dev.madaari at gmail.com>
>>>>>> <mailto:dev.madaari at gmail.com <mailto:dev.madaari at gmail.com>>>> wrote:
>>>>>>>     >     > Hi,
>>>>>>>     >     > I am Udit Agarwal(irc handle: madaari), pursuing my
>>>>>>> undergraduate
>>>>>>>     >     degree in
>>>>>>>     >     > Electronics Engineering from Netaji subash Institute of
>>>>>>>     >     Technology, India.
>>>>>>>     >     > I had some experience of TI-RTOS on MSP432(Cortex M)
>>>>>>>     >     microcontroller, with a
>>>>>>>     >     > pretty good grasp on Beagle Bone Black and other hardware
>>>>>>> stuff.
>>>>>>>     >     I'm pretty
>>>>>>>     >     > excited about the project of improving RTEMS support on
>>>>>>> Beagle
>>>>>>>     >     Bone Black,
>>>>>>>     >     > and would thus like to compete for it.
>>>>>>>     >     > Being already started with RTEMS documentation and the
>>>>>>> Hello World
>>>>>>>     >     project,
>>>>>>>     >     > I would like to know about any other resources i can
>>>>>>> consider
>>>>>>>     >     apart from
>>>>>>>     >     > documentation and RSB guide, for this specific BSP(Beagle
>>>>>>> Bone Black)?
>>>>>>>     >     >
>>>>>>>     >     > Thanks,
>>>>>>>     >     > Udit Agarwal
>>>>>>>     >     >
>>>>>>>     >     > _______________________________________________
>>>>>>>     >     > users mailing list
>>>>>>>     >     > users at rtems.org <mailto:users at rtems.org>
>>>>>>> <mailto:users at rtems.org
>>>>>> <mailto:users at rtems.org>>
>>>>>>>     <mailto:users at rtems.org <mailto:users at rtems.org>
>>>>>> <mailto:users at rtems.org <mailto:users at rtems.org>>>
>>>>>>>     >     > http://lists.rtems.org/mailman/listinfo/users
>>>>>> <http://lists.rtems.org/mailman/listinfo/users>
>>>>>>>     <http://lists.rtems.org/mailman/listinfo/users
>>>>>> <http://lists.rtems.org/mailman/listinfo/users>>
>>>>>>>     >     <http://lists.rtems.org/mailman/listinfo/users
>>>>>> <http://lists.rtems.org/mailman/listinfo/users>
>>>>>>>     <http://lists.rtems.org/mailman/listinfo/users
>>>>>> <http://lists.rtems.org/mailman/listinfo/users>>>
>>>>>>>     >     _______________________________________________
>>>>>>>     >     users mailing list
>>>>>>>     >     users at rtems.org <mailto:users at rtems.org>
>>>>>>> <mailto:users at rtems.org
>>>>>> <mailto:users at rtems.org>>
>>>>>>>     <mailto:users at rtems.org <mailto:users at rtems.org>
>>>>>> <mailto:users at rtems.org <mailto:users at rtems.org>>>
>>>>>>>     >     http://lists.rtems.org/mailman/listinfo/users
>>>>>> <http://lists.rtems.org/mailman/listinfo/users>
>>>>>>>     <http://lists.rtems.org/mailman/listinfo/users
>>>>>> <http://lists.rtems.org/mailman/listinfo/users>>
>>>>>>>     >     <http://lists.rtems.org/mailman/listinfo/users
>>>>>> <http://lists.rtems.org/mailman/listinfo/users>
>>>>>>>     <http://lists.rtems.org/mailman/listinfo/users
>>>>>> <http://lists.rtems.org/mailman/listinfo/users>>>
>>>>>>>     >
>>>>>>>     >
>>>>>>>     >
>>>>>>>     >
>>>>>>>     > _______________________________________________
>>>>>>>     > users mailing list
>>>>>>>     > users at rtems.org <mailto:users at rtems.org>
>>>>>> <mailto:users at rtems.org <mailto:users at rtems.org>>
>>>>>>>     > http://lists.rtems.org/mailman/listinfo/users
>>>>>> <http://lists.rtems.org/mailman/listinfo/users>
>>>>>>>     <http://lists.rtems.org/mailman/listinfo/users
>>>>>> <http://lists.rtems.org/mailman/listinfo/users>>
>>>>>>>     >
>>>>>>>
>>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>> _______________________________________________
>>>>>> users mailing list
>>>>>> users at rtems.org
>>>>>> http://lists.rtems.org/mailman/listinfo/users
>>>>>>
>>>>> _______________________________________________
>>>>> users mailing list
>>>>> users at rtems.org
>>>>> http://lists.rtems.org/mailman/listinfo/users
>>>>
>>>>
>>>>
>>>> _______________________________________________
>>>> users mailing list
>>>> users at rtems.org
>>>> http://lists.rtems.org/mailman/listinfo/users
>>> _______________________________________________
>>> users mailing list
>>> users at rtems.org
>>> http://lists.rtems.org/mailman/listinfo/users
>>>
>>
>> --
>> --------------------------------------------
>> embedded brains GmbH
>> Herr Christian Mauderer
>> Dornierstr. 4
>> D-82178 Puchheim
>> Germany
>> email: christian.mauderer at embedded-brains.de
>> Phone: +49-89-18 94 741 - 18
>> Fax:   +49-89-18 94 741 - 08
>> PGP: Public key available on request.
>>
>> Diese Nachricht ist keine geschäftliche Mitteilung im Sinne des EHUG.
> _______________________________________________
> users mailing list
> users at rtems.org
> http://lists.rtems.org/mailman/listinfo/users
>

More information about the users mailing list