Angstrom Abandoned for BBB? Rumor + a Rant

[Terry Storm]

A little birdy told me that Angstrom has been abandoned for the BBB ??

If this is the case, and seems to be the case as there hasnt been an update for the Angstrom image for the BBB for quite some time, and there are still a number of issues with LCD CAPES which have not yet been resolved...

What is the 'new' standard distribution for the BBB going to be now?

Us LCD CAPE users are having a hard time getting touch working reliably for a wide range of LCD Capes on the market. I know most of us don't know enough to fix problems ourselves or know what the problem is actually caused by, and we don't know if anyone is working on these issues or if they are being ignored or pass over, or if 'the people' who do know what is going on just don't have time to look at them etc?

Starting to really dislike this whole 'open' community based thing especially when there are issues and no one wants to own them. These issues have been reported for many months, and we are no further ahead.

Does anyone know of anyone who is working on the LCD CAPE touch issue which is a problem on Angstrom (and maybe others which use the same driver/source?)

Terry

[John Syne]

From: Terry Storm <terrys...@gmail.com>
Reply-To: <beagl...@googlegroups.com>
Date: Monday, December 23, 2013 at 1:21 PM
To: <beagl...@googlegroups.com>
Subject: [beagleboard] Angstrom Abandoned for BBB? Rumor + a Rant

A little birdy told me that Angstrom has been abandoned for the BBB ??

If this is the case, and seems to be the case as there hasnt been an update for the Angstrom image for the BBB for quite some time, and there are still a number of issues with LCD CAPES which have not yet been resolved...

What is the 'new' standard distribution for the BBB going to be now?

Us LCD CAPE users are having a hard time getting touch working reliably for a wide range of LCD Capes on the market. I know most of us don't know enough to fix problems ourselves or know what the problem is actually caused by, and we don't know if anyone is working on these issues or if they are being ignored or pass over, or if 'the people' who do know what is going on just don't have time to look at them etc?

Starting to really dislike this whole 'open' community based thing especially when there are issues and no one wants to own them. These issues have been reported for many months, and we are no further ahead.

Hi Terry,

I think you are mistaken in treating this community as some sort of commercial product support. Everyone here is a volunteer and give freely of their own time to support users like yourself. I’m not sure that you will get much help by demanding that works gets done. Also, the ARM based Linux community went through dramatic changes (devicetree, display subsystem, etc) since the V3.2 kernel and we are expecting things to settle down after V3.12. There are users working on the issues your mentioned and I expect that you will see many of these issues resolved in the next several weeks. I just think you stepped into this at the wrong time. BTW, you are always welcome to fix anything that is broken since everyone has access to the same tools and source code; but you won’t be able to do that with commercial software. 

Regards,

John

Does anyone know of anyone who is working on the LCD CAPE touch issue which is a problem on Angstrom (and maybe others which use the same driver/source?)

Terry

--
For more options, visit http://beagleboard.org/discuss
---
You received this message because you are subscribed to the Google Groups "BeagleBoard" group.
To unsubscribe from this group and stop receiving emails from it, send an email to beagleboard...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.

[William Hermans]

Not only what John says, but if you have problems with a specific piece of hardware made by a specific person or entity. You really need to be demanding satisfaction from them. Not the open source community.

[William Hermans]

That is, assuming you paid money for it. If you just copied someones open source schematics . . . well then, you're own your own aren't you ?

[APRichelieu]

Den tisdagen den 24:e december 2013 kl. 00:12:40 UTC+1 skrev
Hermans:ö

[APRichelieu]

[Terry Storm]

Hi

Thanks for the reply.

Yes I have tried contacting the manufacturers however one is a hardware only provider (4D) for the Capes (they fully support their other stuff), and the other one (CircuitCo) which 4D Cares are based on according to their datasheet, doesn't seem to support software and points us here, or simply doesn't reply to emails. I'm not the first person to mention this about CircuitCo by the way. At least 4D do reply and tried to help, but they cant support the software.

Sorry if I sounded demanding, just frustrated more than anything.

Fully understand the community is made up of people who do this for no money etc, however who is actually responsible for open source software, I dont know. No one person I assume, and therefore no place for me to find answers? This is the only place I know to raise it.

In terms of fixing stuff myself. I wouldnt know where to start. I purchased the hardware thinking/hoping that the standard software provided would work correctly. In most cases it is fine, but when dealing with a touch screen that doesnt touch due to software, its kinda frustrating. That said it works fine on Android so it proves the hardware is OK, so CircuitCo and 4D get off the hook in that respect as the hardware is fine, its just the software seems to need some improvements.

I guess all I wanted to hear was what you said, that things are being looked at by someone, and their should be fixes coming out soon. Question though, where does the average Joe Bloggs who came in to the BBB from nothing, has no idea about linux and how everything works, find out what is being worked on so a silly Joe Bloggs like me doesnt make a rant post to try and figure out where they stand...?

So is Angstrom Dead, or is that rumor false?

This new 3.12 you mention, I assume that is the kernel. Is a new Angstrom being built based on that? 

I believe Debian, Ubuntu and Angstrom, no doubt more, have the same touch issues, which is I gather based on the kernel having a problem (3.8?).

I'm a linux noob if you haven't guessed.

Thanks

Terry

[Robert Nelson]

Every developer starts out as noob. Sometimes, all it takes is some
software itch you personally want done...

Regards,

--
Robert Nelson
http://www.rcn-ee.com/

[William Hermans]

Sounds like the perfect "excuse" to start learning.

[Travis Estep]

I think most people make the mistake of thinking the BB and BBB are commercial products with commercial-quality software. They also believe that the capes are just plug-and-play. This board is not for novices. If you are interested in learning electronics, both hardware and software, this is not the board to learn on.

Novices need a firm grasp of how electronics work, especially processors, before moving to this platform. An arduino is a good start, or some variation of an STM32.

Linux is another beast altogether, and experience is key. Download the kernel and dig through it. Experiment with it. Get books about driver development. This is not a subject you will pick up in a few days. The Linux kernel is very complex, and aside from the code, there are concepts you must understand. For me, I didn't understand Linux and how it ran on ARM until I dug into bare metal programming. Once I could make things work without an OS, I moved to Linux and had a far better understanding of what it was doing with my hardware. If you haven't played with starterware yet, check it out. TI gives us all the tools we need to develop code without an OS for this platform.

Good luck!

[Anguel]

On Monday, December 23, 2013 10:21:55 PM UTC+1, Terry Storm wrote:

A little birdy told me that Angstrom has been abandoned for the BBB ??

Terry, I will comment on this topic risking to get some angry feedback from other people posting here. My theory is the following: Koen Kooi was the one who did all Angstrom development and answered all related questions in groups, forums and IRC in the past (just look around). According to his public Google profile he is not working for CircuitCo anymore but for Linaro. Since then, nobody really cares about Angstrom. Angstrom server is down for a long time, no updates, etc. A long time ago, I had also posted some issues + bug fixes to meta-beagleboard but there was no reaction. Most people on #beagle IRC have no idea about Angstrom or Openembedded, so no community support there too. Additionally, Angstrom is a "strange" Linux distro and only a few people know about its details. Afaik it was Koen Kooi who chose it for the BBB for some reason. Another big problem that BBB development faces is the need to switch to device tree. Additionally, I personally don't understand why so much has been invested in some stupid fancy on-board development tools etc. instead of getting the basic things done properly. Maybe this was required by some marketing manager to be able to fool new developers that the BBB is very user friendly and to buy the board...

Regarding the BBB kernel, afaik all the work is done by RobertCNelson who is also working on Ubuntu/Debian development.

So what could be our options for the future of the BBB:
1. We can hope that Ubuntu/Debian for BBB get better and better and finally replace Angstrom as the official distro.
2. Another possibility is that TI relases a new version of their EZSDK with support for a new kernel, device tree, latest Qt and 3D. But according to the work done so far, people don't expect much from TI.
3. My hope is that the very active developers at Buildroot soon have good support for the BBB, so we can get access to new software packages and a very good embedded build system and we could finally forget Angstrom.

No matter what people tell you, BBB is very very very far from the Raspberry Pi's software quality and its huge community. CircuitCo and TI advertise the BBB wherever they can but their promises are far from reality. Why does CircuitCo still advertise their LCDs as working with latest Angstrom after all the discussions I had with them here in the groups?

Many people here tell you that everybody should learn Linux at bare metal level and should be able to write his own kernel drivers to get simple things done. I don't agree, I think that manufacturers like TI and CircuitCo should offer you some working drivers and a stable basis to get started with. We are buying their chips! Currently they don't support us. I have even read postings from expert developers who don't get why the BBB kernel is organized in the way it is, so things are very complex. If you are a new developer you are just not able to learn everything, even if you don't do anything else in your job. IMHO this means: Only wealthy big companies can afford to hire dozens of developers to develop one product they are going to sell in masses. And if this is the only possibility, this is a very very bad thing for Embedded Linux. The BBB and R Pi are wonderful boards for single developers or small companies who have good ideas but don't have the money and time to do everything from scratch. Or like me, they develop products which are needed by e.g. universities but are not expected to ship in large quantities. Unfortunately, the obstacles we currently see with the BBB and most other embedded boards simply prevent such good product ideas to become real products. Small developers and hobbyists are just wasting their time believing the marketing promises of big manufacturers.

The BBB has better HW than the R Pi but in the meantime I ask myself - what is it good for if the board does not have working software and if nobody can help you. So I am seriously considering to switch to the RPi.

Regards,
Anguel

[Robert Nelson]

I really miss beagleboard.org days of old when users actually took
some time to work on the issue they were having and shared a patch/fix
with the community.

[Ulf Samuelsson]

2013-12-29 20:41, Anguel skrev:

On Monday, December 23, 2013 10:21:55 PM UTC+1, Terry Storm wrote:

A little birdy told me that Angstrom has been abandoned for the BBB ??

Terry, I will comment on this topic risking to get some angry feedback from other people posting here. My theory is the following: Koen Kooi was the one who did all Angstrom development and answered all related questions in groups, forums and IRC in the past (just look around). According to his public Google profile he is not working for CircuitCo anymore but for Linaro. Since then, nobody really cares about Angstrom. Angstrom server is down for a long time, no updates, etc.

git clone https://github.com/Angstrom-distribution/setup-scripts.git

git checkout -b Yocto1.5 origin/angstrom-v2013.12-yocto1.5

git log

commit 0b3a4809fafebc742e7a2976ee7570a652547b71
Author: Koen Kooi <ko...@dominion.thruhere.net>
Date:   Mon Dec 16 20:53:41 2013 +0100

    layerman: switch OE-core to angstrom branch
   
    Signed-off-by: Koen Kooi <ko...@dominion.thruhere.net>

commit 55e9feb35c3a43474178e7b1be18c17e3128a611
Author: Koen Kooi <ko...@dominion.thruhere.net>
Date:   Sat Dec 7 09:11:03 2013 +0100

    layerman: switch meta-oe to angstrom staging branch
   
    Signed-off-by: Koen Kooi <ko...@dominion.thruhere.net>

commit 466ad02bec29b1499883f5d8b4540b0949cdd9d6
Author: Koen Kooi <ko...@dominion.thruhere.net>
Date:   Fri Dec 6 18:57:45 2013 +0100


Does not sound that dead to me....

A long time ago, I had also posted some issues + bug fixes to meta-beagleboard but there was no reaction.

Most people on #beagle IRC have no idea about Angstrom or Openembedded, so no community support there too.

You want to post to the openembedded mailing lists.

Additionally, Angstrom is a "strange" Linux distro and only a few people know about its details. Afaik it was Koen Kooi who chose it for the BBB for some reason. Another big problem that BBB development faces is the need to switch to device tree.

When I build Angstrom for the Beaglebone, it is using device trees.

Additionally, I personally don't understand why so much has been invested in some stupid fancy on-board development tools etc. instead of getting the basic things done properly.

Maybe this was required by some marketing manager to be able to fool new developers that the BBB is very user friendly and to buy the board...

Regarding the BBB kernel, afaik all the work is done by RobertCNelson who is also working on Ubuntu/Debian development.

So what could be our options for the future of the BBB:
1. We can hope that Ubuntu/Debian for BBB get better and better and finally replace Angstrom as the official distro.
2. Another possibility is that TI relases a new version of their EZSDK with support for a new kernel, device tree, latest Qt and 3D. But according to the work done so far, people don't expect much from TI.

TI Supports their kernel at
git://git.yoctoproject.org/meta-ti
which contains linux-3.12

3. My hope is that the very active developers at Buildroot soon have good support for the BBB, so we can get access to new software packages and a very good embedded build system and we could finally forget Angstrom.

Angstrom is based on openembedded-core and meta-openembedded, but is not using the standard repos.
You need to check out the Yocto-1.5 branch to get access to the latest stuff though.
If you clone the master, you do not see much development,since this is based on Yocto-1.3 which sees little development.

Buildroot is not Yocto compliant, which means that they will not follow mainstream development funded by large companies.

No matter what people tell you, BBB is very very very far from the Raspberry Pi's software quality and its huge community.
CircuitCo and TI advertise the BBB wherever they can but their promises are far from reality.
Why does CircuitCo still advertise their LCDs as working with latest Angstrom after all the discussions I had with them here in the groups?

Many people here tell you that everybody should learn Linux at bare metal level and should be able to write his own kernel drivers to get simple things done. I don't agree,

I think that manufacturers like TI and CircuitCo should offer you some working drivers and a stable basis to get started with.

I agree 100%.

We are buying their chips! Currently they don't support us.
I have even read postings from expert developers who don't get why the BBB kernel is organized in the way it is, so things are very complex.

I do not understand why you refer it as the BBB kernel.
Angstrom meta-beagleboard allows you to compile the linux kernel with some extra patches for the am335x chips,
using a selected kernel configuration.

This is exactly what they are doing in the Buildroot project, exept that they do not keep track of the beaglebone
so the number of available patches are minimal.

Most embedded chip companies are focused on aligning themselves with the Yocto project.
Yocto is barebone, and Angstom allows for much more advanced functionality.

If you are a new developer you are just not able to learn everything, even if you don't do anything else in your job. IMHO this means: Only wealthy big companies can afford to hire dozens of developers to develop one product they are going to sell in masses. And if this is the only possibility, this is a very very bad thing for Embedded Linux.

There are plenty of small companies which use Embedded Linux.

The BBB and R Pi are wonderful boards for single developers or small companies who have good ideas but don't have the money and time to do everything from scratch. Or like me, they develop products which are needed by e.g. universities but are not expected to ship in large quantities. Unfortunately, the obstacles we currently see with the BBB and most other embedded boards simply prevent such good product ideas to become real products. Small developers and hobbyists are just wasting their time believing the marketing promises of big manufacturers.

The BBB has better HW than the R Pi but in the meantime I ask myself - what is it good for if the board does not have working software and if nobody can help you.
So I am seriously considering to switch to the RPi.

There are plenty of people that can help you, but if you want to get help for free, then that is a different deal.

BTW, How are you going to add an LCD to the RPi, since you only have the HDMI, and no LCD connector?
You can't build your own board.
If you are not happy with the Beaglebone, you can find other AM335x boards where the LCD support may be working.

Regards,
Anguel

--
For more options, visit http://beagleboard.org/discuss
---

You received this message because you are subscribed to a topic in the Google Groups "BeagleBoard" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/beagleboard/WYrk-JUAkbM/unsubscribe.
To unsubscribe from this group and all its topics, send an email to beagleboard...@googlegroups.com.

For more options, visit https://groups.google.com/groups/opt_out.

-- 
Best Regards
Ulf Samuelsson

[William Hermans]

"BTW, How are you going to add an LCD to the RPi, since you only have the HDMI, and no LCD connector?"

SPI ? I2C ? UART ? Pretty sure even the rPI has two of the three listed  . .. even if it does only have 16 or so GPIO's exposed.

You received this message because you are subscribed to the Google Groups "BeagleBoard" group.

To unsubscribe from this group and stop receiving emails from it, send an email to beagleboard...@googlegroups.com.

[William Hermans]

So like all the whining about how circuitco and x-y-z needs to get busy doing drivers for x-y-z feature. I dont know where you guys have been for the last 6-7 months but I have yet to find much of anything that does not work with the BBB. SO the SGX/DRM drivers dont work . . .BFD this is not a computer platform but an embedded system platform.

Seems to me that some of you could apply yourselves more to get whatever it is you wish to work with the hardware. especially considering the cost of the hardware is so small.

[Robert Nelson]

On Sun, Dec 29, 2013 at 6:32 PM, William Hermans <yyr...@gmail.com> wrote:
> So like all the whining about how circuitco and x-y-z needs to get busy
> doing drivers for x-y-z feature. I dont know where you guys have been for
> the last 6-7 months but I have yet to find much of anything that does not
> work with the BBB. SO the SGX/DRM drivers dont work . . .

Actually the "SGX" now work, for John's project... Just under QT...
So it's just us kms/drm/xorg guys are still left out. ;)

[William Hermans]

Sure they work now, but not so much for the last 6-7 months .  . . but the point being there is *NOTHING* I have not personally been able to get working on this wonderful board. Nodejs was a bit of a challenge, as were several Nodejs package managers ( seems this caliber of "programmer" only knows about x 86 ), but i have worked out all my own problems. That is by spending a few days working on them for myself.

What is the point I am trying to make with above comments ? Apply yourselves, and it will very likely work just fine.

[Mike Bremford]

TI Supports their kernel at
git://git.yoctoproject.org/meta-ti
which contains linux-3.12

Angstrom is based on openembedded-core and meta-openembedded, but is not using the standard repos.
You need to check out the Yocto-1.5 branch to get access to the latest stuff though.
If you clone the master, you do not see much development,since this is based on Yocto-1.3 which sees little development.

Buildroot is not Yocto compliant, which means that they will not follow mainstream development funded by large companies.

Where is this documented? And why should I care? The above two paragraphs are unintelligible to anyone that hasn't been involved in embedded Linux for some time.

With the greatest respect to Gerald, Robert and the others who are doing a heroic job and whoever else was involved in designing a very, very nice piece of hardware, the software support is poor. I'm sure this will change at some point but as of today I can choose from a kernel that tends to hang with USB devices (3.8, shipping) or a kernel that doesn't work with capes (3.12 or 3.13rcN). Both are from Robert via a wiki (elinux.org) which isn't associated with the beaglebone website, and it appears he's the only guy working on it. Yet there are apparently several thousand Beaglebones shipping a month?!? That's just crazy.

As for Angstrom, every time a question comes in about "how do I configure this" I wonder why a debian-based distribution wasn't chosen: it would halve the number of messages to this list, and reduce most of the rest to a link to someone else's documentation.

 

No matter what people tell you, BBB is very very very far from the Raspberry Pi's software quality and its huge community. 
CircuitCo and TI advertise the BBB wherever they can but their promises are far from reality. 
Why does CircuitCo still advertise their LCDs as working with latest Angstrom after all the discussions I had with them here in the groups?

Many people here tell you that everybody should learn Linux at bare metal level and should be able to write his own kernel drivers to get simple things done. I don't agree,

I agree with this 100%. I really, really don't want to sound sour and I'm very grateful for the support I've received on this list, but it feels like there should be an organization with resources backing the Beaglebone, and they're just not there.

So like all the whining about how circuitco and x-y-z needs to get busy doing drivers for x-y-z feature. I dont know where you guys have been for the last 6-7 months but I have yet to find much of anything that does not work with the BBB. SO the SGX/DRM drivers dont work . . .BFD this is not a computer platform but an embedded system platform.

USB hotplug doesn't work under 3.8.13, and USB devices tend to crash or hang the kernel - my experience, but it appears form this list I'm not alone. USB works for me with 3.12, but capes do not as the pins can't be remuxed (terminology?) Again, my experience but I believe I am correct and posted my testing results here a few weeks ago.

I believe there is no kernel currently available for download that offers reliable USB hotplug and supports capes, and if anyone can correct me on that with a link to one I would be overjoyed.

Seems to me that some of you could apply yourselves more to get whatever it is you wish to work with the hardware. especially considering the cost of the hardware is so small.

The "if you're not happy why don't you fix it yourself" appeal to open-source rebuttal is a very old one and I'm not sure it flies in this case. Debugging the kernel is beyond most. Athough I'll agree 2 minutes with google might have saved a bit of traffic on this list...

Myself, I was hoping to leverage the platform the BB provides and build on it, rather than spend my days debugging the platform. I'm very experienced with Linux and yet I'm still waiting on someone else to supply a fix; every time I see a message that starts "I'm new to Linux and BBB but I want to do X" I wince, as I know there's many months of poorly charted territory ahead for someone.

Hope that didn't come across as too ranty. I really just wanted to agree with Anguel but got carried away.

[William Hermans]

Aside from hotplug there is nothing wrong with USB in kernel 3.8.x. I have even installed, and booted from a USB hard drive using 3.8.13-bone26.

Personally I think the lot of you are blowing it way out of proportion. This is not some Asus laptop or some Dell desktop / server. This is a hobbyist board, that is meant for hobbyist embedded projects. That means *YOU* the purchaser needs to understand certain things before you even attempt to use such. Hotplug, plug and play etc is the realm of the desktop, not an embedded device which is way more flexible.

So again, we can argue about this until we're both blue in the face. Learn how to use what you have, and be very glad you have one.

[Robert Nelson]

On Sun, Dec 29, 2013 at 8:03 PM, Mike Bremford <mi...@bfo.com> wrote:
>> TI Supports their kernel at
>> git://git.yoctoproject.org/meta-ti
>> which contains linux-3.12
>>
>
>> Angstrom is based on openembedded-core and meta-openembedded, but is not
>> using the standard repos.
>> You need to check out the Yocto-1.5 branch to get access to the latest
>> stuff though.
>> If you clone the master, you do not see much development,since this is
>> based on Yocto-1.3 which sees little development.
>>
>> Buildroot is not Yocto compliant, which means that they will not follow
>> mainstream development funded by large companies.
>>
>
> Where is this documented? And why should I care? The above two paragraphs
> are unintelligible to anyone that hasn't been involved in embedded Linux for
> some time.
>
> With the greatest respect to Gerald, Robert and the others who are doing a
> heroic job and whoever else was involved in designing a very, very nice
> piece of hardware, the software support is poor. I'm sure this will change
> at some point but as of today I can choose from a kernel that tends to hang
> with USB devices (3.8, shipping) or a kernel that doesn't work with capes
> (3.12 or 3.13rcN). Both are from Robert via a wiki (elinux.org) which isn't
> associated with the beaglebone website, and it appears he's the only guy
> working on it. Yet there are apparently several thousand Beaglebones
> shipping a month?!? That's just crazy.

Does it make it more crazy for you, if i told you I don't work for TI
or CircuitCo? Just a hobbiest/user ;)

[APRichelieu]

Den måndagen den 30:e december 2013 kl. 01:29:17 UTC+1 skrev William Hermans:

"BTW, How are you going to add an LCD to the RPi, since you only have the HDMI, and no LCD connector?"

SPI ? I2C ? UART ? Pretty sure even the rPI has two of the three listed  . .. even if it does only have 16 or so GPIO's exposed.

Let me rephrase the question:

"BTW, How are you going to add a 7" WVGA (800 x 480) LCD to the RPi, since you only have the HDMI, and no LCD connector?"

-- 
Best Regards
Ulf Samuelsson

[APRichelieu]

Den måndagen den 30:e december 2013 kl. 03:03:53 UTC+1 skrev Mike Bremford:

TI Supports their kernel at
git://git.yoctoproject.org/meta-ti
which contains linux-3.12

Angstrom is based on openembedded-core and meta-openembedded, but is not using the standard repos.
You need to check out the Yocto-1.5 branch to get access to the latest stuff though.
If you clone the master, you do not see much development,since this is based on Yocto-1.3 which sees little development.

Buildroot is not Yocto compliant, which means that they will not follow mainstream development funded by large companies.

Where is this documented? And why should I care? The above two paragraphs are unintelligible to anyone that hasn't been involved in embedded Linux for some time.

Then the BBB is probably not the right board for You.

People were complaining in this thread, that the support should come directly from the Vendor, and the "meta-ti" is just that, the support from the vendor.
I didn't find this in any documentation, I just looked at the Angstrom layers.txt file, and then googled for meta-ti and found it.
I bet you, that if you look at www.ti.com, you will find links to this repo there.
 

With the greatest respect to Gerald, Robert and the others who are doing a heroic job and whoever else was involved in designing a very, very nice piece of hardware, the software support is poor. I'm sure this will change at some point but as of today I can choose from a kernel that tends to hang with USB devices (3.8, shipping) or a kernel that doesn't work with capes (3.12 or 3.13rcN). Both are from Robert via a wiki (elinux.org) which isn't associated with the beaglebone website, and it appears he's the only guy working on it. Yet there are apparently several thousand Beaglebones shipping a month?!? That's just crazy.

Or you can support my Beaglebone crowd-funding project which fixes the hotplugging issue with  USB on 3.8.13.

As for Angstrom, every time a question comes in about "how do I configure this" I wonder why a debian-based distribution wasn't chosen: it would halve the number of messages to this list, and reduce most of the rest to a link to someone else's documentation.

 

Because the world is going Yocto for Embedded Linux.
Anyway, the selection of Angstrom / Debian / Ubuntu won't solve the problem of a bug inside the kernel.
Only someone actively working on the kernel will solve this problem.

No matter what people tell you, BBB is very very very far from the Raspberry Pi's software quality and its huge community. 
CircuitCo and TI advertise the BBB wherever they can but their promises are far from reality. 
Why does CircuitCo still advertise their LCDs as working with latest Angstrom after all the discussions I had with them here in the groups?

Many people here tell you that everybody should learn Linux at bare metal level and should be able to write his own kernel drivers to get simple things done. I don't agree,

I agree with this 100%. I really, really don't want to sound sour and I'm very grateful for the support I've received on this list, but it feels like there should be an organization with resources backing the Beaglebone, and they're just not there.

So like all the whining about how circuitco and x-y-z needs to get busy doing drivers for x-y-z feature. I dont know where you guys have been for the last 6-7 months but I have yet to find much of anything that does not work with the BBB. SO the SGX/DRM drivers dont work . . .BFD this is not a computer platform but an embedded system platform.

USB hotplug doesn't work under 3.8.13, and USB devices tend to crash or hang the kernel - my experience, but it appears form this list I'm not alone. USB works for me with 3.12, but capes do not as the pins can't be remuxed (terminology?) Again, my experience but I believe I am correct and posted my testing results here a few weeks ago.

I believe there is no kernel currently available for download that offers reliable USB hotplug and supports capes, and if anyone can correct me on that with a link to one I would be overjoyed.

My crowd-funding project is at http://igg.me/at/eMagii/x/5581172
 

[Paulo Ferreira]

On 30/12/2013, at 02:03, Mike Bremford <mi...@bfo.com> wrote:

>
> Where is this documented? And why should I care? The above two paragraphs are unintelligible to anyone that hasn't been involved in embedded Linux for some time.
>

It seems you are "barking at the wrong tree".

Unix is a tool. A powerful tool. As all powerful tools, power should come after some knowledge and practice.

Think of a razor sharp kitchen knife, a chainsaw, an arc welding machine, a forklift, a pickup truck. All those are examples of very useful tools, but they only can be used in a productive way, after some practice time, and after having acquired some knowledge about how they work, and how to use them correctly.


The standard phrase is that Unix is very user friendly, but picky about the friends...

You can approach Unix at several levels:

1) User level - command line use of the Unix utilities, and understanding of shell scripts
2) Admin level - know how to manage users, programs, networking
3) Programmer - know the POSIX programming model and all the UNIX programming tools (config, make, gcc, etc...)
4) Kernel developer - all of the previous ones + how to compile a kernel


If you want to work with BeagleBone, you must at least understand that many people are doing all those levels on the cutting edge of technology,
and that knowledge takes time, because you need to make things, to understand how they work.


The saddest thing, is that people want "things done" (or instant gratification) without "being involved". Open Source does not work that way, and most important, life does not work that way.

In order to do things, in order to get what you want, you need to involve yourself.


Happy New Year to All

Paulo Ferreira

[William Hermans]

Well said Paulo, and I have to agree 100% on the instant gratification / getting something for nothing aspect. 

Also the amount of posts on the group that has nothing to do with with the hardware is staggering. Learning how to use Linux has nothing to do with electronics or the hardware provided by circuitco. I guess the acronym R.T.F.M. is lost on the current generation of users ? 

[Bogdan Teodorescu]

50 % true

The best hardware is useless without an operating systems. Most of this forum users are end-apllications developers, not kernel or embedded Linux developers. they expect quick results based on their current skills

A wise producer will allways provide a solid operating system in order to sel his wonder hardware. Because this hardware is mainly aquired by end applications developers (90%) and not by kernel/Linux developers.

[Micka]

Just an idea, but maybe we should create a software, hardware, kernel mailing list ?

[APRichelieu]

Den måndagen den 30:e december 2013 kl. 11:23:23 UTC+1 skrev Bogdan Teodorescu:

50 % true

The best hardware is useless without an operating systems. Most of this forum users are end-apllications developers, not kernel or embedded Linux developers. they expect quick results based on their current skills

A wise producer will allways provide a solid operating system in order to sel his wonder hardware. Because this hardware is mainly aquired by end applications developers (90%) and not by kernel/Linux developers.

 

It is actually not always so. If you make H/W attractive enough, then other people will make the S/W for you.
Smart customers will require that you have as much as possible ready, so it is a chicken race
between the H/W vendor and the customer.
If the H/W does not bring unique capabilities then you open up to competition by not
supporting your stuff.



 

[Mike Bremford]

Aside from hotplug there is nothing wrong with USB in kernel 3.8.x. I have even installed, and booted from a USB hard drive using 3.8.13-bone26.

Nope, not the case. My kernel was hanging, no logs, nothing on serial console under 3.8.13. No hotplugging, good hub etc., no "babble interrupt" - the console cursor just stopped flashing. I upgraded to 3.12, no other changes, and finally started getting uptime > 48hrs (but my cape no longer worked). Conclusion: 3.8.13 has multiple issues with USB, many of which I see raised repeatedly on this list.

And I respectfully disagree that USB can be considered a "desktop feature" in 2013.

I expected I'd annoy a few people with my post, and the replies are all variations on:

* It's a hobbyist product, you expected it to work?

* It's open source, fix it yourself

* Go learn to use Linux

Which are the same arguments I have seen trotted out since I first installed Linux back in 1995 (kernel 1.2.11, eek). The fact it's so much better now than it was then is down to the efforts of many who thought that position was unsatisfactory.

But ultimately you're missing my points. I come not to bury the Beaglebone but to praise it - it's a great board, but there are aspects of it which could be better and which I criticise only in the hope that my feedback will prove constructive, namely:

1. A single board computer selling tens (hundreds?) of thousands of units and designed to run Linux from the outset, apparently has no-one from the company that designed it (and that presumably profits somewhat from its sales) working on Linux. Community support only goes so far. If Robert decides he has better things to do with his time, what then? This is a computer, not an Atmel AVR, and it's reasonable to treat the OS as part of the package.

2. If you're designing a board with the same price point and form factor as the Raspberry Pi you can't claim it should only be brought by "hobbyists", whatever that means. My hobby (and my career) is writing user-space software, but I need a stable OS to do that, just like kernel developers need stable hardware. You can't build on shaky foundations.

3. Angstrom was a poor choice. The BBB is capable of running Debian or Ubuntu, both of which have hundreds of thousands more developer hours behind them and  considerably better documentation and infrastructure as a result. Just think how much quieter this list would be if all the "how do I use ssh/python/configure my network adapter" questions were on the ubuntu forums?

For my part I'm happy to sit on this list and answer the odd question in my field of expertise, I'm happy to test kernels, and I'm happy to nominate Robert C Nelson as man of the year for his voluntary efforts and to fervently hope he's not hit by a bus. Stay safe, Robert :-)

[David Lambert]

Excellent summary Paulo. I would add one level:
0) Novice user level - Expects a shrink wrapped system to just work
like Windows, but without the crashes ;)

A very Happy New Year to all,

Dave.

[Terry Storm]

Great discussion.

My main issue with this whole system is that its designed for Hobbyists and Developers. Yes developers who know what they are doing can work through issues and develop what they require. Hobbyists however, not so much, obviously depending on the skill of the hobbyist though.

I am a hobbyist, never touched linux on embedded devices before, only played with linux on a PC for as many hours as I have fingers. I brought a LCD4 and a BBB to start with thinking they would both work together and the 'advertised' OS would work 'correctly'. I had touch related issues along with many more, basically rendering the combo useless to do anything without an external mouse, which defeated the whole purpose of what I was trying to achieve. Various issues were fixed over the next few months however the touch issues remained. I then brought a LCD7, and then the 4DCAPE-43T and 4DCAPE-70T, but all had the same touch issues, no doubt stemming from a common cause. None of these boards will do what I want, which is simply to have a working OS with touch screen capability. I wouldn't have thought that was too much to ask for.

I know of many people/companies who have brought BBB's and LCD Capes and wanted to develop systems for them, but due to the OS being so buggy and touch not working correctly, along with other issues, the companies have had to move to another platforms. These companies have linux developers, however not developers who could modify the kernel or write improved drivers, they could write applications to suit their company that run on Linux. Like what was mentioned above by Mike, you have to have a stable OS in order for companies/individuals to write applications on. Since we dont have that, and 90% of people dont have the skills or know-how to fix OS related problems, or write drivers or modify the kernel etc to fix issues before they can even get started.

Personally I think that is a bit rough and harsh to expect people to have to do that.

Yes its open source, however you still need a base to work from that actually works.

I would actually say I purchased the BBB and capes with incorrect information, or to an extent 'false advertising', as these really are not suited to hobbyists. If you have a BBB and want to blink an LED or do some simple Arduino type things with IO, then sure, that is all hobbyist level.

Rant over.

Just think the whole situation sucks to an extent. Circuitco/Ti/who ever is making money, should be putting money back into software development so all the thousands of people who purchased these things and expected some sort of stable system out of the box, or their capes to work as advertised, have working systems they can then build on.

No doubt I will get 'flamed' for what I have written also.

Terry

--
For more options, visit http://beagleboard.org/discuss
--- You received this message because you are subscribed to a topic in the Google Groups "BeagleBoard" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/beagleboard/WYrk-JUAkbM/unsubscribe.

To unsubscribe from this group and all its topics, send an email to beagleboard+unsubscribe@googlegroups.com.

[Gerald Coley]

Circuitco, breaks even, barely. Beagleboard.org makes no money at all.

TI, well they make a little of the chips, but provides no funding for BeagleBoard.org.

Anyone willing to donate some funds so we can fund some SW developers?

Gerald

You received this message because you are subscribed to the Google Groups "BeagleBoard" group.
To unsubscribe from this group and stop receiving emails from it, send an email to beagleboard...@googlegroups.com.

[Mike Bremford]

Aha, well that certainly explains it. Yes, I'd chip in to get cape support added to one of Robert's more recent (3.12 or 3.13) kernels with stable USB. Myself, I'd also be willing to pay slightly more per board (reasonable, as the result would be a better product), but I realise that's more contentious.

[Carl-Fredrik Sundström]

I don't know if there is any truth to this new article but it states that there will be a move towards debian in the next few months.

http://linuxgizmos.com/beaglebone-black-sbc-surpasses-100000-units/

[Gerald Coley]

Yes, we are moving to Debian sometime in the future. Work is underway to make this happen.

Gerald

On Mon, Dec 30, 2013 at 3:52 PM, Carl-Fredrik Sundström <audi...@gmail.com> wrote:

I don't know if there is any truth to this new article but it states that there will be a move towards debian in the next few months.

http://linuxgizmos.com/beaglebone-black-sbc-surpasses-100000-units/

[dave]

Allow me to add my support (AKA 2 cents) to this thread. Noting elsewhere that CircuitCo has shipped over 100,000 units to date, if the cost had included a couple of dollars for software development, that money could have been put to good use. Also adding to the comment on TI, if they don't provide support leading to the demand for the chip, the product will die. It is in the interest of both TI and CircuitCo to see that at least one underlying OS gets supported and will be around for a while. (At this point I don't really care if it's Debian, Ubuntu, Fedora, or even Slackware!)

On my part, I just want an embedded system with Linux that "Just Runs!" I place myself at a level slightly higher than novice since I have used (PC based) Linux in previous projects. Admittedly, this platform is new and there will be some pain in getting there, but I don't like seeing comments of the nature that XXX OS is/will no longer be supported. I want to see support from both TI and CircuitCo to ensure the continuing support of their product. I just want to know that my efforts are not headed for the proverbial drain even before I get it running.

[Gerald Coley]

In order to support the $45 price tag, it was a bear bones offering. Circuitco was paying for the Angstrom support but the maintainer left Circuitco and went to another company. 

TI supports mainline Linux,  currently 3.12 and higher. We are moving in that direction as fast as we can from the 3.3 Kernel.

Angstrom and Linux are distributions of the Linux OS. 

Gerald

[Robert Nelson]

On Mon, Dec 30, 2013 at 3:37 PM, Mike Bremford <mi...@bfo.com> wrote:
> Aha, well that certainly explains it. Yes, I'd chip in to get cape support
> added to one of Robert's more recent (3.12 or 3.13) kernels with stable USB.
> Myself, I'd also be willing to pay slightly more per board (reasonable, as
> the result would be a better product), but I realise that's more
> contentious.

Hey Mike,

I've been playing around with a "stupid" cape manager idea for
v3.12/v3.13, which cape do you need?

[Anguel]

On Monday, December 30, 2013 11:08:42 PM UTC+1, dave wrote:

Allow me to add my support (AKA 2 cents) to this thread. Noting elsewhere that CircuitCo has shipped over 100,000 units to date, if the cost had included a couple of dollars for software development, that money could have been put to good use. Also adding to the comment on TI, if they don't provide support leading to the demand for the chip, the product will die. It is in the interest of both TI and CircuitCo to see that at least one underlying OS gets supported and will be around for a while. (At this point I don't really care if it's Debian, Ubuntu, Fedora, or even Slackware!)

On my part, I just want an embedded system with Linux that "Just Runs!" I place myself at a level slightly higher than novice since I have used (PC based) Linux in previous projects. Admittedly, this platform is new and there will be some pain in getting there, but I don't like seeing comments of the nature that XXX OS is/will no longer be supported. I want to see support from both TI and CircuitCo to ensure the continuing support of their product. I just want to know that my efforts are not headed for the proverbial drain even before I get it running.

Expect anything from TI! My experience with TI started with their ARM Cortex M3. They suddenly dropped all Cortex M3 chips without having any replacement - customers went crazy in the forums! Marketing is the only thing TI can do successfully. Don't expect software support, unfortunately most other manufacturers also don't do better. Now I see that CircuitCo has plans to move to Debian? Just because they lost Koen, their only Linux developer? So I expect them to suddenly drop the BBB completely, as soon as the new Arduino Linux board starts selling successfully. Don't trust anybody in this business anymore.

If people here tell me that there is just nobody who is willing to do work for free I ask what are all these people who bought the thousands of BBB doing with their boards? Are they plugging them in, see that something works as long as you don't try to modify anything and then throwing them away? There must be many companies with enough Linux experience prototyping with the BBB. Why aren't they contributing to the kernel? Or are they just stupid, just working for themselves and ignoring the GPL? And if people here tell me that the BBB is just for fun / hobby then I ask myself, how people use TI chips in real Linux products? What is the difference between a BBB and a custom board? Isn't the BBB a reference platform that should work the best? Do you really believe that TI and CircuitCo are non-profit and build the BBB just because they like the Linux community? That's ridiculous!

Angstrom is full of bugs, I am a Linux guru but even looking at the simplest startup scripts I see that they are full of bashisms etc. Some things in Angstrom work just because of some packages being pulled in by accident. The "production" images are a big mess. Yes, I dived into OpenEmbedded to see that everything is a big mess. It somehow works but don't touch it. Koen was great, it is amazing how he managed to do everything himself. But this is no excuse for TI and CircuitCo. Now I hear they want to sponsor them? I can't believe it...

BTW, anything newer than Angstrom 2012.12 does not work at all. E.g. the Qt cross-toolchain in later Angstrom does not work at all and nobody knows why.

Anguel

[William Hermans]

I'll assume you mean Angstrom and Debian. Which just a factiod for those who do not know already. Angstrom is pretty much already based on Debian. Technically Debian -> Open Zaurus -> Angstrom is the "fork path".

Moving to Debian personally makes total sense to me. Granted I've been using Debian since the 90's and am very familiar with it. However, on the same hand when I speak of Debian on this hardware do realize that I am also a newb when it comes to embedded Linux. Before I got our BBB's here ( 2 ) I had zero hands on with Linux in this capacity, and had never done any Linux development period.

I see this USB babble error mentioned on the groups and have no idea where it comes from and have never experienced it. I have experienced zero problems with USB except that with 3.8.x USB hot plugging will cause a fatal crash. Knowing this, I just do not hot plug USB and there are not problems . . . and it is not as though I do not have as much experience with this board as the next person, because I was working on netbooting, USB booting, this board back in June this year ( when the board was first released to the general public ).

Knowing what I do now, I would not call myself an expert, but I have learned quite a bit in the last 6-7 months. For me  personally, I would never dream of using Angstrom on this fine hardware. As the second week after playing with the board I've had Debian running on it in just about every conceivable way SD card, TFTP/NFS netboot, USB, etc.

Has everything worked perfectly from the start ? Hell no, I had to work at it. I took 2 weeks studying the uEnv.txt file alone. Got every bit of information I could find on the web on my own to help me understand uboot, and how it all worked. Then by the time I understood things well enough to ask a sensible question ( at least i hope it was sensible ), I asked my question, and got pointed to the C header file for uboot configuration/initialization I guess it was.

Anyway, the point being is if *you* want something to work on this platform, you had better work on it yourself. Otherwise, if it does not work already you're going to be stuck getting no where. That and people like me will not care if something does not work for *you* especially considering you're  not even willing to take the time to learn Linux, or even attempt to do sometime about it yourselves.

I suppose a perfect example on this list would be all the conversation about QT not compiling on the BBB . . . I find the topic hilarious on multiple levels. Suffice it to say if you do not already understand why, do not bother asking . . .

[Terry Storm]

Gerald, I find this very interesting.

So given the barebones target price of $45, it seems Marketting and Sales have failed on the BBB totally, as if 100K BBB's have been sold now and Circuitco is breakeven and Beagleboard.org doesnt make anything, then to me it sounds like the project was/is a flop. Hardware is great, but its crap if nothing runs on it out of the box. 

I have 2 BBB's which gather dust at this moment, along with 4 Capes, of which I personally can do nothing about. I would have to spend years learning from nothing in order to do anything productive to get anywhere even close to trying to fix the problems I personally care about. Its not about me personally not putting in the effort to learn Linux, its about capability, and I am not capable. If Koen was the only developer on this thing, I imagine he has a fairly impressive CV regarding linux etc, so myself in comparison is N/A, and I imagine that is the case for 90% of users out there. Mutter Mutter. So peoples comments about everyone else not putting in the effort to learn and fix themselves, I find irrelevant, as not everyone is capable of doing this work. Its like trying to get a truck driver to design a bridge across some massive canyon.

I agree with what was said above, I would have happily paid $50 for a BBB if it meant it actually works software wise.

What has happened is essentially like 'Acme Computers' brought processors off 'Beta Processors Ltd' and used 'Gigasoft OS' as their OS, sold 100K computers to the public and said it does xyz, but it turns out it doesn't work as people expected, and actually requires customer to learn how to modify said Gigasoft OS themselves in order to get the hardware working correctly before they even start the development of their software which they are to then sell to make a profit themselves. Crazy.

Would hate to think of the number of customers who are pissed off at the situation. Would be interesting to see the number of customers who use their share of the 100K of BBB's too. I would imagine a large number purchased, couldn't do much, so have put them in the cupboard and are either waiting for a stable OS or have moved on, or given up etc.

Still cant believe the breakeven after 100K. 

Terry.

[Andrew Frazer]

On Tuesday, December 31, 2013 9:52:52 AM UTC+13, Gerald wrote:

Circuitco, breaks even, barely. Beagleboard.org makes no money at all.

TI, well they make a little of the chips, but provides no funding for BeagleBoard.org.

Anyone willing to donate some funds so we can fund some SW developers?

Gerald

gee.. and people used to buy BBW's for $89 and did'nt really complain.    

 

[William Hermans]

Terry,

I do realize you're mostly aiming your comments at Gerald but let me just say that I think you missed a few key points.

1) This is not a Dell or common general purpose PC. So you can not expect the same from such hardware. With that said, there is nothing wrong with the hardware. It all works fine, and actually up until recently the only thing that did not work was the SGX/DRM video drivers. Now, that has been resolved, but still in alpha / beta stage (  ish ). Different people seem to have experienced different problems here and there, but think this is very likely mostly user error.

2) Do you understand the idea of open source hardware ? *.org is usually and possibly always non profit domain name affiliation. What Gerald and his partners have planned I have no idea, but in my own mind the Beaglebone black is far from a flop. Quite the opposite actually when electronics retailers can hardly seem to keep them in stock, because they sell so fast.

3) If you can pick up a book, or read web pages you can do this yourself. Many in this group will help, even me, when you have a reasonable question to ask. 

Anyways it really suck that this does not seem to be working for you. Just now that many of of are perfectly happy with these little boards. Heck there are a few people who own ten's and possibly even hundred's of these very boards . . .

[APRichelieu]

Den måndagen den 30:e december 2013 kl. 23:08:42 UTC+1 skrev dave:

Allow me to add my support (AKA 2 cents) to this thread. Noting elsewhere that CircuitCo has shipped over 100,000 units to date, if the cost had included a couple of dollars for software development, that money could have been put to good use. Also adding to the comment on TI, if they don't provide support leading to the demand for the chip, the product will die. It is in the interest of both TI and CircuitCo to see that at least one underlying OS gets supported and will be around for a while. (At this point I don't really care if it's Debian, Ubuntu, Fedora, or even Slackware!)

I think that a Semiconductor company should be responsible for the boot loader and the kernel, and provide a Yocto compliant layer
which shows how to build the BSP.
The smart Semiconductor company will ensure that all stuff is merged with the mainstream kernel,
and should not like T.I. maintain its own SDK.

When I worked at Atmel, there was a lot of discussions, but they eventually adopted this approach.
The kernel + BSP is maintained on github, and frequently merged with mainlina.

Freescale has been particularity bad on this, and they typically released a kernel, and then only patched
this kernel, so eventually  you end up with a 2-3 year old kernel with hundreds of patches.
They claim that they have adopted a sane approach with the iMX6, but this is priced
way above the AM335x, making it unattractive for cost reasons.

It is not their expertise to provide something similar to Debian,
so I think they should avoid this.

Asking the semiconductor company to support each and every variant of a graphic LCD screen, probably also won't happen
Touchscreen same thing. An internal touch screen controller, yes. Capacitive touch with an external chip: only if it
is available on a development kit.

Most serious customer has a single configuration, and do not need the flexibility of capes.
It is specific to the Beaglebone, and thus it is not the responsibility of TI to support it.
Cape functionality can be handled through writing a new device tree file.
Some help in writing such files, would be nice, but it is not there at  the moment.

It is unlikely that large customers use Beagleboard at all, except for prototyping.
For any decent volume, people are building their own board.
For a company the size of TI, the Beaglebone project is interesting, but not critical.

If you add lets say 2$ for each Beaglebone, this is $200,000, which will pay for 1-2 man years max,
This is not enough to support a full blown operating system, so a board vendor can provide some
support, but not something which is off the shelf fully tested.

If you want to get a tested linux, with warranty, then you have to sign a contract with
a large company like Wind River or Mentor, and then we are talking about orders of magnitude
higher cost than the Beaglebone.
 
Moving to Debian will help somewhat, but it will not fix problems with the kernel,
nor will it handle beagleboard specific things like capes.

 

[Terry Storm]

Hi William

As I said, I have no problem with the BBB hardware.

I am 100% referring to the software. If the software sucks then to the average Joe Bloggs the hardware then also, well, sucks.

If Joe Bloggs buys a BBB and a cape made by the same company, and there is a recommended software to go with it which claims compatibility, then Joe Bloggs would rightly expect that the BBB + CAPE + OS will run as advertised, and this isnt the case.

I personally am using but a fraction of what the BBB is capable of doing, however I wouldn't have thought using a BBB made by CircuitCo and a LCD Cape made by CircuitCo, running the recommended OS with everything as default, and finding that the touch does not work and things like the mouse pointer jumps all over the screen, kinda not fit for purpose?

If someone like me buys a BBB + Cape and uses an OS which is recommended and claims compatibility, and finds that it doesn't work for something as fundamental as the touch screen, is that classed as user error?

I personally think not.

I stand by my comment of it being a flop, as if the company does not make a profit after selling 100K of a popular product, then really... what was the point.

And in terms of sales and people not being able to get stock, you have to start to wonder how many of those are from users reading the marketting and thinking that will be great, to find when they get it they have no idea how to fix problems they encounter (like me) and so cannot use it as they wanted to.

Many of the happy people probably dont use Capes as you said. It just sucks that the people who designed the cape also design the BBB, and they dont work right purely due to the Software. BBB + LCD Cape on Android works fine, so the hardware is perfectly fine, but Android is of no use to many so it doesnt help much.

Terry

[John Syne]

From: Gerald Coley <ger...@beagleboard.org>
Reply-To: <beagl...@googlegroups.com>
Date: Monday, December 30, 2013 at 2:21 PM
To: "beagl...@googlegroups.com" <beagl...@googlegroups.com>

Subject: Re: [beagleboard] Angstrom Abandoned for BBB? Rumor + a Rant

In order to support the $45 price tag, it was a bear bones offering. Circuitco was paying for the Angstrom support but the maintainer left Circuitco and went to another company. 

TI supports mainline Linux,  currently 3.12 and higher. We are moving in that direction as fast as we can from the 3.3 Kernel.

I explained this previously, but this seems to have been missed in the wining and griping present in this thread. Linus required that the ARM community clean up the sprawling board support code in the Linux kernel. Each ARM vendor duplicated each others work and this became a mess when trying to merge each vendors kernel patches. For this reason, the Linux kernel has undergone significant change (devicetree, display subsystem, clock management, power management, etc) since v3.2 and it is expected to settle down sometime after v3.14. This has little to do with BBB and more to do with all ARM vendors. 

It is just unfortunate that this occurred just as the sales volume of BBB exploded towards the end of 2013. What is needed is a little patients as I see most of these issues are being actively addressed by very capable developers in this community. 

Happy new year

Regards,

John

[Gerald Coley]

I designed the BBB. Other people design the capes. No capes come from BeagleBoard.org. So feel free to rant, but facts do have a place in rants.

Gerald

[liyaoshi]

A part of this group interest me is about the rants 

Nest Time ,if I have chance to be abroad , I will use the skill learned from here 

:)

2013/12/31 Gerald Coley <ger...@beagleboard.org>

[David Anders]

just in case some of the folks here would like to educate themselves on what beagleboard.org is:

http://www.beagleboard.org/about

[liyaoshi]

Come on , its just a joke

2013/12/31 David Anders <dande...@gmail.com>

[William Hermans]

Well I for one am very glad of this hardware, including the software that *does* work with it. Which is to say Debian runs flawlessly, GPIO works fine, and all the other hardware through software works just fine. I would not say that it is exactly the same experience you may get on PC Linux( Debian ), but you know what it is close enough for me(actually very dahmed close ).. For the rest that *I* personally need working. I am willing to spend time to make it work as I expect. Perhaps even asking for help/guidance once in a while.

[John Syne]

From: Terry Storm <terrys...@gmail.com>
Reply-To: <beagl...@googlegroups.com>
Date: Monday, December 30, 2013 at 5:56 PM
To: <beagl...@googlegroups.com>

Subject: Re: [beagleboard] Angstrom Abandoned for BBB? Rumor + a Rant

Hi William

As I said, I have no problem with the BBB hardware.

I am 100% referring to the software. If the software sucks then to the average Joe Bloggs the hardware then also, well, sucks.

If Joe Bloggs buys a BBB and a cape made by the same company, and there is a recommended software to go with it which claims compatibility, then Joe Bloggs would rightly expect that the BBB + CAPE + OS will run as advertised, and this isnt the case.

Actually, you could always use the v3.2 kernel where everything works just fine. All the capes work, all the touch screens work. Take a look at Android for example, they mostly use the v2.6 kernel before v4 and KitKat is still using v3.4. BTW, everything works on v3.4 kernel. The problem is you want to use the bleeding edge version of the Kernel and then complain when it doesn’t work. Go back to the v3.2 or v3.4 kernel and everything will work fine.

Regards,

John

I personally am using but a fraction of what the BBB is capable of doing, however I wouldn't have thought using a BBB made by CircuitCo and a LCD Cape made by CircuitCo, running the recommended OS with everything as default, and finding that the touch does not work and things like the mouse pointer jumps all over the screen, kinda not fit for purpose?

If someone like me buys a BBB + Cape and uses an OS which is recommended and claims compatibility, and finds that it doesn't work for something as fundamental as the touch screen, is that classed as user error?

I personally think not.

I stand by my comment of it being a flop, as if the company does not make a profit after selling 100K of a popular product, then really... what was the point.

And in terms of sales and people not being able to get stock, you have to start to wonder how many of those are from users reading the marketting and thinking that will be great, to find when they get it they have no idea how to fix problems they encounter (like me) and so cannot use it as they wanted to.

Many of the happy people probably dont use Capes as you said. It just sucks that the people who designed the cape also design the BBB, and they dont work right purely due to the Software. BBB + LCD Cape on Android works fine, so the hardware is perfectly fine, but Android is of no use to many so it doesnt help much.

Terry

On Tuesday, 31 December 2013 14:07:17 UTC+13, William Hermans wrote:

Terry,

I do realize you're mostly aiming your comments at Gerald but let me just say that I think you missed a few key points.

1) This is not a Dell or common general purpose PC. So you can not expect the same from such hardware. With that said, there is nothing wrong with the hardware. It all works fine, and actually up until recently the only thing that did not work was the SGX/DRM video drivers. Now, that has been resolved, but still in alpha / beta stage (  ish ). Different people seem to have experienced different problems here and there, but think this is very likely mostly user error.

2) Do you understand the idea of open source hardware ? *.org is usually and possibly always non profit domain name affiliation. What Gerald and his partners have planned I have no idea, but in my own mind the Beaglebone black is far from a flop. Quite the opposite actually when electronics retailers can hardly seem to keep them in stock, because they sell so fast.

3) If you can pick up a book, or read web pages you can do this yourself. Many in this group will help, even me, when you have a reasonable question to ask. 

Anyways it really suck that this does not seem to be working for you. Just now that many of of are perfectly happy with these little boards. Heck there are a few people who own ten's and possibly even hundred's of these very boards . . .

[Terry Storm]

Hi John

If I knew how to do that, I would be very happy indeed.

I have tried Debian, Ubuntu, Angstrom, Android 4.2.2, potentially other distros too that had images available for the BBB, and all that I tried had the same touch issue except Android which works for both the TI 3.2 Kernel version, and I believe is ok for Andrew Hendersons 3.8 Kernel version too...

Can you point me to how I can use Debian or Ubuntu, or even Angstrom if that is easier... using an old Kernel, that has working LCD Cape software for both display and touhscreen. 

This would make a huge number of people happy.

This is one of those things that a person from the 'outside' has no idea about... That an old kernel can be used with an existing version of an OS etc.

I was not actually aware that old versions of the kernel didnt have the problem, as I have only used versions of Angstrom which were released on the normal BBB 'Update' page, that stated compatibility with the LCD4 and LCD7.

If I could be pointed to either images that worked for the LCD4 and LCD7, or some kind of instructions on how something is to be built, that would most certainly help and put my ranting to ease, so I can at least start with some working platform and learn from there.

I appreciate it.

Regards

Terry

[John Syne]

From: Terry Storm <terrys...@gmail.com>
Reply-To: <beagl...@googlegroups.com>

Date: Monday, December 30, 2013 at 10:26 PM

To: <beagl...@googlegroups.com>
Subject: Re: [beagleboard] Angstrom Abandoned for BBB? Rumor + a Rant

Hi John

If I knew how to do that, I would be very happy indeed.

I have tried Debian, Ubuntu, Angstrom, Android 4.2.2, potentially other distros too that had images available for the BBB, and all that I tried had the same touch issue except Android which works for both the TI 3.2 Kernel version, and I believe is ok for Andrew Hendersons 3.8 Kernel version too...

Can you point me to how I can use Debian or Ubuntu, or even Angstrom if that is easier... using an old Kernel, that has working LCD Cape software for both display and touhscreen. 

This would make a huge number of people happy.

This is one of those things that a person from the 'outside' has no idea about... That an old kernel can be used with an existing version of an OS etc.

I was not actually aware that old versions of the kernel didnt have the problem, as I have only used versions of Angstrom which were released on the normal BBB 'Update' page, that stated compatibility with the LCD4 and LCD7.

If I could be pointed to either images that worked for the LCD4 and LCD7, or some kind of instructions on how something is to be built, that would most certainly help and put my ranting to ease, so I can at least start with some working platform and learn from there.

Hi Terry,

I would start by contacting the vendors who build the capes you are using and ask them which kernel version do they currently support. BTW, I wouldn’t call v3.2 kernel old; I would call it stable, which is what you want. 

Linux consists of the kernel (zImage), kernel modules and firmware. Usually, zImage is installed on the FAT partition together with MLO (x-loader), u-boot and uEnv.txt. On the second partition, you find the filesystem (OS) and in the /lib folder you will find the modules and firmware folders. If you look in the modules folder, you will find the kernel modules for each Linux kernel installed. If you look at one of these folders, you will find a folder called kernel. This is where you will find the kernel modules (drivers) for that specific Linux kernel version. When you change the zImage on the FAT partition, it will load the drivers from the correct modules folder. If you want to know which modules folder is current, use ‘uname –r’. 

Best place to find the Linux kernels is Robert Nelsons site which you should find from the wiki (hopefully Robert reads this e-mail and suggests the most stable kernel). From there, follow the wiki to install the zImage, kernel modules and firmware. 

I hope this helps.

Regards,

John

I appreciate it.

Regards

Terry

On Tuesday, 31 December 2013 17:44:35 UTC+13, John Syne wrote:

Actually, you could always use the v3.2 kernel where everything works just fine. All the capes work, all the touch screens work. Take a look at Android for example, they mostly use the v2.6 kernel before v4 and KitKat is still using v3.4. BTW, everything works on v3.4 kernel. The problem is you want to use the bleeding edge version of the Kernel and then complain when it doesn’t work. Go back to the v3.2 or v3.4 kernel and everything will work fine.

Regards,

John

--

[Mike Bremford]

Uh, well for me it's the virtual capes - the actual list I'm loading is:

BB-SPIDEV1

BB-I2C1

BB-UART1

BB-UART2

BB-UART4

BB-UART5

BB-ADC

Plus several of the virtual capes from https://github.com/nomel/beaglebone/tree/master/gpio-header so I can turn on/off individual GPIO pull-ups. But it's not the cape manager so much, it's the ability to change the mode of the pins as described here: https://groups.google.com/forum/#!msg/beagleboard/YqONs50YKWU/VuVsbBhQNl0J.

[Robert Nelson]

Well, i got 3 of these working on 3.13 last night..

On Tue, Dec 31, 2013 at 6:23 AM, Mike Bremford <mi...@bfo.com> wrote:
> Uh, well for me it's the virtual capes - the actual list I'm loading is:
>
> BB-SPIDEV1
> BB-I2C1

Should be able to get these working today, just need someone to test..

> BB-UART1
> BB-UART2
> BB-UART4

These work on bbw & bbb..

> BB-UART5
only bbw (need a good way to disable hdmi on bbb)

> BB-ADC

I think i can get this one to work..

> Plus several of the virtual capes from
> https://github.com/nomel/beaglebone/tree/master/gpio-header so I can turn
> on/off individual GPIO pull-ups. But it's not the cape manager so much, it's
> the ability to change the mode of the pins as described here:
> https://groups.google.com/forum/#!msg/beagleboard/YqONs50YKWU/VuVsbBhQNl0J.

Oh the gpio is going to be fun..

Anywho, upgrade to 3.13-rc6-bone2 (use the install-me.sh from rcn-ee.net/deb/ )

Then on your board:

git clone git://github.com/RobertCNelson/rscm.git
cd ./rscm/

patch -p1 < ./3.13.0-rc6-bone2/simple/enable-ttyO1.diff
patch -p1 < ./3.13.0-rc6-bone2/simple/enable-ttyO2.diff
patch -p1 < ./3.13.0-rc6-bone2/simple/enable-ttyO4.diff

sudo ./build.sh
sudo reboot

At this point it's a total hack i dreamt about, so no error checking/etc..

[Mike Bremford]

I see what you're doing, patching am335x which is presumably loaded directly by the kernel on boot - much simpler than loading after the boot I guess.

So they load no problem, and I have UART1, 2 and 4! Nice work!. I have to dig out a complicated wiring harness to actually test them, but the devices are present - will check for correct later. I'm BBB so UART5 will have to keep. 

Testing ADC, I2C and SPI is easy, the devices are soldered to the board so just let me know what you need me to do. FWIW. saw you had an "enable-i2c-1.diff" in that folder too, tried it but although the /dev/i2c-1 was created, a bus scan incorrectly finds no devices.

Thanks Robert!

[Robert Nelson]

On Tue, Dec 31, 2013 at 10:34 AM, Mike Bremford <mi...@bfo.com> wrote:
> I see what you're doing, patching am335x which is presumably loaded directly
> by the kernel on boot - much simpler than loading after the boot I guess.
>
> So they load no problem, and I have UART1, 2 and 4! Nice work!. I have to
> dig out a complicated wiring harness to actually test them, but the devices
> are present - will check for correct later. I'm BBB so UART5 will have to
> keep.
>
> Testing ADC, I2C and SPI is easy, the devices are soldered to the board so
> just let me know what you need me to do. FWIW. saw you had an
> "enable-i2c-1.diff" in that folder too, tried it but although the /dev/i2c-1
> was created, a bus scan incorrectly finds no devices.

Crap, but thanks for testing! Which pins are you using for i2c1?..

P9.18, P9.17: enable-i2c-1.diff
P9.26, P9.24: enable-i2c-1-alt-pins.diff

[Mike Bremford]

It's P9.18, P9.17 - there's nothing in the logs either I'm afraid.

[Robert Nelson]

On Tue, Dec 31, 2013 at 11:19 AM, Mike Bremford <mi...@bfo.com> wrote:
> It's P9.18, P9.17 - there's nothing in the logs either I'm afraid.

Thanks! i just also enabled i2c2 (eeprom bus) now to find an i2c
device to stick on P9.18/19 from my scrap pile..

[Anguel]

Happy New Year everybody!

Well, reading all those comments here about non-profit and break-even calculations my theory is the following:

Big TI saw the great success of the Raspberry Pi and its engaged community so they decided to copy it. They created the "non-profit" Beagleboard.org and let CircuitCo manufacture the boards - yes, have a look who is actually behind Beagleboard.org! IMHO the idea behind it was to make the whole project independent of TI (no liability for anything) and let the community do all the development work for free. So they hired only one (?) HW developer + only one SW developer (who additionally did all the support for the community and now has left the company).

On the other hand it is very interesting for me to know how much this "non-profit" scheme spends for advertising and marketing, I bet it is a very different number in contrast to the money spent for actual Beaglebone HW + SW development. Talking about all this "non-profit" stuff people overlook the fact that the Beaglebone is actually a very effective large marketing machine working for TI and its Linux processors. Just look at all the articles about the Beaglebones, all the Google Adwords and other advertising in magazines you see around, people wrinting in forums, spending time to make YouTube videos just push TI popularity + sales for free and they do it better than any internal marketing department. Also, I still don't get it why TI should make less money when 100.000 BBBs are sold than they do when selling large quantities of a Linux MCU to a big company...

Unfortunately, as we know, the marketing concept did not work as well as expected: The only SW developer at Beagleboard.org decided to choose the niche Angstrom Linux for some reason. For Angstrom there was no community, no existing docs or books, no forums you could find any questions answered. So developers not only had to know Linux at bare-metal level but they also had to learn complex OpenEmbedded and spend too much time to get simple things done in the Angstrom way, if they did not give up immediately. So nobody really contributed anything to Angstrom Beaglebone development. Another problem was the fact, that TI did not even succeed to deliver a proper kernel with working drivers to the community at that time. They also sticked to their parallel Arago (EZSDK) project instead of working together with the community and Angstrom (Angstrom is a real celebrity in contrast to Arago). Additionally, at this time Linus Torvalds decided to force developers to move to device tree. This was another very huge problem that still has not been solved completely.

Reagardless of the problems above, people continued to buy the BBB because it is cheap and advertisements suggested it was "easy" to use. Development boards are a nice business for any manufacturer because everything is expected to be done by the developer (who is always the one who causes the problem) and there are no warranties at all.

So it looks now that TI / Beagleboard / CircuitCo try to correct the Angstrom concept by moving to Debian, probably again without investing in development. If the community will suddenly become as engaded as with the Raspberry Pi is a different question. But at least they would not have to rely on one single Angstrom developer to solve all their problems.

Anguel

[Elena ``of Valhalla'']

On 2014-01-01 at 04:13:55 -0800, Anguel wrote:
> Well, reading all those comments here about non-profit and break-even
> calculations my theory is the following:
>
> Big TI saw the great success of the Raspberry Pi and its engaged community
> so they decided to copy it. They created the "non-profit" Beagleboard.org

This theory has a problem: Beagleboard.org was born in 2008 or so,
much earlier than the Raspberry (which started to be known to the
public in 2011, and was available in 2012).

Of course the success of the Raspberry did influence BB.org's
products: back in 2008 the standard price for this kind of
boards was around 150$ (e.g. the original BeagleBoard) and it had been
slowly coming down to just below 100$ (e.g. the BeagleBone White):
it was Raspberry and its extreme corner cutting that brought
prices down below 50$, and other producers had to adapt their offerings.

--
Elena ``of Valhalla''

[Anguel]

On Wednesday, January 1, 2014 1:37:02 PM UTC+1, Elena Grandi wrote:

This theory has a problem: Beagleboard.org was born in 2008 or so,
much earlier than the Raspberry (which started to be known to the
public in 2011, and was available in 2012).

Ok, I admit I am not much aware of the way Beagleboard.org worked before the lower-cost Beaglebones were introduced, but it has always been driven by marketing, initially aimed at colleges, according to Wikipedia.
I just want to make clear that big companies don't do anything without profit.

 

Of course the success of the Raspberry did influence BB.org's
products: back in 2008 the standard price for this kind of
boards was around 150$ (e.g. the original BeagleBoard) and it had been
slowly coming down to just below 100$ (e.g. the BeagleBone White):
it was Raspberry and its extreme corner cutting that brought
prices down below 50$, and other producers had to adapt their offerings.

I totally agree. Nobody would buy a BBB for $150 when you can get a Raspberry Pi. But prices of other HW components have probably also dropped significantly since the old days.

Anguel

[APRichelieu]

Den onsdagen den 1:e januari 2014 kl. 13:13:55 UTC+1 skrev Anguel:

Happy New Year everybody!

Well, reading all those comments here about non-profit and break-even calculations my theory is the following:

On the other hand it is very interesting for me to know how much this "non-profit" scheme spends for advertising and marketing, I bet it is a very different number in contrast to the money spent for actual Beaglebone HW + SW development. Talking about all this "non-profit" stuff people overlook the fact that the Beaglebone is actually a very effective large marketing machine working for TI and its Linux processors. Just look at all the articles about the Beaglebones, all the Google Adwords and other advertising in magazines you see around, people wrinting in forums, spending time to make YouTube videos just push TI popularity + sales for free and they do it better than any internal marketing department. Also, I still don't get it why TI should make less money when 100.000 BBBs are sold than they do when selling large quantities of a Linux MCU to a big company...

TI probably have a significant number of projects which has 100ku volume or more, which does not require special support.
If on average each customer buys 10 boards and files 1 request to TI support, then definitely it is more expensive
than a single customer which buys 100,000 chips.
If the chip is $5 and the profit margin is 50%, then they earn $25 per customer.
If an engineer costing $100/hour is spending 15 minutes on answering that question, then 100% of the profit is used up.
Simple math.

Unfortunately, as we know, the marketing concept did not work as well as expected: The only SW developer at Beagleboard.org decided to choose the niche Angstrom Linux for some reason. For Angstrom there was no community, no existing docs or books, no forums you could find any questions answered.

Angstrom is the OpenEmbedded/Yocto project repackaged.
 There are two mailing lists for Angstrom, but the main activity is the OpenEmbedded mailing lists.
The core OpenEmbedded project is not concerned with the BSP for specific parts,
but the semiconductor companies involved with embedded linux. 
(I.E: Intel, TI, Freescale and Atmel) all are making Yocto the recommended
way to build a BSP.
The toolchain used by Angstrom/OpenEmbedded is the Linaro toochain which is funded
by most major mobile phone developers.

I think you should check facts before rarting.
 

So developers not only had to know Linux at bare-metal level but they also had to learn complex OpenEmbedded and spend too much time to get simple things done in the Angstrom way, if they did not give up immediately. So nobody really contributed anything to Angstrom Beaglebone development.

The way it works is that peiople contribute to the OpenEmbedded/Yocto project and then the people behine Angstrom
(which incidently are the core people behind OpenEmbedded) make a snapshot which is available in a branch in the
Angstrom Distribution. The default branch is Yocto 1.3 compliant, but there are Yocto 1,4 and 1.5 versions available as well.
 

Another problem was the fact, that TI did not even succeed to deliver a proper kernel with working drivers to the community at that time. They also sticked to their parallel Arago (EZSDK) project instead of working together with the community and Angstrom (Angstrom is a real celebrity in contrast to Arago). Additionally, at this time Linus Torvalds decided to force developers to move to device tree. This was another very huge problem that still has not been solved completely.

Which is totally unrelated to the choice Angstrom and/or Debian.
 

Reagardless of the problems above, people continued to buy the BBB because it is cheap and advertisements suggested it was "easy" to use. Development boards are a nice business for any manufacturer because everything is expected to be done by the developer (who is always the one who causes the problem) and there are no warranties at all.

 

So it looks now that TI / Beagleboard / CircuitCo try to correct the Angstrom concept by moving to Debian, probably again without investing in development. If the community will suddenly become as engaded as with the Raspberry Pi is a different question. But at least they would not have to rely on one single Angstrom developer to solve all their problems.

Moving to Debian will not solve any kernel problems.
It certainly will not solve problems with capes or any other beaglebone specific stuff,
without extra development.

No sane company expecting to deliver significant volyme would choose a development board
based on a part which cannot be purchased, so I expect that they will continue to sell a lot.

Ulf
 

Anguel

[David Anders]

On Wednesday, January 1, 2014 6:55:38 AM UTC-6, Anguel wrote:

On Wednesday, January 1, 2014 1:37:02 PM UTC+1, Elena Grandi wrote:

This theory has a problem: Beagleboard.org was born in 2008 or so,
much earlier than the Raspberry (which started to be known to the
public in 2011, and was available in 2012).

Ok, I admit I am not much aware of the way Beagleboard.org worked before the lower-cost Beaglebones were introduced, but it has always been driven by marketing, initially aimed at colleges, according to Wikipedia.
I just want to make clear that big companies don't do anything without profit.

here is the thing: TI has nothing to do with beagleboard.org other than they sell the processors to circuitco to make the boards - that's it.....

have a quick read - http://beagleboard.org/about
 

 

Of course the success of the Raspberry did influence BB.org's
products: back in 2008 the standard price for this kind of
boards was around 150$ (e.g. the original BeagleBoard) and it had been
slowly coming down to just below 100$ (e.g. the BeagleBone White):
it was Raspberry and its extreme corner cutting that brought
prices down below 50$, and other producers had to adapt their offerings.

I totally agree. Nobody would buy a BBB for $150 when you can get a Raspberry Pi. But prices of other HW components have probably also dropped significantly since the old days.

here is the thing: in 2007 when beagleboard was started, the only other open hardware platform that was available was the arduino. the idea of an open hardware platform was very new. in addition in 2007 you could not purchase an arm development platform for less than $1000USD! beagleboard was the FIRST arm development platform that cost less than $200 and it was the FIRST arm open hardware platform....

so.... beagleboard has been and always will be the driving force for low cost open hardware platforms...


Dave
 

Anguel

[Hugh Johnson]

> Anyone willing to donate some funds so we can fund some SW developers?
>
> Gerald

Yes.

Thanks,
Hugh Johnson

[William Hermans]

You know what I find "funny" ?

That here is one hell of a piece of hardware ( complete system ), that costs $45( very good price ), that works very well if you're willing to spend some time working on it ( software ). And all a bunch of people on this list can do is complain, find conspiracy theories, or otherwise just try to find something wrong with it or the situation. 

There is *NOTHING* better out there right now as a learning platform. PERIOD. In fact, not only is it a very good learning platform, there are people I know for a fact who use MANY of these in commercial type applications. I have talked to 1-2 people who claimed to have purchased 100 or more of these, and have talked with several people who have over 10. Here, we own 2, and eventually we will probably own many more.

Is the BBB perfect ? Well that is debatable, and really depends on your needs, but I would say no. Nothing ever is. But for the price, there really is nothing to complain about. Now if you're incapable of working with the hardware to make it do what you want. You know what, that is your own fault. Compulsive buyer is probably how I would initially label these types. These people can try and twist words from this source and that, but the truth of the matter is that this system is a learning platform, with very good potential. IF YOU'RE WILLING TO WORK AT IT. Or does every one else here think we're able to have a complete Dell like support situation for $45 ? Right . . . keep dreaming.

Now if these people would come on the list asking for help stating that x-y-z does not work with x-y-z platform, using x-y-z software etc. Many people would offer to try and help, even myself. But the same situation with the exception of complaining about it, pointing fingers, trying to find conspiracy . . . yeah I have absolutely zero sympathy. You know what else, I have no idea who works for TI (still ? ), who does not, and what the motivation is for this project for ANY of the companies involved. Go out, take the schematics / gerber files etc. Send them off to some PCB fab company, and TRY to have even just the PCB's made for the same price as what circuitco sells the PCBA for . . . Gratitude for ya . . .

Now to the guy talking about the Cortex M3 MCU's, and how they seemingly went out of production overnight. If you had paid attention to the errata for these processors in the first place you never would have thought about using it. Not for anything serious anyway. TI replaced these with Cortex M4's very quickly, not to mention there are other Cortex "equivalents" ( geared towards safety, automotive etc. ).

Anyway, I need to stop as I am starting to scare myself by sounding too much like a TI Evangelist ( which I am not) But some of you seriously need to wake up to a  fresh cup of reality.

--

[David Lambert]

Well said William! Initially, I adopted Angstrom as it seemed to be the "official" way that the Beagle community was heading. About a month, however, ago I felt that Angstrom was not meeting my needs as far as stability and "support", so I changed to Debian. No big deal. It only took a day or two, with some kind help from Robert Nelson. I have not regretted the change. Over the past few years, I have used Snapgear, uCLinux, Buildroot, as appropriate. IMHO developing embedded systems requires me to have some flexibility to go with the flow, and not to get too upset if the "flavor of the month" changes from my favorite. My experience with the Beagleboards has been generally very positive, and I have learned to live with the few wrinkles I have found.

Nuff said. Happy New Year to all.


Dave.

[Terry Storm]

What is funny is that it still seems to be expected that everyone who buys a BBB should be required to have the capability to make it work correctly. Your argument is valid but still only for a small proportion of buyers. Think of all the people out there who have brought a BBB to maybe be a step up from their Arduino. They have never used linux etc, its a hell of a learning curve.

It find it hard to believe that it should be expected that everyone who buys a BBB should have had at least (say) 5 years of linux experience before they even consider buying one. This is targeted at a wide range of users, which includes hobbyists who have never touch embedded development boards before. The price of the BBB is irrelevant. What is relevant is there are a ton of people stuck, and a bunch of experienced linux people who are capable of fixing issues and making the BBB work as they require, slagging of everyone else who is not capable and saying they shouldn't have purchased in the first place. Or that all these inexperienced users should harden up and take a concrete pill and figure it out themselves.

Ridiculous.

Swap places and see how far you get.

Making GPIO do things, writing basic application code etc, basically turning a BBB into a glorified Arduino, no problem for most people and that is something that can easily be learnt. Rewriting kernels and drivers so supplied hardware even works, this is a problem for most people. Bringing up a discussion to talk about this, how this constitutes as whining or complaining, I don't know.

This topic has gone way off topic anyway.

End of the day we have great hardware, no one is complaining about that. End of the day we have crap supporting software and now it seems no official (paid) developers, and issues relating to kernels which most users are not capable of fixing or understanding themselves. This is what we are talking about.

Terry.

[John Syne]

From: Terry Storm <terrys...@gmail.com>
Reply-To: <beagl...@googlegroups.com>

Date: Wednesday, January 1, 2014 at 7:05 PM

To: <beagl...@googlegroups.com>
Subject: Re: [beagleboard] Angstrom Abandoned for BBB? Rumor + a Rant

What is funny is that it still seems to be expected that everyone who buys a BBB should be required to have the capability to make it work correctly. Your argument is valid but still only for a small proportion of buyers. Think of all the people out there who have brought a BBB to maybe be a step up from their Arduino. They have never used linux etc, its a hell of a learning curve.

It find it hard to believe that it should be expected that everyone who buys a BBB should have had at least (say) 5 years of linux experience before they even consider buying one. This is targeted at a wide range of users, which includes hobbyists who have never touch embedded development boards before. The price of the BBB is irrelevant. What is relevant is there are a ton of people stuck, and a bunch of experienced linux people who are capable of fixing issues and making the BBB work as they require, slagging of everyone else who is not capable and saying they shouldn't have purchased in the first place. Or that all these inexperienced users should harden up and take a concrete pill and figure it out themselves.

Ridiculous.

Swap places and see how far you get.

Making GPIO do things, writing basic application code etc, basically turning a BBB into a glorified Arduino, no problem for most people and that is something that can easily be learnt. Rewriting kernels and drivers so supplied hardware even works, this is a problem for most people. Bringing up a discussion to talk about this, how this constitutes as whining or complaining, I don't know.

This topic has gone way off topic anyway.

End of the day we have great hardware, no one is complaining about that. End of the day we have crap supporting software and now it seems no official (paid) developers, and issues relating to kernels which most users are not capable of fixing or understanding themselves. This is what we are talking about.

Terry,

I don’t know why you keep on going here. As I said before, use Linux Kernel V3.2 and everything will work as you expect. New users shouldn't use Linux Kernel V3.8 or V3.12 because this is bleeding edge and they don’t have the experience to work around some of the missing features. There is no “crap supporting software”, so please stop saying this. You are completely wrong and you don’t know what you are talking about. You need to read and understand what open source software is all about. Understand what the Linux Kernel version numbers mean and stick to stable releases with long term support. Don’t use the latest Linux kernel which is still work in progress and isn’t meant to be free of errors or full featured. 

Regards,

John 

[cwrse...@gmail.com]

When I first got a BBB around the middle of 2013 it took me a week to sort out a working device tree compiler; that's now a straight download from Robert Nelson's site; and there are a lot of other improvements.  However, the BBB software isn't really stable enough for the long-term, albeit hobbyist, data logging I planned to use it for.  Perhaps it will be in future, but not now.

It would have been better if CircuitCo had increased the BBB price by ten dollars or so, still an astounding bargain, and hired someone to develop the software.

Will

[Anguel]

On Thursday, January 2, 2014 1:08:25 AM UTC+1, David wrote:

Well said William! Initially, I adopted Angstrom as it seemed to be the "official" way that the Beagle community was heading. About a month, however, ago I felt that Angstrom was not meeting my needs as far as stability and "support", so I changed to Debian. No big deal. It only took a day or two, with some kind help from Robert Nelson. I have not regretted the change.

David, do you know if USB gadget support (g_multi) is working on Debian as it does on Angstrom? Also, I experience some strange USB problems with latest libusb(x) when doing asynchronous USB transfers in parallel. These problems are not present on my desktop's Ubuntu nor on my Windows PC, only on the BBB. Will try a newer kernel these days.

Regards,
Anguel

[Robert Nelson]

Yes, I set it up that way on debian/ubuntu to mirror Angstrom..

As long as you aren't using any capes.. (i do give you uart/i2c/spidev
options at the moment..) you can also try the v3.13-rc6 based kernel..

[Anguel]

On Thursday, January 2, 2014 8:35:49 PM UTC+1, RobertCNelson wrote:

> David, do you know if USB gadget support (g_multi) is working on Debian as
> it does on Angstrom? Also, I experience some strange USB problems with
> latest libusb(x) when doing asynchronous USB transfers in parallel. These
> problems are not present on my desktop's Ubuntu nor on my Windows PC, only
> on the BBB. Will try a newer kernel these days.

Yes, I set it up that way on debian/ubuntu to mirror Angstrom..

As long as you aren't using any capes.. (i do give you uart/i2c/spidev
options at the moment..) you can also try the v3.13-rc6 based kernel..

Sounds great! Thanks!

Anguel

[William Hermans]

Robert, so 3.12.x is "done" then ?

--

[Robert Nelson]

On Thu, Jan 2, 2014 at 3:03 PM, William Hermans <yyr...@gmail.com> wrote:
> Robert, so 3.12.x is "done" then ?

I'm personally done with it. ;) If someone wants to keep it going for
rt/xenomai... Or they want to port the 2000+ patches from the ti
"3.12" kernel bsp...

idk... ignoring the capes, more things work on mainline now and i have
a few patches i'm getting ready to push to 3.14..

[Robert Nelson]

On Thu, Jan 2, 2014 at 3:13 PM, Robert Nelson <robert...@gmail.com> wrote:
> On Thu, Jan 2, 2014 at 3:03 PM, William Hermans <yyr...@gmail.com> wrote:
>> Robert, so 3.12.x is "done" then ?
>
> I'm personally done with it. ;) If someone wants to keep it going for
> rt/xenomai... Or they want to port the 2000+ patches from the ti
> "3.12" kernel bsp...

ps: i'm tracking the bsp with this script:

https://github.com/RobertCNelson/ti-linux-kernel

ti-linux-3.12.y branch

suspend works.. ;)

http://paste.debian.net/73848/

http://paste.debian.net/73849/

current from 0.29A -> 0.04A

[Anguel]

Just one question: Is there any particular reason to choose Debian over Ubuntu or the other way around?
Also, should I use Debian on my host when planing to use it on the BBB?

Thanks in advance.

Anguel

[Robert Nelson]

On Thu, Jan 2, 2014 at 4:17 PM, Anguel <anguel....@gmail.com> wrote:
> Just one question: Is there any particular reason to choose Debian over
> Ubuntu or the other way around?
> Also, should I use Debian on my host when planing to use it on the BBB?

In general, Debian developers are just easier to work with... ;)

[Anguel]

On Thursday, January 2, 2014 11:26:15 PM UTC+1, RobertCNelson wrote:

In general, Debian developers are just easier to work with... ;)

That's an argument :) Convinced :)
BTW, are there any known issues cross-compiling Qt4-Embedded on Debian?

Anguel

[Robert Nelson]

I usually stay away from gui's...

John has a good thread on qt 5.1.1

https://groups.google.com/forum/#!category-topic/beagleboard/vw_ZQoq1QNM

[John Syne]

From: Anguel <anguel....@gmail.com>

Reply-To: <beagl...@googlegroups.com>
Date: Thursday, January 2, 2014 at 2:35 PM

To: <beagl...@googlegroups.com>
Subject: Re: [beagleboard] Angstrom Abandoned for BBB? Rumor + a Rant

Hi Anguel,

Do you have any references referring to this problem? I’m using QT5.1.1 and it seems to work just fine. 

Regards,

John

[Anguel]

On Thursday, January 2, 2014 11:37:58 PM UTC+1, RobertCNelson wrote:

 

I usually stay away from gui's...

Yes, but sometiomes they are needed :( By using the embedded version of Qt 4.8 I can do without a window manager. On Angstrom I could easily build an embedded cross-toolchain with Qt.
 

John has a good thread on qt 5.1.1

https://groups.google.com/forum/#!category-topic/beagleboard/vw_ZQoq1QNM

Thanks, will have a look at it, although it looks like he does native compilation on the BBB :( Is quemubuilder an option? I have no idea as I just used a cross-compiler so far.

Anguel

[Anguel]

On Thursday, January 2, 2014 11:48:24 PM UTC+1, John Syne wrote:

From: Anguel <anguel....@gmail.com>

Reply-To: <beagl...@googlegroups.com>
Date: Thursday, January 2, 2014 at 2:35 PM
To: <beagl...@googlegroups.com>
Subject: Re: [beagleboard] Angstrom Abandoned for BBB? Rumor + a Rant

That's an argument :) Convinced :)

BTW, are there any known issues cross-compiling Qt4-Embedded on Debian?

Hi Anguel,

Do you have any references referring to this problem? I’m using QT5.1.1 and it seems to work just fine. 

No, just asking because I read that there is no qt4-embedded package on Debian.

Anguel

[Robert Nelson]

> No, just asking because I read that there is no qt4-embedded package on
> Debian.

The full qt4 is available thou. :)

[David Lambert]

On 01/02/2014 05:00 PM, Robert Nelson wrote:
>> No, just asking because I read that there is no qt4-embedded package on
>> Debian.
> The full qt4 is available thou. :)

As a recent convert to Debian on the BBB. I am pleased to find the
"full" versions of applications. With specifically embedded
distributions I found it irksome to have cut-down busybox versions of
programs such as "less". The BBB has plenty of memory and CPU cycles to
handle the full versions. Progress is great, but sometimes confusing ;-) .


Regards,

Dave.
>
> Regards,
>

[John Syne]

From: Anguel <anguel....@gmail.com>
Reply-To: <beagl...@googlegroups.com>

Date: Thursday, January 2, 2014 at 2:53 PM

To: <beagl...@googlegroups.com>
Subject: Re: [beagleboard] Angstrom Abandoned for BBB? Rumor + a Rant

Hi Anguel,

I’m confused. Are you talking about cross compiling the QT framework for Debian or are you talking about cross compiling QT apps? I tried to cross compile QT for the BBB, but I kept getting some strange error which seems to relate to the incompatible compiler used on the BBB and the version used on my host. BBB used Linaro GCC compiler V4.6 and my host was using v4.7. This shouldn’t have been an issue, but I just found it easier to build it natively on the BBB. 

Cross compiling QT apps for Debian is no problem. 

Regards,

John

[John Syne]

From: Anguel <anguel....@gmail.com>
Reply-To: <beagl...@googlegroups.com>
Date: Thursday, January 2, 2014 at 2:53 PM
To: <beagl...@googlegroups.com>
Subject: Re: [beagleboard] Angstrom Abandoned for BBB? Rumor + a Rant

On Thursday, January 2, 2014 11:37:58 PM UTC+1, RobertCNelson wrote:

 

I usually stay away from gui's...

Yes, but sometiomes they are needed :( By using the embedded version of Qt 4.8 I can do without a window manager. On Angstrom I could easily build an embedded cross-toolchain with Qt.

Have you looked a qtwayland? I haven’t got it to work, but it does look interesting.

 

John has a good thread on qt 5.1.1

https://groups.google.com/forum/#!category-topic/beagleboard/vw_ZQoq1QNM

Thanks, will have a look at it, although it looks like he does native compilation on the BBB :( Is quemubuilder an option? I have no idea as I just used a cross-compiler so far.

Anguel

--

[Anguel]

I totally agree, but the advantage of the Qt Embedded is that you don't need the heavy X Window system:
http://qt-project.org/doc/qt-4.8/qt-embedded-linux.html

Anguel

[Anguel]

On Friday, January 3, 2014 1:20:58 AM UTC+1, John Syne wrote:

Have you looked a qtwayland? I haven’t got it to work, but it does look interesting.

Yes, it seems to be a new lightweight X-Window replacement using QPA in Qt 5. Similar to the good old Qt 4 internal lightweight QWS.
I see Debian has some Wayland packages by default but these probably don't work with the setup you have done?
Sorry for my ignorance, but are you actually installing the TI Graphics SDK on top of Robert's Debian? I mean this:

RootFS: omap-image-builder

debian-7.2-machinekit-armhf

Anguel

[pewte...@gmail.com]

I am new to linux and am following this thread as I find it fascinating even though I have no way of knowing what I am reading most of the time:>)  Give me a year or two and I will be up to speed, but in the mean time I have a simple question. I am wanting to use the BBB to retrofit old machines with new controls, Will BBB be a suitable platform for running a graphical interface (touch screen preferable) for this purpose or should I look at PLC's etc.  Thanks Dan Keith an old dog trying to learn new tricks

[Micka]

Well, for the moment there is a Jitter problem with the touch screen .... ( the mouse is moving on his own just after you touch the screen ) But I heard that there is no problem with the Android Kernel ....

On Fri, Jan 3, 2014 at 5:13 PM, <pewte...@gmail.com> wrote:

I am new to linux and am following this thread as I find it fascinating even though I have no way of knowing what I am reading most of the time:>)  Give me a year or two and I will be up to speed, but in the mean time I have a simple question. I am wanting to use the BBB to retrofit old machines with new controls, Will BBB be a suitable platform for running a graphical interface (touch screen preferable) for this purpose or should I look at PLC's etc.  Thanks Dan Keith an old dog trying to learn new tricks

[John Syne]

From: Anguel <anguel....@gmail.com>
Reply-To: <beagl...@googlegroups.com>
Date: Friday, January 3, 2014 at 2:05 AM
To: <beagl...@googlegroups.com>
Subject: Re: [beagleboard] Angstrom Abandoned for BBB? Rumor + a Rant

Hi Anguel,

Yes, I originally built the TI Graphics SDK and installed it on top of Robert’s Debian release, but now Robert has kindly added a SGX script which I can confirm does the same thing. 

Regards,

John

[Andrew Henderson]

I'll just step in here a moment to say something.  The BBB's software has rough edges.  It is improving all of the time.  Many people are tinkering away behind the scenes, trying out different things, finding the rough edges, talking about them, etc.  I understand the frustration of anyone that wants a working base to start from when creating their own projects.  I receive a great deal of mail from many beginnings that are running into all sorts of issues, and I am either able to help them with their project quickly or get them pointed towards the right place to talk about it.  And it isn't just me doing this sort of thing.  The community has a variety of people that are all tinkering with their own projects and are helping others along the way.  We just don't have a "formal" person to hold accountable when things don't work.

I know that this isn't ideal, and I completely understand where you are coming from.  But, I guess that I just accept it for what it is, sigh, and say "I guess I'm going to have to figure out some workaround for this silly thing".  Most people lack the experience to just dig in and fix things, and that's OK.  The kernel for the BBB is quite a patchwork of, well... patches.  It's a mess, really, especially in the 3.8 kernel.  Little pieces work here and there, things break and get fixed, etc.  I think that you're really going to see everything "working and working well" around the time that the 3.13 kernel becomes the standard one for BBB.  3.12 is looking good, anyway, though it still is missing a few bits here and there.

I remember when the BeagleBoard-xM changed from the 2.6 kernel and it couldn't be clocked at 1 GHz anymore because the support wasn't in the kernel.  People complained.  Oh boy, did they complain.  There were statements like "this is false advertising" and "they said it could do 1 GHz and they LIED", etc.  But really, this is just the name of the game when it comes to these hobbyist boards.

I am sorry that you are frustrated with the current state of the BBB's kernel and OS choices, but it is always changing (and usually improving).  In the meantime, the more experienced hobbyists will keep ironing out the kinks as best as we can, documenting what we find, and really putting a lot of sweat and tears into beating the BBB into what we want it to be.  Personally, I do what I do because I am trying to provide good reference platforms (BeagleSNES as a standalone appliance, Android with the 3.8 kernel for tablet prototypers, etc.) that other people can learn from or use as a base for their own projects.  Sometimes, I fix a minor thing and pass it along to everyone else.  Other times, I just do some investigation, figure out what isn't quite right, and then pass that along to people like Robert that are fiddling away on their own interests.

It is what it is.  It isn't ideal and polished, but then again, some people like me find that kind of interesting.  I know it isn't for everyone, though.

Andrew

[Anguel]

On Friday, January 3, 2014 5:13:17 PM UTC+1, pewte...@gmail.com wrote:

I am new to linux and am following this thread as I find it fascinating even though I have no way of knowing what I am reading most of the time:>)  Give me a year or two and I will be up to speed, but in the mean time I have a simple question. I am wanting to use the BBB to retrofit old machines with new controls, Will BBB be a suitable platform for running a graphical interface (touch screen preferable) for this purpose or should I look at PLC's etc.  Thanks Dan Keith an old dog trying to learn new tricks

As you have probably seen in the discussion, even in 2014 it is not a straightforward task to implement a Touchscreen + GUI on a Linux board like the BBB. There are some annoyances you have to expect.
Also, if you require fast response or even real-time control this is a complex task in Linux as you will very probably need to become very familiar with the kernel.

Anguel

[Anguel]

On Friday, January 3, 2014 8:03:32 PM UTC+1, John Syne wrote:

Hi Anguel,

Yes, I originally built the TI Graphics SDK and installed it on top of Robert’s Debian release, but now Robert has kindly added a SGX script which I can confirm does the same thing. 

Sounds good. Today I wrote Robert's Debian 7.3 image to an SD card and successfully booted the BBB. Then I installed Debian's qt-sdk (which is Qt 4.8.2). Qt complained that it cannot find an X Server, so I first tried to install xserver-xfbdev (which is advertised as very lightweight) but this did not work for some reason. Then I installed LXDE and that worked out fine. Haven't tested thoroughly but I suspect that the 3.8.13 kernel in the Debian 7.3 image does not have 3D accelleration.

Of course the native compilation, even at application level, takes a long time on the BBB. And I am still not sure which is the recommended way to cross-compile Qt apps on the BBB.

If I have some time I will also have a look at Qt 5 with TI's Graphics SDK. I think Qt is very important for the BBB and is required by many people. For non-GUI tasks where fast response times are needed I still prefer a Cortex M3 or M4 at bare-metal level :)

Anguel

[John Syne]

From: Anguel <anguel....@gmail.com>

Reply-To: <beagl...@googlegroups.com>
Date: Friday, January 3, 2014 at 12:41 PM

To: <beagl...@googlegroups.com>
Subject: Re: [beagleboard] Angstrom Abandoned for BBB? Rumor + a Rant

On Friday, January 3, 2014 8:03:32 PM UTC+1, John Syne wrote:

Hi Anguel,

Yes, I originally built the TI Graphics SDK and installed it on top of Robert’s Debian release, but now Robert has kindly added a SGX script which I can confirm does the same thing. 

Sounds good. Today I wrote Robert's Debian 7.3 image to an SD card and successfully booted the BBB. Then I installed Debian's qt-sdk (which is Qt 4.8.2). Qt complained that it cannot find an X Server, so I first tried to install xserver-xfbdev (which is advertised as very lightweight) but this did not work for some reason. Then I installed LXDE and that worked out fine. Haven't tested thoroughly but I suspect that the 3.8.13 kernel in the Debian 7.3 image does not have 3D accelleration.

When you install SGX, you have 3D acceleration. The 3D examples are under /opt. To run the QT app, I use the –platform eglfs. 

Of course the native compilation, even at application level, takes a long time on the BBB. And I am still not sure which is the recommended way to cross-compile Qt apps on the BBB.

If I have some time I will also have a look at Qt 5 with TI's Graphics SDK. I think Qt is very important for the BBB and is required by many people. For non-GUI tasks where fast response times are needed I still prefer a Cortex M3 or M4 at bare-metal level :)

Have you tried Xenomai? I looked at using F28M35C Concerto devices and I always seemed to run out of memory when adding all the pieces I needed such as network stack, security, etc. 

[William Hermans]

Robert, so . . . still no cape support, and I at least need I2C, SPI UART, and possibly a few other things ( definitely standard GPIO ). Would enabling these features be a major undertaking, or would someone such as myself who is new to embedded Linux be able to work out the wrinkles ?

I'd be willing to give it a shot either way, but I would need some guidance as how to proceed.

[Robert Nelson]

On Fri, Jan 3, 2014 at 4:19 PM, William Hermans <yyr...@gmail.com> wrote:
> Robert, so . . . still no cape support, and I at least need I2C, SPI UART,
> and possibly a few other things ( definitely standard GPIO ). Would enabling
> these features be a major undertaking, or would someone such as myself who
> is new to embedded Linux be able to work out the wrinkles ?
>
> I'd be willing to give it a shot either way, but I would need some guidance
> as how to proceed.

So right "now" in my am33x-v3.13 branch, the default out of the box
config is going to be: (request of beagleboard.org guys..)

http://elinux.org/Basic_Proto_Cape

SPISPI0_CS0
SPI0_D1
SPI0_D0
SPI0_SCLK
Enabled (untested by me.) someone want to test, just a /dev/spidev1.x

I2CI2C2_SCL
I2C2_SDA
Enabled (untested by me..)

AIN6
AIN5
AIN4
Enabled, looks like they work..

UARTUART1_TXD
UART1_RXD
enabled, tested by me...

PWMEHRPWM1A
EHRPWM1B
ECAPPWM0
driver not fully in mainline...

GPIOGPIO1_16
GPIO2_2
GPIO2_3
GPIO2_4
GPIO2_5
GPIO0_22
not implemented yet.. hopefully next week..

Beyond that, with:

https://github.com/RobertCNelson/rscm

we can enable:
i2c-1, ttyO2, ttyO3...

More to come...

[William Hermans]

Robert, I would not mind helping out, if I can, I just do not know where to start. Actual testing, I would think you'd want some one familiar with a scope / LA, and that person is not me. YET.

Pretty much what I need in the end is all UARTs, all I2C, and all SPI enabled. My buddy and I have a project that uses every single last pin, and then some . . . but anyhow I would like to help If i can. My own motivation would be to get what I want, and the benefit to the community would be every one gets to use whatever I put together.