Nabble - Gumstix

Re: Power for both COM _and_ expansion board?

2009-11-25T14:17:53Z

> I'm not aware of any significant performance issues with tobi.

I recently tested the OMAP3 LAN9221 performance using "iperf" running in a standard Angstrom distribution on an Overo based setup. Running against an Ubuntu Linux Server the results are:
Write: ~95Mbit/s
Read : ~70Mbit/s

All in all quite OK. Running two identical OMAP3 LAN9221 sets against each other gives ~70MBit/s for both read and write, which kind of indicates that the ~70Mbit/s limit on read is set on the OMAP3 LAN9221 configuration...

Using the "new" Overo u-boot with NET-support the performance drops to approximately the half. This drop is (most likely) caused by the GPMC configuration done in this U-boot-version, which seems to be rather low(=safe). I'll look into this in more detail (finding the correct GPMC values) as soon I get a little spare time...

Best regards
Søren

---
SSC Solutions ApS - Denmark - www.ssc-solutions.dk

------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
trial. Simplify your report design, integration and deployment - and focus on
what you do best, core application coding. Discover what's new with
Crystal Reports now. http://p.sf.net/sfu/bobj-july
_______________________________________________
gumstix-users mailing list
gumstix-users@...
https://lists.sourceforge.net/lists/listinfo/gumstix-users

Compiling Kernel Module for 2.6.31-omap1

2009-11-25T14:12:28Z

After following the instructions here: http://www.gumstix.net/Setup-and-Programming/view/Overo-Setup-and-Programming/Setting-up-a-build-environment/111.html and compiling a kernel module as explained here: http://old.nabble.com/gpio-event-not-compiling-td26504812.html I found out that my build environment is building my kernel modules for kernel version 2.6.30-omap1 while my Overo is running kernel version 2.6.31-omap1. This is the likely reason for why the kernel module I compiled above will not load on my Overo. So I was wondering if there is a way to upgrade my build environment to build for the 2.6.31-omap1 kernel. My development environment is described in my first post in the mailing list thread above.

Thanks,

Steven Kane

------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
trial. Simplify your report design, integration and deployment - and focus on
what you do best, core application coding. Discover what's new with
Crystal Reports now. http://p.sf.net/sfu/bobj-july
_______________________________________________
gumstix-users mailing list
gumstix-users@...
https://lists.sourceforge.net/lists/listinfo/gumstix-users

Re: Query on usbnet on the overo with summit

2009-11-25T13:48:37Z

On Wed, Nov 25, 2009 at 1:43 PM, Søren Steen Christensen
<lists@...> wrote:
>> Please note that 2.6.31 still isn't enabled by default, so you will
>> have to comment out the DEFAULT_PREFERENCE = "-1" line.
>
> Steve: Do you know why most OMAP3 boards still stick to 2.6.29 and not
> 2.6.31 as "standard" in OE?
> Any drawbacks of running 2.6.31 compared to 2.6.29? I haven't found any
> yet...

Agreed. I've done enough testing that I am ready to make it the
default in the next day or so.

Steve

------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
trial. Simplify your report design, integration and deployment - and focus on
what you do best, core application coding. Discover what's new with
Crystal Reports now. http://p.sf.net/sfu/bobj-july
_______________________________________________
gumstix-users mailing list
gumstix-users@...
https://lists.sourceforge.net/lists/listinfo/gumstix-users

Re: Query on usbnet on the overo with summit

2009-11-25T13:43:08Z

> Please note that 2.6.31 still isn't enabled by default, so you will
> have to comment out the DEFAULT_PREFERENCE = "-1" line.

Steve: Do you know why most OMAP3 boards still stick to 2.6.29 and not
2.6.31 as "standard" in OE?
Any drawbacks of running 2.6.31 compared to 2.6.29? I haven't found any
yet...
Søren

------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
trial. Simplify your report design, integration and deployment - and focus on
what you do best, core application coding. Discover what's new with
Crystal Reports now. http://p.sf.net/sfu/bobj-july
_______________________________________________
gumstix-users mailing list
gumstix-users@...
https://lists.sourceforge.net/lists/listinfo/gumstix-users

Build failing... Upstream u-boot repo changed again??

2009-11-25T13:36:23Z

When trying to bitbake the Overo console image today on a new dev
machine, it again is failing at the u-boot-omap3 fetch step. I already
tried updating everything via git pull, etc... The recipe I have:

-----------------------------------------------------
require u-boot.inc

FILESDIR =
"${@os.path.dirname(bb.data.getVar('FILE',d,1))}/u-boot-omap3-git/"

SRCREV = "4bc3d2afb380e78fdbb9c501d9a8da6d59eb178e"

PV = "2009.08+${PR}+git${SRCREV}"
PR = "r11"
PE = "1"

SRC_URI = "git://git.denx.de/u-boot-ti.git;protocol=git \
file://fw_env.config \
file://tincan.patch;patch=1 \
"

S = "${WORKDIR}/git"
-----------------------------------------------------

I can't find a matching snapshot on git.denx.de.... And am too much of a
newb to try and update the recipe myself... Hints?

- Jim

__

James M. Pettinato, Jr.
Software Engineer
FMC Technologies Measurement Solutions Inc.

------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
trial. Simplify your report design, integration and deployment - and focus on
what you do best, core application coding. Discover what's new with
Crystal Reports now. http://p.sf.net/sfu/bobj-july
_______________________________________________
gumstix-users mailing list
gumstix-users@...
https://lists.sourceforge.net/lists/listinfo/gumstix-users

Re: Query on usbnet on the overo with summit

2009-11-25T13:24:28Z

On Wed, Nov 25, 2009 at 12:30 PM, ScottEllis
<scottellis.developer@...> wrote:

>
> The OTG USB port isn't configured for USB networking in the stock Gumstix
> kernels.
>
> I started a document here explaining what you need to do or at least one way
> to
> go about doing it.
>
> http://github.com/scottellis/overo-usbnet
>
> It's not ready for prime-time, but you might take something from it.
>
> Any feedback is welcome.

The linux-omap3 recipe for 2.6.31 makes it a bit simpler to switch the
operating mode for the musb port.

See the following commit for details:

http://www.sakoman.net/cgi-bin/gitweb.cgi?p=openembedded.git;a=commitdiff;h=9a97cf6dc270bd415a571d66712cc790811a3655

Please note that 2.6.31 still isn't enabled by default, so you will
have to comment out the DEFAULT_PREFERENCE = "-1" line.

Steve

------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
trial. Simplify your report design, integration and deployment - and focus on
what you do best, core application coding. Discover what's new with
Crystal Reports now. http://p.sf.net/sfu/bobj-july
_______________________________________________
gumstix-users mailing list
gumstix-users@...
https://lists.sourceforge.net/lists/listinfo/gumstix-users

Couple Gumstix questions

2009-11-25T12:47:50Z

Hi! I'm thinking about using the gumstix board for a handheld computer project. Before I buy the board, an expansion board, a LCD screen, I have a couple questions about the following....

1) Does the Chestnut 43 have a vcc and vdd on the board? I'm thinking about building my own 5 volt power supply for the board.

2) Is it possible to install Android on the board? If so, then how can you do so?

3) Is it really difficult to configure the gumstix board to use the LCD screen even with the Chestnut 43?

Thank you!

Re: gpio-event not compiling

2009-11-25T12:31:05Z

Dave,

   It looks like you may be correct about the vermagic.

Here are the commands with their results:

On my build machine:

file gpio-event-drv.ko
> gpio-event-drv.ko: ELF 32-bit LSB relocatable, ARM, version 1 (SYSV), not stripped

modinfo gpio-event-drv.ko

> filename:       gpio-event-drv.ko
> license:        Dual BSD/GPL
> description:    GPIO Event Driver
> author:         Dave Hylands
> srcversion:     F653FD1904BE521B59BE66D
> depends:
> vermagic:       2.6.30-omap1 mod_unload modversions ARMv7

On my overo:

uname -r

> 2.6.31-omap1

So it appears the kernel was built for version 2.30-omap1 but my Overo is actually running 2.6.31. So at this point I am trying to figure out how to update my build environment to 2.6.31. Do you (or anyone else) know if I would have to use a different branch of gumstix-oe? I only see the mainline available here: http://gitorious.org/gumstix-oe. In any event I can start another thread on that topic alone...

Thanks for your help,

Steven Kane

On Wed, Nov 25, 2009 at 11:56 AM, Dave Hylands <dhylands@...> wrote:

Hi Steven,

On Wed, Nov 25, 2009 at 11:11 AM, Steven Kane <steven.kane@...> wrote:
> Dave,
>
> I was able to get 'gpio-event' to compile for the Overo after applying
> some tweaks to 'gpio-event-drv.c', 'gpio-event.bb' and Makefile. Attached
> are these modified files. However, when I installed the package on the
> Overo, I successfully added a pin to be monitored with the command:
> 'gpio-event 144:r:20' as documented here:
> http://docwiki.gumstix.org/GPIO_event, but 'gpio-event' did not show up in
> '/dev' or '/proc'. Also, when I ran the command 'insmod gpio-event-drv.ko' I
> got the error: 'insmod: error inserting 'gpio-event-drv.ko': -1 Invalid
> module format'. Any ideas at this point?

The first thing I would do, is to run the file command on the
gpio-event-drv.ko file. I don't know if the file command is on the
gumstix, but it's most likely on your build machine. Make sure that
the .ko file was built for the ARM architecture.

You should also be able to run modinfo (on your build machine for sure
- maybe on the gumstix).

I believe that there may also be some additional information logged in
the /var/log/messages file.

Typically, the problem with vermagic mismatch is caused because the
version of toolchain/linux that the module was built for doesn't match
the versions of the kernel running on the gumstix (i.e. your build
environment is newer or older than the one used to build the image on
your gumstix).

--

Dave Hylands
Shuswap, BC, Canada
http://www.DaveHylands.com/

------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
trial. Simplify your report design, integration and deployment - and focus on
what you do best, core application coding. Discover what's new with
Crystal Reports now. http://p.sf.net/sfu/bobj-july
_______________________________________________
gumstix-users mailing list
gumstix-users@...
https://lists.sourceforge.net/lists/listinfo/gumstix-users

------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
trial. Simplify your report design, integration and deployment - and focus on
what you do best, core application coding. Discover what's new with
Crystal Reports now. http://p.sf.net/sfu/bobj-july
_______________________________________________
gumstix-users mailing list
gumstix-users@...
https://lists.sourceforge.net/lists/listinfo/gumstix-users

Re: Query on usbnet on the overo with summit

2009-11-25T12:30:29Z

The OTG USB port isn't configured for USB networking in the stock Gumstix kernels.

I started a document here explaining what you need to do or at least one way to
go about doing it.

http://github.com/scottellis/overo-usbnet

It's not ready for prime-time, but you might take something from it.

Any feedback is welcome.

kailoi wrote:

Hi there, this (hopefully) should have a quick fix but I've been
battling with this all afternoon and it's driving me mad.

I've recived my overo and summit board and been having great fun with
it over the last few days. I have a bootable debian SD card and a few
other goodies but now I want to update and add some new packages to
the install.

I thought I'd set up usbnet. I followed the instructions here :
http://docwiki.gumstix.org/index.php/Windows_XP_usbnet

on my XP box and I'm having the following problem.

When I plug in the overo (I am connecting a USB-mini-B into the OTG
port on the overo and the USB A into my XP box) XP sees the new device
plugged in.. thinks for a few seconds then pops up "Unknown Device,
This device is unrecognized by windows" and then turns off the USB
port.

I can't add the drivers to the unknown device, it won't let me.
However. If I reboot the Overo from the root prompt, just after it
power cycles windows for a few seconds sees a "OTG.. something device"
(it goes too fast too read fully) which if I click "add drivers" to
that, lets me add the linux.inf file. And seems to install drivers as
per the above instructions.

_However_ Once the Overo starts booting off the mmc the recognised
device vanishes and the "unrecognized device" comes back.

Has anyone seen this before or has any idea what could be causing this?

Thanks in advance.

Chris

------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
trial. Simplify your report design, integration and deployment - and focus on
what you do best, core application coding. Discover what's new with
Crystal Reports now. http://p.sf.net/sfu/bobj-july
_______________________________________________
gumstix-users mailing list
gumstix-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/gumstix-users

Re: gpio-event not compiling

2009-11-25T11:56:50Z

Hi Steven,

On Wed, Nov 25, 2009 at 11:11 AM, Steven Kane <steven.kane@...> wrote:

> Dave,
>
> I was able to get 'gpio-event' to compile for the Overo after applying
> some tweaks to 'gpio-event-drv.c', 'gpio-event.bb' and Makefile. Attached
> are these modified files. However, when I installed the package on the
> Overo, I successfully added a pin to be monitored with the command:
> 'gpio-event 144:r:20' as documented here:
> http://docwiki.gumstix.org/GPIO_event, but 'gpio-event' did not show up in
> '/dev' or '/proc'. Also, when I ran the command 'insmod gpio-event-drv.ko' I
> got the error: 'insmod: error inserting 'gpio-event-drv.ko': -1 Invalid
> module format'. Any ideas at this point?

The first thing I would do, is to run the file command on the
gpio-event-drv.ko file. I don't know if the file command is on the
gumstix, but it's most likely on your build machine. Make sure that
the .ko file was built for the ARM architecture.

You should also be able to run modinfo (on your build machine for sure
- maybe on the gumstix).

I believe that there may also be some additional information logged in
the /var/log/messages file.

Typically, the problem with vermagic mismatch is caused because the
version of toolchain/linux that the module was built for doesn't match
the versions of the kernel running on the gumstix (i.e. your build
environment is newer or older than the one used to build the image on
your gumstix).

--
Dave Hylands
Shuswap, BC, Canada
http://www.DaveHylands.com/

------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
trial. Simplify your report design, integration and deployment - and focus on
what you do best, core application coding. Discover what's new with
Crystal Reports now. http://p.sf.net/sfu/bobj-july
_______________________________________________
gumstix-users mailing list
gumstix-users@...
https://lists.sourceforge.net/lists/listinfo/gumstix-users

Re: [Gumstix-Users] Use OpenEmbedded or not?

2009-11-25T11:35:55Z

Yes I think MicroSD is the way to go. We won't have Ethernet capabilities on these boards, unless perhaps I can rig something up with USB for development.

I'm not sure if we'll use QT or perhaps GTK or tcl/tk. I was looking for something to supply graphic libraries and event handling between classes similar to .NET or MFC in Windows.

I do have Eclipse set up on a workstation and once OpenEmbedded is installed, I plan to install the OpenEmbedded Tools for Eclipse (http://community.buglabs.net/kgilmer/posts/75-OpenEmbedded-Tools-for-Eclipse-6-) and see how that goes. I'm hoping it will let me just create a project and build it for the Gumstix ARM processor.

I do have an ARM toolchain installed, and Qt-embedded with qt eclipse integration plugins. I just hope everything "sees" everything... so to speak. :)

On Wed, Nov 25, 2009 at 2:11 PM, Scott Ellis <scottellis.developer@...> wrote:

If you develop on the Gumstix you'd be better off with a MicroSD card
or even better with an nfsroot environment. That last assumes you are
using an expansion card with ethernet.

If you are planning on doing QT work, I doubt you are going to have
enough room for builds on the gumstix using just the NAND flash.

For me, nfsroot hosted on my workstation, cross-building using the OE
built tools, but NOT using bitbake or recipes for my code, is the most
convenient.

Build on a workstation with horsepower, run on the Gumstix, no copying
between iterations required.

I don't see much advantage to building on the Gumstix themselves
unless you don't have network connectivity. Gumstix do ethernet, wifi
and usbnet so surely for development you can get some sort of
networking going.

I think OE is great resource for building tools, keeping external
libraries up to date and final packaging.

I wouldn't recommend the bitbake/recipe approach for day to day
development of my own code though. Too slow.

It should be easy enough to point Eclipse running on your workstation
at the OE built cross-tools and libraries. I say this never having
used Eclipse, but it would be a pretty dumb tool if you couldn't.

On 11/25/09, coderdrone@... <coderdrone@...> wrote:
> If we develop directly on the Gumstix, about how much space would the
> development environment take up? I'm more familiar with software
> development on Windows where we can use MFC or .NET for display APIs, event
> handling, but for the Gumstix, we were considering using Eclipse and Qt
> Embedded with OpenEmbedded Tools plugins unless there are better options for
> the Gumstix. Space is limited, at least on NAND flash.
>
> As for the overo-oe install, I'm talking about the build environment. I was
> trying to set it up on a Red Hat box following the 'setting up a build
> environment' instructions on gumstix.net's documentation pages. One of the
> first steps has me do a git clone from gitorious.org for
> org.openembedded.dev, but this fails every time, reporting an 'unexpected
> EOF' as it's trying to receive the files.
>
> On Tue, Nov 24, 2009 at 4:17 PM, richard dorfner <rdorfner@...> wrote:
>
>> Howdy there, as always, the answers are not typically straight
>> forward. First though, a bit of nomenclature,
>>
>> Open Embedded is the particular flavor of linux running on Gumstix
>> processors.
>> The build system is Bitbake. Now, you can use the gumstix like any
>> linux system, and install the os and whatever tools you need,
>> including a compiler and simply develop the code natively on the
>> gumstix itself. There are pros and cons to this, however, the
>> biggest pro is hardly any learning curve, if you already know linux,
>> that is.
>>
>> So no, you don't, strictly speaking, need to use the whole open
>> embedded cross development environment. If you decide to, though,
>> you will find the learning curve a bit steep. At least, for me it has
>> been, and I STILL am not comfortable creating my own user recipes at
>> all. But that's me, other folks find it intuitive and easy to
>> learn/use and understand.
>>
>> Not sure what's going on with your overo-oe install, are you talking
>> about the build environment? Or getting the kernel onto the gumstix?
>> or...???
>>
>> Rick
>>
>>
>> On Tue, Nov 24, 2009 at 2:18 PM, coder drone <coderdrone@...> wrote:
>> > Hello everyone,
>> >
>> > Our project recently acquired some Gumstix boards to develop software
>> > for
>> an
>> > embedded system. I'm really quite new to the whole Embedded Linux
>> > world,
>> > and am trying to get a feel for how everything works. Is it necessary
>> > to
>> > use OpenEmbedded to develop software for Gumstix?
>> >
>> > I'm assuming OpenEmbedded provides a way for you to build the kernel,
>> > filesystem, drivers, etc., but if the board already has all of this,
>> > couldn't you just develop software and build it with the appropriate
>> > compiler for that processor?
>> >
>> > We're looking to use Eclipse to develop C++ code, maybe even using Qt
>> > embedded, to run on the Gumstix. I installed the OpenEmbedded Tools for
>> > Eclipse plugin, but it apparently assumes we already have a working
>> > installation of OpenEmbedded, but the overo-oe install keeps failing.
>> >
>> > Thanks.
>> >
>> >
>> ------------------------------------------------------------------------------
>> > Let Crystal Reports handle the reporting - Free Crystal Reports 2008
>> 30-Day
>> > trial. Simplify your report design, integration and deployment - and
>> focus
>> > on
>> > what you do best, core application coding. Discover what's new with
>> > Crystal Reports now. http://p.sf.net/sfu/bobj-july
>> > _______________________________________________
>> > gumstix-users mailing list
>> > gumstix-users@...
>> > https://lists.sourceforge.net/lists/listinfo/gumstix-users
>> >
>> >
>>
>>
>>
>> --
>> Say you can or say you can't, either way you will be right.
>> Computers are like old testament gods: Lots of rules and no mercy.
>>
>>
>> ------------------------------------------------------------------------------
>> Let Crystal Reports handle the reporting - Free Crystal Reports 2008
>> 30-Day
>> trial. Simplify your report design, integration and deployment - and focus
>> on
>> what you do best, core application coding. Discover what's new with
>> Crystal Reports now. http://p.sf.net/sfu/bobj-july
>> _______________________________________________
>> gumstix-users mailing list
>> gumstix-users@...
>> https://lists.sourceforge.net/lists/listinfo/gumstix-users
>>
>

------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
trial. Simplify your report design, integration and deployment - and focus on
what you do best, core application coding. Discover what's new with
Crystal Reports now. http://p.sf.net/sfu/bobj-july
_______________________________________________
gumstix-users mailing list
gumstix-users@...
https://lists.sourceforge.net/lists/listinfo/gumstix-users

Re: OpenEmbedded for multiple developers

2009-11-25T11:24:13Z

You can keep the build in one location then each user just needs to set
up their environment before using the compiler. As there are now connex,
verdex, overo builds - we tend to make an alias for them in .bashrc like:
alias
overo="/home/gum/overo/tmp/deploy/glibc/addons/arm-angstrom-linux-gnueabi-angstrom-overo-devshell"

a simple nfs mount of the build (/opt was a good suggestion) on each
machine should be OK.

good luck,
Kirk

coderdrone@... wrote:

> I'm currently in the process of installing OpenEmbedded on a Red Hat
> Enterprise Linux workstation. It is installed under my user account's
> home directory. Since there are multiple developers on the team, each
> with their own individual login accounts on the box, what would be a
> good way to make this OE installation available to all of the
> developers? Does overo-oe need to be installed from scratch for each user?
>
> I also want to get this installation of overo-oe onto several
> non-networked workstations. I'm not sure I can just image the hard
> drive since the other workstations use a different RHEL subscription
> than the original box.
>
>
> ------------------------------------------------------------------------
>
> ------------------------------------------------------------------------------
> Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
> trial. Simplify your report design, integration and deployment - and focus on
> what you do best, core application coding. Discover what's new with
> Crystal Reports now. http://p.sf.net/sfu/bobj-july
>
>
> ------------------------------------------------------------------------
>
> _______________________________________________
> gumstix-users mailing list
> gumstix-users@...
> https://lists.sourceforge.net/lists/listinfo/gumstix-users

--
Electronics and Computer Science
University of Southampton, UK
+44 (0)2380 594491
http://www.ecs.soton.ac.uk/~km

------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
trial. Simplify your report design, integration and deployment - and focus on
what you do best, core application coding. Discover what's new with
Crystal Reports now. http://p.sf.net/sfu/bobj-july
_______________________________________________
gumstix-users mailing list
gumstix-users@...
https://lists.sourceforge.net/lists/listinfo/gumstix-users

Re: OpenEmbedded for multiple developers

2009-11-25T11:19:12Z

Consider putting OE in /opt/OpenEmbedded/ (or similar)?

Kyle Fazzari

On Nov 25, 2009, at 11:15 AM, coderdrone@... wrote:

> I'm currently in the process of installing OpenEmbedded on a Red Hat
> Enterprise Linux workstation. It is installed under my user
> account's home directory. Since there are multiple developers on
> the team, each with their own individual login accounts on the box,
> what would be a good way to make this OE installation available to
> all of the developers? Does overo-oe need to be installed from
> scratch for each user?
>
> I also want to get this installation of overo-oe onto several non-
> networked workstations. I'm not sure I can just image the hard
> drive since the other workstations use a different RHEL subscription
> than the original box.
> ------------------------------------------------------------------------------
> Let Crystal Reports handle the reporting - Free Crystal Reports 2008
> 30-Day
> trial. Simplify your report design, integration and deployment - and
> focus on
> what you do best, core application coding. Discover what's new with
> Crystal Reports now. http://p.sf.net/sfu/bobj-july_______________________________________________
> gumstix-users mailing list
> gumstix-users@...
> https://lists.sourceforge.net/lists/listinfo/gumstix-users

OpenEmbedded for multiple developers

2009-11-25T11:15:14Z

I'm currently in the process of installing OpenEmbedded on a Red Hat Enterprise Linux workstation. It is installed under my user account's home directory. Since there are multiple developers on the team, each with their own individual login accounts on the box, what would be a good way to make this OE installation available to all of the developers? Does overo-oe need to be installed from scratch for each user?

I also want to get this installation of overo-oe onto several non-networked workstations. I'm not sure I can just image the hard drive since the other workstations use a different RHEL subscription than the original box.

------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
trial. Simplify your report design, integration and deployment - and focus on
what you do best, core application coding. Discover what's new with
Crystal Reports now. http://p.sf.net/sfu/bobj-july
_______________________________________________
gumstix-users mailing list
gumstix-users@...
https://lists.sourceforge.net/lists/listinfo/gumstix-users

Re: gpio-event not compiling

2009-11-25T11:11:53Z

Dave,

I was able to get 'gpio-event' to compile for the Overo after applying some tweaks to 'gpio-event-drv.c', 'gpio-event.bb' and Makefile. Attached are these modified files. However, when I installed the package on the Overo, I successfully added a pin to be monitored with the command: 'gpio-event 144:r:20' as documented here: http://docwiki.gumstix.org/GPIO_event, but 'gpio-event' did not show up in '/dev' or '/proc'. Also, when I ran the command 'insmod gpio-event-drv.ko' I got the error: 'insmod: error inserting 'gpio-event-drv.ko': -1 Invalid module format'. Any ideas at this point?

Steven

On Tue, Nov 24, 2009 at 4:48 PM, Dave Hylands <dhylands@...> wrote:

HI Steven,

On Tue, Nov 24, 2009 at 2:32 PM, Steven Kane <steven.kane@...> wrote:

> After following the instructions for setting up my build environment found
> here:
>
> http://www.gumstix.net/Setup-and-Programming/view/Overo-Setup-and-Programming/Setting-up-a-build-environment/111.html
>
> and successfully compiling the 'omap3-console-image', I was unable to get
> the 'gpio-event' app to compile under bitbake found in
> 'org.openembedded.dev/recipes/gpio-event' checked out via the instructions
> above.
>

gpio-event is currently PXA255 or PXA270 specific. It needs some
tweaks (and updates) to work on the OMAP.

Here's a post I made a while back:
<http://thread.gmane.org/gmane.linux.distributions.gumstix.general/44916/focus=44928>

...snip...

> If anyone has any information on how to fix this OR if anyone has
> information on how to configure/read/write the GPIO ports on the Overo Earth
> COM on the Chestnut43 expansion board running the console image, your help
> would be much appreciated.

--
Dave Hylands
Shuswap, BC, Canada
http://www.DaveHylands.com/

------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
trial. Simplify your report design, integration and deployment - and focus on
what you do best, core application coding. Discover what's new with
Crystal Reports now. http://p.sf.net/sfu/bobj-july
_______________________________________________
gumstix-users mailing list
gumstix-users@...
https://lists.sourceforge.net/lists/listinfo/gumstix-users

[gpio-event-drv.c]

/**************************************************************************** * * Copyright (c) 2006 Dave Hylands <dhylands@...> * * This program is free software; you can redistribute it and/or modify * it under the terms of the GNU General Public License version 2 as * published by the Free Software Foundation. * * Alternatively, this software may be distributed under the terms of BSD * license. * * See README and COPYING for more details. * **************************************************************************** * * This driver allows multiple GPIO pins to be monitored and allows a user * mode program to be notified when the pin changes. * ****************************************************************************/ /* ---- Include Files ---------------------------------------------------- */ #include <linux/module.h> #include <linux/init.h> #include <linux/list.h> #include <linux/fs.h> #include <linux/spinlock.h> #include <linux/proc_fs.h> #include <linux/sysctl.h> #include <linux/poll.h> #include <linux/interrupt.h> #include <linux/device.h> #include <linux/cdev.h> #include <linux/seq_file.h> #include <linux/spinlock.h> #include <linux/version.h> #include <linux/gpio.h> #include <asm/uaccess.h> #include <asm/ioctls.h> /*#include <asm/arch/hardware.h> #include <asm/arch/pxa-regs.h>*/ #include "gpio-event-drv.h" /* ---- Public Variables ------------------------------------------------- */ /* ---- Private Constants and Types -------------------------------------- */ #define GPIO_EVENT_DEV_NAME "gpio-event" #define DEBUG_ENABLED 1 #if DEBUG_ENABLED # define DEBUG( flag, fmt, args... ) do { if ( gDebug ## flag ) printk( "%s: " fmt, __FUNCTION__ , ## args ); } while (0) #else # define DEBUG( flag, fmt, args... ) #endif /* ---- Private Variables ------------------------------------------------ */ static char gBanner[] __initdata = KERN_INFO "GPIO Event Monitor 0.1 Compiled: " __DATE__ " at " __TIME__ "\n"; static int gDebugTrace = 0; static int gDebugIoctl = 0; static int gDebugError = 1; static int gLostEvents = 0; static struct ctl_table_header *gSysCtlHeader; static struct ctl_table gSysCtlSample[] = { { .ctl_name = 1, .procname = "lost-events", .data = &gLostEvents, .maxlen = sizeof( int ), .mode = 0644, .proc_handler = &proc_dointvec }, { .ctl_name = 101, .procname = "debug-trace", .data = &gDebugTrace, .maxlen = sizeof( int ), .mode = 0644, .proc_handler = &proc_dointvec }, { .ctl_name = 102, .procname = "debug-ioctl", .data = &gDebugIoctl, .maxlen = sizeof( int ), .mode = 0644, .proc_handler = &proc_dointvec }, { .ctl_name = 103, .procname = "debug-error", .data = &gDebugError, .maxlen = sizeof( int ), .mode = 0644, .proc_handler = &proc_dointvec }, { 0 } }; static struct ctl_table gSysCtl[] = { { .ctl_name = CTL_GPIO_EVENT, .procname = "gpio-event", .mode = 0555, .child = gSysCtlSample }, { 0 } }; /* * An instance of GPIO_FileData_t is maintained for file open */ #define GPIO_EVENT_QUEUE_LEN 20 // GPIO_EVENT_BUFFER_SIZE needs to be big enough to hold the ASCII version // of the GPIO_Event_t as well as the binary version of the GPIO_Event_t #define GPIO_EVENT_BUFFER_SIZE 32 typedef struct { struct list_head list; wait_queue_head_t waitQueue; spinlock_t queueLock; GPIO_Event_t queueData[ GPIO_EVENT_QUEUE_LEN ]; volatile int getIndex; volatile int putIndex; volatile int numEvents; GPIO_EventReadMode_t readMode; char buffer[ GPIO_EVENT_BUFFER_SIZE ]; int bufBytes; } GPIO_FileData_t; /* * An instance of GPIO_PinData_t is maintained for each GPIO line which is * monitored, */ typedef enum { PIN_LOW = 0, // Matches level of GPIO line PIN_HIGH = 1, PIN_BOUNCING_LOW, PIN_BOUNCING_HIGH, } PinState_t; typedef struct { struct list_head list; // list of all pins int gpio; // The gpio line being monitored // We maintain two lists, a global list of pins, and a list associated with each open struct timer_list debounceTimer; // Timer to wake u up after an edge uint8_t debounceMilliSec; // debounce time in milliseconds char devName[ 16 ]; // gpio xx event GPIO_EventEdgeType_t edgeType; // Type of edge(s) we're looking for. PinState_t pinState; // Was the GPIO line low or high? } GPIO_PinData_t; static volatile int gReportLostEvents = 1; static struct class *gGpioEventClass = NULL; static struct cdev gGpioEventCDev; static dev_t gGpioEventDevNum = 0; static spinlock_t gFileListLock = SPIN_LOCK_UNLOCKED; static spinlock_t gPinListLock = SPIN_LOCK_UNLOCKED; static LIST_HEAD( gFileList ); static LIST_HEAD( gPinList ); static struct proc_dir_entry *gProcGpioEvent; static struct proc_dir_entry *gProcPins; /* ---- Private Function Prototypes -------------------------------------- */ /* ---- Functions -------------------------------------------------------- */ typedef struct { unsigned long flags; struct list_head *list; } pin_seq_t; /**************************************************************************** * * pin_seq_start * * seq_file iterator which goes through the pins being monitored * ****************************************************************************/ static void *pin_seq_start( struct seq_file *s, loff_t *pos ) { pin_seq_t *ps; loff_t i; s->private = NULL; if (( ps = kcalloc( 1, sizeof( pin_seq_t ), GFP_KERNEL )) == NULL ) { return ERR_PTR( -ENOMEM ); } s->private = ps; spin_lock_irqsave( &gPinListLock, ps->flags ); if ( list_empty( &gPinList )) { DEBUG( Trace, "list_empty\n" ); return NULL; } ps->list = gPinList.next; for ( i = 0; i < *pos; i++ ) { if ( list_is_last( ps->list, &gPinList )) { DEBUG( Trace, "No item @ %llu\n", i + 1 ); return NULL; } ps->list = ps->list->next; } DEBUG( Trace, "ps->list = 0x%08lx, *pos = %llu\n", (long)ps->list, *pos ); return ps->list; } // pin_seq_start /**************************************************************************** * * pin_seq_show * * seq_file iterator which goes through the pins being monitored * ****************************************************************************/ static int pin_seq_show( struct seq_file *s, void *v ) { GPIO_PinData_t *pin = list_entry( v, GPIO_PinData_t, list ); char *edgeTypeStr; DEBUG( Trace, "v = 0x%08lx\n", (long)v ); switch ( pin->edgeType ) { case GPIO_EventRisingEdge: edgeTypeStr = "Rising "; break; case GPIO_EventFallingEdge: edgeTypeStr = "Falling"; break; case GPIO_EventBothEdges: edgeTypeStr = "Both "; break; default: edgeTypeStr = "Unknown"; break; } seq_printf( s, "GPIO: %3d Edge: %s Debounce: %d msec\n", pin->gpio, edgeTypeStr, pin->debounceMilliSec ); return 0; } // pin_seq_show /**************************************************************************** * * pin_seq_next * * seq_file iterator which goes through the pins being monitored * ****************************************************************************/ static void *pin_seq_next( struct seq_file *s, void *v, loff_t *pos ) { pin_seq_t *ps = s->private; DEBUG( Trace, "v = 0x%08lx *pos = %llu\n", (long)v, *pos ); if ( list_is_last( ps->list, &gPinList )) { DEBUG( Trace, "ps->list = 0x%08lx (end of list)\n", (long)ps->list ); return NULL; } (*pos)++; ps->list = ps->list->next; DEBUG( Trace, "ps->list = 0x%08lx\n", (long)ps->list ); return ps->list; } // pin_seq_next /**************************************************************************** * * pin_seq_stop * * seq_file iterator which goes through the pins being monitored * ****************************************************************************/ static void pin_seq_stop( struct seq_file *s, void *v ) { pin_seq_t *ps = s->private; DEBUG( Trace, "v = 0x%08lx\n", (long)v ); if ( ps != NULL ) { spin_unlock_irqrestore( &gPinListLock, ps->flags ); kfree( ps ); } } // pin_seq_stop /**************************************************************************** * * pin_seq_ops * * Ties all of the pin_seq_xxx routines together. * ****************************************************************************/ static struct seq_operations pin_seq_ops = { .start = pin_seq_start, .next = pin_seq_next, .stop = pin_seq_stop, .show = pin_seq_show }; /**************************************************************************** * * pins_proc_open * * Open method for /proc/gpio-event/pin * ****************************************************************************/ static int pins_proc_open( struct inode *inode, struct file *file ) { DEBUG( Trace, "called\n" ); return seq_open( file, &pin_seq_ops ); } /**************************************************************************** * * pin_proc_ops * * File operations for our /proc/gpio-event/pins file * ****************************************************************************/ static struct file_operations pins_proc_ops = { .owner = THIS_MODULE, .open = pins_proc_open, .read = seq_read, .llseek = seq_lseek, .release = seq_release }; /**************************************************************************** * * find_pin * * Searches the list to see if 'gpio' is currently being monitored. * ****************************************************************************/ static GPIO_PinData_t *find_pin( int gpio ) { struct list_head *pin; assert_spin_locked( &gPinListLock ); list_for_each( pin, &gPinList ) { GPIO_PinData_t *pinData = list_entry( pin, GPIO_PinData_t, list ); if ( pinData->gpio == gpio ) { return pinData; } } return NULL; } // find_pin /**************************************************************************** * * gpio_event_queue_event * * Queues an sample event from the bottom half to the top half. This * function queues up the event on every file that's open. * ****************************************************************************/ static void gpio_event_queue_event( const GPIO_Event_t *gpioEvent ) { unsigned long flags; struct list_head *file; DEBUG( Trace, "gpio %d:%c@%ld.%06ld\n", gpioEvent->gpio, gpioEvent->edgeType == GPIO_EventRisingEdge ? 'R' : 'F', gpioEvent->time.tv_sec, gpioEvent->time.tv_usec ); // Queue up the event on all of the open files // // This function is only called from the ISR, with interrupts already // disabled. spin_lock_irqsave( &gFileListLock, flags ); list_for_each( file, &gFileList ) { GPIO_FileData_t *fileData = list_entry( file, GPIO_FileData_t, list ); spin_lock( &fileData->queueLock ); { if ( fileData->numEvents >= GPIO_EVENT_QUEUE_LEN ) { // Queue is full - Only report first event lost if ( gReportLostEvents ) { printk( KERN_ERR "GPIO Event: event lost due to queue full\n" ); gReportLostEvents = 0; } gLostEvents++; } else { fileData->queueData[ fileData->putIndex++ ] = *gpioEvent; if ( fileData->putIndex >= GPIO_EVENT_QUEUE_LEN ) { fileData->putIndex = 0; } fileData->numEvents++; } } spin_unlock( &fileData->queueLock ); wake_up_interruptible( &fileData->waitQueue ); } spin_unlock_irqrestore( &gFileListLock, flags ); } // gpio_event_queue_event /**************************************************************************** * * gpio_event_dequeue_event * * Removes an event from the queue * ****************************************************************************/ static int gpio_event_dequeue_event( GPIO_FileData_t *fileData, GPIO_Event_t *gpioEvent ) { unsigned long flags; int eventAvailable = 0; spin_lock_irqsave( &fileData->queueLock, flags ); { if ( fileData->numEvents > 0 ) { *gpioEvent = fileData->queueData[ fileData->getIndex++ ]; if ( fileData->getIndex >= GPIO_EVENT_QUEUE_LEN ) { fileData->getIndex = 0; } fileData->numEvents--; eventAvailable = 1; if ( fileData->numEvents == 0 ) { // Since somebody is reading the queue now, indicate that we // can report lost events again gReportLostEvents = 1; } } } spin_unlock_irqrestore( &fileData->queueLock, flags ); DEBUG( Trace, "gpio %d:%c@%ld.%06ld\n", gpioEvent->gpio, gpioEvent->edgeType == GPIO_EventRisingEdge ? 'R' : 'F', gpioEvent->time.tv_sec, gpioEvent->time.tv_usec ); return eventAvailable; } // gpio_event_dequeue_event /**************************************************************************** * * gpio_event_irq * ****************************************************************************/ #if ( LINUX_VERSION_CODE <= KERNEL_VERSION( 2, 6, 20 )) #define REGS_PARAM , struct pt_regs *regs #define REGS_UNUSED (void)regs #else #define REGS_PARAM #define REGS_UNUSED #endif static irqreturn_t gpio_event_irq( int irq, void *dev_id REGS_PARAM ) { GPIO_PinData_t *pinData = (GPIO_PinData_t *)dev_id; GPIO_Event_t gpioEvent; /*int currLevel = (( GPLR( pinData->gpio ) & GPIO_bit( pinData->gpio )) != 0 );*/ int currLevel = gpio_get_value(pinData->gpio); // We're called with interrupts disabled. (void)irq; REGS_UNUSED; do_gettimeofday( &gpioEvent.time ); gpioEvent.gpio = pinData->gpio; if ( pinData->debounceMilliSec == 0 ) { // We assume that this is a clean signal pinData->pinState = (PinState_t)currLevel; if ( pinData->edgeType == GPIO_EventBothEdges ) { // There's no register to tell which edge just occurred. So we // assume that it just changed into its current level. if ( currLevel ) { // Pin is currently high, so this must be a rising edge gpioEvent.edgeType = GPIO_EventRisingEdge; } else { // Pin is currently low, so this must be a falling edge gpioEvent.edgeType = GPIO_EventFallingEdge; } } else { // If we're only monitoring one type of edge, then that's the one // that happened. gpioEvent.edgeType = pinData->edgeType; } gpio_event_queue_event( &gpioEvent ); } else { gpioEvent.edgeType = 0; // If we need to debounce, then we need to monitor both edges, and // use the debounce timer to figure out the real state. So we don't // actually know which edge we just got. We use a state machine // to track things. switch ( pinData->pinState ) { case PIN_LOW: { pinData->pinState = PIN_BOUNCING_HIGH; gpioEvent.edgeType = GPIO_EventRisingEdge; break; } case PIN_HIGH: { pinData->pinState = PIN_BOUNCING_LOW; gpioEvent.edgeType = GPIO_EventFallingEdge; break; } default: { break; } } if (( pinData->edgeType & gpioEvent.edgeType ) != 0 ) { // This is an edge that the user is interested in - send it along. gpio_event_queue_event( &gpioEvent ); } // Disable interrupts for our gpio to allow debounce to occur. The // timer will re-enable the interrupt. disable_irq_nosync( irq ); // Since we have no idea when in the current jiffy that the edge // occurred, we add 1 to the calculation to guarantee at least one // whole jiffy. mod_timer( &pinData->debounceTimer, jiffies + msecs_to_jiffies( pinData->debounceMilliSec ) + 1 ); } return IRQ_HANDLED; } // gpio_event_irq /**************************************************************************** * * gpio_event_timer * ****************************************************************************/ void gpio_event_timer( unsigned long data ) { GPIO_PinData_t *pinData = (GPIO_PinData_t *)data; // This function is called when the debounce timer for a gpio expires. // We record the state of the pin so that we can figure out what the // next edge will be. pinData->pinState = gpio_get_value(pinData->gpio); // Turn interrupts back on so we can catch the next edge enable_irq( gpio_to_irq(pinData->gpio )); } // gpio_event_timer /**************************************************************************** * * gpio_event_monitor * ****************************************************************************/ static int gpio_event_monitor( GPIO_EventMonitor_t *monitor ) { int rc = 0; unsigned long flags; GPIO_PinData_t *pinData; unsigned long irqFlags; spin_lock_irqsave( &gPinListLock, flags ); if ( monitor->onOff ) { // Check to make sure we aren't already monitoring the gpio if (( pinData = find_pin( monitor->gpio )) != NULL ) { // We are already monitoring the pin. Unmonitor the pin and then // proceed. monitor->onOff = 0; spin_unlock_irqrestore( &gPinListLock, flags ); gpio_event_monitor( monitor ); spin_lock_irqsave( &gPinListLock, flags ); } if (( pinData = kcalloc( 1, sizeof( *pinData ), GFP_KERNEL )) == NULL ) { DEBUG( Error, "GPIO %d: Out of memory\n", monitor->gpio ); rc = -ENOMEM; goto out; } INIT_LIST_HEAD( &pinData->list ); snprintf( pinData->devName, sizeof( pinData->devName ), "gpio %d event", monitor->gpio ); // Note: // Calling request_irq will automatically set the pin to be an input. irqFlags = 0; if ( monitor->debounceMilliSec == 0 ) { // A clean signal is being presented, so we can just look for // a particular edge if (( monitor->edgeType | GPIO_EventRisingEdge ) != 0 ) { irqFlags |= IRQF_TRIGGER_RISING; } if (( monitor->edgeType | GPIO_EventFallingEdge ) != 0 ) { irqFlags |= IRQF_TRIGGER_FALLING; } } else { // Since we need to debounce, we need to look for both types of // edges, since we get both types of edges whenever a bounce // happens. irqFlags |= IRQF_TRIGGER_RISING; irqFlags |= IRQF_TRIGGER_FALLING; } if (( rc = request_irq( gpio_to_irq( monitor->gpio ), gpio_event_irq, irqFlags, pinData->devName, pinData )) != 0 ) { DEBUG( Error, "Unable to register irq for GPIO %d\n", monitor->gpio ); kfree( pinData ); goto out; } pinData->gpio = monitor->gpio; pinData->edgeType = monitor->edgeType; pinData->debounceMilliSec = monitor->debounceMilliSec; init_timer( &pinData->debounceTimer ); pinData->debounceTimer.data = (unsigned long)pinData; pinData->debounceTimer.function = gpio_event_timer; list_add_tail( &pinData->list, &gPinList ); if (gpio_get_value(pinData->gpio) == 0 ) { pinData->pinState = PIN_LOW; } else { pinData->pinState = PIN_HIGH; } } else { if (( pinData = find_pin( monitor->gpio )) == NULL ) { DEBUG( Error, "GPIO %d isn't being monitored\n", monitor->gpio ); rc = -ENXIO; goto out; } // We've found the gpio being monitored - turn things off. free_irq( gpio_to_irq( pinData->gpio ), pinData ); del_timer_sync( &pinData->debounceTimer ); list_del( &pinData->list ); kfree( pinData ); } out: spin_unlock_irqrestore( &gPinListLock, flags ); return rc; } // gpio_event_monitor /**************************************************************************** * * gpio_event_ioctl * * Called to process ioctl requests * *****************************************************************************/ int gpio_event_ioctl( struct inode *inode, struct file *file, unsigned int cmd, unsigned long arg ) { GPIO_FileData_t *fileData; DEBUG( Trace, "type: '%c' cmd: 0x%x\n", _IOC_TYPE( cmd ), _IOC_NR( cmd )); fileData = file->private_data; switch ( cmd ) { case GPIO_EVENT_IOCTL_MONITOR_GPIO: { GPIO_EventMonitor_t monitor; if ( copy_from_user( &monitor, (void *)arg, sizeof( monitor )) != 0 ) { return -EFAULT; } return gpio_event_monitor( &monitor ); } case GPIO_EVENT_IOCTL_SET_READ_MODE: { fileData->readMode = (GPIO_EventReadMode_t)arg; break; } case TCGETS: { // When cat opens this device, we get this ioctl return -ENOTTY; } default: { DEBUG( Error, "Unrecognized ioctl: '0x%x'\n", cmd ); return -ENOTTY; } } return 0; } // gpio_event_ioctl /**************************************************************************** * * gpio_event_open * ****************************************************************************/ static int gpio_event_open( struct inode *inode, struct file *file ) { unsigned long flags; GPIO_FileData_t *fileData; DEBUG( Trace, "gpio_event_open called, major = %d, minor = %d\n", MAJOR( inode->i_rdev ), MINOR( inode->i_rdev )); // Allocate a per-open data structure if (( fileData = kcalloc( 1, sizeof( *fileData ), GFP_KERNEL )) == NULL ) { return -ENOMEM; } INIT_LIST_HEAD( &fileData->list ); init_waitqueue_head( &fileData->waitQueue ); spin_lock_init( &fileData->queueLock ); fileData->getIndex = 0; fileData->putIndex = 0; fileData->numEvents = 0; fileData->bufBytes = 0; fileData->readMode = GPIO_EventReadModeAscii; file->private_data = fileData; spin_lock_irqsave( &gFileListLock, flags ); { list_add_tail( &fileData->list, &gFileList ); } spin_unlock_irqrestore( &gFileListLock, flags ); return 0; } // gpio_event_open /**************************************************************************** * * gpio_event_read * ****************************************************************************/ static ssize_t gpio_event_read( struct file *file, char *buffer, size_t spaceRemaining, loff_t *ppos ) { int rc; ssize_t bytesCopied = 0; ssize_t bytesToCopy; GPIO_FileData_t *fileData = file->private_data; DEBUG( Trace, "gpio_event_read called, major = %d, minor = %d\n", MAJOR( file->f_dentry->d_inode->i_rdev ), MINOR( file->f_dentry->d_inode->i_rdev )); if ( spaceRemaining == 0 ) { return 0; } // First of all, return any unread data from the previous call if ( fileData->bufBytes > 0 ) { if ( spaceRemaining < fileData->bufBytes ) { bytesCopied = spaceRemaining; } else { bytesCopied = fileData->bufBytes; } if ( copy_to_user( &buffer[0], &fileData->buffer[0], bytesCopied ) != 0 ) { return -EFAULT; } if ( fileData->bufBytes > bytesCopied ) { memmove( &fileData->buffer[ 0 ], &fileData->buffer[ bytesCopied ], fileData->bufBytes - bytesCopied ); } fileData->bufBytes -= bytesCopied; if ( fileData->bufBytes > 0 ) { // We copied some data, but not all of it. Return early. return bytesCopied; } } do { if ((( file->f_flags & O_NONBLOCK ) != 0 ) && ( fileData->numEvents == 0 )) { // File was opened non-blocking and no more data is available // We don't want to wait for an event, so exit from the loop break; } rc = wait_event_interruptible( fileData->waitQueue, ( fileData->numEvents > 0 )); if ( rc != 0 ) { return rc; } if ( fileData->readMode == GPIO_EventReadModeBinary ) { gpio_event_dequeue_event( fileData, (GPIO_Event_t *)&fileData->buffer[0] ); fileData->bufBytes = sizeof( GPIO_Event_t ); } else { GPIO_Event_t gpioEvent; gpio_event_dequeue_event( fileData, &gpioEvent ); // ASCII Mode output: // // nn E tttttttt.tttttt // // Where nn is the base-10 GPIO number // E is R or F (for rising or falling edge) // tttttttt.tttttt is the timestamp with microsecond resolution fileData->bufBytes = snprintf( fileData->buffer, sizeof( fileData->buffer ), "%2d %c %ld.%06ld\n", gpioEvent.gpio, (( gpioEvent.edgeType == GPIO_EventRisingEdge ) ? 'R' : 'F' ), gpioEvent.time.tv_sec, gpioEvent.time.tv_usec ); } if ( spaceRemaining >= fileData->bufBytes ) { bytesToCopy = fileData->bufBytes; } else { bytesToCopy = spaceRemaining; } if ( copy_to_user( &buffer[ bytesCopied ], &fileData->buffer[0], bytesToCopy ) != 0 ) { return -EFAULT; } spaceRemaining -= bytesToCopy; bytesCopied += bytesToCopy; fileData->bufBytes -= bytesToCopy; if ( fileData->bufBytes > 0 ) { // We couldn't copy all of the data out of the buffer. Move the // remaining data to the beginning of the buffer and exit. memmove( &fileData->buffer[ 0 ], &fileData->buffer[ bytesToCopy ], fileData->bufBytes ); return bytesCopied; } } while (( fileData->numEvents > 0 ) && ( spaceRemaining > 0 )); if ((( file->f_flags & O_NONBLOCK ) != 0 ) && ( bytesCopied == 0 )) { // File was opened non-blocking and we didn't copy any data. return -EAGAIN; } return bytesCopied; } // gpio_event_read /**************************************************************************** * * gpio_event_poll - used by select & poll * ****************************************************************************/ static unsigned int gpio_event_poll(struct file *file, poll_table *wait) { unsigned long flags; GPIO_FileData_t *fileData = file->private_data; unsigned int mask = 0; poll_wait( file, &fileData->waitQueue, wait ); spin_lock_irqsave( &fileData->queueLock, flags ); { if (( fileData->bufBytes > 0 ) || ( fileData->numEvents > 0 )) { mask |= POLLIN | POLLRDNORM; // readable } } spin_unlock_irqrestore( &fileData->queueLock, flags ); return mask; } // gpio_event_poll /**************************************************************************** * * gpio_event_release * ****************************************************************************/ static int gpio_event_release( struct inode *inode, struct file *file ) { unsigned long flags; GPIO_FileData_t *fileData = file->private_data; DEBUG( Trace, "gpio_event_release called\n" ); spin_lock_irqsave( &gFileListLock, flags ); { list_del( &fileData->list ); } spin_unlock_irqrestore( &gFileListLock, flags ); kfree( fileData ); return 0; } // gpio_event_release /**************************************************************************** * * File Operations (these are the device driver entry points) * ****************************************************************************/ struct file_operations gpio_event_fops = { owner: THIS_MODULE, ioctl: gpio_event_ioctl, open: gpio_event_open, poll: gpio_event_poll, release: gpio_event_release, read: gpio_event_read, }; /**************************************************************************** * * gpio_event_init * * Called to perform module initialization when the module is loaded * ****************************************************************************/ static int __init gpio_event_init( void ) { int rc; DEBUG( Trace, "called\n" ); printk( gBanner ); // Get a major number if (( rc = alloc_chrdev_region( &gGpioEventDevNum, 0, 1, GPIO_EVENT_DEV_NAME )) < 0 ) { printk( KERN_WARNING "sample: Unable to allocate major, err: %d\n", rc ); return rc; } DEBUG( Trace, "allocated major:%d minor:%d\n", MAJOR( gGpioEventDevNum ), MINOR( gGpioEventDevNum )); // Register our proc entries. gProcGpioEvent = create_proc_entry( "gpio-event", S_IFDIR | S_IRUGO | S_IXUGO, NULL ); if ( gProcGpioEvent == NULL ) { return -ENOMEM; } gProcPins = create_proc_entry( "pins", 0444, gProcGpioEvent ); if ( gProcPins != NULL ) { gProcPins->proc_fops = &pins_proc_ops; } #if ( LINUX_VERSION_CODE <= KERNEL_VERSION( 2, 6, 20 )) gSysCtlHeader = register_sysctl_table( gSysCtl, 0 ); if ( gSysCtlHeader != NULL ) { gSysCtlHeader->ctl_table->child->de->owner = THIS_MODULE; } #else gSysCtlHeader = register_sysctl_table( gSysCtl ); #endif // Register our device. The device becomes "active" as soon as cdev_add // is called. cdev_init( &gGpioEventCDev, &gpio_event_fops ); gGpioEventCDev.owner = THIS_MODULE; if (( rc = cdev_add( &gGpioEventCDev, gGpioEventDevNum, 1 )) != 0 ) { printk( KERN_WARNING "sample: cdev_add failed: %d\n", rc ); return rc; } // Create a class, so that udev will make the /dev entry gGpioEventClass = class_create( THIS_MODULE, GPIO_EVENT_DEV_NAME ); if ( IS_ERR( gGpioEventClass )) { printk( KERN_WARNING "sample: Unable to create class\n" ); return -1; } device_create( gGpioEventClass, NULL, gGpioEventDevNum, NULL, GPIO_EVENT_DEV_NAME ); return 0; } // gpio_event_init /**************************************************************************** * * gpio_event_exit * * Called to perform module cleanup when the module is unloaded. * ****************************************************************************/ static void __exit gpio_event_exit( void ) { struct list_head *next; struct list_head *pin; GPIO_EventMonitor_t monitor; DEBUG( Trace, "called\n" ); // If there are any pins which are currently being monitored, then we // need to unmonitor them. memset( &monitor, 0, sizeof( monitor )); list_for_each_safe( pin, next, &gPinList ) { GPIO_PinData_t *pinData = list_entry( pin, GPIO_PinData_t, list ); monitor.gpio = pinData->gpio; gpio_event_monitor( &monitor ); } // Deregister our driver device_destroy( gGpioEventClass, gGpioEventDevNum ); class_destroy( gGpioEventClass ); cdev_del( &gGpioEventCDev ); if ( gSysCtlHeader != NULL ) { unregister_sysctl_table( gSysCtlHeader ); } remove_proc_entry( "pins", gProcGpioEvent ); remove_proc_entry( "gpio-event", NULL ); unregister_chrdev_region( gGpioEventDevNum, 1 ); } // gpio_event_exit /****************************************************************************/ module_init(gpio_event_init); module_exit(gpio_event_exit); MODULE_AUTHOR("Dave Hylands"); MODULE_DESCRIPTION("GPIO Event Driver"); MODULE_LICENSE("Dual BSD/GPL");

gpio-event.bb (2K) Download Attachment
Makefile (1K) Download Attachment

Re: [Gumstix-Users] Use OpenEmbedded or not?

2009-11-25T11:11:29Z

If you develop on the Gumstix you'd be better off with a MicroSD card
or even better with an nfsroot environment. That last assumes you are
using an expansion card with ethernet.

If you are planning on doing QT work, I doubt you are going to have
enough room for builds on the gumstix using just the NAND flash.

For me, nfsroot hosted on my workstation, cross-building using the OE
built tools, but NOT using bitbake or recipes for my code, is the most
convenient.

Build on a workstation with horsepower, run on the Gumstix, no copying
between iterations required.

I don't see much advantage to building on the Gumstix themselves
unless you don't have network connectivity. Gumstix do ethernet, wifi
and usbnet so surely for development you can get some sort of
networking going.

I think OE is great resource for building tools, keeping external
libraries up to date and final packaging.

I wouldn't recommend the bitbake/recipe approach for day to day
development of my own code though. Too slow.

It should be easy enough to point Eclipse running on your workstation
at the OE built cross-tools and libraries. I say this never having
used Eclipse, but it would be a pretty dumb tool if you couldn't.

On 11/25/09, coderdrone@... <coderdrone@...> wrote:

> If we develop directly on the Gumstix, about how much space would the
> development environment take up? I'm more familiar with software
> development on Windows where we can use MFC or .NET for display APIs, event
> handling, but for the Gumstix, we were considering using Eclipse and Qt
> Embedded with OpenEmbedded Tools plugins unless there are better options for
> the Gumstix. Space is limited, at least on NAND flash.
>
> As for the overo-oe install, I'm talking about the build environment. I was
> trying to set it up on a Red Hat box following the 'setting up a build
> environment' instructions on gumstix.net's documentation pages. One of the
> first steps has me do a git clone from gitorious.org for
> org.openembedded.dev, but this fails every time, reporting an 'unexpected
> EOF' as it's trying to receive the files.
>
> On Tue, Nov 24, 2009 at 4:17 PM, richard dorfner <rdorfner@...> wrote:
>
>> Howdy there, as always, the answers are not typically straight
>> forward. First though, a bit of nomenclature,
>>
>> Open Embedded is the particular flavor of linux running on Gumstix
>> processors.
>> The build system is Bitbake. Now, you can use the gumstix like any
>> linux system, and install the os and whatever tools you need,
>> including a compiler and simply develop the code natively on the
>> gumstix itself. There are pros and cons to this, however, the
>> biggest pro is hardly any learning curve, if you already know linux,
>> that is.
>>
>> So no, you don't, strictly speaking, need to use the whole open
>> embedded cross development environment. If you decide to, though,
>> you will find the learning curve a bit steep. At least, for me it has
>> been, and I STILL am not comfortable creating my own user recipes at
>> all. But that's me, other folks find it intuitive and easy to
>> learn/use and understand.
>>
>> Not sure what's going on with your overo-oe install, are you talking
>> about the build environment? Or getting the kernel onto the gumstix?
>> or...???
>>
>> Rick
>>
>>
>> On Tue, Nov 24, 2009 at 2:18 PM, coder drone <coderdrone@...> wrote:
>> > Hello everyone,
>> >
>> > Our project recently acquired some Gumstix boards to develop software
>> > for
>> an
>> > embedded system. I'm really quite new to the whole Embedded Linux
>> > world,
>> > and am trying to get a feel for how everything works. Is it necessary
>> > to
>> > use OpenEmbedded to develop software for Gumstix?
>> >
>> > I'm assuming OpenEmbedded provides a way for you to build the kernel,
>> > filesystem, drivers, etc., but if the board already has all of this,
>> > couldn't you just develop software and build it with the appropriate
>> > compiler for that processor?
>> >
>> > We're looking to use Eclipse to develop C++ code, maybe even using Qt
>> > embedded, to run on the Gumstix. I installed the OpenEmbedded Tools for
>> > Eclipse plugin, but it apparently assumes we already have a working
>> > installation of OpenEmbedded, but the overo-oe install keeps failing.
>> >
>> > Thanks.
>> >
>> >
>> ------------------------------------------------------------------------------
>> > Let Crystal Reports handle the reporting - Free Crystal Reports 2008
>> 30-Day
>> > trial. Simplify your report design, integration and deployment - and
>> focus
>> > on
>> > what you do best, core application coding. Discover what's new with
>> > Crystal Reports now. http://p.sf.net/sfu/bobj-july
>> > _______________________________________________
>> > gumstix-users mailing list
>> > gumstix-users@...
>> > https://lists.sourceforge.net/lists/listinfo/gumstix-users
>> >
>> >
>>
>>
>>
>> --
>> Say you can or say you can't, either way you will be right.
>> Computers are like old testament gods: Lots of rules and no mercy.
>>
>>
>> ------------------------------------------------------------------------------
>> Let Crystal Reports handle the reporting - Free Crystal Reports 2008
>> 30-Day
>> trial. Simplify your report design, integration and deployment - and focus
>> on
>> what you do best, core application coding. Discover what's new with
>> Crystal Reports now. http://p.sf.net/sfu/bobj-july
>> _______________________________________________
>> gumstix-users mailing list
>> gumstix-users@...
>> https://lists.sourceforge.net/lists/listinfo/gumstix-users
>>
>

Re: IR_UART port as a standard UART

2009-11-25T09:22:44Z

HI Ricardo,

2009/11/25 Ricardo Martinez Monje <RMartinez@...>:
> Hello,
>
> Could I use the IR_UART port on the 40-pin header as a standard UART? Or
> that pins are used as the console port trough the FTDI ic?

I presume that you're talking about the overo?

Looking at the schematics on the summit, the IR_TXD3 and IR_RXD3
signals go to the FTDI (see page 4 of the schematics).

The TXD1 and RXD1 signals on the 40-pin connector should be fine.

--
Dave Hylands
Shuswap, BC, Canada
http://www.DaveHylands.com/

------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
trial. Simplify your report design, integration and deployment - and focus on
what you do best, core application coding. Discover what's new with
Crystal Reports now. http://p.sf.net/sfu/bobj-july
_______________________________________________
gumstix-users mailing list
gumstix-users@...
https://lists.sourceforge.net/lists/listinfo/gumstix-users

IR_UART port as a standard UART

2009-11-25T08:10:33Z

Hello,

Could I use the IR_UART port on the 40-pin header as a standard UART? Or that pins are used as the console port trough the FTDI ic?

Thank you.

Ricardo Martínez Monje

Network & Connectivity.

aurensis | Telespazio, a Finmeccanica/Thales Company

----------------

Rosa Sensat 9-11, 6ª p · 08005 Barcelona

Tel. (+34) 935 830 200 | Fax (+34) 935 830 201

---------------------------------------------------------

mail@...

http://www.aurensis.com

----------------

AURENSIS, S.L.U. le informa de que este mensaje, la información que contiene y cualesquiera archivos que adjunte se dirigen exclusivamente a su destinatario, estando prohibida por Ley su distribución, copia o utilización con cualquier finalidad sin autorización expresa del remitente. Si Vd. recibe este mensaje por error, le rogamos que proceda a su eliminación definitiva y nos lo comunique inmediatamente a través de correo electrónico.

AURENSIS, S.L.U. le informa, asimismo, de que todos los datos recibidos por esta vía serán almacenados, pudiendo ser incluidos en un fichero de nuestra responsabilidad, donde se mantendrán bajo la más estricta confidencialidad, con la finalidad de atender su solicitud y dar cumplimiento a los lícitos objetivos de la empresa. Puede ejercitar sus derechos de acceso, rectificación, cancelación y oposición, en los términos que reconoce la Ley Orgánica 15/1999 de Protección de Datos, dirigiéndose a nuestras oficinas en Barcelona, C/ Rosa Sensat, 9-11. 08005, o enviando un e-mail a lopd@...

Re: NTP server on gumstix

2009-11-25T07:52:18Z

Hi Joshua:

This is something I intend to do. But haven’t gotten to it yet. I’ll let you know if I make progress.

Thanks

Azhar

From: Joshua Garcia [mailto:jgarcia@...]
Sent: Tuesday, November 24, 2009 2:01 PM
To: gumstix-users@...
Subject: [Gumstix-users] NTP server on gumstix

Has anybody successfully installed and ran a NTP server on a gumstix COM using linux? If so, what version of linux did you use?

Thanks

Re: how to give gumstix user permission to reboot?

2009-11-25T06:29:00Z

Hi Dave,

the chmod u+s /sbin/reboot works

brilliant!

Dave Hylands wrote:

Adding your user to the group just allows access to the reboot file,
it won't make reboot run as root.

You could also get the user sticky bit:

chmod u+s reboot, and then reboot will run as root. Of course, using
sticky bits has security implications...

Re: cpufreq on the Overo

2009-11-25T04:29:15Z

ravix wrote:

Philip Balister wrote:

ravix wrote:
I think the low-level support for cpufreq is only in the linux-omap-pm
branch. I've run this on the beagle, but the stock overo kernel will
support more things.

There is a linux-omap-pm recipe that could be made to work for the overo
I think.

Also, when I try to run bitbake -c rebuild linux-omap-pm I get:
"ERROR: Nothing PROVIDES 'linux-omap-pm'"

Thanks

-Eric

Hi,

I've copied only "oe/recipes/images/omap3-console-image.bb" to "user.collection/recipes/images/my-image.bb"
Then I've added a line to the bottom of the file:
PREFERRED_VERSION_linux-omap = "pm_2.6.29"

Then:
bitbake my-image

In your local.conf... make sure, you've got BBFILES pointing to both oe recipes and user.collection recipes.
Also check your /oe/recipes/linux/...
Do you have linux-omap-pm* folders there ?

I baked the imaged OK, but I'm still having problems booting it properly... It gets stuck at "Starting bluetooth..."

Cheers,
Robert

Re: [Gumstix-Users] Use OpenEmbedded or not?

2009-11-25T04:01:30Z

If we develop directly on the Gumstix, about how much space would the development environment take up? I'm more familiar with software development on Windows where we can use MFC or .NET for display APIs, event handling, but for the Gumstix, we were considering using Eclipse and Qt Embedded with OpenEmbedded Tools plugins unless there are better options for the Gumstix. Space is limited, at least on NAND flash.

As for the overo-oe install, I'm talking about the build environment. I was trying to set it up on a Red Hat box following the 'setting up a build environment' instructions on gumstix.net's documentation pages. One of the first steps has me do a git clone from gitorious.org for org.openembedded.dev, but this fails every time, reporting an 'unexpected EOF' as it's trying to receive the files.

On Tue, Nov 24, 2009 at 4:17 PM, richard dorfner <rdorfner@...> wrote:

Howdy there, as always, the answers are not typically straight
forward. First though, a bit of nomenclature,

Open Embedded is the particular flavor of linux running on Gumstix processors.
The build system is Bitbake. Now, you can use the gumstix like any
linux system, and install the os and whatever tools you need,
including a compiler and simply develop the code natively on the
gumstix itself. There are pros and cons to this, however, the
biggest pro is hardly any learning curve, if you already know linux,
that is.

So no, you don't, strictly speaking, need to use the whole open
embedded cross development environment. If you decide to, though,
you will find the learning curve a bit steep. At least, for me it has
been, and I STILL am not comfortable creating my own user recipes at
all. But that's me, other folks find it intuitive and easy to
learn/use and understand.

Not sure what's going on with your overo-oe install, are you talking
about the build environment? Or getting the kernel onto the gumstix?
or...???

Rick

On Tue, Nov 24, 2009 at 2:18 PM, coder drone <coderdrone@...> wrote:
> Hello everyone,
>
> Our project recently acquired some Gumstix boards to develop software for an
> embedded system. I'm really quite new to the whole Embedded Linux world,
> and am trying to get a feel for how everything works. Is it necessary to
> use OpenEmbedded to develop software for Gumstix?
>
> I'm assuming OpenEmbedded provides a way for you to build the kernel,
> filesystem, drivers, etc., but if the board already has all of this,
> couldn't you just develop software and build it with the appropriate
> compiler for that processor?
>
> We're looking to use Eclipse to develop C++ code, maybe even using Qt
> embedded, to run on the Gumstix. I installed the OpenEmbedded Tools for
> Eclipse plugin, but it apparently assumes we already have a working
> installation of OpenEmbedded, but the overo-oe install keeps failing.
>
> Thanks.
>

> ------------------------------------------------------------------------------
> Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
> trial. Simplify your report design, integration and deployment - and focus
> on
> what you do best, core application coding. Discover what's new with
> Crystal Reports now. http://p.sf.net/sfu/bobj-july
> _______________________________________________
> gumstix-users mailing list
> gumstix-users@...
> https://lists.sourceforge.net/lists/listinfo/gumstix-users
>
>

--
Say you can or say you can't, either way you will be right.
Computers are like old testament gods: Lots of rules and no mercy.

------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
trial. Simplify your report design, integration and deployment - and focus on
what you do best, core application coding. Discover what's new with
Crystal Reports now. http://p.sf.net/sfu/bobj-july
_______________________________________________
gumstix-users mailing list
gumstix-users@...
https://lists.sourceforge.net/lists/listinfo/gumstix-users

Re: [Gumstix-Users] Use OpenEmbedded or not?

2009-11-25T02:37:55Z

On 24-11-09 22:17, richard dorfner wrote:
> Howdy there, as always, the answers are not typically straight
> forward. First though, a bit of nomenclature,
>
> Open Embedded is the particular flavor of linux running on Gumstix processors.

Bzzzzt! Wrong! OpenEmbedded is a buildsystem. Angstrom
(http://www.angstrom-distribution.org/) is the (linux) distribution
running on those boards.

------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
trial. Simplify your report design, integration and deployment - and focus on
what you do best, core application coding. Discover what's new with
Crystal Reports now. http://p.sf.net/sfu/bobj-july
_______________________________________________
gumstix-users mailing list
gumstix-users@...
https://lists.sourceforge.net/lists/listinfo/gumstix-users

Re: touch screen problem+chesnut43

2009-11-24T22:52:43Z

I am newbies in this area and I have not solved this problem

I am using standart board water+chesnut43 and gumstix image file (palmtop,console) and boot is
normally but why is correctly working touch screen I dont understand

thanks

ekrem wrote:

my configuration is overo water,chesnut43 and samsung lcd

I have search info in this forum about the touch screen but I dont found clearly info.

I several time booting overo plamtop image and other times Console image whit no errors
but touch screen is not working stable so not working correctly

if I must setting touch resistance reduce to 100 ohm, how I can do it

regards

Re: Module robostix not found?

2009-11-24T20:16:37Z

UPDATE: I've also tried changing the gumstix, rebuilding the image again and again but still the same problem. The kernel I'm using is 2.2.6.24.

Would appreciate any help. Thanks.

Re: Power for both COM _and_ expansion board?

2009-11-24T16:49:27Z

Ah yes... sorry; after doing some looking myself I could really only
find one thread and it was to do with the wireless, not the wired
solution.

On Tue, Nov 24, 2009 at 4:31 PM, Steve Sakoman <sakoman@...> wrote:

> On Tue, Nov 24, 2009 at 3:13 PM, Chris MacDonald
> <chris@...> wrote:
>
>> On the topic of the tobi and tobi duo though, I'd take some time and
>> look though the list archives... I personally don't have experience
>> with the ethernet interface on the tobi but from what I've read here
>> performance isn't as good as it could be, and that may have some
>> bearing on your purchasing decisions.
>
> I'm not aware of any significant performance issues with tobi.
>
> >From a local server I get 4.43 megabytes/sec transfer rate:
>
> root@omap3:~# wget http://192.168.0.220/oe/jeff-map.bin
> --16:27:12-- http://192.168.0.220/oe/jeff-map.bin
> => `jeff-map.bin.3'
> Connecting to 192.168.0.220:80... connected.
> HTTP request sent, awaiting response... 200 OK
> Length: 37,879,101 [application/octet-stream]
>
> 100%[====================================>] 37,879,101 4.43M/s ETA 00:00
>
> 16:27:26 (4.31 MB/s) - `jeff-map.bin.3' saved [37879101/37879101]
>
> The above writes the file to the microSD card. I repeated the test
> writing to ram disk and got 4.86 megabytes/sec:
>
> root@omap3:~# cd /tmp
> root@omap3:/var/volatile/tmp# wget http://192.168.0.220/oe/jeff-map.bin
> --16:27:35-- http://192.168.0.220/oe/jeff-map.bin
> => `jeff-map.bin'
> Connecting to 192.168.0.220:80... connected.
> HTTP request sent, awaiting response... 200 OK
> Length: 37,879,101 [application/octet-stream]
>
> 100%[====================================>] 37,879,101 4.86M/s ETA 00:00
>
> 16:27:42 (4.79 MB/s) - `jeff-map.bin' saved [37879101/37879101]
>
> Steve
>
> ------------------------------------------------------------------------------
> Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
> trial. Simplify your report design, integration and deployment - and focus on
> what you do best, core application coding. Discover what's new with
> Crystal Reports now. http://p.sf.net/sfu/bobj-july
> _______________________________________________
> gumstix-users mailing list
> gumstix-users@...
> https://lists.sourceforge.net/lists/listinfo/gumstix-users
>

Re: gpio-event not compiling

2009-11-24T16:48:50Z

HI Steven,

On Tue, Nov 24, 2009 at 2:32 PM, Steven Kane <steven.kane@...> wrote:

> After following the instructions for setting up my build environment found
> here:
>
> http://www.gumstix.net/Setup-and-Programming/view/Overo-Setup-and-Programming/Setting-up-a-build-environment/111.html
>
> and successfully compiling the 'omap3-console-image', I was unable to get
> the 'gpio-event' app to compile under bitbake found in
> 'org.openembedded.dev/recipes/gpio-event' checked out via the instructions
> above.
>

gpio-event is currently PXA255 or PXA270 specific. It needs some
tweaks (and updates) to work on the OMAP.

Here's a post I made a while back:
<http://thread.gmane.org/gmane.linux.distributions.gumstix.general/44916/focus=44928>

...snip...
> If anyone has any information on how to fix this OR if anyone has
> information on how to configure/read/write the GPIO ports on the Overo Earth
> COM on the Chestnut43 expansion board running the console image, your help
> would be much appreciated.

--
Dave Hylands
Shuswap, BC, Canada
http://www.DaveHylands.com/

------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
trial. Simplify your report design, integration and deployment - and focus on
what you do best, core application coding. Discover what's new with
Crystal Reports now. http://p.sf.net/sfu/bobj-july
_______________________________________________
gumstix-users mailing list
gumstix-users@...
https://lists.sourceforge.net/lists/listinfo/gumstix-users

Re: how to give gumstix user permission to reboot?

2009-11-24T16:45:25Z

Hi Alex,

On Tue, Nov 24, 2009 at 12:58 PM, Alex4 <alex.fielding@...> wrote:

>
> the conventional linux/unix approaches allowing a user to reboot don't seem
> to be working for me, and so part of this may be my learning how to do it on
> the gumstix linux distribution. I even tried adding the user to the root
> group in /etc/group and that doesn't seem to work either.
>
> btw, this is why it seems I need to reboot for the new ip address
> $etc/init.d/networking restart
> Reconfiguring network interfaces... Auto negotiation NOT supported
>
> I tried creating a group called rebooter then adding my user to that in
> /etc/group and it didn't work either
>
> lrwxrwxrwx 1 root rebooter 15 Jan 1 1970 reboot ->
> reboot.sysvinit

Adding your user to the group just allows access to the reboot file,
it won't make reboot run as root.

You could also get the user sticky bit:

chmod u+s reboot, and then reboot will run as root. Of course, using
sticky bits has security implications...

--
Dave Hylands
Shuswap, BC, Canada
http://www.DaveHylands.com/

------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
trial. Simplify your report design, integration and deployment - and focus on
what you do best, core application coding. Discover what's new with
Crystal Reports now. http://p.sf.net/sfu/bobj-july
_______________________________________________
gumstix-users mailing list
gumstix-users@...
https://lists.sourceforge.net/lists/listinfo/gumstix-users

Re: Power for both COM _and_ expansion board?

2009-11-24T16:31:12Z

On Tue, Nov 24, 2009 at 3:13 PM, Chris MacDonald
<chris@...> wrote:

> On the topic of the tobi and tobi duo though, I'd take some time and
> look though the list archives... I personally don't have experience
> with the ethernet interface on the tobi but from what I've read here
> performance isn't as good as it could be, and that may have some
> bearing on your purchasing decisions.

I'm not aware of any significant performance issues with tobi.

>From a local server I get 4.43 megabytes/sec transfer rate:

root@omap3:~# wget http://192.168.0.220/oe/jeff-map.bin
--16:27:12-- http://192.168.0.220/oe/jeff-map.bin
=> `jeff-map.bin.3'
Connecting to 192.168.0.220:80... connected.
HTTP request sent, awaiting response... 200 OK
Length: 37,879,101 [application/octet-stream]

100%[====================================>] 37,879,101 4.43M/s ETA 00:00

16:27:26 (4.31 MB/s) - `jeff-map.bin.3' saved [37879101/37879101]

The above writes the file to the microSD card. I repeated the test
writing to ram disk and got 4.86 megabytes/sec:

root@omap3:~# cd /tmp
root@omap3:/var/volatile/tmp# wget http://192.168.0.220/oe/jeff-map.bin
--16:27:35-- http://192.168.0.220/oe/jeff-map.bin
=> `jeff-map.bin'
Connecting to 192.168.0.220:80... connected.
HTTP request sent, awaiting response... 200 OK
Length: 37,879,101 [application/octet-stream]

100%[====================================>] 37,879,101 4.86M/s ETA 00:00

16:27:42 (4.79 MB/s) - `jeff-map.bin' saved [37879101/37879101]

Steve

------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
trial. Simplify your report design, integration and deployment - and focus on
what you do best, core application coding. Discover what's new with
Crystal Reports now. http://p.sf.net/sfu/bobj-july
_______________________________________________
gumstix-users mailing list
gumstix-users@...
https://lists.sourceforge.net/lists/listinfo/gumstix-users

Re: Power for both COM _and_ expansion board?

2009-11-24T15:13:55Z

Yup, you need to use some kind of expansion board. Gumstix offers a
range of expansion boards (you mentioned the Tobi Duo) and all the
drawings are available online (http://pubs.gumstix.com/boards/) with
the intention that you can either use one of those boards, or with the
help of the drawings and other documentation you can make your own to
suit your needs.

On the topic of the tobi and tobi duo though, I'd take some time and
look though the list archives... I personally don't have experience
with the ethernet interface on the tobi but from what I've read here
performance isn't as good as it could be, and that may have some
bearing on your purchasing decisions.

Chris

On Tue, Nov 24, 2009 at 2:21 PM, Jake Peavy <djstunks@...> wrote:

> On Tue, Nov 24, 2009 at 4:48 PM, Chris MacDonald <chris@...>
> wrote:
>>
>> The overo COMs actually get power from their connection (the two
>> 70-pin connectors) to the expansion board, they don't have a barrel of
>> their own.
>>
>> Chris
>>
>> On Tue, Nov 24, 2009 at 1:35 PM, Jake Peavy <djstunks@...> wrote:
>> > Hi all,
>> >
>> > I'm new around these here parts, but I had a quick (probably stupid)
>> > question.
>> >
>> > I'm looking into creating a device with the Tobi-Duo dual-ethernet
>> > expansion
>> > board. My understanding is that in order to do anything with this I
>> > also
>> > need a COM module which mates with the expansion board via the two
>> > 70-pin
>> > connectors.
>> >
>> > My question is that both the COM module (say the plain ol' Overo Earth)
>> > and
>> > the Tobi-Duo have +5V barrel connectors.
>> >
>> > Does that mean I need two power supplies for my device? Or is the
>> > expansion
>> > board powered by the COM module via the two connectors?
>> >
>
> aha, I looked at the up close pics of the Overo Earth COM. of course you're
> right, no barrel connector.
>
> I guess that means you must use at least one expansion board?
>
> Thanks,
>
> --
> -jp
>
> Just as irrigation is the lifeblood of the Southwest, lifeblood is the soup
> of cannibals.
>
> deepthoughtsbyjackhandey.com
>
>
>
> ------------------------------------------------------------------------------
> Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
> trial. Simplify your report design, integration and deployment - and focus
> on
> what you do best, core application coding. Discover what's new with
> Crystal Reports now. http://p.sf.net/sfu/bobj-july
> _______________________________________________
> gumstix-users mailing list
> gumstix-users@...
> https://lists.sourceforge.net/lists/listinfo/gumstix-users
>
>

gpio-event not compiling

2009-11-24T14:32:37Z

After following the instructions for setting up my build environment found here:

http://www.gumstix.net/Setup-and-Programming/view/Overo-Setup-and-Programming/Setting-up-a-build-environment/111.html

and successfully compiling the 'omap3-console-image', I was unable to get the 'gpio-event' app to compile under bitbake found in 'org.openembedded.dev/recipes/gpio-event' checked out via the instructions above.

My development environment is running in Ubuntu 9.10 and I am trying to compile for the Overo Earth COM on the Chestnut43 expansion board running the image found at: http://www.sakoman.com/feeds/omap3/glibc/images/overo/200910121635/omap3-console-image-overo-200910121635.tar.bz2

Here is what I get when I run 'bitbake gpio-event':

WARNING: No matches in filesystem for /home/steve/overo-oe/user.collection in COLLECTIONS
NOTE: Psyco JIT Compiler (http://psyco.sf.net) not available. Install it to increase performance.
NOTE: Handling BitBake files: | (7489/7489) [100 %]
NOTE: Parsing finished. 7196 cached, 0 parsed, 293 skipped, 2 masked.
NOTE: Cache is clean, not saving.
NOTE: Resolving any missing task queue dependencies
NOTE: Preparing runqueue
NOTE: Executing runqueue
NOTE: Running task 465 of 652 (ID: 8, /home/steve/overo-oe/org.openembedded.dev/recipes/gpio-event/gpio-event.bb, do_compile)
ERROR: function do_compile failed
ERROR: log data follows (/home/steve/overo-oe/tmp/work/overo-angstrom-linux-gnueabi/gpio-event-1.0-r49/temp/log.do_compile.3622)
| NOTE: make CROSS_COMPILE=arm-angstrom-linux-gnueabi- KERNELDIR=/home/steve/overo-oe/tmp/work/overo-angstrom-linux-gnueabi/linux-omap3-2.6.30-r49/git CC=arm-angstrom-linux-gnueabi-gcc -march=armv7-a -mtune=cortex-a8 -mfpu=neon -mfloat-abi=softfp
| make -C /home/steve/overo-oe/tmp/work/overo-angstrom-linux-gnueabi/linux-omap3-2.6.30-r49/git M=/home/steve/overo-oe/tmp/work/overo-angstrom-linux-gnueabi/gpio-event-1.0-r49/gpio-event ARCH=arm CROSS_COMPILE=arm-angstrom-linux-gnueabi- modules
| make[1]: Entering directory `/home/steve/overo-oe/tmp/work/overo-angstrom-linux-gnueabi/linux-omap3-2.6.30-r49/git'
| scripts/Makefile.build:49: *** CFLAGS was changed in "/home/steve/overo-oe/tmp/work/overo-angstrom-linux-gnueabi/gpio-event-1.0-r49/gpio-event/Makefile". Fix it to use EXTRA_CFLAGS. Stop.
| make[1]: *** [_module_/home/steve/overo-oe/tmp/work/overo-angstrom-linux-gnueabi/gpio-event-1.0-r49/gpio-event] Error 2
| make[1]: Leaving directory `/home/steve/overo-oe/tmp/work/overo-angstrom-linux-gnueabi/linux-omap3-2.6.30-r49/git'
| make: *** [kernel-module] Error 2
| FATAL: oe_runmake failed
NOTE: Task failed: /home/steve/overo-oe/tmp/work/overo-angstrom-linux-gnueabi/gpio-event-1.0-r49/temp/log.do_compile.3622
ERROR: TaskFailed event exception, aborting
ERROR: Build of /home/steve/overo-oe/org.openembedded.dev/recipes/gpio-event/gpio-event.bb do_compile failed
ERROR: Task 8 (/home/steve/overo-oe/org.openembedded.dev/recipes/gpio-event/gpio-event.bb, do_compile) failed
NOTE: Tasks Summary: Attempted 464 tasks of which 464 didn't need to be rerun and 1 failed.
ERROR: '/home/steve/overo-oe/org.openembedded.dev/recipes/gpio-event/gpio-event.bb' failed

Following the instructions in the error output I edited "/home/steve/overo-oe/tmp/work/overo-angstrom-linux-gnueabi/gpio-event-1.0-r49/gpio-event/Makefile" and changed all instances of 'CFLAGS' to 'EXTRA_CFLAGS', and running 'bitbake gpio-event' again, I got this new set of errors:

WARNING: No matches in filesystem for /home/steve/overo-oe/user.collection in COLLECTIONS
NOTE: Psyco JIT Compiler (http://psyco.sf.net) not available. Install it to increase performance.
NOTE: Handling BitBake files: | (7489/7489) [100 %]
NOTE: Parsing finished. 7196 cached, 0 parsed, 293 skipped, 2 masked.
NOTE: Cache is clean, not saving.
NOTE: Resolving any missing task queue dependencies
NOTE: Preparing runqueue
NOTE: Executing runqueue
NOTE: Running task 465 of 652 (ID: 8, /home/steve/overo-oe/org.openembedded.dev/recipes/gpio-event/gpio-event.bb, do_compile)
ERROR: function do_compile failed
ERROR: log data follows (/home/steve/overo-oe/tmp/work/overo-angstrom-linux-gnueabi/gpio-event-1.0-r49/temp/log.do_compile.3779)
| NOTE: make CROSS_COMPILE=arm-angstrom-linux-gnueabi- KERNELDIR=/home/steve/overo-oe/tmp/work/overo-angstrom-linux-gnueabi/linux-omap3-2.6.30-r49/git CC=arm-angstrom-linux-gnueabi-gcc -march=armv7-a -mtune=cortex-a8 -mfpu=neon -mfloat-abi=softfp
| make -C /home/steve/overo-oe/tmp/work/overo-angstrom-linux-gnueabi/linux-omap3-2.6.30-r49/git M=/home/steve/overo-oe/tmp/work/overo-angstrom-linux-gnueabi/gpio-event-1.0-r49/gpio-event ARCH=arm CROSS_COMPILE=arm-angstrom-linux-gnueabi- modules
| make[1]: Entering directory `/home/steve/overo-oe/tmp/work/overo-angstrom-linux-gnueabi/linux-omap3-2.6.30-r49/git'
| CC [M] /home/steve/overo-oe/tmp/work/overo-angstrom-linux-gnueabi/gpio-event-1.0-r49/gpio-event/gpio-event-drv.o
| /home/steve/overo-oe/tmp/work/overo-angstrom-linux-gnueabi/gpio-event-1.0-r49/gpio-event/gpio-event-drv.c:41:31: error: asm/arch/hardware.h: No such file or directory
| /home/steve/overo-oe/tmp/work/overo-angstrom-linux-gnueabi/gpio-event-1.0-r49/gpio-event/gpio-event-drv.c:42:31: error: asm/arch/pxa-regs.h: No such file or directory
| /home/steve/overo-oe/tmp/work/overo-angstrom-linux-gnueabi/gpio-event-1.0-r49/gpio-event/gpio-event-drv.c: In function 'gpio_event_irq':
| /home/steve/overo-oe/tmp/work/overo-angstrom-linux-gnueabi/gpio-event-1.0-r49/gpio-event/gpio-event-drv.c:537: error: implicit declaration of function 'GPLR'
| /home/steve/overo-oe/tmp/work/overo-angstrom-linux-gnueabi/gpio-event-1.0-r49/gpio-event/gpio-event-drv.c:537: error: implicit declaration of function 'GPIO_bit'
| /home/steve/overo-oe/tmp/work/overo-angstrom-linux-gnueabi/gpio-event-1.0-r49/gpio-event/gpio-event-drv.c: In function 'gpio_event_timer':
| /home/steve/overo-oe/tmp/work/overo-angstrom-linux-gnueabi/gpio-event-1.0-r49/gpio-event/gpio-event-drv.c:652: error: implicit declaration of function 'IRQ_GPIO'
| /home/steve/overo-oe/tmp/work/overo-angstrom-linux-gnueabi/gpio-event-1.0-r49/gpio-event/gpio-event-drv.c: In function 'gpio_event_init':
| /home/steve/overo-oe/tmp/work/overo-angstrom-linux-gnueabi/gpio-event-1.0-r49/gpio-event/gpio-event-drv.c:1149: error: implicit declaration of function 'class_device_create'
| /home/steve/overo-oe/tmp/work/overo-angstrom-linux-gnueabi/gpio-event-1.0-r49/gpio-event/gpio-event-drv.c: In function 'gpio_event_exit':
| /home/steve/overo-oe/tmp/work/overo-angstrom-linux-gnueabi/gpio-event-1.0-r49/gpio-event/gpio-event-drv.c:1187: error: implicit declaration of function 'class_device_destroy'
| make[2]: *** [/home/steve/overo-oe/tmp/work/overo-angstrom-linux-gnueabi/gpio-event-1.0-r49/gpio-event/gpio-event-drv.o] Error 1
| make[1]: *** [_module_/home/steve/overo-oe/tmp/work/overo-angstrom-linux-gnueabi/gpio-event-1.0-r49/gpio-event] Error 2
| make[1]: Leaving directory `/home/steve/overo-oe/tmp/work/overo-angstrom-linux-gnueabi/linux-omap3-2.6.30-r49/git'
| make: *** [kernel-module] Error 2
| FATAL: oe_runmake failed
NOTE: Task failed: /home/steve/overo-oe/tmp/work/overo-angstrom-linux-gnueabi/gpio-event-1.0-r49/temp/log.do_compile.3779
ERROR: TaskFailed event exception, aborting
ERROR: Build of /home/steve/overo-oe/org.openembedded.dev/recipes/gpio-event/gpio-event.bb do_compile failed
ERROR: Task 8 (/home/steve/overo-oe/org.openembedded.dev/recipes/gpio-event/gpio-event.bb, do_compile) failed
NOTE: Tasks Summary: Attempted 464 tasks of which 464 didn't need to be rerun and 1 failed.
ERROR: '/home/steve/overo-oe/org.openembedded.dev/recipes/gpio-event/gpio-event.bb' failed

If anyone has any information on how to fix this OR if anyone has information on how to configure/read/write the GPIO ports on the Overo Earth COM on the Chestnut43 expansion board running the console image, your help would be much appreciated.

Steven Kane

------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
trial. Simplify your report design, integration and deployment - and focus on
what you do best, core application coding. Discover what's new with
Crystal Reports now. http://p.sf.net/sfu/bobj-july
_______________________________________________
gumstix-users mailing list
gumstix-users@...
https://lists.sourceforge.net/lists/listinfo/gumstix-users

Re: Power for both COM _and_ expansion board?

2009-11-24T14:21:15Z

On Tue, Nov 24, 2009 at 4:48 PM, Chris MacDonald <chris@...> wrote:

The overo COMs actually get power from their connection (the two
70-pin connectors) to the expansion board, they don't have a barrel of
their own.

Chris

On Tue, Nov 24, 2009 at 1:35 PM, Jake Peavy <djstunks@...> wrote:
> Hi all,
>
> I'm new around these here parts, but I had a quick (probably stupid)
> question.
>
> I'm looking into creating a device with the Tobi-Duo dual-ethernet expansion
> board. My understanding is that in order to do anything with this I also
> need a COM module which mates with the expansion board via the two 70-pin
> connectors.
>
> My question is that both the COM module (say the plain ol' Overo Earth) and
> the Tobi-Duo have +5V barrel connectors.
>
> Does that mean I need two power supplies for my device? Or is the expansion
> board powered by the COM module via the two connectors?
>

aha, I looked at the up close pics of the Overo Earth COM. of course you're right, no barrel connector.

I guess that means you must use at least one expansion board?

Thanks,

--
-jp

Just as irrigation is the lifeblood of the Southwest, lifeblood is the soup of cannibals.

deepthoughtsbyjackhandey.com

------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
trial. Simplify your report design, integration and deployment - and focus on
what you do best, core application coding. Discover what's new with
Crystal Reports now. http://p.sf.net/sfu/bobj-july
_______________________________________________
gumstix-users mailing list
gumstix-users@...
https://lists.sourceforge.net/lists/listinfo/gumstix-users

Re: Power for both COM _and_ expansion board?

2009-11-24T13:48:15Z

The overo COMs actually get power from their connection (the two
70-pin connectors) to the expansion board, they don't have a barrel of
their own.

Chris

On Tue, Nov 24, 2009 at 1:35 PM, Jake Peavy <djstunks@...> wrote:

> Hi all,
>
> I'm new around these here parts, but I had a quick (probably stupid)
> question.
>
> I'm looking into creating a device with the Tobi-Duo dual-ethernet expansion
> board. My understanding is that in order to do anything with this I also
> need a COM module which mates with the expansion board via the two 70-pin
> connectors.
>
> My question is that both the COM module (say the plain ol' Overo Earth) and
> the Tobi-Duo have +5V barrel connectors.
>
> Does that mean I need two power supplies for my device? Or is the expansion
> board powered by the COM module via the two connectors?
>
> TYVM
>
> --
> -jp
>
> I think man invented the car by instinct.
>
> deepthoughtsbyjackhandey.com
>
>
> ------------------------------------------------------------------------------
> Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
> trial. Simplify your report design, integration and deployment - and focus
> on
> what you do best, core application coding. Discover what's new with
> Crystal Reports now. http://p.sf.net/sfu/bobj-july
> _______________________________________________
> gumstix-users mailing list
> gumstix-users@...
> https://lists.sourceforge.net/lists/listinfo/gumstix-users
>
>

Power for both COM _and_ expansion board?

2009-11-24T13:35:24Z

Hi all,

I'm new around these here parts, but I had a quick (probably stupid) question.

I'm looking into creating a device with the Tobi-Duo dual-ethernet expansion board. My understanding is that in order to do anything with this I also need a COM module which mates with the expansion board via the two 70-pin connectors.

My question is that both the COM module (say the plain ol' Overo Earth) and the Tobi-Duo have +5V barrel connectors.

Does that mean I need two power supplies for my device? Or is the expansion board powered by the COM module via the two connectors?

TYVM

--
-jp

I think man invented the car by instinct.

deepthoughtsbyjackhandey.com

------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
trial. Simplify your report design, integration and deployment - and focus on
what you do best, core application coding. Discover what's new with
Crystal Reports now. http://p.sf.net/sfu/bobj-july
_______________________________________________
gumstix-users mailing list
gumstix-users@...
https://lists.sourceforge.net/lists/listinfo/gumstix-users

Re: [Gumstix-Users] Use OpenEmbedded or not?

2009-11-24T13:17:28Z

Howdy there, as always, the answers are not typically straight
forward. First though, a bit of nomenclature,

Open Embedded is the particular flavor of linux running on Gumstix processors.
The build system is Bitbake. Now, you can use the gumstix like any
linux system, and install the os and whatever tools you need,
including a compiler and simply develop the code natively on the
gumstix itself. There are pros and cons to this, however, the
biggest pro is hardly any learning curve, if you already know linux,
that is.

So no, you don't, strictly speaking, need to use the whole open
embedded cross development environment. If you decide to, though,
you will find the learning curve a bit steep. At least, for me it has
been, and I STILL am not comfortable creating my own user recipes at
all. But that's me, other folks find it intuitive and easy to
learn/use and understand.

Not sure what's going on with your overo-oe install, are you talking
about the build environment? Or getting the kernel onto the gumstix?
or...???

Rick

On Tue, Nov 24, 2009 at 2:18 PM, coder drone <coderdrone@...> wrote:

> Hello everyone,
>
> Our project recently acquired some Gumstix boards to develop software for an
> embedded system. I'm really quite new to the whole Embedded Linux world,
> and am trying to get a feel for how everything works. Is it necessary to
> use OpenEmbedded to develop software for Gumstix?
>
> I'm assuming OpenEmbedded provides a way for you to build the kernel,
> filesystem, drivers, etc., but if the board already has all of this,
> couldn't you just develop software and build it with the appropriate
> compiler for that processor?
>
> We're looking to use Eclipse to develop C++ code, maybe even using Qt
> embedded, to run on the Gumstix. I installed the OpenEmbedded Tools for
> Eclipse plugin, but it apparently assumes we already have a working
> installation of OpenEmbedded, but the overo-oe install keeps failing.
>
> Thanks.
>
> ------------------------------------------------------------------------------
> Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
> trial. Simplify your report design, integration and deployment - and focus
> on
> what you do best, core application coding. Discover what's new with
> Crystal Reports now. http://p.sf.net/sfu/bobj-july
> _______________________________________________
> gumstix-users mailing list
> gumstix-users@...
> https://lists.sourceforge.net/lists/listinfo/gumstix-users
>
>

--
Say you can or say you can't, either way you will be right.
Computers are like old testament gods: Lots of rules and no mercy.

------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
trial. Simplify your report design, integration and deployment - and focus on
what you do best, core application coding. Discover what's new with
Crystal Reports now. http://p.sf.net/sfu/bobj-july
_______________________________________________
gumstix-users mailing list
gumstix-users@...
https://lists.sourceforge.net/lists/listinfo/gumstix-users