Re: [Gumstix-users] Gumstix Overo SPI read problems

SourceForge Headquarters 225 Broadway Suite 1600 San Diego, CA 92101 +1 (858) 454-5900

Hi Markus, 

i have triggering on my scope and i send the message periodically in the test_program, plus i shortcut miso and mosi to receive the same value i send. instead i only get 0x00.

the bits_per_word and max_speed_hz are set to 8, resp. 1000000 in my kernel driver, its only in the test program in the struct spi_transfer where i set them to NULL or 0 or commented them out, because some people wrote that they had issues with that. A bug or something...

Thanks for your test-program though, i will definitely try that one tomorrow as well.

Regards, 
Norman

-------- Original-Nachricht --------
> Datum: Mon, 25 Oct 2010 16:39:05 -0400
> Von: Markus Svilans <msv...@ae...>
> An: gum...@li...
> Betreff: Re: [Gumstix-users] Gumstix Overo SPI read problems

> Hi Norman,
> 
> I never tried the spidev_test.c program yet. However I have some other 
> tips that may help.
> 
> Instead I used my own program (code is attached).
> 
> I installed the task-native-sdk package onto my Gumstix, and then used 
> the following command to build the program directly on the Gumstix:
> 
> g++ -Wall test-spi.cc -o test-spi
> 
> After I configured SPI in my kernel, as I outlined below for Will, I was 
> able to see activity on the SPI signal lines when reading and writing.
> 
> In your case I would suggest trying bits_per_word=8 and 
> max_speed_hz=1000000 (1 million). I'm not sure what happens when you set 
> these to zero, but I imagine nothing useful.
> 
> Now when you try to read and write from /dev/spidev1.0, you should see 
> the chip select line change and the clock line should have some pulses. 
> Also know that the clock will only have activity while you are 
> reading/writing. Since the SPI clock rate is in the MHz range, you 
> should enable triggering on your scope, otherwise you may not notice the 
> read/write events because they are so far.
> 
> Markus.
> 
> 
> 
> On 10-10-25 04:17 PM, nw236 wrote:
> > Hi Will and Markus,
> >
> > I am working on the same problem right now actually.
> >
> > I have an Overo Air hooked up to a summit board and am building the same
> > kernel-version like Will: linux-omap-2.6.34.
> >
> > I patched my kernel to disable ADS7846 and enable spidev and i see
> > spidev1.0, but for some reason i cant see any signal when connecting the
> > pins directly to a scope when i run the spidev_test program. Neither do
> i
> > receive the messages i send if i shortcut MISO and MOSI.
> >
> > Do you perhaps have any idea what could be the reason for that? I also
> set
> > max_speed_hz and bits_per_word to zero in my spidev_test program as
> > suggested in some posts, nothing helped.
> >
> > Regards,
> > Norman
> >
> >
> > will2km wrote:
> >> Hi Markus,
> >>
> >> Awsome! It's working now. One problem though, I wasn't able to
> >> configure the modules by using "bitbake -c menuconfig
> >> linux-omap3-2.6.34", a window popped up and disappeared very quickly.
> >> So I had to manually change the defconfig file.
> >>
> >> But anyways, it's working. Thank you very much!
> >>
> >>
> >> Regards,
> >>
> >>
> >> Will
> >>
> >> Markus Svilans-2 wrote:
> >>> On 10-10-05 02:19 PM, will2km wrote:
> >>>> Hi Markus,
> >>>>
> >>>> Would you please give me a detailed walkthrough of how you
> successfully
> >>>> enable your SPI? I've tried all the methods I can find in this mail
> list
> >>>> but
> >>>> none of them worked for me.
> >>>>
> >>>> Thank you so much.
> >>>>
> >>>> Regards,
> >>>>
> >>>> Will
> >>>>
> >>>>
> >>>
> >>> Hi Will,
> >>>
> >>> There are a number of steps involved to enable communication to your
> SPI
> >>> device. The information below is based on my own research, and
> probably
> >>> people more experienced than me with Linux and SPI will have some
> >>> corrections and improvements to make.
> >>>
> >>> 1. You will have to make some configuration changes to your kernel.
> Make
> >>> sure the touchscreen is disabled (ADS7846). Make sure the spidev
> module
> >>> is enabled. You can reconfigure your kernel by using this command:
> >>>
> >>>       bitbake -c menuconfig linux-omap3-2.6.34
> >>>
> >>> (This assumes you are using the Linux 2.6.34 kernel.)
> >>>
> >>> This page on the Gumstix Wiki explains in more detail how to configure
> >>> the kernel and rebuild it:
> >>> http://gumstix.net/wiki/index.php?title=Kernel_Reconfiguration
> >>>
> >>> Note that you can use the '/' key to search for things in the
> menuconfig
> >>> program, e.g. the ADS7846 touchscreen.
> >>>
> >>> 2. Modify your board-overo.c file. The file is located here:
> >>>
> ${OVEROTOP}/tmp/work/overo-angstrom-linux-gnueabi/linux-omap3-2.6.34-r81/git/arch/arm/mach-omap2/board-overo.c
> >>>
> >>> There are two major changes to make.
> >>>
> >>> Change #1
> >>> At lines 73 and 74, I made the following changes:
> >>>
> >>> //#if defined(CONFIG_TOUCHSCREEN_ADS7846) || \
> >>> //    defined(CONFIG_TOUCHSCREEN_ADS7846_MODULE)
> >>> #if 0
> >>>
> >>> This basically disables the touchscreen stuff, even if you have them
> >>> enabled in the kernel configuration.
> >>>
> >>> Change #2
> >>> Next, go to line 619.
> >>> This is where you will provide information to the Linux kernel about
> >>> your SPI device, so that it will allow you to communicate with it
> using
> >>> the open(), read() and write() system functions.
> >>>
> >>> In this section, I commented out the existing definitions, and added
> my
> >>> own. The changed file now looks like this:
> >>>
> >>> static struct spi_board_info overo_spi_board_info[] __initdata = {
> >>> /*
> >>> #if defined(CONFIG_TOUCHSCREEN_ADS7846) || \
> >>>       defined(CONFIG_TOUCHSCREEN_ADS7846_MODULE)
> >>>       {
> >>>           .modalias        = "ads7846",
> >>>           .bus_num        = 1,
> >>>           .chip_select        = 0,
> >>>           .max_speed_hz        = 1500000,
> >>>           .controller_data    =&ads7846_mcspi_config,
> >>>           .irq            = OMAP_GPIO_IRQ(OVERO_GPIO_PENDOWN),
> >>>           .platform_data        =&ads7846_config,
> >>>       },
> >>> #endif
> >>> #if defined(CONFIG_PANEL_LGPHILIPS_LB035Q02) || \
> >>>       defined(CONFIG_PANEL_LGPHILIPS_LB035Q02_MODULE)
> >>>       {
> >>>           .modalias        = "lgphilips_lb035q02_panel-spi",
> >>>           .bus_num        = 1,
> >>>           .chip_select        = 1,
> >>>           .max_speed_hz        = 500000,
> >>>           .mode            = SPI_MODE_3,
> >>>       },
> >>> #endif
> >>> */
> >>>       {  // this is the part you need to add for your SPI device
> >>>           .modalias = "spidev",
> >>>           .bus_num = 1,
> >>>           .chip_select = 0,
> >>>           .max_speed_hz = 1000000,
> >>>           .mode = SPI_MODE_3  // clock starts high, and sampling is
> done
> >>> on rising edge
> >>>       }
> >>> };
> >>>
> >>> Some notes:
> >>> * The "bus_num" parameter should be 1, if your SPI device is connected
> >>> to the SPI pins on the 40-pin header on the Tobi and Summit boards.
> The
> >>> other parameters depend on the SPI device you want to communicate
> with.
> >>> * The "chip_select" parameter is sort of like the address of your SPI
> >>> device. IIRC, the Gumstix Overo supports 3 different chip select
> values,
> >>> meaning you can have 3 SPI devices on the bus.
> >>> * The "mode" parameter is explained in the kernel SPI documentation
> (see
> >>> link below).
> >>>
> >>> 3. Rebuild, install and boot your modified kernel. (See the Gumstix
> Wiki
> >>> page, mentioned above, for detailed info on this step.)
> >>>
> >>> 4. When you reboot, if all went well, you should be able to type "ls
> >>> /dev" at the command prompt and see the new "spidev1.0" device in the
> >>> listing. Now you should be able to use the open(), read() and write()
> >>> system calls to send and receive data.
> >>>
> >>>
> >>> As you can see, SPI communication is non-trivial compared to other
> >>> methods, such as RS-232 or I2C for example.
> >>>
> >>> You should become familiar with the Linux kernel SPI documentation
> >>> (http://www.mjmwired.net/kernel/Documentation/spi/spi-summary). There
> >>> are various other parameters that you can use to configure spidev for
> >>> your SPI device. There is also lots of useful background info on
> spidev
> >>> and its limitations.
> >>>
> >>> Hope this helps. I wouldn't mind knowing how it turned out for you.
> >>>
> >>> Best regards,
> >>> Markus.
> >>>
> >>>
> >>>
> >>>> Markus Svilans-2 wrote:
> >>>>
> >>>>> Dear gumstix-users,
> >>>>>
> >>>>> I am trying to use SPI to read data from an ADS7476 analog to
> digital
> >>>>> converter. The sample rate will be slow, on the order of 10 to 100
> >>>>> samples per second. I would like to make the SPI device available
> from
> >>>>> user space, so that I can just read values from the A/D whenever
> >>>>> needed.
> >>>>> The timing requirements are loose enough that this should be ok.
> >>>>>
> >>>>> Data sheet for ADS7476:
> >>>>> http://www.national.com/profile/snip.cgi/openDS=ADCS7476
> >>>>>
> >>>>> In the SPI board info in board-overo.c, I have this:
> >>>>>
> >>>>> static struct spi_board_info overo_spi_board_info[] __initdata = {
> >>>>>        {
> >>>>>            .modalias = "spidev",
> >>>>>            .bus_num = 1,
> >>>>>            .chip_select = 0,
> >>>>>            .max_speed_hz = 1000000,
> >>>>>        }
> >>>>> };
> >>>>>
> >>>>>
> >>>>> I commented out the entries for the various touchscreens (ADS7846,
> >>>>> etc.).
> >>>>>
> >>>>> When I rebuilt the kernel and booted up, /dev/spi1.0 was present. So
> >>>>> far
> >>>>> so good.
> >>>>>
> >>>>> Next, I wrote a little C++ program to read 2 bytes (16 bit sample)
> from
> >>>>> /dev/spi1.0 at 1 second intervals. This was only partially
> successful.
> >>>>> There were 4 or 5 successful reads, and then read() returned -1, and
> >>>>> perror() just said "Input/output error". Then a few more successful
> >>>>> reads followed, and then another failed read.
> >>>>>
> >>>>> During this testing, the actual A/D was not present on the SPI bus.
> I
> >>>>> was using a scope to check that the Gumstix was generating the chip
> >>>>> select and clock signals properly.
> >>>>>
> >>>>> Can anyone help determine why some SPI reads are failing?
> >>>>>
> >>>>> Many thanks,
> >>>>> Markus.
> >>>>>
> >>>>>
> >>>>>
> ------------------------------------------------------------------------------
> >>>>> Start uncovering the many advantages of virtual appliances
> >>>>> and start using them to simplify application deployment and
> >>>>> accelerate your shift to cloud computing.
> >>>>> http://p.sf.net/sfu/novell-sfdev2dev
> >>>>> _______________________________________________
> >>>>> gumstix-users mailing list
> >>>>> gum...@li...
> >>>>> https://lists.sourceforge.net/lists/listinfo/gumstix-users
> >>>>>
> >>>>>
> >>>>>
> >>>>
> >>>
> >>>
> ------------------------------------------------------------------------------
> >>> Beautiful is writing same markup. Internet Explorer 9 supports
> >>> standards for HTML5, CSS3, SVG 1.1,  ECMAScript5, and DOM L2&  L3.
> >>> Spend less time writing and  rewriting code and more time creating
> great
> >>> experiences on the web. Be a part of the beta today.
> >>> http://p.sf.net/sfu/beautyoftheweb
> >>> _______________________________________________
> >>> gumstix-users mailing list
> >>> gum...@li...
> >>> https://lists.sourceforge.net/lists/listinfo/gumstix-users
> >>>
> >>>
> >>
> 

-- 
Neu: GMX De-Mail - Einfach wie E-Mail, sicher wie ein Brief!  
Jetzt De-Mail-Adresse reservieren: http://portal.gmx.net/de/go/demail