2006-05-25 18:18:32     ppifcd_test on BF537-STAMP

Document created by Aaronwu Employee on Aug 7, 2013
Version 1Show Document
  • View in full screen mode

2006-05-25 18:18:32     ppifcd_test on BF537-STAMP

Michael Benjamin (UNITED STATES)

Message: 7963    Does the ppifcd_test Blackfin-Test program work on the BF537-STAMP board? If so, could someone help me?


I currently have the following Kernel settings:

Character devices

[*] Blackfin BF53x Programmable Flags Driver

[*] Blackfin BF53x PPI Camera frame capture driver

I2C support

[*] I2C support

[*] I2C device interface


I've tried the "Generic Blackfin and HHBF533/561 development board I2C support" and "BFIN I2C SDA/SCL Selection-->" with "(2)SDA" and "(1) SCL)"

as described here:


but was told not to use the generic I2C driver.


[Before, I tried SDA=2, SCL=1 as described in the 'Frame Capture driver test(Net camera)' section of "Testing the uClinux Kernel" and also SDA=2, SCL=4 as described in the schematic for the PPI CMOS camera of "Sample PPI Driver."]


When I used the generic I2C driver, I had errors reading and writing (from i2c_write_register and i2c_read_register in ppifcd-test/i2c.c).

Now that I am not using that driver, I have "Error: could not open /dev/i2c-0".


Any help is greatly appreciated,






2006-05-25 18:33:58     RE: ppifcd_test on BF537-STAMP

Skye Sweeney (UNITED STATES)

Message: 7964    I am currently working on a PPI. I started with the ppifcd driver. I am not sure how it ever worked as I needed some fairly significant changes to get it to work. The big one was to get the GPIO registers set properly to allow the PPI to be active. Its a good starting point but should not be considered "shrink wrap".






2006-05-25 19:57:10     RE: ppifcd_test on BF537-STAMP

Michael Hennerich (GERMANY)

Message: 7968    BF537 has a build in TWI peripheral - so you don't need to use the generic I2C support.

This might haven't been included in the last release.


Either wait for the next release - will be available somewhere next week - or get a checkout form our CVS HEAD.






2006-05-25 20:00:50     RE: ppifcd_test on BF537-STAMP

Michael Hennerich (GERMANY)

Message: 7970    I checked BF537 Support into the driver recently.


Either wait for the next release - will be available somewhere next week - or get a checkout form our CVS HEAD.





-* $Id: bfin_ppifcd.c,v 1.2 2006/03/31 12:28:12 hennerich Exp $


+* $Id: bfin_ppifcd.c,v 2006/05/23 09:04:37 hennerich Exp $



* DATE OF CREATION: Sept. 10th 2004



@@ -249,6 +249,7 @@


printk("ppifcd_error_irq: \n");

printk("PPI Status = 0x%X \n", *pPPI_STATUS);





/* Acknowledge DMA Interrupt*/




@@ -257,6 +258,9 @@


*pPPI_CONTROL = pdev->ppi_control & ~PORT_EN;


pdev->done = 1;



+ /* Clear Sticky Bits */




/* Give a signal to user program. */



@@ -269,6 +273,7 @@



DPRINTK("ppifcd_error_irq: return \n");









@@ -580,7 +585,11 @@



request_irq(IRQ_PPI_ERROR,(void *) ppifcd_irq_error,SA_INTERRUPT,"PPI ERROR" ,filp->private_data);





+#if (defined(CONFIG_BF537) || defined(CONFIG_BF534) || defined(CONFIG_BF536))

+ *pPORTG_FER = 0xFFFF; /* PPI[15:0] */

+ *pPORTF_FER |= 0x8300; /* PF.15 PPI_CLK FS1 FS2*/

+ *pPORT_MUX &= ~0x0E00;





DPRINTK("ppi_open: return \n");





2006-05-25 20:05:36     RE: ppifcd_test on BF537-STAMP

Skye Sweeney (UNITED STATES)

Message: 7971    Ahhhh!

The magic FER and MUX register fix I put in myself! Its hard to get the PPI to work when it is not an active peripheral!


Sitting behind a firewall I can't directly see the CVS tree. I guess I may be forced to keep tabs on it from home.






2006-05-26 21:28:07     RE: ppifcd_test on BF537-STAMP

Michael Benjamin (UNITED STATES)

Message: 8014    When I run ppifcd_test on the BF537-STAMP, it outputs the text of "Caculated Times Based on the actual Camera setting" with the Master Clock and times, but it doesn't do anything after that. My command line is "ppifcd_test -c 1 test.BMP".


When I run fcd I get the web page, but never get an image after I specify my GlobalGain and click 'Capture.' The fcd program does not crash, but does report: "PPI wait_event_interruptible ierr".


I tested on a BF533-STAMP and ppifcd_test gives me the same message but fcd works fine.


Also, I've posted a page about the PPIFCD at:




but it still needs work.








2006-05-26 22:42:46     RE: ppifcd_test on BF537-STAMP

Michael Hennerich (GERMANY)

Message: 8018    I don't know your hardware setup. But the Micron Sensor module relies on that someone puts it out of standby - the pflags driver for BF537 is currently broken.


I will fix it early week and modify ppi_fcd and ppifcd_test accordingly - and run tests on BF537.


But to keep you going - PPI wait_event_interruptible ierr means that there was an timeout before you received any or all pixels - I would say any.






2006-05-26 23:02:10     RE: ppifcd_test on BF537-STAMP

Michael Benjamin (UNITED STATES)

Message: 8019    I have a Micron sensor module and the LEDs turn on when I run ppi_fcd and flash once when I capture with fcd on the BF533. On the BF537, they never turn on - so I bet you're right: it's just on standby and timing out.


Thanks so much for your help!




2006-06-19 23:08:02     RE: ppifcd_test on BF537-STAMP

Brian Schmidt (UNITED STATES)

Message: 9121   



i am trying to run ppifcd_test on the BF537-Stamp as well.


I have the latest CVS and everything. My build will never even open. When i run the ppifcd_test i get this:


Could not open dev/ppi : 2


i checked, and the device isnt there. Why isnt the ppi showing up for me? i did everything it said to do in the readme.txt file in the ppifcd-test directory.


any one know?




2006-06-19 23:41:46     RE: ppifcd_test on BF537-STAMP


Message: 9122    you have to add the following line in <uClinux-dist>/vendors/AnalogDevices/BF537-STAMP/device_table.txt



/dev/ppi c 664 0 0 241 0 0 1 1





2008-08-26 12:17:44     Re: ppifcd_test on BF537-STAMP


Message: 61114   


Well, I'm replying to a 2-year old post with question...  I'm working with a BF537 EZKit and a custom card that puts a Micron Sensor on the PPI.  As a result, I wanna play with PPIFCD (or is there something more applicable?).  I see several folks have some success with the PPIFCD test; however, I'm running with uClinux-dist-2008R1.5-RC3  and when I run "make menuconfig" there is no "Blackfin BF53x Programmable Flags Driver" option.  So, I ignore that and build without it.  It builds, but doesn't run 'cause it can't find /dev/pf 27.  I must be doing something fundamentally wrong.  Has anyone built ppifcd-test on a recent (perhaps even uClinux-dist-2008R1.5-RC3 ) distribution?




2008-08-26 13:23:46     Re: ppifcd_test on BF537-STAMP

Mike Frysinger (UNITED STATES)

Message: 61117   


that driver no longer exists and so anything that attempts to use it, will fail


the programmable flags driver has been replaced with the simple gpio driver


you could use the v4l driver that now exists for the micron senseor instead ...




2008-08-26 14:38:17     Re: ppifcd_test on BF537-STAMP


Message: 61120   


Ok, that explains that.  Just for the sake of anyone else that may trek this path, I enabled the simple-gpio-test stuff and it was crashing, but it appears that the crash is caused by some kind of conflict between it and the fact that I had not turned off the Blackfin BF53x PPI Camera frame capture driver (after rebuilding with FCD disabled, the simple gpio stuff worked fine).  Maybe it would be best if the Wiki page "PPI Frame Capture Device" be either removed or at least noted that it no longer works.  Maybe that driver should be removed from the build entirely?


Thanks for your help.




2008-08-27 01:52:20     Re: ppifcd_test on BF537-STAMP

Andras Szemzo (HUNGARY)

Message: 61145   






The ppi-fcd kernel driver and the example working with the latest svn too, after some slight modification on the application. I'm not using the gpio-s for the micron MT9T001, instead triggering with i2c register.


I'm using custom BF532 with a custom Micron sensor board.








2008-09-14 18:50:15     Re: ppifcd_test on BF537-STAMP


Message: 62140   


"Me too" - the ppifcd driver works just fine for me in the 2008R1.5 release.   It's a nice simple example and I don't think it should be removed.