Linux

Forum Post: RE: Syslink error using M3 firmware while streaming live video and dvr

Gunter Schmer — Fri, 17 May 2013 16:50:00 GMT

Gil,

ok we will put at the very beginning of omxrpc_module_init_client() and retest.

We will post traces for BEFORE (without Gate properly located) and AFTER (with Gate_enterModule at very beginning and Gate_leaveModule at very end). Then we should clearly see the difference in the trace as well.

Regards,

--Gunter

Forum Post: RE: Proper Way to Benchmark Timing within OpenMP on C6678

Ryan Radjabi — Fri, 17 May 2013 16:04:00 GMT

This is not a Linux related question. Please move it to the appropriate forum. This application uses SYSBIOS.

Forum Post: Linux kernel compilation for camera

Sekra Invos — Fri, 17 May 2013 13:54:00 GMT

Hai

When i compile the board-omap3beagle-camera.c,i get the following errors in mux file.

arch/arm/mach-omap2/mux.h:96: error: redefinition of 'struct omap_board_data'
arch/arm/mach-omap2/mux.h:113: error: redefinition of 'struct omap_mux_partition'
arch/arm/mach-omap2/mux.h:131: error: redefinition of 'struct omap_mux'

.................................So i go through the mux.h file in location mach-omap2 directory,in that file there are two header files were added for mux.

1)#include <../arch/arm/mach-omap2/mux.h>, 2)#include "mux.h" as below shown in camerea.c file

#include <../arch/arm/mach-omap2/mux.h>

/* Include V4L2 ISP-Camera driver related header file */

#include <../drivers/media/video/omap34xxcam.h>
#include <../drivers/media/video/isp/ispreg.h>

#include "mux.h"

what is the reason of 2 files for mux configuration.Anybody can clear my doubts

Thanks in advance.

Forum Post: where is dsplink logs?

Kalaiyarasan ES — Fri, 17 May 2013 13:08:00 GMT

HI,

I am using omapl138 and TI DVSDK 4.3.0.6.

I have changed "CFG_OMAPL138SHMEM.c" in folder " ${DSPLINK}/config/all " as below

######################################################################################################

STATIC LINKCFG_Log  LINKCFG_logObject = {     TRUE,             /* GDMSGQPUT         : GPP->DSP MSG Transfer  - MSGQ_put call */     TRUE,             /* GDMSGQSENDINT     : GPP->DSP MSG Transfer  - GPP sends interrupt */     TRUE,             /* GDMSGQISR         : GPP->DSP MSG Transfer  - DSP receives interrupt */     TRUE,             /* GDMSGQQUE         : GPP->DSP MSG Transfer  - Message queued at DSP */     TRUE,             /* DGMSGQPUT         : DSP->GPP MSG Transfer  - MSGQ_put call */     TRUE,             /* DGMSGQSENDINT     : DSP->GPP MSG Transfer  - DSP sends interrupt */     TRUE,             /* DGMSGQISR         : DSP->GPP MSG Transfer  - GPP receives interrupt */     TRUE,             /* DGMSGQQUE         : DSP->GPP MSG Transfer  - Message queued at GPP */     FALSE,             /* GDCHNLISSUESTART  : GPP->DSP CHNL Transfer - Entering inside ISSUE call */     FALSE,             /* GDCHNLISSUEQUE    : GPP->DSP CHNL Transfer - ISSUE: Buffer is queued in internal structure on GPP */     FALSE,             /* GDCHNLISSUECOMPL  : GPP->DSP CHNL Transfer - ISSUE call completed */     FALSE,             /* GDCHNLXFERSTART   : GPP->DSP CHNL Transfer - Initiating a buffer transfer by GPP */     FALSE,             /* GDCHNLXFERPROCESS : GPP->DSP CHNL Transfer - Actual transfer of buffer is going to take place */     FALSE,             /* GDCHNLXFERCOMPL   : GPP->DSP CHNL Transfer - Buffer transfer is complete */     FALSE,             /* GDCHNLRECLSTART   : GPP->DSP CHNL Transfer - Entering RECLAIM call */     FALSE,             /* GDCHNLRECLPEND    : GPP->DSP CHNL Transfer - RECLAIM: Wait on a semaphore */     FALSE,             /* GDCHNLRECLPOST    : GPP->DSP CHNL Transfer - RECLAIM: Posting the Semaphore */     FALSE,             /* GDCHNLRECLCOMPL   : GPP->DSP CHNL Transfer - RECLAIM call completed */     FALSE,             /* DGCHNLISSUEQUE    : DSP->GPP CHNL Transfer - ISSUE: Buffer is queued in internal structure on DSP */     FALSE,             /* DGCHNLXFERSTART   : DSP->GPP CHNL Transfer - Initiating a buffer transfer by DSP */     FALSE,             /* DGCHNLXFERPROCESS : DSP->GPP CHNL Transfer - Actual transfer of buffer is going to take place */     FALSE,             /* DGCHNLXFERCOMPL   : DSP->GPP CHNL Transfer - Buffer transfer is complete */     FALSE,             /* DGCHNLRECLPEND    : DSP->GPP CHNL Transfer - RECLAIM: Wait on a semaphore */     FALSE,             /* DGCHNLRECLPOST    : DSP->GPP CHNL Transfer - RECLAIM: Posting the Semaphore */     10,                /* MSGIDRANGESTART   : MSG ID range: lower limit */     20                 /* MSGIDRANGEEND     : MSG ID range: upper limit */ } ;

###################################################################################################

what is this LINKCFG_logObject?

what happen when i change from FALSE to TRUE?

Where can i get information about it?

Forum Post: RE: the error generate No valid Qt Version found in the Qt Creator 2.4.1 in Ubuntu 10.04

Dave Gerlach — Fri, 17 May 2013 11:30:00 GMT

Hello,

Are you exporting the correct environment before running Qt Creator? An east way to do this is to source the environment setup script (source linux-devkit/environment-setup) and then running Qt Creator from the same command line.

Regards,

Dave

Forum Post: Decoder stop working when i create encoder instance

Anshul Maheshwari — Fri, 17 May 2013 11:20:00 GMT

i was trying to run encoder_create below decoder inside sv04 example

and i found that if i do alg_create than previously running decoder is also not running

I tried same, sometime before and i provided scratch memory in different section in encoder create and decoder create.

i have attached siuVctMemMgr.c, where alg_alloc function is not re entrant.

(Please visit the site to view this file)

i think _ALG_allocMemory have one pointer that is allocated again pInternalScratch Memory is initialized again,

if it has been initialized only once then if there is memory left in scratchMemory section then it work fine if not then it give error

not courrupt memory.

i dont if encoder and decoder can use same memory, i have an observation that 2 instance of encoder work on same memory but encoder and decoder dont work simunaneously

thanks

anshul

Forum Post: RE: Configuring and Building Qt 4.7.0 and QWS for cross-compile on TMDXEVM3730

yase sabar — Fri, 17 May 2013 09:09:00 GMT

Forum Post: Can I connet HDMI PHY(SII9022A audio part) to mca0?

bingyi cheng — Fri, 17 May 2013 08:36:00 GMT

dm8168,dvr3.50,Can I connet HDMI PHY(SII9022A audio part) to mca0? The default situation is that Sii9022A is connected to mca1。

Forum Post: RE: GUI touchscreen not working on am335x

jacky xiao — Fri, 17 May 2013 08:21:00 GMT

I solved my issue, There is one hardware probelm.

thanks!

Forum Post: The way to generate shared dynamic library apps ( examples/ti/sdo/ce/examples/apps//)

jason Lee75734 — Fri, 17 May 2013 07:43:00 GMT

I'd like to get a share dynamic library, such as libvideo1_copy.so, then I can use the .so in my own apps on arm platform.

but the CE/examples/ti/sdo/ce/examples/apps/*/package.bld, it only generate app_local.xv5T

I added Pkg.addLibrary(name, targ).addObjects(srcs); to package.bld ( delete the Pkg.addExecutable(...)), then I got the app_local.av5T

I linked to my apps using arm toolchains, but there are many error when compiling the app( many errors about undefined symbols)

I have the follow questions:

1. How to modify the packeage.bld for generate a completed library(for example, libvideo1_copy.so), so that I use it in my apps on ARM

Forum Post: significance of gateMP in sv04

Anshul Maheshwari — Fri, 17 May 2013 05:45:00 GMT

GateMp module of ipc is used inside barriers,

since gateMp is generic infrastructure to use gate.

i am unable to find why those barriers, in some comment i found like it is just for syncronisation

of codec (master core of codec and slave codec)

so do i need to create multiple barrier if using multiple codec, (encoder and decoder)

do i have to pass the number of cores while creating is total (encoder cores + decoder cores )

i am looking in transcoder (ccs based) code,

if any one has already achived this, please share how to place barrier.

thanks

anshul

Forum Post: which DMA channel is used by AM335X GPMC?

Enders Hu — Fri, 17 May 2013 02:01:00 GMT

hi:

I want to confirm which DMA channel the AM335x GPMC is using under the Linux 3.2.0 PSP.

in file "...plat/dma.h", we know that channel 4 is for the GPMC for OMAP2

#define OMAP24XX_DMA_GPMC 4 /* S_DMA_3 */

in file "plat/dma-33xx.h", there is a macro:

#define AM33XX_DMA_GPM 52

is that channel 52 is used for the GPMC for AM335X?

Forum Post: problem about gpmc driver on dm8168

Xavier15128 — Thu, 16 May 2013 23:52:00 GMT

Hi all
I develop on dm8168 evm using dvrrdk3.50
I have a fpga connect to gpmc cs2, and i want to write a linux device driver for it .
I have test my gpmc configure use mmap funtion to init, it works well,
but when i insmod my driver, the function gpmc_cs_request() broken,
it rets null handle used in the kernel, so i find gpmc_cs_insert_mem() uncalled for CS2,fpga spce,
so i add cs2 configure register ,then it works, however i found the when i call gpmc_cs_request(),
the cs7 register is rewrite not what i want it, and when i rmmod gpmc_fpga.ko ,
it can not insmod anymore sees some resource are not released in kernel.

can anyone tell me is there something i have to change in the kernel

best regards
xavier

void __init gpmc_init(void)
{
   u32 l, irq;
   ....

   gpmc_write_reg(GPMC_SYSCONFIG, l);

   gpmc_cs_write_reg(2, GPMC_CS_CONFIG1, gpmc_fpga[0]);
   gpmc_cs_write_reg(2, GPMC_CS_CONFIG2, gpmc_fpga[1]);
   gpmc_cs_write_reg(2, GPMC_CS_CONFIG3, gpmc_fpga[2]);
   gpmc_cs_write_reg(2, GPMC_CS_CONFIG4, gpmc_fpga[3]);
   gpmc_cs_write_reg(2, GPMC_CS_CONFIG5, gpmc_fpga[4]);
   gpmc_cs_write_reg(2, GPMC_CS_CONFIG6, gpmc_fpga[5]);
   gpmc_cs_write_reg(2, GPMC_CS_CONFIG7, gpmc_fpga[7]);

   gpmc_mem_init();

   ....
}

following is my code:
static int __init fpga_init_cdev(void)
{
   int error;
   u32 val;

   fpga_dev.devt = MKDEV(0, 0);

   error = alloc_chrdev_region(&fpga_dev.devt, 0, 1, "fpga");
   if (error) {
       printk(KERN_ALERT "alloc_chrdev_region() failed: %d\n", error);
       return error;
   }

   cdev_init(&fpga_dev.cdev, &fpga_fops);
   fpga_dev.cdev.owner = THIS_MODULE;

   error = cdev_add(&fpga_dev.cdev, fpga_dev.devt, 1);
   if (error) {
       printk(KERN_ALERT "cdev_add() failed: %d\n", error);
       unregister_chrdev_region(fpga_dev.devt, 1);
       return error;
   }

   printk("Getting Chip Select\n");
   u32 l;
   int err;
   l = DM8168_GPMC_BASE;
   gpmc_base = ioremap(l, SZ_16M);
   printk("gpmc_base, address = %lx\n", gpmc_base);

   // val = 0xf64;
   // gpmc_cs_write_reg(GPMC_CS, GPMC_CS_CONFIG7, val);
   //val = gpmc_read_reg(GPMC_REVISION);
   //printk("GPMC revision %d.%d\n", (val >> 4) & 0x0f, val & 0x0f);
   //gpmc_write_reg(GPMC_IRQENABLE, 0);
   //gpmc_write_reg(GPMC_TIMEOUT_CONTROL, 0);

   gpmc_write_cs_reg(GPMC_CS, GPMC_CS_CONFIG1, gpmc_fpga[0]);
   gpmc_write_cs_reg(GPMC_CS, GPMC_CS_CONFIG2, gpmc_fpga[1]);
   gpmc_write_cs_reg(GPMC_CS, GPMC_CS_CONFIG3, gpmc_fpga[2]);
   gpmc_write_cs_reg(GPMC_CS, GPMC_CS_CONFIG4, gpmc_fpga[3]);
   gpmc_write_cs_reg(GPMC_CS, GPMC_CS_CONFIG5, gpmc_fpga[4]);
   gpmc_write_cs_reg(GPMC_CS, GPMC_CS_CONFIG6, gpmc_fpga[5]);
   gpmc_write_cs_reg(GPMC_CS, GPMC_CS_CONFIG7, gpmc_fpga[6]);

   val = gpmc_read_cs_reg(GPMC_CS, GPMC_CS_CONFIG7);
   printk("GPMC_CS_CONFIG7 value 0x%x\n", val);

   if (gpmc_cs_request(GPMC_CS, SZ_16M, (unsigned long *)&mem_base) < 0) {
       printk(KERN_ERR "Failed request for GPMC mem for usrp_e\n");
       return -1;
   }
   printk("Got CS2, address = %lx\n", mem_base);

   if (!request_mem_region(mem_base, SZ_16M, "mem_fpga")) {
       printk(KERN_ERR "Request_mem_region failed.\n");
       gpmc_cs_free(GPMC_CS);
       return -1;
   }

   val = gpmc_read_cs_reg(GPMC_CS, GPMC_CS_CONFIG7);
   printk("GPMC_CS_CONFIG7 value 0x%x\n", val);

   fpga_base = ioremap(mem_base, SZ_16M);
   printk("fpga_base, address = %lx\n", fpga_base);

   fpga_ver_read();

   val = gpmc_read_cs_reg(GPMC_CS, GPMC_CS_CONFIG7);
   printk("GPMC_CS_CONFIG7 value 0x%x\n", val);

   return 0;
}

Forum Post: RE: DM816x DDR3 SW leveling and EZSDK u-boot code updates

Tate Schaar1 — Thu, 16 May 2013 18:39:00 GMT

Renjith,

Would you mind sending me the patch you were referring to?

Thanks in advance,

Tate

Forum Post: RE: EZSDK 5.03.01.15 Missing gdbserver

tamo2 — Thu, 16 May 2013 18:38:00 GMT

Hi,

> follow this thread to install the library. http://e2e.ti.com/support/embedded/linux/f/354/t/47210.aspx

Has anybody really made it work following the instructions?
I can debug a single thread application but still having a problem debugging multi-threaded application, just as described in http://e2e.ti.com/support/embedded/linux/f/354/p/84357/927400.aspx#927400.

I have tried coping gdbserver and libraries from EZSDK/linux-devkit and also from CodeSourcery, but neither solved the problem.

I use ti-ezsdk_dm814x-evm_5_05_02_00 and DM8148.
If there is an older version of EZSDK which is known to work with gdbserver, I would like to try it.

Thanks in advance.

Forum Post: RE: URGENT : What is the correct approach to port or run OpenCV on DM3730

hamid ouanan — Thu, 16 May 2013 17:32:00 GMT

Sirs,

i use opencv-2.4.2

I follow the link below for the cross-compilation to opencv-2.4.2

http://processors.wiki.ti.com/index.php/Building_OpenCV_for_ARM_Cortex-A8

I want to execute facial Dectection implementation based on opencv ARM Cortex a8 in DM3730 but I dont know how to do?

i need a tutorial to follow to implement the opencv . And, an application based on this ( opencv) on ARM Cortex a8 in DM3730 .

thank you in advance

Hamid

Forum Post: RE: Booting Linux dm365 Network Camera in 3.2 seconds

Adam Porter1 — Thu, 16 May 2013 16:45:00 GMT

If you distribute binaries of modified Linux kernel sources, you are bound by the GPL to also distribute your modified sources.

Are you claiming that your modifications to decrease boot time are only based on modified non-GPL sources? Please clarify.

Thank you,

Adam

Forum Post: Correct settings for 1280x800 display on DM3730

Glenn Wainwright — Thu, 16 May 2013 15:20:00 GMT

I am trying to get a Sharp 1280x800 display working on a DM3730. Unfortunately, running any of the Graphics SDK demos (ex: OGLES2Skybox) gives me a repeated GFX_FIFO_UNDERFLOW error. I'm hoping that you can help me adjust my settings to be able to get this to work properly.

Bootargs:

setenv bootargs mem=80M@0x80000000 mem=128M@0x88000000 console=ttyS0,115200n8 noinitrd rw ip=${ipaddr}:${serverip}:${gatewayip}:${netmask}:${hostname}:eth0:on root=/dev/mtdblock4 rootfstype=yaffs2 psplash=false mpurate=1000 vram=40M omapfb.vram=0:20M omapfb.rotate=1 omapfb.vrfb=y omapfb.mirror=n omapfb.debug=y

FBDev settings:

mode "1280x800-61"
    # D: 72.005 MHz, H: 50.530 kHz, V: 61.322 Hz
    geometry 1280 800 1280 1600 32
    timings 13888 120 15 8 8 10 8
    rgba 8/16,8/8,8/0,8/24
endmode

Frame buffer device information:
    Name        : omapfb
    Address     : 0x70000000
    Size        : 13107200
    Type        : PACKED PIXELS
    Visual      : TRUECOLOR
    XPanStep    : 1
    YPanStep    : 1
    YWrapStep   : 0
    LineLength : 8192
    Accelerator : No

I have tried removing the rotation, and reducing the format to RGB565, but that does not seem to solve the error. What suggestions can you provide?

Thank you,
Glenn Wainwright

Forum Post: Can my program check the syslink version?

Markus Grunwald — Thu, 16 May 2013 13:57:00 GMT

Hello,

Is there a way for my software (that is using syslink) to check the version of the syslink module?

It seems that the software is very dependant on the correct syslink version. I would like to avoid problems with a wrong version by checking it in my software.

Of course, I could parse the dmesg output but I would prefer an API call of syslink.