Quantcast
Channel: Processors forum - Recent Threads
Viewing all 17527 articles
Browse latest View live

RTOS/OMAP-L138: Does IPC 3.x Support OMAPL138 ?

$
0
0

Part Number: OMAP-L138

Tool/software: TI-RTOS

Just wish to check if OMAPL138 work on IPC3.x with ARM running on Linux.

Thanks.


CCS/AM5706: Problems with loading DSP part via Linux

$
0
0

Part Number: AM5706

Tool/software: Code Composer Studio

Hello,

I am trying to make a simple test programm for a dsp core in AM5706. I have a Linux created by arago-project on a target machine.

1) I went to CCS 9 and created a Sysbios typical project for C66XX. I choosed simVayu like a platform, because I had had only simVayu or simKepler and I have no idea what is difference.

2) I built the project successfully, renamed it to dra7-dsp1-fw.xe66 and put it to /lib/firmware on the target machine. After that I tried do an unbind/bind sequence but got the next fail - 

root@am5708-p38:/sys/bus/platform/drivers/omap-rproc# echo "40800000.dsp" > unbind
[ 4354.291572] remoteproc remoteproc2: releasing 40800000.dsp
root@am5708-p38:/sys/bus/platform/drivers/omap-rproc# echo "40800000.dsp" > bind
[ 4356.381686] omap-rproc 40800000.dsp: assigned reserved memory node dsp1-memory@99000000
[ 4356.390879] remoteproc remoteproc2: 40800000.dsp is available
[ 4356.406962] remoteproc remoteproc2: powering up 40800000.dsp
[ 4356.414815] remoteproc remoteproc2: Booting fw image dra7-dsp1-fw.xe66, size 1842104
[ 4356.435141] omap_hwmod: mmu0_dsp1: _wait_target_disable failed
[ 4356.441035] omap-iommu 40d01000.mmu: 40d01000.mmu: version 3.0
[ 4356.446966] omap-iommu 40d02000.mmu: 40d02000.mmu: version 3.0
[ 4356.456054] remoteproc remoteproc2: Failed to find resource table
[ 4356.468414] omap_hwmod: mmu0_dsp1: _wait_target_disable failed

3) I added to app.cfg next strings

var Ipc = xdc.useModule('ti.sdo.ipc.Ipc');
var Resource = xdc.useModule('ti.ipc.remoteproc.Resource');
Resource.customTable = false;

But I couldn't build the project because next errors.

"C:/ti/ipc_3_50_03_04/packages/ti/ipc/remoteproc/rsc_table_vayu_dsp.h", line 244: error: identifier "ti_trace_SysMin_Module_State_0_outbuf__A" is undefined
"C:/ti/ipc_3_50_03_04/packages/ti/ipc/remoteproc/rsc_table_vayu_dsp.h", line 249: error: identifier "PHYS_MEM_IPC_VRING" is undefined

4) I decided to make my own platform via RTSC platform manager. I choosed device family - c6000 and device name - dra7xx, set clock speed and memory. Platform was created sucessfully but I coulsn't choose it, I still had only simVayu and simKepler in a combobox. 

So, I don't know what I shoud try next and ask your advice.

Thank you

Linux/66AK2E05: Message queue IPC mechanism requires high cpu load

$
0
0

Part Number: 66AK2E05

Tool/software: Linux

Hi ,

I try to use TI's IPC message queue for Keystone II communication betwenn arm a15 cores and dsp core.

Typically arm cores operate Linux kernel and RTOS is on DSP core.

Arm cores use typical epoll loop to receive various tcp/udp packets, process them and then try to feed some of those to dsp core . 

Code is something like (after having initialized the queue and IPC properly based on TI's examples) :

ARM A15 ==> DSP:

MessageQ_put();

a pthread is spawned  on ARM side to wait response packets from DSP, thread's function looks like :

while (some_runnining_cond_which_is_always_true)
    {
        /* check for return message */
        status = MessageQ_get(&comm_module, (DSP_Comm_Msg_header_t**)&msgrx,0);

        switch (status)
        {
        case MessageQ_S_SUCCESS: //Message successfully returned
            /* extract message payload */
            status = process_rx_message(&msgrx->voice_header);
#ifndef STATIC_MESSAGEQ
            /* free the message */
            dsp_comm_message_free(msgrx);
#endif
            if (status != 0)
                status = -1;

            break;

        case MessageQ_E_TIMEOUT: //MessageQ_get() timed out
            status = 0;
        break;
        default:    // Some other error
            MC_LOG_ERRO("Error while getting message from DSP");
            status =  -1;
        break;
        }
    }

CPU load is close to 20% (top command output) !

timeout is 0 (nonblocking) so i would like ideally to increase timeout to some maximum value to decrease

cpu load provoked by "while" running continuously.

Which is the max timeout value I can use ?

I could not find it in

http://software-dl.ti.com/dsps/dsps_public_sw/sdo_sb/targetcontent/syslink/latest/docs/html/_message_q_8h.html

Is it possible that message queue provides a file descriptor which I could integrate into existing epoll loop?

Thank you!

RTOS/TMDX654IDKEVM: UniFlash Issue

$
0
0

Part Number: TMDX654IDKEVM

Tool/software: TI-RTOS

Hello, I'm trying to flash via Uniflash with the following steps:

    • Switch to UART boot : SW3 = 0101_0000_00;
    • Power On the EVM ;
    • Determine the UART instance that corresponds to the EVM:
    • In Terminal, use command: ls -al /dev/ttyUSB*         ;
    • If /dev/ttyUSB0 is the first instance, then you have to use /dev/ttyUSB1 in further commands;
    • /dev/ttyUSB1 in UART mode will print "C" characters in the minicom console. Verify that;
    • Exit all minicom sessions that are connected to the UARTS;
    • Run the following commands from command line using Uniflash 5.0:
    • ./dslite.sh --mode processors -c /dev/ttyUSB1 -f /(path)/uart_am65xx_evm_flash_programmer.tiimage
    • ./dslite.sh --mode processors -c /dev/ttyUSB1 -f /(path)/tiboot3.bin -d 3 -o 0
    • ./dslite.sh --mode processors -c /dev/ttyUSB1 -f /(path)/sysfw.bin -d 3 -o 40000
    • ./dslite.sh --mode processors -c /dev/ttyUSB1 -f /(path)/app -d 3 -o A0000
  • Power off the board and switch to OSPI boot mode to verify . SW3 = 1000_0001_00

The application works fine on the sd-card , but when flashed with Uniflash nothing boots up, no coom. on UART  , maybe the SW2 switch should also be set ?

Thanks.


Linux/TDA2E: Can the kernel software support the power management of the TF card?

$
0
0

Part Number: TDA2E

Tool/software: Linux

HI,

How does the kernel software do the power management of the TF card:

After the TF card is removed, we need to turn off the TF card power.

When the TF card is reinserted, the kernel will not recognize the TF card, and then the TF card will be powered. The kernel still cannot recognize the TF card.

AM5728: .Bin without rprc

$
0
0

Part Number: AM5728

Hi,

The customer generated .bin with CCSV7 which is not available in uboot and the .bin can be used successfully in SBL. After comparison, he found the SBL has function that deal with rprc file, but he didn't deal with it in his uboot. 

He tried this instruction to convert .out to .bin

${TI_PDK_INSTALL_DIR}/packages/pdkAppImageCreate.bat ${TI_PDK_INSTALL_DIR}/packages ${CG_TOOL_ROOT} ${PROJECT_LOC}/${ConfigName} ${ProjName} AM572x dsp 

So how to generate .bin file without  rprc file?

DRA80XMEVM: starting cores from bare metal code

$
0
0

Part Number: DRA80XMEVM

Hi,

I am working on bare metal code for A53 on DRA80xM IDK board. I am connecting to core-0 through debugger. 

Is there any sample code available to start core-1 from cluster-0 and respective cores from cluster-1?

Regards,

Devendra 

DM385: CSI2_COMPLEXIO_IRQSTATUS meaning

$
0
0

Part Number: DM385

Hello!

I'm trying to start CSI2 interface on DM385 processor.

In connection with that I have the question:

 Does zero in CSI2_COMPLEXIO_IRQSTATUS register field  ERRSOTHSx mean that Start of Transition event had never happened

on proper line? (Provided CSI2 is powered and enabled)

Just in case, related register values

 [m3vpss ] CSI2_CTRL    [29a869]
 [m3vpss ] CSI2_COMPLEXIO_IRQENAB [7ffffff]
 [m3vpss ] CSI2_COMPLEXIO_CFG   [6b000123]
 [m3vpss ] CSI2_COMPLEXIO_IRQSTAT [0]


Linux/AM3358: mtd_debug erase Work Abnormality

$
0
0

Part Number: AM3358

Tool/software: Linux

The following error occurred when using mtd_debug on EVMSK

root@am335x-evm:~# mtd_debug erase /dev/sda 0x1D000 0x1000
MEMERASE: Invalid argument

SDK version :05.03.00.07

OMAP-L138: /sys/class/pwm is empty

$
0
0

Part Number: OMAP-L138

Hi all,

I've change menuconfig to enable EHRPWM, the source codes is being compiled as well, but when I load into OMAPL138 evk, directory /sys/class/pwm is empty

What else do I need to change?

FYI: using linux 4.6

TMS320C6747: Is RVDD regarded as CVDD ?

$
0
0

Part Number: TMS320C6747

Hi,

As you see in the original post, my customer has to move over from DA810 to C6747 on their custom board. They noticed that
1. H6 and H12 pins are RVDD pins on C6748. It seems these pins are Internal ram supply voltage pins
2. These pins are regarded as CVDD on DA810

My customer would like to know if they have to care about these difference. I believe my customer doesn`t have to so and they can regard these pins as other CVDD pins including power up sequence. Is my understanding correct ? Just a confirmation.

Best Regards,
NK 

TDA3XEVM: Adding custom board to Vision SDK project

$
0
0

Part Number: TDA3XEVM

Hi,

I would like to know how to add support for our custom board into both a Vision SDK- and PDK-based project.

Please note that we would like to place these files outside of the SDK installed path (like previously discussed)

Regards

Claes

PS: I will be off for a couple of weeks which means it is not a hurry right now, but will be when I am back in beginning/middle of August.

AM3356: take 1us to write register twice for setting 32 GPIO high or low

$
0
0

Part Number: AM3356

Dear Sir:


I am sorry to trouble you. Can you give me a hand?


My customer complains that it will take 1us to write register twice for setting 32 GPIO high or low just including two steps of writing register.


Is it normal for AM3356? Would you please explain it?


Thank you very much.

RTOS/AM3358: Problem with generating PDK_am335x examples

$
0
0

Part Number: AM3358

Tool/software: TI-RTOS

Hi,

i downloaded the pdk_am335x v1.0.15 with processor SDK software-dl.ti.com/.../index_FDS.html

When i run pdkProjectCreate.bat, I get:

C:\ti\pdk_am335x_1_0_15\packages>pdkProjectCreate.bat
=========================================================================
Configuration:
   SOC             :   AM335x
   BOARD           :   all
   ENDIAN          :   little
   MODULE          :   all
   PROJECT_TYPE    :   all
   PROCESSOR       :   arm
   PDK_SHORT_NAME  :   C:\ti\PDK_AM~1\packages\
=========================================================================
Checking Configuration...
Complete
=========================================================================
   PDK_PARTNO         : AM335
   PDK_ECLIPSE_ID     : com.ti.pdk.am335x
   RTSC_PLATFORM_NAME : ti.platforms.evmAM3359
   RTSC_TARGET        : gnu.targets.arm.A8F
   CCS_DEVICE         : "Cortex A.AM3359.ICE_AM3359"
*****************************************************************************
Detecting all projects in PDK and importing them in the workspace C:\ti\PDK_AM~1\packages\\MyExampleProjects
Detected Test Project: EMAC_BasicExample_evmAM335x_armExampleproject
The system cannot find the path specified.
Copying macro.ini
The system cannot find the path specified.
        0 file(s) copied.
Detected Test Project: EMAC_BasicExample_skAM335x_armExampleproject
The system cannot find the path specified.
Copying macro.ini
The system cannot find the path specified.
        0 file(s) copied.

Do I have to configure the PDK in other way than running pdksetupenv.bat? Or maybe install newer version?

I want to compile NDK ethernet examples for sitara, I am using BeagleBone Black board

TMS320C5517: audio processing algorithm support

$
0
0

Part Number: TMS320C5517

Hi team,

Our customer wants to use TMS320C5517 in their system, and they come to us with 2 question.

  1. Do we get audio processing algorithm support for them? And what debug tools could we offer to them?
  2. Does TMS320C5517 support 2 SD card with each range up to 512M?

Could you give some comments to this?

Thanks,

Will


コンパイラ/ TMS320C6747:The following build errors occur in CCSv9, please tell me the workaround.

$
0
0

Part Number: TMS320C6747

Tool/software: TI C / C ++コンパイラ

Hello
The issue of boot loading that I got advice from everyone found and solved a flash rewriting program developed nine years ago. Transfer this application to C6747 using UARTHOST.EXE, start it, send HEX in Hyper Terminal and the application will write. I apologize to you for a fuss over.

It is another question. I will tell you the condition again.
Developed nine years ago, there is another company's "C6747 + spartan3e + usb" circuit board.
There was a record that CCS at the time of development was IDE V4. There are also source files. This time CCS uses v9. xds200 usb is not yet purchased.
The installation location is C: \ ti \ ccs910
Workspace, project
C: \ Users \ alex \ workspace_v9
C: \ Users \ alex \ workspace_v9 \ LS041
In addition, I installed the following BIOS.
   DSPBIOS5_41_03_17
   C6747_BIOSPSP_01_30_01
   EDMA3_LLD_BIOS5_01_11_00_03
And
C: \ Program Files \ Texas Instruments \ ccsv4 \ emulation \
boards \ evmc6747 \ gel \ evmc6747_dsp.gel
Switched to a new one. The difference is
(1)Setup_System_Config, * _PLL, * _Psc_All_On, * _EMIFA / B
comment out in OnTargetConnect.
(2)100 MHz in Setup_PLL
(3)PLL0_PLLDIV3 / 5/7, KICK0 / 1R, CFGCHIPS3 Comment out.
The above settings were described in the found development document.


So the question is
The following build errors occur in CCSv9, please tell me the workaround.
">> Compilation failure
subdir_rules.mk:9: recipe for target 'syscfg.obj' failed
"../syscfg.c", line 3: fatal error # 1965: can not open source file "ti / pspiom / cslr / soc_C6747.h"
1 catastrophic error detected in the compilation of "../syscfg.c".
Compilation terminated.
gmake: *** [syscfg.obj] Error 1
gmake: Target 'all' not remade because of errors. "

The third line of syscfg.c is
#include "ti / pspiom / cslr / cslr_syscfg_C6747.h"

There are eight other errors of the same type.
Hst.h, iom.h, log.h, que.h can not be opened.
I searched for everything but I can not find it.

TI's compiler understands that only ascii can be used, and because Japanese is included in the comment section of each place of the source file with about 30 files, fatal errors may be caused by this Japanese. However, it should have passed 9 years ago, so there should be a way to avoid NG.
Is the setting I ran short? Is it possible for me to switch compilers?

It is very complicated, but I hope it helps.

Linux/AM3352: Issue with system start (Am335x based platform)

$
0
0

Part Number: AM3352

Tool/software: Linux

Hi,

I needed a little help in debugging an issue relating to system start. Ours is an AM335x based platform, currently running Linux kernel is 4.9.31 (Arago). As you know, the sequence followed for booting the system is :

a) ROM code

b) MLO

c) u-boot 

d) Kernel

After the system is powered on and the uboot has loaded successfully, I see the bootloader print "Starting Linux kernel" on the serial console.. there are times where-in 2 - 4 secs after this (looks like kernel couldn't load), I again see the MLO loading and then the u-boot.img and then the kernel loading via serial logs. This time the kernel loads successfully. How do I go about debugging this issue? Why does the kernel not load sometimes and instead after about 2 - 4  secs, MLO loads again. Any help on why this happen or how to debug would be greatly appreciated.

Regards,

Fariya

AM5728: I have a trouble in accessing some memory-mapped registers.

$
0
0

Part Number: AM5728

Hi,

I'm now using the board:
TMDXIDK5728:
AM572x Industrial Development Kit (IDK)

But I have a problem in accessing some memory-mapped registers.
In more specific, I cannot access L3_INSTR region (0x5400_0000 ~ 0x547F_FFFF). For example, when I read the value from MPU_C0_CS_PTM_MPU (0x5414_C000 ~ 0x5414_CFFF), all the read values were 0x0 even after writing some non-zero value. According to ARM Cortex-A15 TRM document, 0x5414C004 should not be zero, because [31] and [27] must be non-zero at least. But 0x0 is read from that address.

Is there any restriction to access L3_INSTR register region?

I used the branch "processor-sdk-linux-4.19.y" from github and ti_sdk_dra7x_debug_defconfig as the configuration.
I think, you can simply regenerate this issue.

Thanks,
Hyun

Linux/AM4378: 5 wire resistive Touch Screen Connection Check

$
0
0

Part Number: AM4378

Tool/software: Linux

Dear TI Team,

We are evaluating 5-wire resistive touch screen in our custom board. Generally there are 4-wire reference designs, so that the SENSE(5th pin) pin is not included in them

Can you pls. help to check whether our connections are proper or not?

Regards,

Zafer Çalışkan

Linux: TDA2 PROCESSOR_SDK_VISION_03_07_00_00 new-usercase draw Ti logo

$
0
0

Tool/software: Linux

Hi

I builded a new usecase "my_usecase", and I wanted to draw Ti logo first

a. MAKECONFIG?=tda2xx_evm_linux_all

b. my_usecase.txt

UseCase: my_usecase

GrpxSrc -> Display_Grpx

c. grpxSrcLink_my_usecase.c

Int32 GrpxSrcLink_my_usecase_draw(GrpxSrcLink_Obj *pObj)
{
Draw2D_RegionPrm region;
Draw2D_BmpPrm bmpPrm;

/* fill full buffer with background color */
region.color = DISPARITY_BACKGROUND_COLOR;
region.colorFormat = SYSTEM_DF_BGR16_565;
region.startX = 0;
region.startY = 0;
region.height = pObj->info.queInfo[0].chInfo[0].height;
region.width = pObj->info.queInfo[0].chInfo[0].width;

Draw2D_fillRegion(pObj->draw2DHndl,&region);


/* draw bitmap's */

/* TI logo */
bmpPrm.bmpIdx = DRAW2D_BMP_IDX_TI_LOGO_1;
Draw2D_drawBmp(pObj->draw2DHndl,
1600 +80,
DISPARITY_FRAME_THICKNESS,
&bmpPrm
);

return SYSTEM_LINK_STATUS_SOK;
}

I had build and run apps.out,

I only got "Hello" and "test" from Vps_printf of my_usecase.c in usecases.

But nothing output of logo or icon .

It is still show 1920*1080 and FPS from HDMI on my output device

Do any link or display function not set for my usecase to draw logo ?

Viewing all 17527 articles
Browse latest View live


<script src="https://jsc.adskeeper.com/r/s/rssing.com.1596347.js" async> </script>