• State TI Thinks Resolved
  • Locked Locked
  • Replies 12 replies
  • Answers 3 answers
  • Subscribers 98 subscribers
  • Views 2246 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

TDA2EVM5777: How to merge 2 input queue info?

yash trivedi
Prodigy 190 points

Part Number: TDA2EVM5777

Hi

we are using vision sdk3.0.0

my execution flow is

UseCase: chains_lvdsVipMultiCamViewMosaicWithLut

Capture -> Sync -> Alg_DmaSwMsWithLeftLut (A15) -> Dup

Dup -> Alg_CrModule (DSP1) -> Merge

Dup -> Merge

Merge -> Alg_FrameCopy (A15) -> Display_video

GrpxSrc -> Display_Grpx

I have made priv.c and priv.h using usecase genration tool.

my querys  are

1)apart from pric.c which other changes do i need to apply to merge 2 algorithm input ?

2)when i tried with this flow its giving 

[IPU1-0]     21.639053 s:  VIDEO_SENSOR: INST1 : I2C1 : I2C Addr = 0x38
[IPU1-0]     24.162172 s:  VIDEO_SENSOR: VIP 0: DRV ID 1205 (I2C ADDR 0x38): a635:0000:7fa2
[IPU1-0]     24.162660 s:  VIDEO_SENSOR: INST2 : I2C1 : I2C Addr = 0x39
[IPU1-0]     26.690264 s:  VIDEO_SENSOR: VIP 2: DRV ID 1205 (I2C ADDR 0x39): a635:0000:7fa2
[IPU1-0]     26.690752 s:  VIDEO_SENSOR: INST3 : I2C1 : I2C Addr = 0x3a
[IPU1-0]     29.230952 s:  VIDEO_SENSOR: VIP 4: DRV ID 1205 (I2C ADDR 0x3a): a635:0000:7fa2
[IPU1-0]     29.231440 s:  VIDEO_SENSOR: INST4 : I2C1 : I2C Addr = 0x3b
[IPU1-0]     31.758219 s:  VIDEO_SENSOR: VIP 8: DRV ID 1205 (I2C ADDR 0x3b): a635:0000:7fa2
[IPU1-0]     31.766424 s:  CAPTURE: Create in progress !!!
[IPU1-0]     31.766638 s:  CAPTURE: VIP1 Slice0 PortA capture mode is [ 8-bit] !!!
[IPU1-0]     31.886994 s:  CAPTURE: VIP1 Slice1 PortA capture mode is [ 8-bit] !!!
[IPU1-0]     32.007350 s:  CAPTURE: VIP2 Slice0 PortA capture mode is [ 8-bit] !!!
[IPU1-0]     32.127706 s:  CAPTURE: VIP3 Slice0 PortA capture mode is [ 8-bit] !!!
[IPU1-0]     32.248337 s:  CAPTURE: Create Done !!!
[IPU1-0]     32.249405 s:  IPC_OUT_0   : Create in progress !!!
[IPU1-0]     32.249740 s:  IPC_OUT_0   : Create Done !!!
[HOST  ]     32.249923 s:  IPC_IN_0   : Create in progress !!!
[HOST  ]     32.250259 s:  IPC_IN_0   : Create Done !!!
[HOST  ]     32.250350 s:  ALGORITHM: Create in progress (algId = 0) !!!
[HOST  ]     32.254193 s:  Assertion @ Line: 415 in algorithmLink_algPluginSupport.c: status == SYSTEM_LINK_STATUS_SOK : failed !!!

3) which paramerter should be same for merge?

  • 0
    TI__Mastermind 27743 points

    Hello Yash,

    1)apart from pric.c which other changes do i need to apply to merge 2 algorithm input ?

    You dont need to set any app level params for merge link. From the error log looks like use-case is failing due to incorrect/missing parameters for the dma mosaic alg link. This along with other links parameters needs to be set in use-case main file (chains_.c)

    Have you tried creating simple use-case first and adding components one- by one?

    This way it would be easier for you to pin point issue.

  • 0 in reply to Prasad Jondhale
    Prodigy 190 points

    yes i have tried this flow also

    UseCase: chains_lvdsVipMultiCam_Display

    Capture -> Dup -> VPE -> Sync -> Alg_DmaSwMs -> Merge

    Dup -> Alg_FrameCopy (A15) -> Merge

    Merge -> Display_video

    GrpxSrc -> Display_Grpx

    it give me same error    32.254193 s:  Assertion @ Line: 415 in algorithmLink_algPluginSupport.c: status == SYSTEM_LINK_STATUS_SOK : failed !!!

     i have attched file flow_files.tar.gz

  • 0 in reply to yash trivedi
    TI__Mastermind 27743 points

    Hello Yash,

    Thanks for sharing use-case code. From the use-case .txt file, i see only change you made in the existing lvdsvip multicam use-case is adding the frame copy algo right?

    When calling the use-case main function  (vision_sdk\apps\src\rtos\common\chains_main_bios_multi_cam.c), are you setting framecopy core id to intended one?

    gChains_usecaseCfg.algProcId = SYSTEM_PROC_DSP1; (done in vision_sdk\apps\src\rtos\common\chains_main_bios_single_cam.c before callign Chains_vipSingleCameraFrameCopy)

  • 0 in reply to Prasad Jondhale
    Prodigy 190 points
    Thank you for your quick response,
    yes i have just added framecopy algorithm and accordingly i have genrated the priv.c and priv.h file and added algprms in usecase file
    i havent done any change in chains_main_bios_single_cam.c error which i am getting error related this
    AlgorithmLink_putEmptyOutputBuffer(pObj, outputQId, pSystemBuffer);
  • 0 in reply to yash trivedi
    TI__Mastermind 27743 points

    Yash,

    Can you add below line in file(mostly chains_main_bios_multi_cam.c) from which you are calling your use-case main function. Add it before function call.

    gChains_usecaseCfg.algProcId = SYSTEM_PROC_DSP1; <== change core to one you want to use 

  • 0 in reply to Prasad Jondhale
    Prodigy 190 points
    I have done that its stopping at same line
    could you please suggest me how to merge 2 algorithm output and give it to another algorithm ?
    and which modification i need to do
  • 0 in reply to yash trivedi
    TI__Mastermind 27743 points
    Hello Yash,

    The changes you have made looks correct for merging output from two chains.
    Issue mostly is due to algorithm parameters not set correctly for framecopy or dmamosaic links.

    I would suggest you to build SDK in debug mode and put breakpoint in use-case main function and from there trace which set apps params causes assert.
    Other way is to put breakpoint in line 415 of algorithmLink_algPluginSupport.c and see function stack which function caused this assert.

    2.254193 s: Assertion @ Line: 415 in algorithmLink_algPluginSupport.c: status == SYSTEM_LINK_STATUS_SOK : failed !!!
  • 0 in reply to Prasad Jondhale
    Prodigy 190 points
    take this flow
    UseCase: chains_vipSingleCameraFrameCopy

    Capture -> Dup

    Dup -> Alg_FrameCopy (A15) -> Merge

    Dup -> Alg_FrameCopyAfterLut (A15) -> Merge

    Merge -> Display_Video

    GrpxSrc -> Display_Grpx

    both are framecopy algorithm how can i merge input from different queue and display it?
  • 0 in reply to yash trivedi
    TI__Mastermind 27743 points

    Yash,

    Is this issue resolved?

    I see my colleague Sujith helping you in other thread so I guess we can close this thread. Thanks.

    https://e2e.ti.com/support/arm/automotive_processors/f/1021/t/648623

  • 0 in reply to Prasad Jondhale
    Prodigy 190 points

    hey prasad

    could you please help me out about maximum number of lcd supported by tda2xx?

    I have seen lcd.h which stat that max is supported is 2 so would it still support if i change and try to connect 3 lcd altogether?

  • 0 in reply to yash trivedi
    TI__Guru 55685 points
    Hi Yash,

    TDA2xx supports three LCDs. You can change this macro to 3 and make necessary display controller changes (overlay, etc.) and have three LCD displays.

    Regards,
    Rishabh
  • 0
    TI__Genius 12455 points

    Hi Yash,

    The frame-copy support processing of single channel. You will have to update this plugin to support multiple channels.

    Regards, Sujith