Problem about DM365EVM with Leopard 5M02 camera board.

This is likely a difference in the kernel and the overall system.  You see, the capture driver (along with other drivers) need to allocate DDR2 memory to carry out their work.  In addition, depending on size of kernel loaded to DDR2 along with any other application running on DDR2, the total size of 'mem=' will vary from system to system. 

For IPNC, it appears they hace decided their system would require mem=116M.  This larger size may be due because their Linux system has not been optimized to fit in a smaller foot print yet, or maybe becuase it really cannot fit into a smaller memory foot-print.  I am not familiar enough with IPNC design to make this call.  Are they doing any encoding/decoding on this reference design?  If so, what resolutions?

For our EVM, we set mem=76M and made sure that our Linux system is small enough to fit in the memory foot-print.  The motivation for this is that we wanted to support a large enough CMEM memory foot-print to accomodate encoding and decoding of HD video resolutions (which can take much memory, specially if you are doing double and/or tripple buffering like our demos do).

To have both demos and IPNC app working in a single system, you may need to 1) optimize IPNC Linux system as much as feasible 2) rearrange CMEM to fit in smaller foot-print and change demos to use this smaller foot-print (e.g. eliminate tripple buffering, number of active windows, or support for higher resolutions).

Juan,

I think you may have confused the Appro IPNC camera with the setup listed above which is the DM365 EVM with the Leopard Imaging Camera Daughter card with the MT9P031 sensor onboard.  I was having the same problem with CMEM not loading and I never got around the issue.  (See: http://e2e.ti.com/forums/t/10097.aspx ).

I did manage to get video out of the camera so the Leopard code is most likely not using CMEM buffers.  I did not dig into to it after I got the camera to produce video, but would be interested in the solution if it comes up.

Thanks,

Jim

Hi Jmalone,

Seems that u have the same problem with me. I have some troubles about running Leopard programmes.

After insmod the mt9p031 drivers, running the ipipe_aew programme and the messages on the console seems the programme runs successful, but I can't see video display on the monitor.

The message is as below:

root@192.168.0.41:/opt/dvsdk# insmod mt9p031.ko
vpfe ccdc capture vpfe ccdc capture.1: vpif_register_decoder: decoder = MT9P031
root@192.168.0.41:/opt/dvsdk# ./ipipe_aew -v 31 &
davinci_resizer davinci_resizer.2: RSZ_G_CONFIG:0:1:108
[1] 785
opening resize device
Successfully setdavinci_previewer davinci_previewer.2: ipipe_set_preview_config
davinci_previewer davinci_previewer.2: ipipe_set_preview_config

default configuration setting in Resizer successfull
Resizer ivpfe ccdc capture vpfe ccdc capture.1: IMP chained
nitialized
Opervpfe ccdc capture vpfe ccdc capture.1: Resizer present
ating mode changed successfully to continuous in previewerSetting default config
uration in previewer
default configuration setting in previewer successfull
Setting default for LUT Defect Correction
Setting default for OTF Defect Pixel Correction
Setting default for 2-D Noise filter - 1
Setting default for 2-D Noise filter - 2
Setting default for Green Imbalance Correction
previewer -- set wb gain
Setting default for CFA Interpolation
previewer -- set rgb2rgb param
Setting default for Gamma Correction
Setting default for RGB-RGB Conversion - 2
Setting default for 3D LUT
Setting default for RGB-YCbCr conversion
Setting default for Global Brightness,Contrast Control
Setting default for YUV 422 conversion
Setting default for Edge Enhancer
Setting default for Chromatic Artifact Reduction
Setting default for Chromatic Gain Suppression
previewer initialized
input_std_params: name = V4L2_STD_MT9P031_480p_30FPS
input_std_params: scan width = 720
input_std_params: scan height = 480
input_std_params: output width = 720
input_std_params: output height = 480
init_camera_capture(): setting data format, width = 720, height = 480
root@192.168.0.41:/opt/dvsdk# MUX: initialized GPIO40
mt9p031 chip version reg = 1801
mt9p031_set_format_params height = 1919, width = 2559
setting data format, width = 720, height = 480
set_data_format:setting std to 0
camera_input =RAW-1
input.name = RAW
camera_input =RAW-1
input.name = COMPOSITE
camera_input =RAW-1
input.name = SVIDEO
camera_input =RAW-1
input.name = COMPONENT
camera_input =RAW-1
input.name = RAW-1
----------camera_input =RAW-1--------
----------input.name = RAW-1---------
----------input.index = 4--------
mt9p031 chip version reg = 1801
mt9p031_set_format_params height = 1919, width = 2559
mt9p031_set_format_params height = 1465, width = 2207
InitDevice:ioctl:VIDIOC_S_INPUT, selected input

Calling configCCDCraw()
pix_fmt = 0
frm_fmt = 0
fid_pol = 0
vd_pol = 0
hd_pol = 0
data_size = 4
data_shift = 0
test_pat_gen = 0
win.width = 640
win.height = 480
module_params.culling.hcpat_odd = 0xff
module_params.culling.hcpat_even = 0xff

configCCDCraw Done

S_STD Done
vpfe ccdc capture vpfe ccdc capture.1: hpitch = 720, vpitch = 480, bpp = 2

GetSTD Done WITH std = 0
set_data_format:requesting width:720 height:480
Setting format to V4L2_PIX_FMT_UYVYDavinciDisplay DavinciDisplay.1: Before finis
hing with S_FMT:
layer.pix_fmt.bytesperline = 1440,
 layer.pix_fmt.width = 720,
 layer.pix_fmt.height = 480,
 layer.pix_fmt.sizeimage =691200
 at capture

SDavinciDisplay DavinciDisplay.1: pixfmt->width = 720,
 layer->layer_info.config.line_length= 1440
_FMT Done
initiStarting ccdc_config_raw..
alizing capture starting ccdc_reset...buffers

REQBU<7>
End of ccdc_reset...F Done

QUERYB<7>Writing 0x2 to ...CCDCFG
UF Done
captureStarting ccdc_setwin... buffer = 0x8400<7>
End of ccdc_config_ycbcr...
0000

QUERYBUF Done
capture buffer = 0x84800000

QUERYBUF Done
capture buffer = 0x83000000

QUERYBUF Done
capture buffer = 0x81000000

QUERYBUF Done
capture buffer = 0x81800000
Capture initialized
---------init_display_device():width = 720 ,height = 480--------
/sys/class/davinci_display/ch0/output was opened successfully
Current output value is COMPOSITE

Changed output to COMPOSITE

/sys/class/davinci_display/ch0/mode was opened successfully
Current mode value is NTSC

Changed mode to PAL

Setting display pixel format to V4L2_PIX_FMT_UYVY
Initialized display
Queing buffer:0

Q_BUF Done
Queing buffer:1

Q_BUF Done
Queing buffer:2

Q_BUF Done
Queing buffer:3

Q_BUF Done
Queing buffer:4

Q_BUF Done

STREAMON Done
5. Test enque first buffer
6. Test enqueuing of buffers - done
5. Test enque second buffer
6. Test enqueuing of buffers - done
7. Test STREAM_ON

fps = 25.00

Thanks for your help.