Hi Ugnius, Please go through the below e2e thread: e2e.ti.com/.../502276 Let me know if you still have NAND boot questions. Regards, Pavel
↧
Forum Post: RE: Linux/AM3358: NAND boot error
↧
Forum Post: PROCESSOR-SDK-TDAX: TDA2xx IN/OUT
Part Number: PROCESSOR-SDK-TDAX Hello to all, I am trying to send video from camera to VOUT1 on TDA2xx that is connected to VIN1A on second TDA2xx chip, as I can see, that is connected via LCD, so I made some changes in chains_vipSingleCam_DualDisplay so I could get video from cameras and send it to LCD and HDMI. In chains_common.c file, in function ChainsCommon_DualDisplay_StartDisplayCtrl I just set the lcdType to CHAINS_DISPLAY_TYPE_LCD_10_INCH and no signal was sent to LCD but I have video on HDMI. I also tried by setting lcdType to CHAINS_DISPLAY_TYPE_LCD_7_INCH. On second TDA2xx I have made usecase to capture video and display it on HDMI. To set capture params I used ChainsCommon_SingleCam_SetCapturePrms function from chains_common_vision.c, set second and third parameter to LCD width and height (1280 and 800). In chains_common_vision in that function, I set captureSrc to CHAINS_CAPTURE_SRC_ DM388 and in there pInstPrm->vipInstId = SYSTEM_CAPTURE_INST_VIP1_SLICE1_PORTA, pInstPrm->videoIfWidth = SYSTEM_VIFW_24BIT and pInprms->dataFormat = SYSTEM_DF_RGB24_888. Best regards, Bosko.
↧
↧
Forum Post: RE: TDA2SX: TDA2xx A15 frequency
Hello, [quote user="user1427660"]For TDA2xx, can I build with OPPMODE=opp_high ?[/quote] Yes. [quote user="user1427660"]anywhere else need to be modified ?[/quote] I don't think so, not aware, the build argument should be enough. Try it. [quote user="user1427660"]If yes, what will the frequency be ?[/quote] The frequencies vary from part to part number. Refer to the device Data Manual to find what frequencies your part number supports. Thanks Alex
↧
Forum Post: RE: Linux/AM5728: OTA update of the OS
Hi Bin, will this technique eliminate the files that gets different just because of their building time-stamp? Ultimately I want to mark the files those are only generated or modified as a result of the patches applied for that upgrade. Do you think this technique will give same results? Thanks, Divyeshkumar
↧
Forum Post: RE: TMDXIDK5718: IDK board not working
Hi Tom, Sorry for the delay reply. As per the your suggestion,measured the following signals on the EVM board and corresponding values are below: LDO2_OUT is 0.0V LDOVRTC_OUT is 1.7498 V LDOVANA_OUT is 0.0 V VCC1 Input is 4.933V GPIO_7/POWERHOLD and RESET signals are Low. Thanks, Sreedhar
↧
↧
Forum Post: RE: AM3358: NAND not detected on GP EVM
You are not writing the complete arago-base-tisdk-image-am335x-evm.ubi, you are writing 41Mib out of 145Mib. Its size has increased during the SDK development. Change the 0x2950000 to 0x92c0000.
↧
Forum Post: RE: TMDXIDK5718: IDK board not working
Sreedhar, RESET_IN is low? Are you certain that LDOVAVA is 0V? Where did you measure these voltages? Tom
↧
Forum Post: RE: TMDXIDK5718: IDK board not working
Tom, Yes, RESET_IN is Low and measured at R794. LDOVANA is 0V and measured across C302. Thanks, Sreedhar
↧
Forum Post: RE: Linux: AM437x pin toggling at boot time
Our device tree has been modified to account for this so I'm not sure exactly why we can't use this pin for GPIO. It works fine for our usage except during this boot up sequence. When we put a scope on the pins for the USB0_DRVVBUS and USB1_DRVVBUS we notice that the pins do not act in similar fashion. USB1_DRVVBUS stays low on load but USB0_DRVVBUS goes high but not high like it was turned on. The behavior looks like there is a pulldown resistor connected to USB0_DRVVBUS but we know there isn't one on the hardware or configured to be pulldown by the device tree. Once the USB controller software has been loaded by the kernel the USB1_DRVVBUS goes high as expected and USB0_DRVVBUS goes low as expected. This is a strange behavior to us here and we're curious what in the distributed kernel SDK could possibly be telling that PIN to go high and the one we actually configured in the device tree goes low.
↧
↧
Forum Post: RE: Linux/AM3358: NAND boot error
[quote user="Ugnius Subacius"]U-Boot SPL 2016.05-gaeedf80-dirty (Oct 26 2017 - 13:01:07) Trying to boot from NAND The Expected Linux image was not found. Please check your NAND configuration. Trying to start u-boot now...[/quote] This is most probably caused by file size mismatch and/or wrong offset. Refer to the below pointers for more info: [quote user="Ugnius Subacius"]The U-boot by default first tries to boot from SD card and checks ethernet:[/quote] Make sure your boot pins are configured for NAND boot first. Regards, Pavel
↧
Forum Post: RE: TMDXIDK5718: IDK board not working
Sreedhar, Can you probe around U105? What do you measure on VMAIN and VPWRIN_JCK and the CT and MR pins? Is 1.8V (VRTC_OUT) seen on R928? Tom
↧
Forum Post: RE: AM5728: PCIe access error
Hi, Sung-IL, The error is from l3_interrupt_handler() and seems to be caused by DSP1 acessing L4_CFG through MDMA interface. I suspect it is the memory used by DSP still under ARM MMU protection. You may want to configure those area used by PCIe in DSP resource table. I'll have DSP engineer to take a closer look at this one. Rex
↧
Forum Post: RE: RTOS/AM5728: Component versions
Hi, For DSP computation, please check processors.wiki.ti.com/.../Processor_SDK_RTOS_Software_Developer_Guide , there is tab called DSP Optimized Libraries, with link to DSPLIB, FFTLIB , etc. For IPC, starting from processors.wiki.ti.com/.../IPC_Users_Guide. Regards, Eric
↧
↧
Forum Post: RE: Linux/AM3358: NAND boot error
Hi Ugnius, What is your NAND size and change the NAND configuration as per your NAND size in "am335x_evm.h". #define CONFIG_SYS_NAND_PAGE_SIZE #define CONFIG_SYS_NAND_OOBSIZE #define CONFIG_SYS_NAND_BLOCK_SIZE #define CONFIG_SYS_NAND_ECCPOS #define CONFIG_SYS_NAND_ECCSIZE #define CONFIG_SYS_NAND_ECCBYTES #define CONFIG_NAND_OMAP_ECCSCHEME #define MTDPARTS_DEFAULT #define CONFIG_CMD_SPL_NAND_OFS ?? /* os parameters */ #define CONFIG_SYS_NAND_SPL_KERNEL_OFFS ??/* kernel offset */ #define CONFIG_CMD_SPL_WRITE_SIZE ?? Thanks, Anil
↧
Forum Post: RE: RTOS/AM3352: Example project for NDK with interrupts
Michael, While Dave is continuing to inspect the issue from hardware side, I try to summarize what we understand from recent investigation, and would recommend to step back to EMAC unit test example - 'EMAC_BasicExample_evmAM335x_armExampleproject' in the PDK for helping narrow down the root cause. 1. We confirmed that pinmux as well as clock is not done by PDK board driver, and verified your pinmux_init is correct. Also the clock tree for CPSW should have been setup correctly. If it fails, the module idle status check should not pass. 2. We decoded the PHY registers. The PHY @ MDIO ADDR 0x01 is reporting a LINK_UP @ 100Mbs/Half Duplex - Auto-Negotiation Complete, which is somewhat rare. It doesn't appear there are any other issues on the PHY side. 3. From CPSW stats register, we don't see any GOOD_Rx frames, which indicates the remote DHCP server doesn't receive any DHCP discovery message from the client. 4. The code for setting up the DHCP client with NDK in the project should be stable, i.e. can initiate the DHCP discovery by broadcasting messages. 5. Able to configure EMAC instance with the only one port with PHY address 1 existed on the custom board, however the 'negotiated connection' still shows 100Mbs/Half. The EMAC unit test example is a EMAC Loopback test at CPSW switch interface through an external loopback adapter. The example demonstrates loopback capability by sending dummy broadcast packet to CPSW switch. Unit test registers receive packet callback routine with LLD to be called for RX packet. Call back routine will extract packet received, perform simple memory comparison against packet sent for integrity check. The unit test will iterate over configured packet count for packet transmission and reception check, see processors.wiki.ti.com/.../Processor_SDK_RTOS_EMAC Again, due to the PHY difference (Qualcomm's AR8031 on EVM vs. Microchip’s LAN8710A on BBB vs. TI's DP83848 on your board), you might need look closely to check if need merge the PHY configuration functions from your current code to the loopback example project. Regards, Garrett
↧
Forum Post: RE: CCS/AM5728: RTOS boot from eMMC
Anders, Currently the MMC/SD driver in the Processor SDK does support reading and writing from eMMC memory as you can see from the section "eMMC Usage" in the wiki processors.wiki.ti.com/.../Processor_SDK_RTOS_MMCSD The missing pieces are created a eMMC writer to write MLO and app to EMMC and modify the SD boot to also cover eMMC. One thing that we are discussing is whether to support only raw boot or FATFS based boot from eMMC. Yes, u-boot can be used as starting point as it does have eMMC boot but you will need to add the code into uboot to power up slave cores and load and run application on slave cores if that is required. Previously we have noticed that ARM-Linux boot usually power down slave cores until the remoteproc driver loads IPC firmware on DSP/M4 so there may be need for more updates. Regards, Rahul
↧
Forum Post: RE: Linux/DRA725: omapdrmtest
Hi Cesar, did you run 'modetest' to see if it finds second display as suggested by Ramprasad? Thanks, Yordan
↧
↧
Forum Post: RE: RTOS/AM3352: Example project for NDK with interrupts
here is the scope plot (purple is RXER):
↧
Forum Post: TDA2HG: Where can I find documentation about the Device Security Features?
Part Number: TDA2HG Hello, I actually want to use the TDA2+, but the answer might be the same as for TDA2HG .(I was not able to select TDA2+ in the "Part Number" field.) Where can I find documentation about the device security features? I want to make sure that we can protect our software: Software will be stored encrypted in external eMMC Flash and will be loaded by the bootloader. Is there a way to store a key inside the SOC. I read "Secure ROM and boot" in the Data Sheet. But there is no further information about that. Best regards, Michael.
↧
Forum Post: RE: Linux/AM5728: HDMI hot plug issue
Hi Boycchi, we are able to replicate the failure. We will get back with our analysis soon.
↧
