Quantcast
Channel: Processors

Forum Post: RE: TDA4VM: TIDL TFLite delegate build for QNX in PSDK 09.01

0
0
Hi Jakub, The TIDL TFLite on SDK 9.1 is only supported by Linux for the TI SDKs. QNX/Blackberry team may have a port available. Thank you for notification of broken link will let TIDL team know. Regards, kb

Forum Post: RE: AM3356: How to receive CAN data without any drop or overrun even if CPU load is high

0
0
Hi, Did you download the TI RT SDK for the AM335x? Or are trying to patch the kernel tree? Does the command uname -a have this PREEMPT_RT in the response? Best Regards, Schuyler

Forum Post: RE: SK-AM64B: Query related to running led blink example on R5 core through linux

0
0
Hello Amey, This response will need to be brief, as I am on vacation again on Friday. I'll be able to reply again next week. Notes on booting remote cores AM64x Linux SDK with the SPL boot boots remote cores by default during Linux boot. Refer to here: https://dev.ti.com/tirex/explore/node?a=7qm9DIS__LATEST&node=A__AdAyuKWUWVV5j4wBc7C6XA__AM64-ACADEMY__WI1KRXP__LATEST You only need to update the firmware name in the sysfs interface /sys/class/remoteproc/remoteproc1/firmware OR the name of the firmware in the symbolic link ln -sf gpio_led_blink.out am64-main-r5f0_0-fw You don't need to do both. Reference the page I linked above. I haven't watched that youtube video, but it is probably out of date relative to the AM64x academy, which I tested on SDK 9.0. (NOTE!!! There were some significant changes between SDK 9.1 and SDK 9.0, and I haven't had the time to update the academy pages for SDK 9.1 yet. Future customers, refer to Amey's previous thread here for details: https://e2e.ti.com/support/processors-group/processors/f/processors-forum/1323865/sk-am64b-syscfg-ti_drivers_config-o-build-errors-while-trying-to-build-led-blink-example-with-ipc armstrip I am not familiar with this command. I assume it's in the video you referenced? If you give me a timestamp I can take a quick look. Can I prevent the core from executing until a Linux command is executed? Yes you can. There are several options here. The simplest one to program is probably to have your MCU+ code get loaded during boot time, but add a while() loop to your MCU+ code that waits until your Linux application sends an RPMsg telling the MCU+ code to start executing. There should also the option to disable the remote cores in the Linux devicetree file and manually start the cores from the terminal, but I have not played around with that for several years so I cannot give a good description of how to do it at this point in time. Very quick review of the linker.cmd files The biggest difference I see between the linker.cmd files on a quick scan is that your SDK 9.1 linker.cmd file places a lot of data in DDR (DDR_1) that your SDK 9.0 linker.cmd file places in MSRAM. You could check your .map file (generated next to the .out file) to see if you are using more memory than you allocated for that data section in the linker.cmd file. You'll also notice a difference in size for the allocation of the resource table - that is expected. For some reason SDK 9.0 (and ONLY AM64x SDK 9.0) required the resource table allocation to be 1MB, even though the resource table only took up 4kB. SDK 9.1 returned to the previous behavior of only needing 4kB allocation for the 4kB resource table. Regards, Nick

Forum Post: RE: AM68A: edgeai-tidl-tools latest version 09_01_06_00 errors out on host docker container with "source ./scripts/run_python_examples.sh"

0
0
Hi, Thank you for the additional information! I just tried this again and was able to see an almost identical error. Allow me some time to look into this with my team. There may be a delay in response due to tomorrow's holiday. Thank you, Fabiana

Forum Post: SK-AM62A-LP: QNX BSP Ethernet interface issues

0
0
Part Number: SK-AM62A-LP Hey TI-experts, We're having issues getting the ethernet interfaces running with the QNX BSP with the `ti-processor-sdk-qnx-am62axx_09_01_00_01` version. Specifically using: `710_SVN977709_JBN6_am62a`. (We've also tried the exact same with ti-processor-sdk-qnx-am62axx_09_00_00_01). After booting the built SD card we're getting: (Not containing an interface corresponding to the actual ethernet port). Looking through the slog2info we can see some issues such as: link_elf_load_file: devs-libpci.so load fail cause:Library cannot be found Failed to open devs-libpci.so and ignore_pci not set Error: ti_sci_msg_xfer (268): sciclient service call failed on ACK We've tried making various changes to the recipe, but to no avail. For instance relating to the network: Involving `io-pkt-v6-hc` which we've used for other targets (building QNX for tda4al). (However with this we're running into "ifconfig: getifaddrs: Bad message" outputs -- without any indication about what is actually missing). Explicitly added the `cpsw2g` interface. Various configurations for io-sock as outlined here: www.qnx.com/.../index.html Note that the below files are the outcome of using the SDK as-is building the QNX BSP following the instructions in the docs. (Had to append .txt to get the post to accept it): e2e.ti.com/.../0044.am62a_2D00_evm_2D00_ti.build.txt And the corresponding slog2info: e2e.ti.com/.../slog2info.log Assistance in the matter would be greatly appreciated, Thank you Best regards, Oskar

Forum Post: RE: SK-AM62: ILI2130

0
0
Hello Hamsadurga, First off, your compatible field looks wrong. That compatible string does not show up anywhere in the Linux kernel 6.1 docs. I would refer to the bindings documentation and make sure that you are exactly following the template: ti-processor-sdk-linux-am62xx-evm-09.01.00.08/board-support/ti-linux-kernel-6.1.46+gitAUTOINC+247b2535b2-g247b2535b2$ grep -r 'ili213' drivers/input/touchscreen/ilitek_ts_i2c.c: {.compatible = "ilitek,ili2130",}, drivers/input/touchscreen/ilitek_ts_i2c.c: {.compatible = "ilitek,ili2131",}, drivers/input/touchscreen/ilitek_ts_i2c.c: {.compatible = "ilitek,ili2132",}, Documentation/devicetree/bindings/input/touchscreen/ilitek_ts_i2c.yaml: - ilitek,ili2130 Documentation/devicetree/bindings/input/touchscreen/ilitek_ts_i2c.yaml: - ilitek,ili2131 Documentation/devicetree/bindings/input/touchscreen/ilitek_ts_i2c.yaml: - ilitek,ili2132 If you are still seeing issues after fixing your Linux devicetree node, please search your boot logs for any helpful errors and attach those as well. I'm in a bit of a rush today so I was not able to double-check your pinmux settings - if needed, I can check those next week (I'm on vacation again on Friday). Regards, Nick

Forum Post: RE: DRA821U: DDR Tool required for new LPDDR4 due to discontinuance issue of referenced Micron memory

0
0
Hi Jack, We generally request that customers fill in the register configuration tool. https://www.ti.com/lit/pdf/spracu8 If you have specific questions about how to fill in a specific parameter, please let us know. Regards, Kevin

Forum Post: RE: J7200XSOMXEVM: paired with the J721EXCP01EVM common processor board, PCIe X 1 LANE/PCIe X2 LANE not working

0
0
Hi Praveen, I see some concerning reviews on the product page for that card. For example: https://www.amazon.com/Bulk-Intel-Gigabit-Desktop-Adapter/product-reviews/B00BL4PQ9Y/ref=cm_cr_dp_d_show_all_btm?ie=UTF8&reviewerType=all_reviews Looking through the datasheet and at the form factor, it does seem like a PCI-Express card, but I would recommend you obtain a different cheap PCIe interface card to test if the issue is with the card or the TI board/SDK. On my end, I have tested with a PCIe WiFi card (Intel 9260 with adapters to fit the PCIe slot) with 8.6 SDK on J7200 and the card gets detected fine. Regards, Takuma

Forum Post: RE: SK-AM62A-LP: Facing error and not able to run the "ipc_notify_echo" example program on the "SK-AM62A-LP" EVK Kit (A53 and MCU-R5 IPC Communication Program)

0
0
Hello Nisarg, IPC_Notify, and where to find more info about MCU+ IPC Please note that IPC_Notify is ONLY supported between MCU+ cores. If you want to use IPC between Linux and an MCU+ core, you need to use IPC_RPMsg with 512 byte packets (where 496 bytes of data are possible, since there is a 16 byte header). Refer to https://software-dl.ti.com/mcu-plus-sdk/esd/AM62AX/09_01_00_39/exports/docs/api_guide_am62ax/IPC_GUIDE.html Note that the example you pointed to is to run on an A53 core that is running RTOS, NOT on an A53 core that is running Linux. Creating a MCU+ project that can be loaded by Linux remoteproc driver Most of the MCU+ SDK projects CANNOT be loaded by Linux remoteproc driver by default, because they are missing the resource table that the Linux remoteproc driver checks for before Linux will load a remote core firmware. Please modify your MCU+ application by following the steps in our AM62Ax academy, Multicore module, page Application Development on Remote Cores: https://dev.ti.com/tirex/explore/node?node=A__AeO-0ljXOpe7uiPS6ozGxg__AM62A-ACADEMY__WeZ9SsL__LATEST Regards, Nick

Forum Post: RE: SK-AM62A-LP: Facing error and not able to run the "ipc_notify_echo" example program on the "SK-AM62A-LP" EVK Kit (A53 and MCU-R5 IPC Communication Program)

0
0
You will find other helpful pages in the Multicore module that you also might want to look at, like the How to Develop with RPMsg IPC page: https://dev.ti.com/tirex/explore/node?node=A__AUcFeivQIlcT9WsOz3y0VA__AM62A-ACADEMY__WeZ9SsL__LATEST Regards, Nick

Forum Post: RE: TMDS64EVM: Interrupt priority from PRU to R5

0
0
Hello Kiyomasa, This is an MCU+ development question, not a PRU firmware question, so unfortunately I am not the right person to comment. I am reassigning your thread to a team member more familiar with MCU+ interrupts to answer your question. Regards, Nick

Forum Post: RE: PROCESSOR-SDK-J721S2: x86_64 Emulation: C7x address generator (SA) yeilds bad values when the DECDIM1is enabled

0
0
Hi Amine, I am not familiar with any current issues with __SA0ADV and I'll discuss internally with our compiler team. We may have you provide a compiler test case. From an internal C7x development standpoint, we have shifted to using the __vstore_pred instruction to do predicated stores over using _SA0ADV. Let me get back to you on 4/1 about this topic, we have a TI holiday on 3/29. Best, Asha

Forum Post: RE: AM625: Can anyone give me a code example of PWM of MCU_TIMER for M4F?

0
0
Hi Rafael, we will have EPWM for M4 included in 9.2 AM62x MCU-SDK. We are working on releasing 9.2 in a week or two. We will keep you posted. thank you, Paula

Forum Post: RE: AM625: PRU memory error

0
0
Hello Rafael, I'm quickly checking in with you before I'm back on vacation on Friday (I'll still be back the first week of April as promised. The first week of April is when I'll try to port the UART example to AM62x). make from the command line vs make from CCS Have you already gone through the PRU Getting Started Labs? If not, please go through those for me. https://software-dl.ti.com/processor-sdk-linux/esd/AM62X/09_01_00_08/exports/docs/common/PRU-ICSS/PRU-Getting-Started-Labs.html The last time I used CCS to build a PRU project was with AM62x SDK 8.3 while I was writing those labs. Let's see if you are still running into issues after following all the steps to set up your CCS environment. SysConfig settings? With PRU projects, we do not typically do any sysconfig configuration. This isn't too bad if the PRU only controls peripherals within the PRU subsystem - you just need to make sure you add the PRU pinmuxing to the Linux devicetree somewhere within an existing pinmux setting. If the PRU needs to control another peripheral on the processor, it gets more complex since we need a way to tell SYSFW that the PRU is getting ownership of those peripherals. I have not had time to dig into what that looks like for K3 devices like AM62x and AM64x - it is definitely possible and it has been done by some of my team members, but it is not straightforward with our current tools. For M4F, you DO configure the M4F using SysConfig. If you have not yet followed the MCU+ SDK docs, I would start by looking at this page: https://software-dl.ti.com/mcu-plus-sdk/esd/AM62X/09_01_00_39/exports/docs/api_guide_am62x/SYSCONFIG_INTRO_PAGE.html Regards, Nick

Forum Post: RE: TDA4AL-Q1: What is the eye spec of CSI-RX?

0
0
Hello Xingyu, I need to take a closer look at this topic. I will update you by end of business next Monday (1 April 2024). Thanks and regards, Mark

Forum Post: RE: TMDSIDK574: TMDSIDK574

0
0
Hi Teams, I would appreciate if you could response to my below question . >Is TMDSIDK574 compliant with RoHS2 (10 substances)? Best Regards, Kamijo / TED

Forum Post: RE: TMDSIDK574: How to buy this Eval kit? Or provide example to another EtherCat Master device

0
0
Hi Alec, Could you advise on the above question? Thanks! -s

Forum Post: RE: SK-AM68: Inconsistent color output on DisplayPort

0
0
Hi T S, Thanks for letting us know about the issue. This is a known issue which should be fixed in our upcoming 9.2 release next week. My recommendation would be to try out the new release once it comes out, since release date is near. Regards, Takuma

Forum Post: RE: AM62A3: WKUP R5F and remoteproc

0
0
Hello Gunter, Are your observations expected? "Attached" is expected. The Linux remoteproc driver checks to see if a core is running before it tries to initialize the core during kernel boot. If the core is already running (e.g., DM R5F which will always be running, or MCU R5F if the R5F was initialized earlier in the boot process by uboot (link here) or SBL (link here) ), then Linux will simply try to "attach" to the remote core (where it initializes the RPMsg communication interface between Linux and the remote core) instead of trying to re-boot the remote core. You CANNOT shut down the DM R5F during runtime - that means that you are unable to use the sysfs interface to stop the DM R5F, and then load new firmware into it. How to load new firmware into the DM R5F? You'll need to rebuild your uboot binary to build a new DM R5F firmware into it. There is a work-in-progress FAQ here with some additional information: https://e2e.ti.com/support/processors-group/processors/f/processors-forum/1317407/faq-am625-how-do-i-modify-the-ddr-memory-footprint-of-the-dm-r5f I was running into some issues when moving the DDR allocations around, and then my computer died and I have not had time to circle back yet. So ignore the parts about moving memory allocations around for now, but the rest of the information should be valid. Regards, Nick

Forum Post: AM623: Booting from eMMC (bootpart 0:1, mmcdev 0) Kernel begins to boot from eMMC but fails with kernel panic.

0
0
Part Number: AM623 Hi TI Experts Getting a kernel panic, but from u-boot the settings look correct. I think there is a problem with bootargs. I'm not sure if this is related to using GPT partitions .vs. MBR style bootable flag. Error: [ 2.686019] VFS: PARTUUID= is invalid. [ 2.686019] Expected PARTUUID= [/PARTNROFF=%d] [ 2.686024] Disabling rootwait; root= is invalid. [ 2.686650] /dev/root: Can't open blockdev [ 2.686668] VFS: Cannot open root device "PARTUUID=" or unknown-block(0,0): error -6 [ 2.686672] P [ 1.136841] mmc0: SDHCI controller on fa10000.mmc [fa10000.mmc] using ADMA 64-bit [ 1.140322] debugfs: Directory 'pd:182' with parent 'pm_genpd' already present! [ 1.141044] mmc2: CQHCI version 5.10 [ 1.141079] mmc1: CQHCI version 5.10 [ 1.143290] debugfs: Directory 'pd:186' with parent 'pm_genpd' already present! [ 1.155788] ALSA device list: [ 1.155799] No soundcards found. [ 1.183234] mmc2: SDHCI controller on fa20000.mmc [fa20000.mmc] using ADMA 64-bit [ 1.213866] mmc0: Command Queue Engine enabled [ 1.213885] mmc0: new HS200 MMC card at address 0001 [ 1.214587] mmcblk0: mmc0:0001 S0J56X 14.8 GiB mmc part Partition Map for MMC device 0 -- Partition Type: EFI Part Start LBA End LBA Name Attributes Type GUID Partition GUID 1 0x00000022 0x003f8021 "primary" ls mmc 0 4096 . 4096 .. 16384 lost+found 4096 bin 4096 boot 4096 dev => ls mmc $bootpart /boot/dtb 4096 . 4096 .. 4096 ti => ls mmc $bootpart /boot/dtb/ti 4096 . 4096 .. 1990 k3-am625-beagleplay-csi2-ov5640.dtbo 1636 k3-am625-sk-pwm.dtbo 2483 k3-am625-sk-rpi-hdr-ehrpwm.dtbo 59456 k3-am625-sk.dtb echo $mmcdev 0 => echo $bootpart 0:1 And I have executed => echo $name_fdt ti/k3-am625-sk.dtb => echo $fdtfile ti/k3-am625-sk.dtb mmc partconf 0 1 1 1 (This makes boot0 (0 1 1 1) running tiboot3.bin/tispl.bin/u-boot.img boot 1 (0 1 2 1)) HW Partition is the 3rd value, So UDA is (0 1 0 1) and RPMD is (0 1 3 0), partconf is zero indexed. mmc bootbus 0 2 0 0 8 bit eEMMC TO FIX THE " 2.657698] Please append a correct "root=" boot option; here are the available partitions:" I set: setenv root /dev/mmcblk0p1 Then => boot Which fails as above. Once I get U-boot to launch from eMMC u-boot I can update uEnv.txt.emmc to perform these steps automatically. Note, The Ethernet version of U-boot correctly utilizes uEnv.txt.ether to boot and the flash the eMMC boot0, erase eMMC UDA, then create partitions and load the RFS image (with kernel/dtb/ that DOES correctly boot Linux (without kernel panic). Just really close but still not right at this time. Regards, RG




Latest Images