Part Number: TDA4VH-Q1 Other Parts Discussed in Thread: TDA4VH Tool/software: Hi: I recently discovered the following forum topic: https://e2e.ti.com/support/processors-group/processors/f/processors-forum/1283968/tda4vh-q1-cache-coherency I have also been examining the PSDK software but there seems to be a contradiction between what I see in the software and the answer provided in the above topic. For reference, I am looking at the 9.2.0.5 Linux ADAS and PSDK software distributions (which should be the latest). Some background: The TDA4VH has A72 cores (compute cluster), R5f cores and C71 DSP processors. With TI's design, the Linux kernel provides facilities for using RPMSG to talk to the remote processors. The A72 cores and remote processors can also share portions of DDR memory between them. The PSDK distribution contains library and application code that interfaces with the kernel so that application code and interact with the remote processors. As currently structured, the PSDK code has an app_utils layer which provides some abstraction for the tiovx and vision application code. In particular, app_utils provides routines like appMemAlloc(), appMemMap() and appMemCacheWb() and appMemCacheInv(). In the Linux implementation, these routines seem to use a dma_bufs interface, which uses ioctls to allocate DMA-capable memory regions. These routines in turn are used by higher level APIs like tivxMemBufferAlloc(), tivxMemBufferMap() and tivxMemBufferUnmap(). It's important of course that cache coherency be accounted for when using these routines, since the allocated buffers are shared with the remote processors. Now, here's the thing: the forum topic referenced above says this: "For TDA4VH each A72 cluster is direct connected to the MSMC3 so both A72 clusters are fully coherent between each other however each C7x is no longer direct connected to the MSMC3. This means that some SW cache operations are needed when exchanging data between an A72 instance and a C7 instance." Either this is not actually correct, or I'm missing something important. Here's what the tiovxMemBufferMap() function does: { #if defined(SOC_AM62A) appMemCacheWb(host_ptr, size); #else #ifndef A72 appMemCacheWb(host_ptr, size); #endif #endif } Notice that for the A72 processor, it skips the appMemCacheWb() call entirely. The same soft of construction is present in tiovxMemBufferUnmap(). From my analysis, there is also a separate shared memory region used for exchanging object descriptors, which is used with the IPC module. This is used for example when calling ownObjDescSend(). From what I can see, there is no provision in the descriptor management code for cache synchronization at all. Given that it seems the tiovx code seems to be skipping/missing any software cache sync operations for these buffers, there are only two ways I can see how this could possibly work: 1) Linux is mapping these regions with the cache disabled 2) Linux is somehow mapping these regions as cache-enabled, and is also ensuring hardware cache coherency with the remote processors. Here's my problem: I'm trying to implement support for the PSDK library on a different OS (running on the A72 cores), and I want to be able to host it the same way that Linux does. But I'm having trouble figuring out the right way to manage cache coherency. Currently the shared memory regions are mapped as cache inhibited, and that works. When I say "works," this means I'm using the valet parking demo (app_tidl_avp). It processes frames and displays progress on an extrnal monitor via the DP1 display port connector as expected. But this means accesses to the shared memory regions by the A72 cores are much slower than I'd like. If I try to map the regions as cache enabled, the demo won't run at all. This leaves me with a couple of important questions: 1) Is the answer in the previous discussion thread really correct? Are software cache operations really needed on the A72 cores for synchronization with the C71 cores? The PSDK software itself seems to contradict this. 2) If hardware cache coherency is possible, what specific Arm MMU page attributes would be needed to ensure it? Also, would any other configuration be needed, e.g. of the MSMC module? (Please don't reply to this by saying "just look at what TI Linux does" because if I knew where in the source code to look I would have done so by now.) Unfortunately the other thread says: "The software in TI's SDK is designed to follow the rules as described above. Most end users at the A72 and C7x cluster level are probably adding/building applications atop framework API abstractions so they likely are not worrying about the lower layer plumbing." I guess I am not "most users." -Bill

Part Number: AM62P Tool/software: How can you measure the wakeup latency for Low Power Modes on AM62x/A/P? Low Power Mode's wakeup latency is the amount of time required from triggering a wakeup event to resuming normal operation. This FAQ shows to measure this time. This applies to AM62x, AM62A, AM62P.

Hi Haythem, Sorry for the delay. Before I can answer, could you please specify the document you are referring to for the queries. Thanks Yogesh

Hi Zhiyong, Could you start by sharing the output of media-ctl -p? Thank you, Fabiana

Hi Felix, the BOOTMODE pins are actually latched around the point when MCU_PORz_Out goes high. Are the bootmode pins stable at this point? Do you see any activity on PCIe interface after RESETSTATz goes high? Regards, James

Hi, Can you clarify what the issue is? There is an interrupt and callback called when your read a GPIO pin? Thanks, Neehar

It is not really the same as AM335x, as there are multiple processors that get involved in the boot flow. The AM62x will perform an integrity check of the boot image before attempting to boot (This is described in the Figure 5-3 in the TRM). Linux boot flow is also described here: https://dev.ti.com/tirex/explore/node?node=A__AXMgiFB9o6cNWaPJLfbHDw__AM62-ACADEMY__uiYMDcq__LATEST&search=am62x REgards, James

Hi Wilson, I have looked into this and there should not be any significant changes or changes within the SDK that may be causing issues with the API. Are you looking to upgrade to SDK 9.2? Can you slowly integrate the changes from SDK 8.6 to 9.2 to root cause what the issue may be? Thanks, Neehar

Ruggero, any pin designated as reserved for a particular boot mode is not used by ROM. However, they cannot just be left floating, you must either pull them up or down on your board. Regards, James

Using strace strace is a function built in the default SDK. Add the -c flag to generate a report of time spent on system calls to measure the latency of the wakeup sequence. root@am62pxx-evm:~# strace -c echo mem > /sys/power/state [ 21.837337] PM: suspend entry (deep) [ 22.207596] Filesystems sync: 0.366 seconds [ 22.216783] Freezing user space processes [ 22.219505] Freezing user space processes completed (elapsed 0.002 seconds) [ 22.219531] OOM killer disabled. [ 22.219536] Freezing remaining freezable tasks [ 22.220871] Freezing remaining freezable tasks completed (elapsed 0.001 seconds) [ 22.220881] printk: Suspending console(s) (use no_console_suspend to debug) [ 22.262892] omap8250 2800000.serial: PM domain pd:146 will not be powered off [ 22.269272] Disabling non-boot CPUs ... [ 22.271191] psci: CPU1 killed (polled 0 ms) [ 22.274417] psci: CPU2 killed (polled 0 ms) [ 22.277276] psci: CPU3 killed (polled 0 ms) [ 22.278352] Enabling non-boot CPUs ... [ 22.278672] Detected VIPT I-cache on CPU1 [ 22.278742] GICv3: CPU1: found redistributor 1 region 0:0x00000000018a0000 [ 22.278785] CPU1: Booted secondary processor 0x0000000001 [0x410fd034] [ 22.279487] CPU1 is up [ 22.279713] Detected VIPT I-cache on CPU2 [ 22.279751] GICv3: CPU2: found redistributor 2 region 0:0x00000000018c0000 [ 22.279775] CPU2: Booted secondary processor 0x0000000002 [0x410fd034] [ 22.280255] CPU2 is up [ 22.280478] Detected VIPT I-cache on CPU3 [ 22.280520] GICv3: CPU3: found redistributor 3 region 0:0x00000000018e0000 [ 22.280547] CPU3: Booted secondary processor 0x0000000003 [0x410fd034] [ 22.281080] CPU3 is up [ 22.281398] ti-sci 44043000.system-controller: ti_sci_resume: wakeup source: 0x80 [ 22.382049] OOM killer enabled. [ 22.385192] Restarting tasks ... done. [ 22.390920] random: crng reseeded on system resumption % time seconds usecs/call calls errors syscall ------ ----------- ----------- --------- --------- -------------[ 22.396718] PM: suspend exit --- 99.65 0.133407 133407 1 write 0.10 0.000133 33 4 mprotect 0.06 0.000081 27 3 munmap 0.05 0.000072 18 4 close 0.05 0.000070 11 6 mmap 0.03 0.000040 13 3 newfstatat 0.02 0.000024 8 3 brk 0.01 0.000013 13 1 getrandom 0.01 0.000009 9 1 prlimit64 0.01 0.000008 8 1 set_tid_address 0.01 0.000007 7 1 rseq 0.00 0.000006 6 1 set_robust_list 0.00 0.000000 0 1 1 faccessat 0.00 0.000000 0 2 openat 0.00 0.000000 0 1 read 0.00 0.000000 0 1 pread64 0.00 0.000000 0 1 execve ------ ----------- ----------- --------- --------- ---------------- 100.00 0.133870 3824 35 1 total root@am62pxx-evm:~# The table shows 0.1339 seconds so the wakeup sequence took 133.9 ms.

Using /sys/power/pm_print_times This is not enabled by default in the SDK. In menuconfig, go to 'Power managment options --->' Enable 'Power Managment Debug Support' The changes should be reflected in the .config file: CONFIG_PM_DEBUG=y Then build the Linux Image and modules & copy over to the SD Card (or which ever boot method). At the command line, enable pm_print_times by running 'echo 1 > /sys/power/pm_print_times' root@am62pxx-evm:~# echo 1 > /sys/power/pm_print_times root@am62pxx-evm:~# echo mem > /sys/power/state [ 152.534525] PM: suspend entry (deep) [ 152.563291] Filesystems sync: 0.024 seconds [ 152.571367] Freezing user space processes [ 152.577331] Freezing user space processes completed (elapsed 0.005 seconds) [ 152.577402] OOM killer disabled. [ 152.577416] Freezing remaining freezable tasks [ 152.579050] Freezing remaining freezable tasks completed (elapsed 0.001 seconds) [ 152.579065] printk: Suspending console(s) (use no_console_suspend to debug) [ 152.623975] platform mcu_gpio_key: PM: calling platform_pm_suspend+0x0/0x70 @ 1033, parent: platform [ 152.624006] platform mcu_gpio_key: PM: platform_pm_suspend+0x0/0x70 returned 0 after 0 usecs [ 152.624020] platform 2b300050.target-module: PM: calling platform_pm_suspend+0x0/0x70 @ 1033, parent: bus@f0000:bus@b00000 [ 152.624031] platform 2b300050.target-module: PM: platform_pm_suspend+0x0/0x70 returned 0 after 0 usecs [ 152.624101] input input0: PM: calling input_dev_suspend+0x0/0x80 @ 1033, parent: gpio_key [ 152.624116] input input0: PM: input_dev_suspend+0x0/0x80 returned 0 after 1 usecs [ 152.624113] mmcblk mmc1:aaaa: PM: calling mmc_bus_suspend+0x0/0x80 @ 1014, parent: mmc1 [ 152.624136] leds mmc1::: PM: calling led_suspend+0x0/0x60 @ 1033, parent: fa00000.mmc [ 152.624154] leds mmc1::: PM: led_suspend+0x0/0x60 returned 0 after 1 usecs [ 152.625903] mmcblk mmc1:aaaa: PM: mmc_bus_suspend+0x0/0x80 returned 0 after 1774 usecs [ 152.625940] sdhci-am654 fa00000.mmc: PM: calling pm_runtime_force_suspend+0x0/0x14c @ 1033, parent: bus@f0000 [ 152.626038] sdhci-am654 fa00000.mmc: PM: pm_runtime_force_suspend+0x0/0x14c returned 0 after 80 usecs [ 152.626058] regulator regulator.2: PM: calling regulator_suspend+0x0/0x114 @ 1033, parent: regulator-2 [ 152.626072] regulator regulator.2: PM: regulator_suspend+0x0/0x114 returned 0 after 1 usecs [ 152.626085] reg-fixed-voltage regulator-2: PM: calling platform_pm_suspend+0x0/0x70 @ 1033, parent: platform [ 152.626097] reg-fixed-voltage regulator-2: PM: platform_pm_suspend+0x0/0x70 returned 0 after 0 usecs [ 152.626119] regulator regulator.1: PM: calling regulator_suspend+0x0/0x114 @ 1033, parent: regulator-3 [ 152.626131] regulator regulator.1: PM: regulator_suspend+0x0/0x114 returned 0 after 0 usecs [ 152.626129] pca953x 1-0022: PM: calling pca953x_suspend+0x0/0x8c @ 51, parent: i2c-1 [ 152.626144] gpio-regulator regulator-3: PM: calling platform_pm_suspend+0x0/0x70 @ 1033, parent: platform [ 152.626150] pca953x 1-0022: PM: pca953x_suspend+0x0/0x8c returned 0 after 4 usecs [ 152.626155] gpio-regulator regulator-3: PM: platform_pm_suspend+0x0/0x70 returned 0 after 0 usecs [ 152.626173] davinci_gpio 601000.gpio: PM: calling davinci_gpio_suspend+0x0/0xcc @ 1033, parent: bus@f0000 [ 152.626188] davinci_gpio 601000.gpio: PM: davinci_gpio_suspend+0x0/0xcc returned 0 after 2 usecs [ 152.626207] davinci_gpio 600000.gpio: PM: calling davinci_gpio_suspend+0x0/0xcc @ 1033, parent: bus@f0000 [ 152.626222] davinci_gpio 600000.gpio: PM: davinci_gpio_suspend+0x0/0xcc returned 0 after 3 usecs [ 152.626237] cpufreq-dt cpufreq-dt: PM: calling platform_pm_suspend+0x0/0x70 @ 1033, parent: platform [ 152.626247] cpufreq-dt cpufreq-dt: PM: platform_pm_suspend+0x0/0x70 returned 0 after 0 usecs [ 152.626262] omap8250 2800000.serial: PM: calling omap8250_suspend+0x0/0x140 @ 1033, parent: bus@f0000 [ 152.626426] omap8250 2800000.serial: PM domain pd:146 will not be powered off [ 152.626447] omap8250 2800000.serial: PM: omap8250_suspend+0x0/0x140 returned 0 after 169 usecs [ 152.626546] ti-udma 485c0000.dma-controller: PM: calling platform_pm_suspend+0x0/0x70 @ 1033, parent: bus@f0000:bus@48000000 [ 152.626557] ti-udma 485c0000.dma-controller: PM: platform_pm_suspend+0x0/0x70 returned 0 after 0 usecs [ 152.626677] ti-udma 485c0100.dma-controller: PM: calling platform_pm_suspend+0x0/0x70 @ 1033, parent: bus@f0000:bus@48000000 [ 152.626688] ti-udma 485c0100.dma-controller: PM: platform_pm_suspend+0x0/0x70 returned 0 after 0 usecs [ 152.626709] ti-sci-inta 48000000.interrupt-controller: PM: calling platform_pm_suspend+0x0/0x70 @ 1033, parent: bus@f0000:bus@48000000 [ 152.626720] ti-sci-inta 48000000.interrupt-controller: PM: platform_pm_suspend+0x0/0x70 returned 0 after 0 usecs [ 152.626733] ti-sci-intr bus@f0000:interrupt-controller@a00000: PM: calling platform_pm_suspend+0x0/0x70 @ 1033, parent: bus@f0000 [ 152.626744] ti-sci-intr bus@f0000:interrupt-controller@a00000: PM: platform_pm_suspend+0x0/0x70 returned 0 after 0 usecs [ 152.626779] pca953x 1-0023: PM: calling pca953x_suspend+0x0/0x8c @ 51, parent: i2c-1 [ 152.626796] pca953x 1-0023: PM: pca953x_suspend+0x0/0x8c returned 0 after 4 usecs [ 152.626852] ti-sci-reset 44043000.system-controller:reset-controller: PM: calling platform_pm_suspend+0x0/0x70 @ 1033, parent: 44043000.system-controller [ 152.626865] ti-sci-reset 44043000.system-controller:reset-controller: PM: platform_pm_suspend+0x0/0x70 returned 0 after 0 usecs [ 152.626878] platform e000000.watchdog: PM: calling platform_pm_suspend+0x0/0x70 @ 1033, parent: bus@f0000 [ 152.626888] platform e000000.watchdog: PM: platform_pm_suspend+0x0/0x70 returned 0 after 0 usecs [ 152.626901] platform e010000.watchdog: PM: calling platform_pm_suspend+0x0/0x70 @ 1033, parent: bus@f0000 [ 152.626911] platform e010000.watchdog: PM: platform_pm_suspend+0x0/0x70 returned 0 after 0 usecs [ 152.626926] platform e020000.watchdog: PM: calling platform_pm_suspend+0x0/0x70 @ 1033, parent: bus@f0000 [ 152.626935] platform e020000.watchdog: PM: platform_pm_suspend+0x0/0x70 returned 0 after 0 usecs [ 152.626948] platform e030000.watchdog: PM: calling platform_pm_suspend+0x0/0x70 @ 1033, parent: bus@f0000 [ 152.626958] platform e030000.watchdog: PM: platform_pm_suspend+0x0/0x70 returned 0 after 0 usecs [ 152.626973] ti-sci-clk 44043000.system-controller:clock-controller: PM: calling platform_pm_suspend+0x0/0x70 @ 1033, parent: 44043000.system-controller [ 152.626984] ti-sci-clk 44043000.system-controller:clock-controller: PM: platform_pm_suspend+0x0/0x70 returned 0 after 0 usecs [ 152.627000] ti_sci_pm_domains 44043000.system-controller:power-controller: PM: calling platform_pm_suspend+0x0/0x70 @ 1033, parent: 44043000.system-controller [ 152.627011] ti_sci_pm_domains 44043000.system-controller:power-controller: PM: platform_pm_suspend+0x0/0x70 returned 0 after 0 usecs [ 152.627025] ti-sci 44043000.system-controller: PM: calling platform_pm_suspend+0x0/0x70 @ 1033, parent: bus@f0000 [ 152.627035] ti-sci 44043000.system-controller: PM: platform_pm_suspend+0x0/0x70 returned 0 after 0 usecs [ 152.627051] snd-soc-dummy snd-soc-dummy: PM: calling platform_pm_suspend+0x0/0x70 @ 1033, parent: platform [ 152.627061] snd-soc-dummy snd-soc-dummy: PM: platform_pm_suspend+0x0/0x70 returned 0 after 0 usecs [ 152.627076] platform psci-cpuidle: PM: calling platform_pm_suspend+0x0/0x70 @ 1033, parent: platform [ 152.627086] platform psci-cpuidle: PM: platform_pm_suspend+0x0/0x70 returned 0 after 0 usecs [ 152.627098] ti-cpufreq ti-cpufreq: PM: calling platform_pm_suspend+0x0/0x70 @ 1033, parent: platform [ 152.627108] ti-cpufreq ti-cpufreq: PM: platform_pm_suspend+0x0/0x70 returned 0 after 0 usecs [ 152.627129] platform Fixed MDIO bus.0: PM: calling platform_pm_suspend+0x0/0x70 @ 1033, parent: platform [ 152.627139] platform Fixed MDIO bus.0: PM: platform_pm_suspend+0x0/0x70 returned 0 after 0 usecs [ 152.627221] serial8250 serial8250: PM: calling platform_pm_suspend+0x0/0x70 @ 1033, parent: platform [ 152.627235] serial8250 serial8250: PM: platform_pm_suspend+0x0/0x70 returned 0 after 3 usecs [ 152.627319] gpio-keys gpio_key: PM: calling platform_pm_suspend+0x0/0x70 @ 1033, parent: platform [ 152.627333] gpio-keys gpio_key: PM: platform_pm_suspend+0x0/0x70 returned 0 after 3 usecs [ 152.627581] platform l2-cache0: PM: calling platform_pm_suspend+0x0/0x70 @ 1033, parent: platform [ 152.627592] platform l2-cache0: PM: platform_pm_suspend+0x0/0x70 returned 0 after 0 usecs [ 152.627606] pinctrl-single f4000.pinctrl: PM: calling platform_pm_suspend+0x0/0x70 @ 1033, parent: bus@f0000 [ 152.627656] pinctrl-single f4000.pinctrl: PM: platform_pm_suspend+0x0/0x70 returned 0 after 40 usecs [ 152.627670] ti-msgmgr 4d000000.mailbox: PM: calling platform_pm_suspend+0x0/0x70 @ 1033, parent: bus@f0000:bus@48000000 [ 152.627706] ti-msgmgr 4d000000.mailbox: PM: platform_pm_suspend+0x0/0x70 returned 0 after 25 usecs [ 152.627722] simple-pm-bus bus@f0000:bus@48000000: PM: calling platform_pm_suspend+0x0/0x70 @ 1033, parent: bus@f0000 [ 152.627732] simple-pm-bus bus@f0000:bus@48000000: PM: platform_pm_suspend+0x0/0x70 returned 0 after 0 usecs [ 152.627746] phy-gmii-sel 104044.phy: PM: calling platform_pm_suspend+0x0/0x70 @ 1033, parent: 100000.bus [ 152.627757] phy-gmii-sel 104044.phy: PM: platform_pm_suspend+0x0/0x70 returned 0 after 0 usecs [ 152.627769] simple-pm-bus 100000.bus: PM: calling platform_pm_suspend+0x0/0x70 @ 1033, parent: bus@f0000 [ 152.627779] simple-pm-bus 100000.bus: PM: platform_pm_suspend+0x0/0x70 returned 0 after 0 usecs [ 152.627791] k3-chipinfo 43000014.chipid: PM: calling platform_pm_suspend+0x0/0x70 @ 1033, parent: 43000000.syscon [ 152.627802] k3-chipinfo 43000014.chipid: PM: platform_pm_suspend+0x0/0x70 returned 0 after 0 usecs [ 152.627814] platform 43000000.syscon: PM: calling platform_pm_suspend+0x0/0x70 @ 1033, parent: bus@f0000:bus@b00000 [ 152.627824] platform 43000000.syscon: PM: platform_pm_suspend+0x0/0x70 returned 0 after 0 usecs [ 152.627837] simple-pm-bus bus@f0000:bus@b00000: PM: calling platform_pm_suspend+0x0/0x70 @ 1033, parent: bus@f0000 [ 152.627847] simple-pm-bus bus@f0000:bus@b00000: PM: platform_pm_suspend+0x0/0x70 returned 0 after 0 usecs [ 152.627916] simple-pm-bus bus@f0000: PM: calling platform_pm_suspend+0x0/0x70 @ 1033, parent: platform [ 152.627929] simple-pm-bus bus@f0000: PM: platform_pm_suspend+0x0/0x70 returned 0 after 0 usecs [ 152.627942] platform timer-cl0-cpu0: PM: calling platform_pm_suspend+0x0/0x70 @ 1033, parent: platform [ 152.627952] platform timer-cl0-cpu0: PM: platform_pm_suspend+0x0/0x70 returned 0 after 0 usecs [ 152.627965] psci-cpuidle-domain firmware:psci: PM: calling platform_pm_suspend+0x0/0x70 @ 1033, parent: platform [ 152.627975] psci-cpuidle-domain firmware:psci: PM: platform_pm_suspend+0x0/0x70 returned 0 after 0 usecs [ 152.627991] regulator regulator.0: PM: calling regulator_suspend+0x0/0x114 @ 1033, parent: reg-dummy [ 152.628003] regulator regulator.0: PM: regulator_suspend+0x0/0x114 returned 0 after 1 usecs [ 152.628016] reg-dummy reg-dummy: PM: calling platform_pm_suspend+0x0/0x70 @ 1033, parent: platform [ 152.628026] reg-dummy reg-dummy: PM: platform_pm_suspend+0x0/0x70 returned 0 after 0 usecs [ 152.628356] ti-udma 485c0000.dma-controller: PM: calling udma_pm_suspend+0x0/0xac @ 1033, parent: bus@f0000:bus@48000000 [ 152.628377] ti-udma 485c0000.dma-controller: PM: udma_pm_suspend+0x0/0xac returned 0 after 6 usecs [ 152.628465] ti-udma 485c0100.dma-controller: PM: calling udma_pm_suspend+0x0/0xac @ 1033, parent: bus@f0000:bus@48000000 [ 152.628487] ti-udma 485c0100.dma-controller: PM: udma_pm_suspend+0x0/0xac returned 0 after 11 usecs [ 152.629293] sdhci-am654 fa00000.mmc: PM: calling genpd_suspend_noirq+0x0/0x20 @ 1033, parent: bus@f0000 [ 152.629364] sdhci-am654 fa00000.mmc: PM: genpd_suspend_noirq+0x0/0x20 returned 0 after 50 usecs [ 152.629396] davinci_gpio 601000.gpio: PM: calling genpd_suspend_noirq+0x0/0x20 @ 1033, parent: bus@f0000 [ 152.629440] davinci_gpio 601000.gpio: PM: genpd_suspend_noirq+0x0/0x20 returned 0 after 31 usecs [ 152.629457] davinci_gpio 600000.gpio: PM: calling genpd_suspend_noirq+0x0/0x20 @ 1033, parent: bus@f0000 [ 152.629500] davinci_gpio 600000.gpio: PM: genpd_suspend_noirq+0x0/0x20 returned 0 after 31 usecs [ 152.629516] omap8250 2800000.serial: PM: calling genpd_suspend_noirq+0x0/0x20 @ 1033, parent: bus@f0000 [ 152.629560] omap8250 2800000.serial: PM: genpd_suspend_noirq+0x0/0x20 returned 0 after 32 usecs [ 152.629678] simple-framebuffer ff700000.framebuffer: PM: calling genpd_suspend_noirq+0x0/0x20 @ 1033, parent: platform [ 152.632990] simple-framebuffer ff700000.framebuffer: PM: genpd_suspend_noirq+0x0/0x20 returned 0 after 3298 usecs [ 152.633039] omap_i2c 20010000.i2c: PM: calling genpd_suspend_noirq+0x0/0x20 @ 1033, parent: bus@f0000 [ 152.633094] omap_i2c 20010000.i2c: PM: genpd_suspend_noirq+0x0/0x20 returned 0 after 41 usecs [ 152.633126] omap_i2c 20000000.i2c: PM: calling genpd_suspend_noirq+0x0/0x20 @ 1033, parent: bus@f0000 [ 152.633138] omap_i2c 20000000.i2c: PM: genpd_suspend_noirq+0x0/0x20 returned 0 after 1 usecs [ 152.633168] ti-sci 44043000.system-controller: PM: calling ti_sci_suspend+0x0/0xdc @ 1033, parent: bus@f0000 [ 152.633305] ti-sci 44043000.system-controller: PM: ti_sci_suspend+0x0/0xdc returned 0 after 123 usecs [ 152.633665] Disabling non-boot CPUs ... [ 152.635580] psci: CPU1 killed (polled 0 ms) [ 152.638013] psci: CPU2 killed (polled 0 ms) [ 152.640717] psci: CPU3 killed (polled 4 ms) [ 152.642148] Enabling non-boot CPUs ... [ 152.642470] Detected VIPT I-cache on CPU1 [ 152.642538] GICv3: CPU1: found redistributor 1 region 0:0x00000000018a0000 [ 152.642581] CPU1: Booted secondary processor 0x0000000001 [0x410fd034] [ 152.643250] CPU1 is up [ 152.643472] Detected VIPT I-cache on CPU2 [ 152.643509] GICv3: CPU2: found redistributor 2 region 0:0x00000000018c0000 [ 152.643534] CPU2: Booted secondary processor 0x0000000002 [0x410fd034] [ 152.644019] CPU2 is up [ 152.644249] Detected VIPT I-cache on CPU3 [ 152.644290] GICv3: CPU3: found redistributor 3 region 0:0x00000000018e0000 [ 152.644319] CPU3: Booted secondary processor 0x0000000003 [0x410fd034] [ 152.644861] CPU3 is up [ 152.644959] phy-gmii-sel 104044.phy: PM: calling phy_gmii_sel_resume_noirq+0x0/0xcc @ 1033, parent: 100000.bus [ 152.644987] phy-gmii-sel 104044.phy: PM: phy_gmii_sel_resume_noirq+0x0/0xcc returned 0 after 1 usecs [ 152.645145] ti-sci 44043000.system-controller: PM: calling ti_sci_resume+0x0/0xac @ 1033, parent: bus@f0000 [ 152.645229] ti-sci 44043000.system-controller: ti_sci_resume: wakeup source: 0x80 [ 152.645236] ti-sci 44043000.system-controller: PM: ti_sci_resume+0x0/0xac returned 0 after 76 usecs [ 152.645263] omap_i2c 20000000.i2c: PM: calling genpd_resume_noirq+0x0/0xe0 @ 1033, parent: bus@f0000 [ 152.645313] omap_i2c 20000000.i2c: PM: genpd_resume_noirq+0x0/0xe0 returned 0 after 36 usecs [ 152.645340] omap_i2c 20010000.i2c: PM: calling genpd_resume_noirq+0x0/0xe0 @ 1033, parent: bus@f0000 [ 152.645384] omap_i2c 20010000.i2c: PM: genpd_resume_noirq+0x0/0xe0 returned 0 after 32 usecs [ 152.645411] simple-framebuffer ff700000.framebuffer: PM: calling genpd_resume_noirq+0x0/0xe0 @ 1033, parent: platform [ 152.645475] simple-framebuffer ff700000.framebuffer: PM: genpd_resume_noirq+0x0/0xe0 returned 0 after 52 usecs [ 152.645551] omap8250 2800000.serial: PM: calling genpd_resume_noirq+0x0/0xe0 @ 1033, parent: bus@f0000 [ 152.645593] omap8250 2800000.serial: PM: genpd_resume_noirq+0x0/0xe0 returned 0 after 31 usecs [ 152.645610] davinci_gpio 600000.gpio: PM: calling genpd_resume_noirq+0x0/0xe0 @ 1033, parent: bus@f0000 [ 152.645653] davinci_gpio 600000.gpio: PM: genpd_resume_noirq+0x0/0xe0 returned 0 after 32 usecs [ 152.645668] davinci_gpio 601000.gpio: PM: calling genpd_resume_noirq+0x0/0xe0 @ 1033, parent: bus@f0000 [ 152.645711] davinci_gpio 601000.gpio: PM: genpd_resume_noirq+0x0/0xe0 returned 0 after 31 usecs [ 152.645756] sdhci-am654 fa00000.mmc: PM: calling genpd_resume_noirq+0x0/0xe0 @ 1033, parent: bus@f0000 [ 152.645812] sdhci-am654 fa00000.mmc: PM: genpd_resume_noirq+0x0/0xe0 returned 0 after 42 usecs [ 152.646405] ti-udma 485c0100.dma-controller: PM: calling udma_pm_resume+0x0/0xb4 @ 1033, parent: bus@f0000:bus@48000000 [ 152.646432] ti-udma 485c0100.dma-controller: PM: udma_pm_resume+0x0/0xb4 returned 0 after 8 usecs [ 152.646484] ti-udma 485c0000.dma-controller: PM: calling udma_pm_resume+0x0/0xb4 @ 1033, parent: bus@f0000:bus@48000000 [ 152.646501] ti-udma 485c0000.dma-controller: PM: udma_pm_resume+0x0/0xb4 returned 0 after 6 usecs [ 152.646663] reg-dummy reg-dummy: PM: calling platform_pm_resume+0x0/0x6c @ 1033, parent: platform [ 152.646677] reg-dummy reg-dummy: PM: platform_pm_resume+0x0/0x6c returned 0 after 0 usecs [ 152.646688] regulator regulator.0: PM: calling regulator_resume+0x0/0x180 @ 1033, parent: reg-dummy [ 152.646700] regulator regulator.0: PM: regulator_resume+0x0/0x180 returned 0 after 1 usecs [ 152.646712] psci-cpuidle-domain firmware:psci: PM: calling platform_pm_resume+0x0/0x6c @ 1033, parent: platform [ 152.646722] psci-cpuidle-domain firmware:psci: PM: platform_pm_resume+0x0/0x6c returned 0 after 0 usecs [ 152.646733] platform timer-cl0-cpu0: PM: calling platform_pm_resume+0x0/0x6c @ 1033, parent: platform [ 152.646741] platform timer-cl0-cpu0: PM: platform_pm_resume+0x0/0x6c returned 0 after 0 usecs [ 152.646752] simple-pm-bus bus@f0000: PM: calling platform_pm_resume+0x0/0x6c @ 1033, parent: platform [ 152.646761] simple-pm-bus bus@f0000: PM: platform_pm_resume+0x0/0x6c returned 0 after 0 usecs [ 152.646771] simple-pm-bus bus@f0000:bus@b00000: PM: calling platform_pm_resume+0x0/0x6c @ 1033, parent: bus@f0000 [ 152.646781] simple-pm-bus bus@f0000:bus@b00000: PM: platform_pm_resume+0x0/0x6c returned 0 after 0 usecs [ 152.646791] platform 43000000.syscon: PM: calling platform_pm_resume+0x0/0x6c @ 1033, parent: bus@f0000:bus@b00000 [ 152.646800] platform 43000000.syscon: PM: platform_pm_resume+0x0/0x6c returned 0 after 0 usecs [ 152.646811] k3-chipinfo 43000014.chipid: PM: calling platform_pm_resume+0x0/0x6c @ 1033, parent: 43000000.syscon [ 152.646820] k3-chipinfo 43000014.chipid: PM: platform_pm_resume+0x0/0x6c returned 0 after 0 usecs [ 152.646830] simple-pm-bus 100000.bus: PM: calling platform_pm_resume+0x0/0x6c @ 1033, parent: bus@f0000 [ 152.646839] simple-pm-bus 100000.bus: PM: platform_pm_resume+0x0/0x6c returned 0 after 0 usecs [ 152.646850] phy-gmii-sel 104044.phy: PM: calling platform_pm_resume+0x0/0x6c @ 1033, parent: 100000.bus [ 152.646859] phy-gmii-sel 104044.phy: PM: platform_pm_resume+0x0/0x6c returned 0 after 0 usecs [ 152.646869] simple-pm-bus bus@f0000:bus@48000000: PM: calling platform_pm_resume+0x0/0x6c @ 1033, parent: bus@f0000 [ 152.646878] simple-pm-bus bus@f0000:bus@48000000: PM: platform_pm_resume+0x0/0x6c returned 0 after 0 usecs [ 152.646889] ti-msgmgr 4d000000.mailbox: PM: calling platform_pm_resume+0x0/0x6c @ 1033, parent: bus@f0000:bus@48000000 [ 152.646921] ti-msgmgr 4d000000.mailbox: PM: platform_pm_resume+0x0/0x6c returned 0 after 22 usecs [ 152.646932] pinctrl-single f4000.pinctrl: PM: calling platform_pm_resume+0x0/0x6c @ 1033, parent: bus@f0000 [ 152.646980] pinctrl-single f4000.pinctrl: PM: platform_pm_resume+0x0/0x6c returned 0 after 38 usecs [ 152.646990] platform l2-cache0: PM: calling platform_pm_resume+0x0/0x6c @ 1033, parent: platform [ 152.646999] platform l2-cache0: PM: platform_pm_resume+0x0/0x6c returned 0 after 0 usecs [ 152.647075] gpio-keys gpio_key: PM: calling platform_pm_resume+0x0/0x6c @ 1033, parent: platform [ 152.647090] gpio-keys gpio_key: PM: platform_pm_resume+0x0/0x6c returned 0 after 5 usecs [ 152.647122] serial8250 serial8250: PM: calling platform_pm_resume+0x0/0x6c @ 1033, parent: platform [ 152.647134] serial8250 serial8250: PM: platform_pm_resume+0x0/0x6c returned 0 after 2 usecs [ 152.647165] platform Fixed MDIO bus.0: PM: calling platform_pm_resume+0x0/0x6c @ 1033, parent: platform [ 152.647174] platform Fixed MDIO bus.0: PM: platform_pm_resume+0x0/0x6c returned 0 after 0 usecs [ 152.647187] ti-cpufreq ti-cpufreq: PM: calling platform_pm_resume+0x0/0x6c @ 1033, parent: platform [ 152.647196] ti-cpufreq ti-cpufreq: PM: platform_pm_resume+0x0/0x6c returned 0 after 0 usecs [ 152.647207] platform psci-cpuidle: PM: calling platform_pm_resume+0x0/0x6c @ 1033, parent: platform [ 152.647216] platform psci-cpuidle: PM: platform_pm_resume+0x0/0x6c returned 0 after 0 usecs [ 152.647227] snd-soc-dummy snd-soc-dummy: PM: calling platform_pm_resume+0x0/0x6c @ 1033, parent: platform [ 152.647237] snd-soc-dummy snd-soc-dummy: PM: platform_pm_resume+0x0/0x6c returned 0 after 0 usecs [ 152.647248] ti-sci 44043000.system-controller: PM: calling platform_pm_resume+0x0/0x6c @ 1033, parent: bus@f0000 [ 152.647258] ti-sci 44043000.system-controller: PM: platform_pm_resume+0x0/0x6c returned 0 after 0 usecs [ 152.647268] ti_sci_pm_domains 44043000.system-controller:power-controller: PM: calling platform_pm_resume+0x0/0x6c @ 1033, parent: 44043000.system-controller [ 152.647279] ti_sci_pm_domains 44043000.system-controller:power-controller: PM: platform_pm_resume+0x0/0x6c returned 0 after 0 usecs [ 152.647290] ti-sci-clk 44043000.system-controller:clock-controller: PM: calling platform_pm_resume+0x0/0x6c @ 1033, parent: 44043000.system-controller [ 152.647300] ti-sci-clk 44043000.system-controller:clock-controller: PM: platform_pm_resume+0x0/0x6c returned 0 after 0 usecs [ 152.647311] platform e030000.watchdog: PM: calling platform_pm_resume+0x0/0x6c @ 1033, parent: bus@f0000 [ 152.647320] platform e030000.watchdog: PM: platform_pm_resume+0x0/0x6c returned 0 after 0 usecs [ 152.647330] platform e020000.watchdog: PM: calling platform_pm_resume+0x0/0x6c @ 1033, parent: bus@f0000 [ 152.647340] platform e020000.watchdog: PM: platform_pm_resume+0x0/0x6c returned 0 after 0 usecs [ 152.647350] platform e010000.watchdog: PM: calling platform_pm_resume+0x0/0x6c @ 1033, parent: bus@f0000 [ 152.647359] platform e010000.watchdog: PM: platform_pm_resume+0x0/0x6c returned 0 after 0 usecs [ 152.647370] platform e000000.watchdog: PM: calling platform_pm_resume+0x0/0x6c @ 1033, parent: bus@f0000 [ 152.647379] platform e000000.watchdog: PM: platform_pm_resume+0x0/0x6c returned 0 after 0 usecs [ 152.647390] ti-sci-reset 44043000.system-controller:reset-controller: PM: calling platform_pm_resume+0x0/0x6c @ 1033, parent: 44043000.system-controller [ 152.647400] ti-sci-reset 44043000.system-controller:reset-controller: PM: platform_pm_resume+0x0/0x6c returned 0 after 0 usecs [ 152.647424] ti-sci-intr bus@f0000:interrupt-controller@a00000: PM: calling platform_pm_resume+0x0/0x6c @ 1033, parent: bus@f0000 [ 152.647434] ti-sci-intr bus@f0000:interrupt-controller@a00000: PM: platform_pm_resume+0x0/0x6c returned 0 after 0 usecs [ 152.647430] pca953x 1-0023: PM: calling pca953x_resume+0x0/0x224 @ 51, parent: i2c-1 [ 152.647445] ti-sci-inta 48000000.interrupt-controller: PM: calling platform_pm_resume+0x0/0x6c @ 1033, parent: bus@f0000:bus@48000000 [ 152.647456] ti-sci-inta 48000000.interrupt-controller: PM: platform_pm_resume+0x0/0x6c returned 0 after 0 usecs [ 152.647468] ti-udma 485c0100.dma-controller: PM: calling platform_pm_resume+0x0/0x6c @ 1033, parent: bus@f0000:bus@48000000 [ 152.647478] ti-udma 485c0100.dma-controller: PM: platform_pm_resume+0x0/0x6c returned 0 after 0 usecs [ 152.647558] ti-udma 485c0000.dma-controller: PM: calling platform_pm_resume+0x0/0x6c @ 1033, parent: bus@f0000:bus@48000000 [ 152.647568] ti-udma 485c0000.dma-controller: PM: platform_pm_resume+0x0/0x6c returned 0 after 0 usecs [ 152.647602] omap8250 2800000.serial: PM: calling omap8250_resume+0x0/0x84 @ 1033, parent: bus@f0000 [ 152.648024] omap8250 2800000.serial: PM: omap8250_resume+0x0/0x84 returned 0 after 406 usecs [ 152.648050] cpufreq-dt cpufreq-dt: PM: calling platform_pm_resume+0x0/0x6c @ 1033, parent: platform [ 152.648062] cpufreq-dt cpufreq-dt: PM: platform_pm_resume+0x0/0x6c returned 0 after 0 usecs [ 152.648074] davinci_gpio 600000.gpio: PM: calling davinci_gpio_resume+0x0/0xc0 @ 1033, parent: bus@f0000 [ 152.648090] davinci_gpio 600000.gpio: PM: davinci_gpio_resume+0x0/0xc0 returned 0 after 4 usecs [ 152.648105] davinci_gpio 601000.gpio: PM: calling davinci_gpio_resume+0x0/0xc0 @ 1033, parent: bus@f0000 [ 152.648118] davinci_gpio 601000.gpio: PM: davinci_gpio_resume+0x0/0xc0 returned 0 after 2 usecs [ 152.648132] gpio-regulator regulator-3: PM: calling platform_pm_resume+0x0/0x6c @ 1033, parent: platform [ 152.648142] gpio-regulator regulator-3: PM: platform_pm_resume+0x0/0x6c returned 0 after 0 usecs [ 152.648153] regulator regulator.1: PM: calling regulator_resume+0x0/0x180 @ 1033, parent: regulator-3 [ 152.648166] regulator regulator.1: PM: regulator_resume+0x0/0x180 returned 0 after 1 usecs [ 152.648193] pca953x 1-0022: PM: calling pca953x_resume+0x0/0x224 @ 1014, parent: i2c-1 [ 152.649502] pca953x 1-0023: PM: pca953x_resume+0x0/0x224 returned 0 after 2053 usecs [ 152.651120] pca953x 1-0022: PM: pca953x_resume+0x0/0x224 returned 0 after 2912 usecs [ 152.651142] reg-fixed-voltage regulator-2: PM: calling platform_pm_resume+0x0/0x6c @ 1033, parent: platform [ 152.651154] reg-fixed-voltage regulator-2: PM: platform_pm_resume+0x0/0x6c returned 0 after 0 usecs [ 152.651165] regulator regulator.2: PM: calling regulator_resume+0x0/0x180 @ 1033, parent: regulator-2 [ 152.651176] regulator regulator.2: PM: regulator_resume+0x0/0x180 returned 0 after 0 usecs [ 152.651190] sdhci-am654 fa00000.mmc: PM: calling pm_runtime_force_resume+0x0/0x104 @ 1033, parent: bus@f0000 [ 152.651323] sdhci-am654 fa00000.mmc: PM: pm_runtime_force_resume+0x0/0x104 returned 0 after 118 usecs [ 152.651339] leds mmc1::: PM: calling led_resume+0x0/0x70 @ 1033, parent: fa00000.mmc [ 152.651355] leds mmc1::: PM: led_resume+0x0/0x70 returned 0 after 0 usecs [ 152.651374] input input0: PM: calling input_dev_resume+0x0/0x50 @ 1033, parent: gpio_key [ 152.651387] input input0: PM: input_dev_resume+0x0/0x50 returned 0 after 0 usecs [ 152.651405] mmcblk mmc1:aaaa: PM: calling mmc_bus_resume+0x0/0x70 @ 1014, parent: mmc1 [ 152.651423] platform 2b300050.target-module: PM: calling platform_pm_resume+0x0/0x6c @ 1033, parent: bus@f0000:bus@b00000 [ 152.651430] mmcblk mmc1:aaaa: PM: mmc_bus_resume+0x0/0x70 returned 0 after 11 usecs [ 152.651433] platform 2b300050.target-module: PM: platform_pm_resume+0x0/0x6c returned 0 after 0 usecs [ 152.651444] platform mcu_gpio_key: PM: calling platform_pm_resume+0x0/0x6c @ 1033, parent: platform [ 152.651453] platform mcu_gpio_key: PM: platform_pm_resume+0x0/0x6c returned 0 after 0 usecs [ 154.931058] OOM killer enabled. [ 154.934200] Restarting tasks ... done. [ 154.939454] random: crng reseeded on system resumption root@am62pxx-evm:~# [ 154.947981] PM: suspend exit Each module will measure it own individual wakeup latency (measured in microseconds), but there is no cumulative latency. This is useful tool when optimizing wakeup latency.

Hi Burak, Thank you for patience. Can you provide your code where you setup the GPIO instance and interrupt? Thanks, Neehar

Hi Alec Thank you very much for your response on the product TDA4VEN-Q1. I would like to learn more about the product TDA4V Entry.

I should add a follow-up: It turns out there are two memory regions that come into play here. The k3-j784s4-rtos-memory-map.dtsi shows (among other things): vision_apps_memory_region: vision-apps-dma-memory@af000000 { compatible = "shared-dma-pool"; reg = ; no-map; }; vision_apps_shared_region: vision_apps_shared-memories { compatible = "dma-heap-carveout"; reg = ; }; The first region is used for the object descriptors used with IPCs. The second is used for TIOVX buffers, e.g. buffers allocated via tivxMemBufferAlloc(). It seems the first region does require manual cache synchronization. But the TI PSDK code does not actually implement it: instead it seems to rely on this region being mapped as uncached. The second region on the other hand does seem to be hardware coherent, so you can map it as cached. One clue to this is that the vision_apps_shared-memories region doesn't have the "no-map" property specified in it. From what I can tell looking at the carveout-heap.c code in the TI Linux kernel, if "no-map" is absent, the memory is treated as cacheable. It turns out that in my OS, both of these regions were being mapped as uncached, which does work, but which leads to accesses to the buffers obtained via tivxMemBufferAlloc() being needlessly slow. Once I realized that only the vision-apps-dma-memory needed to be mapped as uncached, I enabled caching and coherency for the other region and performance got much better. I think this means that the answer in the other thread was misleading: it implies that you need to use explicit software cache synchronization everywhere in the TIOVX code, but it seems that's only true for the vision-apps-dma-memory region. -Bill

Hi Ramesh, As I remember, the external memory via EMIF can only be accessed in 16-Bit word, but the .text is the code section which need byte addressable. May be that is the issue. Can you put the .text into the internal memory and data sections in external memory? Best regards, Ming

Hi, You suppose to do it in /media/ /rootfs when sd card inserted in pc. Or you can do once after booting of board. Best Regards, Sudheer

HI, [quote userid="565381" url="~/support/processors-group/processors/f/processors-forum/1383901/am69-main_mdio-association-with-cpsw/5301462#5301462"] With the following dts excerpt I'm able to communicate with the phy (main_cpsw0_port2 using cpsw0_port2_phy4). I have to enable main_cpsw1_port1 AND main_phy0 to let cpsw0_port2_phy4 working. [/quote] Yes, this how suppose to use other cpsw instance mdio. [quote userid="565381" url="~/support/processors-group/processors/f/processors-forum/1383901/am69-main_mdio-association-with-cpsw/5301462#5301462"] I do not want to enable hardware that is not present onboard. In particular I have an additional ethernet (eth2) which is not desirable. Have you any suggestions? [/quote] If you are enabling other cpsw instance, mdio is not accessible. So, you must enable other cpsw instance as well even though you are not using cpsw ports. If you don't want to enable other cpsw instance for accessing mdio, you need to redesign your H/W with proper mdio connected to phy corresponding to cpsw instance being used. You have to go with above s/w workaround i.e. changing device tree file or make H/W change. Best Regards, Sudheer

Hi, I don't think any other possibility for this. Let me check internally and get back to you soon. Best Regards, Sudheer

Part Number: TDA4VM Tool/software: Hi, May I ask how to implement custom node functionality using other cores instead of A72? For example, YUV image cropping, matrix operations, etc? SDK:08.06.00.12 vision_apps app_multi_cam_codec Regards， Yueqian

Hi, We can use a timer. And I have done tests, we can use DMTIMER0~DMTIMER7. But now the problem is that I have to rely on some libraries (lib\freertos\ti.osal.ae71, lib\c7x\ti.kernel.freertos.ae71). I don't know what parameters are configured for interrupts in these libraries. Thanks