AM62A7: Suspend/Deep Sleep support on the 09.01 SDK

Thanks. It looks like I got lost in the edgeai documentation and didn't know there was a separate processor-sdk-linux area.  I will give this a shot and get back to you.

Trying the rtcwake command is failing to suspend.  Looks like there is something in the message box?

root@mitysom-am62ax:~# rtcwake -m mem -s 10
rtcwake: wakeup from "mem" using /dev/rtc0 at Fri Apr 12 20:10:36 2024
[  164.069192] PM: suspend entry (deep)
[  164.073234] Filesystems sync: 0.000 seconds
[  164.096130] remoteproc remoteproc0: stopped remote processor 79000000.r5f
[  164.103920] Freezing user space processes
[  164.109597] Freezing user space processes completed (elapsed 0.001 seconds)
[  164.116650] OOM killer disabled.
[  164.119887] Freezing remaining freezable tasks
[  164.125501] Freezing remaining freezable tasks completed (elapsed 0.001 seconds)
[  164.132904] printk: Suspending console(s) (use no_console_suspend to debug)
[  164.157222] am65-cpsw-nuss 8000000.ethernet eth0: Link is Down
[  164.160443] omap8250 2800000.serial: PM domain pd:146 will not be powered off
[  164.161875] omap-mailbox 29010000.mailbox: fifo 2 has unexpected unread messages
[  164.161894] omap-mailbox 29010000.mailbox: PM: dpm_run_callback(): platform_pm_suspend+0x0/0x7c returns -16
[  164.161921] omap-mailbox 29010000.mailbox: PM: failed to suspend: error -16
[  164.161933] PM: Some devices failed to suspend, or early wake event detected
[  164.165490] am65-cpsw-nuss 8000000.ethernet: set new flow-id-base 19
[  164.176044] am65-cpsw-nuss 8000000.ethernet eth0: PHY [8000f00.mdio:03] driver [Microchip KSZ9131 Gigabit PHY] (irq=POLL)
[  164.176074] am65-cpsw-nuss 8000000.ethernet eth0: configuring for phy/rgmii-rxid link mode
[  164.185424] am65-cpsw-nuss 8000000.ethernet eth1: PHY [8000f00.mdio:04] driver [Microchip KSZ9131 Gigabit PHY] (irq=POLL)
[  164.185438] am65-cpsw-nuss 8000000.ethernet eth1: configuring for phy/rgmii-rxid link mode
[  164.190396] usb-conn-gpio usb_microb_connector: repeated role: device
[  164.506884] OOM killer enabled.
[  164.510027] Restarting tasks ... done.
[  164.515836] random: crng reseeded on system resumption
[  164.521189] remoteproc remoteproc0: powering up 79000000.r5f
[  164.526967] remoteproc remoteproc0: Booting fw image am62a-mcu-r5f0_0-fw, size 52148
[  164.536139] rproc-virtio rproc-virtio.2.auto: assigned reserved memory node r5f-dma-memory@9b800000
[  164.546797] virtio_rpmsg_bus virtio0: rpmsg host is online
[  164.547471] virtio_rpmsg_bus virtio0: creating channel ti.ipc4.ping-pong addr 0xd
[  164.552546] rproc-virtio rproc-virtio.2.auto: registered virtio0 (type 7)
[  164.560374] virtio_rpmsg_bus virtio0: creating channel rpmsg_chrdev addr 0xe
[  164.567022] remoteproc remoteproc0: remote processor 79000000.r5f is now up
[  164.581074] PM: suspend exit

Hi Jonathan,

Are you using a TI EVM or custom hardware?

Did you use any of the Pre-Built images in the SDK? Only the Default Image will contain the M4 firmware.

Thanks,

Anshu

Sorry for the slow reply, the e2e was down last friday and I couldn't post a response.

I am using our custom hardware.  I can retest this on the evm, though I'm unsure how this would affect the mailboxes.

I tested this using the edgeai and the default tisdk image, and got the same error on both.

Testing the rtcwake on the evm using the edgeai image did work.

Any idea which device this mailbox belongs to? 29010000.mailbox

root@am62axx-evm:/opt/edgeai-gst-apps# rtcwake -s 10 -m mem -d rtc1
rtcwake: wakeup from "mem" using rtc1 at Thu Jan  1 00:17:33 1970
[ 1091.563681] PM: suspend entry (deep)
[ 1091.567410] Filesystems sync: 0.000 seconds
[ 1091.596693] remoteproc remoteproc1: stopped remote processor 79000000.r5f
[ 1091.604020] Freezing user space processes
[ 1091.609638] Freezing user space processes completed (elapsed 0.001 seconds)
[ 1091.616637] OOM killer disabled.
[ 1091.619856] Freezing remaining freezable tasks
[ 1091.625579] Freezing remaining freezable tasks completed (elapsed 0.001 seconds)
[ 1091.633066] printk: Suspending console(s) (use no_console_suspend to debug)
[ 1091.652615] am65-cpsw-nuss 8000000.ethernet eth0: Link is Down
[ 1091.656160] omap8250 2800000.serial: PM domain pd:146 will not be powered off
[ 1091.662506] Disabling non-boot CPUs ...
[ 1091.663901] psci: CPU1 killed (polled 4 ms)
[ 1091.666756] psci: CPU2 killed (polled 0 ms)
[ 1091.669313] psci: CPU3 killed (polled 0 ms)
[ 1091.670128] Enabling non-boot CPUs ...
[ 1091.670441] Detected VIPT I-cache on CPU1
[ 1091.670509] GICv3: CPU1: found redistributor 1 region 0:0x00000000018a0000
[ 1091.670560] CPU1: Booted secondary processor 0x0000000001 [0x410fd034]
[ 1091.671304] CPU1 is up
[ 1091.671528] Detected VIPT I-cache on CPU2
[ 1091.671568] GICv3: CPU2: found redistributor 2 region 0:0x00000000018c0000
[ 1091.671603] CPU2: Booted secondary processor 0x0000000002 [0x410fd034]
[ 1091.672221] CPU2 is up
[ 1091.672460] Detected VIPT I-cache on CPU3
[ 1091.672504] GICv3: CPU3: found redistributor 3 region 0:0x00000000018e0000
[ 1091.672547] CPU3: Booted secondary processor 0x0000000003 [0x410fd034]
[ 1091.673217] CPU3 is up
[ 1091.673696] ti-sci 44043000.system-controller: ti_sci_resume: wakeup source: 0x50
[ 1091.682217] am65-cpsw-nuss 8000000.ethernet: set new flow-id-base 19
[ 1091.696085] am65-cpsw-nuss 8000000.ethernet eth0: PHY [8000f00.mdio:00] driver [TI DP83867] (irq=POLL)
[ 1091.696109] am65-cpsw-nuss 8000000.ethernet eth0: configuring for phy/rgmii-rxid link mode
[ 1091.949447] OOM killer enabled.
[ 1091.952591] Restarting tasks ... done.
[ 1091.958696] random: crng reseeded on system resumption
[ 1091.963978] remoteproc remoteproc1: powering up 79000000.r5f
[ 1091.969769] remoteproc remoteproc1: Booting fw image am62a-mcu-r5f0_0-fw, size 52148
[ 1091.983205] rproc-virtio rproc-virtio.7.auto: assigned reserved memory node r5f-dma-memory@9b800000
[ 1091.993062] virtio_rpmsg_bus virtio1: rpmsg host is online
[ 1091.993492] virtio_rpmsg_bus virtio1: creating channel ti.ipc4.ping-pong addr 0xd
[ 1091.998692] rproc-virtio rproc-virtio.7.auto: registered virtio1 (type 7)
[ 1092.006421] virtio_rpmsg_bus virtio1: creating channel rpmsg_chrdev addr 0xe
[ 1092.012946] remoteproc remoteproc1: remote processor 79000000.r5f is now up
[ 1092.027113] PM: suspend exit

Hi Jonathan,

29010000.mailbox refers to mailbox0_cluster1: https://git.ti.com/cgit/ti-linux-kernel/ti-linux-kernel/tree/arch/arm64/boot/dts/ti/k3-am62a-main.dtsi?h=ti-linux-6.1.y#n779

We can try to disable the node in the device tree. I'll get back to you on a device tree patch to do this.

Best Regards,

Anshu

This change allowed me to put my device into sleep.

Is this "Hack" okay to keep in the kernel? Will it affect the DSP demos?

Hi Jonathan,

Since the hack only changes the mailbox FIFOs, I wouldn't expect any impact on other functionalities if left enabled.

Thanks,

Anshu