Failed to resume from the SC7 state

Hi,

I am using JetPack 6.2 (Jetson Linux 36.5) on the Jetson Orin NX 8GB Developer Kit. Except for the NVMe SSD, no other peripherals are connected. After resuming from the SC7 state (entered by running systemctl suspend) using the power button, the system reports the following error and then hangs indefinitely.
NOTICE: tegra_soc_pwr_domain_on_finish: exited SC7 successfully. Entering normal world.
OER Exception reas= ydrom*=x**001
ERROR: RAS Uncorrectable Error in IOB, base=0xe010000:
ERROR: Status = 0xe4000612
ERROR: SERR = Error response from slave: 0x12
ERROR: IERR = CBB Interface Error: 0x6
ERROR: MISC0 = 0x44062040
ERROR: MISC1 = 0x6154860000000000
ERROR: MISC2 = 0x0
ERROR: MISC3 = 0x0
ERROR: ADDR = 0x8000003228003000
ERROR: **************************************
ERROR: sdei_dispatch_event returned -1
ERROR: **************************************
ERROR: RAS Uncorrectable Error in ACI, base=0xe01a000:
ERROR: Status = 0x60000d04
ERROR: SERR = Assertion failure: 0x4
ERROR: IERR = SNOC Write Error: 0xd
ERROR: **************************************
ERROR: sdei_dispatch_event returned -1
syndrome=0xbe002411eacor=
ERROR: **************************************
ERROR: RAS Uncorrectable Error in IOB, base=0xe010000:
ERROR: Status = 0xe4000612
ERROR: SERR = Error response from slave: 0x12
ERROR: IERR = CBB Interface Error: 0x6
ERROR: MISC0 = 0x44062040
ERROR: MISC1 = 0x4ed4860000000000
ERROR: MISC2 = 0x0
ERROR: MISC3 = 0x0
ERROR: ADDR = 0x800000003e080800
ERROR: **************************************
ERROR: **************************************
ERROR: RAS Uncorrectable Error in ACI, base=0xe01a000:
ERROR: Status = 0x60000d04
ERROR: SERR = Assertion failure: 0x4
ERROR: IERR = SNOC Write Error: 0xd
ERROR: **********************************
ERROR: sdei_dispatch_event returned -1
ERROR: Powering off core
Admin Task Init
Admin Task Init complete
Print Task Init
RM Task Init
Admin Task Started
DCE SC7 SHA Disabled
RM Task Started
RM Task Running
Print Task Started
Print Task Running
DCE: FW Boot Complete
Admin Task Running

SC7 Resume Failure.log (135.3 KB)
Full logs.

Hi nano.wang,

R36.5 should be included in Jetpack 6.2.2 rather than Jetpack 6.2.

How did you setup the developer kit? With SDK manager?
How about the failed rate for the RAS error during resume?

Do you connect any devices(display, USB, PCIe, SPI, I2C..etc) on the devkit currently?

Hi Kevin,

Yes, it’s Jetpack6.2.2.

I flashed the developer kit using the l4t_initrd_flash.sh script rather than SDK Manager.

The issue is 100% reproducible. The RAS error occurs every time the system resumes from the SC7 state.

The only connected peripherals are the NVMe SSD and a serial console for logging. No display, USB devices, additional PCIe devices, SPI devices, I2C devices, or any other external peripherals are connected.

As you are using the Orin Nano devkit(P3768) with Orin NX 8GB module, please check if there’s the similar issue if you flash it through SDK manager.

We will also validate this on our local setup.

Can the hang condition only be recovered by resetting the board? (i.e. power cycle manually)

Hi Kevin,

The version I burned using SDK Manager indeed does not have this issue. The version I made myself has this issue. I mainly made some kernel modifications and adapted the Wi-Fi 7 module, but during verification, the module was not plugged into the device.
When a problem occurs, the system cannot be restored by pressing the power button, which should be powered off and then powered on.

Could you just flash with the official BSP package (w/o custom kernel customization)to verify?

Or disable it from device tree to check?