SMMU faults after upgrade to 8.0 release

Autonomous Vehicles DRIVE AGX Xavier DRIVE AGX Xavier General
1

After flashing the 8.0 release, I continually see SMMU fault messages on the serial console for both the Xaviers - A & B.

The console continually prints -

�����62050027|HV/c0: �����SMMU instance: 0
�����62050911|HV/c0: �����CB46 SMMU fault. StreamID: 42
�����62051821|HV/c0: �����FSR: 0x80000402; FAR: 0xbff15dc0; FSYND0: 0x220001
�����62053063|HV/c0: �����(255) csr_rcer: EMEM address decode error
�����62054132|HV/c0: �����  status = 0x200640d2; addr = 0xffffffff00
�����62055233|HV/c0: �����  secure: yes, access-type: read
�����62059085|HV/c0: �����vpr base=8:70000000, size=80, ctrl=1, override:(a01a8340, fcee10c1, 1, 0)
�����62068171|HV/c0: �����(255) csw_rcew: MC request violates VPR requirements
�����62075402|HV/c0: �����  status = 0x0ff740d3; addr = 0xffffffff00
�����62081753|HV/c0: �����  secure: yes, access-type: write
�����62087312|HV/c0: �����(255) csr_rcer: non secure access to secure region
�����62094367|HV/c0: �����  status = 0x200640d2; addr = 0xffffffff00
�����62100717|HV/c0: �����  secure: yes, access-type: read
�����62106188|HV/c0: �����(255) csr_dla1rda: EMEM GSC access violation
�����62112713|HV/c0: �����  status = 0x7c800ac2; addr = 0xc807c40440
�����62119064|HV/c0: �����  secure: no, access-type: read
�����62124446|HV/c0: �����SMMU instance: 1
�����62128529|HV/c0: �����CB46 SMMU fault. StreamID: 42
�����62133706|HV/c0: �����FSR: 0x402; FAR: 0xbff00100; FSYND0: 0x1c0011

The above set of messages continually repeat.

Any idea what’s going on? How to get the board in a sensible state again?

Thanks,
Punit

2

Dear punitagrawal,

Did you flash the board via sdkmanager?
Could you let me know the board Aurix FW version like below?

shell> version
Info: Executing cmd: version, argc: 0, args:
SW Version: DRIVE-V5.0.13-E3550-EB-Aurix-ForHyperion-3.01.05
Compilation date: Nov 12 2018, 12:50:16
Command Executed

3

I installed and compiled the image using sdkmanager but flashed it using -

./pdk_flash.sh -b e3550b03-t194-es

This is another doubt I have. The serial number at the front starts with - “E3550-B03-”. Am I using the correct board?

Here is the Aurix FW version as requested -

shell> version
Info: Executing cmd: version, argc: 0, args:
SW Version: DRIVE-V5.0.10-E3550-EB-Aurix-With3LSS-ForHyperion-3.00.07
Compilation date: Jul 20 2018, 09:08:33
Command Executed
4

Also, I haven’t touched the FW on the aurix during the upgrade to 8.0 release. I didn’t have to flip the switch to “PRG” from the “RUN” position.

5

Dear punitagrawal,

Could you please try to flash the board using the latest SDK manager(Beta 0.9.9.2351)?

6

Hi,

I think I have figured out what went wrong with upgrading the images on the board using -

./pdk_flash.sh -b e3550b03-t194-es

The environment where the command was run was missing ‘rsync’ which leads to corrupted images being flashed to the board. This in turn leads to corrupted images being flashed and the subsequent SMMU faults in $SUBJECT.

Installing ‘rsync’ and flashing again (after figuring out the missing USB devices in the other thread) fixed the problem.

One way to make the flashing process more robust would be to check for the presence of ‘rsync’ in nvimagegen and bail out with a helpful message if it’s missing.

Thanks.