Appendix D: Platform-Dependent Workarounds#

Some Grace platforms require temporary (or permanent) alterations to their configurations to work around known issues, such as hardware errata. These workarounds are described in the following sections by the corresponding Grace platform.

D.1 All Grace Platforms#

Starting with linux-nvidia-64k-6.11 and later versions, the ffa-module is integrated into the kernel. The following kernel messages are displayed:
```
[   33.757800] ARM FF-A: Driver version 1.2
[   33.766080] ARM FF-A: Firmware version 1.1 found
[   33.775832] GICv3: [Firmware Bug]: Illegal GSI5 translation request
[   33.788966] ARM FF-A: Failed to create IRQ mapping!
```
The GICv3 Firmware Bug and IRQ mapping failure can be safely ignored. The ffa-module will still load successfully and function normally.
Starting with linux-nvidia-64k-6.11 and later versions, the Generic Hardware Error Source (GHES) driver has been modified to respect the panic= kernel parameter instead of overriding it to trigger a system reboot after fatal hardware error detection. If panic= is disabled, the error will be logged to the console, but the system will not automatically reboot.

To restore previous kernel behavior, the administrator must complete one of the following tasks:
- Set panic=<value> in kernel command line.
- Modify /proc/sys/kernel/panic value.
- Modify kernel.panic sysctl parameter.
The Memory System Resource Paritioning and Monitoring (MPAM) has been enabled in linux-nvidia kernel versions 6.14.0-1012 and later. This feature might be incompatible with earlier firmware versions, and you might see the following errors at boot time:
```
[   15.643073] mpam:__mpam_irq_handler: error irq from msc:1 'Req_PARTID_Range', partid:43, pmg: 0, ris: 0
[   15.652696] mpam:__mpam_irq_handler: error irq from msc:1 'Req_PARTID_Range', partid:43, pmg: 0, ris: 0
[   15.662305] mpam:__mpam_irq_handler: error irq from msc:1 'Req_PARTID_Range', partid:43, pmg: 0, ris: 0
[   15.678235] mpam:__mpam_irq_handler: error irq from msc:1 'Req_PARTID_Range', partid:43, pmg: 0, ris: 0
[   15.696000] mpam:__mpam_irq_handler: error irq from msc:1 'Req_PARTID_Range', partid:43, pmg: 0, ris: 0
[   15.708002] mpam:__mpam_irq_handler: error irq from msc:1 'Req_PARTID_Range', partid:43, pmg: 0, ris: 0
[   15.718295] Unable to handle kernel paging request at virtual address
```
NVIDIA recommends that you update the system firmware to the latest version. If the system still cannot boot after the firmware update, to avoid loading MPAM during boot, deploy the following workaround:
- To temporarily deploy this workaround for the duration of the current boot:
  1. During boot, stop at the grub menu, select the boot entry, and press the e key to edit the entry.
  2. Append initcall_blacklist=mpam_msc_driver_init to the end of the list of kernel boot parameters.
  3. Boot the entry by pressing Ctrl-X or pressing F10.
- To permanently deploy this workaround so that it is always active at boot time:
  1. With administrative privileges, edit /etc/default/grub.
    1. Append initcall_blacklist=mpam_msc_driver_init to the end of the list of kernel boot parameters specified in GRUB_CMDLINE_LINUX_DEFAULT.
    2. Save the file and exit the editor.
  2. Run the following command to update GRUB.
    sudo update-grub
  3. Reboot the system.
- To permanently remove this workaround so that it is not active at boot time:
  1. With administrative privileges, edit /etc/default/grub.
    1. Remove initcall_blacklist=mpam_msc_driver_init from the list of kernel boot parameters in GRUB_CMDLINE_LINUX_DEFAULT.
    2. Save the file and exit the editor.
  2. Run the following command to update GRUB:
    sudo update-grub
  3. Reboot the system.
- To verify the presence of the workaround:
  - Evaluate the kernel boot parameters set for the current boot.
    cat /proc/cmdline | grep mpam_msc_driver_init
    When nothing is returned, the workaround is not active.