Thanks Kevin for the comments.
>>
>> @@ -46,6 +51,7 @@ struct nvgrace_gpu_pci_core_device {
>> struct mem_region resmem;
>> /* Lock to control device memory kernel mapping */
>> struct mutex remap_lock;
>> + bool has_mig_hw_bug_fix;
>
> Is 'has_mig_hw_bug" clearer given GB+ hardware should all inherit
> the fix anyway?
IIUC, are you suggesting an inverted implementation? i.e. use
has_mig_hw_bug as the struct member with the semantics
!has_mig_hw_bug_fix?
>>
>> if (ops == &nvgrace_gpu_pci_ops) {
>> + nvdev->has_mig_hw_bug_fix =
>> nvgrace_gpu_has_mig_hw_bug_fix(pdev);
>> +
>
> Move it into nvgrace_gpu_init_nvdev_struct() which has plenty
> of information to help understand that line.
Ack, will update in the next version.
