If the user has < 65 cores and passes in an affinitized range for the 0th CPU group, if valid, honor it.

[mrsharm]

Summary

For cases where the user is running on a machine with < 65 cores and has passed in a GCHeapAffinitizeRanges for just the 0th CPU group, apply the affinitized range rather than returning a CLR_E_GC_BAD_AFFINITY_CONFIG_FORMAT. And fail if:

• Any other CPU group is passed in if the machine has < 65 cores.
• If a user passes in a range that's > max processors available if the machine has < 65 cores and 0 is passed in as the CPU group.

NOTE: If both the Affinitized Range and the Affinitized Mask are passed, the Affinitized Mask is used if there are no CPU groups available on the machine.

Implementation Details

1. The check to see if we can bypass the Affinitized Ranges is simply to check if the user passed in the 0th CPU Group and if CPU Groups are not available.
2. If CPU Groups are not available, the affinitization will be based on the Affinitized Mask; to facilitate this cleanly, we are applying the CPU Group Affinitized Range on the Affinitized Mask so that down the road, it'll be successfully used.

[ghost]

Tagging subscribers to this area: @dotnet/gc
See info in area-owners.md if you want to be subscribed.

Issue Details

---

Summary

For cases where the user is running on a machine with < 65 cores and has passed in a GCHeapAffinitizeRanges for just the 0th CPU group, apply the affinitized range rather than returning a CLR_E_GC_BAD_AFFINITY_CONFIG_FORMAT. And fail if:

• Any other CPU group is passed in if the machine has < 65 cores.
• If a user passes in a range that's > max processors available if the machine has < 65 cores and 0 is passed in as the CPU group.

NOTE: If both the Affinitized Range and the Affinitized Mask are passed, the Affinitized Mask is used if there are no CPU groups available on the machine.

Implementation Details

1. The check to see if we can bypass the Affinitized Ranges is simply to check if the user passed in the 0th CPU Group and if CPU Groups are not available.
2. If CPU Groups are not available, the affinitization will be based on the Affinitized Mask; to facilitate this cleanly, we are applying the CPU Group Affinitized Range on the Affinitized Mask so that down the road, it'll be successfully used.

Author:	mrsharm
Assignees:	mrsharm
Labels:	area-GC-coreclr
Milestone:	-

[cshung]

Looks good to me, thanks for getting it fixed!

[Maoni0]

@janvorli would you like to take a look?

[janvorli]

LGTM modulo the nits.

[mrsharm]

Thanks for the feedback, @janvorli! I have incorporated them all and will merge these changes once the CI completes.