Re: [PATCH 2/2 v6] mm/mempolicy: Don't create weight sysfs for memoryless nodes

[Gregory Price]

On Mon, Mar 10, 2025 at 09:26:48PM +0900, Honggyu Kim wrote:
> Yeah, the proximity domain detects the node correctly as follows in dmesg.
> 
>  [  0.009915] ACPI: SRAT: Node 0 PXM 0 [mem 0x00000000-0x7fffffff]
>  [  0.009917] ACPI: SRAT: Node 0 PXM 0 [mem 0x100000000-0x207fffffff]
>  [  0.009919] ACPI: SRAT: Node 1 PXM 1 [mem 0x60f80000000-0x64f7fffffff]
>  [  0.009924] ACPI: SRAT: Node 2 PXM 6 [mem 0x2080000000-0x807fffffff] hotplug
>  [  0.009925] ACPI: SRAT: Node 3 PXM 7 [mem 0x64f80000000-0x6cf7fffffff] hotplug
> 
> It is printed even before CXL detection.
> 

I wrote a some documentation on some example configurations last friday

https://lore.kernel.org/linux-mm/Z226PG9t-Ih7fJDL@gourry-fedora-PF4VCD3F/T/#m2780e47df7f0962a79182502afc99843bb046205

This isn't exhaustive, but as I said with Rakie, I think your
configuration is probably ok - if slightly confusing.

What I'm going to guess is happening is you have 1 CFMWS per device that
do not have matching SRAT entries, and then the CFMWS covering these:

>  [  0.009924] ACPI: SRAT: Node 2 PXM 6 [mem 0x2080000000-0x807fffffff] hotplug
>  [  0.009925] ACPI: SRAT: Node 3 PXM 7 [mem 0x64f80000000-0x6cf7fffffff] hotplug

have interleave set up