[4/5] EDAC, dsm_edac: Wrap ACPI DSM methods for address translation

["Luck, Tony" <tony.luck@intel.com>]

> We can put it someplace else if you have a better suggestion.

Any thoughts on this?

I'll note that other users of ACPI DSM calls are scattered around the kernel in the drivers that use them:

$ git grep -l acpi_evaluate_dsm
drivers/acpi/acpi_extlog.c
drivers/acpi/nfit/core.c
drivers/acpi/sleep.c
drivers/acpi/utils.c
drivers/acpi/x86/apple.c
drivers/char/tpm/tpm_crb.c
drivers/char/tpm/tpm_ppi.c
drivers/gpu/drm/i915/intel_acpi.c
drivers/gpu/drm/nouveau/nouveau_acpi.c
drivers/gpu/drm/nouveau/nvkm/subdev/mxm/base.c
drivers/hid/i2c-hid/i2c-hid.c
drivers/iommu/dmar.c
drivers/mmc/host/sdhci-acpi.c
drivers/mmc/host/sdhci-pci-core.c
drivers/net/ethernet/hisilicon/hns/hns_dsaf_misc.c
drivers/pci/pci-acpi.c
drivers/pci/pci-label.c
drivers/platform/x86/intel-hid.c
drivers/usb/dwc3/dwc3-pci.c
drivers/usb/host/xhci-pci.c
drivers/usb/typec/typec_wcove.c
drivers/usb/typec/ucsi/ucsi_acpi.c
include/acpi/acpi_bus.h
include/linux/acpi.h
sound/soc/intel/skylake/skl-nhlt.c

We originally coded the DSM bits directly into the driver. But during
internal review it was obvious that we'd be copy/pasting that code
into future drivers, so better to somehow separate it from the driver.

The registration/callback methods we employed in this patch series
aren't required. Happy to adopt a cleaner/simpler solution.

-Tony