9:30 a.m.
On Fri, 2016-02-12 at 11:41 -0800, Dan Williams wrote:
On Tue, Feb 2, 2016 at 10:55 AM, Toshi Kani
<toshi.kani(a)hpe.com> wrote:
> Change acpi_nfit_register_region() to call iomem_set_desc() with
> IORES_DESC_PERSISTENT_MEMORY for NFIT_SPA_PM ranges found in ACPI
> NFIT table.
>
> When FW sets E820_PMEM in e820 and EFI_PERSISTENT_MEMORY in EFI,
> this code simply sets PMEM type again to "Persistent Memory" entries
> in the iomem table. When FW sets reserved type for persistent
> memory ranges, it sets PMEM type to "reserved" entries covering
> PMEM ranges.
>
> This allows the EINJ driver, which calls region_intersects() with
> IORES_DESC_PERSISTENT_MEMORY to check persistent memory ranges,
> to work continuously even if FW sets reserved type to persistent
> memory in e820 and EFI.
>
> Signed-off-by: Toshi Kani <toshi.kani(a)hpe.com>
> Cc: Rafael J. Wysocki <rjw(a)rjwysocki.net>
> Cc: Dan Williams <dan.j.williams(a)intel.com>
> Cc: Ingo Molnar <mingo(a)kernel.org>
> Cc: Borislav Petkov <bp(a)suse.de>
> Cc: Andrew Morton <akpm(a)linux-foundation.org>
> ---
> drivers/acpi/nfit.c | 6 ++++++
> 1 file changed, 6 insertions(+)
>
> diff --git a/drivers/acpi/nfit.c b/drivers/acpi/nfit.c
> index ad6d8c6..add04f0 100644
> --- a/drivers/acpi/nfit.c
> +++ b/drivers/acpi/nfit.c
> @@ -1781,6 +1781,12 @@ static int acpi_nfit_register_region(struct
> acpi_nfit_desc *acpi_desc,
>
> nvdimm_bus = acpi_desc->nvdimm_bus;
> if (nfit_spa_type(spa) == NFIT_SPA_PM) {
> + rc = iomem_set_desc(spa->address, spa->length,
> + IORES_DESC_PERSISTENT_MEMORY);
> + if (rc)
> + dev_dbg(acpi_desc->dev,
> + "error setting iomem desc: %d\n", rc);
> +
Hmm, if we set the type on driver load, should we clear the type on
driver unload?
I think this type update should stay for the life-cycle of this iomem entry
itself since this range is PMEM even after the driver is unloaded. This is
an extension of the boot-time iomem table initialization from e820/EFI,
which allows ACPI to set a correct type. This is independent from driver's
resource allocations.
Actually it might be more straightforward to specify a type at
request_region() time. That way it gets released at release_region().
We're already setting a resource name at request_region time, adding a
type annotation at the time seems appropriate.
I first considered simply setting "namespaceX.X" as PMEM. However,
region_intersects() and its friends only check the top-level entries, not
their children, of the iomem table. And I think a child should have the
same type as the parent as I fixed it in patch 1/3.
Thanks,
-Toshi