On Mon, 2018-12-03 at 10:30 -0800, Dan Williams wrote:
> A "short" ARS (address range scrub) instructs the platform firmware
> to
> return known errors. In contrast, a "long" ARS instructs platform
> firmware to arrange every data address on the DIMM to be read /
> checked
> for poisoned data.
>
> The conversion of the flags in commit d3abaf43bab8 "acpi, nfit: Fix
> Address Range Scrub completion tracking", changed the meaning of
> passing
> '0' to acpi_nfit_ars_rescan(). Previously '0' meant "not short", now
> '0'
> is ARS_REQ_SHORT. Pass ARS_REQ_LONG to restore the expected scrub-
> type
> behavior of user-initiated ARS sessions.
>
> Fixes: d3abaf43bab8 ("acpi, nfit: Fix Address Range Scrub completion
> tracking")
> Reported-by: Jacek Zloch <jacek.zloch@xxxxxxxxx>
> Cc: Vishal Verma <vishal.l.verma@xxxxxxxxx>
> Signed-off-by: Dan Williams <dan.j.williams@xxxxxxxxx>
> ---
> drivers/acpi/nfit/core.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
Looks good,
Reviewed-by: Vishal Verma <vishal.l.verma@xxxxxxxxx>
>
> diff --git a/drivers/acpi/nfit/core.c b/drivers/acpi/nfit/core.c
> index 14d9f5bea015..5912d30020c7 100644
> --- a/drivers/acpi/nfit/core.c
> +++ b/drivers/acpi/nfit/core.c
> @@ -1308,7 +1308,7 @@ static ssize_t scrub_store(struct device *dev,
> if (nd_desc) {
> struct acpi_nfit_desc *acpi_desc =
> to_acpi_desc(nd_desc);
>
> - rc = acpi_nfit_ars_rescan(acpi_desc, 0);
> + rc = acpi_nfit_ars_rescan(acpi_desc, ARS_REQ_LONG);
> }
> device_unlock(dev);
> if (rc)
>
