ACPI / scan: Prefer devices without _HID for _ADR matching

Commit c2a6bbaf0c5f (ACPI / scan: Prefer devices without _HID/_CID
for _ADR matching) added a list_empty(&adev->pnp.ids) check to
find_child_checks() so as to catch situations in which the ACPI
core attempts to decode _ADR for a device having a _HID too which
is strictly against the spec.  However, it overlooked the fact that
the adev->pnp.ids list for the devices taken into account by
find_child_checks() may contain device IDs set internally by the
kernel, like "LNXVIDEO" (thanks to Zhang Rui for that realization),
and it broke the enumeration of those devices as a result.

To unbreak it, replace the overly coarse grained list_empty()
check with a much more precise check against the pnp.type.platform_id
flag which is only set for devices having a _HID (that's how it
should be done from the start, as having both _ADR and _CID is
actually permitted).

Fixes: c2a6bbaf0c5f (ACPI / scan: Prefer devices without _HID/_CID for _ADR matching)
Link: https://bugzilla.kernel.org/show_bug.cgi?id=194889
Reported-and-tested-by: Mike <mike@mikewilson.me.uk>
Tested-by: Hans de Goede <hdegoede@redhat.com>
Cc: 4.10+ <stable@vger.kernel.org> # 4.10+
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>

diff --git a/drivers/acpi/glue.c b/drivers/acpi/glue.c
index fb19e1c..edc8663 100644 (file)
--- a/drivers/acpi/glue.c
+++ b/drivers/acpi/glue.c
@@ -99,13 +99,13 @@ static int find_child_checks(struct acpi_device *adev, bool check_children)
                return -ENODEV;
 
        /*
-        * If the device has a _HID (or _CID) returning a valid ACPI/PNP
-        * device ID, it is better to make it look less attractive here, so that
-        * the other device with the same _ADR value (that may not have a valid
-        * device ID) can be matched going forward.  [This means a second spec
-        * violation in a row, so whatever we do here is best effort anyway.]
+        * If the device has a _HID returning a valid ACPI/PNP device ID, it is
+        * better to make it look less attractive here, so that the other device
+        * with the same _ADR value (that may not have a valid device ID) can be
+        * matched going forward.  [This means a second spec violation in a row,
+        * so whatever we do here is best effort anyway.]
         */
-       return sta_present && list_empty(&adev->pnp.ids) ?
+       return sta_present && !adev->pnp.type.platform_id ?
                        FIND_CHILD_MAX_SCORE : FIND_CHILD_MIN_SCORE;
 }