acpi: Defer reserving resources for ACPI devices

The goal of reserving firmware-assigned resources is to ensure that
"wildcard" resource allocation requests will not claim an address
range that is actually in use even if no attached driver is actively
using that range.  However, the current approach can break in some
cases.

In particular, ACPI can enumerate devices behind PCI bridges that
don't show up in a normal PCI scan, but those device_t objects can end
up as direct children of acpi0.  Reserving resources for those devices
directly from acpi0 ends up conflicting with later attempts to reserve
the PCI bridge windows.

As a workaround, defer reserving unclaimed resources until after the
initial probe and attach scan.  Eventually this pass of reserving
unclaimed resources can be moved earlier, but it requires changes to
other drivers in the tree to permit enumerating devices and reserving
firmware-assigned resources in a depth-first traversal before
attaching devices whose drivers request wildcard allocations.

PR:		272507
Reported by:	Justin Tocci <justin@tocci.org>
Reported by:	john@feith.com, many others
Tested by:	Oleg Sidorkin <osidorkin@gmail.com>, dch

(cherry picked from commit f2fcb68074a51a8b399dc80d4c03fbe98a0ab92c)

1 files changed, 17 insertions, 5 deletions

diff --git a/sys/dev/acpica/acpi.c b/sys/dev/acpica/acpi.c
index cd390ada512c..49153eedbcc0 100644
--- a/sys/dev/acpica/acpi.c
+++ b/sys/dev/acpica/acpi.c
@@ -1268,8 +1268,17 @@ acpi_sysres_alloc(device_t dev)
 }
 
 /*
- * Reserve declared resources for devices found during attach once system
- * resources have been allocated.
+ * Reserve declared resources for active devices found during the
+ * namespace scan once the boot-time attach of devices has completed.
+ *
+ * Ideally reserving firmware-assigned resources would work in a
+ * depth-first traversal of the device namespace, but this is
+ * complicated.  In particular, not all resources are enumerated by
+ * ACPI (e.g. PCI bridges and devices enumerate their resources via
+ * other means).  Some systems also enumerate devices via ACPI behind
+ * PCI bridges but without a matching a PCI device_t enumerated via
+ * PCI bus scanning, the device_t's end up as direct children of
+ * acpi0.  Doing this scan late is not ideal, but works for now.
  */
 static void
 acpi_reserve_resources(device_t dev)
@@ -2157,9 +2166,6 @@ acpi_probe_children(device_t bus)
     /* Pre-allocate resources for our rman from any sysresource devices. */
     acpi_sysres_alloc(bus);
 
-    /* Reserve resources already allocated to children. */
-    acpi_reserve_resources(bus);
-
     /* Create any static children by calling device identify methods. */
     ACPI_DEBUG_PRINT((ACPI_DB_OBJECTS, "device identify routines\n"));
     bus_generic_probe(bus);
@@ -2168,6 +2174,12 @@ acpi_probe_children(device_t bus)
     ACPI_DEBUG_PRINT((ACPI_DB_OBJECTS, "acpi bus_generic_attach\n"));
     bus_generic_attach(bus);
 
+    /*
+     * Reserve resources allocated to children but not yet allocated
+     * by a driver.
+     */
+    acpi_reserve_resources(bus);
+
     /* Attach wake sysctls. */
     acpi_wake_sysctl_walk(bus);