Merge tag 'driver-core-5.11-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/driver-core

Pull driver core updates from Greg KH:
 "Here is the big driver core updates for 5.11-rc1

  This time there was a lot of different work happening here for some
  reason:

   - redo of the fwnode link logic, speeding it up greatly

   - auxiliary bus added (this was a tag that will be pulled in from
     other trees/maintainers this merge window as well, as driver
     subsystems started to rely on it)

   - platform driver core cleanups on the way to fixing some long-time
     api updates in future releases

   - minor fixes and tweaks.

  All have been in linux-next with no (finally) reported issues. Testing
  there did helped in shaking issues out a lot :)"

* tag 'driver-core-5.11-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/driver-core: (39 commits)
  driver core: platform: don't oops in platform_shutdown() on unbound devices
  ACPI: Use fwnode_init() to set up fwnode
  misc: pvpanic: Replace OF headers by mod_devicetable.h
  misc: pvpanic: Combine ACPI and platform drivers
  usb: host: sl811: Switch to use platform_get_mem_or_io()
  vfio: platform: Switch to use platform_get_mem_or_io()
  driver core: platform: Introduce platform_get_mem_or_io()
  dyndbg: fix use before null check
  soc: fix comment for freeing soc_dev_attr
  driver core: platform: use bus_type functions
  driver core: platform: change logic implementing platform_driver_probe
  driver core: platform: reorder functions
  driver core: make driver_probe_device() static
  driver core: Fix a couple of typos
  driver core: Reorder devices on successful probe
  driver core: Delete pointless parameter in fwnode_operations.add_links
  driver core: Refactor fw_devlink feature
  efi: Update implementation of add_links() to create fwnode links
  of: property: Update implementation of add_links() to create fwnode links
  driver core: Use device's fwnode to check if it is waiting for suppliers
  ...

1 files changed, 9 insertions, 11 deletions

diff --git a/drivers/usb/host/sl811-hcd.c b/drivers/usb/host/sl811-hcd.c
index adaf4063690a..115ced0d93e1 100644
--- a/drivers/usb/host/sl811-hcd.c
+++ b/drivers/usb/host/sl811-hcd.c
@@ -1614,12 +1614,18 @@ sl811h_probe(struct platform_device *dev)
 	void __iomem		*addr_reg;
 	void __iomem		*data_reg;
 	int			retval;
-	u8			tmp, ioaddr = 0;
+	u8			tmp, ioaddr;
 	unsigned long		irqflags;
 
 	if (usb_disabled())
 		return -ENODEV;
 
+	/* the chip may be wired for either kind of addressing */
+	addr = platform_get_mem_or_io(dev, 0);
+	data = platform_get_mem_or_io(dev, 1);
+	if (!addr || !data || resource_type(addr) != resource_type(data))
+		return -ENODEV;
+
 	/* basic sanity checks first.  board-specific init logic should
 	 * have initialized these three resources and probably board
 	 * specific platform_data.  we don't probe for IRQs, and do only
@@ -1632,16 +1638,8 @@ sl811h_probe(struct platform_device *dev)
 	irq = ires->start;
 	irqflags = ires->flags & IRQF_TRIGGER_MASK;
 
-	/* the chip may be wired for either kind of addressing */
-	addr = platform_get_resource(dev, IORESOURCE_MEM, 0);
-	data = platform_get_resource(dev, IORESOURCE_MEM, 1);
-	retval = -EBUSY;
-	if (!addr || !data) {
-		addr = platform_get_resource(dev, IORESOURCE_IO, 0);
-		data = platform_get_resource(dev, IORESOURCE_IO, 1);
-		if (!addr || !data)
-			return -ENODEV;
-		ioaddr = 1;
+	ioaddr = resource_type(addr) == IORESOURCE_IO;
+	if (ioaddr) {
 		/*
 		 * NOTE: 64-bit resource->start is getting truncated
 		 * to avoid compiler warning, assuming that ->start