Intel xhci: Support EHCI/xHCI port switching.

The Intel Panther Point chipsets contain an EHCI and xHCI host controller
that shares some number of skew-dependent ports.  These ports can be
switched from the EHCI to the xHCI host (and vice versa) by a hardware MUX
that is controlled by registers in the xHCI PCI configuration space.  The
USB 3.0 SuperSpeed terminations on the xHCI ports can be controlled
separately from the USB 2.0 data wires.

This switchover mechanism is there to support users who do a custom
install of certain non-Linux operating systems that don't have official
USB 3.0 support.  By default, the ports are under EHCI, SuperSpeed
terminations are off, and USB 3.0 devices will show up under the EHCI
controller at reduced speeds.  (This was more palatable for the marketing
folks than having completely dead USB 3.0 ports if no xHCI drivers are
available.)  Users should be able to turn on xHCI by default through a
BIOS option, but users are happiest when they don't have to change random
BIOS settings.

This patch introduces a driver method to switchover the ports from EHCI to
xHCI before the EHCI driver finishes PCI enumeration.  We want to switch
the ports over before the USB core has the chance to enumerate devices
under EHCI, or boot from USB mass storage will fail if the boot device
connects under EHCI first, and then gets disconnected when the port
switches over to xHCI.

Add code to the xHCI PCI quirk to switch the ports from EHCI to xHCI.  The
PCI quirks code will run before any other PCI probe function is called, so
this avoids the issue with boot devices.

Another issue is with BIOS behavior during system resume from hibernate.
If the BIOS doesn't support xHCI, it may switch the devices under EHCI to
allow use of the USB keyboard, mice, and mass storage devices.  It's
supposed to remember the value of the port routing registers and switch
them back when the OS attempts to take control of the xHCI host controller,
but we all know not to trust BIOS writers.

Make both the xHCI driver and the EHCI driver attempt to switchover the
ports in their PCI resume functions.  We can't guarantee which PCI device
will be resumed first, so this avoids any race conditions.  Writing a '1'
to an already set port switchover bit or a '0' to a cleared port switchover
bit should have no effect.

The xHCI PCI configuration registers will be documented in the EDS-level
chipset spec, which is not public yet.  I have permission from legal and
the Intel chipset group to release this patch early to allow good Linux
support at product launch.  I've tried to document the registers as much
as possible, so please let me know if anything is unclear.

Signed-off-by: Sarah Sharp <sarah.a.sharp@linux.intel.com>

drivers/usb/host/ehci-pci.c

@@ -348,11 +348,50 @@ static int ehci_pci_suspend(struct usb_hcd *hcd, bool do_wakeup)
 	return rc;
 }
 
+static bool usb_is_intel_switchable_ehci(struct pci_dev *pdev)
+{
+	return pdev->class == PCI_CLASS_SERIAL_USB_EHCI &&
+		pdev->vendor == PCI_VENDOR_ID_INTEL &&
+		pdev->device == 0x1E26;
+}
+
+static void ehci_enable_xhci_companion(void)
+{
+	struct pci_dev		*companion = NULL;
+
+	/* The xHCI and EHCI controllers are not on the same PCI slot */
+	for_each_pci_dev(companion) {
+		if (!usb_is_intel_switchable_xhci(companion))
+			continue;
+		usb_enable_xhci_ports(companion);
+		return;
+	}
+}
+
 static int ehci_pci_resume(struct usb_hcd *hcd, bool hibernated)
 {
 	struct ehci_hcd		*ehci = hcd_to_ehci(hcd);
 	struct pci_dev		*pdev = to_pci_dev(hcd->self.controller);
 
+	/* The BIOS on systems with the Intel Panther Point chipset may or may
+	 * not support xHCI natively.  That means that during system resume, it
+	 * may switch the ports back to EHCI so that users can use their
+	 * keyboard to select a kernel from GRUB after resume from hibernate.
+	 *
+	 * The BIOS is supposed to remember whether the OS had xHCI ports
+	 * enabled before resume, and switch the ports back to xHCI when the
+	 * BIOS/OS semaphore is written, but we all know we can't trust BIOS
+	 * writers.
+	 *
+	 * Unconditionally switch the ports back to xHCI after a system resume.
+	 * We can't tell whether the EHCI or xHCI controller will be resumed
+	 * first, so we have to do the port switchover in both drivers.  Writing
+	 * a '1' to the port switchover registers should have no effect if the
+	 * port was already switched over.
+	 */
+	if (usb_is_intel_switchable_ehci(pdev))
+		ehci_enable_xhci_companion();
+
 	// maybe restore FLADJ
 
 	if (time_before(jiffies, ehci->next_statechange))

drivers/usb/host/pci-quirks.c

@@ -69,6 +69,9 @@
 #define	NB_PIF0_PWRDOWN_0	0x01100012
 #define	NB_PIF0_PWRDOWN_1	0x01100013
 
+#define USB_INTEL_XUSB2PR      0xD0
+#define USB_INTEL_USB3_PSSEN   0xD8
+
 static struct amd_chipset_info {
 	struct pci_dev	*nb_dev;
 	struct pci_dev	*smbus_dev;
@@ -673,6 +676,64 @@ static int handshake(void __iomem *ptr, u32 mask, u32 done,
 	return -ETIMEDOUT;
 }
 
+bool usb_is_intel_switchable_xhci(struct pci_dev *pdev)
+{
+	return pdev->class == PCI_CLASS_SERIAL_USB_XHCI &&
+		pdev->vendor == PCI_VENDOR_ID_INTEL &&
+		pdev->device == PCI_DEVICE_ID_INTEL_PANTHERPOINT_XHCI;
+}
+EXPORT_SYMBOL_GPL(usb_is_intel_switchable_xhci);
+
+/*
+ * Intel's Panther Point chipset has two host controllers (EHCI and xHCI) that
+ * share some number of ports.  These ports can be switched between either
+ * controller.  Not all of the ports under the EHCI host controller may be
+ * switchable.
+ *
+ * The ports should be switched over to xHCI before PCI probes for any device
+ * start.  This avoids active devices under EHCI being disconnected during the
+ * port switchover, which could cause loss of data on USB storage devices, or
+ * failed boot when the root file system is on a USB mass storage device and is
+ * enumerated under EHCI first.
+ *
+ * We write into the xHC's PCI configuration space in some Intel-specific
+ * registers to switch the ports over.  The USB 3.0 terminations and the USB
+ * 2.0 data wires are switched separately.  We want to enable the SuperSpeed
+ * terminations before switching the USB 2.0 wires over, so that USB 3.0
+ * devices connect at SuperSpeed, rather than at USB 2.0 speeds.
+ */
+void usb_enable_xhci_ports(struct pci_dev *xhci_pdev)
+{
+	u32		ports_available;
+
+	ports_available = 0xffffffff;
+	/* Write USB3_PSSEN, the USB 3.0 Port SuperSpeed Enable
+	 * Register, to turn on SuperSpeed terminations for all
+	 * available ports.
+	 */
+	pci_write_config_dword(xhci_pdev, USB_INTEL_USB3_PSSEN,
+			cpu_to_le32(ports_available));
+
+	pci_read_config_dword(xhci_pdev, USB_INTEL_USB3_PSSEN,
+			&ports_available);
+	dev_dbg(&xhci_pdev->dev, "USB 3.0 ports that are now enabled "
+			"under xHCI: 0x%x\n", ports_available);
+
+	ports_available = 0xffffffff;
+	/* Write XUSB2PR, the xHC USB 2.0 Port Routing Register, to
+	 * switch the USB 2.0 power and data lines over to the xHCI
+	 * host.
+	 */
+	pci_write_config_dword(xhci_pdev, USB_INTEL_XUSB2PR,
+			cpu_to_le32(ports_available));
+
+	pci_read_config_dword(xhci_pdev, USB_INTEL_XUSB2PR,
+			&ports_available);
+	dev_dbg(&xhci_pdev->dev, "USB 2.0 ports that are now switched over "
+			"to xHCI: 0x%x\n", ports_available);
+}
+EXPORT_SYMBOL_GPL(usb_enable_xhci_ports);
+
 /**
  * PCI Quirks for xHCI.
  *
@@ -732,6 +793,8 @@ static void __devinit quirk_usb_handoff_xhci(struct pci_dev *pdev)
 	writel(XHCI_LEGACY_DISABLE_SMI,
 			base + ext_cap_offset + XHCI_LEGACY_CONTROL_OFFSET);
 
+	if (usb_is_intel_switchable_xhci(pdev))
+		usb_enable_xhci_ports(pdev);
 hc_init:
 	op_reg_base = base + XHCI_HC_LENGTH(readl(base));
 

drivers/usb/host/pci-quirks.h

@@ -8,6 +8,8 @@ int usb_amd_find_chipset_info(void);
 void usb_amd_dev_put(void);
 void usb_amd_quirk_pll_disable(void);
 void usb_amd_quirk_pll_enable(void);
+bool usb_is_intel_switchable_xhci(struct pci_dev *pdev);
+void usb_enable_xhci_ports(struct pci_dev *xhci_pdev);
 #else
 static inline void usb_amd_quirk_pll_disable(void) {}
 static inline void usb_amd_quirk_pll_enable(void) {}

drivers/usb/host/xhci-pci.c

@@ -242,8 +242,28 @@ static int xhci_pci_suspend(struct usb_hcd *hcd, bool do_wakeup)
 static int xhci_pci_resume(struct usb_hcd *hcd, bool hibernated)
 {
 	struct xhci_hcd		*xhci = hcd_to_xhci(hcd);
+	struct pci_dev		*pdev = to_pci_dev(hcd->self.controller);
 	int			retval = 0;
 
+	/* The BIOS on systems with the Intel Panther Point chipset may or may
+	 * not support xHCI natively.  That means that during system resume, it
+	 * may switch the ports back to EHCI so that users can use their
+	 * keyboard to select a kernel from GRUB after resume from hibernate.
+	 *
+	 * The BIOS is supposed to remember whether the OS had xHCI ports
+	 * enabled before resume, and switch the ports back to xHCI when the
+	 * BIOS/OS semaphore is written, but we all know we can't trust BIOS
+	 * writers.
+	 *
+	 * Unconditionally switch the ports back to xHCI after a system resume.
+	 * We can't tell whether the EHCI or xHCI controller will be resumed
+	 * first, so we have to do the port switchover in both drivers.  Writing
+	 * a '1' to the port switchover registers should have no effect if the
+	 * port was already switched over.
+	 */
+	if (usb_is_intel_switchable_xhci(pdev))
+		usb_enable_xhci_ports(pdev);
+
 	retval = xhci_resume(xhci, hibernated);
 	return retval;
 }